Molecular Mechanisms of Spondyloarthropathies
ADVANCES IN EXPERIMENTAL MEDICINE AND BIOLOGY Editorial Board: NATHAN BACK, State University of New York at Buffalo lRUN R. COHEN, The Weizmann Institute ofScienc e ABEL LAJTHA, N.S. Klin e Institute for Psychiatric Research JOHN D . LAMBRIS, University ofPennsylvania RODOLFO PAOLETTI, University ofMilan Recent Volumes in this Series Volume 641 CELLULAR OSCILLATORY MECHANISMS Edited by Miguel Maroto and Nick Monk Volume 642 THE SARCOMERE AND SKEL ETAL MUSCLE DISEASE Edited by Nigel G. Laing Volume 643 TAURINE? Edited by Junich i Azuma Volume 644 TROPOMYOSIN Edited by Peter Gunn ing Volume 645 OXYGEN TRANSPORT TO TISSUE XXX Edited by Per Liss, Peter Hansell, Duane F. Bruley, and David K. Harrison Volume 646 EARLY NUTRITION PROGRAMMING AND HEALTH OUTCOMES IN LATER LIFE Edited by Berthold Koletzko, Tamas Desci, Denes Molnar, and Anne De la Hunty Volume64? THERAPEUTIC TARGETS OF THE TNF SUPERFAMILY Edited by Iqbal Grewal Volume 648 ARTERIAL AND ALLIED CHEMORECEPTORS Edited by Constancio Gonzalez, Colin A . Nurse, and Chris Peers Volume 649 MOLECULAR MECHANISMS OF SPONDYLOARTHROPATHIES Edited by Carlos Lopez -Larrea and Roberto Diaz-Pcfia A Continuation Order Planis availableforthis series. A continuation order will bring deliveryof each new volume immediately upon publication. Volumes are billed only upon actual shipment. For further information please contact the publisher.
Molecular Mechanisms of Spondyloarthropathies Edited by
Carlos L6pez-Larrea, PhD Hospital Universitario Central de Asturias Oviedo, Spain Roberto Diaz-Pefia, BSc Hospital Universitario Central de Asturias Oviedo, Spain
Springer Science+Business Media, LLC Landes Bioscience
Springer Science+Business Media, LLC Landes Bioscience Copyright ©2009 Landes Bioscience and Springer Science+Business Media, LLC All rights reserved . No part ofthis book may be reproduced or transmitted in any form or by any means, electronic or mechan ical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher, with the exception ofany material supplied specifically for the purpose of being entered and executed on a computer system; for exclusive usc by the Purchaser of the work . Printed in the USA . Springer Science+Business Media, LLC, 233 Spring Street, New York, New York 10013, USA http ://www.springer.com Please address all inquir ies to the publishers : Landes Bioscience, 1002 West Avenue, Austin, Texas 7870 I, USA Phone: 512/637 6050; !'AX: 512/6376079 http ://www.Iandesbioscience.com The chapters in this book are available in the Madame Curie Bioscience Database. http ://www.landesbioscience.com/curie Molecular Mechanisms ofSpondyloarthropathies, edited by Carlos Lopez-Larrea and Roberto Diaz-Pcfia. Landes Bioscience / Springer Science+Business Media , LLC dual imprint / Springer series : Advances in Experimental Medicine and Biology
ISBN : 978-1-4419-0297-9
While the authors, editors and publisher believe that drug selection and dosage and the specifications and usage of equipment and devices, as set forth in this book, are in accord with current recommendations and practice at the time of publication, they make no warranty, expressed or implied, with respect to material described in this book . In view of the ongoing research, equipment development, changes in governmental regulations and the rapid accumulation of information relating to the biomed ical sciences, the reader is urged to carefully review and evaluate the information provided herein.
Library of Congress Cataloging-in-Publication Data Molecular mechanisms of spondyloarthropathies / edited by Carlos Lopez-Larrea, Roberto Diaz-Pefia. p. ; cm. -- (Advances in experimental medicine and biology ; v. 649) Includes bibliographical references and index. ISBN 978-1-4419-0297-9 I. Spondyloarthropathies--Molecular aspect s. I. Lopez-Larrea, Carlos, 1953- II. Diaz-Pena, Roberto , M.A. Ill. Series : Advances in experimental medicine and biology ; v. 649 . [DNLM: I . Spondyloarthropathies--genetics. WI AD559 v.649 2009 / WE 725 M718 2009] RC935 .S67M672009 616 .7'3--dc22 2009012986
PREFACE It is now over ten years since we edited the first edition of HLA-B27 in the Development ojSpondyloarthropathies (SpA) . It is obvious that over this period an enormous amount of information concerning SpA and HLA-B27 has accumulated, and this has been reflected in the knowledge of molecular mechanism ofthe spondyloarthropathies. Discussion for such a book took initial form at the outstanding 4th Gent Symposium on SpA in October 2006, but was not formally commissioned by Landes Bioscience until early 2007. Molecular Mechanisms ojSpondyloarthropathies aims to synthesize this growing knowledge and present all the current studies concerning the basic research ofSpA. Over the last decade, enormous progress has been made in the understanding ofthe molecular and cellular processes that lead to disease pathology. Recognition of the pathways involved in the pathogenic mechanisms of disease and the potential to target specific immune effector functions have opened the door to a wide range of innovative treatment opportunities for the SpA. Thus, we believe that the understanding of the specific molecular mechanisms involved would help to design highly specific drugs in the future. The contributors are among the best in their field and reflect the state of the art of research and current opinion of rheumatologists and immunologists. The first part covers the general clinical aspects of SpA . Established classification criteria for ankylosing spondylitis (AS) and SpA are less effective in early disease stages and thus contribute to delay in diagnosis. Major factors for improving the rate ofAS patients diagnosed early are HLA- B27 and imaging of the sacroiliac joints . The book begins with chapters on the available measures of major areas of clinical disease impact (disease activity, structural damage and functioning) in the spondyloarthropathies (Chapter I) and the role of imaging (MRI) in the evaluation and management of SpA (Chapter 2). MRI now permits the detection of patients at an earlier stage of their disease course with the potential for new insights into the pathogenesis of disease. These first chapters are followed by a section describing clinical similarities and differences with common enthesopathic disorders-e-spondylarthritis , chondrocalcinosis and diffuse idiopathic skeletal hyperostosis-which may help in the differential diagnosis (Chapter 3). v
vi
Preface
The second part focuses on the 'enthesis organ concept' and the osteoproliferation mechanisms in SpA (Chapter 4). The microdamage and propensity for bacterial involvment in the context of genetic factors such as HLA-B27 appears to lead to the characteristic inflammatory changes of AS. Understanding the "enthesis organ concept" helps to explain synovitis and osteitis in spondyloarthropathy. It is conceivable that disturbances in the sensing microorganisms at the mucosal interface or biomechanical stress at musculoskeletal sites is important for the initiation ofthe inflammation and the determination of predilection sites (Chapter 5). Bone loss is a common finding in the spondyloarthropathies. There is however increasing evidence to support a role for osteoclasts in bone erosions . Included in the book is an overview (Chapter 6) which contains recent literature related to the role of osteoclasts and the RANKL/OPG system in the various spondyloarthropathies. The inhibition ofRANKL seems to be a rational strategy to treat both localized and generalized bone loss in the inflammatory arthritides (Chapter 7). The interplay between inflammation and ankylosis is best illustrated in AS. Thus, prevention ofankylosis is a therapeutic aim in AS, which may either be achieved by early intervention or before bone proliferation has been established. Further understanding of the regulation of osteoproliferation in AS is thus warranted, and this will help to tailor the therapeutic interventions, which prevent structural remodeling in AS (Chapter 8). Many more biomarkers will need to be discovered and are now being evaluated (Chapter 9). Novel drugs specifically targeting biomarkers and their signaling pathways are expected to revolutionize the treatment of these pathologies. In recent years, clinical trials have proved that anti-TNFa is highly effective in the treatment of SpA . Thus, we believe that understanding of the specific molecular mechanisms involved could help to design highly specific drugs in the future (Chapter 10). The third part of the book focuses on the genetic and biochemical aspects of HLA-B27. The strong association of the human MHC class I allele HLA-B27 with the development ofAS is clear and has been known for over three decades. Besides the specificity of antigen presentation, which could provide a possible pathogenetic mechanism based on molecular mimicry, the peptide binding features ofHLA-B27 subtypes are important for their relationship with other biochemical properties. Yet it is far from clear how HLA-B27 is directly involved in AS. Several studies indicate that other genes may also be involved in SpA susceptibility. In Chapter 11 the genetic epidemiology ofAS and the gene-mapping studies performed to date have been reviewed. The successful identification ofARTS 1 and IL23R should give those involved in AS genetics research considerable encouragement in understanding the potential of this research. HLA-B27 represents a family of 38 closely related cell surface, proteins and the epidemiological data available have shown that some alleles (B*2706 and B*2709) are negatively associated with AS. Chapter 12 reviews the prevalence of HLA-B27 subtypes on large extended worldwide population. Molecular studies demonstrating the effects of the sequence variations on the peptide binding specificity of the molecule may yield clues to disease pathogenesis. The availability of HLA-B27 subtypes differentially associated with ankylosing
Preface
vii
spondylitis provides a unique tool for exploring the relationship between peptide specificity and pathogenetic potential. Subtype-specific polymorphic residues play a key role in determining whether an HLA-B27 subtype is AS-associated or not and opens the possibility for correlating the structural and functional characteristics of a given subtype with the disease association. The antigen peptide binding properties ofHLA-B27 subtypes are described in Chapters 13 and 14. The use of X-ray crystallography and various other biophysical techniques have revealed how several different peptides are accommodated within the binding groove of the molecules. Many studies have focused on defining the nature of subtype-bound repertoires. These aim to identify peptide features that may correlate with association to disease and to find constitutive self-ligands with sequence homology to microbial epitopes . These studies were undertaken on the assumption that molecular mimicry between self and foreign ligands of HLA-B27 might trigger autoimmunity. More recently, theories based on several non-antigen presentation properties have emerged. In this section recent data to determine whether any biochemical features of HLA-B27 can supply clues to its enigmatic role in AS have been reviewed, and comments are also made on future potential directions of biochemical research into HLA-B27. The study of the biochemistry of HLA-B27 has revealed a number of unique observation s, many of which center around the efficiency with which it undergoes folding within the endoplasmic reticulum (ER), leading to the presence of unfolded heavy chains which are predisposed to form dimmers B272 structures (Chapter 15). The observation that the HLA-B27 heavy chain has a tendency to misfold has raised the possibility that associated diseases may belong to a rapidly expanding category ofprotein misfolding disorders. In transgenic rats, overexpression of 132m results in a drastic decrease of the misfolded heavy chain in the ER, demonstrat ing that promoting the formation of HLA-B27/peptide complexes is sufficient to curb misfolding . Enhanced accumulation of misfolded heavy chains during the induct ion of class I expression by cytokines, can cause ER stress resulting in activation of the unfolded protein response (UPR). Adaptation to chronic ER stress is associated with modification of an NFKB-dependent pathway. Effects of UPR activation on cytokine production are beginning to emerge and may provide important missing links between HLA-B27 misfolding and spondyloarthritis. Chapter 16 reviews the current knowledge about HLA-B27 misfolding in human cells and in the transgenic rat model of spondyloarthritis-like disease and the consequences of HLA-B27induced UPR activation. These mechanisms could be relevant to the relationship between HLA-B27 and host-pathogen interaction. One form of SpA is reactive arthritis (ReA), which develops as a complication after certain bacterial infections. Since most of the patients suffering from ReA are HLA-B27 positive, it has been proposed that HLA-B27 may modulate the interaction between ReA-triggering bacteria and the host cell. Homodimer B272 formation and misfolding ofHLA-B27 heavy chain in the endoplasmic reticulum (ER) may trigger ER/stress signaling pathways in the host cell, which in turn may modulate cell signaling in favor of ReA-triggering. Chapter
viii
Preface
17 summarizes the observations of HLA-B27 modulating the interaction between ReA-triggering bacteria and the host cell and discusses the potential mechanisms behind the interaction. Transgenic models have now been developed that will hopefully allow a clearer view of the function of B27 in the pathogenesis of the disease. Chapter 18 evaluates the different models and presents an overview of the most prominent. Areas of investigation to which these models contribute include the role HLA-B27, processes of spinal and peripheral joint inflammation and calcification, immune responses to candidate antigens, and the role ofTNF. Finally, the last part ofthe book contains a section on the immunological aspects of SpA. Most immune responses, both innate and adaptive, involve activation of multiple cell types, as results of which may up-regulate the mechanism of cell activation in patients. SpA inflammation is most likely the result of defects at different levels in the immune system. The classical role ofHLA-B27 is to present peptides from intracellular pathogens for recognition by the T-cell receptor (TCR) of CD8+ T cells . Different studies have focused on a model of cross-reactivity based on self, viral and bacterial peptides that share sequence and structural similarities and trigger T-cell responses in a disease-susceptible context thus giving strong support to this hypothesis. According to this theory, traditionally reported as the "arthritogenic" peptide hypothesis, cytotoxic CD8+ T cell responses against cross-reactive self and foreign antigens presented by the HLA-B27 with AS could be the primary pathogenic event in the autoimmune process . The possible implications of T-cell crossreact ivity between self and viral epitopes (peptides) in AS pathogenesis is discussed in Chapter 19. Chapter 20 explores different aspects of dendritic cell (DC) function in relation to SpA. The response of DC is essential to the initiation and modulation of most antigen-specific inmunoresponse as well as being involved in innate immune responses. DC are particularly well equipped to detect these bacterial products using the immune innate Toll like receptors (TLR) and other bacterial sensors The response of DC to TLR ligands can be strongly influenced by the physiologic state of the endoplasmic reticulum of the cell. Also, NK recognition of non-canonical conformations of HLA-B27, in the form ofB27 2 homodimers by members ofthe killer cell immunoglobulin-like (KlR) and leukocyte immunoglobulin-like (LILR) receptor families, has been reported as a potential contributor to AS development (Chapter 21). Upregulation of B272 in spondyloarthritis and the differential interaction of ~2m-associated HLA-827 and B27 2 with NK immune receptors could be involved in the pathogenesis of 827associated spondyloarthritis (AS). In addition, several association studies based on a model in which KlRs synergize with HLAs have also been reported. This interaction may generate compound genotypes, which provide different levels of activation or inhibition. Furthermore, some of these have been associated with certain SpA, such as ankylosing spondylitis (AS) and psoriatic arthritis (PsA) (reviewed in Chapter 23). Finally, the contribution of other innate immune receptors, such as pathogenassociated molecular patterns (PAMP) and TLR in SpA, constitute an important group for attention. Their potential role in SpA is discussed in Chapter 22.
Preface
ix
I am extremely grateful to all authors for their contribution to this textbook. My hope is that such volume as Molecular Mechanisms ofSpondyloarthropies will help to provide a better knowledge of these complex diseases.
Carlos Lopez-Larrea
ABOUT THE EDITORS...
CARLOS LOPEZ-LARREA is a Professor of Immunology (UCM, Madrid, Spain) and currently Head ofthe Histocompatiblity Unit at the Hospital Universitario Central de Asturias, Oviedo, Spain. He is a world expert on spondyloarthropathies (SpA), in particular genetic (MHC) and immunologic factors that influence the development of SpA. His primary research is focused on understanding the role of HLA-B27 polymorphism in the development of ankylosing spondylitis (AS), and describing relevant associations of B27 subtypes with AS in large extended worldwide populations . In the last decade, the availability of HLA-B27 subtypes differentially associated with the disease has provided a unique tool for exploring the relationship between peptide specificity and pathogenetic potential. The main research interests of his group currently include the role of innate immunity (NKs) in SpA. He is a member ofseveral national and international scientific organizations. It is now over ten years since the first edition of HLA-B27 in the Deve/opement of Spondy/oarthropathies (SpA) edited by C. Lopez-Larrea was published. This current book aims to synthesize the growing knowledge in the field and present all recent studies concerning basic research in SpA.
xi
ABOUT THE EDITORS...
ROBERTO OiAZ-PENA is a predoctoral investigator in the Histocompatibility Unit of the Oepartment of Immunology, Hospital Universitario Central de Asturias. He graduated in biochemistry, at the University of Oviedo, Spain. His main research interests include the genetics ofspondyloarthropathies, especially ankylosing spondylitis, and the role of inhibitory and activating KlR receptors in this disease.
xii
PARTICIPANTS Allen P. Anandarajah Clinical Immunology Research Center Rochester, New York USA Antony N. Antoniou Bute Medical School University of St. Andrews Fife, Scotland
Bruno Filipe Bettencourt Hospital de Santo Espirito de Angra do Heroismo Azores and Institute for Molecular and Cell Biology Porto Portugal
UK Jacome Brogues Armas Immunogenetic Service Hospital de Santo Espirito de Angra do Heroismo Azores Portugal Dominique Baeten Clinical Immunology and Rheumatology Academic Medical Center University of Amsterdam Amsterdam The Netherlands Michael Benjamin School of Biosciences Cardiff University Wales
Miguel Angel B1anco-Gelaz Department of Immunology Hospital Universitario Central de Asturias Oviedo Spain Paul Bowness MRC Human Immunology Unit Weatherall Institute of Molecular Medicine John Radcliffe Hospital Headington , Oxford UK
Jiirgen Braun Rheumazentrum Ruhrgebiet Heme Germany
UK
xiii
Participants
xiv
Matthew A. Brown Diamantina Institute of Cancer Immunology and Metabolic Medicine Princess Alexandra Hospital Woolloongabba Australia and University of Oxford Institute of Musculoskeletal Sciences Botnar Research Center Nuffield Orthopaedic Center Headington, Oxford UK Chun-Hsiung Chen Institute of Clinical Medicine National Yang-Ming University and Veterans General Hospital Taipei Taiwan Chung-Tei Chou Division ofAllergy, Immunology and Rheumatology Veterans General Hospital Taipei Taiwan Robert A. Colbert Division of Rheumatology Cincinnati Children's Hospital Medical Center Cincinnati, Ohio USA Ana Rita Couto Hospital de Santo Espirito de Angra do Heroismo Azores and Institute for Molecular and Cell Biology Porto Portugal Monica L. DeLay Division of Rheumatology Cincinnati Children's Hospital Medical Center Cincinnati, Ohio USA
Roberto Diaz-Pefia Hospital Universitario Central de Asturias Oviedo Spain Maria T. Fiorillo Department of Cell Biology and Development University Sapienza Roma Italy J.S . Hill Gaston Department of Rheumatology University of Cambridge Addenbrooke's Hospital Cambridge UK Jane C. GoodaIl Department of Rheumatology University of Cambridge Addenbrooke's Hospital Cambridge UK Kaisa Granfors Department of Bacterial and Inflammatory Diseases National Public Health Institute Turku Finland Robert D. Inman University of Toronto Toronto Western Hospital Toronto, Ontario Canada Lorna B. Jarvis Department of Rheumatology University of Cambridge Addenbrooke's Hospital Cambridge UK
xv
Participants
Simon Kollnberger MRC Human Immunology Unit Weatherall Institute of Molecular Medicine John Radcliffe Hospital Headington, Oxford UK
Dennis McGonagle University of Leeds Leeds and Calderdale Royal Hospital Halifax UK
Gerlinde Layh-Schmitt Division of Rheumatology Cincinnati Children's Hospital Medical Center Cincinnati, Ohio USA
Herman Mielants Department of Rheumatology University Hospital Ghent Ghent Belgium
Andreas Leibbrandt Institute of Molecular Biotechnology Austrian Academy of Sciences Vienna Austria Bernhard Loll Max-Planck-Institut fllr Medizinische Forschung Abteilung fur Biomolekulare Mechanismen Heidelberg Germany Jose A. Lopez de Castro Centro de Biologia Molecular Severo Ochoa Universidad Autonoma de Madrid Madrid Spain Carlos Lopez-Larrea Hospital Universitario Central de Asturias Oviedo Spain Walter P. Maksymowych Alberta Heritage Foundation for Medical Research University ofAlberta Edmonton, Alberta Canada
Rolf Misselwitz Institut filr Immungenetik Charite-Universitatsmedizin Berlin Freie Universitiit Berlin Berlin Germany Josef M. Penninger Institute for Molecular Biotechnology Austrian Academy of Sciences Vienna Austria Markus A. Penttinen Department of Medical Microbiology and Immunology University ofTurku Turku Finland Simon J. Powis Bute Medical School University of St. Andrews Fife, Scotland UK John D. Reveille Division of Rheumatology and Clinical Immunogenetics University of Texas Health Science Center Houston, Texas USA
Participants
xvi
Anna S. Sahlberg Department of Bacterial and Inflammatory Diseases National Public Health Institute Turku Finland Susana G. Santos Bute Medical School University of St. Andrews Fife, Scotland
UK Georg Schett Department of Internal Medicine 3 Institute for Clinical Immunology University of Erlangen, Nuremberg Erlangen Germany Edward M. Schwarz Clinical Immunology Research Center Rochester, New York USA Rosa Sorrentino Department of Cell Biology and Development University Sapienza Roma Italy Dawn P. Sowders Division of Rheumatology Cincinnati Childrens Hospital Medical Center Cincinnati, Ohio USA Paul P. Tak Division of Clinical Immunology and Rheumatology Department of Internal Medicine Amsterdam The Netherlands
Joel D. Taurog Rheumatic Diseases Division Department of Internal Medicine University of Texas Southwestern Medical Center Dallas, Texas USA Barbara Uchanska-Ziegler Institut ftir Immungenetik Charite-Universitatsmedizin Berlin Freie Universitat Berlin Berlin Germany Bernard Vandooren Clinical Immunology and Rheumatology University ofAmsterdam Ghent Belgium Ruth Wittoek Department of Rheumatology University Hospital Ghent Ghent Belgium David Tak Yan Yu Rheumatology Division Rehabilitation Center University of California Los Angeles, California USA Libin Zhang Department of Rheumatology University of Cambridge Addenbrooke's Hospital Cambridge
UK Andreas Ziegler Institut fur Immungenetik Charite-Universitatsmedizin Berlin Freie Universitat Berlin Berlin Germany
CONTENTS 1. CLINICAL ASSESSMENT IN THE SPONDYLOARTHROPATHIES ••••• 1 Ruth Wittoek and Herman Mielants Abstract 1 Introduction .................••.....................................................•••......••....•...............•.•..•............•1 Ankylosing Spondylitis 2 Psoriatic Arthritis.................•...............................•.............•...•....•..•..............••••.•.........•.....•.9
2. IMAGING IN SPONDYLOARTHRITlS•..........•..•....••••••...•••••.••••••••••••••••••• 17 Walter P. Maksymowych Abstract.........•..•....•..••...........•.....•...............•.....•..................•....•..•.........••..•...••••..••..•.••....... 17 Introduction 17 Plain Radiography .............•............................•...................................•....•.••.•....••....•..•..•... 18 Computed Tomography 21 Isotopic Imaging 21 Ultrasound 22 Magnetic Resonance Imaging .........•....................•..............•..............•....•...........•............. 22 Sacroiliac Joint ......................••..•.........................................•..••....•.••.....•........••.........•........24 Spine 27 Conclusions 33
3. SPONDYLOARTHRITlS, DIFFUSE IDIOPATHIC SKELETAL HYPEROSTOSIS (DISH) AND CHONDROCALCINOSIS •••••••••••••• 37 Jacome Brogues Annas, Ana Rita Couto and Bruno Filipe Bettencourt Abstract Introduction Spoodyloarthritis Diffuse Idiopathic Skeletal Hyperostosis (DISH) Chondrocalcinosis/Crystal Deposition Disease (Cppd CDD) Molecular Genetics Conclusions
37 37 37 40 40 42 47 xvii
xviii
Contents
4. THE ENTHESIS ORGAN CONCEPT AND ITS RELEVANCE TO THE SPONDYLOARTHROPATHIES...•.•............•.....•.•.•••..••..•••.• 57 Michael Benjamin and Dennis McGonagle Abstract 57 Introduction 58 Enthesis Organ Structure 58 Mechanisms for Reducing Stress Concentration at the Enthesis Itself 63 Related Concepts••..................................•...................••........................... .......................•... 66 Conclusions 67
5. SYNOVIAL AND MUCOSAL IMMUNOPATHOLOGY IN SPONDYLOARTHRITIS
71
Bernard Vandooren, Paul P. Tak and Dominique Baeten Abstract•......•......••....................................•.....•••.•....•.....••....................•...•..•........................ 71 Introduction ••..•...•...........•......•...•.............•.•.....•.•........•..•.....•....•.....••..•...•............•...•..•..•....71 Axial Versus Peripheral Spondyloarthritis.........••.....•.....•....•..•.......•........•.....•.....••.....••.• 73 Skeletal Remodeling in Spondyoarthritis ...........•••..•.....•..•.........•...••...••.•.............•.•.•.•.•..73 Synovial Biomarkers in Spondyoarthritis ......•...•.•..•.....•..•.•.......••...•..••••............•..•••.....•.75 Gut Inflammation in Spondyloarthritis ..•.•..•...........•...•.•.....•..•............••.........•.....•......•..• 76 The Pathogenesis of Spondyloarthritis: The Mucosa Hypothesis 77 Innate Immunity and Mucosal Inflammation 77 Adaptive Immunity and Mucosal Inflammation 78 The Pathogenesis of Joint Inflammation 78 Conclusions 79
6. BONE LOSS IN THE SPONDYLOARTHROPATHIES: ROLE OF OSTEOCLAST, RANKL, RANK AND OPG IN THE SPONDYLOARTHROPATHIES
85
Allen P. Anandarajah and Edward M. Schwarz Abstract..•..•...•...•.•.............•.......•......•.•................. ..•..••.....•.............••.....•.•..•..•........•.•.....•....85 Introduction ...•............................•...............................•....•............•••.•......................•.......... 85 Pathogenesis of Bone Loss 86 Bone Loss in the Spondyloarthritis .........•..•.........•........ ................... ................................ 91 Bone Loss in Ankylosing Spondylitis.•.•.......................•...•....•.........•................•.......••.•..... 91 Bone Loss in PsA 93 Bone Loss in Enteropathic Arthritis.................•......••.....•.....•..•......•....•..............••••..•....... 93 Bone Loss in Reactive Arthritis .•..•................................••......................................•.•........ 95 Conclusions.......•.......................•.....•................... .....•.......•.......•...................................... ..•.95
Contents
7. RANKL/RANK AS KEY FACTORS FOR OSTEOCLAST DEVELOPMENT AND BONE LOSS IN ARTHROPATHIES
xix
100
Andreas Leibbrandt and Josef M. Penninger Abstract Introduction Basic Characteristics of the Rankl-Rank-Opg Axis RANKL-RANK-OPG Interactions and Bone Remodeling RANK(L) SignaUng Pathways Modulators of Rankl-Mediated Osteoclastogenesis Rank-Rankl as Therapeutic Targets for Arthropathies Conclusions
100 100 101 l04 104 107 109 110
8. BONE FORMATION VERSUS BONE RESORPTION IN ANKYLOSING SPONDYLITIS
114
Georg Schett Abstract Introduction Lessons from Rheumatoid Arthritis Structural Damage in Ankylosing Spondylitis Mechanism of Joint Formation-Molecular Lessons for Joint Fusion Bony Protrusion as a Stress Response of the Joint Osteoblast-The Bone Forming Cells From the Osteoblast to the Osteophyte Molecular Regulation of Osteophytes Concepts of Osteophyte Formation Why Do Osteophytes Grow in AS and not in RA? Conclus ions
114 114 114 115 115 115 116 116 116 117 118 120
9. BIOMARKERS IN SPONDYLOARTHROPATHIES
122
Chun-Hsiung Chen, David Tak Van Yu and Chung-Tei Chou Abstract Introduction Purpose of Developing Biomarkers in Ankylosing Spondyltis How do Investigators Assess the Degree of Usefulness of SpondyUtis Biomarkers? Matrix Metalloproteinases (MMPs) and Tissue Inhibitors of Metalloproteinases (TIMPs) Expression of MMP-3 in Ankylosing Spondylitis Correlation between Serum MMP-3 and Disease Activity Sensitivity and Specificity of Serum MMP -3 in Evaluation of Disease Acti vity Changes of Serum MMP-3 in Response to Therapies
122 122 122 125 125 126 126 126 127
xx
Contents
Serum MMP-3 as Predictor of Disease Progression Are Molecules Reflecting Cartilage Turnover Biomarkers of Ankylosing Spondylitis? Macrophage Colony Stimulating Factor (M-CSF) Serum Amyloid A (SAA) Interleukin-6 (IL-6) Conclusions
127
10. THERAPY OF SPONDYLOARTHRITIDES
133
128 128 128 129 129
Jiirgen Braun Abstract Introduction Management. Imaging Conclusions
133 133 134 138 143
11. GENOMEWIDE SCREENS IN ANKYLOSING SPONDYLITIS
148
Matthew A. Brown Abstract Introduction Methods for Identifying Disease-Causing Genes Linkage Studies and Ankylosing Spondylitis High Density Association Studies and Ankylosing Spondylitis MHC Genetics of Ankylosing Spondylitis Future Studies Conclusions
148 148 149 150 151 153 154 155
12. SUBTYPES OF HLA-B27: HISTORY AND IMPLICATIONS IN THE PATHOGENESIS OF ANKYLOSING SPONDyLITIS .... 159 John D. Reveille Abstract Introduction HLA-B27 Subtypes Associations with AS The Lack of Disease Association ofHLA-B*2706 and HLA-B*2709 How Amino Acid Differences Between HLA-B27 Subtypes May Impact on Disease Pathogenesis Conclusions
159 159 166 167 167 172
Contents
xxi
13. IMPLICATIONS OF STRUCTURAL AND THERMODYNAMIC STUDIES OF HLA-B27 SUBTYPES EXHIBITING DIFFERENTIAL ASSOCIATION WITH ANKYLOSING SPONDYLITIS ••••••••••••••••••••••••••••••••.•••••.•••••••••..••••••••••••••••••••••••••••••••••. 177 Andreas Ziegler, Bernhard Loll, Rolf Misselwitz and Barbara Uchanska-Ziegler Abstract Introduction General Structural Properties of HLA Class I Molecules HLA-B27 Subtypes Exhibiting Differential Disease Association HLA-B27: Peptide Complexes for Structural and Other Biophysical Studies Structural Studies with the m9 Peptide Structural Studies with the Proven HLA-B27 Self-Peptide Tis Structural Studies of the Viral Peptide pLMP2 HLA-B27 Subtype Dependent Conformational Peptide Dimorphisms HLA-B27 Subtype- and Peptide-Dependent Molecular Mimicry Thermodynamic Analyses of Selected HLA-B27: Peptide Complexes Conclusions
177 178 178 180
186 187 189 192
14. HLA-B27-BOUND PEPTIDE REPERTOIRES: THEIR NATURE, ORIGIN AND PATHOGENETIC RELEVANCE
196
180 181 183 185
Jose A. Lopez de Castro Abstract Introduction HLA-B27 Polymorphism and Association to AS Peptide Specificity of HLA-B27 Subtypes T-Cell Presentation of Shared Ligands by HLA-B27 Subtypes Molecular Mimicry between Self and Foreign Ligands of HLA-B27 The Origin of HLA-B27-Bound Peptide Repertoires: Role of the Proteosome Assisted Loading of the HLA-B27 Peptide Cargo Conclusions
196 196 196 199 202 203 204 204 206
15. BIOCHEMICAL FEATURES OF HLA-B27 AND ANTIGEN PROCESSING
210
Simon J. Powis, Susana G. Santos and Antony N. Antoniou Abstract Introduction The MHC Class I Assembly Pathway A Key Role for Cys-67? HLA-B27 in Dendritic Cells Conclusions
210 210 211 212 213 214
xxii
16. HLA-B27 MISFOLDING AND SPONDYLOARTHROPATHIES
Contents
217
Robert A. Colbert, Monica L. DeLay, Gerlinde Layh-Schrnitt and Dawn P. Sowders Abstract 217 Introduction 217 Usual, Unusual and Unique Features ofHLA-B27 and Their Potential Role in Disease Pathogenesis 218 Importance of Protein Folding 219 220 Consequences of Protein Misfolding HLA-B27 Misfolding 222 Evaluating the Role of HLA-B27 in Disease 223 HLA-B27 Subtypes and Spondyloarthritis 228 Overexpression of Additional h~2m: The Model of SpA...•.......•.........•......................... 229 Conclusions 229
17. HLA-B27 AND HOST-PATHOGEN INTERACTION ..••...••...•...........•.. 235 Anna S. Sahlberg, Kaisa Granfors and Markus A. Penttinen Abstract Introduction The Molecular Characteristics of HLA-B27 HLA-B27 and Host-Pathogen Interaction HLA-B27 and Host Cell-Pathogen Interaction HLA-B27 and Survival of Intracellular ReA-Triggering Bacteria HLA-B27, ER Stress-Induced Signal Transduction Pathways and Host-Pathogen Interaction HLA-B27 and LPS-Induced Inflammatory Response Conclusions
235 235 236 236 237 238 239 240 241
18. ANIMAL MODELS OF SPONDYLOARTHRITIS
245
Joel D. Taurog Abstract Introduction Naturally Occurring Spondyloarthritis in Primates Proteoglycan-Induced Spondylitis and Arthritis in Mice Naturally Occurring Spontaneous Models Ankylosing Enthesopathy HLA-B27 Transgenic Mice TNF Transgenic Mice HLA-B27 Transgenic Rats A New, Improved AS Model in HLA-B27 Transgenic Rats Conclusions
245 245 245 246 246 247 248 248 248 249 250
Contents
xxiii
19. T-CELL RESPONSES AGAINST VIRAL AND SELF-EPITOPES AND HLA-B27 SUBTYPES DIFFERENTIALLY ASSOCIATED WITH ANKYLOSING SPONDyLITIS 255 Maria T. Fiorillo and Rosa Sorrentino Abstract 255 Introduction 255 Searching for the Holy Grail: From B27-Restricted Viral Antigens to Self-Ligands Eliciting Autoreactive Cytotoxic T-Cell Responses in Patients with Ankylosing Spondylitis.••••••••.••••.•.•....••••.....•••......••....••..•••..••.••••...256 Extension of T-Cell Cross-Reactivity to a Third Player: pGR.......••.....•••.••••.....•..••••... 259 T'-Cell Cross-Reactivity between Self and Non-Self: The tCR Point of View ••••••.•••.••260 Conclusions ..•.....•.••.•••••..•.....••.•..••...•....•.•.....•••.....•.•••.•....•.......••........••.......•...••••.....•..••••...261
20. DENDRITIC CELL: T-CELL INTERACTIONS IN SPONDYLOARTHRITIS•.••••......••....•.......•••....•••................•.••.••.••• 263 J.S. Hill Gaston, Lorna B. Jarvis, Libin Zhang and Jane C. Goodall Abstract•...•......................•...................................................•...•••......•.......••.........•.•.•••..••...263 Introduction.•.................•............................................................••......••...••...•....••.•..•........ 263 A Novel Subset of CD8+T Cells 264 The Influence ofthe Unfolded Protein Response on DC Function••......••..................••270 Conclusions................•............••...........•.•.....•.•...•....•.•.•.........•..•.•••...•..••..•.•...•....•..•..••....... 275
21. THE ROLE OF B27 HEAVY CHAIN DIMER IMMUNE RECEPTOR INTERACTIONS IN SPONDYLOARTHRITIS....•... 277 Simon Kollnberger and Paul Bowness Abstract...............................................................................•........................... ••...•..•...•..... 277 Introduction 277 Leukocytes Express Cell Surface B27 Heavy Chain Homodimers.•...•....•.....•...........•. 278 Immune Receptors Interacting with HLA-B27 Heterotrimers and Dimers .......••..... 279 Differential Binding ofB27 2 and B27 Heterotrimers to KIR, LILR and PIR•........•.•281 Involvement of KIRJDL2-Expressing T'-Cells and NK-Cells in Spondyloarthritis 282 Conclusions 282
22. KIR GENES AND THEIR ROLE IN SPONDYLOARTHROPATHIES
286
Roberto Diaz-Pefia, Miguel Angel Blanco-Gelaz and Carlos Lopez-Larrea Abstract Introduction Complexity of KIR Family KIR and Spondyloarthropathies Conclusions
286 286 287 292 295
xxiv
23. INNATE IMMUNITY OF SPONDYLOARTHRITlS: THE ROLE OF TOLL-LIKE RECEPTORS
Contents
300
Robert D. Inman Abstract.••...•..••..........•...........•....................................••......•.......................•....•................. 300 I ntroduction 300 Current Concepts in Innate Immunity 301 Innate Immunity in Host Response to Arthritogenic Bacteria 302 Innate Immunity and Spondyloarthritis 304 Conclusions 307
INDEX
311
ACKNOWLEDGEMENTS This book was supported in part by Spanish Grants SAF-2004 /02669, FYCIT (PC-07/006) and by CAJASTUR and ASHISTO.
xxv
CHAPTER!
Clinical Assessment in the Spondyloarthropathies Ruth Wittoek and Herman Mielants'"
Abstract n order to measure disease activity. progression and response to therapy, it is important to use accurate, reliable and feasible outcome measures that can ideally be used in longitudinal cohorts. clinical trials and clinical practice . With emerging therapies, the focus on the methodology ofoutcome assessment has increased to ensure that discriminant and responsive instruments are used. This chapter reviews available measures of three major areas of disease impact in the spondyloarthropathies (disease activity. structural damage and functioning) and discusses the relevance for use in clinical practice. First. the outcome measures available for the assessment of different domains in ankylosing spondylitis, composite-indice s and response criteria for use in clinical trials and clinical practice in ankylosing spondylitis are discussed. Secondly, the performance ofthe se in psoriatic arthritis and more disease-specific instruments in psoriatic arthritis are discussed.
I
Introduction The spondyloarthropathies (SpAs) are a group ofdisorders with common clinical, radiological and genetic features, clearly distinct from other inflammatory rheumatic d isorders such as rheuma toid arthritis (RA). Ankylosing spondylitis (AS) is the prototype disease belonging to this concept. Other diseases that are considered to be part of this group are psoriatic arthritis (PsA), arthritis associated with inflammatory bowel diseasessuch as Crohn'sdiseaseand ulcerative colitis and reactive arthritis. Patients are classified as having SpA if the y have either inflammatory low back pain or a peripheral arthritis with any ofthe following: positive family history, psoriasis, alternating buttock pain , enrhesopathy or sacroiliitis (the European Spondyloarthropathy Study Group (ESSG)).I Specific classification criteria are made for AS and PsA (the CASPAR criteria for PsA2 and the modified New York criteria for AS3).Patients fulfilling the ESSG criteria but cannot be classified into one of the more specific subgroups are termed undifferentiated spondyloarthropathies. There a major heterogeneity in the clinical landscape of the SpAs. There 's a possibility of combination in any individual patient ofaxial disease (sacroiliids, spondylitis), several peripheral articular disease manifestations, going from peripheral arthritis (generally pauci-articular, asymmetrical and by preference involving the joints ofthe lower limbs) to enrhesitis and dactylitis and extra-articular manifestations (such as uveitis, psoriasis, inflammatory bowel disease). Due to the broad clinical spectrum of the manifestation of the disease, it is obviously difficult to assess the disease activity by a single outcome measure.
' Cor respo nd ing Author: Herman Mielants-Department of Rheumatology, University Hospital Ghent, De Pintelaan 185, 900 0 G hent, Belgium. Email: herman.miel
[email protected]
Molecular Mechanisms ofSpondyloarthropathies, edited by Carlos Lopez-Larrea and Roberto Diaz-Peiia. ©2009 Landes Bioscience and Springer Science+Business Media.
2
Molecular Mechanisms ofSpondyloarthropathies
In order to measure disease activity, progression and change with therapy, it is important to use accurate , reliable and feasible measures that can ideally be employed in longitudinal cohorts, clinical trials and clinical practice. Over the past SO years, multiple outcome measures and composite scoring systems have been developed and are availablefor clinical application. Frequently these measurements have not been standardized or assessedfor reliability, validity, or sensitivity to change. Moreover, performance of some measures is time-consuming and may not be feasible in clinical practices. Therefore the goal would be to identify a minimum number of individual clinically appropriate measures to accurately assess the different manifestations of disease. Therefore, the OMERACT (Outcome measures in rheumatoid arthritis clinical trials) process involves consensus on outcome measures and is based on the 'OMERACT filter: composed ofthree key components: truth, discrimination and feaslbiliry," Truth addresses the issuesofface, content, construct and criterion validity; or does the instrument measure what it is intended to? Is the result unbiased and relevant? Discrimination addresses issues ofreliability (inter- and intraobserver) and sensitivity to change . Does the instrument detect differences between groups? Feasibility addresses whether the measure can be easily applied or used with regard to time, money, training, scoring , interpretability and acceptance by physician and patient. The assessment ofdiseasein SpA is a subject that has attracted increasing attention over the last few years. The advent of biologic treatments, such as tumor necrosis factor (TNF) inhibitors, has provoked a therapeutic breakthrough in inflammatory rheumatic diseases and consequently raised the interest in the development and validation ofobjective and standardized tools to assessdifferent aspects of the disease in order to investigate the efficacy ofthe newer therapeutical options.' In particular, groups such as the Assessments in Ankylosing Spondylitis (ASAS) international working group and later also the group for Research and Assessments in Psoriasis and Psoriatic Arthritis (GRAPPA), in conjunction with the Society for Outcome measures in Rheumatology, have contributed to the development ofsuch instruments and their diffusion amongst the rheumatologic community. The most important question is to decide what to measure in a patient and how to measure it? Therefore it is important to determine which aspects of the disease are the most appropriate to measure.
Ankylosing Spondylitis Disease activity in ankylosing spondylitis (AS) is measured by multiple measures including metrology, radiology, laboratory, functioning and disabiliry," The ASAS group established a core set offive domains and also the choice ofmeasure to represent each domain: the domains are the patient's global assessment, pain, physical function, spinal stiffuess/inflammation and spinal rnobility,?·8 The patient's global assessment is represented by the VAS (visual analogue score) global assessment score (0-100 scale), pain is represented by the VAS pain score (0-100 scale), function is represented by the BASFI score (0-100 scale) and inflammation is represented either by (first choice) the mean ofthe two morning stiffness-related BASDAI VAS scores or by (second choice) morning stiffness duration with a maximum of 120 minutes. With regard to spinal mobility, there is lower consensus about the preferred measure to assess this domain. The ASAS working group has recommended that spinal mobility measures be used as part of the disease controlling antirheumatic therapy criteria."
Symptoms and Disease Activity The predominant symptoms ofthe disease are related to the spine. Acute symptoms, pain and stiffness represent the inflammatory activity of the disease.
Axial Pain Spinal involvement is the characteristic and most prominent feature ofAS. The pain is mostly located in the sacroiliacal joints and the spine. Inflammatory low back pain (IBP) (Table 1) is thought to originate from the active disease process and mechanical back pain is thought to originate from structural damage due to ongoing disease. Classically, IBP is associated with morning
Clinical Assessment in the Spondyloarthropathies
3
Table 1. Criteria for inflammatory low back pain (IBPJ9 Chronic back pain >3 months and Onset of disease <45 years and 1. Morning stiffness >30 minutes 2. Awakening due to pain in the second half of the night 3. Improvement upon exercise, not with rest 4. Alternating buttock pain 3 out of 4 criteria: high likelihood of IBP
stiffness and nocturnal pain (in the second halfofthe night) and improves upon physical activity. Alternating buttock pain is often present in AS. In contrast, mechanical pain worsens by activity and improves by rest. Recently, a new set of criteria for IBP is established and performed better than previous criteria in AS patients with established disease (Table 1).9 A prospective study is needed to validate the diagnostic properties ofthe new candidate criteria set in patients with early disease. The ASAS core set specifies nocturnal spinal pain and general spinal pain as the two most useful measures for monitoring pain in patients with AS. Pain can be measured quantitatively and qualitatively on a number ofscales.The most commonly used scales are visual analogue scales (VAS) on which the patient can indicate the level of pain that was experienced over a period of time on a continuous line of 100 mm length running between 'no pain' and 'unbearable pain' and numerical rating scales (NRS) where the patient rates the experienced pain between 0 ('no pain') and 10 ('unbearable pain').
Stiffness In AS, morning stiffness primarily refers to stiffness ofthe spine due to inflammation. During the development of the ASAS response criteria , the average ofduration and severity ofmorning stiffness performed better than either of the two scales separately. The easiest method to assess morning stiffness is to measure ftom the time of awakening . This can be done quantitatively or by indicating the duration on a VAS scale using 'no spinal stiffness' and 'two hours or more ' as extremities on the 100 mm line. Peripheral Disease Manifestations Although AS is primarily a disease of the axial joints (spine , sacroiliacal and hip joints), extra-axial manifestations ofthe disease are frequent and need separate assessment. Approximately 25% of the patients will have active peripheral joint involvement, typically oligoarticular, asymmetrical and by preference involving the lower limbs. There are no specific assessment tools available for AS, however the existing tools in RA are widely used in clinical trials in AS. A 44-joint count has been proposed to measure peripheral joint involvement in clinical trials, which covers the joints most commonly affected in AS and is thus more specific than the customary 28-joint count in RA. Although a useful tool for baseline assessment in clinical practice, this is clearly insufficient for monitoring the individual patient for disease progression or treatment response in affected joints. The assessment of enthesitis in AS patients is even more problematic. Many potential sites of enthesitis are clinically inaccessible and therefore difficult, if not impossible , to assess. Clinical examination of enthesitis in the lower limbs has a sensitivity of only 22.6% and a specificity of 79.9%, giving a positive predictive value ofIess than 60%.10 There are currently three validated instruments used in clinical trials for measuring enthesitis: the Mander (Newcastle) Enthesitis Index (MEl) was the first standardized instrument to assess the tenderness over entheses in AS patients. I I More recently, two other indices has been described and validated; the Maastricht Ankylosing Spondylitis Enthesitis Scale (MASES) and the Berlin Entheslris Index (BEl). The MASES is the most accepted and takes into account the 13 most specific and sensitive sites from the MEl (first costochondral joint, seventh costochondral joint,
4
Molecular Mechanisms ofSpondyloarthropnthies
posterior superio r iliac spine, anterior superio r iliac spine, iliac crest, fifth lumbar spinous process and proximal insertion of the Achill es tendon). It sco res each site as 0 ('no pain') or 1 ('painful'), giving a total pos sible score of 0-13. It is certainly a more feasible instrument, but it has not been assessed in other diseases ofthe SpA concept including PsA. Moreover, the fact that it does not sco re one of th e m ain enthesitis sites of th e feet (inse rt ion of the plantar fascia on the calcaneus) gives rise to so me co ncern with mo st clinici ans. To overcome this, Braun et aJl2used an enthesis index composed of 12 major entheses reported to be commonly affected in the inflammato ry process in AS in a stu dy on infliximab in AS. It is scored sim ilarly as the MASES, giving a po ssible score of 0-12. The index did not perform better compared to the MASES in th e above -mentioned stu dy, however Gladman et al 13 have reported on the performance of inve stigators from the C anadian Spondyloarthropathy Group in the ir ability to reliably assess four different enthesitis area: the plantar fascia, the Achilles tendon, tibial tuberosity and rotator cuff insertions. Observer agreement was 'mo derate' in the first three locations and 'poo r' in the fourth. Fatigue Fatigue is an important cause of morbidity in AS. It's reported in up to 65% of patients, but oft en under-recognized by physicians. It ha s an important impact on many different aspects ofthe general wellbeing ofa patient, including the ability to perform daily act ivitie s and social contacts. There is no specific instrument or questionnaire that mea sure s fatigue in AS but the general fat igue questionnaires perform well in AS . The simplest tool is a VAS scale ranging from 'no fatigue' to 'very severe fatigue'. The ASAS international working group ha s selected the BASDAI question on fatigue as the preferred instrument to assess fatigue in AS. Mobility It is not possible to differentiate between changes in spinal mobility due to inflammation and those due to spinal ankylosis. Therefore, isolated spinal measurements are not informative, but have to be interpreted alongside other information on the patient's status. Different individual measures are available and are all feasible to perform in clinical practice. Performance of those measurements in practice are de scribed by the ASAS working group on the website."
Cervical Rotation Cervical rotation can be tested in several ways. By using a Myrin inclinometer full rotation from left to right can be measured or with a tape measure the change in difference between the chin and the coronoideus process in maximal cervical rotation from left to right can be measured. Another method using a tape mea sure assesses the d ifference in the distance between a mark in the suprastern al notch and the tragus ofth e right ear when the neck is rotated from left to rlght." The Bath Ankylosing Spondyliti s Metrology index (BASMI) measure s cervical ro tation using a gravity-action goniometer. The inter- and intraobserver reliability is 0.98 and 0.99 respectively." All different ways have shown to be reliable. However, measurements th at require specific instrument may be less feasible in daily practice.
LateralFlexion ofthe Neck Thi s can be me asured by using an incl inometer or by measuring the distance between the tragus and th e coronoideus process ofthe clavicle in maximal lateral bending of the head to the left and right. Both methods correlate moderate with radiological change (r = 0.57 and r = 0.60, respectively) and show excellent inter- and intraobserver reliability (ICC 0.90 and 0.96 , respectively for mea suring by inclinometer and ICC 0.97 and 0.98, resp ectively for tape mea suring)."
Chin to Chest Th e mea sure of the forward flexion of th e ne ck is found to correlate poorl y with radiological di sease. It is, ho wever, found to be reliable, with inter- and intraobserver co rrel ation of 0.92 and 0.95 , respecrively.v It ha s not been included as an outcome measure in clinical trials, so sens itivity to cha nge is not available.
ClinicalAssessment in the Spondyloarsbropatbies
5
Occiput to Wall Distance (OWD) This measures the distance between the occiput and the wallwhen the patient stands with heels and shoulder against the wall with the back straight. It is found to correlate with radiographic change and is also proved to be sensitive to change, with an effect size of0.25. 16 Tragus to Wall Distance (TWD) The distance between the tragus of the ear and the wall is measured when the patient stands with heels and buttocks against the wall, knees straight, shoulders back and places the head as far as possible,keeping the chin in . Correlation with radiographs washigh (r ~ 0.92) and intraobserver reliability is very good (r = 0.99). OWD and TWD are equally reliable in assessing thoracic spine extension. It's recommended to measure the OWD over TWD since in the former a value ofzero easily distinguishes patients with normal thoracic spine extension from kyphotic patients.'? Schober's Test Several variations exist on the Schober's test. The original report in 1937 18 described the Schober's test as the distance between the lumbosacral junction and a mark 10 em above this, recorded when the patient bends maximally forward. A first modification was done by Macrae and Wright,19 which added an additional mark 5 cm below the first mark and showed that both the original Schober and their modification reflected forward flexion (confirmed by radiographs). Moll and Wright further refined the Schober's test by providing an anchor for the lumbosacral junction using a line across the upper limit ofthe dimples ofVenus.20 The subject was asked to bend forward maximally and the new distance between the upper and lower marks was measured. A moderate correlation between spinal mobility and duration ofsymptoms in patients with AS and no correlation with age were found." The modified Schober is included as one ofthe measures of the BASMI. However, in the BASMI the first mark is made as a line across the iliac crests, which is at the level of L5 rather than at the lumbosacral junction. Inter- and intraobserver reliability of the Macrae and Wright modification of the Schober's test is excellent (ICC 0.96 and 0.94) and it correlates well with the radiographic changes in the spine (ICC 0.7l).22The modified Schober is demonstrated to be sensitive to change (ES 0.24), following intensive physiotherapy, although another study could not confirm rhis." Moreover, the Schober's test did not distinguish between patients treated with active drug or placebo in a placebo controlled Randomized Clinical Trial(RCT) with infliximab in AS.24 Finger to Floor Distance Forward flexion of the spine using finger to floor distance was included among the measures stud ied by Viitanen et al l6 and proved to have a good reliability and sensitivity, but did not correlate with radiographic changes. Another study demonstrated that finger to floor distance improved following physiotherapy in patients with AS.23 Lateral Lumbar Flexion The distance between the tip ofthe third finger and the floor when the patient stands upright with heels, buttocks and shoulder against the wall and when the patient maximally bends sideways without lifting the opposite foot offthe ground is measured . This method provides excellent reliability and is easy to perform using only a tape measure and requires minimal training. Chest Expansion This is measured by difference in chest circumference between full expiration and inspiration at the fourth intercostal space. Chest expansion did not correlate with radiographic changes (r = 0.38).16.22 However, intra- and interobserver reliability is good (ICC 0.95 and 0.85, respectively) and sensitivity to change was noted, with an effect size of0.42. Although chest expansion did not correlate with radiographic change, it did provide reliability and discrimination and is included in the BASMI and the ASAS response criteria.
6
Molecular Mechanism s ofSp ondyloarthropathies
Intermalleolar Distance Intermalleolar distance measures abduction ofthe hips. Keeping the knees straight and the legs in contact with the resting surface, the patient is asked co take the legs as far apart as po ssible and the distance between the medial malleoli is measured. Intermalleolar distance provides excellent inter- and intraobserver reliability (ICC 0.98 and 0.99, respectively) and is part of th e BASMI. Although widely used, the occiput-co-wall di stan ce and modified Schober's test primarily reflect irreversible structural damage and demonstrate limited sensitivity co change in studies of intensive physiotherapeut ic interventions. I3,23.25 BASMI (Bath Ankylosing Spondylitis Metrology Index) The ASAS core set initially recommended the use ofthre e measurement s: the occiput-co-wall distanc e, co reflect the cervical and thoracic mobility, chest expan sion , which evaluates the thoracic expansion capacity and the modified Schober index, to measure the lumb ar mobility. At a later stage, the lateral lumba r flexion was added co the ASAS core set. Another possibility is the composite index of several measures, the entire Bath Ankylosing Spondylit is Metrology Index (BASMI) (Table 2). This score includes the measurements of cervical rotation , tragus-to-wall d istanc e (which showed to be equivalent co the OWD in a direct comparative stu dy, I? anterior lumbar flexion (modified Schober test), lateral lumbar flexion and intermalleolar d istance. The HASMI was published with a 0-2 answer ing scale (BAM SI2) and a 0-10 mod ification (BAMSIlO) : co calculate the BASMI2 each of the five assessments is scored between 0 and 2 based on a conversion table ; the BASMI lOuses categories between 0 and 10. BASM12 the sum ofthe five separate assessments, BASMI lOis the sum ofthe five assessments divided by five, thus giving a range between 0 (best possible) and 10 (worst possible) (Table 3). Amongst all five assessments, the lateral lumbar flexion seems co have the highest standardized response mean (thus the best representing the total BASMI),26 followed by the Schober's test." Finally, chest expansion, the only assessment of rib cage mobility, although no t used in the BASMI has proved to be sensitive to change during anti-TNF trialS.28•29
Table 2. Outcome measure to assess movements of the spine and hip. The respective elements of the BASMI and fDASMI Location Cerv ica l spi ne
Lumbar spine
Direction of Movement
BASMI
X Cervical rot ation by grav ity actio n goniometer Cervica l rotation by tape measure between chin and the corono ideus pro cess Lateral flexion Lateral flexion of the neck Forw ard flexion Chin to chest Occi put to w all Tragus to w all X Forward flexion Schober's test
EDASMI
Rot ation
M odified Scho ber Finger to floo r distance Lateral lumbar flexion Chest expansion
X
Abduction Intermalleolar distance Internal rotation
X
Lateral flexion Thoracic spine Hi p
Outcome Measure
X
X
X X
X
7
Clinical Assessment in the Spondyloartbropatbies
Table 3. Conversion table BASMI BASMI
o
Tragus-to-wall (em) Cervical rotation n Lumbarside flexion (em) Modified Schober's (em) Intermalleolar distance (em)
<15 >70 >10 >4 >100
2 15-30 20-70 5-10 2-4 70-100
>30 <20 <5 <2 <70
EDASMI (Edmonton AS Metrology Index) Another composite score was developed by the Canadian group of Maksymowych and is a 4-item composite measure ofspinal and hip mobilitity (Table 2).15 The four mobility measures comprising the EDASMI were the cervical rotation, measured by the distance between the suprasternal notch and the tragus after maximal rotation towards both sides, the chest expansion, the lateral lumbar flexion and the internal rotation ofthe hip. It meets the standards of feasibility, truth and discrimination and a key attribute to this tool is its simplicicty: all measures require a measuring tape only and three of the four measures are conducted with the patient seated on the examining table. Construct validity is comparable with the BASMI, whereas responsivene ss is more evident using the EDASM I.1 5To calculate the score, each of the four assessments is scored between 0 and 4 on a conversion table (Table 4). This results in a scoring range of 0- 16 for th e four measures comprising the EDA SMI.
Functioning Physical Function Physical function is a subset of health-related quality of life and disability is a concept that include s social functioning and adaptation and reflects more than only loss of function. There are a number of available AS-specific instruments, the best validated being the Bath Ankylosing Spondylitis Functional Index (BASFI) (Fig. 1) and the Dougados Functional Index (DFI).30.3ITh e forme r is the preferred disease-specific tool to assess physical function. Both have been shown to perform well with regard to validity, reliability and responsiveness, although the D FI may not be as respon sive to small change s as the BASFJ.32 Quality ofUfe The ASAS core sets for measuring AS do not include the domain 'health related quality of life' (H RQ oL); nevertheless, it is agreed that problems with HRQoL are a natural consequence of impairments in functioning and health and a part of more extensive patient assessment. The Ankylosing Spondylitis Quality ofUfe (ASQol) questionnaire is the best studied disease-specific measure available to assessHRQoL in AS patients, is easy to administer and correlates well with the generic medical outcomes survey SF_36.33.34
Table 4. Conversion table fDASMI EDASMI
o
Cervical rotation (em) Lumbar lateral flexion (em) Chest expansion (em) Hip internal rotation n
;04.0 ;01 8. 0 ;06 .0 ;047
<4.0 and ;02.8 <18 .0 and ;013.8 <6.0 and ;04.5 <47 and ;042
2
3
4
<2.8 and ;02. 0 <13.8 and ;010.2 <4.5 and ;03 .0 <42 and ;036
;01 .4 a nd <2 .0 ;05.7 and <10.2 ;02 .3 and <3.0 ;02 7 and <3 6
<5.7 <2.3
<1.4
<27
8
Molecular Mechanisms tifSpondyloarthropathies
1. Putting on your sock. or tights without help or aids (e.g . sock aid).
Figure 1. Bath Ankylosing Spondylitis Functional Index (BASFI).
Structural Damage The result of disease activity, the 'structural damage: represents the irreversible changes that accompany AS: spinal syndesmophyres and ankylosis, destruction and ankylosis of sacroiliacal and hip joints. Staging ofradiographic damage has been described extensively and several tools are well known. From the three existing validated scoring methods in AS (being BASRI, SASSS and mSASSS) ,35 m-SASSS was selected to be the preferred method by the ASAS international working group." Imaging is not included in the ASAS core set for clinical record keeping because it is insensitive to change in daily practice. The existing tools are reserved for use in clinical trials.
ClinicalAssessment in the Spondyloarthropathies
9
The instruments measuring structural damage will be described extensively in the next chapter ('Imaging in SpA and AS', by D. Vander Heijde),
Composite Index The Bath Ankylosing Spondylitis Disease Activity Index (BASDAI)(Fig. 2) was developed as a composite index consisting of an evaluation offatigue, axial pain, peripheral pain, stiffuess and enthesopathyon a visual analogue scale." Several studies have proved the BASDAI to be a valid and appropriate composite to define disease activity in AS.37 For the purposes ofinclusion in clinical trials, active disease has been defined as a BASDAI of>4 (possible score 0.10). The BASDAI has also a meaning in monitoring therapy ofAS patients: an improvement of the BASDAI of at least 50% from baseline (BASDAI50%) is considered by ASAS experts as an adequate tool for monitoring anti-TNF therapy in daily clinical practice and for considering the discontinuation of this type of expensive treatment in those patients whose response is less than 50% after 6-12 weeks anti-TNF treatment."
Response to Treatment In addition to the development and assessment ofthe several assessment tools and core sets of health domains in AS, the ASAS international working group has also developed specific criteria to assess response to treatment in AS trials (Table 5). The initial response criteria'? were derived using patient data from five trials ofNSAIDs in ASY·40-43 Statistical analysis was used to select the appropriate instruments to assess improvement on each ofthe four domains were defined, as well as the definition ofimprovement. The short-term improvement criteria are known as the ASAS20 % response criteria (ASAS20) and treatment response is defined as an improvement of at least 20% and at least 10 units (scales 0-100 on VAS) at least three domains without deterioration ofa similar magnitude in the potentially remaining domain. Other criteria were developed to use in clinical trials with anti-TNF therapy" and a similar statistical procedure was applied to the data from two randomized controlled trials ofinfliximab and etanercept in AS patients. Two sets of criteria were derived from this analysis: the ASAS4O , which includes the same domains and assessment tools as the ASAS20, but uses 40% and at least 20 units of improvement in at least three domains without any deterioration in the potentially remaining domain and ASAS 5/6, which requires a 20% or lO-unit improvement in five of six selected domains. The concerning six domains include the four domains of the ASAS20 and the additional domains 'sp inal mobility' and 'CRP: since anti-TNF therapy is very likely to have disease controlling properties reflected by improvement in those domains, which are known not to be influenced by NSAIDsY.39 The ASAS response criteria were developed for use in clinical trials and are not validated for use in clinical practice. A Simpler definition of treatment response to apply to daily patient care is the BASDAI50% response.F'"
Psoriatic Arthritis Psoriatic arthritis is manifest in diverse ways including: peripheral joint pain and swelling, spinal pain and stiffness, articular damage, enthesiris, dactylitis, iritis and skin psoriasis. Such disease manifestations may cause major disturbances in a number ofareas oflife. The Group for Research and Assessment ofPsoriasis and Psoriatic Arthritis (GRAPPA) selected 6 key domains: peripheral joint activity, skin activity, pain, patient global assessment, physical function and health-related quality oflife.45 These domains can be assessed by individual and composite measures. A number of measures have been 'borrowed' from the fields ofrheumatoid arthritis, ankylosing spondylitis and psoriasis and adapted to PsA. Others are being developed specifically for PsA. Few are validated but most have been shown to perform well in distinguishing placebo from treatment response.
10
Molecular Mechanisms ojSpondyloarthropathies
1. How would you describe the overall level of fatigueltiredness you have experienced?
none severe
Non severe
Figure 2. Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) .
Symptoms and Disease Activity Skin Psoriasis The most important and most commonly form ofpsoriasis is psoriasis vulgaris. In the specialism of rheumatology palmoplantar pustulosis is also important. The basic characteristics ofpsoriasis lesions-redness, thickness and scaliness - provide a means ofassessing the severity of psoriasis. A number ofmeasures to assessthe extent and severity ofskin involvement have been recently reviewed in detail. The current gold standard for assessment ofextensive psoriasis has been the Psoriasis Area and Severity Index (PASI).46The PASI is a measure of the average redness, thickness and scaliness ofthe lesions (each graded on a 0-4 scale), weighted by the area of involvement. While the PASI has been the most widely used measure, it does have a number oflimitations, one of which is its poor sensitivity to change for relatively small areas of involvement. Another key measure used in clinical trials of psoriasis is the physician global assessment (PGA). It has the advantage that it can be done for extensive disease as well as localized plaques. The static form ofPGA has become the standard. Although the PASI is probably the gold standard for assessment of severe psoriasis (and PGA is also widely used), there have been few validation studies and the construct validity and sensitivity to change are not well characterized." For patients with severe psoriasis, clinicians typically consider at least 75% improvement in disease to be a clinically meaningful improvement indicative ofsuccess.
Clinical Assessment in the Spondyloarthropathies
11
Table 5. Response criteria and criteria for residual activity (partial remission) developed by the ASAS group 1. ASAS response criteria: ASAS20 (ASAS40) A patient is considered as a responder if he/she fulfilled the following: •
•
An improvement of at least 20% (40%) and absolute improvement of at least of 10 (20) on a 0-100 scale in at least 3 of the following domains: i. Patient global, in the last week, measured on a VAS 0-100 mm scale ii. Pain, measured on a VAS 0-100 mm scale iii. Function, measured by BASFI iv. Inflammation, measured as the mean of the last two BASDAI quest ions (intensity and duration of morning stiffness) Absence of deterioration (of at least 20% (40%) and absolute change of at least 10 (20) on a 0-100 scale in the potential remaining domain.
2. ASAS partial remission criteria A patient is considered to be in partial remission if he/she attained: •
A value below 20 on a 0-100 scale in each of 4 domains: i. Patient global ii. Pain iii. Function iv. Inflammation
3. ASAS five out of six response criteria (ASAS 5/6), developed for use in trials with anti-TNF therapy •
An improvement of at least 20% in at least five of the following six domains: i. Pain ii. Patient global iii. Function iv. Inflammation v. Spinal mobility, measured by BASMI vi. Acute phase reactants, by C-reactive protein
The National Psoriasis Foundation (NPF) has developed the NPF Psoriasis Score (NPF-PS), a responder index that includes sixsubdomains: induration at two target sites,current and baseline body surface area, physician global assessment, patient global assessment and patient assessment ofitch. A major component of the assessment of psoriasis now is the measurement of quality of life. Quality of life can be measured at different levels. Nonspecific measures such as the Medical Outcome Survey Short Form 36 (SF -36), the Euro Qol and utility measures assess patients' overall quality of life.48.49There are more specific instruments that focus on aspects ofquality oflife that are affected by skin disease. These include the Dermatology Life Quality Index (DLQI) and the Skindex .P?'
Peripheral Disease/Enthesitis/Dactylitis Although there are no widely validated and accepted measures to assess peripheral joint in volvement in PsA, several instruments have used in recent RCT and these were able to distinguish active treatment from placebo. 52 These measures include the American College ofRheumatology (ACR) response criteria developed for RA,53 the Psoriatic Arthritis Response Criteria (PsARC) developed by Clegg for the sulfasalazine study in PSA54 and the Disease Activity Score (DAS) response criteria, also developed for RA .55 All these measures include assessment of tender joint
12
Molecular Mechan isms ofSpondyloarthropathies
count (TJC), swollen joint count (SJC) , as well as patient and physician assessments of global disease activity. The PsARC was specifically created, although not validated, for a study by Clegg et al and measures tender joint scores (TJS), swollen joint scores (SJS), physician global assessment of disease activity (0-5 point Likert scale) and patient global assessment (0-5 point scale). A response in the joint counts is determined by a reduction of .,,30% in TJS and/or SJS and reduction of at least one in Likert global assessment scores . An overall response is indicated by improvement in two ofthe four items , one ofwhich must be a joint count, without worsening in any of the four remaining items . For some RCTs in PsA studies, the assessed joint counts were increased to 78 in order to include the distal interphalangeal joints (DIP) of the feet. A study compared responsiveness and discriminative capacity ofACR criteria, PsARC, DAS and DAS28 in two Phase-2 PsA trials, one with etanercept and the other with inflixirnab/" When retrospectively applied, DAS28 performed better in discriminating active drug from placebo than the ACR20 improvement criteria, which in turn performed better than PsARC. It was also found that the pooled indices (DAS and DAS28) were generally more responsive and more discriminative than the Single core-set measures. Discussions at OMERACT and GRAPPA meetings recommended that a 68-joint count is used, as it includes the majority of joints affected by PsA. Because concerns existed on possible exclusion ofpatients in ReT due to adaptation ofa lower joint count, it was decided not to include the distal joints of the feet (78 tender joint count) as it may be difficult to distinguish DIP from Proximal interphalangeal (PIP) joint inflammation in the toes. It has been suggested that ifeither the PIP or the DIP of the toe is involved , it will be indicated as a PIP.45 Several instruments exists to assess enthesitis in PsA that have been used in AS. These instruments partially meet the OMERACT filter : the Mander, MASES, Berlin and SPARCC enthesitis indices. The Spondyloarthritis Research Consortium of Canada (SPARCC) study' ! in patients with PsA used another modified index that applied only to eight sites and reliability was variable. A simple enthesitis measure evaluating only the Achilles tendon and plantar fascia was found to be useful and able to discriminate between effective therapy and placebo in the IMPACT and IMPACT2 trials. 57.58 Dactylitis is characterized by swelling ofa whole digit and represents a combination ofsynovitis and inflammation oftendon and ligament insertion. Dactylitis appears to have prognostic significance, as it is associated with more aggressive disease in affected digits. 59 A number of measures exist, ranging from simple counting to more advanced measurements ofthe circumference of the digits. Recently all available measures were compared/" The recently developed objective measure , the Leeds Dactylitis Index , (LDI) was thought to be the best option for clinical trials since it provides the best approximation to 'truth'. The LDI measures the ratio of the circumference of the affected digit to the circumference of the digit on the opposite hand or foot: a minimum difference of 10% is used to define a dacrylitic digit. lfipsilateral and contralateral digits are thought to be involved, a table of normative values is provided to compare. The ratio of circumference is multiplied by a tenderness score, originally based on the Ritchie index (graded 0-3) , but later modification amended this to a binary score (0 for nontender, 1 for tender- this later modification is referred to as the LDI basic). The results ofeach digit are then summed to produce a final score.v The LDI reliability is good, with inter- and intraobserver intraclass correlation coefficient (ICC) reported of 0.9 and 0.84 , respectively. However, the LDI takes roughly I minute to assess three paired digits and also requires an additional tool to measure digital circumference which can intervene with the feasibility.
Nail Involvement Nail changes can be observed in up to 55% of patients with psoriasis and nail psoriasis is estimated to affect 80-90% ofpatients with psoriasis at some point in their lives.62 There is a strong association between the duration ofskin lesions and nail psoriasis and nail involvement has been reported to restrict patients' acrivitles.P Recently, the Nail Psoriasis Severity Index (NAPSI) was developed, which assesses the extent ofthe involvement ofthe psoriatic nail unit.64
Clinical Assessment in the Spondyloarthropathies
13
This measure incorporates all aspects of psoriatic nail disease and has recently been used in several studies as an outcome measure . To grade the nails according to the NAPSI, the nail was divided into quadrants by imaginary horizontal and longitudinal lines. Each quadrant ofthe nail was evaluated for nail-matrix disease (pitting, leukonychia, red spots in the lunula, crumbling) and for nail-bed disease (onycholysis, splinter haemorrhages, subungual hyperkeratosis, salmon-patch dyschromia) and scored 0 for the absence and 1 for the presence of any of these nail-matrix or nail-bed findings. The total nail score (matrix + bed) for all quadrants ofthe 20 nails ofa patient (0-160) is referred to as the total NAPSI score. The interobserver reliability of the total NAPSI score is observed to be 0.78 , which indicates a moderate to good agreemene.f Recently, a modified target NAPSI for target nail assessment was proposed by Parrish et al.66They suggested giving each parameter a degree ofgradation from 0 to 3 (0 - none, 1 = slight, 2 = moderate, 3 = severe), as for the PASI, in order to obtain a more sensitive system for assessing nail changes in response to therapy. The rnNAPSI scores had excellent interobserver reliability (Cronbach's alpha 0.98). and may be a valuable cool."
Fatigue Also in PsA, fatigue is an important system domain. The Functional Assessment of Chronic Illness Therapy-Fatigue Scale (FACIT) has been used in a RCT ofadalimumab in PsA (ADEPT) and demonstrates good responslvenes s/" A recent study shows the FACIT fatigue results are reproducible (ICC 0.95) and correlate with other fatigue measures as well as with disease activity in patients with PSA.69 Therefore the FACIT fatigue is a reliable and valid instrument to measure fatigue in PsA. Spine Assessments The frequency of spinal involvement in PsA has varied between 20% and 70% , depending on the features used to define disease." The instruments to assessaxial disease in AS, are also used inPsA.
Functioning Physical Function Measures ofphysical function, including the HAQdisease Index and the SF-36 physical function subscale, have been validated in PsA.71 Both these measures improve significantly in the context of anti -TNF studies to discriminate well between placebo and effective treatment, although the SF-36 physical function scale appears to show better sensitivity to change.58.7 !
Quality ofLife HRQOL is an important outcome measure to assess the quality of life. Physical function is also considered to be important and both the HAQ-DI and the physical function component of the SF-36 are suitable to measure physical function in PSA.45 The SF-36 questionnaire is reliable and valid for use in PsA, supporting its use as an adjunct outcome mea sure for clinical trials in PSA.72 Recently, a disease-specific patient-derived measure for use in PsA is developed (PsAQol)73 and should now been tested in RCT.
StructuralDamage Radiographic features thought to be specific to PsA include asymmetrical sacroiliitis, nonmarginal syndesmophytes, asymmetrical syndesmophytes, paravertebral ossification and more frequent involvement of cervical spine compared to AS. 74•7 5The same methods used to quantify sacroiliitis and spondylitis in AS are used in PsA. Whether they are valid in PsA is not clear, since they have not been tested.
14
Molecular Mechanisms ofSpondyloarthropathies
References 1. Dougados M. van der Linden S, ]uhlin R et at. The European Spondylarrhropathy Study Group preliminary criteria for the classification of spondylarthropathy, Arthritis Rheum 1991 : 34(10):1218-1227. 2. Helliwell PS, Taylor WJ. Classification and diagnostic criteria for psoriatic arthritis. Ann Rheum Dis 2005: 64(Suppl 2):ii3-8. 3. van der Linden S, Valkenburg HA, Cats A. Evaluation of diagnostic criteria for ankylosing spondylitis. A proposal for modification of the New York criteria. Arthritis Rheum 1984: 27(4):361-368. 4. Boers M. Brooks P. Strand CV er al. The OMERACT filter for outcome measures in rheumatology. J Rheumarol 1998 ; 25(2):198-199. 5. Kavanaugh A, Cassell S. Outcome measures in psoriatic arthritis. Curr Rheumatol Rep 2005: 7(3):195-200. 6. Rigby AS, Silman AJ. Outcome assessment in clinical trials of ankylosing spondylitis. Br] Rheumatol 1991; 30(5):321-322. 7. van der Heijde D, Bellamy N, Calin A et al. Preliminary core sets for endpoints in ankylosing spondylitis. Assessments in ankylosing spondylitis working group.] Rheumatol 1997 : 24(11) :2225-2229. 8. van der Heijde D. Calin A, Dougados M et al. Selection of instruments in the core set for DC-ART. SMARD. physical therapy and clinical record keeping in ankylosing spondylitis. Progress report of the ASAS working group. Assessments in ankylosing spondylitis.] Rheumarol 1999: 26(4):951-954. 9. Rudwaleit M , Metter A, Listing] et al. Inflammatory back pain in ankylosing spondylitis: a reassessment of the clinical history for application as classification and diagnostic criteria. Arthritis Rheum 2006 : 54(2):569-578. 10. Balint PV, Kane D, Wilson H et al. Ultrasonography of entheseal insertions in the lower limb in spondyloarthropathy, Ann Rheum Dis 2002: 61(10):905-910. 11. Mander M , Simpson ]M. McLellan A et al. Studies with an enthesis index as a method of clinical assessment in ankylosing spondylitis. Ann Rheum Dis 1987: 46(3):197-202. 12. Braun]. Brandt J, Listing] er aI. Treatment of active ankylosing spondylitis with infliximab : a randomised controlled multicentre trial. Lancet 2002 : 359(9313):1187-1193. 13. Gladman DD. Cook RJ. Schentag C et al. The clinical assessment of patients with psoriatic arthritis: results of a reliability study of the spondyloarthritis research consortium of Canada.] Rheumatol 2004; 31(6) :1126-1131. 14. www.asas-group.org. 15. Maksymowych WP, Mallon C, Richardson R et al. Development and validation of the edmonton ankylosing spondylitis metrology index . Arthritis Rheum 2006: 55(4) :575 -582. 16. Viitanen j V, Kautiainen H . Suni] er al. The relative value of spinal and thoracic mobility measurements in ankylosing spondylitis. Scand] Rheumato] 1995: 24(2):94-97. 17. Heufi-Dorenbosch L. Vosse D. Landewe Ret al. Measurement of spinal mobility in ankylosing spondylitis : comparison of occiput-to-wall and tragus-to-wall distance.] Rheurnatol 2004: 31(9) :1779-1784. 18. Schober P. The lumbar vertebral column and backache. Munch Med Wschr 1937: 84:336. 19. Macrae IF. Wright V. Measurement of back movement. Ann Rheum Dis 1969: 28(6) :584-589. 20. Moll ]M. Wright V. Normal range of spinal mobility. An objective clinical study. Ann Rheum Dis 1971; 30(4):381-386. 21. Moll ]M, Wright V. The pattern of chest and spinal mobility in ankylosing spondylitis. An objective clinical study of 106 patients. Rheumatol Rehabil 1973: 12(3):115-134. 22. Viitanen ]Y. Heikkila S, Kokko ML et al. Clinical assessment of spinal mobility measurements in ankylosing spondylitis: a compact set for follow-up and trials? Clin Rheumatol 2000 ; 19(2) :131-137. 23 . Heikkila S. Viitanen jV, Kautiainen H et al. Sensitivity to change of mobility tests: effect of short term intensive physiotherapy and exercise on spinal. hip and shoulder measurements in spondyloarthropathy, J Rheumatol 2000; 27(5) :1251 -1256. 24 . van der Heijde D, Dijkmans B. Geusens P et al. Efficacy and safety of infliximab in patients with ankylosing spondylitis: results of a randomized. placebo-controlled trial (ASSERT). Arthritis Rheum 2005: 52(2) :582-591. 25. Auleley GR, Benbouazza K, Spoorenberg A et al. Evaluation of the smallest detectable difference in outcome or process variables in ankylosing spondylitis. Arthritis Rheum 2002: 47(6):582-587. 26. Jauregui E. Conner-Spady B, Russell AS et al. Clinimetric evaluation of the bath ankylosing spondylitis metrology index in a controlled trial of pamidronate rherapy, ] Rheumatol 2004: 31(12):2422-2428. 27 . Brandt]. Listing]. Sieper] et al. Development and preselection of criteria for short term improvement afier anti-TNF alpha treatment in ankylosing spondylitis. Ann Rheum Dis 2004: 63(1l):143ll-1444. 28. Kruithof E, Van den Bosch F, Baeten D et al. Repeated infusions of inflixirnab , a chimeric anti -TNFalpha monoclonal antibody. in patients with active spondyloarthropathy: one year follow up. Ann Rheum Dis 2002; 61(3) :207-212.
ClinicalAssessment in the Spondyloarthropathies
15
29. Breban M, Vignon E, Claudepierre p et al. Efficacy of infliximab in refractory ankylosing spondylitis : results of a six-month open-label study. Rheumatology (Oxford) 2002; 41(11) :1280-1285 . 30. Calin A, Garrett S, Whitdock H er al. A new approach to defining functional ability in ankylosing spondylitis: the development of the bath ankylosing spondylitis functional index. J Rheumatol 1994; 21(12) :2281-2285. 31. Dougados M, Gueguen A, Nakache JP et al. Evaluation of a functional index and an articular index in ankylosing spondylitis. J Rheurnatol 1988; 15(2):302-307. 32. Ruof J, Stucki G. Comparison of the dougados functional index and the bath ankylosing spondylitis functional index. A literature review.J Rheumatol1999; 26(4):955-960. 33. Hdliwdl PS, Marzo-Ortega H, Tennant A. Comparison of a disease-specific and a generic instrument for measuring health-related quality of life in ankylosingspondylitis. Arthtitis Rheum 2002; 46(11) :3098. 34. Doward LC, Spoorenberg A, Cook SA ec al. Development of the ASQoL : a quality of life instrument specific to ankylosing spondylitis . Ann Rheum Dis 2003; 62(1) :20-26. 35. Wanders AJ, Landewe RB, Spoorenberg A et al. What is the most appropriate radiologic scoring method for ankylosing spondylitis? A comparison of the available methods based on the outcome measures in rheumatology clinical trials filrer, Arthritis Rheum 2004; 50(8):2622-2632. 36. Garrett S, Jenkinson T, Kennedy LG et al. A new approach to defining disease status in ankylosing spondylitis: the bath ankylosing spondylitis disease activity index. J Rheumarol 1994; 21(12) :2286-2291. 37. Calin A, Nakache JP, Gueguen A et al. Defining disease activity in ankylosing spondylitis: Is a combination of variables (bath ankylosing spondylitis disease activity index) an appropriate instrument? Rheumatolog y (Oxford) 1999,38(9):878-882. 38. Braun J, Pham T, Sieper J et al. International ASAS consensus statement for the use of anti-tumour necrosis factor agents in patients with ankylosing spondylitis. Ann Rheum Dis 2003; 62(9):817-824. 39. Anderson JJ, Baron G, van der Heijde D et al. Ankylosing spondylitis assessment group preliminary definition of short-term improvement in ankylosing spondylitis. Arthritis Rheum 2001, 44(8) :1876-1886. 40. Dougados M, Behier JM, Jolchine I et al. Efficacy of celecoxib, a cyclooxygenase 2-specific inhibitor, in the treatment of ankylosing spondylitis: a six-week controlled study with comparison against placebo and against a conventional nonsteroidal antiinflammatory drug. Arthritis Rheum 2001; 44(1) :180-185. 41. Dougados M, Nguyen M, Caporal R ec al. Ximoprofen in ankylosing spondylitis. A double blind placebo controlled dose ranging study. Scand J Rheumatol 1994; 23(5) :243-248 . 42. Dougados M, Caporal R, Doury P er al. A double blind crossoverplacebo controlled trial of ximoprofen in as. J Rheumarol 1989; 16(8):1167-1169. 43. Dougados M, Gueguen A, Nakache JP er al. Ankylosing spondylitis: What is the optimum duration of a clinical study? A one year versus a 6 weeks nonsteroidal anti-inflammatory drug trial. Rheumatology (Oxford) 1999, 38(3) :235-244. 44. Braun J, DavisJ, Dougados M et al. First update of the international ASAS consensus statement for the use of ant i-TNF agents in patients with ankylosing spondylitis. Ann Rheum Dis 2006 ; 65(3):316-320. 45. Gladman DD, Mease PJ, Strand V er al. Consensus on a core set of domains for psoriatic arthritis. J Rheumatol2007; 34(5):1167-1170. 46. Fredriksson T, Petrersson U. Severe psoriasis- Oral therapy with a new retinoid . Dermatolog ica 1978; 157(4):238-244. 47. Ashcroft DM , Wan Po AL, Williams HC er al. Clinical measures of disease severity and outcome in psoriasis: a critical appraisal of their quality. Br J Derrnatol 1999; 141(2) :185-191. 48. Anderson RT, Aaronson NK, Wilkin D. Critical reviewof the international assessmentsof health-related quality of life. Qual Life Res 1993; 2(6):369-395. 49. Brazier J, Jones N, Kind P. Testing the validity of the Euroqol and comparing it with the SF-36 health survey questionnaire. Qual Life Res 1993; 2(3):169-180. SO. Finlay AY, Khan GK. Dermatology Life Quality Index (DLQI)-A simple practical measure for routine clinical use. Clin Exp Dermatol1994, 19(3):210-216. 51. Chren MM, Lasek R], FloekeSA et al. Improveddiscriminative and evaluative capabilityof a refinedversionof skindex, a quality-of..lifc instrument for patients with skin diseases. Arch Dermatoll997; 133(11):1433-1440. 52. Mease PJ, Antoni CE, Gladman DD er al. Psoriatic arthritis assessment tools in clinical trials. Ann Rheum Dis 2005 , 64(Suppl 2):ii49-54. 53. Fdson DT, Anderson JJ, Boers M et al. American college of Rheumatology. Preliminary definition of improvement in rheumatoid arthritis. Arthritis Rheum 1995; 38(6) :727-735. 54. Clegg DO, Reda DJ, Mejias E et al. Comparison of sulfasalazine and placebo in the treatment of psoriatic arthritis. A department of veterans affairscooperativestudy. Arthritis Rheum 1996; 39(12):2013-2020. 55. van Gestel AM, Prevoo ML , van't Hof MA et al. Development and validation of the european league against rheumatism response criteria for rheumatoid arthritis. Comparison with the prellminary American college of rheumatology and the world health organization/international league against rheumatism criteria. Arthritis Rheum 1996,39(1) :34-40.
16
Molecular Mechanisms ofSpondyloarthropathies
56. Fransen ). Antoni C. Mease P) ec al. Performance of response criteria for assessing peripheral arthritis in patients with psoriatic arthritis: analysis of data from randomised controlled trials of two tumour necrosis factor inhibitors. Ann Rheum Dis 2006 : 65(10) :1373-1378. 57. Antoni CEo Kavanaugh A. Kirkham B et al. Sustained benefits of Infliximab therapy for derrnatologic and articular manifestations of psoriatic arthritis: results from the infliximab multinational psoriatic arthritis controlled trial (IMPACT). Arthritis Rheum 2005 : 52(4) :1227-1236. 58. Antoni C. Krueger GG. de Vlam K et al. Infliximab improves signs and symptoms of psoriatic arthritis: results of the IMPACT 2 trial. Ann Rheum Dis 2005 ; 64(8):1150-1157. 59. Brockbank jfi, Stein M. Schenrag CT et al. Dactylitis in psoriatic arthritis: a marker for disease severity? Ann Rheum Dis 2005: 64(2):188-190 . 60. Healy Pl. Helliwell PS. Measuring dactylitis in clinical trials: Which is the best instrument to use? ) Rheumatol 2007: 34(6) :1302-1306. 61. Helliwell PS. Firth). Ibrahim GH et al. Development of an assessment tool for dactylitis in patients with psoriatic arthritis.} Rheumatol2005: 32(9):1745-1750. 62. de Berker D. Management of nail psoriasis. Clin Exp Dermacol 2000: 25(5):357-362. 63. de long EM. Seegers BA. Gulinck MK et al. Psoriasis of the nails associated with disability in a large number of patients : results of a recent interview with 1.728 patients. Dermatology 1996: 193(4):300-303. 64. Rich P. Scher RK . Nail psoriasis severity index: A useful tool for evaluation of nail psoriasis.) Am Acad Derrnarol 2003; 49(2) :206-212. 65. Aktan S, Ilknur T. Akin C et al. Interobserver reliability of the nail psoriasis severity index. Clin Exp Dermarol 2007: 32(2) :141-144. 66. Parrish CA. Sobera )0. Elewski BE. Modification of the nail psoriasis severity index.) Am Acad Dermatol 2005 ; 53(4):745-746: author reply 746-747. 67. Cassell SE, Bieber )0. Rich P et al. The modified Nail Psoriasis Severity Index: validation of an instrument to assess psoriatic nail involvement in patients with psoriatic arthritis .) RheumatoI2007; 34(1):123-129. 68. Mease Pl. Gladman DO. Ritchlin CT et al. Adalimumab for the treatment of patients with moderately to severely active psoriatic arthritis: results of a double-blind, randomized, placebo-controlled trial. Arthritis Rheum 2005: 52(10) :3279-3289. 69. Chandran v, Bhella S, Schentag C et al. Functional assessment of chronic illness therapy-fatigue scale is valid in patients with psoriatic arthritis. Ann Rheum Dis 2007; 66(7) :936-939. 70. Gladman DD. Brubacher B. Buskila 0 et al. Differences in the expression of spondyloarthropathy: a comparison between ankylosing spondylitis and psoriatic arthritis. Clin Invest Med 1993; 16(1) :1-7. 71. Husred jA, Gladman DO. Cook RJ et al. Responsiveness of health status instruments to changes in articular status and perceived health in patients with psoriatic arthritis.) Rheumato11998: 25(11):2146-2155. 72. Husred IA, Gladman DO. Farewell VT et al. Validating the SF-36 health survey questionnaire in patients with psoriatic arthritis.) Rheumatol1997: 24(3) :511-517. 73. McKenna SP, Doward LC. Whalley 0 et al. Development of the PsAQoL: a quality of life instrument specific to psoriatic arthritis. Ann Rheum Dis 2004; 63(2):162-169. 74. McEwen C, DITata D. Lingg C et al. Ankylosingspondylitisand spondylitis accompanyingulcerative colitis.regional enteritis. psoriasis and Reiter's disease. A comparative study. Arthritis Rheum 1971; 14(3):291-318. 75. Helliwell PS, Hickling P. Wright V. Do the radiological changes of classic ankylosing spondylitis differ from the changes found in the spondylitis associated with inflammatory bowel disease, psoriasis and reactive arthritis? Ann Rheum Dis 1998; 57(3):135-140.
CHAPTER 2
Imaging in Spondyloarthritis Walter P. Maksymowych*
Abstract
T
he role ofimaging in the evaluation and management ofSpA has experienced a resurgence ofinterest with the introduction of MRI and more sophisticated sonographic technologies. Several approaches have been developed to score plain radiographic abnormalities in the spine and sacroiliac joints of patients with SpA and this approach remains the standard for assessment of structural damage. The modified Stoke AS Spinal Score (mSASSS) is the most responsive outcome instrument for scoring damage in the spine although responsiveness is limited and requires a minimum of 2 years before significant change becomes apparent in patients on standard therapies. Magnetic resonance imaging (MRI) is the most sensitive imaging abnormality for the detection ofinflammation and the advent of fat suppression imaging allows detection of bone marrow inflammation in the sacroiliac joints as one of the earliest abnormalities in AS. Spinal inflammation can now be reliably scored using MRI-based outcome instruments that are highly sensitive to change and this represents a major advance in the objective evaluation of new therapeutic interventions. Moreover, MRI now allows the detection ofpatients at an earlier stage of their disease course with the potential for new insights into the pathogenesis of disease. Ultrasound provides a more feasible and cost-effective approach to the assessment of peripheral inflammation, especially enthesitis.
Introduction Spondyloarthritis (SpA) is a group ofinflammatory disorders that primarily affect the sacroiliac joints and structures ofthe spine and are associated with the HLA-B27 gene. A hallmark ofdisease is the involvement of entheses. Large peripheral joints, especially of the lower limbs, may also be involved, particularly in juvenile boys. Imaging has been an indispensable tool for evaluation of these structures that has been available to clinicians for many decades. Recent advances have been driven by recognition that more information is required to evaluate active inflammation within soft tissues and bone and not merely the structural damage that is a consequence ofinflammation, to recognize and manage early disease, to assist the basic scientist in the examination ofimmunopathological events in early disease and to develop internationally validated outcome assessment tools to assist the clinical researcher in the evaluation of therapeutic interventions. Five principle methods are currently used to evaluate patients with SpA: plain radiography, computed tomography (CT), scintigraphy, ultrasound and magnetic resonance imaging (MRI). This chapter examines the advantages and limitations ofeach imaging modality, presents insights that each method has provided into the pathogenesis ofdisease, scrutinizes the potential ofeach method as an outcome assessment tool and evaluates each method for its ability to influence diagnostic and therapeutic decisions for the practicing clinician. 'Walter P. Maksymowych-Alberta Heritage Foundation for Medical Research, 562 Heritage Medical Research Building, University of Alberta, Edmonton, Alberta T6G 252, Canada. Email:
[email protected]
Molecular Mechanisms ofSpondyloarthropathies, edited by Carlos Lopez-Larrea and Roberto Diaz-Peiia. ©2009 Landes Bioscience and Springer Science+Business Media.
18
Molecular Mechanisms ofSpondyloarthropathies
Plain Radiography Plain radiography is still the cornerstone of diagnosis, evaluation and classification of SpA . It depicts abnormalities of cancellous or cortical bone and so will demonstrate focal or diffuse changes ofosteoporosis, sclerosis, erosion or ankylosis. In the sacroiliac joints, the earliest abnormality is a loss of distinctness of the subchondral bone followed by more obvious erosion and widening ofthe joint space. Progressive sclerosis and joint space ankylosis represent late features of disease. Changes in bone marrow or cartilage are indirectly visualized. For example, inflammation within bone marrow is implied when erosion ofthe calcified bony matrix is visible and cannot be detected in the absence of bone changes. Moreover, erosion is often quite severe in the sacroiliac joint before being detected due to the complex anatomy and overlapping structures. A number ofradiographic approaches have been described for evaluation of the sacroiliac joints although it appears that a Single anteroposterior pelvis radiograph is comparable to dedicated sacroiliac views. In 445 patients with SpA, oblique sacroiliac-joint views and the anteroposterior pelvic view did not differ significantly regarding sensitivity for diagnosing sacroiliitis.' The modified New York criteria for AS require the presence ofradiographic sacroiliitis in order to classify the patient as having AS and, although not intended for this purpose, are also widely used for diagnostic assignment." However, radiographic sacroiliitis is a relatively late finding in the course ofAS and several years may elapse from the onset oflnflamrnarory back pain (IBP) until the appearance of radiographic sacroiliitis. In a study of88 patients with clinical features compatible with early AS (IBP plus additional features such as peripheral arthritis, heel pain, acute uveitis, or elevated levels of acute-phase reactants) but with radiographically normal sacroiliac joints, only 36% had developed radiographic evidence of sacroiliitis after 5 years and 59% had done so after 10 years.l The small sample size and me lack ofaccepted diagnostic criteria for IBP limit the generalizability of this study. Moreover, the interpretation of sacroiliitis on pelvic radiographs is inherently difficult and subject to significant inter-individual variability which does not improve with rraining.t This may contribute to me long lag-time between the onset ofsymptoms and the diagnosis ofAS . Radiography ofthe 51] is therefore increasingly being replaced by more sensitive imaging modalities, such as CT and MRI (see below). Plain radiography can show a variety of features in the spine. In spinal vertebrae me earliest feature is the loss of the cortex at me corner of me vertebral body giving the appearance of an erosion. Bone remodelling and new bone formation lead to the radiographic appearance ofsquaring and sclerosis at me vertebral corner. Further new bone formation from me vertebral corner across me disc space to the adjacent veretebral corner or syndesmophyte may ultimately lead to complete ankylosis (Fig. 1). This occurs not only at me vertebral rim but also in me interior ofthe disc . Spondylodiscitis is radiographically evident as disruption and loss ofme vertebral endplare. Facet joint abnormalities consist oferosions, loss ofjoint space and ankylosis. They are not readily visible in me thoracic spine because ofoverlapping structures. Systematic prospective radiographic studies are limited but clinical observations indicate that abnormalities typically originate in the lumbar spine and ascend cranially although cervical lesions may be preferentially affected in some patients. Typical features occur relatively late in the course of the disease and are in general not contributory to the diagnosis. In a cross-sectional cohort ofpatients with AS and a mean disease duration of almost 12 years, more than 60% of patients had features attributable to AS on their spinal radiographs, but only a minority had syndesmophytes extending over multiple vertebrae.' A "bamboo spine", which reflects an end-stage ofspinal AS, was observed in less man 5%. Radiographic abnormalities are associated with impaired spinal mobility,"The relationship is not linear and the association increases with increasing level ofankylosis. The pace ofradiographic progression in the spine isvariable but until recently was assumed to be very slow requiring several years to demonstrate unequivocal change and therefore of limited use for outcome assessment. Prospective analysis ofradiographic changes and application ofsystematic approaches to the scoring ofthese changes have changed these insights. Three methods have been described in the literature to score abnormalities in the spine . The first is the Bath Ankylosing Spondylitis Radiology Index (BASRI)? This method is a global grading of the lateral cervical spine, the anterior and lateral
Imaging in Spondyloarthritis
19
Figure 1. Radiographic features of ankylosing spondylitis in the spine in well established disease.
lumbar spine combined and the SI]. Each site is scored from 0 (= normal) to 4 (= fusion involving at least 3 vertebrae). The sum of the sites gives the BASRis (spine) ranging from 0 to 12. The minimum score is 2 if it is assumed that patients fulfil the modified New York criteria and consequently have a score of2 for the SI]. The other two scoring methods are based on radiographic abnormalities in the vertebral corners. The Stoke Ankylosing Spondylitis Spinal Score (SASSS) assesses the anterior and posterior corners ofeach lumbar vertebra which are scored for the presence ofsquaring, sclerosis, erosions, syndesmophyres and bridging syndesmophyres.t Ihe maximal score is 72. In the modified SASSS the posterior corners ofthe lumbar vertebrae are not scored and the anterior corners of the cervical vertebrae are added to the scoring method giving rise to the same scoring range (Fig. 2). The developers of this method showed that changes could be detected in 36 of 57 patients over a period of48 weeks," A comparison of the various aspects of the validity of these three scoring methods showed that no method reliably detected change over 1 year and that only the mSASSS reliably detected change scores over a period of2 years. 10 Radiographic progression was observed in 43% and 41% of patients in the lumbar and cervical spine , respectively. Progression in either the cervical or lumbar spine was observed in almost 60% ofpatients over a 4-year period. A major disadvantage of the BASRI method is its susceptibility to ceiling effects as shown by the finding of23% and 15% of patients with a maximal BASRI score for cervical and lumbar spine as compared to only 5% and 4%, respectively, by the mSASSS. The requirement for an AP view ofthe lumbar spine also necessitates increased radiation The preference for the mSASSS was recently endorsed by international consensus and its ability to discriminate between treatment groups was recently demonstrated in
20
Molecular Mechanisms ofSpondyloarthropathies
0
Normal
0 1
Erosion
1
Sclerosis
1
Squaring
2
Obvious
2
Syndesmophyte
3
Total bony
3
Bridge
Figure 2. The modified Stoke Ankylosing Spondylitis Spinal Score (Msasss)measures structural changes in the anterior corners of the cervical and lumbar vertebrae. At each vertebral corner the following scores are possible : sclerosis = 1, erosion = 1, squaring = 1, syndesmophyte - 2. At each disc space complete ankylosis scores a 6. The total scoring range is 0-72.
a trial comparing two strategies ofnonsteroidal anti-inflammatory drugs (NSAIDs); cont inuous versus on -demand use ofNSAIDs.ll Radiographic progression in the continuous NSAID group was statistically significantly lower as compared to the on-demand group. It has also been used to show that the serological biomarker, metalloproteinase 3, is a prognostic factor for further progression ofradiographic changes in the presence ofpre -existing spinal abnormalities. Two-thirds ofpatients in the highest category of risk had radiographic progression detected by the mSASSS with some patients having progression ofover 10 mSASSS units by 2 years.'? Limitations to the mSASSS are both conceptual and methodological. This method records both destructive changes i.e., erosions and reparative changes i.e., sclerosis, syndesmophytes and ankylosis. Erosive changes are uncommon «3% of patients) and could be excluded from assessment. Radiographic progression over 2 yearshas been measured without the blinding to time point that constitutes the standard approach in clinical trials. Adoption ofthis approach results in lower
Imaging in Spondyloarthritis
21
scores for radiographic progression thereby reducing the discriminative power ofthis instrument over 2 years.'? Moreover, training does not appear to improve reliability of change scores over 2 years, principally because ofsubstantial measurement error in the context ofminor progression. It is unlikely that the mSASSS will be used to assess diseasemodifying therapies in SpA because radiographic change detected by this method is too long for the time frame ofclinical trials. However, radiographic abnormalities, especially the presence of syndesmophytes, are of prognostic value. In one prospective study, definite radiographic progression wasfound in 44% ofthe patients with syndesmophyteslankylosis at baseline versus 19% ofthe patients without such changes." Radiographic changes in peripheral SpA occur primarily in the hip joint and at entheses. Concentric narrowing of the hip joint occurs in about 10-15% of patients, particularly those with juvenile-onset SpA. Joint space narrowing can also occur in other lower limb joints. Tarsal involvement with ankylosis is a particularly severe manifestation of juvenile SpA observed in association with enthesitis. The latter is typically manifest as a loss of bone cortex followed by more obvious erosions and then reparation with new bone formation. At the Achilles tendon insertion, erosive changes are more typically observed at the posterior calcaneum superior to the insertion of the tendon. The frequency of involvement by plain radiography varies according to disease duration. At the shoulder, erosions are characteristically observed at the insertion of the supraspinatus. Radiographic changes in the small joints are primarily confined to psoriatic SpA and then include both destructive erosive changes ofthe joint margins as well as reparative changes such as periosteal new bone formation.
Computed Tomography As for plain radiography. observations with CT are restricted to abnormalities ofcancellous or conical bone. However. for the sacroiliac joints CT is easier to interpret. overlapping structures are not a problem and erosion is detectable at an earlier stage. CT has been reported to show greater sensitivity and specificity in the diagno sis of sacroiliitis than plain radiography although these studies included small numbers ofparients.' >'?In a retrospective study of910 patients with back pain where the sacroiliac joints were examined by radiography and CT over a 2 year period. agreement between these methods was only fair but 41% ofradiographic reports provided a false answer," Moreover. definitive diagnosis ofsacroiliitis was over twice as frequently reported with CT (25%) as with plain radiography (11%). Two studies which compared standard radiography. CT and MRI in the same populations found that MRI was more sensitive for detecting sacroiliitis than the other two merhods.P:" The primary limitation of pelvic CT is the higher dose ofradiation . especially in males. although this may be less using a high resolution protocol ofdiscontiguous slices. One group has proposed the use of3 representative 3-mm slices through the SI joints instead ofthe full series which corresponded to halfthe radiation dose ofplain radiography and a 20-fold reduction compared to the full CT series.22 The use of CT in routine practice appears to be limited to the diagnostic evaluation ofinflammatory back pain where plain radiography shows normal sacroiliac joints and access to MRI is limited.
Isotopic Imaging Scintigraphy relies upon abnormal uptake ofradiotracer in areas ofincreased bone turnover. This test isoften positive earlier than x-raybut depends on diseaseactivity. In addition. the complex shape ofthe joint makes interpretation ofplanar images difficult. This is less ofa problem when SPECT images are acquired. Once again , marrow and cartilage are not directly visualised. Quantitative scanning of the sacroiliac joints has been evaluated as a diagnostic tool although cut-offs that reliably distinguish patients from controls have proved difficult to establish.P:" Sensitivity has ranged from 29% to 40% while specificity was usually less than 80%. One prospective study compared quantitative radionuclide scanning with MRI of the sacroiliac joints in patients with inflammatory back pain but normal plain radiographs." Of21 patients. 20 had MRI evidence of sacroiliac inflammation, whereas only 10 had radionuclide scan evidence of sacroiliitis. A recent
22
Molecular Mechanisms ofSpondyloarthropathies
systematic review concluded that isotopic imaging ofthe sacroiliac joints had low diagnostic utility for evaluation ofinflammatory back pain ." Isotopic imaging has been used to evaluate peripheral enthesitis and shown to demonstrate comparable sensitivity to MRI (94%) for detecting symptomatic calcaneal enthesitis. 3o•31
Ultrasound Ultrasound is an inexpensive.widely available, noninvasive and nonradiating technique that can be readily adopted by experienced rheurnarologists. Its primary value has been established in the assessment ofperipheral SpA. especiallyenthesitis. Several studies have shown that ultrasonography may even detect subclinical enthesopaehy.P'" Abnormalities in entheseal vascular supply by power Doppler are of particular diagnostic value. A recent extensive evaluation of2952 entheses from 164 patients with SpA (AS = 104, undifferentiated SpA = 30 . psoriatic arthritis = 21. inflammatory bowel disease arthritis = 6. reactive arthritis = 3),64 controls (34 with mechanical back pain) and 30 with rheumatoid arthritis using power Doppler ultrasound assessed the prevalence . specificity and severity ofenthesirls." Almost all patients (98%) with SpA had enthesitis and 38% ofall entheses were abnormal on ultrasound imaging. The commonest sites were the Achilles (79%) and plantaris fascia (74%) insertion into the calcaneum. patellar tendon insertion into the inferior pole ofthe patella (59%). the greater trochanter (44%). the quadriceps insertion into the patella (28%) and the medial (25%) and lateral (24%) epicondyles. The distribution ofaffected entheses did not vary amongst SpA subtypes or whether inflammation was predominantly axial or peripheral. Clinical examination ofa subset of34 SpA patients revealed a lower frequency ofenthesitis (14.4%) though a similar distribution ofaffected entheses. A sonographic enthesitis index that scores enthesitis in the quadriceps tendon insertion into the superior pole ofthe patella, the proximal and distal insertions ofthe patellar tendon. the Achilles tendon en thesis and the plantar fascia enthesis has been proposed and scores signs ofactive inflammation (hypoechogenicity, increased tendon thickness, peritendinous oedema, bursitis) and chronic changes (insertional bone erosions, intratendinous calcifications, decreased thickness and tears)." Inter-observer reliability was modest and no correlation with clinical measures ofdisease activity or severity was apparent. Although this approach is considered subjective in performance and evaluation, limited studies suggest that ultrasound may be superior to MRI in detecting the early signs ofenthesitis.36-38 Preliminary studies have also used color and duplex Doppler ultrasound (CDDUS) to detect arterial vascularity within and around sacroiliac joints and paraspinal areas in patients with AS. Arterial vascularity as measured by the resistive index was significantly greater in patients with clinically defined active inflammation and decreased in response to anti-TNF therapy," Another report compared MRI with microbubble contrast-enhanced color Doppler ultrasound of the sacroiliac joints in 103 patients with inflammatory back pain, 75% ofwhom turned out to have some form of spondyloarthropathy and showed that this technique had high negative predictive value for detection ofactive sacroiliiris.t" However. results ofCDD US in patients with inactive AS have been similar to controls indicating that this approach is likely to be oflimited diagno stic value though it may be useful as an objective parameter ofdisease activity. particularly for the assessment ofresponse to therapeutic intervention. Further comparative study with MRI is necessary.
Magnetic Resonance Imaging A major advance in the field has been the development ofMRI for spondyloarthritis because ofits ability to permit earlier diagnosis , which until now has relied primarily on plain imaging. It is superior to plain radiography and CT through its ability to visualize soft tissue inflammation and inflammatory lesions within bone in 3 dimensions. It therefore provides objective features of active inflammation which may not be readily assessed by clinical or laboratory evaluation. particularly ifdisease is confined to the axial joints. Quantitative assessment by MRI ofsacroiliac joint and spinal inflammation has now become possible and shown to facilitate evaluation ofnew therapeutics in clinical trials. Despite this, many clinicians and radiologists remain unfamiliar with
Imaging in Spondyloarthritis
23
the spectrum of acute and chronic MRI features in SpA and appropriate indications for the use ofMRI in diagnostic evaluation. Unlike plain radiography and CT, MRI does not involve ionizing radiation but relies on the properties of protons when exposed to an external magnetic field. Hydrogen nuclei gyrate at frequencies in the range ofradiowaves when exposed to strong magnetic fields (0.5 to 1.S Tesla). Application of a radiofrequency pulse excites the nuclei and once the radiofrequency pulse is turned off. the protons return to a low energy state releasing energy detected as an MRI signal which is proportional to the number of excited protons in the sample. Soft tissues are therefore readily detected in relation to their content ofprotons. For example, T l-welghted images readily detect the signal from fat which appears bright, whilst fluid appears grey and bone appears dark. Fat may be replaced by other tissues such as bone, tumour, cyst, inflammatory and fibrous tissue resulting in loss ofthe T I fat signal. T2- weighted MRI sequences detect the signal from water as may be observed in cysts, inflammatory infiltrates and highly vascularized tissues such as tumours. With these sequences fluid is very bright, whilst fat is less bright and bone is black. The STIR (short tau inversion recovery) sequence is particularly useful in SpA where inflammatory lesions are characteristically present in subchondral bone and adjacent to entheses. With this sequence the signal from bone marrow fat issuppressed so that the marrow appears dark whilst fluid remains bright allowing visualization ofedema, cysts, tumours and inflammatory infiltration (Fig. 3). An alternative approach to the detection ofinflammation iswith contrast enhancement after intravenous
Figure 3. Semi-coronal MRI slice through the synovial portion of the right sacroiliac joint. A) TI spin echo sequence demonstrates loss of marrow fat signal in both iliac and sacral bones (arrows) that corresponds with increased signal on B) STIRsequence.
24
Molecular M echanisms ofSpondyloarthropathies
administration ofgadolinium (Gd). Gadolinium accumulates at sites ofincreased vascularity and capillary permeability where it alters the magnetic properties of surrounding soft tissues. This process can be followed over time where both the rate and degree of gadolinium augmentation provides quantitative information on the severity of inflammation. This is termed dynamic MR imaging. Gadolinium augmented MRI therefore offers the possibility of quantifying inflammation in terms of (I) Intensity ofaugmentation (peak), (2) Maximal rate ofaugmentation (slope) , (3). Time to peak and can demonstrate the distribution of inflammation in considerable detail. Delineation of a region of interest in a particular location on the MR scan allows comparison of gadolinium enhancement before and after therapeutic interventions.
Sacroiliac Joint Imaging of the 51] has to address the fact that the region of interest in the sacroiliac joint, the cartilaginous portion, is convex shaped with the apex facing antero-inferiorly. Consequently, images are obtained in the semicoronal orientation along the long axis of the sacral bone which allows visualization of this portion of the joint. Transaxial sections allow superior evaluation of ligamentous structures in the postero-superior portion ofthe joint and are orientated perpendicular to semi-coronal sections. On a Tl spin echo (SE) sequence bone marrow fat in the iliac and sacral bones appears bright whilst subchondral bone lining the joint cavity is dark. On STIR images of the normal sacroiliac joint the bone marrow is dark and bright signal is detected from presacral veins. An initial study should consist of a semicoronal TlSE and either a STIR or fat-saturated T2-weighted sequence . Questionable abnormalities can be further analyzed using fat-saturated Gd-augmented imaging. Limited histopathological studies in early AS provide evidence that the earliest inflammatory lesions are observed within the subchondral bone and synovium of the sacroiliac joint. At least 3 lesions are evident on MRI in patients with early sacroiliitis, namely capsulitis, synovitis and subchondral bone marrow inflammation. 41.42 Although subchondral bone marrow edema was noted in the first detailed description ofthe MR appearances ofsacroiliitis, the significance ofthis observation remained unclear and was largely ignored." However, subsequent studies using fat suppression techniques consistently noted subchondral bone marrow edema as an important component of the early pathology of sacroiliitis. In an attempt to further enhance the sensitivity ofMRI in the detection ofearly sacroiliitis, dynamic MR imaging with gadolinium augmentation was proposed." Examination ofpatients clinically characterized as having inflammatory low back pain but normal plain film radiographs revealed augmentation of the joint capsule, synovium and subchondral bone marrow in a significant proportion ofpatients especially in the postero-caudal portion ofthe synovial part ofthe joint where the capsule attaches to bone. Validation of these early MRI abnormalities in the sacroiliac joints according to clinical and pathological features ofSpA has been limited. One group quantified gadolinium augmentation in the sacroiliac joints by measuring maximal augmentation compared to baseline values and showed that a grading scheme for severity of inflammation correlated with disease activity as determined by clinical pararnerers." Clinical improvement following CT guided intra articular injections with corticosteroid was significantly associated with decreased gadolinium augmentation although it was not specifically stated whether decreased augmentation was noted in both subchondral marrow as well as synovium. Direct CT-guided biopsy of the sacroiliac joints has also demonstrated significant correlations between the degree ofgadolinium enhancement on MRI ofthe sacroiliac joints and the histopathological grade ofinflammanon." Contrast enhanced MRI isexpensive and time consuming and it is not clear whether it offers Significant advantages over T2 weighted fat suppression imaging ofthe sacroiliacjoints in patients with early SpA. One additional observation is relevant to the validation ofsubchondral marrow edema on MRI as a specific component in the pathology ofearly sacroiliitis. Capsulitis and subchondral bone marrow edema involving the iliac bone has been described in 10% ofHLA B27 positive juveniles with synovitis ofthe hips and/or knees despite the absence of back symptoms." There has been very limited study ofthe sensitivity and specificity ofMRI abnormalities ofthe sacroiliac joint in patients with early SpA. This has been problematic because the gold standard
Imaging in Spondyloarthritis
25
for diagnosis requires plain radiographic features ofsacroiliitis which means that such evaluations must be prospective to allow time for radiographic abnormalities to appear. To date , sensitivity has been examined in limited cross-sectional studies at single sites using gadolinium enhanced MRI. One group used dynamic imaging to compare patients with inflammatory back pain (n = 36) according to the European Spondyloarthropathy Study Group (ESSG) criteria, though normal pelvic x-rays,and patients with established AS (n = 36) according to the modified New York criteria, with those who had mechanical causes for low back pain (n = 53).48A positive MRI for sacroiliitis was defined on the basis of maximal enhancement after administration ofgadolinium. On this basis, MRI had sensitivity of83% and specificity of93% for clinically defined inflammatory back pain. In a second study that examined 48 patients with equivocal changes on plain x-ray and clinical features of AS with a mean duration of IBP of 0.8 years, MRI evidence of inflammation using gadolinium enhancement was evident in 76% of patients but specificity was not addressed since no controls were included." A third study reported a sensitivity of95% for MRI using both fat suppression and gadolinium enhancement techniques in 41 patients with inflammatory back pain with median duration ofsymptoms of I 9 months." In a fourth study that described a cohort of68 patients with inflammatory back pain (57 and 14 fulfilled ESSG and modified New York criteria, respectively) ofless than two years duration , inflammation in the SI joints could be detected using fat-suppressed MRI in only about one third of the patients (22/68) .49Ten of these patients also showed no abnormalities on plain rad iography of the sacroiliac joints while virtually all patients with abnormal pelvic x-ray also had abnormalities on MRI. MRI was comparable to plain radiography in the detection ofstructural changes . Another important finding in this stu dy was that it is probably sufficient to look for subchondral bone marrow oedema alone as a sign ofinflammation. The contribution of other sites of the joint to a diagnosis of sacroiliitis was only marginal. Only one patient had inflammation that was restricted to joint capsule and ligaments. The limitations posed by the lack of a suitable diagnostic gold standard are highlighted in additional findings from this study. The percentage ofpatients with MRI abnormalities ofthe sacroiliac joints rose to 58% in those patients who met all 3 clinical criteria sets for AS (Amor, ESSG , Berlin), to 73% in those carrying HLA B27 and to 91 % in those defined as having a good symptomatic response to NSAID.lt is important to point out that inflammatory back pain symptoms are very common in patients with mechanical causes of back pain so that selection of patients for MRI studies based largely on clinical criteria will inevitably result in varying estimates for sensitivity and specificity. The balance of the data indicates that MRI of the sacroiliac joints is most useful diagnostically in those patients with a clinical suspicion of inflammatory back pain who are HLA B27 positive with a good symptomatic response to NSAID and have a normal pelvic x-ray. Longitudinal studies with plain radiographic end points are limited in both numbers of patients studied and duration offollow up. One study evaluated 17 patients with inflammatory back pain of 3-14 months duration but normal pelvic x-ray who were followed for 1Y2 to 2 Y2 years." Virtually all patients had abnormalities on baseline MRI that persisted on follow up MRI 2-30 months later and 11 developed plain radiographic features of sacroiliitis on follow up, suggesting that inflammation observed on MRI is of prognostic significance. Unfortunately, there was insufficient data presented to determine if the addition of B27 testing to clinical evaluation was of similar predictive value . In a second prospective study of25 consecutive HLA-B27 positive patients with inflammatory low back pain and:s grade 2 unilateral sacroiliitis, the positive predictive value of:i!: grade 2 sacroiliitis on baseline MRI for the development of a grade 2 sacroiliitis on plain radiograph after 3 years was 60% (sensitivity 85%, specificity 47%). 28 In contrast to the previous study, gadolinium augmentation was not done, subchondral marrow edema was found in only 20 sacroiliac joints and median duration of symptoms was much longer (4 years) raising questions as to the reliability of the clinical diagnosis. As the lesion evolves, there is erosion ofthe subchondral bony plate followed by bone sclerosis and ultimately disappearance ofthe high marrow signal on fat-suppressed images. Erosions do not develop marginally, as typically seen in peripheral arthritis, but within the joint space and are visible as loss ofsubarticular bone with decreased fat signal on T l-weighted images and increased signal
26
Molecular Mechanisms ofSpondyloarthropathies
on fat-suppressed and Gd-augmented images reflecting invasive inflammatory tissue . Chronic changes are best visible on T l-weighted images and include increased peri-articular marrow signal due to fat replacement. decreased signal due to bone sclerosis and transarticular ankylosis (Fig. 4). Bone and sofi:tissue oedema at ligamentary insertions within the postero-superior extra-synovial portion ofthe joint has also been described and is best viewed on transaxial sections. Several studies have compared the performance ofdifferent imaging modalities for detection of abnormalities in patients diagnosed clinically with inflammatory back pain though with normal plain x-rays of the sacroiliac joims.IS.20.28Abnormalities have been detected in more patients using either Gd-augmented dynamic MRl or STIR when compared with CT and isotope bone scanning. Formal methods for quantifying disease activity in the sacroiliac joints on MRI have been proposed that can be generally categorized as being primarily based on either a global scheme that focuses on the single most severely affected semi-coronal image or on a more detailed method that scores several consecutive semicoronal images that depict the synovial portion of the sacroiliac joint (the Spondyloarthriris Research Consortium of Canada (SPARCC) scoring method)." The presence and extent ofbone marrow edema in the synovial portion of the joint is the primary MRI feature that is scored although one method also scores inflammation in the joint space and the ligamentous portion of the joint.i" In the SPARCC method. each sacroiliac joint is divided into quadrants and the presence of marrow edema in each quadrant is scored on a dichotomous basis with additional weighting for intensity and depthY Anatomical orientation and interpretation of abnormalities is greatly facilitated by simultaneously viewing T'l -weighted images whilst scoring edema on the STIR image s. The ASessments in AS International Working Group (ASAS) recently conducted a multi-reader exercise evaluating the different scoring methods for reliability and sensitivity to change according to the requirements of the OMERACT filter." Agreement between readers and sensitivity to change was somewhat better for the more detailed SPARCC scoring method. Only the SPARCC approach to scoring inflammation in the sacroiliac joints has been evaluated in a randomized placebo-controlled trial. Its high discriminatory capacity was confirmed in a placebo-controlled trial of adalimumab in AS where a sign ificant effect of active
Figure 4. A) T1SE and B) STIRMRI sequences of a semicoronal slice through the synovial portion of the sacroiliac joint. 1) Low T1 and low STIR signal = bone sclerosis . 2) Low T1 signal and increased STIR signal = active inflammation. 3) H igh T1 signal and low STIR signal = fat infiltration. 4) Low T1 signal with loss of subchondral bone and increased STIRSignal = erosion . S) High T1 signal and low STIR signal superiorly and increased STIR signal inferiorly = fat infiltration w ith inflammation inferiorly. 6) Low T1 signal and low STIR signal = bone sclerosis. 7) Low T1 signal and patchy increase in STIR signal = inflammation and bone sclerosis.
Imaging in Spondyloarthritis
27
treatment on sacroiliac joint inflammation visible on MRI was first demonstrated despite the presence of frequent chronic changes in patients with long-standing disease." One method has been reported for scoring chronic changes in the sacroiliac joints. It scores sclerosis and erosions at 8 sites in the iliac and sacral sides of the cartilaginous and ligamentous portions ofthe joints, respectively," It also scores joint space width. Inter-reader reliability ofMRI was only moderate. This imaging modality seems to offer few advantages over plain imaging or CT in [he evaluation of chronic changes in the sacroiliac joints.
