Você está na página 1de 1931


Seventh Edition


MACP, MACR Professor of Medicine
Professor of Medicine and Epidemiology and Public Health Chairman, Department of Medicine III
Head, Division of Rheumatology and Clinical Immunology Division of Rheumatology
Vice Chair, Department of Medicine Medical University of Vienna
University of Maryland School of Medicine Chairman, Second Department of Medicine
Director, Gerontology Research Education and Clinical Center Center for Rheumatic Diseases
Veterans Affairs Maryland Health Care System Hietzing Hospital
Baltimore, Maryland Vienna, Austria


Myles J. McDonough Chair in Rheumatology John R. and Eileen K. Riedman Professor of Medicine
Chief, Division of Rheumatology Harvard Medical School
Department of Medicine R. Bruce and Joan M. Mickey Distinguished Chair in Rheumatology
University of Massachusetts Memorial Health Care Division of Rheumatology, Immunology, and Allergy
Professor of Medicine Brigham and Women’s Hospital
University of Massachusetts Medical School Boston, Massachusetts
Worcester, Massachusetts


Cedars-Sinai Chair in Rheumatology
Professor of Musculoskeletal Health Director, Division of Rheumatology
Nuffield Department off Orthopaedics, Rheumatology and Musculoskeletal Sciences Professor of Medicine
Oxford University Cedars-Sinai Medical Center
United Kingdom Distinguished Professor of Medicine
David Geffen School of Medicine at University of California, Los Angeles
Los Angeles, California
1600 John F. Kennedy Blvd.
Ste 1800
Philadelphia, PA 19103-2899

RHEUMATOLOGY, 7th edition ISBN: 978-0-7020-6865-2

Copyright © 2019 by Elsevier, Inc. All rights reserved.

Previous editions copyrighted 2015, 2011, 2008, 2004, 2001, 1996 by Elsevier.

No part of this publication may be reproduced or transmitted in any form or by any means, electronic or
mechanical, including photocopying, recording, or any information storage and retrieval system, without
permission in writing from the publisher. Details on how to seek permission, further information about the
Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance
Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions.

This book and the individual contributions contained in it are protected under copyright by the Publisher
(other than as may be noted herein).


Knowledge and best practice in this field are constantly changing. As new research and experience
broaden our understanding, changes in research methods, professional practices, or medical treatment
may become necessary.
Practitioners and researchers must always rely on their own experience and knowledge in evaluating
and using any information, methods, compounds, or experiments described herein. In using such
information or methods they should be mindful of their own safety and the safety of others, including
parties for whom they have a professional responsibility.
With respect to any drug or pharmaceutical products identified, readers are advised to check the
most current information provided (i) on procedures featured or (ii) by the manufacturer of each
product to be administered, to verify the recommended dose or formula, the method and duration of
administration, and contraindications. It is the responsibility of practitioners, relying on their own
experience and knowledge of their patients, to make diagnoses, to determine dosages and the best
treatment for each individual patient, and to take all appropriate safety precautions.
To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors,
assume any liability for any injury and/or damage to persons or property as a matter of products
liability, negligence or otherwise, or from any use or operation of any methods, products, instructions,
or ideas contained in the material herein.

Library of Congress Cataloging-in-Publication Data

Names: Hochberg, Marc C., editor.

Title: Rheumatology / [edited by] Marc C. Hochberg, Ellen M. Gravallese, Alan J. Silman,
  Josef S. Smolen, Michael E. Weinblatt, Michael H. Weisman.
Other titles: Rheumatology (Hochberg)
Description: Seventh edition. | Philadelphia, PA : Elsevier, Inc., [2019] |
  Includes bibliographical references and index.
Identifiers: LCCN 2017007320 | ISBN 9780702068652 (hardcover : alk. paper)
Subjects: | MESH: Rheumatic Diseases | Antirheumatic Agents–therapeutic use
Classification: LCC RC927 | NLM WE 544 | DDC 616.7/23–dc23 LC record available at https://lccn.loc.

Content Strategist: Nancy Anastasi Duffy

Content Development Specialist: Jennifer Shreiner
Publishing Services Manager: Patricia Tannian
Project Manager: Ted Rodgers
Design Direction: Bridget Hoette

Printed in China

Last digit is the print number: 9 8 7 6 5 4 3 2 1

To our parents (living or of blessed memory) and our spouses, children, and grandchildren

Susan Hochberg, Francine, Jeffrey, and Eleanor (Nora) Zoe Giuffrida, and Jennifer Hochberg

M. Timothy Hresko, Andrew (Drew) and Gregory Hresko

Ruth Silman, Joanna, Timothy, and Daniel Silman

Alice Smolen, Eva, Daniel, Sami, and Ilona Hruschka, Nina Smolen-Wilson, Etienne, Anna, and James Jandi Wilson,
Daniel Smolen and Catherine McKenzie, and Alexander Smolen and Meeri Parikka

Barbara Weinblatt, Hillary, Jason, Sidney Allen and Annika Gray Chapman, and Courtney and Michael Fasciano

Betsy Weisman, Greg, Nicole, Mia, and Joseph Weisman, Lisa, Andrew, David, and Thomas Cope, and Annie, Bill, Caro-
line, and Dorothy Macomber

Abby G. Abelson, MD Daniel A. Albert, MD, FACR Mary-Carmen Amigo, MD, FACP
Clinical Assistant Professor of Medicine Professor of Medicine and Pediatrics Professor and Head
Chair, Department of Rheumatic and Immunologic Divisions of Rheumatology Rheumatology Service
Diseases Dartmouth Hitchcock Medical Center American British Cowdray Medical Center
Cleveland Clinic Lerner College of Medicine Lebanon, New Hampshire Mexico City, Mexico
Case Western Reserve University Chapter 120, Inflammatory back pain Chapter 148, Antiphospholipid syndrome: pathogenesis,
Cleveland, Ohio diagnosis, and management
Chapter 199, Management of osteoporosis Daniel Aletaha, MD, MSc
Associate Professor of Internal Medicine Nigel K. Arden, MBBS, FRCP, MSc, MD
Steven B. Abramson, MD Division of Rheumatology Professor
Frederick H. King Professor of Internal Medicine Medical University of Vienna Nuffield Department of Orthopaedics,
Chair, Department of Medicine Vienna, Austria Rheumatology, and Musculoskeletal Sciences
Professor of Medicine and Pathology Chapter 101, Assessment of the patient with University of Oxford
New York University Langone Medical Center rheumatoid arthritis and the measurement of Oxford, United Kingdom;
New York, New York outcomes Professor in Rheumatic Diseases and Consultant
Chapter 24, The microbiome in rheumatic diseases Rheumatologist
Ilias Alevizos, DMD, MMSc, MPH, PhD Medical Research Council
Jonathan D. Adachi, BSc, MD, FRCPC Investigator, Sjögren’s Syndrome Clinic University of Southampton
Alliance for Better Bone Health Chair in Molecular Physiology and Therapeutics Branch Southampton, United Kingdom
Rheumatology National Institute of Dental and Craniofacial Chapter 188, Management of osteoarthritis
Professor, Department of Medicine Research
McMaster University Michael G. DeGroote School National Institutes of Health Martin Aringer, MD
of Medicine Bethesda, Maryland Professor of Medicine (Rheumatology)
St Joseph’s Healthcare Chapter 147, Sjögren syndrome Department of Medicine III, Division of
Hamilton, Ontario, Canada Rheumatology
Chapter 200, Glucocorticoid-induced osteoporosis Kavitta B. Allem, MD University Clinical Center Carl Gustav Carus at the
Physician TU Dresden
Michael A. Adams, PhD Division of Rheumatology Dresden, Germany
Professor of Biomechanics Scripps Clinic Chapter 11, Signal transduction in immune cells
Centre for Applied Anatomy La Jolla, California
University of Bristol Chapter 67, Immunosuppressive agents: cyclosporine, Elizabeth V. Arkema, ScM, ScD
Bristol, United Kingdom; cyclophosphamide, azathioprine, mycophenolate Assistant Professor of Clinical Epidemiology
Visiting Professor mofetil, and tacrolimus Department of Medicine
Sir Run Run Shaw Hospital Karolinska Institutet
Zheijang University Mohammed Almehthel, MD, ABIM, FRCPC Stockholm, Sweden
Hangzhou, China Clinical Assistant Professor of Medicine Chapter 25, Principles of epidemiology
Chapter 7, Biomechanics of spinal degeneration Division of Endocrinology
University of British Columbia Dana P. Ascherman, MD
Rohit Aggarwal, MD, MS Vancouver, Canada Associate Professor of Medicine
Associate Professor of Medicine Chapter 51, Dual x-ray absorptiometry and Division of Rheumatology
Division of Rheumatology and Clinical Immunology measurement of bone University of Miami Miller School of Medicine
University of Pittsburgh School of Medicine Miami, Florida
Pittsburgh, Pennsylvania Mohamed Almohaya, MD Chapter 156, Classification, epidemiology, and clinical
Chapter 156, Classification, epidemiology, and clinical Endocrinologist features of inflammatory muscle disease
features of inflammatory muscle disease Obesity, Endocrine and Metabolism Center
King Fahad Medical City Sergei P. Atamas, MD, PhD
Prof. Dr. Thomas Aigner Riyadh, Saudi Arabia Professor of Medicine and Microbiology &
Institut für Pathologie Clinical Fellow, Metabolic Bone Disorders Immunology
Klinikum Coburg Endocrinology Department University of Maryland School of Medicine;
Coburg, Germany University of British Columbia Research Health Scientist
Chapter 183, Pathogenesis and pathology of Vancouver, Canada Baltimore VA Medical Center
osteoarthritis Chapter 51, Dual x-ray absorptiometry and Baltimore, Maryland
measurement of bone Chapter 10, Principles of adaptive immunity
Shizuo Akira, MD, PhD
Professor Elena Riera Alonso, MD Timothy J. Atkinson, PharmD
Laboratory of Host Defense Division of Rheumatology Clinical Pharmacy Specialist, Pain Management
WPI Immunology Frontier Research Center Hospital Mútua Terrassa Pharmacy Service
Osaka University Barcelona, Spain VA Tennessee Valley Healthcare System
Osaka, Japan Chapter 173, Adult-onset Still disease Murfreesboro, Tennessee
Chapter 9, Principles of innate immunity Chapter 62, Principles of pharmacologic pain

Maha A. Azeez, MB BCh BAO, BSc, MD, MRCPI Les Barnsley, B.Med(Hons), Grad.Dip.Epi., PhD, Kim L. Bennell, BAppSci(physio), PhD
Nuffield Institute of Orthopedics, Rheumatology, FRACP, FAFRM(RACP) Professor of Physiotherapy
and Musculoskeletal Sciences Associate Professor of Medicine Director, Centre for Health, Exercise, and Sports
University of Oxford Sydney University Medicine
Botnar Research Centre Sydney, Australia; University of Melbourne
Headington, United Kingdom Head, Department of Rheumatology Melbourne, Australia
Chapter 72, Tumor necrosis factor inhibitors Concord Hospital Chapter 55, Principles of rehabilitation: physical and
Concord, Australia occupational therapy
Pedro M. Azevedo, MD, PhD Chapter 80, Neck pain
Assistant Professor of Rheumatology Michael W. Beresford, MBChB, PhD
Evangelical University Hospital of Curitiba Andrew J. Barr, MBBS, PhD Brough Chair, Professor of Child Health,
Curitiba, Brazil Clinical Research Fellow Institute of Translational Medicine
Chapter 117, Acute rheumatic fever Leeds Institute of Rheumatic and Musculoskeletal University of Liverpool
Medicine Honorary Consultant Paediatric Rheumatologist
Alan N. Baer, MD, FACP University of Leeds Alder Hey Children’s NHS Foundation Trust
Associate Professor of Medicine Chapel Allerton Hospital Liverpool, United Kingdom
Division of Rheumatology Leeds, United Kingdom Chapter 111, Connective tissue diseases in children
Director, Jerome L. Greene Sjögren’s Syndrome Center Chapter 185, Imaging of osteoarthritis
Johns Hopkins University; Brian Berman, MD
Clinical Director Joan M. Bathon, MD Professor of Family and Community Medicine
Johns Hopkins Rheumatology Professor of Medicine Director, Center for Integrative Medicine
Good Samaritan Hospital Division of Rheumatology University of Maryland School of Medicine
Baltimore, Maryland Columbia University College of Physicians and Baltimore, Maryland
Chapter 147, Sjögren syndrome Surgeons Chapter 58, Complementary and alternative medicine
New York, New York
Dominique Baeten, MD, PhD Chapter 103, Management of established rheumatoid Bonnie L. Bermas, MD
Professor of Rheumatology arthritis Assistant Professor of Medicine
Department of Clinical Immunology and Division of Rheumatology
Rheumatology Angela Bauch, PhD Harvard Medical School
University of Amsterdam CeMM Research Center for Molecular Medicine of Director, Lupus Center
Academic Medical Center the Austrian Academy of Sciences Brigham and Women’s Hospital
Amsterdam, The Netherlands Vienna, Austria Boston, Massachusetts
Chapter 122, Pathogenesis and pathophysiology of axial Chapter 22, Proteomics Chapter 59, Drugs and pregnancy
Anne-Christine Bay-Jensen, mMBS, MSc, PhD George Bertsias, MD, PhD
Colin P. Baines, MBChB, MRCP(UK) Head, Department of Rheumatology Assistant Professor in Rheumatology-Clinical
Specialty Doctor Division of Biomarkers and Research Immunology
Departments of General and Vascular Medicine Nordic Bioscience University of Crete;
Ninewells Hospital and Medical School Herlev, Denmark Rheumatology, Clinical Immunology and Allergy
Dundee, United Kingdom Chapter 198, Biochemical markers in bone disease University Hospital of Iraklio
Chapter 154, Raynaud phenomenon Iraklio, Greece
Jill J.F. Belch, MBChB, MD (Hons), FRCP, FRS Chapter 145, Management of renal lupus
Alejandro Balsa, MD, PhD Professor of Vascular Medicine
Head, Department of Rheumatology Division of Clinical and Molecular Medicine Sonal Bhalla, MD
Hospital Universitario La Paz Ninewells Hospital and Medical School Fellow, Department of Rheumatology
Associate Professor of Rheumatology Dundee, United Kingdom Dartmouth Hitchcock Medical Center
Department of Medicine Chapter 154, Raynaud phenomenon Lebanon, New Hampshire
Universidad Autonoma Chapter 120, Inflammatory back pain
Madrid, Spain Nicholas Bellamy, MD, MSc, MBA, DSc, FRCP(C),
Chapter 113, Bacterial arthritis FRCP (Glasg, Edin), FACP, FRACP Yelda Bilginer, MD
Emeritus Professor Associate Professor of Pediatric Nephrology and
Xenofon Baraliakos, MD
Centre for Chronic Disease Rheumatology
Rheumazentrum Ruhrgebiet
The University of Queensland School of Medicine Hacettepe University Faculty of Medicine
Ruhr-University Bochum
Herston, Australia Ankara, Turkey
Herne, Germany
Chapter 26, Principles of clinical outcome assessment Chapter 169, IgA vasculitis (Henoch-Schoenlein
Chapter 125, Imaging of spondyloarthritis
Thomas Bardin, MD Teresita Bellido, PhD
Professor of Rheumatology Professor Julius Birnbaum, MD
Université Paris 7; Departments of Anatomy and Cell Biology and Assistant Professor of Medicine
Head, Department of Rheumatology Internal Medicine (Endocrinology) Assistant Professor of Neurology
Hôpital Lariboisière Indiana University School of Medicine Associate Director, Jerome L. Greene Sjögren’s
Paris, France Roudebush VA Medical Center Syndrome Center
Chapter 214, Miscellaneous arthropathies
Indianapolis, Indiana Johns Hopkins University School of Medicine
Chapter 3, Bone structure and function Baltimore, Maryland
Jammie K. Barnes, MD Chapter 41, The nervous system in rheumatic disease
Assistant Professor of Medicine Michael Benjamin, PhD†
Division of Rheumatology Professor Emeritus Jane F. Bleasel, MB BS, PhD, MHPE, FRACP
University of Texas Health Science Center Houston School of Biosciences Professor of Rheumatology
McGovern School of Medicine Cardiff University Royal Prince Alfred Hospital
Houston, Texas Cardiff, United Kingdom Sydney Medical School
Chapter 149, Classification and epidemiology of Chapter 132, Enthesopathies University of Sydney
systemic sclerosis (scleroderma) Sydney, Australia
Chapter 208, Hemophilia and von Willebrand disease


Joel A. Block, MD Richard D. Brasington, Jr., MD Gerd-Rüdiger Burmester, MD

The Willard L. Wood MD Professor and Director Professor of Medicine Professor of Medicine
Division of Rheumatology Division of Rheumatology Division of Rheumatology and Clinical Immunology
Rush University Medical Center Washington University School of Medicine in St. Charité-Universitätsmedizin Berlin
Chicago, Illinois Louis Berlin, Germany
Chapter 181, Clinical features of osteoarthritis St. Louis, Missouri Chapter 33, Laboratory tests in rheumatic disorders
Chapter 94, Clinical features of rheumatoid arthritis
Ashley W. Blom, MBChB, MD, PhD, FRCS Jane C. Burns, MD
Professor of Orthopaedic Surgery Prof. Dr. Jürgen Braun Professor of Pediatrics
Musculoskeletal Research Unit Rheumazentrum Ruhrgebiet Director, Kawasaki Disease Research Center
University of Bristol Ruhr-University Bochum University of California San Diego School of
Bristol, United Kingdom Herne, Germany Medicine
Chapter 61, Long-term complications of total hip and Chapter 125, Imaging of spondyloarthritis La Jolla, California;
knee replacement Director, Kawasaki Disease Clinic
Elisabeth Brouwer, MD, PhD Rady Children’s Hospital-San Diego
Markus Böhm, MD Department of Rheumatology and Clinical San Diego, California
Professor of Dermatology Immunology Chapter 168, Kawasaki disease
University of Münster University of Groningen, University Medical Center
Münster, Germany Groningen David B. Burr, PhD
Chapter 35, The skin in rheumatic disease Groningen, The Netherlands Distinguished Professor
Chapter 162, Biology and immunopathogenesis of Associate Vice Chancellor for Research (IUPUI)
Michael Bonelli, PhD vasculitis Department of Anatomy and Cell Biology, Indiana
Fellow University School of Medicine
Molecular Immunology and Inflammation Branch Jacques P. Brown, MD, FRCPC Department of Biomedical Engineering, IUPUI
National Institute of Arthritis and Musculoskeletal Clinical Professor of Medicine Indianapolis, Indiana
and Skin Diseases Université Laval; Chapter 3, Bone structure and function
National Institutes of Health Division of Infectious Disease and Immunity
Bethesda, Maryland CHU de Québec Research Centre Frank Buttgereit, MD
Chapter 11, Signal transduction in immune cells Québec City, Canada Professor of Rheumatology and Clinical
Chapter 203, Paget disease of bone Immunology
Grace Borchert Charité University Medicine Berlin
Research Assistant Matthew A. Brown, MBBS, MD, FRACP, FAHMS, Berlin, Germany
Translational Research Institute FAA Chapter 64, Systemic glucocorticoids in rheumatology
The University of Queensland Diamantina Institute Professor of Immunogenetics
Brisbane, Australia Director of Genomics Vivian P. Bykerk, MD
Chapter 124, Animal models of spondyloarthritis Queensland University of Technology Associate Professor of Medicine
Institute of Health and Biomedical Innovation Hospital for Special Surgery
Christina A. Boros, MBBS, PhD, FRACP Translational Research Institute New York, New York
Honorary Research Fellow Princess Alexandra Hospital Chapter 65, Synthetic disease-modifying antirheumatic
Division of Infection, Inflammation and Brisbane, Australia drugs and leflunomide
Rheumatology Chapter 123, Genetics of axial spondyloarthritis
UCL Great Ormond Street Institute of Child Health Leonard H. Calabrese, DO
London, United Kingdom; Ian N. Bruce, MD, FRCP Professor of Medicine
Senior Lecturer Professor of Rheumatology Cleveland Clinic Lerner College of Medicine
Discipline of Paediatrics and Reproductive Health Arthritis Research UK Epidemiology Unit Case Western University
University of Adelaide The University of Manchester Vice Chairman, Department of Rheumatic and
Head, Rheumatology Unit Manchester, United Kingdom Immunological Diseases
Women’s and Children’s Hospital Chapter 128, Clinical features of psoriatic arthritis R.J. Fasenmyer Chair of Clinical Immunology
Adelaide, Australia Theodore F. Classen DO Chair of Osteopathic
Chapter 111, Connective tissue diseases in children Maya H. Buch, MBChB, FRCP, PhD Research and Education
Associate Professor and Honorary Consultant Cleveland Clinic Foundation
Dimitrios T. Boumpas, MD, FACP Rheumatologist Cleveland, Ohio
Professor of Internal Medicine Section Head, Clinical Musculoskeletal Medicine Chapter 115, Viral infections
National and Kapodistrian University of Athens Deputy Director, Leeds Institute of Rheumatic and
Medical School Musculoskeletal Medicine Jeffrey P. Callen, MD
Sotiria General Hospital for Pulmonary and Heart University of Leeds Professor of Medicine (Dermatology)
Diseases; Leeds, United Kingdom Chief, Division of Dermatology
Rheumatology and Clinical Immunology Chapter 102, Management of early rheumatoid arthritis University of Louisville School of Medicine
4th Department of Internal Medicine Louisville, Kentucky
Attikon University Hospital; Christopher D. Buckley, MBBS, DPhil Chapter 170, Cutaneous vasculitis and panniculitis
Affiliated Investigator Professor of Medicine
Immunobiology Rheumatology Research Group Andrew J. Carr, DSc, MA, ChM, FRCS
Biomedical Research Foundation of the Academy of Institute of Inflammation and Ageing Nuffield Professor of Orthopaedic Surgery
Athens University of Birmingham Head, Nuffield Department of Orthopaedics,
Athens, Greece Birmingham, United Kingdom Rheumatology and Musculoskeletal Sciences
Chapter 145, Management of renal lupus Chapter 1, The synovium Director, Botnar Research Centre
University of Oxford
Aline Bozec, PhD William D. Bugbee, MD Oxford, United Kingdom
Professor of Rheumatology and Immunology Attending Physician Chapter 4, Tendons and ligaments
Department of Internal Medicine 3 Division of Orthopaedic Surgery
University of Erlangen-Nuremberg Faculty of Scripps Clinic
Medicine La Jolla, California;
Erlangen, Germany Professor of Orthopaedic Surgery
Chapter 16, Osteoimmunology University of California, San Diego
San Diego, California
Chapter 85, The hip
Graeme J. Carroll, MBBS, MD, FRACP Lorinda Chung, MD, MS Robert G. Cooper, MD, FRCP
Assistant Professor of Medicine Associate Professor of Medicine and Dermatology Professor of Medicine (Muscle and Rheumatology)
University of Notre Dame School of Medicine Division of Immunology and Rheumatology Institute of Ageing and Chronic Disease
Fremantle, Australia; Stanford University School of Medicine University of Liverpool
Consultant Physician Stanford, California Liverpool, United Kingdom
Division of Rheumatology Chapter 158, Clinical significance of autoantibodies in Chapter 42, The muscles in rheumatic disease
Fiona Stanley Hospital inflammatory muscle disease
Perth, Australia Karen H. Costenbader, MD, MPH
Chapter 210, Hemochromatosis Francesco Ciccia, MD, PhD Co-Director, Lupus Center
Dipartimento Biomedico di Medicina Interna e Division of Rheumatology, Immunology and Allergy
Sabrina Cavallo, BSc (OT), MSc Specialistica Brigham and Women’s Hospital;
Ecole de Santé Publique Università degli Studi di Palermo Associate Professor of Medicine
Université de Montréal; Palermo, Italy Harvard Medical School
Occupational Therapist Chapter 166, Polymyalgia rheumatica and giant cell Boston, Massachusetts
Department of Occupational Therapy arteritis Chapter 133, Epidemiology and classification of
Montreal Children’s Hospital of the McGill systemic lupus erythematosus
University Health Centre Daniel J. Clauw, MD
Montréal, Canada Professor of Anesthesiology, Medicine George L.D. Cox, BMedSci, BMBS, MD,
Chapter 112, Rehabilitation and psychosocial issues in (Rheumatology) and Psychiatry FRCS(T&O)
juvenile idiopathic arthritis Director, Chronic Pain and Fatigue Research Center Consultant
University of Michigan Medical School Departments of Trauma and Orthopaedics
Ricard Cervera, MD, PhD, FRCP Ann Arbor, Michigan University Hospital Southampton
Head, Department of Autoimmune Diseases Chapter 91, Fibromyalgia and related syndromes Southampton, United Kingdom
Hospital Clínic Chapter 82, The shoulder
University of Barcelona Jacqui Clinch, MBBS, MRCP, FRCPCH
Barcelona, Catalonia, Spain Consultant in Paediatric Rheumatology Paul Creamer, MD, FRCP
Chapter 143, Management of nonrenal and non– Bristol Royal Hospital for Children Consultant and Senior Clinical Lecturer
central nervous system lupus Bristol, United Kingdom; Department of Rheumatology
Consultant in Paediatric Pain North Bristol Trust
Prateek Chaudhary, DO, MPH Bath Centre for Pain Services Bristol, United Kingdom
Fellow Royal National Hospital for Rheumatic Diseases Chapter 205, Neuropathic arthropathy
Division of Rheumatology and Immunology Bath, United Kingdom
Duke University Medical Center Chapter 216, Hypermobility syndrome Bruce N. Cronstein, MD
Durham, North Carolina Paul R. Esserman Professor of Medicine
Chapter 146, Systemic lupus erythematosus in pregnant Megan E.B. Clowse, MD, MPH Division of Rheumatology
patients and neonatal lupus Assistant Professor of Medicine New York University School of Medicine
Division of Rheumatology and Immunology New York, New York
Hyon K. Choi, MD, DrPH Duke University Medical Center Chapter 79, Precision medicine and pharmacogenomics
Professor of Medicine Durham, North Carolina in rheumatology
Division of Rheumatology, Allergy, and Chapter 146, Systemic lupus erythematosus in pregnant
Immunology patients and neonatal lupus Raymond Cross, MD, MS
Harvard Medical School; Associate Professor of Medicine
Director, Gout and Crystal Arthropathy Center J. Gerard Coghlan, MD, FRCP University of Maryland School of Medicine
Director, Clinical Epidemiology and Health Consultant Cardiologist Baltimore, Maryland
Outcomes Royal Free Hospital Chapter 39, The gastrointestinal tract and rheumatic
Massachusetts General Hospital London, United Kingdom disease
Boston, Massachusetts Chapter 37, The heart in rheumatic disease
Chapter 189, Epidemiology and classification of gout
Maurizio Cutolo, MD
Robert A. Colbert, MD, PhD Professor of Rheumatology and Internal Medicine
Ernest H. Choy, MD, FRCP Senior Investigator Director, Research Laboratories and Academic Unit
Professor of Medicine National Institute of Arthritis, Musculoskeletal and of Clinical Rheumatology
Cardiff University School of Medicine Skin Diseases Director, Postgraduate Academic School of
Section of Rheumatology National Institutes of Health Rheumatology
Institute of Infection and Immunity Bethesda, Maryland University of Genova
Cardiff, United Kingdom Chapter 108, Etiology and pathogenesis of juvenile Genova, Italy
Chapter 74, Inhibitors of T-cell co-stimulation idiopathic arthritis Chapter 15, Neuroendocrine immune crosstalk in
chronic inflammatory systemic diseases
Lisa Christopher-Stine, MD, MPH Philip G. Conaghan, MBBS, PhD, FRACP, FRCP
Associate Professor of Medicine (Rheumatology) Professor of Musculoskeletal Medicine Chris D’Adamo, PhD
and Neurology Leeds Institute of Rheumatic and Musculoskeletal Assistant Professor
Director, Johns Hopkins Myositis Center Medicine Department of Family and Community Medicine
Johns Hopkins University School of Medicine University of Leeds; Assistant Professor
Baltimore, Maryland Deputy Director Department of Epidemiology and Public Health
Chapter 160, Metabolic, drug-induced, and other National Institute for Health Research Director of Research
noninflammatory myopathies Leeds Musculoskeletal Biomedical Research Unit Center for Integrative Medicine
Leeds, United Kingdom University of Maryland School of Medicine
Alvina D. Chu, MD Chapter 185, Imaging of osteoarthritis Baltimore, Maryland
Adjunct Clinical Instructor Chapter 58, Complementary and alternative medicine
Division of Immunology and Rheumatology Cyrus Cooper, FMedSci
Stanford University School of Medicine Professor and Director Vivette D. D’Agati, MD
Palo Alto, California MRC Lifecourse Epidemiology Unit Professor of Pathology
Pharmacyclics LLC, an AbbVie Company University of Southampton Columbia University College of Physicians and
Sunnyvale, California Southampton, United Kingdom Surgeons
Chapter 142, Assessing disease activity and outcomes Chapter 195, Epidemiology and classification of Director, Renal Pathology Laboratory
in systemic lupus erythematosus osteoporosis Columbia University Medical Center
New York, New York
Chapter 137, Immunopathology of systemic lupus

David I. Daikh, MD, PhD Chad L. Deal, MD Paul Emery, MD, MA, FMedSci, FRCP
Professor of Medicine Associate Professor of Medicine Arthritis Research UK Professor of Rheumatology
University of California, San Francisco; Cleveland Clinic Lerner College of Medicine Director, Leeds Institute of Rheumatic and
Chief, Arthritis Section Case Western Reserve University Musculoskeletal Medicine
San Francisco VA Medical Center Cleveland, Ohio University of Leeds;
San Francisco, California Chapter 199, Management of osteoporosis Director, NIHR Leeds Musculoskeletal Biomedical
Chapter 138, Animal models of systemic lupus Research Centre
erythematosus Kevin D. Deane, MD, PhD Leeds Teaching Hospital NHS Trust
Associate Professor of Medicine Leeds, United Kingdom
Stephanie G. Dakin, PhD, BVetMed, MRCVS Division of Rheumatology Chapter 75, Inhibitors of B cells
Oxford UCB Prize Fellow University of Colorado School of Medicine
Nuffield Department of Orthopaedics, Aurora, Colorado Gerard Espinosa, MD, PhD
Rheumatology and Musculoskeletal Sciences Chapter 93, Preclinical rheumatoid arthritis Department of Autoimmune Diseases
University of Oxford Hospital Clínic
Oxford, United Kingdom Paul F. Dellaripa, MD University of Barcelona
Chapter 4, Tendons and ligaments Associate Professor of Medicine Barcelona, Catalonia, Spain
Harvard Medical School; Chapter 143, Management of nonrenal and non–
Nicola Dalbeth, MBChB, MD, FRACP Division of Rheumatology central nervous system lupus
Professor and Rheumatologist Brigham and Women’s Hospital
Department of Medicine Boston, Massachusetts Luis R. Espinoza, MD
University of Auckland Chapter 38, The lung in rheumatic disease Professor and Chief
Auckland, New Zealand Department of Internal Medicine
Chapter 190, Etiology and pathogenesis of gout Elaine Dennison, MB, BChir, MSc, FRCP, PhD LSU Health Science Center
Professor of Musculoskeletal Epidemiology Louisiana State University
Aileen M. Davis, BScPT, MSc, PhD MRC Lifecourse Epidemiology Unit New Orleans, Louisiana
Senior Scientist, Health Care, and Outcomes Southampton University Chapter 114, Mycobacterial, brucellar, fungal, and
Research Southampton, United Kingdom; parasitic arthritis
Toronto Western Research Institute Professor of Clinical Research
University Health Network; School of Biological Sciences Stephen Eyre, PhD
Professor Victoria University Research Associate
Institute of Health Policy, Management, and Wellington, New Zealand Arthritis Research UK Epidemiology Unit
Evaluation Chapter 195, Epidemiology and classification of The University of Manchester
Department of Physical Therapy and Graduate osteoporosis Manchester, United Kingdom
Department of Rehabilitation Science Chapter 28, Principles of genetic epidemiology
University of Toronto Christopher P. Denton, PhD
Toronto, Ontario, Canada Professor of Experimental Rheumatology Antonis C. Fanouriakis, MD
Chapter 186, Assessment of the patient with Consultant Rheumatologist Research Associate
osteoarthritis and measurement of outcomes Centre for Rheumatology Rheumatology and Clinical Immunology
University College London 4th Dept of Internal Medicine
David P. D’Cruz, MD, FRCP Royal Free Hospital Attikon University Hospital
Consultant Rheumatologist London, United Kingdom Athens, Greece
The Louise Coote Lupus Unit Chapter 150, Clinical and serologic features of systemic Chapter 145, Management of renal lupus
St. Thomas’ Hospital sclerosis
London, United Kingdom Joshua Farber, MD
Chapter 37, The heart in rheumatic disease Paul Dieppe, BSc, MD, FRCP, FFPH Senior Investigator
Emeritus Professor of Health and Wellbeing Laboratory of Molecular Immunology, DIR
Karel De Ceulaer, MD, Dipl Rheum University of Exeter Medical School National Institute of Allergy and Infectious Diseases
Consultant Rheumatologist Exeter, United Kingdom National Institutes of Health
Department of Medicine Chapter 57, Placebos, caring, and healing in Bethesda, Maryland
University of the West Indies rheumatology Chapter 12, Cytokines
Kingston, Jamaica
Chapter 209, Sickle cell disease and other Patricia Dolan, PhD Anders Fasth, MD, PhD
hemoglobinopathies Reader in Biomechanics Professor, Department of Pediatrics
Centre for Applied Anatomy University of Gothenburg
Adriana A. de Jesus, MD, PhD University of Bristol Consultant, Department of Pediatric Immunology
Staff Scientist Bristol, United Kingdom; The Queen Silvia Children’s Hospital
Translational Autoinflammatory Diseases Studies Visiting Professor Gothenburg, Sweden
National Institute of Allergy and Infectious Diseases Sir Run Run Shaw Hospital Chapter 107, Clinical features of juvenile idiopathic
National Institutes of Health Zheijang University arthritis
Bethesda, Maryland Hanghzhou, China
Chapter 174, Monogenic autoinflammatory diseases Chapter 7, Biomechanics of spinal degeneration Eugen Feist, MD
Professor of Medicine
Salvatore De Vita, MD Tracy J. Doyle, MD, MPH Division of Rheumatology and Clinical Immunology
Professor of Rheumatology Instructor in Medicine Charité-Universitätsmedizin Berlin
Department of Medical Area Harvard Medical School Berlin, Germany
University of Udine; Pulmonary and Critical Care Division Chapter 33, Laboratory tests in rheumatic disorders
Chief, Rheumatology Clinic Brigham and Women’s Hospital
Academic Hospital ‘S. Maria della Misericordia’ Boston, Massachusetts Candace H. Feldman, MD, MPH, ScD
Udine, Italy Chapter 38, The lung in rheumatic disease Associate Physician
Chapter 171, Cryoglobulinemia Division of Rheumatology, Immunology and Allergy
George S.M. Dyer, MD Brigham and Women’s Hospital;
Maarten de Wit, PhD Assistant Professor of Orthopedic Surgery Assistant Professor of Medicine
Department of Medical Humanities Harvard Medical School Harvard Medical School
VU University Medical Centre Orthopaedic Hand Surgeon Boston, Massachusetts
Amsterdam, The Netherlands Brigham and Women’s Hospital Chapter 133, Epidemiology and classification of
Chapter 52, The patient perspective Boston, Massachusetts systemic lupus erythematosus
Chapter 84, The wrist and hand
Debbie Feldman, PhD Anthony J. Freemont, BSc, MB BS, MD, FRCP, Garry E. Gold, MD
Professor FRCPath Professor of Radiology
Faculty of Medicine, School of Rehabilitation Procter Professor of Pathology Stanford School of Medicine
Université de Montréal Director of MMPathIC Palo Alto, California
Montréal, Canada Faculty of Biology, Medicine and Health Chapter 47, Functional magnetic resonance imaging
Chapter 112, Rehabilitation and psychosocial issues in University of Manchester
juvenile idiopathic arthritis Manchester, United Kingdom Raphaela Goldbach-Mansky, MD, MHS
Chapter 34, Synovial fluid analysis Chief, Translational Auto-inflammatory Disease
Andrew Filer, MBChB, PhD Studies
Senior Lecturer Kevin B. Fricka, MD National Institute of Allergy and Infectious Diseases
School of Immunity and Infection Clinical Instructor National Institutes of Health
The University of Birmingham Adult Reconstructive Hip and Knee Fellowship Bethesda, Maryland
Honorary Consultant Rheumatologist Anderson Orthopaedic Clinic Chapter 174, Monogenic autoinflammatory diseases
Department of Rheumatology Alexandria, Virginia
University Hospitals Birmingham NHS Foundation Chapter 85, The hip Sarah J. Goldingay, MA, PhD
Trust Lecturer
Birmingham, United Kingdom Cem Gabay, MD College of Humanities
Chapter 1, The synovium Professor of Medicine University of Exeter
Head, Division of Rheumatology Exeter, United Kingdom
David F. Fiorentino, MD, PhD University Hospitals of Geneva Chapter 57, Placebos, caring, and healing in
Professor Geneva, Switzerland rheumatology
Department of Dermatology Chapter 70, Interleukin-1 inhibitors
Stanford University School of Medicine José A. Gómez-Puerta, MD, PhD, MPH
Stanford, California Massimo Gadina, PhD Grupo de Inmunología Celular e Inmunogenética
Chapter 158, Clinical significance of autoantibodies in Chief Universidad de Antioquia
inflammatory muscle disease Translational Immunology Section Medellín, Colombia
National Institute of Arthritis and Musculoskeletal Chapter 143, Management of nonrenal and non–
Benjamin A. Fisher, MD(res) MBBS and Skin Diseases, National Institutes of Health central nervous system lupus
Senior Clinical Lecturer National Institutes of Health
Institute of Inflammation and Ageing Bethesda, Maryland Susan M. Goodman, MD
University of Birmingham Chapter 12, Cytokines Professor of Clinical Medicine
Honorary Consultant Rheumatologist Weill Cornell Medical School
University Hospitals Birmingham NHS Trust Saviana Gandolfo, MD Attending Physician
Birmingham, United Kingdom Rheumatology Clinic Division of Rheumatology
Chapter 69, Biologic agents: monoclonal antibodies and Department of Medical Area Hospital for Special Surgery
receptor antagonists Academic Hospital ‘S. Maria della Misericordia’ New York, New York
Udine, Italy Chapter 60, Perioperative care of the rheumatic disease
John D. Fisk, PhD Chapter 171, Cryoglobulinemia patient
Nova Scotia Health Authority Jeremy Gauntlett-Gilbert, PhD, D.Clin.Psy. Caroline Gordon, MD, MA, FRCP
Associate Professor of Psychiatry Principal Clinical Psychologist Professor of Rheumatology
Associate Professor Psychology & Neuroscience Bath Centre for Pain Services Consultant Rheumatologist
Assistant Professor of Medicine Royal United Hospitals Institute of Inflammation and Ageing
Dalhousie University Bath, United Kingdom; University of Birmingham
Halifax, Nova Scotia, Canada Visiting Fellow Birmingham, United Kingdom
Chapter 144, Management of central nervous system Faculty of Health and Applied Sciences Chapter 69, Biologic agents: monoclonal antibodies and
lupus University of the West of England receptor antagonists
Bristol, United Kingdom
Raymond H. Flores, MD Chapter 56, Multi-disciplinary approaches to managing Sharon M. Gordon, DDS, MPH, PhD
Associate Professor of Medicine chronic pain in arthritis Professor and Chair
Division of Rheumatology and Clinical Immunology Department of Foundational Sciences
University of Maryland School of Medicine Steffen Gay, MD Associate Dean for Research
Baltimore, Maryland Professor of Experimental Rheumatology East Carolina University School of Dental Medicine
Chapter 89, Entrapment neuropathies and compartment Lead Physician Greenville, North Carolina
syndromes Center for Experimental Rheumatology Chapter 88, The temporomandibular joint
University Hospital Zürich
Lindsy Forbess, MD, MSc Zürich, Switzerland Rachel Gorodkin, MBChB, PhD, FRCP
Assistant Professor of Medicine Chapter 23, Epigenetics Consultant Rheumatologist
Division of Rheumatology The Kellgren Centre for Rheumatology
Cedars-Sinai Medical Center Lianne S. Gensler, MD Manchester Royal Infirmary;
Los Angeles, California Assistant Clinical Professor of Medicine Honorary Lecturer
Chapter 163, Polyarteritis nodosa and Cogan syndrome University of California San Francisco Faculty of Medical and Human Sciences
San Francisco, California University of Manchester
David A. Fox, MD Chapter 121, Clinical features of axial spondyloarthritis Manchester, United Kingdom
Professor of Medicine Chapter 90, Complex regional pain syndrome (reflex
Division of Rheumatology Danielle M. Gerlag, MD, PhD sympathetic dystrophy)
University of Michigan Medical School Associate Professor
Ann Arbor, Michigan Department of Clinical Immunology and Simon Görtz, MD
Chapter 73, Interleukin-17, interleukin-12, and Rheumatology Chief Resident
interleukin-23 inhibitors Academic Medical Center, University of Amsterdam Department of Orthopaedic Surgery
Amsterdam, The Netherlands; University of California, San Diego
Head, Clinical Unit Cambridge San Diego, California
Department of Research and Development Chapter 85, The hip
Cambridge, United Kingdom
Chapter 45, Minimally invasive procedures

Andrew J. Grainger, BM BS, MRCP, FRCR Karlene Hagley, MBBS, MD Philip J. Hashkes, MD, MSc
Consultant and Honorary Senior Lecturer Consultant in Internal Medicine Head, Pediatric Rheumatology Unit
Department of Musculoskeletal Radiology Spanish Town Hospital Department of Pediatrics
Leeds Teaching Hospitals St. Catherine, Jamaica; Shaare Zedek Medical Center
Leeds, United Kingdom Associate Lecturer Clinical Associate Professor of Pediatrics
Chapter 185, Imaging of osteoarthritis Department of Medicine Hebrew University Hadassah School of Medicine
University of the West Indies Jerusalem, Israel;
Ellen M. Gravallese, MD Kingston, Jamaica Associate Professor of Medicine and Pediatrics
Myles J. McDonough Chair in Rheumatology Chapter 209, Sickle cell disease and other Cleveland Clinic Lerner College of Medicine
Chief, Division of Rheumatology hemoglobinopathies Case Western Reserve University
Department of Medicine Cleveland, Ohio
University of Massachusetts Memorial Health Care Shuhong Han, PhD Chapter 109, Management of juvenile idiopathic arthritis
Professor of Medicine Assistant Scientist
University of Massachusetts Medical School Division of Rheumatology and Clinical Immunology Andrew Bassim Hassan, DPhil, FRCP
Worcester, Massachusetts University of Florida College of Medicine Head, Sarcoma and TYA Oncology Unit
Chapter 100, Pathogenesis and pathology of Gainesville, Florida NHS Department of Oncology
rheumatoid arthritis Chapter 139, Autoantibodies in systemic lupus Oxford Haematology and Cancer Centre
erythematosus Oxford University Hospitals Trust;
Jeffrey D. Greenberg, MD, MPH Professor of Medical Oncology
Assistant Professor of Medicine John G. Hanly, MD, FRCP(C) Sir William Dunn School of Pathology
Division of Rheumatology Professor of Medicine and Pathology University of Oxford
New York University School of Medicine Dalhousie University; Oxford, United Kingdom
New York, New York Attending Staff Rheumatologist Chapter 217, Bone tumors
Chapter 79, Precision medicine and pharmacogenomics Queen Elizabeth II Health Sciences Center
in rheumatology Halifax, Nova Scotia, Canada Lukas Haupt, MD
Chapter 144, Management of central nervous system Second Department of Medicine
Sharon Grieve, MRes Clinical Research, lupus Hietzing Hospital
BSc(Hons) Vienna, Austria
Pain Lead Research Nurse Eric P. Hanson, MD Appendix: Classification and Diagnostic Criteria
Complex Regional Pain Syndrome Service Senior Clinical Fellow
Royal United Hospitals Immunodeficiency and Inflammation Unit, Gillian A. Hawker, MD, MSc
Bath, United Kingdom; Autoimmunity Branch Professor of Medicine
Visiting Research Fellow National Institute of Arthritis Musculoskeletal and Division of Rheumatology
Faculty of Health and Applied Sciences Skin Disease University of Toronto
University of the West of England National Institutes of Health Chief of Medicine
Bristol, United Kingdom Bethesda, Maryland Women’s College Hospital
Chapter 56, Multi-disciplinary approaches to managing Chapter 11, Signal transduction in immune cells Toronto, Ontario, Canada
chronic pain in arthritis Chapter 186, Assessment of the patient with
Boulos Haraoui, MD, FRCPC osteoarthritis and measurement of outcomes
Luiza Guilherme, PhD Associate Professor of Medicine
Professor of Immunology Université de Montréal Philip N. Hawkins, PhD, FRCP, FRCPath, FMedSci
Heart Institute - InCor Montreal, Québec, Canada Professor of Medicine
University of São Paulo School of Medicine Chapter 65, Synthetic disease-modifying antirheumatic University College London
Institute for Immunology Investigation drugs and leflunomide National Amyloidosis Centre
National Institute for Science and Technology Royal Free Hospital
São Paulo, Brazil Sarah A. Hardcastle, MBChB, BSc (Hons), PhD London, United Kingdom
Chapter 117, Acute rheumatic fever Specialist Registrar in Rheumatology Chapter 177, Amyloidosis
Royal National Hospital for Rheumatic Diseases
Ahmet Gül, MD Bath, United Kingdom Peter Heeringa, PhD
Professor of Internal Medicine Chapter 196, Clinical features of osteoporosis Professor of Pathology and Medical Biology
Division of Rheumatology University of Groningen, University Medical Center
Istanbul University John B. Harley, MD, PhD Groningen
Istanbul Faculty of Medicine Professor of Pediatrics Groningen, The Netherlands
Istanbul, Turkey University of Cincinnati College of Medicine; Chapter 162, Biology and immunopathogenesis of
Chapter 167, Behçet disease Director, Center for Autoimmune Genomics and vasculitis
Sarthak Gupta, MD Cincinnati Children’s Hospital Medical Center Turid Heiberg, PhD
Henry Metzger Scholar Cincinnati, Ohio Professor and Research Leader
Systemic Autoimmunity Branch Chapter 136, Genetics of systemic lupus erythematosus Department of Research, Innovation and Education
National Institute of Arthritis and Musculoskeletal Oslo University Hospital
and Skin Diseases Tayseer G. Haroun, MBBS Oslo, Norway
National Institutes of Health Fellow Chapter 104, Multidisciplinary non-pharmacologic
Bethesda, Maryland Division of Rheumatology and Immunology approach to rheumatoid arthritis
Chapter 140, Pathogenesis of systemic lupus Duke University Medical Center
erythematosus Durham, North Carolina Dick Heinegård, MD, PhD†
Chapter 146, Systemic lupus erythematosus in pregnant Professor, Department of Clinical Sciences
Matilda Hagan, MD patients and neonatal lupus Division of Rheumatology–Molecular Skeletal
Division of Gastroenterology Biology
University of Maryland School of Medicine Lund University Faculty of Medicine
Baltimore, Maryland Lund, Sweden
Chapter 39, The gastrointestinal tract and rheumatic Chapter 2, The articular cartilage

Simon M. Helfgott, MD, CM Thomas W.J. Huizinga, MD, PhD Sarada Jaimungal, MBBS
Associate Professor of Medicine Professor of Rheumatology Fellow
Harvard Medical School; Leiden University Medical Center Division of Endocrinology, Diabetes and Nutrition
Director of Education and Fellowship Training Leiden, The Netherlands University of Maryland School of Medicine
Division of Rheumatology Chapter 71, Interleukin-6 inhibitors Baltimore, Maryland
Brigham and Women’s Hospital Chapter 197, Pathophysiology of osteoporosis
Boston, Massachusetts M. Elaine Husni, MD, MPH
Chapter 207, Rheumatoid manifestations of endocrine Staff Physician Judith A. James, MD, PhD
and lipid disease Orthopedic and Rheumatologic Institute Member and Chair
Cleveland Clinic Arthritis and Clinical Immunology Research
Kim Henriksen, PhD Cleveland, Ohio Program
Head of Musculoskeletal Diseases Chapter 127, Classification and epidemiology of Oklahoma Medical Research Foundation
NordicBioscience psoriatic arthritis Professor of Medicine and Adjunct Professor of
Herlev, Denmark Pathology
Chapter 198, Biochemical markers in bone disease Robert D. Inman, MD Oklahoma University Health Sciences Center
Professor of Medicine and Immunology Oklahoma City, Oklahoma
Rana S. Hinman, BPhysio, PhD University of Toronto Faculty of Medicine Chapter 134, Preclinical features of systemic lupus
Professor of Physiotherapy Kremil Research Institute erythematosus
Centre for Health, Exercise, and Sports Medicine University Health Network
University of Melbourne Toronto, Ontario, Cananda M. Kassim Javaid, MBBS, BMedSCI, PhD, FRCP
Melbourne, Australia Chapter 118, Reactive arthritis Associate Professor
Chapter 55, Principles of rehabilitation: physical and Nuffield Department of Orthopaedics,
occupational therapy Dai Inoue, PhD Rheumatology, and Musculoskeletal Science
Assistant Professor of Radiology University of Oxford
Pauline Y.P. Ho, MBBCh (Honours), BSC, MSc, Kanazawa University Hospital Oxford, United Kingdom
PhD, FRCP Kanazawa, Japan Chapter 215, Heritable connective tissue disorders
Consultant Rheumatologist and Honorary Senior Chapter 178, IgG4-related disease
Lecturer Pooya Javidan, MD
The Kellgren Centre for Rheumatology Zacharia Isaac, MD Fellow, Lower Extremity Reconstruction
Manchester Royal Infirmary Director, Interventional Physical Medicine and Department of Orthopaedic Surgery
The University of Manchester Rehabilitation Scripps Green Hospital
Manchester, United Kingdom Department of Physical Medicine and Rehabilitation La Jolla, California
Chapter 128, Clinical features of psoriatic arthritis Harvard Medical School Chapter 85, The hip
Medical Director, Comprehensive Spine Care
Marc C. Hochberg, MD, MPH, MACP Center Rose-Marie Javier, MD
Professor of Medicine and Epidemiology and Public Brigham and Women’s Hospital Senior Lecturer
Health Boston, Massachusetts Medical University Louis Pasteur;
Head, Division of Rheumatology and Clinical Chapter 81, Low back pain Senior Attending Physician
Immunology Rheumatology Unit
University of Maryland School of Medicine John D. Isaacs, BSc (hon), MB BS, PhD University Hospital Hautpierre
Acting Director, Gerontology Research Education Professor of Clinical Rheumatology Strasbourg, France
and Clinical Center Institute of Cellular Medicine Chapter 211, Gaucher disease
Veterans Affairs Maryland Health Care System Newcastle University;
Baltimore, Maryland Consultant Rheumatologist Roy Jefferis, BSc, PhD, FRSC, CChem, MRCP,
Chapter 188, Management of osteoarthritis Musculoskeletal Directorate FRCPath, DSc
Chapter 197, Pathophysiology of osteoporosis Freeman Hospital Professor Emeritus
Newcastle upon Tyne, United Kingdom School of Immunity and Infection
Markus H. Hoffmann, PhD Chapter 76, Emerging therapeutic targets University of Birmingham
Research Group Leader Birmingham, United Kingdom
Department of Medicine 3 Maura D. Iversen, BSc, MPH, DPT, SD Chapter 69, Biologic agents: monoclonal antibodies and
Friedrich Alexander University Erlangen- Professor and Chair receptor antagonists
Nuremberg Department of Physical Therapy
Erlangen, Germany Northeastern University Alyssa K. Johnsen, MD, PhD
Chapter 99, Autoantibodies in rheumatoid arthritis Senior Instructor in Medicine Associate Director
Harvard Medical School Global Clinical Research
Michael F. Holick, MD, PhD Behavioral Scientist Bristol Myers Squibb
Professor of Medicine, Physiology and Biophysics Section of Clinical Sciences Princeton, New Jersey
Section of Endocrinology, Nutrition, and Diabetes Division of Rheumatology, Immunology & Allergy Chapter 66, Methotrexate
Boston University School of Medicine Brigham & Women’s Hospital
Boston, Massachusetts Boston, Massachusetts Joanne M. Jordan, MD, MPH
Chapter 201, Osteomalacia and rickets Chapter 54, Arthritis patient education, self- Joseph P. Archie, Jr. Eminent Professor of Medicine
management, and health promotion Vice Dean, Faculty Affairs and Leadership
Christopher R. Holroyd, BM, MRCP Development
Consultant Rheumatologist Douglas A. Jabs, MD, MBA University of North Carolina at Chapel Hill School
MRC Lifecourse Epidemiology Unit Professor and Chair Emeritus of Ophthalmology of Medicine
University of Southampton Professor of Medicine Chapel Hill, North Carolina
Southampton, United Kingdom Director, New York Eye and Ear Infirmary of Mount Chapter 179, Epidemiology and classification of
Chapter 195, Epidemiology and classification of Sinai osteoarthritis
osteoporosis The Mount Sinai Hospital
New York, New York; Andrew A. Joyce, MD
Cathy Holt, BEng, PhD Adjunct Professor of Epidemiology Resident Physician
Professor of Biomechanics and Orthopaedic The Johns Hopkins University Department of Physical Medicine and Rehabilitation
Engineering Bloomberg School of Public Health Harvard Medical School
School of Engineering Baltimore, Maryland Spaulding Rehabilitation Hospital
Cardiff University Chapter 36, The eyes in rheumatic disease Boston, Massachusetts
Cardiff, Wales Chapter 81, Low back pain
Chapter 6, Biomechanics of peripheral Joints

Michelle Jung, MD, FRCPC Jonathan Kay, MD Kathleen D. Kolstad, MD, PhD
Alberta Health Services Professor of Medicine Fellow
Calgary, Alberta, Canada Division of Rheumatology Division of Immunology and Rheumatology
Chapter 142, Assessing disease activity and outcomes University of Massachusetts Medical School; Stanford University School of Medicine
in systemic lupus erythematosus Director of Clinical Research Stanford, California
Division of Rheumatology Chapter 158, Clinical significance of autoantibodies in
Ruba Kado, MD UMass Memorial Medical Center inflammatory muscle disease
Clinical Instructor in Internal Medicine Worcester, Massachusetts
Division of Rheumatology Chapter 77, Biosimilars in rheumatology Leah Kottyan, PhD
University of Michigan Medical School Chapter 214, Miscellaneous arthropathies Research Associate
Ann Arbor, Michigan Center for Autoimmune Genomics and Etiology
Chapter 73, Interleukin-17, interleukin-12, and Richard M. Keating, MD, MHS Cincinnati Children’s Hospital Medical Center
interleukin-23 inhibitors Program Director, Rheumatology Fellowship Cincinnati, Ohio
Division of Rheumatology Chapter 136, Genetics of systemic lupus erythematosus
Tsuneyasu Kaisho, MD, PhD Scripps Clinic and Scripps Green Hospital
Professor La Jolla, California Virginia Byers Kraus, MD, PhD
Department of Immunology Chapter 67, Immunosuppressive agents: cyclosporine, Professor of Medicine (Rheumatology),
Institute of Advanced Medicine cyclophosphamide, azathioprine, mycophenolate Orthopaedics, and Pathology
Wakayama, Japan mofetil, and tacrolimus Duke Molecular Physiology Institute
Chapter 9, Principles of innate immunity Duke University Medical Center
Jennifer A. Kelly, MPH Durham, North Carolina
Cees G.M. Kallenberg, MD, PhD Research Project Director Chapter 187, Preclinical and early osteoarthritis
Professor of Rheumatology and Clinical Arthritis and Clinical Immunology Research Chapter 211, Rare osteoarthritis: ochronosis and Kashin-
Immunology Program Beck disease
University of Groningen, University Medical Center Oklahoma Medical Research Foundation
Groningen Oklahoma City, Oklahoma Tore K. Kvien, MD, PhD
Groningen, The Netherlands Chapter 136, Genetics of systemic lupus erythematosus Professor of Rheumatology
Chapter 162, Biology and immunopathogenesis of University of Oslo Faculty of Medicine
vasculitis David Kendler, MD Head, Department of Rheumatology
Professor of Medicine Diakonhjemmet Hospital
David Kane, PhD, FRCPI University of British Columbia Olso, Norway
Clinical Professor of Rheumatology Vancouver, Canada Chapter 104, Multidisciplinary non-pharmacologic
Trinity College Dublin Chapter 51, Dual x-ray absorptiometry and approach to rheumatoid arthritis
Consultant Rheumatologist measurement of bone
Adelaide and Meath Hospital Robert Lafyatis, MD
Dublin, Ireland Munther A. Khamashta, MD, PhD, FRCP Professor of Medicine
Chapter 48, Musculoskeletal ultrasound Professor and Consultant Physician Divison of Rheumatology
King’s College London; University of Pittsburgh School of Medicine
Mariana J. Kaplan, MD Director, Graham Hughes Lupus Research Unit Pittsburgh, Pennsylvania
Senior Investigator St. Thomas Hospital Chapter 151, Etiology and pathogenesis of systemic
Chief, Systemic Autoimmunity Branch London, United Kingdom sclerosis
Intramural Research Program Chapter 148, Antiphospholipid syndrome: pathogenesis,
National Institute of Arthritis and Musculoskeletal diagnosis, and management Robert B.M. Landewé, MD
and Skin Diseases Professor of Clinical Immunology & Rheumatology
National Institutes of Health Irfan R. Khan, MD University of Amsterdam
Bethesda, Maryland Assistant Professor of Ophthalmology Academic Medical Center
Chapter 140, Pathogenesis of systemic lupus Johns Hopkins School of Medicine Amsterdam, The Netherlands;
erythematosus Wilmer Eye Institute Consultant Rheumatologist
Baltimore, Maryland Atrium Medical Center
Morten Asser Karsdal, MSc, PhD, mMBA Chapter 36, The eyes in rheumatic disease Heerlen, The Netherlands
Head of Research and Development Chapter 29, Interpreting the medical literature for the
Nordic Bioscience Dinesh Khanna, MD, MS rheumatologist
Herlev, Denmark Professor of Medicine Chapter 50, Use of imaging as an outcome measure in
Chapter 198, Biochemical markers in bone disease Frederick G. l. Huetwell Professor of Rheumatology clinical trials
Director, University of Michigan Scleroderma
Dimitrios G. Kassimos, MD, MSc Program Carol A. Langford, MD, MHS, FACP
Consultant Rheumatologist Division of Rheumatology Director, Center for Vasculitis Care and Research
Director of Education Department University of Michigan Medical School Harold C. Schott Chair in Rheumatic and
401 General Military Hospital of Athens Ann Arbor, Michigan Immunologic Diseases
Athens, Greece Chapter 152, Assessing disease activity and outcome in Cleveland Clinic;
Chapter 205, Neuropathic arthropathy systemic sclerosis Associate Professor of Medicine
Cleveland Clinic Lerner College of Medicine of
Daniel L. Kastner, MD, PhD Mari Klokkerud, OT, PhD Case Western Reserve University
Scientific Director, Division of Intramural Research Researcher Cleveland, Ohio
National Human Genome Research Institute National Advisory Unit on Rehabilitation in Chaper 165, Takayasu arteritis
National Institutes of Health Rheumatology
Bethesda, Maryland Diakonhjemmet Hospital Arthur N. Lau, MD, FRCPC
Chapter 174, Monogenic autoinflammatory diseases Department of Rheumatology Assistant Professor
Oslo, Norway Division of Rheumatology
Mitsuhiro Kawano, MD, PhD Chapter 104, Multidisciplinary non-pharmacologic McMaster University
Director, Division of Rheumatology approach to rheumatoid arthritis Michael G. DeGroote School of Medicine
Department of Internal Medicine Hamilton, Ontario, Canada
Kanazawa University Hospital Chapter 200, Glucocorticoid-induced osteoporosis
Kanazawa, Japan
Chapter 178, IgG4-related disease
Robert A. Lavin, MD, MS Adam P. Lightfoot, PhD C. Ronald MacKenzie, MD
Assistant Professor of Neurology Lecturer Attending Physician
University of Maryland School of Medicine School of Healthcare Science Department of Medicine and Rheumatology
Director of Chronic Pain Management Manchester Metropolitan University Hospital for Special Surgery
Department of Neurology Manchester, United Kingdom Professor of Clinical Medicine and Medical Ethics
VA Maryland Health Care System Chapter 42, The muscles in rheumatic disease Weill Medical College of Cornell University
Adjunct Professor C. Ronald MacKenzie Chair in Ethics and Medicine
Department of Occupational Medicine Geoffrey O. Littlejohn, MBBS(Hons), MD, MPH Hospital for Special Surgery
Johns Hopkins University School of Medicine Clinical Professor of Medicine New York, New York
Baltimore, Maryland Monash University Chapter 30, Ethics in clinical trials
Chapter 62, Principles of pharmacologic pain Emeritus Director, Rheumatology Unit Chapter 60, Perioperative care of the rheumatic disease
management Monash Medical Centre patient
Clayton, Victoria, Australia
Ronald M. Laxer, MDCM, FRCPC Chapter 204, Diffuse idiopathic skeletal hyperostosis Julia Manasson, MD
Professor of Pediatrics and Medicine Fellow
Division of Rheumatology Pilar Lorenzo, PhD Division of Rheumatology
University of Toronto Faculty of Medicine Researcher, Department of Clinical Sciences New York University School of Medicine
The Hospital for Sick Children Division of Rheumatology—Molecular Skeletal New York, New York
Toronto, Ontario, Canada Biology Chapter 24, The microbiome in rheumatic diseases
Chapter 109, Management of juvenile idiopathic arthritis Lund University Faculty of Medicine
Lund, Sweden David Charles Mangham, BSc, MB BS, FRCPath
Thomas J. Learch, MD Chapter 2, The articular cartilage Professor of Applied Molecular Pathology
Chief of Musculoskeletal Imaging Deputy Director of MMPathIC
Department of Imaging Rik J.U. Lories, MD, PhD Faculty of Biology, Medicine and Health
Cedars-Sinai Medical Center Professor of Medicine University of Manchester
Los Angeles, California Skeletal Biology and Engineering Research Center Manchester, United Kingdom
Chapter 96, Imaging of rheumatoid arthritis Katholieke Universiteit Leuven Chapter 34, Synovial fluid analysis
Leuven, Belgium
Cianna Leatherwood, MD Chapter 130, Animal models of psoriatic arthritis Rong Mao, PhD
Rheumatology Fellow Investigator
Harvard Medical School Thomas A. Luger, MD Department of Medicine
Brigham and Women’s Hospital Professor of Dermatology Stanford School of Medicine
Boston, Massachusetts University of Münster Palo Alto, California
Chapter 59, Drugs and pregnancy Münster, Germany Chapter 14, The complement system
Chapter 207, Rheumatoid manifestations of endocrine Chapter 35, The skin in rheumatic disease
and lipid disease Lyn M. March, MBBS, MSc, PhD, FRACP, FAFPHM
Ingrid E. Lundberg, MD, PhD Liggins Professor of Rheumatology and
Jennifer S. Lewis, PhD MSc, Dip COT Professor of Rheumatology Musculoskeletal Epidemiology
Senior Clinical Research Occupational Therapist Department of Medicine, Solna University of Sydney Professorial Rheumatology
Complex Regional Pain Syndrome and Cancer Late Karolinska Institutet Deparment
Effects Rehabilitation Service Stockholm, Sweden Royal North Shore Hospital
Royal National Hospital for Rheumatic Diseases, Chapter 157, Etiology and pathogenesis of St Leonards, New South Wales, Australia
Royal United Hospitals Bath NHS Trust inflammatory muscle disease (myositis) Chapter 53: Treatment recommendations and "treat to
Bath, United Kingdom; target"
Senior Lecturer, Department of Allied Health Raashid A. Luqmani, DM, FRCP, FRCPE
Professions Professor of Rheumatology Alejandro Olivé Marqués, MD, PhD
University of the West of England Nuffield Department of Orthopaedics, Rheumatology Service
Bristol, United Kingdom Rheumatology and Musculoskeletal Science Hospital Universitari Germans Trias i Pujol
Chapter 56, Multi-disciplinary approaches to managing
University of Oxford; Barcelona, Spain
chronic pain in arthritis
Consultant Rheumatologist Chapter 173, Adult-onset Still disease
Nuffield Orthopaedic Centre
George Lewith, MD, FRCP Oxford, United Kingdom Javier Márquez, MD
Professor of Health Research Chapter 164, Antineutrophil cytoplasm antibody– Rheumatologist
Primary Care and Population Sciences associated vasculitis Department of Internal Medicine
University of Southampton Medical School Hospital Pablo Tobon Uribe
Honorary Consultant Physician Hanna M. Lythgoe, MBChB, MRCPCH Medellín, Colombia
Primary Care and Population Sciences Clinical Research Fellow Chapter 114, Mycobacterial, brucellar, fungal, and
Southampton University NHS Hospitals Trust Department of Women’s and Children’s Health, parasitic arthritis
Southampton, United Kingdom Institute of Translational Medicine
Chapter 58, Complementary and alternative medicine
University of Liverpool Emilio Martín-Mola, MD, PhD
Clinical Research Fellow Department of Rheumatology
Katherine P. Liao, MD, MPH NIHR Alder Hey Clinical Research Facility Hospital Universitario La Paz
Assistant Professor of Medicine Alder Hey Children’s NHS Foundation Trust Madrid, Spain
Harvard Medical School Liverpool, United Kingdom Chapter 113, Bacterial arthritis
Associate Physician Chapter 111, Connective tissue diseases in children
Division of Rheumatology, Immunology, and Manuel Martínez-Lavín, MD
Allergy Kristin J. MacDougall, MD Chief, Rheumatology Department
Brigham and Women’s Hospital Division of Rheumatology National Institute of Cardiology;
Boston, Massachusetts Columbia University College of Physicians and Professor of Rheumatology
Chapter 92, Classification and epidemiology of
Surgeons National Autonomous University
rheumatoid arthritis
New York, New York Mexico City, Mexico
Chapter 103, Management of established rheumatoid Chapter 213, Digital clubbing and hypertrophic
arthritis osteoarthropathy

Eric L. Matteson, MD, MPH Lachy McLean, MD, PhD, FRCP Ingrid Möller, MD, PhD
Professor of Medicine Vice President Assistent Professor of Anatomy
Divisions of Rheumatology and Epidemiology Takeda Pharmaceuticals University of Barcelona
Mayo Clinic College of Medicine San Diego, California Director, Department of Rheumatology
Rochester, Minnesota Chapter 190, Etiology and pathogenesis of gout Instituto Poal de Reumatología
Chapter 95, Extraarticular features of rheumatoid Barcelona, Spain
arthritis Andrew Robert Ian Melville, MBBS Lond, MA Chapter 44, Aspiration and injection of joints and
Camb, MRCP periarticular tissues and intralesional therapy
Stephen J. Matzat, BA Core Medical Trainee
Department of Radiology Northwick Park Hospital Paul A. Monach, MD, PhD
Stanford School of Medicine London, United Kingdom Chief, Rheumatology Section
Palo Alto, California Chapter 154, Raynaud phenomenon VA Boston Healthcare System
Chapter 47, Functional magnetic resonance imaging Associate Professor, Section of Rheumatology
Laëtitia Michou, MD, PhD Boston University School of Medicine
Maureen D. Mayes, MD, MPH Associate Professor of Medicine Chapter 100, Pathogenesis and pathology of
Professor of Medicine CHU de Québec-Université Laval rheumatoid arthritis
Division of Rheumatology Endocrinology and Nephrology Unit
University of Texas Health Science Center Houston CHU de Québec Research Centre Barend Mons, PhD
McGovern School of Medicine Québec City, Canada Professor in Bio-Semantics
Houston, Texas Chapter 203, Paget disease of bone Department of Human Genetics
Chapter 149, Classification and epidemiology of Leiden University Medical Centre
systemic sclerosis (scleroderma) Rob Middleton, BMBCh, MA(Cantab), MRCS Leiden, The Netherlands
Botnar Research Centre Chapter 20, Open Science for Systems biology
Timothy McAlindon, MD, MPH, MRCP Nuffield Department of Orthopaedics,
Professor of Medicine Rheumatology and Musculoskeletal Science Elena Myasoedova, MD, PhD
Tufts University School of Medicine University of Oxford Assistant Professor of Medicine
Chief, Division of Rheumatology Oxford, United Kingdom Mayo Clinic College of Medicine
Tufts Medical Center Chapter 86, The knee Rochester, Minnesota
Boston, Massachusetts Chapter 95, Extraarticular features of rheumatoid
Chapter 206, Osteonecrosis Jamal A. Mikdashi, MD, MPH arthritis
Associate Professor of Medicine
Candy S. McCabe, PhD, MSc Division of Rheumatology and Clinical Immunology Jackie L. Nam, MBBCh, FCP, Rheum cert, PhD
Florence Nightingale Foundation Chair in Clinical University of Maryland School of Medicine Rheumatology Clinical Research Fellow
Nursing Practice Research Baltimore, Maryland Leeds Institute of Rheumatic and Musculoskeletal
University of the West of England Chapter 172, Primary angiitis of the central nervous Medicine
Bristol, United Kingdom; system University of Leeds
Royal United Hospitals NHS Foundation Trust Leeds, United Kingdom
Royal National Hospital for Rheumatic Diseases Frederick W. Miller, MD, PhD Chapter 102, Management of early rheumatoid arthritis
Bath, United Kingdom Chief, Environmental Autoimmunity Group
Chapter 56, Multi-disciplinary approaches to managing Program of Clinical Research Gauthier Namur, MD
chronic pain in arthritis National Institute of Environmental Health Sciences Department of Nuclear Medicine
National Institutes of Health Christian Hospital Center
Edward F. McCarthy, MD Bethesda, Maryland Liège, Belgium
Professor of Pathology and Orthopaedic Surgery Chapter 159, Management of inflammatory muscle Chapter 49, Bone scintigraphy and positron emission
Johns Hopkins University School of Medicine disease tomography
Baltimore, Maryland
Chapter 217, Bone tumors Paul D. Miller, MD Esperanza Naredo, MD
Medical Director Department of Rheumatology
Geraldine M. McCarthy, MD, FRCPI Colorado Center for Bone Research Hospital General Universitario Gregorio Marañon
Clinical Professor of Medicine Lakewood, Colorado Madrid, Spain
School of Medicine and Medical Science Chapter 202, Renal osteodystrophy Chapter 44, Aspiration and injection of joints and
University College Dublin; periarticular tissues and intralesional therapy
Consultant Rheumatologist Kirsten Minden, MD
Mater Misericordiae University Hospital Staff Scientist, Epidemiology Unit Barbara Neerinckx, MD
Dublin, Ireland German Rheumatism Research Center Berlin; Skeletal Biology and Engineering Research Center
Chapter 194, Basic calcium phosphate crystal Consultant in Pediatric Rheumatology Katholieke Universiteit Leuven
deposition disease Children´s University Hospital Charité Leuven, Belgium
Berlin, Germany Chapter 130, Animal models of psoriatic arthritis
Michael F. McDermott, MB, BCh, BAO, MRCPI, Chapter 106, Classification and epidemiology of juvenile
DMed idiopathic arthritis Amanda E. Nelson, MD, MSCR
Professor of Experimental Rheumatology Assistant Professor of Medicine
University of Leeds—St James’s University Hospital Jonathan J. Miner, MD, PhD Division of Rheumatology, Allergy, and
National Institute of Health Research—Leeds Instructor of Medicine Immunology
Musculoskeletal Biomedical Research Unit Division of Rheumatology Thurston Arthritis Research Center
Leeds Institute of Rheumatic and Musculoskeletal Washington Unviersity School of Medicine in St. University of North Carolina at Chapel Hill
Medicine Louis Chapel Hill, North Carolina
Leeds, United Kingdom St. Louis, Missouri Chapter 179, Epidemiology and classification of
Chapter 174, Monogenic autoinflammatory diseases Chapter 94, Clinical features of rheumatoid arthritis osteoarthritis

Dennis McGonagle, PhD, FRCPI Rikke Helene Moe, PhD, PT, MSc Philippa J.A. Nicolson, BPhty
Professor of Investigative Rheumatology Researcher Doctoral Candidate
Leeds Institute of Rheumatic & Musculoskeletal National Advisory Unit on Rehabilitation in Centre for Health, Exercise and Sports Medicine
Medicine Rheumatology University of Melbourne
University of Leeds Diakonhjemmet Hospital Melbourne, Australia
Leeds, United Kingdom Department of Rheumatology Chapter 55, Principles of rehabilitation: physical and
Chapter 129, Etiology and pathogenesis of psoriatic Oslo, Norway occupational therapy
arthritis Chapter 104, Multidisciplinary non-pharmacologic
approach to rheumatoid arthritis
César E. Fors Nieves, MD Voon H. Ong, PhD, MRCP Clarissa A. Pilkington, MBBS, BSc, FRCP,
Instructor, Department of Medicine Senior Lecturer in Rheumatology FRCPCH
Division of Rheumatology Centre for Rheumatology Honorary Lecturer
New York University School of Medicine University College London Institute of Child Health
Associate Director Royal Free Hospital University College London Medical School
Rheumatology Fellowship Program London, United Kingdom Consultant in Paediatric and Adolescent
NYU Medical Center Chapter 150, Clinical and serologic features of systemic Rheumatology
Hospital for Joint Diseases sclerosis Great Ormond Street Hospital NHS Trust
New York, New York London, United Kingdom
Chapter 79, Precision medicine and pharmacogenomics Patrik Önnerfjord, PhD Chapter 111, Connective tissue diseases in children
in rheumatology Associate Professor, Department of Clinical Sciences
Division of Rheumatology—Molecular Skeletal Michael H. Pillinger, MD
Ellen B. Nordal, MD, PhD Biology Associate Professor of Medicine and Pharmacology
Consultant, Department of Pediatrics Lund University Faculty of Medicine Director, Rheumatology Training
University Hospital of Northern Norway Lund, Sweden Director, Masters of Science in Clinical
Associate Professor, Department of Clinical Chapter 2, The articular cartilage Investigation Program
Medicine New York University School of Medicine
UIT the Arctic University of Norway Philippe Orcel, MD, PhD Section Chief, Rheumatology
Tromsø, Norway Professor of Rheumatology New York Harbor Health Care System–NY Campus
Chapter 107, Clinical features of juvenile idiopathic Paris-Diderot Faculty of Medicine Department of Veterans Affairs
arthritis University of Paris 7 New York, New York
Chief, Department of Musculoskeletal Diseases Chapter 13, Inflammation and its chemical mediators
Ulrich Nöth, MD, MHBA Rheumatology and Bone Diseases
Senior Consultant and Chief Hospital Lariboisière, Assistance Publique Hôpitaux Carlos Pineda, MD, MSc
Department of Orthopedics and Trauma Surgery de Paris Research Director
Evangelisches Waldkrankenhaus Spandau Paris, France Instituto Nacional de Rehabilitación
Berlin, Germany Chapter 211, Gaucher disease Mexico City, Mexico
Chapter 18, Principles of tissue engineering and cell- Chapter 213, Digital clubbing and hypertrophic
and gene-based therapy Carl Orr, MB BAO, BCh, BMedSci, MSc LMD, osteoarthropathy
Eleana Ntatsaki, MRCP(Rheumatology)(UK), MA The Centre for Arthritis and Rheumatic Disease Nicolò Pipitone, MD, PhD
MedEd, FHEA University College Dublin Unità di Reumatologia
Consultant in Rheumatology Dublin, Ireland Azienda Ospedaliera IRCCS di Reggio Emilia
Ipswich Hospital NHS Trust Chapter 131, Management of psoriatic arthritis Reggio Emilia, Italy;
Ipswich, United Kingdom; Dipartimento Biomedico di Medicina Interna a
Honorary Senior Clinical Lecturer John J. O’Shea, MD Specialistica
University College London Medical School Scientific Director Sezione di Reumatologia
London, United Kingdom National Institute of Arthritis and Musculoskeletal Università degli Studi di Palermo
Chapter 161, Classification and epidemiology of and Skin Diseases Palermo, Italy
vasculitis National Institutes of Health Chapter 166, Polymyalgia rheumatica and giant cell
Bethesda, Maryland arteritis
Chester V. Oddis, MD Chapter 11, Signal transduction in immune cells
Professor of Medicine Chapter 12, Cytokines Denis Poddubnyy, MD
Division of Rheumatology and Clinical Immunology Professor of Rheumatology
University of Pittsburgh School of Medicine Caroline Ospelt, MD, PhD Head, Rheumatology Unit
Pittsburgh, Pennsylvania Senior Researcher Charité University Hospital
Chapter 156, Classification, epidemiology, and clinical Center of Experimental Rheumatology Berlin, Germany
features of inflammatory muscle disease University Hospital Zürich Chapter 126, Management of axial spondyloarthritis
Zürich, Switzerland
Ahmed Omar, MBBch, MRCP Chapter 23, Epigenetics Janet E. Pope, MD, MPH, FRCPC
Clinical Fellow Professor of Medicine
Division of Rheumatology Seza Özen, MD University of Western Ontario
University of Toronto Faculty of Medicine Professor of Pediatric Rheumatology Head, Division of Rheumatology
Toronto, Ontario, Canada Hacettepe University Faculty of Medicine St. Joseph’s Health Care
Chapter 118, Reactive arthritis Ankara, Turkey London, Ontario, Canada
Chapter 169, IgA vasculitis (Henoch-Schoenlein Chapter 142, Assessing disease activity and outcomes
Michael J. Ombrello, MD purpura) in systemic lupus erythematosus
Investigator, Translational Genetics and Genomics
Unit Matthew J.S. Parker, MBChB, MRCP Anna Postolova, MD, MPH
National Institute of Arthritis and Musculoskeletal Advanced Trainee Fellow
and Skin Diseases Department of Rheumatology Division of Immunology and Rheumatology
National Institutes of Health Royal Prince Alfred Hospital Stanford University School of Medicine
Bethesda, Maryland Sydney, Australia Stanford, California
Chapter 19, Principles and techniques in molecular Chapter 208, Hemophilia and von Willebrand disease Chapter 158, Clinical significance of autoantibodies in
biology inflammatory muscle disease
Carlo Patrono, MD
Antonina Omisade, PhD Professor of Pharmacology Andrew Price, MBBS, BSc, Dphil, FRCS
Psychologist Catholic University School of Medicine Professor of Orthopaedic Surgery
Queen Elizabeth II Health Sciences Centre Rome, Italy Nuffield Department of Orthopaedics,
Halifax, Nova Scotia, Canada Chapter 63, Nonsteroidal antiinflammatory drugs Rheumatology and Musculoskeletal Science
Chapter 144, Management of central nervous system University of Oxford
lupus Consultant Orthopaedic Surgeon
Nuffield Orthopaedic Centre
Oxford University Hospitals Trust
Oxford, United Kingdom
Chapter 86, The knee

Luca Quartuccio, MD, PhD Elaine F. Remmers, PhD William H. Robinson, MD, PhD
Rheumatology Clinic Associate Investigator, Inflammatory Disease Associate Professor of Medicine
Department of Medical Area Section Division of Immunology and Rheumatology
Academic Hospital ‘S. Maria della Misericordia’ National Human Genome Research Institute Stanford School of Medicine
Udine, Italy National Institutes of Health Palo Alto, California
Chapter 171, Cryoglobulinemia Bethesda, Maryland Chapter 14, The complement system
Chapter 19, Principles and techniques in molecular
Lars Rackwitz, MD biology Valerie Rogers, MBBS, MRCPCH, BA (Hons) QTS
Senior Consultant Department of Paediatric Rheumatology
Department of Orthopaedics and Trauma Surgery Luis Requena, MD Bristol Royal Hospital for Children
Evangelisches Waldkrankenhaus Spandau Professor of Dermatology Bristol, United Kingdom;
Berlin, Germany Universidad Autónoma Consultant Paediatrician
Chapter 18, Principles of tissue engineering and cell- Chairman, Department of Dermatology Bath Centre for Pain Services
and gene-based therapy Fundación Jiménez Diaz Royal National Hospital For Rheumatic Diseases
Madrid, Spain Bath, United Kingdom
Helga Radner, MD Chapter 170, Cutaneous vasculitis and panniculitis Chapter 216, Hypermobility syndrome
Medical University Vienna
Department of Internal Medicine III Gary Reynolds, BSc, MBBS, PhD Ivan O. Rosas, MD
Division of Rheumatology Academic Clinical Lecturer Associate Professor of Medicine
Vienna, Austria Musculoskeletal Research Group Harvard Medical School
Chapter 43, Multimorbidity Newcastle University Pulmonary and Critical Care Division
Newcastle upon Tyne, United Kingdom Brigham and Women’s Hospital
Soumya Raychaudhuri, MD, PhD Chapter 76, Emerging therapeutic targets Boston, Massachusetts
Associate Professor of Medicine Chapter 38, The lung in rheumatic disease
Harvard Medical School Clio Ribbens, MD, PhD
Associate Physician Head of Clinic Ann K. Rosenthal, MD
Divisions of Rheumatology and Genetics Department of Rheumatology Will and Cava Ross Professor of Medicine
Brigham and Women’s Hospital University Hospital of Liège Chief of Rheumatology
Boston, Massachusetts; Liège, Belgium Medical College of Wisconsin;
Professor Chapter 49, Bone scintigraphy and positron emission Staff Physician
Arthritis Research UK Centre for Genetics and tomography Clement J. Zablocki VA Medical Center
Genomics Milwaukee, Wisconsin
University of Mancehster Bethan Richards, MBBS, MClinEpi, M Sports Med Chapter 193, Calcium pyrophosphate deposition
Manchester, United Kingdom Clinical Senior Lecturer in Medicine disease (pseudogout)
Chapter 97, Genetics of rheumatoid arthritis University of Sydney;
Head, Department of Rheumatology Andrew D. Rowan, PhD
Anthony C. Redmond, PhD Royal Prince Alfred Hospital Professor of Molecular Rheumatology
Professor and Head Sydney, New South Wales, Australia Head of the Musculoskeletal Research Group
Section of Clinical Biomechanics and Physical Chapter 53: Treatment recommendations and "treat to Institute of Cellular Medicine
Medicine target" Newcastle University
Leeds Institute of Rheumatic and Musculoskeletal Newcastle upon Tyne, United Kingdom
Medicine Bruce C. Richardson, MD, PhD Chapter 17, Tissue destruction and repair
University of Leeds; Professor of Internal Medicine
NIHR Leeds Biomedical Research Unit University of Michigan Medical School Martin Rudwaleit, MD
Leeds Teaching Hospitals NHS Trust Chief, Section of Rheumatology Professor of Internal Medicine and Rheumatology
Leeds, United Kingdom Veterans Affairs Hospital Klinikum Bielefeld
Chapter 87, The ankle and foot Ann Arbor, Michigan Bielefeld, Germany;
Chapter 141, Drug-induced lupus Professor
Jonathan L. Rees, MBBS, FRCS (Eng), MD, FRCS Charité University Medicine
(Tr & Orth) Pascal Richette, MD, PhD Berlin, Germany;
Professor Professor of Rheumatology Professor of Rheumatology
Nuffield Department of Orthopaedics, Université Paris Diderot University Hospital Ghent
Rheumatology and Musculoskeletal Science Assistance Publique-Hôpitaux de Paris Ghent, Belgium
University of Oxford Hôpital Lariboisière Chapter 119, Classification and epidemiology of
Oxford, United Kingdom Paris, France spondyloarthritis
Chapter 82, The shoulder Chapter 210, Hemochromatosis
Marina Rull, MD
Westley H. Reeves, MD Adrian Richter, PhD, MSc Chief, Immunology and Rheumatology Department
Marcia Whitney Schott Professor of Medicine Epidemiology Unit Instituto Nacional de Ciencias Medicas y Nutrición,
Division of Rheumatology and Clinical Immunology German Rheumatism Research Centre Salvador Zubirán
University of Florida College of Medicine Berlin, Germany México City, México
Gainesville, Florida Chapter 31, European biologics registers Chapter 44, Aspiration and injection of joints and
Chapter 139, Autoantibodies in systemic lupus periarticular tissues and intralesional therapy
Christopher T. Ritchlin, MD, MPH
Professor of Medicine Marite Rygg, MD, PhD
Linda M. Rehaume, PhD Center for Musculoskeletal Research Professor, Department of Laboratory Medicine,
Research Fellow University of Rochester Medical Center Children’s and Women’s Health
Translational Research Institute Rochester, New York Norwegian University of Science and Technology
The University of Queensland Diamantina Institute Chapter 129, Etiology and pathogenesis of psoriatic Consultant, Department of Pediatrics
Brisbane, Australia arthritis St. Olav’s Hospital
Chapter 124, Animal models of spondyloarthritis Trondheim, Norway
Susan Y. Ritter, MD, PhD Chapter 107, Clinical features of juvenile idiopathic
Associate Physician arthritis
Division of Rheumatology, Allergy and Immunology
Brigham and Women’s Hospital;
Instructor in Medicine
Harvard Medical School
Boston, Massachusetts
Chapter 176, Relapsing polychondritis
Kenneth G. Saag, MD, MSc Jose U. Scher, MD Bernadette C. Siaton, MD
Jane Knight Lowe Professor of Medicine Assistant Professor of Medicine Assistant Professor of Medicine
Division of Clinical Immunology and Division of Rheumatology Division of Rheumatology and Clinical Immunology
Rheumatology, Department of Medicine New York University School of Medicine University of Maryland School of Medicine
University of Alabama at Birmingham Director, Arthritis Clinic Baltimore, Maryland
Birmingham, Alabama NYU-Hospital for Joint Diseases Chapter 89, Entrapment neuropathies and compartment
Chapter 64, Systemic glucocorticoids in rheumatology New York, New York syndromes
Chapter 24, The microbiome in rheumatic diseases
Jane E. Salmon, MD Richard M. Siegel, MD, PhD
Professor of Medicine Georg Schett, MD Chief, Immunoregulation Section
Weill Cornell Medical College Professor of Rheumatology and Immunology Autoimmunity Branch
Collette Kean Research Chair Department of Internal Medicine 3 National Institute of Arthritis and Musculoskeletal
Hospital for Special Surgery University of Erlangen-Nuremberg Faculty of and Skin Diseases
New York, New York Medicine National Institutes of Health
Chapter 137, Immunopathology of systemic lupus Erlangen, Germany Bethesda, Maryland
erythematosus Chapter 16, Osteoimmunology Chapter 12, Cytokines

Donald M. Salter, MBChB, MD Adam I. Schiffenbauer, MD Joachim Sieper, PhD

Professor of Osteoarticular Pathology Staff Clinician Professor of Rheumatology
Centre for Genomics and Experimental Medicine Environmental Autoimmunity Group Charité University Hospital
University of Edinburgh National Institute of Environmental Health Sciences Berlin, Germany
MRC Institute of Genetics and Molecular Medicine National Institutes of Health Chapter 126, Management of axial spondyloarthritis
Edinburgh, United Kingdom Bethesda, Maryland
Chapter 5, Connective tissue responses to mechanical Chapter 159, Management of inflammatory muscle Julia Fridman Simard, ScD
stress disease Assistant Professor
Chapter 183, Pathogenesis and pathology of Departments of Health Research and Policy
osteoarthritis Alan L. Schiller, MD (Epidemiology) and Medicine (Immunology and
Founding Chair of Pathology Rheumatology)
Carlo Salvarani, MD Nova Southeastern University Medical School; Stanford University School of Medicine
Director, Division of Rheumatology Director, Pathology Residency Program Stanford, California;
Azienda Ospedaliera IRCCS di Reggio Emilia HCA Hospitals Assistant Professor of Medicine
Reggio Emilia, Italy; Miami, Florida Clinical Epidemiology Unit
Dipartimento Biomedico di Medicina Interna a Chapter 217, Bone tumors Karolinska Institutet
Specialistica Stockholm, Sweden
Sezione di Reumatologia Naomi Schlesinger, MD Chapter 25, Principles of epidemiology
Università degli Studi di Palermo Professor of Medicine
Palermo, Italy Chief, Division of Rheumatology Barry P. Simmons, MD
Chapter 166, Polymyalgia rheumatica and giant cell Rutgers Robert Wood Johnson Medical School Associate Professor of Orthopaedic Surgery
arteritis New Brunswick, New Jersey Harvard Medical School
Chapter 191, Clinical features of gout Chief, Hand and Upper Extremity Service
Linda J. Sandell, PhD Brigham and Women’s Hospital
Mildred B. Simon Research Professor and Director Benjamin E. Schreiber, MBBS, MA, MRCP Boston, Massachusetts
of Research Consultant in Pulmonary Hypertension and Chapter 84, The wrist and hand
Director, Center for Musculoskeletal Research Rheumatology
Department of Orthopaedic Surgery Royal Free Hospital Robert W. Simms, MD
Professor, Departments of Cell Biology and Honorary Senior Lecturer Professor of Medicine
Physiology and Biomedical Engineering Department of Medicine Boston University School of Medicine
Washington University School of Medicine in St. University College London Chief, Section of Rheumatology
Louis London, United Kingdom Boston Medical Center
St. Louis, Missouri Chapter 37, The heart in rheumatic disease Boston, Massachusetts
Chapter 2, The articular cartilage Chapter 155, Localized scleroderma and scleroderma-
Daniella Muallem Schwartz, MD like syndromes
Sabina Sandigursky, MD, MS Clinical Fellow
Department of Rheumatology Molecular Immunology and Inflammation Branch, Nora G. Singer, MD
New York University School of Medicine Translational Immunology Section Professor of Medicine and Pediatrics
New York, New York National Institute of Arthritis Musculoskeletal and Case Western Reserve University School of
Chapter 13, Inflammation and its chemical mediators Skin Disease Medicine
National Institutes of Health Director, Division of Rheumatology
Tore Saxne, MD, PhD Bethesda, Maryland MetroHealth Medical Center
Professor, Department of Clinical Sciences Chapter 11, Signal transduction in immune cells Cleveland, Ohio
Division of Rheumatology—Molecular Skeletal Chapter 105, Evaluation of the child with rheumatologic
Biology Lauren M. Shapiro, BA complaints
Lund University Faculty of Medicine Department of Radiology
Lund, Sweden Stanford School of Medicine Judith A. Smith, MD, PhD
Chapter 2, The articular cartilage Palo Alto, California Associate Professor of Pediatrics
Chapter 47, Functional magnetic resonance imaging Division of Allergy, Immunology, and
Prof. Dr. Hans-Georg Schaible Rheumatology
Professor of Physiology Leena Sharma, MD University of Wisconsin School of Medicine and
Friedrich Schiller University Jena Professor of Medicine Public Health
Director, Institute of Physiology Northwestern University Feinberg School of Madison, Wisconsin
Jena University Hospital Medicine Chapter 110, The juvenile-onset spondyloarthropathies
Jena, Germany Chicago, Illinois
Chapter 8, Scientific basis of pain Chapter 180, Local and systemic risk factors for
incidence and progression of osteoarthritis

Stacy E. Smith, MD Andre F. Steinert, MD Robert Terkeltaub, MD

Chief and Distinguished Barbara N. Weissman Associate Professor Chief, Rheumatology Section
Chair Department of Orthopedic Surgery VA Healthcare System
Division of Musculoskeletal Imaging and Julius-Maximilians-University San Diego, California;
Intervention Würzburg, Bavaria, Germany Professor of Medicine
Brigham and Women’s Hospital; Head of Department Division of Rheumatology, Allergy, and
Assistant Professor of Radiology Department of Orthopedic and Trauma Surgery Immunology
Harvard Medical School Agatharied Hospital University of California, San Diego
Boston, Massachusetts Haushab, Bavaria, Germany La Jolla, California
Chapter 46, Conventional radiography and computed Chapter 18, Principles of tissue engineering and cell- Chapter 192, Management of gout and hyperuricemia
tomography and gene-based therapy
Beje Thomas, MD
Josef S. Smolen, MD, FRCP George Stojan, MD Assistant Professor of Medicine
Professor of Internal Medicine Instructor in Medicine Division of Nephrology
Chair, Division of Rheumatology Division of Rheumatology University of Maryland School of Medicine
Medical University of Vienna; Harvard Medical School Baltimore, Maryland
Chairman, Center for Rheumatic Diseases Beth Israel Deaconess Medical Center Chapter 40, The kidneys in rheumatic disease
Hietzing Hospital Boston, Massachusetts
Vienna, Austria Chapter 160, Metabolic, drug-induced, and other Ranjeny Thomas, MBBS, MD
Chapter 101, Assessment of the patient with noninflammatory myopathies Professor and Chair
rheumatoid arthritis and the measurement of Department of Rheumatology
outcomes Vibeke Strand, MD, MACR, FACP Translational Research Institute
Adjunct Clinical Professor of Medicine The University of Queensland Diamantina Institute
Daniel H. Solomon, MD, MPH Division of Immunology/Rheumatology Brisbane, Australia
Professor of Medicine Stanford University School of Medicine Chapter 124, Animal models of spondyloarthritis
Harvard Medical School Palo Alto, California
Brigham and Women’s Hospital Chapter 142, Assessing disease activity and outcomes Jennifer E. Thorne, MD, PhD
Boston, Massachusetts in systemic lupus erythematosus Associate Professor of Ophthalmology
Chapter 43, Multimorbidity Johns Hopkins University School of Medicine
Rainer H. Straub, MD Associate Professor of Epidemiology
John Stack, MB, BCh, BAO Professor of Internal Medicine Center for Clinical Trials
Specialist Registrar in Rheumatology University Hospital Johns Hopkins University Bloomberg School of
Mater Misericordiae University Hospital Regensburg, Germany Public Health
Dublin, Ireland Chapter 15, Neuroendocrine immune crosstalk in Baltimore, Maryland
Chapter 194, Basic calcium phosphate crystal chronic inflammatory systemic diseases Chapter 36, The eyes in rheumatic disease
deposition disease
Elizabeth A. Streeten, MD Jonathan H. Tobias, MD, PhD
David Stanley, MB BS, BSc (Hons), FRCS Associate Professor of Medicine Professor of Rheumatology
Consultant Orthopaedic Surgeon Division of Endocrinology, Diabetes & Nutrition Musculoskeletal Research Unit
BMI Thornbury Hospital University of Maryland School of Medicine Bristol Medical School
Sheffield, United Kingdom Baltimore, Maryland Bristol, United Kingdom
Chapter 83, The elbow Chapter 197, Pathophysiology of osteoporosis Chapter 196, Clinical features of osteoporosis

Virginia D. Steen, MD Giulio Superti-Furga, PhD Adriana H. Tremoulet, MD, MAS

Professor of Medicine CeMM Research Center for Molecular Medicine of Associate Professor of Pediatrics
Division of Rheumatology, Clinical Immunology, the Austrian Academy of Sciences Associate Director, Kawasaki Disease Research
and Allergy Vienna, Austria Center
Georgetown University School of Medicine Chapter 22, Proteomics University of California San Diego School of
Washington, D.C. Medicine
Chapter 153, Management of systemic sclerosis Paul P. Tak, MD, PhD La Jolla, California
Professor Chapter 168, Kawasaki disease
Allen C. Steere, MD Department of Clinical Immunology &
Professor of Medicine Rheumatology Leendert A. Trouw, PhD
Harvard Medical School Academic Medical Center, University of Amsterdam Department of Rheumatology
Director, Translational Research in Rheumatology Amsterdam, The Netherlands; Leiden University Medical Center
Massachusetts General Hospital Chief Immunology Officer and Senior Vice Leiden, The Netherlands
Boston, Massachusetts President R&D Pipeline Chapter 99, Autoantibodies in rheumatoid arthritis
Chapter 116, Lyme and other tick-borne diseases Department of Research and Development
GlaxoSmithKline Rocky S. Tuan, PhD
Günter Steiner, PhD Cambridge, United Kingdom Distinguished Professor and Director
Head of Research Lab Chapter 45, Minimally invasive procedures Department of Orthopaedic Surgery
Division of Rheumatology Center for Cellular and Molecular Engineering
Medical University of Vienna Chen Tang, MD University of Pittsburgh School of Medicine
Coordinator Department of Internal Medicine Pittsburgh, Pennsylvania
Ludwig Boltzmann Cluster for Arthritis and Cedars-Sinai Medical Center Chapter 18, Principles of tissue engineering and cell-
Rehabilitation Los Angeles, California and gene-based therapy
Vienna, Austria Chapter 96, Imaging of rheumatoid arthritis
Chapter 99, Autoantibodies in rheumatoid arthritis Carl Turesson, MD, PhD
Peter C. Taylor, MA, PhD, FRCP Professor of Rheumatology
Professor of Rheumatology Skåne University Hospital
Kennedy Institute of Rheumatology Lund University
Nuffield Institute of Orthopedics, Rheumatology, Malmö, Sweden
and Musculoskeletal Sciences Chapter 95, Extraarticular features of rheumatoid
University of Oxford arthritis
Botnar Research Centre
Headington, Oxford
Chapter 72, Tumor necrosis factor inhibitors
Hisanori Umehara, MD, PhD Sebastien Viatte, MD, PhD Robert J. Ward, MD, CCD
Director of Rheumatology and Immunology Lecturer in Genetics Assistant Professor of Radiology
Nagahama City Hospital Arthritis Research UK Centre for Genetics and Chief, Division of Musculoskeletal Imaging
Shiga, Japan Genomics Tufts Medical Center
Chapter 178, IgG4-related disease Centre for Musculoskeletal Research Boston, Massachusetts
Associate Member Chapter 206, Osteonecrosis
Ana M. Valdes, MA, PhD Manchester Collaborative Centre for Inflammation
Associate Professor and Reader in Musculoskeletal Research (MCCIR) Laura Watts, MA (Oxon), BMBCh, MRCP
Genetics Faculty of Biology, Medicine and Health Rheumatology Registrar
Academic Rheumatology The University of Manchester Nuffield Department of Orthopaedics,
University of Nottingham Manchester, United Kingdom Rheumatology, and Musculoskeletal Science
Nottingham, United Kingdom Chapter 97, Genetics of rheumatoid arthritis University of Oxford
Chapter 184, Genetics of osteoarthritis Oxford, United Kingdom
Tonia Vincent, MBBS, PhD, FRCP Chapter 215, Heritable connective tissue disorders
Désirée van der Heijde, MD, PhD Professor of Musculoskeletal Biology
Professor of Rheumatology Nuffield Department of Orthopaedics, Richard A. Watts, MA, DM, FRCP
Leiden University Medical Center Rheumatology, and Musculoskeletal Sciences Consultant Rheumatologist
Leiden, The Netherlands Kennedy Institute of Rheumatology; Ipswich Hospital NHS Trust
Chapter 50, Use of imaging as an outcome measure in Honorary Consultant Rheumatologist Ipswich, United Kingdom;
clinical trials Oxford University Hospitals Trust Clinical Senior Lecturer
Oxford, United Kingdom Norwich Medical School
Ronald F. van Vollenhoven, MD, PhD Chapter 182, Animal models of osteoarthritis University of East Anglia
Professor and Chief, Clinical Therapy Research Norwich, United Kingdom
Inflammatory Diseases (ClinTRID) Edward M. Vital, MB ChB, MRCP, PhD Chapter 161, Classification and epidemiology of
The Karolinska Institute Associate Professor and Honorary Consultant vasculitis
Senior physician Leeds Institute of Rheumatic and Musculoskeletal
Chief, Clinical Trials Unit Medicine Lucy R. Wedderburn, MD, MA, PhD, FRCP
Division of Rheumatology University of Leeds Professor in Paediatric Rheumatology
The Karolinska University Hospital NIHR Leeds Musculoskeletal Biomedical Research Institute of Child Health
Stockholm, Sweden Centre University College London Medical School
Chapter 68: Tyrosine kinase inhibition Leeds Teaching Hospitals NHS Trust Consultant in Paediatric Rheumatology
Leeds, United Kingdom Great Ormond Street Hospital NHS Trust
Joke Vanderstukken, MD Chapter 75, Inhibitors of B cells London, United Kingdom
Department of Immunology Allergology and Chapter 111, Connective tissue diseases in children
Rheumatology Isabell S. von Loga, AB, MBBS
University of Antwerp, Faculty of Medicine and Nuffield Department of Orthopaedics, Michael E. Weinblatt, MD
Health Sciences Rheumatology, and Musculoskeletal Sciences John R. and Eileen K. Riedman Professor of
Antwerp, Belgium; Kennedy Institute of Rheumatology Medicine
Professor of Rheumatology University of Oxford Harvard Medical School
Leiden University Medical Center Oxford, United Kingdom Division of Rheumatology, Immunology and Allergy
Leiden, The Netherlands Chapter 182, Animal models of osteoarthritis Brigham and Women’s Hospital
Chapter 71, Interleukin-6 inhibitors Boston, Massachusetts
Daniel J. Wallace, MD Chapter 66, Methotrexate
John Varga, MD Professor of Medicine
John and Nancy Hughes Professor Division of Rheumatology Matthew R. Weir, Sr., MD
Department of Medicine Cedars-Sinai Medical Center; Professor of Medicine
Northwestern University Feinberg School of Clinical Professor of Medicine Director, Division of Nephrology
Medicine Division of Rheumatology University of Maryland School of Medicine
Chicago, Illinois David Geffen School of Medicine at UCLA Baltimore, Maryland
Chapter 151, Etiology and pathogenesis of systemic Los Angeles, California Chapter 40, The kidneys in rheumatic disease
sclerosis Chapter 135, Clinical features of systemic lupus
Michael H. Weisman, MD
Dimitrios Vassilopoulos, MD Cedars-Sinai Chair in Rheumatology
Professor of Medicine (Rheumatology) Nicola Walsh, PhD, MSc, MCSP Director, Division of Rheumatology
Joint Rheumatology Program Professor of Knowledge Mobilisation and Professor of Medicine
National and Kapodistrian University of Athens Musculoskeletal Health Cedars-Sinai Medical Center
School of Medicine Centre for Health and Clinical Research Distinguished Professor of Medicine
2nd Department of Medicine and Laboratory University of the West of England David Geffen School of Medicine at University of
Hippokration General Hospital Bristol, United Kingdom California, Los Angeles
Athens, Greece Chapter 56, Multi-disciplinary approaches to managing Los Angeles, California
Chapter 115, Viral infections chronic pain in arthritis Chapter 120, Inflammatory back pain
Chapter 135, Clinical features of systemic lupus
Douglas J. Veale, MD, FRCPI, FRCP(Lon) Sara L. Warber, MD erythematosus
Director of Translational Research Professor Emerita of Family Medicine
The Centre for Arthritis and Rheumatic Disease; Former Director, Integrative Medicine Pamela F. Weiss, MD, MSCE
Professor of Medicine University of Michigan Medical School Associate Professor of Pediatrics
University College Dublin; Ann Arbor, Michigan Perelman School of Medicine at the University of
Fellow, Conway Institute of Biomolecular and Chapter 57, Placebos, caring, and healing in Pennsylvania
Biomedical Medicine rheumatology Attending Physician
Dublin, Ireland Division of Rheumatology
Chapter 131, Management of psoriatic arthritis Gary Warburton, DDS, MD, FDSRCS, FACS Children’s Hospital of Philadelphia
Associate Professor of Oral and Maxillofacial Philadelphia, Pennsylvania
Surgery Chapter 110, The juvenile-onset spondyloarthropathies
University of Maryland School of Medicine,
Baltimore, Maryland
Chapter 88, The temporomandibular joint

Claire Y.J. Wenham, BMBS, MRCP, MD Kevin L. Winthrop, MD, MPH Jane Worthington, PhD
Section of Musculoskeletal Disease Associate Professor of Infectious Diseases Professor of Chronic Disease Genetics
Leeds Institute of Molecular Medicine Department of Public Health and Preventive Institute of Inflammation and Repair
University of Leeds Medicine The University of Manchester
Chapel Allerton Hospital Oregon Health & Science University Manchester, United Kingdom
Leeds, United Kingdom Portland, Oregon Chapter 28, Principles of genetic epidemiology
Chapter 185, Imaging of osteoarthritis Chapter 78, Infections and biologic therapy in
rheumatoid arthritis Katherine D. Wysham, MD
Sterling G. West, MD, MACP Clinical Fellow in Rheumatology
Professor of Medicine Claudia M. Witt, MD, MBA Department of Medicine
Division of Rheumatology Professor and Chair University of California San Francisco
University of Colorado School of Medicine Institute for Complementary and Integrative San Francisco, California
Aurora, Colorado Medicine Chapter 121, Clinical features of axial spondyloarthritis
Chapter 175, Sarcoidosis University Hospital Zurich and University Zurich
Zurich, Switzerland; Huji Xu, MD, PhD
Harriet Branford White, MB ChB, DPhil, FRCS (Tr Professor of Medicine Professor and Chairman
&Orth) Center for Integrative Medicine Department of Rheumatology and Immunology
Trauma and Orthopaedic Registrar University of Maryland School of Medicine Shanghai Changzheng Hospital
North West Thames Rotation Baltimore, Maryland Second Military Medical University
St Mary’s Hospital Chapter 58, Complementary and alternative medicine Shanghai, China
Imperial NHS Trust Chapter 123, Genetics of axial spondyloarthritis
London, United Kingdom Gerhard Witzmann, MD
Chapter 217, Bone tumors David A. Young, PhD
Senior Physician Senior Lecturer
Kenneth E. White, PhD Second Department of Medicine Musculoskeletal Research Group
Professor of Medical and Molecular Genetics Hietzing Hospital Newcastle University
Indiana University School of Medicine Vienna, Austria Newcastle upon Tyne, United Kingdom
Appendix: Classification and Diagnostic Criteria
Indianapolis, Indiana Chapter 17, Tissue destruction and repair
Chapter 3, Bone structure and function
John B. Wong, MD Stephen P. Young, BSc(Hons), PhD
David J. Wilkinson, PhD, MRes Professor of Medicine Reader in Experimental Rheumatology
Research Associate Tufts University School of Medicine Institute of Inflammation and Ageing
Musculoskeletal Research Group Chief, Division of Clinical Decision Making University of Birmingham
Institute of Cellular Medicine Tufts Medical Center Birmingham, United Kingdom
Newcastle University Boston, Massachusetts Chapter 22, Metabolomics
Chapter 27, Principles of health economics
Newcastle upon Tyne, United Kingdom
Chapter 17, Tissue destruction and repair Md Yuzaiful Md Yusof, MBChB, MRCP
Anthony D. Woolf, BSc, MBBS, FRCP NIHR Doctoral Research Fellow
David E. Williams, BEng Honorary Professor of Rheumatology Leeds Institute of Rheumatic and Musculoskeletal
Researcher University of Exeter Medical School Medicine
School of Engineering Exeter, United Kingdom University of Leeds
Cardiff University Bone and Joint Research Group NIHR Leeds Musculoskeletal Biomedical Research
Cardiff, Wales The Knowledge Spa Centre
Chapter 6, Biomechanics of peripheral joints
Royal Cornwall Hospital Leeds Teaching Hospitals NHS Trust
Truro, United Kingdom Leeds, United Kingdom
Richard Williams, BSc, MSc, PhD Chapter 32, History and physical examination
Chapter 75, Inhibitors of B cells
Associate Professor of Rheumatology
Nuffield Department of Orthopaedics, Paul B. Wordsworth, MA, MB, BS, FRCP Haoyang Zhuang, PhD
Rheumatology and Musculoskeletal Sciences Professor of Clinical Rheumatology Division of Rheumatology and Clinical Immunology
University of Oxford Nuffield Department of Orthopaedics, University of Florida College of Medicine
Oxford, United Kingdom Rheumatology, and Musculoskeletal Sciences Gainesville, Florida
Chapter 98, Animal models of rheumatoid arthritis
University of Oxford Chapter 139, Autoantibodies in systemic lupus
Oxford, United Kingdom; erythematosus
Hannah Wilson, BMBCh, MA Oxon, MRCS Honorary Consultant Rheumatologist
Nuffield Department of Orthopaedics, Department of Rheumatology Angela Zink, PhD
Rheumatology and Musculoskeletal Science Nuffield Orthopaedic Centre Head of Epidemiology Unit
University of Oxford Headington, United Kingdom German Rheumatism Research Center
Oxford, United Kingdom Chapter 215, Heritable connective tissue disorders Professor of Musculoskeletal Epidemiology
Chapter 86, The knee Charité University Medicine Berlin
Berlin, Germany
Chapter 31, European biologics registers

Rheumatology, Seventh Edition, builds on the success of the previous editions. ensure that the content and format of the book remain consistent and meet
The most notable change is the addition of Professor Ellen Gravallese as the highest possible standards. Each chapter has been updated to incorporate
one of the Editors. Professor Gravallese is internationally recognized for her a broad range of new information. Seventeen completely new chapters cover
basic and translational work in the study of the immune-mediated inflammatory basic biomedical and translational science, disease and outcome assessment,
mechanisms of bone and cartilage destruction in rheumatoid arthritis. Her including new imaging modalities and early emerging disease, clinical
addition to the Editorial team has refocused our efforts to provide a strong therapeutics, and patient management, including rehabilitation. The text
basic and translational science component to the new edition. has been streamlined, ensuring that each chapter contains the most critical
Designed to meet the needs of the practicing clinician, this medical reference and current information in the field, while supplemental materials (including
book provides extensive coverage of rheumatic and musculoskeletal diseases extra tables, figures, and bonus text) are conveniently located online. The
from basic scientific principles to practical points of clinical management index has also been improved, making it easier for the reader to find topics
in a lucid and logical manner. As stated by Professor Jan Dequeker in his of interest.
review of the fourth edition, “Rheumatology is the most comprehensive, The production of this edition of Rheumatology has been a greatly enjoyable
authoritative rheumatology text designed to meet the complete needs of all team effort. We would like to thank the authors who have contributed to
practicing and academic rheumatologists as well as arthritis-related health this and previous editions of the book, as well as the excellent team at
care professionals and scientists interested in disorders of the musculoskeletal Elsevier, including Jennifer Shreiner, Michael Houston, Nancy Duffy, Ted
system. The edition is firmly grounded on modern medical science, integrating Rodgers, Bridget Hoette, and Nichole Beard.
the relevant basic biology with current clinical practice, easily accessible, We look forward to bringing you the eighth edition in another 4 years.
user-friendly, and a beautifully illustrated color publication.” Dr. Harry Brown, Marc C. Hochberg
in his review of the sixth edition, noted that “[m]y lasting impression of Ellen M. Gravallese
this book was the very same as the first time I opened these two lavish Alan J. Silman
volumes and that was – wow. I would suspect this is the nearest you would Josef S. Smolen
get to an encyclopaedia of rheumatology.” Michael E. Weinblatt
For this new edition, every chapter has been either substantially revised Michael H. Weisman
or, in many cases, entirely rewritten, following a rigorous editorial policy to


We would like to acknowledge the tremendous work of the contributors to who continue to provide stimulating challenges to us in our clinical
this edition of Rheumatology, without whom this book would not have been practices.
possible. In addition, we would like to recognize our mentors: Drs. Eva Marc C. Hochberg
Alberman, Ronald J. Anderson, and Laurie Glimcher, and the late Harry Ellen M. Gravallese
Currey, Georg Geyer, Lawrence E. Shulman, Carl Steffen, Alfred D. Steinberg, Alan J. Silman
Mary Betty Stevens, and the late Nathan J. Zvaifler. Josef S. Smolen
We would also like to thank our in-office editorial support (Aida Medina, Michael E. Weinblatt
Robin Nichols, Jacqui Oliver, Marion Skobek) for all of their hard work and Michael H. Weisman
diligence. Last, but certainly not least, we want to acknowledge our patients,



A. Anatomy and Physiology

The synovium
Andrew Filer • Christopher D. Buckley 1  

Key Points tissue. Diseased synovial tissue may lose any recognizable lining structure
and may only be definable by its relation to the joint. These variations
■ The synovium is a mesenchymal membrane that lines diarthrodial joints, tendon
probably reflect the interplay of several factors in synovial embryogenesis
sheaths, and bursae.
and histogenesis.
■ Specialized functions of synovium include nonadherence, control of synovial fluid
production and composition, and providing chondrocyte nutrition.
■ The normal synovium produces very low levels of proinflammatory cytokines and EMBRYOLOGY
some antiinflammatory, proresolving cytokines and eicosanoids. In addition, the low In the early embryonic limb bud, a central core or blastema that will ultimately
levels of expression of RANKL (receptor activator of nuclear factor-κB ligand) with form the skeleton appears. Within this core, foci of cartilage appear, each
high levels of expression of OPG (osteoprotegerin) result in a low RANKL-to-OPG destined to become bone. Blastemal cells around cartilage foci form a peri-
ratio. This homeostatic balance is likely to be important in preventing chondrial envelope showing strong CD44 expression. The area where this
osteoclastogenesis in the normal, noninflamed synovium. envelope lies between cartilage elements is known as the interzone, from
■ Mesenchymal markers expressed by synovial fibroblasts, such as cadherin-11, which the synovium forms. The perichondrium, forming a sleeve around
endosialin (CD248), and podoplanin (gp38), may be critical for the development of each cartilage element, subsequently invades the cartilage to form bone
the synovial lining by facilitating cellular organization, compaction, and matrix marrow. Thus, synovial and bone marrow stromal cells come from the same
development. In pathologic settings, they appear to promote cartilage invasion and embryonic stock and this is reflected in their transcriptional functional
bone destruction. abilities.2,3
■ Synovial fibroblasts carry positional and topographic memory that may provide the Shortly before the joint cavity forms, CD55 expression appears on cells
molecular basis for site-specific differences in the pattern of joint involvement in along the joint line4 followed later by vascular cell adhesion molecule 1
different rheumatologic diseases. (VCAM-1) expression. After cavity formation, the intimal layer also takes
on a higher level of expression of CD44 and β1 integrins compared with
subintima. Expression of these three markers (CD44, CD55, VCAM-1 [CD106])
confirms the strong similarity between synovial and bone marrow stromal
The mechanism of cavity formation is not fully understood; a working
The study of synovial tissue is of major importance in understanding the hypothesis implicates interactions between interzone cells bearing CD44 (a
pathogenesis of inflammatory arthritis, including rheumatoid arthritis (RA) hyaluronan receptor) and hyaluronan itself.5,6 Shortly before cavity formation,
and seronegative spondyloarthritis (SpA). Despite this, our knowledge of the cells of the potential joint line show high uridine diphosphoglucose
the immunohistochemical architecture of the synovial membrane, particularly dehydrogenase (UDPGD) activity, which suggests increased hyaluronan
in normal subjects, is surprisingly limited, mainly because of the lack of synthesis. At the time of cavity formation, high levels of hyaluronan appear
good tissue and cell-specific markers and the difficulty in obtaining synovial along the joint line, saturate CD44, and induce disaggregation; at low
tissue in the early as opposed to later stages of disease. concentrations hyaluronan cross-links CD44 molecules on adjacent cells,
Synovium is the soft tissue lining the spaces of diarthrodial joints, tendon inducing cell aggregation.
sheaths, and bursae. The term includes both the continuous surface layer Cavity formation might be expected to require lysis of matrix fibers.
of cells (intima) and the underlying tissue (subintima). Whereas the intima However, in human joints, cavity formation is not associated with high local
is composed of specialized tissue resident macrophages and fibroblasts, the levels of matrix metalloproteinases (MMPs) at the joint line. In fact, matrix
subintima contains blood and lymphatic vessels, a cellular content of both fibers appear to run only parallel to the joint line before cavity formation;
resident fibroblasts and infiltrating cells in a collagenous extracellular matrix apoptotic cells found in the interzone at this time are not localized to the
(ECM). Between the intimal surfaces is a small amount of fluid, usually rich joint line and are unlikely to contribute to cavity formation. It appears,
in hyaluronan (hyaluronic acid). Together, this structure provides a nonadher- therefore, that development of the joint cavity arises more from differential
ent surface between tissue elements. Unlike serosal surfaces, which also have tissue expansion than through loss of solid elements.
nonadherent properties, synovium is derived from ectoderm and does not
contain a basal lamina.
In normal subjects, the intimal layer is 20 to 40 µm thick in cross-section,
and the areolar subintima can be up to 5 mm in thickness. At many sites, The microscopic anatomy of synovial tissue was first fully described by Key,7
there is no discrete membrane, especially where subintima consists of fat who divided synovium into three main types on the basis of subintimal
pad or fibrous tissue. structure: fibrous, areolar, and adipose (Fig. 1.1, a to c). He also noted that
Synovium is often atypical. Intimal cells may be absent. Superficial bursae subintima may be periosteum, perimysium, or even hyaline or fibrocartilage.
contain little or no hyaluronan-rich fluid.1 Ganglia are herniated sacs containing Areolar synovium is the most specialized form (see Fig. 1.1, a). It is often
hyaluronan-rich fluid but do not occur at sites of mechanical shearing and crimped into folds, which may disappear when stretched. Less often it carries
do not have a typical intima and so may not be considered really to be synovial projections or villi. A more or less continuous layer of cells lies two or three

2 SECTION 1  Scientific Basis of Rheumatic Disease

a b c

FIG. 1.1  (a) Areolar form of synovium (hematoxylin and eosin [H&E]). (b) Adipose form of synovium (H&E). (c) Fibrous form of synovium (H&E). (Magnification ×200.)

Protein disulphide
CD248 isomerase FAP CD68

FIG. 1.2  Stromal markers

100 µm differentially expressed in the
lining and sublining. FAP,
Fibroblast activation protein-α.

CD90 Endo180 Nuclei

Adipose synovium occurs as fat pads and within villi (see Fig. 1.1, b). It
has a complete intimal cell layer and a superficial net of capillaries. The
intima may lie directly on adipocytes, but there is usually a band of collagen-
rich substratum, but the deeper tissue is fat. Villi usually have a central
arteriole and venule but can be avascular. The amount of fat in villi varies
and probably decreases with age, with an increase in fibrous tissue.
Fibrous synovium is more difficult to define, consisting of fibrous tissue
such as ligament or tendon on which lies an intermittent layer of cells (see
Fig. 1.1, c). Fibrous synovium may be indistinguishable from fibrocartilage,
especially in the annular pads found in finger and toe joints.

Two types of intimal cells have been defined by electron microscopy, one
FIG. 1.3  Synovium in rheumatoid arthritis (×400) showing a thickened intimal layer consistent with a macrophage (type A) and the other with a fibroblast (type
containing mainly CD68+ macrophages (red) on the surface and weakly CD55+ B).15 It is now generally accepted—from immunohistochemical studies and
fibroblastic cells beneath (blue). other lines of evidence—that intimal macrophages are true macrophages.
However, whether they are derived from circulating precursors from the
bone marrow or are derived from fetal macrophages remains unclear.16 Intimal
deep on the tissue surface.8,9 Immediately beneath these cells are capillaries. fibroblasts on the other hand are nonhematopoietic cells and are tissue
Further into the tissue, there is a plexus of small arterioles and venules,10,11 derived.14,15,17,18 In normal healthy synovium, synovial fibroblasts are the
often associated with mast cells. Lymphatic vessels can be found in all types dominant cell population.19
of normal synovial tissue, although they are infrequent in the fibrous type Immunohistochemical and cytochemical methods have superseded electron
of normal synovium.12 In normal synovium, most lymphatic vessels are microscopy as tools for cell identification.20 Intimal macrophages can be
found in the deep subintima and fibrous layers, but in synovium from patients distinguished by their nonspecific esterase (NSE) activity and expression of
with inflammatory arthritis, lymphatic vessels are widespread and numerous. surface markers such as CD68 and CD163. Often they are CD45 positive as
Nerve fibers are present, chiefly in association with blood vessels.13 Three well as expressing podoplanin/gp38 (Fig. 1.2). Intimal fibroblasts show intense
different layers of tissue matrix may be distinguished. The intima is associated activity of the enzyme UDPGD and prominent expression of VCAM-1 and
with a fine fibrillar matrix with few type I collagen fibers.14 Beneath this is CD55 (complement decay-accelerating factor [DAF]). In most disease states
a layer relatively rich in type I collagen, which forms a physical membrane. such as RA, intimal cells increase in size and number (Fig. 1.3). This is not
Deeper is a loose layer that allows the membrane to move freely. Beyond just due to hyperplasia but a complex change in cell populations, in terms
the loose layer lies ligament, tendon, or periosteum. of both origin and function, which may be dominated by macrophage influx.19
CHAPTER 1  The synovium 3

diversity of stromal cells, and in particular fibroblast phenotype and function

SYNOVIAL MACROPHAGES and their roles beyond those of space filling and ECM homeostasis, has been
Macrophages are present in both the intima and subintima. Intimal mac- underplayed in the synovium.26,27 We now know that these cells vary phe-
rophages carry typical macrophage lineage markers. They show prominent notypically at different anatomical sites and contribute significantly to the
NSE activity and are strongly positive for CD163 and CD68 but less so for identity of individual tissues, providing the so-called “stromal postcode.”28
CD14 (Fig. 1.4). Macrophages also express the immunoglobulin receptor Furthermore, it is known that, rather than acting as a bystander to the body’s
FcγRIIIa. Strong FcγRIIIa expression is restricted to a subset of macrophages protective mechanisms and to disease processes, the fibroblast is capable of
that correspond closely to sites of macrophage activation in rheumatoid actively participating and indeed orchestrating inflammation and immunity.29
disease: synovial, alveolar, serosal, scleral, and salivary gland; lymphoid The fibroblast communicates with resident and infiltrating cells via cytokines
tissue and bone marrow; and Kupffer cells.21 Subintimal macrophages are and cell contact dependent mechanisms, playing a central role in the
FcγRIIIa dull or negative. Macrophages also express Z39Ig, a recently described pathogenesis of synovial pathology.
inducible cell surface receptor linked to the classic complement pathway; The synovial intima contains cells that are adapted to hyaluronan produc-
Z39Ig expression can occur during macrophage differentiation and induce tion. In normal synovium, CD68-negative intimal fibroblasts express high
activation of the transcription factor nuclear factor-κB (NF-κB) and production enzymatic activity for UDPGD.30 UDPGD converts UDP glucose into UDP
of matrix-degrading MMP-9.22 glucuronate, one of the two substrates required by hyaluronan synthase for
Macrophages make up a minority of cells in normal intima (Figs 1.4 and hyaluronan polymer assembly. Unlike the activity of many other enzymes,
1.5). In disease, the proportion of macrophages increases (see Figs 1.2 and UDPGD activity in intimal fibroblasts is reduced, rather than enhanced, in
1.3). Distribution varies, but a common pattern is a superficial layer of diseased tissue. Synovial intimal fibroblasts express CD55 (see Fig. 1.5), a
macrophages with an intimal phenotype; beneath this a layer of intimal feature distinguishing them from intimal macrophages.31,32
fibroblasts; and further beneath and beyond the limits of the intima, a zone Cells disaggregated from inflamed synovium and grown in tissue culture
of NSE-weak, strongly CD14+ and FcγRI+ macrophages, associated with display fibroblast characteristics and ramifying processes with production
venules. The deep, strongly CD14+ cells may be recently recruited cells and of high levels of MMPs.33 It is not known whether they derive from intimal
the superficial cells tissue resident.23 or subintimal cells. In tissue sections, immunoreactivity for collagenase and
In addition to true macrophages, there may be a small number of antigen- gelatinase is patchy and not necessarily confined to the intima.
presenting interdigitating dendritic cells in normal synovial intima; these Synovial intimal fibroblasts also show prominent expression of several
are more frequent in disease with greater overlap of markers, which confounds adhesion molecules,9,34,35 including VCAM-1, intercellular adhesion molecule
interpretation.24,25 Cells with features of osteoclasts such as expression of 1, CD44, and β1 and β3 integrins. Expression of VCAM-1 (Fig. 1.6) is par-
tartrate-resistant acid phosphatase and the vitronectin receptor also often ticularly unusual, being absent from most other normal fibroblast populations,
appear in inflamed synovium. However, fully typical osteoclasts with calcitonin but CD44 and β1 integrins are present at lower levels. The role of VCAM-1
receptors appear to be restricted to pigmented villonodular synovitis and with respect to intimal fibroblasts is puzzling, reflecting its embryologic
giant cell tumors of tendon sheath. similarities to bone marrow fibroblasts, which are also VCAM-1 positive.
Expression of VCAM-1 may modulate cell trafficking because its ligand, α4β1
integrin, is present on mononuclear leukocytes but not granulocytes. Intimal
SYNOVIAL FIBROBLASTS IN THE STROMA fibroblasts may allow transmigration of polymorphs but not mononuclear
The anatomical term stroma was originally derived from the Greek word cells into synovial fluid, potentially trapping inflammatory cell infiltrates
describing a platform on which to lie and is used to describe the supporting within the synovial membrane in disease states such as RA.
substance of a tissue. Its principal role is to maintain the microenvironment Recently, the presence of both clusterin (a glycoprotein involved in recycling
required by the parenchyma, the important functional elements of each body and apoptosis) and (gp38) podoplanin (a membrane glycoprotein with diverse
system. The stroma includes the cells of mesenchymal origin; the nerves, functions) has been reported in normal synovial fibroblasts; interestingly,
the vessels, and the epithelia that reside in a tissue in steady state; and the podoplanin (which in the setting of neoplasia is associated with poor prognosis
extracellular matrices and fluids that these cells produce. Traditionally, the and metastatic disease) has been shown to be highly expressed in RA synovial
fibroblasts with their attendant migratory and invasive potential36,37
(Fig. 1.7).
Under inflammatory conditions, fibroblasts can act as organ-specific sentinel
cells, where they play a role in the switch from acute resolving to chronic
persisting. In addition to contributing to the recruitment and emigration of
inflammatory cells to and out of the joint, they modulate the survival and
retention of infiltrating leukocytes. Interestingly, new data raise the possibility
of epigenetically programmed aggressive fibroblasts “spreading” arthritis
from inflamed to uninflamed joints in the early stages of arthritis but at the
same time offering the possibility of specifically targeting stromal subpopula-
tions of choice.38
The expression of two other surface molecules by synovial fibroblasts is
noteworthy. Complement receptor 2 (CR2, CD21) is not expressed by normal
intimal fibroblasts but can be induced on synovial fibroblasts in culture, in
contrast to other fibroblast populations.39 DAF, VCAM-1, and CR2 are all
involved in B-lymphocyte survival, as is a bone marrow stromal cell marker,
FIG. 1.4  Synovial macrophages. Normal synovium (×200) stained for CD68+ mac-
BST-1, reported to be expressed on fibroblasts in rheumatoid, but not normal,
rophages (red).
synovial intima.40 Other molecules associated with bone marrow stromal

FIG. 1.5  Normal synovium (×200) stained for CD55+ fibroblasts, which are the
predominant cell in the normal synovium intimal layer (contrast with Fig. 1.3). FIG. 1.6  Normal synovium (×200) stained for vascular cell adhesion molecule 1.
4 SECTION 1  Scientific Basis of Rheumatic Disease
Normal synovium Rheumatoid synovium

FIG. 1.7  Stromal markers differentially expressed in

rheumatoid arthritis. (a) Normal synovium. (b) Rheumatoid

a b

CD45 CD31 GP38 CD248

FIG. 1.9  Normal synovium (×200) stained with factor VIII to demonstrate the vascular

FIG. 1.8  Normal synovium stained for hyaluronan using a histochemical probe derived
from proteoglycan core protein hyaluronan-binding region. Staining is most intense
surrounding the lining cells and decreases further into the tissue. (Magnification

cells such as the chemokine stromal cell–derived factor-1 (CXCL12) and

bone morphogenetic proteins and their receptors,41-43 are expressed by synovial
fibroblasts under various conditions. Moreover, lubricin, otherwise known
as superficial zone protein, a glycoprotein found in synovium and the superficial
zone of articular cartilage,44 derives from the same gene as megakaryocyte-
stimulating factor. A defect of this gene leads to CACP (camptodactyly
arthropathy coxa vara pericarditis) syndrome. As indicated earlier, these
patterns of gene expression may reflect a common embryologic origin for
synovial and bone marrow stromal cells.
Self-renewing mesenchymal stem cells that compare favorably with bone
marrow–derived mesenchymal stem cells in terms of their ability to differentiate FIG. 1.10  Normal synovium (×200) stained with lymphatic vessel endothelial hyaluronan
into bone, cartilage, and adipose tissue have been isolated from the normal receptor-1 antibody to demonstrate the lymphatic network.
synovium; it is unclear which component of the synovial membrane is home
to these cells,45,46 but expression of the mesenchymal stromal cell marker
CD248 in the sublining layer (see Fig. 1.7) suggests that they may derive VASCULAR NET
from this anatomical compartment.47
A rich microvascular net lies beneath the synovial surface.10,11 Capillaries
(prominent in children and decreasing with age) occur just below or within
INTIMAL MATRIX the intima (Fig. 1.9). Some capillaries are fenestrated, and fenestrae tend to
Intimal matrix has an amorphous or fine fibrillar ultrastructure. It is poor face the tissue surface50; 50 to 100 µm beneath the surface, small venules
in type I collagen but contains minor collagens III, IV, V, and VI48,49 as well are prominent. About 200 µm beneath the surface, larger venules, together
as laminin, fibronectin, and chondroitin-6-sulfate–rich proteoglycan, which, with arterioles and lymphatics (Fig. 1.10),12 form an anastomosing quadrilateral
with collagen IV, are components of basement membrane; however, the array. Vessels with lymphatic staining characteristics are prominent in RA
basement membrane is conspicuous by its absence beneath the intimal layer. synovium. It has been proposed that failure of lymphatic drainage of synovial
The looser structure of intimal matrix may be explained by the absence of fluid is a cause of villous proliferation in RA synovial tissue. If this is correct,
entactin, which links other components in basement membrane together. it is likely to be due to overloading of existing lymphatic channels with
Intimal microfibrils are of two types: fibrillin-1 microfibrils form a basketwork hyaluronan-rich extracellular fluid and leukocytes rather than a lack of
around cells, and collagen VI microfibrils form a uniform mesh. lymphatic channels.12
Intimal matrix contains large amounts of hyaluronan (Fig. 1.8), which Apart from the fenestration of superficial capillary endothelial cells, there
tails off 20 to 50 µm deep; this possibly indicates diffusion from the surface is little evidence of specialization in synovial endothelium. Endothelial cells
toward clearing lymphatics. enlarge in inflamed tissue, and microvascular proliferation can occur, but
CHAPTER 1  The synovium 5

these events are common to inflammation at many sites. Tissue-specific Functions of the tissue relating to the synovial cavity may be considered
adhesion molecules, or addressins, have been sought in the synovium, but to be the following:
nothing conclusive has been found unlike the case in the skin and gut, both ■ Maintenance of an intact nonadherent tissue surface
epithelial organs. ■ Lubrication of cartilage
■ Control of synovial fluid volume and composition
■ Nutrition of chondrocytes within joints
Evidence to date indicates that both intimal and subintimal macrophages
derive from bone marrow via circulating monocytes, many of which probably
arrive through subintimal venules and migrate to the intima. Whether tissue Synovial surfaces must be nonadherent to allow continued articular movement.
resident macrophages also contribute remains unclear, but recent studies Animal models suggest that production of hyaluronan by intimal fibroblasts
suggest that nonclassical monocytes or macrophages can contribute to the may be important in inhibiting adhesion.53 Plasminogen activator and DAF
persistence or resolution of arthritis.23 from intimal fibroblasts may also inhibit fibrin formation and scarring. To
Intimal fibroblasts are thought to arise by division within synovium. They retain synovial fluid, the intimal matrix consists of a fibrous mat of a particular
might be a discrete self-replicating population, distinct from subintimal porosity that allows free exchange of crystalloids and proteins but inhibits
fibroblasts, but several pieces of evidence argue against this. Rates of cell rapid transit of the viscous hyaluronan solution that is an important component
division within the intima are very low, even in disease. After arthroplasty of the fluid. The vasculature is likely to be important in both intimal cell
or synovectomy, intimal cells—likely replaced from the subintima rather nutrition and recruitment of new cells. New macrophages are derived from
than arising from intimal rests—reappear and express CD55, UDPGD, and blood monocytes that are thought to enter the tissue through venules, and
VCAM-1. Disaggregated and cultured synovial fibroblasts lose VCAM-1 and perivascular fibroblasts may provide the main pool of intimal fibroblast
CD55 expression, but the majority, apparently including cells of subintimal precursors.
origin, readily express these markers after cytokine stimulation, in contrast
to fibroblasts of dermal or subcutaneous origin. These findings suggest that
synovial fibroblasts, in both the intima and subintima, belong to a specialized
population with a propensity to express VCAM-1 and CD55 and are more The ability of synovial fluid to lubricate cartilage surfaces depends on the
similar to bone marrow than skin fibroblasts.2,3 presence of glycoproteins, especially a glycoprotein known as both lubricin
Two studies8,9 have demonstrated the range of cells that can be found in and superficial zone protein because of its localization to the surface of synovium
the synovial subintima. CD3+ T cells, including CD4+, CD8+, and memory and cartilage.44
T cells, can be found within the normal synovial tissue; although they are Whatever the precise forces acting on fluid volume, the presence of
likely to be simply trafficking through the normal synovium, their role, if hyaluronan is likely to be the main factor responsible for retaining a constant
any, in the homeostasis of synovial tissue is unknown. It is also possible to volume of fluid during exercise.54 This fluid is probably important as a
detect B cells, plasma cells, and granzyme B+ cells in normal synovium, cushion for synovial tissue and as a reservoir of lubricant for cartilage. It is
although they are present in small numbers. likely that mechanical stimulation of intimal fibroblasts dictates the rate of
Although production of inflammatory cytokines, including interleukin-1 synthesis and exportation of hyaluronan into the synovial fluid compartment.
(IL-1), IL-6, and tumor necrosis factor-α (TNF-α),9 can be detected in normal Thus, when synovial fluid volume is high, reduced mechanical stresses on
synovial tissue, expression levels are far lower than those seen in inflamed intimal fibroblasts result in a reduced rate of hyaluronan production and
synovial tissue such as in RA. The amount of antiinflammatory cytokine vice versa.
production, at least in the case of IL-1 receptor antagonist (the naturally Two distinct mechanisms create joint effusions. When synovium is
occurring inhibitor of IL-1), IL10, and other proresolving factors is far greater mechanically irritated by worn bone and cartilage, the composition of the
than the amount of inflammatory cytokine seen (Fig. 1.11).51 This would fluid remains reasonably normal. Excessive production of hyaluronan by
achieve the desired result of suppressing an inflammatory process in the intimal fibroblasts stimulated by frictional forces retains plasma dialysate in
normal synovial tissue. Similarly, the amount of RANKL (receptor activator the synovial cavity; in synovitis, the effusion is an accumulation of exudate
of NF-κB ligand, an essential factor for the development of osteoclasts) seen similar to a pleural effusion (i.e., an overspill from the inflammatory edema
in normal synovial tissue is low.9 The net result of this is to suppress the in synovial tissue created by increased vascular permeability). Recent theories
formation of osteoclasts within the normal synovium and preserve homeostasis about possible low-grade inflammatory and immune reactions contributing
within the normal joint. to the pathogenesis of osteoarthritis suggest that these two mechanisms of
effusion development may not be as distinct as originally thought.55 Addition-
ally, recent proteomic evidence suggests that the increased vascular permeability
FUNCTION as seen in inflammation may be related not only to an increase in interen-
The functions of synovial tissue have proven to be remarkably difficult to dothelial gaps but also to glycocalyceal damage and aquaporin
define.52 Similar to other soft connective tissue, synovium provides a deform- upregulation.56
able packing that allows movement of adjacent, relatively nondeformable
tissues. The difference between synovium and other soft connective tissue
is that it allows most of the movement to occur between rather than within
tissues. Areolar synovium may also have specialized viscoelastic properties The synovium provides the major structure that aids chondrocyte nutrition.
for coping with the stretching, rolling, and folding it undergoes during joint In normal joints, a surprisingly large proportion of hyaline cartilage lies
movement. within 50 µm of a synovial surface. In any one position, only a small proportion
of cartilage is opposed to the other articular surface, and synovium packs
most of the space between less congruent areas. In immature joints, the
incomplete subchondral plate may contribute to nutrition, but in adult joints,
this route is unlikely to be significant. Nutrition of areas of cartilage that do
not come into close contact with synovium (concave articular surfaces in
particular) must take an indirect route. Although a small proportion of
nutrition may be imparted by smearing of a thin film of fluid over these
surfaces during movement, indirect routes through cartilage matrix and the
apposed articular cartilage may be more important.6
Diarthrodial joints (in both cartilage and synovium) have been found to
express high levels of transforming growth factor-β (TGF-β), a latent complex
that requires activation to induce a biologic response. Recent in vivo experi-
ments suggest that shearing of synovial fluid because of physiologic joint
motion may play an important role in TGF-β activation, which may be
essential to maintain the biochemical content and structural integrity of
healthy cartilage.57,58
Even though the blood vessels in synovium provide the most direct route
FIG. 1.11  Normal synovium (×200) stained with an antibody to detect the interleukin-1 for cartilage nutrition, there is no evidence that they are structurally adapted
receptor antagonist. to this function. The fenestrae seen in superficial capillaries are present in
6 SECTION 1  Scientific Basis of Rheumatic Disease
tendon sheath synovium at sites where there is no cartilage (or tendon) A recent study in a mouse model of arthritis has also raised the possibility
dependent on their supply of small molecules. of cadherin-11 expression on synovial fibroblasts as a potential therapeutic
target in the treatment of RA60,61; inhibition of cadherin-11 interactions in
SYNOVIUM AS A TARGET FOR this model interfered with both synovial inflammation and cartilage invasion
by pannus without having any effect on bone erosion (predominantly osteoclast
IMMUNE-MEDIATED DISEASE dependent) or immunosuppression. In this model, inflammation could be
Synovial joints are involved in several immunologic and inflammatory reduced substantially by antibodies to cadherin-11 or a cadherin-Fc fusion
disorders, including RA, systemic lupus erythematosus, and seronegative protein. Furthermore, cadherin-11 expression has been found to promote
SpA. Perhaps the most important reason for studying the biology of the invasive behavior of fibroblasts and is increased by IL-17 and TNF-α, cytokines
synovium is to obtain insight into which immunopathologic processes are very relevant in RA pathophysiology.62-64
likely to be suitable therapeutic targets in inflammatory arthritides and in
particular whether the synovial fibroblast is a therapeutic target.2,3
Autoimmune responses to synovial antigens might theoretically occur,
but despite considerable effort, evidence for a specific synovial antigen is Despite the biologic importance of understanding how the synovium responds
lacking. Moreover, associated targeting of other tissues such as pericardium to damage and drives inflammation, remarkably little is known about how
or uveal tract requires an explanation. Few, if any, synovium-specific antigens stromal cells (as opposed to leukocytes) change during synovial development
are known, and when rheumatic disorders are associated with autoantibodies, and inflammation. Difficulties in accessing synovial tissue from normal subjects
the antigens involved are ubiquitous. and patients with early disease and the lack of good cell markers have proved
Understanding the microarchitecture of the normal synovium, including to be obstacles to such work. However, synovial stromal cells are a functionally
the wide range in microscopic appearances; cellular infiltrates; and production heterogeneous group with some displaying proinflammatory and destructive
of cytokines, enzymes, and other biologically relevant proteins, assists in properties but others being immune regulatory and helping facilitate tissue
understanding the relevant changes in synovial tissue architecture and repair. This has led to a dilemma: Which stromal cells should be targeted,
immunopathology in disease states. Although the architecture of the normal and which should be retained? A clear understanding of the biology and
synovium is not as homogeneous as previously portrayed, consistencies significance of synovial tissue biology is therefore essential to provide a
across the broad spectrum of normal synovial tissues can be contrasted with coherent rationale for targeting stromal cells in the future treatment of patients
those seen in chronically inflamed synovial tissue. The marked increase in with arthritis.
synovial lining layer thickness with a reversal of the normal ratio of type A
to type B intimal cells, which favors type B cells in normal synovium and
type A cells in RA, is an example of this. Numerous other examples can be
given, including the changes in subintimal cell content and cytokine and The authors would like to thank Drs. Malcolm Smith and Mihir Wechalekar,
chemokine production,59 vascular and lymphatic changes, and production whose chapter in the previous edition informed much of the present
of MMPs and stimulators of osteoclast formation. discussion.