Spine Images of the spine are typically obtained in the sagittal orientation with the spine being scanned in two segments, C 1-T 10 and T 10-S2 .This allows visualization ofthe entire spine within a reasonable time frame for patients required to lie still within the magnet but comes at the expense of a small field of view and relatively poor visualization of the cervical vertebrae. The same MRI sequences are obtained as for the SI]. STIR images can be readily distinguished from T l-weighred images by the presence of high signal emanating from cerebrospinal fluid in the spinal canal. Because there have been no sequential MRI studies ofthe spine in patients with early SpA the evolution ofabnormalities is unclear and it is unknown if or how commonly spinal inflammation may precede sacroiliitis. In established disease the most characteristic abnormality observed on fat-suppressed images is increased marrow signal at the anterior and posterior corners ofthe spinal vertebrae adjacent to the insertion ofthe annulus fibrosis which corresponds to the Romanus lesion observed histopathologically (Fig. 5). This then resolves through a process ofcartilage metaplasia and fat infiltration with loss ofthe STIR marrow signal and instead becomes evident as increased signal on T l-weighted images reflecting increased fat content (Fig. 6). Both inflammatory lesions as well as fat infiltration are not visible on plain radiography. Syndesmophytes arising from vertebral corners may be visualized on MRI although it is not known if these follow the development of acute lesions or whether they may arise de novo from vertebral corners (Fig. 6) . A lesion that was underestimated prior to the advent of fat-suppression MRI is the presence ofhigh marrow signal within vertebral bone adjacent to the central portion of the discovertebral junction (Fig. 7). This may be found in up to 20% ofpatients and most likely reflects an inflammatory spondylodiscitis. Occasionally the entire marrow of a single vertebra may demonstrate high sign al. As for acute Romanus lesions resolution ofinflammation is associated with fat infiltration. More extensive discovertebral lesions are seen in advanced disease and include diffuse end-plate destruction, with associated marrow oedema, fat infiltration, sclerosis and intradiscal ankylosis (Fig. 8). These abnormalities are not visible on plain radiography with the exception ofend plate irregularity that may be seen on lateral views of the spine. However, erosion of the endplate is far better visualized on MRI and appears to be far more frequent than erosion at the vertebral corner. It is not known if erosions necessarily follow inflammatory disci tis. Fractures through either the disc space or the vertebra itselfmay be observed in ankylosed spines as increased signal on fat-suppressed and decreased signal on TLweighted images reflecting inflammatory oedema around the fracture site. A variety of lesions are frequently visible on fat-suppressed images within the posterior elements of the spine that include high marrow signal in the facet joints and pedicles as well as ligamentary insertions at spinous and transverse processes (Fig. 9) . High marrow signal is also a common finding in the costovertebral and costotransverse joints. These lesions are frequently undetected for several reasons . First, MRI ofthe spine is typically conducted for neurological and orthopedic indications where most abnormalities are typically visible in central sagittal slices e.g., disc prolapse, rather than the lateral sagittal slices which depict the facet and costo-vertebral joints. As a corollary, radiologists have more training and experience detecting abnormalities related to neurological and orthopedic practice. Second, because ofoverlapping structures interpretation of the anatomy in lateral slices is more difficult. Third, minor degrees of scoliosis are very common and conventional imaging in the sagittal orientation may not cover the lateral structures on both sides of the spinal canal. This can be addressed by increasing the number or thickness of sagittal
28
Molecular Mechanisms ojSpondyloarthropathies
Figure 5. A) T1SE and B) STIR MRI sequences of a sagittal slice through the lumbar spine . Numerous areas of low T1 signal and high STIRsignal denoted by asterisks at anterior vertebral corners representing active inflammation.
slices. The importance of this for diagnostic purposes is underscored by a recent report in which systematic evaluation ofinflammatory lesions was undertaken and showed a higher frequency of inflammatory lesions in lateral as opposed to central slices in the thoracic spine. 55 It is unclear to what degree these abnormalities are sensitive and specific for SpA. For example, discovertebrallesions may also be observed in degenerative disc disease. Validation has also been indirect with several reports describing correlations between changes in acute MRI lesions and
Imaging in Spondyloarthritis
29
Figure 6. T1SE MRI sequence of a sagittal slice through the cervicothoracic spine . Numerous areas of high Tl signal indicated by solid wh ite arrows represent fat infil tratio n at the anterior vertebral corners. The open white arrow represents anterior ankylosis across the disc space.
changes in either self-reported pain and stiffness or C-reactive protein in AS patients receiving anti-TNF therapies. 56.57 Correlation with acute phase reactants has been most consistent." One study reported the results of histopathological analysis of facet joints obtained at the time of corrective surgery for severe kyphosis.l! A moderate correlation was noted between the presence of acute MRI lesions and histopathological scores for inflammation. However, sample size was small and these patients had longstanding disease. Most importantly, no reports have analyzed the predictive validity of acute spinal lesions for structural damage in prospective studies. This is of considerable importance in view of the slow evolution ofplain radiographic findings. Systematic evaluation of established disease has shown that the majority of lesions observed on STIR images occur in the thoracic spine.59 Two primary approaches have been described to quantify the extent ofinflammation in the spine for the purposes ofevaluation ofnew therapeutics in clinical trials. Both are based on the assessment ofa discovertebral unit (DVU) which represents the region between two imaginary lines drawn through the middle oftwo adjacent vertebrae. The first method. the ASspiMRI-a (AS spinal MRI activity) index, scores the severity ofbone edema
30
Molecular Mechanisms ofSpondyloarthropathies
Figure 7. A) T1SE and B) STIR MRI sequences of a sagittal slice through the cervicothoracic spine . An area of low T1 signal and high STIR signal denoted by the open arrow represents active inflammation adjacent to the vertebral end plate.
and erosions at each DVU according to a 0 to 6 scoring scheme with higher values being assigned to the presence oferosions.v The score for edema is based on the total area involved in the DVU according to a <25%, 25-50% and >50% grading scheme. The range ofscores is 0-138. The Berlin method does not score erosions but otherwise is the same as ASspiMRI-a and its scoring range is 0-69 . For both methods all 23 DVUs in the spine are scored in a single sagittal plane ofview. This approach has been shown to be reproducible and to discriminate between treatment groups in trials ofanti-TNF therapies.60•61 The second method, the SPARCC MRI spinal inflammation index, takes advantage of the ability of MRI to visualize lesions in three dimensions and scores a maximum of 6 of the most severely affected DVU.62 Since inflammatory lesions observed on MRI are often asymmetrical
Figure 8. A) T1SE and B) STIR MR I sequences of a sagittal slice through the lumbar spine. Open arrow indicates a complex disc lesion with loss of Tl signal and destruction of veretebral endplates together with increased T1 signal indicating fat infiltration and increased STIR signal indicating ongoing inflammation.
Imaging in Spondyloarthritis
31
Figure 9. STIR MRI sequence of a sagittal slice through the thoracolumbar spine. Numerous areas of high signal intensity indicated by white arrows represent active inflammation in the lateral and posterior segments of the spine.
there are theoretical advantages to using a scoring method that systematically assesses lesions in several dimensions. The rationale for the choice of 6 affected DVU is based on data that shows that the mean number of affected DVU in patients with AS is 3.2 (95% confidence intervals: 1.2-5.2).59 Additional work has shown that reliability and sensitivity to change may be optimal with a scoring scheme that is limited to a maximum of6 affected DVU as opposed to all 23 D VU.57 MRI is subject to artifacts and a potential limitation of mandatory scoring ofall 23 DVU is that reliability and sensitivity to change may be impaired if such artifacts, which are typically small, are recorded as lesions. In the SPARCC method each DVU is divided into quadrants with bone edema in each quadrant being scored on a dichotomous basis as being present or absent with additional weighting for intensity and depth (see www.arthritisdoctor.ca). Lesions in each affected DVU are scored in 3 consecutive sagittal slices so that the extent of the lesion can be defined in both the coronal and the sagittal planes. This method was originally validated for reproducibility and discrimination at 2 Canadian sites and then in multicentre trials ofanti-TNF therapies .54•62.63 A recent validation exercise has been conducted under the auspices ofthe Outcomes Measures in Rheumatoid Clinical Trials (OMERACT) network using multiple readers to determine which method performs best with respect to feasibility, reliability and ability to discriminate between active and control therapies.64.65 Both scoring methods demonstrate high responsiveness after administration of anti-TNFa therapies although the SPARCC method is consistently more
32
Molecular Mechanisms ofSpondyloarthropathies
reproducible, particularly when evaluated by neutral observers with limited experience in either method. In addition, the SPARCC method uses a greater part of the scoring range. This may be an advantage in the evaluation of therapies that are less potent anti-inflammatory agents than anti-TNFa therapies. Comparison ofSTIR with Gd-augmented MRI sequences using the ASspiMRI-a method has shown comparable reliability and sensitivity to change. 66 Since Gd -augmentation MRI is costly and imaging requires more than twice the time of STIR, the latter is preferred for initial MRI examination of the spine in AS. One method has been reported for scoring chronic changes in the spine, the ASspi-MRI-c, which analyzes sclerosis, squaring, syndesmophytes and fusion in all 23 DVU. 56 Inter-reader reliability was poor. Furthermore, a comparative study showed that this approach offered no advantages over plain imaging in the detection and scoring ofchronic lesions in the spine ." There are no comparative reports for CT imaging. Further study is necessary before any conclusions can be drawn regarding the value ofMRI in evaluation ofchronic lesions. Recent progress in magnetic resonance imaging (MRI) technology such as the use of whole body multicoil systems has led to the new imaging modality ofwhole body (WB) MRI. A moving table platform allows scanning of adjacent anatomical regions without repositioning of the patient or the imaging coils.Fusion ofthe images obtained at each table position results in a spatial resolution ofWB MRI which equals that ofconventional MRI and enables imaging of the entire body in a single head-to-toe scan in a relativelyshort period oftime. A potential advantage ofthis technique is the comprehensive visualization oflesions in the sacroiliac joints, entire spine, anterior chest wall, hip and shoulder girdles, peripheral joints and entheses (Fig. 10).
Figure 10. A) T1SE and B) STIR Whole Body MRI sequences of a sagittal slice through the entire spine. Open arrows indicate loss of T1 signal and increased STIR signal in the lateral and posterior segments of the spine (pedicles, costovertebral joints) representing active inflammation. Closed arrows indicate inflammatory spondylodiscitis.
Imaging in Spondy/oarthritis
33
MRI may also be useful in the evaluation of peripheral structures affected in SpA. In the shoulder, a characteristic finding is bone marrow edema on the STIR sequence adjacent to the supraspinatus insertion together with erosion evident on the T 1 sequence." Bone marrow edema may also be visible in degenerative enthesopathy though is typically lessextensive than in SpA and is not associated with erosion . Acromion enthesitis at the insertion of the deltoid appears to be pathognomonic for AS. Involvement ofthe hip presents as bone marrow edema ofthe acetabulum and femoral head extending into the neck and even the shaft ofthe femur. Synovitis and enthesitis at the gluteal muscle insertions into the greater and lesser trochanters may also be evident. One study of early knee synovitis reported that enthesitis with bone marrow edema was much more likely in patient who later developed characteristic SpA as compared to RA.69 Bone marrow edema is often evident at the Achilles enthesis just superior to the attachment ofthe tendon to the calcaneus. Although this may also occur in mechanical enthesopathy, it is much more extensive in SpA,7° Involvement of the midfoot is more characteristic ofjuvenile forms of SpA and tarsitis has been noted in up to 88% ofcases in some series. However, marrow edema may also be seen in degenerative and neuropathic forms ofarthritis." There are several pitfalls to the use ofMRI which can be broadly subdivided into those that are primarily technical in nature and those that relate to interpretation. Technical artifacts include thosedue to patient movement (including breathing, swallowing) and physiological motion (e.g., arterial, venous or CSF pulsation artifact). These may simulate low-grade inflammation and so are particularly relevant to patients with SpA. Detection ofabnormalities depends greatly on viewing conditions which are superior when the image is displayed on a computer monitor (as opposed to hard copy film) allowing the observer to manipulate the display properties. Common pitfalls in interpretation may be related to factors such as a lack of systematic assessment, fixed postural abnormalities (e.g.,scoliosis) in the patient and limitations in anatomical resolution as for the cervical vertebrae. A systematic approach to the detection oflesions in SpA is essential (see diagnostic imaging protocol at www.arthritisdoctor.ca). The clinical utility ofMRI is particularly evident in two circumstances. The first focuses on the diagnostic evaluation of a patient with clinical features of inflammatory back pain who is HLA B27 positive but has normal radiographs ofthe sacroiliac joints. Detection ofacute lesions on MRI increases the probability ofSpA. MRI may also be helpful in the assessment ofa patient with established AS who is still symptomatic despite the use ofstandard therapies because clinical assessment alone may not discriminate between inflammatory and mechanical causes ofback pain.
Conclusions For the assessment ofactive inflammation, particularly in bone marrow and soft tissue, there is little doubt that MRI is more sensitive than other imaging modalities . However, plain radiography continues to be the primary approach for evaluation ofchronic lesions. Ultrasound appears to be particularly useful in the assessment ofperipheral enthesitis, It remains to be determined to what degree MRI findings contribute information ofdiagnostic valuefor routine practice beyond clinical evaluation, assessment ofresponse to NSAIDs and B27 analysis.The clinical utility ofMRI might be greatly enhanced if it could be shown that the inflammatory lesions observed on MRI are of prognostic significance. Further study is therefore necessary to answer the question as to whether MRI can be used as a surrogate for structural damage. Further work is also required to clarify the sensitivity and specificity ofMRI lesions ifthis modality is to playa useful role in early diagnosis. MRI methodology should be systematic and standardized. This will facilitate the use of MRI as an outcome tool in trials of new agents for SpA. These studies will then allow us to address two of the most Significant challenges in the field of SpA, namely, therapeutic approaches to disease modification and early therapeutic intervention prior to the development ofstructural damage.
34
Molecular Mechanisms ofSpondyloarthropathies
References 1. Barrisrone M}, Manaster B}, Reda D} et al. Radiographic diagnosis of sacroiliitis-are sacroiliac views really better? J Rheumatol 1998: 25:2395-401. 2. Van der Linden S, Valkenburg HA, Cats A. Evaluation of diagnostic criteria for ankylosing spondylitis. A proposal for modification of the new york criteria. Arthritis Rheum 1984: 27:361-8. 3. Mau W; Zeidler H, Mau R er al. Clinical features and prognosis of patients with possible ankylosing spondylitis: results of a 10-year follow-up.} Rheumatol 1988: 15:1109-14. 4. Van Tubergen A, Heufi-Dorenbosch 1, Schulpen G er al. Radiographic assessment of sacroiliitis by radiologists and rheumatologists: does training improve quality? Ann Rheum Dis 2003: 62 :519-25 . 5. Wanders A}, Landewe RB, Spoorenberg A er al. What is the most appropriate radiologic scoring method for ankylosing spondylitis? A comparison of the available methods based on the outcome measures in rheumatology clinical trials filter. Arthritis Rheum 2004: 50:2622-32. 6. Wanders A, Landewe R, Dougados M et al. Association between radiographic damage of the spine and spinal mobility for individual patients with ankylosing spondylitis: can assessment of spinal mobility be a proxy for radiographic evaluation? Ann Rheum Dis 2005; 64:988-94. 7. MacKay K. Mack C, Brophy S et al. The bath ankylosing spondylitis radiology index (BASRI) : a new, validated approach to disease assessment. Arthritis Rheum 1998: 41 :2263-70. 8. Averns HL, Oxtoby ], Taylor HG et al. Radiological outcome in ankylosing spondylitis: use of the stoke ankylosing spondylitis spine score (SASSS). Br ] Rheumatol 1996; 35 :373-6 . 9. Creemers MC, Franssen M}, van'r Hof MA et al. Assessment of outcome in ankylosing spondylitis: an extended radiographic scoring system. Ann Rheum Dis 2005: 64:127-9. 10. Spoorenberg A, de Vlam K, van der Linden S et al. Radiological scoring methods in ankylosing spondylitis . Reliability and change over I and 2 years. J Rheumatol 2004: 31:125-32. II. Wanders A, Heijde D, Landewe R et al. Nonsteroidal antiinflammatory drugs reduce radiographic progression in patients with ankylosing spondylitis: a randomized clinical trial. Arthritis Rheum 2005; 52:1756-65 . 12. Maksymowych WP, Landewe R, Conner-Spady B et aI. Serum matrix metalloproteinase 3 is an independent predictor of structural damage progression in patients with ankylosing spondylitis. Arthritis Rheum 2007: 56:1846-53. 13. Powell A, Keeling SO, Lambert RGS et aI. Scoring of radiographic progression over 2 yearswith the mSASSS in ankylosing spondylitis: does training improve reliability? Arthritis Rheum 2007: 56(9):S256 . 14. Baraliakos X, Listing}, Rudwaleit M et al. Progression of radiographic damage in patients with ankylosing spondylitis: defining the central role of syndesmophytes. Ann Rheum Dis 2007; 66:910-5. 15. Bartafarano DF, West SG, Rak KM et aI. Comparison of bone scan, computed tomography and magnetic resonance imaging in the diagnosis of active sacroiliitis . Semin Arthritis Rheum 1993; 23:161-76. 16. Geijer M, Sihlbom H, Gothlin jf-l et al. The role of CT in the diagnosis of sacroiliitis. Acta Radiol 1998; 39 :265-8. 17. Carrera GF, Foley WO, Kozin F et aI. CT of sacroiliitis. A}R Am} Roentgenol 1981; 136:41-6. 18. Elgafy H, Semaan HB, Ebraheim NA er al. Computed tomography findings in patients with sacroiliac pain . Clin Orthop Relat Res 2001: 382:112-8. 19. Geijer M, Gothlin GG, Gothlin }H. The clinical utility of computed tomography compared to conventional radiography in diagnosing sacroiliitis. A retrospective study on 910 patients and literature review. J Rheumatol 2007: 34:1561-5. 20. Puhakka KB, }urik AG, Egund N ct al. Imaging of sacroiliitis in early seronegative spondylarthropathy, Assessment of abnormalities by MR in comparison with radiography and CT. Acta Radiol 2003 ; 44:218-29. 21. Yu W; Feng F, Dion E et aI. Comparison of radiography, computed tomography and magnetic resonance imaging in the detection of sacroiliitis accompanying ankylosing spondylitis. Skeletal Radiol 1998: 27 :311-20. 22. Friedman L, Silberberg P}, Rainbow A et aI. A limited, low-dose computed tomography protocol to examine the sacroiliac joints. Can Assoc Radiol ] 1993; 44:267-72 . 23. Verlooy H, Mortelmans L, Vleugels S et aI. Quantitative scintigraphy of the sacroiliac joints. Clin Imaging 1992: 16:230-3. 24. Goei The HS, Lemmens A}, Goedhard G er aI. Radiological and scintigraphic findings in patients with a clinical history of chronic inflammatory back pain. Skeletal Radiol 1985; 14:243-8. 25. Davis MC, Turner DA, Charrers jR et al. Quantitative sacroiliac scintigraphy. The effect of method of selection of region of interest . Clin Nucl Med 1984; 9:334-40. 26. Ayres}, Hilson A}, Maisey MN et aI. An improved method for sacro-iliac joint imaging: a study of normal subjects, patients with sacro-iliitis and patients with low back pain . Clin Radiol 1981 : 32:441-5. 27. Miron SD, Khan MA, Wiesen E} et aI. The value of quantitative sacroiliac Scintigraphy in detection of sacroiliitis. Clin Rheumatol 1983; 2:407-14 .
Imaging in Spondyloartbritis
35
28. Blum U. Buitrago Tellez C. Mundinger A er al. Magnetic resonance imaging (MRI) for detection of active sacroiliitis-a prospective study comparing conventional radiography. scintigraphy and contrast enhanced MRI.] Rheurnatol 1996; 23:2107-15 . 29. Song IH. Carrasco ]A. Rudwaleit M et aI. The diagnostic value of scintigraphy in ankylosing spondylitis and patients with clinically probable sacroiliitis- a systematic literature search. Arthritis Rheum 2007; 56(Suppl) :S475 . 30. Lin WY, Wang S]. Lang]L et aI. Bone scintigraphy in evaluation of heel pain in Reiter's disease: com pared with radiography and clinical examination. Scand] Rheumatol 1995; 24:18-21. 31. Olivieri I, Barozzi L. Padula A. Enthesiopathy: clinical manifestations. imaging and treatment. Baillieres Clin Rheumatol 1998; 12:665-81. 32. Balint PV. Kane D. Wilson H et aI. Ultrasonography of enrheseal insertions in the lower limb in spondyloarthropathy, Ann Rheum Dis 2002 ; 61:905-10 . 33. Kamel M. Eid H. Mansour R. Ultrasound detection of heel enthesitise a comparison with magnetic resonance imaging .] Rheumatol 2003 ; 30:774-8. 34. D'Agosnno MA. Said-Nahal R. Hacquard-Bouder C ct aI. Assessment of peripheral enthesiris in the spondylarthropathies by ultrasonography combined with power doppler: a cross-sectional study. Arthritis Rheum 2003; 48 :523-33. 35. Alcalde M. Acebes ]C. Cruz M er aI. A sonographic enthesitic index of lower limbs is a valuable tool in the assessment of ankylosing spondylitis. Ann Rheum Dis 2007; 66 :1015-9 . 36. Kamel M. Eid H . Mansour R. Ultrasound detection of knee patellar cnthcsitis: a comparison with magnetic resonance imaging. Ann Rheum Dis 2004 ; 63:213-14 . 37. Kamel M. Eid H . Mansour R. Ultrasound detection of heel enthe sitis : a comparison with magnetic resonance imaging.] Rheumatol 2003; 30:774-8. 38. Olivier i I. Barozzi L. Padula A er aI. Retrocalcanc al bur sitis in spondyloarthropathy. assessment by ultrasonography and magnetic resonance imaging.] Rheum atol1998 ; 25 :1352-7 . 39. Onlu E. Pamuk ON. Cakir N . Color and dupl ex doppler sonography to detect sacroiliitis and spina l inflammation in ankylosing spondylitis. can this method reveal respon se to anti-tumor necrosis factor therap y? I Rheurnatol 2007 ; 34:110-6. 40. Klauser A. Halpern E]. Frauscher F et aI. Inflammatory low back pain : high negative predictive value of contrast-enhanced color doppler ultrasound in the detection of inflamed sacroiliac joints. Arthritis Care Res 2005 ; 53:440-4 . 41. Muche B. Bollow M. Francois R] ec aI. Anatomic structu res involved in early- and late-stage sacroiliitis in spondyloarth ritis. Arthritis Rheum 2003 ; 48:1374-84. 42. Puhakka KB. ]urik AG. Schiottz-Christensen B et aI. Magnetic resonance imaging of sacroiliitis in early seronegative spondylarthropathy. Abno rmalities correlated to clinical and laboratory findings. Rheurnarol 2004 ; 43 :234-7. 43. Ahlstrom H. Peltelius N. Nyman R er aI. Magnetic resonance imaging of sacroiliac joint inflammation. Arthritis Rheum 1990 ; 33:1763-1769. 44. Braun ]. Bollow M. Eggens U et aI. Use of dynamic magnetic resonance imaging with fast imaging in the detection of early and advanced sacroiliitis in spondyloarthropathy patients. Arthritis Rheum 1994 ; 37:1039-45. 45. Braun]. Bollow M. Seyrekbasan F et aI. Computed tomography guided corticosteroid injection of the sacroiliac joint in patients with spondyloarthropathy with sacroiliit is: Clinical outcome and followup by dynamic magnetic resonance imaging.] Rheumatol 1996; 23:659-664 . 46. Bollow M. Fischer T. Reisshauer H et aI. Quantitative analyses of sacroiliac biopsies in spondylarthropathies : T-cells and macrophages predominate in early and active sacroiliitis-cellularity correlates with the degree of enhancement detected by magnetic resonance imaging. Ann Rheum Dis 2000 ; 59:135-40. 47. Bollow M . Biedermann T. Kannenberg] er aI. Use of dynamic magnetic resonance imaging to detect sacroiliitis in HLA-B27 positive and negative children with juvenile arthritides. ] Rheurnarol 1998 ; 25:556-64. 48. Bollow M. Braun ]. Hamm B er aI. Early sacroiliitis in patients with spondyloarthroparhy: evaluation with dynamic gadolinium-enhanced MR imaging. Radiolog y 1995; 194:529-36. 49. Heufi-Dorenbosch L. Landewe R. Weijers Ret aI. Combining information obtained from magnetic resonance imaging and conventional radiographs to detect sacroiliitis in patients with recent onset inflammatory back pain. Ann Rheum Dis 2006; 65 :804-8 . 50. Oo srveen ]. Prevo R. den Boer ] et aI. Early detection of sacroiliitis on magnetic resonance imaging and subsequent development of sacroiliitis on plain radiography. A prospective. longitudinal study. J Rheumatol 1999 ; 26:1953-8. 51. van der Heijde DM. Landewe RB. Hermann KG et aI. Application of the OMERACT filter to scoring methods for magnetic resonance imaging of the sacroiliac joints and the spine . Recommendations for a research agenda at OMERACT 7.] Rheurnacol 2005 ; 32:2042-7.
36
Molecular Mechanisms oJSpondyloarthropathies
52. Maksymowych WP. Inman RD. Salonen 0 et al. Spondyl oarrhritis research consortium of canada magneric resonance imaging index for assessment of sacroiliac joint inflammation in ankylosing spondylitis. Arthritis Rheum 2005; 53:703-9 . 53. Landewe RB. Hermann KG. van der Heijde OM et al. Scoring sacroiliac joints by magnetic resonance imaging. A multiple-reader reliability experiment. J Rheumarol 2005; 32:2050-5. 54. Lambert RGW; Salonen 0 , Rahman P et al. Adalimumab significantly reduces both spinal and sacroiliac joint inflammation in patients with ankylosing spondylitis. Arthritis Rheum 2007; 56:4005-14. 55. Crowther SM, Lambert RGW; Dhillon SS er al. High frequency of inflammatory lesions in the posterior structures of the spine in patients with ankylosing spondylitis (AS): A systematic evaluation by MRI. Arthritis Rheum 2006; 54(Suppl 9) :S793. 56. Braun J. Baraliakos X, Golder W et al. Magnetic resonance imaging examinations of the spine in patients with ankylosing spondylitis, before and afier successful therapy with infliximab : evaluation of a new scoring system. Arthritis Rheum 2003 ; 48:1126-36. 57. Maksymowych WP. Dhillon SS, Park Ret al. Validation of the spondyloarthritis research consortium of canada (SPARCC) MRI spinal inflammation index : Is it necessary to score the entire spine? Arthritis Rheum 2007; 57(3):501-07. 58. Appel H . Loddenkemper C. Grozdanovic Z er al. Correlation of histopathological findings and magnetic resonance imaging in the spine of patients with ankylosing spondylitis. Arthritis Res and Ther 2006; 8:RI43. 59. Baraliakos X. Landewe R. Hermann KG et al. Inflammation in ankylosing spondylitis : a systematic description of the extent and frequency of acute spinal changes using magnetic resonance imaging. Ann Rheum Dis 2005 ; 64:730-4. 60. Braun J, Landewe R. Hermann KGA et al. Major reduction in spinal inflammation in patients with ankylosing spondylitis afier treatment with infliximab. Arthritis Rheum 2006; 54:1646 -1652 . 61. Baraliakos X, Davis J. Tsuji W et al. Magnetic resonance imaging examinations of the spine in patients with ankylosing spondylitis before and afier therapy with the tumor necrosis factor alpha receptor fusion protein etanercept. Arthritis Rheum 2005 ; 52:1216-1223. 62. Maksymowych WP. Inman RD, Salonen 0 er al. Spondyloarthritis research consortium of canada magnetic resonance imaging index for assessment of spinal inflammation in ankylosing spondylitis . Arthritis Rheum 2005 ; 53:502-9. 63. Maksymowych WP, Salonen 0, Inman RD er al. Canadian evaluation of low dosE infliximab in ankylosing spondylitis (CANDLE)-I2 week magnetic resonance imaging evaluation of spinal inflammation with the SPARCC MRI method. Arthritis Rheum 2007; 56:S473 . 64. Lukas C , Braun J. van der Heijde 0 er al. ASAS/OMERACT MRI in AS Working Group. Scoring inflammatory activity of the spine by magnetic resonance imaging in ankylosing spondylitis: a multireader experiment. J Rheumatol 2007 ; 34:862-70 . 65. van der Heijde D. Landewe R, Hermann KG et al. ASAS/OMERACT MRI in AS working group . Is there a preferred method for scoring activity of the spine by magnetic resonance imaging in ankylosing spondylitis ? J Rheumatol 2007; 34:871-3. 66. Baraliakos X, Hermann KG. Landewe Ret al. Assessment of acute spinal inflammation in patients with ankylosing spondylitis by magnetic resonance imaging : a comparison between contrast enh anced T1 and short tau inversion recovery (STIR) sequences. Ann Rheum Dis 2005 ; 64 :1141-4. 67. Braun J, Baraliakos X, Golder W et al. Analysing chronic spinal changes in ankylosing spondylitis : a systematic comparison of conventional x rays with magnetic resonance imaging using established and new scoring systems. Ann Rheum Dis 2004; 63:1046-55. 68. Lambert RG, Dhillon SS, Jhangri GS et al. High prevalence of symptomatic enrhesopathy of the shoulder in ankylosing spondylitis : deltoid origin involvement con stitutes a hallmark of disease. Arthritis Rheum 2004; 5 I :68 1-90. 69. McGonagle 0, Gibbon W; O'Connor P et al. Characteristic magnetic resonance imaging entheseal changes of knee synovitis in spondylarthropathy, Arthritis Rheum 1998; 41 :694-700. 70. McGonagle 0, Marzo-Ortega H , O'Connor P er al. The role ofbiomechanical factors and HLA-B27 in magnetic resonance imaging determined bone changes in plantar fascia enthesopachy, Arthritis Rheum 2002 ; 46:489-93. 71. Burgos-Vargas R. Pacheco-Tens C, Vazquez-Mellado J. A short-term follow-up of enthesitis and art hritis in the active phase of juvenile onset spondyloarthropathies. Clin Exp Rheumatol 2002 ; 20:727-31.
CHAPTER
3
Spondyloarthritis, Diffuse Idiopathic Skeletal Hyperostosis (DISH) and Chondrocalcinosis Jacome Brugues Armas," Ana Rita Couto and Bruno Filipe Bettencourt
Abstract
T
he authors describe the main clinical and radiological findings ofcommon enthesopathic d isorders-spondylarthritis (SpA), chondrocalcinosis/ calcium pyrophosphate dehydrate crystal deposition disease (CPPD C D D) and diffuse idiopathic skeletal hyperostosis (D ISH) , stressing similarities and differences which may help in th e differential diagnosis . They emphasize the clinical presentation ofthe "pseudoankylosing spondylitis" form s ofCPPD CDD. They also review the most relevant genes and molecular mechan isms associated with these conditions and with another enthesopathic disorder with high prevalence in th eJap anese populationossificatio n of the po sterior longitudinal ligament (O PLL).
Introduction D ifficulties on the different ial diagnosis bet ween SpA and DISH may be relevant in the clinical practice, mainly because clin icians are frequent ly unaware ofDISH manifestations. The classical acute Chond rocalcinosis manifestation s-pseudogout, are usually a differential diagnosi s ofother crystal disorders like gout. H owever. oth er clinical and radiological manifestations ofCPPD CDD . not very well known by the majority of the clinicians, namely those related to axial enthesopathic phenomena , may give origin to a wron g diagnosis. In th is chapter we will update the genetics of the se disorders. emphasising DISH and Chondrocalcinosis, because mole cular knowledge of An kylosing Spondylitis (AS), the major SpA subgroup, will be detailed in other chapters in thi s book . We will also evaluate the clin ical and radiological findings in the three disease groups, indicating the similarities and differences between them and providing information which may allow an easier differential diagnosis.
Spondyloarthritis Spondylarthritls, spondylarthropathy, or the spondylarthritides (SpA), are the most common terms to describe a group ofdiseases (Table 1),1 which are characterised by the inflammatory involvement ofthe sacroiliac joints, peripheral inflammatory arthropathy and insertional tendinitis (enthesitis) .2.3The tendency towards familial aggregation is very well known " and the association with HLA-B27 was described for the first time by Brewerton DA et al5 and Schlosstein LP er a16 Several criteria for Ankylosing Spondylitis was proposed, like the Modified New York criteria , 19847 (Table 2). The criteria did not encompass conditions like Undifferentiated Spondylarthritis 'Correspo ndi ng Author: Jacom e Bruges Armas - SEEBMO, Hospital de Santo Espirito de Angra do Hero fsmo, Azor es / Institut e for M olecular and Cell Biol ogy (IBMq, University of Porto, Portugal. Email: jacome.armas@netc. pt
Molecular M echan ism s ofSpondyloarthropathies. edited by Carlos Lop ez-Larre a and Roberto Dia z-Pefia. ©2009 Landes Bioscience and Springe r Scien ce-t-Business Media.
38
Molecular Mechanisms ofSpondyloarthropathies
Table 1. Individual conditions that overlap to form the spondylarthritides • Ankylosing spondylitis • Reiter's syndrome/reactive arthropathy Enteropathic spondylitis • Psoriatic arthropathy • Uveitis • Juvenile ankylosing spondylitis • Undifferentiated spondylitis • Pustulotic arthro-osteitis and criteria for the entire group ofSpA were proposed. The European Spondyloarthropathy Study Group criteria (ESSG) have a sensitivity and a specificity of87% (Table 3 ).8 The prevalence of idiopathic AS, the major SpA group, varies within different populations and the prevalence ofHLA-B27.The diagnosis of idiopathic AS implies always the exclusion of other SpA. The sex related prevalence results always reported higher values in males, but figures of male female of3: 1 or 2:1 are acceptable . Hospital based studies concluded that the prevalence of AS was about 0.10-0.20%.9.10 The prevalence ofAS based in studies of blood donors varies from 0.10-0.20%, II and population surveysshowed the highest prevalence in the male Haidas inhabitants from northern Canada, ofwhom 4-6% have AS. 12 In European populations the surveys showed some variability-from 0.10% to 1.4%,13.15 related possibly to the methodology applied and to the HLA-B27 prevalence. It was also estimated that the prevalence ofAS in the HLA-B27 population may be as high as 5%.16.17 According to the information obtained in hospital based and population surveys, there are milder frequently underdiagnosed forms ofAS in the population characterised by less restriction ofspinal disease and lower incidence ofperipheral arthritis.P'!? Sacroiliac joint involvement is the hallmark ofAS, but the inflammation also involvesfrequently the spine, hips and shoulders (both considered axial joints) and, less frequently, peripheral joints. The age of onset is usually after 24 years, but rarely after 45 years 01d.20 Extra-articular structures may also be affected like the eye (uveitis), the heart and, less commonly, the lung. Chronic low back pain and stiffness is the most common clinical finding in AS. Pain starts insidiously and patients usually complain that it is worse during rest and sleep and that there is improvement with movement and exercise. Pain may radiate into both thighs, simulating sciatica, but usually alternates first between one side and the other, later becoming persistent. The pain intensity is variable and frequently episodes ofback (buttock) pain alternate with periods of mild pain and stiffness: this probably represents the aggression of the sacroiliac joint. Some less severe forms of disease may have milder or no back pain, stiffness and later, minor spinal restriction findings.":" The involvement ofshoulders and hips may occur in one third of patients. The hip is the second most frequent area affected by AS, the aggression occurs mainly in the first ten years after the
Table 2. Modified New York criteria, 1984 Criteria 1. Low back pain for at least three months' duration improved by exercise and not relieved by rest 2. Limitation of lumbar spine motion in sagittal and frontal planes 3. Chest expansion decreased relative to normal values for age and sex 4a. Unilateral sacroiliitis grade 3-4 4b. Bilateral sacroiliitis grade 2-4 Definite ankylosing spondylitis if (4a OR 4b) AND any clinical criterion (1-3)
Spondyloartbritis, Diffuse Idiopathic Skeletal Hyperostosis (DISH) and Cbondro calcinosis
39
Table 3. European spondyloarthropathy study group criteria Inflammatory spinal pain or Synovitis (asymetric, predominantly in lower limbs) And any of the following: • Positivefamily history • Psoriasis • Inflammatory bowel disease • Alternate buttock pain • Enthesopathy Sensitivity, 77%; specificity, 89% Adding: • Sacroiliitis Sensitivity, 87%; specificity, 87% diagnosis and is usually bilateral." Shoulder girdle (glenohumeral, acromioclavicular or sternoclavicular joints) disease may lead to mild pain and restriction ofmovements to severe ankylosis of the joint. 23 Peripheral arthritis is usually asymmetric. oligoarticular, transient. mild, rarely persistent and usually without permanent defurmities or erosions.t" Enthesitis may be found in justa-articular or extra-articular insertion tendon or ligament areas and causespain and tenderness. Costostemal junctions tenderness can cause chest pain and tendinitis of the Achilles tendon may cause severe and longlasting pain.25 The role ofproinflammatory cytokines in the development ofosteoporosis was raised. after confirmation of low lumbar and femoral neck reduction in bone densitometry by dual energy X-ray absorptiometry (DEXA) in early AS.26 Osteoporosis in patients with severe AS oflong duration may possibly result also from the absence ofspinal mobility and is a cause of fracture that occurs predominantly in the cervical spine." The most common extra skeletal manifestation of AS is acute anterior uveitis (AAU) which occurs in about 30% of patients and is much more frequent in the presence ofHLA-B27.The uveitis is usually unilateral. frequently recurrent and subsides if lett untreated in 2-3 months, with no sequelae. Symptoms start abruptly-pain, photophobia and blurred vision are the main manifestations. Heart disease occurs mainly in patients with longstanding AS and peripheral arthriris. " Several manifestations have been described, from the complete set ofcard iac conduction abnormalities to aortitis of the ascending aorta and aortic valveincompetence. The severe forms ofAS usually evolvewithin a period of 10 years and the characteristic thoracic kyphosis. posture and gait make easier the diagnosis. In these later stages the evident limitations ofspine movement are mainly related with ankylosis, following inflammatory involvement ofthe interapophyseal spine joints and spine enthesopathic calcification. Earlier radiographic manifestations of the inflammatory process start usually in lumbar spine, affecting the superficial layers ofthe annulus fibrosus where they attach to the corners ofvertebral bodies, followed by reactive sclerosisand erosive bone resorption, X-rays show squaring ofthe vertebral bodies, ossification of the superficial layers ofthe annulus fibrosus and bridging between vertebrae (syndesmophyres).29 The apophyseal joints are involved frequently in AS and ankylosis is usual and the posterior ligaments may ossify causing also ankylosis of one spine process to another.'? Enthesopathic calcification and ossifications are frequent in AS mainly at sites of osseous attachment of tendons and ligaments. Ischial tuberosities, the iliac crest, but also the calcanei and the femoral trochanters may have evidence ofprevious erosions and ostheitis, Ankylosis ofthe peripheral joints is a rare event in AS. The sacroiliac joint is usually the first affected joint in AS. The earliest radiographic signs consist of bilateral blurring of the subchondral bone plate, followed by erosions and sclerosis in the iliac side ofthe joint. Later there is fibrosis, calcification, interosseous bridging and ankylosis. Some patients may have inflammatory back pain for several years without X-ray findings, mainly women " and in these occasions sacroiliac CT scan 32 or MRIl3 may be helpful.
40
Molecular M echanisms ofSpondyloarthropathies
Diffuse Idiopathic Skeletal Hyperostosis (DISH) Despite being a very frequent disease. whose first detailed description is attributed to Forrestier and Rotes ~eroJ,34 there are still many unknown data about DISH. Clinical data is frequently not identified by clinicians and thi s is probably because their clinical manifestations were always regarded as mild and without important con sequences. The majority of criteria developed to diagnose DISH is related to radiographic findings. The three criteria to diagnose spinal DISH are: (1) The presence of flowing calcification and ossification along the anterolateral aspect of at least four contiguous vertebral bodies with or without associated locali sed pointed excrescences at the intervening vertebral bodies-intervertebral disc junctions. (2) Th e pr esence of relative preservation of intervertebral disc height in the involved vertebral segment and the absence of extensive radiographic changes of"degenerative" disc disease, including vacuum phenomena and vertebral body margin sclerosis. (3) The absence ofapophyseal joint bony ankylosis and sacroiliac joint ero sion . sclerosis, or intra-articular osseous fusion/" Albeit being a disease that affects mainly the axial skeleton and in particular the thoracic spine, it is now recognised that this condition is nor limited to the spine . but involves frequently peripheral joints and for this reason the previous designation ofForestier and Rotcs- Querol disease was changed to DISH. The disease is characterised by ossification of the anterior longitudinal ligament and by the presence of flowing osteophytes involving predominantly the right side of the thoracic spine, with preservation of the intervertebral space. Abnormalities in the sacroiliac joints are not uncommon; the upper part of the sacroiliac joint is ligamentous, while the lower 2/3 is a synovial joint and asymmetric superior intraarticular fusion , osreophyres with and without bridging and vacuum phenomena were described by Durback et al 36There are also reports on the intra-articular ankylosis ofthe sacroiliac joints, but this finding is controversial.'? Pelvic bony proliferation (whiskering) is frequently found at the sites of ligaments and tendons attachment to bone, like the iliac crest. ischial tuberosity and trochanters. The iliolumbar and sacrotuberous ligaments may be calcified and ossified. The disease prevalence varies but data from authopsy showed that there was a frequency of 28 % in a population with an average age of 65 years.38 Th e prevalence is higher in males in the majority of reports. 39,40 Although patients are sometimes asymptomatic , pain and stiffness began usually in the thoracolumbar spine, may be initially intermittent, but can progress to involve other spine segments. Sometimes symptoms are reported many years before a radiographic diagnosis. D ysphagia. probably related with prominent cervical osteophytes is a frequent sympto m in some series of DISH parients." There are several reports of patients between 20 to 40 years old with radiographic aspects that do not fulfill the classic DISH criteria, but the findings were strongly suggestive." Not infrequently these find ings are related to pain over the enrhesis in peripheral joints, associated with radiological abnormalities. Enthesopathic phenomena are mo re frequent at the elbows-medial epicondylitis and olecranon, patella. heel and tibial spines.43.44 There are several reports on the positive association ofDISH and Di abetes Mellitus'? and some reports in the literature claim on the familial aggregation ofDISH and OPLL.46-530 PLL, a common cause of spinal canal stenosis in Japanese population, is characterised by ossification ofthe posterior Iongitudinal Iigament and has a prevalence of 1.9-4.3%.54 DISH and OPLL may occur together and Resnick et al reported that about 50% of cases of DISH are associated with OPLL.55
Chondrocalcinosis/Crystal Deposition Disease (CPPD CDD) Th e term cbondrocalcinosis (CC) refers to the calcification of joint hyaline or fibrocartilage, a frequent radiologic feature ofcalcium pyrophosphate dehydrate crystals deposition (CPPD) and is attributed to Ziman and Sitaj.56Other designation s have been coined like pyrophosphatearthropathy which is widel y used in the United Kingdom.V although presently the most accepted name to describe the clinical syndromes associated with C PPD is CPPD crystal deposition disease." In thi s chapter we have adopted th e designation of CPPD CDD. This condition is usually reported as a di sease ofolder patients which prevalence increases with the incr easing ofage. Three main groups were described-Sporadic forms which are con sidered the largest group, hereditary CPPD CDD
Spondyloartbritis, Diffuse Idiopathic Sdeletal Hyperostosis (DISH) and Cbondrocalcinosis
41
and CPPD CDD associated with metabolic diseases: hemochromatosis, hyperparathyroidism, hypothyroidism, hypophosphatasia, Wilson's disease and other entities. CPPD crystal deposition disease is frequently accompanied by acute episodes ofarthritis involving the larger joints, mainly the knees-the pseudogout syndrome and these episodes are caused by the precipitation ofCPPD crystals in the joint fluid. 59•60 The definite diagnosis is performed after the demonstration of CPPD crystals in tissues or synovial fluid by compensated polarised light microscopy. There is a diagnostic classification based on this assumption and modified to include radiographic findings suggested by Resnick et alGI and Martel et al,G2 (Table 4).58 The finding of CPPD crystals is crucial, because this disease is quite heterogeneous and may resemble other conditions, like gout, rheumatoid arthritis, neuropathic disorders and Ankylosing Spondylitis.58.63 McCarthy describes various clinical presentations ofjoint disease associated with CPPD deposition and several groups are considered-Type A-Pseudogout, Type B-Pseudorheumatoid Arthritis, Type C and D-Pseudo-osteoarthritis, Type E-Lanthanic (asymptomatic) and Type F-Pseudoneuropathic joints. About 75% ofCPPD CDD patients presents with the pseudogout (Type A) pattern (25%),or with the Pseudo-osteoarthritis (Types C and D, 50%).58 In this last group acute attacks occur in approximately halfofthe cases and the main joints affected are the knees, wrists, metacarpophalangeal joints, hips, spine, shoulder, elbows and ankles. Some difficulty to perform a differential diagnosis with osteoarthritis (OA) is usual, but CPPD CDD affects joints that are not frequently involved in primary OA, like the metacarpophalangeal joints, wrists, elbows and shoulders. The pattern of joint involvement is usually symmetric and flexion contractures ofthe affected joints, enthesopathic symptoms mainly in the Achilles, triceps and shoulders tendons and enerapment syndromes ofthe upper limbs have been described. 64•65
Table 4. Diagnostic criteria and categories for CPPD crystal deposition disease Criteria I.
Demonstration of CPPD crystals, obtained by biopsy, necropsy, or aspirated synovial fluid, by defin itive means; e.g., characteristic "fingerprint" by X-ray diffraction powder pattern or by chem ical analysis
II.
A) I dentification of monoclinic or triclinic crystals showing a weakly positive, or a lack of, birefringence by compensated polarised light microscopy. B) Presence of typical calcifications on radiographs
III.
A) Acute arthritis, especially of knees or other larger joints, with or without concomitant hyperuricemia. B) Chronic arthritis, especially of knee, hip, wrist, carpus, elbow, shoulder and metacarpophalangeal joints, particularly if accompanied by acute exacerbations; the chronic arthritis shows the following features helpful in differentiating it from osteoarthritis. 1) Uncommon site for primary osteoarthritis; 2) Radiographic appearance; 3) Subchondral cyst formation; 4) Severe progressive degeneration, with subchondral bony collapse (microfractures) and fragmentation, with formation of intraarticular radiodense bodies . 5) Variable and inconstant osteophyte formation. 6) Tendon calcifications, especially of Achilles, triceps and obturador tendons. 7) Involvement of the axial skeleton and subchondral cysts of apophyseal and sacroiliac joints, multiple levels of disc calcification and vacuum phenomenon
Categories A.
Definite-criter ia I or II (A) and II (B) must be fulfilled
B.
Probable-criteria IIA or liB must be fulfilled
C.
Possible-criteria IliA or IIIB should alert the clinician to the possibility of underlying CPPD deposition
42
Molecular Mechanisms ofSpondyloarthropathies
The radiographic features of CPPD CD D may be divided in articular and periarticular calcifications and cartilage, synovium , capsula, tendons, ligaments, soft tissues, blood vessels and bursae may be affected. Demonstrable articular cartilage calcifications are more frequent in knees rneniscii, symphysis pubis , triangular cartilage ofthe wrists, annulus fibrosus ofthe intervertebral disc and acetabular labra (hips).61 Synovial calcification is also common in CPPD CDD and is more frequent in the wrist particularly about the radiocarpal and inferior radioulnar joints, knees and MCP joints.66.67 Capsular calcification is most frequently observed in the elbows, metatarsophalangeal and metacarpophalangeal joints, as fine or irregular linear calcifications and joint contractures mainly in the elbows are frequently associated.?' The first description of familial CPPD CDD is attributed to Zitnan and Sitaj in 1963,68but many other families have been reported since then, in different countries and conrlnents.":" Two distinct clinical phenotypes were described: a group has a milder phenotype and the symptoms start after 60 years old, the disease is frequently oligoarticular and is indistinguishable from the sporadic form ofCPPD CDD. The other phenotype is characterised by early onset ofsymptoms, usually between 20 and forty, but younger ages were reported." Polyarthritis is a frequent finding , the prognosis is variable but the evolution is frequently characterised by the severity ofthe clinical and radiological picture. 68.72.76.78 A systematic f.'lmilialinvestigation ofall caseswith a diagnosis ofsporadic epPD eDD has not been performed and because of this, an exact figure ofthe prevalence ofboth forms is not known. According to McCarey a diagnosis ofsporadic CPPD CDD should be done only after exploring family data?9The frequency ofthe hereditary forms is probably higher than reported and is due mainly to the lack of appropriate investigations. Rodriguez-Valverde et al80 found a prevalence of 17.1% of hereditary CPPD CDD forms in Spain, McCarthy observed a 25% prevalence of familial disease in his patients with pseudogour?"and Dapica and Gomez-Reino found a 28% in a Spanish urban population." Although the clinical and radiographic features, do not allow an easy distinction between the sporadic and the hereditary cases of CPPD CDD, the age of the first symptoms may be helpful. Furthermore, several reports in the literature point to an association of the CPPD CDD hereditary forms with DISH and ankylosis.49.76.78.82.87Harris] er aP9also describes 6 patients with CPPD CDD in knees and hips, out of34 DISH patients. Although in this report the patients ages ranged from 47 to 83 years old, the author does not refer to the age of the six CC patients. Other reports describe the association of Chondrocalcinosis and Ankylosing Spondylitis, or a clinical presentation ofChondrocalcinosis mimicking Ankylosing Spondylitis, referring this form as "pseudoankylosing spondylitis."88.89
Molecular Genetics Mineral deposition is an extremely balanced process normally restricted to bone formation. It is a tightly controlled process taking place in chondrocyte and osteoblast derived matrix vesicles (MVs). These vesicles contain enzymes, calcium and inorganic phosphate ions (Pi) and it is within these vesicles that the first hydroxyapatite (HA) and CPPD crystals are formed. 90•93 It is known that inorganic pyrophosphate (PPi) antagonizes the formation ofHA crystals hence the importance ofhaving a tight balance between the levels of extracellular Pi and PPi for normal mineral depositlon.tv" It has also been long known that individuals suffering from acute inflammatory reactions caused by calcium pyrophosphate dehydrate deposition have high concentrations of inorganic pyrophosphate in synovial fluids.96 Much work has been performed on this subject given that knowledge of PPi metabolism has been considered of great importance in ectopic calcifications studies. To date, many genes involved in mineral deposition have been ldenrified." Mutations in several of these genes affect the rate of generation, transport and degradation of pyrophosphate (PPi) leading to an unbalanced level which will ultimately lead to ectopic deposition of crystals throughout the skeleton" (Fig. 1).98
43
Spondyloartbritis, Diffuse Idiopathic Sde/eta/ Hyperostosis (DISH) and Cbondrocalcinosis
Gitelman syndrome
WA3
SLC12 lENPP1 HypophosphatasJa
~~ CLCNKB ~
Bartter syndrome
1r---'':''c-ry-s-ta-l-SO-IU-b-i-lity~1
D
CPPD-DD '----_ _I
HFE ATP7B Wilson Disease
~~
I HA-DD I
~OPLL
~
CMD
Figure 1. Mechanisms underlying associations of genetic diseases with ectopic m ineral deposition . CMD-craniometaphyseal dysplasia ; CPPD-CPP-calcium pyrophosphate deposition disease; ENPP1-enzyme ectonucleotide pyrophosphatase; ePPi-extra cellular inorganic pyrophosphate; HA-CDD-hydroxyapatite crystal deposition disease ; Mg-magnesium; OPLL-ossification of the posterior longitudinal ligament; TNAP-Tissue nonspecific alkaline phosphatase (adapted from : Couto AR, Brown MA. Genetic factors in the pathogenesis of CPPD crystal deposition di sease. Curr Rheumatol Rep 2007; 9(3):231-6.
CPPD CDD segregates in an autosomal dominant manner and has been described in numerous families."9.99.lOo Until now two chromosomal regions have been linked to CPPD C D D: chromosom e 8q (C C ALl ) and chromosome 51' (C C AL2). In 1995, Baldwin and his colleagues, lUI reported genetic linkage between markers on human chromosome 8q and the disease in a large family from Maine with early onset CPPD and severe degenerative OA [CCALl, MIM 600668]. The disease causing gene was not mapped yet but there is a strong po ssibility that this gene is primarily related with osteoarthritis and that the CPPD deposition is secondary, enhanced by the degenerative changes in cartilage. A second form of autosomal dominant CPPD CDD CCAL2 [CCAL2, MIM 118600] was initially described and genetically localised in five English families with CPPD CDD. Linkage was established to chromosome 51" with a multipoint LOD score of4.6. 102 Subsequently, in two CPPD-CC unrelated families , one from Argentina and the other from France, another zone in chromosome 51', mapping just centromeric to the first, was further identified through recombination analysis.' :" The identification of th e ank gene and demonstration that loss-of-function mutations in the gene cause ectopic hydroxyapatite deposition in the ank mouse model, '!" led investigators to hypothesize that gain-of-function mutations in the human homologue of this gene may be responsible for the disease in these families. !Os This finding was confirmed in other families with autosomal dominant CPPD CDD.106.108 In 2005, Zhang et al 109 demonstrated that a common promoter region polymorphism is associated with "sporadic" cases ofCPPD CDD. The polymorphism identified in this study, a-4 basepair G -to-A transition, was found in 4% of a 128 patient cohort with widespread premature
44
Molecular M echan isms ofSpondyloarthropathies
chondrocalcinosis and/or pseudogout.!" So it wa s suggested that ANKH variants playa role in both familial and severe form s of ,sporad ic' chondrocalcinosis. The ANKH gene encodes a 492 amino acid multiple-pass transmembrane protein which is thought to be involved in th e transport ofinorganic pyrophosphate (PPi) acro ss plasma membrane to the extra cellular compartment. Loss-of-function mutations in ank lead to an increase in intra cellular PPi and a decrease in extr a cellular PPi , promoting the excess calcium hydroxyapatite formation found in theank/ank mouse.'?' The exact me chanism by whi ch ANKH mutations cause these di seases is unknown. The stru cture ofthe ANK protein has not been determined , pos sibly because oftechnical difficulties involved in stru ctural stud ies oftransmembrane proteins. Strong evidence has been pre sented that ANK is a PPi transmembrane transporter. The strongest evidence for this has been for hydroxyapatite-deposition associated mutations causing either the ank mouse phenotype , or craniometaphyseal dysplasia. Here. transfection studies and studies in transgenic mice have shown that these ank variants are associated with markedly reduced PPi transmembrane transport.'! " With CPPD-associated mutations, the same group did not ob serve any significant difference in PPi tran sport compared with wild-type ank.' !" Transient transfection studies in chondrocyte cell lines have suggested that not all CPPD CDD mutations lead to increased extracellular PPi and some have other effects on chondrocyte maturation and mineralization independent ofeffects on PPi. 109The P5L mutation, associated with CPPU CUD. increases 'Iype X collagen expression in both transient and stable transfection studies, I09.111 suggesting an effect ofthe mutation on chondrocyte maturation. perhaps explaining the association of this mutation with chondrocalcinosis (CC). Part ofthe difficulty in determining the function ofANK and effect s ofA N KH mutations may relate to the technically challenging nature of measurement of extra- and intra-cellular PPi and the fact that available chondrocyte cell line s are not ver y close matches to primary cells. Further, ATDC5 chondrocyt ic cell lines expre ss high levels of alkaline phosphatase. which breaks down PPi and CH-8 chondrocytic cells express high levels ofTNF, which has been show n to inhibit PPi elaboration by chondrocytes.l' ? Primary cho nd ro cytes de -differentiate rapidly in culture, complicating their use. The absence ofa murine model ofCPPD chondrocalcino sis is a particular weakness. Even hypophosphatasia m ice, with severe chronic elevations of ePPi, do not develop CPPD chondrocalcinosis. This may be due to the relatively sho rt life span ofmice com pared with humans. or some characteristic ofmurine cartilage. The development of such a mouse model would greatly advance our ability to investigate the aetiopathogenesis of this di sease. The association of ANKH with AS is also rather controversial. The ank mouse initially came to attention as a potential model of human AS. Further, reports ofa clinical association between AS and another disorder of ectopic hydro xyapatite deposition, OPLL raised the possibility that ANKH may playa role in the pathogenesis ofspinal ossification in AS . While a stu dy performed by Tsui et al ll3 reported a modest association of ANKH-OR and ANKH-TR to AS susceptibilit y. a study performed by Timms et al l 14 showed no evidence of association of ANKH polymorphisms with either susceptibility to or severity of disease. Further genotyping 201 multiplex AS families with nineANKHintragenetic and two flanking m icro satellite markers and family-based association analysis , Tsui et al (2003)115 showed that the region associated with AS in women wa s only signi fican tly disease associat ed in th e te st of interaction among the subset of families with affected individuals of both genders. Th ese findings support th e con cept that ANKH plays a role in genetic susceptibility to AS and reveal a gender-genotype spe cificity in th is interaction. To our knowledge. no other gene s po ssibly associated simultaneously to AS and C PP D CDD were reported. HLA-B27 is a critic al factor on AS susceptibility and 48 subtypes were identified until now. B'2705 ha s been found in mo st populations wo rldwide and is overrepresent ed in th e Circumpolar and Subarctic regions of Eurasia and North America. Th e association of this allele as well as B'2702 and B"2704 with AS ha s been proven in case-control stu dies. I 16.117 A S cases were also reported in individuals with the alleles B'2701 . ' 27 03. '2706. '2708. ' 27 10. ' 27 14. ' 27 15 and ' 27 19, in different populations . but the true meaning of th ese associat ions remain unclear due to insuffic ient data.118· 122 HLA C lass I molecules bind and trigger cell-surface receptors
Spondyloartbritis, Diffuse Idiopathic Sdeletal Hyperostosis (DISH) and Cbondrocalcinosis
45
specified by Killer cell Inununoglobulin-like Receptors (KIR) genes. A recent study aimed at analysing whether KIR3DLI and 3DSI genes and their possible synergistic effect with HLA-B alleles, influenced the susceptibility to AS in B27 patients.!" This investigation showed that the presence ofKIR3DS I or KIR3DL I in combination with HLA- B"27s/HLA-Bw4-180 genotypes may modulate the development ofAS. The susceptibility to AS could be determined by the overall balance ofactivating and inhibitory composite KIR-HLA genotypes. As a result ofwhole genome linkage screens several chromosomal areas possibly involved in disease susceptibility, or related with the age ofsymptom onset or to BASFI (Bath Ankylosing Spondylitis Functional Index) were identified.124- 126 IL-I cluster includes the genes for the cytokines IL-I alpha, IL-I beta and their inhibitor IL -I receptor antagonist. These genes lay on chromosome 2 (2q) and were identified in one genome wide screen as AS associated.!" In 2004 Timms et al showed that IL-I B -5 II SNP and IL-I F I 0.1 were consistently AS associated even controllingother associated markers .127 More recently the results of a metanalysis involving groups from ten countries were published. Strong association was observed with three SNPs in the ILIA gene (rs2856836, rs1756, rs1894399, P = 0.0036, 0.000019, 0.0003 respectively). Two new loci related to AS , ARTS I and IL23R, were reported in 2007 by the Wellcome Trust Case Control Consortium. The authors have genotyped 14,436 nonsynonymous SNPs and 897 major histocompatibility complex (MHC) tag SNPs from 1,000 independent cases ofAS. They also claim that IL23R maybe a common susceptibility factor for the major "seronegative" diseases.! " There are genetic evidences suggesting the involvement of the enzyme ectonucleotide pyro phosphatase I (ENPP I) in ectopic calcification disorders such as the OPLL of the spine 129 and arterial calcification .':"The normal activity ofthis enzyme, a membrane bound glycoprotein which produces inorganic pyrophosphate from ATP, inhibits pyrophosphate related calcification and mineralization. The ENPPI gene maps to chromosome 6 (6q22-q23) and comprise 25 exons that encode a protein expressed in a variety oftissuesincluding matrix vesiclesofbone and cartilage . The mouse model ofOPLL, "tiptoe walking " (ttw) mouse, results from a naturally occurring nonsense mutation in the gene encoding ENPP I (Enppl-l-). It exhibits ossification ofthe spinal ligaments and articular cartilage calcification by hydroxyapatite deposition.' ?' Mutations in ENPPI cause the condition idiopathic infantile arterial calcification, a condition characterised by periarticular and arterial hydroxyapatite deposition, with low activity ofENPP I and low PPi levels.mlt is thus possible that gain -of-function mutations in this gene are involved in the development of CPPD CDD, though preliminary studies do not identify a major effect on CPPD CDD risk from common genetic variants of ENPPI (see below) .133 TNAP, an isozyme ofa family offour homologous human alkaline phosphatase genes, is present in the matrix vesiclesand has the ability ofhydrolyze PPi. In humans, this enzyme is enconded by the gene TNAP, that maps in chromosome 1, containing 12 exons. Deactivating mutations in the TNAP gene cause hypophosphatasia [MIM 241500 (Infantile form) , 146300 (Adult typel] , characterized by poorly mineralized cartilage and bones, spontaneous bone fractures, CPPD CDD and elevated concentrations ofPPi. 134 Using double knockout mice , lacking both TNAP and PC-I (ENPPl) genes, Hessle et al have shown evidences that TNAP and ENPPI are key regulators of bone mineralization suggesting that these genes are antagonistic regulators in this process.m To evaluate the hypothetical role ofENPPI and TNAP in CPPD-CDD, Zhang et al' 33screened variants ofENPPI and TNAP in 128 sporadic Caucasian CPPD CDD patients and 600 healthy controls. No significant difference was observed in the alleles or genotype frequencies between patients and controls suggesting that these genes are not major genetic determinants of sporadic CPPD CDD susceptibility in white Caucasians .P' A recent investigation performed by Bruges Armas et al in the Azores , Portugal, identified a seriesoffamilies characterised by both CPPD CDD and florid ligamentous calcification, consistent with the condition DISH [MIM 106400].49 The identification ofCPPD crystals in all synovial fluids examined suggests that these families have a form ofautosomal dominant CPPD deposition disease; however unlike previous reports of this condition, in these families, the CC was in all
46
Molecular Mechanisms ofSpondyloarthropathies
cases complicated by peripheral and axial enthesopathic calcifications. Linkage to chromosomes 5p15 (CCAL2) and chromosome 8p (CCALl) was excluded by microsatellire marker studies. Further genetic studies are currently being performed. It is well known that CPPD CDD has been considered as one of the great mimickers in Rheumatology. It remains to be seen if whether this is a new disease or, on the contrary, an unusual familial variant ofDISH or Cc. The cause of DISH remains unknown ; however, its frequent association with several metabolic conditions may givesome clues concerning its aetiopathogenesis. Some biochemical facts have been observed in DISH patients which includes a higher insulin and growth hormone levels, possibly explaining the osteoblast cell growth and proliferation,' >" a greater body mass index (BMI), a higher serum level of uric acid and a higher possibility of having diabetes mellitus. 137The association of DISH with diabetes mellitus has been much investigated in recent years.P" It is believed that the consequence of the previously mentioned biochemical facts in the etiopathogenesis of DISH is the abnormal cell growth/activity in the bony -ligamentous region. 4S.139 It is known that insulin like growth factor(IGF)-I stimulates alkaline phosphatase activity and Type II collagen in osteoblasts and that growth hormone can induce the local production ofIGF-I and IGF binding proteins in chondrocytes and osteoblasts. Matrix Gla Protein (MGP) is an 84 residue vitamin K dependent protein expressed in various tissues and upregulated in hone cells. The MGP gene spans 3.9 kb and consists of4 exons separated by 3 large intervening sequences which account for more than 80% of the gene.!" It is believed that M G P action includes the inhibition ofmineralization by suppression ofbone morphogenetic protein 2 (BMP-2) which is a potent osteogenic factor. Sarzi-Puttini et al reported higher serum MGP concentration levelsin male and female DISH patients, when compared with healthy controls. This group suggeststhat MGP may be a marker ofhyperostosis because it isproduced in larger amounts by patients with hyperostosis-inducing osteometabolic syndromes such as DISH.141 DISH has also been frequently associated with another paravertebral ligamentous ossification; the ossification ofthe posterior ligament ofthe vertebral bodies (OPLL) to the point that some studies state that OPLL can be seen in up to 50% of the patients with DISH.SS.142,143The genetic predisposition to OPLL and DISH is supported by several reports offamilial incidence of DISH and by studies relating a relative recurrence risk of up to 26.1 % in parents of OPLL patients and 28 .9% in siblings.r" There have been only a few genetic studies on DISH patients most of them based on its association with OPLL. There are contradictory results concerning the involvement of Human Leukocyte Antigen (HLA) alleles.!" Koga and colleagues performed a genetic linkage study in 91 OPLL affected sib pairs and report a predisposing locus for OPLL, on chromosome 6, close to the HLA complex. After analysis ofgenetic variants it was suggested that the candidate gene COL11A2 may be associated with the disease.P? Maeda et al continued to investigate the association ofCOL11A2 with OPLL and report an association ofa gender-specific COL11A2 haplotype with the disease. However, this association could only be seen in male patients."? To elucidate the genetics basis of OPLL Okawa et al investigated the ttw mouse, a naturally occurring mutant which exhibits similar anomalies in the spinal ligaments to the OPLL patients.148 Using a positional candidate-gene approach this group identified a nonsense mutation (Glycine 568 to a Stop Codon) in the Npps gene as the causation of ttw phenotype. The homologous human gene, currently known as ENPP 1, encodes an enzyme that is responsible for the production of inorganic pyrophosphate, a major inhibitor ofcalcification , thus regulating soft tissue calcification and bone mineralization.P' In 2001, novel mouse genes associated with ectopic calcification were isolated by differential display method using the same mouse model-ttw. 149 The association ofTGB1 with OPLL has been postulated due to its role in the growth and differentiation ofcartilage, endochondral and membrane bone and skin. Some studies have shown significant associations between variants of the gene and the prevalence of 0 PLL in the cervical spine .ISO A later study examined the relation between the previously associated polymorphism (T869 > G) and the radiologic characteristics of 0 PLL.Thisstudyconcluded thatthispolymorphism
Spondyloarthritis , Diffuse Idiopathic Sdeletal Hyperostosis (DISH) and Chondro calcinosis
47
is not a factor associated with the occurrence of 0 PLL but rather a factor related to the area of the ossified lesion."! To identify new genetic loci related to OPLL Tanaka and colleagues performed a genomewide linkage study with 142 affected sib-pairs. After multipoint linkage analysis, evidence oflinkage to OPLL was detected in 6 zones of different chromosomes. The best evidence of linkage was found on a zone of chromosome 21 that was extensively investigated. with single nucleotide polymorphisms (SNPs), for linkage and association studies. After case control comparison COL6Al was the gene most significantly associated with OPLL.152 Recently a large-scalestudy investigating the geneticsofOPLL was performed with the objective ofidentify new susceptibility genes for OPLL and to re-examine previously reported associations. Thirty fivecandidate genes were genotyped in sporadic 0 PLL patients (n = 711) and controls (n = 896). Significant associations were found in three genes; the most significant association was found in TGFB3 gene (P = 0.0004). Previously reported associations of COLI IA2, NPPS (ENPP I) and TGFBI were not reproduced in this study.!"
Conclusions AS, CPPD CDD. DISH and OPLL are four different entities characterised by ectopic calcification and ossification of soft tissues. in particular enthesis and ligaments. Very little is known about the genetic determinants involved in the causation of DISH and CPPD CDD but. due to its high prevalence. the better understanding ofthese processeswill surely improve human health care. This review highlighted what has been learnt from recent genetic studies about the processes involved in these different but somehow related ectopic calcification disorders. Although some clues may be evident in each particular disease. particularly with AS and CPPD CDD, genetic polymorphisms which could explain the clinical enthesopathic manifestations identified in all of them were not yet identified. With the possible exception of the ANKH gene which is CPPD CDD associated and also possibly associated with AS. no other genetic associations common to at least two of these diseases were identified. Previous reports have highlighted the resemblance ofspinal CPPD CDD. DISH and AS 89,154 and it can be difficult in some cases to differentiate these conditions. There are several reports in the literature focusing in the differential and sometimes conflicting diagnosis between AS and DISH and between AS and CPPD CDD. As already stated above. enthesopathic phenomenon are the main manifestations ofthe se diseases and particularly in the axial skeleton the y may be very similar. even in pathological specimens. Furthermore. the radiographic abnormalities that were ident ified by several authors in the axial component ofCPPD CDD. with the exception of disk calcification. have not been enough stressed. For this reason. these radiographic manifestations are frequently confused with other conditions and usually are not considered in the differential diagno sisofaxial enthesopathy. Several investigations. already cited in this chapter. report CPPD familieswith particular phenotypes which can be confounded with AS or with DISH. Ofparticular relevance are the reports ofReginato in the Chiloe Islands from Chile. van der Horst in Holland and Bruges Armas in the Azores. Portugal. It is interesting that patients from Argentina are the descendents ofthe islanders that in the 16th century admixed with Portuguese and Spanish, which came from Iberian Peninsula. The Azores (Terceira Island) where CPPD CDD/enthesopathic calcifications seems to be an endemic manifestation, in a familial context. were a secure port for ships demanding or coming from South America and were populated with people from different parts of Europe . namely an important group of Flemish whose connections with the Dutch are well know in the European history. It is possible that the descriptions in these three papers may be associated with migrations ofpeople across continents. These "pseudoespondillris" forms may show an axial phenotypic expression very similar to AS (Fig. 2) and are frequently diagnosed in a subset of early-onset familial CPPD CDD. These patients are symptomatic, have frequently peripheral arthritis and frequent pseudogout attacks. sometimes with complete articular destruction and ankylosis (Fig. 3).
48
Molecular M echanisms ofSpondyloarthropatbies
Axial di sease of CPPD CDD simulating AS or DISH was very well documented in a series ofpathology reports by Resnick et al. I55 O steolysis and subluxation ofthe cervical spine. changes in the apophyseal joints. "vacuum" phenomena and collapse of multiple disks and calcification of th e odont oid process. inter spinous bur sae. or ligamentum flavum are some of the findings wh ich show the resemblance with AS. According to the se authors. the y documented the widespread identification ofCPPD crystals in the spine and sacroiliac joints and there appears to be a definite relation ship between crystal deposition and adjacent structural joint damage in axial and peripheral joints and enthesis. The initial site ofcrystal depo sition are the outer fibers ofthe annulus fibrosus,
Figure 2. Dorsal spine calci fication simulating AS in a pati ent with CPPD COD.
Spondyloarthritis, Diffuse Idiopathic Sdeletal Hyperostosis (DISH) and Cbondrocalcinosis
49
producing syndesmophytes very similar to AS and further crystal deposition may be responsible for the other phenomenon like extensive disc calcification. The synovium lined apophyseal joints which are very frequently involved by the inflammatory process in AS, are also injured by the pro-inflammatory consequences ofthe deposition of CPPD crystals in these joints. Calcification ofthe cartilage, synovium and capsule, joint space narrowing, bone sclerosis and ankylosis were all described in these joints. The sacroiliac joint invariably involved in AS, allowing the differential diagnosis with other conditions, may have subchondral erosions and reactive sclerosis in CPPD
Figure 3. Ankle ankylosis. The patient had axial "pseudoankylosing spondylitis" and peripheral ankles, knees and wrists arthritis.
50
Molecular Me chanisms of Spondyloarthropatbies
CDD. simulatin g AS.62,155.156 Furthermore, inflammatory back pain has been described in some of these familie s, sometimes earlier in th e disease, when a radiographic diagnosi s is not po ssible. making the differential diagnosis more diflicult.49 In the se cases, peripheral arti cular manifestations are useful to make a differential diagnosis, because peripheral destruction. bone fragmentation and ankylosis are rare in AS. It is int eresting to stress that the majority of the authors that reported CPPD CDD famili al cases, beli eve that the disease is autosomal dominant. Because of this and also because some of the more severe CPPD cases reported in the literature were born in small villages or communities. where inbreed ing was common. it may be hypothesized that th ey repre sent the homozygous manifestations of this disease. Th e differential diagnosis bet ween DISH and AS has been the subject of discussion in multiple publications. A majority ofauthors emphasizes the sim ilarities ofthe spina l radiographic features between both diseases. Indeed. wh en a patient is diagnosed after 60 years old. the differential di agnosis may be difficult . In these ages, when fusion alread y occurred in AS, back pain may not be a relevant symptom and spine X-rays may show a similar picture in both diseases. DISH criteria were set to allow the differential diagnosis. The absence of ankylosis. erosion. sclerosis, or intra-articular osseous fusion of apophyseal joint, is an important criteria to diagnose DISH. Also. another criteria which should allow the differential diagnosis between DISH and AS is the sacro iliac di sease, always present in AS patients on their middle ages. However it is important to st ress that sacroiliac abnormalities may be found in DISH, although erosions and intra-articular bony ankylosis are rarely present . PelvicX-rays may be suspicious ofsacroiliac disease and CT scans may show partial superior bilateral or unilateral bridging osteophytes or fusion. 36 Other authors reported progressive sacroiliac obliteration and ossification of the sacroiliac joints in DISH, bur this findings are not usua1.37.157The peripheral musculoskeletal manifestations of DISH are usually not stressed in the literature. DISH is a diffu se disease and several entheseal extraspinal sites may be affected. They may also help in the differential diagnosis. bec ause there is usually a close clinical-radiographic correlation. New bone form ation is frequ entl y found in places where there is pain and sometimes swelling. These findings have a high pr evalence in heels and elbows where en thesophyres may be ident ified on Xvrays." Although it is clear that recent research revealed different molecular mechanisms underlying the pathogenesis of the se conditions. the molecular basis of the enthesopathic/ossification phenomena which is a common finding ofAS, CPPD CDD and DISH, is not know. Future research will address this issue and may possibly clarify the reason why some radiographic features ofthe se di seases are quite sim ilar.
References I. Calin A. Terminology, introdu ction . d iagnostic criteria and overview. in the spo ndylarthritides, A. Calin, ]. Taurog, ed. Oxford Universit y Press: Oxford 1998 :I -IS. 2. Calin A. Textbook of Rheumatology. 3rd ed. WN Kelly. ED H arris, S Rudy, C D Sledge. Philadelphia: WE Saunders 1985 :1021-1137. 3. C alin A. Oxford Textbook of Rheumatology. ed. Peter Maddis on 01. Woo P. ON Glass. Oxford: Oxford University Press. 1993; Vol. 1:681-90. 4. Calin A, Marder A, Marks S et ai, Familial aggregation of Reiter 's syndrome and ankylosing spond ylitis: a comparative study. J Rheurnatol 1984: 11:672 -7. 5. Brew erton OA, Caffrey M, Hart FO er al. Ankylosing spondylitis and H LA-B27. Lancet 1973; 1:1103. 6. Schlosstein LP. Terasaki PI, Bluestone Ret al. High association of an H LA antigen. W27, with ankylosing spondylitis. N EngJ Med 1973: 288 :744-6 . 7. van der Lind en SM, Valkenburg HA . Cats A. Evaluatio n of d iagnostic criteria for ankylosing spondylitis: a proposal for mod ification of New York criteria. Arthritis Rheum 1984; 27:36 1-8. 8. O ougados M. van der Linden S. Juhlin R et al. The european spondyloarthropathy study group : prelim inar y criteri a for the classification of spondyloarthropathy, Arthritis Rheum 1991: 34: 1218-27. 9. West F. The aeti ology of ankylosing spondylitis. Ann Rheum D is 1949: 8:143-8 . 10. Boyer GS. Lanier AP, Templin OW er al. Spond ylarthroparhy and rheumatoid arthri tis in alaskan, yupik eskimos. J Rheumarol 1990; 17:489-96.
Spondyloartbritis, Diffuse Idiopathic Sdeleta! Hyperostosis (DISH) and Cbondrocalcinosis
51
11. Gran JT. Husby G. Epidemiology of ankylosing spondylitis. Rheumatology. AJS Marc C Hochberg. Josef S Smolen. Michael E Weinblatt. Michael H Weisman. eds. London: Elsevier Limited. 2003 ; Vol. 2:1153-1159. 12. Gofion JP. Robinson HS. Trueman GE. Ankylosing spondylitis in a canadian indian population. Ann Rheum Dis 1966; 25:525-27. 13. van der Linden SM, valkcnburg HA, de Jongh BM ec al. The risk of developing ankylosing spondylitis in HLA-B27 positive individuals . Arthritis Rheum 1984; 27:241-49. 14. Bruges Armas J. Lima C. Peixoto MJ et al. Prevalence of spondyloarthritis in terccira, azores: a population based study. Ann Rheum dis 2002; 61:551-53. IS. Gran JT, Husby G, Hordvik M. Prevalence of ankylosing spondylitis in males and females in a young middle-aged population in Tromso, Northern Norway. Ann Rheum Dis 1985; 44:359-67. 16. Johnsen K, Gran JT, Dale K et al. The prevalence of ankylosing spondylitis in a norwegian sami population . J Rheurnatol 1992; 19:1591-94. 17. Thorsby E. Bratlie A, Teisberg P. HLA polymorphism of norwegian Laps. Tissue Antigens 1971; 1:137-46. 18. Calin A. Fries JE Striking prevalence of ankylosing spondylitis in "healthy" w27 positive males and females. A controled study. N EngJ Med 1975; 293:835-39. 19. Bruges Armas J. Espondilartrites e doencas associadas. Heterogeneidade genetica e sua expressao. PhD Thesis: Porto 2001 :65-67 . 20. Sieper J, Braun J. Rudwaleit M et al. Ankylosing spondylitis : an overview. Ann Rheum Dis 2002; 61 (Suppl) :iii 8-iii 18. 21. Hochberg MC. Borenstein DG, Arnett Fe. The absence of back pain in classical ankylosing spondylitis. Johns Hopkins Med J 1978: 143:181-83. 22. Khan MA. Rheumatology. 3rd ed, AJS Marc C Hochberg , JosefS Smolen. Michael E Weinblatt. Michael H Weisman. London: Elsevier Limited, 2003: 3:1161-1181. 23. Resnick D, Niwayama G. Ankylosing spondylitis, in diagnosis of bone and joint disorders, Resnick D, ed, WB Saunders : Philadelphia 1995:1008-74. 24. Khan MA. Ankylosing spondylitis. clinical aspects. in the spondylarrhritides, A Calin, J Taurog, ed. New York: Oxford University Press, 1998:27-40. 25. McGonagle D, Khan MA. Marzo Ortega H et al. Enthesiris in ankylosing spondylitis and related spondyloarthropathies. Curr Opin ion Rheurnatol 1999: 11:244-250. 26. Will R. Palmer R. Bhalla AK et al. Osteoporosis in early ankylosing spondylitis: a primary pathological event? Lancet 1989; 2:1483-5 . 27. Murray GC. Persellin RH. Cervical fracture complicating ankylosing spondylitis . A report of eight cases and review of the literature. Am J Med 1981; 70:1033-41. 28. O'Neill TW, Bresnihan B. The heart in ankylosing spondylitis. Ann Rheum Dis 1992; 51:705-6. 29. Resnick D.. Niwayama G. Ankylosing Spondylitis . D Resnick ed. WB Saunders Company. 1995: 1008-1074. 30. Salonen DC. Brower Ae. Seronegative spondyloarthropachies imaging. in Rheumatology. AJS Marc C Hochberg. Josef S Smolen, Michael E Weinblatt. Michael H Weisman, ed. London, Elsevier Limited: 2003 :1193-1203. 31. Mau W, Zeidler H, Mau R et al. Clinical features and prognosis of patients with possible ankylosing spondylitis : results of a 10 year follow-up. J Rheumatol 1988; 15:1109 -1114. 32. Kozin F. Carrera GF. Ryan LM et al. Computed tomography in the diagnosis of sacroiliitis. Arthritis Rheum 1981: 24:1479-85. 33. Braun J. Bollow M. Eggens U et al. Use of dynamic magnetic resonance imaging with fast imaging in the detection of early and advanced sacroiliitis in spondyloarthropathy patients . Arthritis Rheum 1994 ; 37:1039-45 . 34. Forrestier J, RQJ. Senile ankylosing hyperostosis of the spine. Ann Rheum Dis 1950; 9:321-30 . 35. Resnick D. Niwayama G. Radiographic and pathologic features of spinal involvement in diffuse idiopathic skeletal hyperosto sis (DISH). Radiology 1976; 119:559. 36. Durback MA. Edelstein E. Schumacher R. Abnormalities of the sacroiliac joints in diffuse idiopathic skeletal hyperostosis: demonstration by computed tomography. J Rheurnatol 1988; 15:1506-10. 37. Weisz GM . Green L. Progressive sacro-iliac obliteration in forescier disease. Intern Orthopedics 1986; 10:47-51. 38. Boachie-Adjei O. Bullough P. Incidence of ankylosing hyperostosis of the spine (forestier's disease) at autopsy. Spine 1987; 12:739. 39. Harris J. Carter AR, Glick EN et al. Ankylosing hyperostosis I. Clinical and radiological features 1974 ; 33:210-15. 40. Bloom RA. The prevalence of ankylosing hyperostosis in a Jerusalem population with description of a method of grading the extension of the disease. Scand J Rheum 1984: 13:181-89.
52
Molecular Mechanisms oJSpondyloarthropathies
41. Mader R. Clinical manifestations of diffuse Idiopathic skeletal hyperostosis of the cervical spine . Sem Arthritis Rheum 2002; 32(2):130-35 . 42 . Resnick D , Shapiro RF, Wiesner KB et al. Diffuse idiopathic skeletal hyperostosis (DISH). Sem Arthritis Rheum 1978(7):153-187. 43 . Mara S. Fortin PRo Fitzcharles M et al. A controlled study of diffuse idiopathic skeletal hyperostosis. Medicine 1997; 76(2):104-17. 44 . Littlejohn GO. Urowitz MB. Peripheral enrhesop athy in diffuse idiopathic skeletal hyperostosis (DISH). J Rheurnatol 1982; 9:568-72. 45 . Daragon A. Czernichow P. Louvel JP et al. Vertebral hyperostosis and diabetes mellitus : a case-control study. Annals of the Rheumatic Diseases 1995; 54 :375-8. 46 . Schramm P. Schm idt KL. Rude]. Familial occurrence of multiple basalomas associated with hyperostotic spondylosis. Z Hauktr 1985 ; 60(10):785-6, 791-2. 47 . Gorman C , Jawad ASM , Chikanza 1. A family with diffuse idiopathic skeletal hyperostosis . Ann Rheum Dis 2005; 64:1794-95. 48 . Beardwell A . Familial ankylosing vertebral hyperostosis with tylosis . Ann Rheum Dis 1969; 28 :518-23. 49 . Bruges-Armas JB, CoutO AR. Timms A et al. Ectopic calcification in the azores-clinical and radiological manifestations of DISH and chondrocalcinosis families. Arthritis Rheum 2006 ; 54(4):1340-1349. 50. Havelka S, Faberova R. Gatterova J et al. Familial incidence of diffuse idiopathic skeletal hyperostosis. Vnitr Lek 1990; 36(7):680-4. 51. De Bandt M , Meyer O. Fusrer ]M ct al. Ossification of the posterior longitudinal ligament , diffuse. idiopathic skeletal hyperostosis. abnormal retinol and retinol binding protein : a familial observation. J Rheumatol 1995 ; 22(7):1395-8. 52. Abiteboul M, Mazieres B Menard H . 2 New familial cases of ankylosing vertebral hyperostosis. Rev Rhum Mal Osteoartic 1985 ; 52(1l):645-7. 53. Svab V. To the extent, development and genetic predisposition to the so-called diffuse idiopathic skeletal hyperostosis (osreodesrnosis). Cas Lek Cesk 1981 ; 120(21):625-8. 54. Matsunaga S, Sakou T. Epidemiology of ossification of the posterior longitudinal ligament. in ossification of the posterior longitudinal ligament, ST Yonenobu K, Ono K , Editor. Springer: Tokyo 1997:11-7 . 55. Resnick D, Guerra JJ. Rob inson CA et al. Association of diffuse id iopathic skeletal hyperostosis (DISH) and calcification and ossification of the posterior longitudinal ligament. Am J Roentgenol 1978; 131 :1049-53. 56. Zitnan D, Siraj S. Mnohopocenma familiarna kalcifikacin articularnych chrupiek. Bratisl Lek Lisry 1958; 38:217-28. 57. McCarty DJ. The heberden oration 1982. Cry stals, joints and consternation. Ann Rheum Dis 1983 ; 42 :243-53. 58. Rosenthal AK, Ryan LM . Calcium pyrophosphate crystal deposition disease. pseudogout and articular chondrocalclnosis, in arthritis and allied conditions. W J Koopman. LW Moreland. ed. Lippincott Williams and Wilkins: Philadelphia 2005:2373-2396. 59. McCarty DJ, Gatter RA. Pseudogout syndrome. Bull Rheum Dis 1964; 14:331. 60 . Kohn NN. Hughes RE. McCarty DJ et al. The significance of calcium phosphate crystals in the synovial fluid of arthritic patients: "the pseudogout syndrome" Identification of crystals II. Ann Int Med 1962; 56:738 . 61. Resnick D , Niwayama G, Goergen TG et al. Clinical, radiographic and pathological abnormalities in calcium pyrophosphate dehydrate deposition disease (CPPD) pseudogout. Diagn Radiol 1977; 122:1-15 . 62. Martel \V, McCarter DK , Solsky MA et al. Further observations on the arthropathy of calcium pyro phosphate crystal deposition disease. Radiology 1981 ; 141 :1-15 . 63. McCarty DJ. Diagnostic mimicry in arthritis: patterns of jo int involvement associated with calcium pyrophosphate dehydrate crystal deposits. Bull Rheum dis 1975 ; 25 :804-9. 64 . Gerster JC, Baud CA. Lagier Ret al. Tendon calcifications in chondrocalcinosis. Arthritis Rheum 1977; 20 :717-22 . 65. Lagier R. Boivin G, Gerster J. Carpal tunnel syndrome associated with mixed calcium pyrophosphate dihydrate and apatite crystal deposition in tendon sheath. Arthritis Rheum 1984; 27:1190-95. 66. Ellman MH. Krieger MI. Brown N. Pseudogout mimicking synovial chondromatosis.] Bone Joint Surg (AM) 1975 ; 57:863. 67 . Hensley CD. Lin JJ. Massive intrasynovial deposition of calcium pyrophosphate in the elbow. J Bone Joint Surg (AM) 1984; 66 :133 . 68. Zitnan D. Siraj S. Chondrocalcinosis articularis, Ann Rheum Di s 1963; 22 :142-70. 69. Gaudreau A. Camerlain M. Pibarot ML et al. Familial articular chondrocalcinosis in ~ebec. Arthritis Rheum 1981; 24(4) :611-615.
Spondyloartbritis, Diffuse Idiopathic Sdeletal Hyperostosis (DISH) and Cbondrocalcinosis
53
70 . Richardson BC, Chafetz NI , Ferrel LD et al. Chondrocalcinosis in a mexican-american family. Arthritis Rheum 1983 ; 26 :1387-96. 71. Hamza M. Meddeb N, Bardin T. Hereditary chondrocalcinosis in a tunisian family. Clin Exp Rheum 1992; 10:43-49 . 72 . Balsa A, Martin-Mola E. Gonzalez T et aI. Familial articular chondrocalcinosis in spain . Ann Rheum Dis 1990 ; 49 :531-535. 73 . Eshel G. Gulik A. Halperin N et aI. Hereditary chondrocalcinosis in an Ashkenazi jewish family. Ann Rheum Dis 1990; 49 :528-530 . 74. Doherty M. Hamilton E, Henderson Jet aI. Familial chondrocalcinosis due to calcium pyrophosphate dih ydrate crystal deposition in English Families. Br J Rheum 1991; 30:10-15 . 75. Bjelle A. Edvinson U. Hagstam A. Pyrophosphate Arthropathy in Two Swedish Families. Arthritis Rheum 1982; 25(1) :66-74. 76. Reginato A. Articular chondrocalcinosis in the Chiloe Islanders. Arthritis Rheum 1976 , 19(5uppl 3) :395-404. 77. Rodriguez-Valverde V, Tinture T. Zuniga M et aI. Familial Chondrocalcinosis-Prevalence in Northern Spain and Clinical Features in Five Pedigrees. Arthritis Rheum 1980, 23(4) :471-478. 78. Van der Korst JK . Gerards J. Driessens FM. A hereditary type of idiopathic articular chondrocalcinosis. Survey of a pedigree . Am J Med Genet 1974,56(3):307-14. 79. McCarty DJ. Calcium pyrophosphate dihydrate cr ystal depo sition disease. Arthritis Rheum 1976; 19(5uppl):272-85 . 80. Rodriguez-Valverde V. Zuniga M. C asanueva B et al. Hereditary art icular chondrocalcinosis. Am J Med 1988; 84:101-106. 81. Dapica MPF. Gomez -Reino JJ. Familial chondrocalcinosis in the Spanish popul ation . J Rheumatol 1986; 13:631-3 . 82. Zitnan D. Sitaj S. Natural course of articular chondrocalcinosis. Arthritis Rheum 1976; 19(5uppl) :363-90. 83. Okazaki T. Saito T. Mitomo T et al. Pseudogour: Clinical observations and chemical analyses of deposyts. Arthritis Rheum 1976; 19:293-305. 84. Magnet JL. Strauss J. Dentan S et al. Joint ankylosis in familial chondrocalcinosis. Presse Med 1986, 15(23):1105. 85. Gaucher A. Pourel J. Faure G et al. Diffuse hereditary articular chondrocalcinosis. Rev Rhum Mal Osteoartic 1977; 44(10) :589-97 . 86. Brem J. Vertebral ankylosis in a pat ient with hereditary chondrocalcinosis: a chance association? Arthritis Rheum 1982 ; 25(10) :1257-63. 87. Sharp J. Heredo-familial vascular and articular calcification. Ann Rheum Dis 1954:15-27 . 88. Ho HH. Yang WE . Luo SF. Ankylosing spondylitis with severe chondrocalcinosis : a case report. C hanggeng Yi Xue Za Zhi 1999; 22(3 ):536-40 . 89. Layfer LF. Katz R. Golde n H . C hondrocalcinosis simulating ankylosing spondylitis. JAMA 1978 ; 240(1):55-6. 90. Ander son He. Molecular biology of matrix vesicles. Clin Orthop Rei Res 1995 ; 314 :266-280. 91. Derfus B. Kranendonk S. Camacho N et al. Human osteoarthritic cartilage matrix vesicles generate both calcium pyroph osphate dihydr ate and apatite in vitro. Ca lcif Tissue Int 1998; 63 :258-262. 92. Picher M. Graff RD. Lee GM. Extracellular nucleotide metabolism and Signaling in the pathophysiology of articular cartilage. Arthritis Rheum 2003 ; 48(10) :2722-2736. 93. Ciancaglini P, Simao AMS. Camolezi FL et al. Contribution of matrix vesicles and alkaline phosphatase to ectopic bone form ation . Braz J Med Bioi Res 2006, 39:603-610. 94. Fleisch H . Bisaz S. Mechanism of calcification : inhibitory role of pyrophosphate. Nature 1962 ; 195:911. 95. Terkelraub RA. Inorganic pyrophosphate generation and disposition in pathophysiology. Am J Physiol Cell Physiol 2001 ; 281 :CI-Cl1. 96. Silcox DC. McCartyJr DJ. Elevated inorganic concentrations in synovial fluids in osteoarthritis and pseudogouc .} Lab Clin Med 1974; 83(4) . 97. C ohen MMJ. The new bone biology : Pathological. molecular and clinical correlated. Am J Med Genet 2006 ; Part A(140A ):2646-2 706. 98. Couto R. Brown MA . Genetic factors in the pathogenesis of CPPD crystal deposition disease. Curr Rheumatol Rep 2007 , 9(3) :231-6 . 99. Reginato A. Valenzuela F, Martinez V er al. Polyarticular and familial chondrocalcinosis. Arthritis Rheum 1970; 13(3) :197 -213. 100. Pons-Estel BA. Gimenez C . Sacnun M et aI. Familial osteoarthritis and Milwaukee shoulder associated with calcium pyrophosphate and apatite crystal deposition . J Rheumarol 1999 ; 27(2) :471-480. 101. Baldwin CT. Farrer LA, Adair R er al. Linkage of early-onset osteoarthritis and chondrocalcinosis to human chromosome 8q. Am J Hum Genet 1995; 56:692-697.
54
Molecular Mechanisms ofSpondyloarthropathies
102. Hughes AE, McGibbon D, Woodward E er al. Localisation of a gene for chondrocalcinosis to chromosome 5p. Hum Mol Genet 1995 ,4(7) :1225-1228. 103. Andrew LJ, Brancolini Y, Serrano de la Pena L er aL Refinement of the chromosome 5p locus for familial calcium pyrophosphate dihydrate deposition disease. Am J Hum Genet 1999, 64: 136-145. 104. Ho AM , Johnson MD, Kingsley DM. Role of the mouse ank gene in control of tissue calcification and arthritis. Science 2000 , 289 :265-289. 105. Timms AE, Zhang Y, Russel RG et al. Genetic studies of disorders of calcium crystal deposition . Rheumatol 2002, 41(7) :725-9 . 106. Williams CJ, Zhang Y, Timms A et aL Autosomal dominant familial calcium pyrophosphate dihydrate deposition disease is caused by mutation in the transmembrane protein ANKH. Am J Hum Genet 2002,71:985-991. 107. Pendleton A, Johnson MD, Hughes AE et al. Mutations in ANKH cause chondrocalcinosis . Am J Hum Genet 2002, 71 :933-940. 108 . Williams CJ , Pendleton A, Bonavita G et al. Mutations in the amino terminus of ANKH in two US families with calcium pyrophosphate dih ydrate crystal depos ition disease. Arthritis Rheum 2003 , 48(9) :262 7-2631. 109. Zhang Y, Johnson K, Russel RGG et al. Association of sporadic chondrocalcinosis with a -4 basepair G -to-A transition in the 5' untranslared region of ANKH that promotes enhanced expression of ANKH protein and excess generation of extracellular inorganic pyrophosphate. Arthritis Rheum 2005 , 52 :1110-1117. 110. Gurley KA . Reimer RJ. Kingsley DM. Biochemical and genetic analysis of ANK in arthritis and bone disease. Am J Hum Genet 2006 , 79 :1017-1029. Ill. Zaka R, Stokes D, Dion AS et al. P5L mutation in ANK results in an increase in extracellular inorganic pyrophosphate during proliferation and nonmineralizing hypertrophy in stably transduced ATDC5 cells. Arthritis Re Ther 2006, 8(6) :RI64. 112. Ryan LM , Kurup 1. Cheung HS. Stimulation of cartilage inorganic pyrophosphate elaboration by ascorbate. Matrix 1991, 11(4) :276-281. 113 . Tsui FWL, Tsui HW; Cheng EY et al. Novel genetic markers in the 5'-Ranking region of ANKH are associated with ankylosing spondylitis. Arthritis Rheum 2003 ,48(3) :791-797. 114 . Timms AE, Zhang Y, Bradbury L et aL Investigation of the role of ANKH in ankylosing spondylitis. Arthritis Rheum 2003, 48:2898-2902. 115. Tsui HW'; Inman RD, Paterson AD et al. ANKH variants associated with ankylosing spondylitis: gender differences. Arthritis Res Ther 2005 ,7(3):513-25. 116. MacLean IL, Iqball S, Woo P et aL HLA-B27 subtypes in the spondarthropathies, Clin Exp Immunol 1993 ,91 :214-9. 117 . Lopez-Larrea C , Sujirachato K, Mehra NK ct al. HLA-B27 subtypes in asian patients with ankylosing spondylitis. Evidence for new associations . T issue Antigens 1995,45:1698-76. 118. Armas JB , Gonzalez S, Martinez-Borra J er al. Susceptibility to ankylosing spondylitis is independent of the Bw4 and Bw6 epitopes of HLA-B27 alleles. Tissue Antigens 1999 : 53:237-243. 119 . Garcia F, Rognan D. Lamas JR er al. An HLA-B27 polymorphism (B'2710) th at is critical for T-cell recognition has limited effects on peptide specificit y. Tissue Antigens 1998 : 51:1-9. 120 . Garcia Fernandez S, Gonzalez S, Martinez-Borra J et aL New insights regarding HLA-B27 diversity in the Asian population. Tissue Antigenes 2001 : 58 :259-62. 121. Gonzalez-Roces S, Alvarez MY, Gonzalez S er aL HLA-B27 polymorphism and worldwide susceptibility to ankylosing spondylitis. Tissue Antigens 1997 , 49 :116-123. 122 . Tamouza R, Mansour I, Bouguacha N et al. A new HLA-B'27 allele (B'2719) identified in a Lebanese patient affected with ankylosing spondylitis. Tissue Antigens 2001: 58(1):30-3 . 123. Lopez-Larrea C, BlancoGeiaz MA, Torre-Alonso JC et aI. Contribution of KIR3DLl /3DSI to ankylosing spondylitis in human leukocyte antigen-B27 Caucasian populations. Arthritis Res Ther 2006: 8(4) :RI01. 124 . Brown MA, Pile KD , Kennedy LG et aL A genome wide-screen for susceptibility loci in ankylosing spondylitis. Arthritis Rheum 1998: 41 :588-95. 125. Brown MA . Brophy S, Bradbury L er aL Identification of major loci controlling clinical manifestations of ankylosing spondylitis. Arthritis Rheum 2003: 43 :2234-9. 126. Laval SH, Timms A, Edwards S et al. Whole-genome screening in ankylosing spondylitis : evidence of nonMHC genetic-susceptibility loci. Am J Hum Genet 2001 : 68 :918- 26. 127. Timms AE. Crane AM , Sims AM et aL The interleukin 1 gene cluster contains a major susceptibility locus for ankylosing spondylitis . Am J Hum Genet 2004, 7 5(4 ):587-95. 128. Wellcome Trust Case control Consortium: Austr alo-Anglo-Arnerican Spondylitis Consortium (TASC), e.a., Association scan of 14,500 nonsynonimous SNPs in four diseases identifies autoimmunity variants . Nat Genet 2007: 39(11 ):1329-37.
Spondyloarthritis, Diffuse Idiopathic Sdeletal Hyperostosis (DISH) and Cbondrocalcinosis
55
129. Nakamura I. Ikegawa S. Okawa A et al, Association of the human NPPS gene with ossification of the posterior longitudinal ligament of the spine (OPLL). Hum Genet 1999; 104:492-497. 130. Ruesch F. Ruf N. Vaingankar S et aI. Mutations in ENPPI are associated with "idiopathic" infatile arterial calcificaction . Nature Genet 2003; 34(4) :379-381. 131. Okawa A. Nakamura I. Goto S et aI. Mutation in Npps in a mouse model of ossification of the posterior longitudinal ligament of the spine. Nat Genet 1998; 19:271-273. 132. Ruesch F. Ruf N. Vaingankar S et aI. Mutations in ENPPI are associated with "idiopathic" infatile arterial calcification . Nat Genet 2003 : 34(4) :379-381. 133. Zhang Y. Brown MA. Peach C et al. Investigation of the role of ENPP 1 and TNAP genes in chondrocalcinosis. Rheumatol 2006: doi:l0.1093/rheumatology/kel338. 134. Mornet E. Taillandier A. Peyramaure S et al. Identification of fifteen novel mutations in the tissue-nonspecific alkaline phosphatase (TNSALP) gene in European patients with severe hypophosphatasia . Eur J Hum Genet 1998; 6:308-314. 135. Hessle L. Johnson !CA. Anderson HC et al. Tissue-nonspecific alkaline phosphatase and plasma cell membrane glycoprotein-l are central antagonistic regulators of bone mineralization. PNAS 2002; 99(14):9445-9449. 136. Denko CW; Boja B. Moskovitz RW. Growth promoting peptides in osteoarthritis and diffuse idiopathic skeletal hyperostosis-insulin. insulin-like growth factor. growth hormone. J Rheumatol 1994: 21:1725-30. 137. Kiss C. Szilagyi M. Paksy A er al. Risk factors for diffuse idiopathic skeletal hyperostosis: a case control study. Rheurnatol 2002; 41 :27-30 . 138. Wyatt L. Ferrance R]. The musculoskeletal effects of diabetes mellitus. J Can Chirop Assoc 2006: 50(1) :43-50 . 139. Miendany E. Wassif G. Baddini M . Diffuse idiopathic skeletal hyperostosis (DISH) is it of vascular (aetiology). Clin Exp Rheum 2000 ; 18:193-200. 140. Cancela L. Hsieh C . Francke U et al. Molecular structure. chromosome assignment and promoter organization of the human matrix Gla protein gene. J Bio Chern 1990; 265 :15040-8. 141. Sarzi-Puttini P. Bevilacqua M, Atzeni E Matrix GLA protein :evidence of persistently increased concentrations in DISH patients. Arthritis Rheum 48:S534. 142. Hukuda S. Mochizuki T. Ogata M er al. The pattern of spinal and extraspinal hyperostosis in patients with ossification of the posterior longitudinal ligament and the ligament Aavum causing myelopathy. Skelet Radiol 1983 : 10:79-85. 143. Ehara S. Shimamura T, Nakamura Ret al. Paravertebral Ligamentous Ossification : DISH. OPLL and OLE Eur J Radiol 1998 ; 27:196-205. 144. Terayama K. Genetic studies on ossifcation of the posterior longitudinal ligament of the spine . Spine 1989; 14:1184-91. 145. Denko CW; Boja B. Moskovitz RW. Growth promoting peptides in osteoarthritis and DISH. J Rheumatol 1994; 21:1725 . 146. Koga H. Sakou T. Taketom i E et al. Genetic mapping of ossification of the posterior longitudinalligamerit of the spine . Am J Hum Genet 1998; 62:1460-7 . 147. Maeda S, Koga H. Matsunaga S et al. Gender-specific haplotype association of collagen alpha2 (XI) gene in ossification of the posterior longitudinal ligament . J Hum Genet 2001 ; 46(1 ):1-4. 148. Okawa A. Ikegawa S. Nakamura I et al. Mapping of a gene responsible for twy (tip-toe walking Yoshimura) , a mouse model of ossification of the poserior longitudinal ligament of the spine (OPLL). Mamm Genome 1998 ; 2:155 -6. 149. Koshizuka Y. lkegawa S. Sano M et al. Isolation of novel mouse genes associated with ectopic ossification by differential display method using trw, a mouse model for ectopic ossification . Cytogenet Cell Genet 2001 ; 94:163-8. 150. Kamiya M. Harada A. Mizuno M et al. Association between a polymorphism of the transforming growth factor-beta 1 gene and the genetic susceptibilit y to ossification of the posterior longitudinal ligament in Japanese patients. Spine 2001 : 26(11):1264-6 . 151. Kawaguchi Y. Furushima K, Sugimori K et al. Association between polymorphism of the transforming growth factor-beta 1 gene with the radiologic characteristic and ossification of the posterior longitudinal ligament. Spine 2003: 28(13 ):1424-6 . 152. Tanaka T. Ikari K. Furushima K et al. Genomewide Linkage and Linkage Disequilibrium Analysis Identify COL6AI. on Chromosome 21. as the Locus for Ossification of the Posterior Longitudinal Ligament of the Spine . Am J Hum Genet 2003 ; 73:812-22. 153. Horikoshi T, Maeda K, Kawagushi Y et al. A large-scale genetic association study of ossification of the posterior longitudinal ligament of the spine. Hum Genet 2006 ; 119(6) :6 11-6. 154. Richards AJ. Hamilton EB. Spinal changes in idiopathic chondrocalcinosis articularis , Rheumatol Rehabil 1976 ; 15(3) :138-42.
56
Molecular Mechanisms ofSpondyloarthropathies
155. Resnick D, Pineda C. Vertebral involvement in calcium pyrophosphate dihydrate crystal deposition disease. Radiology 1984; 153:55-60. 156. Littlejohn GO, Baron M, Urowitz MB. Sacroiliac joint abnormalities in calcium pyrophosphate crystal deposition disease. Rheurnatol Int 1982; 1:195-198 . 157. Brigode M, Francois R], Dory MA. Radiological study of the sacroiliac joints in vertebral ankylosing hyperostosis. Ann Rheum Dis 1982; 41 :225-231.
CHAPTER 4
The Enthesis Organ Concept and Its Relevance to the Spondyloarthropathies Michael Benjamin*and Dennis McGonagle Abstract Characteristic feature ofth e spondyloarthropathies is inflammation at tendon or ligament attachment sites. This has traditionally been viewed as a focal abnormality. even though the inflammatory reaction intrinsic to enthesitis may be quite exten sive. We argue that the diffuse nature of the pathology is best understood in the context of an 'enthesis organ concept: This highlights the fact that stress co ncent ration at an insertion site involves not only the enthesis itself. but neighbouring tissues as well. The archet ypal enthesisorgan is that ofthe Achilles tendon where intermittent contact between tendon and bone immediately proximal to the enthesis leads to the formation offibrocartilages on the deep surface ofthe tendon and on the opposing calcaneal tub erosity. but similar functional modifications are widespread throughout the skeleton. Many enthe seshave bursae and fat near the insertion site and both of these serve to promote frictionless movement. Collectively, the fibro carcilages, bursa, fat pad and the enthesis itself con stitute the enthesisorgan. However, it also includes both the immediately adjacent trabecular bone networks and in some cases deep fascia. The concept of a synovio-entheseal complex (SEC ) and of a 'fun ctional enthesis' are com plimentary to that of an enthesis organ and also have important implications for understanding spondyloarthropathy. The SEC concept emphas izes the interdependence between synovial membran e and enthe seswithin enthesis organs. It draws att ent ion to the fact that one component (the enthesis) is prone to microdamage and th e oth er (the synovium) to inflammation. If an enthesis is dam aged. any ensuing inflammato ry reaction is likely to occur in the synovium. The con cept ofa 'func tional enthe sis' serves to emphasise anatomical, biomechanical and pathological features that are shared between true fibrocartilaginous enrheses and regions proximal to the attachment sites themselves where tendons or ligaments wrap around bon y pulleys. Such 'wrap-aroun d region s' are well documented sites of pathology in SpA-for teno synovitis is a recognized feature . Stress concentration at the enthesis itself is dissipated at many sites by fibrous connections between one tendon or ligament and another, close to the insertion site . At a microscopic level, enthesisfibrocartilage is ofparamount importance in ensuring that fibre bending ofthe tendon or ligament is not focused at the hard tissue interface. Normal enthesis organs are avascular in their fibrocartilaginous regions. but tissue microdamage to entheses iscommon and appears to beassociated with tissue repair responses and vessel ingrowth. This makes the enthesis organ a site where adjuvant molecules derived from bacteria may be preferentially deposited. This microdamage and propensity for bacterial molecule deposition in the context of genetic factors such as HLA-B27 appears to lead to th e characteristic inflammator y changes of AS.
A
' Correspo ndi ng Author: Michael Benjam in-School of Biosciences, Cardiff University, Museum Avenue, Cardiff CFl 0 3US, W ales, UK. Email : benjami
[email protected]. uk
Molecular Me chanisms of Spondyloartbropathies, edited by Carlos Lopez-Larrea and Roberto D iaz-Pefia. ©2009 Landes Bioscience and Springer Science+Business Media.
58
Molecular Mechanisms of Spondyloarthropatbies
Understanding the enthesis organ concept helps to explain synovitis and osteitis in spondyloarthroparhy, An appreciation of the complex anatomy of 'articular enthesis organs' (e.g., that associated with the distal interphalangeal joints) is helpful in understanding disease patterns in psoriatic arthritis. In this chapter, we review the extent and types ofenthesis organs and show how a patho-anatomic appreciation of these structures leads to a new platform for understanding the pathogenesis of SpA.
Introduction The region where a ligament, tendon or joint capsule attaches to bone (l.e., the enthesis) is known to be a region ofhigh stress concentration, because it is the meeting point between two tissues (soft and hard) with contrasting physical properties. Indeed strain levels at tendon or ligament entheses can be up to 4 times those which occur in midsubstance.' Stress concentration is further increased because the angle at which the soft and hard tissues meet (the 'insertional angle ') changes with joint movement. The dynamic changes in this angle occurring with locomotion increase the risk that tendon or ligament collagen fibres will fray at the bony interface. Because stress concentration is such an issue at attachment sites, the whole architecture of enrheses is geared towards stress dissipation. This often necessitates a functional adaptation in the neighbouring structures as well and the term 'enthesis organ' was coined by the present authors in recognition of this .2•3 lt denotes a collection of structures that are related to the attachment site ofa tendon, ligament or joint capsule , which together with the enthesis itselfserve to dissipate stress concentration at the soft-hard tissue interface. The significance ofthe enthesis organ concept for understanding enthesopathies has been quickly recognized by clinicians. As Huber er a14 have remarked, it shifts the traditional view ofenthesopathies from being regarded as focal disorders of entheses, to multifocal problems which affect a much wider area. This is of particular relevance of course to the spondyloarrhropathies (SpA), as enthesitis in SpA is known to be associated with diffuse changes in adjacent rlssucs.Y'Thus, at the Achilles tendon enthesis in SpA patients, there may be changes not only at the enthesis itself, but also effusions in the retrocalcaneal bursa and calcaneal erosions on the neighbouring superior tuberosity? The value of the enthesis organ concept to understanding the pathology of psoriatic arthritis has been particularly well recognised'" and rapid strides have been made towards improving the magnetic resonance imaging of the different components of enthesis organs, through the use of ultrashort TE (UTE) pulse sequences .'? The purpose of the present review therefore is to explain the academic thinking which underpins the enthesis organ concept and to show how this can be used to better conceptualise the pathology of SpA.
Enthesis Organ Structure
The ArchetypalEnthesis Organ Canoso has aptly described the bony attachment ofthe Achilles tendon as being the "premiere enthesis",' and his perceptive awareness ofthe importance ofneighbouring structures to enthesis function, has critically influenced the conceptual understanding ofenthesitis in SpA. The Achilles tendon flattens from front to back, but expands from side to side before attaching to the calcaneus (Fig. 1), so that stress concentration is spread from the medial to the lateral side of the calcaneus by a marked flaring of the tendon. Stress is further reduced by an imperceptible blending oftendon fibres with neighbouring periosteum and hence with the plantar aponeurosis in the sole of the foot." It is of key importance for understanding the enthesis organ concept, to appreciate that the Achilles tendon attaches to the middle and lower thirds of the posterior surface of the calcaneus (Fig. 1). This 'frees' the upper third (the superior tuberosity) to act as a pulley for the tendon, as the insertional angle of the tendon changes with foot movements. The tuberosity is in contact with the tendon at a variable range of the plantar/dorsiflexion movements ofthe foot-depending upon the size and shape ofthe tuberosity. Observations ofroutine MRis (M . Benjamin, unpublished observations) suggest that in many individuals , the Achilles
The Entbesis Organ Concept and Its Relevance to the Spondyloa rthropathies
59
tendon is already in contact with the superior tuberosity when the subject is standing upright in the anatomical position. Clearly, the contact pressure must be further increased when the foot is dorsiflexed-e.g., when walking. Because tendons primarily serve to transmit tensile load, they need to be specially adapted to withstand such compression and shear.n .l3 Hence, the part of the Achilles tendon which presses against the tuberosity is fibrocartilaginous (Fig. 2a). The tissue was first called 'sesamoid fibrocartilage' by Rufai et al,'" by analogy with sesamoid bone-the term indicates that the fibrocartilage lies within the substance of the tendon, rather than at its bony attachment sites. The aggrecan-rich extracellular matrix (ECM) ofthe fibrocartilage ensures that the tendon has high a water content and can thus resist compression. Of course it is not only the tendon which is subject to compression/shear against the superior tuberosity, but the tuberosity as well. Hence, the bone has a modified periosteum which is not fibrous (as in most parts of the skeleton), but cartilaginous. It has thus been dubbed the 'p eriosteal fibrocartilage' by Rufai et al 14_a terminology subsequently adopted by others (e.g., de Palma et all». It indicates that the periosteal fibrocartilage sh ould be regarded as a special form of periosteum (Fig. 2a) . It is critically important to emphasise the role ofthe superior tuberosity in understanding the functional complexity of the Achilles tendon enthesis, Its very presence adjacent to the attachment site, explains the existence of the periosteal and sesamoid fibrocartilages-key elements of the Achilles tendon enrhesis organ. Indeed, the larger the tuberosity, the more prominent are the fibrocartilages in a given indivtdual.f The contact the tuberosity makes with the tendon, reduces stress concentration at the enthesis itself, because it dissipates stress onto the preattachment zone. The' bon y lump' effectively acts as a fulcrum , increasing the moment arm ofthe Achilles tendon. The clinician will know ofcourse that it is exactly in the region ofthis bump that stress concentration is increased in a patient with Haglund's deformity-i.e., in patients with an insertional Achilles rendinopathy that is directly associated with an abnormally large tuberosity and aggravated by footwear." It may be helpful to recognise that many other tendons or ligaments elsewhere in the body either have tuberosities near their entheses or attach to the bottom of bony pits . Indeed, either strategy can be viewed as a general principal for increasing the moment arm of a tendon or ligament at its enrhesis. In either case (I.e., tuberosities or pits) , the common denominator is that th e enthesis itselfand the immediately-adjacent region ofthe tendon or ligament lie at slightly different levels from each other (i.e., raised or lowered) so that there is bone-tendon or bone-ligament contact immediately next to th e enthesis. If the levels of compression are sufficient, this leads to fibrocartilage differentiation.' Some of the clearest additional examples other than the tuberosity adjacent to the Achilles insertion, are the pits from which the collateral ligaments of the small joints of the fingers arise and the pit into which the popliteal tendon attaches. In order to minimize the frictional forces between the periosteal and sesamoid fibrocartilages which accompany foor movements, a bursa (the retrocalcaneal) intervenes between them (Fig . 2a). As the bursa is functionally related to the enrhesis and its fibrocartilages, it too is regarded as a part of the enthesis organ complex. Finally, projecting into the bursa is the wedge-shaped tip of Kager's retromalleolar fat pad (Fig. 2a)-called the 'calcaneal bursal wedge' by Theobald et al " Canoso et al l 9 should be credited with having shown for the first time that the fat pad moves into the bursa as the foot is plantar-flexed and out as it is dorsiflexed-observations which have subsequently been confirmed and extended by Theobald er al." It plays a key role is minimising pressure changes in the bursa during foot movements and Canoso et al'" have likened it to a variable spacer or plunger. Intriguingly, the bursal movements of the fat pad are compromised in patients with SpA.II However, the fat pad also has a number ofother functions within the enthesis organ complex." These include a proprioceptive function in monitoring changes in the insertional angle of the Achilles tendon (the enthesis itself is typically aneural) and an immune function that is suggested by its rich population ofresident macro phages . Its content ofnociceptive fibres'" m eans that it is also likely to be a source ofpain in enthesopathies. The enthesis itself, the sesamoid and periosteal fibrocartilages, th e retrocalcaneal bursa and its associated fat pad are the components of the enthesis organ as originally defined by Benjamin and McGonagle. 2., However, it may now be a useful extension of the concept to include fascia
Molecular Mechanisms ojSpondyloarrhroparhies
60 Figure 1. Please see legend on following page.
Figure 2. Please see legend on following page.
b SF
-1'- -
,.---- PF
T
The Enthesis Organ Concept and Its Relevance to the Spondyloartbropatbies
61
Figure 1, viewed on previous page. A posterior view of the calcaneus, outlining the area over which the Achilles tendon (AT) inserts. Its more fibrocartilaginous part (FC) is attached to the smooth facet in the middle third of the bone and its distal fibrous part (F) is anchored to the roughened region below the broken line . The upper third of the bone is formed by the superior tuberosity and does not serve for the attachment of the tendon. Instead, it acts as a pulley, increasing the moment arm of the tendon. Note how the Achilles tendon must flare out as it approaches its enthesis.
as well. Not only is there a variable soft tissue connection which partly links the Achilles tendon directly to the plantar aponeurosis," but there is also the deep crural fascia which lies adjacent to the attachment site of the tendon in the lower part of the leg," Simple observations on a living subject suggest that the actual change in insertional angle which occurs at the attachment of the Achilles tendon during foot movements must be considerably reduced by a gradual bending of the tendon over a distance of several centimetres proximal to the enthesis. This gradual bending is facilitated by the toughness ofthe adjacent deep fascia (Fig. 2b). Finally, the reader should note that the insertional region ofthe Achilles tendon is also closelyassociated with a superficial, subcutaneous bursa which intervenes between the tendon and the skin at the back of the heel. This has not previously been regarded as part ofthe enthesis organ complex, though its existence must reflect a need for independent movement between the skin and the tendon in the region where the tendon gradually bends.
Other Enthesis Organs Although the concept of an enrhesis organ is easiest to understand in the context of the Achilles tendon, it is also applicable to other sites in the body. In this context, Benjamin et aP have distinguished between articular and extraarticular enthesis organs, according to whether they also form part of a synovial joint or lie outside the joint capsule. In the former case, the synovial component ofthe enthesis organ is the synovial membrane ofthe joint itself. In the latter instance, the synovium is that which lines a subtendinous bursa intervening between tendon and bone at the enthesis. Such bursae are characteristic of many tendon attachment sites which are located close to joints e.g., the insertion ofthe patellar and biceps brachii tendons.' Although the micros copic structure of the attachment of the patellar tendon varies between individuals (largely because Figure 2, viewed on prev iou s page. a) A diagrammatic representation of the archetypal enthesis organ of the Achilles tendon (AT). This comprises the enthesis itself that is characterized by an enthesis fibrocartilage (EF), a sesamoid fibrocartilage (SF) on the deep surface of the tendon immediately adjacent to its attachment, a periosteal fibrocartilage (PF) covering the superior tuberosity, an intervening retrocalcaneal bursa (RB)and protruding into the bursa is the tip of Kager's fat pad (FP). Note the virtual absence of compact bone in the region . It is suggested that the spicules of cancellous bone (CB) are an integral part of the anchorage mechan ism and dissipate stress on the hard tissue side of the tendon -bone junction over a broad, but diffuse area (ind icated by the light shading in part of the bone). Note also that the distal part of the Achilles tendon is depicted as having a more fibrous attachment (F) to the calcaneus. This part of the enthesis attaches to the roughened area of bone labelled F in Figure 1. b) In addition to the classic enthesis organ components illustrated in Figure 2a, the deep fascia (DF) may also be considered a part of the enthesis organ. Its strength and thickness posterior to the Achilles tendon (AT) enables it to act as a retinaculum preventing the tendon from bowstringing as the foot moves from dorsi- to plantarflexion. c) Subtendinous bursae (B) often have walls which are formed distally by the deep surface of the tendon (T) and by the per iosteal surface (P) of the opposing bone. Synovium is restricted to the more proximal regions and to a covering for the protruding fat (arrows). Consequently, any pathological change in the most distal part of the tendon equates to bursal pathology as well. Indeed, it is possible that some casesof retrocalcaneal bursitis may start with degeneration of the opposing tendon and bone surfaces and that the synovial inflammation is a secondary effect. The drawing is modelled on the region where Kager's fat pad protrudes into the retrocalcaneal bursa at the Achilles tendon insertion as in Figure 2a.