1. Canoso JJ, Stack MT, Brandt KD. Hyaluronic acid content 15. Barland P, Novikoff AB, Hamerman D. Electron 27. Naylor AJ, Filer A, Buckley CD. The role of stromal cells
of deep and subcutaneous bursae of man. Ann Rheum Dis. microscopy of the human synovial membrane. J Cell in the persistence of chronic inflammation. Clin Exp
1983;42:171-175. Biol. 1962;14:207-220. Immunol. 2013;171(1):30-35.
2. Filer A, Antczak P, Parsonage GN, et al. Stromal 16. Perdiguero EG, Geissmann F. The development and 28. Buckley CD, Barone F, Nayar S, et al. Stromal cells in
transcriptional profiles reveal hierarchies of anatomical maintenance of resident macrophages. Nat Immunol. chronic inflammation and tertiary lymphoid organ
site, serum response and disease and identify disease 2016;17(1):2-8. formation. Annu Rev Immunol. 2015;33:715-745.
specific pathways. PLoS ONE. 2015;10(3):e0120917. 17. Edwards JC. The nature and origins of synovium: 29. McInnes IB, Buckley CD, Isaacs JD. Cytokines in
3. Filer A, Parsonage G, Smith E, et al. Differential survival experimental approaches to the study of synoviocyte rheumatoid arthritis—shaping the immunological
of leukocyte subsets mediated by synovial, bone marrow, differentiation. J Anat. 1994;184:493-501. landscape. Nat Rev Rheumatol. 2016;12(1):63-68.
and skin fibroblasts: site-specific versus 18. Edwards JC, Willoughby DA. Demonstration of bone 30. Wilkinson LS, Pitsillides AA, Worrall JG, et al. Light
activation-dependent survival of T cells and neutrophils. marrow derived cells in synovial lining by means of giant microscopic characterization of the fibroblast-like
Arthritis Rheum. 2006;54(7):2096-2108. intracellular granules as genetic markers. Ann Rheum Dis. synovial intimal cell (synoviocyte). Arthritis Rheum.
4. Edwards JC, Wilkinson LS. Distribution in human tissues 1982;41:177-182. 1992;35:1179-1184.
of the synovial lining-associated epitope recognised by 19. Henderson B, Revell PA, Edwards JC. Synovial lining cell 31. Medof ME, Walter EI, Rutgers JL, et al. Identification of
monoclonal antibody 67. J Anat. 1996;188:119-127. hyperplasia in rheumatoid arthritis: dogma and fact. Ann the complement decay-accelerating factor (DAF) on
5. Craig FM, Bayliss MT, Bentley G, et al. A role for Rheum Dis. 1988;47:348-349. epithelium and glandular cells and in body fluids. J Exp
hyaluronan in joint development. J Anat. 1990;171:17-23. 20. Edwards JCW, Wilkinson LS. Immunohistochemistry of Med. 1987;165:848-864.
6. Edwards JC, Wilkinson LS, Jones HM, et al. The synovium. In: Henderson B, Edwards JCW, Pettifer ER, 32. Stevens CR, Mapp PI, Revell PA. A monoclonal antibody
formation of human synovial joint cavities: a possible role eds. Mechanisms and models in rheumatoid arthritis. New (Mab 67) marks type B synoviocytes. Rheumatol Int.
for hyaluronan and CD44 in altered interzone cohesion. J York: Academic Press; 1995:133-150. 1990;10:103-106.
Anat. 1994;185:355-367. 21. Bhatia A, Blades S, Cambridge G, et al. Differential 33. Krane SM, Goldring SR, Dayer JM. Interactions among
7. Key JA. The synovial membrane of joints and bursae. In: distribution of Fc gamma RIIIa in normal human tissues lymphocytes, monocytes and other synovial cells
Cowdry EV, ed. Special cytology. New York: PB Hoeber; and co-localization with DAF and fibrillin-1: implications in the rheumatoid synovium. Lymphokines. 1982;7:
1932. for immunological microenvironments. Immunology. 75-87.
8. Singh JA, Arayssi T, Duray P, et al. Immunohistochemistry 1998;94:56-63. 34. Edwards JC. Synovial intimal fibroblasts. Ann Rheum Dis.
of normal human knee synovium: a quantitative study. 22. Lee MY, Kim WJ, Kang YJ, et al. Z39Ig is expressed on 1995;54:395-397.
Ann Rheum Dis. 2004;63:785-790. macrophages and may mediate inflammatory reactions in 35. Connolly M, Veale DJ, Fearon U. Acute serum amyloid A
9. Smith MD, Barg E, Weedon H, et al. Microarchitecture arthritis and atherosclerosis. J Leukoc Biol. regulates cytoskeletal rearrangement, cell matrix
and protective mechanisms in synovial tissue from 2006;80:922-928. interactions and promotes cell migration in rheumatoid
clinically and arthroscopically normal knee joints. Ann 23. Misharin AV, Cuda CM, Saber R, et al. Nonclassical arthritis. Ann Rheum Dis. 2011;70:1296-1303.
Rheum Dis. 2003;62:303-307. Ly6C(-) monocytes drive the development of 36. Boland JM, Folpe AL, Hornick JL, et al. Clusterin is
10. Davies DV. The structure and functions of the synovial inflammatory arthritis in mice. Cell Rep. expressed in normal synoviocytes and in tenosynovial
membrane. Br Med J. 1950;1:92-95. 2014;9(2):591-604. giant cell tumors of localized and diffuse types: diagnostic
11. Wilkinson LS, Edwards JC. Microvascular distribution in 24. Poulter LW, Janossy G. The involvement of dendritic cells and histogenetic implications. Am J Surg Pathol.
normal human synovium. J Anat. 1989;167:129-136. in chronic inflammatory disease. Scand J Immunol. 2009;33:1225-1229.
12. Xu H, Edwards J, Banerji S, et al. Distribution of 1985;21:401-407. 37. Ekwall AK, Eisler T, Anderberg C, et al. The
lymphatic vessels in normal and arthritic human synovial 25. Wilkinson LS, Worrall JG, Sinclair HD, et al. tumour-associated glycoprotein podoplanin is expressed
tissues. Ann Rheum Dis. 2003;62:1227-1229. Immunohistological reassessment of accessory cell in fibroblast-like synoviocytes of the hyperplastic synovial
13. Mapp PI. Innervation of the synovium. Ann Rheum Dis. populations in normal and diseased human synovium. Br lining layer in rheumatoid arthritis. Arthritis Res Ther.
1995;54:398-403. J Rheumatol. 1990;29:259-263. 2011;13:R40.
14. Ghadially FN. Fine structure of joints. In: Sokoloff L, ed. 26. Patel R, Filer A, Barone F, et al. Stroma: fertile soil for 38. Lefèvre S, Knedla A, Tennie C, et al. Synovial fibroblasts
The joints and synovial fluid. New York: Academic Press; inflammation. Best Pract Res Clin Rheumatol. spread rheumatoid arthritis to unaffected joints. Nat Med.
1978:105-176. 2014;28(4):565-576. 2009;15(12):1414-1420.
CHAPTER 1  The synovium 7
39. Leigh RD, Cambridge G, Edwards JCW. Expression of 47. Naylor AJ, Azzam E, Smith S, et al. The mesenchymal arthritic and normal synovial tissue. Arthritis Res.
B-cell survival cofactors on synovial fibroblasts. Br J stem cell marker CD248 (endosialin) is a negative 2002;4:201-208.
Rheumatol. 1996;1:110. regulator of bone formation in mice. Arthritis Rheum. 57. Albro MB, Cigan AD, Nims RJ, et al. Shearing of synovial
40. Lee BO, Ishihara K, Denno K, et al. Elevated levels of the 2012;64(10):3334-3343. fluid activates latent TGF-beta. Osteoarthritis Cartilage.
soluble form of bone marrow stromal cell antigen 1 in the 48. Ashhurst DE, Bland YS, Levick JR. An 2012;20:1374-1382.
sera of patients with severe rheumatoid arthritis. Arthritis immunohistochemical study of the collagens of rabbit 58. Fava R, Olsen N, Keski-Oja J, et al. Active and latent
Rheum. 1996;39:629-637. synovial interstitium. J Rheumatol. 1991;18:1669-1672. forms of transforming growth factor beta activity in
41. Fowler MJ Jr, Neff MS, Borghaei RC, et al. Induction of 49. Revell PA, al-Saffar N, Fish S, et al. Extracellular matrix synovial effusions. J Exp Med. 1989;169:291-296.
bone morphogenetic protein-2 by interleukin-1 in human of the synovial intimal cell layer. Ann Rheum Dis. 59. Parsonage G, Falciani F, Burman A, et al. Global gene
fibroblasts. Biochem Biophys Res Commun. 1995;54:404-407. expression profiles in fibroblasts from synovial, skin and
1998;248:450-453. 50. Suter ER, Majno G. Ultrastructure of the joint capsule in lymphoid tissue reveals distinct cytokine and chemokine
42. Marinova-Mutafchieva L, Taylor P, Funa K, et al. the rat: presence of two kinds of capillaries. Nature. expression patterns. Thromb Haemost.
Mesenchymal cells expressing bone morphogenetic 1964;202:920-921. 2003;90(4):688-697.
protein receptors are present in the rheumatoid arthritis 51. Haworth O, Buckley CD. Pathways involved in the 60. Lee DM, Kiener HP, Agarwal SK, et al. Cadherin-11 in
joint. Arthritis Rheum. 2000;43:2046-2055. resolution of inflammatory joint disease. Semin Immunol. synovial lining formation and pathology in arthritis.
43. Seki T, Selby J, Haupl T, et al. Use of differential 2015;27(3):194-199. Science. 2007;315:1006-1010.
subtraction method to identify genes that characterize the 52. Edwards JCW. Functions of synovial lining. In: 61. Firestein GS. Every joint has a silver lining. Science.
phenotype of cultured rheumatoid arthritis synoviocytes. Henderson B, Edwards JCW, eds. The synovial lining in 2007;315:952-953.
Arthritis Rheum. 1998;41:1356-1364. health and disease. London: Chapman & Hall; 1987:41-74. 62. Kiener HP, Niederreiter B, Lee DM, et al. Cadherin 11
44. Jay GD, Britt DE, Cha CJ. Lubricin is a product of 53. Yagi M, Mitsui Y, Gotoh M, et al. Role of the promotes invasive behavior of fibroblast-like
megakaryocyte stimulating factor gene expression by hyaluronan-producing tenosynovium in preventing synoviocytes. Arthritis Rheum. 2009;60:1305-1310.
human synovial fibroblasts. J Rheumatol. adhesion formation during healing of flexor tendon 63. Park YE, Woo YJ, Park SH, et al. IL-17 increases
2000;27:594-600. injuries. Hand Surg. 2012;17:13-17. cadherin-11 expression in a model of autoimmune
45. Arufe MC, De la Fuente A, Fuentes I, et al. Chondrogenic 54. Levick JR, McDonald JN. Fluid movement across experimental arthritis and in rheumatoid arthritis.
potential of subpopulations of cells expressing synovium in healthy joints: role of synovial fluid Immunol Lett. 2011;140:97-103.
mesenchymal stem cell markers derived from human macromolecules. Ann Rheum Dis. 1995;54:417-423. 64. Vandooren B, Cantaert T, ter Borg M, et al. Tumor
synovial membranes. J Cell Biochem. 2010;111:834- 55. Pelletier JP, Martel-Pelletier J, Abramson SB. necrosis factor alpha drives cadherin 11 expression in
845. Osteoarthritis, an inflammatory disease: potential rheumatoid inflammation. Arthritis Rheum.
46. Sakaguchi Y, Sekiya I, Yagishita K, et al. Comparison of implication for the selection of new therapeutic targets. 2008;58:3051-3062.
human stem cells derived from various mesenchymal Arthritis Rheum. 2001;44:1237-1247.
tissues: superiority of synovium as a cell source. Arthritis 56. Shahrara S, Volin MV, Connors MA, et al. Differential
Rheum. 2005;52:2521-2529. expression of the angiogenic Tie receptor family in
The articular cartilage
Dick Heinegård • Pilar Lorenzo • Patrik Önnerfjord • Tore Saxne • Linda J. Sandell

Key Points VI collagen network may have a role in protecting the cell and guiding
■ The cells in cartilage maintain function of the extracellular matrix (ECM) via matrix assembly. A set of molecules close to the cells have specific functions
controlled turnover in response to minor damage caused by fatigue and altered load in binding to particular cell-surface receptors and thereby provide signaling
by removing malfunctioning matrix constituents by breakdown and producing new of conditions in the matrix.
ones to achieve repair. This chapter focuses on describing the individual cartilage macromolecules
■ The cell obtains feedback on the quality of the ECM via a number of cell-surface and, when possible, their functional properties and their implications for
receptors such as integrins, DDR-2 (discoidin domain receptor 2), hyaluronan tissue assembly. In some instances, candidate enzymes have been implicated
receptors, and proteoglycans with specificity for different matrix molecules. in having roles in the degradation of specific macromolecules and will be
■ The composition of cartilage ECM is different close to cells in the territorial matrix discussed.
than in the more distant interterritorial matrix. Composition also varies between
different types of cartilage and from surface to deep articular cartilage. OVERALL TISSUE ORGANIZATION
■ The ECM of cartilage contains a specific proteoglycan—aggrecan—that provides a
very high fixed charge density and therefore an osmotic environment, with water
The part of the matrix closer to cells, the territorial matrix, has a somewhat
retention being essential for tissue resilience. Aggrecan is also part of a network in
different composition and structure than the matrix at some distance, the
which globular domains interact with other molecules.
interterritorial matrix (Fig. 2.2) (for references, see Heinegård and colleagues1).
Examples of components found in both compartments are collagen type II
■ Fibrillar networks with collagen as the major constituent provide the tensile
fibers and aggrecan; in contrast, collagen type VI is found particularly in
properties essential for load distribution and dissipation. The fibers contain other
the territorial matrix, and cartilage oligomeric matrix protein (COMP) and
matrix proteins (e.g., those bound at their surface) that mediate interactions with
cartilage intermediate layer protein (CILP) are primarily found in the interter-
other tissue structures, including neighboring fibers, which enhances their
ritorial matrix of normal cartilage. There is also a difference in the composition
mechanical qualities.
of cartilage from the superficial to deep layers. In the superficial layer, collagen
■ Cartilage ECM contains growth factors and proenzymes that are sequestered by
fibers are thinner and arranged in parallel with the surface of the tissue; in
binding to matrix macromolecules, and these substances can be released upon
the deeper layer, the fibers are thicker and arranged perpendicular to the
degradation of the carrier molecules.
surface, with a transition zone in between them (Fig. e2.1; also see Fig. 2.1).
■ As a result of degradation of cartilage matrix, the fragments formed are released into The superficial part of cartilage is enriched with a number of noncollagenous
surrounding fluids and can be used as indicators of the ongoing process, the molecules, notably lubricin and asporin. At the same time, other molecules
so-called molecular marker technology. are much less abundant in this part of the tissue, as exemplified by aggrecan,
■ Fragments of ECM components can activate innate immune responses such as which is particularly enriched toward the deeper parts of the tissue.1A Certain
complement. molecules such as CILP are found primarily in the middle portions of articular
cartilage. Although a number of cartilage components have been described
in recent years, our understanding of what specific requirements and functions
are met by molecules with such a restricted localization in tissue is still
Articular cartilage has key roles in the function of joints. A major function very limited.
is to take up and distribute load such that a given point of the underlying Different types of cartilage have markedly different compositions. Articular
bone can handle very high strain. Another role of cartilage is to provide cartilage has major similarities, yet multivariate analyses of proteomics data
low-friction movement. One key feature of joint diseases is deterioration of have demonstrated differences between knee and hip articular cartilage.2
joint function, which results from progressive damage to articular cartilage
by degradation of the structural components important for properties of the
tissue. Progressive joint destruction may eventually lead to total loss of the
cartilage accompanied by alterations in underlying bone in the common The major structure of aggrecan, illustrated in Fig. 2.3,1 consists of approxi-
disease of osteoarthritis (OA) (Fig. 2.1). Mechanisms triggering this tissue mately 100 chondroitin sulfate glycosaminoglycan chains, each built from
destruction are not known in detail, but it is clear that excessive load may a disaccharide unit that is repeated some 50 times but with extensive variability.
induce a remodeling process that fails to restore normal cartilage. Stimulation Each disaccharide contains uronic acid with a negatively charged carboxyl
of chondrocytes by cytokines such as interleukin-1 (IL-1), IL-6, and tumor group and an N-acetylgalactosamine with a sulfate in either the 4 or the 6
necrosis factor-α (TNF-α) induces the cells to degrade their surrounding position. Each chain will therefore contribute around 100 negatively charged
matrix and can, over an extended period, result in total dissolution of cartilage groups. The glycosaminoglycans are linked to a serine residue of the protein
in vitro. The identity of the individual enzyme or enzymes responsible for core of the proteoglycan via their reducing terminal end. The chains are
specific fragmentation of a particular matrix protein is not known, but in clustered, and clustering differs between the two regions referred to as CS
some cases, candidates have been established. A repair response often domain 1 and CS domain 2 (see Fig. 2.3). There is an additional glycos­
accompanies the ongoing tissue destruction. In more serious cases, this aminoglycan, keratan sulfate, that has a disaccharide building block of galactose
response is not sufficient, and tissue failure ensues. and an N-acetylglucosamine with a sulfate in the 6 position. These chains
A prerequisite for understanding the mechanisms of tissue destruction are shorter and particularly enriched closer to the N-terminal end of the
and failed repair is to know the functions of molecules in the extracellular protein core, in the so-called keratan sulfate–rich region. The aggrecan
matrix (ECM) and how they are assembled into larger networks. It is equally molecule is composed of some 30 such chains.
important to understand the mechanisms involved in their degradation. One The proteoglycan core protein contains globular domains flanking the
important and basic clinical observation in joint diseases such as rheumatoid three domains carrying glycosaminoglycan chains. The one most N-terminal,
arthritis (RA) and OA is that after joint replacement, the inflammation recedes, the hyaluronan-binding domain (G1 globe), contributes specific high-affinity
and the symptoms that have been plaguing patients are ameliorated or disap- binding of the aggrecan molecule to hyaluronan (see below). After a short
pear in the vast majority of cases. This brings up the question of whether interglobular domain there is a second globular domain (G2) that has structural
the components released from cartilage actually stimulate the inflammatory similarities to the G1 globe but does not bind hyaluronan and has no known
reaction. function. In the very C-terminal end of the aggrecan molecule, there is a
The main structural entities in cartilage include aggrecan and collagen. G3 globe with a lectin homology domain; it contributes by binding to other
The proteoglycan aggrecan has a primary role in taking up load and resisting proteins (e.g., fibulins and tenascins), which themselves can form molecular
deformation. The collagen network provides tensile properties, and the type complexes involving several such molecules (see Fig. 2.2). Recently, mutations

CHAPTER 2  The articular cartilage 8.e1



fibrillogenesis is a
multistep process.
N-terminal C-terminal
Fibers formed vary in size
propeptide propeptide
Procollagen molecule

Collagen molecule

25 Superficial

20 Gap
18 38 58 78 98 118 138 158 178

25 Intermediate

18 38 58 78 98 118 138 158 178 Articular cartilage


20 Intermediate
18 38 58 78 98 118 138 158 178 Deep


Pericellular Territorial Interterritorial

FIG. E2.1  The organization of collagen fibers at different distances from the surface of articular cartilage is shown, as well as differences in fibril diameters
between the pericellular, territorial, and interterritorial compartments.
CHAPTER 2  The articular cartilage 9



Normal cartilage
0 Col6 Superficial Decreased bone volume
Cartilage depth (mm)

Interterritorial (trabeculae)
TSP-1 Intermediate
Thickened subchondral bone

Deep Pericellular
Cartilage destruction
Calcified Tidemark
2 cartilage
Protein Bone compartments Altered cartilage


FIG. 2.1  As an example, one cartilage protein (cartilage intermediate layer protein [CILP]) shows a distinct change in localization from distribution
in normal cartilage in intermediate parts of the tissue and rather selectively in the interterritorial matrix to a prominence at the superficial parts
and primarily in the pericellular compartment in the diseased cartilage. The different distribution of matrix proteins with depth of the articular
cartilage is illustrated by way of three examples in the left diagram. OA, osteoarthritis.

that abolish binding of the G3 domain to the fibulins and tenascins have -EGE373-COOH sequence formed through cleavage by aggrecanase or via a
been shown to lead to familial osteochondritis dissecans.3 further matrix metalloproteinase (MMP) cleavage that forms a C-terminal
Many aggrecan molecules will bind to a single, very long strand of -PEN341-COOH sequence.10
hyaluronan, thereby forming large aggregates containing more than 100 Even from early experiments by Thomas11 and Fitton-Jackson and col-
glycosaminoglycan chains, each with some 100 negatively charged groups. laborators,12 it was clear that chondrocytes have a remarkable capacity to
These charges are thus fixed in the tissue, and the presence of counter ions replace aggrecan molecules removed from tissue by the use of enzymes
results in an osmotic environment that retains water and has an essential cleaving hyaluronan. It appears that a normal chondrocyte should be able
role in cartilage function by resisting compression and distributing load. to replace even large amounts of proteoglycans lost unless suppressed by
The interaction of aggrecan via its G1 globe with hyaluronan is stabilized cytokines such as IL-1 and TNF-α.
by the link protein having structures similar to the hyaluronan-binding There is increasing evidence that the aggrecan degradation encountered
domain of aggrecan. This link protein will bind to this domain of the pro- in early joint disease, such as OA, is effectively counterbalanced by increased
teoglycan, as well as to hyaluronan. synthesis and deposition of the molecule such that no overall loss takes
In normal cartilage turnover, as well as in pathology, the aggrecan molecule place.13,14
is cleaved by enzymes called aggrecanases (i.e., ADAMTS-4 [a disintegrin
and metalloproteinase with thrombospondin motifs] and ADAMTS-5)4,5 (see
Fig. 2.3). One site of this cleavage is in the interglobular domain between
the hyaluronan-binding G1 globe and the G2 globe. The cleavage occurs A major function of the ropelike collagen fiber networks in cartilage is to
between the amino acids EGE and ARG (for references, see elsewhere6). The provide tensile strength and distribute the load so that excessive local force
new N- and C-terminals formed have been used to develop antibodies that is not applied to the underlying bone. One type of fiber contains a core of
recognize the fragments produced only by the cleavage. These antibodies collagen type II with a minor constituent of collagen type XI connected via
have in turn been used to demonstrate such fragments in body fluids and a number of molecules bound at its surface.
tissue extracts.6,7 The cleavage occurring in the domain carrying the chondroitin The other fibrillar network contains a core of collagen type VI that forms
sulfate chains results in shortening of the aggrecan with ensuing decreased a filamentous network primarily in the territorial matrix. This network is
number of fixed charged groups (see Fig. 2.3). With aging, shorter aggrecan connected by a set of linker molecules to other molecules in the matrix,
molecules accumulate, the extreme being hyaluronan-binding domain with including collagen type II fibers and the major proteoglycans. These networks
no glycosaminoglycan-binding structures remaining.8 Indeed, in adults, a are discussed separately.
major proportion of the aggrecan molecules found in the interterritorial
matrix at a distance from the cells do not contain the G3, C-terminal globular
lectin homology domain (Aspberg and Heinegård, unpublished observation).
On the other hand, the molecules in the territorial matrix close to the cells Collagen type II and collagen type XI
contain this domain, probably a result of the gradual turnover of the aggrecan The major fibrillar network in cartilage contains primarily collagen type II
molecules. At the same time, there is substantial accumulation of G1 domain, with a minor constituent of collagen type XI.15 A major feature of both these
apparently retained by being bound to hyaluronan.9 This fragment can be types of collagen is that the molecule forming the fibers is made up of three
identified in tissue either via its new C-terminal as represented by an parallel, tightly associated polypeptide chains forming a very stable triple
10 SECTION 1  Scientific Basis of Rheumatic Disease




Fibronectin Procollagen II Collagen II/XI
Collagen XIII

Collagen II/XI
Matrillin 1/3

SS Biglycan/decorin

rin PRELP FIG. 2.2  Indicated are the interactions




between extracellular matrix (ECM) proteins

HS-PG Collagen VI

HS-PG and specific receptors at the cell surface.




Collagen IX
(syndecan) Note the different molecular composition of

the territorial matrix closer to the cell than in
Integrin the interterritorial matrix at some distance.
CD44, Receptor for hyaluronan; CHAD,

chondroadherin; CILP, cartilage intermediate

layer protein; COMP, cartilage oligomeric

matrix protein; CS, chondroitin sulfate; HA,
hyaluronan; HS-PG, heparan sulfate
proteoglycan; KS, keratan sulfate; NC4,
N-terminal globular domain of collagen 9.


COM Collage Fibulin Link
P n II protein

helix, the collagen molecule. It is extremely asymmetric—300 nm long and metabolized on degradation of the collagen, they eventually end up in urine
1.5 nm in diameter. An amino acid quite unique for collagen is hydroxyproline, and can be measured as indicators of collagen breakdown.18
which is essential for stability of the molecule because of the hydrogen bonds
formed to the hydroxyl group. Collagen type II is produced as a procollagen Collagen type XI
that does not form fibrils until the propeptides at both ends of the C- and This fibril-forming collagen is an integral component of the fibers forming
N-termini are cleaved off (see Fig. e2.1). After this cleavage, the molecules the main network in cartilage. Collagen type XI consists of a major triple-
form fibers by interactions with other collagen molecules such that a large helical portion, similar in size to that of collagen types I and II, but in
part of the surface of the molecule is engaged and the collagen molecules contrast to these collagens, the N-terminal propeptides are retained with the
are positioned so that they form a so-called quarter stagger arrangement in molecule incorporated into the fiber. Some reports have indicated that the
relation to one another. retained N-terminal parts are exposed at the surface of the fibers, with collagen
The assembly process is regulated by a number of molecules in the matrix, type II being the major constituent and the major triple-helical portion being
including collagen type XI. It appears that the dimensions of the fiber formed located more centrally in the fiber.19 Collagen type XI together with collagen
depend on the relative proportion of collagens type II and XI, with the typical type II appears to form the initial assembly of microfibrils that regulate
ratio being on the order of 50 : 1. This may depend on the presence of a further assembly of the cartilage collagen fiber, at least in skeletal morpho-
central core of microfibrils of collagen type II and XI that direct assembly genesis.16 Interestingly, collagen type XI forms cross-links to primarily other
of the fiber.16 collagen type XI molecules.20 There are examples of mutations in collagen
Collagen fibers in the superficial and deep layers of articular cartilage are type XI chains with ensuing major growth disturbances, thus indicating a
different in dimension and direction. Thus, whereas fibers in the superficial role in cartilage growth and stability.21
layer of articular cartilage are thin and run in parallel, the thicker fibers in
the deeper parts of cartilage run perpendicular to the surface. In the transition Collagen type IX
zone layer, fibers run at an angle (see Fig. e2.1).17 The organization can be This molecule is a member of the FACIT collagens (fibril-associated collagens
seen as Benninghoff arcades on polarized light microscopy. with interrupted triple helices) and is found in tissue bound at the surface
Collagen fiber formation is influenced by a number of matrix molecules, of the fibrils, with collagen type II being the major constituent. Collagen
such as decorin, asporin, fibromodulin, COMP, and a special variant of an type IX contains three different α chains with three triple-helical domains
oversulfated chondroitin sulfate chain. In several of these cases, the molecules (col1, col2, and col3), each surrounded by a noncollagenous domain (NC1,
are also retained bound at the surface of the collagen fiber. This is particularly NC2, NC3, and NC4). The NC4 domain with its adjacent col3 triple helix
evident for collagen type IX, which has part of the molecule extending out protrudes from the fibers and is available for interactions with other molecules
from the fiber. These molecules appear to have roles in providing sites for in the ECM, as schematically illustrated in Fig. 2.2.22 Examples of such
interaction with other matrix molecules, including fibers other than those ligands are COMP and the tyrosine sulfate domain of fibromodulin. Collagen
to which the protein is bound (see Fig. 2.2). type IX often contains a chondroitin sulfate side chain bound at the NC3
An important feature of the collagen fiber network is that the interactions domain. Its role in function of the collagen is not known.
become sealed by covalent cross-link formation after the fibers are assembled Functionally, collagen type IX has been shown to interact with matrilins,
outside the cell. This cross-linking depends on the oxidation of lysine and COMP, and in particular, collagen type II. The collagen is actually covalently
hydroxylysine residues by lysyl oxidase to provide an aldehyde function that cross-linked to collagen type II in the fibers of adults.23 When collagen type
forms a Schiff base with a neighboring lysine amino group. These are then IX is added in vitro to fibril-forming systems of collagen type II, assembly
rearranged to become stable pyridinoline groups that cross-bridge between and fiber formation are retarded.
the molecules and within the molecules of a fiber. These cross-links are Mutations in collagen type IX lead to severe growth disturbances, such
important for mechanical stability of the collagen. Because they are not as pseudoachondroplasia or multiple epiphyseal dysplasia. Some of these
CHAPTER 2  The articular cartilage 11


Link protein MMP
cleavage sites ADAMTS EGF homology
G1 globe cleavage sites (spliced)
binding region
G2 globe

KS rich
G3 globe

CS domain 1 CS domain 2


- SO3
CH2OH CH2OH COO- 6 O 0.5µm
O 6 5 O O
6 5 6 5 4 2 O
4 4 2 O 4 2 5 1
2 O 1 O O
1 HO 3 1
O 3 HO 3
3 NH
SO3- Ac
Ac n n
Keratan sulfate Chondroitin sulfate

FIG. 2.3  Depiction of the structure and organization of the proteoglycan aggregate. Also shown is a rotatory shadowing electron micrograph (courtesy of Matthias Mörgelin)
of an aggregate isolated from tissue. The sites for degradation by the primary active ADAMTS-4 and ADAMTS-5 are indicated, as well as one site for cleavage of matrix
metalloproteinase (MMP). CRP, C-reactive protein; CS, chondroitin sulfate; EGF, epidermal growth factor.

disturbances are similar to those with a mutated COMP molecule, which is It appears that even though the collagen itself does not turn over, molecules
of special interest in view of the high-affinity interaction that this protein on the fiber’s surface are continuously removed and replenished.
shows with all four NC domains of collagen type IX. Furthermore, early Molecules with putative roles in regulating fibril formation are found
lesions of articular cartilage similar to those found in OA develop in mice among those that can bind collagen in vitro.
with knockout of collagen type IX.24
Cartilage oligomeric matrix protein
Cartilage oligomeric matrix protein is a molecule primarily found in cartilage,
MOLECULES REGULATING COLLAGEN FIBER ASSEMBLY where it is quite abundant at a concentration of around 0.1% of the tissue’s
The dimensions and orientation of collagen fibers in tissue vary between wet weight. The molecule is made up of five identical subunits, each with
different layers of articular cartilage. Moreover, fibers in the territorial matrix a molecular weight of around 87,000 Da. The five subunits are held together
close to cells are thinner and have similar dimensions in different layers of by a coil-coiled domain close to the N-terminal end, and disulfide bridges
cartilage. In contrast, fibers in the interterritorial matrix are thicker and have further stabilize the binding. The subunits are made up of several modules,
larger and more variable diameters in the deeper layers. This regulation of including some binding calcium. At the C-terminal end, there is a globular
fiber diameter is achieved by a number of macromolecules that bind to domain that is involved in interactions with other proteins in the matrix.
collagen. The exact role of individual molecules in achieving the final dimen- The molecule can be viewed as a bouquet of tulips tied together at their
sions and direction of the fibers is not clear, although there are a number stalks (see Fig. 2.2).27
of examples in which inactivation of individual genes of the involved proteins Cartilage oligomeric matrix protein, also referred to as thrombospondin-5,
leads to altered dimensions of collagen fibrils. It is notable that asporin, is a member of this family of proteins. Cartilage also contains other throm-
similar to decorin, inhibits collagen fiber formation and that levels of these bospondins that share the same properties, with thrombospondin-1 and
proteins are selectively very high in superficial parts of the articular cartilage, thrombospondin-4 being particularly abundant. These thrombospondins,
where collagen fibers are thin.25 however, contain an extension beyond the coiled-coil domain in the N-terminal
The extremely long half-life of collagen type II, in excess of 100 years,9,26 that has a heparin-binding motif, in this manner adding additional interacting
indicates that very little collagen is eliminated over the life of an individual; sites. Whereas thrombospondin-4 contains five identical subunits,
however, at the same time, fibers defective as a result of fatigue have to be thrombospondin-1 contains only three.1
repaired. It is possible that the very variable dimensions of the fibers in the The three-dimensional structure of the C-terminal domain of
interterritorial matrix result from adding newly synthesized collagen molecules thrombospondin-1 has been resolved, and its organization has been found
at the fiber surface and thus gradually increasing the diameter to provide to be stabilized by a large number of calcium ions.28 Because the C-terminal
mechanical stability. domain of the various thrombospondins shows a great deal of conservation,
A number of molecules bound at the surface of collagen fibers are likely it is likely that its structure is similar in all five members of the family.
to prevent further accretion of collagen. It is plausible that these molecules Cartilage oligomeric matrix protein has been shown to bind to collagens
will need to be removed before new collagen molecules can be added to a type I and II,1 where each individual C-terminal globular domain provides
fiber for remodeling or repair. high affinity in the nanomolar range. Four binding sites are evenly distributed
12 SECTION 1  Scientific Basis of Rheumatic Disease
along the collagen molecule. There is one at each end and two positioned lumican up to four, and keratocan only one.38 PRELP, in contrast, contains
along the filament such that the distance is similar between the four binding a cluster of basic residues that contribute heparin-binding activity for this
sites. Even though each COMP molecule has five identical binding sites, an molecule.39
individual molecule can engage only one binding site on each collagen Collagen binding of fibromodulin has been studied extensively. It has
molecule and not span the distance between two such sites. Therefore, each been shown that the molecule inhibits fibril formation in vitro. Fibromodulin-
COMP molecule has the potential to bind to five different collagen molecules. null mice show altered collagen fibril dimensions, particularly apparent in
The quarter stagger arrangement of collagen in the fiber and the fact that a the tail tendon. Unexpectedly in view of the inhibitory effects observed in
pentameric microfibril unit appears to exist29 may relate to the four similarly vitro, the tendon contains a much larger number of thin fibrils. An explanation
spaced collagen binding sites in the molecule. COMP accelerates and provides appears to be a higher abundance of the related lumican molecule, which
faster collagen fibril formation in vitro. It appears that this effect is mediated could be shown to bind to the same site on the collagen, albeit with somewhat
by the COMP molecule bringing together several collagen molecules to lower affinity. It thus appears that lumican may guide early events in fibril
facilitate their interactions in the forming fiber. The COMP molecule does formation. The molecule may then be competed away by fibromodulin to
not remain bound directly to the surface of the forming fiber. The molecule introduce a different function. Because the mRNA levels and therefore synthesis
thus appears to function as a catalyst to enhance fibril formation. of lumican were lower in null mice, it appears that the higher levels of this
In the cartilage of growing individuals and notably in growth plates, protein were caused by retarded elimination apparently secondary to lack
COMP is primarily localized close to the cells in the territorial matrix, where of competition by fibromodulin.1
it may have a role in stimulating collagen fibril formation.30 These findings illustrate that fibril formation takes place in many steps
Cartilage oligomeric matrix protein has the ability to interact with all involving a set of different molecules with different roles.
four NC domains of collagen type IX with similar high affinity in the nanomolar Fibromodulin is also present on collagen fiber in tissue, where it is bound
range. The interaction is mediated via the C-terminal globular domains (see in the gap region.1 It appears to be bound via its protein core with exposure
Fig. 2.2). of a keratan sulfate chain, as well as the tyrosine sulfate domain, which then
In adults, COMP is localized primarily in the interterritorial matrix and become available to interact with PRELP and the NC4 domain of collagen
may be bound to collagen type IX or one of the matrilins, which in turn type IX located on neighboring fibers. Such interactions are important for
bind to the surface of the collagen fiber (see Fig. 2.2). The role of COMP the stability and properties of the collagen network.
in adult cartilage appears to be stabilization of the collagen fiber network. Fibromodulin is a target in joint disease. In a model of articular cartilage
Mutations in the calcium-binding domain of COMP, as well as in the destruction caused by stimulation with IL-1 in explant culture, we have
C-terminal domain, have been shown to lead to severe growth disturbances been able to show that fibromodulin is degraded after aggrecan and that the
in the form of pseudoachondroplasia or multiple epiphyseal dysplasia. A molecule is initially cleaved by MMP-13 to release almost the entire N-terminal
feature of these conditions is that material retained in the endoplasmic tyrosine sulfate domain while the remainder of the molecule is initially
reticulum of chondrocytes contains both COMP and collagen type IX.31 retained in the tissue, probably bound to the collagen.41 This loss of the
On the other hand, the COMP-null mouse shows no detectable alteration anionic domain is likely to alter the properties and interactions of fibromodulin.
in phenotype.32 It is possible that other molecules compensate for the lack At the same time, MMP-13 can release the entire NC4 domain of collagen
of COMP function in such mice. type IX from cartilage.42 This results in loss of function, with impaired
Cartilage oligomeric matrix protein is significantly upregulated in early noncovalent associations between collagen fibers most likely leading to
stages of OA, even long before diagnosis, in an apparent attempt at repair. impaired maintenance of structure and function of the cartilage. This may
At the same time, the protein already deposited is cleaved and released from represent a central mechanism in the swelling and surface fibrillation observed
the tissue. Indeed, assay for such fragments released into body fluids has in early OA.
been used to measure altered cartilage metabolism as a biomarker for arthritic It is of interest to note that the particular fragment retained in the tissue
disease.33 is found only in pathologic and not in normal tissue, although there is
continuous turnover of matrix constituents in response to altered load,
Decorin including removal of damaged components. This normal turnover appears
Decorin was the first molecule in the leucine-rich repeat (LRR) protein to involve different mechanisms of cleavage.
family to be cloned and sequenced. This family of molecules in the ECM
contains four subclasses with a total of 12 members, all of which appear to Other leucine-rich repeat proteins
share the function of binding to collagen (see Fig. e2.2).1 PRELP is distinguished by having a basic, heparin-binding N-terminal
One functional domain of these molecules is a central LRR region, where domain.43 This mediates binding to molecules with heparan sulfate side
residues, particularly leucine, are found at conserved locations in each repeat chains, including perlecan and cell-surface syndecan and glypican. Simultane-
of some 25 amino acids, albeit somewhat variably long. Most of these molecules ously, the protein binds via its LRR domain to two sites on fibril-forming
have 10 to 11 such repeats, and the entire domain contains a disulfide loop collagen types I and II. Thus, the molecule has the potential to bridge from
structure at each end. One subgroup contains molecules with only six repeats. the collagen network back to the cell surface. It consequently has the potential
For further details on decorin, see ExpertConsult.com. to provide feedback to cells on the condition of the matrix. Little is known
Decorin binds tightly to the fibril-forming collagens with a KD in the of alterations in PRELP in joint disease. An interesting observation is that
nanomolar range. Binding is close to the C-terminus of the collagen as shown the isolated heparin-binding domain of PRELP binds to osteoclast precursors
for collagen type I. Binding occurs via the LRR region and particularly involves via cell-surface proteoglycans. The peptide becomes internalized and trans-
repeats 4 and 5.1 The critical sequence has been identified as SYIRIADTNIT.36 locates to the nucleus, where it inhibits nuclear factor-κB (NF-κB), the target
Via its binding to collagen, decorin inhibits the formation of collagen of receptor activator of NF-κB ligand (RANKL), an important stimulator of
fibers in vitro in a dose-dependent manner. Accordingly, mice lacking decorin osteoclast development. The effect of the peptide is thus to inhibit osteoclast
as a result of knockout technology have irregular collagen fibers with increased formation. Indeed, PRELP has been used to prevent the development of
diameter, particularly prominent in skin.37 They do not show increased early osteoporosis in ovariectomized mice.44
joint pathology, thus indicating that other molecules may compensate for Chondroadherin is a cell-binding protein that forms its own subclass.
the lack of decorin in articular cartilage. The protein is further discussed later.
Decorin is bound to collagen fibers in tissue, with the glycosaminoglycan Asporin is a close relative of decorin but differs in having a variably long
chains being free to interact with other molecules. Thus, decorin can cross- polyaspartate sequence in the N-terminal end. The number of aspartate
bridge to neighboring collagen fibers, as well as to other molecules in the residues varies among individuals.45 It has been demonstrated in studies of
local environment (see Fig. 2.2). several cohorts of individuals in Asia that OA is overrepresented in individuals
with 14 such residues versus those with 13 residues in the asporin N-terminal
Fibromodulin and lumican structure.46 In a different study from the United Kingdom, no such pronounced
Fibromodulin and lumican belong to the same subclass of LRR proteins but relationship could be discerned, thus indicating that other factors are also
with a gene arrangement distinct from that of decorin. Other members of involved.47 One function of the polyaspartate sequence appears to be calcium
this subgroup are keratocan, osteoadherin, and PRELP. All these molecules binding, for which there may be a difference between the 13 and 14 aspartate
except PRELP contain tyrosine sulfate residues in the N-terminal extension. variants.48 At the same time, asporin binds to collagen at the same sites that
Notably, the number of such sulfate residues is variable both with regard to decorin does,48 which can facilitate its fixation to collagen fibers in the tissue.
the relative proportion of candidate tyrosine residues that are sulfated within The molecule may thus have roles in regulating mineralization, which is
a given molecule and with regard to the number of such tyrosine residues relevant to the development of OA. Asporin, similar to decorin, biglycan,
that may carry a sulfate. One extreme is represented by fibromodulin, which and fibromodulin, appears to bind transforming growth factor-β (TGF-β).
contains up to nine sulfate residues; osteoadherin contains up to eight, There is a set of LRR proteins with only six repeats.1
CHAPTER 2  The articular cartilage 12.e1

One variable of almost all the molecules in the family is an N-terminal The function of the domain with clustered tyrosine sulfate residues is
extension of generally less than 20 amino acids, which may contain a variety becoming clearer. The domains in fibromodulin and osteoadherin mimic
of substituents (see Fig. e2.2). Some of the molecules also have a C-terminal heparin in many interactions and bind growth factors (e.g., fibroblast growth
extension. There is also variation in glycosylation of the repeat domain, factor-2 [FGF-2]), cytokines (e.g., oncostatin M and IL-10), MMPs (e.g.,
which usually contains a few N-linked oligosaccharides. In some of the small MMP-13 for fibromodulin), and a number of matrix proteins via their heparin-
LRR proteins, such as fibromodulin and lumican, some of the oligosaccharides binding domains (e.g., PRELP and chondroadherin).40 Thus, the fibromodulin
may contain a variably long array of 6-O-sulfated lactosamine repeat disac- on a collagen fiber appears to extend its tyrosine sulfate domain, which can
charides (see Fig. 2.3) extended to form the glycosaminoglycan keratan bind to molecules with cationic domains. One such charged structure is the
sulfate. NC4 domain of collagen type IX, which has been shown to interact with
The three-dimensional structures of decorin34 and biglycan35 have been the anionic N-terminal extension of fibromodulin,40 as schematically outlined
resolved by x-ray crystallography. These molecules contain one (decorin) in Fig. 2.2.
or two (biglycan) glycosaminoglycan chains bound at the N-terminal extension. All the molecules in this family appear to bind to collagen via their LRR
These chains are chondroitin sulfate or dermatan sulfate, depending on the domain with dissociation constants ranging from 1 to 10 nM.1 Only in a
tissue, though very similar between the molecules in a given tissue. In few cases has it been established exactly where along the collagen that the
articular cartilage, the chain is a low-epimerized dermatan sulfate in which LRR protein binds.
a few of the glucuronic acid residues have been epimerized to iduronic acid, Such proteins include epiphycan, mimecan, and opticin. They all contain
thereby increasing its structural variability. This glycosaminoglycan has the N-terminal extensions carrying glycosaminoglycan chains (epiphycan), tyrosine
capability of specific interactions with other molecules in the matrix, including sulfate residues (mimecan), or O-glycosidically linked oligosaccharides
binding other dermatan sulfate chains.1 (opticin) (see Fig. e2.2). Mimecan has also been named osteoglycin. There
Based on the x-ray crystallographic data presented, the LRR family of is limited knowledge on alterations of these proteins in joint disease. They
molecules forms a curved structure in which two molecules form a dimer do show interesting differences between different types of cartilage. Mimecan
in the crystals that overlap in opposite directions in about 50% of their is present in articular cartilage, menisci, and intervertebral disks but virtually
length such that the N-terminus of one molecule is located in the middle absent in nasal and tracheal cartilage, but epiphycan is prominent in the
of the curved domain of the other. Support for such a dimeric structure was latter cartilage and absent from the others. Opticin is particularly prominent
obtained by other techniques such as gel filtration with online dynamic light in connective tissues other than those of joints.
scattering.34 Electron microscopy indicates that decorin and biglycan, as well
as chondroadherin, may also exist in a monomeric form at very low concentra-
tions (Mörgelin and Heinegård, unpublished work).
The presence of the proteoglycan as a monomer or a dimer has specific
relevance for interactions and functions. Thus, a monomeric molecule has
only one of each interacting site, but a dimer may exhibit two of each binding


CS/DS chain

CS/DS chain

Decorin KS chain
oligosaccharide N-linked
Asporin oligosaccharide
FIG. E2.2  Note the variability in the N-terminal domain

between molecules, which shows negatively charged Tyrosine sulfation Fibromodulin

groups of different character (glycosaminoglycan, CHAD

tyrosine sulfate, polyaspartate, or a cluster of basic Lumican

amino acids). The C-terminal extensions of two
members have distinct features. The dimeric form OSAD
+++ +++++ PRELP
found for decorin and biglycan on x-ray
crystallography is illustrated schematically. CHAD,

Chondroadherin; CS, chondroitin sulfate; ECM, Keratocan

extracellular matrix; LRR, leucine-rich repeat; OSAD,

osteoadherin. OSAD




** *** Opticin
CHAPTER 2  The articular cartilage 13

Heparin and heparan sulfate in the early bone anlagen. Its role in cartilage is becoming clearer, and data
A glycosaminoglycan not prominent in cartilage is heparan sulfate, which indicate roles in collagen network function and integrin binding (for references,
has structural features overlapping those of heparin. Heparan sulfate is found see Klatt and colleagues55). Its two vWFA homology domains (only one in
as side chains of extracellular perlecan (discussed later), which has a very matrilin-3) may mediate the ability of the protein to bind to collagen. Matrilin-1
large protein core with several domains interacting with a number of other was initially isolated because of its apparent ability to bind to aggrecan.
proteins in the ECM. The proteoglycan contains an N-terminal domain with Mutations in matrilin-3 can induce multiple epiphyseal dysplasia or
some three glycosaminoglycan chains and a C-terminal domain with up to pseudoachondroplasia with skeletal malformations, similar to those caused
two chains. These chains may be heparan sulfate or chondroitin sulfate by mutations in the interacting partners of the proteins, such as collagen
(discussed later). IX.56
Heparan sulfate contains two types of disaccharide repeats consisting of Matrilin-1 can be isolated from cartilage as a mixed polymer that contains
either a glucuronic acid and an N-acetylglucosamine or an iduronic acid subunits of matrilin-1, as well as matrilin-3. The functional significance of
and an N-acetylglucosamine. The hexosamine carries an O-sulfate group, this heterocomplex is not understood.57
and some of the residues have an additional N-sulfate instead of the N-acetyl
group. Also, the uronic acid may be sulfated, usually in its 2 position. Thus, A module cross-linking to other matrix constituents
there is extensive variability in the building blocks, which are then assembled Decorin and biglycan appear to also have roles in the completed network.
in variably long stretches with different repeat stretches of low- and high- Collagen type VI isolated from chondrosarcoma cartilage–like tissue contains
sulfated residues. biglycan or decorin bound at the N-terminal globules. The proteoglycan in
Other heparan sulfate–containing proteoglycans are found at the cell turn has a bound member of matrilin-1, -2, or -3. This interaction is tight
surface of the chondrocyte. These include four different syndecans,49 which with a dissociation constant on the order of nanomolar. The matrilin, in
contain a domain intercalated in the cell membrane, and some six glypicans, turn, binds to a procollagen type II molecule or a completed collagen fiber.
which are joined via a glycosylphosphoinositide linkage.50 The syndecans Alternatively, matrilin can bind to an aggrecan molecule (see Fig. 2.2).54
contain an intracellular signaling domain, and as discussed later, they are Thus, collagen type VI seems to be a center in a scaffold that binds to the
involved in a number of cell reactions, and bound molecules can induce other major networks in the matrix. Collagen type VI is found only in the
signal transduction. territorial matrix closer to cells and is absent from the interterritorial matrix
at some distance from the cells. Information on alterations in collagen type
FIBERS AND NETWORKS WITH COLLAGEN TYPE VI AS THE VI in joint disease is limited, but it is important to note that the protein is
found to be particularly enriched in tissue subjected to load.
There is a different fibrillar network in cartilage with a more restricted distribu-
tion in tissue. It has collagen type VI as a major constituent and is primarily MOLECULES INTERACTING AT THE
present in the territorial matrix surrounding the cells (see Fig. 2.1). CELL SURFACE AND MODULATING
Collagen type VI has a distinctive network of beaded filaments. This molecule It is important to realize that chondrocytes have the ability to both degrade
contains three different α chains with a central triple-helical domain flanked the matrix and replenish lost molecules with new constituents in a process
by globular domains. The N-terminal portion, particularly the α3(VI) chain, of remodeling the tissue in response to material fatigue or to altered load.
contains nine von Willebrand factor A (vWFA)-like repeats. In addition, the The cells are guided in this endeavor by receptors at their surface that
C-terminal portions of all three chains contain two vWFA repeats, as well recognize specific molecules in the matrix (see Fig. 2.2). Binding elicits
as other motifs with less clear functions. The vWFA domain is found in specific signals that will either induce cellular spreading and migration by
many proteins, where it is involved in protein–protein interactions.51 engaging the cytoskeleton or lead to alterations in transcription and protein
While still within the cell, two collagen type VI molecules associate in synthesis. Other stimuli that affect cells include mechanical forces, which
an antiparallel fashion such that the dimer is flanked by the N-terminal provide signals that crosstalk with those from other interactions at the cell
domains, and the two C-terminal domains are placed so that they form two surface. Indeed, some data indicate that some of the signals elicited by
interior globular structures along the filament formed by the two triple mechanical load involve integrins.58
helices. Two such dimers associate laterally to form a tetramer, with the
globular domains at each end representing a pair of the vWFA-rich globules.
Internally in the triple-helical central filament are found the two globular
structures representing the C-terminals.51 The tetramer is secreted from the The integrins contain an α chain noncovalently bound to a β chain. The
cell, and its N-terminal structures lay the ground for further associations cells in connective tissues contain either a β1 or a β3 chain in conjunction
with fibrils involving both end-to-end and side-to-side interactions. This with one of many α chains. The various integrins have different preferred
assembly process appears to be governed by other molecules, particularly ECM ligand proteins.59 As an example of the organization of integrins, members
members of the LRR protein family. of one such family with four members bind to collagens. They contain a β1
chain in combination with an α1, α2, α10, or α11 chain. These various integrins
Biglycan and decorin have different tissue distributions such that the one with α10 appears to be
Biglycan and decorin both bind with high affinity to the collagen type VI unique to cartilage, but the others have more ubiquitous distributions. Their
N-terminal domain, independently of their glycosaminoglycan side chains. binding to collagen may elicit different responses and result primarily in
This binding is a prerequisite for formation of the collagen type VI beaded either altered production of enzymes degrading the matrix or production of
filament network in vitro. Regulation of collagen type VI assembly also building blocks such as collagen molecules.
depends on the presence of the chondroitin–dermatan sulfate side chains, A number of factors limit our ability to discern which integrins are present
where the two present in biglycan provide more efficient filament formation on chondrocytes in tissue. One factor is limited accessibility by the antibodies
than the single chain in decorin does. The two closely related proteoglycans used, and another is changes in the presence of integrin during the long
appear to bind to the same site, which interestingly does not seem to involve procedure for cell isolation. Thus, the current information on the presence
the triple-helical domain in this collagen.52-54 of integrins at the cell surface is variable. It is likely that a number of integrins
change their expression on chondrocytes in normal and in pathologic tissue
Matrilins in response to environmental factors.
Cartilage matrix protein (CMP, matrilin-1) was the first of the four matrilins Many of the molecules binding to integrins are not unique for cartilage.
identified. These proteins (for references, see Klatt and colleagues55) contain The collagen-binding integrins do not appear to be specific for a particular
vWFA domains. Matrilin-1 has two such domains in each of the three identical fibril-forming molecule. One exception is chondroadherin. This protein
subunits with a molecular mass of around 50,000 Da; matrilin-3, with four appears to be restricted to cartilage and can bind α2β1 integrin but not other
subunits, has only one such domain. integrins containing the same β chain.
Matrilin-1 and matrilin-3 are quite restricted to cartilage and show similar
distribution between different tissues; the others have a more general distribu- Chondroadherin
tion. Interestingly, matrilin-1 is even further restricted and not found in Chondroadherin is a member of the LRR protein family that is in its own
articular cartilage and intervertebral disks, but it is particularly prominent subclass. This protein differs from other LRR family members in that the
in tracheal cartilage. The protein is present in the more immature cartilage C-terminal cysteine loop is double, and the protein has a short C-terminal
of the femoral head during earlier phases of development and can be seen extension of basic amino acids. The protein has no N-terminal extension.
14 SECTION 1  Scientific Basis of Rheumatic Disease
Chondroadherin specifically binds to cell-surface α2β1 integrin1,60 via a sequence CILP, and PRELP.1 As discussed earlier, isolated heparin-binding domain
in one of its C-terminal disulfide loops. Interestingly, integrin binding elicits fragments can have variable effects on, for example, cells, depending on
signals that do not induce cell spreading. The very C-terminal short extension their fine structure. Effects range from inducing cartilage breakdown in vitro
peptide of chondroadherin contains a different cell-binding sequence that and in vivo to decreasing bone breakdown by targeting osteoclast
engages the heparan sulfate chains of cell-surface proteoglycans, notably precursors.
syndecans.61 Binding leads to cell spreading, as well as enhanced integrin Interestingly, as discussed earlier, the N-terminal tyrosine sulfate domain
activity. The combined engagement of integrins and heparan sulfate leads of fibromodulin can mimic heparin in many interactions and bind other
to cells forming focal adhesion complexes. proteins to crossbridge networks, as well as sequester growth factors.
Chondroadherin is present in articular cartilage but virtually absent from
menisci. The protein is particularly enriched during growth in the prehy-
pertrophic zone of the growth plate where cell multiplication is slowed, in
line with its in vitro effects on cell behavior. Interestingly, the protein appears In joint disease, inflammation is a frequent component causing pain and
to be lost early in the process of joint damage in OA (unpublished limiting function. The inflammation is usually chronic, and one issue is
observations). whether components from cartilage can propagate the inflammatory activity.
It is a long-standing observation that when cartilage is removed in joint
Fibronectin arthroplasty, the inflammation recedes, thus indicating that factors released
Fibronectin62 is present in most tissues and can actually form its own fibrils, from the tissue have a role in propagating inflammation.
which appear to have roles in guiding matrix assembly and cell migration. It has been shown that some patients with RA have circulating antibodies
The protein contains two identical subunits held together by disulfide bonds to collagen type II. Furthermore, an arthritis condition can be elicited by
close to their C-terminal end. injecting mice and rats with collagen type II (collagen-induced arthritis).
Fibronectin contains a collagen-binding domain with preference for The disease can be transferred by antibodies to specific collagen type II
denatured collagen (gelatin). There are integrin-binding domains in which epitopes.70 It is hence possible that antibody binding activates complement
the RGD (arginine–glycine–aspartic acid) sequence represents the classic and inflammation.
motif of integrin binding. This motif in fibronectin preferentially binds α5β1 More recent findings indicate that matrix proteins may activate innate
integrin, although αVβ3 integrin can also interact. Fibronectin also has other immunity. Fibromodulin has been shown to be as active as immune complexes71
integrin-binding domains.63 in activating the classical pathway of complement, although it does not bind
Another motif is represented by the two heparin-binding domains on to the same site. Fibromodulin binds to the head domains of C1q rather
each subunit, each some 20 kDa. These domains can interact with heparan than to the triple-helical stalk region with ensuing deposition and activation
sulfate proteoglycans at the cell surface, including syndecan, which also of C4 and C3.
represent signaling molecules.64 Interestingly, fibromodulin can also recruit factor H by binding to a different
Although the details are not known, fragments of fibronectin (e.g., those site than that for C1q and thereby inhibit further activation of the complement
containing either of the heparin-binding domains), when added to cartilage cascade. Whether different fragments released in disease will have different
in explant culture, as well as injected into the joint, will stimulate chondrocytes roles in complement activation remains to be shown.
to produce proteases and induce cartilage breakdown.65,66 In contrast, the In other experiments, it has been demonstrated that biglycan can also
intact fibronectin molecule will not have such effects on cells. It is possible bind the triple-helical stalk of the complement factor C1q and in doing so
that fragmentation of fibronectin is one mechanism for propagating joint inhibit activation of the classical pathway of complement. It appears that
destruction in disease. The active fragments have not yet been identified in other LRR proteins such as biglycan and decorin can also bind factors involved
body fluids. in regulating the complement cascade. One example is a tight interaction
Fibronectin is already upregulated in articular cartilage at very early stages with C4BP having a role in regulating complement activity.72
of OA in many species, including humans.13 The functional consequences Cartilage oligomeric matrix protein is a potent activator of the alternative
of this upregulation are not known but may represent part of an attempt at pathway via its C-terminal globular domain. Interestingly, this activity can
repair. A result is further availability of fibronectin, which may be fragmented be observed in synovial fluid from subjects with both OA and RA, as well
to further increase tissue degradation. as in blood in the form of COMP-C3b complexes created between the activator
COMP and the C3b fragment formed.73
Collagen has an additional cell receptor for discoidin domain receptor 2
(DDR-2). This receptor has been shown to be upregulated in mice in which The ECM of cartilage contains a number of factors that are sequestered and
OA is developing. DDR-2 will bind collagen type II, thereby inducing thereby bind to specific interaction partners. On degradation of the matrix,
upregulation of MMP-13 production.67 Exposure of collagen type II may be one can predict that these factors are released and affect cellular activities.
enhanced by removal of proteins bound at the fibrillar surface, thus making In particular, a number of growth factors have been found to have the
the molecule available for interactions. capacity to bind to matrix proteins.

Transforming growth factor-β

HYALURONAN Transforming growth factor-β has been shown to bind to a number of matrix
Hyaluronan is an extremely long glycosaminoglycan chain that is distinct proteins, and binding to some members of the LRR protein family has been
from all the others in that it is not bound to a protein core. It does not investigated in particular. All three TGF-β variants bind to bacterially expressed
contain any sulfate groups. The backbone is made up of one to several decorin, biglycan, and fibromodulin, as well as to these proteins with the
thousand repeating disaccharides of glucuronic acid and N-acetylglucosamine. full range of posttranslational modifications. More recently, there are indications
As discussed earlier, hyaluronan specifically binds to a domain in aggrecan that asporin can also bind TGF-β. It has been demonstrated that active
that is essential for the formation of aggregates. TGF-β is released from decorin on treatment with MMP-3. Indeed, cartilage
Hyaluronan also interacts with specific cell-surface receptors. These contains substantial amounts of TGF-β, which has been extracted and purified
receptors are CD44 molecules68 that interact with a minimally long hexasac- from the tissue.
charide sequence of the polymer. The interactions with hyaluronan are
important for organizing the pericellular environment and providing signals Fibroblast growth factor
to the cell. CD44 occurs in many different splice forms with variable presence Some matrix proteins contain heparan sulfate side chains, which are likely
on different cells, but a role for such variability in joint disease in unclear. to bind growth factors within the FGF family. In the case of syndecan at the
Another receptor for hyaluronan is RHAMM,69 also a signaling receptor and, cell surface, these side chains appear to be involved in presenting the growth
similar to CD44, it is present on a variety of cells. factor to its receptor.74


In addition to the heparin-binding domains present in fibronectin, various Perlecan was found to have a central role in cartilage development when a
other matrix proteins contain such domains. Examples are most members mouse with the gene inactivated was developed. Most of these mice die
of the thrombospondin family, with the exception of COMP, many MMPs, during early development as a result of problems with the heart and major
CHAPTER 2  The articular cartilage 15

blood vessels, but those that survive until birth show major growth distur- temperature have the highest level of chondrocyte viability. A study of 60
bances with an extensively altered growth plate lacking a large proportion patients with femoral condylar allografts showed 95% graft survival rate at
of the collagen fibers.75 Further studies have revealed that chondroitin sulfate 5 years and 85% at 10 years.87
side chains specific for perlecan can actually accelerate and catalyze collagen
GP-39 is a protein upregulated in OA. It shares homology with a chitinase,
but the true activity of the protein is not known.77 GP-39 is additionally
expressed in a number of other tissues, particularly in disease. This has become a substantial research area in the field of cartilage repair.
Cartilage matrix also contains a number of proteins that are made elsewhere Two surgical techniques are used, microfracture (MF) of the subchondral
and are normally found primarily in the circulation. There is a preference bone to recruit bone marrow mesenchymal stem cells to populate the cartilage
for certain proteins, and low-molecular-weight basic proteins (e.g., lysozyme) defect and a related technique, autologous chondrocyte implantation (ACI),
in particular appear to bind to the matrix.78 Whether these molecules contribute a two-part procedure in which chondrocytes are isolated from intact non-
specific functions is not known. weight-bearing regions of the joint, expanded and dedifferentiated in the
laboratory and redifferentiated and reimplanted into the osteochondral defect
in the weight-bearing region.88,89 The MF technique is considered the first-line
FRAGMENTS OF MATRIX PROTEINS RELEASED treatment given its minimally invasive nature, technical ease, limited surgical
DURING CARTILAGE BREAKDOWN AS morbidity, and relatively low costs.90 In a study with 7-year follow-up, 80%
of patients rated themselves as improved with patients younger than 35
MOLECULAR INDICATORS OF DISEASE years showing the most improvement.91 Biopsies after MF have noted that
In processes resulting in destruction of cartilage tissue, proteolytic enzymes approximately 10% had hyaline cartilage, with the majority having predomi-
degrade ECM proteins. Some of the fragments formed will no longer be nantly fibrocartilage.92 In general, patients demonstrated clear improvement
retained in the tissue but are released to surrounding body fluids. By using in knee function at 24 months after MF but inconclusive durability and
sensitive immunoassays, such fragments can be quantified in synovial fluid treatment failure beyond 5 years.
or serum. This so-called molecular marker technology (biomarkers) is intended Autologous matrix–induced chondrogenesis combines MF surgery with
to provide new means of assaying ongoing active processes in articular the application of a bilayer collagen membrane that physically stabilized the
cartilage. With further development, such techniques may be used in diagnosis, clot and may guide and enhance marrow-derived repair. In a study using a
for estimation of the activity of the tissue-destroying process, to document BST-CarGel (Piramal Like Sciences, Bio-Orthopaedic Division), treatment
effects of therapeutic intervention, and, most importantly, to discover processes with MF resulted in greater lesion filling and superior repair tissue quality
during the early stage before clinical symptoms become apparent. One example compared with MF treatment alone; however, clinical symptoms were
is the demonstration that COMP can be used to identify patients who will equivalent between groups.93
have the most extensive joint destruction in both OA and RA.33,79 Also, a Autologous chondrocyte implantation is approved for use in the United
number of collagen fragments created by cleavage with collagenases, as well States and is most useful for younger patients who have single defects
as by subsequent gelatinase activity, have been used to monitor disease.80,81 larger than 2 cm. Disadvantages include the need for two-stage procedures
With further development of the technology, we hope to be able to identify and an open arthrotomy, expense, and a significant rate of reoperation
indicators specific to a particular pathologic process in a given tissue. Thus, for graft hypertrophy, specifically with first-generation ACI treatments.
it should be possible to identify the activity of a process in the meniscus on Second- (porcine membrane) and third-generation (matrix-associated) ACI
the one hand and activity in cartilage on the other. treatments show promise but are not approved in the United States at this
As discussed earlier, some of the fragments released will activate immune time. In a series of more than 200 patients treated with ACI for larger
responses with the potential to further enhance disease activity. lesions, ACI provided durable outcomes with a survivorship of 71% at
10 years and improved function in 75% of patients.94 Magnetic resonance
imaging findings confirmed complete defect filling in half of patients at final
Cartilage ECM molecules have extremely long half-lives, and the cells of
cartilage are not renewed. In addition, it is difficult to imagine how newly
synthesized molecules would fully and functionally integrate into the
cartilage matrix. However, anecdotally, it is believed that some individuals The limited expansion capacity of de-differentiated chondrocytes and their
can repair cartilage to a limited extent, and there are animal models that increasing inefficiency at redifferentiation during extended culture has led
demonstrate genetic differences in the ability to repair cartilage.82 Although to the use of mesenchymal stem cells (MSCs) for autologous cartilage repair
the genetic differences are not known for sure, some growth factors involved of larger defects.95 MSCs are found in numerous human tissues, including
in development of cartilage appear to play a role, such as the Wnt family, or bone marrow, adipose tissue, articular cartilage, and synovial membrane.96
TGF-β. Cell-based therapy is becoming an established element of modern health
When cartilage is damaged, a number of repair strategies have been care and is predicted to grow as knowledge and implementation of cell
developed, including (1) replacing bad cartilage with good cartilage from biology, biomaterials, and regenerative medicine increases. MSCs are defined
the same joint, (2) using patient chondrocytes to make new ECM, (3) as adherent self-renewing, fibroblastoid-like cells that can differentiate into
attempting to recruit or implant patient stem cells, and (4) other strategies osteoblasts, adipocytes, and chondrocytes in vitro.97 These characteristics
that attempt to replace the cartilage matrix with an artificial matrix. We will are present in the MSCs derived from the bone marrow, adipose tissue, and
briefly review recent advances in these technologies. synovial joint tissues. These cells can be differentiated in vitro to chondrocytes
and used as chondrocytes or can be injected into the joint, where they may
differentiate into chondrocytes or, more likely, provide factors and immune
REPLACING CARTILAGE inhibitors that facilitate joint repair.98 Bone marrow MSCs were found to be
Intact cartilage from the same joint is used to replace damaged cartilage in at least as effective as chondrocytes for articular cartilage repair in improving
surgical techniques including mosaicplasty or surgery using morcellized symptoms of patients. However, chondrogenically induced bone marrow
cartilage. The primary surgical technique is called osteochondral autologous MSCs hold the inherent risk of either forming transient fibrocartilaginous
transplantation (OAT) mosaicplasty.83 This technique involves an open or repair tissue or undergoing terminal differentiation to form calcified cartilage,
arthroscopic transplantation of multiple cylindrical osteochondral grafts subchondral bone overgrowth, or intralesional osteophyte formation.99 Thus
from the relatively less weight-bearing periphery of the articular surface to alternatives to bone marrow MSCs are sought.
the cartilage defect, thus providing a hyaline cartilage–covered resurfacing.84 Adipose-derived MSCs, now called adipose-derived stem cells (ASCs),
OAT mosaicplasty has demonstrated acceptable long-term clinical outcomes, offer the advantage that they are abundant in adipose tissue or liposuction
given the appropriate indication for surgery, with a limitation being the samples.100 Two recent reviews summarize efforts to fabricate or regenerate
defect size.85 articular cartilage using scaffolds and growth factors with ASCs.101 Cells
In another surgical technique, osteochondral allograft transplantation is from the synovium, synovial fluid, periosteum, the infrapatellar fat pad,
also used. If the cartilage defect is too large for an autograft or a patient has trabecular bone, and muscle have also been used to differentiate chondrocytes,
failed a cartilage repair procedure, then a fresh osteochondral allograft (OCA) but no definitive positive regeneration has been documented. There is the
is a single-stage technique for large osteochondral defects, particularly in a possibility of recruiting these native cells from the tissues to the cartilage
setting of extensive subchondral bone loss.86 Chondrocyte viability seems defect, but this has not yet been productive. The area of cell-based strategies
to be very important in this technique, and fresh OCAs stored at physiologic for cartilage repair is thus under intensive study. However, the field has not
16 SECTION 1  Scientific Basis of Rheumatic Disease
moved as fast as was predicted, primarily because of safety, cost, efficacy, materials have been developed. For a review, please see Boushell
and regulatory hurdles.102 et al.103


A variety of combinations of MSCs with various scaffolds using biomaterials This chapter originates from the sixth edition of Rheumatology, amended
approaches, bioreactors, and combinations of cartilage- and bone-inducing with the section “Repair and regeneration of cartilage” by Linda J. Sandell.

1. Heinegård D, Aspberg A, Franzén A, et al. 17. Aspden RM, Hukins DWL. Stress in collagen fibrils of dimerization is essential for folding and stability of class
Non-collagenous glycoproteins in the extracellular articular cartilage calculated from their measured I small leucine-rich repeat proteoglycans. J Biol Chem.
matrix, with particular reference to cartilage and bone. orientations. Matrix. 1989;9:486-488. 2006;281:13324-13332.
In: Royce P, Steinmann B, eds. Connective tissue and its 18. Lohmander LS, Atley LM, Pietka TA, et al. The release 36. Kalamajski S, Aspberg A, Oldberg A. The decorin
heritable disorders: molecular, genetic, and medical aspects. of crosslinked peptides from type II collagen into human sequence SYIRIADTNIT binds collagen type I. J Biol
New York: Wiley-Liss; 2002:271-291. synovial fluid is increased soon after joint injury and in Chem. 2007;282:16062-16067.
1A.  Müller C, Khabut A, Dudhia J, et al. Quantitative osteoarthritis. Arthritis Rheum. 2003;48:3130-3139. 37. Danielson KG, Baribault H, Holmes DF, et al. Targeted
proteomics at different depths in human articular 19. Gregory KE, Oxford JT, Chen Y, et al. Structural disruption of decorin leads to abnormal collagen fibril
cartilage reveals unique patterns of protein distribution. organization of distinct domains within the morphology and skin fragility. J Cell Biol.
Matrix Biol. 2014;40:34-45. non-collagenous N-terminal region of collagen type XI. J 1997;136:729-743.
2. Önnerfjord P, Khabut A, Reinholt FP, et al. Quantitative Biol Chem. 2000;275:11498-11506. 38. Önnerfjord P, Heathfield TF, Heinegård D. Identification
proteomic analysis of eight cartilaginous tissues reveals 20. Eyre DR. Collagens and cartilage matrix homeostasis. of tyrosine sulfation in extracellular leucine-rich repeat
characteristic differences as well as similarities between Clin Orthop Relat Res. 2004;427(suppl):S118-S122. proteins using mass spectrometry. J Biol Chem.
subgroups. J Biol Chem. 2012;287:18913-18924. 21. Vikkula M, Mariman ECM, Lui VCH, et al. Autosomal 2004;279:26-33.
3. Stattin EL, Wiklund F, Lindblom K, et al. A missense dominant and recessive osteochondrodysplasias 39. Bengtsson E, Aspberg A, Heinegård D, et al. The
mutation in the aggrecan C–type lectin domain disrupts associated with the COL11A2 locus. Cell. amino-terminal part of PRELP binds to heparin and
extracellular matrix interactions and causes dominant 1995;80:431-437. heparan sulfate. J Biol Chem. 2000;275:40695-40702.
familial osteochondritis dissecans. Am J Hum Genet. 22. Vaughan L, Mendler M, Huber S, et al. D-periodic 40. Tillgren V, Önnerfjord P, Haglund L, et al. The tyrosine
2010;86:126-137. distribution of collagen type IX along cartilage fibrils. J sulfate–rich domains of the LRR proteins fibromodulin
4. Glasson SS, Askew R, Sheppard B, et al. Deletion of Cell Biol. 1988;106:991-997. and osteoadherin bind motifs of basic clusters in a
active ADAMTS5 prevents cartilage degradation in a 23. Eyre DR, Pietka T, Weis MA, et al. Covalent variety of heparin-binding proteins, including bioactive
murine model of osteoarthritis. Nature. cross-linking of the NC1 domain of collagen type IX to factors. J Biol Chem. 2009;284:28543-28553.
2005;434:644-648. collagen type II in cartilage. J Biol Chem. 41. Heathfield TF, Önnerfjord P, Dahlberg L, et al. Cleavage
5. Stanton H, Rogerson FM, East CJ, et al. ADAMTS5 is 2004;279:2568-2574. of fibromodulin in cartilage explants involves removal of
the major aggrecanase in mouse cartilage in vivo and in 24. Fässler R, Schnegelsberg PNJ, Dausman J, et al. Mice the N-terminal tyrosine sulfate-rich region by proteolysis
vitro. Nature. 2005;434:648-652. lacking a1(IX) collagen develop noninflammatory at a site that is sensitive to matrix metalloproteinase-13.
6. Sandy JD. A contentious issue finds some clarity: on the degenerative joint disease. Proc Natl Acad Sci U S A. J Biol Chem. 2004;279:6286-6295.
independent and complementary roles of aggrecanase 1994;91:5070-5074. 42. Danfelter M, Önnerfjord P, Heinegård D. Fragmentation
activity and MMP activity in human joint aggrecanolysis. 25. Hedlund H, Mengarelli-Widholm S, Reinholt FP, et al. of proteins in cartilage treated with interleukin-1:
Osteoarthritis Cartilage. 2006;14:95-100. Stereologic studies on collagen in bovine articular specific cleavage of type IX collagen by matrix
7. Hughes CE, Caterson B, Fosang AJ, et al. Monoclonal cartilage. APMIS. 1993;101:133-140. metalloproteinase 13 releases the NC4 domain. J Biol
antibodies that specifically recognize neoepitope 26. Maroudas A, Palla G, Gilav E. Racemization of aspartic Chem. 2007;282:36933-36941.
sequences generated by “aggrecanase” and matrix acid in human articular cartilage. Connect Tissue Res. 43. Bengtsson E, Mörgelin M, Sasaki T, et al. The leucine
metalloproteinase cleavage of aggrecan: application to 1992;28:161-169. rich repeat protein PRELP binds perlecan and collagens
catabolism in situ and in vitro. Biochem J. 27. Oldberg Å, Antonsson P, Lindblom K, et al. COMP and may function as a basement membrane anchor. J
1995;305:799-804. (cartilage oligomeric matrix protein) is structurally Biol Chem. 2002;277:15061-15068.
8. Platt D, Bird JL, Bayliss MT. Ageing of equine articular related to the thrombospondins. J Biol Chem. 44. Rucci N, Rufo A, Alamanou MC, et al. The
cartilage: structure and composition of aggrecan and 1992;267:22346-22350. glycosaminoglycan-binding domain of PRELP acts as a
decorin. Equine Vet J. 1998;30:43-52. 28. Kvansakul M, Adams JC, Hohenester E. Structure of a cell type-specific NF-kappaB inhibitor that impairs
9. Maroudas A, Bayliss MT, Uchitel-Kaushansky N, et al. thrombospondin C-terminal fragment reveals a novel osteoclastogenesis. J Cell Biol. 2009;187:669-683.
Aggrecan turnover in human articular cartilage: use of calcium core in the type 3 repeats. EMBO J. 45. Lorenzo P, Aspberg A, Önnerfjord P, et al. Identification
aspartic acid racemization as a marker of molecular age. 2004;23:1223-1233. and characterization of asporin—a novel member of the
Arch Biochem Biophys. 1998;350:61-71. 29. Holmes DF, Gilpin CJ, Baldock C, et al. Corneal collagen leucine rich repeat protein family closely related to
10. Westling J, Fosang AJ, Last K, et al. ADAMTS4 cleaves fibril structure in three dimensions: structural insights decorin and biglycan. J Biol Chem.
at the aggrecanase site (Glu373-Ala374) and secondarily into fibril assembly, mechanical properties, and tissue 2001;276:12201-12211.
at the matrix metalloproteinase site (Asn341-Phe342) in organization. Proc Natl Acad Sci U S A. 46. Kizawa H, Kou I, Iida A, et al. An aspartic acid repeat
the aggrecan interglobular domain. J Biol Chem. 2001;98:7307-7312. polymorphism in asporin inhibits chondrogenesis and
2002;277:16059-16066. 30. Shen Z, Heinegård D, Sommarin Y. Distribution and increases susceptibility to osteoarthritis. Nat Genet.
11. Thomas L. Reversible collapse of rabbit ears after expression of cartilage oligomeric matrix protein and 2005;37:138-144.
intravenous papain, and prevention of recovery by bone sialoprotein show marked changes during rat 47. Mustafa Z, Dowling B, Chapman K, et al. Investigating
cortisone. J Exp Med. 1956;104:245-252. femoral head development. Matrix Biol. 1995;14:773- the aspartic acid (D) repeat of asporin as a risk factor
12. Hardingham TE, Fitton-Jackson S, Muir H. Replacement 781. for osteoarthritis in a UK Caucasian population. Arthritis
of proteoglycans in embryonic chick cartilage in organ 31. Briggs MD, Chapman KL. Pseudoachondroplasia and Rheum. 2005;52:3502-3506.
culture after treatment with testicular hyaluronidase. multiple epiphyseal dysplasia: mutation review, 48. Kalamajski S, Aspberg A, Lindblom K, et al. Asporin
Biochem J. 1972;129:101-112. molecular interactions, and genotype to phenotype competes with decorin for collagen binding, binds
13. Lorenzo P, Bayliss MT, Heinegård D. Altered patterns correlations. Hum Mutat. 2002;19:465-478. calcium and promotes osteoblast collagen
and synthesis of extracellular matrix macromolecules in 32. Svensson L, Aszodi A, Heinegård D, et al. Cartilage mineralization. Biochem J. 2009;423:53-59.
early osteoarthritis. Matrix Biol. 2004;23:381-391. oligomeric matrix protein–deficient mice have normal 49. Couchman JR. Syndecans: proteoglycan regulators of
14. Stubendorff JJ, Lammentausta E, Struglics A, et al. Is skeletal development. Mol Cell Biol. 2002;22:4366-4371. cell-surface microdomains? Nat Rev Mol Cell Biol.
cartilage sGAG content related to early changes in 33. Saxne T, Månsson B, Heinegård D. Biomarkers for 2003;4:926-937.
cartilage disease? Implications for interpretation of cartilage and bone in rheumatoid arthritis. In: Firestein 50. Bernfield M, Gotte M, Park PW, et al. Functions of cell
dGEMRIC. Osteoarthritis Cartilage. 2012;20:396- G, Panayi G, Wollheim F, eds. Rheumatoid arthritis: new surface heparan sulfate proteoglycans. Annu Rev
404. frontiers in pathogenesis and treatment. Oxford: Oxford Biochem. 1999;68:729-777.
15. Blaschke UK, Eikenberry EF, Hulmes DJ, et al. Collagen University Press; 2006:301-313. 51. Lamande SR, Shields KA, Kornberg AJ, et al. Bethlem
XI nucleates self-assembly and limits lateral growth 34. Scott PG, McEwan PA, Dodd CM, et al. Crystal structure myopathy and engineered collagen VI triple helical
of cartilage fibrils. J Biol Chem. 2000;275:10370- of the dimeric protein core of decorin, the archetypal deletions prevent intracellular multimer assembly
10378. small leucine-rich repeat proteoglycan. Proc Natl Acad and protein secretion. J Biol Chem. 1999;274:21817-
16. Holmes DF, Kadler KE. The 10+4 microfibril structure of Sci U S A. 2004;101:15633-15638. 21822.
thin cartilage fibrils. Proc Natl Acad Sci U S A. 35. Scott PG, Dodd CM, Bergmann EM, et al. Crystal 52. Wiberg C, Hedbom E, Khairullina A, et al. Biglycan and
2006;103:17249-17254. structure of the biglycan dimer and evidence that decorin bind close to the N-terminal region of the
CHAPTER 2  The articular cartilage 17
collagen VI triple helix. J Biol Chem. 2001;276: interpretation of redundancy. Proc Natl Acad Sci U S A. articular cartilage defects in the knee. Knee.
18947-18952. 2004;101:18081-18086. 2013;20(4):287-290.
53. Wiberg C, Heinegård D, Wenglén C, et al. Biglycan 70. Holmdahl R. Dissection of the genetic complexity of 86. Gortz S, Bugbee WD. Allografts in articular cartilage
organizes collagen VI into hexagonal-like networks arthritis using animal models. Immunol Lett. repair. Instr Course Lect. 2007;56:469-480.
resembling tissue structures. J Biol Chem. 2006;103:86-91. 87. Gross AE, Shasha N, Aubin P. Long-term followup of the
2002;277:49120-49126. 71. Sjöberg A, Önnerfjord P, Mörgelin M, et al. The use of fresh osteochondral allografts for posttraumatic
54. Wiberg C, Klatt AR, Wagener R, et al. Complexes of extracellular matrix and inflammation: fibromodulin knee defects. Clin Orthop Relat Res. 2005;435:79-
matrilin-1 and biglycan or decorin connect collagen VI activates the classical pathway of complement by 87.
microfibrils to both collagen II and aggrecan. J Biol directly binding C1q. J Biol Chem. 88. Grande DA, Pitman MI, Peterson L, et al. The repair of
Chem. 2003;278:37698-37704. 2005;280:32301-32308. experimentally produced defects in rabbit articular
55. Klatt AR, Becker AK, Neacsu CD, et al. The matrilins: 72. Happonen KE, Sjöberg AP, Mörgelin M, et al. cartilage by autologous chondrocyte transplantation. J
modulators of extracellular matrix assembly. Int J Complement inhibitor C4b-binding protein interacts Orthop Res. 1989;7(2):208-218.
Biochem Cell Biol. 2011;43:320-330. directly with small glycoproteins of the extracellular 89. Brittberg M, Lindahl A, Nilsson A, et al. Treatment of
56. Suleman F, Gualeni B, Gregson HJ, et al. A novel form of matrix. J Immunol. 2009;182:1518-1525. deep cartilage defects in the knee with autologous
chondrocyte stress is triggered by a COMP mutation 73. Happonen KE, Saxne T, Geborek P, et al. Serum chondrocyte transplantation. N Engl J Med.
causing pseudoachondroplasia. Hum Mutat. COMP-C3b complexes in rheumatic diseases and 1994;331(14):889-895.
2012;33:218-231. relation to anti-TNF-alpha treatment. Arthritis Res Ther. 90. Colby SM, Hintermeister RA, Torry MR, et al. Lower
57. Wu JJ, Eyre DR. Matrilin-3 forms disulfide-linked 2012;14:R15. limb stability with ACL impairment. J Orthop Sports
oligomers with matrilin-1 in bovine epiphyseal cartilage. 74. Allen BL, Filla MS, Rapraeger AC. Role of heparan Phys Ther. 1999;29(8):444-451, discussion 452-454.
J Biol Chem. 1998;273:17433-17438. sulfate as a tissue-specific regulator of FGF-4 and FGF 91. Sterett WI, Hutton KS, Briggs KK, et al. Decreased range
58. Millward-Sadler SJ, Salter DM. Integrin-dependent signal receptor recognition. J Cell Biol. 2001;155:845- of motion following acute versus chronic anterior
cascades in chondrocyte mechanotransduction. Ann 858. cruciate ligament reconstruction. Orthopedics.
Biomed Eng. 2004;32:435-446. 75. Gustafsson E, Aszodi A, Ortega N, et al. Role of collagen 2003;26(2):151-154.
59. Humphries MJ, Travis MA, Clark K, et al. Mechanisms type II and perlecan in skeletal development. Ann N Y 92. McCarthy A, Kerr M, Abadie E, et al. Report on the
of integration of cells and extracellular matrices by Acad Sci. 2003;995:140-150. joint EFPIA, DIA and EMEA pharmacogenetics
integrins. Biochem Soc Trans. 2004;32:822-825. 76. Kvist A, Johnson A, Mörgelin M, et al. Chondroitin workshop: moving toward clinical application.
60. Haglund L, Tillgren V, Addis L, et al. Identification and sulfate perlecan enhances collagen fibril formation. Pharmacogenomics. 2004;5(6):731-739.
characterization of the integrin alpha2beta1 binding Implications for perlecan chondrodysplasias. J Biol 93. Stanish WD, McCormack R, Forriol F, et al. Novel
motif in chondroadherin mediating cell attachment. J Chem. 2006;281:33127-33139. scaffold-based BST-CarGel treatment results in superior
Biol Chem. 2011;286:3925-3934. 77. Recklies AD, White C, Ling H. The chitinase 3–like cartilage repair compared with microfracture in a
61. Haglund L, Tillgren V, Önnerfjord P, et al. The protein human cartilage glycoprotein 39 (HC-gp39) randomized controlled trial. J Bone Joint Surg Am.
C-terminal peptide of chondroadherin modulates stimulates proliferation of human connective-tissue cells 2013;95(18):1640-1650.
cellular activity by selectively binding to heparan sulfate and activates both extracellular signal–regulated 94. Minas T, Von Keudell A, Bryant T, et al. The John Insall
chains. J Biol Chem. 2013;288:995-1008. kinase– and protein kinase B–mediated signalling Award: A minimum 10-year outcome study of
62. Pankov R, Yamada KM. Fibronectin at a glance. J Cell pathways. Biochem J. 2002;365:119-126. autologous chondrocyte implantation. Clin Orthop Relat
Sci. 2002;115:3861-3863. 78. Thonar EJ, Feist SB, Fassbender K, et al. Quantification Res. 2014;472(1):41-51.
63. Beauvais A, Erickson CA, Goins T, et al. Changes in the of hen egg white lysozyme in cartilage by an 95. Wakitani S, Goto T, Pineda SJ, et al. Mesenchymal
fibronectin-specific integrin expression pattern modify enzyme-linked immunosorbent assay. Connect Tissue Res. cell-based repair of large, full-thickness defects of
the migratory behavior of sarcoma S180 cells in vitro 1988;17:181-197. articular cartilage. J Bone Joint Surg Am.
and in the embryonic environment. J Cell Biol. 79. Sharif M, Kirwan JR, Elson CJ, et al. Suggestion of 1994;76(4):579-592.
1995;128:699-713. nonlinear or phasic progression of knee osteoarthritis 96. Fellows CR, Matta C, Zakany R, et al. Adipose, bone
64. Woods A, Longley RL, Tumova S, et al. Syndecan-4 based on measurements of serum cartilage oligomeric marrow and synovial joint-derived mesenchymal stem
binding to the high affinity heparin-binding domain of matrix protein levels over five years. Arthritis Rheum. cells for cartilage repair. Front Genet. 2016;7:213.
fibronectin drives focal adhesion formation in 2004;50:2479-2488. 97. Barry FP, Murphy JM. Mesenchymal stem cells: clinical
fibroblasts. Arch Biochem Biophys. 2000;374:66-72. 80. Garnero P, Charni N, Juillet F, et al. Increased urinary applications and biological characterization. Int J
65. Homandberg GA, Kang Y, Zhang J, et al. A single type II collagen helical and C telopeptide levels are Biochem Cell Biol. 2004;36(4):568-584.
injection of fibronectin fragments into rabbit knee joints independently associated with a rapidly destructive hip 98. Barry F, Murphy M. Mesenchymal stem cells in joint
enhances catabolism in the articular cartilage followed osteoarthritis. Ann Rheum Dis. 2006;65:1639-1644. disease and repair. Nat Rev Rheumatol.
by reparative responses but also induces systemic effects 81. Ameye LG, Deberg M, Oliveira M, et al. The chemical 2013;9(10):584-594.
in the non-injected knee joints. Osteoarthritis Cartilage. biomarkers C2C, Coll2-1, and Coll2-1NO2 provide 99. Beitzel K, McCarthy MB, Cote MP, et al. Rapid isolation
2001;9:673-683. complementary information on type II collagen of human stem cells (connective progenitor cells) from
66. Johnson A, Smith R, Saxne T, et al. Fibronectin catabolism in healthy and osteoarthritic mice. Arthritis the distal femur during arthroscopic knee surgery.
fragments cause release and degradation of Rheum. 2007;56:3336-3346. Arthroscopy. 2012;28(1):74-84.
collagen-binding molecules from equine explant 82. Rai MF, Patra D, Sandell LJ, et al. Transcriptome analysis 100. Gimble J, Guilak F. Adipose-derived adult stem cells:
cultures. Osteoarthritis Cartilage. 2004;12:149-159. of human injured meniscus reveals a distinct phenotype isolation, characterization, and differentiation potential.
67. Xu L, Peng H, Wu D, et al. Activation of the discoidin of meniscus degeneration with aging. Arthritis Rheum. Cytotherapy. 2003;5(5):362-369.
domain receptor 2 induces expression of matrix 2013;65(8). 101. Bielli A, Scioli MG, Gentile P, et al. Adipose-derived
metalloproteinase 13 associated with osteoarthritis in 83. Hangody L, Rathonyi GK, Duska Z, et al. Autologous stem cells in cartilage regeneration: current perspectives.
mice. J Biol Chem. 2005;280:548-555. osteochondral mosaicplasty. Surgical technique. J Bone Regen Med. 2016;11(7):693-703.
68. Knudson CB, Knudson W, Smith RL. Hyaluronan and Joint Surg Am. 2004;86-A(suppl 1):65-72. 102. Bara JJ, Herrmann M, Evans CH, et al. Improving
CD44: modulators of chondrocyte metabolism. Clin 84. Barber FA, Chow JC. Arthroscopic osteochondral translation success of cell-based therapies in
Orthop Relat Res. 2004;427(suppl):S152-S162. transplantation: histologic results. Arthroscopy. orthopaedics. J Orthop Res. 2016;34(1):17-21.
69. Nedvetzki S, Gonen E, Assayag N, et al. RHAMM, a 2001;17(8):832-835. 103. Boushell MK, Hung CT, Hunziker EB, et al. Current
receptor for hyaluronan-mediated motility, compensates 85. Solheim E, Hegna J, Oyen J, et al. Results at 10 to 14 strategies for integrative cartilage repair. Connect Tissue
for CD44 in inflamed CD44-knockout mice: a different years after osteochondral autografting (mosaicplasty) in Res. 2016;1-14.
Bone structure and function
David B. Burr • Teresita Bellido • Kenneth E. White

Key Points growth, the primary spongiosa is composed mostly of disorganized woven
■ Bone is organized differently and for different functions at the organ, tissue, and
bone, or primary lamellar bone surrounding a core of calcified cartilage. It
molecular levels.
is separated from the remodeled, more highly oriented secondary spongiosa
by an arbitrary boundary. The secondary spongiosa reflects patterns of stress
■ Noncollagenous proteins function as structural support in the bone matrix and
and functions largely to funnel stresses to the stronger and more massive
regulate mineral deposition, crystal growth, and cell attachment.
cortical bone. In regions beneath joint surfaces, it also attenuates forces
■ Bone mineralization is effected through hormonal action (fibroblast growth factor 23
generated by mechanical loading and may protect the joint surface from
[FGF-23]) as well as through the bioactivity of intracellular and secreted matrix
loading-related trauma. The cancellous bone itself is composed of plates and
proteins with emerging roles (dentin matrix protein-1, FAM20c, and matrix
rods of bone, each about 200 µm thick, with a porosity of about 75% to
extracellular phosphoglycoprotein).
85%. The marrow in the spaces between the trabecular struts are regions in
■ Growth factors, including insulin-like growth factor 1, the autocrine/paracrine FGFs, which red blood cells are formed (red marrow). In osteoporosis, the differentia-
and transforming growth factor-β, play key roles in bone growth and structure. tion of cells in the bone lineage can be partly diverted to adipocytes, and
■ Osteoclasts are the primary bone-resorptive cells and originate from precursors of the the marrow becomes more fatty (yellow marrow). Because cancellous bone
hematopoietic lineage upon stimulation with receptor activator of nuclear factor-κB has a large surface area in contact with vascular marrow, it is ideal for the
ligand (RANKL) and macrophage colony-stimulating factor. long-term exchange of calcium ions. In osteopenia, regions of cancellous
■ Whereas the rate of osteoclast generation determines the extension of the bone are affected first. This is why the vertebral column, which is mainly
bone-remodeling unit, the life span of osteoclasts determines the depth of composed of cancellous bone, is affected earlier and more severely in
resorption. osteoporosis than even the femoral neck and hip.
■ Osteoblasts are responsible for bone formation and originate from mesenchymal
progenitors that also give rise to chondrocytes, muscle cells, and adipocytes.
■ Osteocytes form a network that senses mechanical and hormonal environmental cues
and orchestrates the function of osteoblasts and osteoclasts. At the microstructural level, bone is organized differently for different functions
■ Osteocytes produce and secrete factors (sclerostin/Sost, RANKL, osteoprotegerin) (Fig. 3.2). In humans, bone tissue can be divided into three broad categories
that affect other bone cells by paracrine/autocrine mechanisms and secrete based partly on the arrangement of the collagen fibers and partly on whether
hormones (FGF-23) that affect other tissues by endocrine mechanisms. it has replaced preexisting bone: (1) woven bone, (2) primary bone, and (3)
■ Four dynamic processes—growth, modeling, remodeling, and repair—control skeletal secondary bone.
development and adaptation, defined by the relationship of bone resorption and Woven bone is characterized by randomly oriented collagen fibers, which
bone formation to each other. tend to be smaller in diameter than those in more highly organized primary
or secondary bone. Woven bone is not lamellar, is very porous and may
become highly mineralized. Woven bone can be deposited de novo without
any bony or cartilaginous substrate or anlage, but it can also be formed as
INTRODUCTION part of the process of endochondral ossification, either at the growth plate
during development or during fracture repair. Woven bone proliferates rapidly
Bone is a complex natural composite material that undergoes millions of because it has a large cell-volume ratio, which makes its role in fracture
loading cycles during a lifetime without failure. Its structure is hierarchical, repair ideal. It is often found associated with osteosarcoma, probably because
being organized differently at the organ (whole bone), tissue (material), and of proliferation especially of the cellular periosteum. The function of woven
molecular (collagen–mineral) levels. bone is primarily mechanical, rapidly providing both temporary strength
The mechanical functions of bone are the most widely recognized and and scaffolding upon which lamellar bone may be deposited. However, it
are often described in terms of strength and stiffness. Although strength and can also be associated with pathologic processes that involve inflammatory
stiffness are important, bone is particularly effective at dissipating energy cytokines.
that could cause a fracture over repeated cycles of loading. Beyond its Primary bone must be deposited on a preexisting substrate and is organized
mechanical function, the marrow cavity and the porous trabecular bone in into lamellar layers. Because of this, trabecular plates, which are composed
the ends of long bones and in the vertebrae and iliac crest are regions in mainly of primary lamellae, cannot be replaced after they are perforated.
which red blood cells are formed and stored. Bone is in fact a primary This accounts for the loss of trabeculae with age and is part of the reason
blood-storing organ. Additionally, bone is the body’s primary storehouse of that it is difficult to reverse osteoporotic changes after trabecular connectivity
calcium and phosphorus; 99% of the body’s calcium is stored in bone. These has been lost. Primary lamellar bone also forms in rings around the endocortical
minerals are necessary for the proper function of a variety of systems in the and periosteal surfaces of whole bone (circumferential lamellae). Primary
body and are essential for enzyme reactions, blood clotting, the proper function lamellar bone itself is not very vascular and can become very dense. Primary
of contractile proteins in muscles, and the transmission of nerve impulses. bone can also be arranged in concentric rings around a central canal—much
like a secondary haversian system, except without a definable cement line.
These primary osteons tend to be small with few concentric lamellae. In
ORGANIZATION OF BONE reality, primary osteons are modified vascular channels that have “filled in”
by the addition of lamellae to the surface of the vascular space.
MACROSCOPIC (ORGAN) LEVEL Secondary bone is the product of the resorption of preexisting bone and
Bone at the organ level consists of the diaphysis (shaft), the metaphysis, and its replacement with new bone. This can occur within dense cortical bone
the epiphysis (Fig. 3.1). In the long bones of growing children, the growth (resulting in a secondary osteon, or haversian system) or begin on the surfaces
plate separates the epiphysis from the metaphysis. The primary component of trabeculae (sometimes called a hemi-osteon). The distinction between
of the diaphysis is cortical or compact bone. The haversian canals in cortical primary and secondary bone is important because it is likely that control of
bone create a porosity of about 3% to 5%, although this increases in older primary bone apposition is different from replacement of preexisting bone
age and with osteoporotic changes to the skeleton. Compact bone is also by secondary bone. A secondary haversian system has a central vascular
found surrounding the spongy bone of the vertebral body and in the skull. canal 50 to 90 µm in diameter. The blood vessels in the canal have the
It is very strong and provides both support and protection. characteristics of capillaries and are generally paired within the canal. Venous
Cancellous (trabecular or spongy) bone is found in the metaphyses of sinusoids and lymphatic vessels are not found in haversian canals, although
the long bones and in the vertebrae, surrounded by cortical bone. During it has been suggested that prelymphatic vessels may exist.1 The vessel walls

CHAPTER 3  Bone structure and function 19

contain no smooth muscle but are fenestrated capillaries lined by an incomplete bone formation (canal diameter). The secondary osteon is bounded by a
layer of endothelial cells, similar to vessels in other blood-forming organs cement line, probably composed largely of sulfated glycosaminoglycans.4
like the spleen and bone marrow.2 The vessel is accompanied about 60% of The cement line forms an effective boundary that can arrest cracks in bone
the time by two to seven unmyelinated or poorly myelinated nerve fibers.3 and stop them from growing to a critical size.
Because the capillaries have no smooth muscle, these nerves do not serve a
vasomotor function but are primary sensory nerves or autonomic nerves
from the sympathetic nervous system. Both primary afferents and sympathetic
nerve fibers express neuropeptides that may play some role in regulating Bone is composed of organic and mineralized components, mainly consisting
bone remodeling. Afferent nerve fibers also express GAP43, which is associated of a matrix of cross-linked type I collagen mineralized with nanocrystalline,
with axonal growth, and may reflect the need for vessel growth and rein- carbonated apatite. Bone matrix incorporates a small fraction of noncollagenous
nervation during bone remodeling. proteins that serve to control collagen assembly and size as well as the
Lining cells (resting osteoblasts) cover the haversian canal, which is process of mineralization and cell attachment. The mineral comprises about
surrounded by a series of concentric lamellae containing bone cells—osteocytes. 65% to 67% of the bone by weight, the organic component about 22% to
The relationship between the size of the osteon and the size of its canal is 25%, and water the remaining fraction (≈10%). Within the organic fraction,
a measure of the balance between bone resorption (osteon diameter) and type 1 collagen makes up about 90%, with the remainder accounted for by
several minor collagens (types III and V) and a variety of noncollagenous
proteins, most of them extracellular, with cell protein accounting for about
15% of the noncollagenous proteins in bone.
Epiphysis Type I collagen in bone is formed by a triple helix composed of two α1
chains and a single α2 chain. At either end of the collagen molecule are an
N-telopeptide and a C-telopeptide, which can be cleaved when bone is resorbed
and which are used to measure bone resorption biochemically. These individual
Epiphyseal plate triple helical collagen molecules self-associate into a periodic arrangement
of parallel molecules spaced in quarter-staggered array at distances of 40 nm
between their ends to form collagen fibrils (Fig. e3.1). The holes plus the
overlap zones give the collagen its banded appearance with a D-periodic
Longitudinal Metaphysis
spacing of 67 nm on average. The spaces between the ends of the collagen
growth Cancellous bone molecules are known as hole zones, and the 35-nm gap zones that run
longitudinally parallel between the molecules are known as pores. Hydroxy-
apatite [Ca10(PO4)6(OH)2], which has small and poorly formed crystals,
Osteoclastic nucleates within these spaces. However, most of the mineral in bone, about
resorption 75%, is in the form of more highly carbonated apatite that forms plates about
300 nm long that lie outside the collagen fibril (extrafibrillar).5,6 The initial
deposition of mineral—or primary mineralization—occurs rapidly after new
Diaphysis bone is deposited, with about two thirds of the eventual mineral content
achieved within about 3 weeks.7 As the bone tissue matures, the mineral
Osteoblastic Cortical bone crystals grow, become more platelike, and orient themselves parallel to one
formation another and to the collagen fibrils. This process of crystal growth is sometimes
called secondary mineralization and occurs over the next year or so until full
FIG. 3.1  Bone consists of the diaphysis (shaft), the metaphysis, and the epiphysis. mineralization is achieved. In cancellous bone, the c-axis or long axis of the
In growing children, the growth (epiphyseal) plate separates the epiphysis from the mineral crystal aligns with the longitudinal axis of the trabeculae; in cortical
metaphysis. During growth, the periosteal surface of the metaphysis must be constantly bone, the mineral orients to the long axis of the osteon. Other elements can
resorbed, while concurrent bone formation occurs on the endocortical surface to easily substitute into the hydroxyapatite crystal, changing its character and
convert the thin-walled but flared metaphysis into the narrower but thicker-walled its ability to withstand mechanical loads. Collagen fibers can become cross-
diaphysis (McNeal tetrachrome stain). linked through the action of enzymes, which results in immature or reducible


FIG. 3.2  The three general microscopic types of bone, defined

morphologically. Woven bone is composed of disorganized, randomly
oriented collagen fibers. It is found at sites of fracture or inflammation.
Lamellar bone can be divided into primary and secondary. Primary bone Primary
is deposited in layers by direct apposition on a substrate. It is found lamellar
circling the endocortical and periosteal circumferences of a long bone bone
and within trabeculae. Secondary bone is formed by replacement of
primary bone through the process of resorption and subsequent new
bone formation. (Woven bone: polarized light. Primary lamellar bone:
left, basic fuchsin; right, polarized light. Secondary osteonal bone:
left, toluidine blue; right, scanning electron microscopic image.)

CHAPTER 3  Bone structure and function 19.e1


67 nm Periodic staggered structure

Hole zones Mineral associated with gaps in stagger

“Young Bone”:
Cross-links associated with
N and C end terminals of
Enzymatic cross-links

Pore zones

“Old Bone”:
Non–end terminal
associated cross-links

FIG. E3.1  In the collagen fibril, molecules are organized in a quarter-staggered array,
with hole and pore zones between them where mineral can deposit. The hole zones
give the collagen fibril its banded appearance when viewed by electron microscopy.
Cross-links are formed either through the action of enzymes (divalent and trivalent
cross-links) or through nonenzymatic processes, the latter forming advanced glycation
end products that can make bone behave in brittle fashion.
20 SECTION 1  Scientific Basis of Rheumatic Disease
(divalent) cross-links that mature through the action of lysyl oxidase to mouse) also suggest that mature versican PG may play a role in directing
irreducible (trivalent) cross-links such as pyridinoline and deoxypyridinoline. limb chondrocyte aggregation depending on the surrounding extracellular
Because divalent cross-links are rapidly converted into mature trivalent matrix (ECM).12
cross-links under normal circumstances, the quantity of divalent cross-links 2. Heparan sulfate PGs (HSPGs) are produced by osteoblast and osteoclast
in bone is an indirect measure of remodeling rate. Divalent and trivalent lineage cells. These molecules play important roles in cell–cell interactions
cross-links are associated with the N- and C-terminals of the collagen molecule. during bone formation by trapping autocrine and paracrine heparin-binding
Cross-links can also be formed between the fibers without the action of fibroblast growth factor (FGF) family members, as well as acting as
enzymes. These cross-links are formed by the condensation of arginine, coreceptors with the FGF receptors. Also, other secreted molecules bind
lysine, and free sugars or through various oxidation reactions, which results HSPGs, such as transforming growth factor-β (TGF-β, betaglycan) and
in the formation of advanced glycation end products. Accumulation of osteoprotegerin (OPG) (syndecans). The bioactivity of these factors is
advanced glycation end products occurs in diabetes, making the bone more modulated by HSPGs, potentially through focusing of concentrations of
brittle and increasing the risk of fracture. these potent molecules near differentiating cells.
The skeletal noncollagenous proteins comprise 10% to 15% of total bone 3. Small leucine-rich PGs are the most abundant of the PGs in bone matrix
protein. These molecules constitute a wide range of polypeptide species and and include decorin, biglycan, fibromodulin, lumican, and osteoadherin.
have roles in multiple skeletal processes in addition to functioning as structural These molecules help to provide the structural organization of the bone
scaffolding within the bone matrix. One important feature of bone that has matrix and interact with specific growth factors and collagen to increase
been largely overlooked until recently is water, which comprises about 10% factor concentration and bioactivity in the matrix (Fig. 3.3). The localization
of bone volume and can be found in either free-flowing or bound forms. of these proteins in mature bone varies; whereas decorin may localize
Water stabilizes collagen structure8 but also allows sliding at the interface with specific matrix areas, biglycan is evenly distributed throughout the
between the mineral and the collagen, which increases the ductility and matrix. Whereas decorin-null animals have alterations in collagen fibril
toughness of bone.9 Because proteoglycans are hydrophilic, they may be size and bone shape,13 biglycan-null animals demonstrate reduced bone
important in retaining the water in some compartments. mass because of lower osteoblast numbers and also show reduced numbers
of osteoclasts.14 Small leucine-rich PGs play an essential role in the regula-
tion of growth factor activity. In this regard, decorin, biglycan, and
PROTEOGLYCANS fibromodulin all possess the ability to bind to TGF-β; however, decorin
Proteoglycans (PGs) are a ubiquitous family of molecules composed of a is the best characterized of these proteins for the ability to bind this
core protein and one or several covalently attached sulfated glycosaminoglycan factor. Decorin enhances TGF-β binding with its cognate receptors and
chains. The glycosaminoglycans are linear polymers of repeated disaccharide enhances its bioactivity and, in concert, may act to sequester TGF-β in
units of hexosamine and hexuronic acid, except for keratan sulfate, in which the collagen fibrils, thus reducing its activity. TGF-β activity is associated
hexuronic acid is replaced by galactose. The core proteins attached to the with increased apoptosis of osteoprogenitors; therefore, biglycan and
glycosaminoglycans are a diverse protein group and range in size from 10 decorin appear to be essential for maintaining mature osteoblast numbers
to 500 kDa. The wide variety of protein structure may aid in directing the through regulation of the proliferation and survival of bone marrow
unique functional roles of each PG family. progenitor cells.
The bone matrix contains PG families of several primary structures,
including the following:
1. Hyaluronan/CD44, chondroitin sulfate–containing PGs are expressed in
several regions of bone. Hyaluronan is expressed in focal regions within Osteocalcin (OC) is a polypeptide posttranslationally modified to carry
periosteum and endosteum and surrounding most of the major bone cell dicarboxylic glutamyl (Gla) residues, which relies on vitamin K for proper
types including osteoblasts, osteoprogenitor cells, osteoclasts, and osteocyte production (another identifier for OC is bone γ-carboxyglutamic acid [Gla]
lacunae. CD44 is a cell-surface hyaluronan receptor that may play roles protein [BGLAP or BGP]). In humans, vitamin K is primarily a cofactor in
in guiding bone development and has been localized to osteoclasts, the enzymatic reaction that converts glutamate residues into γ-carboxyglutamate
osteocytes, and bone marrow cells.10 Versican, a chondroitin sulfate– residues in these vitamin K–dependent proteins including OC but also in
containing PG, may be enriched during early osteoid formation and may proteins involved in blood clotting such as factor IX. These Gla-containing
provide a temporary framework in newly formed cartilage matrix during motifs are thought to enhance calcium binding, which may function to
bone development.11 Results from cultured mesenchymal cells derived control mineral deposition and bone remodeling. A nine-residue domain
from a mouse model carrying a disrupted versican gene (the hdf transgenic proximal to the N-terminal of secreted OC shares high homology with the


FGF PGs increase local FIG. 3.3  Proteoglycans (PGs) regulate growth factors
growth factor at several levels in the bone matrix and in bone cells.
OPG Biglycan Perlecan concentrations These polypeptides trap and locally concentrate
RANKL Decorin endocrine and paracrine growth factors within the
matrix. Soluble PGs such as decorin, biglycan, and
Binding to PG perlecan also modulate activity through binding, thus
controls factor controlling the concentrations of bioavailable factor.
Collagen availability and
Membrane-bound PGs such as betaglycan and
fibrils therefore activity
syndecan modulate ligand–receptor interactions and
thus play roles in regulating intracellular signaling.
FGF, Fibroblast growth factor; FGFR, fibroblast
growth factor receptor; OPG, osteoprotegerin; RANK,
PGs provide receptor activator of nuclear factor-κB; RANKL,
ligand-receptor receptor activator of nuclear factor-κB ligand; TGF-β,
stabilization transforming growth factor-β; TGF-βRI/RII,
transforming growth factor-β receptor types I and II.
Cell (Adapted from Lamoureux F, Baud'huin M, Duplomb
membrane Receptor signaling L, et al. Proteoglycans: key partners in bone cell
modified by PG biology. Bioessays 2007;29:758-71.)
Betaglycan expression
RANK Syndecan FGFR
TGF-RII dimer
CHAPTER 3  Bone structure and function 21

corresponding regions in known propeptides of the γ-carboxyglutamic phosphodiesterase 1 (ENPP1) is a type II transmembrane glycoprotein and
acid–containing blood coagulation factors. This common structural feature a member of the ENPP family. ENPP1 has broad specificity and cleaves
may be involved in the posttranslational targeting of these proteins for phosphodiester bonds of nucleotides and nucleotide sugars as well as
γ-carboxylation. pyrophosphate bonds of nucleotides and nucleotide sugars. This protein
Osteocalcins may also act as a hormone to regulate the activity of osteoclasts may function to hydrolyze nucleoside 5′-triphosphates to their corresponding
and their precursors. In support of this, the skeleton of the OC-null animal monophosphates and may also hydrolyze diadenosine polyphosphates.
manifests osteopetrosis compared with wild-type litter mates. In humans, Loss-of-function mutations in ENPP1 result in autosomal recessive hypophos-
OC is expressed largely by osteoblasts and osteocytes, and the measurement phatemic rickets type 2, characterized by a ricketic phenotype and elevation
of this protein in serum has been used as a marker of bone turnover. OC of FGF-2320 (see later discussion).
messenger RNA (mRNA) is upregulated by vitamin D through interactions
with trans-acting factors in vitamin D response elements in the OC promoter.15
Because of its cell-specific expression, the OC promoter has proven to be
invaluable as an active, functional DNA to drive foreign complementary The thrombospondins (TSPs) are a family of secreted glycoproteins of high
DNAs in osteoblasts in transgenic animals. molecular mass. TSP1 and TSP2 share high homology and form 450-kDa
homotrimers. Both TSP1 and TSP2 are expressed by mesenchymal cells and
chondrocytes during cartilage formation. As osteoblasts replace the mineralizing
OSTEOPONTIN cartilage, TSP2 expression decreases in chondrocytes and increases in the
Osteopontin (OPN), also referred to as secreted phosphoprotein-1, is a member matrix within areas undergoing ossification. TSP1 and TSP2 are strong
of the SIBLING (small integrin-binding ligand N-linked glycoprotein) family, antiangiogenic factors and therefore may also play a role in controlling blood
which is a group of noncollagenous ECM proteins involved in bone mineraliza- vessel organization in forming bone.
tion. The genes coding for these proteins are localized to human chromosome In developing animals, TSP1 and TSP2 are expressed in temporal and
band 4q21-25; have similar exon arrangements; and include those coding spatial patterns distinct for each gene. TSP1 (mouse gene, Thbs1) and TSP2
for dentin matrix protein-1, dentin sialoprotein, dentin phosphoprotein, (Thbs2) have both been disrupted in mice and have unique phenotypes
integrin-binding sialoprotein, and matrix extracellular phosphoglycoprotein associated with each gene. Thbs1 is a regulator of TGF-β in vivo, and null
(MEPE). The SIBLING proteins share common structural features, such as animals have lower viability and prolonged wound healing. For skeletal
multiple phosphorylation sites, a highly acidic nature, the presence of an phenotypes, this model has spine curvature and craniofacial alterations.
arginine–glycine–aspartic acid cell attachment domain, and proteolysis-resistant Thbs2-null mice have increased cortical bone density, higher numbers of
acidic serine aspartate–rich MEPE-associated motif. mesenchymal stem cells, and a resistance to bone loss due to ovariectomy.
Osteopontin has a high sialic acid content and is produced by osteoblasts The fact that the Thbs2-null mice demonstrate less bone resorption than
under stimulation by calcitriol. OPN is expressed within cement lines and wild-type controls after ovariectomy may suggest a role for this molecule in
may thus act as a promoter of adhesion and allow the arrangement of dissimilar estrogen-dependent reductions in bone mass and in the control of osteoclast
tissues together in biologic composites such as teeth and bone.16 OPN binds function.
tightly to hydroxyapatite and may be involved in the anchoring of osteoclasts
to the mineral of bone matrix. The vitronectin receptor, which has specificity
for OPN, is focused within the osteoclast plasma membrane in the regions
involved in the binding process. Long bones from OPN-null mice are Fibroblast growth factor 23
indistinguishable from those from wild-type litter mates by radiography, but Fibroblast growth factor 23 is a phosphaturic hormone produced in bone
the relative amount of mineral in the more mature areas of the bone (central (Fig. 3.4), and the encoding gene was identified as the mutated gene in
cortical bone) of the OPN-null mice is significantly increased, as is the
mineral maturity (mineral crystal size and perfection) throughout all regions
of the bone.17 In vitro, exogenous OPN inhibits inorganic pyrophosphate
(PPi)–dependent mineralization of a cultured osteoblast cell line.18 These BONE MATRIX MINERALIZATION
findings indicate that OPN is a potent inhibitor of mineral formation as well
as crystal growth and proliferation.
Osteonectin Matrix OPN
OC mineralization MEPE Osteocytes
Osteonectin, also referred to as secreted protein acidic and rich in cysteine DMP1
(SPARC), is a phosphoprotein that is the most abundant noncollagenous
polypeptide expressed in bone. The mature protein binds selectively to AP
hydroxyapatite, collagen fibrils, and vitronectin at distinct sites and may FGF-23
PPi Pi
allow proper organization of the bone matrix through contacts with the ENPP
cellular surface. Osteonectin also inhibits cellular proliferation through arrest
of cells in the G1 phase of the cell cycle. It may regulate the activity of Pi
platelet-derived growth factor, vascular endothelial growth factor, and FGF-2. Intracellular Dietary reabsorption Kidney
The osteonectin crystal structure has revealed a novel follistatin-like component PPi Pi
and an extracellular calcium-binding region containing two EF-hand motifs.
Osteonectin-null mice develop severe osteopenia, which indicates that this FIG. 3.4  Bone matrix mineralization involves an interplay of factors and is controlled
gene may have roles in osteoblast proliferation and in mineralization. by the balance between inorganic phosphate (Pi) and inorganic pyrophosphate (PPi)
and the expression of key local and systemic factors. Whereas an excess in Pi induces
ALKALINE PHOSPHATASES AND ECTONUCLEOTIDE mineralization, PPi inhibits it. Proteins expressed in osteocytes and osteoblasts regulate
mineralization. Specifically, whereas PHEX (phosphate-regulating gene with homologies
PYROPHOSPHATASE AND PHOSPHODIESTERASES to endopeptidases on the X chromosome) and dentin matrix protein-1 (DMP1) induce
Alkaline phosphatases are widely distributed and are membrane-bound mineralization, fibroblast growth factor 23 (FGF-23) inhibits mineralization. Inhibition
glycoproteins that hydrolyze monophosphate esters.19 The liver–bone–kidney of mineralization by FGF-23 is believed to be caused by inhibition of Pi reabsorption
alkaline phosphatase, referred to as tissue-nonspecific alkaline phosphatase in the kidney, which reduces blood and bone Pi levels. Induction of mineralization
(encoded by the ALPL gene), acts as a lipid-anchored phosphoethanolamine by PHEX and DMP1 may be secondary to inhibition of FGF-23 and thus increases
and pyridoxal 5′-phosphate ectophosphatase. Loss-of-function mutations in in circulating Pi. The Pi levels in the extracellular matrix depend on dietary intake
the ALPL gene lead to hypophosphatasia, which is characterized by marked and on its rate of synthesis from PPi catalyzed by alkaline phosphatase (AP). Conversely,
defects in bone mineralization and is lethal in the infantile form. The PPi the levels of PPi depend on its conversion from Pi by ectonucleotide pyrophosphatase/
produced by cells inhibits mineralization by binding to crystals, and the phosphodiesterase enzymes (ENPPs). Other bone-derived proteins, such as matrix
presence of PPi may thus prevent the soft tissues from undergoing mineraliza- extracellular phosphoglycoprotein (MEPE), osteopontin (OPN), osteocalcin (OC), and
tion. The degradation of PPi to inorganic phosphate by ALPL in bones and osteonectin, coordinately regulate mineralization, likely through direct interactions
teeth may facilitate crystal growth; therefore, it is thought that loss of function with the mineralized matrix. (Adapted from Bellido T, Plotkin LI, Bruzzaniti A. Bone
of the ALPL gene in hypophosphatasia results in accumulation of PPi and cells. In: Burr DB, Allen MR, editors. Basic and applied bone biology. San Diego:
reduced skeletal mineralization. Ectonucleotide pyrophosphatase/ Academic Press; 2014, p. 27-46.)
22 SECTION 1  Scientific Basis of Rheumatic Disease
autosomal dominant hypophosphatemic rickets (ADHR), a metabolic bone and osteocytes embedded in the mineralized matrix. In vitro studies of human
disorder of isolated renal phosphate wasting.21 Full-length FGF-23 (32 kDa) osteoblast cell cultures indicate that MEPE expression is the highest during
is the biologically active form of the protein and is inactivated upon cleavage the mineralization phase.30 Mepe-null mice display increased trabecular and
into 20- and 12-kDa protein fragments. Intracellular cleavage of FGF-23 cortical bone mass because of increases in both osteoblast number and activity,
occurs between R179 and S180 within a highly charged subtilisin-like proprotein and these mice are also resistant to age-dependent trabecular bone loss. 31
convertase (SPC) proteolytic site (R176H177T178R179/S180AE). The human FGF-23 Taken together, these findings indicate that MEPE likely has a role as an
ADHR mutations R176Q, R179Q, and R179W destroy this site and stabilize important gene for the negative regulation of skeletal mineralization.
the full-length active form of the protein. The production and secretion of
whole-molecule, bioactive FGF-23 is a dynamic process. Within the trans-Golgi
network, UDP-GalNAc transferase 3 (GALNT3) O-glycosylates FGF-23 on
T180. This glycosylation stabilizes bioactive FGF-23 by inhibiting cleavage Multiple growth factors, either produced within bone or circulating to bone,
between residues R179 and S180 via the SPC furin (PCSK3). The importance are critical for skeletal development and function. These factors may be
of this event is underscored by the fact that patients with GALNT3 inactivating sequestered within bone matrix via the bloodstream or may be produced by
mutations do not efficiently produce bioactive FGF-23.22 The GALNT3- the major bone cell types and act as paracrine and autocrine factors.
mediated FGF-23 glycosylation can be sterically hindered through prior
phosphorylation of S180 by the novel kinase FAM20C,23 thus providing a Insulin-like growth factors
molecular interaction that controls the ability of osteoblasts and osteocytes The insulin-like growth factors IGF-1 (somatomedin C) and IGF-2 (somato-
to regulate serum phosphate concentrations and potentially bone mineralization medin A) are produced primarily in the liver but are also produced in bone.
(see Fig. 3.4). FGF-23 acts in the kidney to inhibit phosphate reabsorption These factors predominantly circulate complexed with IGF binding proteins
by reducing expression of the proximal tubule type I sodium–phosphate (IGFBPs) to facilitate their transport to tissues. IGFBPs can either enhance
cotransporters Npt2a24 and Npt2c. The subsequent low serum phosphate or inhibit IGF activity. IGF-1 and -2 act through the IGF-1 receptors (IGFR1
level results in marked osteomalacia and rickets, fracture, and dental anomalies. and IGFR2) and possess bioactivity that promotes cell proliferation and
Application of in situ hybridization to adult trabecular bone revealed the differentiation.
presence of FGF-23 mRNA in osteocytes and flattened bone-lining cells. In The IGF-1-null mouse has reduced cortical bone and femur length; however,
regions of active bone formation, newly formed osteocytes and osteoprogenitor trabecular density is increased. In vitro findings suggest that IGF-1 also
cells also express FGF-23.25 FGF-23 levels are elevated in vivo by increased increases osteoclastogenesis, and IGF-1-null mice have reduced levels of
serum phosphate and 1,25-dihydroxyvitamin D concentrations, and FGF-23 receptor activator of nuclear factor-κB ligand (RANKL) in osteoblasts isolated
then completes the feedback loop by reducing phosphate reabsorption and from bone marrow. Therefore IGF-1 may regulate osteoclastogenesis through
1,25-dihydroxyvitamin D production in the kidney. Evidence also indicates direct and indirect actions. Overexpression of IGF-1 specifically in osteoblasts
that FGF-23 is directly regulated by parathyroid hormone (PTH) in leads to increased bone mineral density and increased trabecular volume,
osteocytes.26 although osteoblast numbers are not increased.32 These studies suggest that
Whether FGF-23 has direct effects on the skeleton is uncertain because IGF-1 acts directly on osteoblasts to enhance their function. Specific removal
the FGF-23 coreceptor α-Klotho is predominantly expressed in the kidney of IGFR1 from osteoblasts results in decreased trabecular number and volume
and parathyroid glands. However, because FGF-23 is produced in bone, and a dramatic decrease in bone mineralization, which further supports the
FGF-23 expression and its actions on serum phosphate concentrations may role of the IGFs with regard to osteoblasts. Less is known regarding the
be coordinated with intraskeletal signals to allow proper bone formation functions of IGF-2 in bone. However, it has been suggested that IGF-2 may
and mineralization. be a local regulator of bone cell metabolism.