62
Molecular Mechanisms ofSpondyloarthropathies
of differences in the shape of the tibial tuberosity), in its most highly differentiated state, there arc well defined periosteal and sesamoid fibrocarcilages.t These tissues arc separated by the deep infrapatellar bursa into which the tip of Hoffa's fat pad protrudes.' Equally, a well differentiated biceps brachii enthesis organ has a sesamoid fibrocartilage on the deep surface of the tendon and a periosteal fibrocartilage covering the bicipital tuberosity,' Between the two is the bicipital bursa. Both bursae (infrapatellar and bicipital) are well documented sites ofbursitis in overuse injuries and authors occasionally comment that it is difficult to distinguish bursitis from tendonitis." Because the bursal walls are only formed by synovium proximally, it seems likely to us that bursitis and tendonitis will inevitably ofien be seen to occur together-not only in traumatic injuries, but also in SpA. In the distal part of many sub tendinous bursae, there is no synovium and thus the bursal walls are formed by the tendon and bone themselves (Fig. 2c). This is a point made previously by Canoso!' and well illustrated by the work of Olivieri et al,25 who found that a high proportion of SpA patients with Achilles enthesitis, also presented with retrocalcaneal burs itis. The enthesis organ found in association with the interphalangeal joints' is the best example of an articular enthesis organ (Fig. 3a). It is 'articular' because it includes a prominent sesamoid fibrocartilage on the deep surface of the extensor tendon facing the joint cavity, immediately proximal to the site where it is inserted onto the distal phalanx. The sesamoid fibrocartilage extends the bony articular surface provided for the head of the adjacent phalanx when the finger is flexed (Fig. 3a) . The phalangeal head serves as a simple pulley, increasing the mechanical advantage of the extensor tendon as it straightens a bent finger and decreasing the stress concentration at the enthesis itself Any pathology affecting either the extensor tendon enthesis or the joint itselfcould well affect the other-just as bursitis can lead to tendonitis and vice versa . The enthesis organs associated with the interphalangeal joint are ofparticular interest in relation to psoriatic arthritis, for this disease is known to be associated with diffuse changes in the region ofthe distal interphalangeal joint. Thus, the pathology can affect not only the joints themselves, but also their associated entheses and the neighbouring nail plate." As patients may present with dactylitis, nail lesions, osteolysis or new bone formation," it is helpful to understand the structural and functional interrelationship between the nail, the joint and the neighbouring enrheses.Ie has long been recognized that the attachment site of the extensor tendon lies very close to the nail plate itself,26.27 but Tan et al 28 have recently shown that the anatomical relationship is even more intimate than previously supposed. At its enthesis, the extensor tendon gives off a slipfs) which embraces the nail root and its associated matrix, so that effectively the nail becomes an extension ofthe tendon insertion site (Fig. 3b) . Consequently, the inflammation noted by Tan et al in MRIs,28 involving entheses and extending over the whole nail bed, is associated anatomically with an extensive enthesis organ. Among the most complex enthesis organs are those at the attachments ofthe tibialis posterior, flexor hallucis and flexor pollicis longus tendons. 29.3o In all cases, the tendons fuse with adjacent structures in the vicinity of their attachments, thus dissipating the stress over a wider area . These adjacent structures (e.g., part of the spring ligament in the case of tibialis posterior'") are thus also part of the enthesis organ.
The Blood Supply ofEnthesis Organs The fibrocartilages ofenthesis organs, like cartilage elsewhere in the body can be regarded as avascular-fat is the only part ofmany enchesis organs that always contains blood vessels.P'Nevertheless, vesselsstill ofien grow into the fibrocartilages as a consequence oftissue damage including age-related changes." Consequently, the fibrocartilage region(s) ofentheses or enthesis organs can be inflamed in SpA. We have suggested that the altered vascularity of th e enthesis is a result of microdamage could make it a site where adjuvant molecules derived from gut bacteria are preferentially deposited in patients with ankylosing spondylitis." This leads to inflammation at those sites. The invading blood vesselseither come from loose connective tissue or fat on the surface ofthe tendon or ligament, or enter fibrocartilage from the bone marrow." In the latter case,vessel invasion is facilitated by a local (microscopic) absence of bone at the enthesis, This is a general feature of many attachment sites in elderly individuals." Limited information on anterior cruciate ligamenr"
The Enthesis Organ Concept and Its Relevance to the Spondyloartbropathies
63
and Achilles tendon (M. Benjamin, unpublished observations) entheses from younger people, suggests that local bone absence is seen at entheses at any age, but this awaits further confirmation at other sites. The subchondral bone defects also provide a key entry point for inflammatory cells-and because the holes directly link the soft tissue of an enthesis to the bone marrow, they help us to understand why enthesitis is associated with osteitis." One of the important corollaries of the common occurrence of subchondral bone defects at entheses, is that there is very little compact bone at fibrocartilaginous attachment sites." Consequently, the pattern of cancellous bone architecture is critical for the integrity of the tendon or ligament attachment. Contrary to popular belief, there are no Sharpey's fibres at many entheses." With the virtual absence of cortical bone at fibrocartilaginous attachments, there is simply insufficient hard tissue for Sharpey's fibres to be 'nailed' to the bone in the manner originally envisaged by Sharpey himself. Instead, the firm attachment of tendon (or ligament) to bone (which of course is the primary function of the enthesis) is promoted by (a) a complex interdigitation of calcified enthesis fibrocartilage and bone at the attachment site." (b) the architectural pattern of the underlying bone spicules." The number and arrangement of the spicules directly reflects the mechanical forces acting on the enthesis and is thus a key element of the anchorage mechanism." Radiographic evidence suggests that at sites of heavy load (e.g., the Achilles and quadriceps tendons), the spicules are orientated along the direction of pull of the tendon ." This is of key importance in relation to SpA, for it means that the network of bone spicules in the immediate vicinity of an enthesis is integrated functionally with the soft tissues at the attachment site and thus intrinsically involved in the mechanical aspects ofenthesis function. ' ! The implication which clinicians should consider is that load transfer between hard and soli: tissues affects a wider neighbourhood than simply the enthesis itself-hence the diffuse changes in the osteitis that accompanies enehesicis ." The spicules ofbone in the immediate vicinity of an enrhesis are truly part of a broader enthesis organ and we suggest that they are analogous to the roots of a tree in the anchorage they provide. The proximity of bone marrow to tendons or ligaments at entheses and the direct contact between them promoted by subchondral bone defects, also allows stem cells in the bone marrow to gain direct access to the soli:tissue side ofan enthesis." As the enthesis is a very common site of microdamage," this provides the opportunity for tissue repair. Evidence that en theses can repair damaged cnthesis fibrocartilage is abundant. even in elderly individuals," The combination of increased vascularity and microdamage in the zone of cnrhesis fibrocartilage could contribute to fibrocartilage autoimmunity-especially in HLA B27 individuals." However, conclusive evidence in support of this theory is still lacking. To summarise key points that we have made thus far : we have noted several mechanisms contributing to the formation ofanchorage organs that lead to a reduction of stress concentration at the attachment sites themselves. These include (a) the presence of tuberosities or pits at or near entheses, increasing the contact of the tendon or ligament with the bone: they in turn lead to the formation of periosteal and sesamoid fibrocartilages (b) fascial extensions of enrheses ensuring that load is partly dissipated to adjacent soli:tissues (c) retinacula (generally some distance from entheses) which reduce the extent ofinsertional angle change that occurs during joint movement (d) bursae and associated fat pads-which provide frictionless movement between tendon and bone (b) the predominantly cancellous nature of the bone at entheses-contributing to stress dissipation on the hard tissue side of the anchorage site.
Mechanisms for Reducing Stress Concentration at the Enthesis Itself Although larger textbooks of anatomy do state that tendons and ligaments frequently have fascial as well as bony attachments," this may not be widely appreciated by many clinicians. The dual existence offibrous and bony attachments is a simple form ofenthesis organ, in the sense that fibrous interconnections between tendons or ligaments dissipate stress concentration away from a local region ofthe skeleton. We have alluded to this briefly above, but will now consider it in more detail. It is almost a general principle that tendons or ligaments do not just attach to bone, but
64
Molecular Mechanisms ofSpondyloarthropathies
ET
(b)
Figure 3. Please see legend on following page.
(a)
.- DF-... . .
~-
U F - --"
,
,, ,, ' ' , , ,, , , , ,, , , ,
e.
- ....--•
••• •
~ ~ --- e •••••
:~ .-. . ------'•
'
,
...
(b)
-~ ' .
. - ...
- . .- : ..... -
TM Figure 4. Please see legend on following page.
'
.. ...
-
.
The Enthesis Organ Concept and Its Relevance to the Spondyloarthropathies
65
Figure 3, viewed on previous page. a) A semi-diagrammatic representation of the 'articular enthesis organ' associated with the extensor tendon of an interphalangeal joint. This consists of the enthesis itself (E), a prominent sesamoid fibrocartilage (SF) in the deep surface of the extensor tendon (ET), the articular cartilage (AC) on the head of the phalanx (HP), a synovial fold (5) and the joint cavity (lC). b) A diagrammatic representation of how the extensor tendon (ET)associated with the distal interphalangeal joint attaches to the base of the terminal phalanx (TP),but also blends with the adjacent periosteum (arrows) and gives off a slip which embraces the proximal end of the nail root (N). This provides an anatomical explanation for the diffuse inflammatory changes in the region that characterize patients with psoriatic arthritis."
to fibrous soft tissue as well. Such fascia of course may ultimately attach to bone elsewhere-but often at more distant locations. Consequently, the load carried by the tendon/ligament may be dissipated over a wide area and this helps to explain why patients presenting with enthesopathies can have symptoms radiating to other sites as well to the enthesis itself There is a clear example ofthis which is highly relevant to SpA. It has long been recognized that nail disease and psoriatic arthritis ofthe distal interphalangeal (D IP) joint are linked ..l6-39The epicenter ofthe diseaseappears to be the various entheses associated with the joint, but conspicuous inflammatory changes spread to adjacent tissues-including the nailbed." It is thus pertinent to recall the comment made above that the extensor tendons ofthe fingers (and presumably the toes as well-though this has yet to be confirmed) attach not only to the distal phalanx , but to the nail root and matrix as well." This gives an anatomical explanation for the link between enthesopathies and nail plate involvement in patients with psoriatic arthritis. Further examples ofsoft tissue links between adjacent enrheses that are known to be targeted in SpA, include those between the plantar fascia and Achilles tendon (see above) and between the quadriceps and patellar tendons." There are mechanisms for reducing stress concentration at entheses that can only be seen at a microscopic level. As most tendons or ligaments approach the bone, they become fibrocartilaginous. 2,12.42.44 Such 'enthesis (entheseal) fibrocartilage ' (a term introduced by Rufai et al'") probably serves a number of functions, but it certainly dissipates bending on the soft tissue side of the interface (Fig. 4a) . This is a role well understood by Schneider" who compared the fibrocartilage to that of a grommet on an electrical lead at the junction of the plug with the lead. The fibrocartilage stiffens the soft tissue at its enthesis, thus enabling it to resist the compressive forces that accompany bending movements. It is pertinent to note that there are a number of studies reporting a greater quantity of fibrocartilage at those entheses where bending of the tendon or ligament commonly occurs .464 8 Clearly however, it is not simply a question ofbending, for there are certain entheses where bending must be pronounced, yet the enrheses are fibrous rather than fibrocartilaginous (e.g., pronator reres"). Tendon or ligament shape is likely to be a key factor as well-it seems to be the more rounded/oval tendons or ligaments rather than the flattened ones , which have fibrocartilaginous entheses . However, it should also be noted that the enthesis fibrocartilage also reduces any risk ofthe tendon or ligament narrowing at such a critical site, as it comes under load." As with other fibrocartilages in enthesis organs, the ability ofenthesis fibrocartilage to tolerate compression is associated with high water content. This again stems from the presence of aggrecan 50.51- a molecule implicated as a potential autoantigen in SpA.35,52,53 Figure 4, viewed on previous page. a) A diagrammat ic representation of a typical fib rocartilaginous enthes is showing its 4 zones-dense fibrous connective tissue (DF), uncalcified fibrocartilage (UF), calcified fibrocartilage (CF) and bone (B). Note how the collagen fibres of the tendon or ligament bend gradually as they pass through the zone of uncalcified fibrocarti lage in order to reach the bone. TM-tidemark separating the two fibrocartilage zones. b) An enlargement of the region enclosed in the rectangle in Figure 4a, showing the difference in character between the superficial and deep boundaries of the zone of calcified fibrocartilage (CF). The zones of calcified and uncalcified fibrocartilage are separated by a tidemark (TM) that is straight (at a healthy insertion), but the calcified fibrocartilage knits into the underlying bone (B) at a much more complex interface. The resulting increased surface area is important in promoting anchorage and resistance to shear.
66
Molecular Mechanisms ofSpondyloarthropathies
In addition to the zone of uncalcified fibrocartilage at an attachment site, there is a thin, yet equally important layerofcalcified fibrocartilage at the very heart ofthe enthesis itself(Fig. 4b, see refs. 33,42,44 for further details). We have always regarded this as a highly specialized terminal part ofa tendon or ligament, though Haines and Mohuiddin" have called it metaplastic bone. It serves to 'knit' the soft tissue of the tendon/ligament into the adjacent bone in the most intimate offashions. 34 In doing so, it promotes the basic anchorage function ofan enthesis, It protects an attachment site against shear 4s-most tendons or ligaments insert at a small insertional angle, so as to reduce the risk of pull-out. However, although tensile loading is reduced. the level of shear is correspondingly increased. It is worth noting that a conspicuous layer ofcalcified fibrocartilage is a particular feature of highly loaded entheses.ss
Related Concepts The Synovio-Entheseal Complex (SEC) Because of the interdependence between synovial membrane and entheses within enthesis organs, we have recently proposed that an enthesis and its neighbouring synovium be viewed as forming a "synovio-entheseal complex" (SEC) within an enthesis organ." The SEC concept emphasizes the point that the synoviurn and enthesis are functionally interdependent in a manner which may be beneficial in health, but deleterious in disease. It draws attention to the fact that within the complex, there is one component which is very prone to microdamage (the enehesis") and another which is vulnerable to inflanunation in disease (the synovium) . In other waystoo, the basic parts ofthe SEC are fundamentally different from each other and it may be helpful to think ofthe SEC region as walking an "immunological righrrope'P" Prior to any mechanical damage, the normal en thesis is avascular, intrinsically anti-inflammatory, lacks macrophages and isdominated by extracellular matrix (making it physically strong). On the other hand, the synovium is richly vascular, pro-inflammatory, contains numerous macrophages and is physically weak-because its stroma is formed of fat or loose connective tissue. Consequently, if the enthesis is damaged, an ensuing inflammatory reaction is likely to occur in the synovium. According to the SEC -centered view, microdamage or tissue repair at the enthesis could trigger innate immunity and fast onset synovitis. Thus, some aspects of the angiogenic phenotype that occurs in PsA could be linked to the kinetics of the synovitis response-which in turn could be related to the SEC. This offers a novel perspective on putative immunopathogenetic mechanisms ofjoint disease related to psoriatic arthritis and indeed on me mechanism ofsynovitis in SpA in general.
FunctionalEntheses A key aspect of many enthesis organs (e.g., that of the Achilles or digital extensor tendons) on which we have placed great emphasis, is that the tendon presses against the bone immediately proximal to the enthesis itself However, a similar region of tendon-bone contact can also occur some considerable distance from the attachment site. Such areas have been called 'functional entheses' by Benjamin and McGonagle 2 in order to draw attention to important anatomical, blomechanical and pathological features that the two regions have in common. Functional entheses are typical of all of the tendons that press against the medial or lateral malleoli as they enter the foot from the leg (Fig. 5).]ust as at the enthesis itself, tendon-bone contact at a functional enthesis means increased levels of shear and/or compression, leading to the differentiation of fibrocartilage." Intriguingly, functional entheses are sites of pathology in SpA , for tenosynovitis is a well recognized feature." Thus, some SpA-targeted sites in tendons/ligaments are situated well away from their bony insertions, yet share anatomical, biomechanical and pathological features with the entheses themselves. This is of immense importance in understanding the pathogenesis of SpA, since ultimately an insertion point per se is not a sine qua non for disease expression-but perhaps there is a unifying biomechanical basis for disease manifestation i? It is also of interest to note that even though many of the most striking functional entheses (e.g., those ofthe peroneal tendons and tibialis posterior at the ankle) are a considerable distance from the tendon insertion site, the tendon-bone contact which occurs at the functional enthesis
The Entbesis Organ Concept and Its Relevance to the Spondyloarthropatbies
67
must inevitably modify the stress concentration 'downstream' at the enthesis itself, to an extent which has yet to be properly considered. Thus from such a perspective. a functional enthesis could be regarded as part of a "spatially extended enthesis organ" The analogy is strengthened by the fact that the opposing surfaces are separated by the synovial cavity ofa tendon sheath 12-Le., the counterpart of a bursa (Fig. 5). Furthermore. osteitis of an identical pattern to that seen in SpA-related enthesitis may be evident adjacent to sites of tendon compression confirming similar pathological mechanisms for bone disease at such sites.57 Finally. it should be noted that several synovial joints targeted in SpA (e.g.• the sacroiliac . sternocostoclavicular, acromioclavicular and temporomandibular (TMJ) joints) are lined by fibrocartilage rather than hyaline cartilage," and are subject to high levels of shear. This may reflect their anatomical orientation (often approaching the vertical) or the joint mechanics-the TMJ for example experiences high shear as a consequence of mandibular movements. The key point however. is that the presence offibrocartilage (as an indicator ofcompression and/or shear) is a common denominator linking many of the sites targeted in SpA.2 Furthermore. as we have outlined earlier. the bone beneath an entheseal fibrocartilage is an intimate part of the enthesis structure. Based on imaging and histological studies offibrocartilaginous joints in SpA . it is clear that these joints are subject to a diffuse osteitis of a magnitude similar to that seen at insertions proper. Furthermore. carriage of the HLA-B27 gene is associated with more severe osteitis and this provides a conceptual link between inflammation in the 51]bone and the HLA-B27 gene in ankylosing spondylitis.58
Conclusions The historical view on the immunopathogenesis ofankylosing spondylitis and the related spondyloarthropathies was that an unknown "arthritogenic" peptide derived from somewhere within the joint. most probably the synovium, was presented to T-cells. This culminated in joint inflammation and especially synovitis at sites such as the synovial part of the sacroiliac joint. However. we think that there is evidence for a unifying biomechanical, rather than an autoimmune basis for disease. since inflammation localizes to sites ofhigh mechanical stress as indicated by the presence offibrocartilage. Therefore. enthesitis, adjacent osteitis and synovitis and even fibrocartilaginous joint disease can be viewed in a different light-and the enthesis organ concept is central to a clearer understanding ofthese pathologies. We would emphasize that the sites to which disease localizes in SpA are also sites ofmicrodamage. This is especially pertinent given the recent claims that the innate immune system is involved in SpA. Ofcourse the innate immune response is involved in all inflammatory reactions. However. there is no compelling evidence, thus far. that genetic polymorphisms in innate immune-related receptors playa role in ankylosing spondylitis in particular or SpA in general. It is more likely that innate immune activation occurs in a normal and appropriate manner at sites oftissue microdamage that is best exemplified by insertions. Indeed an increasing number ofdamage associated molecular patterns (DAMPs), that would be anticipated at sites of microdamage (where they can initiate innate immunity) have now been Identified," Furthermore, viable microbes may not be necessary to trigger innate immunity and the presence ofmicrobial products at an enthesis could be sufficient to lead to d isease initiation via the activation ofToll-like receptors and other pattern recognition receptors. This offers a simple explanation for the failure to incriminate viable microbes in SpA.2.59 The molecular basis for enthesitis is still in its infancy. but recent studies have begun to explore the issue and have shown a role for BMP signalling in disease expression /" What then has the enthesis organ concept told us about SpA? Perhaps its greatest value is that it rationalizes the diffuse nature of entheseal pathology (especially the location of bone oedema and synovitis) and makes sense ofseemingly disparate symptoms with which patients may present. It explains the interconnections between bursitis and tendonitis (rendinosls) and highlights the functional interplay between synovial membranes and entheses that we have formalized via the related concept ofa 'synovio-entheseal complex'. The continued refinement ofthe enthesis organ concept is a clear demonstration of the importance of anatomy to rheumatology in connection
68
Molecular Mechanisms ofSpondyloarthropathies
Figure 5. The region where a tendon or ligament presses against a bone (promoting a change of direction) without attaching to it is referred to as a 'functional enthesis'. Functional entheses are characteristic of the region where tendons in the leg wrap around the malleolus of the tibia (T) before entering the foot. The subtend inous bursa which is typical of many 'true' enthesis organs is here represented by a synov ial tendon sheath (55) which surrounds the tendons of flexor hallucis longus (FHL) and tibialis posterior (TP). Both the tendon and bone have opposing surfaces that are fibrocartilaginous. The retinaculum (R) is a specialised part of the deep fascia and an integral part of the functional enthesis organ . Redrawn after Gerrish FH 'A textbook of Anatomy'. London : Henry Kimpton publishers 1899.
with understanding enthesopathies.'! It also illustrates the importance oftissue specific factors in the phenotypic expression ofimmune mediated inflammation."
References 1. Woo SL. Buckwalter JA. AAOS /NIH /ORS workshop. Injury and repair of the musculoskeletal soli: tissues. Savannah Georgia 1987. J Orthop Res 1988; 6:907-931. 2. Benjamin M, McGonagle D. The anatomical basis for disease localisation in seronegative spond yloarthropathy at enrheses and related sites. J Anac 2001 ; 199:503-526. 3. Benjamin M. Moriggl B. Brenner E et aI. The "enthesis organ " concept: Why enthesopathies may not present as focal insertional disorders. Arthritis Rheum 2004 ; 50:3306-3313. 4. Huber LC. Moritz F, Gay S. Spondylarthrirides and related entities : Entheses and hypotheses. Arthritis Rheum 2007 ; 56:4-8. 5. McGonagle D, Khan MA, Matzo-Ortega H et aI. Enrhesiri s in spondyloarthropathy, Curr Opin Rheurnatol 1999 ; 11:244-250. 6. McGonagle D, Gibbon W; Emery P. Spondyloarthropathy, Lancet 1999; 353 :1526-1527. 7. Alvarez-Nemegyei J. Canoso JJ. Heel pain : Diagnosis and treatment. step by step. Cleve Clin J Med 2006; 73 :465-471. 8. Ritchlin CT. Pathogenesis of psoriatic arthritis. Curr Opin Rheumatol 2005 ; 17:406-412 . 9. Leung IT, Tam LS. Kun EW et aI. Psoriatic arthritis as a distinct disease entity. J Postgrad Med 2007 ; 53 :63-71.
The Entbesis Organ Concept and Its Relevance to the Spondyloarthropatbies
69
10. Benjamin M. Bydder GM. Magnetic resonance imaging of entheses using ultrashort TE (UTE) pulse sequences. J Magn Reson Imaging 2007 ; 25:381-389 . 11. Canoso JJ. The premiere enrhesis. J Rheumatol1998; 25 :1254 -1256. 12. Benjamin M. Ralphs JR. Fibrocartilage in tendons and ligaments-An adaptation to compressive load. J Anat 1998; 193:481-494. 13. Vogel KG. Koob TJ. Structural specialization in tendons under compress ion. Int Rev Cytol 1989 ; 115:267-293. 14. Rufai A. Ralphs JR, Benjamin M . Structure and histopathology of the insertional region of the human Achilles tendon. J Orthop Res 1995; 13:585-593. 15. de Palma L. Marinelli M. Meme L et al. Immunohistochemistry of the enthesis organ of the human Achilles tendon. Foot Ankle Inr 2004; 25:414-418. 16. Theobald P. Biomechanical Analysis of the Achilles Tendon Enthesis Organ. PhD Thesis. Cardilf University. 2006: 32-76. 17. Sella EJ. Caminear DS. McLarney EA. Haglund's syndrome . J Foot Ankle Surg 1998; 37:110-114. 18. Theobald P. Bydder G. Dent C et al. The functional anatomy of Kager's fat pad in relation to retrocalcaneal problems and other hindfoot disorders. J Anar 2006; 208:91-97 . 19. Canoso JJ. Liu N, Traill MR ec al. Physiology of the retrocalcaneal bursa . Ann Rheum Dis 1988; 47:910-912. 20. Shaw HM, Santer RM , Watson AHD et al. Adipose tissue at enrheses: The innervation and cell composition of the retromalleolar fat pad associated with the Achilles tendon. J Anat 2007 ; J Anat 2007: 211 :436-443. 21. Wood Jones F. Structure and Function as Seen in the Foot . London : Bailliere, Tindall and Cox . 1944:60 . 22. Benjamin M. Theobald P. Suzuki D et al. The anatomy of the Achilles tendon. In : Maffulli N. Almekinders L. eds. The Achilles Tendon. Berlin: Springer-Verlag. 2007 ; 5-16 . 23. Stand ring S. Gr ay's Anatomy: The Anatomical Basis of Clinical Practice . 38 ed, 2004 :881. 24. Yamamoto T. Mizuno K. Soejirna T et al. Bicipital radial bursitis: CT and MR appearance . Cornput Med Imaging Graph 2001 ; 25 :531-533. 25. Olivieri I, Barozzi L. Padul a A er al. Retrocalcaneal bursitis in spondyloarthropathy: Assessment by ultrasonography and magnetic resonance imaging. J Rheurnarol 1998; 25:1352-1357. 26. Schweitzer TP. Rayan GM . The terminal tendon of the digital extensor mechanism : Part I, anatomic stud y. J Hand Surg (Am) 2004 ; 29:898-902. 27. Shum C. Bruno RJ. Ristic S et al. Examination of the anatomic relationship of the proximal germinal nail matrix to the extensor tendon insertion . J Hand Surg [Am] 2000; 25:1114-1117. 28. Tan AL, Benjamin M, Toumi H et al. The relationship between the extensor tendon enthesi s and the nail in distal interphalangeal joint disease in psoriatic arthritis-A high-resolution MRI and histological study. Rheumatology (Oxford) 2007; 46:253-256. 29. Moriggl B. Kumai T, Milz S er al. The structure and histopathology of the "enrhe sis organ " at the navicular inserti on of the tend on of tibialis posterior. J Rheurnatol 2003; 30(3 ):508-517. 30. McGonagle D, Lories RJU, Benjamin M. The concept of a "synovio-enrheseal complex" and its impli cations for understanding joint inAammation and damage in psoriatic arthritis and beyond. Arthritis Rheum 2007; 56:2482-2491. 31. Benjamin M, Toumi H , Suzuki D et al. Microdamage and altered vascularity at the enthesis-bone interface provides an anatomic explanation for bone involvement in the HLA-B27-associated spondylarthrirides and allied disorders. Arthritis Rheum 2007; 56:224-233. 32. McGonagle D. Gibbon W, O 'Connor P et al. Characteristic magnetic resonance imaging entheseal changes of knee synovitis in spondylarthropathy, Arthritis Rheum 1998 ; 41 :694-700. 33. Benjamin M . Kumai T. Milz S et al. The skeletal attachment of tendons-tendon 'entheses. Comp Biochem Physiol A Mol Integr Physiol 2002 ; 133:931-945. 34. Milz S, Rufai A. Buettner A cr al. 'Three-d imensional reconstructions of the Achilles tendon insert ion in man . J Anar 2002; 200 :145-152. 35. Maksymowych WP. Ankylos ing spondylitis -At the interface of bone and cartilage. J Rheurnarol 2000; 27:2295-2301. 36, Baker H, Golding DN. Thompson M. The nails in psoriatic arthritis. Br J Derrnarol 1964; 76:549-554. 37. Cohen MR, Reda DJ. Clegg DO. Baseline relationships between psoriasis and psoriatic arthritis: analysis of221 patients with active psoriatic arthritis . Department of Veterans Affairs Cooperative Study Group on Seronegative Spondyloarrhropathies. J Rheumatol 1999; 26:1752-1756. 38. Jones SM. Armas JB. Cohen MG et al. Psoriatic arthritis : Outcome of disease subsets and relationship of joint disease to nail and skin disease. Br J Rheumarol 1994; 33 :834-839.
70
Molecular M echanisms of Spondyloarthropathies
39. Wright V. Roberts Me. Hill AG. Dermatological man ifesrarions in psoriat ic arthritis: A follow-up stud y. Acta Derrn Venereol 1979 ; 59:235-240. 40 . Tan AL. Grainger A]. Tanner SF er aI. A high-resolurion magnetic resonance imaging stud y of distal interphalangeal joint arthropathy in psoriatic arthritis and osteoarthritis: Are the y the same? Arthritis Rheum 2006; 54:1328-1333. 41. Toumi H. Higashiyama I, Suzuki 0 et aI. Regional variations in human patellar trabecular architecture and the stru cture of the prox imal patellar tendon enthesis.] Anat 2006; 208 :47-57. 42 . Benjamin M . Toumi H . Ralphs JR et aI. Where tendons and ligaments meet bone: attachment sites ('entheses') in relation co exercise and/or mechanical load. ] Anat 2006; 208 :471-490 . 43. Benjamin M . Kumai T. MHz S et aI. The skeletal attachment of tendons-Tendon "en theses", Comp Biochem Physiol A Mol Integr Physiol 2002; 133:931-945. 44. Benjamin M. Evans E]. Copp L. The hiscology of tendon attachments to bone in man.] Anac 1986; 149:89 -100 . 45. Schneider H . Zur Strukrur der sehnenansarzzonen. Z Anat Enrwi cklung 1956 ; 119:431 -456. 46. Benjamin M , Evans E]. Rao RD ce aI. ~antitative differences in the hiscology of the attachment zones of th e meniscal horn s in the knee joint of man . J Anat 1991 ; 177:12 7-134. 47. Benjamin M . Newell RL. Evans EJ et aI. The structure of the insertions of the tendons of biceps brachii, triceps and brachialis in elderly dissecting room cadavers. J Anar 1992 ; 180:327-332. 48. Evans E]. Benjamin M. Pemberton OJ. Fibrocartilage in the attachment zone s of the quadriceps tendon and patellar ligament of man.] Anar 1990; 171:155-162. 49 . Knese K-H. Biermann H . Die Knochenbildung an sehnen-und Bandsarzen im beriech ursprunglich chondraler apophysen , Z Zellforsch 1958; 49:142-18 7. 50. Waggett AD . Ralphs ]R. Kwan AP et aI. Characterization of collagen s and proreoglycans at the insertion of the human Achilles tendon. Matrix Bioi 1998; 16:457-470. 51. Milz S. Schluter T. Putz R et aI. Fibrocartilage in the transverse ligament of the human atlas. Spine 2001; 26 :1765 -1771. 52. Zou ] . Zhang Y. Thiel A er aI. Predominant cellular immune response to the cartilage auroantigenic G 1 aggrecan in ankylosing spond ylitis and rheumatoid arthritis. Rheumatology (Oxford) 2003 ; 42 :846-855. 53. Giant T. Csongor ], Szucs T. Immunopathologic role of proteoglycan antigens in rheumatoid joint disease. Scand] Immunol 1980 ; 11:247-252 . 54. H aines RW; Mohuiddin A. Metaplastic bone. J Anar 1968; 103 :527-538. 55. Evans E]. Benjamin M. Pemberton OJ. Variations in the amount of calcified tissue at the attachments of the quadriceps tendon and patellar ligament in man .] Anat 1991; 174 :145 -151. 56. Myerson M. Solomon G. ShereffM. Posterior tibial tendon dysfuncrion : Its association with seronegative inflammatory disease. Foot Ankle 1989; 9:219 -225. 57. McGonagle D. Marzo-Ortega H . Benjamin M et aI. Report on the second International enthes itis workshop. Arthritis Rheum 200 3; 48:896-905. 58. McGonagle D . Marzo-Ortega H . O 'Connor P er aI. The role ofbiomechanical factors and HLA-B27 in magnetic resonance imaging-determined bone changes in plantar fascia enchesopathy, Arthritis Rheum 2002 ; 46:489-4 93. 59. McGonagle D. Srockwin L. Isaacs] et aI. An enrhesins based model for the pathogenesis of spond yloarrhroparhy, additive effects of microbi al adjuvant and biornechanical factors at disease sites. J Rheumarol 2001 ; 28:2155··2159 . 60. Lories R]. Derese I, Luyten FP. Modulation of bone morphogenetic protein signaling in hibits the onset and progression of ankylosing eruh esins.} Clin Invest 2005 ; 115:1571-1579. 61. McGonagle D. McDermon ME A proposed classification of the immunological diseases. PLoS Med 2006; 3:e297.
CHAPTERS
Synovial and Mucosal Immunopathology in Spondyloarthritis Bernard Vandooren, Paul P. Tak and Dominique Baeten"
Abstract
C
hronic inflanunation ofmusculoskeletal structures isthe most prominent diseasemanifestation ofSpA. More specifically, the axial disease affects the spine , the sacroiliac joints and the hips. Peripheral disease includes peripheral arthritis,with a preference for asymmetrical inflammation ofjoints ofthe lower limbs and enthesitis, which is the presence ofinflanunation at the sites were ligaments and tendons attach to the bone . Additionally, SpA is often characterized by subclinical inflammation of the gut which partially resembles inflammatory bowel disease.' Here . we will review the immunopathology of these different disease manifestations and relate them to clinical applications as well as emerging pathogenic concepts .
Introduction TheImmunopathology ofAxial Spondyloarthritis In the spine, one should disrinpuish the inflammation of the vertebral body from that of the zygapophyseal joints. In the vertebral bodies and intervertebral spaces. the inflammatory cells are mostly seen in the annulus fibrosus, in association with erosion of the bone that makes contact with the annulus. A prominent neovascularization can be seen in these structures.' Besides signs ofinflammation and bone erosion. different authors also describe the presence ofnew bone forma tion th at will lead to the formation ofsyndesmophytes.2.3 The exact nature ofthis bone formation proce ss is unclear : one manuscript describes the predominant presence ofmembranous or reactive new bone formation,which is the direct transdifferentiation offibroblastoid cells into osteoblasts ,! while the other study describes the predominance ofendochondral bone formation,' which is the formation ofbone out ofcartilage . Almost 40 yearslater, this controversy has still not been solved. As far as we know. there are no published studies of detailed immunohistochemical analysis of the vertebral bodie s in SpA. The zygapophyseal joints are synovial joints that form the articulation between the vertebral bodies. These joints are frequently affected in SpA, often leading to bony ankylosis.. However, such bony ankylosis can also be seen in the absence of pathology-The bony bridges in zygapophyseal joints can form in the granulation tissue, which penetrates into the joint from the bone marrow space after erosion of the bone-end plate , as well as in the periarticular capsula.' The synoviurn and the subchondral bone marrow are heavily infiltrated with inflanunatory cells.' Identification of these inflammatory cells in subchondral marrow space showed a predominant infiltration with T-cells (CD4- as well as CDS positive). but also Bvlymphocytes." Organised lymphocyte aggregates were present in 6 out of 16 samples that were studied. Apart from the presence ofbone ' Corresponding Author: Dominique Baeten-Clinical Immunology and Rheumatology Academ ic M edical Center/University of Amsterdam, Mei bergdreef 9, 1105 AZ Amsterdam, The Netherl ands. Email : d.l.b
[email protected]
Molecular M echanisms ofSpondyloarthropathies. edited by Carlos Lopez-Larrea and Roberto Dfaz-Peiia. ©2009 Landes Bioscience and Springer Science+Business Media.
72
Molecular Mechanisms ofSpondyloartbropathies
marrow oedema, there was also a prominent neovascularization in these samples. Unfortunately, the synovial membrane was not investigated in this study. The sacroiliac joint is the most typical disease site in SpA . In a detailed histomorphological study ofthe sacroiliac joints ofAS patients, Francois er al 6 describes the presence ofinAamrnatory cells in the subchondral bone marrow as well as the synovial membrane, which can even form a small pannus. In comparison with normal sacroiliac joints, there was significantly more superficial cartilage erosion, bone marrow oedema, new bone formation and bony ankylosis in the diseased joints. In attempt to perform an analysis of the kinetics of the histopathological changes, the author describes the presence of synovial inAammation and bone marrow oedema as the earliest changes. This was followed by the formation of strings of connective tissue and new bone which eventually lead to complete bony ankylosis of the joints. Here, the author describes the presence ofmembranous as well as enchondral new bone formation. A second study used tissue obtained from the sacroiliac joint by CT-guided needle biopsies. ?Although it remains unclear ifthe biopsies represent genuine synovium as the characteristic intimal lining layer was not described, it clearly showed that the inflammatory infiltrate consisted of both T-cells (CD4 as well as CD8 positive) and macrophages. Finally, an immunohistochemical study offemur heads ofAS patients described a pronounced infiltration of the subchondral bone marrow by T-Iymphocytes, with a predominance for CD4 positive cells over CD8 positive cells." In all samples, lymphocytic aggregates were present. Interestingly, the presence ofT-lymphocytes as well as osteoclasts was closely associated with the presence ofcartilage in the close vicinity. Unfortunately, the authors did not investigate the presence of B-cells and macrophages in these samples. Also, the changes in the synovial membrane were not described.
The Immunopathology ofPeripheral Spondyloarthritis In the past years, our group investigated intensively the immunopathology ofperipheral arthritis in SpA. An important difference with the studies described above, which were almost exclusively performed on the surgical resection specimens, is that the use of the minimally invasive needle arthroscopy technique allowed us to study synovial tissue samples from patients with early and active disease. The Inflamed synovium contains a mixed inflammatory infiltrate which consists of macrophages, T-Iymphocytes, B-Iymphocytes, plasma cells and a few neurrophils.t" The macrophages are the most prominent cellular population and the majority of these cells express CD 163, a scavenger receptor for haem-haptoglobin complexes." An important role ofthese cells in the local inAammatory process is suggested by the close correlation ofthe number ofthese cells with the disease activity, both cross-sectionally and during effective treatment. 15.16 Similarly to most axial disease sites , the inflamed synovial membrane is characterized by a pronounced neovascularization. 17 •18 No signs of new bone formation were observed in these synovial membrane biopsies obtained by needle arthroscopy (own unpublished observations). We can, however, not exclude that this process would still occur deeper in the tissues. Moreover, this technique does not allow to sample and study other joint tissues such as the subchondral bone marrow and the enthesis. Enthesitis is the inAammation ofthe insertion of tendons or ligaments on the bone. There are 2 kinds of entheses: fibrous and fibrocartilaginous entheses. Fibrocartilaginous entheses, which contain fibrocartilage, develop at the insertion of ligaments on bones that arise by enchondral osteogenesis, for instance the calcaneum. Fibrous entheses develop at the insertion ofligaments on bones that arise by membranous osteogenesis, for instance the iliac bone. In SpA, inAammation can occur in both types ofentheses .' ? At the microscopical level, one can see small lesions consisting ofa mixed inAammatory infiltrate in the fibrous and fibrocartilaginous part ofthe enthesis as well as in the subentheseal bone marrow. At the site of the insertion, one can also see small bone ero sions which contain inAammatory cells as well numerous small blood vessels. In some samples, th e authors also described the presence ofnew bone formation.' Immunophenotyping ofthe infiltrate in SpA enthesitis was performed by two groups. McGonagle et al 20 prelevated fibrocartilaginous tissue from the entheses of SpA patients by needle biopsy. CD68 positive macrophages were the
Synovial and Mucosal Immunopathology in Spondyloarthritis
73
predominant cellular population in the inflammatory infiltrate of these samples. Laloux et al2l prelevared the entheses from the crucial ligaments and vastus lateralis of SpA patients by open surgery. The authors focussed on the changes in the subentheseal bone marrow. They describe the presence of bone marrow oedema together with an inflammatory infiltrate that predominantly consisted of CD8 positive T-Iymphocytes. There was no increase in the numbers ofmacrophages in this tissue. Both studies describe a pronounced neovascularization in the tissue.
Axial Versus Peripheral Spondyloarthritis Taken together, the histopathological studies ofthe musculoskeletal inflammation in SpA show a number ofstriking similarities between axial and peripheral diseasesites. Firstly, both sites the inflammation develops in close proximity ofthe cartilage or the fibrocartilaginous enthesis. Secondly, the infiltration with immune cells occurs in both the soft tissues that attach to the skeleton as well as the underlying bone marrow at all the localisations that have been investigated so far. Thirdly, the inflammatory infiltration is mostly mixed.with a predominance ofmacrophages in the soft tissues such as the synovial membrane or the enthesis and a predominance ofT-lymphocytes in the bone marrow. Fourthly, there is a very pronounced neovascularization in all inflamed tissues. However, there is also clear evidence that the pathophysiological mechanisms underlying peripheral SpA are distinct from those involved in exclusivelyaxial SpA. Firstly, patients with peripheral arthritis clustered in specific families in a study that included 329 SpA patients belonging to multiplex families." As this was independent from HLA-B27 status ofthe family members and disease duration, this strongly suggests that the development of peripheral synovitis is linked to the specific genetic background of the patients. This is in line with earlier studies reporting that ReA was inherited differently than AS.23 Secondly, therapeutic studies demonstrated that the inflammatory back pain in patients with exclusive axial disease was more sensitive to treatment with NSAID24 and sulphasalazine (SSZ)25 than in patient with concomitant peripheral arthritis. However, the effect of SSZ is controversial because an earlier study demonstrated that there was no improvement of axial complaints after treatment with SSZ in patients with AS.26 Thirdly, we found that patients with peripheral SpA had much higher serum levelsofmatrix-metalloproteinase 3 (MMP-3) than patients with exclusive axial SpA, despite similar levels of serum CRP or ESR values.F'It is not clear, however, ifthis is simply the result ofleakage ofMMP-3 from the inflamed joint or a reflection of a fundamental difference in the type of inflammation between peripheral and axial SpA. Interesting in this context was the recent observation that serum MMP-3 is an independent predictor of structural remodelling of the spine in AS patients, although the interpretation ofthese results is complicated by the inclusion ofpatients with both axial and peripheral joint disease." In conclusion, emerging genetic, therapeutic and serological evidence suggests that peripheral SpA could be fundamentally different from exclusive axial SpA. Due to the extreme difficulty to obtain axial samples, however, there are unfortunately no systematic comparisons of structural and inflammatory features ofthe spine versus the peripheral joint in SpA or, alternatively, ofthe inflamed spine in patients with peripheral versus exclusiveaxial SpA. Additional genetic and immunopathological studies should be designed and performed to address this issue.
Skeletal Remodeling in Spondyoarthritis SpA is not only characterized by articular inflammation but also by a unique and complex pattern of structural remodeling of the skeletal tissue. Radiology shows both signs of cartilage and bone degradation and signs ofnew bone formation. Most typically, the progression of these phenomena is very slow, taking many years before reaching the terminal stage of complete bony ankylosis of the joint.29The changes in the sacroiliac joints have been best described and are incorporated in classification criteria ofSpA.30The earliest changes are blurring ofthe subchondral bone , followed by the formation of bone erosions and sclerosis of the apposing bone . Later, the joint is bridged by radio-opaque strings that can lead to complete replacement ofjoint space by mineralized tissue. A similar sequence of changes can also be seen in the spine, leading to the development ofa characteristic bamboo-spine, or at the enthesis.Ieading for example to a typical
74
Molecular Mechanisms ofSpondy loarthropathies
heel spur at the calcaneum." As described above, parallel change s are seen in the tissues when evaluated by histopathology. In the past years, a lot of progress has been made in the understanding of cartilage and bone degradation in experimental models. A central concept in this process is the fact tha t the skeletal structures are continuously remodeled and therefore the homeostasis is dependent on a strict regulation between tissue formation and degradation pathways . So far the structural remodeling process in chronic joint inflammation has been best studied in the human TNF-alpha transgenic mouse. These mice spontaneously develop a chronic polyarthritis that leads to rapid destruction of the articular cartilage and bone . The destruction ofthe cartilage in these mice is the result ofboth increased chondrocyte cell death," impaired matrix synthesis by resident chondrocytes and increased matrix destruction by degradative enzymes." Importantly, the TNF-driven cartilage destruction, in contrast to the synovial inflammation, is completely dependent on myeloid cell derived IL-I Y The bone destruction in the mice is the result ofboth increased differentiation and activation of specialized bone resorbing cells, osteoclasrs, from myeloid progenitor cells," and impaired bone repair responses.P The differentiation and activation ofosteoclast from myeloid progenitors cells is primarily dependent on the stimulation of RANK on the progenitor cells by RANKL,36but is also directly enhanced by TNF alpha'?and IL-I beta .38The suppression ofreparative responses of the bone is mediated by dickopf-I (DKK-I).This med iator is synt hesized by FLS in response to TNf alpha and is capable ofinhibiting canonical Wnt-signaling. In the absence ofDKK-I , TNF transgenic mice have overzealous reactive bone formation instead of bone erosions." Apart from the Wnt-signaling pathway, also the SMAD-signaling pathway plays an important role in the embryogenesis and homeostasis ofthe articular cartilage and bone." Lories et al investigated the role ofbone morphogenic proteins (BMP) in the structural remodeling ofthe joint in aging male DBA/I mice.'? These mice spontaneously develop arthritis and enthesitis of the paw and ankle joints and, in contrast to TNF alpha transgenic mice, also develop reactive enchondral bone formation that leads to bony ankylosis of the affected joints and entheses, BMP-2 and BMP-6 were over-expressed in the enthe ses of the mice and when the ir function was blocked by adenoviral overexpression of noggin, a BMP antagonist, the pathology was inhibited. Similarly, intra-articular injection ofBMP-2,41but also TGF bera," in healthy joints of mice leads to the formation ofosreophytes. Recently,intercellular adhesion proteins have also been identified as important players in joint remodeling in chronic arthritis. More specifically, cadherin-Ll was shown to be a critical anchor for the intercellular adhesion of FLS and formation of a normal intimal lining layer.43.44 When arthritis was induced by transfer of arthritogenic serum , cadherin-ll deficient mice were partly protected from synovial inflammation and cartilage lesions were completely absent." These joint remodeling proces ses and their key mediators are only fragmentary described in SpA. As already mentioned above, SpA patients, not only develop cartilage degradation and bone erosions , but also reactive new bone formation. We have stud ied determinants of cartilage integrity in SpA patients, analyzing synovial membrane biopsies and synovial fluid. Our studies were focused on biomarkers of cartilage metaboli sm on the one hand and known mediators of inflammatory cartilage and bone remodelingon the other hand. As a strategy we used the systematic comparison of these parameters between SpA and RA patients, who develop a more destructive form ofarthritis. Focusing first on cartil age, we found equal amounts ofthe cartilage degradation biomarker COMP in synovial fluid ofSpA as in RA.27Similarly, the synovial expression ofMMPs and their inhibitors (TIMPs)27aswell as cadherin-l l (Vandooren et al. Manuscript in Preparation) were equal between RA and SpA patients. In contrast, the con centration of MIA, a biomarker of chondrocyte anabolism, was significantly higher in the SF ofSpA than in RA patients." As to bone degradation, we have also studied the expression ofosteoclasts and the RANK/RANKL cytokine system in SpA patients and again found no significant differen ces with RA parients." This study confirmed a pre vious study that demonstrated the abundant pr esence of osteoclast precursor cells and RANKL in the synovial membrane ofPsA patients."
Synovial andMucosal Immunopathology in Spondyloarthritis
75
Altogether these data indicate that the differences in joint remodeling between RA and SpA patients most likely are not caused by a differential activation ofdegradative pathways, but more likely by differences in tissue repair responses. In this context, Lories et al studied the expression of BMPs in the synovial membrane of SpA patients. A clear up-regulation of BMP-2 and -6 expression, together with an activation of their SMAD signaling pathway, was observed in the SpA patients, but this was not different from RA parienes." These data indicate that also BMPs can not explain the differences in tissue remodeling between SpA and RA joints. An interesting observation in this context is that in inflammatory conditions cells are desensitized for all sorts of anabolic stimuli, including BMP_2.49 In addition, it was recently demonstrated that reactive enchondral new bone formation induced by BMP-2 was dependent upon beta-catenin signaling and that this process can be inhibited by the canonical Wnt-antagonist DKK-1. 50 The expression of DKK-l is increased in the synovial membrane ofRA patients and the serum levels of DKK-l are significantly higher in RA than SpA paelenrs." Future studies will have to address the expression and function ofDKK-l in the joints ofSpA patients.
Synovial Biomarkers in Spondyoarthritis An interesting clinical spin -offofthe study ofthe immunopathology and molecular pathways of joint inflammation and remodeling in SpA is the discovery of novel synovial biomarkers. The value of blood analysis in the diagnosis of SpA is very limited at the moment, as there are no disease-specific autoantibodies and as the classical acute phase response parameters, C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR), are elevated in only half of the SpA patients. Several molecules produced and /or released during joint inflammation and remodeling may, however, provide additional information. A prototypic examples is MMP3, which serum levels are elevated in SpA patients compared to healthy persons." Serum MMP3 is especially elevated in AS patients which have concomitant peripheral arthritis but not in pure axial disease and originates from the inflamed joint." Accordingly,serum MMP-3levels correlated with clinical diseaseactivity parameters 52.53 and decreased strongly anr' rapidly upon effective treatment by TNF alpha blockade ." Although Maksymowich and coworkers reported that serum MMP3, but not CRP or ESR, is an independent predictor ofstructural damage ofthe spine in AS panenrs," it is clear from the previously mentioned studies that serum MMP3 reflects the inflammatory process rather than tissue damage and remodeling. Because this skeletal remodeling is a prominent feature ofSpA, different studies have focused on serum biomarkers that specifically assesscartilage and bone turn-over. Putative cartilage synthesis biomarkers, C-propeptide of collagen Type 2 (Cpn) and aggrecan 846 epitope, as well as the cartilage degradation marker, C2C, were found to be increased in the serum ofAS patients.54 However, it is not clear how these biomarkers change upon effective treatment with TNF blockers as one study found no changes at all,54 while another study showed an increase in the aggrecan 846 epitope and a decrease in C2C.55 Surprisingly, however, neither the Cf'H and aggrecan 846 epitope nor C2C were found to predict structural damage in an independent study by the same authors." Because the significance of the cartilage neosynthesis biomarkers was not clear, cpn being released from cartilage by protease degradation 56 and the 846 epitope being an epitope from an embryonal aggrecan isoform.>' we carefullyvalidated the useofa new cartilage anabolism biomarker, MIA .45 When MIA was measured in the serum of SpA patients with peripheral synovitis, there were no overall differences compared to healthy persons or rheumatoid arthritis. There was, however, a broad inter-individual variation in serum MIA levelsin SpA patients and these levels were inversely correlated with the number ofclinically affected joints. When the patients were treated with TNF-blockade the serum MIA levels increased. Similarly, the bone formation biomarkers, osteocalcin and bone specific alkaline phospatase, which were found to be increased in the serum ofAS patients,57 were further upregulared by TNF blockade." In the past years, intensive research on the evaluation ofsynovial tissue analysis as a biomarker in SpA was performed. In a diagnostic setting, it was found that no single histopathological tra it
76
Molecular Mechanisms ofSpondyloarthropathies
ofthe synovium reliably discriminates SpA from other forms ofchronic arthritis. When performing a multi-parameter analysis, however, a majority ofthe SpA patients were identified. The traits that made the diagnosis ofSpA more likely were: (1) the pronounced neovascularization of the SpA synovial membrane, (2) the modest hyperplasia of the intimal lining layer and (3) the relative abundance ofpolymorphonuclear cells in the synovial tissue." In addition, there is a strong positive correlation between the number of infiltrating CD 163 positive macrophages as well as polymorphonuclear cells and the clinical disease activity/" In a follow-up study, it was found th at synovial membrane analysis,on a relativelysmall number oftissue samples, reliably discriminated clinical effective treatment from ineffective treatment." The traits that best discriminated, were the therapy induced changes in synovial MMP3 expression and the number ofCD163 positive macrophages.The SpA biomarkers described here, including MMP3, MIA and synovial CD 163 , are unlikely to reach sufficient discriminative power to be used as diagnostic or prognostic marker in single patients in daily practice. However, they may provide useful information to optimize observational and imerventional trial design in SpA.
Gut Inflammation in Spondyloarthritis About 5% ofSpA patients sufferfrom inflammatory bowel disease (IRD), 62 while in the general population the prevalence ofthis condition is less than 0,5%.63 Genetic studies demonstrated that SpA patients with IBD are genetically distinct from the other SpA patientsY·64 Although the presence ofHLA-B27 strongly predisposes IBD patients to develop musculoskeletal manifestations of SpA, the frequency of HLA-B27 in the IBD associated SpA (25-85%) is lower than in the general SpA population (75-95%).65 Because these findings suggested that other genes than HLA-B27 were important for diseasepathogenesis in IBD associated SpA, many other candidates are currently under investigation. The CARD 15 gene encodes nucleotide binding oligomerization domain 2 (NOD2), which is a cytosolic receptor for conserved microbial patterns. Recently described polymorphisms of this gene are strongly linked with the development ofIBD in the general population.f Although the same polymorphisms in CARD 15 are not linked to the development ofSpA, one study suggests that microscopic signs of gut inflammation are more frequently observed in SpA patients with CARD15 polymorphisms than in other SpA patients." Interesting in this context is the observation that 41 % offirst degree relatives ofpatients with AS have subclinical gut inflammation together with radiological sacroileitis independently of the HLA status of the subjects," It remains unclear ifthis phenomenon is related to CARD 15 mutations as an early report ofan association between CARD 15 mutations and sacroiliitis in IBD patients could not be reproduced by the same authors." More recently, a genome wide association study has shown that polymorphisms in the gene of the IL-23 receptor are strongly linked to IBD. 70 More specifically, an infrequent mutation which leads to an Arg-Gly substitution at position 381 strongly protects against the development of Crohns disease. Recently, it was discovered that the same mutation also strongly protected against the development ofskin psoriasis" as well as AS.72 Taken together, these data strongly suggest that IL-23 plays an important role in the pathogenesis of SpA and related conditions. Although the underlying pathophysiologic pathways related to IL-23 still need to be deciphered in these conditions, it is noteworthy that monoclonal antibodies directed against the p40 subunit that is shared between IL-12 and IL-23 have a favourable therapeutic effect in IBD 7 3and psoriasis." Besides the overt gut inflammation in IBD associated SpA, histopathological analysis of gut biopsies from an unselected cohort of SpA patients showed that 50 to 60% ofSpA patients have signs ofsubclinical inflammation ofthe gut mucosa." Gut inflammation in SpA can be present in the ileum as well as the colon and can have the histological aspect ofacute inflammation, in which neutrophils predominate, aswell as chronic inflammation, that is characterized by a mixed inflammatory infiltrate together with structural remodelling of the gut mucosa . Immunohistochemical studies also show that, even in the absence of histopathological signs of inflammation, there are alterations in immune cell subsets in the gut of SpA parients." The noninflamed ileum of SpA
Synovial and Mucosal Immunopathology in Spondyloarthritis
77
patients shows a strong and selective increase in the number ofsub-epithelial CD 11c positive cells, while in the colon there is a slight increase in the number ofCD11a and VCAM-l positive cells. Intriguing is the fact that CD 11c positive dendritic cells in terminal ileum make massive amounts ofIL-23.77 The colon ofSpA patients was also characterized by an increase in the number ofCD68 and CD 163 positive macrophages. Interestingly, the CD163 positive macrophage subset was also increased in the colon ofpatients with Crohn'sdisease,but not colitis ulcerosa," and the synovium ofSpA patients." Together with changes in myeloid cell subsets, there are also changes in lymphoid cell subsets with an increase in IL-2 expanded CD 103 positive T-cells in the gut mucosa of SpA patients/" Also the cytokine profile ofT-lymphocytes from the gut,8! the joint,82 and peripheral blood 83.84 is disturbed in SpA: the T-cells produce less IFN gamma, IL-2 and TNF alpha after polyclonal stimulation,8l.83.84 indicating an impaired Thl response. Regulatory T-cell subsets and IL-l7 producing T-cells have not yet been investigated in SpA.
The Pathogenesis ofSpondyloarthritis: The Mucosa Hypothesis The frequent presence of gut inflammation in SpA patients has led to the hypothesis that the gut is the primary site of pathology and that recirculation of activated immune cells from the gut co musculoskeletal sites leads to joint inflammarlon.v The archetypical example of this is enterogenic ReA. In this SpA subtype, patients develop arthritis and/or spondylitis following acute bacterial enteritis. Supporting the recirculation hypothesis, May et al,86 demonstrated the clonal expansion ofthe same T-cell clones in the gut and joint ofa patient with ReA. The strongest argument supporting the concept of SpA as a primary mucosal pathology, however, comes from animal experimental research. HLA-B27 transgenic rats develop a systemic inflammatory syndrome that strongly resembles human SpA: the animals develop ileocolitis but also spondylitis, arthritis and psorias is-like skin and nail lesions." However, when the animals are held in a sterile environment they don't develop any inflammation ofthe gut and the joinr." Surprisingly considering the link with an MHC class I molecule, the disease is independent of the presence of CD8 positive T-lymphocytes, but dependent on CD4 Tvcells." More recently, i~ was discovered that antigen presenting cells from these animals were poor T-cell stimulators due to a disturbed engagement of costimulatory molecules ."
Innate Immunity and Mucosal Inflammation The inflammation predilection sites in SpA can roughly be divided in two categories : tissues with intimate contact with the outside world, more specifically the mucosae and the skin and sites that are exposed to great biomechanical stress, like the weight-bearing joints, the entheses, the insertion ofthe lens-fibers (anterior uvea) and the ascending aorta. A characteristic feature of the skin and mucosae is that the immune system is continuously exposed co potential pathogens. If the response co potentially dangerous organisms is coo weak this can lead to life-threatening bacreraemia, but an overzealous reaction can cause uncontrolled mucosal inflammation. Most likely, mucosal inflammation in SpA is the result of an excessive response of the immune system against non ho stile micro-organisms. The sensing ofmicro-organisms at the mucosal interface is performed by specialized receptors that are expressed by immune cells as well as structural cells,such as epithelial cellsand fibroblasts." There are many classesofthese "innate immune receptors" but the Toll-like recepcors (TLR) have been most thoroughly studied in recent years. The TLR recognize conserved molecular patterns that are expressed by micro-organisms. After activation ofTLR-2 or -4 by components of the bacterial cell wall, an intracellular Signaling cascade is activated that eventually leads to nuclear translocation ofnuclear factor kappa B (NF-kappaB), followed by the transcription ofinflammacory response genes such as cytokines, chemokines leukocyte growth factors and antimicrobial peptides." A novel hypothesis proposes that either a defect in the threshold for TLR activation or impaired dampening after the activation ofthese "innate immune receptors" plays a role in SpA inflammanon.P The regulation ofthe threshold for activation ofTLR's occurs at many levels, for instance the amount ofreceptor that is expressed at the cell surface, but also by cytosolic proteins
78
Molecular Mechanisms ofSpondyloarthropathies
that interfere with the transmission ofthe signal from the cell surface to the nucleus." Important examples ofsuch proteins are MyD88 adaptor like protein (Mal) 95 and silencer ofcytokine signaling (SOCS)-I,96 but also A20 97•98and CYLD.99.!OO Ofmajor relevance for SpA is the recent identification of a strong association between polymorphisms in the gene that codes for TLR2I01 and AS. In addition it was also shown that the levels ofTLR2 are increased on the surface of monocytes from SpA patiencs.':" Besides TLR, many other classes of "innate immune receptors" may playa role in SpA. A particular example is the family of NOD proteins that, in contrast to TLR, are localized in the cytosol. As already mentioned above, polymorphisms in the gene encoding NOD2 are related to gastrointestinal inflammation in Crohn's disease as well as in SpA. Interestingly, NO D2 is capable of regulating the threshold for activation of TLR2 but no other TLR's.103 Apart from these 2 examples, many other "innate immune receptors" are involved in the sensing and interpretation of danger signals at the mucosal interface and the challenge for the future will be to identify the role of these receptors in the SpA inflammatory process .
Adaptive Immunity and Mucosal Inflammation In human beings, the innate immune system never acts on its own but simultaneously activates and regulates the adaptive arm ofthe immune system, 104 and vice versa.!" Antigen presenting cells (APC) playa particular role in this process. After the phagocytosis of antigen, the APC mature, upregulare MHC class II and costimularory molecules at the cell surface and secrete T-cell polarizing factors. How the APC responds to a particular stimulus depends not only on the nature of the stimulus but also the micro-environment in which the APC resides.106 For instance, in normal circumstances gut dendritic cells induce an inflammation-suppressive T-cell phenotype while this mechanism isdefective in Crohn's disease.!" In Crohn'sdisease, mucosal dendritic cellsexpressmore TLR-2 and -4 on their cellular surface and secrete more IL-6 and IL-I2/IL-23 p40 subunit.!" The T-cells are polarized to the secretion ofIFN gamma and IL-I7 and need IL -23 to survive.l'" The specific blockade ofIL-23, but not IL-I2, prevents and cures T-cell-mediated colitis. I10 Because a polymorphism in the IL-23 receptor gene strongly protects against the development ofIBD,7° it is possible that a disturbed APC-T-Iymphocyte communication is a primary defect that, together with disturbed sensing and interpreting ofthe microbial gut content, leads to the development and perseverance ofgut inflammation in IBD patients. As already mentioned above, the same mutation in the IL-23 receptor also offers potent protection against AS.n Of relevance in this context, is also the strong linkage ofthe IL-I cytokine cluster with the presence ofSpA ,III.Il2 with IL-I beta being a very potent inducer Th 17 differentiation in the human immune system.I13
The Pathogenesis ofJoint Inflammation Together with mucosal inflammation, SpA patients also suffer from musculoskeletal inflammation. An important question here is ifthe mechanisms contributing to mucosal inflammation in IBD and potentially in SpA could also drive musculoskeletal disease in the latter. Although the bacterial exposure of these tissues is minimal in comparison with the gut and the skin, they are exposed to a tremendous biomechanical stress. Biomechanical stress can initiate an inflammatory program in 2 ways. Tissue damage, as found at the en theses of healthy subjects,'!" can induce an inflammatory response mediated by 'innate immune receptors". For instance, biglycan, which is a prominent ECM in tendons.!'?is released from the ECM when the tissue is damaged and is able to directly activate TLR-2 and TLR-4. 116 A second mechanism is the direct biomechanical stress on structural cells themselves. While low-intensity biomechanical stress inhibits inflammatory responses,117high-magnitude biomechanical stress induces the nuclear translocation ofNF-kappaB and the transcription of a pro inflammatory gene program.!" After initiating an inflammatory program, local structural cellssuch as chondrocytes and tendon fibroblasts will start to synthesize and secrete proinflamrnatory cytokines, chemokines and leukocyte growth factors , which in turn can attract and activate T and B-Iymphocytes, macrophages and neurrophils and thereby lead to the development of chronic inflammatory lesions.
Synovial and Mucosal Immunopathology in Spondyloarthritis
79
Conclusions In sununary, SpA inflammation is most likely the result of defects at different levels in the immune system. It is conceivable that disturbances in the sensing ofeither danger under the form of micro-organisms at the mucosal interface or biomechanical stress at musculoskeletal sites are important for the initiation ofthe inflammation and the determination ofpredilection sites. On the other hand, defects in APC-T-Iymphocyte interactions could be more important for determining the chronicity ofthe inflammation. We believe that the most important genetic alterations in SpA patients will be concentrated in the genes that code for proteins that are involved in the transmission and control of 2 important signal transduction axes: the NF-kappaB pathway that plays a crucial role in pro-inflammatory signal transduction and the STAT3 axis. The STAT3 pathway is essential for IL-23 receptor signal transductlon!" and may playa critical role in polarizing the immune system to SpA-like inflammatory responses. The final but crucial question is how HLA-B27 fits in this picture? Despite more than 30 years of intensive research, still nobody knows the precise role ofHLA-B27 in SpA pathogenesis. However, different hypotheses have been suggested. The first and, at first sight, most logical explanation is the "arthritogenic peptide hypothesis". This hypothesis states that an arthrirogenic peptide binds HLA-B27, but no other HLA molecule and is presented to cytotoxic Tdyrnphocytes. After many years of research, such a peptide has not yet been convincingly demonstrated and, as indicated, CD8 positive T-cells does not appear to be crucial players in the HLA-B27 transgenic rat model. A second hypothesis is based on the observation that a HLA-B27 transfected human monocyte cell line is less efficient in killing bugs such as Salmonella Spp.120 It is, however, unlikely that this mechanism would be the cause ofarticular disease because no viable bacteria can be found in the joints of SpA patients. A third hypothesis is based on the observation that HLA-B27 easily misfolds in the endoplasmatic reticulum, which leads to the induction of unfolded protein response and subsequent activation of a pro-inflammatory response program."! Although a generalized unfolded protein response is also unlikely to lead to the specificinflammatory pathology ofSpA,122 a basal unfolded protein response could modulate the threshold for NFKB activation by other stressors, such as biomechanicalloading. In this context, it is clear that future research will have to concentrate on the interaction between different 'stress' triggers and their associated Signaling programs rather than on a single pathophysiological pathway.
References I. Rudwaleir M, Baeten D. Ankylosing spondylitis and bowel disease. Best Pract Res Clin Rheumatol
2006 ; 20(3):451-71. 2. Cawley MI , Chalmers TM , Kellgren JH et al. Destructive lesions of vertebral bodies in ankylosing spondylitis. Ann Rheum Dis 1972; 31:345-58. 3. Francois RJ. Microradiographic study of the intervertebral bridges in ankylosing spondylitis and in the normal sacrum. Ann Rheum Dis 1965; 24:481-9. 4. de Vlam K, Mielants H, Verstaete KL et al. The zygapophyseal joint determines morphology of the enrhesophyre. J Rheumatol 2000; 27:1732-9 . 5. Appel H, Kuhne M , Spiekermann S et al. Immunohistologic analysis of zygapophyseal joints in patients with ankylosing spondylitis. Arthritis Rheum 2006 ; 54:2845 -51. 6. Francois RJ, Gardner DL, Degrave EJ er al. Histopathologic evidence that sacroiliitis in ankylosing spondylitis is not merely enthe sins, Arthritis Rheum 2000 : 43:2011-24. 7. Braun J, Bollow M, Neure L et al. Use of immunohistologic and in situ hybridization techniques in the examination of sacroiliac joint biopsy specimens from patients with ankylosing spondylitis. Arthritis Rheum 1995 ; 38 :499 -505. 8. Appel H, Kuhne M, Spiekermann S et al. Immunohistochemical analysis of hip arthritis in ankylosing spondylitis: evaluation of the bone-cartilage interface and subchondral bone marrow. Arthritis Rheum 2006; 54:1805-13. 9. Baeten D, Demetrer P, Cuvelier C et al. Comparative study of the synovial histology in rheumatoid arthritis, spondyloarthropathy and osteoarrhritis: influence of disease duration and activity, Ann Rheum Dis 2000; 59 :945-53. 10. Kruithof E, Baeten D, De Rycke L er al. Synovial histopathology of psoriatic arthritis, both oligo- and polyarticular, resembles spondyloarthropachy more than it does rheumatoid arthritis. Arrhriris Res Ther 2005; 7:R569-80.
80
Molecular Mechanisms ofSpondyloarthropathies
II. Smeets TJ. Dolhain RJ. Breedveld FC et al. Analysis of the cellular infiltrates and expression of cytokines in synovial tissue from patients with rheumatoid arthritis and reactive arthritis. J Parhol 1998; 186:75-81. 12. van Kuijk AW, Reinders-Blankert P. Smeets TJ et al. Detailed analysis of the cell infiltrate and the expression of mediators of synovial inflammation and joint destruction in the synovium of patients with psoriatic arthritis: implications for treatment. Ann Rheum D is 2006; 65 :1551-7. 13. Cunnane G. Bresnihan B. FitzGerald O. Immunohistologic analysis of peripheral joint disease in ankylosing spondylitis. Arthritis Rheum 1998; 41:180 -2. 14. Kristiansen M. Graversen JH, Jacobsen C ec al. Identification of the haemoglobin scavenger receptor. Nature 2001; 409:198-201. 15. Baeten D. Kruithof E. De Rycke Let al. Infiltration of the synovial membrane with macrophage subsets and polymorphonuclear cells reflects global disease activity in spondyloarthropathy, Arthritis Res Ther 2005; 7:R359-69. 16. Kruirhof E, De Rycke L. Vandooren B er al. OMERACT Special Interest Group on Synovial Analysis in Clinical Trials. Identification of synovial biomarkers of response to experim ent al treatment in early-phase clinical trials in spondylarthritis. Arthritis Rheum 2006; 54:1795-804. 17. Reece RJ, Canete JD, Parsons WJ et al. Distinct vascular patterns of early synovitis in psoriatic, reactive and rheumatoid arthritis. Arthritis Rheum 1999; 42 :1481-4. 18. Canete JD. Rodriguez JR , Salvador G et al. Diagnostic usefulness of synovial vascular morphology in chronic arthritis. A systematic survey of 100 cases. Semin Arthritis Rheum 2003 ; 32:378-87. 19. Francois RJ. Braun J. Khan MA . Entheses and enthesitis: a histopathologic review and relevance to spondyloarthricides. Curr Opin Rheumatol 2001; 13:255-64. 20. McGonagle D. Marzo-Ortega H , O'Connor P et al. Histological assessment of the early enthesitis lesion in spondyloarthropathy, Ann Rheum Dis 2002 ; 61 :534-7 . 21. Laloux L, Voisin MC . Allain J ec al. Immunohistological study of enrheses in spondyloarrhroparhies: comparison in rheumatoid arthritis and osteoarthritis. Ann Rheum Dis 2001 ; 60:316-21. 22. Said-Nahal R, Miceli-Richard C , D'Agostino MA er al. Phenotypic diversity is not determined by independent genetic facto rs in familial spondylarthropathy, Arthritis Rheum 2001 ; 45 :478-84 . 23. Calin A. Marder A. Marks S et al. Familial aggregation of Reiter's syndrome and ankylosing spondylitis : a comparative study. J Rheumatol1984; 11:672-7. 24. Gossec L, van der Heijde 0, Melian A er al. Efficacy of cyclo-oxygenase-2 inhibition by etoricoxib and naproxen on the axial manifestations of ankylosing spondylitis in the presence of peripheral arthritis. Ann Rheum Dis 2005 ; 64:1563-7. 25. Braun J, Zochling J, Baraliakos X et al. Efficacyof sulfasalazine in patients with inflammatory back pain due to undifferentiated spondyloarthritis and early ankylosing spondylitis: a mulricentre randomised controlled trial. Ann Rheum Dis 2006 ; 65 :1147-53 . 26. Clegg DO. Reda OJ, Weisman MH et al. Comparison of sulfasalazine and placebo in the treatment of ankylosing spondylitis. A Department of Veterans Affairs Cooperative Study. Arthritis Rheum 1996; 39 :2004 -12. 27. Vandooren B. Kruithof E, ¥u DT er al. Involvement of matrix meralloproreinases and their inhibitors in peripheral synovitis and down-regulation by tumor necrosis factor alpha blockade in spondylarrhropathy, Arthritis Rheum 2004; 50:2942-53. 28. Maksymowych WP, Landewe R. Conner-Spady B et al. Serum matrix metalloproteinase 3 is an independent predictor of structural damage progression in patients with ankylosing spond ylitis. Arthritis Rheum 2007; 56:1846-53. 29. Brophy S, Mackay K, Al-Saidi A er al. The natural history of ankylosing spondylitis as defined by radiological progression. J Rheumatol 2002; 29:1236-43. 30. Dougados M, van der Linden S, Juhlin R et al. The European Spondylarthropathy Study Group preliminary criteria for the classification of spondylarthropathy, Arthritis Rheum 1991 ; 34:1218-27. 31. van der Linden S. Spondyloarth ropathies.Jn. Kelley. Harris, Ruddy and Sledge. Textbook of Rheumatology. 5th Ed. 1997; 969-82. 32. Polzer K, Schett G. Zwerina J. The lonely death : chondrocyte apoptosis in TNF-induced arthritis. Autoimmunity 2007; 40 :333-6. 33. Zwerina J. Redlich K, Polzer K et al. TNF-induced structural joint damage is mediated by IL-l . Proc Nad Acad Sci USA 2007 ; 104:11742-7. 34. Redlich K, Hayer S, Ricci R et al. Osteocla sts are essential for TNF-alpha-mediated joint destruction. J Clin Invest 2002 ; 110:1419-27. 35. Diarra D. Stolina M. Polzer K et al. Dickkopf-l is a master regulator of joint remodeling Nat Med 2007 ; 13:156-63 . 36. Pettit AR. Ji H, von Stechow 0 er al. TRANCE/RANKL knockout mice are protected from bone erosion in a serum transfer model of arthritis. Am J Parhol 2001 ; 159 :1689-99.
SynovialandMucosal Immunopathology in Spondyloarthritis
81
37. Lam J. Takeshita S. Barker JE et al. TNF-alpha induces osteoclaseogenesis by di rect stimulation of macrophages exposed to permissive levels of RANK ligand. J Clin Invest 2000, 106:1481-8. 38. Wei S. Kitaura H . Zhou P et al. IL-l mediates TNF-induced osreoclasrogenesis. J Clin Invest 2005, 115:282-90 . 39. Rountree RB. Schoor M. Chen H et al. BMP receptor signaling is required for postnatal maintenance of articular cartilage . PLoS Bioi 2004, 2:e355. 40. Lories RJ. Derese I. Luyten FP. Modulation of bone morphogenetic protein signaling inhibits the onset and progression of ankylosing enthesitis, J Clin Invest 2005 , 115:1571-9 . 41. van Beuningen HM. Glansbeek HL. van der Kraan PM et al. Differential effects of local application of BMP-2 or TGF-beta 1 on both articular cartilage composition and osteophyte formation. Osteoarthritis Cartilage 1998, 6:306-17. 42. van Beuningen HM. van der Kraan PM . Arntz OJ er aI. Transforming growth factor-beta 1 stimulates articular chondrocyte proteoglycan synthesis and induces osteophyte formation in the murine knee joint. Lab Invest 1994 , 7 1:279-90. 43. Valencia X. Higgins JM . Kiener HP ct al. Cadherin-11 provides specific cellular adhesion between fibroblast-like synoviocyte s. J Exp Med 2004 ; 200:1673-9. 44. Lee DM. Kiener HP, Agarwal SK ec al. Cadherin-11 in synovial lining formation and pathology in arthritis. Science 2007,315:1006-10. 45. Vandooren B, Cantaert T, van Lierop MJ et al. Melanoma Inhibitory Activity. a biomarker related to chondrocyte anabolism. is reversibly suppressed by proinRammatory cytokines in rheumatoid arthritis. Submitted for publication. 46. Vandooren B, Cantaert T. Noordenbos T er al. The abundant synovial expression of the RANK-RANKL-OPG system in peripheral spondyloarthritis is partially disconnected from inflamrnation. Arthritis Rheum . In press. 47. Ritchlin CT. Haas-Smith SA. Li P et al. Mechanisms ofTNF-alpha- and RANKL-mediated osteoclastogenesis and bone resorption in psoriatic arthritis. J Clin Invest 2003, 111:821 -31. 48. Lories RJ. Dere se 1, Ceuppens JL et al. Bone morphogenetic proteins 2 and 6. expressed in arthritic synoviurn, are regulated by proinflammatory cyrokine s and differ entially modulate fibroblast -like synoviocyte apoptosis. Arthritis Rheum 2003 ; 48:2807-18. 49. Glansbeek HL, van Beuningen HM. Vitters EL et al. Bone morphogenetic protein 2 stimulates articul ar cartilage proteoglycan synthesis in vivo but does nor counteract interleukin -I alpha effects on proteoglycan synthesis and content. Arthritis Rheum 1997,40:1020 -8. 50. Chen Y, Whetstone HC, ¥oun A et al, Bera-carenin signaling pathway is crucial for bone morphogenetic protein 2 to induce new bone formation . J Bioi Chern 2007 , 282:526-33. 51. Ribben s C. Martin y Porras M . Franchimont N et al. Increased matrix meralloproeeinase-S serum levels in rheumatic diseases: relat ionship with synovitis and steroid treatment. Ann Rheum Dis 2002 , 61:161-6. 52. Chen CH. Lin KC. ¥u DT et al. Serum matrix meralloproreinases and tissue inhibitors of meralloproteinases in ankylosing spondylitis: MMP-3 is a reproducibly sensitive and specific biomarker of disease activity. Rheumatology 2006, 45 :414-20. 53. YangC. Gu J. Rihl M er al. Serum levels of matrix metalloproteinase 3 and macrophage colony-stimulating factor 1 correlate with disease activity in ankylosing spond ylitis. Arthritis Rheum 2004, 51:69 1-9. 54. Kim TH. Stone M. Payne U et al. Inman Cartilage biomarkers in ankylosing spondylitis: relationship to clin ical variables and treatment response. Arthritis Rheum 2005 , 52:885-91. 55. Maksymowych WP, Poole AR. Hiebert Let al. Eranercepr exerts beneficial effects on articular cartilage biomarkers of degradation and turnover in patients with ankylosing spondylitis. J Rheurnatol 2005 , 32:1911-7. 56. Olsen AK. Sondergaard BC. Byrjalsen I er al. Anabolic and catabolic function of chondrocyte ex vivo is reflecced by the metabolic processing of type II collagen. Osteoarthritis Cartilage 2007, 15:335-42. 57. Grisar J. Bernecker PM. Aringer M et al. Ankylosing spondylitis. psoriatic arthritis and reactive arthritis show increased bone resorption. but differ with regard to bone formation . J Rheumarol 2002, 29:1430-6. 58. WOO JH. Lee HJ. Sung IH er al. Changes of clinical response and bone biochemical markers in patients with ankylosing spondylitis taking etanercept. J Rheumatol2007, 34:1753-9. 59. Baeten D, Kruithof E, De Rycke L er al. Diagnostic classification of spondylarrhropathy and rheumatoid arthritis bysynovial histopathology: a prospective study in 154 consecutive patients. Arthritis Rheum 2004 , 50:2931-41. 60. Baeten D. Kruithof E, De Rycke L et al. Infiltration of the synovial membrane with macrophage subsets and polymorphonuclear cells reflects global disease activity in spondyloarthropathy, Arthritis Res Ther 2005 ; 7:R359-69.
82
Molecular M echanism s ofSpondyloarth ropathies
6 1. Kruithof E, De Rycke L, Vandooren B et al. OMERACT Special Interest Group on Synovial Anal ysis in Clinical Trials. Identification of synovial biomarkers of respons e to experimenta l treatment in early-phase clinical tr ials in spondylarthritis. Arthritis Rheum 2006; 54 :1795 -804. 62 . Breb an M. Genetics of spon dyloarthri tis. Best Pract Res Clin Rheurnatol 2006; 20 :593 -9. 63 . Loftu s EV]r. Clin ical epidemiology of inflammato ry bowel d isease: Incid ence, prevalence and environmental influences. Gastroenterology 2004 ; 126:1504 -17. 64. Porcher R, Said-Nahal R, D'Agostino MA er al. Two major spondylarthropathy ph enotypes are distin gui shed by pattern analysis in mult iplex families. Arthritis Rheum 2005; 53:263-71. 65. de Vlam K, Mielanrs H , Cuvelier C et al. Spondyloarrhropath y is undere stim ated in inflammatory bowel d isease: p revalence and HLA association. ] Rh eurnatol 2000; 27:2860 -5. 66. Hugot ]P, C hamaillard M, Zouali H et al. Associat ion of NOD2 leucine-ric h repeat variant s with susceptibility to Crohn's disease. Natur e 200 I ; 4 11:599-603. 67. Lauken s D, Peeter s H , Marichal D et al. C ARD I 5 gene pol ymorphisms in patients with spondyloarthropath ies ident ify a specific phenotype previou sly related to C rohn's disease. Ann Rheum D is 2005 ; 64 :930-5. 68 . Bjarn ason I, Helgason KO, Geirsson A] et al. Subcl inical inte stinal inflamm ation and sacroiliac change s in relatives of patients with ankylosing spondylitis. Gastroenterology 2003 ; 125:1598-605 . 69. Peeters H , Vander Cru yssen B, Mielants H er al. Clinical and genetic facto rs associated with sacroiliitis in C rohn's disease.] Gasrroenrerol Hepatol 2007 ; [Epub ahead of print] . 70 . Duerr RH, Taylor KD , Brant SR et al. A genome-wide association stu dy identifi es IL23R as an inflam matory bowel disease gene. Science 2006 : :'11 4:1461 ·3. 7 1. Capon I', Di Meglio P, Szaub ] et al. Sequence variant s in the genes for th e imerleukin-23 receptor (IL23R) and its ligand (IL12B) confer protection against psoria sis. Hum Gene r 2007; 122:201-206. 72. Brown MA . Nature Genetics . In Press. 73. Mannon PJ, Fuss I], Mayer L et al. Anti -IL-12 Crohn's D isease Stud y Group. Anti -imerleukin-12 ant ibod y for active Crohn's disease. N Engl] Med 2004 ; 351 :2069-79 . 74. Krueger GG, Langley RG, Leonardi C et al. CNTO 1275 Psor iasis Study Group. A human interleukin -12/23 monoclonal ant ibody for the treatment of psoriasis. N Engl ] Mcd 2007; 356 :580-92. 75. D e Vos M, Cuvelier C, Mielants H et al. I1eocolonoscopy in seronegative spondylarthroparhy, Gastroente rology 1989; 96:339-44. 76. Demetter P, Van Huysse ]A , D e Keyser F et al. Increase in lymph oid follicles and leuko cyte adhe sion mole cules empha sizes a role for th e gut in spondyloarrhropathy pathogenesis. ] Parhol 200 2; 198:517-22. 77. Becker C . Wi rt z S. Blessing M et al. C onst itutive p40 promote r activation and IL-23 product ion in the term in al ileum med iated by dendritic cells. ] C lin Invest 2003; 112:693-706. 7 8. D ernerter P, D e Vos M, Van Hu ysse]A et al. Colon mucosa of pat ients both with spond yloarrhriris and Crohns d isease is enriched with macroph ages expressing the scavenger receptor CO 163. Ann Rheum D is 2005 ; 64 :321-4. 79. Baeten D, Demerter P, Cuvelier CA et al. Macrophages expressing the scavenger receptor CD 163 : a link between immune alterations of th e gut and synovial inflammation in spondyloarthropathy; ] Pathol 2002; 196 :343-50. 80. Van Darnme N, Elewaur D, Baeten 0 et al. Gut mucosal T-cell lines from ankylosing spondylitis patients are enriched with alphaEbeta 7 integrin. Clin Exp Rheumatol 2001 ; 19:681- 7. 81. Van Damrne N, De Keyser F, Demetter P ct al. The proportion ofThI cells, wh ich prevail in gut mucosa , is decrea sed in inflammator y bowel synd rome. C lin Exp Immunol 2001 ; 125:383 -90. 82. Canete ]0. Martinez SE. Farres] et al. Differential ThI /Th2 cyrokine patterns in chronic arthritis: interferon gamma is highly expressed in synovium of rheumatoid arthritis compared with seronegative spondyloarthroparhies, Ann Rheum Dis 2000; 59:263-8. 83. Baeten D , Van Damme N. Van den Bosch F er al. Impaired ThI cytokinc produ ction in spondyloarthropathy is restored by anti-TNFalpha. Ann Rheum Dis 2001 ; 60 :750-5 . 84. Rudwaleit M , Siegert S. Yin Z et al. Low T-cell production ofTNFalph a and IFNgamma in ankylosing spondylitis: its relation to HLA-B27 and influence of the TNF-308 gene polymorph ism. Ann Rheum Di s 200 I ; 60( I):36-42. 85. Baeten D. De Keyser F, Mielants H er al. Imm une linkages between inflamm atory bowel disease and spo ndyloarthropath ies. Curr Op in Rheurnatol 2002 ; 14:342-7. 86. May E. Marker-Hermann E, Wittig BM er al. Ident ical T- cell expansion s in th e colon mucosa and the synovium of a patient with enrerogeni c spo ndyloart hropathy, Gastroent erology 2000: 119:1745 -55. 87. H amm er RE, Maika SO. Richardson ]A et al. Spont aneous inflamm ator y disease in tran sgenic rats exp ressing HLA-B27 and human beta 2m: an anim al model of HLA-B27-associated hum an disorders. C ell 1990 ; 30 :1099-112.
Synovial andMucosal Immunopathology in Spondyloarthritis
83
88. Taurog JD. Richardson JA. Croft JT er al. The germfree State prevents development of gut and joint inAammatory disease in HLA-B27 transgenic rats. J Exp Med 1994, 180:2359-64. 89. Breban M. Fernandez-Sueiro JL. Richardson JA et al. T-cells. but not thymic exposure to HLA-B27. are required for the inAammatory disease of HLA-B27 transgenic rats. J Immunol 1996; 156 :794-803. 90. Hacquard-Bouder C . Chimenti MS. Giquel B et al. Alteration of antigen-independent immunologic synapse formation between dendritic cells from HLA-B27-transgenic rats and CD4+ T-ceIls : selective impairment of costimulatory molecule engagement by mature HLA-B27. Arthritis Rheum 2007; 56:1478-89 . 91. Janeway CA Jr. Medzhitov R. Innate immune recognition. Annu Rev Immunol 2002, 20 :197-216. 92. Li Q, Verma 1M. NF-kappaB regulation in the immune system. Nat Rev Immunol 2002, 2:725 -34 . 93. De Rycke L. Kruithof E. Vandooren B et al. Pathogenesis of spondyloarthritis: insights from synovial membrane studies. Curr Rheumatol Rep 2006; 8:275-82. 94. Liew FY. Xu D . Brint EK ec al. Negative regulation of toll-like receptor-mediated immune responses. Nat Rev Immunol 2005 ; 5:446-58. 95. Fitzgerald KA. Palsson-McDermott EM. Bowie AG er al. Mal (MyD88-adapter-like) is required for Toll-like receptor-s signal transduction. Nature 2001 , 413 :78-83. 96. Mansell A. Smith R. Doyle SL et al. Suppressor of cytokine signaling 1 negatively regulates Toll-like receptor signaling by mediating Mal degradation. Nat Immuno12006; 7:148-55. 97. Lee EG. Boone DL. Chai S et al. Failure to regulate TNF-induced NF-kappaB and cell death responses in A20-deficient mice. Science 2000; 289 :2350-4. 98. Boone DL. Turer EE. Lee EG er al. The ubiquitin-modifying enzyme A20 is requ ired for termina tion of Toll-like receptor respon ses. Nar Immunol 2004; 5:1052-60. 99. Trompouki E, Hatzivassiliou E. Tsichritzis T et al. CYLD is a deub iquitinating enzyme that negatively regulates NF-kappaB activation by TNFR famil y members. Nature 2003; 424:793-6. 100. Yoshida H . Jono H, Kai H et al. The tumor suppressor cylindromatosis (CYLD) act s as a negative regulator for toll-like receptor 2 signaling via negative cross-talk with TRAF6 AND TRAF7. J BioI Chern 2005, 280 :41111-21. 101. Inman RD . Rohekar S, Reena R et at. TLR2. but not TLR4. variants are associated with reactive arthritis. Poster pre sentation American College of Rheumatology meeting 2007. 102. De Rycke L. Vandooren B. KruithofE et aI. Tumor necrosis factor alpha blockade treatment down -modulates the increased systemic and local expression of Toll-like receptor 2 and Toll-like receptor 4 in spondylarrhropathy, Arthritis Rheum 2005; 52:2146-58. 103. Watanabe T. Kitani A. Murray PJ et al. Nucleotide binding oligomerization domain 2 deficiency lead s to dysregulated TLR2 signaling and induction of antigen-specific colitis. Immunity 2006; 25 :473-85. 104. Iwasaki A . Medzhirov R. Toll-like recept or control of the adaptive immune responses. Nat Immunol 2004, 5:987-95. 105. Kim KD. Zhao J. Auh Set at. Adaptive immune cells temper initial innate responses . Nat Med 2007; 13 :124 8-52 106. Pulendran B. Palucka K, BanchereauJ. Sensing pathogens and tuning immune responses. Scien ce 2001 , 293 :253-6. 107. Rimoldi M , C h ieppa M . Salucci V er at. Intestinal immune homeostasis is regulated by the crosstalk between epithelial cells and dendritic cells. Nat Immunol 2005 ; 6 :507-14. 108. Hart AL . AI-Hass i HO. Rigby RJ er al. Characteristics of intestinal dendritic cells in inflammatory bowel diseases. Gastroenterology 2005 ; 129:50-65. 109. Annunziato F, Cosmi L. Santarlasci V et al. Phenotypic and functional feature s of human Th 17 cells. J Exp Med 200 7 ; 204:1849-61. 110. Elson CO. Cong Y. Weaver CT er al. Monoclonal anti-interleukin 23 reverse s active colitis in a Tvcell-mediared model in mice . Gastroenterology 2007, 132:2359-70. Ill. Timms AE , Crane AM. Sims AM et al. The inte rleukin 1 gene cluster contains a major susceptibility locu s for ankylosing spondylitis. Am J Hum Genet 2004; 75 :587-95. 112. Maksymowych WP. Rahman P. Reeve JP et al. Association of the III gene cluster with susceptibility to ankylosing spondylitis: an analysis of three Canadian populations. Arthritis Rheum 2006; 54 :974-85. 113. Wilson NJ, Boniface K. Chan JR et at. Development, cytokine profile and function of human interleukin 17-producing helper T-cells. Nat Immunol 2007; 8:950-7. 114. Benjamin M. Toumi H. Suzuki D et at. Microdamage and altered vascularity at the enrhesis-bone int erface provides an anatomic explanation for bone involvement in the HLA-B27-associated spondylarthrirides and allied disorders. Arthritis Rheum 2007; 56:224-33. 115. Corps AN, Robinson AH, Movin T et al. Increased expression of aggrecan and biglycan mRNA in Achilles rendinoparhy, Rheumatology 2006 ; 45 :291-4. 116. Schaefer L. Babelova A. Kiss E et al. The matrix component biglycan is proinAammatory and signals through Toll-like receptors 4 and 2 in macrophages. J Cl in Invest 2005 , 115:2223-33.
84
Mo lecular Mechanisms ofSpondyloarthropathies
117. Dossumbekova A, Anghelina M, Madhavan S et al. Biomechanica l signals inhibit IKK activity to attenuate NF-kappaB transcription activity in inflamed chond rocytes. Art hri tis Rheum 2007; 56:3284-3296. 118. Agarwa l S, Deschne r ], Long P et al. Role of NF -kapp aB transcripti on facto rs in ant iinRammatory and proinRammatory actions of mechanical signals. Arthritis Rheum 2004 ; 50: 3541-8. 119. Yang XO , Panopoulos AD , Nurieva R et al. STAT3 regu lates cytok ine-rncdiated generation of inRammato ry helper T-cells. ] Bioi Chern 2007; 282 :9358-63. 120. Laitio 1', Vir tala M, Salmi M er al. HLA-B27 modu lates intracellular sur vival of Salmonella enteritidis in human monocytic cells. Eur] Irnmun ol 1997; 27:1331-8. 121. Turner M], Sowders 01', DeLay M L et al. HLA-B27 misfold ing in tra nsgenic rats is associated with activation of the unfo lded prot ein response. I Immun ol 2005 ; 175 :2438-48. 122. Tran TM, Dorris ML, Sarumtira N er al. Additional human beta2-microglobulin curbs HLA-B27 misfo lding and promotes art hri tis and spondylitis without coliti s in male HLA-B27-transgenic rats. Arthri tis Rheum 2006; 54:13 17-27.
CHAPTER 6
Bone Loss in the Spondyloarthropathies: Role ofOsteoclast, RANKL, RANK and OPG in the Spondyloarthropathies Allen P. Anandarajah* and Edward M. Schwarz
Abstract
B
one loss is a common find ing in the spondyloarthropathies.It may be localized and present as erosions or be generalized and cause osteoporosis. The pathogenesis ofbone loss in the spondyloarthropathies is yet to be fully understood. There is however increasing evidence to support a role for the osteoclasts in bone erosions. Similarly a balance between the receptor activator of nuclear factor-xls ligand (RAN KL) and osteoprotegerin (OPG) levels is thought to regulate osteoclastic activity and therefore bone loss in the inflammatory arthritides. In this chapter we will review the recent literature on the role of osteoclasts and the RANKL/OPG system in the various spondyloarthropathies and try to formulate a hypothesis for the possible mechanism ofbone loss in th is group of inflammatory rheumatic disease.
Introduction The spondyloarthropathies comprise a group ofinterrelated. chronic inflammatory rheumatic disorders that share several common clinical, epidemiologic and genetic factors. Included among this group are ankylosing spondylitis (AS), psoriatic arthritis (PsA), reactive arthritis (ReA) , enteropathic arthritis and undifferentiated spondyloarthropathy. In addition to peripheral joint diseasewhich may be symmetric or asymmetric and often is oligoarticular, these arthropathies have the propensity to often involve the axial skeleton. The SpA are characteristically associated with inflammation of the enthesis, synovium, cartilage and bone. Additionally, bone loss is an increasingly well recognized finding in the spondyloarthropathies, While the pathophysiology ofbone loss in the spondyloarthropathies is yet to be clearly understood. the mechanisms underlying the process ofbone loss has been the focus of several recent studies. Bone loss can be localized and manifest as erosions or be generalized and present as osteoporosis. Although initially thought to have distinct pathophysiological mechanisms, recent studies have suggested that both types ofbone loss are mediated by osteoclast activation. It is now known that osteoclast activation in turn is regulated by the receptor activator ofnuclear factor-xf ligand (RANKL)/receptor antagonist of nuclear faceor-xll (RANK)/osteoprotegerin (OPG) system. Indeed in patients with rheumatoid arthritis, as association between the presence ofjoint erosions and generalized osteoporosis has been described.' Bone loss in the spondyloarthropathies, as with rheumatoid arthritis, is thought to be a consequence ofdistorted bone remodeling in the context of chronic inflammation.
'Corresponding Author: Allen P. Anandara jah-Clinical Immunology Research Center, 601 Elmwood Avenu e, Box 695 , Rochester, New York 14642 , USA. Email: aIlen_anandaraja h@urmc. roc hester.edu
Molecular Mechanisms ofSpondyloanbropathies, edited by Carlos Lopez-Larrea and Roberto Dfaz-Pefta. ©2009 Landes Bioscience and Springer Science+Business Media.
86
Molecular Mechanisms ofSpondyloarthropathies
Bone erosion often signifieslossofmineralized tissue and is commonly associated with a break in the cortical lining. Initial descriptions ofbone erosions date back to the 19th century-The advances in radiographic imaging since then have made plain radiographs a valuable tool in the assessment ofjoint damage in inflammatory arthritides. Indeed, changes on plain radiographs are an essential diagnostic criterion for several inflammatory arthritides including the spondyloarthroparhles. The modified New York criteria for the diagnosis of ankylosing spondylitis rely mainly on the detection of sacroilitis on plain radiographs.' In recent years the use musculoskeletal ultrasound and magnetic resonance imaging have further highlighted the importance of bone erosions as a diagnostic and prognostic marker in the assessment of spondyloarthritis. Bone erosions in the spondyloarthropathies are most commonly seen in the sacroiliac joints ofpatients with AS and in the peripheral and/or sacroiliac joints ofpatients with PsA. The critical mechanisms that underlie pathologic bone resorption in the spondyloarthropathies however are not well understood. Most of the current knowledge of the pathogenesis of bone erosions in the inflammatory arthritides comes from studies in patients with rheumatoid arthritis. There are however increasing reports to support an essential role for osteoclasts and the RANKL/RANK/OPG system in the pathogenesis oferosions in spondyloarthropathies. Osteoporosis and consequent fragility fractures are also well recognized complications of the spondyloarthropathies. Several factors including immobility, smoking, malabsorption and drug therapy contribute to generalized bone loss in the spondyloarthropathies. In addition to these, recent data demonstrate an important association between systemic inflammation and diffuse diffuse bone loss, in inflammatory arthritis. Generalized bone loss or osteoporosis is particularly common in patients with AS and chronic inflammatory bowel diseases." While it has been less well studied in the other spondyloarthroparhies, recent studies have suggested that osteopenia and osteoporosis may also be more common than previously reported in PsA.6.7While the exact cause of bone loss in these disease are not known, recent reports indicate that the RANKL/RANK/ OPG system and activated osteoclasts may be key players in the pathogenesis ofdiffuse bone loss in the spondyloarthropathies. The discovery ofthe RANK/RANKL/OPG pathway has helped relate the interaction between inflammation and bone lossin the inflammatory arthritides. The elucidation ofthe RANKL/ 0 PG system has also helped support the hypothesis that osteoporosis and bone erosions may share a common pathophysiology in the inflammatory arthritides. In this article we will discuss emerging data that support the role of the RANK/RANKL/OPG system and osteoclasts in pathogenesis ofbone loss in AS, PsA, inflammatory bowel diseases and reactive arthritis.
Pathogenesis ofBone Loss Bone remodeling is a tightly regulated, continuous process in the adult skeleton that allows for the repair ofmicro damage while maintaining the mechanical strength and structure ofbone . This process involves bone resorption by osteoclasts, specialized cells ofmonocyte lineage followed by bone formation by the osteoblasts. Under physiologic conditions, the amount ofbone resorption is exactly matched by the amount ofbone formation . An imbalance in these processes that favors increased osteoclastic activity results in bone loss that can be focal and present as bone erosion or be generalized and lead to osteoporosis.
Osteoclast Osteoclasts are large multinucleated cells (50-100 mm in diameter) that are derived from the mononuclear cell precursors ofthe monocyte/macrophage Iineage .t They are a fusion product of up to 20 single cells that form a syncytium and have abundant mitochondria, numerous lysosomes and free ribosomes. Osteoclasts are the principal cells involved in bone resorprlon.t Ir is uniquely adapted to degrade mineralized matrix through the expression ofintergrin s, matrix metallopteinases and various enzymes including cathepsins and mast cell proteinases'? (Fig. 1).The cyclicalfunction ofosteoclast starts with the migration and attachment of the cell to the bone. The attachment of the osteoclast to bone is mediated by intergrins which are heterodimeric transmembrane matrix
87
Bone Loss in the Spondyloarthropatbies
Hematopoietic precursor
1 Myeloid progenitor
j Pre-osteoclast! CD 14 monocyte
j Osteoclast
Mature Osteoclast Figure 1. Osteoclast.
receptors. The resorptive process then follows . When the osteoclast comes in contact with bone it forms a complex villous-like structure. the rumed membrane. which greatly increases the surface extent of the cell. 11 At the same time. the cell's fibrillar actin organize to from a circular structure. the actin ring or sealing zone, that isolates the resorptive microenvironment from the rest of the cell space .' ? The acidic environment. created by the action of a proton pump. helps dissolve the mineral component of the bone matrix. Once resorption is complete. the actin ring and ruffled membrane are disassembled. The cell then detaches and migrates to the next area ofbone remodeling. Death of the osteoclast is quickly followed by their removal by phagocytes. The average lifespan of osteoclasrs is about 2 weeks. Osteoclasts playa pivotal role in the pathogenesis of bone loss in inflammatory arthritides. Recent studies revealed that functional ostcoclasts are formed within a days of onset of inflammation in animal models of chronic immune mediated arthritis.n Other studies demonstrated that osteoclast-like multinucleated cells are detected in areas ofextensive bone resorption, in rats with collagen induced arthritis." Furthermore. it was shown that mice lacking osteoclasts were resistant to arthritis induced bone erosion despite evidence for significant inflammation. I 5.16 Studies with potent bisphosphonates have also demonstrated a remarkable reduction in extent of bone erosions in animal rnodels.F:" Taken together these studies provide ample support for the role of osreoclasrs in the pathogenesis of bone erosions.
88
Molecular Mechanisms ofSpondyloarthropathies
Initial evidence implicating the role of osteoclasrs in bone resorption in the inflammatory arthritides was provided by Bromley et al when they demonstrated the presence ofacid phosphatase positive multinucleated cells at the bone pannus junction and subchondral bone marrow ofpatients with rheumatoid arthritis.19 .2oSince then several other studies have identified osteoclasts at sites of bone erosion in rheumatoid arthritis,21.22while other reports have demonstrated that osteoclast-like cells can also be detected in the synovium of RA patients. 23.24The discovery that macrophages can be induced to differentiate into mature, active osteoclasts has led to the speculation that the multinucleated cells at the bone pannus junction are derived from synovial macrophages." This hypothesis is further supported by findings that macrophages from rheumatoid synovium can differentiate into osteoclast like cells in the presence of 1, 2S dihydroxyvitamin D 3 and macrophage colony stimulating factor (M-CSF) .26 Similarly a loss of bone homeostasis, with resultant increased bone loss compared with bone formation, results in a reduction in skeletal mass or generalized bone loss. This increased bone loss is thought to be secondary to increased number ofosteoclastic cells rather than increased activity of these cells in postmenopausal osteoporosis." In addition to growth factors and cytokines, several hormones and mechanical signals, are thought to playa role in the regulation ofosteoclasts. Osteoclastogenesis, the process of differentiation of myeloid progenitor cells into cells of osteoclast lineage, is influenced by several hormones and cytokines (Fig. 2). These cytokines and hormones include tumor necrosis factor-a (TNF-a), interleukin-I (IL-I), interleukin-6 (IL-6), transforming growth factor-a (TGF-a, transforming growth factor-B (TGF-13) , prostaglandin E2 (PGE2), 1, 2S hydroxyvitamin D3, calcitonin and parathyroid hormone. Although these factors affect osteoclastogenesis at distinct stages of development they are thought to act via a common final pathway involvingRANKL, RANK and OPG pathway," Indeed, there is mounting evidence that RANKL may be the central mediator ofosteoclast function and therefore plays an essential role in the pathophysiology of bone loss pathologies."
RANK/RANKL/OPG System In addition to the osreoclasts, bone resorption is influenced by several cytokines and factors. Indeed, osteoclast differentiation, recruitment and function are regulated by these cytokines and factors that act directly on osteoclast precursors or on osteoclasrs. The essential and direct acting factor for osteoclast differentiation has been cloned and identified as RANKL, also known as osteoclast differentiation factor (ODF), osteoprotegerin ligand (OPG-L) and tumor necrosis factor-related activation-induced cytokine (TRANCE). It has been suggested that RANKL is necessary and
Membrane transport ers
• • ••• Acidified resorbing
zone
Figure 2. Osteoclastogenesis.
Multiple nuclei
I Ruffiedborder
Bone Loss in the Spondy/oarthropathies
89
sufficient for osreoclastogenesis.P The balance between RANKL and its decoy receptor, OPG, is now known to play central role in the regulation ofbone remodeling and bone loss. The identification ofRANKL and the elucidation ofthe RANK/RANKL/OPG pathway have led to a better understanding of the process of bone remodeling in the inflammatory arthritides."
RANKL RANKL is a 317 amino acid peptide that belongs to the tumor necrosis factor (TNF) superfamily. It exits in soluble or membrane bound forms. RANKL plays a crucial role in osteoclast activation and differentiation as well as in the inhibition ofosteoclast apoptosis. Duringphysiologic remodeling, the source ofRANKL is the osteoblasts and their precursors. RANKL is expressed at high levels in the lymph nodes, thymus and lung and at lower levels in spleen and bone marrow. In inflammatory arthritis, the T-lymphocytes in inflamed synovial tissue express RANKL which modulates dendritic cell function to directly activate osteoclasts.P Many proinflammatory cytokines, including TNF, IL-l, IL-6 and IL-17 and hormonal factors, including vitamin D3, PGE-2 and PTH, are important inducers ofRANKL expression. P In animal models RANKL knock-out mice develop severe osteopetrosis with a total loss of osteoclasts and RANKL -/ - mice develop arthritis but do not exhibit bone erosion.34•35The identification of mRNA for RANKL in fibroblasts and activated CD4+ and CD8+ lymphocytes derived from RA synoviurn , provides compelling evidence for the role ofRANKL in RA .36 •37 RANKL is similarly thought to be the mediator of bone loss in patients with ankylosing spondylitis, psoriatic arthritis, animal models of septic arthritis and periodontal Infection."
RANK RANK, a natural receptor of RANKL, is a transmembrane protein of 616 amino acids that also belongs to the TNF receptor superfamily. It is expressed primarily by osteoclast precursors, B and T -cells, dendritic cells and fibroblasts. 39 •40 RANK protein expression has also been reported in mammary gland epithelial cells and in multiple epithelial cell lines. 41 .42 The binding ofRANKL to RANK, stimulates differentiation ofosteoclastic precursors into mature osteoclasts and activates mature osteoclasts, emphasizing the pivotal role of RANK in the control of osteoclastogenesis. The important role of RANK in osteoclast formation was confirmed by experiments which demonstrated that mice deficient for RANK develop a severe osteopetrosis characterized by the absence ofosteoclasts. 43.44 OPG A major advance in the field of bone biology was the discovery of OPG in 1997.45 Osroeprotegerin ("protector ofthe bone"), also called osreodastogenesis inhibitory factor is a 40 1 amino acid peptide that also belongs to the TNF receptor superfamily. It is expressed by stromal cells and osreoblasts. OPG mRNA has also been isolated from other tissues including lung, heart, kidney, liver, stomach, intestine, thyroid gland , brain and spinal cord. OPG is a soluble decoy receptor for RANKL that therefore acts as natural inhibitor of RANKL, as it competes with RANK to bind with RANKL. Furthermore, it binds with high affinity to RANKL, thereby down regulating signaling through the RANKL-RANK interaction. A decrease in biological effects ofRANKL in turn results in a decrease in osteoclast differentiation, suppression of osteoclast activation and induction of osteoclast apoptosis. Treatment of a rat model with arthritis, using OPG, demonstrated a complete block in development ofmarginal erosions. F Additionally, high systemic levelsof 0 PG in 0 PG transgenic mice are known to cause osteopetrosis. 46ln contrast, 0 PG knockout mice are osteoporotic as a result ofincreased osteoclast differentiation and function ."
Other Important Cytokines Involved in Bone Remodeling The transcriptional program that controls osteolcastogenesis is regulated by several cytokines. A list ofcytokines is shown in Table 1. To understand the pathophysiology ofbone loss in inflammatory arthritis, it is important to understand the role ofsome ofthese cytokines in this intricate
90
M olecular M echan isms ofSpondyloarthropathies
mechanism. Most of the current knowledge comes from stud ies in rheumatoid arthritis with dat a emerging in recent years to suggest a similar concept in th e spondyloarthropathies. T N F-a is one of the most potent osteoclastogeneic cyrokine produced by inflammato ry tissue. Increased TNF-a levels result in activation and recruitment of leukocytes and production of pro inflammatory cyto kines, including IL-I, IL-6 and IL-8, in rhe umato id arthr itis. TNF-a also stimulates osreoblasts and T-lymphocytes to produce RANKL and M-CSF. Furthermore, TNF-a , in the presence of M-CSF, independ ent of RANKLIRANK, may directly induc e the formation of osteoclasts." Elevated levels of T N F-a are central to the path ogenesis of the spon dyloarthropathies." TNF has been demonstrated in the synovial tissue. entheses and in sacroiliac joints of pati ents with spondyloarthritis. The pivotal role ofTNF-a in the pathogenesis ofspondyloarthroparhies has been highlighted further by the efficacy ofanti-TNF agents in the treatment ofthe se diseases. Clinical stu dies with anti-T N f agents have demonstrated a significant decrease in the progression of joint erosions in patients with PsA.50 Interestingly, anti-TNF agents have also been shown to improve bone mineral den sity in patients with AS and inflammatory bowel disease.51.52 IL-I is a potent stimulus for bone resorption. IL-I blo ckade prevents bone and cartilage destruction in animal models with arthritis. 53•54 IL-I has been localized in synovial tissue and cartilage-pannus junction ofpatients with rheumatoid arthritis." The role for IL-I in spondyloarthritis is less clear. Open label studies using IL-I blockade indicated some benefit in patients with ankylosing spondylitis56.S7 but another study using anakinra (biologic agent that block IL-I) in pati ents with psoriatic arthritis revealed little or no benefit." IL -6 support s osteoclast differentiation and can stimulate osteoclast activity. Osteoblastic cells and stromal cells are the main source ofIL-6. Patients with active AS and bone loss have been shown to have higher levels ofIL-6 than patients with inactive AS.58M-CSF is essential in the regulation ofosteocl astogenesis and plays an important part in TNF-a induced osteolysis. Mice that lack M-CSF develop osteoclast-deficient osreopctrosis.l? Indeed, M-CSF inhibition can completely arrest oste ocl ast formation. A recent study indicated that M-CSF levels are markedly increased in patients with AS compared with healthy controls/" Parathyroid hormone, par athyroid hormone-related peptide, 1,25 dih ydroxyvitamin D3 , pro staglandin E2, Inrerlu ekin -l l (IL- I I) , onco statin M, estrogen and glucocrti coid also increase
Table 1. Cytokines that regulate osteoclast function and differentiation Cytokines that stimulate osteoclast differentiation and activity Interleukin-l (IL-l) Interl eukin (IL-3) Interl euk in (IL-6) Interl eukins (IL-ll ) Interl eukin (IL-17) Tumor necro sis factor (TNF) Gr anul ocyt e macrophage co lony sti mulating factor (GM -CSF) Leukem ia inhibitory factor (L1 F) O ncostatin M (O SM) Ciliary neutropic factor (CNTF) c-ki t ligand
Cytokines that Inhibit osteoclast differentiation and activity Interl eu kin (IL-4) Interleuki n (IL-10) Interl euk in-18 (IL-18) Inter feron -y
Bone Loss in the Spondy/oarthropathies
91
bone loss. These factors and hormones increase osteoclastic bone resorption indirectly via their effects on osteoblasts or bone lining cells which then release products that act on osteoclast precursors." Chemokines such as macrophage inflammatory peptide 1, IL-8 and IL-I6 promote lymphoproliferation'S which in turn express RANKL supporting osteoclastogenesis. Similarly, T-cell derived cytokines such as IL -I7 and IL-IS are also thought to promote osteoclast activity,63.64 in patients with rheumatoid arthritis. The role ofthese cytokines and hormones is less clear in the spondyloarthropathies.
Bone Loss in the Spondyloarthritis Localized bone loss in the form of erosions is a common clinical finding in the axial and/or peripheral skeleton of patients with spondyloarthritis. Osteoporosis is also a well recognized complication especially of AS and inflammatory bowel diseases. While the past decade has seen major advances in the field ofosteoimrnunology, the mechanisms ofbone loss in the spondyloarrhropaehies is still not fully understood. Increased osteoclastic activity has been reported in PsA and AS. Increasing evidence also supports the hypothesis that local and/or systemic inflammatory cytokines are major contributors to bone loss in the spondyloarthropathies. A negative relation between serum osteocalcin and serum inflammatory markers and a positive relation between bone resorption markers and serum inflammatory markers has also been described." These reports have helped highlight the important role of osteoclasts and the RANKL/OPG system in the pathogenesis of bone loss in the spondyloarthropathies.
Bone Loss in Ankylosing Spondylitis While radiographic damage is a frequent finding in AS, bone formation rather than bone erosions, are the more prominent findings. Erosions however are described at the sacroiliac joint and in the vertebral bodies. They are less commonly seen in peripheral joints but when present at the hips they signify a poor prognosis.f" The pathogenesis of bone erosions in AS is yet to be fully elucidated. Few studies have addressed this issue, essentially due to the difficulty in accessing tissue from the sacroiliac joints. In recent years some studies have also identified a role for several proinflammatory cytokines in the pathophysiology ofbone resorption in AS but to date the role of the RANKL/RANK/OPG pathway is not known. Evidence supporting a role for osteoclasts in bone erosions , in AS, comes from studies that have ident ified these cells in areas close to bone resorption. Biopsiesofthe femoral head from AS patients have demonstrated the presence ofan increased number ofosteoclasts, 67.68 Other studies detected proinflammatory cyrokines involved in osreoclastogenesis, including TNF -a, IL-6 and IL-I ~ in the sacroiliac joint.67.69 Immunohistological studies ofthe zygapophyseal joints revealed no difference in the number ofosteoclasts in AS patients when compared with that of healthy controls,"? Furthermore, no difference in osteoclast numbers was noted between patients with persistent inflammation and patients with less active AS. Interestingly, the same study detected significantly elevated number of CD4+ and CD8+ T-cells and an increase in the number of microvessels in AS patients, even in those with ankylosis ofmost areas ofthe spine, when compared with healthy controls. The reason for the lack ofdifference in osteoclast numbers between the patients with AS and controls is not clear it should be noted that the number ofsubjects in this study was small and the subjects with ankylosing spondylitis were at a late stage ofthe disease process. The detection of increased number ofT-cells in this study however supports evidence for persistent inflammation in the bone marrow of the zygapophyseal joints. Erosions in the peripheral joints in AS are less common. Biopsies ofthe femoral head from AS patients have also reported the presence of an increased number of osteoclasts. These cells were detected in areas ofbone resorption close to the osteochondral junction." The osteoclast numbers were higher in AS patients when compared with patients who had osteoarthritis and rheumatoid arthritis. Another interesting finding was that the number of osteoclastic foci was significantly higher in AS patients, in areas with cartilage on the surface compared with areas without surface cartilage. Furthermore, the osteoclasts were detected close to T-cell aggregates. Taken together
92
Molecular Mechanisms ofSpondyloarthropathies
these findings support the hypothesis that osteoclasts play an important role in the mechanism of bone loss and that they in turn are probably stimulated by inflammatory cells and cytokines secreted by T-cells. Synovial biopsies ofknee joints ofAS patients have demonstrated large quantities of CD 68+ macrophages in the lining and sublining layers of the synoviurn." In addition a large number ofT-cells, B-cells and plasma cells were also seen suggesting a role for these cells in the pathogenesis ofperipheral joint damage in AS . While the above studies suggest a role for osteoclasts and inflammatory cytokines in the pathogenesis ofbone erosions in AS, to date , the data are limited to few studies. The contribution ofthe RANKL/RANK/OPG pathway, in bone loss in AS is not fully understood. Indirect evidence supporting the role of inflammatory cyrokines, especially TNF-a, in bone erosions in patients with AS, is provided by trials that have demonstrated a decrease in radiographic progression in AS patients treated with anti-TNF agents. 73•74 On the other hand, the Outcome in Ankylosing Spondylitis International Study (OASIS) cohort suggests that there in no relationship between radiographic progression and clinical disease activity parameters" suggesting that there may be no direct association between inflammation and bone remodeling in AS. The recent identification of the role ofwingless proteins (Wnt) and the Dickkopf-I (D KK-I) protein in the spondyloarthropathies have helped make some inroads into understanding of bone remodeling in these diseases. Increased activation of the Wnt proteins in AS is known to stimulate bone formation but it also activates OPG which then is thought suppress RANKL-induced osteoclast formation. The current hypothesis therefore is that TNF induced inflammation results in an initial bone resorptive phase but then is followed by profound bone formation.
Generalized Bone Loss in AS Incidence ofdecreased bone mineral density (BMD) in AS is estimated at 41-62% in spine and 46-86% in the femur while the incidence ofnon-traumatic fractures is estimated to be 10-40%76•77 Bone loss in AS occurs early in the disease process and the degree of bone loss appears to be related to disease activity.78.79 While osteoporosis in AS is a well known phenomenon the cause and mechanisms of bone loss are not fully identified yet. Recent studies suggest that a relationship with inflammation and more specifically the balance of the RANKL and OPG levels appears to be an important component ofdiffuse bone loss in AS. An imbalance and a disturbed interaction ofRANKL and OPG may be one ofthe causes for reduced bone mineral density in patients with AS. Serum levels of soluble RANKL (sRANKL) in AS patients are significantly higher, as in the case with patients with rheumatoid arthritis, compared with patients with osteoarthritis and normal controls. 8o•81 In contrast, there is no consensus to the serum level ofserum OPG levels in AS patients. One study suggested that the serum levels ofOPG in AS were not different from those in normal conrrols.f Another stu dy of264 patients with AS reported low serum levels of OPG as well a lack of compensatory age-related increase of serum OPG levels.P Furthermore, the low OPG levels were associated with increased bone loss. These results are in contrast with findings of increased serum OPG levels in patients with AS in two other small studies.84•85 Despite the conflicting findings, it is now thought that the altered interaction ofRANKL and OPG rather than serum levelsofthese cytokines are the cause for bone loss in AS. Indeed, in one study the sRANKL/OPG ratio was noted be higher in AS patients than in healthy controls and was also related to changes in the bone mineral densiry." Interestingly, the sRANKL: OPG ratio was noted to be higher in patients with higher radiographic scores based on the New York criteria for sacroilitis. Similarly, a higher sRANKL: OPG ratio was also seen in patients with osteoporosis or osteopenia compared with those who had normal bone density. Other cytokines are also thought to play important role in bone loss. Support for the role of TNF-a in AS. albeit indirect, is provided by studies that have demonstrated an increase in bone density in AS patients treated with anti-TNF agenrs." Serum RANKL levels have also been demonstrated to correlate with serum levelsofTNF-a and IL-17 in patients with AS. IL-17, Like
Bone Loss in the Spondy/oarthropathies
93
TNF-a, is a potent stimulator ofosteoclast activation in bone remodeling and is up-regulated in synovial fluid ofRA parienrs ."
Bone Loss in PsA Erosions are a frequent finding in psoriatic arthritis. It is reported that an estimated 40-57% ofpatients with PsA develop bone erosions.86,87 Furthermore, a recent study suggested that these radiographic changers occur within two years of diagnosis." While a dysregulation in bone remodeling, similar to that in rheumatoid arthritis, was thought to be the cause for the erosions until recently there was little evidence to support this hypothesis. In the past few years however compelling evidence has emerged to implicate osteoclasts and the RANKL/RANK/OPG system in the pathogenesis ofbone erosions in PsA. Ritchlin et al demonstrated the presence ofosteoclasts at the bone pannus junction and subchondral bone areas ofpatients with PSA.89 Parallel studies reported an increase in the frequency ofosteoclast precursors in the peripheral blood ofpatients with PsA but not in healthy controls." Furthermore, the increase in osteoclast precursors was even more striking in patient with erosive disease compared with psoriatic patients without erosions on plain radiographs. Additionally, treatment with anti-TNF therapy resulted in a decrease in the osteoclast precursor frequency and this decline was associated with a decrease in bone marrow edema, lesions considered to be predictive ofbone erosions, on conventional and three-dimensional MR!. These studies provide ample suppor t for the critical role ofosteoclasts in the pathogenesis ofbone erosions in PsA. The study by Ritchlin et al also highlighted the role for the RANKL/RANK system in bone erosions. Immunohistochemical studies of the synovium ofPsA patients revealed the expression of RANK in synovial tissue adjacent to the bone pannus junction." In contrast, no RANKL was detected in the synovial tissue from osteoarthritis patient. OPG expression, on the other hand, was limited to the endothelium and essentially was not seen in the synovium ofPsA patients. Further support for the important role offRANKL is provided by the demonstration that blocking RANKL with 0 PG substantially decreased the number ofosteoclast precursors. The process oflocalized bone loss in PsA is therefore thought to be similar to that ofrheumatoid arthritis. It however has been proposed that mechanism may be bidirectional.89 The elevated levels ofTNF in psoriatic joints increase the frequency ofosteoclast precursors. The osteoclast precursors in the peripheral blood enter the synovium where they are induced to become osteoclasts by the RANKL expressed in the synovium . Meanwhile the osteoclast precursors in the subchondral bone are stimulated to undergo osteoclastogenesis by RANKL secreted by osteoblasts and stromal cells. Recent studies have suggested that contrary to previous beliefs patients with PsA may also have decreased BMD.6.7Another recent study reported that the bone density changes were similar for patients with PsA were similar to that ofrheumatoid patients density, when adjusted for weight." Despite these observations, to date, there are no published studies on the possible mechanisms underl ying diffuse bone loss in psoriatic arthritis.
Bone Loss in Enteropathic Arthritis Diffuse bone loss is a common finding in patients with inflammatory bowel disease. It is estimated that up to 50% ofpatients with Crohn's disease may have osteopenia while about 15% are thought to have osteoporosis.S'Ihe role ofosteoclasts and the RANKL/RANK/OPG system in the process ofbone loss in patients with inflammatory bowel diseaseis not clearlyunderstood. A strong relationship is known to exist between gut and joint inflammation in the spondyloarthropathies." Recent evidence also suggests that inflammation related factors rather thanmalabsorption, as previously reponed in the literature, may be the major contributor in the genesis ofosteoporosis. Initial evidence for the role ofinflammatory cytokines and specifically the RANKL/RANK/ o PG system, in generalized bone loss in inflammatory bowel disease came from animal studies. In a murine colitis model, dramatic bone losswas noted to occur with the onset ofcolitis but the bone loss was recovered during periods ofdisease remission secondary to increased bone formation."