PHEX Bone morphogenetic protein family

X-linked hypophosphatemia, a disorder of rickets and osteomalacia, is caused Bone morphogenetic proteins (BMPs) are members of the TGF-β superfamily.
by inactivating mutations in PHEX (phosphate-regulating gene with homologies There are now more than 20 BMP-related proteins, which are classified into
to endopeptidases on the X chromosome).27 PHEX encodes a protein that subgroups based on structure and function. These factors play important
is similar to the M13 family of membrane-bound metalloproteases such as roles in skeletal development by directing the fate of mesenchymal cells,
neutral endopeptidase and endothelin-converting enzymes 1 and 2 (see Fig. through differentiation of these precursor cells into cells of the osteoblastic
3.4). These proteases are known to cleave small peptide hormones. Mutations lineage, and by inhibiting their differentiation into myoblastic lineage cells.
in PHEX lead to dramatic over expression of FGF-23; however, the PHEX BMPs also increase osteoclastogenesis, which is tightly coordinated with
substrate and molecular mechanisms underlying this increase are currently osteoblastogenesis. BMPs activate specific receptors and induce cell signaling
unknown. by phosphorylating cytoplasmic receptor–regulated Smads, which enter the
nucleus to recruit transcription factors and enhance gene expression. The
Dentin matrix protein-1 human disorder fibrodysplasia ossificans progressiva is a disease of dramatic
Dentin matrix protein-1 (DMP1), similar to OPN, is a member of the SIBLING ectopic bone formation, which can be accelerated after blunt trauma. A
gene family. DMP1 is highly expressed in osteocytes and is composed of 513 recurrent mutation in activin receptor IA/activin-like kinase 2, a BMP type
residues but is secreted in bone and dentin as 37-kDa N-terminal (residues I receptor, was reported as the molecular cause of fibrodysplasia ossificans
17–253) and 57-kDa C-terminal (residues 254–513) fragments from a 94-kDa progressiva.33 These findings underscore the potent effects of BMP signaling
full-length precursor (see Fig. 3.4). Recombinant DMP1 binds calcium- on bone formation.
phosphate ions and the N-telopeptide region of type I collagen with high
affinities, so in vivo DMP1 may regulate local mineralization processes in Fibroblast growth factors
bone and teeth. The C-terminal portion of DMP1 has been implicated in Members of the FGF family of proteins primarily act as paracrine and autocrine
DNA binding, in gene regulation, and as an integrin-binding protein. Inactivat- factors and bind to one or several of four FGF receptors (FGFRs). FGFRs
ing mutations in DMP1 result in the metabolic bone disease autosomal normally exist as an inactivated monomer. With FGF binding in the presence
recessive hypophosphatemic rickets, which is associated with elevated FGF-23 of heparin/heparan sulfate, the FGFRs dimerize, which leads to the autophos-
levels in these patients. As shown in the Dmp1-null mouse (and in the Hyp phorylation of tyrosine residues. The FGF family has potent effects on bone
mouse model of X-linked hypophosphatemia), the primary cellular defect development. This is clearly evident by the fact that activating mutations in
caused by loss of Dmp1 may be an alteration in osteoblast to osteocyte FGFR1 and FGFR2 are responsible for disorders of craniosynostosis and
maturation, leading to inappropriate expression of typically “osteoblastic” limb patterning, and FGFR1 and FGFR3 mutations result in disorders of
or “early osteocyte” genes such as type I collagen, alkaline phosphatase, and hypochondroplasia and achondroplasia. The FGFs interact with HSPGs and
FGF-23 in mature embedded osteocytes. The relationship of DMP1 to cell are sequestered within the mineralizing matrix. In addition, the HSPG syndecan
differentiation is currently unknown. Interestingly, FAM20c, a novel secreted may stabilize FGF-FGFR interactions and promote FGF signaling and
kinase,28 has been shown to phosphorylate the SIBLING proteins DMP1 and bioactivity.
MEPE (see next section). Loss of FAM20c in mice results in a Dmp phenotype,29 The FGF family members play important roles in bone development and
which indicates a complex relationship between the phosphorylation of ECM formation. Expression of several FGF ligands, including FGF-2, FGF-5,
proteins and their normal functions. FGF-6, FGF-7, and FGF-9, has been observed in mesenchyme surrounding
the initial congregations of cells that proliferate and differentiate to form
Matrix extracellular phosphoglycoprotein bone. In limb bud, FGFR1 and FGFR2 are expressed in condensing mesen-
Another member of the SIBLING family found in the mineralizing matrix chyme. In rat growth plates, mRNAs encoding all four FGFRs and FGF-2
is MEPE (see Fig. 3.4). MEPE is predominantly expressed in odontoblasts can be detected, and FGF-2 is also present in osteoblasts. FGF-2 treatment
CHAPTER 3  Bone structure and function 23

of osteoblasts enhances the binding of Runx2 to the Cbfa1 consensus sequence bone healing. This factor also destabilizes blood vessels during healing to
in the OC promoter and may therefore have a role in differentiation. allow sprouting of new vessels. VEGF is produced by many cell types including
fibroblasts, hypertrophic chondrocytes, and osteoblasts. VEGF may act not
Transforming growth factor-β only in bone angiogenesis and vascular differentiation but also in aspects of
Transforming growth factor-β controls proliferation, differentiation, and other development, such as chondrocyte and osteoblast differentiation, as well as
functions in many cell types. TGF-β1, TGF-β2, and TGF-β3 all function osteoclast recruitment.
through the type I and type II TGF receptors. The type I TGF-β receptor
forms a heterodimer with the type II TGF-β receptor. TGF-β stimulation
leads to activation of SMAD2 and SMAD3, which form complexes with
SMAD4 that accumulate in the nucleus and regulate the transcription of Bone development and the adaptation of the adult skeleton to mechanical
target genes. needs and hormonal changes depend on the ability of bone cells to resorb
Transforming growth factor-β is the most abundant growth factor in and form bone in the right places and at the right time. Bone growth, modeling,
human bone; it is localized within the bone matrix and has functions both and remodeling are defined by the spatial and temporal relationship between
during embryonic development and in mature bone. During embryonic bone resorption and bone formation. Osteoclasts resorb bone, osteoblasts
development, TGF-β1 plays a role in cell migration, controlling epithelial– form bone, and osteocytes detect the need for bone augmentation or reduction
mesenchymal interactions, and the formation of cellular condensations, which and coordinate the activity of osteoclasts and osteoblasts.
influence bone shape. This factor also plays a key role in inducing mesenchymal
cell differentiation to either chondrocytes or osteoblasts. In adult bone,
TGF-β1 controls osteoblast differentiation, which affects matrix formation
and mineralization. TGF-β1 inhibits the expression of the differentiation Osteoclasts are the primary bone-resorptive cells. They are needed for bone
markers Runx2 and OC in osteoblast cell lines, and its functions interplay modeling, which leads to changes in the shape of bones, and for bone
with those of other systems in bone such as the PTH and the Wnt/β-catenin remodeling, which maintains the integrity of the adult skeleton. Osteoclasts
systems. originate from precursors of the monocyte/macrophage family of the hema­
The skeletal disorder Camurati-Engelmann disease (CED) highlights the topoietic lineage that differentiate to multinucleated cells upon stimulation
importance of TGF-β1 in skeletal formation. CED is a progressive diaphyseal with RANKL and macrophage colony-stimulating factor (M-CSF) (Fig. 3.5).
dysplasia characterized by hyperostosis and sclerosis of the diaphyses of the Upon completing bone resorption, all osteoclasts undergo programmed cell
long bones. The TGFB1 gene was screened and three different heterozygous death or apoptosis and disappear from the bone surface.
missense mutations were found in exon 4 in the nine families examined.
All of the mutations in the CED patients were located either at C225 or near Osteoclast morphology and function
R218, which suggests the importance of this region in activating TGF-β1 in Osteoclasts adhere firmly to bone through the interactions established between
the bone matrix. integrins expressed in the osteoclast membranes with collagen, fibronectin,
and other bone matrix proteins. Expression of αV and α3 integrin is induced
Platelet-derived growth factor and vascular endothelial during osteoclast differentiation, and the integrin binds to the amino acid
growth factor sequence Arg-Gly-Asp present in OPN and bone sialoprotein. The importance
All platelet-derived growth factors (PDGFs) and vascular endothelial of these events for osteoclast activity is underscored by the inhibition of
growth factors (VEGFs) are dimers of disulfide-linked polypeptide chains, resorption with competitive Arg-Gly-Asp ligands34 and a progressive increase
encoded by nine different genes that direct production of four different in bone mass caused by osteoclast dysfunction in mice null for β 3
PDGF chains (PDGF-A, PDGF-B, PDGF-C, and PDGF-D) and five different integrin.
VEGF chains (VEGF-A, VEGF-B, VEGF-C, VEGF-D, and placental growth The intimate contact between the osteoclast and the bone matrix creates
factor). All members of these families carry a growth factor core domain that a space called the sealing zone. There is also polarization of the osteoclast
is necessary for receptor activation. PDGFs mediate their bioactivity through fibrillar actin into a circular structure called the actin ring, containing
two receptors, PDGFR-α and PDGFR-β. These receptors both have five podosomes composed of an actin core surrounded by αVβ3 integrins and
extracellular immunoglobulin loops for ligand binding and an intracellular associated cytoskeletal and signaling proteins. Thus, the area in which the
tyrosine kinase domain. The VEGFs act through a homologous family of osteoclast apposes the bone is isolated from the general extracellular space
receptors, VEGFR1, VEGFR2, and VEGFR3. PDGFs act primarily as paracrine and becomes acidified by the activity of a proton pump and a chloride
growth factors. channel.34 The low pH in this area dissolves the mineral and exposes the
Platelet-derived growth factor is chemotactic and mitogenic for osteoblasts organic matrix, which is subsequently degraded by the activity of lysosomal
and osteoprogenitor cells, and it upregulates cytokines that are crucial to cathepsin K and matrix metalloproteases. These degrading enzymes are


FIG. 3.5  Osteoclast differentiation is governed by Stromal

osteoblastic Osteocytes
receptor activator of nuclear factor-κB ligand
(RANKL) and macrophage colony-stimulating factor
(M-CSF) secreted by osteoblasts and osteocytes,
which control various steps of the osteoclast
differentiation process, including precursor
proliferation, commitment, differentiation, and
maturation. Osteoprotegerin (OPG), which is also
secreted by osteoblasts and osteocytes, acts as a
decoy receptor for RANKL and reduces osteoclast
differentiation. (From Bellido T, Plotkin LI, Bruzzaniti Hematopoietic
A. Bone cells. In: Burr DB, Allen MR, editors. Basic monocyte/
and applied bone biology. San Diego: Academic
Press; 2014, p. 27-46.)
Monocytes Preosteoclast Fusion Polarization Resorption

Determination Proliferation Commitment Maturation Apoptosis

survival differentiation

24 SECTION 1  Scientific Basis of Rheumatic Disease
transported into acidified vesicles that fuse with the osteoclast plasmalemma, Osteoclast apoptosis
forming a villous structure referred to as the ruffled border. This structure All osteoclasts undergo apoptosis and disappear from the bone surface after
and the actin ring are essential features of a resorbing osteoclast, and abnor- completing bone resorption. High concentrations of extracellular calcium,
malities of either structure lead to arrested bone resorption. similar to the ones present in resorption cavities, induce osteoclast apoptosis
The cytoplasmic domains of integrins serve as platforms for signaling in vitro and may be the triggering event. Fas ligand stimulates osteoclast
proteins involved in osteoclast function, such as the kinase Src, which is apoptosis, and Fas-deficient mice exhibit more osteoclasts and decreased
crucial for osteoclast attachment and resorption. Src regulates podosome bone mass, which suggests that this pathway controls osteoclast life span in
disassembly and ruffled membrane formation by its ability to interact with vivo. Osteoclast apoptosis might also result from loss of survival signals
the focal adhesion–related kinase Pyk2 and the proto-oncogene c-Cbl. Rho, provided by integrin interactions with the matrix or by changes in the produc-
Rac, and the guanine nucleotide exchange factor Vav3, which activates tion of cytokines or growth factors that preserve osteoclast viability. Potential
guanosine diphosphatases into guanosine triphosphatases, also play a central antiapoptotic factors are M-CSF and RANKL, the same cytokines that induce
role in modifying the resorptive capacity of osteoclasts by modulating the osteoclast differentiation. TNF-α and IL-1 also delay osteoclast apoptosis.
actin cytoskeleton. Osteoclast resorption products are transported in vesicles All of these cytokines activate the extracellular signal–regulated kinases
through the cytosol to the basolateral surface and discharged to the extracellular (ERKs), the activation of which is required for osteoclast survival. Phos-
milieu or directly released to the surrounding fluid after osteoclast retraction phatidylinositol 3′-kinase (PI3-K) and its target the kinase Akt are required
from the resorption pits. for osteoclast differentiation but not for survival. Instead, mammalian target
of rapamycin (mTOR), another PI3-K target, is required for the antiapoptotic
Osteoclast formation and differentiation actions of M-CSF, RANKL, and TNF-α in osteoclasts. Because mTOR is also
Mature, multinucleated osteoclasts are formed by fusion of mononuclear activated by ERKs, it appears to be a point of convergence in the action of
precursors of the monocyte/macrophage lineage (see Fig. 3.5). The earliest prosurvival kinases in osteoclasts.
recognized osteoclast precursor is the granulocyte-macrophage colony-forming RANKL, TNF-α, and IL-1 also activate NF-κB, a transcription factor
unit, which also gives rise to granulocytes and monocytes. Osteoclast precur- shown to inhibit apoptosis in various cell types. Downregulation of NF-κB
sors proliferate in response to growth factors such as interleukin-3 (IL-3) mRNA inhibits IL-1–dependent survival, and blockade of NF-κB binding to
and colony-stimulating factors like granulocyte-macrophage colony-stimulating DNA with specific oligonucleotides induces apoptosis. However, osteoclast
factor (GM-CSF) and M-CSF to form postmitotic, committed mononucleated precursors lacking NF-κB subunits have normal survival rates, and inhibition
osteoclast precursors, which differentiate and fuse to form multinucleated of NF-κB activation via a dominant-negative IKK2 does not affect the ability
osteoclasts under the influence of RANKL, a member of the tumor necrosis of IL-1 to promote osteoclast survival. Therefore, the relevance of NF-κB
factor (TNF) family of ligands. signaling for osteoclast survival is still controversial.
M-CSF and RANKL are critical for osteoclastogenesis, and deletion of
M-CSF, RANKL, or RANK (the receptor for RANKL expressed by osteoclasts Regulation of osteoclast generation and survival
and their precursors) inhibits osteoclast differentiation, leading to osteopetrosis In the bone-remodeling unit, whereas the rate of osteoclast generation
in mice. Both M-CSF and RANKL are expressed by bone marrow stromal determines the extension of the bone-remodeling unit, the life span of
cells and osteoblastic cells, as well as T lymphocytes and other cell types osteoclasts determines the depth of resorption. Although both genesis
in pathologic settings. Importantly, osteocytes have now been found to be and apoptosis of osteoclasts lead to changes in osteoclast number
a major source of M-CSF and RANKL, as well as OPG, the RANKL decoy and bone resorption, alteration of osteoclast life span might represent
receptor,35,36 and deletion of RANKL from osteocytes leads to osteopetrosis,36 a more effective mechanism to accomplish rapid changes in bone
demonstrating a central role of osteocytes in osteoclastogenesis (see Fig. resorption rate.
3.5). Whereas M-CSF contributes to osteoclast differentiation, migration, Sex steroids have profound effects on osteoclasts. Both estrogens and
and survival by binding to its receptor c-Fms on osteoclast precursors, androgens inhibit osteoclast generation by regulating the production of
RANKL facilitates osteoclast formation via direct binding to the receptor pro-osteoclastogenic cytokines (e.g., IL-6 and IL-1) by cells of the stromal/
RANK. RANKL is expressed on the cell surface and is also secreted as a osteoblastic lineage. Estrogens also induce apoptosis of mature osteoclasts.
soluble form. Although the soluble form of RANKL is found in the circula- This, together with an inhibitory effect of the hormones on osteoblast genera-
tion and its presence is sufficient to induce differentiation of osteoclast tion, leads to attenuation of the rate of bone remodeling.
precursors in vitro, its actual role in osteoclast formation in vivo remains Mice receiving excess glucocorticoids exhibit reduced osteoclast progenitors,
unproven. but cancellous osteoclast number does not decrease in the early phases of
RANKL expression is upregulated by hormones and cytokines known to the disease because glucocorticoids prolong the life spans of preexisting
induce osteoclast generation. This explains the long-observed property of osteoclasts. This effect may account for the early transient increase in bone
primary osteoblastic cells or osteoblastic cell lines that, upon treatment with resorption in patients with hyperglucocorticoidism. In contrast to the rapid
vitamin D, PTH, or IL-11, IL-6, TNF, and IL-1, support osteoclast develop- prosurvival effect of glucocorticoids on mature osteoclasts, glucocorticoids
ment when co-cultured with osteoclast precursors derived from spleen or induce a decrease in osteoclast formation caused by a reduction in the pool
bone marrow. RANKL mediates several aspects of osteoclast differentiation, of osteoblastic cells that support osteoclastogenesis. This effect leads to the
including fusion of mononucleated precursors into multinucleated cells, typical low remodeling rate observed in chronic glucocorticoid-induced
acquisition of osteoclast-specific markers, attachment of osteoclasts to osteoporosis.
the bone surfaces, stimulation of resorption, and promotion of osteoclast
survival. Although M-CSF contributes to RANKL effects, RANKL appears
to play a dominant role in bone resorption. Thus, whereas M-CSF–null
mice recover with time from the decreased osteoclast number and activity, Osteoblasts are the cells responsible for bone formation. They originate from
RANKL knockout mice do not. Furthermore, RANKL appears to stimulate mesenchymal progenitors, which also give rise to chondrocytes, muscle cells,
osteoclast formation and resorption in mice even in the absence of functional and adipocytes (Fig. 3.6). Commitment of mesenchymal cells to the osteoblastic
M-CSF.37 lineage depends on the specific activation of transcription factors induced
RANKL activates several signal transduction pathways involving the by morphogenetic and developmental proteins that carry out the functions
recruitment of the adapter protein TRAF6 (TNF receptor–associated factor of bone matrix protein secretion and bone mineralization. Upon completion
6) to the intracellular domain of the receptor RANK. TRAF6 in turn activates of bone matrix formation, some mature osteoblasts remain entrapped in
kinase-dependent signaling as well as transcription factors. Among them, bone as osteocytes, some flatten to cover quiescent bone surfaces as bone-lining
NF-κB has been shown to undergo nuclear translocation, leading to upregula- cells, and most die by apoptosis.
tion of c-Fos. c-Fos, in turn, binds to nuclear factor of activated T cells,
cytoplasmic 1 (NFATc1) and upregulates genes crucial for osteoclast differentia- Osteoblast function
tion and function. Although other signaling pathways are activated by RANKL The main function of osteoblasts is to synthesize collagen type I and other
in osteoclasts, the evidence that deletion of NF-κB, c-fos/AP1, and NFATc1 specialized matrix proteins that serve as a template for the subsequent mineral
leads to osteoclast dysfunction demonstrates the crucial role of these genes deposition in the form of hydroxyapatite. Mature osteoblasts actively engaged
in osteoclasts.37 in this process are recognized by their location on the bone surface and by
Osteoprotegerin is an inhibitor of RANK activation and osteoclastogenesis their morphologic features typical of cells secreting high levels of proteins:
that also belongs to the TNF family of receptors. OPG is a secreted protein cuboidal shape with large nucleus, enlarged Golgi apparatus, and extensive
with no transmembrane domain, and therefore it has no direct signaling endoplasmic reticulum. Osteoblasts express high levels of alkaline phosphatase
capabilities. OPG suppresses osteoclast formation and resorption by binding and OC, and the level of these proteins in blood reflects the rate of bone
to RANKL, thereby impeding RANKL interaction with RANK. formation.

CHAPTER 3  Bone structure and function 25


Runx2 Rbp-Jk
Smad 1/4/5
FIG. 3.6  Osteoblastogenesis is controlled by Dlx3,5,6 Rbp-Jk
transcription factors that affect the proliferation and Transcription Rbp-Jk Osx FRA-1/ATF4
differentiation of osteoblast precursors. Mature factors Twist1
-catenin NFAT2 -catenin
osteoblasts can surround themselves by bone matrix Dermo1 -catenin Osteocyte
and differentiate further to become osteocytes, flatten
to cover the quiescent bone surface as lining cells, or
die by apoptosis. (From Bellido T, Plotkin LI,
Bruzzaniti A. Bone cells. In: Burr DB, Allen MR, Lining cell
editors. Basic and applied bone biology. San Diego: Mesenchymal Osteochondral Preosteoblast Immature Mature
Academic Press; 2014, p. 27-46.) stem cell precursor osteoblast osteoblast


Interaction of osteoblasts among themselves, with lining cells, and with Examination of the nuclear morphology of cells transfected with fluorescent
bone marrow cells is established by adhesion junctions, tight junctions, and proteins containing a nuclear localization sequence has proven a particularly
gap junctions. Adhesion junctions mainly mediated by cadherins and tight useful tool for studying apoptosis in cells co-transfected with genes of interest.
junctions serve to join cells and facilitate their anchorage to the ECM. Changes Cell detachment from the substrate, changes in the composition of the plasma
in the expression level of the major cadherins expressed in osteoblasts, membrane, and changes revealing cell shrinkage are also features that have
N-cadherin and cadherin 11, influence osteoblast differentiation and survival. been used to detect and quantify apoptotic cells.
Intercellular communication among osteoblasts and neighboring cells is
maintained by cell coupling via gap junctions. Opening of gap junction Regulation of osteoblast generation and apoptosis
channels contributes to coupling and the coordination of responses within Most major regulators of skeletal homeostasis influence both generation and
a cell population. The major gap junction protein expressed in bone cells survival of osteoblasts. The BMP and Wnt signaling pathways promote
is connexin 43. Its absence or dysfunction leads to impaired osteoblast osteoblast differentiation, but whereas BMPs induce osteoblast apoptosis,
differentiation, premature apoptosis of osteoblasts and osteocytes, and deficient Wnts inhibit it. BMPs induce apoptosis of mature osteoblasts as well as of
response to hormones and pharmacotherapeutic agents.38 Furthermore, gap mesenchymal osteoblast progenitors in interdigital tissues during the develop-
junction communication is fundamental for the maintenance of a continuum ment of the hands and feet. Wnt signaling has a profound effect on bone as
from bone, where osteocytes reside, through bone surface cells, osteoblasts shown by the high-bone-mass phenotype of mice and humans with activating
and osteoclasts, bone marrow cells, and endothelial cells of the blood vessels.39 mutations of low-density lipoprotein receptor–related protein 5 (LRP5),
This functional syncytium might be responsible for the coordinated response which together with Frizzled proteins are receptors for Wnt ligands. Wnts
of the bone tissue to changes in physical and chemical stimuli, as will be stimulate differentiation of undifferentiated mesenchymal cells toward the
discussed later. Interactions between osteoblasts and the bone matrix via osteoblastic lineage and stimulate differentiation of preosteoblasts. Canonical
integrins also modulate osteoblast differentiation, function, and survival. In Wnt signaling in osteoblasts also affects osteoclasts by enhancing the expression
particular, loss of antiapoptotic signals provided by the ECM causes apoptosis, of the RANKL decoy receptor OPG, which leads to inhibition of osteoclast
a phenomenon referred to as anoikis. development. In addition, Wnt signaling inhibits apoptosis of mature osteo-
blasts and osteocytes.42 The increased bone formation exhibited by mice
Osteoblast formation and differentiation lacking the Wnt antagonist known as secreted Frizzled related protein-1
The process of osteoblastogenesis can be divided into steps comprising (sFRP-1) is associated with decreased osteoblast and osteocyte apoptosis.
proliferation, ECM development and maturation, mineralization, and apoptosis. The prevalence of osteoblast and osteocyte apoptosis is also decreased in
Each stage is characterized by activation of specific transcription factors and mice expressing the high bone mass–activating mutation of LRP5 (G171V),
genes leading to a succession of osteoblast phenotypic markers (see Fig. which exhibit reduced ability to bind the Wnt antagonist sclerostin secreted
3.6). Transcription factors of the helix-loop-helix family (Id, Twist, and by osteocytes. Consistent with this, sclerostin induces osteoblast apoptosis
Dermo) are expressed in proliferating osteoblast progenitors and are responsible in vitro. Moreover, reduction of sclerostin levels by PTH and mechanical
for maintaining the osteoprogenitor population by inhibiting the expression loading increases osteoblast number and activity as a result of stimulation
of genes that characterize the osteoblast mature phenotype. Transcription of osteoblast differentiation and increased survival.43,44 Activation of Wnt
factors of the activating protein family, such as c-fos, c-jun, and junD, are signaling in vitro by ligands known to activate the so-called canonical as
expressed during proliferation as well as later in the differentiation pathway well as noncanonical pathways also prevents apoptosis of osteoblast progenitors
and may activate or repress transcription. Runx2 and osterix are essential and differentiated osteoblasts through a mechanism that involves the Src/
for establishing the osteoblast phenotype. Their absence from the mouse ERK and PI3/AKT pro-survival kinases.45
genome results in lack of skeletal mineralization and perinatal lethality. Glucocorticoids induce rapid bone loss resulting from a transient increase
Runx2 and osterix regulate the expression of other genes that control bone in resorption caused by delayed osteoclast apoptosis. This initial phase is
formation and remodeling, including OC and RANKL. Runx2 regulates followed by a sustained and profound reduction in bone formation and
differentiation, survival, and function of osteoblasts by affecting several turnover caused by decreased osteoblast and osteoclast generation and
signaling pathways, including those activated by Wnts, BMPs, integrins, and increased osteoblast apoptosis.
the PTH receptor. Both persistent excess of PTH, as in hyperparathyroidism, and intermittent
elevation of PTH (by daily injections) increase the number of osteoblasts.
Osteoblast apoptosis Sustained PTH elevation inhibits the expression of sclerostin, with a consequent
Upon completing the process of bone formation, 60% to 70% of osteoblasts increase in Wnt signaling and in differentiation of osteoblast precursors. A
die by apoptosis; the rest become lining cells or osteocytes. Apoptosis occurs major effect of intermittent elevation of PTH is inhibition of apoptosis of
throughout all stages of osteoblast life.40 The prevalence of osteoblast apoptosis osteoblasts, which thereby prolongs their life span and ability to form bone.
in bone sections can be quantified by measuring fragmented DNA. Apoptosis
of cultured osteoblasts has been extensively studied using several methods, 41
including increased activity of initiator or effector caspases, the presence of
cleaved genomic DNA by TUNEL or ISEL assay, and nuclear fragmentation Osteocytes are former osteoblasts that become entombed during the process
and chromatin condensation using fluorescent dyes that bind to DNA. of bone deposition and are regularly distributed throughout the mineralized

26 SECTION 1  Scientific Basis of Rheumatic Disease
bone matrix. Osteocyte bodies are individually encased in lacunae and exhibit rapid activation of the Src/Shc/ERK signaling pathway through nongeno-
cytoplasmic dendritic processes that run along narrow canaliculi excavated tropic actions of the classical receptors for sex steroids. This effect requires
in the mineralized matrix. Osteocytes communicate with each other, with only the ligand-binding domain of the receptor, and unlike the classical
cells on the bone surface, and with cells of the bone marrow through genotropic action of the receptor protein, it is eliminated by nuclear
gap junctions established between cytoplasmic processes of neighboring targeting.
cells. Today it is accepted that osteocytes are the mechanosensory cells. Excess of glucocorticoid activity in bone may also contribute to induction
Osteoblasts and osteoclasts are present on bone only transiently, in low of osteocyte (and osteoblast) apoptosis because aged mice exhibit higher
number, and in variable locations. On the other hand, osteocytes are the serum levels of corticosterone, elevated adrenal weight, and increased expres-
most abundant resident cells and are present in the entire bone volume. sion in bone of 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1), the
Osteocytes are also the core of a functional syncytium that extends from enzyme that amplifies glucocorticoid action. The apoptotic effect of gluco-
the mineralized bone matrix to the bone surface and the bone marrow corticoids is reproduced in cultured osteocytes and osteoblasts in a manner
and all the way to the blood vessels. This strategic location permits the strictly dependent on the glucocorticoid receptor.40 Induction of osteocyte
detection of variations in mechanical signals as well as in levels of circu- and osteoblast apoptosis by glucocorticoids can result from the direct action
lating factors and allows amplification of the signals leading to adaptive of the steroids, because overexpression of the enzyme that inactivates glu-
responses. cocorticoids, 11β-HSD2, specifically in these cells abolishes the increase in
apoptosis. Strikingly, in the osteocytic MLO-Y4 cell line, the proapoptotic
Osteocyte apoptosis: consequences and regulation effect of glucocorticoids is preceded by cell detachment caused by interference
Osteocytes are long-lived cells. However, similar to osteoblasts and osteoclasts, with FAK-mediated survival signaling generated by integrins. In this mecha-
osteocytes die by apoptosis, and decreased osteocyte viability accompanies nism, Pyk2 (a member of the FAK family) becomes phosphorylated and
the bone fragility syndromes that characterize glucocorticoid excess, estrogen subsequently activates proapoptotic JNK signaling. The proapoptotic actions
withdrawal, and mechanical disuse.34 Conversely, preservation of osteocyte of glucocorticoids may involve suppression of the synthesis of locally produced
viability might explain at least part of the antifracture effects of bisphospho- antiapoptotic factors, including IGF-1- and IL-6-type cytokines, as well as
nates, which cannot be completely accounted for by changes in bone mineral matrix metalloproteins, and stimulation of the proapoptotic Wnt antagonist
density.46 SFRP-1.

Preservation of osteocyte viability by mechanical stimuli Regulation of bone formation by osteocytes: sclerostin
Osteocytes interact with the ECM in the pericellular space through discrete Osteocytes express sclerostin, the product of the Sost gene, which antagonizes
sites in their membranes, which are enriched in integrins and vinculin, as several members of the BMP family of proteins and also binds to LRP5/LRP6,
well as through transverse elements that tether osteocytes to the canalicular preventing canonical Wnt signaling. Loss of Sost in humans causes the
wall. Loading of the bones induces ECM deformation and fluid flow through high-bone-mass disorders van Buchem syndrome and sclerosteosis. In addition,
the canaliculi, producing tension in the tethering elements and strain on administration of an antisclerostin antibody increases bone formation and
osteocyte membranes. The consequent integrin engagement leads to intracel- restores the bone lost after ovariectomy. Conversely, transgenic mice over-
lular signaling. Physiologic levels of mechanical strain imparted by stretching expressing Sost exhibit low bone mass.49 These lines of evidence demonstrate
or pulsatile fluid flow prevent apoptosis of cultured osteocytes. Mechanotrans- that sclerostin derived from osteocytes exerts a negative feedback control at
duction is accomplished by molecular complexes assembled at caveolin-rich the earliest step of mesenchymal stem cell differentiation toward the osteoblast
domains of the plasma membrane and composed of integrins, cytoskeletal lineage. Moreover, PTH and mechanical loading downregulate the expression
proteins, and kinases, including the focal adhesion kinase FAK and Src, of sclerostin in osteocytes, which reveals a novel mechanism of bone anabolism
which results in activation of the ERK pathway and osteocyte survival. triggered by osteocytes.
Intriguingly, a ligand-independent function of the estrogen receptor is
indispensable for mechanically induced ERK activation in both osteoblasts Osteocytes as mediators of the anabolic actions of canonical Wnt
and osteocytes. This observation is consistent with reports that mice lacking signaling in bone
estrogen receptor-α and estrogen receptor-β exhibit a poor osteogenic response Bone anabolic stimuli activate Wnt signaling, and human mutations of
to loading. components along this pathway underscore its crucial role in bone accrual
In vivo mechanical forces also regulate osteocyte life span. Apoptotic and maintenance. However, the cell responsible for orchestrating Wnt anabolic
osteocytes are found in unloaded bones or in bones exposed to high levels actions had remained elusive because genetic activation or deletion of
of mechanical strain. In both cases, increased osteocyte apoptosis is observed components of the pathway in osteoblasts or their precursors only affect
before any evidence of increased osteoclast resorption. Apoptotic osteo- bone resorption without evident effects on bone formation.50,51 A recent
cytes accumulate in areas subsequently removed by osteoclasts. Targeted study demonstrates that activation of canonical Wnt signaling exclusively
ablation of osteocytes in transgenic mice is sufficient to induce osteoclast in osteocytes increases (not decreases) resorption and induces bone anabolism,
recruitment and resorption, leading to bone loss. These findings led to the leading to high bone remodeling with bone gain. This effect is due to the
notion that dying osteocytes become the beacons for osteoclast recruitment fact that activation of the pathway in osteocytes not only decreases OPG as
to the vicinity and the resulting increase in bone resorption.47 Whether in osteoblasts but also increases RANKL, leading to a higher RANKL-to-OPG
living osteocytes continually produce molecules that restrain osteoclast ratio that favors resorption, and in this setting, Wnt signaling also favors
recruitment or whether in the process of undergoing apoptosis osteocytes osteoblast–osteocyte differentiation.52 Thus, these findings demonstrate
produce pro-osteoclastogenic signals remains to be determined. Taken together disparate outcomes of β-catenin activation in osteocytes versus osteoblasts
with the evidence that osteocyte apoptosis is inhibited by estrogens and and identify osteocytes as central target cells of the anabolic actions of
bisphosphonates,46,48 these findings raise the possibility that preservation canonical Wnt/β-catenin signaling in bone.
of osteocyte viability contributes to the ability of these agents to inhibit
remodeling. Regulation of bone resorption by osteocytes: RANKL
and osteoprotegerin
Osteocyte apoptosis and aging The cues that signal bone resorption are not completely understood. Apoptotic
One of the purported functions of the osteocyte network is to detect micro- osteocytes could regulate the recruitment of osteoclast precursors and their
damage and trigger its repair. During aging, there is an accumulation of differentiation in two ways. Osteocyte apoptosis may indirectly stimulate
microdamage and a decline in osteocyte density accompanied by decreased osteoclastogenesis by inducing stromal/osteoblastic cells to secrete RANKL.
prevalence of osteocyte-occupied lacunae, an index of premature osteocyte In addition, osteocytes can directly secrete RANKL. Indeed, in vitro, purified
death. Age-related loss of osteocytes caused by apoptosis could be partially osteocytes express higher levels of RANKL than osteoblasts and bone marrow
responsible for the disparity between bone quantity and quality that occurs stromal cells. The severe osteopetrotic phenotype observed in mice lacking
with aging. The decline in physical activity and thus reduced skeletal loading RANKL in osteocytes and their resistance to bone loss induced by tail suspen-
with old age is a potential mechanism for the increased prevalence of osteocyte sion strongly suggests that osteocytes are a major source of RANKL in vivo.
(and osteoblast) apoptosis, as is the loss of estrogen in women during and Osteocytes also secrete OPG, which, as in osteoblasts, is regulated by the
after menopause. Wnt/β-catenin pathway. Mice lacking β-catenin in osteocytes are osteoporotic
because of increased osteoclast numbers, but their osteoblast function is
Hormonal regulation of osteocyte life span normal. Emerging evidence also points to osteocytes as an additional source
Estrogen and androgen deficiency both lead to increased prevalence of of secreted M-CSF in bone. Together, these new findings suggest that osteocytes
osteocyte apoptosis. Conversely, estrogens and androgens inhibit apoptosis control the bone-remodeling process by regulating osteoclast and osteoblast
of osteocytes as well as osteoblasts.48 This antiapoptotic effect is due to differentiation and function.

CHAPTER 3  Bone structure and function 27



Osteoclast precursors Osteoblast precursors

Marrow capillary

Lining cells Bone remodeling compartment

RANKL Sclerostin
M-CSF Dkk-1



FIG. 3.7  Osteocytes sense the need for bone resorption and send signals to lining cells, which retract from the bone
surface to allow the formation of a canopy under which remodeling occurs, called the bone-remodeling compartment
(BRC). Osteoclast precursors are transported to the BRC by marrow capillaries, differentiate to mature osteoclasts
under the influence of pro-osteoclastogenic and antiosteoclastogenic cytokines (receptor activator of nuclear factor-κB
ligand [RANKL], macrophage colony-stimulating factor [M-CSF], and osteoprotegerin [OPG]) derived from osteocytes,
and initiate bone remodeling. Osteoblast precursors from the bone marrow or the circulation differentiate into
mature, bone-synthesizing cells in response to factors released from the bone matrix by resorption. Differentiation
and function of osteoblasts are controlled by molecules derived from osteocytes, including sclerostin and Dkk-1.
(From Bellido T, Plotkin LI, Bruzzaniti A. Bone cells. In: Burr DB, Allen MR, editors. Basic and applied bone biology.
San Diego: Academic Press; 2014, p. 27-46.)

Table 3.1  the canalicular system and secretion of the other Wnt antagonist Dkk-1 by
osteocytes as well as osteoblasts influence the rate of bone formation, providing
Dynamic Processes of Skeletal Development and Adaptation an additional level of control of osteoblast activity. Based on these lines of
Process Mechanism Morphology Function evidence, the BRC might provide a supportive environment for differentiation
Growth F Woven, lamellar Increased mass of osteoclast and osteoblast progenitors. Thus, regulation of the bone-
Modeling A-F Primary lamellar Net increased mass remodeling rate by hormonal and mechanical stimuli could be accomplished
or by controlling the balance between resorption and formation within the BRC
Adaption of architecture through the regulation of osteocytic molecules, including sclerostin, RANKL,
A-R Control of drift and and OPG.
Remodeling A-R-F Secondary lamellar Bone maintenance
(osteons, hemi-osteons) Repair of microdamage GROWTH, MODELING, REMODELING,
Prevention of bone loss AND REPAIR
Repair F Woven Repair of fractures
Rapid mechanical Four dynamic processes are involved in skeletal development and adaptation.
adaptation These are defined by the relationship of bone resorption and bone formation
to each other (Table 3.1). These mechanisms include a coordinated system
A, Activation; F, formation; R, resorption. that first involves the activation (A) of cell populations followed by the
resorption (R) of preexisting tissue and/or the formation (F) of new or
replacement tissue. Bone growth serves to increase bone mass through bone
formation. Resorption of bone is not part of the growth process, and the
Osteocytes and the bone-remodeling compartment function of growth is only to increase mass, not to adapt the developing
Lining cells play an important function in initiating bone remodeling by structure to its mechanical needs. Growth can occur on a substrate but may
retracting from quiescent bone surfaces and allowing the formation of a involve ossification directly from fibrous tissue (intramembranous bone
canopy over osteoclasts and osteoblasts in the bone multicellular unit.53 On formation) or by formation of a model with cartilage first and then replacement
the endocortical surface, this canopy presumably encases bone marrow of the cartilage with bone (endochondral ossification). Modeling uses the
osteoblast precursors and is penetrated by blood vessels that provide hema­ tissue formed during the growth process to further increase bone mass and
topoietic osteoclast progenitors. The canopy, associated capillaries, osteocytes, to shape its geometry to mechanical needs. Modeling occurs through the
osteoclasts, and osteoblasts form a compartment, the bone-remodeling activation of cells followed by either formation or resorption. Formation
compartment (BRC), which is separated from the rest of the marrow and and resorption are coordinated processes in modeling but do not occur
which can sequester molecules that regulate the cells that remodel bone sequentially on the same surface of bone. Remodeling, on the other hand,
(Fig. 3.7). The signals that trigger lining cell detachment in a particular is defined by the sequential processes on the same bone surface of activation–
bone area are unknown. Premature apoptosis of osteocytes has been shown resorption–formation (the A-R-F sequence). The function of remodeling is
to precede osteoclast accumulation and resorption,47 which raises the possibility bone maintenance, not increase in bone mass, and the removal of microdamage.
that osteocytes release molecules that induce lining cell retraction facilitating Bone’s repair function, which restores its mechanical properties after a complete
access of osteoclast precursors to bone surfaces. However, the molecular fracture or trabecular microfracture, usually occurs through the process of
entities responsible for this purported osteocytic function remain unknown. endochondral ossification, which forms a cartilage callus that also includes
As discussed earlier, osteocytes express M-CSF,54 which stimulates proliferation woven bone to bridge the fracture gap. This is eventually replaced through
of preosteoclasts, and RANKL,55,56 the master cytokine inducer of osteoclast remodeling with replacement by lamellar bone.
differentiation, both of which could reach the BRC. Growth factors released Growth, modeling, and remodeling are present concurrently in all growing
from the bone matrix upon resorption, in turn, stimulate osteoblastogenesis. children. When skeletal maturity is reached, growth naturally stops. Modeling
It is also likely that osteocyte-derived sclerostin reaching the BRC through slows down or stops but may still be present at a reduced rate on trabecular

28 SECTION 1  Scientific Basis of Rheumatic Disease
surfaces and on the periosteal surface of the bone. At maturity, the predominant the wider joint surface continually be reshaped and narrowed as it moves
process is bone remodeling, which maintains the bone that has been formed down into the metaphysis and diaphysis. Modeling is a continuous and
and repairs microscopic damage that may be sustained in bone during normal prolonged process—unlike remodeling, which is episodic—and involves a
daily activities. Dysfunction in the remodeling process is associated with the coordinated process of bone resorption on some surfaces, while other surfaces
loss of bone found in osteoporosis. undergo bone formation. Bone modeling occurs on both periosteal and
endocortical envelopes and sculpts bone shape while allowing for expansion
of the marrow cavity and periosteal diameter of the diaphysis. At the
GROWTH metaphysis, this occurs by osteoclastic resorption on the periosteal surfaces.
Growth of bone occurs through two different skeletal processes, one involving As growth continues, however, bone is added to the periosteal surface by
formation of bone from fibrous membranes and the other involving the osteoblasts and simultaneously removed from the endocortical surface by
formation of a skeletal anlage, or model. Intramembranous bone formation osteoclasts. This increases whole-bone diameter and expands the marrow
occurs at centers of ossification via direct mineralization in highly vascular cavity, necessary for the formation of blood. It serves a second purpose: to
fibrous tissues through the action of mesenchymal cells. The calvaria of the increase the mechanical strength of the bone while at the same time not
skull is the best example of intramembranous bone formation, with the increasing its mass or weight at the same rate. Bone curvature is also adjusted
individual bones of the skull acting as centers that eventually grow together during growth through a process known as drift, in which the periosteal
at the sutures. Apposition of bone on the periosteal surface of long bones surface on one side of the bone undergoes apposition while the opposite
also occurs through intramembranous ossification. periosteal surface undergoes resorption. Likewise, different portions of the
In the long bones, development generally occurs by the initial condensation endocortical surface form or resorb bone in coordination to maintain the
of mesenchyme or hyaline cartilage in the form of the eventual skeletal cortical thickness of the diaphysis.
structure. This cartilage model mineralizes over time and becomes detectable
as a primary center of ossification. Some bones are formed from a single
ossification center, although most of the long bones form secondary centers
of ossification at the ends (epiphyses), which eventually fuse to the bone
that developed from the primary ossification center (diaphysis) (Fig. e3.2). The quantum concept of bone-remodeling states that bone is replaced in
The secondary centers allow growth to occur at a cartilaginous growth plate packets through the coupled activity of osteoclasts and osteoblasts. Coupling
until skeletal maturity in the late teens or (for the vertebral bodies) in the between osteoclasts and osteoblasts is the reason that it was difficult for so
early part of the third decade of life. The growth plate slowly converts into many years to control the processes involved in bone loss. When bone
primary spongiosa that becomes remodeled into lamellar trabecular plates resorption is suppressed, formation is also suppressed because these activities
in the metaphysis of the bone. are linked by intercellular signaling mechanisms that are not fully understood.
In remodeling, resorption and formation are coupled but in fact may not be
balanced. Coupling and balance are not the same; whereas balance refers to
MODELING the relationship between the amount of bone resorbed and the amount
Long bones must grow both in length and in diameter (Fig. 3.8). At the formed, coupling denotes only that the processes are linked in some way.
ends of the long bones—near the epiphyses—growth of bone demands that Resorption and formation are in balance in the healthy skeleton, but when
these are out of balance, the amount of bone that is resorbed can be either
greater or less than the amount that is subsequently formed. Thus, even
though cells are coupled, bone can be lost or added in several different ways
based on the altered balance of resorption and formation. In actuality, one
THE METAPHYSEAL CUTBACK THAT OCCURS THROUGH MODELING almost never finds a balance in favor of bone formation in a remodeling
PROCESSES DURING GROWTH TO SHAPE THE BONE system, although this has been shown to occur with anabolic treatments for
osteoporosis, such as the intermittent administration of recombinant human
Growth PTH(1-34) (Teriparatide).57 More often, the balance is in favor of resorption.
This is the case in osteoporosis, in which global resorption is increased but
formation at each of the erosion sites is normal or reduced, which leads to
a deficit in bone mass.
This coupled system is termed a bone multicellular unit (BMU) because
different cell populations are involved. A BMU typically consists of about
10 osteoclasts and several hundred osteoblasts. When cut in longitudinal
This bone removed section, the BMU shows the sequential aspects of the A-R-F system and the
by modeling various cell populations that are involved (Fig. e3.3). Each of the phases in
this sequence is location, magnitude, and rate specific, so that alterations in
the magnitude or timing of one can produce morphologic features characteristic
of specific skeletal abnormalities (Fig. 3.9). Activation is initiated by chemical
or mechanical signals but actually involves a series of events that include
recruitment of precursor cells, differentiation and proliferation of cells, and
migration to the site of activity. In humans, these processes take about 5 to
10 days. Bone resorption by mature osteoclasts takes about 3 weeks at a
Diaphysis given site, although osteoclasts moving longitudinally through bone at a
rate of about 40 µm/day may live much longer than this. There is a period
Metaphyseal cutback of reversal during which there is neither bone resorption nor formation; this
may represent a period like the activation period during which osteoblasts
are undergoing differentiation and proliferation from their precursors. This
is followed by a period of bone formation that lasts about 3 months. As
unmineralized bone—or osteoid—is laid down, it subsequently begins to
mineralize, quickly at first, and then more slowly over the following year.
This sequence of events occurs on all four skeletal envelopes (periosteal,
Diaphyseal enlargement Diaphyseal drift endocortical, trabecular, and intracortical).
Changes in bone mass can occur simply through changes in activation
FIG. 3.8  During growth, metaphyseal bone is removed by modeling processes to frequency. Changes in activation frequency may lead only to transient changes
shape the bone. Subsequent enlargement of the diaphysis occurs through direct in bone mass if resorption and formation are in balance. Early losses of bone
periosteal apposition, which is often accompanied by resorption on the endocortical mass solely caused by increased activation frequency may resolve after several
surface to enlarge the marrow cavity. Bone can also change its location and curvature months as the newly resorbed sites are refilled. Likewise, bone mass may
through “drift.” Arrows indicate the direction of drift, with smaller circular cells change because some aspect of the recruitment, proliferation, migration, or
representing osteoblasts and bone formation and larger ellipsoidal cells representing differentiation of either osteoclasts or osteoblasts is interrupted. These changes
osteoclasts and bone resorption. (From Martin RB, Burr DB, Sharkey NA. Skeletal may be manifested as alterations of resorption–formation balance but may
tissue mechanics. New York: Springer; 1998, p. 62, with permission.) be caused by activation defects during cell maturation.

CHAPTER 3  Bone structure and function 28.e1

Primary spongiosa

FIG. E3.2  At the organ level, bone consists of compact L2, 37 y.o. male
(cortical) bone, which forms a shell around the more
porous, cancellous (trabecular) bone, or spongiosa. In
the cancellous regions of the long bones, such as the
proximal tibia shown here, the primary spongiosa is
separated from the secondary spongiosa by an Secondary
arbitrary boundary. Primary spongiosa is composed of spongiosa
primary bone, often laid down during growth on a
calcified cartilage core that is subsequently
remodeled. Secondary spongiosa is remodeled,
reflects patterns of stress, and directs these stresses
to the cortical shell.

Cortical bone
cortical bone


Capillary bud


Erosion surface

FIG. E3.3  A bone multicellular unit (BMU). This system shows the sequential aspects
of the activation–resorption–formation (A-R-F) system, and the various cells that are
involved. At the head of the resorption front (also called the cutting cone), there is
a capillary bud, which supplies nutrients to the multicellular osteoclasts that are
decalcifying and resorbing bone matrix. Behind the resorption front, teams of osteoblasts
are lined up along the wall of the BMU, laying down new bone, or osteoid, that will
subsequently become mineralized. Some of these osteoblasts will eventually embed
themselves and become terminally differentiated osteocytes (McNeal tetrachrome

CHAPTER 3  Bone structure and function 29


remodeling sequence, not
including mineralization, takes
about 4 months in humans. Many
cellular processes occur during
each part of the activation– Migration
resorption–formation (A-R-F)
sequence of events. The amount
of resorption and formation are
determined both by the individual
activity of the cells and by the
duration of the cells’ lifetimes. In
cross-section, formation requires Nucleation Maturation
about four times longer than the
resorption of the same amount of
bone, and consequently there are

many more osteoblasts in bone

than osteoclasts. (From Burr DB.
Orthopedic principles of skeletal
growth and remodeling. In:
Carlson DS, Goldstein SA, editors. Duration Duration
Bone biodynamics in orthodontic
and orthopedic treatment.
Craniofacial Growth Series, vol. 27. 10 21 days 5 90 days ~200 days
days days
Ann Arbor, MI: Center for Human
Growth and Development of the ion
n ion l n zat
io pt sa io ali
University of Michigan; 1992, at r r at r
p. 15-50.) tiv so ve rm ine
Ac Re Re Fo M


Regeneration, not repair

Three sequential, yet overlapping stages

Normal bone Healed fracture

FIG. 3.10  An injury to bone is

typically followed by the
development of a hematoma with
associated inflammatory
responses. This phase is followed
by the development of a periosteal
bridging callus (at least in
separated and unstable fractures)
that is composed of calcified
cartilage. Over time, this cartilage
callus remodels to become bone,
and eventually the bone is
reshaped through modeling
processes to achieve something
close to its original dimensions. Inflammation Regeneration Remodeling
(With permission from Dr. Stuart
Warden.) Injury Hematoma/inflammation Regeneration Remodeling

Days Weeks Months/years

REPAIR on periosteal surfaces through a process typical of intramembranous

Fracture healing involves several stages in the repair (or regeneration) ossification. This unites the two ends of the bone to stabilize the fracture,
process (Fig. 3.10). The injury is typically followed by the development but the bone is still not mechanically adapted to normal loads and may
of a hematoma with associated inflammatory responses. This phase is fol- be quite weak. The callus subsequently matures to primary bone over a
lowed within a week or so by a regenerative phase characterized by the period of months via processes similar to those involved in endochondral
formation of a cartilage callus. Also, there is direct woven bone apposition ossification.

30 SECTION 1  Scientific Basis of Rheumatic Disease
Over time, the cartilage callus remodels via the A-R-F sequence of events facilitate healing. A variety of local factors can influence healing. Nonunions
to become lamellar bone, and eventually the bone is reshaped through modeling can occur if early signals for repair (e.g., from TGF-β and other growth
processes to achieve something close to its original dimensions. When this factors and cytokines) are not received, if there is compromise to the local
process is complete, the bone will be geometrically and mechanically similar blood supply, or if there are complicating factors caused by infection or
to its prefracture state. bone death caused by radiation or thermal injuries. Comorbid conditions
The rate and extent of healing depend on the fracture and on the (e.g., poor diet, smoking, calcium or vitamin D deficiencies, and exces-
mechanical environment. Large amounts of motion increase the size of sive alcohol intake) can prolong the time required to heal a fracture. Any
the cartilage callus and interrupt its remodeling to bone. Small amounts agent that suppresses bone turnover (e.g., bisphosphonates or RANKL
of motion and juxtaposition of the fractured ends of the bone can allow inhibitors) will prevent the cartilage callus from fully remodeling to lamellar
healing without the development of much, or any, callus. Small loads also bone.

1. Casley-Smith JR. The functioning and interrelationships 19. Weiss MJ, Henthorn PS, Lafferty MA, et al. Isolation and 38. Civitelli R. Cell-cell communication in the osteoblast/
of blood capillaries and lymphatics. Experientia. characterization of a cDNA encoding a human liver/bone/ osteocyte lineage. Arch Biochem Biophys.
1976;32:1-12. kidney-type alkaline phosphatase. Proc Natl Acad Sci 2008;473:188-192.
2. Bennett HS, Luft JH, Hampton JC. Morphological U S A. 1986;83:7182-7186. 39. Aarden EM, Burger EH, Nijweide PJ. Function of
classifications of vertebrate blood capillaries. Am J Physiol. 20. Lorenz-Depiereux B, Schnabel D, Tiosano D, et al. osteocytes in bone. J Cell Biochem. 1994;55:287-299.
1959;196:381-390. Loss-of-function ENPP1 mutations cause both generalized 40. Jilka RL, Bellido T, Almeida M, et al. Apoptosis in bone
3. Cooper RR. Nerves in cortical bone. Science. arterial calcification of infancy and autosomal-recessive cells. In: Bilezikian JP, Raisz LG, Martin TJ, eds. Principles
1968;160:327-328. hypophosphatemic rickets. Am J Hum Genet. of bone biology. San Diego: Academic Press; 2008:237-261.
4. Schaffler MB, Burr DB, Frederickson RG. Morphology of 2010;86:267-272. 41. Bellido T, Plotkin LI. Detection of apoptosis of bone cells
the osteonal cement line in human bone. Anat Rec. 21. White KE, Evans WE, O’Riordan JLH. Autosomal in vitro. In: Westendorf JJ, ed. Osteoporosis. Totowa, NJ:
1987;217:223-228. dominant hypophosphataemic rickets is associated with Humana Press; 2007:51-75.
5. Sasaki N, Tagami A, Goto T, et al. Atomic force mutations in FGF23. Nat Genet. 2000;26:345-348. 42. Bodine PV. Wnt signaling control of bone cell apoptosis.
microscopic studies on the structure of bovine femoral 22. Topaz O, Shurman DL, Bergman R, et al. Mutations in Cell Res. 2008;18:248-253.
cortical bone at the collagen fibril-mineral level. J Mater GALNT3, encosing a protein involved in O-linked 43. Bellido T, Ali AA, Gubrij I, et al. Chronic elevation of
Sci Mater Med. 2002;13:333-337. glycosylation, causes familial tumoral calcinosis. Nat PTH in mice reduces expression of sclerostin by
6. Schwarcz HP. The ultrastructure of bone as revealed in Genet. 2004;36:579-581. osteocytes: a novel mechanism for hormonal control of
electron microscopy of ion-milled sections. Sem Cell Dev 23. Tagliabracci VS, Engel JL, Wiley SE, et al. Dynamic osteoblastogenesis. Endocrinology. 2005;146:4577-4583.
Biol. 2015;46:44-50. regulation of FGF23 by Fam20C phosphorylation, 44. Robling AG, Niziolek PJ, Baldridge LA, et al. Mechanical
7. Fuchs RK, Allen MR, Ruppel ME, et al. In situ GalNAcv-T3 glycosylation, and furin proteolysis. Proc stimulation of bone in vivo reduces osteocyte expression
examination of the time-course for secondary Natl Acad Sci U S A. 2014;111:5520-5525. of Sost/sclerostin. J Biol Chem. 2008;283:5866-5875.
mineralization of Haversian bone using synchrotron 24. Shimada T, Mizutani S, Muto T, et al. Cloning and 45. Almeida M, Han L, Bellido T, et al. Wnt proteins prevent
Fourier transform infrared microspectroscopy. Matrix Biol. characterization of FGF23 as a causative factor of apoptosis of both uncommitted osteoblast progenitors
2008;27:34-41. tumor-induced osteomalacia. Proc Natl Acad Sci U S A. and differentiated osteoblasts by beta-catenin-dependent
8. Masic A, Bertinetti L, Schuetz R, et al. Osmotic pressure 2001;98:6500-6505. and -independent signaling cascades involving Src/ERK
induced tensile forces in tendon collagen. Nat Comm. 25. Riminucci M, Collins MT, Fedarko NS, et al. FGF-23 in and phosphatidylinositol 3-kinase/AKT. J Biol Chem.
2015;6:5942. fibrous dysplasia of bone and its relationship to renal 2005;280:41342-41351.
9. Katti DR, Pradhan SM, Katti KS. Directional dependence phosphate wasting. J Clin Invest. 2003;112:683-692. 46. Plotkin LI, Weinstein RS, Parfitt AM, et al. Prevention of
of hydroxyapatite-collagen interactions on mechanics of 26. Rhee Y, Bivi N, Farrow EG, et al. Parathyroid hormone osteocyte and osteoblast apoptosis by bisphosphonates
collagen. J Biomech. 2010;43:1723-1730. receptor signaling in osteocytes increases the expression and calcitonin. J Clin Invest. 1999;104:1363-1374.
10. Noonan KJ, Stevens JW, Tammi R, et al. Spatial of fibroblast growth factor-23 in vitro and in vivo. Bone. 47. Bellido T. Osteocyte apoptosis induces bone resorption
distribution of CD44 and hyaluronan in the proximal 2011;49:636-643. and impairs the skeletal response to weightlessness.
tibia of the growing rat. J Orthop Res. 1996;14:573-581. 27. The Hyp Consortium. A gene (PEX) with homologies to Bonekey Osteovision. 2007;4:252-256.
11. Shibata S, Fukada K, Imai H, et al. In situ hybridization endopeptidases is mutated in patients with X-linked 48. Kousteni S, Bellido T, Plotkin LI, et al. Nongenotropic,
and immunohistochemistry of versican, aggrecan and link hypophosphatemic rickets. Nat Genet. 1995;11:130-136. sex-nonspecific signaling through the estrogen or
protein, and histochemistry of hyaluronan in the 28. Tagliabracci VS, Engel JL, Wen J, et al. Secreted kinase androgen receptors: dissociation from transcriptional
developing mouse limb bud cartilage. J Anat. phosphorylates extracellular proteins that regulate activity. Cell. 2001;104:719-730.
2003;203:425-432. biomineralization. Science. 2012;336:1150-1153. 49. Loots GG, Kneissel M, Keller H, et al. Genomic deletion
12. Williams DR Jr, Presar AR, Richmond AT, et al. Limb 29. Wang X, Wang S, Li C, et al. Inactivation of a novel of a long-range bone enhancer misregulates sclerostin in
chondrogenesis is compromised in the versican deficient FGF23 regulator, FAM20C, leads to hypophosphatemic Van Buchem disease. Genome Res. 2005;15:928-935.
hdf mouse. Biochem Biophys Res Commun. rickets in mice. PLoS Genet. 2012;8:e1002708. 50. Glass DA, Bialek P, Ahn JD, et al. Canonical Wnt
2005;334:960-966. 30. Siggelkow H, Schmidt E, Hennies B, et al. Evidence of signaling in differentiated osteoblasts controls osteoclast
13. Corsi A, Xu T, Chen XD, et al. Phenotypic effects of downregulation of matrix extracellular differentiation. Dev Cell. 2005;8(5):751-764.
biglycan deficiency are linked to collagen fibril phosphoglycoprotein during terminal differentiation in 51. Holmen SL, Zylstra CR, Mukherjee A, et al. Essential role
abnormalities, are synergized by decorin deficiency, and human osteoblasts. Bone. 2004;35:570-576. of beta-catenin in postnatal bone acquisition. J Biol Chem.
mimic Ehlers-Danlos-like changes in bone and other 31. Gowen LC, Petersen DN, Mansolf AL, et al. Targeted 2005;280(22):21162-21168.
connective tissues. J Bone Miner Res. 2002;17:1180-1189. disruption of the osteoblast/osteocyte factor 45 gene 52. Tu X, Delgado-Calle J, Condon KW, et al. Osteocytes
14. Xu T, Bianco P, Fisher LW, et al. Targeted disruption of (OF45) results in increased bone formation and bone mediate the anabolic actions of canonical Wnt/b-catenin
the biglycan gene leads to an osteoporosis-like phenotype mass. J Biol Chem. 2003;278:1998-2007. signaling in bone. Proc Natl Acad Sci U S A.
in mice. Nat Genet. 1998;20:78-82. 32. Zhao G, Monier-Faugere MC, Langub MC, et al. Targeted 2015;112(5):E478-E486.
15. Staal A, Van Wijnen AJ, Birkenhager JC, et al. Distinct overexpression of insulin-like growth factor I to 53. Eriksen EF. Cellular mechanisms of bone remodeling. Rev
conformations of vitamin D receptor/retinoid X osteoblasts of transgenic mice: increased trabecular bone Endocr Metab Disord. 2010;11:219-227.
receptor-alpha heterodimers are specified by dinucleotide volume without increased osteoblast proliferation. 54. Harris SE, MacDougall M, Horn D, et al.
differences in the vitamin D-responsive elements of the Endocrinology. 2000;141:2674-2682. Meox2Cre-mediated disruption of CSF-1 leads to
osteocalcin and osteopontin genes. Mol Endocrinol. 33. Shore EM, Xu M, Feldman GJ, et al. A recurrent mutation osteopetrosis and osteocyte defects. Bone. 2012;50:42-53.
1996;10:1444-1456. in the BMP type I receptor ACVR1 causes inherited and 55. Nakashima T, Hayashi M, Fukunaga T, et al. Evidence for
16. McKee MD, Nanci A. Osteopontin at mineralized tissue sporadic fibrodysplasia ossificans progressiva. Nat Genet. osteocyte regulation of bone homeostasis through RANKL
interfaces in bone, teeth, and osseointegrated implants: 2006;38:525-527. expression. Nat Med. 2011;17:1231-1234.
ultrastructural distribution and implications for 34. Teitelbaum SL, Ross FP. Genetic regulation of osteoclast 56. Xiong J, Onal M, Jilka RL, et al. Matrix-embedded cells
mineralized tissue formation, turnover, and repair. development and function. Nat Rev Genet. control osteoclast formation. Nat Med. 2011;17:
Microsc Res Tech. 1996;33:141-164. 2003;4:638-649. 1235-1241.
17. Boskey AL, Spevak L, Paschalis E, et al. Osteopontin 35. Kramer I, Halleux C, Keller H, et al. Osteocyte Wnt/ 57. Dempster DW, Cosman F, Kurland ES, et al. Effects of
deficiency increases mineral content and mineral beta-catenin signaling is required for normal bone daily treatment with parathyroid hormone on bone
crystallinity in mouse bone. Calcif Tissue Int. homeostasis. Mol Cell Biol. 2010;30:3071-3085. microarchitecture and turnover in patients with
2002;71:145-154. 36. O’Brien CA, Nakashima T, Takayanagi H. Osteocyte osteoporosis: a paired biopsy study. J Bone Miner Res.
18. Addison WN, Azari F, Sorensen ES, et al. Pyrophosphate control of osteoclastogenesis. Bone. 2013;54(2):258-263. 2001;16:1846-1853.
inhibits mineralization of osteoblast cultures by binding 37. Kearns AE, Khosla S, Kostenuik PJ. Receptor activator of
to mineral, up-regulating osteopontin, and inhibiting nuclear factor κB ligand and osteoprotegerin regulation of
alkaline phosphatase activity. J Biol Chem. bone remodeling in health and disease. Endocr Rev.
2007;282:15872-15883. 2008;29:155-192.

Tendons and ligaments
Stephanie G. Dakin • Andrew J. Carr 4  

Key Points enthesopathy, and desmopathy are used throughout to describe disease affecting
■ Tendinopathy, enthesopathy, and desmopathy are common causes of
tendons, the enthesis, and ligaments, respectively.
musculoskeletal pain and disability.
Musculoskeletal soft tissue disease is a common and significant health
care problem both in sports injuries and in the aging population, causing
■ Genetic predisposition, metabolic disorders, aging, cumulative mechanical stress,
pain and impeding activities of daily living. The etiology of intrinsic tendi-
and injury are important contributing factors.
nopathy, enthesopathy, and desmopathy is complex and not fully understood.
■ Inflammation is present in both tendinopathic and ruptured tendons and is
Multiple factors are implicated, including an interplay among the effects of
associated with pain.
repetitive wear and tear, daily exercise, and aging. Genetic susceptibility is
■ Inflammatory signatures differ between different stages of disease. an additional predisposing factor, demonstrated by individuals who have
■ Diseased tendons and ligaments heal by fibrosis resulting in tissue with reduced sustained damage to multiple tendons and ligaments within their bodies.
structural and functional integrity. Studies of siblings suggest that genetic factors have a role in the progression
■ Development of targeted therapies that address the underlying disease-stage specific of full-thickness tears of the rotator cuff,1 and variations within the Tenascin
mechanisms is essential. C gene have been reported to be associated with Achilles ruptures.2 The
importance of inflammation as a contributor to the pathogenesis of tendon
and ligament injury has been contentious in recent years. Although the
INTRODUCTION phenotypes of the key cells orchestrating inflammation have not been fully
characterized, there is a growing body of recent evidence to support the
Tendons and ligaments are connective tissue structures that facilitate the contribution of inflammation to the onset and progression of tendinopathy
stabilization and movement of joints. Both tendons and ligaments have a and enthesopathy as shown in Table 4.2.
highly organized ultrastructure, which confers the mechanical properties
required for their function. Tendons and ligaments share a similar structure, PATHOPHYSIOLOGY OF TENDON
consisting of a hierarchical arrangement of collagen fiber bundles oriented
along the long axis of the tissue. Other major constituents are water and
noncollagenous matrix. Although tendons and ligaments both respond to Histopathologic evaluation of diseased tendons shows increased cellularity,
mechanical load, exercise, and stress deprivation, there are some differences increased vascularity, and matrix disorganization compared with healthy
between these structures. These include compositional differences in extracel- tendon tissues (Fig. 4.1). Increased cellularity can be attributed to fibroblast
lular matrix (ECM) and site-specific adaptations of some tendons, which proliferation, with local infiltration of inflammatory cells to the damaged
can be characterized, broadly speaking, according to their function as either site. Calcific deposits and lipid droplets may be found in tendons with
positional or energy storing. Sites commonly affected by tendinopathy include established or end stage disease.
the shoulder (supraspinatus), elbow, knee (patellar tendon), and ankle Immunohistochemical techniques facilitate improved in situ identification
(Achilles). Desmopathy (disease affecting ligaments) frequently affects high- of the cell types present in diseased tendons. In the absence of a panel of
motion joints such as the knee (collateral and cruciate ligaments) and ankle validated markers for healthy and diseased tendon stromal cells, markers of
(collateral ligaments). myeloid and lymphoid cells have been investigated in these tissues. Sections
of diseased tendon tissues show significantly increased numbers of mac-
rophages, T cells, mast cells, and natural killer cells compared with healthy
CLASSIFICATION OF DISEASE tendons.3-5 However, the phenotypes of these immune cells have not been
Both tendons and ligaments are susceptible to damage and the effects of fully characterized. Recent studies also suggest a key role for several alarmins,
aging. Disease affecting these structures can arise in the form of acute trauma including hypoxia-induced elements, cytokines, and heat shock proteins,
resulting from an impact or penetrating object (extrinsic) or more commonly affecting tissue rescue mechanisms in tendon pathology. 6 Elevated
the result of cumulative microtrauma that occurs over a prolonged period interleukin-33 mRNA and protein has been shown to be a feature of early-stage
(intrinsic). The efficacy of the ensuing repair depends on numerous factors, disease in human shoulder tendinopathy.7
including the location of the injured tendon (intra- vs extrasynovial) or
ligament (intra- vs extraarticular). Clinical presentations of intrinsic tendon INFLAMMATION ACTIVATION PATHWAYS SHOW PLASTICITY
pathology can be variable. Patients may experience pain and loss of function
with a structurally intact tendon or present with a “spontaneous” tendon
rupture without preexisting pain or perceived loss of function. There is also Macrophages play an essential role orchestrating inflammation and tissue
considerable individual variation between the degree of pain experienced repair. The signaling pathways underpinning activation of macrophages to
by a patient and the extent of structural tendon disease present. It is not become M1 or M2 subtypes have been recently revised to highlight receptors
clear whether these distinct clinical presentations represent isolated disease and key signaling mediators in common and distinct pathways. These include
entities or divergent ends of a spectrum of disease process and progression. proinflammatory pathways regulated by type I interferons (IFNs) and nuclear
The site of tendon pathology can also vary, affecting the midbody of the factor κB (NF-κB), profibrotic pathways containing signal transducer and
tendon (e.g., Achilles) or occurring at the enthesis (e.g., supraspinatus and activator of transcription 6 (STAT6), and inflammation resolving pathways
Achilles). It is not presently known whether common mechanisms are shared involving glucocorticoid receptor activation.8 These macrophage activation
between tendon disease affecting the midbody and enthesis or whether the pathways have recently been investigated in tissue biopsies from patients
presence of pathology at one site is likely to influence the other. Furthermore, with supraspinatus tendon disease. Tendon biopsies from patients with intact
it is not clear whether inflammatory disease at immune-privileged sites such tendinopathy showed expression of genes and proteins induced by type I
as the gastrointestinal tract may trigger development of localized tendon or Interferons and NF-κB.3 Conversely, biopsies from patients with large to
entheseal pathology. massive tendon tears showed expression of genes and proteins induced by
STAT6 and glucocorticoid receptor activation pathways. This transition in
ETIOLOGY OF TENDINOPATHY, ENTHESOPATHY, inflammation signature suggests that the type of inflammation differs between
tendinopathic (intact) and large to massive supraspinatus tendon tears.
AND DESMOPATHY Whereas tendinopathic shoulder tendons show a markedly proinflammatory
Terminology commonly used to describe musculoskeletal soft tissue pathology signature, torn shoulder tendons exhibit low-level chronic inflammation.
is listed in Table 4.1. For the purpose of this chapter, the terms tendinopathy, The persistence of chronic inflammation is an important feature of tendon

32 SECTION 1  Scientific Basis of Rheumatic Disease

Table 4.1  Table 4.2 

Commonly used terms to describe pathology affecting tendons Inflammatory and nociceptive molecules in tendinopathy
and ligaments and enthesopathy
Terminology Meaning Molecule Potential role
Tendinopathy or Used to describe disorders affecting tendons and IL-1 Inflammation in surrounding tissues38,39
desmopathy ligaments, respectively; does not assume IL-6 Inflammation within tendon40
knowledge of underlying pathology IL-21R Inflammation within tendon41
Tendonitis or desmitis Implies disease of tendon or ligament and is IL-33 Inflammation and alarmin within tendon7
accompanied by an inflammatory response COX-2 Prostaglandin production42,43
Tendinosis Implies disease develops because of a primarily TGF-β Tissue repair and fibrosis44
degenerative process in the absence of Substance P Nociceptive neuromodulator10
inflammation Glutamate Nociceptive neuromodulator11
Tendon or ligament Tendon pathology develops because of effects of
disease repetitive cyclic loading and aging, resulting in COX2, Cyclooxygenase-2; IL-1, interleukin1; IL-6, interleukin 6; IL-21, interleukin 21 receptor; IL-33,
interleukin 33; TGF-β, transforming growth factor-β.
cumulative microdamage (intrinsic)
Tendon or ligament Occurs as a consequence of traumatic injury, e.g.,
injury laceration or cutaneous injury (extrinsic)


FIG. 4.1  Histologic features of healthy and diseased supraspinatus (shoulder) tendons. A, Hematoxylin and eosin–stained sections of healthy supraspinatus tendon showing
organized collagen fiber bundles and relative low cellularity. B, Biopsy of a patient with painful supraspinatus tendinopathy. C, Tissue section from a patient with a painful
supraspinatus tendon tear. Diseased tendons show increased cellularity and vascularity compared with healthy tendon.

disease, demonstrated by the retention of inflammatory cells, failure of interactions between components of the ECM. Key elements of this are the
clearance of apoptotic cells, and disorganized ECM (Fig. 4.2). Failure to dense fibrillar network of parallel-orientated collagen fibers, proteoglycans
resolve tendon inflammation may be conducive to development of fibrosis. and glycoproteins. Collagens are the major proteins of the tendon ECM and
Improved understanding of tendon inflammation signatures and the phenotypes constitute approximately 65% of the tissue dry weight. Ninety-five percent
of lymphoid and stromal cells populating these tissues will inform development is type 1 collagen, with the remaining proportion composed of collagen
of disease-stage specific therapeutic targets to moderate tendon types II, III, IV, V, VI, IX, XII, and XIV. The noncollagenous portion of tendon
inflammation. is composed of glycoproteins and proteoglycans, which have important roles
in assembly, organization, and maintenance of the ECM and confer the ability
to resist tensile forces associated with loading by binding water.
PAIN AND NEUROPEPTIDES Tendon disease is associated with numerous molecular changes to the
The peripheral nervous system plays an important role in the regulation of ECM, including increased amounts of collagen III mRNA and protein.14
tendon function and repair. In healthy tendons, neuromediators are pre- Collagen III fibrils have a smaller average diameter than type I fibrils, resulting
dominantly found in the paratenon, the loose connective tissue layer sur- in formation of a collagen mesh rather than highly aligned fibrils, impacting
rounding the tendon.9 During tendon repair, there is peripheral nerve ingrowth the strength of the repair. Increased amounts of fibronectin, tenascin C,
into the tendon, and neurotransmitters, including substance P and glutamate, aggrecan, and biglycan found in samples of diseased tendons are consistent
are present.10,11 This peripheral neural phenotype is maintained in cells derived with increased matrix turnover and remodeling associated with the condition.15
from diseased tendons, which show increased mRNA expression of nociceptive Homeostasis and remodeling of the ECM is mediated by matrix metallo-
neuromodulators, including mGluR2, N-methyl-D-aspartate receptor proteinase (MMP) enzymes and their respective tissue inhibitors (tissue
(NMDAR1), and kainite receptor 1, compared with cells derived from healthy inhibitors of metalloproteinases [TIMPs]). In tendon disease, expression and
tendons.12 Consequently, antagonists of neuropeptides and neurotransmitters activity of numerous MMPs and TIMPs are altered as a consequence of
have been investigated as potentially useful therapies for the management increased proteolytic activity and turnover within the tendon ECM, summarized
of painful tendon disorders. However, a proportion of patients with significant in Table 4.3.
radiologic and histopathologic evidence of tendon disease do not experience Damaged tendons and ligaments repair and remodel by fibrosis, defined
pain, suggesting a mismatch between the presence of structural tendon disease as the scarring or hardening of tissues. During fibrosis, the composition of
and pain perception. This may be attributable to the activation of central as the ECM changes, and increased production of proteoglycans, glycosami-
well as peripheral pain pathways13 and the poorly understood interactions noglycans, and collagen type III is associated with matrix disorganization.
between these systems. This results in the formation of tendon or ligament that is inferior in structure
and function compared with healthy tissue and prone to recurrent disease.
EXTRACELLULAR MATRIX TURNOVER IN HEALTH Fibrosis is the product of chronic inflammatory reactions induced by a variety
of stimuli, including tissue injury.16 Fibrotic pathways are poorly studied in
AND DISEASE diseased musculoskeletal soft tissues. Important regulators of fibrosis in
The ability of tendons and ligaments to act as elastic energy stores and resist other connective tissues include Th2 cytokines such as interleukin-13 and
tensile loads is attributable to their highly organized structure and the transforming growth factor-β1 (TGF-β1), angiogenic factors (vascular

CHAPTER 4  Tendons and ligaments 33

FIG. 4.2  Expression of inflammation activation proteins in tendon disease. Representative immunofluorescence images of sections of tendinopathic supraspinatus tendon
(intact tendinopathy) stained for inflammation activation markers including those of the STAT6 pathway (CD206, green), the glucocorticoid receptor pathway (CD163, red),
the interferon pathway (IRF5, purple), and the NF-κB pathway (IDO1, red). MerTK (purple) represents Mer tyrosine kinase, a macrophage apoptotic cell receptor. Nuclei
are stained with POPO-1 nuclear counterstain. Scale bar, 20 µm.

Table 4.3  including platelet-rich plasma (PRP) and stem cell therapies; high-volume
injections; and dry needling. However, few have shown clinical benefit in
Enzymes in extracellular matrix turnover and remodeling placebo-controlled randomized clinical trials. Controlled-motion exercise
Molecule Potential significance therapies are frequently used for the treatment of lower limb tendinopathies.20,21
MMP-3, MMP-10, TIMP-3 Reduced in painful tendons45 Prolonged immobilization is known to have deleterious effects on the quality
ADAM-12, MMP-23 Increased in painful tendons45 of tendon and ligament repair and local tissue mechanics. Further controlled
MMP-7, TIMP-2, TIMP-3, TIMP-4 Reduced in ruptured tendons45 studies tailored toward distinct tendons and ligaments are required to identify
MMP-1, MMP-2, MMP-3, MMP-9, Increased in ruptured tendons45,46 the optimum exercise regimes to promote improved reparative function.
MMP-19, MMP-25, TIMP-1, ADAM-8, NSAIDs are frequently used to manage tendon and ligament pain. The effects
ADAM-12 of cyclooxygenase (COX) inhibition on tissue healing are inconclusive. Total
inflammatory blockade may be potentially deleterious given that inflammation
ADAM, A disintegrin and metalloproteinase; MMP, matrix metalloproteinase; TIMP, tissue inhibitor of is necessary for the local debridement of tissues and initiation of healing.
metalloproteinase. Inflammation also stimulates resolution, which promotes the restoration of
tissue homeostasis after injury.22 COX-2–selective NSAIDs have been shown
to diminish endogenous resolution responses,23,24 which may impair the
tissues’ innate ability to heal. Furthermore, prolonged use of NSAIDs is
endothelial growth factor), growth factors (platelet-derived growth factor), known to have a deleterious effect on collagen synthesis,25 which is likely
and caspases, which have been investigated as potential targets of antifibrotic to influence tissue repair processes. Local injection of corticosteroids is a
drugs.16-19 Further studies are required to improve understanding of the commonly used treatment for tendinopathy with limited evidence to support
mechanisms driving fibrosis in diseased tendons and ligaments to inform its use. Reported deleterious effects associated with corticosteroid treatment
therapeutic target discovery. include reduced tendon cell viability, proliferation, and adverse effects on
tissue mechanical properties.26,27
The use of biologic therapies for tendon and ligament disease has grown
TREATMENT OF TENDON AND significantly in recent years. Biologic therapies such as PRP and stem cells
aim to promote tissue regeneration and improve the quality of repair. Studies
LIGAMENT PATHOLOGY investigating the use of PRP to treat tendinopathy have shown mixed results.
Treatment should be tailored according the type of disease and severity of Although positive outcomes have been reported for prospective and randomized
injury. The presence of bilateral disease should be assessed. The goals of controlled clinical trials,28,29 other randomized controlled clinical trials have
therapy for disorders of tendons and ligaments should be to: suggested that the use of PRP did not significantly improve clinical out-
■ Relieve pain. comes.30,31 The composition of biologic products such as PRP is not standard-
■ Restore function. ized and varies among individuals and preparation methods. Reported
■ Address the underlying tissue pathobiology. deleterious tissue effects of PRP on treated tendons include reduced cellularity
■ Optimize tissue repair and promote return of homeostasis. and vascularity and increased apoptosis.32 The ability of bone marrow–derived
■ Reduce the likelihood of recurrent disease. mesenchymal stem cells to promote tendon repair has been investigated in
athletic horses with tendinopathy.33,34 However, large-scale placebo-controlled
randomized clinical trials are required to investigate the efficacy of these
MEDICAL TREATMENTS treatments in human patients. The current lack of effective treatments
Numerous treatments have been advocated for patients with chronic painful combined with the prolonged rehabilitation and high risk of recurrent disease
tendinopathy. These include physiotherapy; nonsteroidal antiinflammatory compound the difficulties associated with the successful medical management
drugs (NSAIDs); local corticosteroid or sclerosing injections; biologics, of tendon and ligament disorders.

34 SECTION 1  Scientific Basis of Rheumatic Disease
abnormality, which increases susceptibility to recurrent injury. Tendons and
SURGICAL TREATMENTS ligaments are capable of modulating their synthetic activity in response to
Treatment of extrinsic musculoskeletal soft tissue injury as a consequence their chemical and physical environment. The etiology of intrinsic tendon
of external trauma necessitates appropriate surgical repair. Ruptures that disease is complex and multifactorial, including an interplay among the
develop as a result of intrinsic tendon disease may be treated conservatively effects of repetitive wear and tear, daily exercise, and aging. Pathology results
with rest, stabilization, and rehabilitative physiotherapy or may require surgical in substantial and permanent change in the tissue, including increased
repair. Surgical repair of tendon tears such as those in the shoulder are cellularity, increased vascularity, matrix disorganization, and altered composi-
associated with high failure rates.35 Consequently implantation of scaffolds tion. Identification of the phenotypes and interactions among cell populations
to augment tendon repair is common, and there are a variety of scaffolds in in diseased tendons derived from well-defined patient cohorts is essential
current clinical use. These commercially available scaffolds can be biologic to improve understanding of the biologic basis of disease and inform thera-
or synthetic in composition, have a broad range of applications, and are not peutic target discovery. Enhanced understanding of the perceived mismatch
tailored for the repair torn tendons. Novel electrospun nanofiber scaffolds between the presence of structural tendon disease and pain perception is
have been developed to enhance endogenous repair mechanisms through vital to improve patient management. Finally, multicenter placebo-controlled
biophysical cues and are now entering clinical trials. Such bioactive scaffolds randomized clinical trials are required to determine the efficacy of established
have the potential to be used in combination with growth factors and cells.36 treatments for patients with tendon and ligament disorders.
Of critical importance is the effective translation of new therapies from the
laboratory to the clinic, which will require a well-designed program of clinical
trials that are designed to assess both their safety and efficacy.37
The authors acknowledge the contributions of Professor Graham Riley, who
was the author of this chapter in the previous edition.
Injured and aging tendons and ligaments attempt to repair through an
inflammatory process. This invariably results in fibrosis and persistent tissue

1. Gwilym SE, Watkins B, Cooper CD, et al. Genetic 19. Parsons CJ, Takashima M, Rippe RA. Molecular 33. Smith RK, Werling NJ, Dakin SG, et al. Beneficial effects
influences in the progression of tears of the rotator cuff. mechanisms of hepatic fibrogenesis. J Gastroenterol of autologous bone marrow-derived mesenchymal stem
J Bone Joint Surg Br. 2009;91(7):915-917. Hepatol. 2007;22(suppl 1):S79-S84. cells in naturally occurring tendinopathy. PLoS ONE.
2. Collins M, Raleigh SM. Genetic risk factors for 20. van der Plas A, de Jonge S, de Vos RJ, et al. A 5-year 2013;8(9):e75697.
musculoskeletal soft tissue injuries. Med Sport Sci. follow-up study of Alfredson’s heel-drop exercise 34. Godwin EE, Young NJ, Dudhia J, et al. Implantation of
2009;54:136-149. programme in chronic midportion Achilles tendinopathy. bone marrow-derived mesenchymal stem cells
3. Dakin SG, Martinez FO, Yapp C, et al. Inflammation Br J Sports Med. 2012;46(3):214-218. demonstrates improved outcome in horses with overstrain
activation and resolution in human tendon disease. Sci 21. van Ark M, Cook JL, Docking SI, et al. Do isometric and injury of the superficial digital flexor tendon. Equine Vet J.
Transl Med. 2015;7(311):311ra173. isotonic exercise programs reduce pain in athletes with 2012;44(1):25-32.
4. Kragsnaes MS, Fredberg U, Stribolt K, et al. Stereological patellar tendinopathy in-season? A randomised clinical 35. Carr AJ, Cooper CD, Campbell MK, et al. Clinical
quantification of immune-competent cells in baseline trial. J Sci Med Sport. 2016;19(9):702-706. effectiveness and cost-effectiveness of open and
biopsy specimens from Achilles tendons: results from 22. Serhan CN, Hamberg M, Samuelsson B. Lipoxins: novel arthroscopic rotator cuff repair [the UK Rotator Cuff
patients with chronic tendinopathy followed for more series of biologically active compounds formed from Surgery (UKUFF) randomised trial]. Health Technol
than 4 years. Am J Sports Med. 2014;42(10):2435-2445. arachidonic acid in human leukocytes. Proc Natl Acad Sci Assess. 2015;19(80):1-218.
5. Millar NL, Hueber AJ, Reilly JH, et al. Inflammation is U S A. 1984;81(17):5335-5339. 36. Hakimi O, Mouthuy PA, Zargar N, et al. A layered
present in early human tendinopathy. Am J Sports Med. 23. Gilroy DW, Colville-Nash PR, Willis D, et al. Inducible electrospun and woven surgical scaffold to enhance
2010;38(10):2085-2091. cyclooxygenase may have anti-inflammatory properties. endogenous tendon repair. Acta Biomater.
6. Millar NL, Murrell GA, McInnes IB. Alarmins in Nat Med. 1999;5(6):698-701. 2015;26:124-135.
tendinopathy: unravelling new mechanisms in a common 24. Gilroy DW, Lawrence T, Perretti M, et al. Inflammatory 37. Stace ET, Dakin SG, Mouthuy PA, et al. Translating
disease. Rheumatology (Oxford). 2013;52(5):769-779. resolution: new opportunities for drug discovery. Nat Rev regenerative biomaterials into clinical practice. J Cell
7. Millar NL, Gilchrist DS, Akbar M, et al. MicroRNA29a Drug Discov. 2004;3(5):401-416. Physiol. 2016;231(1):36-49.
regulates IL-33-mediated tissue remodelling in tendon 25. Christensen B, Dandanell S, Kjaer M, et al. Effect of 38. Gotoh M, Hamada K, Yamakawa H, et al. Interleukin-1-
disease. Nat Commun. 2015;6:6774. anti-inflammatory medication on the running-induced induced glenohumeral synovitis and shoulder pain in
8. Murray PJ, Allen JE, Biswas SK, et al. Macrophage rise in patella tendon collagen synthesis in humans. rotator cuff diseases. J Orthop Res. 2002;20(6):1365-1371.
activation and polarization: nomenclature and J Appl Physiol. 2011;110(1):137-141. 39. Ko JY, Wang FS, Huang HY, et al. Increased IL-1beta
experimental guidelines. Immunity. 2014;41(1):14-20. 26. Dean BJ, Lostis E, Oakley T, et al. The risks and benefits expression and myofibroblast recruitment in subacromial
9. Ackermann PW, Salo PT, Hart DA. Neuronal pathways in of glucocorticoid treatment for tendinopathy: a systematic bursa is associated with rotator cuff lesions with shoulder
tendon healing. Front Biosci. 2009;14:5165-5187. review of the effects of local glucocorticoid on tendon. stiffness. J Orthop Res. 2008;26(8):1090-1097.
10. Ackermann PW, Franklin SL, Dean BJ, et al. Neuronal Semin Arthritis Rheum. 2014;43(4):570-576. 40. Legerlotz K, Jones ER, Screen HR, et al. Increased
pathways in tendon healing and tendinopathy—update. 27. Metcalfe D, Achten J, Costa ML. Glucocorticoid injections expression of IL-6 family members in tendon pathology.
Front Biosci. 2014;19:1251-1278. in lesions of the achilles tendon. Foot Ankle Int. Rheumatology (Oxford). 2012;51:11611165.
11. Schizas N, Weiss R, Lian O, et al. Glutamate receptors in 2009;30(7):661-665. 41. Campbell AL, Smith NC, Reilly JH, et al. IL-21 receptor
tendinopathic patients. J Orthop Res. 28. Randelli P, Arrigoni P, Ragone V, et al. Platelet rich plasma expression in human tendinopathy. Mediators Inflamm.
2012;30(9):1447-1452. in arthroscopic rotator cuff repair: a prospective RCT 2014;2014:481206.
12. Dean BJ, Snelling SJ, Dakin SG, et al. Differences in study, 2-year follow-up. J Shoulder Elbow Surg. 42. Tsuzaki M, Bynum D, Almekinders L, et al. ATP
glutamate receptors and inflammatory cell numbers are 2011;20(4):518-528. modulates load-inducible IL-1beta, COX 2, and MMP-3
associated with the resolution of pain in human rotator 29. Rha DW, Park GY, Kim YK, et al. Comparison of the gene expression in human tendon cells. J Cell Biochem.
cuff tendinopathy. Arthritis Res Ther. 2015;17:176. therapeutic effects of ultrasound-guided platelet-rich 2003;89(3):556-562.
13. Dean BJ, Gwilym SE, Carr AJ. Why does my shoulder plasma injection and dry needling in rotator cuff disease: 43. Tsuzaki M, Guyton G, Garrett W, et al. IL-1 beta induces
hurt? A review of the neuroanatomical and biochemical a randomized controlled trial. Clin Rehabil. COX2, MMP-1, -3 and -13, ADAMTS-4, IL-1 beta and
basis of shoulder pain. Br J Sports Med. 2013;27(2):113-122. IL-6 in human tendon cells. J Orthop Res.
2013;47(17):1095-1104. 30. Castricini R, Longo UG, De Benedetto M, et al. 2003;21(2):256-264.
14. Riley G. The pathogenesis of tendinopathy. A molecular Platelet-rich plasma augmentation for arthroscopic rotator 44. Goodier HC, Carr AJ, Snelling SJ, et al. Comparison of
perspective. Rheumatology (Oxford). 2004;43(2): cuff repair: a randomized controlled trial. Am J Sports transforming growth factor beta expression in healthy and
131-142. Med. 2011;39(2):258-265. diseased human tendon. Arthritis Res Ther.
15. Riley G. Tendinopathy–from basic science to treatment. 31. Kesikburun S, Tan AK, Yilmaz B, et al. Platelet-rich 2016;18(1):48.
Nat Clin Pract Rheumatol. 2008;4(2):82-89. plasma injections in the treatment of chronic rotator 45. Jones GC, Corps AN, Pennington CJ, et al. Expression
16. Wynn TA. Fibrotic disease and the T(H)1/T(H)2 cuff tendinopathy: a randomized controlled trial with profiling of metalloproteinases and tissue inhibitors of
paradigm. Nat Rev Immunol. 2004;4(8):583-594. 1-year follow-up. Am J Sports Med. 2013;41(11): metalloproteinases in normal and degenerate human
17. Li MO, Wan YY, Sanjabi S, et al. Transforming growth 2609-2616. achilles tendon. Arthritis Rheum. 2006;54(3):832-842.
factor-beta regulation of immune responses. Annu Rev 32. Carr AJ, Murphy R, Dakin SG, et al. Platelet-rich plasma 46. Riley GP, Curry V, DeGroot J, et al. Matrix
Immunol. 2006;24:99-146. injection with arthroscopic acromioplasty for chronic metalloproteinase activities and their relationship with
18. Wynn TA. IL-13 effector functions. Annu Rev Immunol. rotator cuff tendinopathy: a randomized controlled trial. collagen remodelling in tendon pathology. Matrix Biol.
2003;21:425-456. Am J Sports Med. 2015;43(12):2891-2897. 2002;21(2):185-195.

Connective tissue responses to mechanical stress
Donald M. Salter 5  

Key Points In adult bone, peak strain magnitudes several thousand times the magnitude
■ Bone, cartilage, tendon, skeletal muscle, and other connective tissue require
are encountered during vigorous exercise compared with that seen during
exposure to physiologic levels of mechanical stress to remain healthy.
normal activity.8 Cartilage and tendon are exposed to loads many times
greater than body weight and that may be up to five times that seen with
■ Moderate exercise improves the strength and function of connective tissue.
normal daily activity.9
■ Withdrawal of mechanical stimuli or prolonged exposure to excessive mechanical
load results in tissue degeneration and the development of diseases such as
osteoarthritis. MECHANOCOUPLING
■ Mechanotransduction is the process by which a mechanical signal is recognized by Connective tissue cells are not directly exposed to mechanical forces because
connective tissue cells and translated into biochemical and molecular responses. they are surrounded by extracellular matrix (ECM), which acts to absorb,
■ Connective tissue cells express mechanoreceptors that allow the recognition of transmit, and dissipate these forces. Chondrocytes, for instance, are embedded
physiochemical changes that occur within connective tissue during mechanical within a specialized, type VI collagen pericellular matrix, and the interterritorial
loading. matrix of cartilage is able to withstand high compressive and shearing loads
■ Activation of mechanoreceptors generates secondary messenger molecules and because of a type II collagen network that imparts tensile strength and a
activation of signaling events in cells that lead to tissue remodeling. high proteoglycan content that resists compression. As such, the mechanical
forces encountered directly by cells in most connective tissues are not known
exactly, but they are likely to result in a range of physiochemical changes,
including hydrostatic pressure gradients, fluid flow, streaming potentials,
INTRODUCTION alterations in matrix water content, fixed charge density, mobile ion concentra-
tions, pH and osmotic pressure, and cell deformation.10 All of these changes
Exposure to physiologic mechanical forces is important for the maintenance have been shown to influence cell behavior. Loading of bone induces shear,
of most, if not all, connective tissues, including bone, cartilage, tendon, and strain, and compressive forces, with up to 5000 microstrain generated at the
skeletal muscle.1-4 Movement and mechanical stimulation are essential for osteocyte cell surface11 and the development of pressure gradients that stimulate
normal embryonic development and morphogenesis of the skeleton and the flow of interstitial fluid and drag forces through the osteocytic lacunar
joints. Wolff’s law states that bone in a healthy person will adapt to the loads network.12 In tendon, the complex three-dimensional structure13 results in
under which it is placed. When loading increases, bone remodels and becomes a tensile load at the macroscopic level being converted into tensile, shear,
stronger, but when loading decreases, bone atrophies because the stimulus and compressive strains encountered by the tenocyte.
for continued remodeling, required to maintain bone mass, is reduced. This
ability to alter structure in response to mechanical loading is shared with
all connective tissue and is referred to as tissue mechanical adaptation.
Mechanical loading of connective tissue within a physiologic range met Connective tissue cells express mechanoreceptors that recognize the physi-
during normal, everyday activity is sufficient to keep connective tissue ochemical changes that occur within connective tissue during mechanical
healthy. Clinical observations and in vivo studies show that moderate exercise loading. Activation of these receptors results in the stimulation of intracellular
increases connective tissue performance. Changes in skeletal muscle bulk and signaling cascades that regulate gene expression, changes in protein expression,
strength with exercise are obvious, but other connective tissues are similarly and tissue remodeling (Fig. 5.2). Molecules and molecular complexes that
affected. Differences in bone mass in the dominant and nondominant arms recognize physiochemical changes, including alterations in cell shape, cell
of tennis players are a consequence of differences in applied mechanical load membrane deformation, local ion concentration, and fluid flow, are candidate
to either limb. The mechanical strength and collagen content of tendons are mechanoreceptors. These include integrins, mechanosensitive ion channels,
increased by physical exercise, and exercise promotes proteoglycan synthe- connexins, and the LINC (linker of nucleoskeleton and cytoskeleton)
sis and increases the thickness of articular cartilage and hence improves complex.14-16 These are considered in detail below. In addition, mechanical
load-bearing capacity. In contrast, withdrawal of mechanical stimuli results forces can also be transmitted directly from the cell membrane to the nucleus
in connective tissue atrophy, as evident in such diverse clinical situations without the intermediate step of activation of intracellular signal cascades,
as paraplegia, patients undergoing prolonged bed rest, and astronauts. allowing more immediate sensing and response to mechanical
The bones in paralyzed and underused limbs become architecturally and stimulation.17
mechanically inferior to those subjected to normal mechanical loading.
Similarly, immobilization results in the major loss of mechanical properties
of tendons, and, significantly, both overloading and underloading of articular
cartilage are associated with loss of cartilage matrix and the development Integrins
of osteoarthritis.5 Integrins are heterodimeric transmembrane glycoproteins that contain α and
For tissue adaptation to occur in response to applied mechanical forces, β subunits, each consisting of an extracellular domain, a single transmembrane
cells in connective tissue need to receive information from their environment region, and a cytoplasmic tail. The extracellular domain provides the binding
and modify tissue structure appropriately by initiating anabolic or catabolic site for ligands, including fibronectin and collagen. The cytoplasmic tail is
pathways which result in changes in cell structure and extracellular matrix coupled to the cytoskeletal network by focal adhesions, or specialized large
content to best withstand these mechanical stresses (Fig. 5.1). This process macromolecular assemblies that include actin-associated proteins and signaling
is readily divided into four major phases from initial mechanical loading of molecules such as focal adhesion kinase, allowing integrins to transduce
tissues to the ultimate goal of tissue adaptation (Table 5.1).6 Most connective mechanical signals transmitted through the matrix into biochemical responses
tissue cells are mechanosensitive. In cartilage, chondrocytes both sense within the cell.18 The cytoskeleton also responds to forces channeled through
mechanical stimuli and act as effector cells that produce matrix macromolecules integrins by rearranging its interlinked actin microfilaments, microtubules,
and the proteases that are required for tissue turnover and homeostasis. and intermediate filaments. Mechanoreceptor roles for α5β1, αVβ5, and αVβ3
Tenocytes have a similar role in mechanical tissue adaptation in tendons.7 integrins have been demonstrated in vitro, but β1 integrins appear to be the
In bone, osteocytes appear likely to be the most important mechanosensitive major integrin mechanoreceptors in connective tissue.19,20 Integrins may be
cells for bone mechanical adaptation through paracrine effects on cells of part of larger mechanoreceptor complexes, including accessory molecules
the osteoprogenitor and osteoclast lineage.1,2 Mechanical loads applied to such as CD47, that control integrin activation21 and regulate intracellular
and withstood by different tissues subject to the level of activity (Table 5.2). signaling and gene expression.

36 SECTION 1  Scientific Basis of Rheumatic Disease



Mechanical Mechanical ·Integrins
loading and stress signal ·Ion channels

Intracellular signaling Autocrine/paracrine

·Intracellular calcium mediators
·cAMP, PLC, PKC, ·Prostaglandins, ATP
Connective ·PI3-kinase/Akt ·Neuropeptides
tissue ·MAP kinases ·Cytokines, growth factors

Transcription factor activation

Tissue Recognition of
adaptation mechanical load
Regulation of gene expression

FIG. 5.2  Mechanical stimuli are recognized by mechanoreceptors that activate a

Extracellular matrix range of intracellular signal pathways, either directly or through autocrine and paracrine
Anabolic or catabolic
and cellular signaling. This leads to changes in gene expression and tissue remodeling. ATP,
remodeling Adenosine triphosphate; cAMP, cyclic adenosine monophosphate; MAP, mitogen-
activated protein; PI3K, phosphatidylinositol-3′-kinase; PKC, protein kinase C; PLC,
phospholipase C.

FIG. 5.1  Processes involved in the response of connective tissue to mechanical

Table 5.2 
Tissue loading and activity

Table 5.1  Tissue Mechanical force applied

Sequential connective tissue responses to mechanical loading
Normal activity Peak strain <10 microstrain
Mechanocoupling Translation of macroscopic forces through tissue to Vigorous activity Peak strain 2000–3500 microstrain
the surface of a mechanosensitive cell
Mechanotransduction Hip joint
  Mechanoreceptors Recognition of a mechanical stimulus by a cell Walking 3–4× body weight
mechanoreceptor Running 5× body weight
  Intracellular pathways Activation of intracellular signal pathways to Stumbling 7–10× body weight
regulate gene transcription and protein
production Achilles tendon
Signal propagation and Autocrine or paracrine signaling between Walking 4× body weight
regulation of response mechanosensitive cells and effector cells Running 16–20× body weight
Tissue adaptation Extracellular matrix modeling and remodeling
Intervertebral disk

Standing 0.5 MPa
Lifting a 20-kg weight (flexed back) 2.3 Mpa
Squat lifting body weight 6–10× body weight
Stretch-activated or stretch-sensitive ion channels
Stretch-activated or stretch-sensitive ion channels (SACs) open on mechanical
deformation of the cell membrane. 15 Potassium-selective channels, the
Shaker-IR K+ channel, the N-type Ca2+ channel, ionotropic N-methyl-D-aspartate
receptors (NMDARs), and Ca2+-dependent BK channels have all been shown protein gene expression in bone cells. NMDARs are ligand-gated ion channels
to act as SACs. SACs are directly activated by mechanical forces applied that may be involved in chondrocyte mechanotransduction.22
along the plane of the cell membrane. These forces induce membrane tension
and distortion of the cell membrane lipid bilayer, which results in confor- Primary cilia
mational changes that alter opening or closing rates and allow ion flux. Primary cilia are solitary, immotile cilia present in most cells, including
Hydrostatic pressure, in which mechanical forces are applied perpendicular chondrocytes and bone cells, which grow from the centrosome and extend
to the cell membrane, appears to be less effective in activating SACs.19 Opening from the cell surface. They contain a variety of cell membrane receptors
of calcium-permeable SACs increases intracellular calcium levels and stimula- such as integrins and function as chemosensors or mechanosensors (or
tion of downstream calcium-dependent intracellular signal cascades. SACs both).23,24 Matrix or cell deformation induces bending of the cilium and
sensitive to gadolinium are necessary for load-related increases in prostaglandin activation of integrins, SACs, and connexin hemichannels,25 thereby leading
and nitric oxide (NO) production, cytoskeletal reorganization, and changes to downstream signaling and changes in gene expression.
in gene expression in response to fluid flow of bone cells and responses of
chondrocytes to both strain and compression. The effects of different forms The LINC complex
of mechanical stress may be regulated by the activity of voltage-gated or The LINC complex is a complex of nesprins and SUN (Sad1 and UNC-84
ligand-gated ion channels. Whereas L-type calcium channels have roles in homology domain) family of proteins that physically link the nucleus and
bone cell responses to fluid flow rather than to strain, K+ channels of the the cytoskeleton. Nesprins span the outer nuclear membrane and interact
TREK family but not gadolinium- or nifedipine-sensitive ion channels are with SUN proteins within the inner nuclear membrane and cytoskeletal
important for stretch-induced elevation of parathyroid hormone–related proteins, including actin and intermediate filaments in the cytoplasm.26

CHAPTER 5  Connective tissue responses to mechanical stress 37

NF-κB appears to be relevant for control of inflammatory responses to

INTRACELLULAR SIGNALING PATHWAYS mechanical loads. Dynamic biomechanical signals of low physiologic magnitude
After recognition of the mechanical stimulus by mechanoreceptors, secondary block NF-κB activity and are potent antiinflammatory signals in bone and
messenger molecules are generated, and cascades of downstream signaling cartilage.30,31 Conversely, high-amplitude dynamic stimulation (a catabolic
events are initiated that ultimately lead to changes in gene transcription and stimulus) induces rapid nuclear translocation of NF-κB subunits p65 and
expression of matrix molecules and proteases involved in tissue remodeling p50 in a similar manner to interleukin-1β (IL-1β).
and hence tissue structure and function. A number of pathways are involved
in the signal transduction response, including G proteins, phosphatidyinositol-3′-
kinase, protein kinase B, protein kinase C, and mitogen-activated protein
kinases (Fig. 5.3). Mechanical forces applied directly to the nucleus via the Cellular mechanotransduction induces tissue remodeling activity in a cell
LINC complex may also transduce mechanical signals into biochemical by direct activation of gene-modifying signal cascades in that mechanosensitive
signaling pathways that regulate phosphorylation of nucleoskeletal proteins cell. However, it is clear that mechanotransduction also results in stimulation
such as emerin and lamin A to C. of release of a range of autocrine and paracrine signaling molecules, including
prostaglandins, NO, cyclic nucleotides, cytokines, and growth factors, which
can propagate the mechanical response within tissues and by regulating the
G PROTEINS degree and nature of the response of resident cells help orchestrate the overall
G proteins are associated with a variety of cell-surface (G protein–coupled) tissue response to mechanical stimulation (Fig. 5.4).
receptors and, depending on the subclass, activate different intracellular
signal cascades. Whereas Gs induces the formation of cyclic adenosine
monophosphate and activation of protein kinase A, Gq and Go activate
phospholipase C, which leads to the production of inositol–triphosphate Connexins, especially Cx43, are widely expressed in bone, tendon, and
and diacylglycerol with release of calcium from intracellular storage sites meniscus.32-34 By forming gap junctions and hemichannels, connexins allow
and activation of protein kinase C (PKC). diffusion of ions, metabolites, and small signaling molecules such as cyclic
nucleotides and inositol derivatives between the interior of cells. This allows
propagation of a mechanical stimulus between a network of interconnected
PROTEIN KINASE B/AKT cells, facilitating a tissue-level response to mechanical load as seen in shear
Protein kinase B (PKB)/Akt is a serine/threonine kinase activated by stress–induced remodeling in bone and collagen production in response to
phosphatidylinositol-3′-kinase (PI3K). Mechanical forces activate PI3K through strain in tendons.35,36
integrin stimulation (see Fig. 5.3). Inactivation of the PI3K/PKB pathway
may be important in the deleterious effects of mechanical overloading of
cartilage and bone or tissue atrophy in response to withdrawal of loading.27
Prostaglandins, NO, and cyclic nucleotides are produced after mechanical
MITOGEN-ACTIVATED PROTEIN KINASES AND NUCLEAR stimulation of most connective tissue cells and tissues. In bone, prostaglandin
E2 (PGE2) and NO are rapidly produced and are required for an anabolic
FACTOR κB response. In contrast, in cartilage, PGE2 production is catabolic, although
Mitogen-activated protein kinases and nuclear factor κB (NF-κB) regulate production is only seen under pathologic loading, with physiologic mechanical
cellular activities such as gene expression, mitosis, differentiation, and cell loading inhibiting synthesis of both PGE2 and NO.37 Mechanical loading–
survival and apoptosis. ERK1/2, JNK, and p38 are activated in chondrocytes, induced release of adenosine triphosphate by bone cells and chondrocytes
bone cells, and fibroblasts2,28,29 after mechanical stimulation and are of critical acts through the metabotropic P2Y receptors and P2X receptors expressed
importance in the regulation of matrix protein and protease gene expression. by the cells to further propagate and regulate the cellular response.


Mechanical force
Extracellular matrix Integrins Mechanical
Ca Mechanosensitive cell
Integrins ATP ATP
CD47 Stretch activated NK1
ion channel Cell Hemichannel
Connexin IP3


Ca2+ Gap junction


Cytoskeleton acting molecules

PKC PI3K MAPK NF-κB Effector cell


IL-4 IL-1β
Substance P
complex TGF-β PGE2
Nuclear laminis Gene Effector cell
Nucleus NO ATP
chromatin transcription IL-4R TGF-βR

FIG. 5.3  Cell surface integrins and stretch-activated ion channels stimulated by a FIG 5.4  A mechanosensitive cell can propagate a mechanical response to adjacent
mechanical force initiate intracellular signaling events by activating focal adhesion cells through the production and release of a range of molecules. Some, such as
kinase (FAK) and increasing intracellular calcium. These in turn activate a variety inositol triphosphate (IP3), may be passed from the mechanosensitive cells directly
of downstream pathways, including protein kinase C (PKC), phosphatidylinositol- into the cytoplasm of adjacent cells via gap junctions, but others, including interleukin-4
3′-kinase (PI3K) mitogen-activated protein kinases (MAPKs), and nuclear factor κB (IL-4), interleukin-1β (IL-1β), and prostaglandins, are secreted and have both autocrine
(NF-κB)–dependent cascades, that regulate gene transcription. Mechanical deformation and paracrine activity. ATP, Adenosine triphosphate; IP3, inositol triphosphate; NK1,
of the actin cytoskeleton can directly regulate cell behavior via the LINC (linker of Neurokinin 1; NO, nitric oxide; P2R, ATP receptor; PGE2, prostaglandin E2; TGF,
nucleoskeleton and cytoskeleton) complex. transforming growth factor.

38 SECTION 1  Scientific Basis of Rheumatic Disease
layer of complexity by which the cellular and tissue response to mechanical
CYTOKINE AND GROWTH FACTORS loading can be controlled.
An increasing array of cytokines and growth factors, which are locally produced
after mechanical stimulation, are now recognized as having pivotal roles in
connective tissue adaptation to mechanical loading. Interleukin 4 (IL-4) and
IL-1β autocrine and paracrine activity are seen in the integrin-dependent
mechanotransduction cascade of chondrocytes (IL-4 and IL-1β) and bone Individuals are exposed to a wide range of mechanical stress during daily
cells (IL-1β). In chondrocytes, release of IL-4 relies on secretion of the activity, which is required to maintain connective tissue, including bone,
neuropeptide substance P, which binds to its NK1 receptor. Both IL-4 and cartilage, and tendons, in a healthy state. Moderate exercise can improve
substance P are necessary but not sufficient for the increased expression of the strength and function of connective tissue, but withdrawal of mechanical
aggrecan mRNA and the decrease in matrix metalloproteinase-3 mRNA stimuli or prolonged exposure to excessive mechanical loads results in tissue
induced by the mechanical stimulus, suggesting crosstalk with other mecha- degeneration and the development of diseases such as osteoarthritis. The
nosensitive signaling pathways. IL-1β is involved in the early mechanotransduc- mechanisms by which mechanical loading regulates connective tissue cell
tion pathway of both osteoarthritic chondrocytes and human trabecular function are beginning to be understood. Identification of mechanoreceptors
bone–derived cells. Loading of connective tissues may also influence tissue and mechanotransduction pathways will increase our knowledge of how
structure through production of proinflammatory chemokines such as abnormal mechanical environments have detrimental effects on connective
monocyte chemoattractive protein-1 and by inducing release or activation tissue. This increased knowledge will facilitate the development of novel
of ECM sequestered growth factors, including transforming growth factor-β1 therapeutic interventions such as mechanomimetics and pathway inhibitors
and fibroblast growth factor-2. Regulation of growth factor receptor activation to mitigate the adverse mechanical effects on connective tissue structure
and signaling, whether by integrins or other mechanisms, creates a further and function that occur with aging and in disease.

1. Turner CH. Bone strength: current concepts. Ann N Y 15. Martinac B. Mechanosensitive ion channels: molecules of signaling. Am J Physiol Endocrinol Metab.
Acad Sci. 2006;1068:429-446. mechanotransduction. J Cell Sci. 2004;17:2449-2460. 2008;294:E794-E801.
2. Rubin J, Rubin C, Jacobs CR. Molecular pathways 16. Müller U. Cadherins and mechanotransduction by hair 27. Fitzgerald JB, Jin M, Chai DH, et al. Shear- and
mediating mechanical signaling in bone. Gene. cells. Curr Opin Cell Biol. 2008;20:557-566. compression-induced chondrocyte transcription requires
2006;367:1-16. 17. Katsumi A, Orr AW, Tzima E, et al. Integrins in MAPK activation in cartilage explants. J Biol Chem.
3. Grodzinsky AJ, Levenston ME, Jin M, et al. Cartilage mechanotransduction. J Biol Chem. 2008;283:6735-6743.
tissue remodeling in response to mechanical forces. Annu 2004;279:12001-12004. 28. Jeon YM, Kook SH, Son YO, et al. Role of MAPK in
Rev Biomed Eng. 2000;2:691-713. 18. Lee HS, Millward-Sadler SJ, Wright MO, et al. Integrin mechanical force–induced up-regulation of type I collagen
4. Wang JH. Mechanobiology of tendon. J Biomech. and mechanosensitive ion channel–dependent tyrosine and osteopontin in human gingival fibroblasts. Mol Cell
2006;39:1563-1582. phosphorylation of focal adhesion proteins and Biochem. 2009;320:45-52.
5. Jortikka MO, Inkinen RI, Tammi MI, et al. Proteoglycan beta-catenin in human articular chondrocytes after 29. Knapik DM, Perera P, Nam J, et al. Mechanosignaling in
alterations in rabbit knee articular cartilage following mechanical stimulation. J Bone Miner Res. bone health, trauma and inflammation. Antioxid Redox
physical exercise and immobilization. Connect Tissue Res. 2000;15:1501-1509. Signal. 2014;20:970-985.
1983;11:45-55. 19. Phillips JA, Almeida EA, Hill EL, et al. Role for beta1 30. Anghelina M, Sjostrom D, Perera P, et al. Regulation of
6. Duncan RL, Turner CH. Mechanotransduction and the integrins in cortical osteocytes during acute biomechanical signals by NF-kappaB transcription factors
functional response of bone to mechanical strain. Calcif musculoskeletal disuse. Matrix Biol. 2008;27:609-618. in chondrocytes. Biorheology. 2008;45:245-256.
Tissue Int. 1995;57:344-358. 20. Orazizadeh M, Lee HS, Groenendijk B, et al. CD47 31. Jeffrey JE, Aspden RM. Cyclooxygenase inhibition lowers
7. Screen HRC, Shelton JC, Lee DA, et al. Cyclic tensile associates with alpha 5 integrin and regulates responses prostaglandin E2 release from articular cartilage and
strain upregulates collagen synthesis in isolated tendon of human articular chondrocytes to mechanical reduces apoptosis but not proteoglycan degradation
fascicles. Biochem Biophys Res Commun. 2005;336:424-429. stimulation in an in vitro model. Arthritis Res Ther. following an impact load in vitro. Arthritis Res Ther.
8. Rubin CT, Lanyon LE. Regulation of bone formation by 2008;10:R4. 2007;9:R129.
applied dynamic loads. J Bone Joint Surg Am. 21. Salter DM, Wright MO, Millward-Sadler SJ. NMDA 32. Ralphs JR, Benjamin M, Waggett AD, et al. Regional
1984;66:397-402. receptor expression and roles in human articular differences in cell shape and gap junction expression
9. Giddings VL, Beaupré GS, Whalen RT, et al. Calcaneal chondrocyte mechanotransduction. Biorheology. in rat Achilles tendon: relation to fibrocartilage
loading during walking and running. Med Sci Sports 2004;41:273-281. differentiation. J Anat. 1998;193:215-222.
Exerc. 2000;32:627-634. 22. Jiang JX, Siller-Jackson AJ, Burra S. Roles of gap junctions 33. Hellio Le Graverand MP, Ou Y, Schield-Yee T, et al.
10. Nuki G, Salter DM. The impact of mechanical stress on and hemichannels in bone cell functions and in signal The cells of the rabbit meniscus: their arrangement,
the pathophysiology of osteoarthritis. In: Sharma L, transmission of mechanical stress. Front Biosci. interrelationship, morphological variations and
Berenbaum F, eds. Osteoarthritis: A companion to 2007;12:1450-1462. cytoarchitecture. J Anat. 2001;198:525-535.
rheumatology. St. Louis: Elsevier; 2007:33-52. 23. Poole CA, Jensen CG, Snyder JA, et al. Confocal analysis 34. Thi MM, Kojima T, Cowin SC, et al. Fluid shear stress
11. Han Y, Cowin SC, Schaffler MB, et al. of primary cilia structure and colocalization with the remodels expression and function of junctional proteins
Mechanotransduction and strain amplification in Golgi apparatus in chondrocytes and aortic smooth in cultured bone cells. Am J Physiol Cell Physiol.
osteocyte cell processes. Proc Natl Acad Sci U S A. muscle cells. Cell Biol Int. 1997;21:483-494. 2003;284:C389-C403.
2004;101:16689-16694. 24. Malone AM, Anderson CT, Tummala P, et al. Primary cilia 35. Waggett AD, Benjamin M, Ralphs JR. Connexin 32 and
12. Burger EH, Klein-Nulend J. Mechanotransduction in mediate mechanosensing in bone cells by a 43 gap junctions differentially modulate tenocyte
bone—role of the lacuno-canalicular network. FASEB J. calcium-independent mechanism. Proc Natl Acad Sci response to cyclic mechanical load. Eur J Cell Biol.
1999;13:S101-S112. U S A. 2007;104:13325-13330. 2006;85:1145-1154.
13. Screen HRC, Lee DA, Bader DL, et al. Development of a 25. Alam S, Lovett DB, Dickinson RB, et al. Nuclear forces 36. Whitfield JF. The solitary (primary) cilium—a
technique to determine strains in tendons using the cell and cell mechanosensing. Prog Mol Biol Transl Sci. mechanosensory toggle switch in bone and cartilage
nuclei. Biorheology. 2003;40:361-368. 2014;126:205-215. cells. Cell Signal. 2008;20:1019-1024.
14. Millward-Sadler SJ, Salter DM. Integrin-dependent signal 26. Dufour C, Holy X, Marie PJ. Transforming growth 37. Salter DM, Millward-Sadler SJ, Nuki G, et al. Integrin-
cascades in chondrocyte mechanotransduction. Ann factor-beta prevents osteoblast apoptosis induced by interleukin-4 mechanotransduction pathways in human
Biomed Eng. 2004;32:435-446. skeletal unloading via PI3K/Akt, Bcl-2, and phospho-Bad chondrocytes. Clin Orthop Relat Res. 2001;391:S49-S60.

Biomechanics of peripheral joints
David E. Williams • Cathy Holt 6  

Key Points or stereophotogrammetric techniques. This motion can involve one joint
■ Diarthrodial joints in the body act as levers with varying levels of mechanical
surface that is just rolling on the other, one surface sliding on another, or a
advantage depending on their dynamic nature and load-bearing capacity.
combination of the two (gliding). Single radiographs can provide the descrip-
tion of a relative uniplanar motion of two adjacent bones.
■ Surface joint motion is a combination of rolling, sliding, and gliding.
As one bone rotates about another bone, at any instant, there is a stationary
■ Joints can be classified based on rotational and translational degrees of freedom of
point (i.e., zero velocity), which is defined as the instantaneous center of
rotation (ICR). The ICR can be found by identifying displacement of two
■ Joint lubrication is a combination of boundary lubrication, fluid film lubrication, or a points on a bone as it moves from one position to another in relation to the
mixture depending on the joint loading and relative movement of the joint surfaces. other bone, which is considered stationary. Therefore, the ICR of a joint can
■ Fluid film lubrication supports contact loads through the pressure developed in the be identified using a series of successive radiographs of the joint in different
film. poses. The ICR pathway can also be used to describe the relative movement
■ Boundary lubrication protects surfaces via a layer of lubricant, preventing surface-to- between contact points of bones, or joint contacts.1 For a particular joint in
surface contact. a single plane of movement, the ICR is a parameter that can describe the
■ Many tissues in the body are nonhomogeneous and anisotropic, behaving differently effect of a pathology on joint function (e.g., the screw home mechanism of
when loaded in different locations and directions. the knee), and this may also be indicative of stress concentrations in the
■ Many tissues in the body are viscoelastic in response to load, and this depends on AC and joint distraction, leading to focal damage and stretching of
structural composition and fluid content. ligaments.
■ A characteristic stress–strain curve of collagenous tissues (e.g., tendon and A joint’s ROM is defined by its form and function. The joint may provide
ligament) is an initial toe region followed by a steep linear region before failure. musculoskeletal stability during motion and act as a fulcrum between two
lever arms (e.g., the knee between the femoral and tibial long bones). It
allows mobility in terms of joint rotations and translations and can be a
contributor to conservation of momentum of ADLs (e.g., the knee flexes
and extends to limit the up and down movement of a person’s center of
INTRODUCTION mass during walking).
Biomechanics is the application of mechanical engineering principles to Range of motion can be described in engineering terms with respect to
living organisms. It can be examined at different levels, including the cellular the degree of freedom (DOF) of movement of a joint. The DOF in the context
level (e.g., response of cells to an externally applied force or deformation), of whole-joint movement refers to the translations and rotations of a rigid
tissue level (e.g., strain of the anterior cruciate ligament during normal gait), body in anatomic planes. Generally, the relative motion of one human body
and whole-joint level (e.g., joint movement and contact forces during activities segment (e.g., bone), with regard to an adjacent one can involve translation,
of daily living [ADLs]). rotation, or both. For example, a hip joint is a ball and socket joint that is
Engineering principles may be used to understand the causes and progres- inherently stable and has three rotational DOF; alternatively, the tibiofemoral
sions of many musculoskeletal diseases. A basic understanding of these joint of the knee is much more complex and unstable with three rotational
principles is beneficial for clinicians and medical professionals, and it facilitates and three translational DOF (6DOF in total) (Fig. 6.1). However, if the
a multidisciplinary approach to understanding the biologic mechanisms of object of the study is gross human motion, such as walking or exercise, the
disease that may be driven by mechanical factors. This chapter presents a translations can often be disregarded because of their small magnitude (which
comprehensive introduction of engineering mechanics pertaining to whole-joint can be difficult to quantify accurately), compared with rotations, thus allowing
and tissue mechanics in the human body. joint motion to be analyzed as pure rotation. In describing joint ROM, it is
important to understand the joint configuration. Simplifying assumptions
can be made when joint motions are considered as pure rotations around
BIOMECHANICS OF WHOLE JOINTS fixed axes, that is, the axes of rotation intersect at one point (if the joint
The effects of a musculoskeletal disease are often seen at the microscopic exhibits rotation around more than one axis), coinciding with a joint reference
level, but the symptoms of the disease are commonly found at the whole-joint coordinate system which usually matches the anatomic axes (e.g., for the
level. For example, the symptoms of osteoarthritis, including joint pain, knee, the flexion–extension axis should coincide with the frontal axis passing
stiffness, and reduced range of motion, are accompanied by the formation through joint center2,3).
of clefts and fissures in articular cartilage (AC) along with bony spurs Methods used to measure and quantify joint configuration differ primarily
(osteophytes) and joint space closing. To understand and quantify functional in the manner in which the local reference frame is fixed within a body
capabilities of a joint and how they change during the onset and progression segment. All local reference systems are classified as either technical or
of rheumatic disease and in response to treatment, it is important to examine somatic.4,5
the movement and transmission of forces through the synovial joint. This ■ Technical reference systems are fixed with technical devices (i.e., goni-
must include an appreciation of the construct stiffness and the response of ometer, inertial movement sensor, or accelerometer), located externally
individual tissues in the joint. This chapter concentrates on the biomechanics on the body surface rather than fixed directly at the origin of the body
of the whole joint in terms of movement and loading. Biomechanics can be segment local reference frame.
applied to understand the movement of the person, the individual joints, ■ Somatic reference systems can be defined in three ways (Fig. 6.2). Also,
or the bones, along with the effects of the forces and moments acting on a movement pattern that is considered as a pure rotation around one
them. Two types of analyses can be applied: reference axis in one frame may be expressed as a complex general
1. Kinematics, which is the study of movement of a rigid body motion in another reference system.6 The Clinical Reference System
2. Kinetics, which is the study of forces applied to the body (CRS) is most often used by practitioners and is defined as the sagit-
tal, frontal, and transverse planes defined in the anatomic position.
Segmental motion can be defined as (1) motion in the sagittal plan
WHOLE-JOINT MOVEMENT (flexion–extension), (2) motion away from or toward the midsagittal
Kinematics can be broken down further into surface joint motion (SJM) plane (abduction–adduction), (3) rotation about the long axis (internal–
and range of motion (ROM), and both of these can be defined in the three external rotation or supination–pronation depending on the joint), or (4)
anatomic planes. SJM is used to measure the motion between articulating a combination of flexion–extension and abduction–adduction resulting in
surfaces of joints in any plane and can be measured using simple radiography circumduction.7

40 SECTION 1  Scientific Basis of Rheumatic Disease




FIG. 6.3  The included angle is located between longitudinal axis of two segments
defining a joint—150 degrees and the anatomic angle is the angle through which
the joint would have to be moved to take it from anatomic position to the position
of interest. (From Zatsiorsky VM. Kinetics of human motion. Human Kinetics; 1998,
p. 97.)

FIG. 6.1  (a) Six degrees of freedom (DOF) of movement of the tibiofemoral joint of
the knee involves three translations and three rotations. (b) The more constrained
and therefore more stable hip joint is a ball and socket joint with only three rotational a b
DOF; however, hip distraction is a translational DOF and is not present in healthy
hip function but has been measured in total hip arthroplasty using fluoroscopy25
and occurs in hip dislocation.


Dependent posture 90° flexion
Center Center
of femoral of femoral
c d
head head

Posterior points Trans-
Most distal
on femoral epicondylar
points on
condyles line
XF condyles XF
ZT ZT ZT YF 90° abduction Adduction
Center of tibial FIG. 6.4  A subject performs three shoulder joint motions in succession: (a) starting
intercondylar from a dependent posture with arms by the side and the palms of the hand facing
a b c
notch medially: (b) flexion of 90 degrees and (c) horizontal extension (abduction) of 90
degrees. (d) Adduction. The hand is now in a pronated position (anatomic posture),
FIG. 6.2  Somatic reference system for a knee joint: mechanical-based system, (a)26
although pronation was not performed. This phenomenon, known as the Codman’s
system based on anatomic axes of the tibia and femur27 (b), and system based on
paradox,29 occurs because rotations in a clinical system are not defined in accordance
the transepicondylar line and the tibial mechanical axis (c).28 Experimentally measured
with the requirements of kinematics.
differences of abduction and adduction at maximum flexion in leg swing during the
gait cycle are 4 degrees between a and c and 7 degrees between b and c. (From
Zatsiorsky VM. Kinetics of human motion. Human Kinetics; 1998, p. 83.)
body in equilibrium (i.e., at rest or constant speed). An example of a static
analysis is the evaluation of forces in the knee during standing. In vector 
Geometrically included angles can be viewed as internal joint angles, and notation, if the system isin equilibrium, the sum Σ of the all the forces F
anatomic angles can be viewed as external, to designate joint configuration and moments (torques)
 M acting on the body should be equal to zero; this
(Fig. 6.3). The advantages of using the CRS are that it provides anatomically is written as ∑ F = 0, ∑ M = 0. A dynamic analysis examines the forces and
meaningful definitions of the main segmental movements and is convenient moments acting on a body in motion (i.e., accelerating or decelerating) and
when a joint moves from a standard anatomic or neutral position. The includes descriptions of the inertial properties
  of the body being analyzed.

disadvantages are that it is not suitable when a joint rotation commences In vector notation, this is written as ∑ F = ma and ΣM = Iα, where a and α
from a non-neutral configuration and is not appropriate to describe complex are the linear and angular accelerations, respectively, of the body; m is the
movement patterns (Fig. 6.4). mass of the body; and I represents the inertial properties (mass distribution)
of the body.
STATICS AND DYNAMICS Example of static analysis
Whole-joint force analyses are commonly performed to estimate the force An example of a static force analysis is that performed for the shoulder joint
and moments that a diarthrodial joint may experience during ADLs. However, (Fig. 6.5, a). Although this is performed in two dimensions, for ease of
most activities are dynamic (even standing still!). The main forces that must analysis, important generalized biomechanical characteristics of whole joints
be considered during this type of analysis are due to body weight, muscle are evident from the results. The goal of this analysis is to calculate the static
forces, other forces related to soft tissues constraints, and other externally in vivo muscular forces and contact force between the proximal humerus
applied loads (e.g., carrying the weight of an object). Friction forces also and glenoid fossa of the glenohumeral joint. We examine a shoulder at 90
exist within a joint when two articulating surfaces interact; however, for a degrees of abduction with a ball held in the hand.
healthy joint, these are negligible and can be ignored for a dynamic analysis. To perform a force analysis (static or dynamic), it is necessary to construct
A static analysis examines the forces and moments (torques) acting on a a free-body diagram (FBD), which displays all intrinsic forces and moments

CHAPTER 6  Biomechanics of peripheral joints 41


∑F x = 0 → FJ ⋅ cos(θ J ) − FM ⋅ cos(θM ) = 0 [Eqn. 1]
Deltoid Weight ∑ Fy = 0 ↑ FM ⋅ sin(θM ) − FJ ⋅ sin(θ J) + WA + WB = 0 [Eqn. 2]
muscle Radius in hand
Humerus In a static analysis, the sum of moments (ΣMz) may be taken at any
coordinate location. In the current analysis, the sum of moments was taken
about the point of application of the contact force on the humerus. The sum
of moments was taken at this point to preclude the presence of the
articulating surface contact force in the moment equation, yielding:
Scapula Ulna

∑M Z = 0 FM ⋅ sin(θM ) ⋅ a − WA ⋅ b − WB ⋅ c = 0 [Eqn. 3]

The preceding equations can be solved:

y FJ Fm From Eqn. 3:
θJ θm Wa ⋅ b + Wb ⋅ c
FM = [Eqn. 4]
x sin(θM ) ⋅ a
From Eqn. 1:
FJ ⋅ cos(θJ ) = FM ⋅ cos(θM ) [Eqn. 5]
b c From Eqn. 2:

FIG. 6.5  (a) Person holding a ball in the hand at 90 degrees of shoulder abduction. FJ ⋅ sin(θJ ) = FM ⋅ sin(θM ) − Wa − Wb [Eqn. 6]
Relative positions of the bones in the arm are indicated. (b) Free-body diagram of
Divide Eqn. 6 by Eqn. 5:
the arm in 90 degrees of shoulder abduction. (The humerus, radius, and ulna are
grouped together for simplification of analysis.) a, b, c, Distances from the glenohumeral FM ⋅ sin(θM ) − Wa − Wb
joint center to forces acting on the limb; FJ, joint contact force; FM, deltoid muscle tan(θJ ) =
FM ⋅ cos(θM )
force; θJ, angle of joint contact force; θM, angle of deltoid muscle force; WA, weight
of arm; WB, weight of ball. The z axis is acting out of the page.  F ⋅ sin(θM ) − Wa − Wb 
θ J = tan−1  M 
 FM ⋅ cos(θM ) 
acting on a body or segment of interest. An FBD of a simplified shoulder
joint and the key muscle group acting to maintain the joint in its current The resultant FJ is found by the following equation (using a known
posture is shown in Fig. 6.5, b. For ease of analysis, a single rigid body is trigonometric identity):
used to represent a combined radius, ulna, humerus, and hand. All forces
and moments acting on the selected body segment must be included in the FJ2 = (FM ⋅ cos(θM ))2 + (FM ⋅ sin(θM ) − Wa − Wb )2
FBD. Forces included may be separated into four classes: FJ = (FM ⋅ cos(θM ))2 +(FM ⋅ sin(θM ) − WA − WB )2
1. Contact force from the humerus acting on the glenoid fossa
2. Muscle forces from muscle groups that flex, extend, or rotate the shoulder We now substitute values for known variables from anthropometric
joint; only the deltoid is modeled here measurements24 a = 15 cm, b = 30 cm, c = 60 cm, θM = 15 degrees,
3. Intrinsic weight of the arm WA = 40 N, and WB = 60 N to solve for the unknown variables FM, FJ and NJ,
4. External forces, such as the weight of the ball calculated as FM = 1236.4 N, FJ = 1214.4 N, and θJ = 10.4 degrees.
In the FBD, it is important to record the lines of action of the forces and
moments plus the point of application, direction (orientation), and magnitude
of each force (as it is a vector). The vector (arrow), representing the contact
force, is placed at the proximal end of the humerus, representing the contact This simple analysis shows characteristics common to diarthrodial joints
region in the joint between the humeral head and the glenoid; however, its throughout the human body: (1) muscle force is the main contributor to
direction (θJ) and magnitude (FJ) are unknown. The vector (arrow), represent- the force acting through the joint, and (2) poor mechanical advantage is
ing the deltoid muscle force, is placed at its point of attachment to the shown for the shoulder when the weight of the ball in the hand is compared
humerus at a distance “a” from the glenohumeral joint rotation center, which with the muscle force required to maintain the ball position (the calculated
in this case is assumed to coincide with the glenohumeral joint contact muscle force is more than 20 times the weight of the ball!). Altering three
location, for simplicity. The magnitude of the muscle force (FM) is unknown. factors of the analysis may increase the mechanical advantage of the system
Lines of action of muscle forces (θM) are assumed on the basis of detailed by decreasing the muscle force required to hold the ball: (1) An increase of
anatomic dissections of joints. θM would increase the component of FM that resists the downward force
The force vector representing the arm weight (WA) is assumed to be of the combined weight of the arm and ball. (2) Increase the moment arm
known and is placed at the center of mass of the arm at a distance “b” from of FM, distance “a,” but this is often not practical. (3) Flex the elbow and
the joint center. The force vector representing the arm weight is acting bring the effect of the combined weight of the arm and ball closer to the
downward, in the direction of gravity. The same is true for the ball of known center of rotation of the shoulder. This would change the resulting value
weight held in the hand (WB and distance c). The unknown variables are from the moment equation and reduce the effective muscle force and resulting
the magnitude of the contact force (FJ), the direction of the contact force contact force.
(θJ), and the magnitude of the muscle force (FM). These are calculated using
static force analysis. Limitations of static analysis
A body is considered to be in static equilibrium when all forces and all A number of limitations to the static analysis exist because of the number
moments acting on the body sum to zero. This summation may be performed of unknown variables. The number of static equilibrium equations available
around any point on the FBD. For this two-dimensional model, static for use prevents the inclusion of additional muscle or ligamentous forces.
equilibrium vector equations may be written in terms of their individual x, That is, for a two-dimensional problem, we could only use two force equations
y, and z components: (ΣFx = 0, ΣFy = 0) and one moment equation (ΣMz = 0) to solve the unknown
variables (i.e., only three unknown variables may be determined in a two-
∑F x = 0, ∑F Y = 0, ∑M z =0 dimensional analysis). If the analysis were performed in three dimensions,
it would be possible to solve nine unknown variables. Owing to complex
human anatomy and numerous muscle forces, we often have a greater number
The force equilibrium equations are applied to the FBD in Fig. 6.5, b, of unknown variables to solve than we have equations to use in the analysis.
and the result displayed in Box 6.1. This results in an indeterminate problem.