94
Molecular Mechanisms ofSpondyloartbropatbies
This study supports a link between bone formation, bowel inflammation and overproduction of inflammatory cytokines. Another study demonstrated that bone loss in the IL-2 deficient colitis mouse model was mediated by activated T-cells secreting RANKL,95 Even more interesting was the finding that exogenous 0 PG administration to these mice, resulted in reversal ofbone loss and amelioration of intestinal inflammation. These findings support the hypothesis that the inflammatory cytokines, acting via the RANKL/RANK/OPG pathway, are the major contributors to bone loss in inflammatory bowel disease. A similar concept implicating the role ofthe inflammatory cytokines and the RAN KL/RANK/ OPG system has been proposed in humans. Elevated mucosal and serum levels ofIL-la, TNF-a and IL-6, cytokines that can stimulate osteoclastic activity as discussed earlier, are recognized in Crohns disease and ulcerative colitis. 96•97 Initial, support for the hypothesis that these cytokines are involved in bone loss in inflammatory bowel disease is provided by reports that demonstrated a marked improvement of bone turnover markers and bone mineral density in patients treated with anti-TNF agents. 98.99 As with rheumatoid arthritis, these inflammatory cytokines in turn are thought to act via the RANL/RANK/OPG pathway to cause bone loss in inflammatory bowel disease . Evidence for this is provided by a recent study by Moschen er al loo They demonstrated that OPG levels were increased in patients with Crohn's disease and ulcerative colitis when compared with healthy controls. They were also able to demonstrate that the intestine was the source ofthe OPG. Also noted was a significant negative correlation between OPG and BMD at the femoral neck and lumbar spine. In contrast, soluble RANKL levels were not significantly different in patients with inflammatory bowel disease compared with healthy controls. Furthermore, they showed that the OPG/sRANKL ratio was significantly higher in patients with Crohns disease and ulcerative colitis, with Crohns patient having the highest ratios. These findings are the reverse of what is seen in rheumatoid and PsA.lt however has been speculated that the elevated OPG levels may represent a homeostatic response, to try and reverse the bone loss. The fact that a gradual increase in OPG levels in patients with normal bone density, osteopenia and osteoporosis was detected, may support this hypothesis. More studies are however required to confirm this proposal. Although, ofien not considered in the context ofspondyloarthropathies, celiac disease is also an inflammatory bowel disease. Osteoporosis is also a well documented finding in patients with celiac disease.'?' Again the underlying pathogenetic mechanisms have not yet been fully elucidated. There however, is increasing evidence to support a role for inflammatory cytokines, osteoclasts and the RANKL/RANK/OPG pathways. Proinflammarory cytokines, including TNF-a, IL-ll3 and IL-6, have been demonstrated in the intestinal mucosa of celiac parients.l'" A significant inverse correlation between BMD z-scores and serum IL-6levels in patients with celiac disease has also been reported.!" Support for the role ofthe RANKL/RANK/OPG pathways comes from a recent study by Taranta er al104 This study compared the levels ofvarious cytokines involved in bone remodeling between celiac patients on a gluten free diet (n = 25), subjects on a gluten free diet (n = 17) and healthy controls (n = 21) . RANKL and OPG levels were increased in all patients with celiac disease but the RANKL/OPG ratio was significantly higher (2.5-fold) in patients not yet on a gluten free diet while no difference was noted in the ratio between patients on a gluten free diet and controls. Levels of IL-6 were also significantly increased in patients not on gluten free diet compared with patients on a gluten frec diet. Additionally, urinary N-terminal telopeptide (NTx) ofprocollagen Type I (NTx), a marker ofbone resorption, was increased in patients not on a gluten free diet while it was within normal limits in patients on a gluten free diet. A significant correlation was noted between the NTx levels and IL-6 and the RANKL/OPG ratio . Also a low BMD Z-score was negatively correlated with serum IL-6levels and the RANKL/OPG ratio. These studies implicate a role for the RANKL/RANK/OPG pathway in bone loss in celiac disease. This study by Taranra et al also suggests a role for osteoclasts in the pathogenesis of bone loss in celiac disease. An incre ase in number ofosreoclasts was demonstrated by adding sera ofpatients with celiac disease not on a gluten free diet to peripheral blood mononuclear cells from healthy donors.'?" Lesser increases were noted when sera from celiac patients on a gluten free diet were added. Interestingly, although increased pit indice s, a marker ofbone resorption activity by mature
Bone Loss in the Spondyloartbropathies
95
osteoclasts , was observed in the presence ofsera from celiacpatient on the gluten free diet and not on the gluten free diet, no statistically significant difference wasdetected between the two groups. As the authors conclude this might suggest that the proinflammatory cytokines are likely to affect osteoclastogenesis more than bone resorption activity ofmature osteoclasts. Until recently diffuse bone loss in inflammatory bowel diseasewas thought to be a direct consequence ofmalabsorption. Recent findings discussed above however suggest that inflammation is most likely a major player in the parhophysiology ofbone loss in inflammatory bowel diseases. An intriguing interplay ofinflammatory cytokines, especiallyTNF-a. and IL-6, osteoclasts and the balance between RANKL and OPG is thought to be the mechanism for bone loss in the inflammatory bowel diseases. No studies however have been reported in those patients with arthritis in addition to the inflammatory bowel disease. Future challenges would include studies to determine whether an even more marked bone loss occurs in patients with enteropathic arthritis and to detect if there is a correlation between generalized and localized bone loss in these patients Peripheral arthritis occurs in 10-15% ofpatients with ulcerative colitis and 15-20% ofpatients with Crohns disease while spondylitis is seen in 5% and 10% respectively. Although, erosions have been described in the sacroiliac joints of these patients to date there is a paucity of data in this field. Not surprisingly, little is therefore known about the mechanism ofjoint inflammation in enteroparhic arthritis. Indirect evidence for the role of inflammatory cytokines comes from recent studies that detected increased numbers ofCD 163+ macrophages, a member ofthe group B scavengerreceptor cysteine-rich superfamily, in synovium and in colonic mucosa ofpatients with enteropathic arthritis. 105These macrophages are known to produce TN Fa. and may help shed some light on potential mechanisms for synovial inflammation.
Bone Loss in Reactive Arthritis Verylittle is known about bone loss, both localized and generalized, in reactive arthritis. The role of TNF-a. in reactive arthritis has been established in animal models and in humans .106•107 Reports of efficacy with anti-TNF agents in patients with reactive arthritis give credence to the role played by this cytokine in the pathogenesis ofreactive arthritis. While recent trials have also shown an increased serum levels of crosslinked telopeptide of collagen -I (CTX-I) and urinary deoxypyridinoline, markers of bone resorption in patients with reactive arthritis." there are no studies to date that have attempted to explore a role for osteoclastsand the RANKL/OPG pathway as a cause for bone loss in reactive arthritis.
Conclusions Bone loss in the inflammatory arthritis is a result of activation of hormones and cytokines which then increases the number and activity of the osteoclasts. Osteoclasts are multinucleated cells that are responsible for bone resorption. Osteoclast differentiation and function is regulated by severalcytokines that act by controlling the balance between RANKL and OPG. Delineating the pathway that eventually results in bone resorption by osteoclasts has provided novel therapeutic targets for the prevention and treatment of bone loss in the inflammatory arthritis. The inhibition ofRANKL seems to be a rational strategy to treat both localized and generalized bone loss in the inflammatory arthritides. Indeed, this concept is supported by preliminary results from clinical trials using a monoclonal antibody against RANKL. While great strides have been made in understanding the pathogenesis of bone loss in rheumatoid arthritis, it is only in recent years that studies have attempted to identify the cause for bone loss in the spondyloarthropathies. A better comprehension of the pathogenesis of bone loss in the spondyloarthropathies could help prevent progression of joint damage and osteoporosis, two common causes of morbidity in the spondyloarthropathies.
96
Molecular Mechanisms ofSpondyloarthropathies
References 1. Sinigaglia L, Nervetti A, Mela Q et aI. A multicenter cross sectional study on bone mineral density in rheumatoid arthritis . Italian study group on bone mass in rheumatoid arthritis. ] Rheumatol 2000 ; 27:2583-89. 2. Weichselbaum A. The finer changes of cartilage with fungus synovitis and caries of the joint ends (German). Archiv Pathol Anat Physiol Klin Med 1878; 73 :461-75. 3. van der Linden S, Valkenburg HA, Cats A et al. Evaluation of diagnostic criteria for ankylosing spon dylitis . A proposal for modification of the New York criteria. Arthritis Rheum 1984; 27:361-68. 4. Sinigaglia L, Varenna M, Girasole G et al. Epidemiology of osteoporosis in rheumatic diseases. Rheum Dis Clin N Am 2006; 32:631-58. 5. Bernstein CN, Blanchard]F, Leslie Wet aI. The incidence of fracture among patients with inflammatory bowel d isease. A population based cohort study, Ann Intern Med 2000; 133:795-99. 6. Harrison B], Hutchinson CE, Adams] er al. Assessing periarticular bone mineral density in patients with early psoriatic arthritis or rheumatoid arthritis. Ann Rheum Dis 2002; 61:1007-11. 7. Frediani B, Allegri A, Falsetti P et aI. Bone mineral density in patients with psoriatic arthritis.] Rheumarol 2001 ; 28 :138-43. 8. Massey HM, Flanagan AM. Human osteoclasrs derived form CD 14-positive monocytes . Br] Haematol 1999; 106 :167-70. 9. Teitelbaum SL. Bone resorption by osteoclasts. Science 2000:1504-1508. 10. Holliday LS, Welgus HG, Fliszar C] et al. Initiation of osteoclast bone resorption by interstitial collagenase.] Bioi Chern 1997; 272:22053-58 . 11. Teitelbaum SL. Osteoclasts : What to they do and how do they do it? Am ] Parhol 2007; 170:427-435. 12. Vaananen HK, Horton M. The osteoclast clear zone is a specialized cell-extracellular matrix adhe sion structure.] Cell Sci 1995; 108:2729-32 . 13. Schett G, Stolina M, Bolon B et al. Analysis of the kinetics of osteoclastogenesis in arthritic rats. Arthritis Rheum 2005 ; 52 :3192-201. 14. Suzuki Y, Nishikaku F, Nakatuka M et al. Osteoclast-like cells in murine collagen induced arthritis. ] Rheumatol 1998; 25:1154-60. 15. Pettit AR, ]i H , von Stechow D er al. TRANCE/RANKL knockout mice are protected from bone erosion in a seru m transfer model of arthritis. Am] Parhol 2001 ; 159:1689-99. 16. Redlich K, Hayer S, Ricci Ret al. Osteoclascs are essential for TNF-medicated joint destruction.] Clin Invest 2002; 110:1419-27 . 17. Sims NA , Green ]R, Glatt M et al. Targeting osteoclasts with zoledronic acid presents bone destruction in collagen induced arthritis . Arthritis Rheum 2004; 50:2338-46. 18. Herrak P, Gortz B, Hayer S et al. Zoledronic acid protects against local and systemic bone loss in tumor necrosis factor-mediated arthritis. Arthritis Rheum 2004; 50:2327-37. 19. Bromley M, Woolley DE . Chondroclasts and osteoclases at subchondral sites of erosions in the rheumatoid joint. Arthritis Rheum 1984; 27:968-75. 20. Bromley M, Woolley DE . Histopathology of the rheumatoid lesion; identification of the cell types at sites of cartilage erosion. Arthritis Rheum 1984; 27:857-63. 21. Gravallese EM, Harada Y, Wang]T et al. Identification of cell types responsible for bone resorption in rheumatoid arthritis and juvenile rheumatoid arthritis. Am ] Pathol 1999; 152:949-51. 22. Gravellese EM, Manning C, Tsay A er aI. Synovial tissue in rheumatoid arthritis is a source of osteoclast differentiation factor. Arthritis Rheum 2000; 43 :250-56 . 23. Fujikawa Y, Shingu M, Torisu T er al. Bone resorption by tartrate-resistant acid phosphatase-positive multinuclear cells isolated from rheumatoid synovium . Br] Rheurnatol 1996; 35:213-17. 24. Suzuki Y, Tsutsum i Y, Nakagawa M et al. Osteoclast-like cells in an in-vitro model of bone destruction by rheumatoid synovium. Rheumatology (Oxford) 2001 ; 40 :673-82. 25. Udagawa N, Takahashi N, Akatsu T et al. Origin of osreo clasrs: mature monocytes and macrophages are capable of differentiating into osteoclasrs under a suitable microenvironment prepared by bone marrow-derived stromal cells. Proc Natl Acad Sci USA 1990 ; 87:7260-64. 26. Fujikawa Y, Sabokbar A, Neale S er al. Human osteoclast formation and bone resorption by monocytes and synovial macrophages in rheumatoid arthritis. Ann Rheum Dis 1996; 55:816-22. 27. Manolagas SC, ]ilka RL. Bone marrow, cytokines and bone remodeling. N Eng] Med 1995 ; 332:305-11. 28. Hofbauer LC , Khosla S, Dunstan CR et aI. The roles of osteoprotogerin and osteoprotegerin ligand in the paracrine regulation of bone resorption. ] Bone Miner Res 2000 ; 15:2-12. 29. Crotti TN, Smith MD, Weedon H et al. Receptor activator NF-kB ligand (RANKL) expression in synovial tissue from patients with rheumatoid arthritis, spondyloarthropathy, osteoarthritis and from normal patients: semiquantitative and quantitiarive analysis. Ann Rheum Dis 2002 ; 61:1047-54.
Bone Loss in the Spondyloartbropathies
97
30. Lacey DL. Timms E. Tan H -L er al. Osteoprotegerin (OPG) ligand is a cytokine that regulates osteoclast differentiation and activation. Cell 1998; 93;165-76. 31. Anandarajah AP. Schwarz EM. Anti-RANKL therapy for inAammatory bone disorders: Mechanisms and potential clinical applications.] Cell Biochem 2006; 97:226-32. 32. Kong IT. Feige U. Sarosi I et al. Activated T-cells regulate bone loss and joint destruction in adjuvant arthritis through osteoproregerin ligand. Nature 1999; 402 :304-09. 33. Lam ]. Takeshita S. Barker ]E er al. TNF-alpha induces osteoclastogenesis by direct stimulation of macrophages exposed to permissive levels of RANK ligand.] Clin Invest 2000 ; 106:1481-88. 34. Pettit AR. ]i H. von Stechow 0 et al. TRANCE/RANKL knockout mice are protected from bone erosion in a serum transfer model of arthritis. Am] Pathol 2001 ; 159:1689-99. 35. Li P. Schwarz EM. O'Keefe et al. RANK signaling is not required for TNF-mediated increase in CD 11 (hi) osteoclast precursors but is essential for mature osteoclast formation in TNF alpha mediated inAammatory arthritis.] Bone Miner Res 2004 ; 19:207-13. 36. Gravallese EM. Manning C . Tsay A ec al. Synovial tissue in rheumato id arthritis is a source of osteoclast d ifferentiation factor. Arthritis Rheum 2000; 43 :250-58. 37. Takayanagi H. Lizuka H . ]uji T et al. Involvement of receptor activator of nuclear factor leB ligand/ osteoclast differentiation factor in osteoclastogenesis from synoviocytes in rheumatoid arthritis. Arthritis Rheum 2000; 43 :259-69 . 38. Hofbauer LC . Schopper M. Clinical implications of the osteoprotegerin/RANKL/RANK system for bone and vascular diseases. ]AMA 2004 ; 292:490-95. 39. Anderson MS. Maraskovsky E. Billingsley WL et al. A homologue of the TNF receptor and its ligand enh ances T-cell growth and dendritic cell function . Nature 1997; 390:175-79. 40. Hsu H. Lacey DL. Dunstan C R ct al. Tumor necrosis factor receptor family member RANK mediates osteoblast differentiation and activation by osteoproregerin ligand. Proc Nad Acad Sci USA 1989 ; 96:3540-45. 4 1. Fata ]E, Kong IT, Li] et al. The osteoclast differentiation factor osteoprotegerin-ligand is essential for mamm ary gland development. C ell 2000 ; 103:41-50. 42. Ch en G. Sircar K. Aprikian A et al. Expression of RA NK /RANKL/OP G in pr imar y and metastat ic human prostate cancer as markers of disease stage and functional regulation. Cancer 200 6; 107:289-98. 43. Dougall WC, Glaccum M . Charrier K et al. RANK is essential for osteocla st and lymph node develop ment. Genes Dev 1999; 13:2412 -24. 44. Kapur RP. Yao Z. Lida MH er al Malignant autosomal recessive osteopetrosis caused by spontaneous mut at ion of murine Rank.] Bone Miner Res 2004 ; 19:1689 -97. 45. Simonet WA. Lacey DL. Dunstan CR et al. Osteoprotegerin: A novel secreted protein involved in the regulati on of bone density. Cell 1997; 89:309 -19. 46. Kapur RP. Yao Z , Lida MH ct al Malignant autosomal recessive osteopetrosis caused by spo nt aneous mutation of mur ine Rank.] Bone Miner Res 2004; 19:1689-97. 47. Mizun o A. Kanno T, Hoshi M et al. Transgenic mice over expressing soluble osteoclast d ifferentiation facto r (sO D F) exhibit severe osteopo rosis. ] Bone Miner Metab 2002 ; 20:337-44 . 48. Azum a Y, Kaji K, Kacogi R et al. Tumor necrosis factor alpha stimulates osteocla st differentiation of and bone resorp tion by osreoclascs.} BioI C hem 2000 ; 275:4858-64. 49. Fitzgerald 0 , Mcinnes 1. Spondyl oarrhroparhy: disease at the crossroads of immunity. Best Pract Res C lin Rheurnatol 2006 ; 20:949 -67. SO. Anandarajah AP. Ritchlin CT. Treatment update on spondyloarrhro parhy, Curr Opin Rheumarol 2005 ; 17:247-56. 51. Demis E. Roux C. Breban M er al. InAiximab in Spondyloarthropathy-inAuence on bone density. Clin Exp Rheumatol 2002 ; 20:5185-86. 52. Allali F, Breban M. Porcher Ret al. Increase in bone mineral density of patients with spondyloarthropathy treated with ami-tumour necrosis factor alpha. Ann Rheum Dis 2003 ; 62 :347-49. 53. Joosten LAB. Helsen MMA. van de Loo FA] et al. Anticyokine treatment of established type II collagen induced arthritis in DBAll mice: a comparative stud y using anti -TNF, anti-IL-l alpha /beta and IL-1R a. Arthritis Rheum 1996; 39:797-809. 54. Joosten LAB, Helsen MMA. Saxne T et al. IL-1 alpha blockade prevents cartilage and bone destruction in murine type II collagen-induced arthritis. whereas TNF alpha blockade only ameliorates joint inAammation.] Immunol 1999; 163:5049 -55. 55. Delauren BW; Chu CQ. Filed M et al. Localization of interleukin-Io type 1 imerleukin 1 receptor and imerleukin 1 receptor antagonist in the synovial membrane and cartilage/pannus junction in rheumatoid arthritis. Br] Rheumarol 1992; 31:80 1-9. 56. Tan AL. Marzo -Ortega H , O 'Connor P et al. Efficacy of anakima in active ankylosing spondylitis: a clini cal and magnet ic resonan ce imaging study. Ann Rheum Dis 2004; 63 :1041 -45.
98
Molecular Mechanisms ofSpondyloarthropathies
57. Haibel H. Rudwaleit M. Listing J et al. Open label trial of anakinca in active ankylosing spondylitis over 24 weeks. Ann Rheum Dis 2005 : 64:296-98. 58. Gratacos J. Collado A. Pons F et al. Significant loss of bone mass in patients with earlyactive ankylosing spondylitis. Arthritis Rheum 1999; 42 :2319-24. 59. Yoshida H. Hayashi S-I. Kunisada T et al. The murine mutation osteopetrosis is in the coding region of the macrophage colony stimulating factor gene. Nature 1990: 345 :442-44. 60. Yang PT. Kasai H. Xiao WG er al. Increased expression of macrophage colony stimulating factor in ankylosing spondylitis and rheumatoid arthritis. Ann Rheum Dis 2006; 65:1671-72. 61. Manolagas Sc. Birth and death of bone cells: basic regulatory mechanisms and implications for the pathogenesis and treatment of osteoporosis. Endocr Rev 2000 ; 21:115-37 . 62. Chabaud M. Durand JM. Buchs et al. AT-cell-derived proinflammatory cytokine produced by the rheumatoid synovium. Arthritis Rheum 1999; 42 ;963-70. 63. Korake S. Udagawa T. Takahashi N et al. IL-17 in synovial fluids from patients with rheumatoid arthritis is a potent stimu lator of osceoclasrogenesis. J Clin Invest 1999; 103:1345-52. 64. McInnes lB. Leung BP. Sturrock RD er al. Interleukin-15 mediated Tvcell-dependent regularion of tumor necrosis factor alpha production in rheumatoid arthritis. Nat Med 1997; 3:189-95. 65. Allali F. Breban M. Porcher Ret al. Increase in bone mineral density of patients with spondyloarthropathy treated with anti-tumor necrosis factor a. Ann Rheum Dis 2003; 62:347-49. 66. Amor B. Santos RS. Nahal Ret al. Predictive factors for the longterm outcome of spondyloarthroparhies. J Rheumatol1994; 21:1883-87. 67. Francois RJ. Neure L, Sieper J er al. Immunohistological examination of open sacroiliac biopsies of patients with ankylosing spondylitis: detection of tumor necrosis factor a in two patients with early disease and transforming growth factor b in three more advanced cases. Ann Rheum Dis 2006 ; 65:713-20. 68. Marzo-Ortega H. O'Connor P, Emery P ct al. Sacroiliac joint biopsies in early sacroiliris, Rheumatol 2007; 46 :1210-11. 69. Braun J. Bollow M. Konig H er al. Use of immunohistologic and in situ hybridization techniques in the examination of sacroiliac joint biopsy specimens from patients with ankylosing spondylitis. Arthritis Rheum 1995; 38 :499-05. 70. Appel H. Kuhne M. Spiekermann S er al. Immunohistologic analysis of zygapophyseal joints in patients with ankylosing spondylitis. Arthritis Rheum 2006 ; 54:2845-51. 71. Appel H. Kuhne M. Spiekermann S et al. Immunohistochemical analysis of hip arthritis in ankylosing spondylitis. Arthritis Rheum 2006 ; 54:1805-13. 72. Cunnane G. Bresnihan B. Fitzgerald O. Immunohisrologic analysis of peripheral joint disease in ankylosing spondylitis. Arthritis Rheum 1998; 41 :180-82 . 73. Baraliakos X. Listing J. Brandt H et al. Radiographic progression in patients with ankylosing spondylitis after 4 yrs of treatment with the anti-TNF-alpha antibody infliximab. Rheumatology (Oxford) 2007; 46 :1450-3. 74. Baraliakos X. Listing J. Rudwaleir M et al. Radiographic progression in patients with ankylosing spondylicis after 2 years of treatment with the tumour necrosis factor alpha antibody infliximab. Ann Rheum Dis 2005 ; 64 :1462-6 . 75. Schett G. Landewe R. van der Heidje D. Tumor necrosis factor blockers and structural remodeling in ankylosing spondylitis: what is reality and what is fiction. Ann Rheum Dis 2007; 66 :709-11. 76. Mitra D. Elvins OM , Spenden OJ et al. The prevalence of vertebral compression fractures in mild ankylosing spondylitis and their relationship to bone mineral density. Rheumatology (Oxford) 2000 ; 39 :85-9. 77. EI Maghraoui A. Osteoporosis and ankylosing spondylitis. Joint Bone Spine 2004; 71 :291-95 . 78. Will R, Bhalla AK. Palmer Ret al. Osteoporosis in early ankylosing spondylitis: a primary pathological event? Lancet 1989; 11:1483-85. 79. Gratacos J. Collado A, Pons F er al. Significant loss of bone in patients with early active ankylosing spondylitis. Arthritis Rheum 1999; 42 :2319-24. 80. Kim H-R. Kim H-Y, Lee SoH. Elevated levels of soluble receptor activator of nuclear factor-xls ligand (sRANKL) and reduced bone mineral density in patients with ankylosing spondylitis (AS). Rheumatology 2006; 45 :1197-00. 81. Crotti TN, Smith MD. Weedon H er al. Receptor activator NF-kB ligand (RANKL) expression in synovial tissue from patients with rheumatoid arthritis, spondyloarthropathy, osteoarthritis and from normal patients: semiquantitative and quantitative analysis. Ann Rheum Dis 2002; 61 :1047-54. 82. Haynes 0 R, Barg E, Crotti TN et al. Osreoprotegerin expression in synovial tissue from patients with rheumatoid arthritis. spondyloarthropachies and osteoarthritis and normal controls. Rheumatol 2003 ; 42 :123 -34.
Bone Loss in the Spondyloarthropathies
99
83. Franck H , Meurer T, Hofbauer LC. Evaluation of bone mineral densiry, hormones, biochemical markers of bone metabolism and osreoproregerin serum levels in patients with ankylosing spondyliris. J Rheumarol 2004; 31 :2236-41. 84. Grisar J, Bernecker PM, Ar inger M et aI. Ankylosing spondylitis, psoriatic arrhriris and reactive arthritis show in creased bone resorption, but differ with regard to bone formation . J Rheumatol 2002; 29:1430-36 . 85 . Golmia RP, Sousa BD, Scheinberg MA . Increased oseeoproregerin and decreased pyridinoline levels in patients wirh ankylosing spondyliris: comment on the article by Graracos er aI. Arthritis Rheum 2002; 46:3390-91. 86. Gladman DD, Schukett R, Russell ML et aI. QJ Med 1987; 62 :127-41. 87. Torre Alonso JC, Perez RA , Castrillo AJM er aI. Psoriatic arthritis (PA) : a clinical, immunological and radiological study of 180 patients. Br J Rheumatol 1991; 30 :245-50. 88. Kane D , Stafford L, Bresn ihan B er aI. A prospective, clinical and radiological study of early psoriatic arthritis: an early synovitis clinic experience. Rheuamtology (Oxford) 2003; 42 :1460-68. 89. Rirchlin CT, Haas-Smith SA, Li P er aI. Mechanisms of anri-TNF-alpha and RANKL mediated osreoclascogenesis and bone resorption in psoriatic arthritis. J Clin Invest 2003; III :82 1-3 1. 90. Anandarajah AP, Schwarz EM, Totterman S et aI. The effect of eranercepr on osteoclast precursor frequency an d enhancing bone marrow edema in parients with psoriatic arthritis, Ann Rheum Dis (in press) . 91. Redd y SM , Reed G, Anandarajah AP et aI. Comparative analysis of disease activity, radiographic features and bone density in psoriatic and rheumatoid arthritis. Arthritis Rheum 2005; 52:S1726 . 92. Dinca M, Fries W; Luisetto G et aI. Evolut ion of osreopenia in inRammatory bowel disease. Am J Gastroenrerol 1999 ; 94 :1292-97 . 93. De Keyser F, Elewant D, De Vos M er aI. Bowel inRammarion and the spondyloarrhroparhies. Rheum Dis Clin North Am 1998; 24 :785-13. 94. Lin C L, Moniz C. Chambers TJ et aI. Coliris causes bone loss in rat s through suppression of bone formation . Gastroenrerology 1996 ; Ill : 1263-71. 95. Ashcroft AJ, Cruickshank SM, Croucher PI er aI. Colonic dendritic cells, intestinal inRammarion and T-cell mediated bone destruction are modulated by recombinant osteopcrtogerin. Immunity 2003; 19:849-61. 96. Fuss 1]. Cytokine network in inRammatory bowel disease. Curr Drug Targets InRamm Allerhy 2003; 2:101-12. 97. Wesrerholm-Ormio M, Garioch J, Ketola 1 et aI. Inflammatory cytokines in small intestinal mucosa of parients with potential coeli ac disease. Clin Exp Immunol 2002 : 128 :94-01. 98. Ryan BM. Russel MG . Schurgers L er aI. Effect of antitumor necro sis factor-alpha therapy on bone turnover in patients with active Crohn's disease : a pro spect ive study. Aliment Pharmacol Ther 2004; 20 :851 -57 . 99. Paziana s M. Rhim AD. Weinberg AM et aL The effecr of anti -TNF-a therapy on spinal bone mineral density in pat ients with C rohn's disease. Ann NY Acad Sci 2006 ; 1068 :543-46. 100. Moschen AR . Kaser A. Enrich B et aI. The RANKLIOPG system is act ivated in inRammatory bowel disease an d relate s to th e state of bone loss. Gut 2005 : 54:479 -487 . 101. Kemppainen T, Kroger H . janatuinen E et aI. Osteoporosis in adult patients with celiac disease. Bone 2002 ; 24 :249-255 . 102. Koneakou M , Przerniosio RT, Sturgess RP et aI. Expression of tumor necrosi s factor-alpha. interleukin-6 and inrerleukin-Z mRNA in the jejunum of patients with celiac disease. Scand J Gastroenterol 1996; 30 :456-463. 103. Fornan MC, Pedreira S, Niveloni S et aI. Pre- and posttrearment serum levels of cytokines IL-l bera, IL-6 and IL-l receptor antagonist in celiac disease. Are they related to the associated osreopenia? Am J Gasrroenrerol 1998 ; 93 :413-418. 104. Taranta A, Fortunati D. Longo M et aI. Imbalance of osreoclastogenesis-regulanng factors in patients with celiac dise ase. J Bone Miner Res 2004 ; 19:1112-1121. 105. Baeten D . Dernetter P. Cuvelier CA et al. Macrophages expressing the scavenger receptor CD163: a link between immune alterations of the gut and synovial inRammation in spondyloarthropathies, J Parhoi 2002; 196:343-350. 106. Inman RD. Chiu B. Early cytokine profiles in the joint define pathogen clearance and severity of arthritis in Chlamydia-induced arthritis in rats. Arthritis Rheum 2006; 53:499-507. 107. But rimiene 1, Jarmalaite S. Ranceva J et aI. Different cytokine profiles in patients wirh chronic and acute reactive arthritis. Rheurnatol 2004; 43 :1300-1304.
CHAPTER 7
RANKL/RANK as Key Factors for Osteoclast Development and Bone Loss in Arthropathies Andreas Leibbrandt andJosefM. Penninger' Abstract steoporosis or rheumatoid arthritis are bone diseases affecting hundreds of millions of people worldwide and thus pose a tremendous burden to health care. Ground-breaking discoveries made in basic science over the last decade shed light on the molecular mechanisms of bone metabolism and bone turnover. Thereby, it became possible over the past years to devise new and promising strategies for treating such diseases. In particular, three molecules, the receptor activator ofNF-KB (RANK), its ligand RANKL and the decoy receptor ofRANKL, osteoproregerin (0 PG) , have been a major focus ofscientists and pharmaceutical companies alike, since experiments using mice in which these genes have been inactivated unanimously established their pivotal role as central regulators ofosteoclast function. RANK(L) signaling not only activates a variety ofdownstream signaling pathways required for osteoclast development, but crosstalk with other signaling pathways also fine-tunes bone homeostasis both in normal physiology and disease. Consequently, novel drugs specifically targeting RANK-RANKL and their signaling pathways in osteoclasts are expected to revolutionize the treatment ofvarious bone diseases,such as cancer metastases, osteoporosis, or arthropathies.
O
Introduction For all its rigidity, bone is constantly remodeled throughout adult life. Bone remodeling involves resorption by osteoclasts and the synthesis ofnew bone matrix by osteoblasts. Ifanything disturbs this intricate balance between resorption and synthesis of bone, skeletal abnormalities, such as osteoporosis or osteopetrosis, develop and become a severe burden to patients. 1·4 Osteoporosis is a disease characterized by a global decline in bone mineral density and structural deterioration of bone tissue, subsequently leading to bone fragility and an increased susceptibility to fractures especially ofthe hip, spine and wrist. It is estimated that in the US 10 million people already have osteoporosis, while 34 million are predicted to be osteopenic, dramatically increasing their risk for osteoporosis. Osteoporosis manifests predominantly in older people around 50 and older, with a strong gender preference for women (80% of those affected by osteoporosis) and accoun ts for more than 1.5 million fracture s annually. By contrast, osteopetrosis, or abnormally increased bone density, is way less prevalent in our population and occurs mainly as a result ofrare hereditary disorders.' 'Corresponding Author: Josef M . Penninger-In stitute of Molecul ar Biot echn ology of the Austri an Academy of Sciences, Dr. Bohr-Gasse 3, A-l 030 V ienna, A ustria. Email : Josef.Penn in ger@imb a.oeaw.ac.at
Molecular M echanisms ofSpondyloarthropathies, edited by Carlos Lopez-Larrea and Roberto Diaz-Pena. ©2009 Landes Bioscience and Springer Science+Business Media.
RANKL/RANK as Key FactorsfOr Osteoclast Development and Bone Loss in Arthropathies
101
The landmark discoveries of three molecules, the receptor activator of NF-KB (RANK),6its ligand RANKL6-9 and the decoy receptor for RANKL, osteoprotegerin (OPG),IO-12 had a huge impact on how we think about osteoclast function and as a consequence ofosteoporosis and other bone diseases. It needs to be stressed that although various calciotropic hormones and cyrokines, such as PTHrP, Vitamin D3, IL-lb, or TNF-a, have all been shown to affect osteoclastogenesis at distinct stages ofdevelopment, 2only RANK(L) has proven to be absolutely required for osteoclast development in vivo as evidenced by the complete absence ofosteoclasts in RANKL and RANK knockout mice .13- IS Osteoblasrs express RANKL and binding to its cognate receptor RANK on osteoclastic precursors is crucial for the development and activation of mature osteoclasts from hematopoietic progenitor cells.' OPG also binds RANKL, thereby preventing RANKL binding to RANK and thus inhibits RANK signaling and bone turnover by osteoclasts.' Since in patients suffering from bone diseases such as osteoporosis, metastases to bone, or rheumatoid arthritis all show an increased activity ofosteoclasts, it appears that the RANKL-RANK-OPG axis is the most relevant therapeutic target for osteoclast-regulated bone diseases. Here we discuss the importance of the RANKL-RANK-OPG axis in bone metabolism and review various signaling pathways known to be activated by RANK signaling in osreoclasts. This knowledge is of utmost importance for the development ofnovel therapeutic approaches to treat diseases ofthe bone that affect millions of people.
Basic Characteristics ofthe RANKL-RANK-OPG Axis RANKL RANKL (also known as osteoprotegerin ligand OPGL, osteoclast differentiation factor ODF, TNFSFll, TRANCE, CD254) was independently cloned by four groups.v?RANKLis a member ofthe tumor necrosis factor (ligand) superfamily ofcytokines. Structurally, TNF family cytokines are Type II transmembrane proteins, each containing a membrane-anchoring domain, a connecting stalk and a receptor-binding ecrodornain, Human and murine RANKL encode glycoproteins of317 and 316 amino acids, respectively and are highly similar. The crystal structure ofthe extracellular, biologically active domain ofmurine RANKL has recently been solved and revealed that functional RANKL protein self-assembles into stable, noncovalently associated trirners, which is in line with all TNF family cytokines examined to date .": " RANKL is most highly expressed in skeletal and lymphoid tissues that are active in mediating the immune response, but RANKL mRNA expression can also be detected in keratinocytes ofthe skin, mammary gland heart, skeletal muscle, lung, spleen, thymus, stomach, placenta, thyroid gland and brain .6,x,9,IH-21 Importantly, not only is RANKL biologically active in its membrane-bound form (40-45 kDa) , but also in a soluble form (31 kDa) which is derived either from proteolytical cleavageor alternative splicing. Analysis ofRankl mRNA expression revealed two alternative transcripts, encoding a membrane-bound isoform of287 amino acids with a shorter intracellular domain and a soluble version of 199 amino acids lacking both the transmembrane and intracellular domain ofcanonical RANKL.22 Biochemical studies showed direct interactions of alternatively spliced RANKL isoforms and homo- or heterotrimerization ofisoforms. Although all isoforms are functional in in vitro osteoclastogenesis assays, there seems to be an inhibitory effect of the shortest isoform when co-expressed in cells with the other isoforms in that it suppresses the formation of multinucleated osteoclasts by inhibiting the fusion of preosteoclasts.i':" These findings leave open the possibility that regulated expression ofRANKL isoforms might contribute to the control of osteoclast development. Apart from transcriptional regulation, posttranslational modification of membrane-bound RANKL by proteolytic cleavage also seems to occur. Both members ofthe disintegrin and metalloproteases domain (ADAM) family and matrix rnetalloproteases (MMPs) have been shown to contribute to ectodomain shedding of RANKL.2S,26 However, the biological effects ofproteolytic cleavage remain somewhat contradictory. In a murine model ofprostate cancer, it has been shown that increased MMP-7 expression by osteoclasts at the tumor-bone interface can convert membrane-bound RANKL to soluble RANKL,
Figure 1. Legend viewed on following page.
osteoclast
osteoblast
1
~D
•
T
••
l~' - - J
M-CSF
~
~.
1:;-
S.
~ ss:.
l:!
~ ~ <:>
~.
l:!
8s:.
~
I::"" ...
Ii
~
~
...
RANKL/RANK as Key Factorsfor Osteoclast Development and BoneLoss in Arthropathies
103
Figure 1, viewed on prev ious page. Schematic diagram of RANK(L) signaling pathways that control lineage commitment and activation of osteoclasts . RANK stimulation activates various different signaling pathways , such as the MAPK, PI3K and NF-KB pathway, to control osteoclastogenesis. TRAFs and other adaptors, such as Gab2, bound to the cytoplasmic tail are key mediators of RANK(L) signaling. Both the canonical (IKK13) and alternative (NIK-IKKa) NF-KB pathway are activated by RANK stimulation and contribute to transcriptional regulation of target genes. MAPK pathway leads to activation of AP-1 family member wh ich are also crucial for osteoclastogenesis. PI3K signaling links RANK stimulation to activation of Akt/ PKB and Ca2+ signaling via PLCy, which is crucial for calcineurin-mediated NFATcl activation. Besides RANK activation, other costimulatory signals from the osteoblasts are also required for osteoclast differentiation. M-CSF-cFMS signaling is crucial for the proliferation and survival of osteoclastic precursor cells. The ITAM-bearing adaptors DAP12 or FcRy are essential for RANK-mediated Ca2+ induction and osteoclastogenesis via PLCy, calcineurin and NFATcl induction, respectively. There is also positive and negative crosstalk with other signaling pathways and molecules, such as TNF-a, which acts positively on NF-KB, or interferons which negatively influence RANK signaling by promoting accelerated TRAF6 degradation in the case of IFN-y or by interfering with RANKL-induced cFos expression in the case of IFN-13. 4
thereby promoting osteoclast activation and thus osteolysis." Another recent study reported that MMP-14 plays an important role in RANKL shedding both in vitro and in vivo." However, this study suggested that membrane-bound RANKL induces osteoclastogenesis more efficiently than soluble RANKL and MMP-14-mediated ectodomain shedding of RANKL would rather negatively regulate osteoclastogenesis. This observation is corroborated by the osteoporotic phenotype of MMP-14-deficient mice which the authors contributed to increased osteoclast numbers due to the increase in membrane-bound RANKL,28Thus it seems that the biological outcome-be it positive or negative-of RANKL ectodomain shedding by various MMP or ADAM family members on osteoclastogenesis strongly depends on the biological context and certainly needs to be addressed further in the future .
RANK The receptor for RANKL is RANK (receptor activator ofNF-KB, also known as TNFRSF l1A, OFE,ODFR,TRANCE-R,ODAR,CD265),amemberoftheTNFreceptorsuperfamily.Human and mouse RANK eDNA encode Type 1transmembrane glycoproteins of616 and 625 amino acids, respectively. They are comprised ofa 29 and 30 amino acid signal peptide, an extracellular domain of 183 and 184 amino acids, a tran smembrane domain of21 and 20 amino acids and a large cytoplasmic domain of383 and 391 amino acids, respectively. SinceTNF receptors commonly assemble into trimeric complexes on the cell surface prior to ligand binding, as shown for FAS, TNFRl , or TNFR2, it is inferred that RANK trimerization is a prerequis ite for RANKL binding and signal rransmission.V 'F RANK mRNA is expressed with highest levelsin dendritic cells, bone, skeletal muscle, thymus, liver, colon, small intestine and adrenal gland.6.33.34 Moreover, RANK protein can be detected on the surface ofdendritic cells,6.18.34 CD4+ and CD8+ T-cells,35 Langerhans cells (A. Leibbrandt,J. Penn inger, unpublished)" and on mammary epithelial cells where expression is regulated throughout pregnancy, with highest levels at day P15.5 ofpregnancy.20.36
OPG The third molecule in the axis is 0 PG (also called TNFRSF 11B, osteoprotegerin, OCIF, TRl, or FDRCl), which was initially isolated as a secreted TNF receptor family member regulating bone." Human and murine OPG are 85% identical and both are synthesized as 401 amino acid precursor proteins, ofwhich a 21 amino acid signal peptide iscleaved offto give rise to the mature peptide. Since OPG lacks a hydrophobic transmembrane-spanning domain, OPG is secreted as a soluble protein. OPG is synthesized as a monomer (55-62 kDa) and is finally secreted as a homodimeric glycoprotein of-l1O kDa .IO-12,37.38 OPG mRNA can be detected in the brain, liver, lung, heart, kidney, skeletal muscle, skin, intestines, calvaria, stomach, testis and placenta.10-12.37.38
104
Molecular M echanisms ofSpondyloarthropathies
RANKL-RANK-OPG Interactions and Bone Remodeling OPG was the first protein ofthe triad to be identified as a factor that would potentially inhibit osteoclastogenesis.!':" A5 expected for an inhibitory factor ofosreoclastogenesis, transgenic mice overexpressing 0 PG or mice which were treated with recombinant 0 PG both exhibited a marked increase in bone density and osteopetrosis. respectively,' ? Not surprisingly. mice in which OPG has been inactivated by targeted deletion developed early-onset osreoporosis.P'" These studies not only established for the first time the critical requirement for osreoproregerin in the maintenance of postnatal bone mass, but also suggested that OPG might act on another TNF-related factor that would do the opposite. namely stimulating osteoclast development.This postulated positive factor was soon to be found in RANKL by expression cloning: RANKL specifically bound to OPG, enhanced differentiation of bone marrow cells into osteoclasts in vitro and activated mature osteoclasrs to resorb bone both in vitro and in vivo.8•9 In a breakthrough study. the genetic inactivation of RANKL in mice was shown to lead to osteopetrotic mice with defects in tooth eruption as a result from the complete absence of osteoclasts and unanimously proofed the essential function ofRANKL in osreoclasrogenesis.r' The circle was later closed by showing that the receptor for RANKL. RANK. is also essential for osteoclast differentiation and activation induced by RANKL.33.41 Most importantly. RANK-I- mice phenocopy RANKL-I- mice . i.e.. they are osteopetrotic. have a defect in tooth eruption and lack osteoclasrs.P:'! In summary. these find ings unambiguously established the essential role of RANKL-RANK interactions in positively regulating osreoclasrogenesis, which is counteracted and balanced by OPG in vivo by binding to RANKL thereby serving as a decoy receptor for RANKL. Most importantly. the functions ofthe RANKL-RANK-OPG axis in bone remodeling-as established primarily by different mouse models-have also direct relevance to human bone diseases . For example. duplications in the signal peptide of RANK have been linked to four families with Familial expansile osteolysis or Paget disease ofthe bone. rare autosomal dominant bone dysplasia characterized by focal areas of increased bone remodeling. Both insertion mutations (a 18 base pair and a 27 base pair tandem duplication in exon 1 of RANK) result in reduced expression levels and increased constitutive RANK-mediated NF-KB signaling in vitro." Moreover. in patients suffering from expansile skeletal hyperphosphatasia, a familial metabolic bone disease characterized by expanding hyperostotic long bones. early onset deafness . premature tooth loss and episodic hypercalcemia. an insertion in exon 1 ofRANK was identified as the cause ofthe disease.43.44 In patients with an osteoclast-poor form of autosomal recessive osteopetrosis (ARO). various mutations in RANKL have been identified as the cause ofthe disease." Although these ARO patients did not respond to hematopoietic stem cell transplantation. they could form functional osteoclasts from monocyte-lineage cells upon exogenous RANKL application, suggesting that these individuals could profit from a RANKL therapy," Finally. several mutations in OPG. frequently affecting the ligand binding domain. have been ascribed to Juvenile Paget disease . an autosomal recessive osteopathy characterized by rapidly remodeling woven bone. osteopenia, fractures and progressive skeletal deformity.44.46.47
RANK(L) Signaling Pathways Binding of RANKL to RANK results in the activation of signaling cascades that control lineage commitment and activation ofosteoclasts. Given the pleiotropic effects of RANKL and RANK in vivo. it is obviously of key interest to study RANK(L) signaling pathways to identify pathways specific for osteoclast development and to understand their crosstalk to other receptor systems (Fig. 1).
Adaptor Proteins RANK. as member ofthe TNFR family. does not have any kinase motifin its cytoplasmic tail and thus needs associated proteins to transduce Signals. TN FR associated factors. or TRAFs, bind to the cytoplasmic tail ofTNF receptors and are important mediators ofTNFR signaling. v The cytoplasmic domain of RANK contains binding sites for several TRAFs that cluster in di stinct cytoplasmic domains ofRANK: a region from amino acids 235-358 as well as amino acids 359-531
RANKL/RANK as Key Factorsfor Osteoclast Development and Bone Lossin Arthropathies
105
bind to TRAF6, whereas the region spanning amino acids 532-625 contains multiple binding sites for TRAFs 2, 5 and 6.49.51These TRAF binding domains were shown to be functionally important for NF-KB and c-Iun NH2-terminal kinase (JNK) activities in response to RANK stimulation. When the membrane-proximal TRAF6 interaction domain is deleted, RANK-mediated NF-KB signaling is completely inhibited while residual ]NK activation is still possible, suggesting that interactions with TRAFs are necessary for NF-KB activation but not essential for activation of the]NK pathway.49'so,s2.S3 The key importance ofTRAF6 in functional RANK signaling has been substantiated in Traf6 mutant mice that exhibit bone phenotypes similar to Rankl-/- and Rank-" mice due to a partial block in osteoclastogenesis and defective activation ofmature osreoclascs.l"? However, there is some controversy with respect to the impact ofTRAF6 on osteoclastogenesis, since one Traf6-1- strain still has TRAp· osreoclasts," while another independently generated Traf6+ strain is devoid of osteoclasts." Since cell-permeable peptides with the TRAF6-binding motifinhibit TRAF6 signaling and can arrest osteoclastogenesis in vitro," the idea that TRAF6 is indeed essential for osteoclast differentiation is strongly supported despite the controversial phenotypes ofTraf6-1- strains." Compared to TRAF6, the contributions ofTRAF2 and TRAF5 to osteoclastogenesis are minor. For instance, fetal liver derived Traf2-deficient progenitor cells show only slightly (20%) reduced multinuclear osteoclasts and activation ofNF-KB and ]NK by RANKL was cornparable/" TRAF5 deficient cells also show mildly reduced osreoclastogenesis and again NF-KB and]NK activation was not apparently affected upon RANK stimulation.6o.61 Thus, TRAF6 seems to be the main adaptor molecule to link RANK signaling to NF-KB for the activation ofmature osteoclasrs, but other TRAFs (and possibly other molecules) seem to at least partially compensate for Traf6-deficien cy during osteoclast development. Recently, another molecular adapter was found to be important for RANK Signaling. Grb2 associated binder 2 (Gab2) associates with RANK and mediates RANK-induced NF-KB , Akr and ]NK activation. Genetic inactivation ofGab2 in mice results in osteopetrosis and decreased bone resorption due to defective osteoclast differentiarlon.f Importantly, the contribution of Gab2 to osteoclasrogenesis is relevant not only to mice but also to humans, since siRNA-mediated inactivation of Gab2 in human peripheral blood derived progenitor cells likewise prohibited oseeoclasrogenesis." Various receptors important for osreoclastogenesis such as integrins, c-Frns, FcRy, RANK, or G -protein coupled receptors can act through Gab2 and Gabs can bind to a variety ofsignaling molecules such as PLCy, the p85 subun it ofPI3K, Grb2, or SHP2. Thus, Gab2 (and po ssibly the Gab2 family members Gabl and Gab3) might integrate stimulation ofvarious receptors in osteoclasts.
NF-KB Signaling As indicated above, RANK stimulation triggers the activation ofNF-KB, dimeric transcription factors belonging to the Rel family.63 Mammals express five Rel (NF-KB) proteins that belong to two classes. Proteins of the first class comprise RelA (p65), c-Rel and RelB , are synthesized as mature products and do not require further proteolytic processing.v Conversely, proteins of the second class, NF-KBI and NF-KB2 , are first synthesized as large precursors, p105 and plOO , respectively and proteolytically processed to produce the mature p50 and p52 NF-KBproreins .v NF-ICB dimers containing RelA or c-Rel are retained in the cytoplasm through association with IKB inhibitor proteins. Upon stimulation, IKB becomes rapidly phosphorylated (and thereby marked for ubiquitin-mediared degradation) by the IKK complex, which consists oftwo catalytic subunits, IKKa and IKK13 and the regulatory subunit, IKKy (NEMO).63 In general, activation of p50:RelA and p50:c-Rel dimers is referred to as the canonical pathway which depends mainly on IKK13 activity. In an alternative pathway, the IKKa subunit is required for processing of the NF-KB2/plOO protein complexed with Rem in order to release p52:RelB dirners from inhibition required for nuclear translocation.f Importantly, expression ofboth NF-KB p50 and p52 proteins is required for osteoclast formation, as p50/p52 double knockout mice develop osteopetrosis due to a defect in osteoclast differentiation. 64•65In vitro, the NF-ICB activating kinases IKKa and IKK13 have both been implicated in RANKL-RANK signaling and osteoclasrogenesis. However, the
106
Molecular Mechanisms ofSpondyloarthropathies
analysis of IKKa- and IKK~-deficient mice revealed that IKK~, but not IKKa, is essential for osteoclastogenesis in vivo.66Thus, it seems that RANKL-RANK signaling exerts its downstream effects on osteoclastogenesis mainly through IKK~ and the classical NF-KB activation pathway. This is further supported by the observation that mice in which the upstream activating kinase of IKKa, NIK, has been disrupted are not osteopetroric." Taken together, RANKL-RANK seems to primarily signal through IKK~ to activate the classical NF -KBpathway to control osteoclastogenesis in vivo.
MAPKSignaling Mitogen-activated protein kinases (MAPK) are a family of Ser/Thr protein kinases consisting of extracellular signal-regulated kinases (Erkl/2), p38-MAPKs (a/f3/y/b), c-jun N-terminal kinases (JNKl ,2,3) and 'big' MAPKs (ErkS,7,8).68 Several MAPKs are activated downstream of RANK and help to integrate RANK activation to a cellular response . As for p38-MAPKs, it has been shown that by inhibiting p38a and p38f3with the pharmacological blocker SB203S80 RANKL-induced osteoclastdifferentiation wasabrogated.69These results indicate that p38 MAPK activation plays an important role in RANKL-induced osteoclast differentiation of precursor bone marrow cells." In addition to p38-MAPKs, JNKs and their direct upstream kinase MKK7 have also been shown to be involved in osteoclastogenesis in cell culture.JNKl, but not JNK2, is specifically activated by RANKL and required for osteoclastogenesis in vitro." Moreover, c-jun, a component ofthe dimeric AP-l transcription factors , is activated byJNK and essential fur efficient osteoclasrogenesis.?? Other AP-l family members andJNK targets, namely JunB, c-Fos and Pra, but not JunD, have also genetically been shown to control osteoclascogenesis, thereby strongly supporting the critical function ofMAPK-JNK signaling in RANKL-mediated osteoclast forma tion,7Q-73 In line with these findings,overexpressionofa dominant-negative form ofMKK7 impaired RANKL-mediatedJNK activation and consequently RANKL-mediated osteoclast formation." Finally, ERK p421p44 phosphorylation was reported to be increased by RANK stimulation but was dispensable for RANKL-mediated osteoclast differentiation in vitro .69 Interestingly, another report suggested that specificMEK inhibitors markedly enhanced RANKL-mediated osteoclastogenesis in vitro and moreover implied a crosstalk between the p38-MAPK and ERK pathways during RANKL-mediated osteoclasrogenesls.Ylhe possible involvement ofother ERKs, ERKS, ERK7 and ERK8, in RANK(L) signaling and osteoclastogenesis remains to be tested.
ce:/Calcineurin/NFAT Signaling In search for further factors regulated by RANKL during osteoclast differentiation, the transcription factor NFATc1 was Identified." NFATc1 expression was shown to be dependent on the NF-KB and c-Fos pathways and ectopic expression ofNFATc1 resulted in efficient osteoclast differentiation in vitro even in the absence of RANKL in the culture system." Likewise, NFATc1-deficient ES cellscould not be differentiated into osteoclasts ." Since NFAT family members require the Ca 2+ / calmodulin-dependent Ser/Thr phosphatase calcineurin for activation and nuclear translocation, it wasalso intriguing to test whether RANKL-mediated NFATc1 activation would depend on Ca 2+ Signalingand calcineurin, respectively. Both chelation of Ca 2+ ions in the medium or specific inhibition ofcalcineurin resulted in a block ofosteoclast differentiation in the presence ofRANKL in virro." Moreover, RANKL induced Ca 2+ oscillations are required NFATc1 activation. Collectively, these results argue that RANKL-mediated Ca 2+ oscillation is critical for the terminal differentiation ofosteoclasts,whereby NFATc 1would be constantly activated through the Ca2+-dependent calcineurin pathway," :"
Src, PKB and P/3KKinase Activation The tyros ine kinase c-Src has been implicated in osteoclast function by gene targeting experiments since mice deficient in c-Src develop osteopetrosis." Subsequently, c-Src was connected with RANKL signaling by showing that c-Src is pivotal for RANKL-induced activation of the anti -apoptotic Ser/Thr kinase Akt /PKB in osreoclasts and osteoclast survival." c-Srcand TRAF6 directly interact with each other and with RANK following receptor engagement. TRAF6 seems
RANKL/RANKas Key FactorsfOr Osteoclast Development and Bone Loss in A rthropathies
107
to enhance the kinase activity ofc-Src leading to tyrosine phosphorylation ofdownstream signaling molecules such as c-CbF9 Moreover, PKB activation also requires PI3-kinase activity and PI3-kinase activity was also shown to be important in osteoclastogenesis using specific inhibitors in vitro." The role ofPDK1-which phosphorylates and activates PKB-in RANK(L) signaling has not been reported yet. PIP3 production by PI3-kinase is negatively regulated by two key 3' and S'lipid phosphatases, namely PTEN and SHIP I, which remove a phosphate group from PIP3.81 While overexpression ofPTEN suppressed RANKL-mediated osteoclast differentiation and overexpression ofdominant-negative PTEN constitutively induced osteoclast differentiation in viero." there is still no genetic evidence that PTEN plays an important role in osteoclastogenesis in vivo. By contrast, the importance ofSH IP 1 in osteoclastogenesis has been shown both in vitro and in vivo. In vitro osteoclast cultures demonstrated increased osreoclastogenesis in SHIP I-deficient cells due to hypersensitivity to RANKL and M-CSF stimulation. Most importantly, SHIP 1-1- mice exhibit osteoporosis." Thus , SHIP1 seems to be an essential negative regulator for RANK(L)-induced osteoclast differentiation. Phospholipase C (PLC) is another phosphatidylinositol-relared enzyme that catalyzes the production of the second messengers IP , and diacylglycerol (DAG).84 RANKL acts through PLC to release Ca 2 • from intracellular stores. Subsequently, the increased Ca 2·levels activate the essential transcription factor NFATc 1 via calcineurin. Blockade ofPLCy enzymatic activity impairs early osteoclast development and function ." Importantly, PLCy2-'- mice are osteopetrotic and PLCy2, independent ofPLCy 1, is required for RANKL-induced osteoclastogenesis by differentially regulating NFATc 1, AP-1 and NF-KB.8s Interestingly, PLCy2 can associate with the regulatory adapter molecule GAB2 in osteoclasts, is required for GAB2 phosphorylation and appears to modulate GAB2 recruitment to RANK.-('·8s The activation of PLCy by RANK requires Syk and ITAM-bearing molecules such as DAP12 and FcRy.c(,.8(, Mice lacking DAP12 and FcRy are severely osteopetrotic and DAP 12-I -FcRy-l- double mutant bone marrow cells fail to differentiate into multinucleated osteoclasts in vitro and exhibit impaired phosphorylation ofthe Syk tyrosine kinase." Moreover, Syk-I- progenitors are similarly defective in osteoclast development and bone resorption.87These data indicate that recruitment of Syk to phosphorylated ITAMs is critical for osteoclastogenesis." Mechanistically, FcRy associates with PIR-A and OSCAR, whereas DAP 12 associates with TREM-2 and SIRP~l and the se receptors via DAP12 and FcRy then mediate costimulatory signals from osreoblasts.Ieading to enhanced Ca 2• signaling and NFATc 1 activation in cooperation with RANK signaling.":" Whether osteoblasts are indeed essential for activation ofTREM-2 and SIRP~l is, however, not yet known.
Protein Kinase C (PKC) Signaling Besides the calcineurin-NFATcl pathway, protein kinase C (PKC ) proteins are also activated by Ca 2". Recently, mutations in the atypical PKC (aPKC) scaffolding protein p62 have been shown to be the cause ofthe Sq3S-linked Paget 'sdisease ofbone, a genetic disorder characterized by aberrant osteoclastic activity." P62, like TRAF6, becomes up regulated during RANKL-induced osteoclastogenesis and genetic inactivation ofp62 in mice leads to impaired osteoclastogenesis in vitro and in vivo, as well as inhibition ofIKK activation and NF-KB nuclear rranslocarion." Moreover, RANK Signaling induces the formation ofa ternary complex involving TRAF6, p62 and atypical PKCs. These observations demonstrate that aPKC/p62 signaling is important during osteoclastogenesis and bone remodeling. As for the aPKCs, ).,,/LPKC might well be the key aPKC required for osteoclastogenesis, since PKC~-I- bone marrow-derived macrophages (BMDMs) do not have osteoclastogenic defects ." However, in vivo confirmation has to wait until conditional knockout mice will be available to circumvent the early embryonic lethality ofPKC~ mutant mice,"?
Modulators ofRANKL-Mediated Osteodastogenesis Although RANK(L) Signaling is essential for osteoclastogenesis, crosstalks with other signaling molecules fine-tune this differentiation pathway. For example , RANKL is also expres sed on activated T-cells , but T-cells also secrete a factor that negatively influences RANK signaling. This
Molecular Mechanisms ofSpondyloarthropathies
108
osteoporosis arthritis arthropath ies Figure 2. Activated T-cells produce inflammatory cytokines (IL-l , TNF-a, etc .) which together with calciotropic factors (Vitamin 03, PCE2, etc .) stimulate RANKL expression in osteoblasts (OBs). Activated T-cells-which directly express RANKL-and OBs induce OC differentiation from progenitors via RANKL-RANK signaling , which results in bone resorption by mature osteoclasts (OCs). The soluble decoy receptor for RANKL, OPC, blocks both pathways. Inhibition of RANKL via OPC might be useful to treat osteoporosis, crippling in arthritis, or osteopenic disorders such as Paget's dlsease.?
T-cell-derived negative factor was identified as IFN _y.9O IFN -y promotes accelerated degradation of TRAF6 via the ubiquitin-proteasorne pathway resulting in strong inhibition ofRANKL-induced NF-KB and ]NK activation." Moreover, RANKL also induces IFN-13 in osteoclast precursor cells and IFN-13 inhibits osteoclast differentiation by interfering with the RANKL-induced expression ofc-Fos, a transcription factor essential for the formation ofosteoclasts." This ob servation would suggest an autoregulatory loop in which RANKL-induced c-Fos induces its own inhibitor IFN-13. In line with these observations, IFN-13-/-mice exhibit severe osteopenia accompanied by enhanced osreoclastogenesis.?' Sex hormones are tightly linked to bone disease. Postmenopausal women, for instance, lose estrogen production in their ovaries resulting in loss of bone mass and old-age osteoporosis. In fact, sex hormones such as estrogens an d androgens suppress RANKL-induced osteoclast differentiation by down-regulating the ]NK-c-]un pathway." Moreover, estro gens and androgens can control expression of the decoy receptor OPG, thereby shifting the balance of RANK(L) signaling." Indeed, OPG treatment can revert increased osteoclast numbers and bone loss after ovarectomy in female rats. 12 Whether ]NK-c-]un regulation is an additional pathway for sex hormone regulated bone metabolisms in vivo needs to be demonstrated by genetic rescue experiments. Recently, the mechanism by which estrogen acts to prevent osteoporotic
RANKL/RANKas Key Factorsfor Osteoclast Development and Bone Lossin Arthropathies
109
bone loss has been further elucidated. When estrogen receptor a (ERa) was specifically deleted in differentiated osteoclasts in mice, ERaAQ-yjA~ females exhibited trabecular bone loss, similar to the osteoporotic bone phenotype in postmenopausal women." Importantly, estrogen specifically induced apoptosis in trabecular bones ofwild type but not ERaAQ-yjA~ mice by upregulation ofFas ligand expression." These results suggest that estrogen regulates the life span ofmature osteoclasts via the induction ofthe Fas/FasL system, thereby providing an explanation for the osteoprotective function ofestrogen. Besides hormones, multiple cytokines affect osteoclastogenesis in different ways. For instance, IL-4 abrogates osteoclasrogenesis through STAT6-dependent inhibition of NF-JCB.94 TGF family members, TGF~1 and activin A, but not BMP-2, induce 1unB expression and enhance osteoclastogenesis induced by RANKL in vitro ," M-CSF, a cytokine which signals through the receptor tyrosine kinase c-FMS, is crucial for the proliferation and survival of osteoclastic precursor cells as well as macrophages." M-CSF mediates its effects mainly by activating ERK through GRB2 and Akt/PKB through PI3K, but also stimulates RANK expression in monocyte-macrophage precursor c-FMS+RANK- cells. M-CSF-mediated RANK induction renders these cells responsive to RANKL and induces RANKL-mediated commitment of late-stage precursor cells into osteoclasts .Y" In summary, while RANKL/RANK stimulation is absolutely essential for the induction of osreoclastogenesis, multiple additional receptors and signaling pathways can positively and negatively modulate RANK-mediated bone metabolism. Further analysis of these pathways and their crosstalks might provide novel strategies to control RANK{L) signaling specifically in osteoclasts.
RANK-RANKL as Therapeutic Targets for Arthropathies Bone-related diseases, such as osteoporosi s or rheum atoid arthritis, affect millions of people worldwide and pose a tremendous burden on health care. With the discovery ofRANK{L) and the subsequent key experiments, i.e., genetic proofthat both RANKL and RANK are absolutely essential for osteoclastogenesis and osteoclast activation in vivo, new doors have been opened for bone research and drug development. Our group was the first to demonstrate that RANKL is the key mediator ofosteoclast activation and joint destruction in a rat model ofarthritis." inflammatory cells produce RANKL which then trigger local development and activation of osteoclasts, a finding that now has become the basis for osteoimmunology and which was reproduced in multiple laboratories using various model systems (Fig. 2).2Although cytokines such as IL-l and TNF-a have also been proposed as potential therapeutic targets to control bone loss in arthritis, IL-l and TNF-a alone or in combination are much less potent than RANKL in terms of the induction of active osteoclasts. " Thus, controlling RANKL and RANK at the receptor level or RANK{L) signaling seems to be most promising in the treatment of arthropathies. In particular inhibition ofRANKL is a rational therapeutic strategy to develop novel drugs that block inappropriately enhanced bone resorption-which in severalcaseshas already been shown effective in vivo, for example by using recombinant OPG,IO.12 fully human anti-RANKL blocking antibodies and RANK-Fc. IOO Interestingly, in a mouse parathyroid hormone-regulated protein {PTHrP)-mediated bone resorption model , inhibition ofRANKL using 0 PG causesa greater suppression ofbone resorption and hyper-calcemia than that by bisphosphonsees," the current standard therapy for the treatment ofbone loss. Thus, great hope lies in a fully human monoclonal IgG 2antibody to human RANKL, denosumab (AMG 162), to treat postmenopausal osteoporosis or rheumatoid arthritis (RA). In a randomized, placebo-controlled, dose-ranging Phase 2 study of 412 postmenopausal women with low bone mineral density (BMD), subcutaneous application of denosumab at 3-month or 6-month intervals over a period of 12 months resulted in a sustained decrease in bone turnover and a rapid increase in BMD.101 In an ongoing study with 227 patients with mild or moderately active RA, RANKL inhibition by denosumab also increased BMD.102- 104 In all cases,denosurnab administration was appeared to be well tolerated and at least as good or superior to current standard medication, but further clinical trials are required to substantiate the indicative benefits of RANKL inhibition on suppressing bone destruction in arth ritis.
110
Molecular Mechanisms ofSpondyloarthropathies
Conclusions The identification ofRANKL. its receptor RANK and the decoy receptorofRANKL. OPG. has been ofkey importance for our understanding of osteoclast development and activation . By elucidating their exact functions and signaling pathways. it became possible over the years to devise new and promising strategies to treat bone loss in arthropathies. Although there are other potential targets such as kinases, adaptor molecules and transcription factors to interfere with osteoclastogenesis, targeting the extracellular factor RANKL and its cell surface receptor RANK are currently the most promising and advanced strategies to offer treatment to millions ofpatients suffering from arthropathies in the near future .
References 1. Karsenty G, Wagner EF. Reaching a genetic and molecular understanding of skeletal development. Dev Cell 2002; 2(4) :389-406. 2. Theill LE, Boyle WJ. Penninger JM. RANK-L and RANK : T-cells, bone loss and mammalian evolution. Annu Rev Immunol 2002; 20:795-823. 3. Boyle WJ. Simonet WS. Lacey DL. O steoclast differentiation and activation. Nature 2003; 423(6937):337-342. 4. Wada T, Nakashima T, Hiroshi N er al. RANKL-RANK signaling in osteoclastogenesis and bone disease. Trends Mol Med 2006; 12(1):17-25. 5. 'Io lar J, Teitelbaum SL. Orchard PJ. Osteopetrosis. N Engl J Med 2004 ; 351(27) :2839-2849. 6. Anderson DM. MaraskovskyE, BillingsleyWL et al. A homologue of the TNF receptor and its ligand enhance T-cell growth and dendritic-cell function. Nature 1997; 390(6656) :175-179. 7. Wong BR. Rho J. Arron J er al. TRANCE is a novel ligand of the tumor necrosis factor receptor family that activates c-jun N-terminal kinase in T-cells. J Bioi Chern 1997; 272(40) :25190-25194. 8. Lacey DL, Timms E, Tan HL et al. Osreoprotegerin ligand is a cytokine that regulates osteoclast differentiation and activation. Cell 1998; 93(2) :165-176. 9. Yasuda H, Shima N, Nakagawa N er al. Osteoclast differentiation factor is a ligand for osteoprotegerinl osteoclastogenesis-inhibitory factor and is identical to TRANCE/RANKL. Proc Nat! Acad Sci USA 1998; 95(7) :3597-3602 . 10. Yasuda H, Shima N, Nakagawa N et al. Identity of osteoclastogenesis inhibitory factor (OCIF) and osteoprotegerin (OPG) : a mechanism by which OPG/OCIF inhibits osteoclastogenesis in vitro. Endocrinology 1998; 139(3):1329-1337. 11. Tsuda E, Goro M. Mochizuki S et al. Isolation of a novel cytokine from human fibroblasts that specifically inhibits osteoclastogenesis. Biochem Biophys Res Commun 1997; 234(1):137-142. 12. Simonet WS. Lacey DL. Dunstan CR et al. Osteoprotegerin: A novel secreted protein involved in the regulation of bone density. Cell 1997; 89(2):309-319. 13. Dougall WC. Glaccum M. Charrier K et al. RANK is essential for osteoclast and lymph node development . Genes Dev 1999; 13(18):2412-2424. 14. Kong YY, Yoshida H , Sarosi I et al. OPGL is a key regulator of osceoclastogenesls, lymphocyte development and lymph-node organogenesis. Nature 1999; 397(6717) :315-323. 15. Li J, Sarosi I, Yan XQ et al. RANK is the intrinsic hematopoietic cell surface receptor that controls osteoclasrogenesis and regulation of bone mass and calcium metabolism . Proc Nat! Acad Sci USA 2000; 97(4): 1566-1571. 16. Lam J, Nelson CA. Ross FP et al. Crystal structure of the TRANCE/RANKL cytokine reveals determinants of receptor-ligand specificity. J Clin Invest 2001 ; 108(7) :971-979 . 17. Ito S. Wakabayashi K, Ubukara 0 et al. Crystal structure of the extracellular domain of mouse RANK ligand at 2.2-A resolution . J Bioi Chern 2002; 277(8) :6631-6636. 18. Wong BR, Josien R. Lee SY et al. TRANCE (tumor necrosis factor [TNF]-related activation-induced cytokine), a new TNF family member predominantly expressed in T-cells, is a dendritic cell-specific survival factor. J Exp Med 1997; 186(12):2075-2080 . 19. Kartsogiannis V. Zhou H, Horwood NJ et al. Localization ofRANKL (receptor activator ofNF kappa B ligand) mRNA and protein in skeletal and extraskeletal tissues. Bone 1999; 25(5):525-534. 20. Fata JE, Kong YY, Li J ct al. The osteoclast differentiation factor osteoprotegerin-ligand is essential for mammary gland development. Cell 2000; 103(1):41-50. 21. Loser K, Mehling A. Loeser Set al. Epidermal RANKL controls regulatory Tvcell numbers via activation of dendritic cells. Nat Med 2006; 12(12):1372-1379. 22. Ikeda T, Kasai M, Utsuyama M et al. Determination of three isoforms of the receptor activator of nuclear factor-kappaB ligand and their differential expression in bone and thymus. Endocrinology 2001; 142(4) :1419-1426.
RANKL/RANKas Key FactorsfOr Osteoclast Development and Bone Loss in Arthropathies
111
23. Ikeda T, Kasai M, SuzukiJ er aI. Multimerization of the receptor activatorof nuclear factor-kappaB ligand (RANKL) isoforms and regulation of osteoclasrogenesis, J Bioi Chern 2003 ; 278(47):47217-47222. 24. Suzuki J, Ikeda T, Kuroyama H et al. Regulation of osreoclastogenesis by three human RANKL isoforms expressed in NIH3T3 cells. Biochem Biophys Res Commun 2004; 314(4):1021-1027. 25. SchlondorffJ, Lum L, Blobel CPo Biochemical and pharmacologicalcriteria define two shedding activities for TRANCE/OPGL that are distinct from the tumor necrosis factor alpha convertase, J Bioi Chern 2001; 276(18):14665-14674. 26. Chesneau V. Becherer JD , Zheng Y er at. Catalytic properties of ADAMI9. J Bioi Chern 2003 ; 278(25) :22331-22340. 27. Lynch CC, Hikosaka A, Acuff HB er at. MMP-7 promotes prostate cancer-induced osteolysis via the solubilization of RANKL. Cancer Cell 2005 ; 7(5) :485-496. 28. Hikita A, Yana I, Wakeyama H er at. Negative regulation of osreoclastogenesis by ecrodornain shedding of receptor activator ofNF-kappaB ligand. J Bioi Chern 2006; 281(48) :36846-36855. 29. Chan FK, Chun HJ, Zheng L ec at. A domain in TNF receptors that mediates ligand-independent receptor assembly and Signaling. Science 2000; 288(5475):2351 -2354. 30. Siegel RM, Frederiksen JK, Zacharias DA er aI. Fas preassociation required for apoptosis Signaling and dominant inhibition by pathogenic mutations . Science 2000; 288(5475):2354-2357. 31. Chan KF, Siegel MR, Lenardo JM . Signaling by the TNF receptor superfamily and T-cell homeostasis . Immunity 2000; 13(4):419-422. 32. Locksley RM, Killeen N, Lenardo MJ. The TNF and TNF receptor superfamilies: integrating mammalian biology. Cell 2001; 104(4) :487-501. 33. Nakagawa N, Kinosaki M, Yamaguchi K et at. RANK is the essential Signaling receptor for osteoclast differentiation factor in osteoclastogenesis. Biochem Biophys Res Commun 1998; 253(2) :395-400. 34. Williamson E, Bilsborough JM, Viney JL. Regulation of mucosal dendritic cell function by receptor activator of NF-kappa B (RANK)/RANK ligand interactions: impact on tolerance induction. J Immunol 2002 ; 169(7) :3606-3612. 35. Josien R, Wong BR, Li HL et at. TRANCE, a TNF family member, is differentially expressed on T-cell subsets and induces cytokine production in dendritic cells. J Immunol 1999; 162(5) :2562-2568. 36. Gonzalez-Suarez E, Branstetter D, Armstrong A et at. RANK overexpression in transgenic mice with mouse mammary tumor virus promoter-controlled RANK increases proliferation and impairs alveolar differentiation in the mammary epithelia and disrupts lumen formation in cultured epithelial acini. Mol Cell Bioi 2007; 27(4) :1442-1454. 37. Tan KB, Harrop J, Reddy M et al. Characterization of a novel TNF-like ligand and recently described TNF ligand and TNF receptor superfamily genes and their constitutive and inducible expression in hematopoietic and nonhernaropoieric cells. Gene 1997; 204(1-2):35-46. 38. Kwon BS, Wang S, Udagawa N et at. TRI, a new member of the tumor necrosis factor receptor superfamily, induces fibroblast proliferation and inhibits osteoclastogenesis and bone resorption. FASEB J 1998; 12(10) :845-854. 39. Bucay N, Sarosi 1, Dunstan CR et al osreoproregerin-deficienc mice develop early onset osteoporosis and arterial calcification. Genes Dev 1998; 12(9):1260-1268. 40. Mizuno A, Amizuka N, Irie K et al. Severe osteoporosis in mice lacking osreoclastogenesis inhibitory factor /osteoprotegerin. Biochem Biophys Res Commun 1998; 247(3) :610-615. 41. Hsu H, Lacey DL, Dunstan CR er at. Tumor necrosis factor receptor family member RANK mediates osteoclast differentiation and activation induced by osteoprotegerin ligand. Proc Natl Acad Sci USA 1999; 96(7):3540-3545. 42. Hughes AE, Ralston SH, Marken J et al. Mutations in TNFRSF11A, affecting the Signal peptide of RANK , cause familial expansile osteolysis. Nat Genet 2000; 24(1) :45-48. 43. Whyte MP, Mills BG, Reinus WR et al. Expansile skeletal hyperphosphatasia : A new familial metabolic bone disease. J Bone Miner Res 2000; 15(12):2330-2344. 44. Whyte MP, Hughes AE. Expansile skeletal hyperphospharasia is caused by a IS-base pair tandem duplication in TNFRSF 11A encoding RANK and is allelic to familial expansile osteolysis. J Bone Miner Res 2002 ; 17(1):26-29. 45. Sobacchi C, Frattini A, Guerrini MM et aI. Osteoclast-poor human osteopetrosis due to mutations in the gene encoding RANKL. Nat Genet 2007; 39(8):960-962. 46. Cundy T, Hegde M, Naot D er al. A mutation in the gene TNFRSF11B encoding osteoprotegerin causes an idiopathic hyperphosphatasia phenotype. Hum Mol Genet 2002; 11(18) :2119-2127. 47. Chong B, Hegde M, Fawkner M et at. Idiopathic hyperphosphatasia and TNFRSF11B mutations: relationships between phenotype and genotype . J Bone Miner Res 2003 ; 18(12):2095-2104. 48. Inoue J, Ishida T, Tsukamoto N et at. Tumor necrosis factor receptor-associated factor (TRAF) family: Adapter proteins that mediate cytokine signaling. Exp Cell Res 2000 ; 254(1) :14-24.
112
Molecular Mechanisms ofSpondyloarthropathies
49. Damay BG. Haridas V. Ni] et aJ. Characterization of the intracellular domain of teceptor activator of NF-kappaB (RANK). Interaction with rumor necrosis factor receptor-associated factors and activation ofNF-kappab and c-jun N-terminal kinase.] BioI Chern 1998: 273(32) :20551-20555. 50. Wong BR. ]osien R. Lee SY et aJ. The TRAF family of signal transducer s mediates NF-kappaB activation by the TRANCE recepror.} BioI Chern 1998: 273(43):28355-28359 . 51. Wong BR. ]osien R. Choi Y. TRANCE is a TNF family member that regulates dendritic cell and osteoclast function.] Leukoc BioI 1999: 65(6) :715-724 . 52. Galiberc L. Tomersko ME, Anderson DM et al, The involvement of multiple tumor necrosis factor receptor (TNFR)-associated factors in the signaling mechanisms of receptor activator of NF-kappaB. a member of the TNFR superfamily.] BioI Chern 1998: 273(51) :34120-34127. 53. Lee ZH, Kwack K. Kim KK et aJ. Activation of c-jun N-terminal kinase and activator protein 1 by receptor activator of nuclear factor kappaB. Mol Pharmacol2000: 58(6) :1536-1545 . 54. Lomaga MA, YehWC . Sarosi I et aJ. TRAF6 deficiency results in osteopetrosis and defective interleukin-I, CD40 and LPS signaling. Genes Dev 1999: 13(8) :1015-1024. 55. Naito A, Azuma S. Tanaka S er aJ. Severe osteopetrosis, defective interleukin-I signalling and lymph node organogenesis in TRAF6-deficient mice. Genes Cells 1999: 4(6) :353-362 . 56. Kobayashi N. Kadono Y, Naito A et aJ. Segregation ofTRAF6-mediated Signalingpathways clarifies its role in osteoclastogenesis. EMBO] 2001: 20(6) :1271 -1280. 57. Kobayashi T. Walsh PT. Walsh MC et al, TRAF6 is a critical factor for dendritic cell maturation and development. Immunity 2003: 19(3):353-363. 58. Ye H, Arron ]R, Lamothe B et al, Distinct molecular mechanism for initiating TRAF6 signalling. Nature 2002 ; 418(6896) :443-447. 59. Teitelbaum SL, Ross FP. Genetic regulation of osteoclast development and function . Nat Rev Genet 2003 ; 4(8) :638-649. 60. Kanazawa K, Kudo A. TRAF2 is essential for TNF-alpha-induced osceoclastogenesis. ] Bone Miner Res 2005 : 20(5) :840-847. 61. Kanazawa K, Azuma Y, Nakano H er al, TRAF5 functions in both RANKL- and TNFalpha-induced osreoclasrogenesis.} Bone Miner Res 2003; 18(3):443-450. 62. Wada T, Nakashima T. Oliveira-dos-Santos A] et aJ. The molecular scaffold Gab2 is a crucial component of RANK signaling and osreoclasrogenesis, Nat Med 2005 : 11(4) :394-399 . 63. Karin M. Lin A. NF-kappaB at the crossroads of life and death . Nat Immunol 2002; 3(3) :221-227. 64. Iorsova V, Caarnano ], Loy] et aJ. Osteopetrosis in mice lacking NF-kappaBl and NF-kappaB2 . Nat Med 1997; 3(11) :1285-1289. 65. Franzoso G, Carlson L, Xing L ec al. Requirement for NF-kappaB in osteoclast and B-cell development. Gene s Dev 1997: 11(24):3482-3496 . 66. Ruocco MG, Maeda S. Park]M et al, l{kappa}B kinase (IKK){beta}, but not IKK{alpha}, is a critical mediator of osteoclast survival and is required for inflammation-induced bone loss. ] Exp Med 2005 : 201 (10) :1677-1687. 67. Novack DY, Yin L. Hagen-Stapleton A et aJ. The IkappaB function of NF-kappaB2 plOO cont rols stimulated osteoclastogenesis.] Exp Med 2003; 198(5) :771-781. 68. Wada T, Penninger ]M . Mitogen -activated protein kinases in apoptosis regulation . Oncogen e 2004; 23( 16):2838-2849. 69. Matsumoto M, Sudo T. Saito T et aJ. Involvement of p38 mitogen-activated protein kinase signaling pathway in osreoclastogenesis mediated by receptor activator of NF-kappa B ligand (RANKL). ] BioI Chern 2000: 275(40):31155-31161. 70. David ]P. Sabapathy K. Hoffmann 0 et aJ. ]NKI modulates osceoclastogenesis through both c-jun phosphorylation-dependent and -independenr mechanisms.] Cell Sci 2002: 115(Pt 22):4317-4325 . 71. Wagner EF. Functions of API (Fos/]un) in bone development. Ann Rheum Dis 2002: 61(Suppl 2) :ii40-42. 72. Yamamoto A. Miyazaki T. Kadono Y et aJ. Possible involvement of IkappaB kinase 2 and MKK7 in osreoclastogenesis induced by receptor activator of nuclear factor kappaB ligand. ] Bone Miner Res 2002: 17(4):612-621. 73. Kenner L, Hoebercz A. BeilT et aJ.Mice lacking]unB are osteopenic due to cell-autonomous osteoblast and osteoclast defecrs] Cell Bioi 2004: 164(4):613-623. 74. Hotokezaka H, Sakai E, Kanaoka K et aJ. U0126 and PD98059. specific inhibitors of MEK, accelerate differentiation of RAW264.7 cells into osteoclast-like cells.] BioI Chern 2002 : 277(49) :47366-47372. 75. Takayanagi H , Kim S. Koga T er al, Induction and activation of the tran scription factor NFATcl (NFAT2) integrate RANKL signaling in terminal differentiation of osteoclasts. Dev Cell 2002: 3(6) :889-901. 76. Takayanagi H. Osteoimmunology : Shared mechanisms and crosstalk between the immune and bone systems. Nat Rev Immunol 2007: 7(4) :292-304.
RANKL/RANK as Key Factorsfor Osteoclast Development and Bone Loss in Arthropathies
113
77. Asagiri M. Sato K. Usami T et aI. Autoamplification ofNFATcl expression determines its essential role in bone homeo stasis. J Exp Med 2005; 202(9) :1261-1269. 78. Soriano P. Montgomery C. Geske R et aI. Targeted disruption of the c-src proto-oncogene leads to osteopetrosis in mice. Cell 1991: 64(4) :693-702. 79. Wong BR. Besser D. Kim N et aI. TRANCE. a TNF family member. activates Akr/PKB through a signaling complex involving TRAF6 and c-Src, Mol Cell 1999: 4(6) :1041-1049. 80. Nakamura 1, Takahashi N. Sasaki T er aI. Wortmannin. a specific inhibitor of phosphatidylinositol-3 kinase. blocks osteoclast ic bone resorption . FEBS Lett 1995; 361(1) :79-84. 81. Kalesnikoff J. Sly LM. Hughes MR er aI. The role of SHIP in cyrokine-induced signaling. Rev Physiol Biochem Pharmacol 2003; 149:87-103. 82. Sugatani T. Alvarez U. Hruska KA. PTEN regulates RANKL- and osteopcntin-stimulated signal transduction during osteoclast differentiation and cell motility. J Bioi Chern 2003 ; 278(7) :5001-5008. 83. Takeshita S. Namba N. Zhao JJ et aI. SHIP-deficient mice are severely osteoporotic due to increased numbers of hyper-resorptive osteoclasts, Nat Med 2002: 8(9):943-949. 84. Rhee SG. Regulation of phosphoinosiride-speciflc phospholipase C. Annu Rev Biochem 2001 : 70:281-312. 85. Mao D. Epple H . Uthgenannt B et aI. PLCgamma2 regulates osteoclasrogenesis via its interact ion with ITAM proteins and GAB2. J Clin Invest 2006; 116(11) :2869-2879. 86. Takayanagi H . Mechanistic insight into osteoclast differentiation in osreoimmunology, J Mol Med 2005 ; 83(3):170-179. 87. Mocsai A. Humphrey MB. Van Ziffie JA et aI. The immunomodulatory adapter proteins DAP12 and Fc receptor gamma-chain (FcRgamma) regulate development of functional osceoclascs through the Syk tyrosine kinase. Proc Natl Acad Sci USA 2004; 101(16) :6158-6163 . 88. Laurin N. Brown JP. Mor issette J et al. Recurrent mutation of the gene encod ing sequestosome 1 (SQ ST M l / p62) in Paget disease of bone. Am J Hum Genet 2002: 70(6) :1582-1588 . 89. Duran A. Serrano M. Leitges M er aI. The atypical PKC -interacting protein p62 is an important mediator of RANK- activated osteoclasrogenesis. Dev Cell 2004; 6(2) :303-309. 90. Takayanagi H. Ogasawara K. Hida S et aI. T-cell-mediated regulation of osteoclasrogencsis by signalling cross-talk between RANKL and IFN-gamma. Nature 2000 ; 408(6812):600-605. 91. Takayanagi H. Kim S. Matsuo K er aI. RANKL maintains bone homeostasis through c-Fos-dependenr induction of interferon-beta. Nature 2002; 416(6882) :744-749 . 92. Syed F. Khosla S. Mechanisms of sex steroid effects on bone. Biochem Biophys Res Commun 2005 ; 328(3) :688 -696. 93. Nakamura T. Imai Y. Matsumoto T et aI. Estrogen prevents bone loss via estrogen receptor alpha and induction of Fas ligand in oseeoclasts. Cell 2007: 130(5):811-823. 94. Abu-Amer Y. IL-4 abrogates osreoclasrogenesis through STAT6-dependent inhibition of NF-kappaB. J Clin Invest 2001; 107(11 ):1375-1385. 95. Koseki T. Gao Y. Okaha shi N er aI. Role ofTGF-beta family in osreoclasrogenesis induced by RANKL. C ell Signal 2002; 14(1) :31-36. 96. Ross FP. Teitelbaum SL. Alphavbeta3 and macrophage colony-stimulating factor : partners in osteoclast biology. Immunol Rev 2005; 208:88-105. 97. Arai F. Miyamoto T. Ohneda 0 er aI. Commitment and differentiation of osteoclast precursor cells by the sequential expression of c-Fms and receptor activator of nuclear factor kappaB (RANK) receptors . J Exp Med 1999; 190(12):1741-1754. 98. Kong IT. Feige U. Sarosi I et aI. Activated T-cells regulate bone loss and joint destruction in adjuvant arthritis through osteoprotegerin ligand. Nature 1999; 402(6759) :304-309. 99. Kostenuik PJ. Osteoprotegerin and RANKL regulate bone resorption. density. geomctr y and strength. Curr Opin Pharmacol 2005; 5(6) :618-625. 100. Wittrant Y. Theoleyre S. Chipoy C et aI. RANKLIRANK/OPG : new therapeutic targets in bone tumours and associated osteolysis. Biochim Biophys Acta 2004; 1704(2) :49-57. 101. McClung MR . Lewiecki EM. Cohen SB et aI. Denosumab in postmenopausal women with low bone mineral density. N Engl J Med 2006; 354(8) :821-831. 102. Lane NE. Iannini M. Atkins C er aI. RANKL in hi bition with denosumab decreases markers of bone and cartilage turnover in patients with rheumatoid arthritis. Arthritis Rheum 2006; 54(Suppl 9) :S225-226. 103. Dore R. Hurd E. Palmer W er aI. Denosumab increases bone mineral density in patients with rheumatoid arthritis. Arthritis Rheum 2006 ; 54(Suppl 9):S240. 104. Cohen SB. Valen P. Ritchlin C et aI. RANKL inhibition with denosumab reduces progression of bone erosions in patients with rheumatoid arthritis : Month 6 MRI results. Arthritis Rheum 2006; 54(SuppI9) :S831-832.
CHAPTERS
Bone Formation Versus Bone Resorption in Ankylosing Spondylitis Georg Scherr"
Abstract nkylosin g spondylitis (AS) and other forms of seronegative spondylarthriris (SpA) are characterized by two major processes in joints-the first is chronic inflammation and the second is progressive ankylosis. Both features go hand-in-hand and determine the clinical picture ofdisease, which is joint pain, progressive stiffness and, in case ofperipheral joint involvement also joint swelling.The interplay between inflammation and ankylosis isbest illustrated in AS, where chronic inflammation ofthe spine leads to progressive stiffness, reduced spinal mobility and kyphosis . AS may thus be considered as a synthesis ofinflammatory disease and bone disease.
A
Introduction It is commonly accepted that inflammation is the driving force for structural damage in inflammatory arthropathies. Inflammation in AS is particularly found in the sacroiliac and small facet joints, at the insertion sites of the tendons (enthesis) and along the spinal ligaments. Thu s, the localization ofinflammatory lesions differs profoundly between AS and rheumatoid arthritis (RA). Whereas RA is considered as a disease primarily affecting the synovial membrane, inflammation in AS is much more pronounced at periarticular sites and has a close relation to insertion sites of ligaments and tendons. Inflammation clearlydominates the clinical picture ofAS and precipitates pain, stiffness and loss of function, all ofwhich are reduced by effective therapy. Current knowledge suggest that tumor necrosis factor (TNF) is a pivotal cytokine fueling inflammation in AS, which is backed by the success of treatment ofAS by all 3 TNF-neutralizing agents currently on the market."! The role ofother cytokines in AS is much less established at the moment. Moreover there is a central contribution ofcyclooxygenase in the clinical features ofAS suggesting that the production ofsmall inflammatory lipid mediators like prostaglandin £2 significantly contributes to inflammation in this disease. The consequences of chronic joint inflammation on articular architecture are entirely different in RA and AS, with the latter showing a much more complex reaction." However, despite these overt differences it is evident that the architectural changes of joints are a consequence of the initial inflammatory attack in both disease entities, since there is clear colocalization between inflammation and structural remodeling.
Lessons from Rheumatoid Arthritis RA is a prototype ofan inflammatory joint disease,which links chronic inflammation with bone
damage,"The skeletal lesions in RA are basicallyholes screwed into the juxta-articular bone and show parallels to bone metastases. These lesions form in consequence of chronic inflammatory synovitis, 'Georg Schett-Department of Internal Medicine 3 and Institute for Clinical Immunology, University of Erlangen-Nuremberg, Erlangen, Germany. Email:
[email protected] Molecular Mechanisms ojSpondyloarthropathies, edited by Carlos Lopez-Larrea and Roberto Diaz-Peiia. ©2009 Landes Bioscience and Springer Science+Business Media.
Bone Formation Versus Bone Resorption in AnkylosingSpondylitis
us
which invades and degrades mineralized tissue. Bone destruction in RA is exerted by osteoclasts, which are raised from monocytes/macrophages infiltrating the synovial tissue and differentiate into mature osteoclasts upon cytokine challenge," In RA, bone damage continuously progresses leading to the destruction ofthe joints by bone erosion, leading to instability and loss ofjoint function." It is an important feature ofRA, that there are no major signsofbone and cartilage response towards the inflammatory attack, which allows bone destruction to progress over time . In case of inactive disease these lesions tend to show "sclerosis" on radiographs suggesting that there is at least some reactivity ofskeletal tissue. Still bony responses are not an essentialpart ofthe clinical picture ofRA and instability rather than ankylosis and stiffening is the faith of a joint affected by RA .
Structural Damage in Ankylosing Spondylitis The hallmark of structural damage in AS is bony ankylosis. Ankylosis is the end stage of an excessive bone formation starting from periosteal sites close to the joint and the intervertebral spaces. Focal bone growth, which starts from the periosteal lining, is termed an "osteophyte', which isa bony spur. Whereas such lesions are virtually absent in RA , they constitute a frequently found pathology in AS, psoriatic arthritis and osteoarthritis. Osteophytes are built up at specific sites, which are usually the edges ofbones next to the intra-articular space or to the intervertebral spaces (then termed spondylophytes and syndesmophytes). Syndesrnophytes, vertical bony spurs , ultimately leading to a bridge between vertebrae , are a hallmark ofAS. Similar lesions, now more horizontally oriented, are also found in degenerative joint diseasessuch as OA, psoriatic arthritis or hemochromatosis arthropathy. Osteophytes also emerge at insertion sites of the tendons, especially the Achilles tendon and cause pain upon movement, AS teaches us that bone spurs can bridge the entire distance between neighboring skeletal structures, fusing the joint and creating bony ankylosis. How does this process work? There is yet no full molecular explanation ofankylosis but novel insight in the developmental process of joint formation may help to better understand its basic principles .
Mechanism ofJoint Formation-Molecular Lessons for Joint Fusion Joints and intervertebral spaces form gaps between bones, which allow motion and flexibility. These gaps are actively formed during early development, when chondrogenic formation of the vertebral column and limbs start to branch and build segments. Formation ofthese gaps depends on the expression ofproteins involved in mesenchymal cell differentiation, like cartilage-derived morphogenic proteins- 1 (CDMP-l , also called GDFS) as well as bone morphogenic protein-S (BMP-S)9,10 Without these proteins no joints are formed since the appropriate differentiation of cells, which form the synovial membrane, are lacking. Wingless (Wnt) proteins, such as Wnt-14 (also known as Wnt-9a) are also crucial for the initiation of joint formation in the limbs.'? Joint formation can thus be considered as an active differentiation process, which replaces chondrogenic matrix by specified fibroblast like cells that form the synovial membrane, the periosteum and the joint capsule.
Bony Protrusion as a Stress Response ofthe Joint Joints allow maintaining motion, which, however, requires a structurally intact joint space for smooth gliding ofarticular surfaces. Inflammation leads to joint damage, which causes pain, swelling, stiffness and functional impairment in patients with chronic inflammatory and degenerative joint disease. Under these conditions motion is an exacerbating factor, which enhances symptoms and worsens inflammation. This stress factor can be circumvented by turning on the programs, which regulate bone growth to attempt a bridging ofthe inflamed and instable joint. Induction of these programs requires the differentiation ofcells,which induce bone formation- the osteoblasts. Osteoblasrs originate from resident mesenchymal cells, which undergo a series ofdifferentiation steps before becoming major osreoblasts. Upon final differentiation these cellsproduce bone matrix and facilitate the deposition ofhydroxyapatite crystals. Osteoblasts are the natural counterparts of
116
Molecular Mechanisms ofSpondyloarthropathies
osteoclasts, the primary bone resorbing cells. Osteoblasts and osteoclasrs are functionally coupled and mutually regulate their function. During physiological conditions bone resorption and bone formation are coupled .Osteophyte formation indicates a state, where bone formation by osteoblasts by far outweighs bone resorption by osteoclasts. The attempt of closing the joint and providing stability is thus dependent on a sufficient number ofcells, which enter the osteoblast lineage and built up new bony structures .
Osteoblast-The Bone Forming Cells Differentiation of osteoblasts from mesenchymal cells requires a series of key signals such prostaglandin £2 (PG£2), IGF-l, parathyroid hormone (PTH), bone morphogenic proteins (BMPs) as well as wingless proteins (wnr) . All these signals help to drive mesenchymal cells into pre osteoblast and finally into mature osteoblasts, which are cuboid- shaped cells producing the bone matrix. Osteoblastogenesis is regulated by specific transcription factors such as Cbfa-l (Runx-Z) and Osterix, which are required to built up bone. Under physiological conditions very few osteoblasts are found in a joint, which is usually a region ofslow bone remodeling, particularly ifperipheral joints with a high proportion ofcortical bone are examined. Osteoblasrs also provide signals to the osteoclast allowing their differentiation form hematopoietic precursors and providing a cellular and molecular link between the bone forming and hone resorbing cells.
From the Osteoblast to the Osteophyte Osteophyte formation requires bone deposition, which is linked to the osteoblasts. There are two pathways ofbone formation, one is the direct differentiation ofmesenchymal cells into bone forming osteoblasts and the other more frequent form is endochondral bone formation, which first leads to differentiation ofhypertrophic chondrocytes and proteoglycan rich matrix deposition before bone is built up.Formation ofosteophytes is usually linked to endochondral bone formation and osteophytes start to grow from periosteal sites close to the articular cartilage (Fig. 1). Insertion sites of tendons are predilection sites for osteophyte formation and point to a mechanical factor, which supports enhanced mesenchymal cell differeneiarion.!"!' Alternatively some of the bony spurs observed in AS, especially calcification of the longitudinal ligaments may not be based on endochondral bone formation but rather involve the differentiation of mesenchymal cells into bone forming cells allowing the remodeling ofthe ligament into a bony bridge. It is not yet clear, which mechanisms dominates bone formation in AS. Morphological studies of the small facet joint of the vertebral column suggest that endochondral ossification participates in joint fusion, building a bridge of hypertrophic chondrocytes filling the joint gap.14These cells are then likely replaced by bone when the remodeling ofcartilage starts.
Molecular Regulation of Osteophytes BMPs and Wnt proteins a currently considered as key components, which induce osteophyte formation. Immunohistochemical analysis of osteophytes of the Achilles tendon has demonstrated the expression of phosphorylated and thus activated Smad 3, which is a key component of BMP signaling.13This suggests that BMPs induce the differentiation of mesenchymal cells into osteoblasts and drive bone formation, which builds up bony protrusions. IS Indeed experimental animal models of arthritis, which are characterized by a prominent bone proliferative response such as male DBA mice,exhibit an activation ofBMP signaling. 16•17 Moreover, noggin an inhibitor of the BMP pathway can block this proliferative response .'! Recently, the role ofWnt proteins in osteophyte formation has been investigated. Wnt proteins act synergistic with BMPs in bone formation and are expressed in human jointS.18.19 Activation of the Wnt pathway appears to significantly contribute to bone formation . This is reflected by the effects of blockade ofDickkopf (DKK)-l, an antagonist ofWnt, by neutralizing antibodies." DKK-l blockade increased Wnt signaling by enhancing the activation ofthe intracellular signaling molecule beta-catenin and leads to osteophyte growth even in models, where osteophyte formation is normally absent. Increased activation ofDKK-l is observed in RA and may explain the failure of skeletal response towards
117
Bone Formation Versus Bone Resorption in Ankylosing Spondylitis
B.
A.
EARLY PROLIFERATION
REMODELING
Figure 1. Sequence of osteophyte growth . A ) Mesenchymal cell proliferation (o range) at the periosteum; B) D ifferentiation to hypertrophic chondrocytes (green ) in the center and grow th by mesenchymal cell proliferation at the top . C) Bone growth (gr ey) start ing when remodel ing of primary matrix is initiated ; D ) Remodeling of the osteophyte by replacing the primary cho ndrocy tic matrix by bone. Vascularization and osteoclast-based bone remodel in g (red) .
th e attack of syn ovial inflammatory tissu e." In AS the levels of D K K-1 are very low. even below normal indi vid uals supporting th at Wnt signaling is active in AS. D KK -I . as a major Wnt inhibitor. is regulated by TNF and th e failur e ofTNF to up regulate DKK-I in AS in contrast to RA remains elusive. The se data suggest that BMPs and Wnts are the main driver for the formation of bony spurs. Th ese protein families are involved in mesenchymal cell differentiation and suppo rt osteogenic commitment of cells. Other Wnt inhibitors such as FrzB might act similar and affect ost eophyte formation by down regulating Wnt signaling."
Concepts of Osteophyte Formation Osteophyt e o rigin from the pe riosteum. which is the outer lining layer of bone (Fig. 2). Th e periosteum covers bone and consists ofdense collagen fibers, wh ich are in close contact to mesenchymal cells and microvessels. These cells obviously contain a great potential to proliferate and to differentiate into the various mesenchymal lineages particularly into chondroblasts and osteoblasts. Th is proliferative burst of mesenchymal cells and their differentiation into chondroblasts and osteoblasts results in a cell cluster close to the bone surface. which then remodels into cartilage and bone based on the matrix deposition by chondroblasts and ost eoblasts. Chondroblasts in osteophytic lesions are usually metabolically active hypertrophic chondrocytes similar as they are found in the growth plate. Proteoglycan deposition can occur within osteophytes although proteoglycan content does not meet the one found in articular cartilage. Deposition of bone matrix by osteoblasts is the pr erequisite for giving st ability to the osteophyte. which then appears as solid bony "swelling" typically seen when peripheral joints are affected. Newly formed bone within the osteophyte contains cartilage and ha s strong similarities with th e primar y spon giosa underneath the growth plate. Upon invasion of ost eoclasts int o osteophytes, which require the
Molecular Mechanisms ofSpondyloarthropatbies
JJ8
"Bone marrow"
Periosteum
Remodeling Zone
Growth front
Hypertrophic chondrocytes
Figure 2. Photomicrograph of an osteophyte. Tartrate-resistant acid phosphatase staining of a section through an osteophyte in a murine model of arthritis. Osteoclasts appears purple and are linked to a newly formed "bone marrow cavity". They are localized at the remodeling front bordering hypertrophic chondrocytes. The periosteum forms the basis of the bony spur; the apex is composed of proliferating mesenchymal cells (growth front).
vascularization of these structures, this primary bone is then rebuilt into secondary spongiosa, which is the mature bone and does not contain cartilage remnants any more. Large osteophytes can even contain bone marrow, which is a consequence ofcavity formation within the osteophyte due to osteoclast influx.
Why Do Osteophytes Grow in as and not in RA? Although RA and AS are both considered as chronic inflammatory joint diseases and share common mediators ofinflammation, such as TNF, they exhibit an entirely different clinical picture." AS is a primary osteoproliferative disease, whereas RA is primarily osteodestructive.1t is noteworthy that systemic bone is degraded in both RA and AS resulting in osteoporosis and increased fracture risk. 23.24 Systemic bone loss in RA and AS is most likely a consequence of inflammation, which is an independent risk factor for osreoporosis.P The reason for the profoundly different pattern of bone remodeling between AS and RA is unclear although there are several hypotheses (Fig. 3) .
DifferentialDistribution Whereas RA is a disease, which predominantly involves the small joints and spares the vertebral column, except the upper cervical spine, AS is mostly affecting the lower spine and the sacroiliac joints. It may well be, that these different compartments react differently among stress and show a differential susceptibility to osteophyte formation . It is also conceivable that joints like that sacroiliac joint and the intervertebral spaces show a different reaction pattern towards stress as compared
B one Formation Versus Bone R esorption in A nky losing Spondylitis
act ivity
..
· .. ··. o'
00
•
0
•
0
•
0 00
o• •
.
.,
"
00
o•
119
0
:
:
~ ~
·. ·· .~ o
•
........'
.'.'
.'.'
. RANKL< OPG Wnt
> DKK
-<
RANKL> OPG Wnt
< DKK
BMP > NOGGIN BMP > NOGGIN
Figu re 3. H yp oth eses fo r di fferent ial jo int remodeling in rheumato id arthritis (RA J and anky los ing spo ndy li tis (AS). Center: An k yl o sing sp ondy litis lAS) is an osteo pro lii erati ve co nd itio n, w hereas rh eumatoid arthritis (RA ) is an osteo destructive process. U p pe r left: Different di stri bution of disease with AS in th e spine and RA in th e hand. U p pe r right: D ifferent co m partme nts invo lved in jo in ts: Pro mi nent bon e marrow invol vement in AS with the po tential for an " inside-to -the-o utside" process and synoviti s with bone erosions in RA as a potential "o utsid e-to -t he-insid e" process. Lowe r left: Intermittent disease co urse with flares of high activity in AS and chron ic progressive course in RA. Lower right: Diiierences in bone remodeling with high osteoblast activity and increased expression of OPG -, Wnt- and BMP prot ein s in AS. to peripheral joints. However. in case ofinvolvement ofperipheral joints in A S. the aspect ofjoint swelling differs from synovitis in RA . involving periarticular stru ctures and th e insert io n site for ligaments and tendons. Thi s could mean that the articular compartments primarily affected by AS and RA differ from each other even in the same joints.
Bone Marrow Inflammation As d iscussed . the compartments of the joint affected by AS and RA d iffer from each other. Most prominently, this is seen in case of bone marrow involvement. Bone marrow changes are prominent and Widespread in AS. The sacral and iliac bones as well as the vertebral bodies show exten sive bone marrow inflammation (osteitis). whi ch are a hallmark of AS and detected in MRI scans." In RA . area s ofbone marrow ch anges are usually smaller and more lo calized. In most cases they are linked to an ero sion of the co rt ical bon e layer. which results in b on e marrow chan ges.2~
120
Molecular Mechanisms ofSpondyloarthropathies
Bone marrow inflammation is an important trigger for osteosclerosis. Mixed aggregates ofB-cells and T-cells for instance are present in such bone marrow infiltrates and their production ofBMPs drives osteoblastogenesis in the bone marrow, " Bone marrow inflammation might be a primary event in AS affecting bone formation at periosteal sites. In contrast, RA is dominated by synovitis, which starts outside the cortical bone barrier and invades bone from the outside.
Dijferent Time Courses In contrast to RA, which is mostly a chronic progressive condition, AS is of a much more episodic character with flares followed by periods oflow disease activity. This intermittent course with phases of high and low disease activity in AS might allow joints to respond to the inflammatory attack by new bone formation during periods of low disease activity. This means that disease course in AS has its breaks. which are used for rebuilding of bone as a kind of a stress response. This is an interesting concept since chronic inflammation as it is evident in RA induces the continuous differentiation of osteoclasts, which resorb bone and do not allow an adequate regenerative response.
Molecular Dijferences The obvious imbalance between bone formation and bone resorption in AS and RA, with the former dominating in bone formation and the latter being a primarily resorptive disease, is based on differential activation ofsignaling pathway regulating bone homeostasis. The balances between RANKL and osteoprotegerin (0 PG). BMPs and noggin as well as Wnts and D KKs may essentially influence bone remodeling in AS and RA. For instance, Wnt activity is high in AS but not RA and the natural regulators ofWnt, the D KK proteins are blunted in AS but not in RA. Moreover, there is a tight cross talk between the se protein families . For instance Wnt proteins induce the activation ofOPG, which blocks RANKL activity and thus bone resorption by osteoclasts (20,29) .
Conclusions Although bony proliferations in form of osteophytes, spondylophytes and syndesmophytes can be considered as a response-to-stress strategy of the joint, these lesions affect the function of joints and thus the quality of life. Bone ankylosis is the final con sequence of osteoproliferation, which fuses and immobilizes the joint. Thus prevention of ankylosis is a therapeutic aim in AS , which may either achieved by early intervention before bone proliferation has been established or by directly inhibiting the proliferative response by novel drugs, which specifically interfere with bone formation . Both strategies have not been established in AS, although the fact that continuous blockade of cyclooxygenase has proven to be superior to an on-demand use with respect to structural progression in AS.30 This is particularly interesting since cyclooxygena se is involved in bone formation through generation of PGE2, which is a potent stimulator for bone formation. Thus nonsteroidal anti -inflammatory drugs could indeed have their role in controlling structural damage in AS . Further understanding in the regulation ofosteoproliferation in AS is thus warranted and will help to tailor the therapeutic interventions, which prevent structural remodeling in AS .
References 1. Braun]. Landewe R. Hermann KG er al. Major reduction in spinal inflammation in patients with ankylosing spondylitis after treatment with inflixirnab : Results of a multicenter. randomized. double-blind . placebo-controlled magnetic resonance imaging stu dy. Arthritis Rheum 2006; 54 :1646-52 . 2. Davis]. van der Heijde D. Braun] et al. Recombinant human tumor necrosis factor receptor (Etanercepr) for treating ankylo sing spond ylitis A rand omi zed. controlled trial. Arthritis Rh eum 2003; 48 :3230-6; 3. van der Heijde D. Kivitz A. Schiff M et al. Efficacy and safety of adalimumab in patients with ankylosing spondylitis: Results of a multicenter. random ized. double-blind. placebo-controlled trial. Arthritis Rheum 2006:54 :2136-2. 4. Scherr G, Landewe R. Desiree van der Heijde D. TNF blockers and structural remodeling in ankylo sing spondylitis- what is reality and what is fiction ? Ann Rheum Dis 200 7; 66 :709 -11. 5. Scheer G. Redlich K. Smolen ]S. Inflammation-induced bone loss in the rheumatic diseases. In : Favus M]. Ed . Primer on the Metabolic Bone Diseases and Di sorders of Mineral Metabolism. 6th Edition. Washington D.C. : ASBMR . 2006; 6:310-313.
Bone Formation Versus Bone Resorption in Ankylosing Spondylitis
121
6. Gravallese EM. Harada Y. Wang JT er al. Identification of cell types responsible for bone resorption in rheumatoid arthritis and juvenile rheumatoid arthritis. Am J Parhol 1998; 152:943-951. 7. Boers M . Kostense PJ. Verhoeven AC er a!' Inflammation and damage in an individual joint predict further damage in that joint in patients with early rheumatoid arthritis. Arthritis Rheum 200 I ; 44:2242-6. 8. Scott DL. Pugner K. Kaarela K er al. The links between joint damage and disability in rheumatoid arthritis. Rheumatology (Oxford) 2000 ; 39:122-132. 9. Storm EE. Kingsley DM. Joint patterning defects caused by single and double mutations in members of the bone morphogenetic protein (BMP) family. Development 1996; 122:3969-79. 10. Hartmann C . Tabin C]. Wnt-14 plays a pivotal role in inducing synovial joint formation in the developing appendicular skeleton. Cell 2001 ; 104:341-51. 11. Benjamin M. McGonagle D. The anatomical basis for disease localization in seronegative spondylarthroparhy at entheses and related sites. J Anat 2001; 199:503-526. 12. Ball J. Enthesopathy of rheumatoid and ankylosing spondylitis. Ann Rheum Dis 1971; 30(3):213-23. B . Lories RJ. Derese I, Luyten FP. Modulation of bone morphogenetic protein signaling inhibits the onset and progression of ankylosing enthesiris. J Clin Invest 2005 ; 115:1571-9. 14. Appel H. Kuhne M, Spiekermann S et aI.Immunohistologic analysis of zygapophyseal joints in patients with ankylosing spondylitis. Arthritis Rheum 2006 ; 54(9):2845-51. 15. Scharsruhl A, Vitters EL, van der Kraan PM et al. Reduction of osteophyte formation and synovial thickening by adenoviral overexpression of tran sforming growth factor beta/bone morphogenetic protein inhibitors during experimental osteoa rthr itis. Arthritis Rheum 2003 ; 48 :3442 -51. 16. Lories R], Derese 1, de Bari C et a!' Evidence for uncoupling of inflammation and joint remodeling in a mouse model of spondylarthritis. Arthritis Rheum 2007 ; 56(2) :489-97 . 17. Lories RJ, Matthys P, de Vlam K et al. Ankylosing enthesiris, dactylitis and onychoperiostitis in male DBAII mice : a model of psoriatic arthritis . Ann Rheum Dis 2004 ; 63(5) :595-8. 18. Miller JR . The Wnts . Genome Bioi 2002 ; 3(1 ):3001. 19. D iarra D, Srolina M, Polzer K et al. Dickkopf-I is a master regulator of joint remodeling. Nat Med 2007 ; 13:156-163 . 20. Lories R. Peeters J. Bakker A et a!' Deletion of Frzb affects articular cartilage and biomechanical poperties of the long bones . Arthritis Rheum 2007 , in press. 21. Mcinnes 1, Scheu G. Cytokines in the pathogenesis of rheumatoid arthritis. Nat Immunol 2007; 7:429-44 2. 22. Gough AK, Lilley J, Eyre S er al. Generalized bone loss in patients with early rheumatoid arthritis. Lancet 1994 ; 344 :23-27 . 23. Ralston SH, Urquhart GD , Brzeski M et al. Prevalence of vertebral comp ression fractures due to osteoporosis in ankylosing spondylitis. BMJ 1990; 300:563-565. 24. Scherr G, Kiechl S, Weger S et al. High -sensitivity C vreakrive protein and risk of nontraumatic fracture s in the bruneck study. Arch Int Med 2006 ; 166:2495 -501. 25. Appel H , Loddenkernper C, Grozdanovic Z et a!' Correlation of histopathological findings and magnetic resonance imaging in the spine of patients with ankylosing spondylitis. Arthritis Res Ther 2006; 22:RI43. 26. Jimenez-Boj E, Nobauer-Huhmann I, Hanslik-Schnabel B er a!' Bone erosions and bone marrow edema as defined by magnetic resonance imaging reflect true bone marrow inflammation in rheumatoid arthritis. Arthritis Rheum 2007 ; 56:1118-24. 27. Gortz B, Hayer S. Redlich K er al. Arthritis induces lymphocytic bone marrow inflammation and endosteal bone formation. J Bone Miner Res 2004 ; 19(6) :990-8. 28. Glass DA, Bialek P, Ahn JD et al. Canonical Wnt signaling in differentiated osteoblasrs controls osteoclast differentiation . Dev Cell 2005 ; 8:751-64. 29. van der Heijde DM. Landewe RB, Ory P et aI. Two-year etanercept therapy does not inhibit radiographic progression in patients with ankylosing spondylitis . Ann Rheum Dis 2006; 65(SuppllI):81.
CHAPTER 9
Biomarkers in Spondyloarthropathies Chun-Hsiung Chen, David Tak Yan Yu and Chung-Tei Chou"
Abstract
T
he study ofbiom arkers in spondyloarthroparhy (SpA) has emerged to be a very important field ofresearch . This is particularly because the two commonly used biomarkers, erythrocyte sedimenration rate (ESR) and C-reactive protein (CRP), are ofvery low sensitivity and specificity.The second reason is, with advances in the treatment ofSpA by the very expensive tumor necrosis factor-a (TNF-a) blockers , for cost-effectiveness, clinicians need to be much more accurate in predicting disease progression, evaluating disease activity and monitoring therapeutic efficacy. This review focuses on several biomarkers of promise: matrix metalloproteinases 3 (MMP-3), Type II collagen neoepitope (C2C and Cl-2C), C -propeptide ofType II collagen (CPII), aggrecan 846 epitope, macrophage colony stimulating factor (M-CSF), serum amyloid A (SAA) and Interleukin-6 (IL-6). The results summarized in Table 1 call for a co-ordinated effort for systematic studies ofexisting biomarkers and for search for new candidates.
Introduction A review of literature on biomarkers in spondyloarthritis (SpA) and ankylosing spondylits (AS) would be an easy task if there are already universally acknowledged biomarkers derived from multi-center co-ordinated studies, validated methodologies both in assays and in statistical evaluations and commonly accepted degrees ofthreshold sensitivities and specificities. Other than HLA-B27, ofhigh sensitivity but low specificity, such ideal biomarkers do not exist at this point. What is agreed upon so far would be the definition of SpA, 1 the necessity to use Bath Ankylosing Spondylitis Activity Index (BASDAI) (http://www.asas-group.org) as the golden measure of disease activity, that BASDAI score of~4.0 on a scale ofO-l 0 represents high disease activity,2.3 that erythrocyte sedimentation rate (ESR) and c-reacrive proteins (CRP) are poor biornarkersv' and that disease activity indices respond well to treatment with TNFa blockers.s? A stringent review ofcurrent candidate biomarkers (Table l ) demonstrate that the field ofbiomarker research in SpA has not reached a stage ofmaturity yet. The following is a discu ssion ofwhat type ofbiomarkers we need in SpA, review of major candidates which have been studied and a projection of the future. Because most of such studies have focu sed on AS, it will be the primary target ofthis review.
Purpose ofDeveloping Biomarkers in Ankylosing Spondyltis For osteoarthritis (OA) , biomarkers have been defined as molecules that reflect a specific biological or pathological process, consequence ofa process, or a response to therapeutic intervention. The purposes of identifying biomarkers in osteoarthritis would be to serve as aids for diagnosis or prognosis, for measuring the burden of disease, or the degree of response to treatment. Finally there are biomarkers which cannot serve the above purposes, but provide new information for the pathogenesis of the disease." To a large extent, the purpose of developing biomarkers for AS *Corresponding Author: Chung-Tei Chou -Institute of Clinical M edicin e, Nation al Yang-M ing University, and Veterans General Hospital-Taipei, Taipei, Taiwan . Email : ctchou @vghtpe.gov.tw
Molecular Mechanisms ofSpondyloarthropathies, edited by Carlos Lopez-Larrea and Roberto Diaz-Peiia. ©2009 Landes Bioscience and Springer Science+Business Media.
Wilcoxon matched-pairs signed- ranks lest, 2-tailed
p - 0.01
r - 0.693', P < 0.05 w ith Bonferron i correctio n
p - 0.005
2-tai led Spearman rank cor relation
Wi lcoxo n matched-pairs signed-ranks test,
Response to therapy etanercept
Disease activi ty
Wilcoxon matched-pairs signed ranks test M ultiva riate linear regression analysis ROC Linear regression Spearman rank correlation
Infl iximab progression of disease
continuedon next page
33
44
33
44
29 36 36 33
34 33 32 35
!;.
30 30,32
r - 0.80', P < 0.01 p.O.04 p = 0.013 r • 0.446 ', P • 0.022; r = 0.449 ", P • 0.021 p = 0.007 13 - 0.291, P • 0.004 AUC = 0.703' r - 0.48 ', P • 0.048 ; r - 0.51", P - 0.04 r - 0.693', P < 0.05 with Bonferron i correc tio n
~
W
I ..... N
~.
I:l
30 31
r • 0.366', P • 0.0 17 r = O.sW, p • 0.002 AU C • 0.74', P • 0.01 AUC • 0.75', P - 0.009
Spearman rank corre latio n Spearman rank corre latio n ROC ROC
~
:.
.,.'"
~
S·
3
32
References
r • 0.48', P • 0.0007
Statistical Values, P Value
I ~
Spearman rank correlatio n
Calculation
Pearson correlatio n Wil coxo n rank-sum test, 2-tailed W ilcoxo n matched-pairs signed ranks test Spearman rank cor relation
C-propeptide of Type II Disease activity collagen (CPII) Respon se to therapy etanercepl Aggrecan 846 epi tope
C2C neoepitope
Disease activ ity
M atrix metalloproteinases 3
Response to therapy Infliximab Infliximab Infli ximab Etanercept
Purpose
Biomarkers
Table 1. Serum laboratory tests as potential biomarkers in spondyloarthropathy
b:l
0;' ;;l
Response to therapy infli ximab
Disease activity
Spearman rank correlation
Spearman rank correlation
Spearman rank correlation
r
=
60 60
58
57
Values are biomarkers associated with *BASDAI; *CRP; ttESR; ' mSASSS; *Correlation between CPII/C2C and CRP ROC, Receive Operator Characteristics; AUC, area und er the curve.
Interleukin 6 Spearman rank correlation
35 55,56
P - 0.035
paired t-test Spearman rank correlation; Pearson correlation
response to therapy Disease activity
0.431*, P - 0.007; r = 0.78*, P < 0.0001 r = 0.005*, P < 0.0011; r - 0.57 tt , P < 0.0011 r = 0 .253~, P - 0.034 ; r = 0.3854tt , P = 0.001 r - 0.134*, P = 0.0327 r = 0.689*, P < 0.001
32,47
r = 0.41*, P = 0.004; r = 0.62*, P < 0.05
correlation
Disease activity
Macrophage colony stimulating factor Amyloid A
References
Statistical Values, P Value
Calculation
Purpose
Biomarkers
Table 1. Continued
~
~.
~
~
~ ;:,.. ~
s;:.
~
l!
~
~
~
1::.
l!
~
~
...
l:l ;:;-
~
..... ~
Biomarkers in Spondy/oarthropathies
125
follows the same strategy as for OA. In addition, unlike OA, many patients with AS respond dramatically to treatment with TNF-a blockers. Biomarkers which can predict the degree and duration ofresponse are much needed," Further, for practical purposes, for AS, biomarkers which can betested with blood or urine samples would bemuch more useful compared to those requiring synovial fluids or tissue biopsies. This is because many AS patients do not have swollen peripheral joints for aspiration or biopsy. For example, although biomarkers for response to treatment have been reported in 20 patients with SpA, only 10 ofthose patients have AS.lO
How Do Investigators Assess the Degree ofUsefulness ofSpondylitis Biomarkers? How do investigators assess the usefulness of a candidate biomarker in e.g., diagnosing AS? First, they need to test the candidate biomarker in a cohort ofAS patients in whom the diagnosis is certain. For this, a gold standard for diagnosis is necessary. Currently, this would be the 1984 Modified New York Criteria. Then, they need to compare the results to a cohort of normal subjects, or patients with other types ofdiagnoses such as mechanical spinal pain. Once these data are available, for preliminary evaluation, the investigators will calculate whether there are statistically significant differences between the AS group and the control group , using e.g., the Student's t test or the chi square test. If there is no statistically significant difference, the candidate biomarker is probably useless, although the investigators might try to refine the test by applying it to only a subgroup ofpreselected subjects, e.g., those with peripheral joint swelling. If the statistical results are promising, it does not necessarily mean that the candidate biomarker is useful in clinical practice. To test the degree ofusefulness in a clinical setting, the investigators need to calculate for the likelihood ratio or even better the sensitivity and specificity." The statistical method commonly used to test for sensitivity and specificity is called the Receive Operator Characteristics (ROC). A highly sensitivity biomarker is positive in almost all patients with AS. A highly specific biomarker is negative in almost all subjects not having AS. Statistical evaluation by ROC is sometimes expressed as "area under the curve " (AVC), which is a plot of sensitivity against 1 minus specificity. The highe st AVC is 1.0 indicating perfect sensitivity and specificity.An AVC of 0.5 means that the usefulness of the test is as good as purely by chance alone. In general, a test would be acceptable if the AVC is more than 0.75. This concept is important as we examine the ROC results ofsome of the tests to be described in later sections. The above discussion concerns candidate biomarkers for the purpose ofdiagnosis. Candidate biomarkers for assessing the degree of disease activity, progression of disease, response to treatment, would be based on the same principle. For each ofthese parameters, a gold standard would be required. For example, the commonly accepted gold standard for disease activity would be the "Bath Ankylosing Spondylitis D iseaseActivity Index" (BASDAI) . Since BASDAI is a continuous variable, initial statistical assessment will be by a regression analysis.However, ultimately, the ROC test needs to be used to show the sensitivity and specificity ofthe biomarker.
Matrix Metalloproteinases (MMPS) and Tissue Inhibitors ofMetalloproteinases (TIMPS) Degradation ofthe extracellular matrix (ECM) components is a pathological feature ofchronic inflammatory arthritis. MMPs are zinc-dependent endopeptidases which playa major role in the degradation and remodeling ofthe ECM. 12.15 MMPs are essentialin normal physiological activities, but also participate in many pathological conditions, such as arthritis, cancer and cardiovascular diseases.P:" In arthritis, MMPs are frequently generated by fibroblasts, rnacrophages," synovial cells,18.20 endothelial cells, neurrophils and chondrocytes 21,22 in response to proinflammatory cytokines stimulation such as inrerleukin-I (IL-l) and TNF_a.23.24 Structurally, the MMPs share two common domains: zinc-binding catalytic domain and propeptide domain. The catalytic domain is the active one, while the propeptide domain blocks the catalytic domain, so that the enzyme is inactive until the propeptide domain is removed.25The activity ofMMPs is dependent on the activation ofthe pro-enzymes and is also regulated by inhibitors, such as a2-macroglobulins and
126
Molecular Mechanisms ofSpondyloarthropathies
TIMPs.16 TIMPs inhibit MMPs by forming a 1:1 molar ratio non covalent complexes, blocking the access of substrates to MMPs.16Disruptions of the balance between MMPs and their inhibitors can lead to ECM degradation in the inflamed jointS. 13.16 MMPs have been well studied in rheumatoid arthritis (RA) and are considered to be biomarkers ofRA disease activity. Recently, several studies have indicated that MMPs, particularly MMP-3 (scromelysin-L) are potential candidate biomarkers in SpA. MMP-3 hydrolyzes a number ofECM components, including aggrecan, fibronectin, laminin and collagens26,27 and also activates several proMMPs, such as proMMP-1 and proMMP_9 16,28 and pro-TNF.
Expression ofMMP-3 in Ankylosing Spondylitis Immunohistochemical examination of synovial tissue by Vandooren et aP9from 41 SpA patients showed that both the lining layer and the sublininglayerwere positive for MMP-1 , MMP-2, MMP-3, MMP-9, TIMP-l and TIMP-2. While, MMP-3 was more prominent in the sublining layer, MMP-9 was more prominent in the peri- and intravascular area. Statistically, the expression ofMMPs correlated with the levelofglobal inflammatory cellinfiltration, polymorphonuclear cells and vascularity, suggesting a strong association ofMMPs with inflammatory process. The synovial fluid (SF) levels ofMMP-3 also correlated strongly with the expression ofMMP-3 in the lining layer. Moreover, the SF levels ofMMP-3 were about lOOO-fold higher than the serum MMP-3 levels in paired samples and there was a strong correlation with the serum MMP-3 levels. A study of42 AS patients by Chen et al showed that serum MMP-3 levels were significantly higher in AS patients than in healthy controls." Similarly, a cohort of33 uSpA patients demonstrated higher serum MMP-3 levels than those in healthy individuals." In summary, the above studies show that MMP-3 is increased in the synovial tissue, synovial fluid and serum in the patients with AS. In contrast to MMP-3 and as a comparison, the SF levelsofMMP-9 were not significantly correlated with the synovial expression ofMMP-9 or the serum MMP-9levels and the SF levels ofMMP-9 were only mild higher than those in the serum levels.29
Correlation between Serum MMP-3 and Disease Activity In a study of41 AS patients in a Beijing cohort, a high correlation ofserum MMP-3 levelswith ESR (r = 0.54 , p = 0.0001) and BASDAI (r = 0.48, p = 0.0007) scores was observed ." Similarly in a Taiwan cohort of42 AS patients, serum MMP-3 levels significantly correlated with disease activity (BASDAI) (r = 0.366, p = 0.017), functional ability (BASH) (r = 0.344, p = 0.026) and acute-phase reactants [ESR (r = 0.383, p = 0.012) and CRP (r = 0.416, p = 0.006)].30In addition, change in serum MMP-3levels showed a close correlation with change in BASDAI in a longitudinal follow up of 14 AS patients. In a completely independent study, significant correlations were noted between serum MMP-3levels with CRP (r = 0.73, p < 0.0001) or human cartilage glycoprotein-39 (YKL-40) (r = 0.71, P < 0.001)33 There have been comparatively very few studies in uSpA. One study of 33 uSpA patients showed that serum MMP-3levels correlated significantly with CRP (r = 0.514, p = 0.002) and a weak correlation with ESR.31 Together, these papers show that serum MMP-3levels demonstrate a positive correlation with current clinical and laboratory disease activity indices.
Sensitivity and Specificity ofSerum MMP-3 in Evaluation ofDisease Activity Chen et al were the first to evaluate the exact degree of usefulness ofESR, CRP and MMP-3 for detecting AS patients with high disease activity defined as BASDAI? 4. A threshold cucoff value for each of these measurements was selected by using Two-Graph Receiver Operating Characteristic (TG-ROC) analysis." The degree ofusefulness of the 3 biomarkers are as follows: ESR: sensitivity = 61.5%, specificity = 62.5%, CRP : sensitivity = 61.5% , specificity = 62.5%, MMP-3: sensitivity = 69.2%, specificity = 68.8%. The AUC ofMMP-3 in ROC plot was higher than ESR and CRP at each BASDAI cutoffs. The AUC (p value) ofMMP-3, CRP and ESR in determininghighdiseaseactivity(BASDAI~4) in 42 AS patients is0.74 (0.01) , 0.631 (0.158) and
Biomarkers in Spondyloarthropathies
127
0.612 (0.228), respectively. Similar results were reproduced in another cohort, the AU C (p value) ofMMP-3, CRP and ESR in diagnosing high disease activity (BASDAI ~ 5) in 41 AS patients is 0.75 (0.009),0.645 (0.132) and 0.668 (0.081), respectively.30·32 Atleastwith this particular study, MMP-3 provided a better sensitivity and specificity to detect high disease activity in AS patients than the traditional biomarkers, such as ESR and CRP.
Changes ofSerum MMP-3 in Response to Therapies The first report concerning response to therapies used a cohort of21 AS patients treated with infliximab at the dosage of 3mg/kg at 0, 2 and 6 weeks and then every 2 monehs.r' There were significant reductions ofBASDAI, Bath Ankylosing Spondylitis Functional Index (BASFI), Bath Ankylosing Spondylitis Patient Global Score (BASG), ESR and CRP at 14 weeks. MMP-l and MMP-3levels were measured in 11 ofthe 21 patients. Although, the reductions in serum MMP-l and MMP-3 did not show statistically significant in the patients, there were significant correlations between changes in serum MMP-3levels with changes in BASDAI (r = 0.80, p < 0.01), ESR (r = 0.64, p < 0.05) and CRP [r = 0.83, p < 0.005). In a subsequent study with a cohort of 13 Bejing AS patients treated for 14 weeks with infliximab, infliximab induced significant decreases in serum MMP-3 levels (p = 0.013). As expected, infliximab also induced significant decreases in BASDAI values (p =0.00007).32Before the infliximab treatment, there werehigh degree correlations between serum MMP-3 levelsand ESR and CRP. However,correlations became insignificant after infliximab treatment. Unlike the smaller cohort reported in Maksymowych et al study." there was no statistical correlation between changes in BASDAI scores and changes in serum MMP-3Ievels. As for etanercept, an initial study of9 AS patients showed serum levelsofMMP-l and MMP-3 did not show significantly change after etanercept therapy.33 The negative values might be because of the small number of patients involved. In a subsequent and larger study of 26 refractory AS patients with a 12-week etanercept therapy in Korea, serum MMP-3 levels decreased significantly and the change ofMMP-3 showed a high correlation with the changes ofCRP (r = 0.446, P = 0.022) and ESR (r = 0.449, P = 0.021).35 These and other studies to be described underscore the importance ofco-ordinated multi-center studies in the future. There are very few studies on the effect oftreatment on MMP-3 in the synovial tissues, except for a Belgium study of a 12-week infliximab therapy cohort. A significant down-regulation of MMP-3 in the lining and sublining layer was observed (p = 0.007) .29
Serum MMP-3 as Predictor ofDisease Progression Since disease progression varies considerably among AS patients, biomarkers to predict disease progression are much in need. Maksymowych et al demonstrated that serum MMP-3 is an independent predictor of structural damage progression in the patients with AS.36 In this study radiographic progression over two years was scored by using modified Stoke AS Spine Score (mSASSS)Y Serological biomarkers which reflect different aspect of synthesis and degradation ofECM were measured including cartilage oligomeric matrix protein (CaMP), human cartilage gp-39 (YKL-40), Type II collagen degradation neoepitopes (C2C and CI-2C), C-propeptide ofType II collagen (CPII), aggrecan 846 epirope, osteoprotegerin and MMP-3. Ofthese candidate biomarkers, only serum MMP-3 levelswas significantly associated with 2-year radiographic progression by using multivariate linear regression analysis, with adjustment for age, sex, disease duration, baseline mSASSS and CRP levels (~= 0.29 , P = 0.004) .36 Using baseline MMP-3levels for prediction of any radiographic progression over 2 years, the AUC was 0.703 by ROC plot analysis. If the selected cutoffMMP-3level was set at 68 ng/rnl, the sensitivity = 70% and specificity = 70% . It is remarkable that this study is carried out with a cohort before TNF-a blocker therapies were widely available. Now, with the much wider use of such therapies, assessment in such patients will be necessary. This is particularly because it is not certain how much of current therapies can affect disease progression.
128
Molecular Mechanisms oJSpondy loarthropathies
Are Molecules Reflecting Cartilage Turnover Biomarkers ofAnkylosing Spondylitis? Because OA is a diseasemainly ofthe articular cartilage. molecules involved in cartilage turnover are major candidate biomarkers in OA. Although a disease with usually more inflammation than OA, the cartilage is also a target in SpA.38 Biomarkers of cartilage synthesis and degradation are potential indices ofdiseaseactivity in SpA. Degradation ofType II collagen, the major composition ofarticular cartilage, generates the neoepitopes Co12-3/4Clong mono (C2C) and Co12-3/4Cshort (Cl_2C).39 .40 The assay of neoepitopes ofC2C and C1 -C2 are commonly regarded as markers of degradation. After the degradation of Type II collagen. chondrocytes would up-regulate the synthesis of procollagen.t' :" The assay for serum procollagen Type II C vrerminal propeptide (CPII) is a measurement ofType II collagen synthesis. Aggrecan is one of the main components ofcartilage proteoglycan. The assayfor its 846 epitope has been used as a marker ofproteoglycan synthesis. 42.43Kim et al evaluated these cartilage biomarkers in AS patients and reported that serum levels of CPII, aggrecan 846 epitope and the CPII-to-C2C (CPII :C2C) ratio were significantly elevated in AS patients than in healthy controls. 44 Ofthese cartilage turnover biomarkers. only the CPII:C2C ratio significantly correlated with the CRP level in the subgroup of AS patients who responded to infllxirnab therapy (r = 0.693, p < 0.05). The CPII:C2C ratio could be considered to be the balance between collagen synthesis and degradation and regarded as marker ofdisease activity. In another study of9 AS patients with etanercept therapy, a significant reduction in levelsof C2C (p = 0.005) and a significant increasein the aggrecan 846 epitope (p = 0.01) were observed.P There was also a significant correlation between changes in C2C and changes in ESR (r = 0.51, p = 0.04) and CRP (r = 0.48. p = 0.048). TGF-~, which was detected in sacroiliac joint biopsy from AS patients. might play an important role in the new bone formation.45.46 The results of Kim's study also showed a trend of correlation between the aggrecan 846 epitope and TGF-~ . suggesting the possible role of this cartilage biomarker as an index of new bone formarion.f Therefore , joint cartilage turnover biomarkers are potentially useful parameters for evaluating disease activity and therapeutic efficacy in AS. However, none of these studies have addressed the usefulness as vigorously as in MMP-3 by testing both the sensitivity and the specificity.
Macrophage Colony Stimulating Factor (M-CSF) Several recent studies have addressed the role ofM-CSF in AS. In a cohort of41 AS patients, serum concentrations of M-CSF were significantly increased as compared to those in healthy controls." In addition, serum level of M-CSF also significantly correlated with BASDAI (r = 0.62, p < 0.05), ESR (r = 0.61, p < 0.05) and serum immunoglobulin A (IgA) (r = 0.68, p < 0.05) in the AS patients. M-CSF mRNA expression in peripheral blood monocytic cells (PBMCs) is 5.6-fold increased in the AS patients as compared to that in controls. In a different study with a different cohort of 41 AS patients, there were also significant correlation between the serum levels ofM-CSF and BASDAI values (r = 0.41. p = 0.004).32 However. in this study, there is no significant difference in the serum M-CSF levelsbetween the 41 AS patients and healthy controls." There are two studies addressing effect of therapy on serum M-CSF levels. In one study of 13 AS patients. there was no change in serum M-CSF with infllxlmab therapy.YIn contrast. in another study using 26 AS patients. serum M-CSF levelsdecreased significantly afrer 12-week etanercept therapy (p = 0.035).35
Serum Amyloid A (SAA) SAA is a family ofapolipoprotein. They are synthesized mostly in the liver and synovia by the activated monocyte and rnacrophage.v?? in response to pro-inflammatory cytokines, particularly IL-1 , IL-6 and TNF_a.51,52 SAA in turn can induce production ofMMPs.53.54Patients with AS had a significantly higher mean SAA level than controls and the BASDAI score of the patients with elevated SAA levels was significantly higher than those ofpatients with normal SAA levels.55 SAA also significantly correlated with BASDAI (r = 0.78, P < 0.0001), ESR (r = 0.894 . p < 0.0001) and CRP (r = 0.862, P < 0.0001)56in a cohort of72 AS patients. This is in agreement with another
Biomarkers in Spondyloarthropatbies
129
study ofa 38 AS patient cohort, in which SAA correlated well with BASDAI (r = 0.431, P = 0.007), ESR (r = 0.521, P = 0.001) and CRP [r = 0.648 , p < 0.001).55
Interleukin-6 (IL -6)
Serum levels ofIL-6 was reportedly higher in AS than in healthy controls57.58 and even higher in AS patients with active disease." Serum IL-6levels significantly correlated with ESR (r = 0.57, p < 0.0011) , CRP (r = 0.005, P < 0.0011), platelet count (r = 0.43, p < 0.00 11) and visual analogue stiffness score (VAS) (r = 0.37, p = 0.022) in a study of 60 AS patients." Another cohort of70 AS patients also demonstrated significantly correlation between serum IL-6levels with ESR (r = 0.3854 , P = 0.001) and CRP (r = 0.2537, P = 0.034) .58One recent study showed mild correlation between serum IL-6levels and BASDAI scores (r = 0.134, P = 0.0327).60 The serum IL-6levels also correlated with the degree of physical mobility, such as occiput-to -wall distance, cervical rotation, finger-floor distance and Schober test and with morning pain.57.61 Immunohistological study showed that IL-6 is involved in early, active lesion area of the sacroilitic joints.46 Significant reductions of serum IL-6, together with vascular endothelial growth factor (VEGF) and CRP, was observed in the AS patients as compared to the placebo group after inAiximab therapy (p < 0.001) .61.62 In a cohort of201 AS patients with inAiximab therapy by Visvanathan S er al showed that change in serum IL-6levels correlated well with change in CRP (r = 0.689 , p < 0.001) and change in VEGF (r = 0.445 , p < 0.001) at week 24. High levelsof baseline IL-6 and CRP were associated with better clinical response (ASAS 20 and BASDAI SO) after inAiximab treatment. Multiple regression analysis showed early reductions in IL-6 were significantly associated with improvements in BASDAI scores and the spinal MRI activirv'"
Conclusions From the briefsurvey ofcurrent biomarkers ofAS, it is clear that the field is only at its infancy. What it lacks most are co-ordinated studies among multiple centers, with agreement in methodology ofassaysand in statistical thresholds. Within these limitations, the most thoroughly evaluated biomarker is MMP-3. However, many more biomarkers will need to be discovered and evaluated. The discovery ofnew biomarkers will probably relyon massivescreening techniques: genome wide association for gene polymorphism, micro array for gene expression and proteomics at the protein level. In addition, there will be screening for epigenetic control and signaling pathways. However, AS is a complex disease. In spite of all these technical advances, it appears unlikely that a single biomarker will be ofsufficient discriminatory ability. There is a high likelihood that a combination of biomarkers will be much more useful than a single one alone. Use ofstat istical methods which can test the usefulness ofcombinations ofbiomarkers will be very important for future research.
References 1. Zochling J, Brandt J, Braun J. The current concept of spondyloarthriris with special emphasis on undifferentiated spo ndyloarrhriris. Rheumatology (Oxford) 2005 ; 44:1483-91. 2. Garrett S, Jenkinson T, Kennedy LG et aI. A new approach to defining disease status in ankylosing spondylitis: the bath ankylosing spondylitis disease activity index. J Rheumatol 1994; 21 :2286-9 1. 3. Braun J. Pham T, Sieper J et al. International ASAS consensus statement for the use of anti-tumour necrosis factor agents in patients with ankylosing spondylitis . Ann Rheum Dis 2003; 62 :817-824. 4. Sheehan NJ, Slavin BM, Donovan MP et aI. Lack of correlation between clinical disease activity and erythrocyte sedimentation rate, acute phase proteins or protease inhibitors in ankylos ing spondylitis. Br J Rheumatol1986; 25:171-4. 5. Ruof], Stucki G. Validity aspects of erythrocyte sedimentation rate and C-reactive protein in ankylo sing spondylitis: a literature review. J Rheumatol 1999; 26:966-70. 6. Gorman JD, Sack KE, Davis JC Jr. Treatment of ankylosing spondylitis by inhibition of tumor necros is factor alpha. N Engl J Med 2002; 346 :1349-56. 7. Braun ], Brandt ], ListingJ et aI. Treatment of active ankylosingspondylitis with infliximab : A randomised controlled multicentre t rial. Lancet 2002; 359 :1187-93. 8. Bauer DC, Hunter DJ, Abramson SB er al. Osteoarthritis biomarkers network. Cla ssification of osteoarthritis biomarkers : a proposed approach. Osteoarthritis Cartilage 2006; 14:723-7.
130
Molecular Mechanisms ofSpondyloarthropathies
9. Huang F. Zhu J. Zhang L et al. Response to one infu sion predicts subsequent improvement as well as the rate of relapse of ankylosing spondylitis infused with three pulses of intllximab. Clin Rheumatol 2007; 26 :920-6 . 10. Kruithof E. De Rycke L. Vandooren B et al. Identification of synovial biomarkers of response to experimental treatment in early-phase clinical trials in spondylarthriris. Arthritis Rheum 2006; 54:1795-804. 11. Pepe MS. Janes H. Longton G et al. Limitations of the odds ratio in gauging the performance of a diagnostic, prognostic. or screening marker. Am J Epidemiol 2004 ; 159:882-90 . 12. Nagase H . Woessner JF Jr. Matrix meralloproreinases, J Bioi Chern 1999; 274 :21491-4. 13. Ishiguro N, Ito T. Miyazaki K et al. Matrix rneralloproreinases, tissue inhibitors of metalloproteinases and glyeosaminoglycans in synovial fluid from patients with rheumatoid arthritis. J Rheumarol 1999; 26 :34-40. 14. Ishiguro N, Ito T, Obara K et al. Determination of strornelysin-I, 72 and 92 kDa type IV collagenase, tissue inhibitor of metalloproteinase-I (TIMP-l) and TIMP-2 in synovial fluid and serum from patients with rheumatoid arthritis. J Rheumarol 1996; 23:1599-604. 15. Yoshihara Y, Nakamura H. Obara K et al. Matrix meralloproteinases and tissue inhibitors of met alloproteinases in synovial fluids from patients with rheumatoid arthritis or osteoarthritis. Ann Rheum Dis 2000; 59:455-61. 16. Visse R, Nagase H . Matrix metalloproteinases and tissue inhibitors of metalloproteinases: Structure. function and biochemistry. Circ Res 2003; 92 :827-39. 17. Grillet B, Dequeker J, Paemen L et al. Gelatinase B in chronic synovit is: inununolocalization with a monoclonal antibody. Br J Rheurnarol 1997; 36:744-7 . 18. Hembry RM , Bagga MR. Reynolds JJ et al. lmmunolocalisation studies on six matrix meralloprorelnases and their inhibitors. TIMP-l and TIMP-2, in synovia from patients with osteo- and rheuatoid arthritis. Ann Rheum Dis 1995; 54:25-32. 19. Okada Y. Takeuchi N. Tomita K et al. Immunolocalization of matrix meralloprorelnase 3 (stromelys in) in rheumatoid synovioblasts (B-cells): Correlation with rheumatoid arthritis. Ann Rheum Dis 1989; 48 :645-53. 20. Ahrens D. Koch AE, Pope RM et al. Expression of matrix meralloproreinase 9 (96-kd gelatinase B) in human rheumatoid arthritis . Arthritis Rheum 1996; 39: 1576-87. 2 I. Cole AA, ChubinskayaS, Schumacher B er al. Chondrocyte matrix meralloproreinase-B, Human articular chondrocytes express neutrophil collagenase. J Bioi Chern 1996; 271 :11023-6. 22. Mohtai M, Smith RL. Schurman DJ et al. Expression of 92-kD type IV collagenase/gelatinase (gelatinase B) in osteoarthritic cartilage and its induction in normal human articular cartilage by interleukin 1. J Clin Invest 1993; 92:179-85. 23. MacNaul KL , Chartrain N. Lark M et al. Discoordinate expression of stromelysin, collagenase and tissue inhibitor of meralloproteinases-I in rheumatoid human synovial fibroblasts. Synergistic effects of inrerleukin-I and tumor necrosis factor-alpha on stromelysin expression. J Bioi Chern 1990 ; 265 :17238 -45. 24 . Zhang Y, McCluskey K, Fujii K er al. Differential regulation of monocyte matrix metalloproteinase and TIMP-l production by 'I'Nf--alpha, granulocyte-macrophage, CSF and IL-l beta through prostaglandin-dependent and independent mechanisms. J Immunol 1998; 161:3071-6 . 25. Visse R, Nagase H. Matrix metalloproteinases and tissue inhibitors of metalloproteinases: structure, function and biochemistry. Circ Res 2003 ; 92:827-839. 26. Manicourt DH, Fujimoto N, Obara K et al. Levels of circulating collagenase. stromelysin-I and tissue inhibitor of matrix rneralloproreinases 1 in patients with rheumatoid arthritis. Relationship to serum levels of antigenic keratan sulfate and systemic parameters of inflammation. Arthritis Rheum 1995; 38 :1031-9. 27. Manicourt DH, Fujimoto N, Obara K er al. Serum levels of collagenase. stromelysin-I and TIMP-1. Age- and sex-related differences in normal subject s and relationship to the extent of joint involvement and serum levels of antigenic keratan sulfate in patients with osteoarthritis. Arthritis Rheum 1994; 37 :1774-83. 28. Nagase H . Activation mechanisms of matrix mecalloproreinases. Bioi Chern 1997; 378:151-60. 29 . Vandooren B, Kruithof E. Yu DT et al. Involvement of matrix met alloproteinases and their inhibitors in peripheral synovitis and down-regulation by tumor necrosis factor alpha blockade in spondylarthropathy, Arthritis Rheum 2004 ; 50:2942 -2953. 30. Chen CH, Lin KC , Yu DT et al. Serum matrix metalloproteinases and tissue inhibitors of metalloproteinases in ankylosing spondylitis : MMP- 3 is a reproducibly sensitive and specific biomarker of disease activity. Rheumatology (Oxford) 2006; 45 :414-20. 31. Chen CH, Liao HT. Chen HA et al. Serum levels of matrix meralloproteinase-B in undifferentiated spondyloarthropathy, Scand J Rheumatol 2007 ; 36 :326-8 .
Biomarkers in Spondyloartbropathies
131
32. Yang C, Gu], Rihl M et aI. Serum levels of matrix metalloproteinase 3 and macrophage colony-stimulating factor 1 correlate with disease activity in ankylosing spondylitis. Arthritis Rheum 2004; 51:691-699. 33. Maksymowych WP, Poole AR, Hiebert L er aJ. Etanercepr exerts beneficial effects on articular cartilage biomarkers of degradation and turnover in patients with ankylosing spondylitis. ] Rheumatol 2005; 32 :1911-1917. 34. Maksymowych WP, ]hangri GS, Lambert RG er aJ. InAiximab in ankylosing spondylitis: A prospective observational inception cohort analysis of efficacy and safety. ] Rheumatol 2002 ; 29 :959-965. 35. Woo ]H, Lee H], Sung IH et al, Changes of clinical response and bone biochemical markers in patients with ankylosing spondylitis taking etanercept. ] Rheumarol: in press. 36. Maksymowych WP, Landewe R, Conner-Spady B et al, Serum matrix metalloproreinase 3 is an independent predictor of structural damage progression in patients with ankylosing spondylitis. Arthritis Rheum 2007; 56:1846-53. 37. Creemers MC, Franssen M], van'r Hof MA er aJ. Assessment of outcome in ankylosing spondylitis: An extended rad iographic scoring system. Ann Rheum Di s 2005 ; 64:127-9. 38. Rousseau ]C, Delmas PD. Biological markers in osteoarthritis. Nat Clin Pract Rheumatol 2007; 3:346-56. 39. Billinghursr RC, Dahlberg L, Ionescu M er aJ. Enhanced cleavage of type II collagen by collagenases in osreoarthritic articular carnlage, ] Clin Invest 1997; 99:1534-45. 40. Kojima T, Mwale F, Yasuda T et aJ. Early degradation of type IX and type II collagen with the onset of experimental inflammatory arthritis. Arthritis Rheum 2001 : 44:120-7. 41. Nelson F, Dahlberg L, Laverty, S et aI. Evidence for altered synthesis of typ e II collagen in patients with osreoarrhricis.} Clin Invest 1998; 102:2115-25. 42. Poole AR . Cartilage in health and disease. In : Koopman W; ed, Arthritis and Allied Conditions: A Textbook of Rheumatology. 14th ed. Philadelphia: Lippincott, Will iams and Wilkins, 2001 ; 226-84. 43. Poole AR , HoweIl DS . Etiop athogenesis of osteoarthritis. In: Moskowitz RY, Howell DS , Goldberg VM, Mankin H], eds. Osreoarthriris: Diagnosis and Management. 3rd ed. Baltimore: Saunders, 2001 ; 29-47. 44. Kim T H , Stone M, Payne U et al, Cartilage biomarkers in ankylosing spondylitis: relationship to clinical variables and treatment respon se. Arthritis Rheum 2005 ; 52:885-91. 45. Braun], BoIlow M , Neure L et al, Use of immunohistologic and in situ hybridization techniques in the examination of sacroiliac joint biopsy specimens from patients with ankylosing spondylitis. Arthrit is Rheum 1995 ; 38 :499-505. 46. Francois R], Neure L, Sieper ] et al, Immunohistological examination of open sacroiliac biopsies of patients with ankylosing spondylitis: detection of tumour necrosis factor alpha in two patients with early disease and transforming growth factor beta in three more advanced cases. Ann Rheum Dis 2006; 65:713-20. 47. Yang PT, Kasai H , Xiao WG et al, Increased expression of macrophage colony-stimulating factor in ankylosing spondylitis and rheumatoid arthritis. Ann Rheum Dis 2006; 65 :1671-2. 48. Uhlar e M , Wh itehead AS. Serum amyloid A, the major vertebrate acute-phase reactant. Eur] Biochem 1999; 265 :501-23. 49. O 'Hara R, Murphy EP, Whitehead AS et aJ. Acute-phase serum amyloid a production by rheumatoid arthritis synovial tissue. Arthritis Res 2000 ; 2:142-4. SO. Vallon R, Freuler F, Desta-Tsedu N et aJ. Serum amyloid a (apoSAA) expression is up-regulated in rheumatoid arthritis and induces transcription of matrix meealloproreinases. ] Immunol 200 1; 166 :2801 -7. 51. Strissel K], Girard MT, West-Mays]A et aJ. Role of serum amyloid A as an intermediate in the IL-l and PMA-stimulated signaling pathways regulating expression of rabb it fibroblast collagenase. Exp Cell Res 1997 ; 237:275-87. Regulation of serum amyloid a gene expression in syrian hamsters by 52. Dowton SB, Peters CN, jesrus cytokines. Intlamrnation 1991; 15:391-7. 53. Migita K, Kawabe Y, Tominaga M et aJ. Serum amyloid a protein induces production of matrix metalloproreinases by human synovial fibroblasrs. Lab Invest 1998 ; 78:535-9. 54. O 'Hara R, Murphy EP, Whitehead AS et aJ. Local expression of the serum amyloid a and formyl peptide receptor-like 1 genes in synovial tissue is associated with matrix metalloproteinase production in patients with inflammatory arthritis. Arthritis Rheum 2004 : 50 :1788-99. 55. lung SY, Park MC, Park YB et aJ. Serum amyloid a as a useful indicator of disease activity in patients with ankylosing spondylitis. Yonsci Med] 2007 ; 48 :218-24. 56. Lange U, Boss B, Teichmann] ct aJ. Serum amyloid A-An indicator of inAammation in ankylosing spondyliti s. Rheumatol Int 2000 ; 19:119-22 . 57. Gratacos ], Collado A, Fildla X er aJ. Serum cytokines (IL -6, TNF-alpha, lL -l beta and IFN-gamma) in ankylosing spondylitis: A close correlation between serum IL-6 and disease activity and severity. Br] Rheumatol1994; 33 :927 -31.
n.
132
Molecular Mechanisms ofSpondyloarthropathies
58. Bal A. Unlu E. Bahar Get aI. Comparison of serum IL-l beta . sIL-2R, IL-6 and TNF-alpha levels with disease activity parameters in ankylosing spondylitis . Clin Rheumatol 2007; 26:211-5. 59. Graracos ], Collado A. Pons F et aI. Significant loss of bone mass in patients with early. active ankylosing spondylitis: a followup study. Arthritis Rheum 1999; 42:2319-24. 60. Visvanarhan S. Wagner CL. Marini]C et aI. Inflammatory biomarkers. disease activity and spinal disease measures in patients with ankylosing spondylitis after treatment with infliximab. Ann Rheum Dis in press. 61. Falkenbach A. Herold M. In ankylosing spondylitis serum interleukin-6 correlates with the degree of mobility restriction, but not with short-term changes in the variables for mobility. Rheumatol Int 1998; 18:103-6 . 62. Brandt ]. Haibel H. Comely D et aI. Successful treatment of active ankylosing spondylitis with the anti-tumor necrosis factor alpha monoclonal antibody infliximab . Arthritis Rheum 2000; 43:1346-52 .
CHAPTER
10
Therapy ofSpondyloarthritides Jiirgen Braun"
Abstract nkylOSin g spondylitis (AS) is the major subtype and a major outcome of an interrelated group ofrheumatic diseases now named as spondyloanhritides (SpA). The most important clinical features of this group are inflammatory back pain (IBP), asymmetric peripheral oligoarthritis, predominantly of the lower limbs, enthesitis and specific organ involvement such as anterior uveitis, psoriasis and chronic inflammatory bowel disease. Aortic root involvement and conduction abnormalities are rare complications of AS. For clinical purposes, five subgroups are differentiated: AS, psoriatic SpA (PsSpA), reactive SpA (ReSpA), SpA associated with inflammator y bowel disease (SpAIBD) and undifferentiated SpA (uSpA). The SpA are genetically linked, the strongest known contributing factor is the MHC class I molecule HLA B27 , ARTS-7, and IL-23R, others still remain to be identified. Most frequently and characteristically, AS staresin the sacroiliacjoints at a mean age of26 years affecting men only slightly more frequent than women. In about 80% of the patients the disease spreads to the spine where all three segments are affected, most frequently thc thoracic spine. Osteodestructive structural changes such as erosions occur less frequently than osteoproliferative changes which are pathognomonic for AS being clinically impressive by their appearance as syndesmophyres and ankylosis. Established classification criteria for AS and SpA perform less well in early disease stages. This partly contributes to the delay of diagnosis which is in the range of 5-10 years-mainly due the high frequency of back pain in the population. Major factors to improve the rate ofAS patients diagnosed early are HLA B27 and imaging of the sacroiliac joints . International recommendations for the management ofAS have been published. The conventional treatment is mainly based on NSAIDs, patients with peripheral arthritis may be treated with sulfasalazine and patients with persistcntly active diseasebenefit from therapy with anti -TNF agents. Physiotherapy is of major importance in the general approach to patients with AS.
A
Introduction Ankylosing spondylitis (AS) is the major subtype and a major outcome ofan interrelated group of rheumatic diseases now named as spondyloarthritides (SpA).! The most important clinical features of this group are inflammatory back pain (IBP), asymmetric peripheral oligoarthriris, predominantly of the lower limbs, enthesitis and specific organ involvement such as anterior uveitis, psoriasis and chronic inflammatory bowel disease. Aortic root involvement and conduction abnormalities are rare complications of AS. For clinical purposes, five subgroups are differentiated: AS, psoriatic SpA (PsSpA), reactive SpA (ReSpA), SpA associated with inflammatory bowel disease (SpAIBD) and undifferentiated SpA (uSpA) . The SpA are genetically linked, the strongest known contributing factor is the MHC class I molecule HLA B27, ARTS-7, IL-23R others are currently being Identified. ' *Jurgen Braun-Rheumazentrum Ruhrgebiet, Landgrafenstrasse 15, 44652, Hern e, Germ any. Email : verw altung@rheumazentrum-ruhrgeb iet.de
Molecular MechanismsofSpondyloartbropathies, edited by Carlos L6pez-Larrea and Roberto Diaz-Pena. ©2009 Landes Bioscience and Springer Science+Business Media.
Molecular Mechanisms ofSpondyloartbropatbies
134
Most frequently and characteristically,AS starts in the sacroiliac joints at a mean age of26 years affecting men only slightly more frequent than women. In about 80% of the patients the disease spreads to the spine where all three segments are affected, most frequently the thoracic spine. Osreodestructive structural changes such as erosions occur less frequently than osteoproliferative changes which are pathognomonic for AS being clinically impressive by their appearance as syndesmophytes and ankylosis. Vertebral fractures occur more frequently in patients with AS as compared to the general population. Standard assessments for patients with AS are available and International recommendations for the management of AS have recently been published.'
Management Ten main recommendations for the management ofAS have been proposed very recently by a combined 'ASessment in Ankylosing Spondylitis working group (ASAS)/European League Against Rheumatism (EULAR) task force (Fig. 1).2 Briefly,the treatment ofAS should be tailored according to the current manifestations of the disease, the level of current symptoms and several other features including the wishes and expectations ofthe patient. The disease monitoring ofAS patients should include patient history, clinical parameters, laboratory tests and imaging. The frequency ofmonitoring should be decided on an individual basis depending on symptoms, severity and medication. The optimal treatment ofAS requires a combination ofnonpharmacological and pharmacological treatment modalities includingpatient education and physicaltherapy. Anti-1NF therapy should be given according to the ASAS recommendations.' Joint replacement has to be considered in patients with radiographic evidence ofadvanced hip involvement who have refractory pain and disability. Spinal surgery is useful in selected patients with symptoms and disability due to disabilitating posture or instable spine.
ASAS/EULAR Recommendations for the Management ofAS Education, Exercise, Physical, Therapy, Rehabilitation, Patient Associations, SelfHelp Groups
NSAIDs
g e
TNF Blockers
r
Figure 1. ASAS/EULAR recommendations for the management of AS. Modified from: Zochling
J et al. Ann Rheum Dis 2006; 65(4):442-5V
Therapy ofSpondyloartbritides
135
Table 1. Measurement domains in AS Domain
Disease Activity
Function
Damage
Patient global assessment Spinal pain Spinal stiffness Spinal mobility Physical function Peripheral joints and en theses Fatigue Disease activity Quality of life Acute phase reactants Imaging
x x x x x x x x x x x
x
x x x
x x x x
x
x
x
x x
Assessments The main outcomes in AS are disease activity which includes pain, morning stiffness and patient's global assessment, function, quality oflife and structural damage." When assessing patients with AS, it is useful to think of measuring disease activity, physical function and structural damage as separate facets of the AS process: disease activity reflecting acute inflammation and rate of change. physical function reflecting the impact the disease has on the patient's ability to perform in his/her daily life and structural damage being the end result of the AS process on anatomical structures. Accepted measurement domains within AS can be seen to cover one or more ofthese general concepts (Table 1). The assessments in Ankylosing Spondylitis (ASAS) group, an international collaboration of clinicians, researchers and industry representatives with particular interest and expertise in AS, wasestablished in 1995 with a goal to rationalize the assessment ofthis debilitating disease. ASAS has developed practical, concise core sets ofconcepts important for patient monitoring (Table 2), both in a clinical practice setting and for assessing treatment response in clinical trials." The core sets were put together using a combination of expert consensus, research evidence and statistical approaches and can be thought of as a standard framework for patient assessment. ASAS has subsequently reviewed the extensive literature on different outcome measures and instruments which have been used in AS clinical trials and selected the most appropriate measures for each core set domain based on evidence ofvalidity and consensus opinion," The ASAS group recommends that the measures put forward in the core sets should be used in all research projects in AS, to standardize outcome measurement, to ensure that meaningful patient outcomes are not overlooked and to facilitate comparisons of response across studies. It is stressed that although the core sets describe the minimum set of domains that should be assessed and monitored in AS patients, they are not exclusive or exhaustive; other concepts such as health-related quality oflife can also add important information and each situation should be individually assessed with regard to the specific aim of the assessment.
Disease Activity Disease activity in AS is measured by the BASDA!,7 composite index evaluating fatigue. axial and peripheral pain, stiffness and enthesopathy (Table 3). The self-administered instrument is made up of6 questions, each to be answered on a VAS scale (0-10 em), where 0 = none (or 0 hours for morning stiffness) and 10 ern = very severe (or 2 or more hours for morning stiffness), regarding the patient's symptoms in the previous week. It is easy and quick to complete, the final score is a simple sum ofits components and the BASDAI has been extensively validated in clinical trials8·9 and translated into several languages. A BASDAI score of> 4 is internationally accepted to indicate
136
Molecular Mechanisms ofSpondyloarthropathies
Table 2. ASAS core sets for assessment in ankylosing spondylitis Core Set Domain
CR SMARD/PT DC-ART Instruments
Patient global assessment
x
x
x
VAS in the last week
Spinal pain
x
x
x
VAS pain at night, average in the last week and VAS, average in the last week
Spinal stiffness
x
x
x
VAS morning stiffness
Spinal mobility
x
x
x
Chest expansion and modified schober index and occiput-to-wall distance
Physical function x
x
x
Bath ankylosing spondylitis functional index, or dougados functional index
x
Number of swollen joints No preferred instrument for entheseal disease
x
No preferred instrument
x
ESR
x
AP and lateral X-rays lumbar spine, lateral cervical spine, pelvis (SI and hip joints)
Peripheral joints and entheses
x
Fatigue Acute phase reactants
x
Imaging
Abbreviations: CR:clinical record keeping; DC-ART:d isease-controlling antirheumatic therapy; SMARD: symptom-modifying antirheumatic drugs; PT: physical therapy; VAS: Visual analogue scale; ESR: Erythrocyte sedimentation rate; AP: antero-posterior; 51; sacroiliac. From: Van der He ijde D et al. J Rheumatol 1999; 26(4):951-954.5 active disease and most clinical trials oftherapy in AS now require patients to have active disease as defined by a BASDAI >4 before inclusion. The BASDAI is one of the most commonly used outcome measures in clinical trials and is simple enough to be implemented in daily practice.
PhysicalFunction There is no single parameter that adequately measures the concept of physical function. A number ofpatient-assessed AS-specific instruments are available which cover a range ofphysical functions and activities of daily living, in order to summarize how a well a patient functions in daily life and to quantify 'disability' (Table 3). The most commonly used are the Bath Ankylosing Spondylitis Functional Index (BASFI, Table 3) which has been shown to perform well in regards to reliability, validity and responsiveness across a range of settings. 10.11 The Bath AS Metrology Index is most commonly used to measure spinal mobility (Table 3), the BASMi-lO and BASMi linear versions are preferable because they are much more sensitive to change.
PeripheralJoints andEntheses Peripheral joint involvement occurs in approximately 25% ofpatients with AS, usually in the form of oligo -articular, asymmetrical large joint involvement. The formal joint counts in use for rheumatoid arthritis are therefore not necessarily as useful in this setting. The ASAS group suggests using a 44-joint count, which includes the sternoclavicular joints, acromioclavicular joints, shoulders, elbows, wrists, knees, ankles, metacarpophalangeal and metatarsophalangeal joints and the proximal interphalangeal joints of the hands. The core set advocates measuring only swollen
X-ray scoring system for the lateral cervical spine, AP and lateral lumba r spine and hips, using the New York system to grade the sacroi liac jo ints X-ray scoring system for the lateral cervi cal and lateral lumb ar spi ne, score range 0-72
Function
M obil ity (spine and hip)
Structural damage
Structu ral damage
BASFI
BASM I
BASRI
mSASSS
Bath ankylosing spondylitis Metrology index (16)
Bath ankylosi ng spondylitis radiolo gy index (19)
Modified stoke ankylosing spondylitis spinal score (18)
- mod ified Schober's test - intermalleolar distance
A composite index made up of 5 cli nical measurements: - cervi cal rotation - tragus to wa ll distance - lumbar side flexion
A composite index made up of 10 questions, coveri ng basic dail y functi o ns such as bendi ng and stand ing, each measured on a 0-100 mm VAS
A composit e index made up of 6 questions, each measured o n a 0-100 mm visual analogue scale (VAS): - fatigue • neck, back or hip pain • pain/swelling in other join ts (not neck, back or hip) - overall discomfort fro m tender areas - overall level of mo rning stiffness (intensity) - duration of morning stiffness
Bath ankylosing spond ylitis functio nal index (8)
Disease activity
BASDA I
Description
Bath anky losing spo ndylitis disease activity index (7)
Measures
Abbrev.
Instrument
Table 3. Validateddisease-specific instruments used for measurement in ankylosing spondylitis
~
....
'-l
w
!t ::;. ~ ....
~
I:.
S""
~
";,s
~ ~
~
~
138
M olecular M echanisms ofSpondyloarthropathies
joints in thi s way; arguments can be made for recording painful joints also. Peripheral joint d isease reflects both disease activity (acute inflammation) and physical fun ction, but rarely progresses to significant struc t ural damage. Th ere is no specific instrument recommen ded in th e ASA S core sets for the assessment ofenth esitis in AS. The Mander Enthesitis Index (ME l ) was the first but is not feasible in clinical trials, the Maastricht Ankylosing Spondylitis Enthesitis Score (MASES) includes only 13 entheses and uses th e d ichotomous responses 'no pain' and 'painful' for each site, resulting in a score between o and 13.J2 Another simplified enthesitis score has been developed in Berlin which requires th e assessment of 12 different enth eses for disease involvement concentrating on the clinically most relevant sites such as the heel and the trochanter and is expressed as a simple score between 0 and 12;13this method ha s not yet been formally validated.
Fatigue An important source of morbidity in AS patients, fatigue seems to be associated with di sease activity, fun ctional ability and global well-being." There are no specific disease-related measurement instruments for fatigue in AS; one question of the BASDAF asks about the overall level of fatigue/ tiredness in the past week (VAS, 0-100 mm), but none ofthe other composite instruments address this domain. This item has been recently shown to be sensitive and specific for fatigue in AS patients with a cut-offof70 mm 15and is the recommended instrum ent for measuring fatigue in the ASAS core sets.
Imaging Structural changesofAS including syndesmophytes,erosions, sclerosisand ankylosis can be seen on sp ina l radiographs. Plain X-rays ofthe spine should include AP and lateral views ofthe lumbar spine and lateral views of the cervical spine. There are currently thre e validated scoring systems in use to assessspinal struc t ural damage in clinical trials in AS: the original Stoke Ankylosing Spondylitis Spinal Score (SA SSS,17),a modified SASSS (mSASSS ,18)and the BASRI. 19 The mSASSS has been identified as the most reliable method." Magnetic resonance imaging (MRI) of the sacroiliac joints and the spin e is increasingly used to assess disease activity in AS. Although it has not been inco rporated in the ASAS core set to date it seems likely on the basis ofrecent data that MRI will have a role both in clinical trials and in daily care of the patients, because it is advantageous to have so me objective evidence of spinal inflammation. While conventional radiography of the spine and the sacroiliac joints is mainly used to detect chronic structural changes in AS, MRI is indicated to detect active axial inflammation, including sacroiliitis, spondylitis and spondylod iscitls":" which are not well seen with oth er imaging tech niques. In clinical practice a combination ofX-rays and MRI appe ared most useful." Ultrasonography can be a useful tool to detect enthesiti s and bursitis in patients with spondyloarthrirides.P'" Ultrasound is much more sensitive than clinical examination for detecting these changes and in clinical pract ice can assist in diagnosis and with ultr asound-directed aspiration or corticosteroid injection.
Laboratory Tests While HLA B27 is only useful for diagnosis, especially in early disease, C- reactive protein (C RP ) and erythrocyte sedimentation rate (ESR) are frequently used to assessdisease activity but th eir usefulness in this regard was shown to be limited.' ?
Measuring Treatment Response The ASAS group has taken th e core sets and their respect ive measurement instrum ents to co nstruc t specific compo site respon se criteria for use to measure treatm ent respon se in AS trials (Table 4) .28Initially derived from 5 short-term trials ofNSAIDs in AS, th e initial improvement criteria consist offour outcome dom ains: physical fun ction, spinal pain, patient global assessment and inflammation. Improvement is defined as a 20% improvement from baseline, or a 10 mm
139
TherapyofSpondyloarthritides
Table 4. ASAS response criteria Instrument
Abbrev.
Description
ASAS improvement criteria (28)
ASAS-IC
Four domains, based on the discrimination between NSAID treatment and placebo - Physical function, measured by the BASFI - Spinal pain, measured on a 0-100 mm VAS - Patient global assessment in the last week , on a 0-100 mm VAS - Inflammation, measured as the mean of the last 2 BASDAI questions (intensity and duration of morning stiffness)
ASAS 20% ASAS20 response criteria (28)
Treatment response is defined as: - ~20% and ~1O mm VAS on a 0-100 scale in at least 3 of the 4 ASAS-IC domains and - no worsening of ~20% and ~10 mm VAS on a 0- 100 scale in the remaining 4th domain
ASAS40 ASAS 40% response criteria (30)
Treatment response is defined as: - ~40% and ~20 mm VAS on a 0-100 scale in at least 3 of the 4 ASAS-IC domains and - no worsen ing of ~40% and ~20 mm VAS on a 0-100 scale in the remaining 4th domain
ASAS 5 out of 6 ASAS 5/6 response criteria (30)
Developed for use in trials of anti-TNF therapy, six domains were incl uded: - pain - patient global assessment - function - inflammation - spinal mobility - C reactive protein (acute phase reactant) Treatment response is defined as improvement in 5 of 6 domains without deterioration in the 6th domain, using predefined % improvements.
From: Van Tubergen A et al. Ann Rheum Dis. 2003; 62(3):215-221 .92
improvement from baseline for VAS measures on a 0-100 mm scale,in at least 3 ofthe 4 domains. There can not be deterioration of20% or more, or of 10 rom or more on a VAS scale, in the corresponding 4th domain. The response criteria show high specificityand moderate sensitiviry" and have been validated in anti-TNF studies. 29 ,3o
Basic Principles ofTreatment The standard treatment ofspinal symptoms ofpatients with AS has consisted ofnonsteroidal anti-inflammatory drugs (NSAIDPI and structured exercise programs" for decades. Whether and to which extent physical therapy and exerciseare beneficial in every stage of the disease (for example in very active disease) is not known. Disease activity, especially the degree of spinal inflammation, function and damage are likely to influence the outcome of physical therapy and regular exercise. Non-pharmacologic therapy comprises besides physical therapy also SpA therapy, patient education and patient selfhelp groups. A recent Cochrane review" concluded that evidence-based medicine isdifficult to provide for that indication but there isa strongly positive expert opinion, the
140
Molecular Mechanisms ofSpondyloarthropathies
effect size is probably small. However, intensive SpA therapy has been shown to be more efficacious than standard prescriptions ofexercisesin an outpatient setting, also after several months."
Pharmacologic Therapy with NonsteroidalAnti-Inflammatory Drugs (NSAIDs) In general, NSAIDs work rather well in patients with AS. A good response to NSAIDs has even been identified as a diagnostic sign for SpA,35 while a state ofnonresponsiveness to NSAIDs may identify patients with a bad prognosis. Clinical experience suggests that active patients should be continuously treated with NSAIDs in a dosage sufficient to control pain and stiffness.36 A recent study has even suggested that continuous dosing with NSAIDs rather than the usual on demand prescription decelerates radiographic progression over 2 years." However, NSAIDs including coxibs are known to have gastrointestinal and potentially also cardiovascular toxicity which may limit their use at least patients in patients at risk." Furthermore, about half of the patients with AS report insufficient control of their symptoms by NSAIDs alonc"
Pharmacologic Therapy with Disease-ModifyingAnti-Rheumatic Drugs (DMARDs) The use ofDMARDS for the treatment ofaxial disease in the SpA has been rather disappointing. Therapies which are effective in suppressing disease activity and slowing disease progression in rheumatoid arthritis (RA) have notably failed to impress in the SpA, particularly for spinal disease.t" Sulfasalazine has been shown to improve SpA-associated peripheral arthritis, but not spinal pain." However, there are differences between the trials related to disease duration and the proportion ofpatients with peripheral arthritis. Thus, the efficacyofsulfasalazine in earlier disease stages may be different. Indeed, in a recent controlled trial of sulfasalazine in uSpA and early AS some efficacy on spinal pain was noted since patients with IBP but no peripheral arthritis had a significantly larger improvement in disease activity than the placebo group despite of using less NSAIDs.42 However, all patients improved and definite conclusions are difficult . Methotrexate (MTX) is commonly used in RA with good results, improving symptoms and slowing the progression of erosive disease. This is different in AS, suggesting another pathomechanism. In a systematic review on the use of MTX in AS,43the conclusion was that there is no evidence for an effect on IBP and inconclusive evidence ofefficacy for peripheral joint disease. The only randomized controlled trial ofMTX in AS44 failed to show a significant effect of7.5 mg oral MTX weekly on spondylitis but there was some improvement ofperipheral arthritis. An open label trial treating 20 active AS patients with 20 mg MTX s.c, for 16 weeks did also not show any efficacy on axial and limited improvement on peripheral symptoms." Many rheumatologists are still using MTX for AS because there used to be no other options. The differences in response between peripheral and axial symptoms may be due to the entheseal pathology. Similarly, leflunomide (LFL) is effective in treating the symptoms and slowing radiographic change in RA. In AS, LFL was not effective to treat axial manifestations.t" but patients with peripheral arthritis had some benefit." LFL is effective in psoriatic arthrltis." Bisphosphonates may be useful to treat spinal symptoms of AS patients as suggested by one study," However, other studies with pamidronare failed to show a similar effect size in patients with AS.50 Thalidomide was also used with some success in patients with AS51 but it is considered to be too toxic for widespread use.
Tumor Necrosis Factor (TNF) Blockers The introduction ofTNF blockers has been the most substantial development in the treatment ofAS and other SpA in the last years.1 Three such agents are now approved for AS: the monoclonal chimeric antibody infliximab, which is given in a dosage of3-5 mg/kg every 6-8 weeks i.v, (approved is the 5 mg/kg every 6-8 weeks regimen) , the fully humanized monoclonal adalimumab which is given in a dosage of 40 mg s.c, every other week and the 75 kD TNF receptor fusion protein etanercept given in a dosage of 50 mg s.c. once or 25 mg s.c, twice a week. The success
Therapy ofSpondyloarthritides
141
ofanti-TNF treatment is very likely to be a class effect. There is some evidence that this therapy works even better in SpA than in RA.52 Large randomized controlled trials of Inflixirnab,13.53 etanercept54,55 and adalimumab'" in AS have shown impressive short term improvements in spinal pain, function and inflammatory markers with therapy as compared to placebo. As shown by MRI, spondylitic changes are largely reduced with these therapies.23.57-61 As experience with these therapies grows out to 2-5 year trials now,61-64 efficacy may persist with ongoing treatment and more than one third of the patients are in remission . The trials show significant improvement ofpain, function and disease activity in AS patients with active disease compared to placebo. Indeed, all outcome measures including BASDAI, BASH, BASMI and the physical component of the SF-36 improved significantly after 24 and 102 weeks. The improvement usually starts within 2 weeks of therapy and CRP levels also tend to decrease rapidly. Alongside the demonstrated long-term efficacyand safety ofTNF blockers in AS, it is important to note the loss of response after cessation ofcontinuous therapy with infliximab for 3 years 65but readministration was successful and did not cause problems, This was similar with etanerccpt/" The studies on infliximab in ASl3,53 differ from etanercept and adalimumab'v" in the fact that about 30% of patients were allowed to continue D MARD and corticosteroid therapy because they were used to that. Adalimumab also showed efficacy in AS in a pilot srudy." and an RCT in which the pain of so me patients with advanced sp in al ankylosis also improved.56 While anti-TNF therapy was shown to decrease spinal inflammation as detected by MRI,57.61 no major inhibition ofradiological progression ofdisease as assessed by the mSASSS 18was seen in a small number ofAS patients treated with infliximab for 2 and 4 years. 67,68 Bone mineral density was shown to improve in patients on anrl-TNf therapy already after 6 months/" Clinical disease activity and spinal inflammation as detected by MRI are significantly reduced by TNF blockers, as shown after short- and long-term anti-TNF therapy. Whether anti-TNF treatment is capable to at least partly halt radiographic progress ion is as yet unclear, uncoupling of inflammation and new bone formation has been discussed as one possibility why there could be a difference between AS and RA. However, structural damage may playa less important role in a disease with significant long-term functional disab ility which is only in part due to structural damage but also due to persistent disease activity which seem s to be effectivel treated with anti-TNF agents. The first pilot studies ofinfliximab and etanercept in undifferentiated SpA have been successfuFO,71 The largest study with adalimumab in this indication showed that a high proportion of patients had major improvement ofsigns and symptoms." Similarly, infliximab, etanercept and adalimumab have been shown to be effective for peripheral joint and skin symptoms in patients with psoriatic arthritis. 72•75 Etanercept is effective for rheumatic manifestations in inflammatory bowel diseases (IBD) regarding joint and spine but not gut symptoms." In line with that, etanercept has no effect on IBDn and does not prevent flares of IBD.78 This is in contrast to infliximab which is approved for Crohn's disease (CD)79 and ulcerative colitis (UC).80 Adalimumab is now also approved for CD. Thus, etanercept is not recommended for the comparatively small SpA subgroup with concomitant IBD . There may also be a difference in the prevention and treatment ofanterior uveitiS.81.83 Recommendations on which AS patients should be treated with TNF-blockers are especially needed on the background ofpossible side effects and the relatively high costs ofthese drugs. Thus, patients with the best risk/benefit ratio should be treated preferentially. An international ASAS consensus statement for the use ofanti-TNF agents in AS patients was published in 2003 and updated in 2006.84A summary ofthese recommendations for the initiation ofanti-TNF alpha therapy are shown in Table 5. A prediction of response to anti-TNF therapy is difficult, patients early in the course oftheir disease, with elevated CRP,85 positive MRI findings, or less structural damage are more likely to respond but overall all patient subgroups may benefit from this treatment.
142
Molecular Mechanisms ofSpondyloarthropathies
Table 5. ASAS consensus for anti-TNF therapy Patient Selection • Diagnosis: Patients 'normally' fulfilling modified New York Criteria for definitive AS (90) • Modified New York criteria 1984 • Radiological criterion • Sacroiliitis, grade > 11bilater ally or grade III to IV unilaterally • Clinical criteria (2 out of the following 3) • Low back pain and stiffness for >3 months that improves with exercise but is not relieved by rest • Limitation of mot ion of the lumbar spine in both the sagitt al and front al plan es • Limitation of chest expansion relative to norm al values correlated for age and sex • Disease activity • Active disease for >4 weeks • BASlJAI >4 (0-10) and an expert' opinion that anti -Tt-Jl- treatment should be startedt • Treatment failure All pati ents must have had adequate therapeutic trials of at least 2 NSAIDs. - An adequate therapeutic trial is defined as: • Treatment for >3 months at maximal recommended or tolerated anti-inflammatory dose unless contraindicated • Treatment for < 3 months where treatment was withdrawn because of intolerance, toxicity, or contraindications • Patients with symptomatic peripheral arthritis (normally having a lack of response to a local steroid injection for those with oligoarticular involvem ent) must have had adequ ate therapeutic trial of both NSAIDs and sulfasalazlnef • Patients with symptomatic enthesit is must have had an adequate th erapeutic trial of at least two local steroid injections unless contraindicated • Contraindications (90 and as locally defined) Assessment of disease • ASAS core set for daily practice • Physical function (BASFI) • Pain (VAS, past week, spine at night, due to AS and VAS, past week, spine due to AS) • Spinal mobility (chest expansion and modified Schober test and occiput to wall distance and lateral lumbar flexion) • Patient 's global assessment (VAS, past we ek) • Stiffness (duration of morning stiffness, spine, past week) • Periph eral joints and entheses (number of swollen joints (44 joint count), enthesiti s score such as developed in Maastricht, Berlin, or San Francisco) • Acute ph ase reactants (ESR or CRP) • Fatigue (VAS) • BASDAI • VAS overall level offatigue/t iredn ess past w eek • VAS overall level of AS neck, back, or hip pain past w eek • VAS overall level of pain/swelling in joints other than neck, back or hips past w eek • VAS overall discomfort from any areas tend er to touch or pressure past week • VAS overall level of morning stiffness from time of awakening past week • Duration and intensity (VAS) of morning stiffness from tim e of awa kening (up to 120 minutes)
continued on next page
Therapy ofSpondyloarthritides
143
Table 5. Continued Patient Selection Assessment of response • Respondercriteria BASDAI 50% relative changeor absolutechange of 2 (scale 0-10) and expert opinion : Continuation yes/no • Time of evaluation Between 6 and 12 weeks VAS, visual analogue scale; all VAScan be replaced by a numerical rating scale (NRS). ·The expert is a doctor, usually a rheumatologist, with expertise in inflammatory back pain and the use of biological agents. Experts should be locally defined; tan expert opinion comprises clinical features (history and exam ination), serum acute phase reactant levels, or imaging results, such as radiographs demonstrating rapid progression or MRI scans indicating inflammation; *sulfasalazine: treatment for > 4 months at standard target dose of 3 g/day, or maximally tolerated dose unless contraindicated or not tolerated. Treatment for <4 months, where treatment was withdrawn because of intolerance or toxicity or contraindicated. Braun Jet al. Ann Rheum Dis. 2006; 65(3):316-320. 3
Directed at a different cytokine in the inflammatory response than the TNF blockers, anakinra is a recombinant human interleukin-I receptor antagonist.In contrast to TNF, it is unclear whether IL-l is present in sacroiliac joints. Two open studies ofanakinra in AS have been published in the last year,showing partly conflicting results.86•87 Other biologics have not been tested so far-studies with rituximab, abatacept and tocilizumab are currently being planned.
Conclusions Cost-effectiveness is an issue when expensive therapies are discussed. Despite the high costs, it was recently demonstrated that the clinical benefits and the improvements in quality oflife in patients with AS treated with infliximab result in lower disease-associated costs than standard care, which translates to short-term cost ofapproximately £3 5,000 per quality-adjusted life year (QALY) gained 88-an amount societies may be willing to pay. However, the calculated costs were higher in other analyses. When modeling for long-term therapy, using annual disease progression of0.07 of the BASFI in the sensitivity analysis, the cost per QALY gained is reduced to < £10000. It seems that the costs for anti-TNF therapy fall well inside what is considered to be 'cost-effective'. Much in line with this, it was recently shown that the daily productivity ofpatients with active AS which was significantly associated with functional impairment and disease activity greatly improved with infliximab and this was associated with reduced workday loss among employed parienrs."
References I. Braun J, Sieper]. Ankylosing spondylitis . The Lancet 2007; 369:1379-1390. 2. Zochling J, van der Heijde D, Burgos-Vargas Ret al. ASAS/EULAR recommendations for the management of ankylosing spondylitis. Ann Rheum Dis 2006; 65(4) :442-52. 3. Braun J, Davis J, Dougados M er al. First update of the international ASAS consensus statement for the use of anti-TNF agents in patients with ankylosing spondylitis. Ann Rheum Dis 2006; 65(3):316-20. 4. Zochling J, Braun]. Assessments in ankylosing spondylitis. Best Pract Res Clin Rheumatol 2007 ; 21(4) :699-712. 5. van der Heijde D, Bellamy N, Calin A et al. Assessments in ankylosing spondylitis working group . Preliminary core sets for endpoints in ankylosing spondylitis. J Rheumatol 1997; 24:2225-29. 6. van der Heijde D, Calin A, Dougados M et al. ASAS working group. Selection of instruments in the core set for DC-ART, SMARD, physical therapy and clinical record keeping in ankylosing spondylitis. Progress report of the ASAS group . J Rheurnarol 1999; 26:951-54. 7. Garrett S, Jenkinson T, Kennedy G et al. A new approach to defining disease status in ankylosing spondylitis: The Bath ankylosing spondylitis disease activity index (BASDAI) . J Rheurnatol 1994; 1:2286-91.
144
Molecular Mechanisms ofSpondyloarthropathies
8. Calin A. Nakache J-P. Gueguen A ct al. Defining disease activ ity in ankylosing spondylitis: is a combination of variables (Bath Ankylosing Spondylitis Disease Activity Index) an appropriate instrument ? Rheumatology 1999; 38 :878-82. 9. Calin A. Garrett S, Whitelock H et al. A new approach to defining functional ability in ankylosing spondylitis: the development of the bath ankylosing spondylitis fun ctional index. J Rheumatol 1994 ; 21:2281-85. 10. Spoorenberg A, van der Heijde D, de K1erk E et al. A comparative study of the usefulness of the bath ankylosing spondylitis functional index and the dougados functional index in the assessment of ankylosing spondylitis. J Rheumatol1999; 26 :961-65. II. Haywood KL. Garratt AM. Dawes PT. Patient-assessed healrh in ankylosing spondylitis: a structured review. Rheumarology 2005 ; 44 :577-86. 12. Heufi-Dorenbosch L. Spoorenberg A, van Tubergen A et al. Assessment of enthesitis in ankylosing spondylitis. Ann Rheum Dis 2003; 62 :127-32. 13. Braun J. Brandt J, ListingJ et al. Treatment of active ankylosing spondylitis with infliximab : a randomised controlled multicentre trial. Lancet 2002; 359 :1187-93 . 14. Dagfinrud H. Vollesrad NK, Loge JH et al. Fatigue in patients with ankylosing spondylitis: a comparison with the general population and associations with clinical and self-reported measures . Arthritis Rheum 2005; 53 :5-11. IS. van Tubergen A. Coenen J, Landewe R et al. Assessment of fatigue in patients with ankylosing spondylitis: a psych ometric analysis. Arthritis Rheum 2002 ; 47 :8-16 . 16. Jenkinson TR, Mallorie PA. Whitelock HC et al. Defining spinal mobility in ankylosing spondylitis (AS) . The bath AS metrology index . J Rheumarol 1994; 21(9):1694-98. 17. Taylor HG, Wardle T, Beswich EJ er al. The relationship of clinical and laboratory measurements to radiological change in ankylos ing spondylitis. Br J Rheumatol 1991 ; 30 :330-35. 18. Creemers MCW'; Franssen M]AM, van 't HofMA et al. Assessment of outcome in ankylosing spondylitis: an extended radiographic scoring system. Ann Rheum Dis 2005 ; 64 :127-29. 19. Calin A. Mackay K, Santos H ct al. A new dimension to outcome: application of the bath ankylosing spondylitis radiology index.] Rheurnatol 1999 ; 26 :988-92. 20 . Wanders AJ. Landewe RB, Spoorenberg A et al. What is the most appropriate radiologic scoring method for ankylos ing spondylitis? A comparison of the available methods based on the outcome measures in rheumatology clinical trials filter. Arthritis Rheum 2004; 50(8):2622-32. 21. Baraliakos X, Landewe R. Hermann KG er al. Inflammation in ankylosing spondylitis: a systematic description of the extent and frequency of acute spinal changes using magnetic resonance imaging. Ann Rheum Dis 2005; 64:730-34. 22. Braun ]. Baraliakos X, Golder W er al. Analysing chronic spinal changes in ankylosing spondylitis: a systematic comparison of conventional x rays with magnetic resonance imaging using established and new scoring systems. Ann Rheum Dis 2004; 63 :1046-55. 23 . Braun], Baraliakos X, Golder W et al. Magnetic resonance imaging examinations of the spine in patients with ankylosing spondylitis. before and atter successful therapy with infliximab: evaluation of a new scoring system. Arthritis Rheum 2003; 48:1126-36. 24. Heufi-Dorenbosch L. Landewe R, Weijers Ret al. Combining information obtained from magnetic resonance imaging and conventional radiographs to detect sacroiliitis in patients with recent onset inflammatory back pain . Ann Rheum Dis 2006; 65(6) :804-8. 25 . Balint PY, Kane D. Wilson H et al. Ultrasonography of enrheseal insertions in the lower limb in spondyloarthroparhy, Ann Rheum Dis 2002; 61 :905-10. 26 . D 'Agostino MA, Said-Nahal R. Hacquard-Bouder C et al. Assessment of peripheral enthesitis in the spondyloarthropathies by ultrasonography combined with power doppler: a cross-sectional study. Arthritis Rheum 2003; 48 :523-33 . 27. Spoorenberg A, van der Heijde D, de K1erk E et al. Relative value of erythrocyte sedimentation rate and Cvreactive protein in assessment of disease activity in ankylosing spondylitis. J Rheumatol 1999; 26 :980-84. 28 . Anderson JJ, Baron G, van der Heijde D er al. Ankylosing spondylitis assessment group preliminary definition of short-term improvement in ankylosing spondylitis. Arthritis Rheum 2001 ; 44 :1876-86. 29 . Stone MA. Inman RD. Wright JG et al. Validation exercise of the ankylosing spondylitis assessment study (ASAS) group response criteria in ankylosing sp ondylitis patients treated with biologics . Arthritis Rheum 2004; 51 :316-20 . 30. Brandt], Listing]. Sieper] et al. Development and pre selection of criteria for short term improvement atter anti-TN Fa treatment in ankylosing spondylitis. Ann Rheum Dis 2004; 63 :1438-44. 31. Dougados M. Dijkmans B. Khan M et al. Conventional treatments for ankylosing spondylitis. Ann Rheum Dis 2002; 61(SuppI3) :iii40-50.
TherapyofSpondyloarthritides
145
32. Kraag G, Stokes B. Groh J et al. The effects of comprehensive home physiotherapy and supervision on patients with ankylosing spondylitis-a randomized controlled trial. J Rheumatol 1990; 17(2) :228-33. 33. van Tubergen A, Landewe R, van der Heijde D er al, Combined spa-exercise therapy is effective in patients with ankylosing spondylitis: a randomized controlled trial. Arthritis Rheum 2001; 45(5):430-8. 34. Dagfinrud H. Kvien TK. Hagen KB. The Cochrane review of physiotherapy interventions for ankylosing spondylitis. J Rheumatol 2005 ; 32(10) :1899-906. 35. Amor B. Dougados M, Mijiyawa M. Criteria of the classification of spondylarthropathies. Rev Rhum Mal Osteoartic 1990; 57(2) :85-9. 36. Dougados M, Behler JM, Jolchine I et al. Efficacy of celecoxib, a cyclooxygenase 2-specific inhibitor, in the treatment of ankylosing spondylitis : a six-week controlled study with comparison against placebo and against a conventional nonsteroidal antiinflammatory drug. Arthritis Rheum 2001 ; 44(1):180-5. 37. Wanders A. Heijde D, Landewe R et al. Nonstero idal antiinflammatory drugs reduce radiographic progression in patients with ankylosing spondylitis: a randomized clinical trial. Arthritis Rheum 2005; 52(6) :1756-65. 38. Ward MM. Kuzis S. Medication toxicity among patients with ankylosing spondylitis. Arthritis Rheum 2002; 47(3) :234-41. 39. Zochling ], Bohl-Buhler MH. Baraliakos X et al. Nonstero idal anti-inflammatory drug use in ankylosing spondylitis-a population-based survey. Clin Rheurnatol 2006; [Epub ahead of print] . 40. Dougados M. vam der Linden S. Leirisalo-Repo M et al. Sulfasalazine in the treatment of spondylarrhroparhy, A randomized. multicenter. double-blind. placebo-controlled study. Arthritis Rheum 1995; 38(5) :618-27. 41. Chen J . Liu C . Sulfasalazine for ankylosing spondyliti s. Cochrane Database Syst Rev 2005(2) :CD004800. 42. Braun J. Zochling J, Baraliakos X et at. Efficacy of sulfasalazine in patients with inflammatory back pain due to undifferentiated spond yloarthritis and early ankylosing spondylitis : a multicentre randomized controlled trial. Ann Rheum Dis 2006; 65(9) :1147-53. 43. Chen J. Liu C . Methotrexate for ankylo sing spondylitis. Cochrane Database Syst Rev 2004(3) :CD004524. 44. Gonzalez-Lopez L. Garcia-Gonzalez A. Vazquez-Dei-Mercado M et al. Efficacy of methotrexate in ankylosing spondylitis: a randomized. double blind , placebo controlled trial. J Rheumatol 2004; 31(8):1568-74. 45. Haibel H , Brandt HC, Song IH er at. Results of an open label pilot study with 20 mg methotrexate parenterally for the treatment of active anklyosing spondylitis. Ann Rheum Dis 2006 ; in press. 46. Haibel H . Rudwaleit M, Braun J. Six months open label trial of leflunomid e in active ankylosing spondylitis. Ann Rheum D is 2005 ; 64(1 ):124-6. 47. Van Denderen JC. Van der Paardt M. Nurmohamed MT er al. Double-blind. randorniscd, placebo-controlled study of leflunomide in the treatment of active ankylosing spondylitis. Ann Rheum Dis 2005 ; 64 :1761-1 764 . 48. Kalrwasser JP, Nash P. Gladman D et at. Efficacy and safety of leflunomide in the treatment of psoriatic arthritis and psoriasis: a multinational. double-blind. randomized. placebo-controlled clinical trial. Arthritis Rheum 2004 ; 50(6) :1939-50. 49. Maksymowych WP. Jhangri GS. Fitzgerald AA er al. A six-month randomized. controlled, double-blind. dose-response comparison of intravenous pamidronate (60 mg versus 10 mg) in the treatment of nonsteroidal antiinflammatory drug-refractory ankylosing spondylitis. Arthritis Rheum 2002 ; 46(3):766-73. 50. Haibel H . Brandt J. Rudwaleit M et al. Treatment of active ankylosing spondylitis with pamidronare. Rheumatolog y (Oxford) 2003 ; 42(8):1018-20. 51. Huang F. Gu J, Zhao Wet a1. One-year open-label trial of thalidomide in ankylosing spondylitis. Arthritis Rheum 2002 ; 47(3) :249-54. 52. Heiberg MS, Nordvag BY, Mikkelsen K er al. The comparative effectiveness of tumor necrosis factor-blocking agents in patients with rheumatoid arthritis and patients with ankylosing spondylitis: a six-month. longitudinal. observational. multicenter study. Arthritis Rheum 2005; 52(8) :2506-12. 53. van der Heijde D. Dijkmans B. Geusens P er al. Efficacyand safety of infliximab in patients with ankylosing spondylitis : results of a randomized. placebo-controlled trial (ASSERT) . Arthritis Rheum 2005 ; 52(2) :582-91. 54. Davis JC Jr. Van Der Heijde D, Braun J et al. Recombinant human tumor necrosis factor receptor (etanercepr) for treating ankylosing spondylitis: a randomized . controlled trial. Arthritis Rheum 2003 ; 48(11) :3230-6 . 55. Calin A, Dijkmans BA. Emery P er al. Outcomes of a mulncenrre randomised clinical trial of etanercept to treat ankylosing spondylitis. Ann Rheum Dis 2004; 63(12 ):1594-600.
146
Molecular Mechanisms ofSpondyloarthropathies
56. van der Heijde D . Kivitz A. Schiff M et al. Adalimumab therapy results in significant reduction of signs and symptoms in subjects with ankylosing spondylitis: The ATLAS Trial. Arthritis Rheum 2006, 54:2136-46. 57. Braun J, Landewe R, Hermann KG er al. Major reduction in spinal inflammation in patients with ankylosing spondylitis after treatment with infliximab: results of a multicenter, randomized, double-blind, placebo-controlled magnetic resonance imaging study. Arthritis Rheum 2006: 54(5) :1646-52 . 58. Baraliakos X , Davis J, Tsuji W. Magnetic resonance imaging examinations of the spine in patients with ankylos ing spondylitis before and after therapy with the tumor necro sis factor a receptor fusion protein etanercept. Arthritis Rheum 2005 , 52:1216 -23. 59. Haibel H , Rudwaleit M, Brandt HC et al. Adalimumab reduces spinal symptoms in active ankylosing spondylitis- clinical and magnetic resonance imaging results of a fifty-two week open label trial. Arthritis Rheum 2006, 54:678-81. 60. Sieper J, Baraliakos X, Listing J et al. Persistent reduction of spinal inflammation as assessed by magnetic resonance imaging in patients with ankylosing spondylitis after 2 yrs of treatment with the anti-tumour necrosis factor agent infliximab. Rheumatology (Oxford) 2005 ,44:1525 -30. 61. Baraliakos X, Brandt J, Listing J er al. Outcome of patients with active anklyosing spond ylitis after 2 years of therapy with etanercept-clinical and magnetic resonance imaging data. Arthritis Res Ther 2005 ,7(3):113. 62. Braun J, Baraliakos X, Brandt J et al. Persistent clinical response to the ant i-TNF-alpha antibody infllximab in patients with ankylosing spondylitis over 3 years. Rheumatology (Oxford ) 2005 : 44(5) :670 -6. 63. Davis JC, van der Heijde DM, Braun J et al. Efficacy and safety of up to 192 weeks of etanercepr therapy in patients with ankylosing spondylitis . Ann Rheum Dis 2007: [Epub ahead of print] . 64. Braun J, Baraliakos X, Listing J et al. Persistent clinical efficacy and safety of anti-TNF{alpha} therapy with infliximab in patients with ankylosing spondylitis over 5 years-evidence for different types of response . Ann Rheum Dis 2007: [Epub ahead of print]. 65. Baraliako s X , Listing J, Brandt J et al. Clinical response to discontinuation of ant i-TNF therapy in patients with ankylosing spondylitis after 3 years of continuous treatment with infliximab, Arthritis Res Ther 2005 , 7(3):R439-44. 66. Brandt J, Kaeiouzov A, Listing J et al. Six months results of a German double-blind placebo controlled clinical trial of etanercept in active ankylosing spondylitis. Arthritis Rheum 2003 , 48(6) :1667-75. 67. Baraliakos X, Listing J, Rudwaleit M et al. Radiographic progression in patients with ankylosing spondylitis after 2 years of treatment with the TNF-a antibody infliximab . Ann Rheum Dis 2005 , 64(10) :1462-6. 68. Baraliakos X , Listing J, Brandt J et al. Radiographic progres sion in patients with ankylosing spondylitis after 4 yrs of treatment with the anti-TNf-n antibody infliximab. Rheumatology (Oxford) 2007 , [Epub ahead of print]. 69. Allali F, Breban M, Porcher R et al. Increase in bone mineral density of patients with spondyloarthropathy treated with anti-tumour necrosis factor a . Ann Rheum Dis 2003, 62:347-49. 70. Brandt J, Haibel H, Reddig J et al. Successful short term treatment of severe undifferentiated spondyloarthropathy with the anti -tumor necrosis factor-alpha monoclonal antibody infliximab, J Rheurnatol 2002,29(1):118-22. 71. Brandt J, Khariouzov A, Listing J et al. Successful short term treatment of patients with severe undifferentiated spondyloarthritis with the anti-tumor necrosis factor -alpha fusion receptor protein etanercept . J Rheumatol 2004, 31(3) :531-8. 72. Feldman SR, Gordon KB, Bala M er al. Infliximab treatment results in significant improvement in the quality of life of patients with severe psoriasis: a double-blind placebo-controlled trial. Br J Derrnatol 2005: 152(5):954-60 . 73. Mease PJ, Kivitz AJ, Burch FX et al. Etanercept treatment of psoriatic arthritis: safety, efficacy and effect on disease progression. Arthritis Rheum 2004, 50(7):2264-72. 74. Mease P, Gladman D, Ritchlin C et al. Adalimumab therapy in patients with psoriatic art hritis: 24-week results of a phase III study. Arthritis Rheum 2004 : 50:4097. 75. Antoni C , Krueger GG, de Vlam K et al. Infliximab improves signs and symptoms of psoriatic arthritis: results of the IMPACT 2 trial. Ann Rheum Dis 2005 : [Epub ahead of print]. 76. Marzo-Ortega H, McGonagle D, O'Connor P er al. Efficacy of etanercept for treatment of Crohn's related spondyloarthritis but not colitis. Ann Rheum Dis 2003, 62( 1):74-6. 77. Braun J, Baraliako s X, Listing J et al. Differences in the incidence of flares or new onset of inflammatory bowel diseases in patients with ankylosing spondylitis exposed to therapy with anti-tumor necrosis factor alpha agents . Arthritis Rheum 2007, 57(4) :639-47. 78. Sandborn WJ, Hanauer SB, Katz S er al. Etanercept for active crohn's disease: a randomi zed, double-blind, placebo-controlled trial. Gastroenterology 2001 , 121(5): 1088-94.
1herapyofSpondyloarthritides
147
79. Hanauer SB. Feagan BG. Lichtenstein GR ec al. Maintenance infliximabfor crohns disease: the ACCENT I randomised trial. Lancet 2002 ; 359(9317):1541-9. 80. Rurgeerts P. Sandborn WJ. Feagan BG er aI. Infliximab for induction and maintenance therapy for ulcerative colitis. N Engl J Med 2005 ; 353(23):2462-76. 81. Braun ] . Baraliakos X. Listing J. Decreased incidence of anterior uveitis in patients with ankylosing spondylitis treated with the anti-tumor necrosis factor agents infliximab and etanercept. Arthritis Rheum 2005; 52(8) :2447-51. 82. Guignard S. Gossec L. Salliot C et al. Efficacy of tumour necrosis factor blockers in reducing uveitis flares in patients with spondylarthropathy: a retrospective study. Ann Rheum Dis 2006 ; 65(12):1631-4. Epub 2006. 83. Lim LL. Fraunfelder FW; Rosenbaum JT. Do tumor necrosis factor inhibitors cause uveitis? A registry-based study. Arthritis Rheum 2007; 56(10) :3248-52. 84. Braun J, Davis J. Dougados M et al. First update of the international ASAS consensus statement for the use of anri-rnf agents in patients with ankylosing spondylitis. Ann Rheum Dis 2005; [Epub ahead of print]. 85. Rudwaleit M. Listing]. Brandt] ec al. Prediction of a major clinical-response (BASDAI 50) to tumour necrosis factor alpha blockers in ankylosing spondylitis. Ann Rheum Dis 2004; 63(6):665-70. 86. Tan AL, Marzo-Ortega H. O'Connor P et al. Efficacy of anakinra in active ankylosing spondylitis : a clinical and magnetic resonance imaging study. Ann Rheum Dis 2004 ; 63(9):1041-5. 87. Haibel H. Rudwaleit M, Listing J et al. Open label trial of anakinra in active ankylosing spondylitis over 24 weeks. Ann Rheum Dis 2005; 64(2) :296-8. 88. Kobelr G. Andlin-Sobocki P. Brophy S et al. The burden of ankylosingspondylitis and the cost-effectiveness of treatment with inAiximab (Remicade). Rheumatology (Oxford) 2004 ; 43(9) :1158-66. 89. van der Heijde D. Han C. Devlam K et al. InAiximab improves productivity and reduces workday loss in patients with ankylosing spondylitis : Results from a randomized. placebo-controlled trial. Arthritis Rheum 2006 ; 55(4):569-574. 90. Van der Linden S. Valkenburg HA. Cats A. Evaluarion of diagnostic criteria for ankylosing spondylitis. A proposal for modification of the New York criteria. Arthritis Rheum. 1984; 27(4) :361-8 . 91. Haibel H. Rudwaleit M, Listing J er al. Efficacy of adalimumab in the treatment of axial spondylarthritis without radiographically defined sacroilitis: results of a twelve-week randomized. double-blind, placebo-controlled trial followed by an open-label extension up to week fifty-two. Arthritis Rheum . 2008; 58(7) :1981-1991. 92. van Tubergen A. van der Heijde D, Anderson Jet al. Comparison of statistically derived ASAS improvement criteria for ankylosing spondylitis with clinically relevant improvement according to an expert panel. Ann Rheum Dis. 2003 ; 62(3) :215-221.
CHAPTER
11
Genomewide Screens in Ankylosing Spondylitis Matthew A. Brown"
Abstract ffo rts to id entify gen es other th an HLA-B27 in AS have been driven by the strength o f the evidence from genetic epi de m iology stud ies indicating that HLA-B27 , altho ugh a major gene in AS, is clearly not the only sign ifica nt gene op erating. Thi s is th e case for both genetic determinants ofdi sease-susceptibility and phenotypic characteristics such as disease severity and associated disease features. In this chapter the genetic epidemiology of AS and the gen e-mapping studies performed to date will be reviewed and the future direction of research in this field di scu ssed .
E
Introduction Fo r m ore than 4 0 years. AS has been known to run in fam ilies. Familiality o ccurs bec ause of sh ared suscep tibility factors amongst family m embers. which m ay be environmental (random or common) o r genetic. Eviden ce that AS is triggered by exposure to a common environ me ntal p athogen includes: • The condition has a (nearly) worldwide distribution. indicating that whatever environmental exposure triggers the disease in a genetically susceptible person occurs from th e Arctic to Asia. • Ep id emics of AS have not been reported , in contrast to react ive arth ri ti s. anoth er B27 -associated disease, th at do es occur in common source outbreaks. • B27-transgenic rats develop arthritis only when exposed to no rmal ent eric commensal bacteria. but when maint ained in germ-free conditions.' • Twin stu diesind icate that the risk ofdeveloping AS is determined nearly ent irely by genetic differences between individuals. indicating that th e environm ental trigger is ubiquitous. The heritabilit y of AS has been assessed in two studies as >90%.2.3 Heritability of clinical manifestations of d isease is also significant. Family and twin studies indicate heritability ofwith age of symptom onset of40%, di sease activity m easured by the Bath Ankylosing Spondylitis Disease Activity Index of 5 1% (BASDAI). functional impairment due to AS measured by the Bath Ankylosing Spondylitis Functional Index (BASFI) at 7 6% and radiographic change measured by the Bath Ankylosing Spondylitis Radiographic Index (BASRI) at 62%.4.61hese stud ies involved families where all cases were B27 -po sitive and thus the heritability reflects nonB27 linked genetic effect s. Such high heritability is usu ally associated wi th monogenic d iseases and the identification of B2 7 led many inve stigators to believe that AS was essentially monogenic. Evidence from family "Matth ew A. Brown-Diamantina Institute of Ca ncer, Immun ology a nd Metabolic Med icine, Princess Alexa ndra Hospital, Ipswich Road, Woolloongabb a, Q ld, 41 02, Australia . Ema il: matt.brown @uq.ed u.au
Molecular Mechanisms of Spondyloarthropathies, edited by Carlos Lopez-L arr ea an d Roberto D iaz -Peiia. ©2009 Landes Bio science an d Springer Science+Business Media.
Genomewide Screens in Ankylosing Spondylitis
149
and twin studies suggest that at least a moderate number ofother genes were involved. Monogenic diseases tend to be quite uncommon, because even minor selection pressure/evolutionary disadvantage leads to adverse genetic effects being eliminated from the general gene pool quite rapidly. It ispossible that B27 has been maintained at fairly high frequency because it protects against some disease-for example, there is some evidence that it may partially protect against HIV infection," However, the high prevalence ofAS argues against it being a monogenic disease. Further, studies ofthe recurrence risk ofthe condition in relatives ofcases indicate that it does not segregate as a monogenic condition. The recurrence risk ofany disease depends on the genetic model. 8 Monogenic diseases reduce in frequency by halfwith each increase in distance ofrelationship to the proband. Polygenic diseases reduce in frequency by roughly the square root with each increase in distance ofrelationship to the proband. AS probably lies somewhere in between these models; B27 is almost essential for the inheritance ofAS, but its penetrance is greatly modified by other genes . Family studies in AS suggest that a moderate number of genes with significant effect are involved in AS, the best fitting models suggesting 5 genes, but with a wide range ofpossible fitting models ranging from 3-9 genes Involved." It is likely that in addition to B27 there will be a small number of genes with moderate effect and a large number of genes with small effect. Certainly, large multicase families, typical ofdiseases with monogenic aetiology, are extremely rare with AS. The findings ofgenomewide association studies reported below support the findings of these segregation analyses.
Methods for Identifying Disease-Causing Genes To a great extent, genetic research has been technologically driven. In the 1980's,the discovery ofPCR and the development ofrestriction fragment length polymorphism approaches to linkage mapping in families enabled a genetics revolution which led to the identification of most genes involved in monogenic diseases. The success in monogenic diseases led to the suggestion that the same approach would work for common, complex, heritable diseases.'? Unfortunately, it has to be acknowledged that family-based approaches have made little progress in common heritable diseases, including AS. This was partly because the technological leap required when this approach was first tried was too great. With no human genome map in place, to identify even the major genes in the linked region required an extensive and expensive mapping and sequencing project. We now also understand that the whole concept ofwhat constituted a gene, that being that it was DNA sequence that encoded a protein, was simplistic. Much ofthe genome is actually transcribed but not translated and is not what was thought to be 'junk DNA: having an increasing range ofknown functions. Genotyping error rates were too high ; error rates> 1% were common, whereas rates >0.1% are now considered insufficiently stringent. Genotypingwas inefficient with costs per genotype exceeding $US 1, whereas nowadays they are less than a percent of that. We had little understanding of the complexity of genetic variation within populations, or methods to identify or control for it . Secondly, the genetic effects involved were mostly too small to be detected by family-based linkage approaches. Studies ofa few hundred families were at the time considered large, but actually had inadequate power to find genes contributing <10% ofthe risk to common diseases. Very few success stories in polygenic diseases exist, but there have been some notable exceptions. These include CARDlS/NOD2, the major susceptibility gene for Crohn's disease ;" sequestosome mutations causing Paget's disease ;12 LRPS, a major gene for oseeoporosis'<" and a handful of others. These examples all highlight the potential of genetics research, providing breakthroughs in the diseases concerned that decades ofhypothesis-driven research had failed to produce. Most genes , though, were below the sensitivity of the linkage approaches. The field has been revolutionized in recent times by technological, bioinformatic and computational advances. Microarray genotyping technologies have enabled high throughput genotyping at <1p/genotype with <0.1% error rates. Studies involving billions ofgenotypes are now commonplace, whereas linkage studies typically involved fewer than a million genotypes. This has removed the need to focus on families (which required a lower marker density) and enabled ultra-high density
150
Mole cular Mechanisms ofSpondyloarthropathies
studies across the entire genome of apparently unrelated cases and controls. The human genome map and International HapMap project provided public resources that enabled intelligent design ofmapping projects. This greatly accelerated the leap from low to high resolution genetic studies and led to improved understanding ofthe complexity ofpopulation structure ofgenetic variation, arising from population migrations and from selection pressures such as infection and malnutrition. Statistical advances enabled this structure to be identified and controlled for. Initially this consisted ofwithin-family designs such as the parent-case trio transmission-disequilibrium association tests, but these have now been extended to unrelated case-control data. These frame-shift changes have brought us to the point where we can realistically survey >80% of the genetic variation in human populations to determine its involvement in common diseases and to a surge in successful identification of human disease genes. Such studies involve genotyping >300,000 SNPs selected across the genome according to the extent ofvariation observed in populations, in thousands ofunrelated casesand controls. These 'genome-wide association studies' thus involve billions ofgenotypes, creating unprecedented analytical challenges and also unprecedented success. For example, the recently completed Wellcome Trust Case-Control Consortium (WTCCC) study of7 common diseases, including rheumatoid arthritis, has identified 25 genes with significant effects definitely associated with disease and a further 58 which have high probability of being involved, many times the number of genes that had been hitherto successfully idendfied.v There had been a concern that ifgenes involved in disease-susceptibility harboured multiple variants associated with disease, then genomewide association studies would not be able to identify them. The WTCCC study shows that for many important disease-susceptibility genes, this is not the case and the approach works. The hypothesis-driven researchers now have a solid, priceless, foundation ofevidence as to the aetiopathogenesis ofthese diseases upon which to base their research henceforth. Because of the public funding of these studies, the data are available to all bonafide researchers free of charge (https :llwww.wtccc .org.uk/). Similar databases have been established in North America with a range ofdata available from cases and healthy controls across a range of diseases and it is to be hoped that data from genomewide association studies in rheumatic diseases will also be deposited there .
Linkage Studies and Ankylosing Spondylitis Three groups have published linkage studies in ankylosing spondylitis;17-20each ofthese studies was underpowered to detect the likely genetic effect sizes we now know operate in the disease. To improve the statistical power of the linkage data , the genotypes from these studies have been combined in formal meta-analysis to include data from 488 multiply affected families involving 589 affected sibling pairs." Even this data set has modest power to detect genes ofsmall to moderate effect in AS and only one region, the MHC, achieved 'sign ificant ' linkage ie linkage expect to occur less frequently by chance in 1/20 genomewide linkage studies. In addition, regions on chromosome 1Oqand 16q achieved 'suggestive' evidence oflinkage (likely to occur once by chance per genome screen) and regions on chromosomes Iq, 3q, 5q, 6q, 9q , 17q and 19q showed at least nominal linkage in 2 or more scans. Two regions where association had previously been reported (chromosome 2, the IL-l gene cluster;22-25 chromosome 22, CYP2D6 26.27 ) showed nominal linkage. The linkages observed were quite broad and th ere is a risk that any particular linkage might be a false positive, given the moderate evidence for linkage. When these scans were performed, it had been planned to follow them up by systematic screens oflinked regions by high density association mapping, but the development ofgenomewide association studies has made this approach obsolete. Nonetheless, these linkage scans remain helpful in prioritization ofgenes for further study, either for candidate gene association studies or following genomewide association studies. It is possible that genomewide association studies will not detect genes that have multiple variants (i.e., that do not fit the 'common-variant , common-disease' hypothesis) and that these will be more easily identified by linkage studies. However the general record in common human diseases suggests that in the absence ofa major gene effect, the power oflinkage studies is too low to robustly identify regions containingdisease causinggenes. Further linkage studies should therefore probably
Genomeiaide Screens in Ankylosing Spondylitis
151
be restricted to studies of families with prior significant evidence of a monogenic effect, such as large families with multiple affected cases. Even in such families it is likely that a high prevalence ofHLA-B27 amongst the founders will eventually be demonstrated to be the explanation for the observed segregation in the family. Advances in sequencing technology are bringing to the forefront the possibility ofwhole genome resequencing in the near future. This approach is likely to be the method ofchoice for identifying genes with multiple disease-associated variants and rare variants with major effects, rather than linkage studies.
High Density Association Studies and Ankylosing Spondylitis No genome-wide association study has yet been completed in AS. The WTCCC AS study, involving genotypingof14,436 nonsynonymous (coding) and 897 MHC SNPs, in 1000 casesand 1500 controls, is by far the largest study to date in AS genetics, but nonetheless has only investigated a small proportion ofthe overall genetic diversity in relation to the d isease." The findings of the MHC component ofthis study are discussed below. The AS component ofthis study has identified two major genes involved in the disease which represent major breakthroughs in our understanding ofthe aetiology ofAS and which will stimulate new research efforts in the disease. These two genes areARTSI and IL23R, for both ofwhich strong, independently replicated association has been identified. These two genes respectively are responsible for 26% and 9% of the population attributable risk ofAS; by contrast the population attributable risk of PTPN22 in rheumatoid arthritis and Type 1 diabetes mellitus is 8%.16 IL23R encodes a critical cytokine receptor in the T H 17 subset of T-cells . T H 17 cells were originally identified as a distinct subset ofT-cells expressing high levels of the proinflammarory cytokine IL-17 in response to stim ulation." This subset ofeffector T -cells was identified by immunologists researching paradoxical effects ofInterferon-y blockade in the mouse models extrinsic autoimmune encephalomyeliris'? and collagen-induced arthritis," models most closely resembling multiple sclerosis and rheumatoid arthritis respectively. In the past 12 months, genetic variation in IL23R has been demonstrated to substantially affect susceptibility to Crohns disease,32.33 psorias is'? and AS.35 It is known that there is a close clinical association between inflammatory bowel disease (IBD) and AS. Up to 70% ofAS patients have subclinical terminal ileal inflammation resembling Crohn's disease," -10% ofAS patients have clinical IBD,9and 3-25% ofIBD cases have AS.36-40An Icelandic genetic database study demonstrated that first-degree relatives ofAS patients were 3.7 and 2.9 times more likely to develop CD or UC respectively than the general population." Increased familiality may occur due to either shared genetic or environmental risk factors, but the fact that spouses ofAS patients were not at increased risk ofCD or U C suggests that shared environmental risk is not a major contributor to the familiality reported with first-degree relatives. IL23R is not only a major gene for AS , but also at least partially explains the clinical association ofAS with Crohns disease and psoriasis. The association in AS wasstronger in patients lacking clinical IBD, indicating that the genetic effects independently influence AS-susceptibility. Th is is thus the first seronegative pleiotropic gene (affecting more than one disease), as the previously identified major genes for these conditions such as CARD 15 for Crohns disease, HLA-B27 for AS and HLA-Cw6 for psoriasis, are disease-specific. Although further research is required to determine the primary associated variant and whether the variants involved are gain- or loss-of-function, the therapeutic potential of this finding is immediately apparent. The strong association with susceptibility to disease suggests that IL23/IL23R targeted therapies may even be capable ofdisease prevention, a particularly exciting potential. The association of the gene ARTSI and disease equally represents a major breakthrough in research ofthe aetiology ofAS. The gene has two known functions, either ofwhich may explain its association with AS. Within the endoplasmic reticulum, ARTS 1 is involved in trimming peptides to the optimal length for MHC Class I presentaeion.P'? AS is primarily an HLA-Class I mediated autoimmune disease, with >90% ofcases carrying the HLA-B27 allele. How B27 causes AS is unknown, but if the mechanism ofassociation ofARTSI with the disease is through effects on peptide pre sentation, this would have significant implications for research into the mechanisms
152
Molecular Me chanisms ofSpondyloarth ropathies
ofassociation ofthe B27 with AS, substant ially favouring mechanisms that involve abnormalities of peptide presentation rather than those involving abnormal form s of B27. The second known function ofARTS 1 is that it cleavescell surface receptors for the proinflarnmarory cytokines IL-l (IL -lR2),44 IL -6 (IL-6Ra)45and TNF (TNFRl ),46thereby down-regulating their signaling. Loss of functionARTSl variants could thus have proinfiammarory effects through this mechanism. Further research will be required as co the mechanism, but one clue in this research is that no evidence was seen for association with ARTS1 in a large stu dy of ulcerative colitis (n = 10 II) and Crohns disease (n = 755) cases compared with healthy controls (n = 633). Neither of these conditions is HLA Class I mediated, whereas up-regulation of pro inflammatory cytokines such as IL-I, IL-6 and TNF is a feature of these conditions. This suggests that it is the peptide trimming function ofARTS I that explains the mechanism ofthe association with AS, though of course much more direct evidence will be required to confirm this theory. These two genes are of sufficient magnitude of effect size that they are predicted to be of diagnostic utility. Figure I presents the posterior probability ofa diagnosis ofAS, given the prior probability ofthe disease and the findings ofgenotypingARTS I and IL23R variants. In a patient with a pretest probability ofAS ofI 0% (for example, a randomly selected IBD case), ifthe patient is B27 / ARTS 1/IL23R positive, then the predicted posttest probability ofAS is 89%, compared with 50% ifonly B27 is tested . If these tests are both negative, the predicted posttest probability of AS is 0.3%. These calculations , based on confirmation study findings by the TASC group." require evaluation in clinical settings before they can be adopted, but are strongly suggestive that limited additional genetic testing using currently known associations is already ofclinical relevance. Other genes for which association with AS has been confirmed include the IL-I gene complex and CYP2D6. 26:J.7 Several reports of association with IL-I gene complex members have been made.22.25.47.48 Like the major histocompatibility complex (MHC), this is a very complex locus to study, being subject to strong evolutionary pressure and with marked preservation of genetic sequence in haplorypes , making it very difficult to identify true di sease-associated variants.The most consistent associationshavebeen with IL1A variants,which are alsoassociatedwith psoriatic arthritis."
100 90 . 80
/
/
."
---
___ -
-
-
--=:-::";;;;;;;;;;;0-"1'
,,, ,,, ,
(
, , ,
70 - ,
60
,
50 ·, 40 ·'
1
30 "
I
20
/
o .........r:=::::;:::i~ o 20
40
60
-
B27 P(D+fT+)
- - B27/ARTS1/1L23R P(D+fT-) - - B27/ARTS1/1L23R ,_ _ _..:,P_.=D_+!!!L. _ __._..
'"
r
10
I
-B27 P(D+fT-)
80
100
PRE-TEST PROBABILITY Figure 1. Post test probability of AS given pretest probability fo r different B27/ARTS1/1L23R genoty pes. P(D+/T +) indicates the poster ior probability of disease if the genotype test is positive; P(D+/T-) indicates the posterior probability of disease if the genotyp e test is negative.
Genomewide Screens in Ankylosing Spondylitis
153
A recent prospective meta-analysis ofthe AS-associated IL-l variants performed in 2675 AS cases and 2592 healthy controls recruited in 12 different centres in 10 countries confirmed association ofIL-IA variants with AS and weaker association ofother IL-l complex polymorphisms in individual datasets.? " Strong association was observed with three SNPs in the ILIA gene (rs2856836, rs17561, rs1894399, P = 0.0036, 0.000019, 0.0003 respectively). The population attributable risk fraction ofthese variants in Caucasians is estimated at 4-6%. This indicates that it is unlikely that the ILIA associations previously reported are false positives and provides a model for a form of study that has the required power to replicate the magnitude ofgenetic effects likely to operate in AS. The relative inefficacy ofIL-l inhibition with anakinra has unfortunately discouraged further research ofIL-l as a potential therapeutic target in AS, despite these genetic findings. Unconfirmed reports of association have been published with several other candidate genes includingANKH, 51,52 TLR4, 53-56 CARDISS7 -65 and KIR genes with AS.66 Of these, TLR4 and CARDIS were investigated in the WTCCC study and neither showed any association with disease (P> 0.1) and thus are very unlikely to be truly associated with the condition. Insufficient markers were studied in the other genes , which await further studies to determine their true significance.
MHC Genetics ofAnkylosing Spondylitis The MH C , situated on chromosome 6 (6p21.3), extends over 3.6Mb and contains about 220 genes, many of which have immunoregulatory functions. There is compelling evidence that the MHC contains several other nonB27 determinants ofdisease-susceptibility, including the HLA-B allele, HLA-B60 and nonHLA-B genes. The association ofHLA-B60 with AS is much weaker than the association with B27, with an odds ratio of3.6.67It is uncertain as to whether HLA-B60 is also disease-causing itself, or a marker ofan MHC haplotype bearing other disease causing genes. The association ofHLA-B60 with disease is well established in B27 -positive cases67 ,68 and there is data suggesting a role in B27 -negative AS. 67.69 To identify other MHC genes involved in AS, investigators have studied association ofother MHC Class II and III genes with disease. We have reported association ofHLA-D RB 1*01 with AS in a B27 -matched case-control study" and in twins." Carriage ofHLA-DRBl*08 in B27 -positive A S cases is associated with younger age of symptom onset and the occurrence of uveitis ,?l,n and weakly with susceptibility to AS,?osuggesting that nonB27 MHC gene s may also influence severity and clinical manifestations of AS. Several small association studies have implicated other MHC gene s in A S, although the studies have been too small and targeted to determine whether these are primary associations or due to linkage disequilibrium with other loci (reviewed in ref. 73). A recent study investigating the protective effect of HLA-B*2709 from AS in Sardinians has suggested that this may be due to linkage disequilibrium with a nonsynonymous SNP in HLA-E (rsI264457, P = 6 x 10-6) , which encodes a R128G substitution." Strangely, comparing B*2705 and B*2709 haplotypes for thi s SNP, no significant difference is noted, so it is hard to see how it can explain the protective effect of B*2709, unless B*2705 Sardinians have different MHC haplotypes to those seen elsewhere in Italy and Europe. This data is interesting but as it involves only a small sample size and for some control sets genotype data was available from fewer than half of the cohort, clearly further studies are required. Controlling for linkage disequilibrium acros s the MH C is also challenging and in many studies not properly handled. Analyses that involve imputation of haplorypes in unrelated individuals should determine the posterior probability of correct haplotype assignment. Many studies use the EM algorithm for haplotype imputation; this algorithm depends on the presence of significant linkage disequilibrium between markers for accuracy. In many implementations of the algorithm, imputed haplotype outputs can even be produced for markers where no linkage disequilibrium exists between them (e.g., even on separate chromosomes) which is obviously nonsense. Clearly analysis ofthis region, which is characterised by extreme allelic diversity yet major haplotypic preservation, requires a sophisticated and skilled analysis to avoid erroneous findings. The WTCCC AS study genotyped 897 MHC SNPs in 1000 AS cases and 1500 healthy controls. These studies showed extremely strong and broad association of the MHC with AS,
154
Molecular M echanism s ofSpondyloarthropathies
CHROMOSOME 6
50 45 40
35
~ ro
30
~ 25 ,9; ~ 20
I
15 10
5
o 30
31
31
32
32
33
33
34
34
Distance (Mb)
Figure 2. MHC findings from WTCCC AS study. The y-axis is -logt p-values) and x-axis distance fro m the p-telomere of chromosome 6 (March 20 0 6 hum an reference sequence (NCBI Build 36 .1 )). HLA-B lies at 31.4Mb a nd HLA-DRB1 at 32.7 Mb. with association with p-values <10->0 pre sent from 30.9 Mb to 32.5 Mb from th e p -relomere of chromosome 6 (see Fig. 2) . As th e cont rols in this anal ysis are not matched for HLA-B27 with th e cases, th is association probably reflect s both linkage disequilibrium with HLA-B27 and the pre sence of nonB27 MHC associated genes. To test for the pre sence of nonB27 MHC associations of AS, we recently completed a stu dy of B27-matched MHC haplorypes in cases and controls." The program 'PH ASE' was used to impute haplotypes and only haplotypes with >90 % posterior probability of correct assignment were carried forward in analysis. Comparing B27-rnarched case and control haplotypes, strong association was observed with DRBI irrespective of whether the haplotype carried HLA-B27 (B27-positive stran d P = 4 X 10-4, B27 -negative strand P = 5 X 10-8) . Specific MH C haplotypes were demonstrated to have highly Sign ificant association with AS, controlling fully for the carriage of B27. The effect size of these associations is substantial. The p opulation attributable fractions from spe cific MHC haplotypes range from 16.5% for the B27-/DR1+ haplotype, down to 3.5% for the B27 + / D RB 1*07;76 assuming an additive model, the attributable risk from these haplotypes is 34%. Thi s strongly suggests that further stu dies of the MHC for AS- susceptibility genes other than B27 are likely to be quite fruitful.
Future Studies In an ongoing study, a conso rtium of Australian, Briti sh and North Am erican investigators (the TA SC group) is performing a genom ewide association stu dy in AS. It is expected that thi s will be completed early in 200 8 and will involve 2000 AS cases and >6000 health y cont rols. The record of pre viou s genom ewide associati on stu dies suggests that thi s scan is likely to produce a small number ofclear-cut hits (low-hanging fru it ) and a moderately large number ofintermediate streng th associations (higher-hanging fru it ) wh ich will require further studies to determine th eir true sign ifican ce. Thi s scan will also be underpo wered to dete ct gen es invo lved in disease severity,
Genomewide Screens in Ankylosing Spondylitis
155
although arms ofthe consortium are investigatinggenetic determinants ofclinical and radiographic severity.Thus there will remain a need for scanning further cases, particularly those characterised for 'disease-severity' measures and AS-associated features, such as uveitis. Other areas which will require further research include the role ofcopy number variation and methylation patterns and the genetic determinants of the disease in other populations. There is a notable paucity of nonMHC genetic data about both AS and rheumatoid arthritis in Asian populations, despite clear evidence in the case of rheumatoid arthritis at least that the genetic determinants are different in Asian ethnic groups. Thus there will be a need for further studies involving even more cases. Studies in different ethnic groups may also be required to identify genes specific to those ethnic groups and to assist with defining the true associated variants in involved genes. Successful follow-up studies to genomewide association studies in other common diseases have required the study of thousands ofcases to produce convincingly clear cut findings, generally requiring the collaboration ofmany research and clinical groups. Thus whilst the TASC study will provide an unprecedented view of the genetic determinants ofAS susceptibility and clinical manifestations, it should be considered the foundation of future research programs into AS aetiopathogenesis rather than answering all the questions itself. Unfortunately the record in AS genetics and in other diseases is that the field has been considerably muddied by publication of numerous follow-up studies with inadequate sample sizes. Small cohorts can be ofuse when combined in prospective meta -analysis, such as in the meta-analysis of IL-l variants discussed above.
Conclusions The successful identification of ARTS] and IL23R should give those involved in AS genetics research great encouragement of the potential of this research. Hypothesis-free genetics research can identify the genes which are the main determinants ofwho develops AS and its clinical manifestations . The early experience in genomewide association studies indicates that some diseases are much more tractable to the approach than others. This relates to issues such as heritability, phenocopies and diagnostic uncertainty and the underlying genetic model. AS appears, like Crohns disease, to be highly tractable to investigation by this approach. The challenge will then be for the hypothesis-driven researchers to work out why these genes are associated with AS and what to do about it. Each additional piece of the jigsaw that genetic studies turns up brings us a little closer to cracking the aetiopathogenesis ofthe disease and some pieces,such as the IL2 3R finding, point directly to treatments. The findings we already have are ofdiagnostic utility. Given the rapid pace of advance of th is field over the last two years, the future of this field looks extremely promising.
Acknowledgements MAB is funded by an NHMRC (Australia) Principal Research Fellowship . Research presented in this review was largely funded by the Arthritis Research Campaign (U.K.),Wellcome Trust and National Institute ofArthritis, Musculoskeletal and Skin Diseases (USA).
References 1. Taurog JD, Richardson JA, Crofi JT et al. The germfree state prevents development of gut and joint
inflammatory disease in HLA-B27 transgenic rats. J Exp Med 1994; 180(6) :2359-64. 2. Brown MA, Kennedy LG, MacGregor AJ et al. Susceptibility to ankylosing spondylit is in twins : the role of genes, HLA and the environment. Arthritis Rheum 1997; 40(10) :1823-8. 3. Pedersen 0, Svendsen A, Ejstrup L er al. Heritabilit y estimates on ankylosing spondylitis. Clin Exp Rheumaro] 2006 ; 24(4) :463. 4. Brown MA, Brophy S, Bradbury L et al. Identification of major loci controlling clinical manifestations of ankylosing spondylitis. Arthritis Rheum 2003; 48(8) :2234-9. 5. Hamersma J, Cardon LR, Bradbury L et al. Is disease severity in ankylosing spondylitis genetically determined? Arthritis Rheum 2001 ; 44(6 ):1396-400. 6. Brophy S, Hickey S, Menon A et al. Concordance of disease severity among family members with ankylosing spondylitis? J Rheurnatol 2004; 31(9):1775 -8. 7. Rowland-Jones S, Colbert RA, Dong T et al. Distinct recognition of closely-related HIV-l and HIV-2 cytotoxic T-cell epiropes presented by HLA-B'2703 and B'2705 . AIDS 1998; 12(11) :1391-3.
156
Molecular Mechanisms ofSpondyloarthropathies
8. Risch N . Linkage strategies for genetically complex traits 1. Multilocus models. Am J Hum Genet 1990; 46(2):222-8. 9. Brown MA, Laval SH, Brophy S et aL Recurrence risk modelling of the genetic susceptibility to ankylosing spondylitis. Ann Rheum Dis 2000; 59( 11):883-6. 10. Lander ES, Botstein D. Strategies for studying heterogeneous genetic traits in humans by using a linkage map of restriction fragment length polymorphisms. Proc Nad Acad Sci USA 1986; 83(19):7353-7. 11. Hugor JP, Chamaillard M, Zouali H er aL Association of NOD2 leucine-rich repeat variants with susceptibility to Crohn's disease. Nature 2001; 411(6837):599-603. 12. Laurin N , Brown JP, Morissette J et aL Recurrent mutation of the gene encoding sequestosome 1 (SQSTMl/p62) in Paget disease of bone. Am J Hum Genet 2002; 70(6) :1582-8 . 13. Ferrari SL, Deutsch S. Choudhury U ct aL Polymorph isms in the low-density lipoprotein receptor-related protein 5 (LRP5) gene are associated with variation in vertebral bone mass. vertebral bone size and stature in whites. Am J Hum Genet 2004 ; 74(5):866-75. 14. Koay M , Woon P-¥, Zhang Y et aL Influence of LRP5 polymorphisms on normal variation in BMD . Journal of Bone and Mineral Research 2004; 19(10):1619-1627. 15. Little R, Carulli J, Del Mastro R et aL A mutation in the LD I. receptor-related protein 5 gene results in the autosomal dominant high-bone-mass trait. Am J Hum Genet 2002 ; 70:11-19. 16. Consortium WTCC. Genomewide association study of 14,000 cases of seven common diseases and 3000 controls. Nature 2007 ; 447:661-83. 17. Zhang G, Luo J, Bruckd Jet aL Genetic studies in familial ankylosing spondylitis susceptibility. Arthritis Rheum 2004; 50(7):2246-54. 18. Miceli-Richard C , Zouali H, Said-Nahal Ret aL Significant linkage to spondyloarthropathy on 9q31-34. Hum Mol Genet 2004; 13(15) :1641-8 . 19. Laval SH, Timms A. Edwards S et aL Whole-genome screening in ankylosing spondylitis : evidence of nonMHC genetic-susceptibility loci. Am J Hum Genet 2001; 68(4) :918-26. 20. Brown MA , Pile KD, Kennedy LG er aL A genome-wide screen for susceptibility loci in ankylosing spondylitis. Arthritis Rheum 1998; 41(4) :588-95 . 21. Carter KW; Pluzhnikov A, Timms AE er aL Combined analysis of three whole genome linkage scans for Ankylosing Spondylitis . Rheumatology (Oxford) 2007; 46(5):763-71. 22. Timms AE, Crane AM, Sims AM ct aL The interleukin 1 gene cluster contains a major susceptibility locus for ankylosing spondylitis . Am J Hum Genet 2004; 75(4) :587-95 . 23. van der Paardt M, Crusius JB, Garcia-Gonzalez MA er al. Inrerleukin-Ibeta and ineerleukin-I receptor antagonist gene polymorphisms in ankylosing spondylitis. Rheumatology (Oxford) 2002; 41(12) :1419-23. 24. McGarry F, Neilly J, Anderson N et al. A polymorphism within the interleukin 1 receptor antagonist (IL-IRa) gene is associared with ankylosing spondylitis. Rheumatology (Oxford) 2001 ; 40(12):1359-64. 25. Chou CT, Timms AE, Wei JC et al. Replication of association of 11.1 gene complex members with ankylosing spondylitis in taiwanese chinese. Ann Rheum Dis 2006; 65(8):1106-9. 26. Brown MA, Edwards S, Hoyle E et aL Polymorphisms of the CYP2D6 gene increase susceptibility to ankylosing spondylitis. Hum Mol Genet 2000; 9( 11): 1563-6. 27. Beyeler C, Armstrong M, Bird HA er aL Relationship between genotype for the cytochrome P450 CYP2D6 and susceptibility to ankylosing spondylitis and rheumatoid arthritis. Ann Rheum Dis 1996; 55(1):66-8. 28. WTCCC, TASC. A genome-wide scan of 14,000 nonsynonymous coding SNPs in 5.500 individuals : The Wellcome Trust Case Control Consortium. Nat Genet 2007; 39(11):1329-37. 29. Park H, Li Z, Yang XO et aL A distinct lineage of CD4 T-cells regulates tissue inflammation by producing interleukin 17. Nat Immuno12005; 6(11):1133-41. 30. Cua DJ, Sherlock J. Chen Y et aL Interleukin-23 rather than interleukin-12 is the critical cytokine for autoimmune inflammation of the brain. Nature 2003; 421(6924):744-8. 31. Murphy CA, Langrish CL. Chen Y et aL Divergent pro- and anti-inflammatory roles for 11.-23 and 11.-12 in joint autoimmune inflammation . J Exp Med 2003; 198(12):1951-7. 32. Duerr RH , Taylor KD, Brant SR et aL A genome-wide association study identifies IL23R as an inflammatory bowel disease gene. Science 2006; 314(5804):1461-3. 33. Tremelling M, Cummings F, Fisher SA et al. IL23R variation determines susceptibility but not disease phenotype in inflammatory bowel disease. Gastroenterology 2007; 132(5):1657-1664. 34. Cargill M, Schrodi S, Chang M er aL A large-scale genetic association study confirms ILl2B and leads to the identification of IL23R as psoriasis-risk genes. Am J Hum Genet 2007 ; 80(2):273-290. 35. Reveille JD, Zhou X. McGinnis Ret aL lnterleukin-23 receptor polymorphisms are a major determinant of susceptibility to ankylosing spondylitis. Nat Genet 2007; Submitted.
Genomewide Screens in Ankylosing Spondylitis
157
36. de Vlam K, Mielants H, Cuvelier C et al. Spondyloarthropathy is underestimated in inflammatory bowel disease: prevalence and HLA association. J Rheumatol 2000; 27( 12):2860-5. 37. Palm 0, Mourn B, Ongre A et al. Prevalence of ankylosing spondylitis and other spondyloarthroparhies among patients with inflammatory bowel disease: a population study (the IBSEN study). J Rheumatol 2002; 29(3):511-5. 38. Salvarani C , Vlachonikolis IG, van der Heijde DM et al. Musculoskeletal manifestations in a population-based cohort of inflammatory bowel disease patients. Scand J Gastroenterol 2001; 36(12):1307-13. 39. Scarpa R, del Puente A, D'Arienzo A er aI. The arthritis of ulcerative colitis: clinical and genetic aspects . J Rheumatol 1992; 19(3) :373-7. 40. Steer S, Jones H, Hibbert J et al. Low back pain, sacroiliitis and the relationship with HLA-B27 in Crohn's disease. J Rheurnatol 2003; 30(3) :518-22. 41. Thjodleifsson B, Geirsson AJ, Bjornsson S et al. A common genetic background for inflammatory bowel disease and ankylosing spondylitis: a genealogic study in Iceland. Arthritis Rheum 2007; 56(8):2633-9. 42. Hammer GE, Gonzalez F, Champsaur M er aI. The aminopeptidase ERAAP shapes the peptide repertoire displayed by major histocompatibility complex class I molecules. Nat Immunol 2006; 7(1) :103-12. 43. Kanaseki T, Blanchard N, Hammer GE er al. ERAAP synergizes with MHC class I molecules to make the final Cut in the antigenic peptide precursors in the endoplasmic reticulum. Immunity 2006; 25(5):795-806. 44. Cui X, Rouhani FN, Hawari F et al. Shedding of the type II IL-l decoy receptor requires a multifunctional aminopeptidase, aminopeptidase regulator of TNF receptor type 1 shedding. J Immunol 2003; 171(12):6814-9. 45. Cui X, Rouhani FN, Hawari F er al. An aminopeptidase, ARTS-I, is required for interleukin-6 receptor shedding. J Bioi Chern 2003; 278(31):28677-85. 46. Cui X, Hawari F, Alsaary S et al. Identification of ARTS-l as a novel TNFRl-binding protein that promotes TNFRI ecrodomain shedding. J Clin Invest 2002 ; 110(4) :515-26. 47. Vazquez-Dei MM, Garcia-Gonzalez A, Munoz-Valle JF er al. Inrerleukin Ibeta (IL-lbeta), IL-I0, rumor necrosis factor-alpha and cellular proliferation index in peripheral blood mononuclear cells in patients with ankylosing spondylitis. J Rheumatol 2002; 29(3):522-6. 48. Maksymowych WP, Rahman P, Reeve JP er al. Association of the III gene cluster with susceptibility to ankylosing spondylitis : an analy sis of three canadian populations. Arthritis Rheum 2006; 54(3) :974-85. 49. Rahman P, Sun S, Peddle L er al. Association between the inrerleukin-I family gene cluster and psoriatic arthritis. Arthritis Rheum 2006 ; 54(7) :2321-5. 50. Sims A-M, Timms A, Bruges Armas Jet al. Prospective meta-analysis of IL-l gene complex polymorph isms confirms associations with ankylosing spondylitis . Ann Rheum Dis 2007; Submitted. 51. Timms AE, Zhang Y, Bradbury L er al. Investigation of the role of ANKH in ankylosing spondylitis. Arthritis Rheum 2003 ; 48(10):2898-902. 52. Tsui FW; Tsui HW; Cheng EY et al. Novel genetic markers in the 5'-flanking region of ANKH are associated with ankylosing spondylitis. Arthritis Rheum 2003; 48(3):791-7. 53. Adam R, Sturrock RD, Gracie JA . TLR4 mutations (Asp299Gly and Thr399lle) are not associated with ankylosing spondylitis. Ann Rheum Dis 2006 ; 65(8):1099-101. 54. Gergely P jr, Blazsek A, Weiszhar Z et al. Lack of genetic association of the Toll-like receptor 4 (TLR4) Asp299Gly and Thr399lle polymorphisms with spondylarthropathies in a Hungarian population. Rheumatology (Oxford) 2006; 45(10):1194-6. 55. Snelgrove T, Lim S, Greenwood C et al. Association of toll-like receptor 4 variants and ankylosing spondylitis: a case-control study. J Rheurnarol 2007 ; 34(2):368-70. 56. van der Paardt M, Crusius JB, de Koning MH et al, No evidence for involvement of the Toll-like receptor 4 (TLR4) A896G and CD14-C260T polymorphisms in susceptibility to ankylosing spondylitis. Ann Rheum Dis 2005 ; 64(2):235-8. 57. Crane AM, Bradbury L, van Heel DA et al. Role ofNOD2 variants in spondylarthritis. Arthritis Rheum 2002 ; 46(6) :1629-33. 58. Miceli-Richard C, Zouali H, Lesage S et al. CARDI5/NOD2 analyses in spondylarchroparhy, Arthritis Rheum 2002; 46(5):1405-6. 59. van der Paardt M, Crusius JB, de Koning MH et aI. CARD15 gene mutations are not associated with ankylosing spondylitis. Genes Immun 2003; 4(1) :77-8. 60. van Heel DA, McGovern DP, Cardon LR et al. Fine mapping of the IBD 1 locus did not identify Crohn disease-associated NOD2 variants: implications for complex disease genetics. Am J Med Genet 2002; 111(3):253-9.