42 SECTION 1  Scientific Basis of Rheumatic Disease
Several computational methods have been developed to increase the
number of variables that may be solved. One is the exclusion of selected HYDRODYNAMIC LUBRICATION
forces. For example, in the previous analysis, the teres minor muscle, an
adducting muscle of the shoulder, was excluded because the shoulder was Weight
in pure abduction. A second is to reduce the number of unknown forces by
assuming a force relationship between the different muscle units. For example,
the supraspinatus muscle is also active during shoulder abduction, but it
was excluded to limit the number of unknown variables. We may have
assumed a relationship between the deltoid and supraspinatus, such as F2
= A • F1, where F2 is the muscle force of the supraspinatus, F1 is the muscle
force of the deltoid, and A is a constant. The relationship between different Fluid Pressure
pressure distribution
muscular units not included in the analysis may be formatted in a linear or
nonlinear manner.
A third method to decrease the number of unknown variables would be
to incorporate representations of muscle electromyographic (EMG) measure-
ments into the analysis. (Note: EMG does not equate directly to muscle FIG. 6.6  Fluid film lubrication. During motion at sufficiently high velocities, the weight
force.) A fourth method to increase the number of unknown variables is to tilts and forms a wedge shape of lubricant entrained into the contact. Because of
use numerical optimization. Using state-of-the-art techniques in clinical the viscous properties of the fluid, a pressure will be created within the fluid to
movement analysis, joint moments can be reliably estimated from recorded support the weight.
movement and external force data. Optimization is performed by minimizing
a mathematical function that is defined as the “cost” of performing an ADL.
Various cost functions that have been used include minimization of muscular
forces, muscle stress, and muscle energy. These cost functions have been COMPARISON OF LUBRICATION
successful in evaluating muscle forces during various ADLs. Static optimization
methods that transform joint moments into estimates of individual muscle Rigid bearings
forces using musculoskeletal models, have been used for several decades.8
1 Load Pressure 2 Load
Examples include OpenSim, an open source software that allows the user distribution
to input limb movement, joint DOF, kinematics, and lines of action and Squeeze Squeeze
force-generating parameters of the muscle. This allows the function of each Rolling Rolling film film
muscle to be analyzed.9 The AnyBody Modeling system is another example
of this type of software to analyze the musculoskeletal system of humans or
other creatures as rigid body systems.10

Human joints are exposed to great variation in loading conditions and Deformable bearings
movement; thus, they are expected to respond dynamically at all times to
allow for ADLs, sports, and functional alterations (e.g., caused by pain, 3 Load 4 Load
deformity, or muscle weakness). There can be high-impact, short-duration Squeeze Squeeze
loads such as in running; moderately low loads with a prolonged loading Rolling Rolling film film
time such as standing; and low loads with rapid motion in the swing phase
of walking. Over a lifetime, there is relatively little wear of the joint articulating
surfaces, which indicates a highly effective lubricating system. The coefficient
of friction between AC lubricated by synovial fluid in healthy joint is 0.005,
where friction is the resistance to movement between two surfaces in contact.
This is lower than found in highly polished machine bearing and the interaction
of steel on ice (i.e., ice skating) (0.01). Much of the experimental research FIG. 6.7  Load carrying by lubricated bearings. A comparison of hydrodynamic lubrication
into understanding the role of synovial fluid and AC in human joints was (1) and squeeze film lubrication (2) of rigid surfaces, and elastohydrodynamic
undertaken in the early days of biotribology research. The following informa- lubrication of deformable bearing surfaces under a hydrodynamic (sliding) action
tion is based on these early studies, which still provide the basis of our (3) and a squeeze film action (4). Surface deformation of elastohydrodynamically
understanding of human and artificial joint tribology today. lubricated bearings increases the contact area, thus increasing the load-carrying
capacity of these bearings and spreading the load over a larger contact area. Shearing
Modes of lubrication takes place between the layers of the lubricant.
Two types of lubrication exist: boundary lubrication and fluid film lubri-
cation. Boundary lubrication is caused by a single layer (monolayer) of
lubricant adsorbed on each bearing surface. In the case of a joint, boundary
lubrication is achieved by a macromolecular monolayer attached to each Lubrication in diarthrodial joints
articular surface. These layers carry loads and are effective in reducing In diarthrodial joints, a mixed mode of lubrication can occur, with the joint
friction. Fluid film lubrication is caused by a thin film of lubricant that is surface loads being sustained by fluid film pressures in areas of noncontact
entrained into, or trapped and retained in between, the contact between and by boundary lubrication in areas of contact (Fig. 6.8). In addition,
the joint surfaces during their motion and impact. This produces a greater cartilaginous joint surfaces differ from typical engineering bearings. Articular
bearing–surface separation. The pressure developed in the lubricating fluid cartilage can be treated as a biphasic material with an interstitial fluid phase
carries the loads applied to the joint (Fig. 6.6), and shearing in the fluid and porous-permeable solid phase. The material properties and porosity of
takes place. the solid matrix define the interstitial fluid pressure, which supports load
The lubricating characteristics depend on the lubricant properties (e.g., bearing and lubrication and contributes to the protection of AC from wear.
viscosity), the shape of the gap between the two bearing surfaces, and the The surfaces can exude and imbibe a lubricating fluid. As the joint moves
relative velocity of the surfaces. In a human joint, the bearing materials (i.e., and the surfaces slide, fluid is exuded in front of and beneath the leading
the AC) are not rigid and stiff, which results in elastohydrodynamic lubrication. half of the load. When the peak stresses decrease, fluid is reabsorbed back
As the joint surfaces move relative to each other and entrain a layer of into the AC, and it returns to its original dimensions.11-15
lubricant (i.e., synovial fluid) into the contact, the fluid pressure that develops The viscosity of a lubricating fluid is important. Synovial fluid undergoes
causes deformation the surfaces (i.e., flattening of the AC) (Fig. 6.7). This large changes in viscosity with changes in both temperature and velocity
changes the film geometry by increasing the surface area (and thus the area gradient. For very low velocities, a thinner lubricating film is desirable.
over which the joint contact force is transmitted), reducing the escape of Because synovial fluid is thixotropic, which means it can become fluid when
lubricant from between the bearing surfaces and generating a longer lasting agitated and then settle when left at rest, it can meet these requirements. If
film. These factors produce a lower stress concentration within the joint a joint effusion is present, the velocity-dependent properties may be lost,
surfaces.11 which results in reduced lubrication and subsequent joint surface wear.

CHAPTER 6  Biomechanics of peripheral joints 43


boundary Pressurized Macromolecular
lubricant fluid monolayer

~0.3 mm

Cartilage Boundary lubricated Articular

asperity contact surface Unloaded Tension Compression Bending

FIG. 6.8  Mixed lubrication operates in articular cartilage bearings: boundary lubrication
in which the lubricating film is as thick as the roughness of the bearing surfaces
(order = 0.3 mm19), and fluid film lubrication in which surfaces are more widely
separated in the troughs between the surface asperities.

Hyaluronan, lubricin, and phospholipid molecules exist in synovial joints,

such as hips and knees, and more recently have been suggested to act as Shear Torsion Combined loading
lubricants responsible for the remarkable lubrication of AC. However, they
cannot alone explain the extremely low friction. Each may play a different FIG. 6.9  Schematic representation of various loading modes.
role, acting together; hyaluronan is attached at the AC surface by lubricin
molecules and phospholipid complexes that provide lubrication via the
hydration–lubrication mechanism.16
Understanding how the various tissues in the joint (i.e., cancellous, corti-
cal, and subchondral bone; AC; ligaments; meniscus) respond to the loads
that they experience during ADLs, overload, and traumatic events provides Yield D
point C
insight into the progression and treatment of musculoskeletal diseases.
When a force is applied to a tissue, the tissue will deform in response to B Fracture
the applied load. If a displacement is applied to the tissue (i.e., by apply- point
Force (Newtons)

ing a force), it will produce a resistive force as a reaction to the applied

displacement and deform. The amount of deformation is proportional to
a combination of the stiffness of the tissue, which is in turn related to its Stiffness
composition; its shape and size; and the properties of the surrounding,
supporting tissues. These material and structural properties vary for all of the ∆F
tissues within a joint. There are a number of ways in which a structure can ∆D
be loaded; Fig. 6.9 provides a simplified approach to understanding these.
Note that these may be experienced, often in combination, by all tissues in Elastic Plastic
the body. region region


Tissues can be tested as isolated individual samples (mechanical testing), to Displacement (millimeters)
understand the material properties, or within an entire structural complex
(structural testing). The structural properties of a tissue incorporate not
only the material composition of the tissue but also its geometric configuration FIG. 6.10  Force–displacement curve for a structure composed of a pliable material.
and mechanical environment.17 Output from structural testing, which involves If a load is applied within the elastic region (A–B) and removed, no permanent
measuring the response of the tissue in terms of its deformation, to an deformation occurs. If loading continues past the yield point (B) and into the plastic
applied load, is shown as a force–displacement curve (Fig. 6.10). The units region (B–D) and is then released, permanent deformation results. The amount of
of force, F, are Newtons (N) or pounds force (lbf), and the units of displace- permanent deformation that occurs if the structure is loaded to point C in the plastic
ment, D, are typically millimeters (mm). Stiffness is a structural property of region and then unloaded is represented by the distance between A and E. Structural
a body and is calculated as the slope of the linear elastic region of the stiffness is calculated as the slope of the linear portion of the elastic region (i.e.,
force–displacement curve. ΔF/ΔD).
The material properties of a tissue encompass the inherent material
(chemical and physical) composition of the tissue itself (e.g., contribu-
tion of collagen fibrils in a tendon).17 Output from mechanical testing of
a tissue sample is shown as a stress–strain curve. The units of stress are
typically Pascals or Newtons/meter2 (Pa, N/m2); strain has no assigned Stress (σ) is calculated as the applied force divided by the initial cross-sectional
units. The mechanical property of the modulus of elasticity (Young’s area of the tissue sample: σ = F/Ao. Strain (ε) is calculated as change of
modulus) is calculated from the linear section of the stress–strain curve. length of the tissue sample divided by the initial length of the sample: ε =
Stress and strain are considered normalized values of load and displace- ΔL/Lo. Stress and strain and force and displacement may be positive or negative
ment, respectively, on the basis of the geometry of the sample being tested values (i.e., putting the structure or sample in tension invokes a tensile force;
(Fig. 6.11). conversely, compressing the sample invokes a compressive force, and this

44 SECTION 1  Scientific Basis of Rheumatic Disease


A2 T1, T2, T3
L2 F FIG. 6.11  Different continuous structures composed
Ti =
A1 Ai of the same pliable material but with different


F F ∆L geometries are tested. When individual forces are

Ei = i
L1 Li converted to stress and displacements are converted
i = 1, 2, 3 to strains, all loading curves are superimposed onto
one another.

Displacement Strain
L3 > L1 = L2 A2 > A 1 = A 3 E1, E2, E3


Brittle 200
material Apparent density
Ductile 0.30 g/cc
material 150 Cortical bone 0.90 g/cc

Stress (MPa)
1.85 g/cc

Trabecular bone
0 5 10 15 20 25
Strain (%)

FIG. 6.13  Examples of typical compressive stress–strain behavior of trabecular and

cortical bone for different apparent demises. (Adapted from Keaveny TM, Hayes
WC. Mechanical properties of cortical and trabecular bone. In: Bone, vol. VII: Bone
Growth-B. Boca Raton, FL: CRC Press; 1992, p. 285-344.)

FIG. 6.12  Ductility is a measure of the amount of plastic deformation at failure and
is shown in this figure. Resilience is the ability of a material to absorb energy when
deformed elastically and to release the energy when unloaded. Resilience is calculated
as the area under the stress–strain curve line. The presence of shear strain in a Tensile loading Compressive loading Shear loading
structure loaded in tension and in compression is indicated by angular

requires +/- representation). Additional information about the tissue may be

extracted from the stress–strain curve (Fig. 6.12). For example, ductility is
a measure of the amount of plastic deformation of a sample when it reaches
failure, and resilience is the ability of a material to absorb energy when
deformed elastically and to release the energy when unloaded. An example
of this can be seen in Fig. 6.13 where the typical stress–strain qualities of FIG. 6.14  Stress distribution within a material with tensile loading, compressive loading,
cortical and cancellous bone with different bone densities are tested under and shear loading. In three dimensions, six independent stress components are
similar conditions. required to describe the state of stress at each point in a body. Three of the stress
The stress distribution in a body is a quantitative description of the components are normal stresses (tension–compression, which act perpendicular to
distribution of the internal forces that arise in the body as a result of the the cross-sectional plane), and three stress components are shear stresses that act
external forces at its surfaces (Fig. 6.14). When an external force acts on along the plane of the cross-sectional area.
the body, the body deforms to resist the applied load; this deformation is
called strain (Fig. 6.15). We must assume a stress–strain relationship to
calculate the stress within a tissue on the basis of experimental strain measure- tissue is isotropic, the mechanical properties of the tissue do not change as
ments. The stress in a body is related to the strain by the modulus of the the orientation of the tissue sample is changed. Most tissues in the body are
body’s material. This relationship may be complex and may consist of numerous inhomogeneous and anisotropic. Stress–strain plots of tissue samples from
modulus values. The exact number of modulus values depends on the an inhomogeneous and anisotropic tissue (AC) are shown in Fig. 6.16.
homogeneity and isotropy of the test sample.18 Because of this variation, the orientation and source location of the tissue
If a tissue is homogeneous, the mechanical properties of the tissue do sample significantly influence the stiffness of the tissue. An example of this
not differ depending on the source location of the tested tissue sample. If a is well understood for AC samples.19 AC samples that are aligned with the

CHAPTER 6  Biomechanics of peripheral joints 45


Unloaded Under tensile Under compressive F y

loading loading

Ey = ∆L = Li – Lo
Lo Lo

Li Lo Ex = ∆W = Wi – Wo
Wo Wo
ν= Ex = (Wi – Wo) Lo
Ey (Li – Lo) Wo

θ Wi
= Initial configuration
= Final configuration

θ = 90 θ > 90 θ < 90 F

FIG. 6.15  Tension and compression strain are indicated by longitudinal elongation FIG. 6.17  Poisson’s ratio relates the longitudinal elongation of the material to the
or compression of the body. When considering the unloaded shape the square will lateral contraction of the material. In an ideal incompressible material, ν = 0.5.
elongate under tension or flatten out under compression. In three dimensions, six
independent strain components are required to describe the state of strain at each
point in a body. Three of the strain components represent longitudinal elongation
or compression of the body; the other three represent shear. Whenever a structure
is subjected to tensile or compressive loading, shear stress is produced. In response Creep: the deformation response of a tissue sample under a constant applied
to shear loading, the diamond in the unloaded body deforms in an angular manner load (see Fig. 6.18)
with the angle changing depending on compression or tension (i.e., more acute or Stress–relaxation: the stress response of a tissue sample under a constant
more obtuse). applied displacement (see Fig. 6.19)
Hysteresis: the difference in the response curve between loading and
unloading of the tissue sample on a stress–strain diagram (see Fig. 6.20)


D C Articular surface force–displacement response of a tissue to mechanical loading: Young’s modulus

(Ey) and Poisson’s ratio (ν). In one dimension, Young’s modulus is calculated
Surface zone Collagen as the slope of the linear (elastic) portion of the stress–strain curve. Poisson’s
A B fibrils C
D ratio relates the longitudinal elongation (εy) of the material to the lateral
Middle zone contraction (εx) of the material (Fig. 6.17). This varies depending on the
A B internal structure of the material; also, soft tissues have a viscoelastic
Deep zone component response.
a = Tissue sample
For a viscoelastic tissue, the stress–strain response depends not only on the
D magnitude of the applied stress or strain but also on the rate of the applied
C stress or strain. Viscoelastic tissues have three characteristic stress–strain
responses: creep, stress–relaxation, and hysteresis as seen in Box 6.2.20

B These responses can be shown diagrammatically on a ligament sample.

A We assume a ligament has been removed from a joint as a bone–ligament–bone
(BLB) segment. One end of the BLB has been anchored to the mechanical
testing apparatus through a load cell to measure applied force, and the other
end of the BLB has been attached to the cross-head of the testing system,
Strain which can be moved to apply a tensile load (pulling the two ends of the test
segment away from each other), which induces a tensile stress distribution
FIG. 6.16  (a) Schematic diagram of collagen fiber orientation through the depth of within the test segment along its test length, or compressive load (bringing
articular cartilage. Superficial fibers are tangent to the articular surface. Fibers in the two ends of the test segment toward each other), which induces a compres-
the middle zone have no preferred orientation. Fibers in the deep zone are oriented sive stress distribution within the test segment along its test length.
radially to the subchondral surface. (b) Stress–strain diagram of cartilage testing in The viscoelastic responses of ligament are due to the interaction of fluid
tension. Tissue sample A is from the same region as tissue sample B but is oriented and solid components of the ligament: water and collagen, respectively. This
90 degrees to tissue sample B. Tissue sample C is from the same region as tissue can be described as biphasic. In a relaxed state before loading, the ligament
sample D but is oriented 90 degrees to tissue sample D. Differences in stress–strain is approximately 60% to 80% water and contains collagen bundles that have
response indicate the anisotropic and inhomogeneous material nature of cartilage. a wavy appearance.21 When a tensile force or displacement is applied to the
ligament, the fluid is exuded through pores in the tissue and the collagen
fibers begin to straighten and align in the direction of the applied load.
preferred direction of collagen fibrils are stiffer than samples that are oriented In a creep test, an instantaneous step or ramp load is applied to the tissue
perpendicular to the preferred direction. In addition, tissue samples from sample and held constant for an extended period of time (Fig. 6.18). This
the articular surface are stiffer than tissue samples from deep within the is considered a load control test, and the resulting tissue displacement is
tissue. Researchers often assume a tissue to be homogeneous and isotropic. measured. The majority of the fluid is exuded from the tissue, and the applied
This results in a requirement for only two variables to fully describe the load is balanced by the straightened collagen bundles. The displacement

46 SECTION 1  Scientific Basis of Rheumatic Disease


= Energy loss
= Direction of loading


0 Time




FIG. 6.20  The loading portion of the hysteresis curve (A) is always above the unloading
0 Time portion of the curve (B). The area between the loading and unloading portions of
the curve indicates relaxation of the collagen fibers within the ligament and the
FIG. 6.18  In creep testing, an instantaneous stress (σo) is applied to a tissue sample exuding of fluid from the ligament.
at time t0 and maintained while the resulting tissue strain is recorded. Whereas
solid-like materials exhibit an equilibrium deformation in response to the applied
load, fluid-like materials elongate continuously. shows an initial elastic response of the tissue followed by a gradual increase
of lengthening of the tissue. As the test proceeds, fluid continues to exude
from the tissue, and the solid components of the tissue support the applied
load. The “solid-like” material will be distracted to a point at which the
solid components of the tissue will balance the applied load and will not
STRESS–RELAXATION TESTING elongate further. The modulus of the tissue is calculated as the stress of the
tissue divided by its end displacement. The “fluid-like” material will not be
able to balance the applied load and will continue to elongate.
In a stress–relaxation test, an instantaneous step or ramp displacement
is applied to the tissue sample and held constant for an extended period of

ε0 time (Fig. 6.19).22 This is considered a displacement control test; the resulting
tissue force is measured. The collagen bundles are straightened by the imposed
displacement, and fluid is initially exuded in response to the rapid increase
in interstitial fluid pressure. With time, the interstitial fluid pressure comes
0 Time to equilibrium, and fluid is no longer exuded. An initial increase in force
measured by the load cell is followed by a gradual reduction. The force
eventually approaches an equilibrium value. The magnitude of this value
depends on the solid or fluid nature of the tissue. A solid-like material will
have a final stress that is not zero. A fluid-like material will have a final
stress that is close to or equal to zero.

Solid-like Finally, hysteresis of a soft tissue is displayed while the tissue is cycled
material to a known force or displacement and back to its initial position (Fig. 6.20).23
A combination of the fluid–solid effects is seen in a hysteresis loop. As shown
Fluid-like in the figure, the loading portion of the force–displacement curve is always
material higher than the unloading portion of the curve. The area between the loading
Time and unloading curves represents nonrecoverable energy that is lost during

FIG. 6.19  In stress–relaxation testing, an instantaneous deformation (εo) is applied

to a tissue sample at time t0 and maintained while the resulting tissue stress is
recorded. Tissues in the body have a solid component that resists the applied The authors acknowledge the contributions of Matthew F. Koff, who was
deformation after interstitial fluid has been exuded from the tissue. the author of this chapter in the previous edition.

1. Nordin M, Frankel V. Basic biomechanics of the 5. Soderkvist I, Wedin P-A. Determining the movements of 9. Delp SL, Loan JP. A graphics-based software system to
musculoskeletal system. 3rd ed. Philadelphia: Lippincott the skeleton using well-configured markers. J Biomech. develop and analyze models of musculoskeletal
Williams & Wilkins; 2001. 1993;26(12):1473-1477. structures. Comput Biol Med. 1995;25(1):21-34.
2. Kinzel GL, Hall AS, Hillberry BM. Measurement of the 6. Zatsiorsky V. Kinematics of human motion. Champaign, IL: 10. Damsgaard M, Rasmussen J, Christensen ST, et al.
total motion between two body segments. I. Analytical Human Kinetics. 1998. Analysis of musculoskeletal systems in the AnyBody
development. J Biomech. 1972;5(1):93-105. 7. American Academy of Orthopaedic Surgeons. Joint Modeling System. Simul Model Pract Theory.
3. Woltring HJ. Representation and calculation of 3-D joint motion: Method of measuring and recording. Chicago: 2006;14(8):1100-1111.
movement. Hum Mov Sci. 1991;10(5):603-616. American Academy of Orthopaedic Surgeons; 11. Unsworth a, Dowson D, Wright V. Some new evidence on
4. Cappozzo A, Gazzani F, Macellari V Skin marker artifacts 1965. human joint lubrication. Ann Rheum Dis.
in gait analysis. In: Abstracts of the Sixth Biannual 8. Erdemir A, McLean S, Herzog W, et al. Model-based 1973;32(6):587-588.
Meeting of the European Society of Biomechanics. 1990. estimation of muscle forces exerted during movements. 12. Radin EL, Paul IL. A consolidated concept of joint
p. 1990. Clin Biomech (Bristol, Avon). 2007;22(2):131-154. lubrication. J Bone Joint Surg Am. 1972;54(3):607-616.

CHAPTER 6  Biomechanics of peripheral joints 47
13. Katta J, Jin Z, Ingham E, et al. Biotribology of articular 19. Mow VC, Ratcliffe A, Poole AR. Cartilage and diarthrodial 25. Dennis DA, Komistek RD, Northcut EJ, et al. “In vivo”
cartilage—A review of the recent advances. Med Eng Phys. joints as paradigms for hierarchical materials and determination of hip joint separation and the forces
2008;30(10):1349-1363. structures. Biomaterials. 1992;13(2):67-97. generated due to impact loading conditions. J Biomech.
14. Walker PS, Dowson D, Longfield MD, et al. “Boosted 20. Callister WD, Rethwisch DG. Materials science and 2001;34(5):623-629.
lubrication” in synovial joints by fluid entrapment and engineering: An introduction. New York: John Wiley & 26. Grood ES, Suntay WJ. A joint coordinate system for the
enrichment. Ann Rheum Dis. 1968;27(6):512-520. Sons; 2007. clinical description of three-dimensional motions:
15. McCutchen CW. The frictional properties of animal 21. Weiss JA, Gardiner JC. Computational modeling application to the knee. J Biomech Eng.
joints. Wear. 1962;5(1):1-17. of ligament mechanics. Crit Rev Biomed Eng. 1983;105(2):136-144.
16. Seror J, Zhu L, Goldberg R, et al. Supramolecular synergy 2001;29(3):303-371. 27. Lafortune MA, Cavanagh PR, Sommer HJ, et al.
in the boundary lubrication of synovial joints. Nat 22. Kennedy JC, Hawkins RJ, Willis RB, et al. Tension studies Three-dimensional kinematics of the human knee
Commun. 2015;6:6497. of human knee ligaments. Yield point, ultimate failure, during walking. J Biomech. 1992;25(4):347-357.
17. Mow V, Gu W, Chen F. Structure and function of and disruption of the cruciate and tibial collateral 28. Pennock GR, Clark KJ. An anatomy-based coordinate
articular cartilage. In: Mow V, Huiskes R, eds. Basic ligaments. J Bone Joint Surg Am. 1976;58(3):350-355. system for the description of the kinematic displacements
orthopedic biomechanics and mechano-biology. 23. Fung Y. Biomechanics: Mechanical properties of living in the human knee. J Biomech. 1990;23(12):1209-
Philadelphia: Lippincott Williams & Wilkins; 2005: tissues. Berlin, Germany: Springer Science & Business 1218.
181-258. Media; 1993. 29. Codman E. The shoulder: Rupture of the supraspinatus
18. Lai W, Rubin D, Krempl E. Introduction to continuum 24. Winter DA. Biomechanics and motor control of human tendon and other lesions in or about the subacromial bursa.
mechanics. 3rd ed. Oxford: Pergamon Press; 1993. movement. New York: John Wiley & Sons; 2009. Boston: Thomas Todd Company; 1934.

Biomechanics of spinal degeneration
Michael A. Adams • Patricia Dolan


■ This chapter describes how mechanical loading can promote skeletal tissue
pathology in the spine.
Some of the underlying mechanical principles are explained more fully in
Chapter 6 and in the present authors’ previous work.1
■ Moderate mechanical loading strengthens vertebrae and (eventually) intervertebral
disks and articular cartilage.
■ Severe spinal loading arises mostly from muscle tension and from inertial forces COMPRESSION, SHEAR, BENDING, AND TORSION
generated during accelerations and falls.
Compressive forces on the spine act down its long axis at right angles to the
■ Excessive loading disrupts spinal tissues, but injuries need not be traumatic if midplane of the intervertebral disks (Fig. 7.1). The curvature of the spine
tissues are weakened by genetic inheritance and by age (as in osteoporosis). ensures that the direction of the compressive force varies between spinal
■ Injuries to cartilage are difficult to detect and slow to heal. levels, and this is consistent with the origins of this force, which is primarily
■ Compressive injury to the spine damages the vertebral endplates, decompresses the tension in the paraspinal muscles. Compressive force is measured in Newtons
adjacent disk nucleus, and can cause internal disk disruption. (N) with 9.81 N = 1 kg. Force per unit area is stress, measured in mega-Pascals
■ Injuries in combined bending and compression create radial fissures in the disk (MPa). Shear acts at right angles to the compressive force (see Fig. 7.1) and
anulus and can cause disk prolapse even in macroscopically normal disks. causes vertebrae to slide forward, backward, or sideways relative to each
■ Disk injury disturbs cellular metabolism and creates a vicious circle of weakening, other. Bending moments (measured as a force multiplied by a lever arm, in
reinjury, and frustrated healing; this is “disk degeneration.” units of Newton meters [Nm]) cause the spine to bend, usually relative to
■ Two disk degeneration phenotypes can be recognized, depending on whether the the centers of rotation within each intervertebral disk. Torsional moments,
initiating injury affects the endplate or anulus. or torques, have the same units as bending moment and cause the spine to
■ Severe disk degeneration and narrowing can initiate a “degenerative cascade” twist about its long axis (axial rotation).
involving instability, osteophytosis, osteoarthritis, stenosis, and senile
Gravity exerts a vertical force on each body segment: its weight. The weight
of the body that acts on each spinal level ranges from 55% of whole-body
weight at the fifth lumbar vertebral level to 7% at the first cervical level.
INTRODUCTION These forces generally represent only a small proportion of the total force
acting on the spine during vigorous activity.
This chapter is written for rheumatologists rather than engineers. It explains
how mechanical loading can initiate degenerative changes in the spine and
drive them to end-stage disease. Essentially, excessive loading can physically
disrupt avascular intervertebral disks, which are unable to heal effectively. When a body segment is accelerated, the force acting on it is amplified
Instead, the disruption spreads by mechanical means (e.g., crack propagation) according to Newton’s second law of motion: Force = Mass × Acceleration.
and by biologic means (disturbed cellular metabolism), and disk narrowing Exceptionally high acceleration occurs when the human body is ejected
initiates a degenerative cascade involving adjacent structures. The concept from an aircraft, and the compressive force on the lumbar spine can
of injury-driven degeneration has been neglected recently, possibly because then be enough to cause vertebral fracture. More typically, the human
of a preoccupation with genes and cell signaling and because injury has spine is subjected to high deceleration during a fall, especially when the
been equated with trauma, which is not a common cause of spinal degenera- impact velocity is high and the surface is so hard that velocity is brought
tion. In fact, injury can occur when normal mechanical loading is applied to zero in a short time. Long legs, lean buttocks, and hard slippery
to abnormally weak tissues, with a prime example being osteoporotic fracture. floors can lead to high inertial forces on the spine when the body hits
The role of vertebral fracture in spinal degeneration and pain is undeniable the ground.
because the evidence is plain to see on radiographs. Cartilage injury, however,
is harder to detect, but there is now abundant evidence that intervertebral
disks and articular cartilage are often injured and that degeneration follows
injury. Tension in a contracting muscle compresses the bones that lie between the
The purpose of this chapter is to explain these mechanisms and muscle’s tendinous insertions. Muscle forces are hidden in the sense that
to integrate them with other important influences to give a holistic they are internal to the body, but they often reach high levels for the reason
account of spinal degeneration and pain. Genetic inheritance and aging illustrated in Fig. 7.2. Essentially, muscles act on short internal lever arms
will be treated as important predisposing factors that lead to vulnerable and consequently must exert forces that are several times greater than the
tissues and structures. Some details of “mechanobiology” are provided in external loads that often act on much greater lever arms. Typically, the
Chapter 5. external load is dominated by the weight of the upper part of the body,
The first section, “Forces Acting on the Spine,” explains how high which is why lifting an object as small as a pen from the floor can generate
forces can arise during seemingly innocuous activities, a topic that has a peak tensile force in the back muscles of greater than 200 kg.2 The highest
medicolegal significance. “Mechanical Function of the Spine” describes the forces tend to arise when muscles accelerate or decelerate body segments
mechanical role of structures such as intervertebral disks, ligaments, and during jerky movements, for example, when playing squash or during a
apophyseal joints and leads naturally into “Injury,” which explains how stumble. Muscle forces can rise even higher during eccentric (lengthening)
they can be damaged. Disk herniation is essentially an injury, although contractions when the collagenous tissue within muscle is severely stretched.
there are important biologic predisposing factors and sequelae. “Biome- A good example is trying to prevent a carried patient from falling, when a
chanics of Disk Degeneration” explains how injury can disturb disk cell second lifter has “dropped her end.” The common need to stabilize the upper
metabolism and break down internal barriers, leading to disk degeneration body during dynamic activities often requires high co-contraction of paraspinal
and pain. Finally, “Spinal Degenerative Cascade” suggests how mechanical and abdominal muscles.
failure of intervertebral disks can affect adjacent tissues and lead to spinal High muscle forces have a number of practical implications. In the
instability, osteophytosis, osteoarthritis, stenosis, degenerative scoliosis, and medicolegal arena, many clinicians underestimate muscle forces by supposing
senile kyphosis. (quite wrongly) that forces on the spine or neck must be low when the

CHAPTER 7  Biomechanics of spinal degeneration 49


AT dw




EM = F × d = W × D + w × dw
C = F + (W + w) × cos α

FIG. 7.1  The lumbar spine showing the directions of compressive (C) and shear (S) FIG. 7.2  During manual handling, high tensile force (F) must be created by the back
forces acting on the lumbosacral disk. Spinal compression acts perpendicular to the
muscles to generate an extensor moment (EM) large enough to lift up an external
midplane of the disk, and thus its direction varies with the spinal level. A bending
weight (W) and the weight of the upper part of the body (w). The back muscles act
moment (BM) causes the spine to bend in flexion, extension, or lateral bending.
on a short internal lever arm (d) as opposed to the long lever arms (D, dw) of the
Axial torque (AT) causes axial rotation about the long axis of the spine. (Reproduced
objects being lifted. In practice, muscle force F is often much greater than the weight
from Adams MA, Bogduk N, Burton K, Dolan P. The biomechanics of back pain.
being lifted, so the compressive force acting on the spine (C) rises to approximately
3rd ed. Edinburgh: Churchill Livingstone; 2013.)
500 kg when a weight of 20 kg is lifted. (Reproduced from Adams MA, Bogduk N,
Burton K, Dolan P. The biomechanics of back pain. 3rd ed. Edinburgh: Churchill
Livingstone; 2013.)

external loads are light. The potential danger of muscle contractions is MEASUREMENT OF SPINAL LOADING IN VIVO
illustrated by the finding that epileptic fits can cause the back muscles to
crush vertebrae even when no falls are involved.3 The total compressive force acting on the spine of a living volunteer can be
measured by inserting a pressure-sensitive needle into an intervertebral disk.
The compressive force on the L4 to L5 disk of healthy subjects ranges from
INTRAABDOMINAL PRESSURE 14 kg when lying prone to more than 200 kg in the stooped, standing position,5
People struggling to lift a heavy weight often hold their breath and go red as shown in Fig. 7.3. Relaxed sitting increases disk compression more than
in the face, thus indicating a common if unwitting strategy to stabilize the standing does because sitting flexes the lower lumbar spine and generates
spine and protect it from high force. Contracting trunk muscles generate additional tension in the posterior ligaments, but a lordotic standing posture
high pressure in the abdominal cavity, which can then transmit a compressive causes some of the compressive force to be resisted by the neural arch.6
force directly from the thoracic spine and ribs to the pelvis, bypassing the Direct measurements of spinal loading are unsafe when the spine is moved
lumbar spine. The mechanism is difficult to quantify because tension in forcefully. An alternative approach is to quantify spinal compression from
longitudinal abdominal muscles acts to compress the spine further. However, electromyographic (EMG) signals recorded from the skin surface overlying
raising intraabdominal pressure is most beneficial when the spine is flexed; the back muscles. EMG signals must be calibrated against moments and
when a thick belt is worn to provide lateral support to the abdomen; or forces generated during standardized exertions in the laboratory.2 Peak
when the pressure is raised primarily by contraction of transversus abdominis, compressive forces on the lumbar spine vary between 250 and 500 kg when
which does not compress the spine.1 lifting objects weighing up to 30 kg from the floor. Rapid lifting increases
peak loading more than 60% compared with slow lifting, and forces of this
magnitude can cause fatigue (“wear-and-tear”) damage to accumulate in
DIURNAL VARIATION some spines. Similar peak forces are reported when mathematical models
In the early morning, intervertebral disks are swollen with water, so the are used to calculate spinal loading from optical measurements of the move-
already-stretched anulus and intervertebral ligaments resist bending strongly ment and acceleration of various body parts.7
and are vulnerable to injury.1 As the day progresses, disks lose 20% of their Shear forces and torques have not been measured reliably in vivo, but
water and height, so the spine becomes more supple, but the neural arches mathematical models show that peak anterior shear forces reach 150 to
then resist more of the compressive force on the spine. These diurnal changes 200 kg in the lower lumbar spine when heavy weights are lifted.7
explain why backache often increases after several hours of standing and The peak bending moments acting on the spine of a living person can
why recurrent back pain can be reduced by avoiding flexion movements be estimated by measuring the moment required to bend a cadaveric spine
during the early morning.4 to the same angle. This technique has been used to show that bending

50 SECTION 1  Scientific Basis of Rheumatic Disease


load (N)



0 2 4 6 8 10 12 14 16 18 20 22 24
Angle of motion segment
FIG. 7.3  The compressive force acting on the lumbar spine increases to 2000 N
(≈200 kg) with stooped, standing postures. Forces were calculated from pressure
in the L4 to L5 intervertebral disk nucleus measured in vivo by a miniature pressure
transducer. “Angle of motion segment” refers to the relative orientation of the upper
and lower endplates of the L4 to L5 disk. (Reproduced from Sato K, Kikuchi S,
Yonezawa T. In vivo intradiscal pressure measurement in healthy individuals and in
patients with ongoing back problems. Spine 1999;24:2468-74.)

moments on the lumbar spine rise to 10 to 25 Nm during heavy lifting.2

Cadaveric spines can be flexed more than this before they are damaged, so Lumbar
in life, back muscles must normally protect the spine from excessive flexion. lordosis
However, protective muscle reflexes can be impaired if the spine is flexed
repeatedly or for periods of an hour or more. 8 This could explain why
activities such as gardening and long-distance driving are often associated
with back pain.


Cervical lordosis appears when infants first lift their heads up, and lumbar FIG. 7.4  The S-shaped curves of the spine can assist in shock absorption during
lordosis develops when walking begins. Thoracic kyphosis appears to be a locomotion. When the center of gravity of the body descends at “heel strike,” the
compensatory mechanism to maintain a level line of sight and to increase cervical and lumbar lordosis both increase, and they decrease again as the body
the volume of the thoracic cavity. Spinal curves (Fig. 7.4) probably play a rises at “toe-off.” This enables energy to be stored and then released by deformed
shock-absorbing role during locomotion because the natural tendency for intervertebral disks, tendons, and ligaments.
the curves to flatten and accentuate as the body rises and falls is resisted by
the muscles of the trunk. The tendons of these muscles have a great capacity
to store energy so that they can act as shock absorbers and minimize vertical
accelerations of the head. A similar mechanism allows the quadriceps tendons load being spread evenly on the adjacent vertebral bodies even when the
to absorb energy when the knees are flexed to soften the landing from a vertebrae are angled in flexion or extension.5,10 The layers (lamellae) of the
jump. anulus fibrosus act in tension to retain this pressurized nucleus, but the
adult human anulus is sufficiently fibrous and stiff to resist direct compressive
loading as well. The internal mechanical functioning of loaded cadaveric
SPINAL MOVEMENTS disks has been investigated by pulling a needle-mounted pressure transducer
Intervertebral movements in living people combine angular rotations with along their midsagittal diameter. The resulting “pressure profiles” (see later)
small gliding movements (translations) in the plane of the disk. Angular show the extent of the fluid-like region and the pressure within it. Both
rotations stretch and compress the disk anulus in such a manner that the decrease with age, but the compressive stress (force per unit area) resisted
center of rotation (the theoretical pivot point) moves around within the disk by the anulus increases.10 Cervical disks behave in essentially the same way.11
nucleus. The oblique surfaces of the apophyseal and uncovertebral joints
ensure that certain movements are mechanically “coupled”; for example,
attempted lateral bending normally creates a small axial rotation as well.
The cervical spine is the most mobile region because cervical intervertebral Compressive loading of the spine is resisted mostly by the disks and vertebral
disks are relatively thick in comparison to the height of adjacent vertebral bodies. As the compression increases, the disk anulus bulges radially outward,
bodies. Conversely, the thoracic spine is the least mobile because its disks and the vertebral endplates bulge vertically into the vertebral bodies. The
are narrow, and thoracic movements are inhibited by the ribs and dipping relatively high stiffness of the disks and vertebral endplates (in comparison
spinous processes. Mobility of the lumbar spine declines by approximately with tendons, for example) ensures that they absorb little energy and so are
50% between the ages of 16 and 85 years.1 This is partly due to disk narrowing, poor shock absorbers. A small proportion of the compressive force is normally
which brings the neural arches of adjacent vertebrae closer together and resisted by the neural arch; this rises to 20% after sustained loading of the
causes the center of rotation for flexion and extension to migrate posteriorly disk and when the spine is bent backward slightly to simulate a lordotic
toward the apophyseal joints.9 standing posture.6


The central nucleus pulposus of an intervertebral disk normally has such a The apophyseal joints resist the forward shearing movements of adjacent
high water content that it behaves like a pressurized fluid, with compressive vertebrae (see Fig. 7.1), with the more frontal plane orientation of the lower

CHAPTER 7  Biomechanics of spinal degeneration 51

lumbar joints, making them particularly well suited to prevent any forward endplate into the vertebral body (Fig. 7.5, b), it eventually becomes surrounded
slip of the overlying vertebra. by a calcified layer known as a Schmorl’s node (Fig. 7.5, a). Endplate fracture
causes an immediate and large drop in pressure in the adjacent nucleus
pulposus (Fig. 7.6, b), especially in middle-aged thoracolumbar disks.18 Other
RESISTANCE TO TORSION types of vertebral fracture are considered later under “Vertebral Deformity
Similarly, the articular surfaces of the apophyseal joints are well orientated and Senile Kyphosis.”
to resist axial rotation. In the lumbar spine, only 1 to 2 degrees of rotation
is permitted before the articular surfaces make firm contact, but considerably
more movement is allowed in the cervical spine.1 Stretching of the apophyseal
joint capsule and ligaments on the tension side and deformation of the disk Cadaveric experiments have shown that apparently normal disks can be
anulus also contribute substantially to the spine’s resistance to torsion. Loss made to herniate posteriorly by loading them simultaneously in bending
of apophyseal joint articular cartilage as in osteoarthritis12 can allow more and compression. A single application of severe loading can cause bulk
free play in these joints and increase the range of axial rotation. The first extrusion of nucleus pulposus (Fig. 7.5, d), especially in lower lumbar disks
two cervical vertebrae provide negligible resistance to torsion. in those aged 30 to 50 years.1 In these experiments, either the bending or
the compression must exceed normal limits but not both. More moderate
but repetitive loading of disks aged 20 to 40 years can create radial fissures
RESISTANCE TO BENDING (Fig. 7.5, c) that allow nuclear material to migrate into and through the
Ligaments of the neural arch reorientate when the spine moves into flexion. posterior anulus.1 The underlying mechanism of disk herniation, which has
They then resist end-range movement strongly, with the interspinous and been validated in experiments on animal tissues19 and by mathematical
supraspinous ligaments being the first structures to be damaged in hyperflexion. models,20 is that spinal flexion stretches and thins the posterior anulus so
The strong capsular ligaments of the apophyseal joints resist flexion the that it (rather than the vertebral endplate) fails first when compressive loading
most followed by the disk.1 The ligamentum flavum has such a high content increases nuclear pressure. Herniating anulus fibrosus often takes with it
of elastin that it can be stretched by up to 100% in full flexion even though fragments of bone or hyaline cartilage from the endplate,21 and composition
it is the only intervertebral ligament to be prestressed in the upright “neutral” of the herniation is likely to influence subsequent symptoms. In the laboratory,
position. The posterior longitudinal ligament is weaker than the anulus to typical herniated disk material swells by 200% to 300% in just a few hours
which it adheres and therefore offers little protection from hyperflexion. A before shrinking again over a period of several days as it loses proteoglycans.1
plexus of the mixed (autonomic–sympathetic) sinuvertebral nerve lies within This transient phenomenon could explain why sciatica sometimes develops
the posterior longitudinal ligament, thus suggesting that its primary function several hours after a back injury and then regresses naturally. Degenerative
may be to serve as a “nerve net” that is particularly sensitive to deformation changes seen in disk herniations retrieved at surgery appear to follow herniation
or abnormalities in the underlying disk. Backward bending (spinal extension) rather than precede it,22 suggesting that prior disk degeneration is not essential
is resisted by the bony surfaces of the neural arch, with most resistance for prolapse to occur. This has a medicolegal significance.
coming from either the facet joints or the spinous processes, depending on
individual variations in anatomy.1 The fact that most disks are wider from
side to side than from front to back suggests that they resist lateral bending
strongly, together with the apophyseal joint on the side that is being com- Excessive flexion tears the interspinous and supraspinous ligaments followed
pressed. In life, lateral bending is often combined with flexion when individuals by the capsular ligaments of the apophyseal joints. Extreme hyperflexion is
bend awkwardly to pick up something that is not directly in front of her or required to tear the posterior wall of the disk or the ligamentum flavum at
him. Such bending movements involve extra stretching of one posterolateral a bending moment of approximately 60 Nm.1 For comparison, injury in
corner of the disk. backwards bending (extension) can occur at 35 Nm, and equivalent values
for the cervical spine are 7 and 8 Nm, respectively. Little is known about
injuries to the iliolumbar ligaments or their role in back pain.
It is often supposed that all mechanical loading is harmful to the spine even Spondylolysis represents a fracture of the pars interarticularis that occurs
though repetitive mechanical loading is well known to strengthen muscles either unilaterally or bilaterally in the lower lumbar spine. It can be repro-
and bones according to the principles of adaptive remodeling.13 Articular duced in cadaveric spines by bending the neural arch cranially relative to
cartilage and intervertebral disks also can become mechanically conditioned the rest of the vertebra or by simulating spinal flexion movements that
if given enough time.14,15 Evidently, it is not mechanical loading that should cause ligaments and muscles to pull the neural arch caudally.1 Alternating
be avoided but mechanical overload, which physically disrupts tissues. A movements in extension and flexion therefore have the potential to bend
clinically relevant injury such as a ligament sprain is probably marked by the neural arch up and then down,23 creating stress reversals, which are
the “elastic limit” when the structure begins to undergo nonreversible problematic for bone metabolism because compression stimulates mineral
deformation. Repetitive loading can cause microscopic damage to accumulate, deposition, but tension stimulates collagen deposition. This could explain
so that “fatigue failure” occurs eventually even though the maximum applied why spondylolysis is associated with sports such as cricket and gymnastics,
forces may remain below 50% of the structure’s known strength. From a which involve alternating lumbar flexion and extension. Adolescents are
biomechanics standpoint, moderate and repetitive spinal loading should be most affected because the reduced mineralization in their vertebrae allows
encouraged, provided that it is not sufficient to cause injury. In the words much greater angular movements at the pars.23 The L5 vertebra is affected
of Nietzsche, “What does not kill him makes him stronger!” most because it joins the mobile spine to the relatively immobile sacrum,
and it often lies at a steep angle to the horizontal such that the gravitational
shearing force (S in Fig. 7.1) contributes to bending the neural arch backward
Vertebral endplates are comprised of a thin layer of hyaline cartilage weakly Spondylolisthesis (a forward slip of a vertebra) often follows bilateral
bonded to a supporting plate of cortical bone. The cartilage layer acts like spondylolysis in adolescents. This is not inevitable, however, suggesting that
a biologic filter and helps to maintain pressure in the adjacent disk nucleus strong back muscles can keep the vertebrae in place.
by strongly resisting fluid flow out of it. The bone plate must be thin and
porous so that sufficient metabolites can be transported from blood vessels
in the vertebral bodies into the adjacent avascular disks.16 Unfortunately,
this compromises the endplate’s mechanical function, so that it bulges markedly These small synovial joints are most likely to be overloaded by excessive
into the vertebral body whenever the hydrostatic pressure in the adjacent shear or torsional loading or by compressive loading following intervertebral
nucleus rises to high levels. Not surprisingly, the central vertebral endplate disk narrowing.6 Torsion can be particularly damaging because it concentrates
is the “weak link” of the vertebral column and is the first structure to be loading on only one of the two joints. The inferior and superior margins of
damaged when compressive loading becomes excessive.1 This occurs at a the articular surfaces are most frequently affected by cartilage loss and by
compressive force between 2 and 12 kN, depending on specimen gender, osteophytes,12 suggesting that damage can also occur when the spine is
size, and age. Cranial endplates (relative to the vertebra) are fractured more heavily loaded in flexion or extension. Hyperextension of the spine can cause
often than caudal endplates because they are thinner and supported by less the inferior articular processes to bend and probably damage the apophyseal
dense trabecular bone.17 If nucleus pulposus is expressed through a fractured joint capsules.23

52 SECTION 1  Scientific Basis of Rheumatic Disease


a *

b d

FIG. 7.5  (a) and (b) show endplate-driven degeneration, and (c) and (d) show anulus-driven degeneration. All images show midsagittal
sections through lumbar disks and vertebrae. (a) Radiograph of a midsagittal section of a vertebral body, anterior on right. Note the
calcified Schmorl’s node within the vertebral body (asterisk), probably in response to prior vertical herniation of nucleus through the
endplate. Arrows indicate large osteophytes on the anterior vertebral cortex. (b) Endplate damage created in vitro has resulted in a
vertical herniation of nucleus pulposus (asterisk) and inward collapse of the inner anterior anulus (arrow). (c) Large radial fissure
(asterisk) in the posterior anulus of this degenerated disk. (d) Compressive and bending overload in vitro can cause some nucleus
pulposus (asterisk) to herniate through a radial fissure in the posterior anulus. (Adapted from Adams MA, Dolan P. Intervertebral disc
degeneration: evidence for two distinct phenotypes. J Anat 2012;221:497-506.)

WHIPLASH genetic inheritance. Current estimates show that the heritability of interver-
tebral disk degeneration is approximately 50%, falling to 30% in the lower
Whiplash is a painful bending injury to the neck that is typically caused by lumbar spine, in men, in whom mechanical loading is most severe,27 and
a low-velocity insult, such as a car “shunt.” Low velocity does not mean rising to more than 70% in some female populations, who presumably load
that low force is involved. Laboratory simulations24 show that a rear impact their spines less. These details highlight the complementary role of genetic
thrusts the shoulders forward relative to the head such that the neck is and environmental influences. Gene polymorphisms associated with disk
initially distorted into an S shape, involving hyperextension of the lower degeneration code for important matrix constituents such as collagens and
cervical levels and variable flexion and extension of the upper cervical levels proteoglycans, various signaling molecules and receptors, and matrix-degrading
(Fig. 7.7). This initial distortion is followed by general hyperextension. enzymes, suggesting that they influence disk strength and healing ability.28
Finally, depending on the nature of the impact and the position of the Also, genes influence physical factors such as body weight and the length
headrest, the neck can be thrown into hyperflexion. Significant movements of internal lever arms (see Fig. 7.2).
occur in less than 50 msec, before the neck muscles can resist them, so the
head is thrown around relative to the slender neck like a “balloon on a
stick.” The cervical spine has approximately 45% of the compressive strength
of the lumbar spine but only 20% of its strength in bending,25 which suggests Loss of proteoglycans and water from aging intervertebral disks mostly affects
that the neck is particularly vulnerable to bending injury. the nucleus, where it leads to a smaller hydrostatic region exhibiting lower
The frequent involvement of both flexion and extension ensures that pressure, coupled with high concentrations of compressive stress (force per
practically any structure in or around the cervical spine can be injured during unit area) within the anulus.10 Collagen cross-linking increases with age,
whiplash.26 The site of injury probably moves away from the sagittal midline especially by nonenzymatic glycation, so that cartilage becomes stiffer and
(perhaps to an apophyseal joint) if the victim turns his or her head at the more vulnerable to injury.29 The accumulation of small defects in the anulus
time of impact. Any warning sounds would allow the victim to contract his ensures that its strength does not increase with stiffness; in fact, strength
or her neck muscles in alarm, thereby causing a higher proportion of compres- can fall slightly.15 The vertebral endplates and their supporting trabeculae
sion to bending to act on the cervical spine and increasing the risk of injury also weaken with age, in common with most bone structures, and endplates
to muscles and intervertebral disks. As with any other human behavior, develop increasing concavity on the disk side. This further reduces nucleus
reports of symptoms persisting after whiplash are subject to a variety of pressure and concentrates stress in the anulus. Hence, increasing stress
psychosocial influences.26 concentrations, increasing stiffness, and accumulating structural defects
make middle-aged disks vulnerable to injury. Injury risk appears to decrease
in old age, possibly because older people are less active and have weaker
back muscles.30
Studies on human twins can reveal the “heritability” of a specific disease, Injuries to the anulus or endplate generate regions of very high and low
expressed as the percentage of the disease variance that is associated with stress within the disk (see Fig. 7.6, b). Disk cells respond to very high and

CHAPTER 7  Biomechanics of spinal degeneration 53


Compressive 3
stress (MPa)
2.5 1.5

1.5 1
1 (hydrostatic)


0 P
0 10 20 30 40 0
Distance across disk (mm) 0 10 20 30 40
Posterior Anterior
anulus anulus Distance across disk (mm)
a Vertical Horizontal b

FIG. 7.6  (a) Photograph showing a mature nondegenerated human intervertebral disk cut through in the midsagittal plane. The graph below it shows how the
horizontal and vertical components of compressive stress vary within such a disk. There is a central region of hydrostatic pressure (the “functional nucleus”)
and a small concentration of compressive stress in the posterior anulus. (b) Internal disruption of the anulus lamellae and the damaged endplate indicate that
this young adult disk is degenerated. Stress distributions in such a disk show a decompressed nucleus and high stress concentrations (arrows) in the anulus.
The generally low disk stress suggests that some compression has been transferred to the neural arch. Approximately 1 MPa = 10 kg/cm2. (Adapted from
Adams MA, Bogduk N, Burton K, Dolan P. The biomechanics of back pain. 3rd ed. Edinburgh: Churchill Livingstone; 2013.)

their local mechanical environment, and injury uncouples this local environ-
STAGES OF REAR-IMPACT WHIPLASH ment from the physical demands placed on the whole disk. (In bone, the
integrated network of osteocytes seems able to overcome this problem.) The
concept of injury-driven disk degeneration is well supported by evidence
from large-animal experiments,32 human organ culture,33 and longitudinal
observations on living humans.34 Metabolite transport difficulties play a role
in this process by limiting disk cell density35 so that healing is frustrated.
However, the mechanism outlined in Fig. 7.8 depends on abnormal disk cell
metabolism after injury rather than reduced metabolism resulting from
inadequate nutrition and cell death.36 It implies that degeneration is an active
process and is consistent with the recent finding that endplate permeability
and porosity increase with age and degeneration.16
Disk injury also breaks down important barriers. Endplate damage could
explain how anaerobic bacteria enter the disk,37 possibly from blood vessels
in the vertebral body, and how cytokines from a degenerated disk create
painful inflammatory-like (“Modic”) changes in the adjacent vertebral body.38
0 50 msec 100 msec Complete radial fissures in the anulus (see Fig. 7.5, c) allow displaced nucleus
tissue to elicit pain from the peripheral anulus and nerve roots,39 facilitate
FIG. 7.7  During a rear-impact “whiplash” injury, the neck adopts an S-shaped deformity ingrowth of nerves and blood vessels,40 and provide a route for increased
after only 50 msec, before it can be protected by muscle action. Subsequently, the movement of bacteria and cytokines.
neck is thrown into hyperextension, possibly followed by hyperflexion if the head
bounces off a headrest. (Reproduced from Adams MA, Bogduk N, Burton K, Dolan TWO DISK DEGENERATION PHENOTYPES?
P. The biomechanics of back pain. 3rd ed. Edinburgh: Churchill Livingstone; 2013.)
It may be useful to distinguish between two disk degeneration “pheno-
types”41 as shown in Fig. 7.5. Endplate-driven degeneration (when the
initiating injury is to the endplate) typically involves inward collapse of
low stress by synthesizing fewer proteoglycans and more matrix-degrading the anulus, has a high heritability, mostly affects disks in the upper lumbar
enzymes.31 This cellular response is “aberrant” in the sense that it is the and thoracic spine, and starts to develop before age 30 years. Anulus-driven
reverse of what would be required to repair and repressurize the injured degeneration (when the initiating injury involves radial fissure formation
disk, and it provides a mechanism for progressive disk degeneration in the anulus) typically involves nucleus migration, has a low heritability,
(Fig. 7.8). The underlying problem for disk cells is that they can detect only mostly affects disks in the lower lumbar spine, and develops progressively

54 SECTION 1  Scientific Basis of Rheumatic Disease
These mechanical changes are sufficient to create “segmental instability”
SUMMARY OF THE BIOMECHANICS OF DISK DEGENERATION in the motion segment, as manifested by reduced resistance to bending
and an increase in horizontal shearing movements.9 In effect, the segment
“wobbles” and has an increased region of free play (or “neutral zone”).
The situation is more serious if the endplate becomes damaged because
High loading Weak tissues
this causes immediate and gross decompression of the nucleus.18 Segmental
instability is associated with back pain, although the supporting evidence
is slight. Pain could arise from the stress concentrations in the anulus (and
Structural failure apophyseal joints) that accompany nuclear decompression rather than from
of matrix the small movements themselves. Instability should not be confused with
hypermobility, which is a systemic condition involving increased ranges
of movement.
Weaker Abnormal “Frustrated
matrix matrix stress healing” VERTEBRAL BODY OSTEOPHYTES
Segmental instability is probably a transient phenomenon that is corrected
by the growth of osteophytes on the margins of the vertebral bodies (see
Abnormal Fig. 7.5, a) adjacent to the “wobbling” disk. Osteophytes can be created by
metabolism an injury to the anulus47 and grow in conjunction with disk narrowing,45
presumably because increased bulging of the anulus pulls on the periosteum.
The primary mechanical effect of vertebral body osteophytes is to increase
FIG. 7.8  Disk structural failure, resulting from high mechanical loading or weak the disk’s resistance to bending,48 thereby reversing the initial destabilizing
tissues, creates abnormal matrix stresses that disturb disk cell metabolism. This influence of disk degeneration. In this light, osteophytes can be viewed as
further weakens the matrix, leading to a “vicious circle” of further disruption and adaptive rather than degenerative, although they may provoke symptoms
impaired metabolism. The process can be likened to “frustrated healing.” (Reproduced from structures in the intervertebral foramen.
from Adams MA, Bogduk N, Burton K, Dolan P. The biomechanics of back pain.
3rd ed. Edinburgh: Churchill Livingstone; 2013.)
Anulus collapse in advanced disk degeneration brings the adjacent neural
arches closer together so that they resist up to 90% of the compressive
force acting on the lumbar spine.6 Abnormal load bearing creates high
after age 30 years. The structural defects that initiate the two processes concentrations of compressive stress in the inferior margins of the joint,
both act to decompress the disk nucleus, making it less likely that the especially in habitual upright postures.1 In addition, direct bone-on-bone
other defect could occur subsequently, and in this sense the two phenotypes contact between the inferior articular processes and the laminae below
are distinct. can lead to gross bone remodeling and osteophyte formation. A close
relationship has been noted in cadaveric apophyseal joints between high
load-bearing and osteoarthritic changes, including cartilage loss, bone eburna-
DISKOGENIC PAIN tion, and sclerosis.49 Imaging studies in vivo have long suggested such a
Back pain is associated with structural defects in the anulus39 and endplate42 relationship.
rather than with age-related proteoglycan loss. These defects allow blood
vessels and nerves to penetrate the normally avascular disks in patients with
back pain,43 probably because they represent focal regions of low mechanical
pressure and low concentration of proteoglycans.40 In a healthy disk, pro- Anulus collapse decreases the height of the intervertebral foramen. Its
teoglycans act to inhibit nerve ingrowth, and the pressure is sufficiently high anteroposterior diameter is also decreased by radial bulging of the disk and
to collapse capillaries. Ingrowing nerves could be sensitized by inflammatory by osteophyte growth. These changes can combine to cause a severe reduction
cytokines or by bacteria and provoked by the high stress concentrations in the cross-sectional area of the foramen and subsequent trapping of the
found in disrupted disks (see Fig. 7.6, b). exiting nerve. The consequences of spinal stenosis (leg or buttock pain on
walking that is relieved by spinal flexion) are usually most severe at L5 to
S1, which has a particularly small foramen.
High stress gradients in the degenerated anulus44 may be able to break the
cross-ties that hold lamellae together, allowing them to collapse inward (see
Fig. 7.6, b) and outward so that disk height reduces by 3% to 4% per year.45 The two apophyseal joints at each spinal level often show pronounced left–right
This suggests a timescale of 20 to 30 years for human disk degeneration to asymmetry, a condition known as tropism. If the affected joints become
run its full course. heavily load bearing as a result of disk narrowing (see earlier), their oblique
articular surfaces will create an unbalanced torque about the long axis of
the spine that would promote axial rotation. Axial rotation is mechanically
DETECTING AND QUANTIFYING DISK DEGENERATION “coupled” to lateral bending and can also be coupled to flexion and extension,
Magnetic resonance imaging (MRI) can predict disk proteoglycan and water so a “triplanar” spinal deformity could develop that involves permanent
content,46 but this is no more a measure of disk degeneration than bone twisting, lateral bending, and flexion of the thoracolumbar spine. This is a
mineral density (BMD) values are a measure of osteoporosis. In both cases, plausible explanation for degenerative scoliosis, but the mechanism has not
individual measurements must be compared with “normal values” (possibly been proved.
stratified for age, gender, and spinal level) before any disease process can
be inferred. It remains to be seen if T-scores or Z-scores derived from spinal
MRI scans can quantify disk degeneration as accurately as BMD T-scores
characterize osteoporosis. Old vertebrae become vulnerable to fractures or deformities, which are
conventionally classified as “anterior wedge” (when the anterior vertebral
body loses more height than the posterior), “biconcave” (when both endplates
SPINAL DEGENERATIVE CASCADE bulge markedly into the vertebral body), or “crush” (when both the anterior
Intervertebral disks play such an important role in spinal mechanics that and posterior cortices collapse). The first of these types, the wedge fracture,
disk failure usually initiates a cascade of adverse changes in adjacent tissues. is the most common, and when it affects a number of adjacent vertebrae,
the spine develops a kyphotic deformity (“senile kyphosis”), which is common
in very old women.
SEGMENTAL INSTABILITY The underlying cause is systemic loss of BMD, which is partly a result
Loss of nucleus pressure and volume removes tension from the surrounding of declining levels of sex hormones. Physical activity also declines with age
anulus, and loss of disk height creates slack in the intervertebral ligaments. as muscles weaken,30 and this provides a reduced mechanical stimulus to

CHAPTER 7  Biomechanics of spinal degeneration 55

maintain bone mass.13 But these systemic influences do not explain why the
anterior vertebral body is so often, and so severely, affected. ABNORMAL LOAD SHARING IN A DEGENERATED SPINE
Local mechanical factors are involved, and they depend on disk degenera-
tion. A healthy disk presses evenly on the entire vertebral body, even when
the spine is moderately flexed or extended, but a degenerated disk does not.
In habitual upright postures, a narrowed intervertebral disk allows the neural 10% 26% 63%
arch to resist a high proportion of the compressive force acting on the spine,
so that the anterior anulus and anterior vertebral body become “stress
shielded.”50 This leads to chronic anterior bone loss, as shown in Fig. 7.9.
Unfortunately, flexion movements can then disengage the neural arches and
transfer most of the compressive force onto the weakened anterior vertebral
body,50 thereby creating an anterior wedge fracture and leading to kyphotic
It is sometimes assumed that kyphotic posture is a cause of anterior wedge
deformity rather than a consequence, but this is mistaken. If increased habitual
loading of the anterior vertebral body was threatening it with collapse, this
region of bone would have increased bone density rather than the reduced Erect posture
density actually observed.

FIG. 7.9  Origins of vertebral anterior “wedge” fractures adjacent to a degenerated

disk. (Left) The three percentages show how an applied compressive force is distributed
between the anterior half of the vertebral body and disk, the posterior half of the
vertebral body and disk, and the neural arch: average data for cadaveric specimens
with severely degenerated disks tested in the simulated erect standing posture.
(Right) Plain radiograph of such a specimen showing how image intensity (which
reflects bone mineral density) is very low anteriorly and very high in the apophyseal
joints. (Image and data adapted from Adams MA, Pollintine P, Tobias JH, et al.
Intervertebral disc degeneration can predispose to anterior vertebral fractures in the
thoracolumbar spine. J Bone Miner Res 2006;21:1409-16.)