158
Molecular Mechanisms ofSpondyloarthropathies
61. van Heel DA , Dechairo BM, Dawson G er al. The IBD6 Crohn's disease locus demonstrates complex interactions with CARD15 and IBD5 disease-associated vari ants. Hum Mol Genet 2003: 12(20):2569-75. 62. D'Amato M . The Crohn's associated NOD2 3020InsC framesh i/i mutation does not confer susceptibility to ankylosing spondylitis. J Rheumatol 2002 : 29( 11):2470-1. 63. Ferreiros-Vidal I. Amarelo J. Barros F et al. Lack of association of ankylosing spondylitis with the most common NOD2 susceptibility alleles to Crohn's disease. J Rheumatol 2003 : 30(1):102-4. 64. Peeters H . Vander Ccuyssen B, Laukens D et al. Radiological sacroiliitis, a hallmark of spondylitis, is linked with CARD 15 gene polymorphisms in patients with Crohn's disease. Ann Rheum Dis 2004: 63(9):1131-4. 65. Kim TH. Rahman P.Jun JB et al. Analysis of CARD 15 polymorphisms in Korean patients with ankylosing spondylitis reveals absence of common variants seen in western populations. J Rheumarol 2004: 31(10) :1959-61. 66. Lopez-Larrea C, Blanco-GelazMA. Torre-Alonso JC et al. Contribution of KIR3DU /3DSI to ankylosing spondylitis in human leukocyte antigen-B27 Caucasian populations . Arthritis Res Ther 2006: 8(4):RI01. 67. Brown MA , Pile KD. Kennedy LG et al. HLA class I associations of ankylosing spondylitis in the white population in the United Kingdom . Ann Rheum Dis 1996: 55(4):268-70. 68. Robinson WP. van der Linden SM, Khan MA et al. HLA-Bw60 increases susceptibility to ankylosing spondylitis in HLA-B27+ patients. Arthritis Rheum 1989: 32(9) :1135-41. 69. Wei JC, Tsai WC. Lin HS er al. HLA-B60 and B61 are strongly associated with ankylosing spondylitis in HLA-B27-negative Taiwan Chinese patients. Rheumatology (Oxford) 2004: 43(7):839-42. 70 . Brown MA. Kennedy LG, Darke C et al. The effect of HLA-DR genes on susceptibility to and severity of ankylo sing spondylit is. Arthritis Rheum 1998; 41(3):460-5 . 71. Ploski R. Maksymowych W; Forre O. HLA-DR8 and susceptibility to acute anter ior uveitis in ankylosing spondylitis: comment on the article by Monowacul Islam er al. Arthritis Rheum 1996: 39(2):351-2. 72. Monowacul Islam SM. Numaga J. Fujino Y et al. HLA-DR8 and acute anterior uveitis in ankylosing spondylitis. Arthritis Rheum 1995: 38(4):547-50. 73. Sims AM, Wordsworth BP. Brown MA. Genetic susceptibility to ankylosing spondylitis. Curr Mol Med 2004; 4( I) :13-20. 74 . D'Arnaro M. Fiorillo MT, Carcassi C er al. Relevance of residue 116 of HLA-B27 in determining susceptibility to ankylosing spondylitis. Eur J Immunol1995: 25(11):3199-201. 75. Sims AM. Barnardo M, Herzberg I et al. Non B27 MHC associations of ankylosing spondylitis. Genes Immun 2007; 8(2) :115-23. 76 . Miettinen OS . Proportion of disease caused or prevented by a given exposure, trait or intervention. Am J Epidemiol 1974; 99(5) :325-32.
CHAPTER
12
Subtypes ofHLA-B27: Historyand Implications in the Pathogenesis ofAnkylosingSpondylitis John D. Reveille" and Rashmi M. Maganti
Abstract
H
LA-B27 represents a family of38 closely related cell surface proteins (encoded by the alleles HLA-B'2 70 J -39) called subtypes ofHLA-B27, mo st ofwhich have evolved from the ubiquitous HLA-B'2705 (specifically the B '27052 allele).1-2\ HLA-B27 subtypes are largelycharacterized by nucleotide substitutions (mostly nons ynonymous) in exons 2 and 3 which encode a 1and a2 domainsofthe peptide binding groove respectively. Table 1shows the description ofsequences ofHLA-B27 allele sequences. The subtypes could have arisen from B"2705 by point mutation (B"2703, B"2709 . B"2704 ). gene conversion (B"270 1. B"2702, B"2708) and reciprocal recombination(B"2707 )26B"270 6 could have arisen by interlocus gene conversion. Studies from different parts of the world reveal differences in the population distribution.
Introduction The Migrations ofHomo Sapiens and the Evolution ofHLA-B27 Subtypes Anthropologic and geneti c data strongly point to Eastern Africa as the site of origin Homo sapienssapi ens approximately 200 .000 years ago (Fig. 1).27After expanding into surrounding parts of Africa. H omo sapiens sapiens left Africa by successive migration s between 50,000 and 80 ,000 years ago, assumedly via two major routes-from the north across the Sinai peninsula, travelling th ough the middle East to Europe and Cent ral and Eastern Asia and via a southern route between th e horn of Africa and Arabia (where th e lower sea levels during periods of polar ice cap expansion made such travel po ssible). The latter group gave rise to the populations of southern India. the aboriginal groups ofSoutheast Asian and Indonesia and the Australian Aborigines.P''Ihe first group , travelling the northern route. migrated across northern and central Asia and crossed the Bering strait, which formed a land bridge due to the lower sea levels called Beringia, which gave rise to th e groups which would be event ually called Native Americans. HLA-B*270S is present nearly all populations ofthe world. It is the dominant subtype among the native populations ofeastern Siberia and North America and is present in approximately 90% of the B27 -positive individuals of northern European origin. B*2705 is further subdivided into B '270S2, B*270S3 and B*27054 by single silent nucleotide substitutions." Given its ubiquitous prevalence, it is likelyHLA-B'27052 originated, like Homo sapiens, in Africa. Further splits appear to have developed later, either through point mutation or gene conversion events and the resultant subtypes seem to have largely followed three patterns corresponding to directions of migrations from Africa (Figs. 1 and 2).1-25 "Correspond ing Author: John D. Reveille-Division of Rheumatology and Clinical Immunogenetics, University ofTexas Health Science Center, 5656 Kelly St., Houston, TX 77026, USA. Ema il:
[email protected] Molecular M echanisms ofSpondyloarthropathies, ed ited by Carlos Lop ez-Larrea and Roberto Diaz-Pefia. ©2009 Landes Bioscience and Springer Science+Business Media.
160
M olecular Mechanisms ofSpondyloarthropathies
Table 1. First description of sequences of HLA-B27 allele sequences (http://www.ebi. ac.uklimgtlhla/alle/e.html) HLA-B Allele
Year Assigned
Reference
8* 2705 8*2702 8*2707 8*2703 8*2704 8*2706 8*2707 8*2708 8*2709 8*2770 8*2777 8*2772 8*2773 8*2774 8*2775 8*2776 8*2777 8*2778 8*2779 8*2720 8*2727 8*2722 8*2723 8*2724 8*2725 8*2726 8*2727 8*2728 8*2729 8*2730 8*2737 8*2732 8*2733 8*2734 8*2735 8*2736 8*2737 8*2738 8*2739
1985 1986 1987 1987 1987 1987 1991 1993 1994 1996 1996 1997 1998 1998 1998 1999 1999 1999 1999 2000 2000 2000 (revoked 2002) 2000 2001 2001 2004 2004 2005 2005 2005 2005 2005 2006 2006 2006 2007 2007 2007 2007
1 2,3 4 5 6 7 8 9 10
http://www.ebi.ac.uk / imgt/hla/allele .html 11
12 13 14 15 16 17 18 19 20 21 http://www.ebi .ac.uk/imgt/hla/allele.html 22 23 24 25 http ://www.eb i.ac.uk/imgt/hla/all ele .html 26 http ://www.ebi. ac .uk/imgt/hla/allele.html 27 http ://www.ebi .ac.uk/imgt/hla/allele.html http ://www.ebi .ac.uk/imgt/hla/all ele .html http://www.ebi .ac.uk/imgt/hla/allele.html 28 http ://www.ebi .ac.uk/imgt/hla/allele.html 29 http ://www.ebi .ac.uk / imgt/hla/allele.html http://www.ebi .ac.uk/imgt/hla/allele.html http://www.ebi .ac.uk/imgt/hla/allele .html
Subtypes with Substitutions in the First Domain (Group 1) The first pattern is characterized by specific amino acid sub stitutions in the a l domain and a mostly uniform a2 domain. The B27 subtypes following this pattern are found largely in African, Middle Eastern and European group (Figs. 2 and 3) . HLA- B *2702 is the most frequently occurring allele ofthis group (Fig. 3), occurringprimarily in those ofMiddle Eastern (where it achieves the highest prevalence (Table 2)) , Northern African and
SubtypesofHLA-B27: Hi story and Implications in the Pathogenesis ofAnkylosing Spondylitis
161
Figure 1. The migration of modern Homo sapiens and presumed origins of HLA-B27 subtypes. The scheme outl ined above begins with a radiation from EastAfrica to the rest of Africa about 100 ,000 years ago and is fo ll ow ed by an expansion from the same area to Asia, probably by two routes, southern and northern between 60,000 and 40,000 years ago . Oceania, Europe and America were settled from Asia in that order. Listed are only geographic origins HLA-B27 subtypes of known population specificity. Those whose orig ins are not known (or originating from North American Caucasians of unspec ified ancestry, such as HLA-B*2701 , B*2710, B*2717, B*2721 , B*2728, B*2731 , B*2733, B*2734, B*2737, B*2738) are not listed.
European ancestry, being more frequent in Southern European groups. It is restricted to Caucasoids and accounts for 5-10% ofHLA-B27 subtypes in Europeans.HLA-B'2702 has also been described in Mestizo Mexicans and South American populations, probably reflecting European admixture. Its absence in subSaharan Africa, relatively high frequency in some groups and occasional occurrence in even Asian populations suggest that this subt ype is quite ancient, (perhaps occurring first soon after the initial migrations out of Africa) . It likely arose from HLA-B '2705 by a gene conversion event from other HLA-B alleles." The next most commonly occurring subtype in this group, HLA-B'2703, differs from B '2705 by only one amino acid, a substitution of histidine to tyrosine at position 59 in the u l domain. HLA-B'2703 likely evolved in West Africa, in where it occurs in up to 45% of B27 positive individuals." It is been observed outside Africa primarily among people of African descent," though ha s been reported in Lebanon, Greece, the Azores and even as far away as Taiwan (Table 2) . HLA-B'2708, although first identified in the British population.'? occurs most frequently in Asian Indian and Venezuelan controls (Table 2).38.39 It has also been reported less commonly in Turkey;" northwestern Spain" and was found to be associated with AS in a large family from the Azores, a Portuguese archipelago in the Atlantic ocean. This seemingly disparate occurrence is more difficult to explain, though the otherwise uncommon occurrence ofthis allele could be consistent with Portuguese colonization of western India in the seventeenth century. Alternatively, its widespread distribution and relationship to other rarer and more geographically discrete HLA-B27 subtypes (B'2712, B'27I8, B'2723, B'2733) suggest B'2708 may be a very ancient subtype that could have originated soon after early Homo sapiens sapiens left Africa (Fig. 1). HLA-B'2712 was first reported in a healthy family in Spain on the MHC haplotype A2-Cw2-B '2712- DRI 5-DQ6,42The presence ofthis subtype has also been reported in a Caucasian family from the British population on a haplotype containing HLA-A'29; B'27I2; Cw'1203;
162
M olecular M echanisms ofSpondyloarthropathies
HLA-B* 270S Afn calMtddle EastfEurope
MKtdi e East/Southern As ia
Eastern ASIa
Group
01 02 8 '2713
0102
Cauc.lan
0
0
8'2703
1
0
8 '2717
1 0
CaUCilSlan
8'2737
1 0
CaUca&13l"I
8'2739
1
Atncan
01 02
0
HispaniC
8 '2731
0
Caucasian
8'2701
0
Ca uc
8 '2702
0
Caucas.lan
0
CaUCAlatl
0
caucasian
8'2716 8'2708
4
8 '2726
4 0
8'2729 8 '2712
7
8 '2723
Afr
Afncan
0
Cauea&lan
0
Cauca$lan
0
caucasian
8 '2730
3
1
Caucasian
8 '2733
3
4
Unknown
Group
8 '2736
1 0
A.... n
8 '2704
1
1
lI.,.n
8 '2715
1
2
ASian
8'2708
1
3
A.... n
8 '2725
1 3
8'2721
1 4
lI.,.n caucasian
8 '2711
1 5
ASian
8'2720
1 5
lI.,.n
8'2724
1
7
ASian
8'2718
9
1
A5lOn
Group
1
Caucasian
1
Caucasian
8'2709
0
8 '2710
0
8 '2732
0
8'2738
0
1
UnknONn
8 '2726
0
2
caucasian Caucasian
caucasian
8 '2735
0
2
8'2727
0
3
Hlspamc:
8'271 4
0
3
Caucasian
MtddleEas.t
8 '2719
0
3
8'2734
0
4
8 '2707
0
caucasian We5/ A$lO
Figure 2. Subt ypes of HLA-B2 7 developed over Three Ancestral Pathways. PDai,t.l.on
10
'0
8'270502
mr rrv
.'2701 .'2702
-
. -.703 . '2704
.'270' . '"'2707 . '2708 . "270' 8 ·2710 ."2711 . '"'2712 . '2713 8 "271. . '2715 8 '"'271' ."2717 .'2718
8'271' .'2720 8'2721 8'2723 . '"'27:24 . '272& . '272. . ' 272 7 9 "2728
. ' 272' 8 ·2730 8'"'2 7 31 .'"' 273 2 . '2733
. ' 2734 8 '2735 8 '"'273'
8 '2737 8'"'2738
. '"'273'
eo
70
:m:~aw: ~
- - - - - - -y-~-
so 100 110 =nNcsn ~ or.P=u.
120 131 151 160 172 =aQ=:X: SSIlT1 ",~n.=C','nILllllrI.
A---- ---- ---- ---- -- ---------
J.---- ---- ---- ---- -- --------- - - - - - - - - - - - - - - - - - -1:-------- - - - - - - - - - - - -:>-Y-- - - -1:-------- - - - - - 5 - - - - - - --Hlf-Y--;c- - - ---- ---- - - - - - - --------- - - - - - - - - - - - --11-- - - - - - - - - - - -- ---- ------------- ---- ---- -- -1:-------- - - - - - 5 - - - - - - - - 5 - - - - - - --Hlf-Y--
-- - - - - --N--I ---- ---- ---- ---- ---- ---- -- ---------
-
-8------
-
- - - - -Dr T- -
-
- - - - - - - - - - - - - - --tf-T-- - L - - - - - - - - - - - - - - - - - -
----------------
--5---5----- - 5--11 ----
;>-- - - - - - - - - -
;>-- - - - - - - - -5- -II - ; 0 : - - - - - - - - - - - - - - - - - - - - - - - - - -
- -- ---- --5-- ---- ---- ---- ---- -- -I----T----
-
-- - - -nz
T--- - ---- ------- - - - - -- ----------- ---- ---- -- ---------
-F- - - - - - - - - - - - -iIC---- - - - - - - - - - - - -- - - - - --1-- - - - -
-- - - s -nr T-Y-S--N
---- ---- ---- -- -[--------II-II-- - - - - - - - - - - - - - - - - - - - - - - - - - - --HN-Y-- i l - -
- -- - - - - --5-- - -- - - - ll-- - - - -
-:>-Y-- -
-
-1--------(--------
-
- - -H-T-Dl T-Y-5-- - - - - - - - - - - - - - - - - - - - - - - - - - - --Hlf-Y-- ;>--
-
-- - - - - --5-- - - - - - - - - - - - - - - - - - - -l-If----L----
-
-- ----(l - - 5 - - -;0:--- - - - - - - - - - - - - - - - - - - - - - - -
- -- ---- --5-- ---- --5-- ----
-1--------
- - - - - - - - - - - - - - - - - - - - -ilIf-Y-- - - - -- - - - - - --- ------ ------- ---- ---- ---- ---- -- -----T---R- - S - Dr T - Y - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
- - - - - - - -11--1 A - - - - - I I - I I - - - - - - - - - - - - - - - - - - - - -
-
-- ----(l
-
-- - - - - - - - - - - - - - - - - - - - - -
--5-- - - - - - - - - - - - - - - - - - - - - - - - - - - -- -- - -- -----5--- ---------- -----5-- ----a;- - - -5-- ------- -----HN--- a-- - - - - - - - - - -- ---- ---- ---- --5-- ---- --11>-- a-- - - - - - - - -... - - - ---- --5-- ---- ---- ---- ---- - - ---------Y- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -- - - - - -- ------- ------- ------- ----- -------------L----- ----- -x,------- ------- -If--- - - - - - - - - - - -
Figure 3. Am ino acid sequence of human leuko cyte anti gen HLA-B2 7 subt ypes in the di versit y regions of the first and second dom ains. + demonstrating onl y the po sitions show ing amino acid var iability. **HLA-B2713 differs from other B*270 5 alleles onl y in the leader pept ide of the first domain .
Northern Europe Denmark Sout hern Europe Spain (Galici a) Northern Spain Az ores Italy Sardini a Greece Cyprus (Greeks) Turkey Lebanon Jew ish Siber ia Nor thern Indi a Western Ind ia Japan Chi nese Singapore (Chinese) Taiwan Taiwan (Han- Chinese) Taiwan (Aborigi nes) Chinese Indonesian Native Indo nesians Ma lays Thailand M aor is Brazil N orth Af rica W est Africans
Populations
0.05
7
*2701
6
10 50
32
3
12
8
7
*2703
0.5
2
30 3 34 52 30 24 48 14
10 10 20 18 7
*2702
89 87 94 100 38 6 19 42 36
66
33 34 82
3
*2704
6 6 5 64 80 50 68
90 90 80 80 91 86 65 77 50 32 43 35 38 84 61 34 18 31 2 4 6
*2705
53
72
62 89
9 7
*2706
3
2
2
6 18
13
8 17 14 30
1 7 2
*2707
3
0.02
12
5
3
*2708
3 20
*2709
Table 2. HLA-B27 subtype frequencies % in different world populations. Only first 14 subtypes are shown
0.02
*2711
1
*2713
a
<1
~
2
1&
~ ~ ::;.
<> ;"
~
~.
a-
~
~
~
~ ::;.
S
S-
"~
;;;' S-
....
;"
::'. <>
~
:::-:-
-s~
l:>..
;"
I:l
..:!
~
~
~
~
~
2
*2714
~
IS -s
164
Molecular Mechanisms ofSpondyloarthropathies
DRB1 *1301; DQBl *0603 as well as in Chinese controls." Both HLA-B*2708 and B '2712 carry the sequence that specifies the Bw6 supertypic epitope (because of the Ser77, Asn80, Arg82, Gly83 motif), in contrast to all the other known B27 subtypes which carry the Bw4 epitope at these positions. HLA-B*2712 has the closest homology with B'2708, with three mismatches in exon 2 at positions 204, 209 and 210. An association ofHLA-B'2712 with AS or related spon dyloarthritis (SpA) has not been reported. The P2 binding pocket ofHLA-B'2712 is comprised ofThr69, Asn70 and Thr71 instead ofthe usual Ala69, Lys70 and Ala71 observed in B*2705 and other disease-associated subtypes. HLA-B*2701 occurs very rarely and was initially described in a North American Caucasoid kindred associated with disease." We have also observed it in a healthy North American African-American control (] Reveille, unpublished data) and it has been reported to occur rarely in Israeli and Chinese controls,45.46 a distribution that favors an ancient development of this subtype , perhaps before man left Africa (unless a similar mutation occurred independently in different locales). It differs from other subtypes with respect to its peptide binding ability. In addition to the peptides with arginine at P2, it can also bind peptides with glutamine at this position. The substitution (D74Y) responsible for the altered selectivity ofB *2701 lies outside pocket BY The HLA-B'2713 subtype initially was identified in a healthy control family in Northern Spain." More recently is has also been described in Chinese controls48,49 and in a Brazilian patients with reactive arthritis." It is identical toB'270S2 except for a single base change (from C to A) in the coding sequence ofexon 1, which encodes the leader or signal peptide. HLA-B *2713isexpected to be disease associated, unless the change in the leader alters the expression or has other effects. HLA-B*2716, first described in Caucasians, differs from B*27052 at three amino acids at positions 69, 70 and 7l-and has no known disease association to date." Other members of this group are quite rare and probably evolved recently, characterized by a single nucleotide mutation, including HLA-B*2717, B*2737 and B*2739, which likely evolved directly from B*2705. 14 HLA-B*2717 shows one mismatch with B'27052 at position 248 (A->T), considered to be a conserved position in nearly all B alleles, which results in a phenylalanine for tyrosine subsirurion." HLA-B*2737, described in a North American Caucasian only in 2007. differs from HLA-B*27052 by a single substitution in the first diversity region of the al domain, resulting in a tyrosine substitution for a histidine at position 9. Finally, HLA-B*2739 was recently described in a Mexican individual ofunknown disease status. It differs from B*270S by a leucine substitution for a glutamine at position 72 (http://www. ebi.ac.uk/irngr/hla/allele.html). HLA-B*2723, described in an individual of Welsh Caucasian ancestry and found to have a phenotype frequency of0.0067% in Welsh blood donors, differs from B *27052by nine nucleotides which encode seven amino acid changes at positions 63 (Glu to Asn) , 67 (Cys to Phe), 69 (Ala to Thr) , 70 (Lys to Asn) , 71 (Ala to Thr) , 74 (Asp to Tyr) and 77 (Asp to Ser) in the alpha1 helix. All these substitutions are possessed by HLA-B*35 alleles suggesting that B*2723 was created by a gene conversion-like event involvingB*27052 and a B*35 allele." HLA-B*2726was described in an East African individual ofunknown disease status It is identical to HLA-B*2708 at exon 2 and exon 3, except for a difference (AAGCAG) at codon 70.21 HLA-B*2729 was described originally in a Swedish Caucasian, although is shares much ofits third diversity region of the first domain with HLA-B*2718, which was described in a Chinese individual(http://www.ebi.ac.uk/imgt/hla/allele.html). Whether these two alleles evolved from each other or independently in antiquity in unknown. HLA-B '2730 was found in a Slovene Caucasian patient with IgA nephropathy and his mother. Its exon 2 sequence is identical to that ofHLA-B*2702 and the exon 3 sequence is identical to that ofHLA-B *2719. This subtype may have arisen as a gene conversion event." HLA-B*2731, identified recently in a North American Caucasian, is most homologous to B*2726 and probably also evolved from B*2708 (http: //www.ebi.ac.uklimgt/hla/allele.html). The recently described HLA-B*2733 is most homologous to B*2708 and was described in a person ofunspecified ancestry. (http://www.ebi.ac.uk/imgt/hla/allele.html)
SubtypesofHLA-B27: History and Implications in the Pathogenesis ofAnkylosing Spondylitis
165
Subtypes with a Uniform Substitution in the First Domain and Variable Substitutions in the Second Domain (Group 2) The second group (Fig. 2) is characterized by a serine substitution for aspartic acid at position 77 in the first domain and a specific pattern of substitutions in the second, most typically a glutamic acid at position 1S2 in the second domain. This group occurs primarily in those ofEastern Asian ancestry. HLA-B*2704 is the most prevalent subtype in Eastern Asians, occurring among Koreans, Siberian Chukchis and Eskimos , Alaskan Eskimos, Inuits,52 Polynesians and Native American tribes. 53.54 It is the predominant subtype among Chinese," japanese'" and Thai 57 AS patients populations and also occasionally occurs among Indians ." It is most similar to HLA-B*27052, is likely the common ancestor ofmost of the rest in group 2 (other than B*3736). The rare and recently described HLA-B*2736 is most homologous to B*27052. differing by the ser77 substitution that defines this second subgroup and may have been a common ancestor'" or perhaps a more recent recombinant event between HLA-B"2704 and B*2705. Ofparticular note is HLA-B*2706, which is found primarily in southeastern Asia (and in fact in Malaysia represents the most common B27 subtype in controls. 59This subtype is not disease associated and differs from B *2704 by two amino acid substitutions in the second domain. It occurred probably later than HLA-B*2704 but also is ancient . Other B27 subtypes in this group are rare and probably evolved more recently. HLA-B*2711 was described initially in was identified in a healthy Japanese wornan'" but recently has also been described in Chinese concrols'" and probably resulted from a recombination between B*2704 (first domain) and B *2707 (second doma in). HLA-B*2720 was described in a Korean of unknown disease status and B*2724. which was observed in a Chinese juvenile AS patienr.f HLA-B*2711 , B*2720 and B *2724 are closelyrelated in sequence, compatible with their Eastern Asian origin . HLA-B*2715, first described in a sample of oriental origin, was found to be represented in two samples from Thais and associated with AS. B*2715 has a unique nucleotide substitution at position 163 which results in an amino acid change of the a2 domain (threonine) compared to that of the glutamic acid encoded by HLA-B*2704 and most other B27 subtypes.f' HLA-B*2721 was described in a Caucasian individual ofunspecified disease status . It is closely related to HLA-B"2706. differing only in having an arginine at pos ition 97 in the second domain instead of a serine. HLA-B*2718 was first described in a Chinese individual/" Compared with other B27 alleles, HLA-B "2718 is polymorphic at multiple sites along the pockets of the peptide binding groove predicted to make contact with the T-cell receptor and is most homologous to B *2712 in the first domain and the B*2704 family in the second domain.The amino acid sequence variation suggests that these alleles had evolved from B "2708.65 HLA-B*2725 has been described in an individual from Northern China,66 sharing the serine and glutamic acid substitions at positions 77 and 1S2 typical ofthis group, but also a unique "W" at position 1S6 and a leucine at position 16S. HLA-B*2736 was also seen in a Chinese individual, differing from B*2705 only with the serine-?? typical ofthis group.f'
Subtypes with a Uniform First Domain and Variable Substitutions in the Second Domain (Group 3) The third group, characterized by a first domain identical to HLA-B*2705 and a second domain with a variable pattern of substitutions. has a distribution primarily in the Middle East (typified by the most commonly occurring member, HLA-B*2707. which probably arose as a result of a reciprocal recombinadon'" and is most common in Middle Eastern Semitic groups but also occurs in Cyprus. Turkey, Israel and Greece 40·69.70 (HLA-B*2707has been reported also in Asian populations (Indtans.Chinese and Thais) but alwaysat a low frequency. Also, this subtype has also been reported in one healthy individual from Spain," perhaps refleceingthe medieval Arab conquest of Spain. Like the non-AS associated allotypes B *2706andB*2709, it lacks an asparginine at position
166
Molecula r M echanisms ofSpondyloarthropathies
116 and shows preference for peptides with nonpolar C-terminal residues. In Greek Cypriots this polymorphism was found only in healthy controls. Another HLA-B27 allele from this group of special interest is HLA-B'2709,n who se lack of association with AS will be discussed further below. It appears have originated in Sardinia, where it accounts for 20% of the HLA-B subtypes and probably evolved directly from
HLA-B'2705. HLA-B'2719 wasfirst described in a Lebanese patient with AS?3 It differs from HLA-B'2705 DNA sequence at five different nucleotide positions leading to three amino acid differences in the a2 domain; Thr to He at position 94 , Leu to He at position 95 and Asn to Arg at position 97. HLA-B'2719 probably evolved as a result of an inter-allelic recombination event between B'2705 and either B'0714 or (more likely) one of the B44 alleles, found at a high frequency in Lebanon. HLA-B'2710 was first reported in a Caucasian American family,?4 It differs from B'2705 by a single substitution at position 152 and has been shown to bind a similar set of peptides as B'2705'?s HLA-B'2714 was first identified in North American Indians and is reported in two Siberians. one ofwhom was a patient with AS,?6 It has also been described in India.W HLA-B '2714 differs from 8*27052 at residues 95, 97 and !03.The change at position 97 (Asn971hr) has been shown to affect the C/F pocket. However, this change appears to be conservative. HLA-B'2714 was found to be the most distinct and distant from the remaining B27 subtypes and seems to have evolved from B'27052 by gene conversion specifying the change ofthe Leu95Trp, Asn971hr and Vall03Leu. This region is also shared with B'13, B'54, '55, '56, '59 and probably represents a recombinant event . Native Siberians are of evolutionary interest, since Siberia appears to be the former homeland ofAmerindians. The first migrants bearingB'2714 could have come from the regions ofNorth or Central Siberia, arrived in Beringia and introduced this allele into the North American Indian population. This relates the native North American B '2714 alleleto the immediate ancestor shared with present-day Siberians. HLA-B'2727was described in a Mexican Hispanic individual, with the first domain like that of HLA-B'2705 and the second domain sharing the "HNQY" motifat po sitions 113-116 common to HLA-B'2707, B'2711 , B'2720 and B'2724, suggesting an Asiatic origin . (http://www. ebi.ac.uk/Imgr/hla/allele.html) HLA-B'2728 was described in a North American Caucasian individual of unknown disease status in 2005.22 Like HLA-B'2715, it has a threonine substitution at position 163 as well as a unique histidine substitution for tyrosine at position 171. It probably arose by gene conversion from other HLA-B alleles and is not related toB'2707. HLA-B'2732 was described in 2005 in a Northern European Caucasian. It shares a serine at position 97 with B'2707, B'2711, B'2733 and B'2734. HLA-B'2734, initially described in a German Caucasoid." differs from B'2707 at two amino acids at positions 114 ("N" instead of an arginine) and 116 ("D" instead of tyrosine). HLA-B*2735 was described in 2006 in a Swedish Caucasoid of unknown disease status. It is probably not related to B'2707, differeing only from B'2705 by a "N " at position 114. Finally, HLA-B'2738, described in 2007 from an individual of unknown ethnicity and disease status, differs from B'2705 only by a leucine at position 163. (http://www.ebi.ac.uk/imgt/hla/allele.html).
HLA-B27 Subtypes Associations with AS Of the HLA-B27 subtypes identified, B'2705, B*2702, B*2704 and B*2707 are definitely known to be disease associated, but the presence ofAS or related SpA has also been documented in individuals possessingB'2701 , B'2703, B'2706, B'2708, B*2710, B'2713, B'2714, B'2715, B'2719 and B*2724.'8.79 HLA-B'2706 and B'2709 were reported as being not associated with AS in Southeast Asians and Sardinians, respectively. The rest of the newly described subtypes are too rare to make any substantial statement about disease association.
SubtypestfHLA-B27: History and Implications in the Pathogenesis ofAnkylosing Spondylitis
167
The Lack ofDisease Association ofHLA-B*2706 and HLA-B*2709 Over the past decade, two subtypes, HLA-B*2706 and HLA-B*2709, have been the subject ofinterest because of their apparent lack ofassociation with AS.80 As is noted above, HLA-B*2706 is a common B27 subtype in Southeast Asia , comprising more than half of all HLA-B27 subtypes found in Indonesia, Malaysia and Thailand. Despite its high prevalence, this subtype has not been described in AS patients in these countries. Two Chinese AS patients with this subtype were reported although no clinical details were available." HLA-B*2709 has been described in only Sardinia and southern Italy. It comprises 20% ofthe B27 alleles on the island ofSardinia and 3% ofB27 alleles in southern Italy-the other prominent allele here being HLA-B*270S. The Sardinian prevalence of HLA-B27 is 4% in the general populaeion.f In this group, previous studies have shown that most of the HLA-B*270S alleles are inherited on one haplotype (HLA-A2; Cw2; B27; DRI6), whereas most of the B*2709 are inherited on another (HLA-A32 or A30; Cw1; B27; DRI2).83 Until recently, extensive studies failed to identify a single AS patient with HLA-B*2709, although four patients with spondyloarthritis not fulfilling criteria for AS have been described in B*2709-positive Italians." However, more recently, two cases of patients with AS who carry HLA-B*2709 have been reported. One had coexisting ulcerative colitis from Lucania in southern ltaly8sand the other from Cagliari in Sardinia" carried HLA-B*1403, which, like B27, contains a free Cys-67 residue and has been found to be associated with AS in a small study ofB27-negative Togolese patients."
How Amino Acid Differences Between HLA-B27 Subtypes May Impact on Disease Pathogenesis Dijference in the Antigen Binding Cleft and Pockets Therein Among HLA-B27 Subtypes The apparent lack ofassociation ofHLA-B *2706andB *2709with AS is particularly intriguing because ofthe close structural similarity of these subtypes to other strongly associated ones. Both HLA-B *2706andB *2709 differ from AS-associated subtypes at residue 116 in the second domain." HLA-B*2706 differs from HLA-B*2704, which is highly disease-associated in Chinese, at only two amino acid positions, 114 and 116, both residing in the floor of the F pocket of the peptide-binding groove. This pocket accommodates the carboxy-terminus of bound peptide and the populations of peptide sequences binding HLA-B*2704 and B*2706 differ by 10-12%, with peptide sequence differences largely confined to the C-terminal residues and two more proximal positions." HLA-B *2709 differs from B*270S by a Single substitution at amino acid position 116. HLA-B*270S and B*2709 share 79% and 88%, respectively, of their peptide repertoires, with sequence differences largely confined to the Ceterrninus,"? Peptides with C-terminal arginine or lysine represent 27% of the natural HLA-B*270Sligands. Crystallographic data have shown that even peptides that are shared between HLA-B*270S and B *2709 may be bound differently. The changes in the F pocket presumably account for the differences in peptides bound by the HLA-B*2706 versus B*2704 and B*2709 versus B*270S encoded antigens. It should be noted, however, that HLA-B*2707, which has been found in patients with AS, also has a substitution position 116 resulting in an altered peptide binding preference. These differences in the peptide repertoire may have profound implications in disease pathogenesis. If the pathogenic role ofHLA-B27 is related to presentation ofarthrirogenic peptides to T-cells, the differential association ofB27 subtypes with ankylosing spondylitis would be based on differences in their peptide-binding specificity or in their capacity to present bound peptides to T-cells rather than to altered antigenicity of shared ligands." Despite that the exact role of HLA-B27 in causing SpA is not known, recent data are point more and more to peptide presentation as the major role . These include a recent study showing that the peptide-binding region alpha 1 and alpha 2 domains of HLA-B27, but not the alpha 3 domain, contributes to spontaneous
168
Molecular Mechanisms ofSpondyloartbropatbies
inflammatory arthritis in transgenic mice." Nearly all HLA-B27 subtype polymorphism is located in this region. This hypothesis is further supported by recent data implicating the ARTS-l gene in AS pathogenesis (otherwise known as Endoplasmic reticulum aminopeptidase-l-ERAP-l), whose function it to trim peptide precursors to lengths appropriate for binding in the grooves of MHC class I molecules by a "molecular ruler" mechanism. Most of HLA-B27 subtypes only differ from each other by a very few amino acids, which alter the peptide binding properties of the molecule. The B pocket tends to be conserved among HLA-B27 subtypes, with changes mostly in the other pockets. The only B27 subtypes with changes in the B pocket of the molecule are HLA-B*2712, B*2716 and B*2718 (which do not have a cysteine at position 67) and no information is available about their disease association. This is particularly important, since this lack of a cysteine at position 67 modifies the B pocket and consequently the peptide binding preferences. HLA-B*1403 (Arg-62-Thr-163-Trp-167) has been associated with AS,S? whereas B*1402 (differing only in residue 156 from B*1403) is not. HLA-B *2712, B*2716 and B *2718 each encode an asparagine residue at position 70. In contrast, most other HLA-B*27 alleleshave lysine at this site, which is located in the vicinity ofthe B pocket (see below). In other HLA-B27 alleles, a salt bridge is predicted to form between lysine at position 70 and asparagine at 74.11.13.15 Residue 74 is substituted with tyrosine in HLA-B*2718, but not in the closely related alleles. HLA-B*2718 and its related alleles display variable substitutions involving positions 70 and 74, but it is not yet known whether the peptide binding properties of the B pocket is altered in these alleles. Interestingly, the HLA allele B *2701 has a substitution ofa tyrosine at position 74, which has been shown to alter the selectivity of the P2 residue in favor of glutamine. This property has been shown to be mediated through an interaction with the lysine at position 70. Another amino acid substitution in HLA-B*2718 is an arginine at position 82, where most other allelesin the B'27 series have a leucine at this site. This substitution replaces an aliphatic residue with a positively charged residue, suggesting that it may be significant. Another polymorphism involves amino acid residue at position 152. This site can also interact with the T-cell receptor and also lines pocket E.46 Most other HLA-B27 alleles have valine at this position whereas the B*2718 allelehas glutamic acid at this position. This substitution represents a switch from an aliphatic residue to a charged residue. Seven other members of the B27 series share this alteration. The presence of serine at residue 67 is another potentially important substitution in the B*2718 allele. Almost all other B27 alleles have a cysteine at this position. This amino acid residue is located above the B pocket, which might influence binding of the P2 anchor residue, arginine. HLA-B*2718 contains a serine at position 67, although it is not yet known whether a disease association exists with this allele. Therefore, it appears unlikely that residues of the B pocket or those responsible for "locking" the peptide near the N terminus like Arg-62, Glu-163 and Trp 167 playa role in AS pathogenesis (Fig. 4).93 The F pocket accommodates the C terminus of the bound peptide and consequently is inaccessible for a TCR. Residues forming this pocket exhibit considerable polymorphism. In pocket F the critical residues involved are at positions: 77, 80, 81, 97 and 116. Clustering ofdifferences in the C/F pockets ofHLA-B27 influences the peptide binding. The other binding pockets act as secondary peptide anchor positions and their contribution are results of the interaction with Band F pockets. Crsytallographic and thermodynamic analysisofsubtypes differentially associated with disease, especially HLA-B*270S and B*2709 have been studied extensively to explain how the structure alters peptide binding. The crystal structures of the disease-associated HLA-B*270S and the nonassociated B*2709 subtypes differ in only one residue (by the presence of an aspartic acid versus a histidine at position 116, respectively) in the F-pocket that accommodates C terminus of the bound peptide. Several different effects of the Asp 116 to His 116 replacement are observed. The bulkier His 116 induces a movement of peptide C-terminal pLys9, allowing the formation of a novel salt bridge to Asp77, whereas the salt bridge between pLys9 and Asp 116 is converted into a hydrogen bond
SubtypesofHLA-B27: History and Implications in the Pathogenesis ofAnkylosing Spondylitis
169
T-ceu receptor
P1
Antigen ic peptideN
/II
P4
~ . . . .",t'..,
P8
.....,.......,
Figure 4. Schematic representation of binding of antigenic peptides to the peptide binding groove of HLA-B27 molecule and recognition by T-cell receptor. Copyright© 1999 Cambridge University Press. Reprinted with permission .
with His 116. His 116 but not Asp 116 adopts two alternative conformations, one ofwhich leads to breakage ofhydrogen bonds. Water molecules near residue 116 differ with regard to number, position and contacts made . It has been shown that there is increased flexibility of the peptide in the binding groove of subtype HLA-B'2709 due to weaker interactions in the F pocker.tt The heavy chain of the B27 molecule is more flexible at binding groove in the HLA-B'270S than in the B '2709 subtype. This difference correlates with an opening ofthe HLA-B'2705 binding groove, accompanied by a partial detachment ofthe C-terminal peptide anchor. The deeply embedded polymorphic heavy-chain residue 116 influences the flexibility ofthe peptide binding groove in a subtype-dependent manner, a feature that could also influence the recognition of the HLA-B27 complexes by effector cells." HLA-B*2705, but not B*2709, is able to bind several peptides in two distinct conformations that show different T-cell reactivities. The subtypes also generate differential T-cell repertoires as exemplified by distinct T-cell responses against the self-peptide pVIPR (RRKWRRWHL). Crystallographic data show that pVIPR binds in an unprecedented dual conformation only to HLA-B*2705 molecules. In one binding mode, peptide pArgS forms a salt bridge to Asp 116 , connected with drastically different interactions between peptide and heavy chain, contrasting with the second, conventional conformation, which is exclusivelyfound in the case ofB*2709. These subtype-dependent differences in pVIPR binding link the emergence ofdissimilar T-cell repertoires in individuals withHLA-B*270S or B*2709to the buried Asp116/His116 polymorphism." The dissociation ofa model peptide with a C-terminal Lysfollows a two-step mechanism. Final peptide release occurs in the second step for both HLA-B27 subtypes. However, thermodynamics and kinetics ofpeptide-HLA interactions reveal different molecular mechanisms underlying the first step, as indicated by different activation energies. In HLA-B*2709, partial peptide dissociation probably precedes fast final peptide release,while in B'2705 an allostearic mechanism based on long-range interactions between ~2-microglobulin and the peptide binding groove controls
170
Molecular Mechanisms ofSpondyloartbrop.ubies
the first step. The resulting peptide presentation mode lasts for days at physiological temperature and determines the peptide-HLA-B*2705 conformation, which is recognized by cellular ligands such as T-cell receptors ." HLA-B'2709 andB'2705 complexes differ significantly in their thcrmostabilitics and this may explain the preference ofHLA-B27 for a peptide with an N -terminal arginine as secondary anchor and the lack ofpreference for tyrosine as peptide C terminus in B*2709. 98 These changes induce subtle peptide conformational alterations that may be responsible for the immunobiological differences between these HLA-B27 subtypes."
Di1ferentialLigand Repertoires ofHLA-B27 Subtypes Comparative peptide studies carried out with subtypes differentially associated with AS indicate that they share about 80% of their peptide repertoires, but differentially bound ligands show some specific motifsYlO.I02 The most conspicuous difference is that the two subtypes not associated with AS, HLA-B*2706 and B*2709, have a high restriction for nonpolar C-terminal residues, including aliphatic ones and phenylalanine. In contrast, disease-associated subtypes are similar in their capacity to bind peptides with C-terminal tyrosine. However, the simplicity ofthis correlation is challenged by the finding that HLA-B*2707, a subtype that is associated with AS in many ethnic groups, does not bind peptides with C-terminal tyrosine and shares with the il27 subtypes not associated with AS a similar restriction for C-terminal nonpolar resldues.'?' However, HLA-B*2707 is not associated to AS in Greek Cypriots and appeared protective. In spite of the common C-terminal motifs, this subtype differs from HLA-B27 alleles not associated with AS due to complex specificity differences at secondary anchor residues. HLA-B*2707differs in peptide specificity from B *2705andB*2704as much as from HLA-B27 subtypes not associated with spondyloarthritis. Like the nonassociated alleles HLA-B*2706 and B*2709, it lacks an aspartic acid residue at position 116 and shows preference for peptides with nonpolar C-terminal residues . The HLA-B*2707-bound repertoire is as different from that of B*2705 as from those of B*2706 and B*2709 (or the two latter subtypes from each other). Differences between B*2707 and B*2705 were based on their C-terminal residue specificity and a subtle modulation at other positions. Differential usage ofsecondary anchor residues explained the disparity between the B*2707-, B*2706- and B*2709-bound repertoires. Similar differences in residue usage is found between B*2707 and both B*2704 and B*2706, as expected from the high peptide overlap between the two latter subtypes. Many shared ligands conserve their alloantigenic features on distinct subtypes.'?' While the differential peptide binding between HLA-B27 subtypes presents a compelling hypothesis for disease association, data thus far cannot confirm obvious peptide motifs or higher peptide sharing among disease-associated subtypes as a pathogenic mechanism.
Impact ofHLA-B27 Subtypes on Molecular Mimicry
HLA-B27 itselfcan be recognized as autoantigenic by CD4+ T-cells 105and can by presented by HLA-class II (DR, DQand DP) heterodimers as an auroanrigen to CD4 positive Tdymphocyres. This was part of the "molecular mimicry" hypothesis that was popular in the late 1980's and early 1990's, wherein homology of peptides from the HLA-B27 molecule shared striking sequence homology with those from bacterial sources. A promising study in this direction was the identification ofa self-ligand ofHLA-B27, derived from the cytoplasmic tailofits own and other HLA class I molecules, that is constitutively presented by three AS-associated subtypes, B*2705,B*2702 and B*2704, but was not found in B*2706 or B *2709.This peptide shows striking similarity with a sequence ofthe DNA primase of Chlamydia trachomatisl'" An alternative approach to a putative arthritogenic peptide is that a self-ligand of HLA-B27 that binds better to those subtypes not associated to AS. Thus, a cross-reactive antigen would more easily overcome tolerance against that self-peptide only when it is sub-optimally presented by AS-associated subtypes.
Subtypes ofHLA-B27: History and Implications in the Pathogenesis ofAnkylosing Spondylitis
171
Using x-ray crystallography, it has been shown that the viral peptide pLMP2 (RRRWRRLTV, derived from latent membrane protein 2 [residues 236-244] of Epstein-Barr virus) is presented by the HLA-B"2705 and B "2709 molecules in two drastically deviating conformations. Extensive structural similarity between pLMP2 and the self-peptide pVIPR (RRKWRRWHL, derived from vasoactive intestinal peptide Type 1 receptor (residues 400-408)) is observed only when the peptides are presented by B"2705 because ofa salt bridge between Arg( 5) ofboth peptides and the subtype-specific heavy chain residue Asp( 116). This demonstrate that molecular mimicry between pLMP2 and pVIPR in the HLA-B27 context is an allele-dependent property,' :" For peptides with certain sequence motifs, molecular and functional mimicry is subtype-dependent and is more relevant when the peptide is presented by the disease-associated HLA-B"2705, than by the nonassociated B"2709 subtype.
Heavy Chain Misfolding One proposed mechanism ofHLA-B27 molecules pathogenicity involves the formation of a disulfide bond between the cysteine residue at position 67 oftwo separate heavy chain molecules , creating homodimers without the participation of132-microglobulin.lo8Thisleads to an endoplasmic reticulum stress and an unfolded protein response .F' ln vitro studies ofhuman B27 alleles show that the substitution ofserine at position 67 can prevent the formation ofheavy chain homodimers. The tendency of HLA-B27 heavy chains to misfold and form homodimers and higher oligomers, with the potential for downstream biochemical and/or immunologic effects have been demonstrated with studies involving with HLA-B"2705.l1 0 Subtypes not associated with AS, HLA-B"2 706 and B "2709, fold much more efficiently than the disease-associated B "2702,B "2704 and B"2705 subtypes, suggesting a correlation between folding properties and propensity to AS. However, HLA-B"2707, which is associated with AS in most populations where it occurs, folds as efficiently as the nondisease associated subtypes, suggesting that the correlation between folding and disease association is not straightforward. Ifthe hypothesis is correct, a logical prediction would be that the peculiar misfolding behavior would correlate with subtype disease susceptibility. That is, one would expect that the behavior would be more pronounced in the B27 subtypes that are strongly associated with AS than in those that are not. However, in a study using cell imaging techniques with fluorescent markers HLA-B'2709 and B '2705 behaved similarly with regard to heavy chain homodimer and oligomer formation. I I I Inappropriately disulfide bonded HLA-B27 heavy chains can be found at both the cell surface and in the endoplasmic reticulum (ER) ofHLA-B27 expressing cells. The heavy chain misfolding does not depend on any unpaired cysteine residue per se when HLA-B27 is highly expressed as thought previously.t'? Major differences exist in the disulfide-dependent conformations of two HLA-B27 subtypes,HLA-B2704andHLA-B2705. The residues 77 , 152 and/or 21 1 influence the redox potential ofthe MHC class I heavy chain and in vitro manipulating the redox environment can alter the conformational state ofHLA-B27 subtypes and aid homodirnerization.! " HLA-B27 subtypes showing differential associations with AS are similar in their extent of 132-microglobulin dissociation and cell surface expression offree heavy chains. I 14
The Role ofTapasin Tapasin is a chaperone protein in the endoplasmic reticulum (ER) involved in peptide loading to MHC class I proteins. Both HLA-B'2704 and B"2705 have an intrinsic tendency to misfold and accumulate in the ER, which is revealed specially in the absence of Tapas in and a relatively slow export rate. In contrast, HLA-B'2706, which is not associated to AS , folds efficiently and is exported quickly in the absence ofTapasin, so that this subtype has an intrinsic folding capacity that is higher than the AS-associated subtypes.!'! However, HLA-B"2709, which is also not associated with AS matures more likeB"2704 andB"2705 than like B"2706, at least in the presence ofTapasin.!"
172
Molecular Mechanisms ofSpondyloarthropatbies
Conclusions Despite the remarkable polymorphism now known [Q exist in HLA-B27, there is still little [Q be gained from performing HLA-B27 subtyping in clinical practice, at least in Caucasians. The only HLA-B27 subtype relevant [Q disease association (or lack thereof) is rare outside Sardinia and its lack ofdisease association is coming under question. It is possible mat in Asian populations there may be more relevance, given me frequent occurrence ofHLA-B '2706 and its demonstrated lack ofa disease association. While new advances in genetics ofAS are occurring in the non MH C region, undemanding the major contribution ofHLA-B27 and its alleles needs [Q elucidated. It is important to investigate which ofthe newer subtypes ofB27 are associated with AS and related SpA and whether there is any preferential association with some ofme clinical features or forms of these diseases. Given fact that most ofthese polymorphism were only recendy demonstrated, the varied distribution of me subtypes in epidemiological studies in various ethnic and racial populations are needed. Molecular studies to learn the effects of me sequence variations on the peptide binding specificity of the molecule may yield a clue [Q disease pathogenesis.
References 1. Weiss EH, Kuon W, Dorner C et al. Organization, sequence and expression of the HLA-B27 gene: a molecular approach to analyze HLA and disease associations. Immunobiology 1985; 170(5) :367-80. 2. Seemann GH, Rein RS, Brown C S et al. Gene conversion-like mechanisms may generate polymorphism in human class I genes. EMBO J 1986; 5(3):547-552. 3. Rojo S, Aparicio P, Choo SY et al. Structural analysis of an HLA-B27 population variant, B27£ Multiple patterns of amino acid changes within a single polypeptide segment generate polymorphism in HLA-B27. J Immunol 1987; 139(3):831-836. 4. Rojo S, Aparicio P, Choo SY er al. Structural analysis of an HLA-B27 population variant, B27£ Multiple patterns of amino acid changes with in a single polypeptide segment generate polymorphism in HLA-B27. J Immunol 1987; 139(3):831-836. 5. Rojo S, Aparicio P, Hansen JA et al. Structural analysis of an HLA-B27 functional variant , B27d, detected in American blacks. J Immunol1987; 139(10) :3396-3401. 6. Vega MA, Wallace L, Rojo S et al. Delineation of functional sites in HLA-B27 antigens. Molecular analysi s of HLA-B27 variant Wewak I defined by cytolytic T-Iymphocytes. J Immunol 1985 ; 135(5) :3323-3332. 7. Vega MA, Bragado R, Ivanyi P er al. Molecular analysis of a functional subtype ofHLA-B27. A possible evolutionary pathway for HLA -B27 polymorphism. J Immunol 1986; 137(11) :3557-3565. 8. Choo SY, Fan LA, Hansen JA. A novel HLA-B27 allele maps B27 allospecificity to the region around position 70 in the alpha I domain. J Immuno11991 ; 147(1) :174-180. 9. Hildebrand WH, Domena JD, Shen SY et al. The HLA-B7~ antigen is encoded by a new subtype of HLA-B27 (B'2708). Tissue Antigens 1994; 44(1) :47-51. 10. Del PP, D'Amaco M, Fiorillo MT et al. Identification of a novel HLA-B27 subtype by restriction analysis of a cytotoxic gamma delta T-cell clone. J Immunol 1994 ; 153(7) :3093-3100. II. Hasegawa T, Ogawa A, Sugahara Y er al. A novel HLA-B27 allele (B'2711) encoding an antigen reacting with both B27- and B40-specific ant isera. Tissue Antigens 1997; 49(6) :649-652. 12. Balas A, Santos S, Garcia-Sanchez F et al. Complete cod ing sequence of HLA-B'2712: a serologic B27-negative antigen associated to Bw6. Tissue Antigens 1998; 51(4 Pt 1):394-397. 13. Seurynck K, Baxter-LoweLA. Novel polymorphism detected in exon 1 ofHLA-B'2713 . Tissue Antigens 1998; 52(2) :187-189. 14. Steiner NK, Jones P, Kosman C et al. Novel HLA-B alleles associated with antigens in the 7C CREG . T issue Ant igens 2001; 57(5) :486-488. IS. Voorter CE , Swelsen WT, van den Berg-Loonen EM. B'27 in molecular diagno stics: impact of new alleles and polymorphism outside exons 2 and 3. Tissue Ant igens 2002; 60(1):25-35. 16. Steiner NK , Jones P, Kosman C et al. Novel HLA-B alleles associated with antigens in the 7C CREG . Tissue Antigens 2001 ; 57(5) :486-488. 17. Voorrer C E, Swelsen WT, van den Berg-Loonen EM. B'27 in molecular diagnostics: impact of new alleles and polymorphism outside exons 2 and 3. Tis sue Antigens 2002 ; 60(1 ):25-35. 18. Feldman D, Kearns J, WuJ et al. Identification and scquencing ofHLA-B*0714 and B'2718 alleles and novel exon 1 sequences of B'0709 and B'2714 alleles in pot ent ial bone marrow donors. Tissue Antigens 2002 ; 59(5):426-429. 19. Tamouza R, Mansour I, Bouguacha N et al. A new HLA-B*27 allele (B*2719) identified in a Lebanese patient affected with ankylosing spondylitis, Ti ssue Antigens 2001 ; 58(1) :30-33.
SubtypesofHLA-B27: History and Implications in the Pathogenesis ofAnkylosing Spondylitis
173
20. Kim MA et al. Hum Immunol 61(SI) :50-50 2000 (abstract) . 21. Gans CPo Mitton W; Ng] er al. Seven new HLA-B alleles associated with antigens in the B7 CREG . Tissue Antigens 2002; 59(3) :229-231. 22. Darke C. Street]. Hammond L et al. Immunogenetic study of a new HLA allele. B'2723. Tissue Antigens 2002; 60(5) :400 -403. 23. Steiner NK, Gans C. Baldassarre L et al. Twenty-five novel HLA-B alleles. Tissue Antigens 2003 : 62(3) :263-266. 24. Steiner NK. Gans C. Baldassarre L et al. Twenty-five novel HLA-B alleles. Tissue Antigens 2003: 62(3):263-266. 25. Luo M. Mao X. Plummer FA. Identification offour novel HLA-B alleles. B'1590. B'1591. B'2726 and B'4705. from an East African population by high-resolution sequence-based typing. Tissue Antigens 2005 : 65(2):187-191. 26. Lazaro AM . Cao K. Masaberg C er al. Twenty-three novel HLA-B alleles identified during intermediate-resolution testing. Tissue Antigens 2006: 68(3) :245-248 . 27. Vidan-]eras B. Kunilo S. Fae I er al. A novel HLA-B'2730 allele found in a Slovene patient affected with 19A nephropathy. Inr ] Immunogenet 2006: 33(5) :371-373. 28. Witter K. Albert T. Volgger A er al. Routine HLA sequence-specific-oligonucleotide typing and confirming sequence-based typing of a prospective bone marrow donor identified a novel B'27 allele. HLA -B'2734. Tissue Antigens 2006; 68(6) :527-528. 29. Li Z. Zou HY. Shao CP er al. Identification of a novel HLA-B'56 allele. B'5618 and an extension of B'2736 by sequence-based typing. Tissue Antigens 2007 : 69(4):365-366. 30. Lopez-Larrea C . Gonzalez-Roces S. Alvarez V. HLA-B27 structu re. function and disease association. Curr Op in Rheumatol 1996: 8(4) :296-308. 31. Spencer Wells: The Journey of Man : A Genetic Odyssey. Princeton University Press. 2002. 32. Diamond]. Gun s germs and steel. The Fates of Human Societies. WW Norton and Company. New York. NY 1999. 33. Seemann GH. Rein RS. Brown C S et al. Gene conversion-like mechanisms may generate polymorphism in human class I genes. EMBO] 1986: 5(3) :547-552. 34. Seemann GH. Rein RS. Brown CS et al. Gene conversion-like mechanisms may generate polymorphism in human class I genes. EMBO] 1986: 5(3):547-552. 35. Rojo S. Aparicio P. Choo SY et al. Structural analysis of an HLA-B27 population variant. B27f. Multiple patterns of amino acid changes within a single polypeptide segment generate polymorphism in HLA-B27.] Immunol1987: 139(3) :831-836. 36. Khan MA . Mathieu A. Sorrentino Ret al. 'The pathogenetic role ofHLA-B27 and its subtypes. Autoirnmun Rev 2007: 6(3) :183-189. 37. Hildebrand WHoDomena ]D, Shen SY er al. The HLA-B7~i antigen is encoded by a new subtype of HLA -B27 (B'2708 ). Tissue Antigens 1994; 44(1) :47-51. 38. Cipriani A, Rivera S, Hassanhi M et al. HLA-B27 subtypes determination in patient s with ankylosing spondyliris from Zulia, Venezuela. Hum Immunol 2003 : 64(7 ):745-749. 39. Shankarkumar U. HLA-B27 allele diversity in Indians: impact of ethnic origin and the caste system. Br] Biomed Sci 2003; 60(4 ):223-226. 40. Oguz FS. Ocal L, Oiler AS ec al. HLA B-27 subt ypes in turkish patients with spondyloarthropathy and health y controls . Dis Markers 2004 : 20(6) :309-312. 41. Fernandez-Sueiro ]L . Alonso C, Blanco F] et al. Prevalence of HLA-B27 and subtypes of HLA-B27 associated with ankylosing spondylitis in Galicia. Spain. Clin Exp Rheurnarol 2004: 22(4) :465-468. 42. Balas A. Santos S. Garcia-Sanchez F et al. Complete coding sequence of HLA-B'2712: a serologic B27-negative antigen associated to Bw6. Tissue Antigens 1998: 51(4 Pr 1):394-397. 43. Yang G. Deng Y]. Yan CX et al. [Frequencies distribution of human leukocyte antigen-B27 subtype s in healthy Chinese]. Zhongguo Yi Xue Ke Xue Yuan Xue Bao 2006 : 28(2) :240-243. 44. MacLean L. HLA-B27 subtypes: implications for the spondyloarthropathies. Ann Rheum Dis 1992: 51(8) :929-931. 45. Alaez C . Gazit E, Ibarrola B et al. Distribution of HLA-B27 subtypes in ankylosing spondylitis in an Israeli population. Arch Med Res 2007: 38(4) :452-455 . 46. Yang G. Deng Y] . Yan CX et al. [Frequencies distribution of human leukocyte antigen-B27 subtypes in health y Chinese] . Zhongguo Yi Xue Ke Xue Yuan Xue Bao 2006 : 28(2):240-243. 47. TaurogJD. Immunology. genetics and animal models of the spondyloarthropathles. Curr Opin Rheumatol 1990: 2(4) :586-591. 48. Ma H]. Hu FP. Diversity of human leukocyte antigen-B27 alleles in han population of hunan province. southern China. Tissue Antigens 2006: 68(2) :163-166. 49. Yang G. Deng Y]. Yan CX et al [Frequencies distribution of human leukocyte antigen-B27 subtypes in healthy Chinese]. Zhongguo Yi Xue Ke Xue Yuan Xue Bao 2006; 28(2) :240-243.
174
Molecular Mechanisms ofSpondyloarthropathies
50. Leffell MS. Fallin MD. Hildebrand WH er al. HLA alleles and haplotypes among the lakota sioux: report of the ASHI minority workshops. parr III. Hum Immunol 2004, 65(1) :78-89. 51. Steiner NK. Jones P. Kosman C et al. Novel HLA-B alleles associated with antigens in the 7C CREG. Tissue Antigens 2001, 57(5) :486-488. 52. Leffell MS. Fallin MD . Hildebrand WH er al. HLA alleles and haplotypes among the lakota sioux: report of the ASHI minority workshops. part III. Hum Immunol 2004; 65(1) :78-89. 53. Dhaliwal JS. Too CL. Lisut M et al. HLA-B27 polymorphism in the Malays. Tissue Antigens 2003, 62(4):330-332. 54. Ma HJ. Hu FP. Diversity of human leukocyte antigen-B27 alleles in han popularion of hunan province. southern China. Tissue Antigens 2006, 68(2):163-166. 55. Ma HJ . Hu FP. Diversity of human leukocyte antigen-B27 alleles in han population of hunan province. southern China. Tissue Antigens 2006 , 68(2) :163-166. 56. Yamaguchi A. Ogawa A. Tsuchiya N et al. HLA-B27 subtypes in Japanese with seronegative spondyloarthropathies and healthy controls. J Rheurnarol 1996; 23(7) :1189-1193. 57. Lopez-Larrea C. Sujirachato K, Mehra NK er al. HLA-B27 subtypes in Asian patients with ankylosing spondylitis. Evidence for new associations. Tissue Antigens 1995; 45(3):169-176. 58. Li Z. Zou HY. Shao CP et al. Identification of a novel HLA-B'56 allele. B'5618 and an extension of B'2736 by sequence-based typing. Tissue Antigens 2007, 69(4) :365-366. 59. Dhaliwal JS. Too CL. Lisut M et al. HLA-B27 polymorphism in the Malays. Tissue Antigens 2003 , 62(4):330-332. 60. Hasegawa T. Ogawa A. Sugahara Y er al. A novel HLA-B27 allele (B'2711) encoding an antigen reacting with both B27- and B40-specific ant isera. Tissue Antigens 1997; 49(6):649-652. 61. Yang G. Deng YJ. Yan CX ct al [Frequencies distribution of human leukocyte antigen-B27 subtypes in healthy Chinese]. Zhongguo Yi Xue Ke Xue Yuan Xue Bao 2006; 28(2) :240-243. 62. Ma HJ. Hu FP. Diversity of human leukocyte antigen -B27 alleles in Han population of Hunan province. southern China. Tissue Antigens 2006 ; 68(2) :163-166. 63. Garcia-Fernandez S. Gonzalez S. Mina BA et al. New insights regarding HLA-B27 diversity in the asian population. Tissue Antigens 2001 ; 58(4) :259-262. 64. Feldman D. Kearns J. WUJet al. Identification and sequencing ofHLA-B'0714 and B'2718 alleles and novel exon 1 sequences ofB'0709 and B'2714 alleles in potential bone marrow donors . Tissue Antigens 2002 ; 59(5):426-429. 65. Chhaya SUO HLA-B27 polymorphism in Mumbai, Western India . Tissue Antigens 2005 ; 66(1) :48-50. 66. Lazaro AM. Cao K. Masaberg C et al. Twenty-three novel HLA-B alleles identified during intermediate-resolution testing. Tissue Antigens 2006 ; 68(3) :245-248. 67. Li Z. Zou HY, Shao CP et al. Identification of a novel HLA-B'56 allele. B'5618 and an extension of B'2736 by sequence-based typing. Tissue Antigens 2007; 69(4):365 -366. 68. Lopez-Larrea C . Gonzalez-Roces S. Alvarez V. HLA-B27 structure. function and disease association. Curr Opin Rheumatol 1996; 8(4) :296-308. 69. Alaez C. Gazit E. Ibarrola B et al. Distribution of HLA-B27 subtypes in ankylosing spondylitis in an Israeli population. Arch Med Res 2007 ; 38(4) :452-455. 70. Varnavidou-Nicolaidou A. Karpasirou K. Georgiou D et al. HLA-B27 in the Greek Cypriot population: distribution of subtypes in patients with ankylosing spondylitis and other HLA-B27-related diseases. The possible protective role ofB'2707. Hum Immunol2004; 65(12) :1451 -1454. 71. Lopez-Larrea C. Gonzalez-Roces S. Alvarez V. HLA-B27 structure. function and disease association. Curr Opin Rheum arol 1996; 8(4) :296-308. 72 . Del PP. D'Amato M. Fiorillo MT et al. Identification of a novel HLA-B27 subtype by restriction analysis of a cytotoxic gamma delta T-cell clone. J Immunol 1994; 153(7) :3093-3100. 73. Tamouza R. Mansour I. Bouguacha N et al. A new HLA-B'27 allele (B'2719) identified in a lebanese patient affected with ankylosing spondylitis. Tissue Antigens 2001 ; 58(1) :30-33. 74. Garcia F. Rognan D. Lamas JR et al. An HLA-B27 polymorphism (B'2710) that is critical for T-cell recogn ition has limited effects on peptide specificity. Tissue Antig ens 1998; 51(1) :1-9. 75. Garcia F. Rognan D. Lamas JR et al. An HLA-B27 polymorphism (B'2710) that is critical for T-cell recognition has limited effects on peptide specificity. Tissue Antigens 1998 ; 51(1):1-9. 76. Garcia-Fernandez S. Gonzalez S. Mina BA et al. New insight s regarding HLA-B27 diversity in the asian population. Tissue Antigens 2001 , 58(4) :259-262. 77. Witter K, Albert T. Voigger A et al. Routine HLA sequence-specific-oligonucleotide typing and confirming sequence-based typing of a prospective bon e marrow donor identified a novel B'27 allele. HLA-B'2734. Tissue Antigens 2006; 68(6):527-528. 78. Lopez-Larrea C. Gonzalez-Roces S. Alvarez V. HLA-B27 st ructu re. function and disease association. Curr Opin RheumatoI1996; 8(4) :296-308.
Subtypes ofHLA-B27: History and Implications in the Pathogenesis ofAnkylosing Spondylitis
175
79. Gonzalez-Roces S, Alvarez MV, Gonzalez S er al. HLA-B27 polymorphism and worldwide susceptibility £0 ankylosing spondylitis. Tissue Antigens 1997; 49(2) :116-123. 80. Nasution AR, Mardjuadi A, Kunmartini S er al. HLA-B27 subtypes positively and negatively associated with spondyloarthropathy, ]' Rheurnatol 1997; 24(6) :1111-1114. 81. Gonzalez-Roces S, Alvarez MV, Gonzalez S et al. HLA-B27 polymorphism and worldwide susceptibility £0 ankylosing spondylitis . Tissue Antigens 1997; 49(2):116-123. 82. D'Ama£O M, Fiorillo MT, Galeazzi M et al. Frequency of the new HLA-B"2709 allele in ankylosing spondylitis patients and healthy individuals. Dis Markers 1995; 12(3):215-217. 83. Fiorillo MT, Cauli A, Carcassi C et al. Two distinctive HLA haplotypes harbor the B27 allelesnegatively or positively associated with ankylosing spondylitis in Sardinia: implications for disease pathogenesis. Arthritis Rheum 2003 ; 48(5):1385-1389. 84. Ramos M, Lopez de Castro ]A . HLA-B27 and the pathogenesis of spondyloarrhritis. T issue Antigens 2002; 60(3) :191-205. 85. Olivieri I, D'Angelo S, Scarano E et al. The HLA-B"2709 subtype in a woman with early ankylosing spondylitis. Arthritis Rheum 2007; 56(8) :2805-2807. 86. Cauli A, Vacca A, Mameli A er al. A Sardinian patient with ankylosing spondylitis and HLA -B"2709 co-occurring with HLA-B"I403. Arthritis Rheum 2007; 56(8) :2807-2809. 87. Lopez-Larrea C, Mijiyawa M, Gonzalez S er al. Association of ankylosing spondylitis with HLA-B"1403 in a West African popularion. Arthritis Rheum 2002; 46(11) :2968-2971. 88. D'Amaro M, Fiorillo MT, Carcassi C ct al. Relevance of residue 116 of HLA-B27 in determining susceptibilit y to ankylosing spondylitis . Eur] Immunol1995; 25(11) :3199-3201. 89. Sesrna L, Montserrat V, Lamas ]R et al. The peptide repertoires of HLA-B27 subtypes differentially associated to spond yloarthropathy (B"2704 and B"2706) differ by specific changes at three anchor positions.] Bioi Chern 2002 ; 277(19):16744-16749. 90. Ramos M, Paradela A, Vazquez M er al. Differential association ofHLA-B"2705 and B"2709 to ankylosing spondylitis correlates with limited peptide subsets but not with altered cell surface stability. ] Bioi Chern 2002; 277(32):28749-28756. 91. Montserrat V, Marti M, Lopez de Castro ]A . Allospecific T-cell eplrope sharing reveals extensive conservation of the antigenic features of pepride ligands among HLA-B27 subtypes differentially associated with spondyloarrhritis.] Immunol 2003 ; 170(11):5778-5785. 92. Khare SO, Lee S, Bull M] er al. Peptide binding alphalalpha2 domain ofHLA-B27 contributes to the disease pathogenes is in transgenic mice. Hum Immunol1999; 60(2) :116-126. 93. Lopez-Larrea C, Mijiyawa M, Gonzalez S er al. Association of ankylosing spond ylitis wirh HLA-B"1403 in a West African population. Arthritis Rheum 2002; 46(11 ):2968-2971. 94. Srarikov EB, Nilsson L, Hulsmeyer M. A single residue exchange between two HLA-B27 alleles triggers increased peptide flexibility. Eur Biophys ] 2004; 33(7):651-655. 95. Fabian H, Huser H , Narzi 0 et al. HLA-B27 subtypes differentially associated with disease exhibit conformational differences in solut ion. ] Mol Bioi 2008; 376(3 ):798-810. 96. Hulsmeyer M. Fiorillo MT, Bettosini F er al. Dual, HLA-B27 subt ype-dependent conformarion of a self-pepride.} Exp Med 2004 ; 199(2):271-281. 97. Winkler K, Winter A, Rueckert C er al. Narural MHC class I polymorphism controls the pathway of peptide dissociation from HLA-B27 complexes. Biophys] 2007; 93(8) :2743-2755. 98. Hillig RC, Hulsmeyer M, Saenger W er al. Thermodynamic and structural analysis of peptide- and allele-dependent properties of two HLA-B27 subtypes exhibiting differential disease association.] Bioi Chern 2004 ; 279(1) :652-663. 99. Hulsmeyer M, Welfle K, Pohlmann T et al. Thermodynamic and structural equivalence of two HLA-B27 subtypes cornplexed with a self-pepride. ] Mol Bioi 2005; 346(5) :1367-1379. 100. Fiorillo MT, Meadows L, D'Amato M et al. Susceptibiliry to ankylosing spondylitis correlates with the C-rerminal residue of peprides presented by various HLA-B27 subtype s. Eur] Immunol 1997; 27(2):368-373 . 101. Garcia F, Marina A, Lopez de Castro ]A . Lack of carboxyl-terminal tyrosine distinguishes the B"2706-bound peptide repertoire from those ofB"2704 and other HLA-B27 subtypes associated with ankylosing spondylitis . Ti ssue Antigens 1997; 49(3 Pr 1):215-221. 102. Lopez de Castro ]A, Alvarez I, Marcilla M et al. HLA-B27 : a registry of consriturive peptide ligands. Tissue Antigens 2004 ; 63(5):424-445. 103. Tieng V, Dulphy N, Boisgeraulr F et al. HLA-B"2707 peptide morif: Tyr C-terminal ancho r is not shared by all disease-associated subrypes. Immunogenerics 1997; 47(1) :103 -105. 104. Gomez P, Montserrat V, Marcilla M et al. B"2707 differs in pepnde specificity from B"2705 and B"2704 as much as from HLA-B27 subtypes nor associated to spondyloarthritis. Eur ] lmmunol 2006 ; 36(7) :1867-1881.
176
Molecular Mechanisms ofSpondyloarthropathies
105. Boyle LH. Goodall]C. Opat SS et al. The recogn ition ofHLA-B27 by human CD4(+) T-lymphocytes. ] Immunol 2001; 167(5) :2619-2624. 106. Ramos M . Alvarez I. Sesma L er al. Molecular mimicr y of an HLA-B27-derived ligand of arthritis-linked subtypes with chlamyd ial proteins.] Bioi Chern 2002; 277(40):37573-37581. 107. Fiorillo MT. Ruckert C. Hulsme yer M et al. Allele-dependent similarity between viral and self-peptide presentation by HLA-B27 subtypes.] Bioi Ch ern 2005 ; 280(4) :2962-2971. 108. Kollnberger S. Bird L. Sun MY et al. Cell-surface expression and immune receptor recognition of HLA-B27 homodimers. Arthritis Rheum 2002; 46(11) :2972-2982. 109. Turner M]. DeLay ML. Bai S er al. HLA-B27 up-regulation causes accumulation of misfolded heavy chains and correlates with the magnitude of the unfolded protein response in transgenic rats: Implications for the pathogenesis of spondylarrhriris-like disease. Arthritis Rheum 2007; 56(1) :215-223. 110. Mear ]P. Schreiber KL. Munz C et al. Misfolding of HLA-B27 as a result of its B pocket suggests a novel mechanism for its role in suscept ibility to spondyloarthropathies. ] Immunol 1999; 163(12):6665-6670. Ill. Giquel B. Carmouse S. Denais C et al. Two HLA-B27 alleles differently associated with spondylarth rin s, B*2709 and B*2705. display similar intracellular trafficking and oligomer formation . Arthritis Rheum 2007; 56(7):2232-2242. 112. Edwards ]C. Bowness P. Archer ]R. Jekyll and Hyde: th e tran sformation of HLA-B27. Immunol Today 2000; 21(6):256-260. 113. Saleki K, Hartigan N, Lith M et al. Differenti al oxidation ofHLA-B2704 and HLA-B2705 in lyrnphoblasroid and rran sfecred adherent cells. Aurioxid Redox Signal 2006; 8(3-4) :292-299 . 114. Vazquez MN, Lopez de Castro ]A . Similar cell surface expression of beta2 -microglobulin-free heavy chains by HLA-B27 subtypes differentially associated with ankylosing spondylitis. Arthritis Rheum 2005; 52(10):3290-3299. 115. Montserrat V, Galocha B, Marcilla M et al. HLA-B*2704. an allotype associated with ankylosing spondylitis. is critically dependent on transporter associated with antigen processing and relativelyindependent of tapasin and immunoproteasome for maturation. surface expression and T-cell recognition : relationship to B*2705 and B*2706.] Immunol 2006 ; 177(10):7015-7023. 116. Goodall]C, Ellis L. Hill Gaston ]S. Spondylarrhritis-associated and nonspondylarthriris-associared B27 subtypes differ in their dependence upon tapasin for surface expression and their incorporation into the peptide loading complex. Arthritis Rheum 2006 ; 54(1) :138-147.
CHAPTER
13
Implications ofStructural and Thermodynamic Studies ofHLA-B27 Subtypes Exhibiting Differential Association with Ankylosing Spondylitis Andreas Ziegler," Bernhard Loll, RolfMisselwitz and Barbara Uchanska-Ziegler Abstract tructu ral and thermodynam ic prop erties of H LA-B27 molecules provide the basis for their function within the immune system and are probably also cent ral for the understanding of the pathology of HL A-B27-associated diseases such as ankolysing spondylitis (AS). Several HLA-B27 alleles are Afi-associated , whereas some are not, although the protein encoded by the former maydiffer in only a single amino acid exchange from tho sespecified by the latter. This indi cates that subtype-specific polymorphic residues play a key role in determining whether an HLA-B27 subtype is AS-associated or not and open the po ssibilit y to correlate struc tural, thermodynamic and functional characteristics ofa given subtype with the diseaseassociation. Our studies involved X-ray crystallography and various other biophysical techn iques to examine how several different peptides are accommodated within th e binding groove of the molecules. The HLA -B'270S and HL A -B"2709 subtypes. who se products differ in only a single amino acid residue of their heavy chains from each other. were primarily cho sen for the se analyses. but our stu d ies have recently also been extended to the closely related subtypes HLA-B "2703. HLA-B"2704 and HLA-B'2706.The analyses reveal that structural and thermodynamic differences between HLA-B27 complexes may exist. depending on the peptide that is displayed. Furthermore.a viral peptide and two self-peptides were found that exhibit HLA-B27 subtype-dependent molecular mimicry, thereby providing a molecular basis to account for the subtype-dependent presence ofautoreactive T-cells. Although the se result s do not exclude other theories for the pathogenesis of AS, they support the arthrirogenic peptide hypothesis which envisages molecular mimicry between HLA-B27-presented foreign and self-peprides to explain the cross-reactivity of auto reactive T-cells that are found in HLA-B"270S-positive individuals. in particular when the y suffer from A S.
S
' CorrespondingAuthor: Andreas Ziegler-Institut fUr Immungenet ik, Charite-U niversitatsrned izi n Berlin, Freie U niversitat Berlin, Thielall ee 73, D-141 95 Berlin, Germ any. Email :
[email protected]
M olecular M echanisms of Spondyloarthropathies. edited by Carl os Lopez-Larrea and Roberto D iaz-Peiia. ©2009 Landes Bioscien ce and Springer Science+ Business Media.
178
Molecular Mechanism s ofSpondylorthropathies
Introduction As detailed also in several other chapters ofthis volume, Ankylosing Spondylitis (AS) is clearly a multi-factorial disease in which not only the HLA-B27 gene, but also genes outside ofthe HLA complex as well as environmental factors are involved."? However, the association of HLA-B27 with AS has early been recognized as being so striking, that a number ofhypotheses in which this gene or its product play the dominant role, were proposed over the last twenty years to provide an explanation. These theories can be grouped into at least two broad categories: either they attribute the causation or prevention ofAS to an HLA-B27-linked gene ,6.7or they focus on the HLA-B27 molecule itself. 8-13 Although not excluding the former, the relevance ofthe latter group oftheories was directly demonstrated by the creation ofrats transgenic for HLA-B27 heavy chains (HC) and human ~2-microglobulin(~2m) that show AS-like symptoms." Over the years, perhaps the most fruitful ofthe theories focussing on the HLA-B27 antigen was the "arthritogenic peptide hypothesis"! which suggests that particular properties ofirs peptide binding groove enable this molecule to display peptides ofbacterial or viral origin that exhibit molecular mimicry with specific arthrirogenic sclf-peptides. Molecular mimicry, which may be defined as similarity in overall shape and charge for an interaction surface, IS will permit HLA-B27 -restricted, cross-reactive cytotoxic T-lymphocytes (CTL) to mount a response, initially directed again st a foreign peptide, to turn towards self-peptides as well, with an ensuing autoimmune response and chronic inflammation. On the other hand, a conformational reorganization of the HC has been suggested" to involve the generation ofa ~2m-free, improperly folded H C aggregate that presents part ofits own polypeptide chain, possibly from a neighbouring HLA-B27 HC, on the cell surface, permitting its recognition by autoreactive T-cells. In contrast, the HC homodimer theory? and the mis-folding hypothesis'? do not concentrate on the peptide presenting function ofHLA-B27 molecules, but envisage particular biochemical properties specific to its H C as being fundamental in causing the disease, while the ~2m-deposition hypothesis'! suggests an involvement ofthe light chain ofHLA-B27 molecules as primary cause ofAS. Here, ~2m is supposed to be released from a sub population ofcell surface-expressed HLA-B27 molecules as a consequence oftheir dissociation due to the binding of low-affinity peptides . In this way, the initiation of chronic inflammatory reactions and possibly even deposition of ~2m within joints is envisaged, finally culminating in destructive spondyloarthropathy. In addition, it has been proposed that HLA-B27 expressing cells might possess an unusual ability to respond to certain microbial infections, thereby altering and impairing the immune response.'! All these theories have their merits, but in one way or another, all have also been challenged. Therefore, as surprising as it may be, after more than thirty years of research on HLA-B27 and AS, there is currently not even a generally accepted working hypothesis to explain the reasons for this association. Here, we will summarize evidence from X-ray crystallographic studies of selected HLA-B27:peptide complexes indicating that the assumptions underlying the arthritogenic peptide hypothesis" are indeed not unfounded. Furthermore, we present results from spectroscopic and microcalorimetric experiments which suggest that thermodynamic properties must not be neglected either when trying to obtain an integral picture of the molecular nature ofHLA-B27 antigens.
General Structural Properties ofHLA Class I Molecules Following the seminal structural studies ofD. Wiley's group on HLA-A2 ,16 HLA-B27 was amongst the first major histocompatibility complex (MHC) class I molecules whose structure was derermined.V'P The choice ofthis antigen was at least partly due to the very strong association of HLA-B27 with AS19.20 and other spondyloarthropathies (for reviews see refs. 21,22). Although no defined peptide was present in the peptide-binding groove ofthe molecule, these initial X -ray crystallographic studies revealed that HLA class I molecules (Fig . 1) consist ofan HC that folds into two a-helices (a1 and (2) which, together with a table-like extended ~-pleatedsheet structure, form a groove that is able to accommodate a peptide derived from foreign (e.g., viral) or cellular proteins. The polymorphic residues that distinguish a given MHC class I molecule from another are nearly exclusively located in the o l- and a2-domains and serve the anchoring of a peptide
Implications ofStructural and Thermodynamic Studies ofHLA-B27 Subtypes
179
f~-pleated
sheet
Figure 1. Architecture of an MHC class I molecule in complex with a peptide . The structure of B*2705 :T1S is depicted. with the peptide N-terminus in front. ~,m (cyan) and heavy chain (green) with the al- and a2-helices (orange) that flank the peptide binding groove are drawn in cartoon representation with the bound nonapeptide TIS (red ) in stick representation. The anchoring residues pArg2 (p2) and pLeu9 (pQ) are drawn in black.
within the groove as well as the allele-specific docking ofCTL that recognize the membrane-bound target molecule through their Tcell receptor (TCR) . The MH C class I molecule is composed of a further, membrane-proximal HC domain (a3) that is recognized by invariant receptors on the surface ofeffector cells such as CD8 23and molecules ofthe immunoglobulin-like transcript (IiT) family." In addition, a noncovalently associated chain ofM, -12,000, 132m, is part ofa typical class I molecule." The entire complex is anchored on the cell membrane via a single transmembrane a-helix and a short intracytoplasmic segment ofthe H C. However, both are not part ofthe soluble, peptide-cornplexed MHC molecule that is the object ofX-ray crystallographic, spectroscopic and other biophysical studies. HLA-B27 differs from nearly all other MHC class I molecules by the characteristic anchoring ofthe peptide through an arginine residue at peptide position 2 (pArg2) (Fig. 1). The firm anchoring of this amino acid within the molecule's B pocket is a consequence of the interaction of the positively charged guanidinium group ofpArg2 with two negatively charged glutamate residues of the HC, Glu45 and Glu63. 1R•2s Consequently, this characteristic property ofHLA-B27 molecules has been proposed to be primarily responsible for the association ofthe HLA-B27 gene with AS 10 and there is biochemical evidence that an exchange of B pocket residues does indeed affect the folding behaviour of the H C. 26 Another unusual property ofHLA-B27 molecules is the presence of the free thiol group ofCys67, through which HC dirnerization" can proceed.
Molecular Mechanisms ofSpondylorthropathies
180
HLA-B27 Subtypes Exhibiting Differential Disease Association A list ofthe HLA-B27 alleles (also termed subtypes) can be found in the chapter by].A. Lopez de Castro in this volume and a continuously updated version is provided by the HLA database website at http: / /www.ebi.ac.uk/imgt/hla/ . Most ofthe se subtypes are very rare, with unknown association to AS . However, two pairs of subtypes deserve particular mention, HLA-B*270S/HLA-B *2709 (in short ,B*270S and B *2709) andHLA-B*2704/HLA-B*2706 (in short,B*2704 and B *2706). B*270S and B*2704 are clearly associated with AS , while B*2709 and B*2706 are at best rarely associated, although the antigens belonging to the B*270S /B*2709 and the B*2704/B*2706 pair differ by only one or two amino acid residues, respectively.F"? In both cases, these exchanges affect the peptide binding groove, more specifically the F pocket which accommodates the peptide C-terminus (pO): Aspl16 (B*270S) is replaced by Hisl16 in B*2709, while His l l-i and Aspl16 (B*2704, identical to the respective residues in B*270S) are changed to Asp1l4 and Tyr1l6 in case ofB*2706. The finding ofthese closely related pairs ofsubtypes has given functional, genetic, biochemical and biophysical studies ofthe HLA-B27 alleles and the encoded products a tremendous stimulus, since a comparison was expected to shed light on the reasons for their differential association to AS . For example, the finding that peptide repertoires were largely overlapping, but nevertheless HLA-B27 subrype-speclfic'":" (Lopez de Castro, this volume), allowed the prediction that the Aspl16His exchange in B*270S /B*2709 or the two amino acid exchanges distinguishingB*2704 and B*2706 will also influence the emergence ofT-cell repertoires in the respective individuals. This might in turn be connected with the subtypes' differential susceptibility to AS . Consequently, all hypotheses seeking to explain the association between HLA-B27 and AS must take the existence ofdifferentially AS-associated HLA-B27 alleles into account.
HLA-B27: Peptide Complexes for Structural and Other Biophysical Studies The B*270S/B*2709 pair appeared to be particularly promising to support or reject any of the hypotheses that have been put forward to explain the association ofselected HLA-B27 alleles with AS . Although some doubts have recently been expressed," there is considerable evidence showing that B*2709 is generally not associated with AS in Sardinia, where about 20% of the HLA-B27-positive individuals carry B*2709 (Fiorillo and Sorrentino, this volume), in contrast to the remaining 80%, who are B *270S-positive. AS patients in Sardinia carry nearly exclusively the B*270S allele ? ·33 Our initial work was carried out with these two subtypes, but we later extended our studies also to B*2703 (differing from B*270S only by a TyrS9His exchange affecting the A pocket which accommodates the peptide N-terminus), B*2704 and B*2706 molecules. However, comparative functional studies employed nearly exclusively cells from individuals t yped as B *2 705 or B*2709. 34 The HC of all sub t ypes was expressed in E. coli as inclusion bodies containing the truncated HC comprising the three external o l , a2 and a3 domains, but lacking the membrane-intrinsic and cytoplasmic domains. ~2m was expressed in parallel using the same system. Soluble, trimeric HLA-B27 complexes were then reconstituted by incubating HC, ~2m and a synthetic peptide for 2-3 weeks at 4 C in a buffer system promoting the assembly of the three components in native form." After concentration and purification by size exclusion chromatography, the complexes were either crystallized for X-ray diffraction studies36.44 or used for other biophysical experirnents.V'" :" The HLA-B27:peptide complexes referred to in this article, together with the sequences of the peprides, their origin, the resolution obtained, Protein Data Bank accession codes and the respective reference, are shown in Table 1. The peptides were either totally artificial (m9), a proven self-p ep tide (TIS), derived from a viral protein (pLMP2), known to be derived from cellular proteins without formal evidence of being naturally presented by HLA-B27 antigens (pVIPR, pGR), or a modified self-peptide (sIOR) . Apart from sIOR, which is a decapepride, all other peptides are nonapeptides. 0
Implications o/Structural and Thermodynamic Studies ofHLA-B27 Subtypes
181
Structural Studies with the m9 Peptide The m9 peptide (GRFAAAIAK; Table 1) was synth esized to posse ss preferred amino acid s for binding to B'Z705 at five positions: pGly 1, pArgZ, pPhe3, pIle7 and pLys9.48 While pArgZ is a primary anchor, pGly1 and pLys9 are secondary anchors . B'Z709 shares the preference for pGly I , pArgZ and pPhe3 with B'Z705, but pIle 7 is not preferred and Lys is never found at pQ in peptides eluted from this subtyp e." Nevertheless, complexes were obtained also for B'Z709 and crystals were produced that diffracted with synchrotron radiation to high resolution in both cases.V The complexes exhibit the typical stru ct ure ofMHC classI molecules (Fig . 1) and the peptide is bound by both subtypes in an elongated conformation (Figs. ZA,B; 3A,B) , similar to that obser ved for th e composite peptide model in the initial stu dies ofHLA-BZ7 molecule s.F:'" While nearly all HC and fi2m residu es occupied identical positions, a detailed comparison of the m9 binding modes of B"Z705 and B'Z 709 revealed minute differences at the sur face of the molecules, around positions p4-p6. In addition, as expected from the kn own differential preference ofth e two subtypes for C -term inal peptide residue s," peptide co -ordi nation in the F po cket was distinguishable between the subcyp es" whi ch is due to the Asp 116His polymorphism. In B'Z 705 , AsplI6 contacts pLys9 via a salt bridge, while in B'Z709, pLys9 issh ifted 1.1 A by the bulky HisII6 towa rds the a I -helix and forms a salt bridge to Asp77. In addition, the position of Lys70 is also affected according to HLA-BZ7 subtype (a shift of 1.1 A as well), although th e salt bridge to Asp74 is ret ained in both struc tures. Furthermore, th e number and positioning of wat er molecules within and in th e vicin ity of th e F pocket are different. These stru ctural alterations are confined to the interior of th e peptide binding groove in th e respective subtype and it app ears very unlikel y that the affected residu es can be contacted by any of th e prot ein ligands ofHLA-BZ7 mole cules such as TCR or NK cell receptors. Overall, th e distinct ive feature s of F pocket arch itecture might contribute to weaken ing of the binding of peptides that contain Lys as pQ in the B'Z 709 subtype," :"
Table 1. HLA-B27:peptide complexes with solved structures that are referred to in this article Subtype
Peptide
Sequence
Origin
Resolution [AI
POB-Entry
Reference
B'2705
m9
G RFAAAIA K
2.10
ljge
36,49
B*2709
m9
G RFAAA IAK
1.09
lk5n
36,49
B*2705 B*2709 B*2705 B*2709 B*2705 B*2709 B*2705 B*2709 B*2703 B*2703 B*2704 B*2704 B*2706 B*2706 B*2709
TIS TIS p LMP2 pLMP2 pV IPR pV IPR pG R pG R p LM P2 pV IPR pLMP2 pVIPR p LMP2 pVI PR slO R
RRLPIFSRL RRLPIFSRL RRRWRRLTV RRRWRR LTV RRKW RRWH L RRKW RRWH L RRRWHRWR L RRRWHR W RL RRRW RRLTV RRKWR RWH L RRRWRRLTV RRKWRRW HL RRRW RRLTV RRKWRRWHL RRLLRGH NQY
M od el pep tide M odel pepti de Self Self V iral V iral Self Self Self Self V iral Self V iral Self Vira l Self M odified self
2.27 2.10 1.55 1.72 1.47 2.20 1.50 1.2 2.00 1.83 1.90 1.60 2.70 1.83 1.90
l w Ov l w Ow l ux s l uxw logt lof2 2a83 3c2 f
37 37 38 38 39 39 40 Thi s pape r 41 41 42 42 43 43 44
' to be d ep o sit ed in th e Pro tein D at a Ban k .
l jgd
Molecular Mechanisms ofSpondylorthropathies
182
B
pArg2
pArg2
k
p116
J
Figure 2. Figure legend viewed on following page.
Implications ofStructural and Thermodynamic Studies ofHLA-B27 Subtyp es
183
Figure 2, viewed on previous page. Comparison of the binding modes of the peptides m9, TIS, pLMP2 , pVIPR and pGR. The peptides are seen from the side of the n2-helix. (A) B*2705:m9 in lemon, (B) B*2709 :m9 in light cyan , (C) B*2705 :TI5 in orange, (D) B*2709:TI5 in purple, (E) B*2705 :pLMP2 in green, (F) B*2709:pLMP2 in olive, (G) B*2705:pVIPR in NC conformation in yellow and in CC conformation in deep purple, (H ) B*2709:pVIPR in grey, (I) B*Z705:pGR in A conformation in magenta and in B conformation in cyan and (J) B*Z709:pGR in light blue. The polymorphic residue 116 of the HC (Asp in case of B*Z705) is drawn if it is in direct contact with the bound peptide. This is only the case in B*2705:pLMPZ (E) and in B*Z705 :pVIPR-NC (G).
but X -ray crystallography cannot provide definitive prooffor this asswnption. Therefore, we have determined a variety ofthermodynamic properties ofthe complexes to verify this hypothesis and the results (see below) show that m9 is indeed bound with much higher affinity to B*270S than to B*2709. 44 .46 In addition to the original description ofthe X-ray strucrures" a preliminary summary ofresults obtained with the m9 peptide has recently been presented."
Structural Studies with the Proven HLA-B27 Self-Peptide TIS The TIS peptide (RRLPIFSRL; Table I) is derived from epidermal growth factor response factor I and has been eluted from B*270S and B*2709 molecules. " :" Again, both protein complexes exhibit the characteristic fold of HLA class I molecules (Fig. I) and the very small structural differences that can be observed are restricted to solvent-exposed side chains at the protein sur face such as pArg8 (see e.g., Figs. 2C,D; 3C,D).37 Slight rearrangements ofatomic contacts within the F pocket are du e to water molecules occupying different positions near the polymorphic residue at H C position 116. The positioning ofthe TIS peptide is also very similar to that ofthe m9 peptide (Figs. 2A,B,C,D ; 3A,B, C,D), since the Co-backbones overlay nearly perfectly. Furthermore, the four structures emphasize the importance of the primary anchor pArg2, as the contacts of this side chain are completely pre served. Nevertheless, an interesting difference is notable between the structures with m9 and TIS "above" the region ofthe A pocket: the peptide residues pGlyI and pArgl induce a rearrangement of several He residues on the surface of the complex, affecting Arg62, Glul63 and Trp167. Thi s rearrangement leads to networks of tight contacts that have previously been obeserved by us in case ofthe peptides m9 and sI OR (Table 1).%,44pGly I (m9 peptide) may be preferred by B*270S and B*2709, because Trp 167 rearranges to cover the side chain-lacking pGly I in a lid-like fashion , th ereby assuring its buried state, while the side chain ofArg62 abandons a water-mediated contact with Glul63 and re-orientates towards Trp167.44However, this explanation is not in line with the notable decrease in affinity that is observed when pArgl is replaced by pGlyl in the slOR p eptide (see below) .44 In contrast, the se contacts above the A and B pockets could explain the preference for pArgl-containingpeptides that has been found for B*270S,48 although an alternative, not mutually exclusive, explanation has been advanced.soOn the other hand, we have proposed a structure-based explanations 1 for the lack ofacidic residues (Asp, Glu) at p 1.481he replacement ofpArgl by pAsp I would result in lack ofn-stacking of the pi side chain with Arg62 and Trp167, reduced van der Waals interactions, loss of the salt bridge to Glu163, electrostatic repulsion between pAspl and Glul63 and loss of the Arg62-Glu163 interaction, resulting in a major, energetically unfavourable repositioning ("domino effect") of residues in the vicinity of the N-terminal portion of the peptide. These considerations are not purely academic, because a pAsp I-containing human type VI collagen-derived peptide (C34) has already been proposed to possess arthrirogenlc potential.V In contrast, our structure-based considerations lead us to suggest that the C34 peptide is not a good candidate arthritogenic pepnde." Finally, the structural studies of the TIS peptide in B*270S and B*2709 permit a number of more general comments to be made regarding the antigenicity of HLA-B27 subtypes. Extensive CTL cross-reactiviry " between the subtypes is due to the fact that 79-80% ofthe bound peptides are shared between B*270S and B*2709 .31,48 Without doubt, the TIS peptide belongs into this
184
Molecular Mechanisms ofSpondylorthropathies
A
B
E
F
H
J
Figure 3. Figure legend viewed on following page.
Impli cations ofStructural and Thermodynamic Studies ofHLA-B27 Subtypes
185
Figure 3, viewed on previous page. TCR-view of the peptides m9, TIS, pLMP2, pVIPR and pGR. The peptides are rotated by 90' towards the viewer in comparison to Figure 2, as an approching TCR would "see" them . (A) B*2705:m9 in lemon, (B) B*2709:m9 in light cyan, (C) B*2705:TISin orange, (D) B*2709:TISin purple, (E) B*2705:pLMP2 in green, (F) B*2709 :pLMP2 in olive, (G) B*2705:pVIPR in NC conformation in yellow and in CC conformation in deep purple, (H) B*2709:pVIPR in grey, (I) B*2705:pGR in A conformation in magenta and in B conformation in cyan and (J) B*2709:pGR in light blue.
by far largest group ofligands that are unlikely to account for differential AS associations because of their functional, structural and thermodynamic (see below) equivalence. We have recently determined the structures of B*Z705 and B*Z709 with a further member of this group, a cathepsin A-derived peptide (pCatA) and found that the two structures are again indistinguishable (Kumar et al, manuscript in preparation). The complexes ofthe subtypes with TIS and pCatA will serve as a useful reference for further experiments with naturally occurring peptides that exhibit allele-dependent differential functional and physical characteristics when bound to HLA-BZ7 subtypes, as described in the next sections.
Structural Studies ofthe Viral Peptide pLMP2 Although pLMP2 (RRRWRRLTV; Table I), a fragment of latent membrane protein 2 of Epstein-Barr virus, is not an immunodominant peptide in case of B*2705,S4 CTL are readily produced by B*270S- as well as by B*2709-positive individuals and these cells recognize pLMP2 also in the context ofthe other subtype. Such cross-reactive CTL have also been used to probe the display of several sequence-related foreign and self-peptldes, including pVIPR and pGR (see the next sections). 34.38-40 These experiments suggested that pLMP2 might not be presented identically by B*2705 and B*2709 and we employed X-ray crystallography to provide a structural foundation for these results. P High quality diffraction data sets were obtained from crystals of both subtypes, permitting a direct comparison ofthe seructures-" Again, the typical HLA class I architecture is observed, with indistinguishable H C and 132m (root mean square deviation 0.2 A), except fur the polymorphic Asp or His 116. For each complex, the peptide could be modelled unambiguously. When complexed to B*2709, pLMP2 is bound in the conventional, canonical (CC) conformation (Fig. I), which we have also termed p4a because the main chain
186
Molecular M echanisms ofSpondylorthropathies
HLA-B27 Subtype Dependent Conformational Peptide Dimorphisms Similar to pLMP2, the sequence-related pVIPR peptide (RRKWRRWHL, derived from vasoactive intestinal peptide Type 1 receptor; Table 1) can be presented by B*2705 and B*2709, but the resulting Tvcell repertoires are different : B '2709 individuals very rarely possesspVIPR-specific CTL, while they are abundant in peripheral blood of healthy B'2705-positive persons and are very frequently present in AS patients. " This suggests that efficient thymic selection removes pVIPR-specific CTL in B'2709-, but not in B*2705-positive individuals. To shed light on the reasons for the presence of these auroreactive CTL in persons typed as B'2705, we crystallized pVIPR in complex with both B*2705 and B*2709. The results show that pVIPR binds to B*2709 in the CC conformation (Figs. 2H ; 3H), while a hitherto unobserved, dual binding mode is seen in B*2705 (Figs. 2G; 3G). One of the two conformations (CC) is identical to that found in B*2709, while the other (NC) is noncanonical and exploits the presence of Asp116. The high resolution ofthe B*2705:pVIPR structure (Table 1) made the unambiguous identification ofboth conformations within the B*2705 binding groove possible. The HLA-B27 subtype-dependent pVIPR binding modes are clearly a consequence ofthe Asp116His polymorphism, because crystallographic artefacts could be excluded as a possible source ofthese findings . A detailed account of these results has already been given.39 However, our crystallographic studies could not provide an answer to the obvious question whether the simultaneous occurrence ofthe pVIPR-CC and pVIPR-NC binding modes isstatic or dynamic. Ifstatic, the crystals used for the diffraction studies would be composed ofroughly equal amounts oftwo species, one in the pVIPR-CC and the other in the pVIPR-NC conformation. If dynamic, the peptide would be expected to change continuously from one conformation to the other within the binding groove, resembling the movement ofa skipping rope. The pivot points might be located near the connections ofp2-p3 and p7 -p8, respectively (Figs. 2G; 3G). However, such a dynamic scenario would very likelyhave led to diffuse electron density for the peptide in the B*2705:pVIPR structure which clearlywas not the case in our studies. Thermodynamic experiments have so far not been able to solve this problem, but we have recently employed infrared spectroscopic measurements of B*2705:pVIPR and B*2709:pVIPR complexes containing BC-labeled 132m combined with molecular dynamics simulations to shed light on the nature of the pVIPR binding mode.'?The results indicate the presence ofthe CC conformation ofpVIPR at physiological temperature and another study employing fluorescence depolarization and pK. calculations suggests that pVIPR is bound in the NC binding mode to B*2705 under physiological conditions (Narzi et al, manuscript submitted for publication). When combined, these two studies provide circumstantial evidence, but still no definitive proof, for the co-existence of the pVIPR-CC and the pVIPR-NC conformations in the B*2705 subtype under physiological conditions. The structural dimorphism ofpVIPR provides a novel explanation for the emergence ofdistinct pVIPR-specific T-cell repertoires in B'270S and B*2709 individuals. T-cell selection might be impaired in the former because ofthe highly unusual dual binding mode ofth is peptide , possibly by disturbing efficient TCR recognition of the B*2705:pVIPR complex. In contrast, negative selection ofpVIPR-reactive Tscells would be expected to remain unobstructed in B*2709 individuals, where pVIPR is presented conventionally by HLA-B27 molecules. A more detailed account ofCTL reactivity against B*2705 and B*2709 is provided by M.T. Fiorillo and R. Sorrentino in another chapter of this volume. We analyzed also an additional peptide with a high degree of sequence similarity to pLMP2 and pVIPR, termed pGR (RRRWHRWRL; Table 1). It is derived from glucagon receptor which is, like vasoactive intestinal peptide Type 1 receptor, another member of the family of secretin receptors. A feature distinguishing it from the other two peptides is the presence of pHis5, replacing pArg5. We had the, in retrospect, naive expectation that this amino acid exchange would preclude presentation ofpGR in the NC conformation by the B*2705 subtype. However, to our surprise, this peptide is displayed not only by B*2705 but also by B*2709 in this unusual binding mode (Figs. 21,J; 31,]).
187
Implications ofStruetural and Thermodynamic Studies ofHLA-B27 Subtypes
In addition, extensive conformational dimorphism ofpGR is again a feature characteristic for the B*Z705structure (Figs. ZI; 31). although pGR exhibits also a moderate degree ofconformational dimorphism when bound to B*Z709 (Figs. Z); 3)). In B*Z705. it affects both the Co-backbone as well as several side chains of the peptide. while in B*Z709. only the side chains of pTrp4. pArg6. pTrp7. and pArg8 are involved. Furthermore, the conformation ofthe pGR peptide in B*Z709 is an unusual "mosaic" ofthe two binding modes seen in B*Z705:while the side chains ofpArg3 resembles B*Z705:pGR in the B conformation. pTrp4 and pTrp7 both occur in double conformations (Figs ZI,); 31,)) . We observed also that pHis5 is differently co-ordinated within the binding groove in the A and B conformations ofB*Z705:pGR. although two water molecules mediate the contact to Asp 116 in both cases.' ? In contrast, there are neither direct nor indirect (water-mediated) contacts between pHis5 and His116 in the B*Z709 subtype. For both complexes , the high (B*Z705: 1.50 A; B*Z709: 1.10 A) resolutions permitted a clear-cut identification of all binding modes. Therefore. there can be no doubt that two distinct conformations of the Co-backbone are observed in the B*Z705. but not in the B*Z709 subtype. In conclusion, conformational dimorphism of certain bound peptides appears to be a feature characterizing B*Z705, but not the B*Z709 subtype. This is clearly a consequence of F pocket polymorphism: the presence of Asp116. but not His116. at the floor of the binding groove must be responsible for this highly unusual, subtype-dependent structural feature.
HLA-B27 Subtype- and Peptide-Dependent Molecular Mimicry We have pointed out in the introduction to this review that the existence ofmolecular mimicry between a foreign and a self-peptide is a central postulate ofthe arthritogenic peptide hypothesis. because it will allow CTL cross-reactivity. The results obtained by analyzing the pLMPZ, pVIPR and pGR peptides in complex with B*Z705 clearly argue in favour ofextensive molecular mimicry. but only when pVIPR-NC and pGR in the B conformation are considered (Fig. 4A.C,E). The most important prerequisite for molecular mimicry is probably that all three peptides exhibit the unusual N C conformation in which Asp 116 is contacted either directly (pLMPZ, pVIPR-N C) or indirectly (pGR-B) (Table Z). Furthermore. TCR-accessible. exposed side chains of the peptide that exhibit structural equivalence between the three peptides include at least pArgl and pTrp4 but possibly also pArg6 because of its considerable flexibility in pGR. In addition. the surfaces above the peptide binding groove-embedded residu es pArg3 /Lys3 . pArg5 /pHis5 and pTrp7/Leu7 are comparable." As observed previously.JM the similarity is most pronounced around the N-terminal halves of the pVIPR and pLMPZ peptides and thi s is also true for the pGR pepride.?" In addition. the electrostatic surface properties ofthe three complexes in B*Z705are similar as well, particularly for the regions sur round ing the N-terminal halves of the three peptides (Fig . 4B.D.F) .
Table 2. Comparison of the conformations of selected peptides in HLA-B27 subtypes HLA-827 Subtype
Peptide
8*2705
8*2709
8*2704
8*2706
8*2703
m9
CCCC NC,d
CC CC CC CC NC-A , NC-B
n.d." n.d.
n.d. n.d.
n.d. n.d.
CC CC
CC CC
CC NC
n.d.
n.d.
n.d.
TIS pLMP2 pVI PR pGR
50% CC, 50% NC NC-A , NC-8"
-Canonlcal conformation of the peptide. bnot done. -noncanonical conformation of the peptide. dpeptides showing extensive molecular mimicry in a given HLA-B27 subtype are highlighted in bold script. eN C-A and -B denote the two noncanonical conformations of pGR that are observed in B*2705 and B*2709.
188
Molecular Mechanisms ofSpondylorthropathies
Figure 4. HLA-BZ7 subtype-dependent molecular mimicry. Surface representation of B*Z705 complexed with pLMP2 in green (A), pVIPR-NC in yellow (C) and pGR-B in magenta (E), as "viewed" by an approaching TCR. The electrostatic surface potential of B*2705 in complex with pLMPZ (B), pVIPR-NC (0) and pGR-B (F), is also shown, coloured red and blue for negative and positive potential , respectively, while grey areas are uncharged .
In contrast, a comparison ofthe three peptides in the context ofB *2709 (Loll et al, manuscript in preparation; see also ref. 38 for a detailed comparison of pLMP2 and pVIPR in complex with B*2709) reveals that molecular mimicry is much less pronounced, in particular because pLMPZ and pVIPR are presented canonically (CC conformation), while pGRis noncanonically displayed (NC binding mode) by this subtype (Fig. 2F,H,J) (Table 2). Furthermore, the side chains of the
Implications ofStructural and Thermodynamic Studies ofHLA-B27 Subtypes
189
residues at pS and p6 point into entirely different directions, making CTL cross-reactivity much less likely than in B*270S (Figs. 2F,H,]; 3F,H,J). Although these studies lend molecular support to the existenceofHLA-B27 subtype-dependent molecular mimicry, preliminary structural analysesofHLA-B27:peptide complexes from the other pair ofsubtypes exhibiting differential association with AS, B*2704 and B *2706 (refs. 42,43 and unpublished observations), indicate that the situation may be more intricate (Table 2), as pLMP2 and pVIPR are presented by both subtypes in a canonical binding mode. In case ofB*2704, the differences between the two structures are most notable at positions p6 and p7, while the surface encompassing the region around pI to pS is similar. Therefore, in principle, the AS-associated B*2704 subtype could permit molecular mimicry between pLMP2 and pVIPR. In B*2706, a comparison of the corresponding two structures is made difficult since the side chains ofpArgS and pArg6 are not visible in case ofbound pVIPR, possibly because ofincreased flexibility,precluding a comparison ofthis region ofthe molecule with that ofthe B*2706:pLMP2 complex. Although the region from pI to p4 appears similar, resembling that seen in case of the B*2704 structures, we noted, however, that Asp114 contacts pArg3 (in pLMP2) via a short salt bridge, while Asp 114 is only loosely connected by a water molecule to the amino group ofpLys3 in one of the two conformations exhibited by this residue in B*2706:pVIPR. These structural differences in the two B*2706 complexes could lead to a higher flexibility ofthe N-terminal half of the pVIPR peptide than in case ofpLMP2, but it is impossible to predict whether functional consequences will ensue . Therefore, we cannot provide a definitive answer regarding molecular mimicry in the B*2704/B*2706 pair of subtypes (refs. 42,43 and unpublished observations). The unpredictability even of gross peptide conformations in HLA-B27 subtypes is further exemplified by our preliminary studies ofthe complexes ofB*2703 with pLMP2 and pVIPR (ref 41 and unpublished observations).Although this subtype differsfrom B*270Sonly in the exchange ofTyrS9 by HisS9, both pLMP2 as well as pVIPR are differently presented (Table 2). Contrary to our expectation, pLMP2 is found in the CC binding mode, while pVIPR is displayed in a single N C conformation. In light ofthe results with B*270S, these binding modes can currently not be explained since B*2703 shares the HisII4/AspII6 constellation which was envisaged by us to be instrumental in leading pArgS to engage in a strong salt bridge to AspII6 (Fig. 2E,G). Perhaps equally surprisingly, the TyrS9His exchange does not appear to have major consequences for the binding of the N-terminal amino acid of the peptide, since the highly conserved architecture of the A pocket 18.36 is nearly perfectly retained: the hydroxyl group ofTyrS9 is merely replaced by a water molecule which contacts H isS9 but otherwise engages in comparable contacts.
Thermodynamic Analyses ofSelected HLA-B27: Peptide Complexes X-ray diffraction data are obtained from crystals that are kept at very low temperature (100 K) during the analyses. Therefore, the question arises whether structural findings (e.g., a dual peptide conformation in B*270S:pVIPR complexes) correlate with CTL cross-reactivity which is assessed at 3T C. It is principally possible that pLMP2, pVIPR or pGRexhibit conformations at physiological temperature that are not identical to those found in the crystals grown at IS'C. However, as comparative diffraction studies do not support this view,55.56 we regard it as most unlikely. The issue of structural identity or similarity of these molecules at different temperatures and states might eventually be resolved with dynamical studies at physiological temperature. Nuclear magnetic resonance analyses ofthe complexes could provide a very comprehensive answer, but such experiments are extremely demanding in terms of the required amounts of material and the necessary differential labelling ofthe components. As a first approach, we have used a variety of spectroscopic techniques with which soluble HLA-B27:peptide complexes can be analyzed 37.44-47.49 These included circular dichroism (CD) specrroscopy'" to study changes in secondary structural elements such as a-helices and (3-pleated sheets,37M.58 time-resolved fluorescence depolarization and the temperature-dependent fluorescence emission of endogenous rryprophans" for the analysis of real-time peptide dynamics and the primary event in complex dissociation 37.45,46.49 and isotope-edited infrared spectroscopy'v"
190
M olecular M echanisms ofSpondy lorthropathies
to assess the conformational dynamics of the component protein chains of the complexes." In addition, we empl oyed differential scanning calorimetry (DSC)62to gain an insight into the melting characteristics of the different chains of HLA-B27 molecules37•44 and an in silico technique, molecular dynamics simulations, to probe the dynamics ofHLA-B27:peptide complexes on the nano second scale.45.47 All analyses were carried out in a comparative fashion with th e B*2705 and B'2709 subtypes complexed with selected peptides. Here, we will restrict ourselves mainly to a brief description of CD and DSC experiments with free 132m and compl exes bound to the peptides m9, TIS, pLMP2, pVIPR and pGR as well as slOR and sl OG (replacement of pArgl in slOR by pGlyl) (Fig. 5: Table 3) to clarify some of the information th at can be gained from th ermodynamic studies. MHC class I molecules are compli cated entities in terms of thermodynamic analyses, but in particular th eir stability is of fundament al significance for their fun ction in the context of immune respon ses. One objective of our studies has therefore been to destabilize HLA-B27:peptide complexes by elevating the temp erature and to assess how this affects structu ral features of the complexes such as their content ofa-helices and 13-pleated sheets as well as peptid e binding. The principle find ings ofsuch analysescan be explained parti cularly easily when free 132m is considered, as it con sists ofjust a single domain exclusivelycomposed of13-strands and was expected to unfold
-5
l N§ .
25 ' g> 20
-0
't>
-7
)
Ol
10 ~ 5
"I
~
-9
~
15
'5
-8
~
I
0"
/. -10 10 20
30
40 50 60 70 Temperature ('C)
eo
30 B'2705:pGR
'g> 20
90 100
'5
I
30
40 50 60 70 Temperature ('C)
eo
30
T..2
90
D
' g> 20 't>
15
)
15
10
'5
10
5 0
}
10
20
~25
't>
~
0 -5
C
~25
)
B
~
't>
0
30
A
~
20
30
40 50 60 70 Temperature ('C)
eo
90 100
5 0 10
20
30
40 50 60 70 Temperature ('C)
eo
90 100
Figure 5. Thermal unfold ing of HLA-B27:peptide complexes. Circular dichroism spectroscopy (A) and differential scann ing calor imetry (B-D) were employed to assess the unfold ing behaviour of B*2 705:pGR and B*2709:pGR. A) Changes of the molar mean residue ellipt icity [el measure d at 218 nm in dependence on temperature ; B) Experimental excessive heat capacity cur ves of B*2709 :pGR (soli d line), B*2705 :pGR (dashed li ne) and Ihm (dotted line); C,D) Deconvolution of the experimental excessive heat cap acity curves of the B*2705:pGR (C) and B*2709:pGR (D) complexes. Decon volution result s in two (C) or three (D) two-state transitions fo r B*2705 :pGR or B*2 709:pGR, respect ivel y, indica ted by Tm1, Tm2 and Tm3 . The th ick solid line and the dashed thin line represent the experiment al excessive heat capacity cur ves and the theoretical fit cur ves, respective ly.
96 ± 7
87 ± 1
53.3 ± 0.1
42.7 ± 0.1
B*2705:s10G
B*2709:s10G
50.9 ± 0.9
57.9 ± 1.1
51.9 ± 0.3
58.9 ± 1.3
55 ± 5
142 ± 7
95 ± 24
105 ± 3
135
114
121
114
126
140
83 ± 12
76 ± 4
128 ± 18
134
6H2
40 .0 64.6 ± 0.8
273 301 259 246 326 211 75 ± 9
70 154 ± 5 77 ± 6 88±2 69 ± 2 75 ± 9
64.0 ± 1.0 63.8 ± 0.5
66.2 ± 0.6
63.8 ± 0.4
62.7 ± 1.8
62.0
49.0
45.0
58.0
54.1 ±0.3
59.4 ± 0 .5
59.2 159
64.1
62.8
280
65.5
51.1
55.1
159
285
294
61.3 ± 2.4
58.2 ± 1.8
49.0
63.0
r, (CD)
299
71
67
231
203
127 ± 1 148 ± 27
303
316
6H ,o'
79 ± 2
182
6H3
185
66.8
62.0
63.3
66.0 ± 1.0
62.9 ± 0.2
64.9 ± 0.2
70.4
1m3
Measurements we re done by di fferential scanning c alo rimetry or circular di chroi sm spec trosc o py (CD). Tm-valu es are give n in 'C, enthalpy (&H) -va lues in kcal/mo l; buff er: 10 mM so di um pho sph ate, p H 7.5, 150 mM NaCl ; Ccomple, - 3-4 !-1M (see refs. 37,44 fo r fur th er exp erimental d etai ls).
~m
74 ± 7
49.4 ± 2.0
B*2709:s10R
B*2705:s10R
58.0
B*2709:pG R
96
58.8
B*2705:pG R
52.7
57.4
B*2709:pVIPR
53.8 62.8
88
48.3
B*2709:pLMP2
56.6
B*2705:pVIPR
87
52.6
59.1 ± 4.7
B*2709:TI5
B*2705:pLM P2
58.5 ± 2.1
B*2705:TI5
96 ± 15
56.2 ± 0.4
52.0 ± 0.8
1m2
B*2709:m9
6H l 67.0
1ml
B*2705:m9
Complex
Table 3. Thermodynamic properties of selected HLA-827:peptide complexes as assessed by differential scanning calorimetry and CD spectroscopy
~
~
I .....
~
~ -s
i:' ....
'l
N
6:,
::...
~
~
~
~
;:
~
II ;:;.
;.
~
"II
"~
;;J
l:>...
II
..... ;.
I:;
~
;:
~
I ~...
<;'
"':::-g"
192
Molecular Mechan isms ofSpondylorthropatbies
at high temperature in a single two-state transition (state 1: folded; state 2: unfolded). CD spectroscopy reveals that this protein possessesa melting point of64.6 "C , in excellent agreement with the value of63.8" C obtained by DSC (Table 3). The enthalpy (AH) value of75 kcal/rnol reflects the amount ofenergy needed to unfold this protein under the conditions employed. When such analysesare carried out with trimeric HLA-B27: fhm :peptide complexes,dissociation of a given complex and melting of H C and ~ 2m may occur sequentially or simultaneously and HC domains could unfold in separate or interdependent transitions. Therefore, the results are difficult to interpret. Figure 5 provides a characteristic scenario (see also Table 3 for details) . CD spectroscopy (Fig. SA) gave T, (half transition temperature) values of54.1"C or 59.4"C for the complexes ofB*2709 or B*2705 with the pGR peptide, respectively, indicating lower stability of B*2709:pGR than ofB*2705:pGR. DSC supports this conclusion, but provides a more extensive picture of the processes governing unfolding in the two complexes. In case ofB*2709:pGR, the melting behaviour is characterized by a curve exhibiting two partly overlapping peaks (Fig. 5B), with the melting temperature ofthe first peak considerably below that of ~2m. Deconvolution of this experimental excessive heat capacity curve (Fig. 5D) (see refs. 37,44 for more information) demonstrates that the B'2709:pGR complex melts in at least three phases. In the first step (Tm 1), a partial unfolding of the HC permits the dissociation of the bound peptide, while the second transition ('[,112) leads to complete unfolding of the HC and separation from the still folded ~2m. Finally, the last transition, supported also by the characteristic Tm3 and AH3 value, is due to unfolding of ~2m (compare with the respective values for free ~2m). In caseofthe more stable B*2705:pGR complex, the melting behaviour is characterized by only two partly overlapping transitions (Fig. 5C; Table 3): initially (Tm2) , at least the HC partly unfolds, while the second transition (Tm3) leads to complete unfolding of the entire trimeric complex, as indicated by the considerably higher value ofAH3 (compare with that offree ~2m). These scenarios are supported by various additional results and considerarlons.Fov" Table 3 shows that the unfolding pathways ofthe various HLA-B27:peptide complexes follow distinct routes. Whereas complexes with TIS show thermodynamic equivalence, there are several cases in which the differences are subtype-dependent (m9, pGR, sI0R).In spite of very similar structures (Figs. 2A,B ; 3A,B), we were able to show that the complexes ofB*2705 and B*2709 with the m9 peptide exhibit completely different thermodynamic properties," a consequence of the different co-ordination ofpLys9by Aspl16 (B*2705) or Asp77 (B*2709).36.44-46 However, there are also notable differences that are clearlypeptide-dependent. For example, the complexeswith pLMP2 are always considerably lessstable than complexes with pVIPR (Table 3) and this has recently also been confirmed by an analysisofthe B*2703, B*2704 and B*2706 subtypes (unpublished results). Our structural analyses did not provide an obvious answer, but recent spectroscopic experiments and pK. calculations showed that the relative instability ofHLA-B27:pLMP2 complexes is probably due to long-range electrostatic interactions between a peptide residue and glutamate residues within the B pocket of the molecule, leading to selective destabilization of pArg2 (Narzi et al, manuscript submitted). Furthermore, a comparison ofthe unfolding pathways ofcomplexes with s1ORand slOG reveals that the exchange ofpArgl to pGly 1 affects the unfolding behaviour ofthe respective complexes in a subtype-independent manner (Table 3 ).44 In some cases,the unfolding of a complex begins already near physiological temperatures. These results demonstrate that it is important not to rely on a single technique such as X-ray crystallography when biochemical and biophysical properties ofMHC class I complexes are to be correlated with functional studies. P
Conclusions Our results with the differentially AS-associated subtypes in complex with pLMP2, pVIPR and pGR have also implications for the initial stages ofAS pathogenesis. The crystallographic studies reveal that there is clearly a structural basis for molecular mimicry. In turn, this provides also the rational to explain the pattern of CTL cross-reactivity that has been observed when studying the B*270S andB*2709 subtypesY .38.4Q It must be pointed out, however, that there is as yet no proof for the involvement of the three peptides in AS pathogenesis since it is principally possible that
Implications 0IStructural and Thermodynamic Studies tfHLA-B27Subtypes
193
the cross-reactive CTL which appear as "specific" for pVIPR, in particular in patients suffering from AS,34 are indeed directed against a hitherto unknown peptide that might share structural properties, but not necessarily sequence similarity, with pV1PR.38.40 Nevertheless, apart from HLA haplotype-dependent eff"ects,6.7it seems to us that two theories, the arthritogenic peptide hypothesis and the 132m-deposition hypothesis,"!' are best in line with theHLA-B27 subtype-dependency ofAS and the results presented here, while the other suggested scenarios of AS pathogenesis9.10.12.13 consider exclusively the B"270S subtype and fail to explain why e.g., B "2709 should not be associated with AS, despite its extreme similarity with the former. The likely involvement of a peptidase encoded by theARTSI gene in AS pathogenesis'> further supports the arthrieogenic peptide hypothesis ofAS pathogenesis, since the trimming ofpep tides within the endoplasmic reticulum is an important prerequisite for their efficient display by HLA class I molecules. In addition, new findings in transgenic rats ascribing an important role to 132m expressions' or deposition65.66 in the emergence ofarthropathy lend support to the 132m-deposition hypothesis. However, as the elucidation ofthe pathogenesis ofAS and other spondyloarthropathies is such a multi-faceted and intricate problem, we regard it as possible that a patchwork ofseveral hypotheses might in the end explain the underlying pathomechanisms.
Acknowledgements Our studies on structural and thermodynamic properties of human histocompatibility antigens were supported by Deutsche Forschungsgemeinschafi:(SFB 449,TP B6; NA1167 112-2), European Union (EFRE 2000-2006 2iil s), VolkswagenStiftung (1/79983) and Sonnenfeld-Stiftung, Berlin. We express our gratitude to Prof W. Saenger, Berlin for his continuing interest and support ofour work on HLA-B27 molecules.
References 1. WTCCC TASe. A genome-wide scan of 14.000 nonsynonymous coding SNPs in 5.500 individuals : The wellcome trust case control consortium. Nat Genet 2007; 39:1329 -1337. 2. Khan MA, Mathieu A, Sorrentino R er al. The pathogenetic role of HLA-B27 and its subtypes. Autoimmun Rev 2007; 6:183-189 . 3. Lopez de Castro JA. HLA-B27 and the pathogenesis of spondyloarthropathies. Immunol Lett 2007 ; 108:27-33 . 4. Taurog JD . The mystery of HLA-B27: If it isn't one thing , it's another. Arthritis Rheum 2007; 56:2478-2481. 5. Brown MA. Breakthrough s in genet ic studies of ankylosing spondylitis. Rheurnarol 2008 ; 47 :132-37. 6. Fiorillo MT. Cauli A, Carcassi C ct al. Two distinct ive HLA haplotypes harbor the B27 alleles negatively or positively associated with ankylosing spondylitis in Sardinia: Implications for disease pathogenesis. Arthritis Rheum 2003 ; 48:1385 -1389. 7. Cascino I. Paladini F. Belfiore F er al. Identification of previously unrecognized predi sposing factors for ankylosing spondylitis from analysis of HLA-B27 extended haplotypes in sardinia. Arthritis Rheum 2007; 56:2640 -2651. 8. Benjamin R, Parham P. Guilt by association: HLA-B27 and ankylosing spondylitis. Immunol Today 1990; 11:137-142 . 9. Allen RL, O 'Callaghan CA , McMichael AJ er al. Cutting edge: HLA-B27 can form a novel beta 2-microglobulin-free heavy chain homodimer structure. J Immuno11999; 162:5045-5048. 10. Colbert RA. HLA -B27 misfolding: A solution to the spondyloarrhropathy conundrum? Mol Med Today 2000 ; 6:224 -230. II. Uchanska-Ziegler B. Ziegler A. Ankylosing spondylitis : A ~2m-deposition disease? Trends Immunol 2003; 24:73-76 . 12. Lurhra-Guptasarma M, Singh B. HLA-B27 lacking associated ~2-microglobulin rearranges to auto-display or cross-display residues 169-181: A novel molecular mechanism for spondyloarthropachies. FEBS Lett 2004; 24: 1-8. 13. Penrtinen MA, Heiskanen KM, Mohaparra R er al. Enhanced intracellular replication of Salmonella enteridis in HLA-B2 7-expressing human monocytic cells: Dependency on glutamic acid at position 45 in th e B pocket of HLA-B27. Arthritis Rheum 2004 ; 50:2255-2263. 14. Hammer RE, Maika SD, Ricjardson JA et al. Spontaneous inflammatory disease in transgenic rats expressing HLA-B27 and human beta 2m: An animal model of HLA-B27-associated human disorders . Cell 1990; 63:1099 -1112.
194
Molecular Mechanisms ofSpondylortbropathies
15. Lang HL , Jacobsen H, Ikemizu S er al. A functional and structural basis for TCR cross-reactivity in multiple sclerosis. Nat Immunol 2002; 3:940-943. 16. Bjorkman PJ, Saper MA, Samraoui B er al. Structure of the human class I histocompatibility antigen, HLA-A2 . Nature 1987; 329:506-512 . 17. Madden DR, Gorga JC , Strominger JL et al. The structure of HLA-B27 reveals nonamer self-peptides bound in an extended conformation. Nature 1991; 353:321-325. 18. Madden DR , GorgaJC, Strominger JL et al. The three-dimensional structure ofHLA-B27 at 2.1 A resolution suggests a general mechanism for tight peptide binding to MHC. Cell 1992; 70:1035-1048 . 19. Brewerton DA , Hart FD, Nicholls A et al. Ankylosing spondylitis and HLA-B27. Lancet 1973; 1:904-907. 20. Schlosstein L, TerasakiPI, Bluestone Ret al. High association of an HL-A antigen, W27, with ankylosing spondylitis . N Engl J Med 1973; 288:704-706. 21. Gaston H. Mechanisms of disease: The immunopathogenesis of spondyloarthropathies. Nat Clin Pract Rheumatol 2006; 2:383-392. 22. Braun J, Sieper J. Ankylosing spondylitis. Lancet 2007; 369 :1379-1390. 23. Natarajan K, Li H, Mariuzza RA et al. MHC class I molecules, structure and function . Rev Immunogenet 1999; 1:32-46. 24. Borrego F. Kabat J. Kim DK er al. Structure and function of major histocompatibility complex (MHC) class I specific receptors expressed on human natural killer (NK) cells. Mol Immunol 2002; 38:637-660. 25. Madden DR . The three-dimensional structure of peptide-MHC complexes. Annu Rev Immuno11995; 13:587-622. 26. Dangoria NS. DeLay M, Kingsbury DJ er al. HLA-B27 misfolding is associated with aberrant intermolecular disulfide bond formation (dimerization) in the endoplasmic reticulum. J Bioi Chem 2002; 277 :23459-23468. 27. L6pez-Larrea C. Sujirachato K, Mehra NK et al. HLA-B27 subtypes in asian patients with ankylosing spondylitis. Evidence for new associations. Tissue Antigens 1995; 45:169-176. 28. Gonzalez-Roces S, AlvarezMV, Gonzalez S ec al. HLA-B27 polymorphism and worldwide susceptibility to ankylosing spondylitis. Tissue Antigens 1997; 49 :116-123. 29. D'Amaro M. Fiorillo MT, Carcassi C er al. Relevance of residue 116 of HLA-B27 in determining susceptibility to ankylosing spondylitis. Eur J Immunol 1995; 25:3199-3201. 30. Nasution AR, Mardjuati A, Kunmartini S et al. HLA-B27 subtypes positively and negatively associated with spondyloarthropathy, J Rheumatol 1997; 24:1111-1114. 31. Ramos M, Paradela A, Vazquez M et al. Differential association ofHLA-B*2705 and B*2709 to ankylosing spondylitis correlates with limited peptide subsets but not with altered cell surface stability. J Bioi Chem 2002; 277:28749-28756. 32. Sesma L, Montserrat V, Lamas JR er al. The peptide repertoires of HLA-B27 subtypes differentially associated to spondyloarthropathy (B*2704 and B*2706) differ by specific changes at three anchor positions. J Bioi Chem 2002; 277:16744-16749. 33. Cauli A, Vacca A, Mamcli A et al. A Sardinian patient with ankylosing spondylitis and HLA-B*2709 co-occurring with HLA-B*1403. Arthritis Rheum 2007 ; 56:2807-2809. 34. Fiorillo MT, Maragno M. Butler R et al. CD8+ T-cell auroreactiviry to an HLA-B27-restricted self-epitope correlates with ankylosing spondylitis. J Clin Invest 2000; 106:47-53 . 35. Garboczi DN, Hung DT. Wiley DC. HLA-A2-peptide complexes: Refolding and crystallization of molecules expressed in Escherichia coli and complexed with single pepcides. Proc Natl Acad Sci USA 1992; 89:3429-3433. 36. Hiilsmeyer M, Hillig RC, Volz A et al. HLA-B27 subtypes differentially associated with disease exhibit subtle structural alterations. J Bioi Chern 2002; 277:47844 -47853 . 37. Hiilsmeyer M, WelRe K, Pohlmann T er al. Thermodynamic and structural equivalence of two HLA-B27 subtypes complexed with a self-peptide. J Mol Bioi 2005; 346:1367-1379. 38. Fiorillo MT, Riickert C. Hiilsmeyer M et al. Allele-dependent similarity between viral and self-peptide presentation by HLA-B27 subtypes. J Bioi Chern 2005; 280:2962-2971. 39. Hiilsmeyer M, Fiorillo MT, Bettosini F et al. Dual, HLA-B27 subtype-dependent conformation of a self-peptide. J Exp Med 2004; 199:271-281. 40. Riickert C, Fiorillo MT, Loll B et al. Conformational dimorphism of self-pepcides and molecular mimicry in a disease-associated HLA-B27 subtype. J Bioi Chem 2006; 281:2306-2316. 41. Loll B. Zawacka A, Biesiadka J et al. Purification, crystallization and preliminary X-ray diffraction analysis of the human major histocompatibility antigen HLA-B*2703 complexed with a viral peptide and with a self-peptide. Acta Cryst F 2005; 61:372-374. 42. Loll B. Zawacka A. BiesiadkaJ er al. Preliminary X-ray diffraction analysis of crystals from the recombinantly expressed human major histocompatibility antigen HLA-B*2704 in complex with a viral peptide and with a self-peptide. Acta Cryst F 2005 ; 61:939-941.
Implications ofStrueturaland Ihermodynamic Studies ofHLA-B27 Subtypes
195
43. Zawacka A. Loll B. Biesiadka J et aJ. X-ray diffraction analysis of crystals from the human major histo compatibility antigen HLA-B'2706 in complex with a viral peptide and with a self-peptide. Acta Cryst F 2005 : 61:1097-1099. 44. Hill ig RC. Hiilsmeyer M. Saenger W er aJ. Thermodynamic and structural analysis of peptide-and allele-dependent properties of two HLA-B27 subtypes exhibit ing differential disease association. J BioI Chern 2004: 279:652-663. 45. Pohlmann T. Beckmann RA. Grubmiiller H et aJ. Differential peptide dynamics is linked to MHC polymorphism. J BioI Chern 2004 : 279:28197-28201. 46. Winkler K. Winter A. Riickert C ec aI. Natural MHC class I polymorphism controls the pathway of peptide dissociation from HLA-B27 complexes. Biophys J 2007; 93:2743-2755 . 47. Fabian H . Huser H . Narzi D et al, HLA-B27 subtypes differentially associated with disease exhibit conformational differences in solution. J Mol BioI 2008; 376:798-810. 48. Lopez de Castro JA . Alvarez I. Marcilla M et aI. HLA-B27 : A registry of constitutive peptide ligands. Tissue Antigens 2004 ; 63:424-445 . 49. Uchanska-Ziegler B. A1exiev U. Hillig R et aI. X-ray crystallographyand dynamic studies of HLA-B'2705 and B'2709 molecules complexed with the same peptide. In : Hansen JA. ed. Immunobiology of the Human MHC: Proceedings of the 13th International Histocompatibility Workshop and Congress. Vol. I. Seattle : IHWG Press. 2006 :138-147. 50. Herberts CA. NeijssenJJ. de Haan J et aI. Cutting edge: HLA-B27 acquires many N-terminal dibasic peptides: coupling cytosolic peptide stability to antigen presentation. J Immunol 2006; 176:2697-2701. 51. Ziegler A. Loll B. Kellermann T et aJ.A cartilage-derived self peptide presented by HLA- B27 molecules? Comment on the article by Atagunduz and colleagues. Arthritis Rheum 2005 : 52:2581-2582 . 52. Atagunduz P. Appel H. Kuon Wet aI. HLA-B27-restricted CD8+ T-cell response to cartilage-derived self peptide s in ankylosing spo ndylitis. Arthritis Rheum 2005 : 52:892-901. 53. Garda-Peydro M. Marti M. L6pez de Castro JA. High Tvcell epitope sharing between two HLA-B27 subtypes (B'2705 and B'2709) differentially associated to ankylosing spondylitis. J Immunol 1999: 163:2299-2305. 54. Brooks JM . Murray RJ. Thomas WA et aI. Different HLA-B27 subtypes present the same immunodominant Epstein-Barr virus peptide. J Exp Med 1993: 178:879-887. 55. Earnest T. Fauman E. Craik CS et aI. 1.59 A structure of trypsin at 120 K: Comparison of low temperature and room temperature structure s. Proteins 1991: 10:171-187. 56. Tilton RF. Dewan JC. Petsko GA. Effects of temperature on protein structure and dynamics: X-ray crystallographic studies of the protein ribonuclease-A at nine different temperatures from 98 to 320 K. Biochem istry 1992: 31:2469-2481. 57. Pace CN. Scholtz JM. Measuring the conformational stability of a protein. In: Creighton TE. ed. Protein Structu re. A Practical Approach. 2nd ed. Oxford : Oxford lRL Press. 1997:299-321. 58. Dedier S. Reinelt S. Reitinger T et aI. Thermodynamic stability of HLA-B'2705 peptide complexes. Effect of peptide and major histocompatibility complex protein mutations. J BioI Chern 2000: 275:27055-2706 1. 59. Alexiev U. Rimke I. Pohlmann T. Elucidation of the nature of the conformational changes of the EF-interhelical loop in bacteriorhodopsin and of the helix VIII on the cytoplasmic surface of bovine rhodopsin : A time-resolved fluorescence depolarization study.J Mol Bioi 2003: 328:705-719 . 60. Fabian H. Mantele w. Infrared spectroscopy of proteins. In : Chalmers JM . Griffiths PRo eds, Handbook of Vibrational Spectroscopy. Chiche ster: Wiley. 2002:3399-3425. 61. Barth A. Zscherp C. What vibrations tell us about proteins . Q Rev Biophys 2002: 35:369-430. 62. Privalov PL. Potekhin SA. Scanning microcalorimetry in studying temperature-induced changes in proteins. Methods Enzymol 1986; 131:4-51. 63. Rudolph MG. Luz JG. Wilson IA. Structural and thermodynamic correlates of T-cell signaling. Annu Rev Biophys Biomol Strucr 2002 : 31:121-149. 64. Tran TM. Dorris ML. Sarumtira N et aI. Additional human beta2m curbs HLA-B27 misfolding and promotes arthritis and spondylitis without colitis in male HLA-B27 transgenic rats. Arthritis Rheum 2006 : 54:1317-1327. 65. Fukunishi S. Yoh K. Kamae S et aI. Beta 2-microglobulin amyloid deposit in HLA-B2 7 transgenic rats. Mod Rheumat ol 2007 ; 17:380-384. 66. Uchanska-Ziegler B. Ziegler A. HLA-B27-transgenic rats. amyloid deposits, and spondyloarthroparhies. Mod Rheumatol2008: doi : 1O.1007/ s 10165-008-0066-8 .
CHAPTER
14
HLA-B27-Bound Peptide Repertoires: Their Nature, Origin and Pathogenetic Relevance Jose A. Lopez de Castro'
Abstract eptide binding is a central biological property of HLA-B27. The availability of HLA-B27 subtypes differentially associated to ankylosing spondylitis provides a unique tool to explore the relationship between peptide specificity and pathogenetic potential. Many studies have focused on defining the nature ofsubtyp e-bound repertoires, aiming to identify peptide features that may correlate with association to disease and to find constit utive self-ligands with sequence homology to microbial epitopes. These studies were pursued on the assumption that molecular mimicry between self and foreign ligands of HLA-B27 might trigger autoimmunity. A second level ofinvolvement ofpeptide reperto ires in the biology and immunopathology ofHLA-B27 is through their critical influence on folding, maturation and cell surface expression and stability. Recent studies have emphasized the mechanisms ofpeptide loading and optimization, the interactions ofHLA-B27 with other components ofthe peptide-loading complex and the contribution of these interactions to shaping HLA-B27-bound peptide repertoires. A novel, more comprehensive and integrative, view is emerging in which the peptide binding specificity is a critical determinant of the whole HLA-B27 biology. A proper understanding of the relationships between peptide specificity and other molecular and functional features of HLA-B27 should provide the key to unveiling its pathogenetic role in spondyloarthritis.
P
Introduction HLA class I proteins are peptide-presenting molecules that constitutively bind large peptide repertoires arising mainly from the metabolic degradation of endogenous proteins, including self and viral components and present them at the cell surface. There, they can be recognized by cytolytic T-Iymphocytes (CTL) and NK cells,allowing for tolerance to self-antigens and killing of infected or otherwise modified (Le., tumor) cells. Th e class I molecule, a heterodimer of an MHC-encoded polymorphic heavy chain (He) and ~2microglobulin (~2m), cannot be properly folded and exported to the cell surface unless it is stably bound to a peptide ligand. Thus, peptide binding critically determines the biochemical and immunological properties ofHLA class I molecules. For the se reasons, analysesofthe pathogenetic role ofHLA-B27 in ankylosing spondylitis (AS) and other spondyloarthropathies (Spa) have largely focused on the peptide binding features ofthis molecule , which will be discussed below.
HLA-B27 Polymorphism and Association to AS Like other serologic specificities, HLA-B27 includes a large number of allelic variants, or subtype s, unequally distributed among human populations, that differ among each other by 1 or 'Jose A. Lopez de Castro-Centro de Biologia molecular Severo Ochoa, Consejo Superior de Invest igacion es Cientificas a nd Universidad autonorn a de Madrid Cantoblanco, 28049, Madrid, Spain. Email: aldeca stroescbrn.uam.es
Molecular Mechanisms ofSpondyloarth ropathies, edited by Carlos Lopez-Larrea and Roberto Dfaz-Pefia . ©2009 Landes Bioscience and Springer Science+Business Media .
B2719
B2718
B271 7
B2716
B271 5
B2714
B2713 E
B2711 B271 2
B2710
B2709
B2708
F
5
N
T
T
T
N
N
T
T
Y 5
5
5
5
5
5
B2706
B2707
5
T
D D
B2704
B2703
H
A
N
C A K
B2702
E
Y N
H Y
N
N
N
I
T
A
A
L
R
R
R
L
G
G
G
R
I
T
5
5
N
I
R
W T
L
L
V
H
H
Y
N
N
D
H
Y
H
Y
Y
D
R
R
5
T
W
E
E
E
E
E
V
L
T
E
E
T
Q
G
G
A
continued on next page
Y
59 63 67 69 70 71 74 77 80 81 82 83 94 95 97 103 113 114 116 131 143 147 152 156 163 171 177 178 180 211
B2701
B2705 A
-20 9
Table 1. Amino acid sequence polymorphism among HLA-B27 subtypes
'-l
1''0
~
::I
~
"1>-
":>::J
"~.
s
~
~
~
.....
::I
l:>
a ~:::I
i:
~ .il
":;.
~
~
~.
"~..,
:>::J ~
~
...
"0:)
~
.....
I ~::I
I ~ ~
::...
~
Y
Y
R G
I
I
R
5
Adapated from the HLA database at: http://www.ebLac.uk/imgt/hla/.
82738
82736 82737
82735
82734 S
S
N
G
82733
A
R
S
S
I
N
S
R
82732
S
Q
T
82731
N
N
T
82730
82729
82728
S
5
82726
82727
5
82725
Q
5
Y 5
N
T
82724
T
82723
F
5
N
S
82721
H
H
H
H
H
N
D
N
N
N
D
N
F
Y
Y
Y
Y
Y
Y
R
R
R
R
R
5
S
L
E
E
E
E
W
L
T
L
H
D
K
E
59 63 67 69 70 71 74 77 80 81 82 83 94 95 97 103 113 114 116 131 143 147 152 156 163 171 177 178 180 211
82720
-20 9
Table 1. Continued
I
....
~
~.
S-
~
Ol
S-
...~
~
~
"
~
'" ~
~.
~
... ~ ~
~ 1:;-
~ ~
~
HLA-B27-Bound Peptide Repertoires: Their Nature, Origin and Pathogenetic Relevance
199
few amino acid changes (Table 1) and have probably evolved from the prototype B"Z705 by gene conversion and possibly other genetic events, such as point mutations. The former mechanism, which among HLA genes frequendy involves limited regions ofthe coding sequence, can introduce multiple synonymous and nonsynonymous base changes in a singleevent, resulting in the generation ofwide allelic repertoires inter-related by patchwork-like patterns ofpolymorphism (Table 1). Most ofthe HLA-BZ7 subtypes are found at very low frequency and their association to AS is unknown. Most studies have focused on relatively few subtypes that are present, at least in some populations, with sufficient frequency for case-control studies. B"Z705, B"Z70Z, B"Z704 and B"Z707are AS-associated subtypes. B"Z705 has a wide ethnic distribution and is the predominant subtype in Caucasians, Eskimo and North American Indians. B"Z70Z is predominandy found in Caucasians, where it accounts for about 5-10% or more of the HLA-BZ7-positive individuals. B"Z704 is the predominant subtype in Far Eastern Asian populations. B"Z707 is found with relatively low frequency in multiple European Caucasians, Indian and Jewish populations and is generally associated to ASI with a reported exception.' B"Z703, B"Z706 and B"Z709are generally not found associated to AS. However, B"Z703 is only prevalent in Black African populations where B"Z705 is also not associated to AS, presumably due to the concomitant presence ofnonBZ7 protective genes.' Therefore the nature ofB"Z703 as a pathogenetic factor for AS is unclear. B"Z706 is mainly found in populations ofSoutheast Asia and the Pacific only among healthy individuals whereas other subtypes are associated to AS in these populations.v" B"Z709 is only prevalent in Sardinia and found at much lower frequency in continental Italy. It is generally considered as a nonpathogenetic subtype, based on population srudies. ?" but some recent reports are compatible with a role ofnonBZ7 genes in modulating the association ofthis subtype to AS. For instance, in Sardinia, where B"Z705 , but not B"2709, is associated to AS both subtypes are in distinct haplotypes ," Some polymorphisms mapping at HLA-E and other NK-related genes were reported to be selectively increased in the B"2709 haplotype. These polymorphisms were also increased, in the other chromosome, among B"Z705 healthy Sardinian individuals, relative to patients.'? However, the protective role ofnonBZ7 MHC genes on B"2705 should be confirmed in other populations where these polymorphisms are likely to occur. B"2709 was found in a few patients with Spa lacking axial Involvement'v':" and, more recently in 2 AS patients,13.14 although both of them carried additional genetic factors predisposing to AS. Although the current genetic and epidemiological evidence is not conclusive, the possibility that lack ofassociation ofB*Z709 to AS in Sardinia may be determined or largel y influenced by nonB27 protective genes cannot be ruled out.
Peptide Specificity ofHLA-B27 Subtypes The availability ofclosely related subtypes differentially associated to AS stimulated studies on the nature ofsubtype-bound peptide repertoires, aiming at correlating peptide binding specificity with disease association. These studies were based on the assumption that the peptide specificity of HLA-BZ7 could be a critical feature determining its pathogenetic role, either through differential presentation ofan arthritogenic peptide I; or by differentially modulating other molecular features such as folding, assembly and export. The peptide binding features ofmajor HLA-B27 subtypes are described below.
B*2705 This is the prototype HLA-B27 molecule and the one for which more information regarding its peptide repertoire is available. Mass spectrometry analyses have provided a reliable estimation ofthe size distribution ofB"Z705ligands as well as the amino acid sequence ofnumerous natural ligands. Together these analyses draw a neat picture ofthe peptide specificity of this allotype and the nature ofits constitutive peptide repertoire. An MS-based estimation ofthe molecular mass of more than 1000 B"Z705 ligands showed an approximately Gaussian distribution ofpeptide sizes with an average molecular mass of 1176 Da,16corresponding well to the known restrictions in peptide size imposed on HLA class I molecules by their peptide binding groove, which optimally binds ligands of9-1O residues with a range of essentially S-13-mers. From early studies'? it was
200
Molecular Mechanisms ofSpondyloarthropathies
established that a major peptide motif of B*2705 ligands was R2, which was imposed by the B pocket ofthe molecule, in particular by the E45 residue, at the bottom ofthis pocket. I S Two other anchor motifs were identified: aliphatic and aromatic residues at position (P)3 and basic, aliphatic and aromatic C-terminal (PC) residues. The structural basis for these restrictions was established from X-ray studies, 19.20 which are the subject ofanother chapter in this book (see the article by A. Ziegler and B. Uchanska-Ziegler) and will not be discussed here. A much lower, ifany, restriction was found at other peptide positions. This was interpreted as a consequence of their exposure at the molecular surface and low contribution to peptide binding. Extensive peptide sequencing ofB*2705ligands allowed for a refinement ofthe peptide motifs and a reassessment ofpeptide binding features." A somewhat surprising outcome ofthis analysis is that skewed residue usage was found at all peptide po sitions, suggesting that the nature of secondary and "non-anchor" residues was actually contributing to peptide binding and selection. These studies confirmed the nearly absolute (98%) restriction for R2 and refined the previously observed restrictions at P3 and PC. The joint frequency of aliphatic and aromatic P3 residues among 173 constitutive ligands was 73.4%. The frequency of aliphatic, aromatic and basic PC residues was 37.8%, 32.5% and 28 .5%, respectively. The allowance for basic and nonpolar PC residues is a remarkable feature ofB*270S whose basis is in the plasticity oftheir interactions with pocket F residues, which include both hydrophobic (Le., L81) and acidic ones (D74, D77, D 116). This makes possible that a variety of chemically distinct peptide residues may adopt alternative binding modes." Other positions also show significant residue skewing. For instance, at PI the frequencies ofG and R were 22% and 22.5% respectively among 147 nonamers and decanters." The preference for Rl is probably due to the interactions ofits side chain with A pocket residues, as described in detail from X-ray studies." However, peptides with dibasic N-terminal sequences (i.e., RR, KR) are particularly resistant to amino-peptidases and it was suggested that the increased cytosolic stability may contribute to their high frequency among HLA-B27 ligands." The preference ofHLA-B27 for peptides with Gl has not been satisfactorily explained. It is po ssible that this residue is simply lessdisfavored than most other P1 residues, except basic ones. Some residues, notably Pro and acidic ones, are strongly disfavored at this position (frequency <0.6%). Residue skewing at "non-anchor" positions (t.e., 31.5% of I«Vat P6)21 is also not properly explained and presumably reflects a significant contribution ofthese positions to peptide binding, which remains to be defined at a molecular level. However, the contribution of positions other than the main anchor ones to peptide binding has long been known, since the predictive value of main anchor residues for HLA-A2ligands was only about 30%, but it increased up to 70% when secondary anchor positions were taken into account."
B*2702 This subtype differs from B*2705 by 3 amino acid changes (D77N, T80I, L81A) that affect the structure ofthe F pocket and modulate the peptide specificity mainly by affecting PC residue preferences. Rather few B*2702ligands have been sequenced.21.2S.26The most notorious difference with B*270S is the restriction in B*2702 for peptides with C-terminal aliphatic and aromatic, but not basic, residues. As expected from identity with B*270S in its B pocket, B*2702 ligands, like most other B27 subtypes, also have the R2 motif. An MS-based comparison of the B*2705 and B*2702-bound peptide repertoires" showed that they overlapped by about 70-80 %, which is consistent with the percentage of B*2705 ligands with basic PC residues . This suggests that residue preferences at other peptide positions might be unaltered in B*2702, relative to B*2705, although this should be confirmed by sequencing ofmany more B*2702ligands. An implication ofthe peptide specificity ofB*2702 is that, if peptide presentation is involved in AS pathogenesis it would either not be mediated by a peptide with a C-terminal basic residue or different peptides would be pathogenetic on both subtypes .
B*2703 This subtype differs from B*2705 only by the Y59H change, located in the A pocket. The most notorious influence of this mutation on peptide specificity is a significantly increased preference
HLA-B27-Bound Peptide Repertoires: Their Nature, Origin and Pathogenetic Relevance
201
for basic PI residues. These account for 77% of31 B*2703 reported ligands ." In particular, Kl accounts for 39% of the ligands. The frequency ofG 1is correspondingly decreased in this subtype (6.5%), relative to B*270S. Putative alterations in residue frequencies at other positions need to be substantiated with further sequencing ofB*2703ligands. Early studies suggested that the contribution of the peptidic R2 to B*2703 binding was larger than on B*270S. This was interpreted as a consequence ofweaker interaction ofthe N-terminus in the A pocket as a consequence ofthe B*2703 mutation, which would increase the relative contribution ofR2 to peptide binding in this subrype." The increased preference for basic PI residues, whose side chain contributes to peptide anchoring through interactions with E163, W167 and other pocket A residues.P suggests that the preference ofHLA-B27 for this N-terminal motifis dictated by structural complementarity rather than by the resistance of these peptides to cytosolic amino pepridases.P
B*2704 andB *2706 These two subtypes are closely related in structure and ethnic distribution, but are differentially associated to AS. B*2704 differs from B*270S by three amino acid changes: D77S, VlS2E and A211G. B*2706 is like B*2704 in these positions, but has two additional changes: H1l4D and D 116Y. The most important difference in peptide specificity between B*2704 and B*270S is the decreased preference for basic PC residues (6.1% vs 28.5%) , which presumably results from the D77S change in the F pocket. Additional differences in residue usage between both subtypes at secondary anchor positions were suggested from a comparison of B*2704 and B'270S-bound nonamers-! but they did not reach statistical significance and they need to be substan tiated with further peptide sequences . From early studies using pool Edman sequencing it became clear that B*2706 differed from B"2704 by an apparently absolute restriction for nonpolar C-terminal residues." The influence of B*2706 polymorphism on peptide specificity was examined in more detail in a comparative study with B·2704.28An MS-based comparison ofnearly 850 B*2704 and B*2706ligands showed an overlap close to 90% between both peptide repertoires. Sequencing of 38 B*2706 ligands confirmed the restricted use of non polar aliphatic and aromatic residues , with exclusion of Y and basic ones." This feature is attributable to the Dll6Y change, which significantly increases the hydrophobicity of the F pocket. Due to the high overlap with B'2704, the effect ofB*2706 polymorphism on modulating residue preferences at other peptide positions could only be assessed through the analysis ofB'2706ligands not bound to B·2704. Only 5 such ligands were sequenced." A distinctive feature of these peptides was the presence of a polar, mainly basic, P3 residue in all of them and the concomitant presence of a basic P(C-2) residue in all but one of them. These effects are probably due to the H 114D change in B'2706, which presumably favors basic residues at both P3 and P(C-2).
B*2709 This subtype differs from B'270S by a single amino acid change, D 116H, in the F pocket. Due to the drastic chemical change and to its location in the peptide binding site, this polymorphism has a significant influence on peptide specificity. The earliest peptide studies established that this subtype differs from B'270S by its strong preference for nonpolar PC residues ." A comparative analysis revealed that B'270S and B'2709 shared with each other around 79% and 88% of their peptide repertoires, respeccively." Although the preference for nonpolar PC residues in B'2709 was confirmed, this subtype also had some allowance for C -terminal R and Y, albeit with much smaller frequency than B·270S. The suboptimal binding ofbasic PC residues (K) and Y to B'2709 has been directly confirmed by X-ray crystallographic and thermostabiliry analyses." It was remarkable that some B'2709 ligands with nonpolar PC residues were not found in the B'270S-bound pool in spite ofthe fact that such residues are also appropriate for B·270S. This suggests that the binding features of these peptides to both subtypes might be different, but the basis for this differential behavior remains to be elucidated.
202
Molecular Mechanisms ofSpondyloarthropathies
B*2707 This subtype differs from B02705by 5 amino acid changes: N97S, YIl3H. H114N, D II6Y and Sl3IR. Although it is generally found associated to AS, it shares with non-AS associated subtypes the lack of D 116 and has the same change as B02706 at this position. Like in this subtype, the effect ofthe D II6Y change in B02707 is to confer a highly restricted specificity for peptides with nonpolar PC residues. a feature that was noted in the first peptide studies on this subrype." Thus, the simple correlation between restriction for C-terminal nonpolar residues and lack ofassociation to AS among HLA-B27 subtypes revealed itselfuntenable. Indeed. a systematic comparison ofthe peptide repertoire ofB02707 with B02705. B02706 and B02709 revealed further complexity. I First , the B02705 and B02707-bound peptide repertoires differed from each other by about 50%. Both peptide repertoires differed mainly by the mentioned differences in PC residue usage. However, minor differences at other positions hinted at a subtle modulation of residue preferences at secondary anchor positions whose confirmation and further definition would require many more peptide sequences than those currently available. In spite ofthe similar specificity for non polar PC residues, B02707 showed significant differ ences in its peptide repertoire with both B°2706 (39% overlap) and B02709 (around 50% overlap). This was due to a significant influence ofB02707 polymorphism on residue preferences at multiple secondary anchor positions,' a modulation that could be explained by the nature of the amino acid changes and their location in pockets ofthe peptide binding site involved in interaction with secondary anchor po sitions.
B*2701 This subtype differs from B02705 by 3 amino acid changes: D74Y, D77N and L8IA. It might by a susceptibility factor for AS. since this subtype has been found in individual patients. but, due to its low frequency, no case control studies have been carried out. From a biochemical standpoint this subtype has special interest because, unlike B02705 and other major subtypes, B0270I has a more relaxed residue specificity in the B pocket. allowing not only for R2, but also, to a significant extent, for Q2 .31 The molecular basisfor this feature are in a long-range effect ofK70. as determined by molecular modeling and site-directed mutagenesis." In B02705 and most other subtypes, the side chain ofK70 is kept away from the B pocket through its interaction with D74. In the absence ofthis residue, a new rotameric state ofK70 is possible in which the side chain is switched towards the B pocket, allowing for binding of the peptidic Q2 residue. Among the known HLA-B27 subtypes, B0270I is unique in having K70 and lacking D74 at once. The D74Y change occurs in 3 other subtypes: B027I8, B02723 and B02729, but all of them lack K70 (Table 1). The peptide specificity ofthese subtypes is unknown. Although few B0270 I ligands have been identified," the PC residue specificity ofthis subtype seems to be similar to that ofB02702, allowing for aliphatic/ aromatic residues , bur no basic ones.
T-Cell Presentation ofShared Ligands by HLA-B27 Subtypes Since HLA-B27 subtypes bind overlapping peptide repertoires. an issue ofgreat immunological relevance is to what extent the shared ligands are presented by different subtypes in an antigenically similar conformation. This question has been approached in two ways: by correlating alloreactive T-cell crossreaction with peptide sharing and by directly comparing the three-dimensional structure of single peptides in complex with various subtypes. Whereas the first approach is indirect. it allows a global assessment of epitope conservation among subtypes because alloreactive CTL recognize a highly diverse set of constitutive peptide ligands.33.34 Thus , virtually each CTL clone can be assumed to recognize a distinct peptide. Among allotypes with significant peptide sharing. the majority ofcross-reactive T-cell epitopes may be assumed to be mediated by recognition ofthe same peptide in the various subtypes. Although the peptides involved in clonal allorecognition are generally unknown and very difficult to identify, this assumption was formally demonstrated at least in one insrance.v The second approach is much more precise and direct, but it can only be applied to few individual ligands. Early studies from our laboratory showed a significant degree of
HLA-B27-Bound Peptide Repertoires: 1heir Nature, Origin and Pathogenetic Relevance
203
Table 2 . Peptide and alloreactive T-cell epitope sharing among HLA-B27 subtypes Subtypes Compared
Peptide Sharing (%)'
T-Cell Epitope Sharing (%)'
B*2 705/B*2701
nd
29 36/nd
B*2705/B*2702
73/8]2 6
B*2705/B*2 703
nd
3336/5326 48/78 36
B*2705/B*2704
nd
1236/ 36 82
B*2705/B*2706
nd
836/nd
B*2705/B*2707
59/ 72 1
B*2 705/B*2 709
40/45 ' 79/8 8 16
B*2702/ B*2 709
nd
80/9083
B*2704/B*270 2
nd
4326/nd 63 82/nd
B*2 704/B*2706
88/90' 8
72 82/nd
B*2706/B*2709
39/3 21
nd
B*2707/B*2706
39/3 9' 56/461
2184/nd
B*2 707/B* 270 9
47 1/nd
' First vs Second/Second vs First. The co rrespondi ng references are give n; nd : not done.
alloreactive T-cell cross-reaction among HLA-B27 subrypes." In later studies , the percentage of alloreactive T -cell clones cross-reacting among particular subtype pairs was close to the percentage ofpeptide sharing ,1.16,26.28 suggesting that peptide ligands tend to adopt similar conformations when bound to distinct subt ypes (Table 2). Whereas this conclusion is supported for individual peptides by X-ray studies," subtype dependent conformations are also observed" (for a complete di scussion of subtype-dependent peptide conformation s see the article by A.ziegler and B. Uchan ska-Ziegler in this book).
Molecular Mimicry between Selfand Foreign Ligands ofHLA-B27 The analysesofsubtype-bound peptide reperto ires revealedthat there isenough peptide overlap and Tvcell epitope conservation among subtypes to account for peptides with arthritogenic potential whose presentation might correlate with association to AS. However, the complex modulation ofpeptide specificity by subtype polymorphism currently precludes a reliable prediction ofpeptide motifs that correlate with disease association . An alternative approach to the identification of putative arrhrirogenic peptides has been to look for molecular mimicry between self and foreign ligands ofHLA-B27. Following the observation that the amino acid sequence ofHLA-B27 has a particularly high homology with proteins from Gram-negative bacreria.P a region in the a2 domain ofthe molecule, spanning residues 168-176, was identified as showing high homology to bacterial sequences/? Although CTL from AS patients recognized the corresponding p( 168-176) peptide," this has not been found in the B27-bound peptide pool. However, two related peptides from thi s region , spanning residues 169-179 and 169-181 were identified as natural ligands of B*2705 and other HLA-B27 subtypes,42-44 but none of them showed a subtype binding specificity that correlated with association to AS. Another peptide, derived from residues 309 -320 ofthe cytoplasmic region of th e HLA-B27 HC, was repo rted as a natural ligand ofB*2705 and other subtypes. This peptide has significant homology with residues 211-222 from the DNA primase of Chlamydia tracbomatis' ? and is a natural ligand oB AS-associated subtypes, B*2705, B*2702 and B*2704, whereas it was not found in B*2706 , B*2707 or B*2709. Thus, except for its absence in B*2707, presentation of p(309-320) correlates with subtype association to AS. In recent studies
204
Molecular Mechanisms ofSpondyloarthropathies
from our laboratory, a closelyrelated chlamydial peptide wasendogenously processed and presented in vivo by B*270S in transfectant cells expressing a fusion protein derived from the chlamydial DNA primase.t" The occurrence of molecular mimicry between chlamydial and self-epitopes presented by HLA-B27 is suggested by studies in which immunization of rats with HLA-B27 endowed these animals with the capacity to elicit a CD8 T-cell response against Chlamydia trachomatis upon in vitro stimulation with this bacteria." In the reverse situation, transgenic rats exposed to Chlamydia trachomatis developed a B27-directed auto-reactive CTL response." An alternative approach to searching for microbial peptides possibly involved in B27-mediated autoimmunity is to consider that a putative arthrirogenic peptide would be suboptimally presented by AS-associated subtypes. One such example, which has been extensively studied by Sorrentino and colleagues, is VIPR(400-408) a presumed natural self-ligand ofHLA-B27 with high homology to an HLA-B27-restricted EBV-derived epitope, LMP2(236-244). VIPR(400-408) binds better B*2709 than B*270S. Both B*270Sand B*2709 individuals elicit CTL responses against the viral peptide, but only the former ones, specially AS patients, responded to the self-pepride." The functional and structural issuesconcerning these peptides are discussed in detail in other chapters of this hook (see those by R. Sorrentino and A. Ziegler) and will not be further discussed here. These studies provide sound immunological and structural evidence for the existence of subtype dependent molecular mimicry between self and foreign HLA-B27 ligands .
The Origin ofHLA-B27-Bound Peptide Repertoires: Role ofthe Proteosome The proteasome is the main protease involved in the generation of MHC class I ligands," although not the only one.50 Multiple nonproteasomal enzymes have been reported to generate particular epitopes in a proreasome-independent way.These include, for instance, tripepridylpeptidase II (TPPII),51.52 furin,53.55 cathepsin S,56 or cystein proteases." For many class I molecules nonproteasomal mechanisms playa limited role in the generation of their constitutive repertoires." However, different allotypes differ in their capacity to be expressed at the cell surface in the presence ofproteasome inhibitors, suggesting a variable degree ofproteasome dependency for peptide presentation. HLA-B27 is among the most proteasome-independent class I molecules by this crireriurn." In a recent study'" we used stable isotope tagging to distinguish between HLA-B27 ligands whose generation was impaired with proteasome inhibitors and those that were not. About 30% ofthe peprides analyzed were generated and presented by HLA-B27 in the presence of epoxomicin, a potent and specific proteasome inhibitor. Identification of multiple proteasome-dependent and independent B27ligands failed to reveal any significant difference in peptide structure, flanking sequences of the peptides or subcellular distribution of the parental proteins. A majority ofthe proreasome-lndependent ligands, other than few derived from signal sequences, arose from small and basic proteins. This was in strong contrast with proteasome-dependent ligands, whose parental proteins showed a size and pI distribution that reflected those in the human proteome. Since proteins with high content of basic residues may be degraded with significant involvement of the tryptic-like activity of the proreasorne'" this might simply suggest lack of proper inhibition of this proteasomal activity. That this was not the case was supported by the fact that both the percentage ofinhibitor-insensitive peptides and the lack ofinhibition of nearly all ofthese pep tides was unchanged even at 2.S u.Mepoxomicin, a concentration known to inhibit both the chymotryptic and tryptic activities ofthe proteasome in intact cells (Fig. 1).61 The nature of the processing pathway responsible for the generation of the proteasorne-independent ligands ofHLA-B27 and its significance for the antigen-presenting and pathogenetic features of this molecule are yet unknown.
Assisted Loading ofthe HLA-B27 Peptide Cargo HLA class I molecules acquire their peptide repertoires in the endoplasmic reticulum (ER) during their biosynthesis. This process is assisted by several chaperones that playa critical role in
HLA-B27-Bound Peptide Repertoires: Their Nature, Origin and Pathogenetic Relevance
205
Epoxomicin 0.2 JJM
1 JJM Proteasome independent
<,
2.5 JJM Proteasome dependent
/
Figure 1. Partial inhibition of the proteasome does not account for the proteasome-independent HLA-B27 ligands . The percentages of epoxomicin-sensitive (white sectors) and insensitive (gray sectors) HLA-B27 ligands were not significantly altered within a lO-fold range of the inh ib itor. At 0.2 !!M epoxomicin only the chymotryptic-like act ivity of the proteasome is significantly inhibited. At 2.5 !!M both the chymotryptic and the tryptic activity are inh lblted." The percentages are derived from analysis of 53 (0.2 !!M), 91(1 !!M) and 54(2.5 !!M) molecular species. Data are from reference 60.
the folding process and help to ensure that optimal peptides are bound before the molecule is exported to the cell surface. The peptide loading complex (PLC) is composed ofthe empty MHC classI-~2m heterodirner, calreticulin, a lectin-like chaperone, TAP, which transports peptides from the cytosol to the lumen ofthe ER , tapasin, an MHC-dedicated chaperone critically involved in peptide editing and ERpS7, a thiol oxidoreducrase." Tapasin is covalently bound to ERpS7 in the PLC. 63 In turn, this complex may form a disulfide-linked trimeric complex with the class I molecule ." The tapasin-ERPS7 conjugate is the critical structure determining optimization of peptide loading,65.66 possibly by ensuring the oxidation state ofthe disulfide bond in the a2 domain, which is located in the peptide binding groove and spatially close to the A pocket. This disulfide bond is presumably essential for the correct establishment ofthe hydrogen bonding network that stab ilizes N-terminal peptide anchoring to the class I molecule/" HLA classI allotypes differ widely in their tapasin dependency for peptide loading, a feature that has been related to the hydrophobicity ofthe peptide-bindinggroove, since it would thermodynamically favor peptide interactions, obviating the requirement for tapasin-mediated opdmizarion." HLA-B27 is, compared to other allotypes, relatively independent of tapasin." However, there are differences among the peptide repertoires ofB*270S bound in the absence or in the presence of tapasin: some peptides are predominantly bound in the presence of the chaperone and others in its absence,"? Indeed, the B*270S-bound peptide repertoire loaded in the absence oftapasin is suboptimal in terms of the overall affinity of the peptide cargo." Few studies have yet addressed tapasin dependency among HLA-B27 subtypes . B*2704 was more dependent on this chaperone for surface expression than other subtypes.72.73 However, tapasin influenced the folding, but not the export rate ofboth subtypes. In the absence ofthis chaperone B*2706 showed efficient folding and fast export rate , compared with other subtypes. This is consistent with the high hydrophobicity ofits F pocket, which might favor tapasin independency. Yet, B*2706 showed a greater tendency than other subtypes to dissociate at the cell surface in rapasin-proficient cells," suggesting that its consti tutive peptide repertoire might be lessoptimized than those ofother subtypes. It has been proposed that tapasin independence, by allowing export of suboptimal peptides, might favor peptide interchange in the secretory compartments, facilitating loading ofexternal an tigens onto HLA-B27 .69However, no correlation hasbeen found between tapasin dependency and
206
Molecular Mechanisms ofSpondyloartbropatbies
subtype association to ASn and the relationship ofthis featur e to the pathogenicity ofHLA-B27, if any, is unclear.
Conclusions Besides the specificity of antigen presentation, which could provide a possible pathogenetic mechanism based on molecular mimicry, 1S the peptide binding features ofHLA-B27 subtypes are important for their relationship to other biochemical properties. Peptide binding is a constituve feature of the MHC class I molecules. This means that proper folding and export to the surface requires the stable binding of high affinity ligands.Thus, folding , peptide binding and tapasin dependency are intimately related. HLA-B27 folding has attracted much attention in recent years due to the putative pathogenetic effects of misfolded HC74 (see the chapter by R. Colbert in this book). Suboptimal peptide loading may exacerbate HLA-B27 misfolding, leading to its accumulation in the ER, formation ofHC hornodirners" and activation ofstress responses." In transgenic rats, over-expression of~2m results in a drastic decrease ofmisfolded HC in the ER,77 demonstrating that promoting the formation ofHLA-B27 I peptide complexes issufficient to curb misfolding. In addition, HLA-B27 homodimers are expressed at the cell surface and are recognized by leukocyte receptors.":" what suggested their possible pathogenetic role (see the article by P. Bowness in this book). Formation ofHLA-B27 HC homodimers at the cell surface appears to follow dissociation ofthe B27 I peptide complexes upon endosomal recycling ,so a process that may be strongly influenced by peptide binding stability. HLA-B27 dissociation at the cell surface has also been proposed to promote ~2m release leading to its deposition in the joints , with putative pathogenetic effects.sl Thus, peptide binding is Criticallyinter-related to other molecular features that together shape the immunology and overall biology ofHLA-B27. Only a proper understanding ofthese relationships, both at the molecular and functional level, is likely to provide the key to the pathogenetic role ofHLA-B27 in spondyloarthritis.
Acknowledements The work ofthe author is supported by grant SAF200S/03188 from the Ministry of Science and Technology and an institutional grant of the Fundacion Ramon Areces to the Centro de Biologia Molecular Severo Ochoa.
References 1. Gomez P, Montserrat V, Marcilla M et aI. B'2707 differs in peptide specificity from B'2705 and B'2704 as much as from HLA-B27 subt ypes not associated to spondyloarthriris. Eur J Immunol 2006 ; 36:1867-81. 2. Varnavidou-Nicolaidou A, Karpa sitou K, Georgiou D et al. HLA-B27 in the Greek Cypriot population: Distribution of subtypes in patients with ankylosing spondylitis and other HLA-B27-related diseases. The possible protective role of B·2707. Hum Imm uno l 2004; 65 :1451-54. 3. Brown M, Jepson A, Young A er aI. Spondyloarthritis in West Africa : evidence for a nonB27 protective effect. Ann Rheum Dis 1997; 56:68-70. 4. Lopez-Larrea C, Sujirachato K, Mehra NK er aI. HLA-B27 subtypes in Asian patients with ankylosing spondylitis. Evidence for new associations. Ti ssue Ant igens 1995 ; 45 :169-76. 5. Ren EC, Koh WH, Sim D er aI. Possible protective role of HLA-B'2706 for ankylosing spond ylitis. Tissue Antigens 1997; 49 :67-69. 6. Nasution AR , Mardjuadi A, Kunmarrini S et aI. HLA-B27 subtypes positively and negatively associated with spondyloarthropathy, J Rheumato] 1997; 24:1111-14. 7. D'Arnato M, Fiorillo MT, Carcassi C et aI. Relevance of residue 116 of HLA-B27 in determining susceptibility to ankylosing spondylitis. Eur J Immun ol1995; 25 :3199-201. 8. Paladini F, Taccari E, Fiorillo MT er aI. Distribution ofHLA-B27 subtypes in Sardinia and continental Italy and their association with spond ylarthropathies. Arthritis Rheum 2005 ; 52:3319-21. 9. Fiorillo MT, Cauli A, Carcassi C er al. Two distinctive HLA haplotypes harbor the B27 alleles negatively or positively associated with ankylosing spond ylitis in Sardinia: implications for disease pathogenesis. Arrhritis Rheum 2003 ; 48 :1385-89.
HLA-B27-Bound Peptide Repertoires: Their Nature, Origin and Pathogenetic Relevance
207
10. Cascino I. Paladini F, Belfiore F et aI. Identification of previously unrecognized predisposing factors for ankylosing spondylitis from analysis of HLA-B27 extended haplotypes in Sardinia. Arthritis Rheum 2007 ; 56 :2640-51. 11. Olivieri I. Padula A, Cianco G et aI. The HLA-B"2709 subtype in a patient with undifferentiated spondarthritis. Ann Rheum Dis 2000 ; 59:654-55 . 12. Olivieri 1, Ciancio G, Padula A et aI. The HLA-B"2709 subtype confers susceptibility to spondylarrhroparhy, Arthritis Rheum 2002; 46 :553-54. 13. Cauli A, Vacca A. Mameli A et aI. A Sardinian patient with ankylosing spondylitis and HLA-B"2709 co-occurring with HLA-B"1403. Arthritis Rheum 2007 ; 56:2807-09. 14. Olivieri I. D'Angelo S. Scarano E ec aI. The HLA-B"2709 subtype in a woman with early ankylosing spondylitis. Arthritis Rheum 2007; 56:2805-07. IS. Benjamin R. Parham P. Guilt by association : HLA-B27 and ankylosing spondylitis. Immunol Today 1990; 11:137-42. 16. Ramos M. Paradela A. Vazquez M et aI. Differential association ofHLA-B"2705 and B"2709 to ankylosing spondylitis correlates with limited peptide subsets but not with altered cell surface stability. J BioI Chern 2002 ; 277 :28749-56. 17. Jardetzky TS. Lane WS , Robinson RA et aI. Identification of self peptides bound to purified HLA-B27. Nature 1991; 353 :326-29. 18. Villadangos JA, Galocha B, Garcia F er aI. Modulation of peptide binding by HLA-B27 polymorphism in pockets A and B and peptide specificity of B"2703. Eur J Immunol 1995; 25 :2370-77. 19. Madden DR. Gorga JC, Strominger JL er aI. The structure of HLA-B27 reveals nonamer self-peptides bound in an extended conformation. Nature 1991; 353 :321-25. 20. Madden DR, Gorga JC, Strominger JL et aI. The three-dimensional structure of HLA-B27 at 2.1 A resolution suggests a general mechan ism for tight peptide binding to MHC. Cell 1992; 70:1035-48. 21. Lopez de Castro JA , Alvarez I, Marcill a M er aI. HLA-B27: a registry of con stitutive peptide ligands. Tissue Antigens 2004; 63 :424-45. 22. Hillig RC , Huelsmeyer M , Saenger W et aI. Thermodynamic and structural anal ysis of peptide-and allele-dependent properties of two HLA-B27 subtypes exhibit ing differential disease association. J BioI Chern 2004 ; 279 :652-63. 23. Herberts CA, Neijssen JJ, de Haan Jet al. Cutting edge: HLA-B27 acquires many Nvrerminal dibasic peptides: coupling cytosolic peptide stability ro ant igen presentation. J Immunol 2006; 176:2697-701. 24. Ruppert J, Sidne y J, Celis E ec aI. Prominent role of secondary anchor residue s in peptide binding to HLA- A2.1 molecules . Cell 1993; 74 :929-37. 25. Rotzschke 0 , Falk K, Stevanovic S er aI. Dominant aromatic/aliphatic C-terminal anchor in HLA-B"2702 and B"2705 peptide motifs. Immunogenetics 1994; 39:74-77. 26. Marti M, Alvarez I, Montserrat V et aI. Large sharing of T-cell epitopes and natural ligands between HLA-B27 subtypes (B"2702 and B"2705 ) associated with spondyloarrhriris. Tissue Antigens 2001 ; 58:351-62. 27. Garcia F. Marina A, Lopez de Castro JA . Lack of carboxyl-rerrnmal tyrosine d istinguishes the B'2706-bound peptide repertoire from those of B"2704 and other HLA-B27 subtypes associated to ankylosing spondylitis. T issue Antigens 1997; 49:215-21. 28. Sesma L, Montserrat V, Lamas JR et aI. The peptide repertoires of HLA-B27 subtypes differentially associated to spondyloarrhropachy (B"2704 and B"2706) differ by specific changes at three anchor positions. J Bioi Chern 2002 ; 277 :16744-49. 29. Fiorillo MT. Meadows L, D'Amaro M et aI. Susceptibility to ankylosing spondylitis correlates with the C vterminal residue of peptides presented by various HLA-B27 subtypes. Eur J Immunol 1997 ; 27:368-73. 30. Tieng V, Dulphy N, Boisgerault F et aI. HLA-B"2707 peptide motif: Tyr C-terminal anchor is not shared by all disease-associated subtypes. Immunogenetics 1997; 47:103-05. 31. Garcia F, Galocha B, Villadangos JA er aI. HLA-B27 (B"2701) specificity for peprides lacking Arg2 is determined by polymorphism outside the B pocket . Tissue Antigen s 1997 ; 49 :580-87. 32. Krebs S, Rognan D , Lopez de Castro JA . Long-range effects in protein-ligand interactions mediate peptide specificity in the human major histocomparibilty antigen HLA-B27 (B"2701). Protein Sci 1999; 8:1393-99. 33. Wang \v, Man S, Gulden PH et aI. Class l-restricced alloreactive cytotoxic T-Iymphocytes recognize a complex array of specific MHC-associated peptides. J Immunol1998; 160 :1091-97. 34. Heath WR, Kane KP, Mescher MF et aI. Alloreactive T'-cells discriminate among a diverse set of endogenous peptides. Proc Nad Acad Sci USA 1991; 88 :5101-05 .
208
Molecular Mechanisms ofSpondyloartbropatbies
35. Paradela A, Garcia-Peydro M, VazquezJ et al. The same natural ligand is involved in allorecognition of multiple HLA-B27 subtypes by a single T-cell clone : role of peptide and the MHC molecule in alloreactivity. J Immunol 1998; 161:5481-90. 36. Lopez D, Garcia Hoyo R, Lopez de Castro JA. Clonal analysis of alloreactive T-cell responses against the closely related B'2705 and B'2703 subtypes. Implication s for HLA- B27 association to spondyloarthropathy. J Immunol 1994; 152:5557-71. 37. Hulsmeyer M, WeiHe K, Pohlmann T er al. Thermodynamic and structural equivalenceof two HLA-B27 subtypes complexed with a self-peptide. J Mol BioI 2005 ; 346 :1367-79 . 38. Hiilsmeyer M, Fiorillo MT, Bettosini F er al. Dual , HLA-B27 subtype-dependent conformation of a self-peptide . J Exp Med 2004; 199:271-81. 39. Scofield RH, Warren WL, Koelsch G et al. A hypothesis for the HLA-B27 immune dysregulation in spondyloarrhroparhy: contributions from enteric organisms, B27 structure, peptides bound by B27 and convergent evolution . Proc Natl Acad Sci USA 1993; 90:9330-34 . 40. Scofield RH, Kurien B, Gross T et al. HLA-B27 binding of peptide from its own sequence and similar peptides from bacteria: implications for spondyloarthropathies. Lancet 1995; 345:1542-44. 41. Frauendorf E, von Goessel H , May E et al. HLA-B27-restricted T-cells from patients with ankylosing spondylitis recognize peptides from B'2705 that are similar to bacteria-derived pepndes, Clin Exp Immunol2003; 134:351-59. 42. Boisgeraulr F, Tieng V, Stolzenberg MC ec al. Differences in endogenous peptides presented by HLA-B'2705 and B'2703 allelic variants: implications for susceptibility to spondylarthropathies, J Clin Invest 1996; 98:2764-70. 43. Garcia F, Marina A, Albar JP er al. HLA-B27 presents a peptide from a polymorphic region of its own molecule with homology to proteins from arthritogenic bacteria . Tissue Antigens 1997; 49:23-28. 44. Alvarez I, Sesma L, Marcilla M et al. Identification of novel HLA-B27 ligands derived from polymorphic regions of its own or other class I molecules based on direct generation by 20S proteasome. J BioI Chern 2001; 276 :32729 -37. 45. Ramos M, Alvarez I, Sesma L et al. Molecular mimicry of an HLA-B27-derived ligand of arthritis-linked subtypes with chlamydial proteins. J Bioi Chern 2002 ; 277 :37573-81. 45a. Cragnolini JJ, Lopez de Castro JA. Identification of endogenously presented peptides from Chlamydia trachomatis with high homology to human proteins and to a natural self-ligand ofHLA-B27. Mol Cell Proteomics 2008 ; 7:170-80. 46. Popov I, Dela Cruz CS, Barber BH er al. The effect of an anti-HLA-B27 immune response on CTL recognition of Chlamydia. J Immunol 2001; 167:3375-82. 47. Popov I, Dela Cruz CS, Barber BH er al. Breakdown of CTL tolerance to self HLA-B'2705 induced by exposure to Chlamydia trachomaris. J Immunol 2002 ; 169:4033-38 . 48. Fiorillo MT, Maragno M, Butler R et al. CD8' T-cell autoreactiviry to an HLA-B27-restricted self-epitope correlates with ankylosing spondylitis. J Clin Invest 2000 ; 106:47-53. 49. Rock KL, York lA, Saric T et al. Protein degradation and the generation of MHC class l-presented pepcides. Adv Immunol 2002; 80:1-70. 50. Del Val M, Lopez D. Multiple proteases process viral antigens for presentation by MHC classI molecules to CD8(+) T-Iymphocytes. Mol Immunol2002; 39:235-47 . 5 I. Seifert U, Maranon C, Shmueli A et al. An essential role for tripeptidyl peptida se in the generation of an MHC class I epitope. Nature Immunology 2003 ; 4:375-79. 52. Guil S, Rodriguez -Castro M, Aguilar F et al. Need for rripepridyl-pepridase II in major histocompatibility complex class I viral antigen processing when proteasomes are detrimental. J BioI Chern 2006; 281 :39925-34. 53. Gil-Torregrosa BC, Castano AR, Del Val M. Major histocompatibility complex class 1 viral antigen processing in the secretory pathway defined by the trans-Golgi network prot ease furin. J Exp Med 1998; 188:1105-16. 54. Gil-Torregrosa BC, Castano AR. Lopez D et al. Generation ofMHC class I peptide antigens by protein processing in the secretory route by furin. Traffic 2000 ; 1:641-51. 55. Zhang Y, Kida Y, Kuwano K et al. Role of furin in delivery of a CTL epitope of an anthrax toxin-fusion protein. Microbiol Immunol 2001; 45:119-25. 56. Sheri L. Sigal LJ, BoesM et al. Important role of cathepsin S in generating peptides for TAP-independent MHC class I crosspresentarion in vivo. Immunity 2004; 21:155-65 . 57. Lopez D. Del Val M. Selective involvement of proteasomes and cysteine proteases in MHC class I antigen presentation. J Immunol 1997; 159:5769-72. 58. Rock KL. Gramm C, Rothstein Let al. Inhibitors of the proteasome block the degradation of most cell proteins and the generation of peprides presented on MHC class I molecules. Cell 1994; 78:761-71. 59. Luckey CJ, Marro JA. Partridge M er al. Differences in the expression of human class 1 MHC alleles and their associated peptides in the presence of proteasome inhibitors. J Immunol 2001; 167:1212-21.
HLA-B27-Bound Peptide Repertoires: Their Nature, Origin and Pathogenetic Relevance
209
60. Marcilla M. Cragnolini ]]. Lopez de Castro ]A. Proteasome-independenr HLA-B27 ligands arise mainly from small basic proteins . Mol Cell Proteomics 2007; 6:923-38. 61. Kisselev AF. Callard A. Goldberg AL. Importance of the different proteolytic sites of the proteasome and the efficacyof inhibitors varies with the protein substrare, ] Bioi Chern 2006 ; 281 :8582-90. 62. Cresswell P. Ackerman AL. Giodini A ct al. Mechanisms of MHC class I-restricted antigen processing and cross-presentation . Immunol Rev 2005; 207:145-57. 63. Peaper DR. Wearsch PA. Cresswell P. Tapasin and ERp57 form a stable disulfide-linked dimer within the MHC class I peptide-loading complex. EMBO ] 2005 ; 24:3613-23. 64. Santos SG. Campbell EC. Lynch S er al. MHC class I-ERp57-tapasin interactions within the peptide -loading complex.] Bioi Chern 2007;282:17587-93. 65. Wearsch PA. Cresswell P. Selective loading of high-affinity peptides ontO major histocompatibility complex class I molecules by the tapasin-ERp57 heterodimer, Nat Immunol 2007; 8:873-81. 66. Kienast A. Preuss M. Winkler M et al. Redox regulation of peptide receptivity of major histocompatibility complex class I molecules by ERp57 and rapasin. Nat Immunol 2007; 8:864-72. 67. Dick TP. Assembly of MHC class I peptide complexes from the perspective of disulfide bond formation . Cell Mol Life Sci 2004; 61:547-56. 68. Zemich D. Purcell AW, Macdonald WA et al. Natural HLA classI polymorphism controls the pathway of antigen presentation and susceptibility to viral evasion. ] Exp Med 2004 ; 200:13-24. 69. Peh CA. Burrows SR. Bamden M et al. HLA-B27-restricted antigen presentation in the absence of tapasin reveals polymorphism in mechanisms ofHLA class I peptide loading. Immunity 1998; 8:531-42. 70. Purcell AW, Gorman ]]. Garcia-Peydro M et al. Quantitative and qualitative influences of tapasin on the class I peptide reperroire.} Immunol 2001; 166:1016-27. 7 1. Williams AP. Peh CA. Purcell AWet al. Optimization of the MHC class I peptide cargo is dependent on tapasin. Immunity 2002; 16:509-20. 72. Goodall ]C. Ellis L. Hill Gaston is. Spondylarthntis-associated and nonspondylarthriris-associared B27 subtypes differ in their dependence upon rapasin for surface expression and their incorporation into the peptide loading complex. Arthritis Rheum 2006 ; 54:138-47. 73. Montserrat V. Galocha B. Marcilla M er al. HLA-B'2704. an allotype associated with ankylosing spondylitis. is critically dependent on transporter associated with antigen processing and relatively independent of tapasin and immunoproteasome for maturation . surface expression and TCell recognition : relationship to B'2705 and B'2706.] Immunol 2006 ; 177:7015-23. 74. Colbert RA. HLA-B27 misfolding: a solution to the spondyloarthroparhy conundrum? Mol Med Today 2000 ; 6:224-30. 75. Dangoria NS. DeLay ML. Kingsbury D] et al. HLA-B27 misfolding is associated with aberrant intermolecular disulfide bond formation (dimerization) in the endoplasmic reticulum. ] Bioi Chern 2002 ; 277:23459-68. 76. Turner M]. Sowders DP. DeLay ML er al. HLA-B2 7 misfolding in transgen ic rats is associated with activation of the unfolded protein response. ] Immunol 2005; 175:2438-48 . 77. Tran TM. Dorri s ML. Satumtira N et al. Additional human f\2-microglobulin curbs HLA-B27misfolding and promotes arthrits and spondylitis without colitis in male HLA-B27 transgenic rats. Arthritis Rheum 2006 ; 54:1317-27. 78. Kollnberger S. Bird L. Sun MY et al. Cell-surface expression and immune receptor recognition of H LA-B27 homodimers. Arthritis Rheum 2002; 46 ;2972-82. 79. Allen RL. Trowsdale J. Recognition of classical and heavy chain forms of HLA-B2 7 by leukocyte receptors. Cur r Mol Med 2004; 4:59-65. 80. Bird LA. Peh CA. Kollnberger S er al. Lymphoblasroid cells express HLA-B27 homodimers both intracellularly and at the cell surface following endosomal recycling. Eur] Immunol 2003 ; 33 :748-59. 81. Uchanska-Ziegler B. Ziegler A. Ankylosing spondylitis : a beta2m-deposition disease? Trends Immunol 2003 ; 24:73-76. 82. Montse rrat V. Mart i M. Lopez de Castro ]A. Allospecific T-cell epirope sharing reveals extensive conservation of the antigenic features of peptide ligands among HLA-B27 subtypes differentially associated with spondyloarthritis. J Immunol 2003: 170:5778-85. 83. Garcia-Peydro M. Marti M. Lopez de Castro ]A . H igh Tcell epirope sharing between two HLA-B27 subtypes (B'2705 and B'2709) differentially associated to ankylosing spondylitis. ] Immunol 1999; 163:2299-305. 84. Montserrat V. Presentacion anrigenica por subripos de HLA-B27 y HLA-BI4 asociados diferencialmente a enfermedad. Tesis Doctoral, Universidad Aut6noma de Madrid 2006;
CHAPTER
15
Biochemical Features ofHLA-B27 and Antigen Processing Simon J. Powis," Susana G. Santos and Antony N. Antoniou
Abstract
T
he strong association of the hum an MH C class I allele HLA-B27 with the developm ent of th e chronic inflammat or y disease ankylosing spondylitis (A S) is clear and has been known for over three decades. Despite thi s. it is far from clear how HLA-R27 is directly involved in AS . In recent years considerable progress has been made in defining the assembly pathway and the protein components involved in successfully folding MHC class I molecules in the environment ofthe endoplasmic reticulum. This process involves a number ofcritical interactions. which may influence how HLA-B27 molecules fold and what peptides become loaded. The impact of the unpaired Cys-67 residue in the peptide-binding groove upon the behaviour ofboth correctly fold ed and misfolded HLA-B27 molecules , especially its ability to allow the formation ofB27 heavy-chain oligomers or dimers, wh ich may form novel targets for immune receptors. or be an indicator ofintracellular stress, has also been the focus ofmuch research. In this chapter we aim to review recent data to determine wheth er any biochemical features ofHLA-B27 can supply clues as to its enigmatic role in AS and will also comment on future potential directions ofbiochemical research into HLA-B27.
Introduction At the time ofwriting, over 1,900 HLA class I alleles have been characterised, with over 900 of the se being HLA-B alleles (www.ebLac.uklimgt/hla/stats.html). Amongst thi s plethora of pol ymorphism. HLA-B27 stands out because ofits asso ciation with inflammato ry arth ritis, most notably ankylo sing spondylitis (A S), wher e 90 % or more of patients exp ress this MHC class I allele. Desp ite thi s strong disease association. HLA-B27 does not seem to confer any gross selective disadvantage, being well represented in many populations around th e world. Furthermore. it is clear that HLA-B27 is perfectly capable of performing the functions required of a normal MHC class I allele in binding and presenting short pep tides to T-Iymphocytes within the immune system,' thus leading to the generation of effective cytotoxic T -cell populations in the adaptive antigen-specific immune response. It also acts as a ligand for receptors such as killer cell Ig-like receptors (KIR) and leucocyte Ig-like receptors (LIR or ILT) expressed on natural killer (NK) cells and some Tvcell subsers.P Given HLA-B27's ability to act as a normal MH C class I molecule, are there any clues from the biochemical study ofHLA-B27 that might indicate its role in AS ? The apparent answer to thi s question is a somewhat qualified yes. HLA-B27 does indeed display some unique characteristics when compared to other MHC class I alleles. In thi s chapter we will review some of the se key observations.
"Correspo nd ing Author: Simon Powis- Bute Medical School, University of St Andrews, Fife,
Scotland, KY1 69TS, UK. E-mail: sjplO @st-andrews.ac.uk Molecular M echan isms ofSpondyloarthropathi es, edited by Carlos Lopez-Larre a and Rob erto Diaz-Peiia. ©2009 Landes Bioscience and Springer Science+Business Media.
Biochemical Features ofHLA-B27andAntigen Processing
211
The MHC Class I Assembly Pathway To understand how HLA-B27 might differ from other non-AS associated MHC class I alleles,it would be beneficial to have a full knowledge ofthe normal assembly pathway ofthis family of molecules . In recent years there has been considerable progress in this area. MHC class I molecules are comprised ofa heavy chain ofapproximate 40 kDa, noncovalently associated with ~2-microglobulin(~2m). Assembly is completed by the binding ofa short peptide, of8-10 residues, within the peptide-bindinggroove on the upper surfaceofthe molecule .The assemblyprocess takes place within the environment ofthe endoplasmic reticulum (ER) and initially involves ER-based chaperones such as calnexin, calreticulin and immunoglobulin bindingprotein (BiP).4-6 However, this requirement to bind a short peptide before exit from the ER imposes an extra level ofquality control that many other proteins do not undergo. For most MHC class I molecules this quality control process is undertaken within the peptide-loading complex (PLC), a multimolecular structure composed ofthe MHC classI molecule, calreticulin , the oxidoreductases ERpS7 and PDI, the accessory protein tapasin and the TAP peptide transporrcrf'
212
Molecular Mechanisms ofSpondyloarthropathies
Figure 1. The MHC class I peptide-loading complex (PLC) comprises the MHC class I molecule heavy (HC) associated with fl2m, cal reticulin (CRT), tapasin, the oxidoreductases ERpS7 and protein disulfide isomerase (POI) and the transporter associated with antigen processing (TAP), which is not shown above, but which interacts directly through tapasin. ERpS7 and POI appear to control the status of the conserved disulfide bond between cys-101 and cys-164 . The additional presence in the pept ide groove of cys-67 in HLA-B27 raises the possibility of aberrant interaction occurring when HLA-B27 interacts with the PLC.
A Key Role for cys-67? As previously mentioned HLA-B27 contains within its peptide-binding groove an unpaired cysteine residue.Whilst this is rare, it isimportant to note that cys-67 is not exclusiveto HLA-B27, as other subtypes ofHLA-B7, -BI5, -B35 and B38 for example also carry this residue. However in the case ofHLA-B27 this residue seems responsible for several of its unique characteristics. The study of cys-67 has vastly increased since the findings of Allen and colleagues" in 1999, when in vitro folding experiments revealed the formation of a heavy chain dimeric structure, formed ostensibly between the two cys-67 residues of the heavy chains. These heavy chain homodimers, usually referred to as B272, can be recognised by a group ofNK family receptors and this area will be reviewed in detail elsewhere. In this chapter we will focus on their biochemistry. B27 2 molecules have been reported to form in cell lines of human and rat origin,23.24 in AS patient derived cells" and in cells from B27 transgenic rats.26.27There appears to be two populations ofB27 2molecules , one type ofwhich forms within the ER and a second that forms at the cell surface. The data so far suggest that they are separate populations because pulse-chase experiments have so far failed to show exit ofB27 2from the ER .23.28 To what extent these two populations are similar is not clear, thus in the case ofthe ER-B27 2form, there may be the involvement ofcys-l64 in their formation," whereas the cell surface forms are thought to be cys-67 linked. How and why B27 2 structures form in the ER is not known. High levels of HLA-B27 heavy chain expression, as often occurs in transfectant cell lines is a likely contributing factor and it is notable that the
Biochemical Features ofHLA-B27 and Antigen Processing
213
disease-prone rat transgenic lines also carry high copy numbers ofthe heavy chain.29 Thus inefficient folding ofthe heavy chain appears to result in a population ofB272molecules. The involvement of the normal chaperones and PLC in the process remains to be determined, although in a series of preliminary experiments we have detected B27 2 molecules associated with the PLC (Antoniou, Santos and Powis, unpublished). The fate of the ER-B27 2 molecules is also unclear, it is possible that some might be re-used to fold as monomers, but most are likely to undergo ER-associated degradation. The presence ofan excess ofunfolded HLA-B27 heavy chains in the ER, in the form ofmonomers or B27 2, may be recognised by the cell as a form of stress.24.28 It is also somewhat unclear as to how the cell surface B27 2molecules form. The evidence to date suggests that internalisation into an endosomal compartment may be required, potentially involving a tyrosine-based recycling morifin the cytoplasmic tail ofHLA-B27. 25.30 Delivering HLA-B27 molecules to an acidic endosomal compartment may encourage unfoldingofmolecules that do not contain high-affinity peptides, thus exposing the unpaired cys-67 residue. Again, overexpression of HLA-B27 in transgenic and in vitro cell lines may encourage unstable cell surface HLA-B27 molecules, thus predisposing towards their unfolding. Interestingly, the presence ofincreased numbers ofunfolded HLA-B27 molecules on the cell surface ofAS patients has been reported." Although a significant amount ofdata has accumulated in the study ofmisfolded monomers and B27 2 molecules. a key problem remains due to the lack of reagents that are available that will distinguish between different unfolded forms ofHLA-B27. It is very difficult to determine whether unfolded monomers and B27 2molecules represent a homogenous or heterogenous population. We rely heavily on the use of the monoclonal antibody H C 10,32 which recognises most HLA-B and -C alleles in a region at the top ofthe al helical domain." For biochemical studies H C lOis extremely useful, allowing immunoprecipitation ofunfolded molecules from detergent lysates, flow cytometry on the cell surface and also the detection of MHC class I heavy chains by immunoblotting. However, its limitations are that it cannot distinguish between what might be various different types of unfolded molecules. For example, depending on the detergent used, heavy chains weakly associated with ~2m can be detected and it cannot distinguish HLA-B27 monomers from B27 2 molecules, except by immunoblotting. This important area ofresearch will take a quantum leap forward when we have reagents that can specifically distinguish these various forms ofHLA-B27. We have focussed here on the role that cys-67 may play in B272formation. However, the study of cys-67 is not new. In the early 1990s Archer and colleagues studied the reactivity ofthe exposed cys-67 residue in considerable detail. These studies clearly demonstrated variable access to cys-67 in the groove by various thiol-modifying agents,34-36 key observations that are relevant to current research in the field because the prevention ofartefactual B27 2molecule formation during detergent cell lysis involving the free cys-67 is a constant risk. The inclusion ofalkylating agents in lysis buffers is absolutely critical. Most researchers now appear to use N-ethylmaleimide (NEM), but incomplete access ofcys-67 to NEM in peptide-loaded HLA-B27 molecules has not been formally excluded. A final caveat to the study ofcys-67 comes in the form ofa transgenic rat line in which cys-67 is mutated to ser-67. 29 This line develops inflammatory arthritis when made homozygous, presumably by increasing the burden ofheavy chains beyond a crucial threshold level. Therefore, B27 2structures formed via cys-67 may be involved, but may not be the only faeror that predisposes to inflammatory arthritis and AS.
HLA-B27 in Dendritic Cells Dendritic cells (DC) represent one of the key cells in the immune response. They acquire antigen in the tissue in which they reside and then migrate to secondary lymphoid tissue where they playa central role in initiating antigen specific immune responses. Recent evidence in the HLA-B27 transgenic rat model indicates that the presence ofHLA-B27 molecules on the surface ofDC impairs their ability to act as efficient stimulatory cells for T-Iymphocytes.37The potential implications of these observations are highly significant. As yet we understand very little of the biochemistry ofHLA-B27 in DC. The pathway ofMHC class I assembly in DC is thought to be
214
Molecular Mechanisms ofSpondyloarthropathies
similar to other cells within the ER , with DC containing all the elements of the PLC. However, things differ in the later stages of the lifespan ofan MH C class I molecule. Some DC are known to possess an endosomal store of mature MHC class I molecules, as antigens that are captured through endocytosis and that were thought to be presented in a class II context are now known to be loaded and presented by class I molecules, via a process named cross-presentation. A couple of mechanisms for cross-presentation have been proposed: DC may be able to deliver antigens to MHC class I molecules in ways that do not occur in other cell types. They may be able to acquire large amounts of antigen by direct fusion ofthe ER compartment with material undergoing phagocytosis, creating large phagolysosomes that have been reported to contain most of the PLC ;40.41and/or DC may be able to deliver antigen to the ER by a direct retrograde transport of material from the cell surface." Despite the uncertainty about the exact mechanism behind peptide loading via cross-presentation, it appears that in dendritic cells, optimisation of most class I molecules takes place soon after activation. This would allow for most class I molecules at the surface ofmature dendritic cells being loaded with antigenic peptides derived from the source of the activation. Examples of this are the accumulation of class I molecules in the Golgi region, which are rapidly relocated to the cell surface upon DC activation, as well as the up-regulation of MHC class I molecules and class I loading machinery.38.39The impact ofboth these mechanisms on the behaviour of HLA-B27 is unknown at present. In the most recent reports about the rat models of SpA, where the formation of conjugates between dendritic cells expressing B27 and T-cells seems to be impaired and where addition of citric acid, modifying the closed versus open class I conformers brings conjugate formation to control values, it would be tempting to speculate that alteration of the closed versus open conformers balance at the cell surface of dendritic cells influences conjugate formation. In view ofsuch exciting recent advances and with the availability of new technologies allowing for the characterization of individual immunologic synapses, DC are likely to be an exciting new area of study in the context of SpA. In preliminary experiments, we have seen similar intracellular accumulations of HLA-B27 in a DC-like cell line (Santos and Powis, unpublished). The status of these endosomal sources of HLA-B27, whether fully folded and peptide-optimised, or a mixture offolded, unfolded and B27 2 remains to be established. Exposure to high levels of unfolded HLA-B27 molecules at a critical stage of DC-T-cell interaction could have a dramatic influence on the subsequent behaviour of aT-lymphocyte.
Conclusions The study of the biochemistry of HLA-B27 has revealed a number of unique observations, many ofwhich centre around the efficiency with which it undergoes folding within the ER, leading to the presence of unfolded heavy chains which are predisposed to form B27 2 structures. Similar structures forming at the cell surface may present novel ligands to the immune system. Defining the precise mechanisms by which HLA-B27 assembles in comparison to nondisease associated alleles remains a highly valid way of analysing the role that HLA-B27 plays in AS. With the development ofnew reagents and the analysis ofcells such as DC, it is likely that more insights into AS will be achieved in the future.
Acknowledgements The authors wish to acknowledge funding support from the UK Arthritis Research Campaign and the Portugese Foundation for Science and Technology.
References 1. Ramo s M, Paradela A, Vazquez M er al. Differencial association ofHLA-B'2705 and B'2709 co ankylosing spondylitis correlates with limited peptide subsets but noc with altered cell surface stability. J Bioi Chern 2002; 277(32):28749-28756. 2. Allen RL, Raine T. Haude A et al. Leukocyte recepcor complex-encoded immunomodulacory recepcors show differing specificity for alternative HLA-B27 structures. J Immun ol 2001 ; 167(10):5543-5547. 3. Kollnberger S, Bird L, Sun MY et al. Cell-surface expression and immune recepcor recognition of HLA-B27 hornodimers. Archritis Rheum 2002 : 46(11) :2972 -2982 .
Biochemical Features ofHLA-B27and Antigen Processing
215
4. Degen E, Williams DB. Participation of a novel 88-kD protein in the biogenesis of murine class I histocompatibility molecules. J Cell Bioi 1991; 112(6) :1099-1115 . 5. Nossner E, Parham P. Species-specificdifferences in chaperone interaction of human and mouse major histocompatibility complex class I molecules. J Exp Med 1995: 181(1) :327-337. 6. Sadasivan B, Lehner PJ, Ortmann B et al. Roles for calreticulin and a novel glycoprotein, tapasin, in the interaction of MHC class 1 molecules with TAP. Immunity 1996: 5(2) :103-114. 7. Antoniou AN, Powis SJ, Elliott T. Assembly and export of MHC class I peptide ligands. Curr Opin Immunol2003: 15(1) :75-81. 8. Hughes EA, Cresswell P. The thiol oxidoreductase ERp57 is a component of the MHC class I peptide-loading complex. Curr Bioi 1998: 8(12):709-712. 9. Lindquist JA , Jensen ON, Mann M et al. ER-60, a chaperone with thiol-dependent reductase activity involved in MHC class I assembly. Embo J 1998; 17(8):2186-2195 . 10. Morrice NA, Powis S]. A role for the thiol-dependenc reductase ERp57 in the assemblyofMHC class I molecules. Curr Bioi 1998; 8(12) :713-716. 11. Ortmann B, Androlewicz MJ, Cresswell P. MHC class IIbeta 2-microglobulin complexes associate with TAP tran sporters before peptide binding . Nature 1994; 368(6474) :864-867. 12. Ortmann B, Copeman J, Lehner PJ et al. A critical role for tapasin in the assembly and function of rnultirneric MHC class I-TAP complexes. Science 1997: 277(5330) :1306-1309 . 13. Park B. Lee S, Kim E et al. Redox regulation facilitates optimal peptide selection by MHC class I during antigen processing. Cell 2006; 127(2) :369-382. 14. Williams AP, Peh CA , Purcell AWet al. Optimization of the MHC class 1 peptide cargo is dependent on tapasin. Immunity 2002; 16(4):509-520. 15. Kienast A, Preuss M, Winkler M et al. Redox regulation of peptide receptivity of major histocompatibility complex class I molecules by ERp57 and tapasin. Nat Immunol 2007 ; 8(8) :864-872. 16. Santos SG, Campbell EC, Lynch S et al. Major histocompatibility complex class1-ERp57-tapasin interactions with in the peptide-loading complex. J Bioi Chem 2007 ; 282(24) :17587-17593. 17. Wearsch PA. Cresswell P. Selective loading of high-affinit y peptides onto major histocompatibility complex class I molecules by the tapasin-ERp57 hetcrodimer, Nat Immunol 2007: 8(8):873-881. 18. Park B, Lee S, Kim E et al. A Single polymorphic residue within the peptide-binding clefi of MHC class I molecules determines spectrum of tapasin dependence . J Immunol 2003; 170(2) :961-968 . 19. Sesma L, Galocha B, Vazquez M et al, Qualitative and quantitative differences in peptides bound to HLA-B27 in the presence of mouse versus human tapasin define a role for tapasin as a size-dependent peptide editor. J Immunol 2005 ; 174(12) :7833-7844 . 20. Tourdot S, Gould KG. Competition between MHC class 1 alleles for cell surface expression alters CTL responses to influenza A virus. J Immunol 2002: 169(10) :5615-5621. 21. Tourdot S, Nejmeddine M. Powis SJ er al. Different MHC class I heavy chains compete with each other for folding independently of beta 2-microglobulin and peptide . J Immunol 2005; 174(2):925-933 . 22. Allen RL. O'Callaghan CA , McMichael AJ er al. Cutting edge: HLA-B27 can form a novel beta 2-microglobulin-free heavy chain homodimer structure. J Immunol 1999; 162(9):5045-5048. 23. Antoniou AN, Ford S, TaurogJD er al. Formation of HLA-B27 homodimers and their relationship to assembly kinetics. J Bioi Chem 2004; 279(10) :8895-8902 . 24. Lemin AJ, Saleki K, van Lith M er al. Activation of the unfolded protein response and alternative splicing of ATF6alpha in HLA-B27 positive lymphocytes. FEBS Lett 2007; 581(9) :1819-1824. 25. Bird LA, Peh CA, Kollnberger S et al. Lymphoblastoid cells express HLA-B27 homodimers both intracellularly and at the cell surface following endosomal recycling. Eur J Immunol 2003; 33(3) :748-759. 26. KoIlnberger S, Bird LA, Roddis M et al. HLA-B27 heavy chain homodimers are expressed in HLA-B27 transgenic rodent models of spondyloarthritis and are ligands for paired Ig-like receptors. J Immunol 2004 ; 173(3) :1699-1710. 27. Tran TM, Satumtira N, Dorris ML er al. HLA-B27 in transgenic rats forms disulfide-linked heavy chain oligomers and multimers that bind to the chaperone BiP. J Immunol2004; 172(8) :5110-5119 . 28. Turner MJ, Sowders DP, DeLay ML et al. HLA-B27 misfolding in transgenic rats is associated with activation of th e unfolded protein response. J Immunol 2005 ; 175(4):2438-2448. 29. Taurog JD, Maika SD, Satumr ira N et al. Inflammatory disease in HLA-B27 transgenic rats. Immunol Rev 1999; 169:209-223. 30. Santos SG, Antoniou AN, Sampaio P et al. Lack of tyrosine 320 impairs spontaneous endocytosis and enhances release of HLA-B2 7 molecules. J Immunol 2006 : 176(5):2942-2949 . 31. Tsai WC, Chen CJ, Yen JH et al. Free HLA class I heavy chain-carrying monocytes-a potential role in the pathogenesis of spondyloarthropathies. J Rheumacol 2002; 29(5) :966-972. 32. Starn NJ, Spits H, Ploegh HL. Monoclonal antibodies raised against denatured HLA-B locus heavy chains permit biochemical characterization of certain HLA-C locus products. J Immunol 1986; 137(7):2299-2306.
216
Molecular Mechanisms ofSpondyloarthropathies
33. Perosa F. Luccarelli G. Prete M ct al. Beta 2-microglobulin-free HLA class I heavy chain epitope mimicry by monoclonal antibody HC-10-specific peptide . J Immunol 2003; 171(4) :1918-1926 . 34. Archer JR. Whelan MA. Badakere SS et al. Effect of a free sulphydryl group on expression ofHLA-B27 specificity. Scand J Rheumacol 1990; 87(Suppl):44-50. 35. MacLean L. Macey M. Lowdell M et al. Sulphydryl reactivity of the HLA-B27 epitope : accessibility of the free cysteine studied by flow cytometry. Ann Rheum Dis 1992; 51(4) :456-460 . 36. Whelan MA. Archer JR. Chemical reactivity of an HLA-B27 thiol group. Eur J Immunol 1993; 23(12) :3278-3285 . 37. Hacquard-Bouder C. Chimenti MS. Giquel B et al. Alteration of antigen -independent immunologic synapse formation between dendritic cells from HLA-B27-transgenic rats and CD4+ T-cells: selective impairment of costimulatory molecule engagement by mature HLA-B27. Arthritis Rheum 2007; 56(5) :1478-1489. 38. Ackerman AL. Cresswell P. Regulation of MHC class I transport in human dendritic cells and the dendritic-like cell line KG-I. J Immunol 2003; 170(8) :4178-4188. 39. MacAry PA. Lindsay M. Scott MA et al. Mobilization of MHC class I molecules from late endosornes to the cell surface following activation of CD34-derived human Langerhans cells. Proc Nat! Acad Sci USA 2001 ; 98(7) :3982-3987 . 40. Guermonprez P. Saveanu L. Kleijrneer M er al. ER-phagosome fusion defines an MHC class I cross-presentation compartment in dendritic cells. Nature 2003; 425(6956) :397-402. 41. Houde M. Berrholet S. Gagnon E er al. Phagosomes are competent organelles for aurigen cross-presentation. Nature 2003; 425(6956) :402-406. 42. Ackerman AL. Kyritsis C. Tampe R et al. Access of soluble antigens to the endoplasmic reticulum can explain cross-presentation by dendritic cells. Nat Immunol 2005; 6( 1):107-113.
CHAPTER
16
HLA-B27 Misfolding and Spondyloarthropathies Robert A . Colbert," Monica L. DeLay, Gerlinde Layh-Schmitt and Dawn P. Sowders
Abstract
H
LA -B27 plays a central role in the pathogenesis of many spondyloarthropathies and in particular ankylosing spondylitis. Th e observation that the HLA-B27 heavy chain has a tendency to m isfold ha s raised the possibility that associated disease s may belong in a rapidly expand ing category ofprotein mi sfolding d isorders. The syn th esis of the HLA-B27 heavy chain . assembly with 132m and th e loading of peptide cargo, occurs in the endoplasmic reticulum (ER) before tr ansport to the cell surface. The evidence indicates that misfolding occurs in the ER pri or to b 2m association and peptide optim ization and is manifested in the formation ofaberrant inter- and intra-ch ain di sulfide bonds and accum ulation of heavy chain bound to the chaperone BiP. En han ced accumulation ofm isfolded heavy chains during the induction ofclass I expre ssion by cytokines, can cause ER str ess resulting in activation of the unfolded protein response (UPR). Effects ofUPR activatio n on cyto kine production are beginning to em erge and may provide important m issing links between HLA-B27 m isfolding and spo ndyloarthriris. In this chapter we will review what ha s been learned about HLA-B27 m isfolding in human cells and in the transgenic rat model of spo ndyloarthritis-like di sease, considering it in the context ofother protein misfolding diso rders. These studi es provide a framework to support mu ch need ed tr an slational work assessing H LA-B27 misfolding and UPR activation in pat ient-derived material, its consequences for disease path ogene sis and ult imately how and where to focus intervent ion strategies.
Introduction Ankylosing spondylitis (AS) is a complex genetic trait with an estimated 4 to 10 genes responsible for the majority of susceptibility.' Spondyloarthropathies (SpA) comprise several disorders that are more heterogeneous clinically and where genetic susceptibility is likely to be more complex and variable. Defining genetic loci and ultimately genes that influence susceptibility, is an area of intense investigation and is discussed in detail in other chapters. Family-based linkage studies using low-density microsatellite markers have been somewhat disappoinnng.P However, single nucleotide polymorphism (SNP) identification and mapping has provided a detailed framework on which to perform whole genome association studies. This approach has already provided valuable information on genes involved in susceptibility to other complex genetic diseases,"and studies on AS are now emerging.' In addition to providing markers that will be useful in identifying AS patients earlier in their disease course, it is anticipated that a more complete picture of genetic susceptibility will inform us on pathways that are important in pathogenesis and identify new th erapeutic targets. 'Co rrespo nd ing Author: Rob ert A Colbe rt- Divisio n of Rheumatology, Cincinnati Children's Hospital Medical Center, Cinci nnati, Ohio 45229-3039, USA. Email: bob. colb ert@cchm c.org
Molecular M echanisms ofSp ondyloarthropathies, ed ited by Carlos Lopez-Larre a and Roberto Diaz-Pefia. © 2009 Landes Bioscience and Spr inger Science+Business Media.
218
Molecular Mechanisms ofSpondyloarthropathies
Unlike most complex genetic diseases, a single gene (HLA-B) plays a dominant role in AS. The B27 allele contributes as much as 40% of the overall genetic load and is a major factor for many other SpA.6.7Although the role ofHLA-B27 has been the focus ofintense investigation for over 30 years, none of the postulated mechanisms has been proven or eliminared," While it has been tacitly assumed that a single feature ofHLA-B27 is responsible, it is conceivable that this is not the case,"and that the answer to the HLA-B27 conundrum will be even more complex than initially anticipated. A detailed understanding ofpathogenesis requires animal models that phenocopy the human condition and are amenable to genetic manipulation and experimenration.'? combined with translational studies ofhuman material that are confirmatory. The development ofanimal models has been attempted over the years through the generation ofHLA-B27 transgenic mice and rats. Transgenic mouse models ofSpA havebeen disappointing for several reasons. Initially,no spontaneous inflammatory disease was observed,'! and although attempts to induce disease with infection revealed some differences in susceptibility," the SpA phenotype was not observed . Subsequent studies suggested that HLA-B27 transgenic mice developed spontaneous arthritis ifyou deleted the endogenous gene for (mouse) 132m.BoI5 However, the use ofa mixed genetic background may have confounded these studies making reproducibility and thus interpretation problemaric." Controlling for mixed backgrounds is very difficult and eventual genetic drift can result in loss of the phenotype. Inbred strains ofanimals are the genetic equivalent ofa single human and thus it is not surprising that different genetic backgrounds would influence susceptibility. In humans less than 5% ofHLA-B27 positive individuals develop SpA . The whole genome association studies mentioned above are being used to identify other human genes that affect susceptibility. It may be possible to exploit strain differences that are important determinants ofdisease in HLA-B27 transgenic rodents and use similar genetic approaches to identify the responsible genes.While the rodent genes may not be the same as those found in humans, by definition they will be involved in pathways that are important for pathogenesis ofHLA-B27-associated disorders and substantial overlap with pathways identified in human genetic studies would be expected. The production of transgenic rats expressing HLA-B27 and human 132m (B27/h131m) that develop spontaneous inflammation resembling SpA signified a major advance in this area.17This demonstrated that under certain conditions overexpression ofHLA-B27 was sufficient to cause disease, providing the first evidence that the gene product itselfwas involved. The SpA-like phenotype includes gastrointestinal tract inflammation (e.g., colitis) , arthritis and other inflammatory lesions in the skin and testicles. The colitis is highly penetrant, while arthritis is lessfrequent and depends more on the strain of rat used. Although axial inflammation can occur, it does not recapitulate the spinal inflammation and ankylosis seen in humans." However, recently Tran et al have reported that overexpressing additional hb .m can alter the phenotype in transgenic rats that already overexpress HLA-B27 and h132m.19 High copy number B27/h131m transgenics with additional hb .m develop more severearthritis and significant axial disease with no apparent change in colitis. Interestingly, rats with low copy number B27/h131mthat normally do not develop spontaneous disease, develop axial and peripheral arthritis without colitis, when additional hb.rn isoverexpressed.These observations and their significance for understanding the role ofHLA-B27 in disease are further discussed later in this chapter and other chapters of this book . In this chapter we will focus on one mechanism that may be the basis for the striking relationship between HLA-B27 and spondyloarthritic diseases. We will explain the general concept and consequences of protein misfolding and then provide a detailed assessment of the special case ofHLA-B27 misfolding and how it may be linked to disease through an auroinflarnmatory stimulus.
Usual, Unusual and Unique Features ofHLA-B27 and Their Potential Role in Disease Pathogenesis Extensive polymorphism at the HLA-B locus results in considerable sequence variation in the HLA-B-encoded heavy chain across the human population. Over 900 alleles have been reported
HLA-B27 MisfOlding and Spondyloarthropathies
219
to date, encoding over 800 different proteins (www.anthonynolan.com/HIG/). These amino acid differences affect a nwnber ofproperties ofclassI heavychains, including peptide binding specificity and affinity, T -cell recognition (both as a result ofand independent from bound peptide), lJ2m bindingaffinity, folding and assembly efficiencyand chaperone interaction (e.g., tapasin) (reviewed in ref 20) . There are also polymorphisms in the promoter region ofHLA-B at the 5' end of the gene, which could affect baseline expression level and inducibility," Features ofHLA-B27 that distinguish it from other alleleshave provided the basis for several hypotheses concerning its contribution to disease. It is convenient to classify these ideas based on whether they invoke immunological recognition of some form of the heavy chain versus intracellular effects.s Immunological recognition by antibodiesf or autoreactive T-cells 23 supposes that the form(s) of HLA-B27 being recognized are typical for HLA class I complexes. More recently the detection of other forms ofHLA-B27, such as heavy chain homodimers.P or unusual unfolded monomers," has led to ideas about recognition by leukocyte receptors on NK cells and other leukocytes. 26' 3o In contrast, the tendency ofHLA-B27 heavy chains to misfold in the intracellular compartment known as the endoplasmic reticulwn (ER)31.32 has led to the notion that intracellular effects ofHLA-B27 might be involved in disease pathogenesis. Misfolding was hypothesized to result in activation ofan intracellular stress response pathway known as the unfolded protein response (UPR),33 which has been shown to occur in B27/hlJ2m transgenic rats.34.3S The consequences of HLA-B27-induced UPR activation will be discussed in detail later in this chapter. Finally, the observation that cell lines tran sfected with HLA-B27 but not other alleles exhibit increased bacterial survival 36 .37 could be important for pathogenesis, particularly in reactive arthritis. Recent evidence suggests that bacterial replication is increased'" and that the p38 MAP kinase pathway may be disrupted." This most likely represents an intracellular or at least nonaneigen-presenting effect ofHLA-B27.40 Experiments usingsite-directed mutants ofHLA-B27 show that the biological effect correlates with heavy chain rnisfolding, but is not associated with acute UPR activation and therefore the molecular mechanism remains to be defined. It will be important to determine whether the expression of heavy chains that misfold is responsible for this effect, since a related phenomenon has been observed for a mutant of surfactant protein-C that misfolds." These authors demonstrated that adaptation to chronic ER stress was associated with modification of an NF-kB-dependent pathway, reminiscent of what has been observed in HLA-B27-transfected cells." Hypotheses related to immunological recognition ofusual and unusual forms ofHLA-B27 are considered in other chapters ofthis book and will not be further discussed here . In this chapter we will focus on HLA-B27 misfolding, considering it in the context ofother proteins that misfold, the cause of misfolding and more importantly, what we have begun to learn about its consequences. There has been a tendency to assume that only one hypothesis,or one aspect ofthe immunobiology ofHLA-B27, will eventually be tied to its role in pathogenesis. However, this may not be correct, particularly when one considers phenotypic differences in the diseases associated with HLA-B27 such as reactive arthritis, uveitis, AS and other forms of undifferentiated SpA.
Importance ofProtein Folding The information required for a protein to fold into its native conformation is contained within its primary sequence, yet a great deal of energy is expended to ensure that this occurs efficiently and without error (reviewed in ref 43) . For multi-subunit proteins or those that transport cargo, the process is even more complex, with many opportunities for errors in the formation of stable, properly folded complexes. It has become increasingly apparent over the last decade that many genetic diseases result in protein misfolding, either due to inherent properties ofthe mutated gene product, or in some cases as a consequence of abnormalities in the cellular pathways that handle misfolded proteins.
220
Molecular Mechanisms ofSpondyloarehropatbtes
HLA Class I Folding andAssembly HLA class I complexes of heavy chain, ~2m and peptide represent an example of a protein (heavy chain) that transports 'cargo' (~2m and peptide) to the cell surface. To perform this function and display self-peptides or pathogen-derived cargo to T-cells during an immune response , HLA class I heavy chains must fold properly, bind ~2m and then load peptide prior to exiting the ER compartment (reviewed in ref 44). High stability of the trimolecular complex is critical for efficient transport through the Golgi, a long half-life on the cell surface and ultim ately a productive immune response . The stability ofHLA class I complexes is critically dependent on early events in the folding and assembly process,including the formation oftwo intrachain disulfide bonds.v The a3 domain folds very rapidly and is stabilized by an intradomain disulfide between Cys-203 and Cys-2S9. The al and a2 domains fold more slowly and this is not complete until peptide is stably bound." A second disulfide between the al and a2 domains (Cys-lOl-Cys-164) maintains the integrity ofthe peptide-binding groove47as the heavy chain/Ben heterodimer interacts with tapasin, ERpS7 and the transporter associated with antigen processing (TAP) to form the peptide loading complex (PLC). Although there are allelic differences in the need to interact with tapasin (and thus the PLC), in general this process facilitates the binding ofhigh affinity peptides . For example, HLA-B27 (the B'2705 subtype) is expressed relatively efficiently in rapasin-deficient cells" and is frequently referred to as a tapasin-independent allele. However, it interacts with tapasin when present and this affects the peptide repertoire." It is possible that the ability ofHLA-B27 to be expressed at high levels on tapasin-deficienr cells may reflect its tendency to fold slowly and be retained in the ER in a peptide-receptive state without tapasin-PLC interaction. This could favor the binding and optimization ofavailable peptides without tapasin-mediated retention. ERpS7 binds to tapasin via a disulfide (ERpS7 -Cys-S7-Cys-9S-Tapasin) and playsan important role in disulfide bond isomerization in the heavy chain during class I assembly.50Recent evidence indicates that formation ofthe ERpS7-tapasin conjugate prevents ERpS7 -rnediared reduction of the a l -a2 int erdomain disulfide in the class I heavy chain, thus maintaining the peptide binding groove in a receptive state." When tapasin is missing or mutated at Cys-9S and thus unable to form a complex with ERpS7, the class I heavy chain al-a2 disulfide is reduced until suitable peptide cargo can bind. Free ERpS7 (or ERpS7-calreticulin complexes) appears to catalyze this reduction in the absence of tapasin leading to the concept that tapasin performs its function by sequestering ERpS7. The final stages of peptide binding to HLA class I molecules includes trimming by the ER aminopeptidase associated with antigen processing (ERAAP or ERAPl).51-56 Peptides appear to be nestled into the peptide-binding groove at their C -terminus first with ERAAP-mcdiated N -terminal trimming to their final sizeof8-11 amino acids. In humans, L-RAP or ERAP2 may also playa role in this process. In addition to peptide trimming for presentation by class I molecules , ERAAP appears to have another role in the immune system. It was discovered independently as aminopeptidase regulator ofTNF receptor (TNFRl) shedding (ARTS -I), but also regulates shedding ofIL-6 and IL-l decoy receptors.57-59
Consequences ofProtein Misfolding The vast majority of proteins are made in the cytosol, or cotranslationally inserted into the ER in the case of membrane bound and secreted proteins. In these two compartments, there are parallel molecular chaperone systems that assist and monitor the folding process to ensure high fidelity production ofproteins that can function properly. When protein folding goes awry, due to mutations or polymorphisms that alter the amino acid sequence , or abnormalities in components of the chaperone systems, misfolding can result (reviewed in ref 43) . The consequences of misfolding depend on the site of synthesis of the protein, the nature and severity of the folding defect, the relative importance of the gene product and whether protein quality control (PQC) processes have intervened sufficiently. Many misfolded and even incompletely folded ER proteins can be eliminated efficientlyby ER-associated degradation (ERAD) if they have remained in the ER for a sufficient time . Diseasesthat ensue are typically due to loss-of-function with classic examples
221
HLA-B27 Misftlding and Spondyloarthropathies
being hemophilia (Factor VIII mutations) and hereditary emphysema (a-I-antitrypsin deficiency)6O (Fig. 1). Gain-of-function phenotypes are more common and more varied. Misfolded proteins that accumulate within (e.g., forming aggresomes or inclusion bodies) or outside the cell (e.g., amyloid fibrils) can be toxic either to the involved cell or surrounding cells. Alpha-I-antitrypsin mutations can also cause pathology due to ER retention, aggregation and mitochondrial injury in the liver,61 providing a striking example ofphenotypic variation due to cell-specific differences in the handling of misfolded proteins. The cellular response to ER protein misfolding referred to as the UPR (unfolded protein response) , is part ofa more global homeostatic response to ER stress/" The UPR also plays a key role in ER expansion during the differentiation of certain cell types, such as plasma cells that become highly specialized to produce and secrete large amounts of inununoglobulins.62The UPR can also contribute to the pathogenesis ofcertain diseases with the most clear-cut examples being situations where UPR-induced apoptosis results in the lossofimportant cells,such as pancreatic 13-cells in the Akita mouse model ofdiabetes 63,64 or neural tissue in Pelizaeus-Merzbacher disease (proteolipid protein 1 in spastic paraplegia)." Another interesting example may be idiopathic
INormal Function I
Properly Folded
• I
I I I I
Galn-ol-Functlon Toxicity ·Incluslon bod ies -Aggresomes ·Amylold fibrils
n) u \". ~'!.g!~~!t~~ "--= :
Incompletely Fol ded
~o<:o,'
ILoss of Function I
"
~'Ij , '
f,'
-ER expansion ·Apoptosls -Cellular dysfunction -Enhanced Innate Immune response
Figure 1, Consequences of Protein Misfolding. Proper fold ing of newly synthesized proteins is critical for normal function. Protein misfolding has been linked to a number of diseases that can be broadly categorized as loss-of-function or gain-of-function, Loss-of-function phenotypes result from destruction of partially folded or misfolded proteins by elaborate qual ity control processes. Gain-of-function phenotypes can result from toxicity if the gene product accumulates and/or activation of cellular stress response pathways such as the UPR. HLA -B27 misfolding is hypothesized to result in gain-of-function abnormal ities through sensitization of immune response cells such as macrophages to other exogenous stimuli as reviewed in this chapter.
222
Molecular Mechanisms ofSpondyloarthropathies
inflammatory myositis. In certain forms of the disease muscle tissue (rnyocytes) exhibits robust UPR activation.f" This has been associated with caspase-I2 activation and it has been postulated that the UPR plays a role in myositis pathogenesis by promoting apoprosis, although additional mechanisms are possible. 66•67 Enforced class I upregulation (H-2Kb) via a tetracycline-regulated transgene driven by a muscle-specific promoter can result in an inflammatory phenotype that recapitulates much of the pathology seen with human disease." This is interesting and may represent an example of inappropriate expression of a class I protein, perhaps with insufficient concomitant expression of peptides, 132m and/or other chaperones such as tapasin, leading to ER stress and UPR activation. While gain- and loss-of-function classification schemes are useful, disease pathogenesis is often complex and may result from more than one consequence of protein misfolding. This is best exemplified by a-I-antitrypsin mutations that result in both types of sequelae.
HLA-B27 Misfolding The first indication that HLA-B27 had a tendency to misfold came from mutagenesis experiments where the entire 'B pocket' was changed by substituting residues from the HLA-A2 allele (creating a hybrid referred to as B27.A2B).69Remarkably, this dramatically altered the folding and assembly characteristics ofthe heavy chain with B27.A2B behaving more like HLA-A2 and other alleles that exhibit rapid folding and assembly kinetics ." Evidence that the heavy chain was in fact rnisfolding, came from experiments looking at where it was being degraded. Normally, HLA class I heavy chains are internalized from the cell surface and broken down in lysosomes. However, a proportion of HLA-B27 heavy chains were found to be dislocated from the ER membrane shortly after synthesis and before becoming associated with b 2m (and probably peptide) and then degraded in the cytosol by proteasomes. This ERAD pathway is used to eliminate ER-synthesized proteins that misfold and/or fail to assemble rapidly?" B27.A2B,as well as other naturally occurring HLA alleles that were examined, did not exhibit this behavior, thus tying misfolding to B pocket composition and the slow folding characteristic of HLA-B27.32 ERAD of HLA-B27, but not the other expressed alleles, was also detected in EBV-transformed human B-cells indicating that it occurs when there is only a single copy of the HLA-B27 gene and is not merely a consequence ofoverexpression." Interestingly, the B pocket was also found to have an unexpected dramatic effect on peptide binding affinity, in addition to its predicted effect on peptide bindingspecificiry," Since this pocket binds the side chain ofthe second amino acid in the peptide (P2), the specificity conferred by the HLA-A2-like B pocket was almost identical to what is found in peprides that bind to HLA-A2 (Leu/Met), rather than the Arg P2 specificity of HLA-B27. However, what was surprising was that HLA-B27 required a 3D-fold higher concentration of peptide (on average) to achieve the same half-maximal binding as B27.A2B. This suggests that the folding abnormality exhibited by HLA-B27 may be related to peptide binding. In other words, this allele might require more peptide to achieve release from the PLC and exit from the ER. It would follow that in situations where the supply of peptides into the ER is restricted and/or the synthesis ofheavy chains is increased, HLA-B27 folding might be disproportionately adversely affected in comparison to other alleles. Further exploration ofevents occurring in the ER for HLA-B27 revealed that the heavy chain has a tendency to form disulfide-linked complexes with itself (and possibly other proteins; unpublished observations) and exhibit prolonged association with the ER chaperone BiP (Grp78/ Hspa5).32.71.72 These features also map to the B pocket and are not exhibited by B27.A2B or other naturally occurring alleles examined to date. Further mutagenesis experiments have defined two B pocket residues that are key for HLA-B27 misfolding; Glu-45 and Cys-67 (reviewed in ref. 73). The single substitution of Met for Glu at position 45 restores rapid folding to the HLA-B27 molecule and eliminates the formation ofdisulfide-linked complexes and prolonged BiP binding (misfolding) even in the presence of Cys-67. Furthermore, the single substitution ofAla for Cys at position 67 also prevents misfolding, even when Glu-45 is intact. These observations suggested a model where two features of the HLA-B27 heavy chain might be required for misfolding to be
HLA-B27Misftlding and Spondyloarthropathies
223
prominent; slow folding and the ability to form aberrant disulfide-linked dimers via Cys-67.73 While Glu-45 and Cys-67 are not unique to HLA-B27, they are very uncommon among other alleles. In addition, there is a Lys residue at position 70 that has been reported to increase the reactivity ofthe sulfhydryl group on Cys-67,74 although it has not been studied in the context of misfolding. These three residues (Glu-45, Cys-67 and Lys-70) are virtually unique to HLA-B2T5 (www.anthonynolan.com/HIG/) and thus would support the idea that misfolding is extremely uncommon ifnot unique to this allele. Additional support for this model comes from the observation that an imposed reduction in folding rate caused by incubating cells at reduced temperature also exacerbates dimer formation and BiP binding to heavy chains," In this study, evidence was provided that Cys-I64, in addition to Cys-67, was involved in dimer formation. This observation has several possible implications since Cys-l64 is involved in the intrachain disulfide bridge between the al and a2 domains of the class I heavy chain (Cys-l 0 l-Cys-I64), which normally forms quite rapidly after heavy chain synthesis and is important for maintaining the integrity ofthe peptide-binding groove (see above) . The involvement ofCys-l64 residue in oxidative dimerization ofHLA-B27 heavy chains is potentially important as it suggests two possible scenarios related to HLA-B27 misfolding. First, if the Cys-l0l-Cys-l64 disulfide bridge forms quickly in HLA-B27 as in other alleles, then it must not be completely protected from reduction/isomerization ifit is eventually involved in dimerization, since the latter process requires it to reform a disulfide with another HLA-B27 heavy chain . Since protection ofthe Cys-IO l -Cys-l64 disulfide from reduction is a key function oftapasin-ERp57,45 the formation ofdimers via Cys-l64 could reflect HLA-B2? not interacting efficiendy with this complex in the ER. Alternatively, it may be that the al-aZ domain disulfide does not form normally in HLA-B2?, making Cys-l64 available to enter into an interchain disulfide linkage . Additional studies are needed to fully delineate the earliest events in HLA-B2? folding that lead to misfolding and its cellular consequences.
Evaluating the Role ofHLA-B27 in Disease A major advance toward understanding the role of HLA-B27 in SpA was made in the 1990s when Taurog and colleagues first produced transgenic rats overexpressing HLA-B2? and human ~ 2m (h~2m) (B27 /h~2m) .J7 High copy number B2? tran sgenic rats were found to develop a 'spont aneou s' inflammatory disease that includes gastrointestinal inflammation (colitis), arthritis, alopecia with psoriasis-like skin lesions , dystrophic nails and testicular inflammation." These phenotypic features are only partially penetrant and are variable in frequency with the exception of colitis , which occurs in 100% of transgenics. The arthritis is predominandy peripheral, although rats overexpressing additional h~2m were shown recendy to develop more severe arthritis with axial involvemenr'" (discussed below). While transgenic rats do not provide a precise phenocopy of human disease, B27 /h~2m transgenics with and without extra h~2m provide reproducible animal models that can be used to investigate pathogenic mechanisms that are likely to have relevance to human disease. Unfortunately rats are not as amenable to experimental manipulation as mice. For example, targeted gene deletion is not currently possible due to the lack ofembryonic rat stem cells.Production oftransgenics is more expensive and labor intensive, ex vivo transduction ofbone marrow cells with retroviruses is not readily achievable and many antibodies useful to visualize and/or block the function ofmouse proteins are not available for rats. Nevertheless, a great deal has been learned about the cellular requirements for disease in high copy B27 /h~2m transgenic rats (reviewed in ref 18). HLA-B2? must be expressed in the bone marrow compartment for the colitis/peripheral arthritis phenotype to occur and ubiquitous expression is not necessary," In addition, the spontaneous inflammatory disease appears to be mediated by CD4 rather than CD8 T-cells.77.78 The presence of gastrointestinal flora is also required, yet normal flora is sufficient to trigger inflammation, especially bacteroides Spp.79.80 These findings have provided strong evidence against a role for arthritiogenic (or 'colirogenic') peptides playing a central role in pathogenesis, but rather suggest that HLA-B27-expressing cellsarising from the bone marrow are either targeted by CD4 Tvcells or alternatively serve as a stimulus for these cells to become pathogenic.
224
Is HLA~B27 Recognized by CD4
Molecular Mechanisms rifSpondyloarthropathies
r.caur
Reports that CD4 T-cells can recognize normal and abnormal forms of HLA-B27 have emerged," raising the question of whether this might explain the importance of these cells for SpA-like disease in B27 /h~ 2m transgenic rats and also be involved in human disease. For human studies, CD4 T-cells were raised by stimulation with T2 cells transfected with HLA-B27.81T2 cells are missing a large region of the major histocompatibility complex (MHC) including TAP 1 and TAP2 genes and thus are unable to transport peptides into the ER from the cytosol. They havebeen reported to express HLA-B27 homodlmers.f although this was not observed in other studies.32.72 Evidence supporting the idea that CD4 T-cells could recognize HLA-B27 came from experiments using a monoclonal antibody (MEl) that recognizes HLA-B27 and could block recognition. When cells with an intact antigen presentation pathway were used, including patient-derived Bvcells, HLA-B27 was poorly recognized. In a follow-up study CD4 'T-cells from two more AS patients were raised using similar methods." These T-cells failed to recognize HLA-B27 on T2 cells, but instead appeared to be reacting to other HLA class I alleles expressed at low levels on these cells, perhaps presenting pep tides derived from degraded HLA-B27 heavy chains. In separate studies, double transgenic mice expressing HLA-B27 and a human T-cell receptor (TCR) that recognizes the HLA-B27-restricted NP383-391 flu peptide, developed CD4 as well as CD8 T-cells capable ofrecognizing this peptide presented by HLA-B27.83 If CD4 T-cells that can recognize HLA-B27 develop in rats and humans , this could have implications for disease. However, these transgenic mice represent an unusual situation where there is forced expression of the same TCR on every T-cell regardless of the costimulatory CD4/8 molecule and thus the question of whether this might occur with TCRs directed against other alleles needs to be addressed . The possibility that CD4 T-cells capable of recognizing HLA-B27 exist in transgenic rats has not, to our knowledge , been examined.
What are the Consequences ofHLA~B27 Misfolding? The observation that HLA-B27 had a propensity to misfold, as defined by the formation of disulfide-linked heavy chains and stable BiP binding, was confirmed and extended in B27/hb 2m transgenic rats." Using severaltransgenic lines with variable transgene copy number and phenotype, Tran et al demonstrated a quantitative correlation between the biochemical features ofHLA-B27 misfolding in splenocytes and the development ofSpA-like disease (colitis and arthritis). This correlation was further supported by the absence of disease in HLA-B7 (B7 /h~2m) transgenic rats, consistent with the evidence that this allele does not misfold, even when overexpressed."