1. Adams M, Bogduk N, Burton K, et al. The biomechanics of 16. Rodriguez AG, Slichter CK, Acosta FL, et al. Human disc 29. DeGroot J, Verzijl N, Wenting-Van Wijk MJ, et al.
back pain. 3rd ed. Edinburgh: Churchill Livingstone; nucleus properties and vertebral endplate permeability. Accumulation of advanced glycation end products as a
2013. Spine. 2011;36:512-520. molecular mechanism for aging as a risk factor in
2. Dolan P, Earley M, Adams MA. Bending and compressive 17. Zhao FD, Pollintine P, Hole BD, et al. Vertebral fractures osteoarthritis. Arthritis Rheum. 2004;50:1207-1215.
stresses acting on the lumbar spine during lifting usually affect the cranial endplate because it is thinner 30. Sinaki M, Nwaogwugwu NC, Phillips BE, et al. Effect
activities. J Biomech. 1994;27:1237-1248. and supported by less-dense trabecular bone. Bone. of gender, age, and anthropometry on axial and
3. Vascancelos D. Compression fractures of the vertebra 2009;44:372-379. appendicular muscle strength. Am J Phys Med Rehabil.
during major epileptic seizures. Epilepsia. 18. Dolan P, Luo J, Pollintine P, et al. Intervertebral disc 2001;80:330-338.
1973;14:323-328. decompression following endplate damage: implications 31. Handa T, Ishihara H, Ohshima H, et al. Effects of
4. Snook SH, Webster BS, McGorry RW, et al. The reduction for disc degeneration depend on spinal level and age. hydrostatic pressure on matrix synthesis and matrix
of chronic nonspecific low back pain through the control Spine. 2013;38:1473-1481. metalloproteinase production in the human lumbar
of early morning lumbar flexion. A randomized 19. Wade KR, Robertson PA, Thambyah A, et al. “Surprise” intervertebral disc. Spine. 1997;22:1085-1091.
controlled trial. Spine. 1998;23:2601-2607. loading in flexion increases the risk of disc herniation 32. Holm S, Holm AK, Ekstrom L, et al. Experimental disc
5. Sato K, Kikuchi S, Yonezawa T. In vivo intradiscal due to annulus-endplate junction failure: a mechanical degeneration due to endplate injury. J Spinal Disord Tech.
pressure measurement in healthy individuals and in and microstructural investigation. Spine. 2004;17:64-71.
patients with ongoing back problems. Spine. 2015;40:891-901. 33. Alkhatib B, Rosenzweig DH, Krock E, et al. Acute
1999;24:2468-2474. 20. Schmidt H, Kettler A, Rohlmann A, et al. The risk of disc mechanical injury of the human intervertebral disc: link
6. Pollintine P, Przybyla AS, Dolan P, et al. Neural arch prolapses with complex loading in different degrees of to degeneration and pain. Eur Cell Mater. 2014;28:
load-bearing in old and degenerated spines. J Biomech. disc degeneration—a finite element analysis. Clin Biomech 98-111.
2004;37:197-204. (Bristol, Avon). 2007;22:988-998. 34. Kerttula LI, Serlo WS, Tervonen OA, et al. Post-traumatic
7. Bazrgari B, Shirazi-Adl A, Arjmand N. Analysis of squat 21. Lama P, Zehra U, Balkovec C, et al. Significance of findings of the spine after earlier vertebral fracture in
and stoop dynamic liftings: muscle forces and internal cartilage endplate within herniated disc tissue. Eur Spine young patients: clinical and MRI study. Spine.
spinal loads. Eur Spine J. 2007;16:687-699. J. 2014;23:1869-1877. 2000;25:1104-1108.
8. Sanchez-Zuriaga D, Adams MA, Dolan P. Is activation of 22. Lama P, Le Maitre CL, Dolan P, et al. Do intervertebral 35. Junger S, Gantenbein-Ritter B, Lezuo P, et al. Effect of
the back muscles impaired by creep or muscle fatigue? discs degenerate before they herniate, or after? Bone Joint limited nutrition on in situ intervertebral disc cells under
Spine. 2010;35:517-525. J. 2013;95-B:1127-1133. simulated-physiological loading. Spine.
9. Zhao F, Pollintine P, Hole BD, et al. Discogenic origins of 23. Green TP, Allvey JC, Adams MA. Spondylolysis. Bending 2009;34:1264-1271.
spinal instability. Spine. 2005;30:2621-2630. of the inferior articular processes of lumbar vertebrae 36. Urban JP, Smith S, Fairbank JC. Nutrition of the
10. Adams MA, McNally DS, Dolan P. ‘Stress’ distributions during simulated spinal movements. Spine. intervertebral disc. Spine. 2004;29:2700-2709.
inside intervertebral discs. The effects of age and 1994;19:2683-2691. 37. Albert HB, Lambert P, Rollason J, et al. Does nuclear
degeneration. J Bone Joint Surg Br. 1996;78:965-972. 24. Kaneoka K, Ono K, Inami S, et al. Motion analysis of tissue infected with bacteria following disc herniations
11. Skrzypiec DM, Pollintine P, Przybyla A, et al. The internal cervical vertebrae during whiplash loading. Spine. lead to Modic changes in the adjacent vertebrae? Eur
mechanical properties of cervical intervertebral discs as 1999;24:763-769, discussion 770. Spine J. 2013;22:690-696.
revealed by stress profilometry. Eur Spine J. 25. Przybyla AS, Skrzypiec D, Pollintine P, et al. Strength of 38. Albert HB, Manniche C. Modic changes following lumbar
2007;16:1701-1709. the cervical spine in compression and bending. Spine. disc herniation. Eur Spine J. 2007;16:977-982.
12. Tischer T, Aktas T, Milz S, et al. Detailed pathological 2007;32:1612-1620. 39. Videman T, Nurminen M. The occurrence of anular tears
changes of human lumbar facet joints L1-L5 in elderly 26. Bannister G, Amirfeyz R, Kelley S, et al. Whiplash injury. and their relation to lifetime back pain history: a
individuals. Eur Spine J. 2006;15:308-315. J Bone Joint Surg Br. 2009;91:845-850. cadaveric study using barium sulfate discography. Spine.
13. Burr DB, Robling AG, Turner CH. Effects of 27. Battie MC, Videman T, Levalahti E, et al. Genetic and 2004;29:2668-2676.
biomechanical stress on bones in animals. Bone. environmental effects on disc degeneration by phenotype 40. Stefanakis M, Al-Abbasi M, Harding I, et al. Annulus
2002;30:781-786. and spinal level: a multivariate twin study. Spine. fissures are mechanically and chemically conducive to
14. Eckstein F, Hudelmaier M, Putz R. The effects of exercise 2008;33:2801-2808. the ingrowth of nerves and blood vessels. Spine.
on human articular cartilage. J Anat. 2006;208:491-512. 28. Mayer JE, Iatridis JC, Chan D, et al. Genetic 2012;37:1883-1891.
15. Skrzypiec D, Tarala M, Pollintine P, et al. When are polymorphisms associated with intervertebral 41. Adams MA, Dolan P. Intervertebral disc degeneration:
intervertebral discs stronger than their adjacent vertebrae? disc degeneration. Spine J. 2013;13:299- evidence for two distinct phenotypes. J Anat.
Spine. 2007;32:2455-2461. 317. 2012;221:497-506.

56 SECTION 1  Scientific Basis of Rheumatic Disease
42. Wang Y, Videman T, Battie MC. ISSLS Prize winner: 45. Hassett G, Hart DJ, Manek NJ, et al. Risk factors for 48. Al-Rawahi M, Luo J, Pollintine P, et al. Mechanical
lumbar vertebral endplate lesions: associations with disc progression of lumbar spine disc degeneration: the function of vertebral body osteophytes, as revealed by
degeneration and back pain history. Spine. Chingford Study. Arthritis Rheum. 2003;48:3112- experiments on cadaveric spines. Spine. 2011;36:770-777.
2012;37:1490-1496. 3117. 49. Robson-Brown K, Pollintine P, Adams MA. Biomechanical
43. Freemont AJ, Peacock TE, Goupille P, et al. Nerve 46. Johannessen W, Auerbach JD, Wheaton AJ, et al. implications of degenerative joint disease in the
ingrowth into diseased intervertebral disc in chronic back Assessment of human disc degeneration and proteoglycan apophyseal joints of human thoracic and lumbar
pain. Lancet. 1997;350:178-181. content using T1rho-weighted magnetic resonance vertebrae. Am J Phys Anthropol. 2008;136:318-326.
44. Stefanakis M, Luo J, Pollintine P, et al. ISSLS Prize imaging. Spine. 2006;31:1253-1257. 50. Adams MA, Pollintine P, Tobias JH, et al. Intervertebral
winner: mechanical influences in progressive 47. Lipson SJ, Muir H. 1980 Volvo award in basic science. disc degeneration can predispose to anterior vertebral
intervertebral disc degeneration. Spine. Proteoglycans in experimental intervertebral disc fractures in the thoracolumbar spine. J Bone Miner Res.
2014;39:1365-1372. degeneration. Spine. 1981;6:194-210. 2006;21:1409-1416.

Scientific basis of pain
Hans-Georg Schaible 8  

Key Points or disease of nociceptive neurons in the peripheral or the central nervous
■ Nociception is the encoding and processing of noxious stimuli in the nervous
system. This pain does not primarily signal tissue damage. It often has
system; the subjective sensation of pain is evoked by nociception and influenced by
an abnormal burning or electrical character and can be persistent or epi-
psychological and social factors.
sodic (e.g., trigeminal neuralgia). It may be combined with hyperalgesia
and allodynia. Thus, even touching the skin with a soft brush can cause
■ Clinically relevant pain is classified as nociceptive, when tissue is inflamed or
intense pain. Causes of neuropathic pain include nerve or plexus damage,
damaged, or as neuropathic, when nerve fibers or nerve cells are affected.
metabolic diseases such as diabetes mellitus, herpes zoster, and others.
■ Peripheral nociceptors are slowly conducting, thinly myelinated or unmyelinated
Damage to central pain processing neurons (e.g., in the thalamus) can cause
nerve fibers that respond to tissue-damaging mechanical, thermal, and chemical
central pain.1,3
Usually pain is called “chronic” when it lasts longer than 6 months.
■ The central nociceptive system consists of neurons in the spinal cord, the brain Chronic pain may result from persistent nociception (with a chronic disease),
stem, the thalamus, and cortical areas that process noxious stimuli; the conscious but it is often significantly influenced by psychological and social factors.
pain sensation with its sensory-discriminative and affective components is generated It can be accompanied by neuroendocrine dysregulation, fatigue, dysphoria,
at the thalamocortical level. and impaired physical and even mental performance.4
■ During inflammation, peripheral nociceptors are sensitized to mechanical or thermal
stimuli (or to both) by inflammatory mediators (peripheral sensitization).
■ Damaged nerve fibers may show action potentials that are evoked at the site of THE PERIPHERAL BASIS OF PAIN
lesion and in the neuronal cell bodies (ectopic discharges).
■ Peripheral sensitization and ectopic discharges may induce a state of STRUCTURE AND DUAL FUNCTION OF
hyperexcitability in the central nociceptive system (central sensitization) that
increases the gain of central nociceptive processing.
■ The brain areas involved in pain processing may show atrophy or reorganization Nociceptors are sensory neurons with thinly myelinated Aδ or unmyelinated
■ Nerve fibers originating in the brain stem can enhance or inhibit the spinal C fiber axons. The cell bodies of nociceptors are located in the dorsal root
nociceptive processing (descending facilitation or inhibition). ganglia. Their peripheral branches form sensory endings (“free nerve endings”)
in the innervated tissue, and their central branches terminate in the dorsal
horn of the spinal cord or in the brain stem, where they activate synaptically
nociceptive dorsal horn neurons (see Fig. 8.1).5
Many nociceptors have a dual function. They encode noxious stimuli
and transmit this information to the spinal cord (sensory function). In addition,
NOCICEPTION AND PAIN they transport neuropeptides such as substance P and calcitonin gene-related
Pain is an unpleasant sensory and emotional experience that is evoked by peptides (CGRP) from the cell body to the periphery and release these
actual or potential noxious (i.e., tissue damaging) stimuli or by tissue injury mediators in the tissue on stimulation. There, these neuropeptides induce
or is described in such terms. Because pain is a subjective experience, it vasodilatation, plasma extravasation, attraction of macrophages or degranula-
cannot be measured objectively. Nociception is the encoding and processing tion of mast cells, or other processes that elicit a neurogenic inflammation.
of noxious stimuli in the nervous system and can be measured objectively Such neurogenic components contribute significantly to many inflammatory
(e.g., with electrophysiologic recordings). Under normal conditions, the diseases.6
relationship between nociception and pain is relatively precise and predictable.
A stimulus that is noxious will usually evoke a subjective pain response.
However, under clinically relevant conditions, the relationship between
nociception and pain may not be strict, particularly when the pain is chronic Most nociceptors respond to noxious mechanical stimuli (painful pressure,
(see later). squeezing or cutting the tissue), noxious thermal stimuli (heat or extreme
Neurons in the peripheral and central nociceptive system that encode cold), and chemical stimuli and are therefore called polymodal. Noxious
noxious stimuli form the nociceptive system. A simplified scheme of the stimuli evoke a sensor potential in the sensory ending (transduction). When
nociceptive system is shown in Fig. 8.1. Peripheral nociceptive neurons the depolarization is sufficiently strong, action potentials are triggered and
(nociceptors) innervate the skin, deep tissue, and most visceral organs. They conducted by the axon to the spinal cord or the brain stem (Fig. 8.2).5
encode noxious stimuli applied to the tissue. The central nociceptive system In joints, nociceptors innervate mainly the fibrous capsule, ligaments,
consists of sensory nociceptive neurons in the spinal cord, spinal reflex adipose tissue, menisci, and the synovial layer. The cartilage is normally not
pathways, and ascending tracts that activate the brain stem and supraspinal innervated. A typical joint nociceptor is activated by strong pressure to the
structures in the thalamus and cortex. Corticothalamic networks produce joint (e.g., hitting the joint) and by noxious movements (i.e., painful rotation
the conscious pain response.1,2 against the resistance of the tissue). It is not usually activated by movements
and positions in the working range.7 Nociceptors in the muscle are located
in the muscle belly and in the tendon. They respond to noxious (painful)
TYPES OF PAIN compression of the muscle, and they may be activated by muscle contraction
Application of a noxious stimulus to normal tissue elicits acute physiologic under ischemic conditions, and some of them are activated by noxious
nociceptive pain. This pain protects tissue from being further damaged because thermal stimuli. Normally, they do not respond to innocuous pressure and
withdrawal reflexes are usually elicited. Inflammation or injury cause to contractions of the muscle.5 Cutaneous nociceptors respond to noxious
pathophysiologic nociceptive pain. It may appear as spontaneous (resting) heat (in the range of 42°C to higher than 50°C), and they encode noxious
pain, hyperalgesia, allodynia, or any combination of these types of pain. mechanical stimuli such as squeezing.5 Visceral nociceptors respond to a
Hyperalgesia is a higher pain intensity felt on noxious heat stimulation variety of mechanical, thermal, and chemical stimuli.5
(thermal hyperalgesia) or noxious mechanical stimulation (mechanical Notably, not all nociceptors are polymodal. An important group of nocicep-
hyperalgesia). Allodynia is pain that is elicited by stimuli that are normally tors is relatively mechanoinsensitive and, in the skin, heat insensitive. Because
below the pain threshold.1 these nociceptors do not respond to noxious stimuli applied to normal tissue,
Whereas nociceptive pain is elicited by noxious stimulation of sensory they were called initially mechanoinsensitive or silent nociceptors.1,7 These
endings of nociceptors in the tissue, neuropathic pain is caused by injury nociceptors are “recruited” during inflammation (see later).

58 SECTION 1  Scientific Basis of Rheumatic Disease

AND THEIR SYNAPSES IN THE DORSAL HORN OF THE SPINAL CORD, With repetitive or strong noxious stimulation of the tissue, and in particular
AND ASCENDING TRACTS during inflammation, nociceptors are sensitized to stimuli such that they are
hyperexcitable. The excitation threshold of polymodal nociceptors decreases
such that even light, normally innocuous stimuli activate the fibers, and
Anterior cingulate silent nociceptors become excitable by innocuous and noxious stimuli.1,7
cortex (ACC) This peripheral sensitization produces enhanced input to the spinal cord and
Insula induces central sensitization (Fig. 8.3). Both peripheral and central sensitization
Prefrontal cortex (see later) cause primary hyperalgesia and allodynia at the site of inflammation.
In addition, central sensitization causes secondary hyperalgesia, enhanced
Cortical areas pain sensitivity in healthy tissue surrounding the site of inflammation
Medial thalamus SI, SII in (Fig. 8.4).
Lateral thalamus somatosensory Peripheral sensitization is a hallmark of many painful diseases of the
musculoskeletal system. It is a major pathophysiological pain mechanism
of primary inflammatory joint diseases such as rheumatoid arthritis (RA),
and it is also involved in osteoarthritic pain, various forms of myositis, and
many others. Because of the peripheral and central sensitization, movements
in the working range and palpation of the joints are painful, and the patients
may even experience pain in the absence of stimulation.
Peripheral tissue Medial
C fiber The sensory endings of nociceptors express ion channels and membrane
receptors that transduce the mechanical, thermal, and chemical stimuli into
Aδ fiber a sensor potential (see Fig. 8.2). Some sensor molecules have been identified.8
The best known is transient receptor potential vanilloid 1 (TRPV1), a ligand-
gated ion channel that is expressed in nociceptors but not in other peripheral
neurons. On opening, cations, in particular Ca2+, flow into the cell and
depolarize it. TRPV1 is opened by temperatures higher than 43°C, which
Sympathetic are felt as painful heat by humans; by chemicals that elicit burning pain,
axon such as capsaicin and ethanol applied to a wound; and by low pH (<5.9),
Motor axon
which occurs in inflamed tissue. TRPV1 is activated by metabolites of ara-
chidonic acid produced by lipoxygenases such as 12-hydroperoxyeicosaenoic
FIG. 8.1  Scheme of the nociceptive system with nociceptive free nerve endings in acid (12-HPETE) and by endocannabinoids such as anandamide and
the peripheral tissue, afferent nerve fibers, and their synapses in the dorsal horn N-arachidonoyl-dopamine (NADA). Furthermore, TRPV1 is sensitized via
of the spinal cord. From there, the medial and lateral spinothalamic tracts ascend second messengers by the inflammatory mediators bradykinin, prostaglandin
to the medial and lateral thalamus, and interneurons project into motor and sympathetic E2 (PGE2), extracellular adenosine triphosphate, glutamate, proteases,
reflex pathways. Note: Other ascending pathways such as the spinoreticular tract cytokines, and nerve growth factor (NGF). Thus, TRPV1 is one transducer
and dorsal column pathways are not displayed. SI and SII, Primary and secondary of noxious heat, and it is the crucial ion channel for inflammatory thermal
somatosensory cortices. hyperalgesia.8
Other TRP channels such as TRPV2, TRPV3, TRPV4, TRPM8, and
TRPA1 are also expressed in proportions of primary sensory neurons.
Some of them are coexpressed with TRPV1, but others are expressed in


inflammatory Cytokines Neurotrophins Others
FIG. 8.2  Sketch of the enlarged sensory ending of a
nociceptor in the tissue. At the bottom are ion channels
for the transduction of thermal, mechanical, and
chemical stimuli (they produce a sensor potential) and
Neuro- voltage-gated ion channels (opening of Na+ channels
peptides produces action potentials). The top and the left sides
show metabotropic receptors for chemical mediators.
Classical mediators are bradykinin, prostaglandin E2, and
others. ASIC, Acid-sensing ion channel; ATP, adenosine
Second messenger systems
triphosphate; P2X, P2 Purinoceptor X; TRP, transient
Adrenergic receptor protein. (From Schaible H-G, Ebersberger A,
mediators Natura G: Update on peripheral mechanisms of pain:
beyond prostaglandins and cytokines. Arthritis Res Ther

ASICs for P2X-receptor TRP for Channel for K+ Ca2+ Na+

protons for ATP thermal mechanical Voltage-gated
transduction transduction ion channels

CHAPTER 8  Scientific basis of pain 59

non-nociceptive neurons. For example, TRPM8 is opened by temperatures pathways involving protein kinase A, protein kinase C, sphingomyelinase,
that are felt to be cool; hence, TRPM8 is assumed to be the transducer in calmodulin, and p38 mitogen-activated protein kinase. The excitability of
non-nociceptive cold fibers. In general, the role of these other TRP chan- neurons may also be controlled by K+ channels (e.g., of the KCNQ family)
nels is less clear. TRP ion channels such as TRPA1 and TRPV4 may also and Ca2+ channels. Excitability is increased when voltage-gated K+ channels
be involved in the transduction of mechanical stimuli and in mechanical are closed (this evokes sustained depolarization of neurons) or when Ca2+
hyperalgesia, which is in many diseases of the joint more important than flows into the neuron through voltage-gated T-type channels.10
thermal hyperalgesia. However, the ionic basis of mechanonociception is The chemosensitivity of nociceptive neurons has two purposes. Chemo-
poorly understood.8 sensitivity either enables mediators to activate the neurons (to elicit action
When the sensory ending is sufficiently depolarized through the ion potentials) directly, or it renders neurons more excitable for mechanical and/
channels of transduction, voltage-gated Na+ channels are opened and action or thermal stimuli. Although some mediators such as adenosine triphosphate
potentials are triggered (see Fig. 8.2). Whereas nociceptive sensory neurons (ATP) may directly open ion channels (see Fig. 8.2), most mediators act on
express mainly the sodium channel types Nav1.7, Nav1.8, and Nav1.9, large- metabotropic receptors in the membrane. Sensitization of primary afferent
sized nonnociceptive neurons express mainly Nav1.1, Nav1.6, and Nav1.7, neurons during inflammation is generated by inflammatory mediators that
and some Nav1.8. Nav1.7, and Nav1.8 channels are directly involved in genera- act on specific membrane receptors (see Fig. 8.2). The ligand binding of
tion of the action potential. Whereas Nav1.7 channels are opened by these membrane receptors activates second-messenger systems, which then
depolarization near the resting membrane potential and are thus responsible affect both transduction molecules and voltage-gated ion channels such that
for the initial triggering of the action potential, Nav1.8 channels are opened they are more excitable (e.g., by phosphorylation). In the long term, some
at more depolarized levels. Nav1.9 channels do not contribute to the action of them may also regulate the expression of such molecules in the membrane.
potential itself but influence the threshold for the elicitation of action In addition, sensory nerve endings possess Toll-like receptors that signal
potentials.9 These channels can be up- or downregulated by second messenger molecular patterns of infectious agents, thus indicating an important role
of the nociceptive nerve fibers in the context of innate immunity.
“Classical” inflammatory mediators such as bradykinin and prostaglandins
activate or sensitize neurons within minutes. For instance, PGE2 acts on G
protein–coupled EP receptors to increase cyclic adenosine monophosphate.
FLOWCHART OF THE GENERATION OF PAIN This activates protein kinase A, which finally leads to a phosphorylation of
IN DIFFERENT PAIN STATES TRPV1 receptors and voltage-gated Na+ currents. Upon a single application,
the effects are quite short (minutes only). By contrast, injection of proinflam-
matory cytokines such as interleukin-6 (IL-6) and tumor necrosis factor-α
(TNF-α) into the joint leads to a slowly developing but persistent sensitization
Sensory-discriminative and of nociceptive afferents (i.e., an increase in the number of action potentials
affective pain response in response to a stimulus) that lasts at least several hours. In addition,
cytokines regulate the expression of receptors (e.g., TNF-α upregulates TRPV1
Thalamocortical systems receptors).11 In this respect, it is interesting that during inflammatory states,
macrophages invade the dorsal root ganglia, where their mediators (e.g.,
Brain stem
cytokines) can directly act on the somata. These effects of cytokines can be
hyperexcitability blocked by anticytokine therapy with biologic agents, and it is expected that
this also inhibits release of proinflammatory substance P and other neuro-
+ – peptides.11 Nerve growth factor is an essential neurotrophin for the develop-
ment of sensory nerve fibers. In adults, a large proportion of sensory nerve
Spinal cord hyperexcitability fibers remain dependent on the trophic effect of NGF. These neurons express
tyrosine kinase receptor A (TrkA) (specific for NGF), and NGF is required
for their structural and functional integrity. However, substantial amounts
Peripheral Pathologic ectopic
of NGF are also produced at inflammatory sites, where NGF acts as an
sensitization activity
inflammatory mediator. It enhances currents through TRPV1 channels and
reduces the threshold of thermal excitation. Long-term exposure to NGF
Trauma increases the expression of TRPV1, bradykinin and P2 Purinoceptor X (P2X)
Nerve damage
inflammation receptors, and Na+ channels, as well as synthesis of substance P and CGRP.
In addition, NGF stimulates inflammatory cells to release inflammatory
compounds. Thus, NGF is a key molecule for nociceptor biology, and
FIG. 8.3  Flowchart of the generation of pain in different pain states. Spinal cord neutralization of it has proved to be highly analgesic in humans.10
hyperexcitability (central sensitization) can result from both peripheral sensitization Nociceptors also express receptors for neuropeptides. Whereas substance
and from ectopic pathologic discharges in the afferent nerve fiber. The brain stem P and CGRP sensitize part of the neurons, opioid peptides and somatostatin
(also rendered hyperexcitable) provides a feedback to the spinal cord that is either inhibit the neurons. Thus, actual receptor sensitivity is due to a balance
inhibitory or facilitatory. between excitatory and inhibitory influences.10


FIG. 8.4  Central sensitization. (a) Regions of primary and hyperalgesia Central
secondary hyperalgesia produced by peripheral and sensitization 3 4
central sensitization. (b) Mechanisms of spinal
sensitization. 1, enhanced release of transmitters from Primary Peripheral
sensitized nociceptors; 2, enhanced excitability of hyperalgesia and central
postsynaptic neurons; 3, suprathreshold synaptic activation sensitization 1
by afferents from remote regions (in the normal,
nonsensitized state insufficient); and 4, spinal release of Secondary
mediators from glial cells and other cells. hyperalgesia
a b

60 SECTION 1  Scientific Basis of Rheumatic Disease
innocuous and noxious somatosensory stimuli. A role in the process of
PERIPHERAL MECHANISMS OF NEUROPATHIC PAIN memory formation and access has also been discussed.14
In healthy sensory nerve fibers, action potentials are generated in the sensory As shown in Figs 8.3 and 8.4, a, nociceptive processes in the periphery
endings on stimulation of the innervated tissue. By contrast, damaged nerve cause numerous central changes that determine the phenotype of pain. Such
fibers often show pathologic ectopic discharges—action potentials generated central processes were found at all levels of the neuraxis and are summarized
at the site of nerve injury or in the cell bodies of impaired fibers in dorsal in the following paragraphs.
root ganglia. Neuropathic pain may also be generated by intact nerve fibers
in the vicinity of injured nerve fibers that are affected by the process of CENTRAL SENSITIZATION (ESPECIALLY SPINAL
wallerian degeneration.3 Pathologic discharges may cause a state of central
sensitization (see Fig. 8.3).
Different mechanisms can produce ectopic discharges. First, after nerve Strong pathologic nociceptive input causes spinal cord hyperexcitability
injury, changes in the expression of Na+ channels may alter the membrane (central sensitization) (see Fig. 8.3). Central sensitization amplifies the synaptic
properties of the neuron such that rapid firing rates (bursting ectopic dis- processing of nociceptive input in the spinal cord, and thus the thalamocortical
charges) are favored. Changes in the expression of K+ channels in neurons system is more strongly activated. Together with peripheral sensitization at
have also been shown. Second, injured axons of primary sensory neurons the site of inflammation, the spinal hyperexcitability produces primary
may be directly excited by inflammatory mediators (e.g., by bradykinin, hyperalgesia at the inflamed region. Furthermore, central sensitization generates
nitrous oxide, and cytokines), released from white blood cells and Schwann a zone of secondary hyperalgesia in regions adjacent to and remote from the
cells around the damaged nerve fibers. Third, injured nerve fibers may be inflamed region even though the latter areas are in a healthy condition (see
affected by the sympathetic nervous system. Although the latter does not Fig. 8.4, a). For example, patients with advanced stages of osteoarthritis
activate primary afferents in normal tissue, injured nerve fibers may become (OA) often report widespread pain far beyond the OA joint and exhibit
sensitive to adrenergic mediators because neuronal adrenergic receptors are lowered pressure pain thresholds in cutaneous and subcutaneous structures
upregulated. Furthermore, after nerve injury, sympathetic nerve fibers may of the entire leg.15 Thus, central sensitization leads to an increase and spread
sprout in the dorsal root ganglion.1,3 Currently, the best treatment for peripheral of pain beyond the site of origin.
neuropathic pain consists of drugs that reduce the excitability of neurons Central sensitization can last as long as the nociceptive input persists
(e.g., carbamazepine or gabapentin). and disappear when the peripheral input is reduced. However, central sensitiza-
tion may also outlast the peripheral nociceptive process. In these cases,
nociceptive input has probably triggered a long-term potentiation, a persistent
THE CENTRAL BASIS OF PAIN increase of synaptic efficacy.16 Such a process could account for pain states
that persist even when the peripheral nociceptive input has disappeared. It
NOCICEPTIVE SPINAL CORD NEURONS is likely that persistent pain syndromes such as fibromyalgia are an expression
Nociceptors activate synaptically nociceptive dorsal horn neurons (see Fig. of persistent central sensitization.
8.1). The latter are either spinothalamic or spinoreticular tract neurons that Important mechanisms of spinal sensitization are displayed in Fig. 8.4,
ascend to supraspinal sites or are local interneurons that are part of segmental b. First, the sensitized nociceptors from the inflamed site create stronger
motor or vegetative reflex pathways. A noxious stimulus applied to normal input into the spinal cord; they release more transmitters at synapses at the
tissue causes a withdrawal reflex that removes the threatened part of the spinal neuron.1,7 This is a presynaptic mechanism of spinal sensitization. The
body from the damaging source. Under inflammatory conditions, the inflamed main transmitter of sensory afferents is glutamate, and peptidergic C fibers
tissue is kept in a position that activates nociceptors as little as possible. in addition release substance P and CGRP.5 Second, enhanced transmitter
Typically, nociceptive spinal cord neurons receive convergent inputs from release from sensitized nociceptors renders the postsynaptic neuron more
numerous sensory neurons in one organ or different organs. Parts of the excitable. This constitutes a postsynaptic mechanism of spinal sensitiza-
nociceptive dorsal horn neurons are activated by input only from the skin; tion. Mechanistically, glutamate activates N-methyl-D-aspartate (NMDA)
others are activated by input from deep tissue (i.e., muscles, tendons, and and non-NMDA receptors (both are ion channels) and metabotropic
joints). A large proportion of dorsal horn neurons receives input from both glutamate receptors in spinal cord neurons.5 During noxious stimulation,
skin and deep tissue, and others are excited by cutaneous, deep tissue, and both non-NMDA and NMDA receptors are opened by glutamate, and in
visceral stimulation. As a consequence, pain may be poorly localized (especially particular, influx of Ca2+ through the calcium-permeable NMDA receptors
in deep somatic tissue) and may even be projected. In particular, noxious triggers biochemical changes in the neurons that enhance the sensitivity of
stimulation of the viscera leads to a projection of pain into skin and deep the neurons. Administration of NMDA receptor antagonists prevents central
tissue that is supplied by nociceptors of the same segment.1,5 sensitization and reduces established hyperexcitability. In addition, many
The spinal cord is under the influence of descending tracts that inhibit spinal nociceptive neurons express receptors for substance P and CGRP, and
or facilitate the spinal nociceptive processing. These descending pathways through activation of these receptors, the enhanced release of substance P
originate from brain stem nuclei (in particular the periaqueductal gray matter and CGRP contributes to spinal sensitization.1,5 Third, the hyperexcitability
and nucleus raphe magnus) and descend in the dorsolateral funiculus of the of the spinal neuron makes it possible for the neuron to now be activated
spinal cord.12,13 even by primary afferents that are normally not able to cause suprathreshold
excitation of the neuron (dashed lines in Fig. 8.4.). Hence the total receptive
GENERATION OF THE CONSCIOUS PAIN RESPONSE IN THE field of the neuron is enlarged. As a consequence of this sensitization, a
higher proportion of neurons responds to stimulation of peripheral tissue.1,7
THALAMOCORTICAL SYSTEM Together, these changes produce the areas of secondary hyperalgesia or
The conscious pain response is produced in the thalamocortical system (see spreading pain.
Fig. 8.1). Different aspects of pain are produced by different networks. Analysis Fourth, synaptic processing is further modified by mediators that are
of the noxious stimulus with respect to its location, duration, and intensity produced in the spinal cord and create a change of the milieu there. Numerous
(the sensory-discriminative aspect of pain) is performed in the lateral thalamo- spinal neurons (and other cells) produce prostaglandins (mainly PGE2 and
cortical system, which consists of relay nuclei in the lateral thalamus and prostaglandin D2 [PGD2]).17 During peripheral inflammation, more PGE2 is
primary and secondary somatosensory cortices (SI and SII) in the postcentral released in the spinal cord and contributes to spinal sensitization. In particular,
gyrus. In these regions, innocuous and noxious stimuli are discriminated.2 cyclooxygenase 2–selective inhibitors reduce spinal hyperexcitability by
The affective aspect of pain (i.e., the noxious stimulus is felt to be unpleasant inhibition of spinal prostaglandin synthesis and, as proposed, by enhancing
and causes aversive reactions) is produced in the medial thalamocortical spinal endocannabinoids.18 By contrast, spinal PGD2 instead inhibits spinal
system, which consists of relay nuclei in the central and medial thalamus sensitization during peripheral inflammation.7 Glial cells in the spinal cord
and the anterior cingulate cortex, the insula, and the prefrontal cortex.2 can significantly contribute to the generation of pain states, particularly
These brain structures are part of the limbic system, and the insula may be during neuropathic pain19 but also in experimental models of OA.15 On
an interface between the somatosensory and the limbic system. Notably, the stimulation, neuroglia secrete a number of mediators such as nitrous oxide,
limbic regions are involved not only in pain processing. Specifically, the neurotrophins, and proinflammatory cytokines, and neutralization of such
anterior cingulate cortex is activated during different emotions, including mediators may attenuate hyperalgesia.19
sadness and happiness, and parts of the anterior cingulate cortex are also Importantly, inhibitory transmitters such as γ-aminobutyric acid (GABA)
involved in the generation of autonomic responses (they have projections and opioids are able to reduce spinal sensitization by decreasing transmitter
to regions that command autonomic output systems). Other cingulate regions release and depolarization of postsynaptic neurons. An important mechanism
are involved in response selection (they have projections to the spinal cord of spinal sensitization may be the loss of segmental inhibition that is normally
and the motor cortices) and in orientation of the body toward or away from provided by GABA-ergic interneurons.5

CHAPTER 8  Scientific basis of pain 61

In addition to the thalamocortical system, the ascending nociceptive

CHANGES IN THE BRAIN DURING CHRONIC PAIN information also activates nuclei in the brain stem (see Fig. 8.3). Pathways
from the brain stem support the excitation of the thalamocortical system,
but, in particular, the parabrachial nucleus of the brain stem forms a projection
Activation of cortical areas
into the amygdala. These structures belong to the limbic system, and their
Induction of changes activation generates fear.20
in the brain Induction of neuroplastic changes
(reorganization, brain atrophy)
(e.g., reduction of descending Fig. 8.3 shows that the brain stem is also activated at the origin of tracts
conditional pain modulation) which descend to the spinal cord. Most of these descending systems have
their origin in brain stem nuclei, which form the so-called descending
inhibition. Experimental evidence suggests that descending inhibition of neurons
Spinal sensitization Peripheral sensitization with input from inflamed areas is increased, at least in the acute stage of
inflammation, keeping spinal sensitization under control.12 However, in severe
FIG. 8.5  Changes in the brain during chronic pain induced in the course of chronic pain states such as OA or RA, some forms of descending inhibition,
musculoskeletal diseases.
called conditioning pain modulation (CPM) are almost absent, indicating
that endogenous pain control systems become insufficient (see Fig. 8.5).13,15
Thus, another mechanism of chronic pain is the lack of endogenous pain
control. Interestingly, this form of descending inhibition is restored after
NEUROPLASTICITY AT THE THALAMOCORTICAL LEVEL replacement of the osteoarthritic hip joint.15
In contrast, descending facilitation may support the expansion of receptive
As already mentioned, the activation of the thalamocortical system produces fields into healthy areas and thereby promote secondary hyperalgesia.12 In
the conscious pain response (see earlier). In chronic pain conditions, the the case of neuropathic pain, mainly descending facilitation has been observed.
thalamocortical system may show different types of changes (Fig. 8.5). In Thus, spinal–supraspinal–spinal loops are quite important in the generation
neuropathic conditions after a peripheral nerve lesion but also during chronic of pathologic pain.12
low back pain, the cortex may show some “reorganization.” This means that
the cortical maps in which the whole body is somatotopically represented
may show spatial changes. After nerve lesions, the extent of such changes
was found to be correlated to pain severity. More often another phenomenon Pain is an important component of suffering from disease; therefore, treatment
is reported in chronic pain states, a so-called “brain atrophy.” Often patients of pain is an absolute necessity in clinical medicine. In a wider sense, pain
with chronic pain show in functional MRI (fMRI) studies a “thinning” of is also an important element of the so-called illness response. The symptoms
the cortical areas involved in pain processing. The structural basis for this of diseases result not only from the local disease processes but also from
thinning is unknown. Interestingly, if it is possible to treat the pain successfully mechanisms in the central neurons system that respond to challenges such
(e.g., in OA by arthroplasty), the cortical atrophy is reversed. It is an open as inflammation, infection, and injury in a coordinated fashion. In fact,
question, therefore, whether local brain atrophies are a cause or a consequence plenty of evidence has shown autonomic–endocrine–immune interactions
of chronic pain. Nevertheless, brain imaging shows that pain in the to be related to acute and chronic pain. In this respect, pain therapy may
musculoskeletal system has substantial consequences at all levels of the not only just attack a clinical symptom but also modify the further course
nervous system.2 of disease.

1. Schaible H-G, Richter F. Pathophysiology of pain. 7. Schaible H-G, Richter F, Ebersberger A, et al. Joint pain. 14. Vogt BA. Pain and emotion: interactions in subregions of
Langenbecks Arch Surg. 2004;389:237-243. Exp Brain Res. 2009;196:153-162. the cingulate gyrus. Nat Rev Neurosci. 2005;6:533-544.
2. Bushnell MC, Ceko M, Low LA. Cognitive and 8. Basbaum AI, Bautista DM, Scherrer G, et al. Cellular 15. Schaible H-G. Mechanisms of chronic pain in
emotional control of pain and its disruption in and molecular mechanisms of pain. Cell. 2009;139: osteoarthritis. Curr Rheumatol Rep. 2012;14:549-556.
chronic pain. Nature Rev Neurosci. 2013;14:502- 267-284. 16. Sandkühler J. Learning and memory in pain pathways.
511. 9. Waxman SG, Zamponi GW. Regulating excitability of Pain. 2000;88:113-118.
3. Campbell JN, Meyer RA. Neuropathic pain: from the peripheral afferents:emerging ion channel targets. Nature 17. Vanegas H, Schaible H-G. Prostaglandins and
nociceptor to the patient. In: Merskey H, Loeser JD, Neurosci. 2014;17:153-163. cyclooxygenases in the spinal cord. Progr Neurobiol.
Dubner R, eds. The paths of pain 1975-2005. Seattle: IASP 10. Schaible H-G, Ebersberger A, Natura G. Update on 2001;64:327-363.
Press; 2005:229-242. peripheral mechanisms of pain: beyond prostaglandins 18. Telleria-Diaz A, Schmidt M, Kreusch S, et al. Spinal
4. Chapman CR, Gavrin J. Suffering: the contributions and cytokines. Arthritis Res Ther. 2011;13:210. antinociceptive effects of cyclooxygenase inhibition
of persistent pain. Lancet. 1999;353:2233-2237. 11. Schaible H-G. Nociceptive neurons detect cytokines in during inflammation: involvement of prostaglandins and
5. Willis WD, Coggeshall RE. Sensory mechanisms of the arthritis. Arthritis Res Ther. 2014;16:470. endocannabinoids. Pain. 2010;148:26-35.
spinal cord. 3rd ed. New York: Kluwer Academic/Plenum 12. Vanegas H, Schaible H-G. Descending control of 19. McMahon SB, Malcangio M. Current challenges in
Publishers; 2004. persistent pain: inhibitory or facilitatory? Brain Res glia-pain biology. Neuron. 2009;64:46-54.
6. Schaible H-G, Del Rosso A, Matucci-Cerinic M. Rev. 2004;46:295-309. 20. Neugebauer V, Li W. Processing of nociceptive mechanical
Neurogenic aspects of inflammation. Rheum Dis Clin 13. Ossipov MH, Dussor GO, Porreca F. Central modulation and thermal information in central amygdala neurons
North Am. 2005;31:77-101. of pain. J Clin Invest. 2010;120:3779-3787. with knee-joint input. J Neurophysiol. 2002;87:103-112.

B. Immunology and Inflammation

Principles of innate immunity

Tsuneyasu Kaisho • Shizuo Akira 9  


■ Innate immunity senses and responds to a variety of microorganism- or host-derived
molecules. Signaling sensors are expressed on the cell surface, in the endosome, or in
■ Innate immune sensors are functionally categorized into three classes: signaling,
the cytosol, where they recognize their corresponding ligands. On recognition,
internalizing, and soluble sensors.
these sensors can stimulate signaling pathways leading to activation of various
signaling molecules, such as nuclear factor κB (NF-κB) or mitogen-activated
■ Signaling sensors, such as Toll-like receptors, trigger signaling pathways for
protein kinases (MAPKs). This then leads to expression of a variety of immune
activating immune response genes.
response genes, including inflammatory cytokines or costimulatory molecules
■ Internalizing sensors incorporate microorganisms and degrade or process them for
such as CD40 and CD86.
presentation to T cells.
■ Soluble sensors opsonize microorganisms, thus marking them for internalization, and Toll-like receptors
activate protease cascades. The Toll-like receptor (TLR) family is a group of typical signaling sensors.2,3
■ Innate immunity induces inflammation, triggers adaptive immune responses, and TLRs are expressed on the plasma membrane or in the endosome and play
provokes antiviral immunity. major roles in activating antigen-presenting cells (APCs). TLRs are type I
■ Innate immune sensors are involved in the pathogenesis of not only infectious and transmembrane proteins. Their extracellular domain includes a repetitive
autoimmune diseases but also various inflammatory conditions in metabolic structure rich in leucine residues, called leucine-rich repeat (LRR), that is
diseases. involved in the recognition of a variety of TLR ligands. The intracellular
region contains a common structure in TLR and interleukin-1 receptor (IL-1R)
family members called Toll/IL-1R homologous (TIR) domain that is essential
for signal transduction through TLRs.
Toll-like receptors can recognize a variety of MAMPs. TLR ligands
INTRODUCTION can be categorized as lipid, protein, and nucleic acid components (Fig.
9.2). Phylogenetically related TLRs can recognize similar types of ligands.
Host defense consists of two types of immunity: innate and adaptive immunity All TLR ligands are potent immune adjuvants that can trigger a vigorous
(Table 9.1). Adaptive immunity is found only in jawed vertebrates and is immune response. The most widely investigated TLR ligand is lipopoly-
mediated by B and T lymphocytes. These cells possess the RAG genes that saccharide (LPS).4 LPS is found in the outer cell walls of gram-negative
mediate a somatic recombination system and thereby create a large repertoire bacteria and is recognized by TLR4. LPS is first bound to a soluble factor,
of antigen receptors. Antigen receptors with certain specificity are clonally LPS-binding protein, in the serum and transferred to target cells such as
expressed, and owing to allelic exclusion, each lymphocyte bears only one macrophages. Macrophages express a phosphatidylinositol-anchored cell
antigen receptor. Thus, groups of lymphocytes are highly heterogeneous in surface molecule, CD14, which can capture and retain LPS. LPS then
terms of antigen specificity. During the immune response, receptors with activates TLR4. A small secreted molecule, myeloid differentiation factor
high affinity for an invading antigen are clonally selected, and lymphocytes 2 (MD-2), is associated with TLR4 and critically involved in forming an
bearing such receptors later stay on as memory cells. LPS-recognizing complex.5
Although this defense system is highly beneficial for the host, establishing Other lipid-containing components of cell walls of a variety of micro-
adaptive immunity and mounting an adaptive immune response take time. organisms are recognized by TLR2 and related TLRs, such as TLR1 and
Especially early in the course of an infection, an immediate response is TLR6. Heterodimerization is critical for TLR2-mediated recognition. For
required for efficient host defense. This immediate response is mediated by example, TLR2 can recognize mycoplasmal macrophage-activating lipopeptide
innate immunity. Innate immunity is an evolutionarily ancient system that 2 (MALP-2) when associated with TLR6. Meanwhile, a TLR2–TLR1 het­
is found in all multicellular organisms, including plants and insects. In erodimer is involved in recognizing bacterial lipopeptides. MALP-2 and
mammals, innate immunity depends on macrophages or dendritic cells (DCs), bacterial lipopeptides carry a diacylated and triacylated cysteine residue at
which do not possess rearranged receptors. Insects can develop resistance their N-terminals, respectively, and this subtle difference is discriminated by
to infection, which indicates that innate immunity is sufficient to eradicate TLR2-containing heterodimers.
pathogens effectively. In contrast to adaptive immunity, innate immunity is TLR5 is involved in recognizing a protein, flagellin, that is a component
mediated by a group of germline-encoded receptors with a fairly limited of the bacterial flagella used for propulsion in a liquid medium. Flagellin can
repertoire of antigen receptors. These innate immune sensors can recognize elicit mucosal immune responses by acting on epithelial cells or macrophages.
various molecular structures derived from microorganisms, which can be Although flagellin is a protein, its amino acid structure is highly conserved,
referred to as pathogen- or microorganism-associated molecular patterns.1 They which suggests that it is a target for innate immune recognition.
are called microorganism-associated molecular patterns (MAMPs) hereafter Nucleic acids are equally important MAMPs recognized by TLRs. Bacterial
because such structures are found irrespective of pathogenicity. Only one DNA has long been known to be a strong immune adjuvant.6-8 This adjuvant
antigen receptor generated after gene rearrangement is expressed in each activity depends on an unmethylated CpG (cytosine phosphate guanine)
adaptive immune cell, but a group of innate immune sensors is expressed motif. The CpG motif is more abundant in bacterial DNA than in mammalian
in each innate immune cell. Therefore, innate immune cells are simultaneously DNA. Furthermore, mammalian CpG DNA is often methylated. Therefore,
or sequentially activated through a variety of sensors. Importantly, innate unmethylated CpG DNA can be regarded as nonself and is recognized by
immune sensors also recognize host-derived endogenous molecules, called TLR9. Viral DNA is also rich in the CpG motif, and DNA virus infection
danger or damage-associated molecular patterns (DAMPs), and are involved does indeed trigger TLR9 signaling.
in a variety of noninfectious inflammatory conditions. This chapter discusses Several small synthetic molecules, including imidazoquinoline derivatives
the innate immune system as a sensor system for activating immune or and several anticancer drugs, have long been known for their antiviral activity.
inflammation responses. This activity depends on TLR7. Single-stranded RNA (ssRNA) from influenza
or human immunodeficiency virus is also a ligand for TLR7 and its close
relative, TLR8. This interaction is critical for sensing RNA virus infection.
SENSING BY INNATE IMMUNITY RNA virus infection also induces the production of double-stranded RNA
Innate immune sensors can be functionally divided into three categories: (dsRNA) in infected cells, and these dsRNAs can act as immune adjuvants
signaling, internalizing, and soluble sensors (Fig. 9.1). and are recognized by TLR3.

64 SECTION 1  Scientific Basis of Rheumatic Disease

Table 9.1 
Innate and adaptive immunity
Innate immunity Adaptive immunity LPS Protein ligand
Lipid ligand
Evolution Ancient Jawed vertebrates or higher
Cells Macrophages T and B lymphocytes Flagellin
Peptidoglycan TLR4
Dendritic cells Bacterial lipopeptide
(Antigen-presenting cells) TLR5
Receptors Encoded in germline Created by rearrangement TLR2 dsRNA
Limited repertoire Large repertoire
Nonclonal Clonal expression
Memory No Yes TLR3
Ligands Molecular structures conserved in Fine structures (peptides) TLR10
a group of microorganisms N
(MAMPs; e.g., LPS, CpG DNA), N
host-derived damaged TLR1
Mycoplasma TLR8
molecules (DAMPs) lipopeptide Imidazoquinolines
Responses Immediate Slow TLR9 ssRNA
CpG, Cytosine phosphate guanine; DAMP, danger/damage-associated molecular pattern; LPS,
lipopolysaccharide; MAMP, microorganism-associated molecular pattern. Nucleic acid ligand

FIG. 9.2  The phylogenetic relationships of the human Toll-like receptors (TLRs) are
THREE CATEGORIES OF SENSORS shown by the dotted lines based on an analysis of their amino acid structures.
Branch lengths are proportional to evolutionary distances. CpG, Cytosine phosphate
Microorganisms guanine; dsRNA, double-stranded RNA; LPS, lipopolysaccharide; ssRNA, single-
stranded RNA.
Ag presentation
Soluble sensor to T cells
sensor from microorganisms including viruses. The representative members are
Signaling sensor Protease retinoic acid–inducible gene I (RIG-I)–like receptors (RLRs) such as RIG-I
cascade and melanoma differentiation–associated gene 5 (MDA5).9,10 RIG-I and MDA5
TLR carry tandem caspase recruitment domains (CARDs) and an RNA helicase
domain at the N-terminal and C-terminal, respectively, and the RNA helicase
Cell membrane domain is involved in recognizing the ligands. Both RIG-I and MDA5 can
recognize virally derived RNAs, but they sense distinct molecular structures
and play differential roles in viral infection.10 For example, RIG-I is essential
Cytosolic for recognizing various ssRNA viruses, including paramyxoviruses, influen-
sensor zavirus A, and Japanese encephalitis virus, and MDA5 is critical for sensing
picornaviruses. RIG-I can recognize 5′-triphosphate ssRNA, which is a typical
structure found in viral, nonself RNA. Host-derived self-RNA, on the other
Gene expression Degradation MHC hand, has a cap structure to mask 5′-triphosphate RNA. Thus RIG-I can
(e.g., cytokines) distinguish viral RNA from self-RNA.
Double-stranded DNA (dsDNA) is also sensed in the cytoplasm of mam-
malian cells. This sensing heavily depends on cyclic GMP-AMP (cGAMP)
FIG. 9.1  Signaling sensors trigger signal transduction pathways upon recognition.
synthase (cGAS), which is a cytoplasmic nucleotidyltransferase that belongs
Internalizing sensors directly incorporate microbes into the cell. Soluble sensors
to the class of template-independent polymerases. When dsDNA is sensed
attach to the microbes and support the function of internalizing sensors. Soluble
by cGAS, a second messenger, cGAMP, is generated from ATP and GTP and
sensors also activate protease cascades. Ag, Antigen; MHC, major histocompatibility
leads to expression of type I interferon (IFN) and antiviral genes. cGAS plays
complex; TLR, Toll-like receptor.
key roles in infection with various kinds of DNA viruses or retroviruses.11
Nucleotide-binding oligomerization domain (NOD)–like receptors (NLRs)
Toll-like receptors can also recognize DAMPs. DAMPs are host-derived also are cytosolic sensors (Fig. 9.3).12 NLRs consist of more than 20 members
endogenous molecules, hidden within cells and ignored by the immune in humans and mice. They carry a CARD, a pyrin domain, or a baculoviral
system in the steady state. After tissue damage or cell death, they are released inhibitor of apoptosis repeat (BIR) domain as a protein interaction domain
and detected by the innate immune sensors. Nucleic acids are typical DAMPs at the N-terminal, which is followed by a NACHT nucleotide-binding domain
and are recognized by TLR3, TLR7, TLR8, and TLR9. These TLRs fail to (NBD) and an LRR domain. NOD1 and NOD2 are representative NLR members
distinguish between microorganism- and host-derived nucleic acids (see involved in recognizing bacterial peptidoglycan-derived molecules.13 NLR
later). RNAs damaged by ultraviolet radiation or mitochondrial DNA are family apoptosis inhibitory protein 5 (NAIP5) can sense flagellin derived
sensed by TLR3 and TLR9, respectively. TLR4 also recognizes a variety of from intracellular bacteria such as Salmonella or Legionella spp. An NAIP5-
DAMPs. These include not only lipids such as saturated fatty acids or oxidized related molecule, NAIP2, can detect conserved rod proteins of the type III
low-density lipoprotein (LDL) cholesterol but also proteins such as heat secretion system from Salmonella or Burkholderia spp. Notably, NLR family
shock proteins or S100 family proteins. pyrin domain–containing 3 (NLRP3) is required for responses against a
Importantly, whereas TLR2 and TLR4 are expressed mainly on the plasma variety of MAMPs and DAMPs. These include bacterial products such as
membrane, nucleic acid–recognizing TLRs are expressed in the endoplasmic RNA or toxin, host-derived nucleic acid metabolites such as adenosine
reticulum. Upon activation, chaperone molecules such as UNC93B1 escort 5′-triphosphate, uric acids, and oxidized mitochondrial DNA. Furthermore,
nucleic acid–recognizing TLRs to the endosome, where nucleic acids are environmental substances such as silica crystals or asbestos and a widely
released from virus or virally infected cells. Thus, lipid and nucleic acid TLR used chemical immune adjuvant, alum, can also activate NLRP3-mediated
ligands are recognized in distinct cellular compartments. signaling.
Members of the pyrin and hematopoietic interferon-inducible nuclear
Cytosolic signaling sensors (HIN) domain–containing protein (PYHIN) family also function as cytosolic
Signaling sensors also exist in the cytosol. DExD/H-box helicase family sensors. This family carries an N-terminal pyrin domain and two DNA-binding
members, which are characterized by motifs that include the DExD/H amino HIN domains. Both absent in melanoma 2 (AIM2) and interferon-γ–inducible
acid sequence (D, E, x, and H represent aspartic acid, glutamic acid, and protein 16 (IFI16) belong to this family and are involved in detecting dsDNA.
amino acid and histidine, respectively) and are highly conserved throughout It should also be noted that some MAMPs are recognized by both trans-
species, function as cytosolic sensors for a variety of nucleic acids, mainly membrane and cytosolic signaling sensors. For example, peptidoglycans are

CHAPTER 9  Principles of innate immunity 65


Cell membrane
DAMPs Fungi Gram+ bacteria
Salmonella Spaetzle
Peptidoglycan Bacterial Uric acid Oxidized
N mitochondrial
Muramyl- N
O DNA Environment-
dipeptide O N N
TTSS rod Protease
Meso- derived cascade
diaminopimelic Toxin P P P
H Silica Flagellin proteins
acid H Asbestos Peptidoglycan
Chemical NAIP2 Fungi
NOD1 NLRP3 NLRC4 recognition protein
Alum recognition
ASC NAIP5 Caspase - 1 protein?
NOD2 Caspase - 1 NLRC4 Anthrax
lethal toxin
Caspase - 1 Toll TLR
Pro - IL - 1 NLRP1 Cell membrane
CARD Pro - IL - 18 ASC
Pyrin domain
BIR domain
Caspase -1
NBD AIM2 dsDNA Nucleus
HIN domain IL - 18 Caspase - 1 Antimicrobial peptide Cytokine

FIG. 9.3  Nucleotide-binding oligomerization domain (NOD)–like receptors (NLRs) FIG. 9.4  Whereas mammalian Toll-like receptors (TLRs) directly recognize microbial
are cytosolic signaling sensors and recognize a variety of microorganism-associated components, Drosophila Toll detects a host-derived molecule. In Drosophila spp.,
and danger/damage-associated molecular patterns. Importantly, NLRs can activate microbes are discriminated by soluble factors in the hemolymph, which leads to the
inflammasome and induce production of active interleukin-1β (IL-1β) and IL-18. activation of protease cascades.
AIM2, Absent in melanoma 2; ASC, apoptosis-associated speck-like protein containing
a caspase recruitment domain; ATP, adenosine 5′-triphosphate; BIR domain, bacu-
lovirus inhibitor of apoptosis repeat domain; CARD, caspase recruitment domain; SOLUBLE SENSORS
DAMP, danger/damage-associated molecular pattern; dsDNA, double-stranded DNA;
Soluble sensors are produced by macrophages or hepatocytes and bind to
HIN domain, hematopoietic, interferon-inducible nuclear protein domain; IL, interleukin;
the cell wall of microorganisms, thus designating them as targets for phago-
LRR, leucine-rich repeat; NAIP2/5, nucleotide-binding oligomerization domain-like
cytosis (see Fig. 9.1). This coating process is termed opsonization, and the
receptor family of apoptosis inhibitory proteins 2/5; NBD, nucleotide binding domain;
coating substances are called opsonins. The serum levels of certain soluble
NLRC4, nucleotide-binding oligomerization domain-like receptor family caspase
sensors, such as C-reactive protein and serum amyloid P component, increase
recruitment domain-containing 4; NLRP1/3, nucleotide-binding oligomerization
in response to inflammatory cytokines; therefore, these sensors are also
domain-like receptor family, pyrin domain–containing 1/3; NOD1/2, nucleotide-binding
referred to as acute-phase proteins.15 Soluble sensors are linked to the comple-
oligomerization domain 1/2; TTSS, type III secretion system.
ment system and can activate protein cascades that are capable of eradicating
pathogenic microbes.

sensed by TLR2 and NOD1/2. Flagellin is sensed by TLR5 and NAIP5. Complement system
Furthermore, LPS is sensed by TLR4 and caspases (see later). The complement system consists of more than 35 soluble or membrane
proteins (see Chapter 15, The Complement System).16 Microbial infection
activates three distinct pathways—the classical, alternative, and lectin
INTERNALIZING SENSORS pathways—for complement activation. All pathways activate a common
Internalizing sensors are expressed on the surface of neutrophils, macrophages, pathway that involves cleavage of C3 into C3a and C3b. C3b attaches to the
or DCs. They bind to and internalize MAMPs or microorganisms, which are microbial surface, and microbes opsonized by C3b are then eliminated by
subsequently transported to lysosomal compartments (see Fig. 9.1). This phagocytosis or killed by membrane attack complex.
process is termed phagocytosis, and the group of sensors involved can thus
also be referred to as phagocytic sensors.14 After phagocytosis, organism-derived
proteins are degraded or processed for presentation to T cells. Internalizing PHYLOGENETIC COMPARISON OF THE
sensors mainly function to incorporate the organisms or their components,
but some internalizing sensors can also transduce the activation signals.
Toll-like receptors have been so named because their molecular structure is
Lectin similar to that of Drosophila Toll.17,18 Toll is a type I transmembrane protein
Lectins generically represent carbohydrate-recognizing proteins, and some carrying an LRR repeat and a TIR domain. Toll-deficient flies display increased
lectins function as internalizing sensors. The mannose receptor is a type I susceptibility to fungal infection, which indicates that Toll is essential for
transmembrane protein that can bind terminal mannose and fucose residues antifungal immunity in insects. Furthermore, both Toll and TLRs can activate
of glycoproteins or glycolipids, which are found on microbial cell walls. The signaling pathways leading to the release of antimicrobial substances.
macrophage scavenger receptor SR-A is a type II transmembrane protein. The Drosophila Toll system, however, is distinct from the mammalian
Although originally defined by its ability to bind and mediate endocytosis TLR system (Fig. 9.4). Toll does not directly recognize fungi-derived products.
of oxidized or acetylated LDL, it can also bind a variety of microorganisms. The pathogens are recognized by certain still-unidentified soluble factors.
Macrophage receptor with collagenous structure (MARCO) is structurally In fungal infection, protease cascades are activated to induce cleavage of a
and functionally similar to SR-A and can bind to gram-positive and gram- secreted protein, Spaetzle. Processed Spaetzle then binds to Toll. Thus, Toll
negative bacteria. recognizes a host-derived product.
Dectin-1 is a C-type lectin and is involved in recognizing β-1,3–linked Toll is also involved in immunity against gram-positive bacteria.19 A soluble
or β-1,6–linked glucans found in fungal or other microbial cell walls. Dectin-1 molecule, peptidoglycan recognition protein SA (PGRP-SA), recognizes
carries an immunoreceptor tyrosine-based activation motif (ITAM) in the peptidoglycans from bacteria, which leads to activation of protease cascades
intracytoplasmic region and also functions as a signaling sensor. Dectin-1 that also cleave Spaetzle. PGRP-SA–deficient flies are susceptible to infection
is colocalized with TLR2 in the phagosome and activates macrophages in with gram-positive bacteria but retain immune responses against fungi,
synergy with TLR2. Macrophage-inducible C-type lectin (Mincle) also carries whereas Toll-deficient flies succumb to infection with both organisms. This
an ITAM and can induce macrophage activation by sensing a mycobacterial indicates that the discrimination of fungi and bacteria happens upstream of
glycolipid, trehalose-6,6′-dimycolate, also called cord factor. Mincle also detects Toll. Thus, flies detect pathogens in the hemolymph not on the plasma
a nuclear ribonucleoprotein that belongs to the DAMP group. Another membrane. In mammals, the complement system, rather than the TLR system,
ITAM-carrying C-type lectin, CLEC9A, also known as DNGR-1, senses exposed is similar to Toll when one considers recognition in the fluid and activation
actin filaments, which also are DAMPs. of protease cascades.

66 SECTION 1  Scientific Basis of Rheumatic Disease


INDUCTION OF INFLAMMATION Innate immunity responds to viral infection by producing type I IFNs. 21
Type I IFNs consist of more than 10 IFN-α proteins and a single IFN-β
Innate immunity detects microbial infection and induces inflammatory protein. However, all type I IFNs use IFN-α/β receptor (IFNAR), composed
reactions in local peripheral tissues such as the skin (Fig. 9.5). TLRs expressed of IFNAR1 and IFNAR2 subunits, as a common receptor. The signaling can
on macrophages or DCs play major roles in this process. TLR signaling can induce expression of a number of antiviral molecules including 2′-5′-oligoad-
induce robust production of proinflammatory cytokines, including IL-6 and enylate synthases or IFNs themselves. It can also upregulate expression of
tumor necrosis factor-α (TNF-α), chemokines, and adhesion molecules, and the major histocompatibility complex (MHC) and induce DC maturation.
thereby recruit or activate various inflammatory cells at the sites of infection. These effects contribute to antiviral immune responses.
Recruited and activated macrophages or neutrophils then ingest, through Type I IFN production is triggered by innate immune signaling sensors.
internalizing sensors, and subsequently kill invading pathogens by producing Cytosolic sensors such as RLRs are potent type I IFN–inducing sensors and
nitric oxide, reactive oxygen species, or defensins. NLR signaling can induce are expressed in a variety of cells, including macrophages and fibroblasts.
generation of proinflammatory cytokines such as IL-1β and IL-18 through Among TLRs, TLR7 and TLR9 can induce both IFN-α and IFN-β. The other
the activation of inflammasome (see later) and contributes to inflammatory TLRs, except TLR3 and TLR4, fail to induce type I IFNs. TLR3 and TLR4
reactions. Inflammation is a critical local response to resolve infection. can induce only IFN-β but not IFN-α. TLR4-induced IFN-β plays a critical
However, inflammation is a double-edged sword because excessive amounts role in shock induction and bacterial infection, but as a principle, in viral
of cytokines can be lethal for the host, as is the case in endotoxin shock. infection, TLRs that recognize nucleic acids are more critical than TLR4 in
production of type I IFNs.
Plasmacytoid dendritic cells (PDCs) are a DC subset and also are known
ACTIVATION OF ADAPTIVE IMMUNITY ON INFECTION as IFN-α–producing cells.22 PDCs look like plasma cells and exhibit poor
Locally activated APCs mature to express a distinct set of chemokine receptors. antigen-presenting activity owing to low levels of expression of MHC class
Then they are transported to the lymph nodes, where they interact with and II and costimulatory molecules. PDCs express TLR7 and TLR9 exclusively
activate T cells (see Fig. 9.5). Thus, an innate immune response subsequently among TLRs and are involved in antiviral immunity by secreting copious
leads to the activation and establishment of adaptive immunity. Clonal T-cell amounts of type I IFNs, especially IFN-α, in a TLR7/9-dependent manner.
expansion requires not only T-cell receptor–mediated but also costimulatory Cross-presentation is also important in antiviral immunity.23 Cross-
molecule–mediated signaling (Fig. 9.6). Internalizing sensors facilitate antigen presentation is defined as exogenous antigen presentation through MHC class
presentation, and signaling sensors such as TLRs can enhance expression I to generate CD8+ T-cell responses (Fig. 9.7). Virally infected cells are
of costimulatory molecules, including CD80 and CD86. Thus, innate immune ingested by APCs. Antigens are processed and transported to MHC class
sensors directly contribute to the activation of adaptive immunity. II–containing endolysosomes, where antigens associate with MHC class II
A CD4+ T-cell differentiates into a type 1 helper T (Th1), type 17 helper molecules. The antigen–MHC class II complex then becomes competent for
T (Th17), or type 2 helper T (Th2) cells.20 Th1 cells produce interferon-γ presentation to CD4+ T cells. Exogenous antigens can also be subjected to
(IFN-γ) and mediate antiviral or antibacterial immunity. Th17 cells produce the association with MHC class I. Antigens can be transported to the cytosol,
IL-17 and are involved not only in antibacterial immunity but also in various processed by proteasomes, and then incorporated into the endoplasmic
inflammatory conditions such as autoimmune disorders. Th2 cells secrete reticulum in a transporter associated with antigen processing (TAP)–dependent
IL-4 or IL-13 and are involved in immunity against helminths, but excessive manner. Then antigens bind to MHC class I molecules and are presented to
activation of these cells may cause allergic reaction. APCs are critically involved CD8+ T cells. Alternatively, antigens can be transferred directly to the MHC
in regulating Th cell differentiation. This activity depends on tissue origin, class I–containing compartment, not via the cytosol. The cross-presentation
cell subsets, or the maturation stage of the APCs. But the nature of the system ensures that virally infected cells, which do not always possess the
stimuli activating the APCs is most important. Most TLR ligands can activate ability to present antigen by themselves, can be targets for cytotoxicity.
APCs to produce Th1- or Th17-inducing cytokines such as IL-12, IL-18, Cross-presentation also should be critical in cancer surveillance because
IL-6, or IL-23. Although Th2-inducing sensors are not well characterized most cancer cells show poor antigen presentation ability. Several immune
yet, certain helminths or their products can enhance the ability of APCs to sensors such as TLRs can facilitate the cross-presentation, and one DC subset
support Th2 cell differentiation. defined by the presence of CD8α, CD103, or the chemokine receptor XCR1
features a high ability to phagocytose dying cells and cross-present the



Lymph node
Lymph duct
Skin DC
T cell Th2 IL-4

Cytokines Class II
T cell