The Unfolded Protein Response One ofth e consequences ofprotein misfoldingin the ER can be activation ofthe UPR (reviewed in ref 84). Some of the earliest cellular events that mark this response are phosphorylation and activation ofPERK (PKR-like ERkinase) and IRE1 (inositol requiring-I) and proteolytic cleavage ofATF6 (activating transcription factor-e) . Immediate downstream events include IRE I-mediated splicing ofthe mRNA encodingXBP-1 (X-box binding prorein-L), PERK-mediated phosphorylation of eIF2a (eukaryotic initiation factor 2 alpha) and increased transcription ofUPR target genes (e.g., BiP, CHOP and many others). The transcriptional response is mediated by activated (cleaved) ATF6, ATF4 (produced in response to eIF2a phosphorylation) and the gene product translated from the spliced XBP-I mRNA (XBP -I s), all ofwhich are active transcription factors. Several reagents used to measure proximal UPR activation (e.g., antibodies to ATF6 and phosphorylated forms of PERK and IRE1) are not available for rats. Furthermore, since the response is transient, it is more convenient to assessinduction of mRNAs encoding BiP and CHOP and splicing ofXBP-1 transcripts (XBP-1s). Using these markers, we found that spleen and thymus cells isolated from B27/h~2m transgenic rats (F344 33.3 line) exhibited little or no evidence of Uf'Racrivation.l? Similarly,bone marrow (BM)-derived macrophages from premorbid rats showed minimal differences in BiP mRNA (50% or 1.5-fold increase), whereas when BM macrophageswere prepared from animals with disease, a robust UPR was observed (5-fold increase in BiP mRNA and up to a lO-fold increase in CHOP). Microarray analyses revealed the UPR to be accompanied
HLA-B27Misfolding and Spondyloarthropathies
225
by an interferon (IFN) signature raising the question of whether IFN exposure is causing UPR activation via HLA-B27 upregulation.The conversewas alsopossible: the UPR might cause IFN upregulation. It was conceivable that both events were occurring simultaneously.
lEN Regulation ofHLA-B27Expression and UPRActivation Subsequent studies have revealed a dual role for IFNs in UPR activation in BM macrophages from B27/h132m transgenic rats. First, BM macrophagesexpressingHLA-B27 that exhibit no UPR at 'baseline' (I,e., without stimulation) will activate the UPR in response to IFN-y treatment. 34,35 This is temporally related to heavy chain upregulation and accompanied by a striking increase in the accumulation of BiP-bound heavy chains and disulfide-linked heavy chain complexes." In contrast, IFN-y does not activate the UPR in cells from nontransgenic (wild type) or B7/h132m transgenic animals. (It should be noted that there is low-levelBiP induction and XBP-l splicing «2-fold increase) with IFN-y treatment ofmacrophages from these animals, but the response in B27/h132m transgenics is substantially higher,") IFN-y has been reported to cause ER stress in oligodendrocytes, but this response was also quantitatively small (-2-fold BiP induction) and required prolonged (48 h) stimularion." This is not surprising given that cytokines and other exogenous stimuli can upregulate hundreds ofproteins, including membrane bound and secreted proteins that are produced in the ER . This low level UPR is likely part ofthe normal physiologic response that enables cells to handle the increased load. It is worth emphasizing that exacerbated HLA-B27 misfolding and UPR activation occur in the face of IFN-y-mediated upregulation of multiple components ofthe classI assembly pathway including TAP 1/2, tapasin, proteasome subunits LMP2 and LMP 7, ERAP 1 and 132m. This implies that HLA-B27 misfolding and ER stress occur despite an increased source of cargo (132m and peptide) as well as equipment necessary to load the cargo. This seems paradoxical and could indicate that one or more ofthese components exacerbate HLA-B27 misfolding, although an alternative explanation is that they may merely be insufficient to prevent misfolding. When splenocytes are treated with IFN-y, we seeonly low-levelupregulation ofHLA-B27 and minimal UPR activation . Examination ofinflamed colon tissue reveals evidence for UPR activation, although the magnitude ofincreases in BiP and CH0 P transcripts are smaller (<3-fold) than observed in isolated cells such as BM macrophages (ref 34 and manuscript submitted). Together, the se data indicate that UPR activation occurs in cells and inflamed tissues from B27/h132m transgenic rats, is specific for HLA-B27 and is temporally related to and strongly correlated with HLA-B27 misfolding. Macrophages are particularly affected by HLA-B27 misfolding in terms of UPR activation, while splenocytes, whole spleen and whole thymus tissue are nor." These results are consistent with HLA-B27 upregulation being a key component of robust UPR activation . In preliminary studies we have observed UPR activation in BM-derived dendritic cells (DCs) from B27/h132m transgenic rats treated with IFN-y, but it does not appear to be as robust as in macrophages. However, since additional stimul i can contribute to classI upregulation and we have not exhaustively examined other cell types, our understanding of the extent ofUPR activation in th ese rats remains incomplete.
lEN Induction by the UPR The second part of the IFN story, is the question of whether IFN expression is upregulated by the UPR. We found low-level induction of the Type I IFN, IFN-I3, in BM macrophages undergoing a UPR, either due to HLA-B27 upregulation or in cells treated with pharmacologic agents (tunicamycin or thapsigargin) that cause ER seress," consistent with a previous report of low-level induction in tunicamycin-treated fibroblasts. 87 IFN-13 has well-recognized autocrine effects at low concenerarionsP"? and thus UPR-induced IFN -13 may have immunological consequences including a pro -survival effect on macrophages." However, perhaps more important is the response observed when macrophages undergoing a UPR are exposed to ligands for pattern recogn ition receptors (e.g., Toll-like receptors or TLRs). TLR4 and TLR3 agonists such as LPS and dsRNA , that upregulate IFN-13 via the TRIF (Toll-like receptor/IL-l receptor related adaptor protein inducing IFN-I3)-dependent pathway, cause robust synergistic IFN-13 production in
226
Molecular Mechanisms ofSpondyloarthropathies
cells exhibiting ER stress. The synergistic response appears to require XBP-ls, but not PERK or ATF6 activation. These results suggest a fundamental relationship between ER stress and innate immune signaling with implications beyond HLA-B27 and disease, as well as a novel function of XBP-l in the convergence of these important signaling pathways.
The UPR and Cytokine Production Links between the UPR and cytokine induction have been reported in the literature. IL-6 production from plasma cells after activation by LPS or CD40 ligation is influenced by XBP-l , although this effect is considerablydelayed and may be secondary to other changes/" Macrophages loaded with cholesterol exhibit UPR activation and increased production of TNF-a and IL-6, effects that appear to be secondary to NF -kB,]NK1I2, p38 and/or Erk1l2 aceivaeion." Using a microarray-based screening approach , we identified IL-23p19 (the unique subunit of the active IL-23 cytokine), as being synergistically induced by LPS-treatment of cells with an active UPR (ref 86 and manuscript submitted). We have found IL-23p19 upregulation in inflamed tissue and myeloid cells derived from the tissue, in B27/h132mtransgenic rats. Il-23p19 is up regulated in a temporal and spatial manner that is consistent with it being involved in the development of colon inflammation. In addition, there is robust upregulation ofIL-17 in the inflamed colon that localizes to CD4 T-cells in the lamina propria and draining mesenteric lymph nodes (ref. 93 and manuscript in preparation). These findings are of interest in the context of several recent developments in our understanding ofT-cell biology, as well as new evidence for genes involved in susceptibility to AS.5
ThI, Th2 and ThI7 Upon antigenic stimulation, naive CD4 T-cells differentiate into T helper (Th) cells with specialized cytokine production profiles and effector functions. The ThllTh2 paradigm established over 20 years ago was that Thl cells produce large quantities ofIFN-y and are essential for clearing intracellular pathogens, while Th2 cells produce IL-4, 5 and 13 and are important for clearance of extracellular organisms and robust humoral immunity.94.95 Key cytokines that drive these two pathways are IL-12 (IL-12p70) and IL-4 . IL-12 induces Thl differentiation through STAT4 activation in T-cells and IL-4 promotes Th2 development through STAT6 and GATA-3 activation, promoting more IL-4 producrion." IFN-yfrom an initial innate immune response (e.g., activated NK cells) is also important for activating the T-bet transcription factor through STATI signalling, which in turn activates Th I-specific genes. Recently, a third subset ofeffector CD4 T-cells characterized by IL-I7 production ('Th 17') has been discovered.F"? Th17 cellsmay have evolved as another arm of the adaptive immune response for enhanced protection against extracellular bacteria (i.e., Klebsiella pneumoniae), protozoa and fungi (e.g., Pneumocystiscarinii) by recruiting neutrophils. However, additional roles for Th 17 in immune defense are possible. What has become very clear, is that Th 17 cells playa crucial role in chronic inflammation in animal models ofhuman auroimrnune/auroinflarnmatory diseases such as RA , MS,IOO.IOI IBD 102.10 3 and psoriasis'?' and there is growing evidence that IL-I7 is a crucial pro-inflammatory cytokine in the human disease counterparts. In addition to IL-I7, ThI7 cells can produce TNF-a and IL_6.100•102 IL-I7 can act on several cell types including macrophages, fibroblasts, endothelial cellsand epithelial cells, to up regulate TNF-a, IL-6, IL-I , as well as several chemokines and metalloproteases (including MMP-3 which has been shown to be a good biomarker for AS).105.107 Thus , downstream effects ofIL-I7 are diverse and highly pro-inflammatory. Several cytokines play key roles in Th 17 development and the balance between Th 17 and regulatory T-cells (Treg) in mice. For example, the combination ofTGF-13 and IL-6 drives naive CD4 T-cells to become Th 17-committed 108.109 through induction of the retinoic acid orphan receptor (RORyt) in naive 'Tcells, which then leads to upregulation ofthe IL-23 receptor (IL-23R)Yo IL-23 can then act on Th 17-competent cells stimulating robust and prolonged IL-17 up regulation 111.112 (and reviewed in ref. 113). In addition, TCRstimulation by MHC class Il-restricred antigens can induce IL-I7 production without IL-23.
227
HLA-B27Misfolding and Spondyloarthropathies
In mice it appears that CD4 T-cells producing IFN-y (Th l) and IL-17 (Th17) are distinct populations, while in humans CD4 T-cells producing IFN-y and IL-17 (ThllTh17) have been documented in the gut of humans with Crohns disease."! In addition, the factors that regulate Th 17 development in humans appear to be different from micewith IL-23 and IL-l13 playing a more important role than IL_6.115 In addition to the predominant form ofIL-17 (IL-17A or CTLA-8) produced by CD4 T-cells, there is an extended family with fiveadditional IL-17 molecules whose cellular source and regulation need to be further defined.':" Other cells that have been reported to produce IL-17 include CDS and gamma/delta T-cells, neutrophils and even macrophages and lymphocytes at sites ofinfection.
HLA-B27 Misfolding, the UPR and the IL-23/IL17Axis: Refining the Hypothesis Preliminary results linking HLA-B27 misfoldingand the UPR to enhanced IL-23 induction in macrophages in response to TLR agonises, together with evidence for activation ofthe Th 17 axis in transgenic rats, suggests a novel paradigm for the development ofHLA-B27-associared colitis (Fig. 2). In the gastrointestinal tract, a low level immune response to bacterial colonization could result in increased expression ofIFNs (Type I and/or Type II) , perhaps via innate immune stimuli (IFN-13) and/or NK cell activation (IFN-y). This would result in upregulation of class I expression and, in cells expressing HLA-B27, activation ofthe UPR. Macrophages would then become sensitized to TLR agonists such as LPS and other bacterial products, polarizing them toward increased production ofIFN-13 and IL-23 and possibly more IL-6. IL-23 would then drive IL-17 production from CD4 'T-cells that have become committed to the Th17lineage. While IL-6 and TGF-13 have been shown to be important for the development ofTh 17 T-cells,99 there is evidence that cells with the capacity to produce IL-17 are normally present in the colon . 102,103 Thus , in this unique mucosal environment, increased IL-23 expression could be a sufficient stimulus for chronic gastrointestinal inflammation. This issupported by the observation that IL-23p19 transgenic mice TLR Agonlsls
"~ • t IL-6, TNF-u, rIL-17 . , / IL·1
-: ~ IL.1 7 "'IL·17
Macrophage
I
..
,,
IL·12
~
rn,
Figure 2. Proposed Paradigm Linking HLA-B27 Misfolding to Innate Immune Activation. The tendency of HLA-B27 to misfold and activate the UPR when upregulated sensitizes cells to certain pathogen-associated molecular patterns and possibly damage -associated molecular patterns, many of which signal through pattern recognition receptors such as the Toll-like receptors (TLR Agonists). Enhanced upregulation of IL-23 promotes IL-17 production from CD4 T-cells of the Th17 lineage. Th17 cells can produce TNF-a and IL-6 and IL-17 is also a potent pro-inflammatory cytokine that acts on many tissue cell types and further induces TNF-a, IL-6 and IL-l as well as chemokines and metalloproteinases. IL-17 is hypothesized to be a key pro-inflammatory cytokine in the immunopathology that develops in the colon of HLA-B27 transgenic rats.
228
Molecula r Mechanisms ofSpondyloarthropathies
develop widespread inflammation without any other additional stimulus.'!" In the HLA-B27 transgenic rats, increased IFN-~ expression might serve to promote HLA-B27 up regulation and also activate NK cells.'!" It is also possible that unusual forms ofHLA-B27 expressed on the cell surface might engage leukocyte receptors and serve as an activating stimulus for NK cells." It is interesting to note that IFN-y can inhibit the Th17 axis.97 •98 In the model we propose for the development of colitis, IFNs would play an important role in promoting IL-23 production via class I upregulation and subsequent UPR activation, but could conceivably inhibit the IL-23/ IL-17 axis through effects on Th17 development. We and others, have documented IFN-y overexpression in the inflamed colon,18.34 but since CD4 T-cells with the capacity to produce IL-17 may already exist in this location, IFN-ymay have little effect on their development. Furthermore, we do not know the relative importance of Type I vs Type II IFNs in HLA-B27 upregulation in rats in vivo, nor whether Type I IFNs have the same inhibitory effect on Th17 development, although this might be expected given the overlap in effects of Type I and Type II IFNs. These and other questions, including the relative importance ofthe Th 1 axis in transgenic rats, need to be further addressed.
HLA-B27 Subtypes and Spondyloarthritis There is heterogeneity within the HLA-H27 group of alleles referred to as subtypes (www. anthonynolan.com/HIG/). The numerical classification for subtypes is to designate them with an asterisk preceding the number (e.g., B"2701, B"2702, etc.). More than 30 subtypes have been reported for HLA-B27 and since most occur infrequently, little is known about their association with AS or SpA. While most of the relatively common subtypes (e.g., B"2705 , B"2702 , B"2704) have been associated with disease, for some time B"2706 and B"2709 have been thought to be exceptions. Since hypotheses explaining how HLA-B27 might cause disease have been driven by our understanding of how it differs from other HLA-B alleles, it should be possible to refine our ideas based on properties of subtypes differentially associated with disease. However, the caveat with this approach is that incomplete or incorrect information about disease associations may lead to incorrect conclusions. There are now new data suggesting that B"2709 may be associated with disease, or at the very least the situation is more complex than previously thought.I 18 Patients with B"2709 who developed SpA were reported several years ago,119.120 and now there are reports ofthis subtype in AS patients.121.122 A recent examination ofthe existing data suggests that B"2709 occurs in these individuals at a greater frequency than by chance alone, ll8thus supporting the idea that this subtype may indeed be associated with disease. The occurrence of8*2709 on a distinct haplotype from B"2705 in the same population,123 along with genetic evidence that additional MHC-encoded genes influence susceptibility, 3.124raises the possibility that the offending alleles are not present on the B"2709 haplotype.!" It is well known that most individuals with HLA-B27 (and B"2705 by inference) do not develop AS/SpA and HLA-B27-positive family members of patients with AS are at much higher risk for disease than the general HLA-B27-positive population. One recently proposed hypothesis isthat B"2709 arose by a single mutation from B"2705 on a low-risk haplotype and that it may be the low-risk haplotype that is more important for disease predisposition than the immunobiological differences between the B"2705 and B"2709 proteins. I 18Additional genetic differences between populations with B"2709 and B"2705 might also contribute. The casefor a lack ofassociation between B"2706 and disease is more compelling , in part because it is present in a much larger and probably more genetically diverse population.!" However,patients with AS and this subtype have also been reported l 26 with two additional casesdescribed recently'" Subtype associations (or lack thereof) need to be extended to larger populations and investigated for the possible existence of MHC haplotypes such as those uncovered in Sardinia.!" Another pitfall ofusing the genetic association data to drive hypotheses about disease causation is that subtyping ofHLA-B27 has traditionally focused on coding sequence variation, with little attention to the promoter region ofthe gene. Promoter polymorphisms, which are known to exist," could have consequences for baseline and inducible HLA-B27 subtype expression.
HLA-B27 M isfolding and Spondyloarthropathies
229
Overexpression ofAdditional h132m: The Model ofSpA The phenotype exhibited by high copy B27/hfJ2m transgenic rats. where colitis and peripheral arthritis predominate. does not include an important component ofAS-axial inflammation and ankylosis. Recently Tran et al found that overexpressing more hb.m by introducing an additional 35 copies ofthe hfJ2mtransgene altered the phenotype ofhigh copy B27/hfJ2m transgenic rats (55 copies ofHLA-B27 and 66 copies ofhfJ2m)}9Rats with 55 copies ofHLA-B27 and 101 copies of hb .m had more frequent and more severe arthritis involving the axial skeleton. while colitis was not affected . In addition. the extra 35 copies ofhfJ2m were able to induce arthritis in intermediate copy B27/hfJ2m transgenic rats (20 copies ofHLA-B27 and 15 copies ofhfJ2m) that normally remain free ofany spontaneous disease. Thus. spondylitis was induced by additional hfJ2m even in the absence ofcolitis . These observations are potentially important as they provide a model system that may be relevant to the pathogenesis ofaxial inflammation. The mechanism by which additional hfJ2m modifies the phenotype of B27 /hl32m transgenic rats is not clear. Based on observations that the additional hfJ2m increased the folding kinetics of HLA-B27. reduced the formation of aberrant disulfide linked heavy chain complexes and resulted in a reduction ofBiP mRNA expression (-25-30%) in splenocytes, the authors concluded that while HLA-B27 misfolding was still associated with intestinal inflammation. it was not critical to the development of HLA-B27-associated arthropathy. However, this conclusion is premature, since UPR activation was not examined after upregulation ofHLA-B27. which we have shown is critical for this response.r v" In addition, since there is some cell type specificity to HLA-B27-induced UPR activation, it will be important to examine cells that are likely to be relevant to disease pathogenesis. Preliminary experiments suggest that while the additional h132m reduces the magnitude ofUPR activation when HLA-B27 is upregulated, it does not eliminate it (unpublished observations) and thus the role ofHLA-B27 rnisfolding in the spondyloarthritis phenotype will require further investigation. It is also important to consider that UPR activation might be a 'double-edged' sword in the pathogenesis of inflammatory disease. Its consequences could depend on the magn itude of the response. For example, it is well known that a strong and unresolved UPR can lead to apoptosis. IfUPR activation in macrophages drives an inflammatory process due to abnormal cytokine production, one could envision downstream effects being different if the cells causing the problem are destined to undergo UPR-induced apoptosis.The consequences ofinappropriate in vivo UPR activation in the immune system are relatively unexplored and it is also likely that we do not yet appreciate precisely what needs to be examined. Our ability to approach these qu estions would be aided greatly by the development of a mouse model, where many more tools are available to address these complex issues.
Conclusions Recent advances in deciphering genetic susceptibility to AS point toward the lL-23 receptor
(IL23R) gene.l This gene encodes a protein that combines with another subunit lL-12RbI to form the active IL-23 receptor expressed on developing ThI7 Tvcells,'!' making them responsive to IL-23. IL23R polymorphisms have also been implicated in susceptibility to Crohns disease and psoriasis, other diseases that have phenotypic overlaps with spondyloarthriris.P''{" Preliminary data indicating that HLA-B27 misfolding may be a stimulus for activating the IL-23/IL-I7 axis, suggests a novel mechanism that may explain . at least in part, the role ofHLA-B27 in colitis in transgenic rats. The striking convergence of the human genetic data and results from HLA-B27 transgenic rats provides a compelling argument that this axis needs to be further examined in SpA and AS.
References 1. Brown MA , Laval SH. Brophy S et al. Recurrence risk modelling of the genetic susceptibility to ankylosing spondylitis. Ann Rheum Dis 2000; 59:883-886 . 2. Laval SH , Timms A, Edwards S et al. Whole-genome screening in ankylosing spondylitis: Evidence of NonMHC genetic-sus ceptibility loci. Am J Hum Genet 2001; 68:918-926.
230
Molecular Mechanisms ofSpondyloarthropathies
3. Zhang G, Luo J, Bruckel J et al. Genetic studies in familial ankylosing spondylitis susceptibility. Arthritis Rheum 2004: 50:2246-2254. 4. Sladek R, Rocheleau G, Rung J er al. A genome-wide association study identifies novel risk loci for type 2 diabetes. Nature 2007, 445 :881-885. 5. Burton PR, Clayton DG, Cardon LR et al. Association scan of 14,500 nonsynonymous SNPs in four diseases Identifies autoimmunity variants. Nat Genet 2007, 39:1329-1337. 6. Reveille JD. Major histocompatibility genes and ankylo sing spondylitis. Best Pract Res Clin Rheumatol 2006: 20 :601-609. 7. Brown MA. Non-major-histocompatibility-complex genetics of ankylosing spondylitis . Best Pracr Res Clin Rheurnatol 2006 , 20:611-621. 8. Smith JA, Marker-Hermann E, Colbert RA. Pathogenesis of ankylosing spondylitis: Current concepts . Best Pract Res Clin Rheumatol 2006 : 20:571-591. 9. Lopez de Castro JA . HLA-B2 7 and the pathogenesis of spondyloarrhropathies. Immunol Lett 2007: 108:27-33. 10. Breban M , Hacquard-Bouder C, Falgarone G. Animal models of HLA-B27-associated diseases. Curr Mo l Med 2004: 4:31-40. 11. Kievits F, Ivanyi P, Krimpenfort P et al. HLA-restricted recognition of viral antigens in HLA transgenic mice. Nature 1987 : 329:447-449. 12. Nickerson CL, Luthra HS, Savarirayan S er al. Susceptibility of HLA-B27 transgenic mice to yersinia enterocolitica infection . Hum Immunol 1990: 28 :382-396. 13. Khare SD, Luthra HS, David CS. Spontaneous inflammatory arthritis in HLA-B27 transgenic mice lacking fI,-microglobulin : a model of human spondyioarrhroparhies. J Exp Med 1995, 182: 1153-11 58. 14. Khare SD, Hansen J, Luthra HS et al. HLA-B27 heavy chains contribute to spontaneous inflammatory disease in B27/human fI,-microgiobulin (fI,m) double transgenic mice with disrupted mouse fI,m. J Clin Invest 1997, 98 :2746-2755. IS. Khare SD, Hansen J, Luthra HS et al. Spontaneous inflammatory disease in HLA-B27 transgenic mice is independent of MHC class II molecules: a direct role for B27 heavy chains and not B27-derived peptides. J Immunol1998: 160:101-106 . 16. Kingsbury DJ, Mear JP, Witte DP er al. Development ofspontaneous arthritis in fI,-microglobulin-deficient mice without expression of HLA-B27 : associarion with deficiency of endogenous major histocompatibility complex class I expression. Arthritis Rheum 2000: 43 :2290-2296. 17. Hammer RE, Maika SD, Richardson JA et al. Spontaneous inflammatory disease in transgenic rats expressing HLA-B27 and human fI,-m: an animal model of HLA-B27-associated human disorders . Cell 1990: 63 :1099-1112. 18. Taurog j D, Maika SD. Sarumrira N er al. Inflammatory disease in HLA-B27 transgenic rats. Immunol Rev 1999 : 169:209-223. 19. Tran TM, Dorris ML, Satumtira N et al. Additional human beta(2)-microglobulin curbs HLA-B27 misfolding and promotes arthritis and spondylitis without colitis in male HLA-B27-transgenic rats. Arthritis Rheum 2006: 54:1317-1327. 20. Hildebrand WHoTurnquist HR, Prilliman KR et al. HLA class I polymorphism has a dual impact on ligand binding and chaperone interaction. Hum Immunol 2002: 63 :248-255. 21. Yao Z. Volgger A, Scholz S et al. Sequence polymorphism in the HLA-B promoter region. Immunogenetics 1995 : 41 :343-353 . 22. Yu DY, Choo SY, Schaack T. Molecular mimicry in HLA-B27-related arthritis. Ann Int Med 1989 ; 111:581-591. 23. Benjamin RJ, Parham P. Guilt by association : HLA-B27 and ankylosing spondylitis. Immunol Today 1990; 11:137-142. 24. Allen RL , O'Callaghan CA , McMichael AJ er al. Cutting edge : HLA-B27 can form a novel beta 2-microglobulin-free heavy chain homodimer structure. J Immunol 1999; 162:5045-5048. 25. Malik P, K1imovitsky P, Deng LW et al. Uniquely conformed peptide-containing beta 2-microglobulin-free heavy chains of HLA-B2705 on the cell surface. J Immunol 2002; 169:4379-4387. 26. Edwards JCW, Bowness P, Archer JR . Jekyll and H yde: The transformation of HLA-B27 . Immunol Today 2000; 21:256-260. 27. Kollnberger S, Bird LA, Sunm M-Y et al. Cell surface expression and immune receptor recogmion of HLA-B27 homodimers. Arth Rheum 2002 ; 46 :2972-2982. 28. Bird LA, Peh CA , Kollnberger S et al. Lymphoblastoid cells express HLA-B27 homodimers both imracellularly and at the cell surface following endosomal recycling. Eur J Immunol 2003; 33 :748-759. 29. Allen RL, Trowsdale J. Recognition of classical and heavy chain forms ofHLA-B27 by leukocyte receptor s. Curr Mol Med 2004 ; 4:59-65.
HLA-B27Misftlding and Spondyloartbropathies
231
30. Kollnberger S. Bird LA. Roddis M er aJ. HLA -B27 heavy chain homodimers are expressed in HLA-B27 transgenic rodent models of spondyloarthrids and are ligands for paired Ig-like receptors. J Immunol 2004 ; 173:1699-1710. 31. Mear JP. Schreiber KL. Munz C et al. Misfolding of HLA-B27 as a result of its B pocket suggests a novel mechanism for its role in susceptibility to spondyloarthropathies. J Immunol 1999; 163:6665-6670. 32. Dangoria NS. DeLay ML . Kingsbury DJ et aJ. HLA-B27 misfolding is associated with aberrant intermolecular disulfide bond formation (dirnerlzatlon) in the endoplasmic reticulum. J Bioi Chern 2002 ; 277 :23459-23468. 33. Colbert RA. HLA-B27 misfolding: A solution to the spondyloarrhropathy conundrum? Mol Med Today 2000; 6:224-230. 34. Turner M]. Sowders DP. DeLay ML er al. HLA-B27 misfolding in transgenic rats is associated with activation of the unfolded protein response. J Immunol 2005; 175:2438-2448. 35. Turner MJ, Delay ML, Bai S et al. HLA-B27 up-regulation causes accumulation of misfolded heavy chains and correlates with the magnitude of the unfolded protein response in transgenic rats: Implications for the pathogenesis of spondylarthritis-like disease. Arthritis Rheum 2007 ; 56:215-223. 36. Laitio P, Virtala M, Salmi M er al. HLA-B2 7 modulates intracellular survival of salmonella enteritidis in human monocytic cells. Eur J Immunol 1997; 27:1331 -1338. 37. Virtala M. Kirveskari J. Granfors K. HLA-B27 modulates the survival of salmonella enteritidis in transfect ed L cells, possibly by impaired nitric oxide production. Infect Immun 1997; 65:4236-4242. 38. Pentrinen MA. Heiskanen KM. Mohapatra R er al. Enhanced intracellular replication of Salmonella enteritidis in HLA-B2 7-expressing human monocytic cells: Dependency on glutamic acid at position 45 in the B pocket of HLA-B27. Arthritis Rheum 2004; 50:2255-2263. 39. Sahlberg AS. Penrtinen MA, Heiskan en KM et al. Evidence that the p38 MAP kinase pathway is dysregulated in HLA-B27-expressing human monocytic cells: Correlation with HLA-B27 misfolding. Arthritis Rheum 2007 ; 56:2652-2662. 40. Penttinen MA, Ekman PGranfors K. Non-antigen presenting effectsof HLA -B27. Curr Mol Med 2004; 4:41-49 . 41. Bridges JP, Xu Y. Na CL er al. Adaptation and increased susceptibility to infection associated with constitutive expression of misfolded src. J Cell Bioi 2006; 172:395-407. 42. Penttinen MA , Holmberg CI . Sistonen L et al. HLA -B27 modulates nuclear factor kB activation in human monocytic cells exposed to lipopolysaccharide. Arthritis Rheum 2002 ; 46:2172-2180. 43. Gregersen N. Bross P, Yang S et al. Protein Misfolding and Human Disease. Annu Rev Genomics Hum Genet 2006 ; 7:103-124 . 44. Hamme r GE, Kanaseki T. Shastri N. The final touches make perfect the peptide-MHC class I repertoire . Immunity 2007 ; 26:397-406 . 45. Kienast A, Preuss M. Winkler M er al. Redox regulation of peptide receptivity of major histo compat ibility complex class I molecules by ERp57 and tapasin. Nat Immunol 2007; 8:864-872. 46. Bouvier M. Accessory protein s and the assembly of human class I MHC molecules : a molecular and structural perspective. Mol Immunol 2003 ; 39:697-706. 47. Dick TP. Assembly of MHC class I peptide complexes from the perspective of disulfide bond formation . Cell Mol Life Sci 2004; 61:547-556. 48. Peh CA. Burrows SR, Barnden M et al. HLA-B27-restricted antigen presentation in the absence of tapasin reveals polymorphism in mechanisms of HLA class I peptide loading. Immunity 1998; 8:531-542. 49. Purcell AW, Gorman JJ, Garcia-Peydro M et al. Q!:antitative and qualitat ive influences of rapasin on the class I peptide repert oire. J Immunol 2001 ; 166:1016-1027. SO. Dick T. Bangia N. Peaper DR er al. Disulfide bond isomerization and the assembly of MHC class l-pepnde complexes. Immunity 2002 ; 16:87-98 . 51. Serwold T. Gaw S, Shastri N. ER aminopeptidases generate a unique pool of peprides for MHC class I molecules. Nat Immunol 2001 ; 2:644-651. 52. Serwold T. Gonzalez F, Kim J er al. ERAAP customizes peprides for MHC class I molecules in the endoplasmic reticulum . Nature 2002; 419 :480-483 . 53. Hammer GE. Gonzalez F. Champsaur M et al, The aminopeptidase ERAAP shapes the peptide reperto ire displayed by major histocompatibility complex class 1 molecules. Nat Immunol 2006 ; 7:103 -112. 54. York lA, Brehm MA, Zendzian S et al. Endoplasmic reticulum aminopeptidase 1 (ERAPl) trims MHC class Lpresenred peptides in vivo and plays an important role in immunodominance. Proc Nat! Acad Sci USA 2006; 103:9202-9207 . 55. Kanaseki T. Blanchard N, Hammer GE et al. ERAAP synergizes with MHC class I molecules to make the final cut in the antigenic peptide precursors in the endoplasmic reticulum . Immunity 2006 ; 25:795-806. 56. Hammer GE, Gonzalez F, James E et al. In the absence of aminopeptidase ERAAP, MHC class I molecules present many unstable and highly immunogenic peptides. Nat Immunol 2007; 8:101-108.
232
Molecular Mechanisms ofSpondyloarthropathies
57. Cui X. Hawari F. Alsaaty S er al. Identification of ARTS-l as a novel TNFRl-binding protein that promotes TNFRI eccodornain shedding. J Clin Invest 2002; 110:515-526. 58. Cui X. Rouhani FN. Hawari F et al. Shedding of the type II IL-l decoy receptor requires a multifunctional aminopeptidase, aminopeptidase regulator of TNF receptor type 1 shedding. J Immunol 2003 ; 171:6814-6819. 59. Cui X, Rouhani FN, Hawari F et al. An aminopeptidase, ARTS-I , is required for interleukin-6 receptor shedding. J Bioi Chern 2003 ; 278 :28677 -28685 . 60. Schroder M, Kaufman RJ. The mammalian unfolded protein respons e. Annu Rev Biochem 2005; 74 :739-789. 61. Perlmutter DH. Liver injury in alpha l-anritrypsin deficiency : An aggregated protein induces mitochondrial injury. J Clin Invest 2002 ; 110:1579-1 583. 62. Iwakoshi NN, Lee AH, Vallabhajosyula P et al. Plasma cell differentiation and the unfolded protein response intersect at the transcription facror XBP-1. Nat Immunol 2003; 4:321-329. 63. Oyadomari S. Araki E, Mori M. Endoplasmic reticulum stress-mediated apoprosis in pancreatic beta-cells. Apoptosis 2002; 7:335-345. 64. Ron D. Proreoroxicity in the endoplasmic reticulum: Lessons from the Akita diabetic mouse. J Clin Invest 2002; 109:443-445. 65. Southwood CM. Garbern J. Jiang W er al. The unfolded protein response modulates disease severity in Pelizaeus-Merzbacher disease. Neuron 2002 ; 36:585-596. 66. Nagaraju K. Casciola-Rosen L, Lundberg I er aI. Activation of the endoplasmic reticulum stress response in autoimmune myositis: potential role in muscle fiber damage and dysfunction. Arthritis Rheum 2005 ; 52:1824-1835. 67. Griffin TA, Reed AM. Pathogenesis of myositis in children. Curr Opin Rheumatol 2007 ; 19:487-491. 68. Nagaraju K, Raben N. Loeffler L et al. Conditional up-regulation of MHC class I in skeletal muscle leads to self-sustaining autoimmune myositis and myositis-specific autoantibodies. Proc Nat! Acad Sci USA 2000; 97:9209-9214. 69. Colbert RA, Rowland-Jones SL, McMichael AJ er al. Allele-specific B pocket transplant in class I major histocompatibility complex protein changes requirement for anchor residue at P2 of peptide. Proc Nat! Acad Sci USA 1993; 90:6879-6883. 70 . Meusser B, Hirsch C. Jarosch E er al. ERAD : the long road ro destruction. Nat Cell Bioi 2005 ; 7:766-77271. Tran TM, Satumtira N, Dorris ML et al. HLA-B27 in transgenic rats forms disulfide-linked heavy chain oligorners and multimers that bind to the chaperone BiP. J Immunol 2004; 172:5110-5119. 72. Antoniou AN. Ford S. TaurogJD et al. Formation of HLA-B27 homodimers and their relationship to assembly kinetics . J Bioi Chern 2004 ; 279 :8895-8902. 73. Colbert RA. The immunobiology of HLA-B27: variations on a theme. Curr Mol Med 2004; 4:21-30 . 74 . Whelan MA. Archer JR . Chemical reactivity of an HLA-B27 thiol group. Eur J Immunol 1993 ; 23 :3278-3285. 75. Kostyu DD. Hannick LI, Traweek JL er al. HLA class I polymorphism : Structure and function and still quest ions. Hum Immunol 1997; 57:1-18 . 76. Breban M, Hammer RE. Richardson JA ec al. Transfer of the inflammatory d isease of HLA-B27 transgenic rats by bone marrow engrafimenr, J Exp Med 1993; 178:1607 -1616. 77. Breban M, Fernandez-Sueiro JL . Richardson JA et al. Tvcells, but not thymic exposure to HLA-B27, are required for the inflammarory disease of HLA-B27 transgenic rats. J ImmunoI1996; 156:794-803. 78. May E. Dorris ML. Satumtira N et al. CD8all T-cells are not essential to the pathogenesis of arthritis or colitis in HLA-B27 transgenic rats. J Immunol2003; 170:1099-1105. 79 . Rath HC, Herfarth HH. Ikeda JS et al. Normal luminal bacteria. especially bacteroides species. mediate chronic colitis, gastritis and arthritis in HLA-B27/human Il, microglobulin transgenic rats. J Clin Invest 1996; 98:945-953. 80. Rarh HC, Wilson KH, Sartor RE. Differential induction of colitis and gastritis in HLA-B27 transgenic rats selectively coloni zed with Bacteroides vulgatus or Escherichia coli. Infect Immun 1999; 67:2969-2974. 81. Boyle LH, Goodall JC, Opac SS et al. The recognition of HLA-B27 by human CD4+ T-Iymphocytes. J Immunol 2001; 167:2619-2624. 82. Boyle LH, Goodall JC, Gaston JS. Major hlstocornpatibiliry complex class I-restricted alloreactive CD4+ T-cdls. Immunology 2004; 112:54-63 . 83. Roddis M , Carter R\V, Sun MY et al. Fully functional HLA B27-restricted CD4+ as well as CD8+ T-cell responses in TCR transgenic mice. J Immunol 2004; 172:155-161. 84. Ron D. Walter P. Signal integration in the endoplasmic reticulum unfolded protein response. Nat Rev Mol Cell Bioi 2007; 8:519-529.
HLA-B27Misfolding and Spondyloarthropathies
233
85. Lin W. Harding HP. Ron D et al. Endoplasmic reticulum stress modulates me response of myelinating oligodendrocyres to the immune cyrokine interferon-gamma. J Cell BioI 2005 ; 169:603-612. 86. Smith JA. Turner MJ. DeLay ML . K1eck EI. Sowders DP. Colbert RA. Endoplasmic reticulum stress-induced and the unfolded protein response are linked to synergistic IFN-II induction via X-box bind ing protein-L Eur J Immunol 2008 ; 38:1194-1203. 87. Lee AH. Iwakoshi NN. Glimcher LH . XBP-l regulates a subset of endoplasmic reticulum resident chaperone genes in the unfolded protein response. Mol Cell BioI 2003; 23:7448 -7459 . 88. Taniguchi T. Takaoka A. A weak signal for strong responses: Interferon-alpha/beta revisited . Nat Rev Mol Cell BioI 2001 ; 2:378-386. 89. Montoya M. Schiavoni G. Mattei F ec al. Type I interferons produced by dendritic cells promote their phenotypic and functional activation . Blood 2002; 99:3263-3271. 90. Gautier G. Humbert M. Deauvieau F et al. A type I interferon autocrine-paracrine loop is involved in Toll-like receptor-induced interleukin-12p70 secretion by dendritic cells . J Exp Med 2005 ; 201 :1435-1446. 91. Seimon TA . Obstfeld A. Moore KJ et al. Combinatorial pattern recognition receptor signaling alters the balance of life and death in macrophages. Proc Nat! Acad Sci USA 2006 ; 103:19794-19799. 92. Li Y. Schwabe RF. DeVrics-Seimon T et al. Free cholesterol-loaded macrophages are an abundant source of tumor necros is factor-alpha and interleukin-6: model of NF-kappaB- and map kinase-dependent inflammation in advanced athero sclerosis. J BioI Chern 2005; 280 :21763-21772 . 93. Colbert RA. Turner MJ. DeLay ML et al. HLA-B27 misfolding activates the I1-23/I1-17 axis via the unfolded pro tein response in transgenic rats: evidence for a novel mechanism of inflammation. Arrh Rheum 2007; 54:S515 . 94. Mosmann TR. Coffman RL. TH I and TH2 cells: Different patterns of Iymphokine secret ion lead to different functional properties. Annu Rev Immuno11989; 7:145-173. 95. Bottomly K. A functional dichotomy in CD4+ T vlymphocyre s. Immunol Today 1988; 9:268-274. 96. Langrish CL . McKenzie BS. Wilson NJ et al. IL-12 and IL-23 : Master regulators of innate and adaptive immun ity. Immunol Rev 2004; 202 :96-105 . 97. Harrington LE. Hatton RD. Mangan PR et al. Interleukin 17-producing CD4+ effector T-cells develop via a lineage di stinct from the T helper type 1 and 2 lineages. Nat Immunol 2005 ; 6:1123-1132. 98. Park H. Li Z. Yang XO et al. A distinct lineage of CD4 T-cells regulates tissue inflammarion by producing interleukin 17. Nat Immunol 2005 ; 6:1133 -1141. 99. Weaver CT. Hatton RD. Mangan PR et al. IL-17 family cyrokines and the expanding diversity of effector T-cell lineages. Annu Rev Immunol 2007 ; 25:821-852. 100. Cua D], Sherlock J, Chen Yet al. Interleukin-23 rather than interleukin-12 is the critical cytokine for autoimmune inflamm ation of the brain. Nature 2003 ; 421 :744-748. 101. Chen Y, Langri sh C L, McKenzie B et al. Anti -IL-23 therapy inh ibit s multiple inflamm atory pathways and amelio rates autoimmune encephalom yelitis. J Clin Invest 2006 ; 116:1317-1326. 102. Yen D. CheungJ, Scheerens H er al. IL-23 is essential for T-cell-mediated colitis and promotes inflammation via IL-17 and IL-6. J Clin Invest 2006 ; 116:1310-1316. 103. Uhlig HH. McKenzie BS, Hue S er aI. Differential activity ofIL-12 and IL-23 in mucosal and systemic innate immune pathology. Immunity 2006 ; 25:309-318. 104. Zheng Y. Danilenko DM. Valdez P er aI. Inrerleukin-Zz, a T(H)17 cytokine, mediates IL-23-induced dermal inflammation and acanthosis . Nature 2007 ; 445 :648-651. 105. Chen CH. Lin KC. Yu DT et aI. Serum matrix meralloproteinases and tissue inhibitors of metalloproteinases in ankylosing spondylitis: MMP-3 is a reproducibly sensitive and specific biomarker of disease activity. Rheumatology (Oxford) 2006 ; 45:414-420. 106. Nakae S. Nambu A. Sudo K ct aI. Suppression of immune induction of collagen-induced arthritis in IL-17-deficient m ice. J Immunol 2003; 171:6173-6177. 107. Kolls JK, Linden A. Interleukin-I? family members and inflammation. Immunity 2004; 21:467-476. 108. Mangan PR . Harrington LE. O'Quinn DB et aI. Transforming growth factor-beta induces development of the T(H) 17 lineage. Nature 2006 ; 441 :23 1-234. 109. McGeachy MJ, Bak-Jensen KS. Chen Yet aI. TGF-beta and IL-6 drive the production of IL-17 and IL-I0 by T -cells and restrain T(H)-17 cell-mediated pathology. Nat Immunol 2007 ; 8:1390-1397. 110. Ivanov. II. McKenzie BS. Zhou L et al. The orphan nuclear receptor RO Ryt directs the differentiation program of pr oinflammatory IL-17+ T helper cells. Cell 2006 ; 126:1121 -1133 . Ill. Aggarwal S. Ghilardi N. Xie MH ec aI. Interleukin-23 promotes a distinct CD4 T-cell activation state characte rized by the production of interleukin-17. J BioI Chern 2003 ; 278 :1910-1914. 112. Langrish CL. Chen Y. Blumenschein WM er aI. IL-23 drives a pathogenic T-cell population that induces autoimmune inflammation. J Exp Med 2005 ; 201 :233-240. 113. Kastelein RA. Hunter CA. Cua DJ. Discovery and biology ofIL-23 and IL-27 : related but funct ionally distinct regulators of inflammation. Annu Rev Immunol 2007 ; 25:221-242.
234
Molecular Mechanisms ofSpondyloarthropathies
114. Annunziato F, Cosmi L, Santarlasci V et al. Phenotypic and functional features of human Th 17 cells. J Exp Med 2007 : 204 :1849-1861. 1I5. Wilson NJ, Boniface K, Chan JR er al. Development , cyrokine profile and function of human inrcrleukin 17-producing helper T-cells. Nat Immunol2007: 8:950-957. 116. Wiekowski MT, Leach M\V, Evans EW et al. Ubiquitous transgenic expression of the IL-23 subunit p 19 induces mulriorgan inflammation, runting, infertility and premature death. J Immunol 200 I : 166:7563-7570. 117. Biron CA, Nguyen KB, Pieri GC et al. Natural killer cells in antiviral defense : Function and regulation by innate cytokines. Annu Rev Immunol 1999: 17:189-220. 118. Taurog JD. The mystery of HLA-B27: if it isn't one thing, it's another. Arthritis Rheum 2007 ; 56:2478-248 I. 119. Olivieri I, Padula A, Ciancio G et al. The HLA-B*2709 subtype in a patient with undifferentiated spondarthritis. Ann Rheum Dis 2000: 59:654-655 . 120. Olivieri I, Ciancio G. Padula A er al. The B*2709 subtype does not give absolute protection against spondyloarthroparhy, Arthritis Rheum 2000: 43:S265. 121. Olivieri I. D'Angelo S. Scarano E et al. The HLA-B*2709 subtype in a woman with early ankylosing spondylitis. Arthritis Rheum 2007 : 56:2805-2807. 122. Cauli A, Vacca A. Marnell A et al. A Sardinian patient with ankylo sing spondylitis and HLA-B*2709 co-occurring with HLA-B*1403. Arthritis Rheum 2007: 56:2807-2809. 123. Fiorillo M'I, Cauli A. Carcassi C ct al. Two distinctive HLA haplotypes harbor the B27 alleles negatively or positively associated with ankylosing spondylitis in Sardinia: implications for pathogenesis. Arch Rheum 2003 : 48 :1385-1389. 124. TaurogJD. HLA-DR4 and the spondyloarrhropathies. Ann Rheum Dis 2002: 61:193-194. 125. Lopez-Larrea C. Sujirachato K, Mehra NK et al. HLA-B27 subtypes in Asian patients with ankylosing spondylitis: evidence for new associations. Tissue Antigens 1995: 45 :169-176. 126. Gonzalez-Roces 5, Alvarez MY, Gonzalez S et al. HLA-B27 polymorphism and worldwide susceptibility to ankylosing spondylitis. Tissue Antigens 1997: 49 :1I6-123. 127. Hou TY, Chen HC. Chen CH et al. Usefulness of human leucocyte antigen-B27 subtypes in predicting ankylosing spondylitis: Taiwan experience. Intern Med J 2007; 37:749-752. 128. Duerr RH, Taylor KD, Brant SR et al. A genome-wide association study identifies IL23R as an inflammatory bowel disease gene. Science 2006: 314:1461-1463. 129. Cargill M. Schrodi SJ, Chang Met al. A large-scale genetic association study confirms ILI2B and leads to the identification of IL23R as psoriasis-risk genes. Am J Hum Genet 2007: 80:273-290.
CHAPTERl7
HLA-B27 and Host-Pathogen Interaction Anna S. Sahlberg, Kaisa Granfors and Markus A. Penttinen'"
Abstract
H
LA-B27 is a risk factor closely associated to spondyloarthropathies (SpA). One form of SpA is reactive arthritis (ReA). which develops as a complication after certain bacterial infections (e.g., Salmonellae. Yersiniae,Shigellae, Campy/obacteriae and Chlamydiae).1he development ofinfection-triggered complication is a complex train ofevents between the triggering bacteria and the ho st. Since mo st ofthe patients suffering from ReA are HLA-B27 posit ive. it has been proposed that HLA-B27 may modulate the interaction between ReA-triggering bacteria and ho st cell. Besides antigen presenting fun ction . HLA-B27 displays other unusual properties that might be ofimportance in th e development of ReA.1hese properties (homodimer formation and misfolding of HLA-B27 heavy chain in the endoplasmic reticulum (ER)) may trigger ER-stress signaling pathways in host cell. wh ich in turn may modulate cellsignaling in favor ofReA-triggering bacteria. Here we summarize the observations ofHLA-B27 modulating the interaction between ReA-triggering bacteria and host cell and discuss potential mechanisms behind the interaction.
Introduction The development ofinfection is a complex train ofevents influenced by host-pathogen interactions . The susceptibility and severity of infectious diseases are influenced by a variety of factors. among them host genetics. One of the best characterized risk facto rs for disease susceptibility is human leukocyte ant igen (H LA)-B27 . which is strongly associated to the development ofspondyloarthropathies (SpA). some of th em being triggered by infection.l'
Molecular Mechanisms ofSpondyloarthropathies. edited by Carlos Lopez-Larrea and Roberto Diaz-Pefia. ©2009 Landes Bioscience and Springer Science+Business Media.
236
Molecular Mechanisms ofSpondyloarthropathies
antigen presenting molecule whereas others favour the idea that HLA-B27 possessadditional functions unrelated to antigen presentation,which contribute to disease susceptibility (Table 1). At the moment none ofthese theories completely explains the role ofHLA-B27 in disease pathogenesis. The natural function ofHLA-B27, as MHC class I molecule, is antigen presentation to T-cells. 13 Thus , T-cells have been implicated as triggering cells in the pathogenesis ofSpA for several decades . Despite active research the critical role ofT-cells in the development ofSpA has not been proved. In fact, recent evidence obtained from studies concerning HLA-B27 transgenic rats suggest that antigen presenting cells (dendritic cells and macrophages) playa role in the development of the disease in HLA-B27 transgenic rats. ' 4-' 6Therefore, dendritic cells and macrophages are also of interest, although yet less studied than T-cells. In this chapter, we summarize the observations suggesting that HLA-B27 modulates the interaction between ReA-triggering bacteria and host cell and discuss potential mechanisms by which HLA-B27 may interfere activation ofintracellular signaling pathways in host cell triggered by ReA -triggering bacteria .
The Molecular Characteristics ofHLA-B27 MHC class I molecules are constructed in the endoplasmic reticulum (ER) and consist of MHC l-encoded heavy chain (He), sz microglobulin (g2m) and a small peptide (usually 8-10 amino acids in length).'? Normally, the complexes are exported to the cell surface. However, ifthe complex is not properly formed, it will never reach the cell surface and will be degraded intracellularly by process called endoplasmic reticulum (ER) -associated degradation (ERAD ).18 Thus, the formation of stable complexes in proper three dimensional structure in the ER is necessary for exporting MHC class I molecules to the cell surface (chapter by Colbert et al.). HLA-B27 is actually a family of approximately 30 subtypes'" and the alleles HLA-B2702, -B2704 and -B2705 are associated to the disease susceptibility of SpA (chapter by Reveille et al.).20 In contrast, alleles like HLA-B2706 and -B2709 are not considered to associate to the susceptibility despite some alleging results. The subtypes differ from each other only by amino acid substiturionls).":" For example, HLA-B2705 and HLA-B2709 differ by one amino acid at the peptide binding groove."
HLA-B27 and Host-Pathogen Interaction The Antigen PresentingFunctions ofHLA-B27andHost-Pathogen Interaction The classical function ofMHC classI molecules like HLA-B27 is peptide presentation to CD8+ T-cells. 18There are several theories suggesting that antigen presenting function ofHLA-B27 is altered and that leads to modulated inflammatory reaction and possibly to the development ofthe disease . These theories are based on a similarity between foreign (arthritogenic peptide derived
Table 1. Hypotheses for the role of HLA-B27 in the pathogenesis of spondyloarthropathies Antigen presenting functions Arthritogenic peptide Aberrant cell surface heavy chains Autodisplay
Non-antigen presenting functions HLA-B27 Misfolding and dimerization in the ER Modulation of intracellul ar survival of ReA-triggering bacteria Modulation of host cell signaling
HLA-B27 and Host-Pathogen Interaction
237
from external antigens) and self-peptide." A peptide from an external antigen, such as the intracellular bacteria or a ubiquitous virus, might be presented by HLA-B27 and trigger a B27-restricted cytotoxic T-Iymphocyte (CTL) response. If this pathogenic peptide would show molecular or an tigenic mimicry with a constitutive self-ligand ofHLA-B27, some activated CTL might overcome tolerance against that peptide, leading to autoimmunity, tissue injury and inflammation. This idea issupported by finding that HLA-B27-restricted CD8+ T-cell clones with specificity for bacteria or possibly self-peptides are formed both in synovial fluid and peripheral blood ofpatients with ReA and AS25 as well as from the finding ofrecurrent structural motifs among the T-cell receptors (TCR) ofHLA-B27-specific CTL isolated from unrelated individuals." Furthermore, CD8+ T-cells reacting with collagen-derived self-peptides in a B27-dependent way were detected in the synovial fluid ofpatients with ankylosing spondylitis (AS).27 Several findings support the arthritogenic peptide theory suggesting that the HLA-B27 sequence itself includes sequences with homology to proteins from arthritogenic gram-negative bacteria.28.29This also lead to identification ofHLA-B27 ligands derived from its own molecule, presenting such homology.28.3o.31 A promising study in this direction was the identification of a self-ligand of HLA-B27, derived from the cytoplasmic tail of its own and other HLA class I molecules, that is constitutively presented by three AS-associated subtypes, HLA-B270S, -B2702 and -B2704, but was not found in HLA-B2706 or -B2709. This peptide shows striking similarity with a sequ ence ofthe DNA primase of Chlamydia tracbomatis/" Unlike other MHC classI molecules, HLA-B27 (HLA-B270S) heavy chain (He) has adistinet feature to misfold in the endoplasmic reticulum (ER )32 and form disulfide-linked heavy chain homodimers in ER (chapter by Colbert et al.).33 These dimers are also expressed on cell surface. 34.35 It has been suggested that these aberrant HLA-B27 HC dimers on cell surface could be recognized by MHC class I receptors on various type ofleukocytes.24.35.36 This raises the possibility that instead of CD8+ T-cells rather CD4 + T-cells or NK cells may playa role in the pathogenesis of SpA .24 In line with this idea, study by May et al indicates that HLA-B27-resrricted CD8a~ T-cells may not serve as effector cells in the transgenic rat model of HLA-B27-associated disease.'? Another recent theory is auto- or cross-display." Based on this theory, HLA-B27 molecule's backbone rotates and this results on occupying its own peptide binding cleft.
HLA-B27 and Host Cell-Pathogen Interaction ReA-TriggeringBacteria The most common causative bacteria triggering chronic form of ReA include Salmonellae, Yersiniae, Shigellae, Campylobacteriae and Chlamydiae. 3The interaction between ReA-triggering bacteria and host cell is a fine-tuned complex event. For example, Salmonellauses Type III secretion system (TTSS) to mediate the transfer ofa battery ofbacterial proteins into host cells." Through molecular and functional mimicry ofhost cell proteins, these bacterial effectors ofvirulence can manipulate a variety ofcellular activities.t?Such effector proteins can interfere Rho family GTPases which can regulate actin dynamics, cell adhesion, macropinocytosis and phagocytosis.t' " The stimulation of Rho family GTPases also leads to activation of the Jun N-terminal Kinase (JNK) and p38 mitogen activated protein kinase (MAPK) signaling pathways as well as transcription factors such as nuclear factor kappa B (NF-KB), resulting in the reprogramming ofhost cell gene expression and the production ofproinflammatory cytokines.44.45 The interaction between another ReA-triggering bacteria, Yersinia enterocolitica and host cell is also widely studied. Y.enterocolitica contains a virulence plasmid (pYV) . Besides Type III secretion apparatus it encodes Yersinia outer protein (yop) family.46 These effectors proteins can modulate general inflammatory response. For example, Hoffmann et al47 have shown that in murine macrophages pYV-carrying yersinia both suppresses the induction of several proinflammatory genes and induces sustained expression of genes with silencing functions whereas infection with pYV-cured yersinia induces several genes evoking phenomenon resembling general inflammatory response." In conclusion, these detailed
238
Molecular Mechanisms ofSpondyloarthropatbies
studies have significantly increased our understanding how delicate systems intracellular bacteria uses to adapt in intracellular milieu in host cell.
HLA-B27 and Survival ofIntracellular ReA-Triggering Bacteria The causative bacteria for ReA have common characteristics: all are facultative or obligative intracellular bacteria. Moreover, all these bacteria express LPS on their outer membrane.v' It has been shown that bacterial antigens such as LPS, persist in the joints of patients suffering from chronic form of ReA. 3,5·12 Also high IgA and IgG antibody concentrations against the triggering microbes have been reported to persist in patients suffering from ReA.3.48,49 These observations together have raised the question whether the interaction between the intracellular bacteria and HLA-B27 positive ReA patients is abnormal. The ability to survive intracellularly is crucial to Salmonella for successful disease pathogenesis." Therefore, we aimed to study whether HLA-B27 modulates the interaction between host cell and Salmonella in favour ofpathogen. Two steps are probably required for intracellular pathogen to cause a chronic infection: the bacteria can actively invade and survive in the host cell. Some studies have proposed that HLA-B27 may modulate the invasion ofReA-triggering bacteria in host cell51·54 but results seem to be dependent on cell type and methods used in the study.55,56 In contrast, studies by us using cell lines as a model suggest that HLA-B27 modulates the survival of Salmonella in host cell,57.6O We observed impaired elimination of Salmonella enteritidis in B27-transfected human monocytic cells and mouse fibroblasts. 57,6O Also others have reported a 14-fold increase in intracellular Yersinia 7 days afier infection in HLA-B27-positive human fibroblasts when compared with fibroblasts obtained from HLA-B27 negative individuals." However, there are several studies indicating that HLA-B27 has no direct impact on the survival of ReA-triggering bacteria in several cell types .53.54.62 As a conclusion, it seems that HLA-B27 may modulate the behaviour of certain host cells e.g., macrophages and fibroblasts. There are several mechanisms ReA-triggering bacteria are activating while persisting in macrophages. For example, toll-like receptor 4 (TLR4) -mediated signaling pathway is activated in phagolysosornes/" TLR4-dependent pathway activates mitogen-activated protein kinases (MAPKs) P38,]NK and extracellular-regulated kinases (ERKs) an d NF-KB leading to the induction ofseveral cytokines and inflammatory mediators, which are subsequently released from macrophages (Fig. 1). As discussed earlier, to manipulate the activation ofthese important signaling pathways, ReA-triggering bacteria (Salmonella, Yersinia) release bacterial effector proteins to the host cell cytosol. Thus, one potential mechanism by which HLA-B27 may interfere the interaction between macrophages and/or fibroblasts and ReA-triggering bacteria is to modulate the crosstalk between bacterial effector proteins and crucial signaling molecules such as MAPKs and NF-KB.This assumption is supported by our recent finding showing that HLA-B27 -expressing human monocytic cells display a defect in the regulation ofp38 in response to S. enteritidis infection." In addition other evidence exists supporting this assumption, since we observed that HLA-B27-expressing mouse fibroblasts show impaired nitric oxide (NO) production'" and others reported that S. typhimurium promotes the induction ofc-fos in HLA-B27-transfected Hel.a-cells.v' The development of the disease in HLA-B27 transgenic rat model seems to require high transgene copy number and high expression levels of HLA-B27 together with the presence of resident bacterial flora.65 This provides evidence for both the role of HLA-B27 and bacteria in the disease pathogenesis.65,66 Interestingly, these rats have reported to show enhanced susceptibility to the infection caused by an intracellular pathogen Listeria monocytogenes suggesting that handling ofintracellular bacteria may be impaired in macrophages ofHLA-B27 transgenic rats.67 Taken together, these results imply that HLA-B27 transgenic rats may have compromised ability to manage infection caused by intracellular bacteria. However, detailed studies are required to elucidate these mechanisms.
239
HLA-B27 and Host-Pathogen Interaction
Macrophage cytoplasm
Bacteria containing LPS
I ERK
Extracellular space Cell membrane
Figure 1. TLR4-dependent signaling in macrophages.
HLA-B27, ER Stress-Induced Signal Transduction Pathways and Host-Pathogen Interaction The underlying mechanism by which HLA-B27 modulates me interaction between Salmonella and monocytic cells is unclear. 58Our studies, however, revealed that me dysregulation ofp38 observed in HLA-B27 monocytic cells is dependent on misfolding feature ofHLA-B27, since normal regulation of p38 was observed in cells expressing nonmisfolding mutant form of HLA-B27. 59 Furthermore, results indicate that the ability ofthe host cell to resist intracellular replication ofS. enteritidis was depending on misfolding ofHLA-B27 H C. S. enteritidis replication was observed in cells expressing wild-type HLA-B27, whereas no significant replication was detected in cells expressing nonmisfolding mutant ofHLA-B27. 58.59Studies are ongoing in our laboratory to reveal the exact mechanism by which misfolding ofHLA-B27 impairs the ability ofhuman monocytic cells to resist the replication of intracellular S. enteritidis . The most plausible assumption is that the accumulation of misfolding HLA-B27 HCs in monocytic cells would trigger ER-stress termed unfolded protein response (UPR) (for details, see chapter by Colbert), which in turn would modulate intracellular signaling pathways contributing to the regulation of the inflammatory response. UPR is activated to help the cell to manage to tolerate the increase ofmisfolded proteins in the ER,68.69 which are in high excess toxic to the cell and may trigger apoptosis. To date, our results however suggest that UPR is not activated in U937 cells, as we did not observe increase in mRNA levels in UPR-inducible genes CHOP and BiP in HLA-B27-expressing U937 cells.58 Studies are ongoing to revealwhether other signaling molecules reported to be activated and/or induced by ER stress are altered in U937 HLA-B27-expressing cells. One potential candidate is double stranded RNA-activated kinase (PKR), a cellular kinase reported to be activated by ER stress," There is evidence that PKR also modulates apoptosis triggered by Salmonella and Yersini« in mouse macrophages." In addition, PKR has been reported
240
Molecular Mechanisms ofSpondyloarthropathies
to modulate the activation ofp38, possibly by binding to mitogen-activated protein kinase kinase (MKK6), the upstream activator ofp38 (Fig. 2 ).72 Therefore altered regulation ofPKRwould offer a possible explanation to observations in HLA-B27 expressing U937 cells. When compared with the stably rransfected celllines (e.g., HLA-B27-expressing U937 cells), the situation in HLA-B27 transgenic rats may be different since an activation of the UPR in macrophages in HLA-B27 transgenic rats has been reported."
HLA-B27 and LPS-Induced Inflammatory Response Culturable bacteria have not been detected from synovial samples of ReA-patients. Antigens such as LPS derived from the causative bacteria (Salmonella , Yersinia) have been found in the affected joints several years after infection. 3.5.6 LPS found in joints is in highly degraded and processed form 5.6 which is known to be a strong antigen capable ofstimulating the production of various inflammatory mediators thereby activating immune reactions.Y'Ihus, LPS derived from ReA-triggering bacteria may trigger ReA symptoms in the affected joints. TNFa is considered to be in a central role in the disease pathogenesis, since anti- TNFa therapies are higWy effective in the treatment of SpA,?5-77 Moreover, TNFa is known to be a key mediator in the pathogenesis of several models of inflammatory arthritis." These results raise the question whether HLA-B27 would influence LPS-induced TNFa production in HLA-B27 positive individuals. In fact, there is evidence that HLA-B27 may directly interfere LPS-induced TNFa production in HLA-B27 positive individuals." Monocyte/macrophages are mostly responsiveofLPS-induced TNFa production in the body.80 The induction ofTNFa synthesis in monocyte-macrophages upon LPS stimulation involves LPS binding to CD 14-TLR4-MD2-complex, which in turn triggers signaling events thereby activating downstream targets MAPKs ERKs,p38.]NK and NF-KB (Fig. 1).80Based on the results suggesting that HLA-B27 may interfere LPS-induced TNFa production,we aimed to study more specifically whether HLA-B27 modulates LPS-induced cascades leading to the activation ofTNFa in U937 monocytic cells. We observed that in HLA-B27-expressing human mononcytic cells the activation ofNF-KB was enhanced and prolonged upon LPS stimularion." In addition, the secretion Endoplasmic
reticulum
~
....... ~
\
---::::::::::::::::::: Figure 2. The schematic presentation how double stranded RNA-activated kinase (PKR) may modulate Salmonella activated signaling in macrophages.
HLA-B27 and Host-Pathogen Interaction
241
ofTNFa upon LPS stimulation was more rapid when compared with the control cells.81 Instead, in a recent study by Goodall et al82 no effect ofHLA-B27 on TNFa production was found in HLA-B27 positive U937 transfectants. Moreover, they did not observe any aberrant regulation of NF-KB-responsive genes in HLA-B27 transfectants at mRNA level.Therefore, they concluded that any effect HLA-B27 expression might have on the response to LPS, is rather subtle and HLA-B27 is not likely to have any regulatory effects on NF-KBsignaling.Taken together, controversial results exist whether HLA-B27 could directly modulate LPS-induced TNFa production in monocyte/ macrophages. To further elucidate the significance of these findings, the molecular basis for the observations requires to be revealed.
Conclusions The role ofHLA-B27 as a risk factor for the development of ReA and several other forms of SpA has been known for several decades. Despite extensive research the molecular mechanism by which HLA-B27 confers disease susceptibility remains to be elucidated. Results obtained from HLA-B27 transgenic rats and in vitro studies using cell lines support direct role ofHLA-B27 in the pathogenesis. The function of HLA-B27 as an antigen presenting molecule has been widely studied. More recently, unusual properties ofHLA-B27 (misfolding in the ER, homodimerization) have been linked to the disease progression." To date, it is unclear whether antigen presentation or these unusual properties ofHLA-B27 are important in diseasepathogenesis. Therefore, future studies should be conducted to aim to understand the importance ofHLA-B27 at the molecular level.
Acknowledgements Our original studies were supported financially by the Academy of Finland and the Sigrid juselius Foundation.
References 1. Aho K, Ahvonen P, Lassus A et aI. HL-A 27 in reactive arthritis . A study ofYersinia arthritis and Reiter's disease. Arthritis Rheum 1974; 17(5) :521-6. 2. Brewerton DA , Hart FD, Nicholls A er aI. Ankylosing spondylitis and HL-A 27. Lancet 1973 ; 1(7809) :904-7 . 3. Granfors K. Do bacterial antigens cause reactive arthritis ? Rheum Dis Clin North Am 1992; 18(1) :37-48 . 4. Penttinen MA , Ekman P, Granfors K. Non-antigen presenting effects of HLA-B27. Curr Mol Med 2004 : 4(1):41-9. 5. Granfors K, Jalkanen S, Lindberg AA er aI. Salmonella lipopolysaccharide in synovial cells from patients with reactive arthritis. Lancet 1990; 335(8691):685-8. 6. Granfors K, Jalkanen S, von Essen R et aI. Yersinia antigens in synovial-fluid cells from patients with reactive arthritis. N Engl J Med 1989 : 320(4) :216-21. 7. Granfors K. Merilahti-Palo R, Luukkainen R er aI. Persistence of Yersinia antigens in peripheral blood cells from patients with Yersinia enrerocolinca 0:3 infection with or without reactive arthritis. Arthritis Rheum 1998 ; 41(5):855-62. 8. Gerard HC, Branigan PJ, Schumacher HR, Jr et aI. Synovial Chlamydia trachomatis in patients with reactive arthriris/Reiter's syndrome are viable but show aberrant gene expression . J Rheumatol 1998 ; 25(4) :734-42. 9. Gerard HC, Schumacher HR, EI-Gabalawy H et aI. Chlamydia pneumoniae present in the human synovium are viable and metabolically active. Microb Pathog 2000; 29(1) :17-24. 10. Villareal C. Whittum-Hudson JA, Hudson AP. Persistent Chlamydiae and chronic arthritis. Arthritis Res 2002; 4( 1):5-9. 11. Ekman P, Nikkari S, Purto-Laurila A er aI. Detection of Salmonella infantis in synovial fluid cells of a pat ient with reactive arthritis. J Rheumarol 1999: 26(11) :2485-8. 12. Gaston JS, Cox C, Granfors K. Clinical and experimental evidence for persistent Yersinia infection in reactive arthritis. Arthritis Rheum 1999; 42(10):2239-42. 13. Gaston H . Mechanisms of Disease: The irnmunoparhogenesis of spondyloarthropathies. Nat Clin Pract Rheumatol 2006; 2(7):383-92.
242
Molecular M echanisms ofSpondyloarthropatbies
14. Hacquard-Bouder C, Chimenti M, Giquel B et al. Alteration of antigen-independent immunologic synapse formation between dendritic cells from HLA-B27-transgenic rats and CD4+ T-cells: Selective impairment of cosrimularory molecule engagement by mature HLA-B27. Arthritis Rheum 2007 ; 56(5) :1478-89. 15. Turner MJ, Delay ML, Bai S et al. HLA-B27 up-regulation causes accumulation of misfolded heavy chains and correlates with the magnitude of the unfolded protein response in transgenic rats : Implications for the pathogenesis of spondylarthritis-like disease. Arthritis Rheum 2007 ; 56(1):215-23. 16. Hacquard-Bouder C, Falgarone G, Bosquet A et al. Defective cosrimularory function is a striking feature of antigen -presenting cells in an HLA-B27-transgenic rat model of spondylarrhropachy, Arthritis Rheum 2004; 50(5): 1624-35. 17. Cresswell P, Bangia N, Dick T et al. The nature of the MHC class 1 peptide loading complex . Imrnunol Rev 1999; 172:21-8. 18. Pamer E, Cresswell P. Mechanisms ofMHC class I-Restricted antigen processing . Annu Rev Immunol 1998 ; 16:323-58. 19. Taurog JD. The mystery of HLA-B27: If it isn't one thing, it's another. Arthritis Rheum 2007; 56(8) :2478-81. 20. Ramos M, Alvarez 1, Sesma L et al. Molecular mimicry of an HLA-B27-derived ligand of arthritis-linked subtypes with chlamydial proteins. J BioI Chern 2002 ; 277(40):37573-81. 21. Ramos M, Paradela A, Vazquez M et al. Differential association of HLA-B*2705 and B*2709 to ankylosing spondylitis correlates with limited peptide subsets hilt not with altered cell surface stability. J BioI Chern 2002; 277(32):28749-56. 22. Sesma L, Montserrat V, Lamas JR et al. The peptide repertoires of HLA-B27 subtypes differentially associated to spondyloarrhroparhy (B*2704 and B*2706) differ by specific changes at three anchor positions. J Bioi Chern 2002; 277(19):16744-9. 23. Fiorillo MT, Greco G, Sorrentino R. The Asp1l6-His1l6 substitu tion in a novel HLA-B27 subt ype influences the acceptance of the peptide C-terminal anchor. Immunogenetics 1995; 41(1):38-9. 24. Smith J, Marker-Hermann E, Colbert R. Pathogenesis of ankylosing spondylitis: current concepts . Best Pract Res Clin Rheumatol 2006 ; 20:571-91. 25. Hermann E, Yu DT, Meyer zum Buschenfelde KH er al. HLA-B27-restricted CD8 T-cells derived from synovial fluids of patients with reactive arthritis and ankylosing spondylitis. Lancet 1993 ; 342(8872) :646-50. 26. May E, Dulphy N, Frauendorf E et al. Conserved TCR beta chain usage in reactive arthritis; evidence for selection by a putative HLA-B27-associated autoanrigen. T issue Antigens 2002 ; 60(4):299-308. 27. Atagunduz P, Appel H, Kuon Wet al. HLA-B27-restricted CD8+ T-cell response to cartilage-derived self peptides in ankylosing spondylitis. Arthritis Rheum 2005 ; 52(3):892-901. 28. Scofield RH , Kurien B, Gross T et al. HLA-B27 binding of peptide from its own sequence and similar peptides from bacteria: implications for spondyloarrhroparhies. Lancet 1995 ; 345(8964) :1542-4. 29. Schwimmbeck PL, Yu DT, Oldstone ME. Autoantibodies to HLA B27 in the sera ofHLA B27 patients with ankylosing spondylitis and Reiter 's syndrome. Molecular mimicry with Klebsiella pneumoniae as potential mechanism of autoimmune disease. J Exp Med 1987; 166(1):173-81. 30. Alvarez I, Sesma L, Marcilla Met al. Identification of novel HLA-B27 ligands derived from polymorphic region s of its own or other class I molecules based on direct generation by 20 S proteasome. J BioI Chern 2001; 276(35):32729-37. 31. Frauendorf E, von Goessel H, May E et al. HLA-B27-restricted T-cells from patients with ankylosing spondylitis recognize peptides from B*2705 that are similar to bacteria-derived peptides. Clin Exp Immunol 2003 ; 134(2):351-9. 32. Mear JP, Schreiber KL, Munz C er al. Misfolding of HLA-B27 as a result of its B pocket suggests a novel mechanism for its role in susceptibility to spondyloarthropathies. J Immunol 1999; 163(12) :6665-70. 33. Dangoria NS, DeLay ML, Kingsbury OJ et al. HLA-B27 misfolding is associated with aberrant intermolecular disulfide bond formation (dimerizatlon) in the endoplasmic reticulum. J BioI Chern 2002; 277(26) :23459-68. 34. Allen RL, O 'Callaghan CA , McMichael AJ et al. Cutting edge : HLA-B27 can form a novel beta 2-microglobulin-free heavy chain homodimer structure. J Immunol 1999 ; 162(9):5045-8. 35. Kollnberger S, Bird L, Sun M et al. Cell-surface expression and immune receptor recognition ofHLA-B27 homodimers. Arthritis Rheum 2002 ; 46(11) :2972-82. 36. Kollnberger S, Bird L, Roddis M er al. HLA-B27 heavy chain homodimers are expressed in HLA-B27 transgenic rodent models of spondyloarthritls and are ligands for paired Ig-like receptors . J Immunol 2004; 173(3):1699-710. 37. May E, Dorris ML, Satumtira N er al. CD8 alpha beta Tvcells are not essential to the pathogenesis of arthritis or colitis in HLA-B27 tran sgenic rats. J Immunol 2003 ; 170(2) :1099-105.
HLA-B27 and Host-Pathogen Interaction
243
38. Luthra-Guprasarma M. Singh B. HLA-B27 lacking associated beta2-microglobulin rearranges to auto-display or cross-display residues 169-181: a novel molecular mechan ism for spondyloarthroparhies. FEBS Lett 2004 ; 575(1 -3):1-8. 39. Galan ]E . Wolf-Watz H . Protein delivery into eukaryotic cells by type III secretion machines. Nature 2006 ; 444(7119):567-73. 40. Stebbins CE oGalan ]E. Structural mimicry in bacterial virulence. Nature 2001; 412(6848):701-5. 41. Rosenberger CM . Finlay BB. Phagocyte sabotage: Disruption of macrophage signalling by bacterial pathogens . Nat Rev Mol Cell Bioi 2003; 4(5) :385-96. 42. Patel ]C. Galan ]E. Differential activation and function of Rho GTPases during Salmonella-host cell int eractions. ] Cell Bioi 2006 ; 175(3):453-63. 43. Bishop AL. Hall A. Rho GTPases and their effector proteins . Biochern ] 2000 ; 348(Pt 2):241-55. 44. Chen L. Hobbie S. Galan ]. Requirement of CDC42 for Salmonella-induced cyroskeleral and nuclear responses. Science 1996 ; 274(5295):2115-8. 45. Hobbie S. Chen L. Davis R et al. Involvement of mitogen-activated protein kinase pathways in the nuclear responses and cytokine production induced by Salmonella typhimurium in cultured intestinal epithelial cells.] Immunol 1997; 159(11):5550-9. 46. Rihl M. K10s A. Kohler L et al. Infection and musculoskeletal conditions: Reactive arthritis . Best Pract Res Clin Rheumatol 2006; 20(6) :1119-37. 47. Hoffinann R. van Erp K. Trulzsch K et al. Transcriptional responses of murine macrophages to infection with yersinia enrerocol irica. Cell Microbiol 2004; 6(4):377-90. 48. Granfor s K. Toivanen A. IgA-anti-yersinia antibodies in yersinia triggered reactive arthritis . Ann Rheum Dis 1986; 45(7) . 49. Maki-Ikola O. Leirlsalo-Repo M. Kantele A er al. Salmonella-specific antibodies in reactive arthritis. ] Infect Dis 1991; 164(6) :1141-8. 50. Fields PI. Swanson RY, Haidaris CG et al. Mutants of Salmonella typhimurium that cannot survive within the macrophage are avirulent. Proc Nad Acad Sci USA 1986; 83(14) :5189-93. 51. Kapasi K. Inman RD. HLA-B27 expression modulates gram-negative bacterial invasion into rransfected L cells.] Immunol 1992; 148(11) :3554-9. 52. Kapasi K. Inman RD. MEl epitope of HLA-B27 confers class l-rnediated modulation of gram-negative bacterial invasion.] Immunol 1994; 153(2):833-40. 53. Saarinen M. Ekman P. Ikeda M et al. Invasion of Salmonella into human intestinal epithelial cells is modulated by HLA-B27. Rheumatology (Oxford) 2002; 41(6 ):651-7. 54. Huppertz HI. Heesernann ]. Invasion and persistence of Salmonella in human fibroblasts positive or negative for endogenous HLA B27. Ann Rheum Dis 1997; 56(11):671-6. 55. Vaharniko S. Penrrinen M. Granfors K. Aetiolog y and pathogenesis of reactive art hritis: role of non-ant igen-presenting effects of HLA -B27. Arthritis Res Ther 2005 ; 7(4) :136-4 1. 56. Granfors K. Ho st-microbe inter action in reactive arthritis: does HLA-B27 have a direct effect ? ] Rheumatol 1998; 25(9 ):1659-61. 57. Laitio P. Virtala M. Salmi M et al. HLA-B27 modulates intracellular survival of Salmonella enteritidis in human monocyt ic cells. Eur] Immunol 1997; 27(6 ):1331-8. 58. Penttinen MA. Heiskanen KM. Mohapatra R er al. Enhanced intracellular replication of Salmonella enteritidis in HLA-B27-expressing human monocytic cells: dependency on glutamic acid at position 45 in the B pocket of HLA-B27. Arthritis Rheum 2004; 50(7 ):2255-63 . 59. Sahlberg AS. Penttinen MA . Heiskanen KM et al. Evidence that the p38 MAP kinase pathway is dysregulated in HLA-B27-expressing human monocytic cells: correlation with HLA-B27 misfolding. Arthritis Rheum 2007 : 56(8) :2652-62 . 60. Virtala M. Kirveskari ], Granfors K. HLA-B27 modulates the survival of Salmonella enteritidis in tran sfected L cells. possibly by impaired nitric oxide production. Infect Immun 1997; 65(10) :4236-42 . 61. Huppertz HI. Heesemann ]. The influence of HLA B27 and interferon -gamma on the invasion and persistence ofyersinia in primary human fibroblasts. Med Microbiol Immunol 1996; 185(3):163-70. 62. Young]L . Smith L. Matyszak MK et al. HLA -B27 expression does not modulate intracellular Chlamydia trachomatis infection of cell lines. Infect Immun 2001; 69(11 ):6670-5. 63. Blander ]M. Medzhitov R. Regulation of phagosome maturation by signals from toll-like receptors . Science 2004 ; 304(5673):1014-8 . 64. lkawa T. Ikeda M. Yamaguchi A er al. Expression of arthritis -causing HLA-B27 on Hela cells promotes induction of c-fos in response to in vitro invasion by Salmonella typhimurium. ] Clin Invest 1998; 101(1):263-72. 65. Taurog]D. Maika SO. Satumtira N er al. Inflammatory disease in HLA-B27 transgenic rats. Immunol Rev 1999; 169:209-23. 66. Taurog ]0. Richardson ]A. Croft ]T er al. The germfree state prevents development of gut and joint inflammatory disease in HLA-B27 transgenic rars.} Exp Med 1994; 180(6) :2359-64.
244
Molecular Mechanisms ofSpondyloarthropatbies
67. Warner TF, Madsen J, Starling J et aI. Human HLA-B27 gene enhances susceptib ility of rats to oral infection by Listeria monocyrogenes. Am J Parhol 1996; 149(5):1737-43. 68. Mori K . Tripartite management of unfolded proteins in the endoplasmic reticulum. Cell 2000; 101(5) :451-4. 69. Ron D . Translational control in the endoplasmic reticulum stress response. J Clin Invest 2002; 11O(10):1383-8 . 70. Lee E, Yoon C, Kim Y et aI. The double-strand RNA-dependent protein kinase PKR plays a significant role in a sustained ER stress-induced apoptosis, FEBS Len 2007; 581(22) :4325-32. 7 I. Hsu LC, Park JM, Zhang K et aI. The pro rein kinase PKR is required for macrophage apoptosis afier activation of Toll-like receptor 4. Nature 2004; 428(6980):341-5. 72. Silva AM, Whitmore M, Xu Z et aI. Protein kinase R (PKR) interacts with and activates mitogen-activared protein kinase kinase 6 (MKK6) in response to double-stranded RNA stimulation. J Bioi Chern 2004; 279(36):37670-6. 73. Turner MJ, Sowders DP, DeLay ML et aI. HLA-B27 misfolding in transgenic rats is associated with activation of the unfolded protein response. J Immunol 2005; 175(4):2438-48. 74. Unanue ER, Allen PM. The basis for the immunoregularory role of macrophages and other accessory cells. Science 1987; 236(4801) :551-7 . 75. Baeten D, Kruirhof E, Van den Bosch F er al. Immunomodularory effects of anti-tumor necrosis factor alpha therapy on synovium in spondylarthroparhy: histologic findings in eighr patients from an open-label pilot study. Arthritis Rheum 2001 ; 44(1) :186-95. 76. De Keyser F, Baeten D, Van den Bosch F et al. InAiximab in patients who have spondyloarrhropathy: clinical efficacy, safety and biological immunomodulation. Rheum Dis Clin North Am 2003; 29(3):463-79. 77. Van den Berg WB . Anri-cyrokine therapy in chronic destructive arthritis. Arthritis Res 2001; 3(1) :18-26 . 78. Repo H, jaattela M, Leirisalo-Repo M er al. Production of tumour necrosis factor and interleukin 1 by monocyres of patients with previous Yersinia arthritis, Clin Exp Immunol 1988; 72(3):410-4. 79. Miyake K. Innate recognition of lipopolysaccharide by Toll-like receptor 4-MD-2. Trends Microbiol 2004; 12(4):186-92. 80. Pentrinen MA , Holmberg CI , Sistonen L er aI. HLA-B27 modulates nuclear factor kappaB activation in human monocytic cells exposed to lipopolysaccharide. Arthritis Rheum 2002; 46(8):2172-80. 81. Goodall JC, Ellis L, Yeo GS et aI. Does HLA-B27 influence the monocyte inAammatory response to lipopolysaccharide? Rheumatology (Oxford) 2007; 46(2):232-7. 82. Colbert RA. The immunobiology of HLA-B27: variations on a theme. Curr Mol Med 2004; 4(1) :21-30.
CHAPTER
18
Animal Models ofSpondyloarthritis Joel D. Taurog*
Abstract nimal models are available for the study ofseveral different aspects ofspondyloarthritis. The models include naturally occurring spontaneous disorder s in primates and rodents, spon taneous disorders in transgenic or gene-deleted rodents and induced disorders in rodents. Areas of investigation to wh ich these models contribute include the role HLA-B27, processes of spinal and peripheral joint inflammation and calcification, immune responses to candidate antigens and the role ofTNF.
A
Introduction Despite recent major advances in diagnosis and therapyof ankylosing spondylitis (AS) and related spondyloarth ritides, the etiology and pathogenesisofthesedisorders is still relatively poorly understood. The role ofHLA-B27 has yet to be defined and the mechanism by which TNF is cent rally involved has yet to be worked out. No target antigen has been ident ified with any degree ofcertainry. The dramatic benefit ofanti -TNF therapy was first demonstrated in patients with spondyloarthritis as a direct extension ofthe experience with rheumatoid arthritis and inflammatory bowel disease, without any Significant preclinical data. Nonetheless, animal models exist from which much can be learned about etiology, pathogenesis and treatment of spondyloarthritis. This chap ter presents an overview ofth e most prominent ofthese models.
Naturally Occurring Spondyloarthritis in Primates Spontaneously occurring d isorders similar to reactive arthritis and AS have been reported to occur sporadically in gorillas and other nonhuman prim ates. The disorder recognized mo st com monly is reactive arth ritis following int estinal infect ion with Shigellaflexneri, de scribed in rhesus macaqu es and gorillas.P Asymmetric large joint arthri tiswith lameness and nonweight-bearing are the most common findings , with occasional dactylitis. Most animals recover without treatment or respond to NSAIDs or sulfasalazine . Classic ankylosing spondylitis has been reported in a gorilla, a gibbon, two baboons and two rhesus macaques. ' ? A high prevalence ofspondyloarthropathy has been inferred from ape skeletal remains by Rothschild and his colleagues.V Given the high degree of homology between HLA and primate MHC alleles, it has been of interest to investigate whether the primates susceptible to spondyloarthritis carry MHC-I alleles similar to HLA-B27. Urvarer et al 2 cloned and sequenced all ofthe MHC class I molecules from 4 gorillas affected with spondyloarthritis. The Gogo-A, B and C loci are homologs ofHLA-A, B and C. Three ofthe four affected gorillas carried an allele, Gogo -B'O101, that shows noticeable similarity to HLA-B27, particularly with respect to peptide binding repertoire. Like B27 , Gogo-B'O10 1 'Corresponding Author : Joel D. Taurog-Rheum at ic Diseases Division, Dep artm ent of Intern al Medic ine, University ofTexas So uthwe stern Medi cal Center, 53 23 Harry Hines Blvd., Dallas, TX 75390-888 4, USA. Email: [oel .taurogwutsou thwestern .edu
Molecular M echanisms ofSpondyloarthropathies, edited by C arlos Lopez-Larrea and Roberto Diaz-Pcfia. ©2009 Landes Bioscience and Springer Science-s-Business Media.
246
Molecular M echanisms ofSpondyloarthropathies
binds a predominance ofpeptides carrying arginine at position 2, with other similarities also noted in the bound peptide sequence motif. The same group also analyzed MHC -I alleles in rhesus macaques affected by reactive arthritis.' Seven of 19 sequenced Mamu-A and Marnu -B alleles carried a Cys residue at position 67 , a feature that distinguishes HLA-B27 from almost all other HLA alleles. Several of the Mamu-A alleles showed close homology to HLA- B'3901,which is associated with spondyloarthritis inJapanese and shares an overlapping peptide repertoire with HLA-B27. 8 No strong correlation was found between anyone allele and susceptibility to posrShigella reactive arthritis in the 24 animals studied, but the prevalence of the Cys67 -carrying alleles was high in both affected and unaffected animals. Whether these MHC-I alleles in gorillas and macaques that share features of HLA-B27 contribute to the pathogenesis ofspondyloarthritis in these species is still a matter ofspeculation. On the one hand, there was no a priori reason to suspect that alleles so closely homologous to HLA-B27 would be found in these species, so the fact that they were found suggests significance. Supporting this notion is the observations that (1) there are no chimpanzee alleles that show close sequence homology to HLA-B27, despite the fact that chimpanzee MHC alleles overall show a closer homology to HLA alleles than do those of gorillas or macaques? and (2) there are also apparently no reports of spondyloarthritis clinically identified in chimpanzees. It is interesting to note, however, that both HLA-B27 and the chimpanzee MHC-I allele Patr-B'03 bind an identical HIV gag peptide that is associated with slow progression ofHIV disease.'?
Proteoglycan-Induced Spondylitis and Arthritis in Mice BALB/c mice repeatedly immunized with several different proteoglycan preparations in Freund's adjuvant develop peripheral arthritis, spondylitis, or both. In the original description, immunization with proteoglycan purified from human fetal articular cartilage that was digested with chondroitinase induced inflammatory peripheral arthritis and spondyliris.U'P'Ihis digestion exposes the protein core termed aggrecan. Unlike human AS, the spondylitis in these mice was associated with destruction ofthe intervertebral dies, Both humoral and cellular immunity to the inciting antigen could be demonstrated. It was subsequently shown that the globular G 1 domain from calf nasal articular cartilage aggrecan that was partially digested with keratanase was sufficient to induce a similar phenotype.'! Immunization with cartilage matrix link protein induced peripheral arthritis but not spondylieis." On the other hand, immunization with the Gl domain of the proteoglycan core protein versican, which is found in the annulus of intervertebral discs, spinal ligaments and entheses, induces spondylitis and sacroiliitis without peripheral arthritis.'? These mice also develop uveitis.Several publications have reported immunity to aggrecan G 1 both in AS patients and patients with RA, 16 but it is not known whether this immunity is important in disease pathogenesis. Evidencefor an inflammatory response to immunization with B27-presented aggregan-derived peptides in HLA-B27 transgenic mice has been reported." Separate genetic loci for arthritis and for spondylitis appear to influence the response to aggrecan in mice.'" With regard to the proteoglycan aggrecan-induced spondylitis, both MHC and nonMHC genes control incidence and severity. 18 Inflammation occurs in the sacroiliac joints and enrheses and subsequently in intervertebral discs in a caudal to rostral progression." Unlike human AS, the intevertebral disceventually becomes destroyed. For reviews ofthe arthritis and spondylitis induced by immunization with cartilage components in mice, see references 20 and 21.
Naturally Occurring Spontaneous Models The Ank Mouse Murine progressive ankylosis is an autosomal recessive disorder due to a mutation in the ank gene on chromosome IS, first reported in 1981. Affected mice undergo deposition of hydroxyappatite in articular cartilage and develop progressive bony ankylosis of the spine and peripheral joints, yielding a phenotype bearing some resemblance to ankylosing spondylitis.P Inflammatory changes are found onlyin earlylesions.The ank gene was found to encode a 492-amino acid multipass
Animal Models ofSpondyloarthritis
247
transmembrane protein that is very highly conserved among vertebrates." The ANK (or, in humans. ANKH) protein was initially suspected to be a transporter ofinorganic pyrophosphate and this view has been supported by subsequent experiments showing that ANK transports PPi ions across membranes from intracellular to extracellular spaces." Other phenotypes related to disordered pyrophosphate handling are seen in the spontaneous mutant PC-] gene (tiptoe walking mice) and in Enppl knockout mice." The naturally occurring ank mutation (the An~ allele) produces a truncated protein lacking the C-terminal43 amino acids. More recently, targeted gene deletion oftheAnk gene has shown thatAnk- l - mice have a phenotype essentially identical to that ofthe naturally occurring mutation. 26 Moreover, conditional deletion ofAnk in joint tissue alone was sufficient to produce the same phenotype, despite the widespread expression ofAnk in other tissues." Although ank/« mice are phenotypically normal, the human disorders that have been found to be associated with mutations in ANKH on chromosome 5p15.1 show an autosomal dominant inheritance. Four ANKH mutations have been found in familial chondrocalcinosis (CCAL2), in which there is progressive deposition ofcalcium pyrophosphate dihydrate in articular cartilage." All 4 are toward the N -rerminal end ofthe protein. Six mutations in ANKH clustered in a small region toward the C -terminus have been found associated with craniometaphyseal dysplasia (CMD). a rare condition characterized by hyperostosis and sclerosis of the skull and abnormal modeling ofthe metaphyses ofthe long bones . but with normal joints. 28 Kingsleyet al have proposed that CCAL2 results from gain offunction mutations that produce elevated levels ofextracellular pyrophosphate in joints. whereas CMD may reflect dominant negative mutations associated with deficient extracellular pyrophosphate." ANKH influence on normal variations of bone size and geometry has also been suggested in one study." With regard to the role of ANKH in AS, the data to date are conflicting. A study in the UK by Timms er al examined polyrnorphisms in the 5' promoter region, exons and splice regions of of ANKH in patients with AS, controls and in multiplex AS families." They found no evidence of significant linkage. In contrast, two studies by Tsui et al of multiplex AS families in Canada and the US found significant linkage.31•32 In the first of these, linkage was found to two 5' -flanking markers. In the second, larger study, linkage was found in men with AS to a 5'-flanking SNP haplotype and in women with AS to a 3' -in tronic SNP haplotype. It thus remains to be determined whether ANKH exerts a genetic influence on the susceptibility to or course of AS. It also remains to be determined whether the disease of the ank mouse . which appears to be related to decreased synovial concentrations of inorganic pyrophosphate. has any relevance to the pathogenesis ofAS.
Ankylosing Enthesopathy Two forms ofspontaneous ankylosing enthesopathy were described in group-housed male mice in the early 1990s. Murine ankylosing enthesoparhy, also called ANKENT. was first described by investigators in Amsterdam and Prague in 1991.33 It is a spontaneous, progressive stiffening ofthe tarsal joints, primarily in group-housed male C57BL/I0 and certain B10 congenic mice . Age of onset is between 2 and 10 mo . Prevalence reaches approximately 10% in C57BL/1O and 35% in C57BL/IO.BR (H-2 k congenic) mice .r' Initially there is inflammation followed by chondrocyte proliferation (enthesius) and then ossification." The spine is not involved. The prevalence was reported to be increased in HLA-B27 transgenic mice." The germfree state is protective and specific pathogen free conditions dramatically reduce the prevalence.Y'" Spontaneous ankylosing enthesitis in group-housed 0 BAIl male mice was first described by Swedish investigators.Pv"The digits are most commonly involved,with dactylitis, PIP and MTP1M CP arthritis, toe stiffness. Ankle arthritis is also seen, but less commonly. Onset depends on the timing ofgroup housing but can occur as early as 2-3 mo ofage and prevalence ranges from SO to 100%. Histologically, the predominant lesions are fibroblast and chondrocyte proliferation with enthesitic new bone formation. There is relatively little inflammatory infiltrate.38•39 Indeed, experiments with gene-deleted mice indicate that neither a~ nor y6 T-cells are necessary" and anti-TNF treatment does not block progression of enrheseal cartilage and bone proliferation." This latter finding is consistent with the in vitro
248
Molecular Mechanisms ofSpondyloarthropathies
observation that TNF appears to inihibit chondrogenesis and stimulate bone resorption.t'v'l The entheseal chondro- and osteogenesis appears to be promoted by bone morphogenetic protein, since signaling events downstream of BMP can be detected early in the course of the disease and inhibition ofBMP inhibits the process." No direct comparison ofthe two models ofankylosing enthesitis in B 10 and DBA/ 1 mice has been reported. The 5-10 fold higher prevalence and earlier onset of pathology in DBA/1 mice makes this model more amenable to investigation. The male predominance in both models is intriguing. Castration and replacement experiments in B1O.BR mice suggest that the testis, though not testosterone per se, is essential for development ofenthesitis in this Iine ."
HLA-B27 Transgenic Mice HLA-B27 transgenic mice have been produced by many investigators. As noted above, the HLA-B27 transgene was reported to increase the prevalence of ankylosing enthesitis BlO and B10 congenic mice." However, most investigators have encountered a lack of any spontaneous B27 -specific arthropathy in mice. The major exception was a series of reports from one group of a high prevalence of peripheral arthritis in B27-transgenic mice lacking mouse ~2-microglobulin (~2m).44,45 Since HLA-B27 encodes an MHC-I heavy chain and ~2m is the invariant light-chain for MHC-I molecules, these mice express only misfolded forms ofthe B27 heavy chain. The genetic background of these mice was evidently a hybrid of the C57BLlI0 and 129 strains that was not fixed. Whether genetic or environmental factors or both account for the inability ofother investigators to reproduce these findings is not clear. It appears, however, that lack offunctional MH C- I expression per se conferred by the absence of ~2m is associated with spontaneous arthritis in mice, irrespective ofHLA-B27.46
TNF Transgenic Mice Mice transgenic for a 3'-rnodified human TNF gene that permits overexpression ofTNF develop a chronic inflammatory polyarthritis that has been studied as a model of rheumatoid arthritis." However, these mice have also been shown to develop bilateral sacroiliitis, with pannus formation, bone erosion and cartilage desrrucdon." This is directly relevant to the inflammatory component of AS which responds dramatically to anti-TNF therapy. These hTNF transgenic mice were one ofthe arthritis models in which it was shown that inhibition ofDKK-l, an inhibitor ofWnt signalling, inhibits bone erosion and promotes osteophyte formation while having little effect on inflammation.49 These same authors showed that TNF promotes D KK-l production by synovial fibroblasts in vitro. They also showed that serum DKK-llevels are dramatically suppressed in patients with AS, irrespective of disease activity. In contrast, serum DKK-llevels were elevated in patients with RA in direct relation to disease activity. This remarkable finding provides a rationale for the impression of some clinical investigators that anti-TNF treatment does not inhibit syndesmophyte progression in AS. 41
HLA-B27 Transgenic Rats Rats are more susceptible than mice to a variety ofinduced forms ofarthritis. This provided the rationale for producing HLA-B27 transgenic rats. Human ~2m (h132m)pairs with HLA-B27 better than does rodent 132m and permits better cell surface expression50 and this provided the rationale for including h132malong with HLA-B27 in producing B27 transgenic rats . It was found that rats transgenic for high transgene copy numbers ofboth B27 and h132mdevelop a spontaneous multisystem inflammatory disease that includes colitis,gastritis, peripheral arthritis, psoriasiform dermatitis, spondylitis, epididymo-orchitis and carditis.51.52 LEW; F344 and PVG inbred backgrounds are permissive for this phenotype, whereas DA is prorective.P The germfree state is protective against gastrointestinal inflammation and arthritis, but not against epididymo-orchiris or dermatitis.54 The entire phenotype requires T-cells and can be largely transferred by bone marrow transplant.55.56 CD4 T-cells specific for bacterial antigens appear to playa role in the pathogenesis of colitis."
Animal Models ofSpondyloarthritis
249
It has so far nor been possible to identify a specific arthritis-related T-cell response, either in the transgenic rats or in AS. In particular, the hypothesis that a specific HLA-B27-restricted T-cell response to a particular B27 -presented peptide contributes to spondyloarthritis still lacks direct experimental support. For this reason , a variety of other phenomena peculiar to B27 have been investigated. It was found that compared with other MHC-I molecules, in both human cell lines and transgenic rat cells B27 assembles slowly in the ER and tends to misfold and form heavy chain homodirners.V'" It has been proposed that these heavy chain homodimers may be involved in recognition by NK cell receptors60 ,6 1 and triggering the unfolded protein response (UPR).59.62
A New, Improved AS Model in HLA-B27 Transgenic Rats The hypothesis that B27-misfolding contributes to spondyloarthritis was tested in B27 transgenic rats bybreeding in additional hfl2m under the rationale that the extra hfl2m would rescue B27 heavy chains from misfolding. Additional transgenic hfl2m was indeed found to reduce B27 heavy chain misfolding and UPR triggering in rats. Unexpectedly however, F1 male offspring ofthe healthy 21-3 line (20 copies ofHLA-B27 and IS copies ofhfl2m transgenes) crossed with the healthy 283-2 line (35 copies ofhfl2m), both on the inbred LEW background, showed a high prevalence, severity and duration of arthritis and spondylitis, in the absence of any colitis or gastritis." Crosses of control lines carrying HLA-B7 or a mutant B27 gene with the 283-2 line confirmed the specificity ofthis phenotype for the (21 -3 x 283-2)FI rats. In the (21- 3 x 283-2)F I rat s,70% ofthe males develop peripheral arthritis that affects the ankles , wrists and digits (Fig. 1A,B) . At least 40% develop spondylitis in the tail , which is often associated in its later stages with sites ofvisible swelling and deformity (Fig. I C). Histologically, the spondylitis resembles AS with extensive inflammation, fibrosis and new bone formation. 63,64 Almost all of the females remain healthy-arthritis has been seen in only one female out ofhundreds. These recent data suggest that the (21-3 x 283 -2)F1 male rat is a faithful experimental animal model for AS.The only experimental perturbation is the introduction ofthe HLA-B27 gene and its associated light chain hfl2m and the disease clinically and histologically resembles AS. The onset of th e peripheral arthritis is predictable-ofthe 70% that develop arthritis, most do so between 100 and 200 days of age. The arthritis is persistent and progressive, almost involving ankle joints and often wrists and fore- and hindpaw digits (Fig. 1A,B). Spondylitis in the tail is detectable by MRI, allowing investigation to be done on early lesions. Moreover, the rat s are the offspring of a cross between two healthy parental lines , so the y can be bred in large numbers relatively efficiently. As noted above, the gut in the (21-3 x 283 -2)FI rats shows no signs of inflammation. It is of interest, however , that the males develop severe epididymo-orchitis before the onset of arthritis. Although this particular lesion is not typically a part ofhuman spondyloarthritis, it is not unheard of in this context." Orchiectomy before onset of orchitis appears to prevent the development of arthritis and spondylitis and this effect has so far not been reversed by reconstitution with testos terone (our unpublished findings). It was previousl y shown that the disease-prone high copy HLA-B27 transgenic lines had defective dendritic cell stimulation of a primary mixed leukocyte response'? and this has recently been confirmed and extended.66,67 The (21- 3 x 283-2)F I rats were found to share this apparently primary DC cell defect , unlike a large series ofhealthy control lines and crosses in which D C function was norrnal/" Whether B27 contributes to spondyloarthritis by presenting specific peptides to T vcells ha s remained an open question for decades. Recent data using (21 -3 x 283-2)FI rats homozygous for an induced null mutation ofCD8 suggest that absence ofCD8 does not prevent arthritis and spondylitis.69The CD8nullfnull rats show a profound decrease in cytolytic T -cell function. Thus , although CD8 is not absolutely required for all T -cell recognition ofMHC-I, these results suggest that the most fundamental role of B27 in spondyloarthritis involves something other than recognition of specific B27-present ed peptides by C D 8 Tvcells.
250
Molecular Mechanisms ofSpondyloarthropathies
Conclusions The spondyloarthritides are chronic inflammatory disorders involving the axial skeleton and other sites. The prototype disease. AS. has a very strong genetic basis. with one predominant susceptibility gene . HLA-B27 and contributions from several other genomic regions.The pathogenesis is not known. but TNF is clearly centrally involved . Calcification of spinal ligaments with spinal fusion ensues in severe cases. The more useful animal models of spondyloarthritis can be grouped into five categories. For models of calcification. mice with spontaneous or induced mutations in the ank, PC-I and Enppl genes are probably the most useful. For models examining immunity to candidate antigens. the various forms of proteoglycan-induced arthritis in BALBlc mice are cur rently the most useful. These are apparently also useful for genetic investigations. For investigation ofthe role ofHLA-B27 and as an overall model of the spontaneously occurring disease. the (21-3
Figure 1. Representative lesions of the spondyloarthritis in male (21-3 x 283-2)F1 transgenic rats carrying 20 transgene copies of HLA-B27 and 50 transgene copies h~2m (A) Chronic arthritis in the hind paws (B) Dactylitis in digits of the forepaw (C) Spondylitis in the tail (arrowhead). Histology corresponding to these lesions is shown in references 63 and 64.
251
Animal Models ofSpondyloarthritis
Table 1. A summary of rodent models of spondyloarthritis Model
Pros
Cons
Best Use of Model
Refs.
HLA-B 27/
Defin ite relationship to HLA-B27; male predominance
Rats lack the genetic flexibility available in mice
Studyi ng role of HLA-B27 and other aspect s of patho genesis in AS; precl inical drug testing
63,64
Proteogl ycaninduced arthritis in mice
Defin ed immune response
Requ irement for antigen preparation; Unknown relationship to AS pathogenesis
20,21 Relating specifi c immune responses to induction of spondylitis and arthri tis; geneti c contro l of response to specific antigen
TN F transgenic m ice
TN F known to be relevant to AS; Good mod el of sacroiliiti s
Unrelated to events upstream ofTN F in AS pathogenesis
Studying agents that inhibi t TN F 47,48 activi ty; studyi ng sacroili itis
Ank mutant m ice
Predic table jo int and spi ne anky losis
Basic underl ying causes prob abl y unrelated to AS
Studyi ng jo int ankylosis; studying pyrophosphate regulation of min eraliz ation
24-2 6
Ankylosing enthesitis
Spontaneous mod el in mice; involved tissue accessible
Unclear wh ether related to human spo ndy loart hrit is
Studying effect of genetic and immuno logic manipul ation s on spo ntaneous perip heral enthesitis
34,70
h~2m
transgenic rats
x 283 -2)F1 male rat transgenic for HLA-B27 and h~2m shows great promise. It may also turn out to be very useful for examining im m unity to candidate ant igens. The TNF transgenic mouse may be the most useful for the direct investigation ofthe role ofTNF in sacroiliitis. For investigation of spontaneous peripheral arthritis and enthesitis in the mouse, which allows by far the most flexibility in terms of genetic and immunologic manipulation than the rat, ankylosing enrhesitis in 0 BAil m ice may be the most useful. Although anti-TNF therapy provides dramatic benefit to many patients with AS , it is not an ideal treatment. In addition to the expense and increased risk ofserious infection, the effects of chronic long term administration are not known. Since anti-TNF therapy is not curative and since AS is a chronic disease that develops primarily in adolescents and young adults and since the course ofAS is not easily predictable, the prospect ofusing anti-TNF therapy for decades in the face of almost completely unknown long terms risks should be approached by both phys icians and patients with caution. It is hoped that the animal models described here will contribute to a better understanding of the pathogenesis ofAS and to continued improvements in therapy. A summary of the rodent models of spondyloarthritis is presented in Table 1.
252
Molecular Mechanisms ofSpondyloarthropathies
References 1. Urvater JA. McAdam SN. Loehrke JH er al. A high incidence of Shigella-induced arthritis in a primate species: major histocompatibility complex class I molecules associated with resistance and susceptibility and their relationship to HLA-B27. Immunogenetics 2000; 51 :314-325. 2. Urvarer JA. Hickman H. Dzuris JL et al. Gorillas with spondyloarrhropachies express an MHC class I molecule with only limited sequence similarity to HLA-B27 that binds peptides with arginine at P2. J Immunol 2001; 166:3334-3344. 3. Sokoloff L. Snell KC. Stewart HL. Spinal ankylosis in old Rhesus monkeys. Clin Orthop Relat Res 1968 ; 61 ;285-293. 4. Adams RF. Flinn GS. Jr. Douglas M . Ankylosing spondylitis in a nonhuman primate; a monkey tale. Arthritis Rheum 1987; 30:956-957. 5. Swezey RL. Cox C. Gonzales B. Ankylosing spondylitis in nonhuman primates: the drill and the siamang. Semin Arthritis Rheum 1991 ; 21 :170-174. 6. Rothschild BM. Woods RJ. Spondyloarthropathy as an Old World phenomenon. Semin Arthritis Rheum 1992; 21 :306-316. 7. Rothschild BM. Ruhli FJ. Comparison of arthritis characteristics in lowland Gorilla gorilla and mountain Gorilla beringei. Am J Primatol 2005; 66 :205-218. 8. Sobao Y, Tsuchiya N. Takiguchi M et al. Overlapping peptide-binding specificities of HLA-B27 and B39 : evidence for a role of peptide supermotif in the pathogenesis of spondylarthroparhies. Arthritis Rheum 1999; 42 :175-181. 9. Lawlor DA. Warren E. Taylor P er al. Gorilla class 1 major histocompatibility complex alleles: comparison to human and chimpanzee class 1. J Exp Med 1991; 174 :1491-1509. 10. Balla Jhagjhoorsingh SS. Koopman G. Mooij P er al. Conserved CTL epitopcs shared between HIV-infected human long-term survivors and chimpanzees. J Immunol 1999; 162:2308-2314. 11. GIant TT. Mikecz K. Arzoumanian A er al. Proteoglycan-induced arthritis in BALB/c mice. Clinical features and histopathology. Arthritis Rheum 1987; 30:201-212. 12. Leroux JY. Guerassimov A. Cartrnan A er al. Immunity to the Gl globular domain of the cartilage proteoglycan aggrecan can induce inflammatory erosive polyarthritis and spondylitis in BALB/c mice but immunity to G 1 is inhibited by covalently bound keratan sulfate in vitro and in vivo. J Clin Invest 1996; 97:621-632. 13. Shi S. Ciurli C. Cartrnan A er al. Experimental immunity to the G 1 domain of the proteoglycan versican induces spondylitis and sacroiliitis, of a kind seen in human spondylarthropathies. Arthritis Rheum 2003; 48:2903-2915. 14. Szabo Z. Szanto S. Vegvari A er al. Genetic control of experimental spondylarthropathy, Arthritis Rheum 2005; 52:2452-2460. IS . Zhang Y. Guerassimov A. Leroux JY et al. Induction of arthritis in BALB/c mice by cartilage link protein: involvement of distinct regions recognized by T and B-lymphocytes. Am J Parhol 1998; 153:1283-1291. 16. Zou J. Appel H. Rudwaleit M et al. Analysis of the CD8+ T-cell response to the G 1 domain of aggrecan in ankylosing spondylitis. Ann Rheum Dis 2005; 64:722-729. 17. Kuon W; Kuhne M. Busch DH et al. Identification of novel human aggrecan T-cell epitopes in HLA-B27 transgenic mice associated with spondyloarthropathy, J Immunol 2004; 173 :4859-4866. 18. Vegvari A. Szabo Z. Szanto S et al. Two major interacting chromosome loci control disease susceptibility in murine model of spondyloarthropathy, J Immunol 2005; 175 :2475-2483. 19. Bardos T. Szabo Z. Czipri M et al. A longitudinal study on an autoimmune murine model of ankylosing spondylitis. Ann Rheum Dis 2005; 64:981-987. 20 . Zhang Y. Animal models of inflammatory spinal and sacroiliac joint diseases . Rheum Dis Clin North Am 2003; 29 :631-645. 21. Adarichev VA. GIant TT. Experimental spondyloarthropathies: animal models of ankylosing spondylitis. Curr Rheumarol Rep 2006; 8:267-274. 22. Mahowald ML. Krug H . TaurogJ. Progressive ankylosis in mice. An animal model of spondylarrhropathy 1. Clinical and radiographic findings. Arthritis Rheum 1988; 31 :1390-1399. 23 . Ho AM. Johnson MD. Kingsley DM. Role of the mouse ank gene in control of tissue calcification and arthritis. Science 2000; 289:265-270. 24. Gurley KA. Reimer RJ. Kingsley DM. Biochemical and genetic analysis of ANK in arthritis and bone disease. Am J Hum Genet 2006; 79 :1017-1029. 25. Hanney D. Hessle L. Narisawa S et al. Concerted regulation of inorganic pyrophosphate and osreopontin by akp2 . enppl and ank : an integrated model of the pathogenesis of mineralization disorders. Am J Patho12004; 164:1199-1209. 26. Gurley KA . Chen H . Guenther C et al, Mineral formation in joints caused by complete or joint-specific loss of ANK function . J Bone Miner Res 2006; 21:1238-1247.
Animal Models ofSpondyloarthritis
253
27. Williams CJ, Pendleton A, Bonavita G er al. Mutations in the amino terminus of ANKH in two US families with calcium pyrophosphate dihydrate crystal deposition disease. Arthritis Rheum 2003; 48 :2627-2631. 28. Reichenberger E, Tiziani V, Watanabe S er al. Autosomal dominant craniometaphyseal dysplasia is caused by mutations in the transmembrane protein ANK. Am J Hum Genet 2001 : 68 :1321-1326. 29. Malkin I, Dahm S, Suk A er al. Association of ANKH gene polymorphisms with radiographic hand bone size and geometry in a Chuvasha population. Bone 2005 ; 36:365-373. 30. Timms AE, Zhang Y, Bradbury L er al. Investigation of the role of ANKH in ankylosing spondylitis. Arthritis Rheum 2003: 48 :2898-2902. 31. Tsui FW; Tsui HW; Cheng EY et al. Novel genetic markers in the 5'-f1anking region of ANKH are associated with ankylosing spondylitis. Arthritis Rheum 2003: 48 :791-797 . 32. Tsui HW; Inman RD, Paterson AD er al. ANKH variants associated with ankylosing spondylitis : gender differences. Arthritis Res Ther 2005 : 7:R513-525. 33. Ivanyi P. Eulderink F. van Alphen L et al. Joint disease in HLA-B27 transgenic mice. In: Lipsky P, TaurogJD, eds, HLA-B27+ Spondyloarthropathies. New York: Elsevier. 1991 :71-78 . 34. Weinreich S, Eulderink F, Capkova J et al. HLA-B27 as a relative risk factor in ankylosing enthesoparhy in transgenic mice. Hum Immunol 1995; 42:103-115 . 35. Eulderink F, Ivanyi P, Weinreich S. Histopathology of murine ankylosing enthesopathy, Parhol Res Pract 1998; 194:797-803. 36. Rehakova Z. Capkova J, Stepankova R et al. Germ-free mice do not develop ankylosing enthesopathy, a spontaneous joint disease. Hum Immunol 2000 : 61 :555-558 . 37. Nordling C. Karlsson-Parra A. Jan sson L er al. Characterization of a spontaneously occurring arthritis in male DBAIl mice. Arthritis Rheum 1992: 35:717-722. 38. Corrhay A. Hansson AS. Holmdahl R. Tvlymphocyres are not required for the spontaneous development of entheseal ossification leading to marginal ankylosis in the DBAIl mouse. Arthritis Rheum 2000 ; 43 :844-851. 39. Lories RJ, Matthys P. de Vlam K et al. Ankylosing enchesitis, dactylitis and on ychoperiostitis in male DBAIl mice: a model of psoriatic arthritis. Ann Rheum Dis 2004; 63 :595 -598 . 40. Lories RJ, Derese I, de Bari C et al. Evidence for uncoupling of inflammation and joint remodeling in a mou se model of spondylarrhrins. Arthritis Rheum 2007 : 56:489-497. 41. Scherr G. Landewe R, van der Heijde D. Tumour necrosis factor blockers and structural remodelling in ankylosing spondylitis: what is reality and what is fiction ? Ann Rheum Dis 2007 : 66:709-711. 42. Lories RJ, Derese I, Luyten FP. Modulation of bone morphogenetic protein signaling inhibits the onset and progression of ankylosing enthesitis. J Clin Invest 2005 : 115:1571-1579. 43. Capkova J. Ivanyi P. Reh akova Z. Sexual dimorphism . but not testosterone itself, is responsible for ankylosing enrhesitis of the ankle in BIO.BR (H-2k) male mice. Ann Rheum Dis 2006 : 65:130-132 . 44. Khare SO, Luthra HS , David CS. Spontaneous inflammatory arthritis in HLA-B27 transgenic mice lacking beta 2-microglobulin: a model of human spondyloarthropathies. J Exp Med 1995: 182:1153 -1158 . 45. Khare SO, Lee S. Bull MJ et al. Spontaneous inflammato ry disease in HLA-B27 transgenic mice does not require transporter of antigenic peptides. Clin Immunol 2001: 98:364-369. 46. KingsburyOJ, MearJP, Witte DP et al. Development of spontaneous arthritis in beta2-microglobulin-deficient mice without expression of HLA-B27: association with deficiency of endogenous major histocompatibility complex class I expression. Arthritis Rheum 2000 ; 43 :2290-2296. 47. Keffer J, Probert L, Cazlaris H et al. Transgenic mice expressing human tumour necrosis factor . a predictive genetic model of arthritis. Embo J 1991; 10:4025-4031. 48. Redlich K. Gertz B, Hayer S et al. Overexpression of tumor necrosis factor causes bilateral sacroiliitis . Arthritis Rheum 2004; 50:1001-1005. 49. Diarra 0 , Stolina M. Polzer K et al. Dickkopf-l is a master regulator of joint remodel ing. Nat Med 2007 ; 13:156-163. 50. Krimpenfort P. Rudenko G, Hochstenbach F et al. Crosses of two independently derived transgenic mice demonstrate functional complementation of the genes encod ing heavy (HLA-B27) and light (beta 2-microglobulin) chain s of HLA class I antigens. Embo J 1987; 6:1673-1676. 51. Hammer RE. Maika SO , Richardson JA et al. Spontaneous inflammatory disease in transgenic rat s expressing HLA-B27 and human beta 2m : an animal model of HLA-B27-associated human d isorders . Cell 1990: 63:1099-1112. 52. TaurogJD, Maika SO, Simmons WA er al. Susceptibility to inflammatory disease in HLA-B27 transgenic rat lines correlates with the level of B27 expression. J Immunol 1993: 150:4168-4178. 53. Taurog JD , Maika SO. Sarumtira N ec al. Inflammatory disease in HLA-B27 transgenic rats. Immunol Rev 1999; 169:209-223. 54. Taurog JD . Richardson JA , Croft JT et al. The germfree state prevents development of gut and joint inflammatory disease in HLA-B27 transgenic rats. J Exp Med 1994; 180:2359-2364.
254
Molecular Mechanisms ofSpondyloarthropathies
55. Breban M. Hammer RE, Richardson JA et al. Transfer of the intlammarory disease of HLA-B27 tran sgenic rats by bone marrow engrafiment. J Exp Med 1993 , 178:1607 -1616. 56. Breban M , Fernandez-Sueiro JL, Richardson JA et al. T-cell s, but not thymic exposure to HLA-B27, are requir ed for the inAammatory disease of HLA-B27 transgenic rats. J Immunol 1996, 156:794-803 . 57. Qian BF, Tonkonogy SL, Hoentjen F et al. Dysregulated luminal bacterial antigen-specificT-cell responses and antigen-presenting cell function in HLA-B27 transgenic rats with chronic colitis. Immunology 2005 , 116:112 -121. 58. Dangoria NS, DeLay ML, Kingsbury OJ et al. HLA-B27 misfolding is associated with aberrant intermolecular disulfide bond formation (dimerization) in the endoplasmic reticulum. J Bioi Chern 2002, 277:23459-23468. 59. Tran TM, Satumtira N, Dorris ML er al. HLA-B27 in tran sgenic rats forms disulfide-linked heavy chain oligomers and mulrimer s that bind to the chaperone BiP. J Immunol 2004, 172:5110-5119. 60. Kollnberger 5, Bird LA, Roddis M et al. HLA-B27 heavy chain homodimers are expressed in HLA-B27 transgenic rodent models of spondyloarthritis and are ligands for paired Ig-like receptors. J Immunol 2004; 173:1699-1710. 61. Kollnberger 5, Chan A, Sun MY et al. Interaction of HLA-B27 homodimers with KIR3DLl and KIR3DL2, unlike HLA-B27 heterotrimers, is independent of the sequence of bound peptide . Eur J Immunol 2007; 37:1313-1322. 62. Turner MJ, Delay ML, Bai 5 et al. HLA-B27 up-regulation causes accumulation of misfolded heavy chains and correlates with the magnitude of the unfolded protein response in transgenic rats: Implications for the pathogenesis of spondylarrhritls-Iike disease. Arthritis Rheum 2007 ; 56:215-223 . 63. Tran TM, Dorris ML, Saturntira N et al. Additional human beta2-microglobulin curbs HLA-B27 misfolding and promotes arthritis and spondylitis without colitis in male HLA-B27-transgenic rats. Arthritis Rheum 2006; 54:1317-1327 . 64. Taurog JD. Ankylosing spondylitis : new improved treatment, new improved models. Drug Discovery Today 2006; 3:27-31. 65. Stagg AJ, Breban M, Hammer RE et al. Defective dendritic cell (DC) function in a HLA-B27 transgenic rat model of spondyloarthropathy (SpA). Adv Exp Med Bioi 1995; 378 :557-559. 66. Hacquard Bouder C, Falgarone G. Bosquet A et al. Defective costimulatory function is a striking feature of ant igen-presenting cells in an HLA-B27 -transgenic rat model of spondylarthropathy, Arthritis Rheum 2004; 50:1624-1635. 67. Hacquard Bouder C, Chimenti MS. Giquel B et al. Alteration of antigen-independent immunologic synapse formation between dendritic cells from HLA-B27-transgenic rats and CD4+ T-cells: selective impairment of costimulatory molecule engagement by mature HLA-B27 . Arthritis Rheum 2007; 56:1478-1489. 68. Fert I. Glatigny S. Poulain C et al. Dendritic cells (DCs) functional defect correlates with spondyloarthritis (SpA) phenotype in HLA-B27/human 132-microglobulin (hI32m) transgenic rat lines. Arthritis Rheum 2007; 56:5247 (abstract). 69. TaurogJD. Dorris ML, Sarumtira N. Neither deletion of CD8 nor alteration of the B27 peptide repertoire prevents spondyloarthritis in HLA-B27/hb2m rats. Arthritis Rheum 2007; 56:5530 -531 (abstract). 70. Lories RJ. Animal models of spondyloarthritis. Curr Opin Rheurnatol 2006 ; 18:342-346.
CHAPTER
19
T-Cell Responses Against Viral and Self-Epitopes and HLA-B27 Subtypes Differentially Associated with Ankylosing Spondylitis Maria T. Fiorillo and Rosa Sorrentino"
Abstract LA- BZ7family comprehends some allelesstrongly associated with Ankylosing Spondylitis (AS) and some others that are not. A comparative analysis at genetic and functional level is likely to give a clue to the understanding ofdisease pathogenesis. Here, we summarize our recent studies on the functional differences between B'Z70S, the most frequent and worldwide AS-associated allele and B'Z709, an allele found in Sardinia where it accounts for ZO% of all BZ7 alleles and where it is not associated with AS. The two BZ7 alleles are distinguished by a single amino acid change, located in the peptide binding groove, that correlates with relevant structural and functional differences in presenting viral and self peptides to T-cells. In particular, B'Z709 individuals lack in their T-cell repertoire of CD8+ T -cells specific for a self-epitope (pVIPR) derived from the vasoactive intestinal peptide Type 1 receptor (VPACl). This peptide shares extensive homology with a viral epirope, pLMPZ, derived from EBV,toward which, both B'Z70S and B'Z709 individuals mount a vigorous CTL response. A likely explanation to this finding, also supported by crystallographic data, is that the auto reactivity present in the disease-prone B'Z70S individuals can be unleashed by a molecular mimicry mechanism which does not occur in the B'Z709 indiv iduals. The possible implications of the T-cell cross-reactivity between pLMPZ, pVIPR and other related peptides in AS pathogenesis are discussed.
H
Introduction For over 30 yeats, the search ofan explanation for the strong association between Ankylosing Spondylitis (AS) and a HLA class I antigen, HLA-BZ7,1.2has revealed some unusual features of the biology and ofthe antigen presenting properties ofthe HLA-BZ7 molecules which, although not univocally, could contribute to the pathogenesis ofAS. Indeed, the disclosure ofeach ofthese unexpected features has brought along the claim to have caught the guilty. However. as an Arabian phoenix, the mystery of this association was coming back unsolved. All these failures have led to what seems the most reasonable conclusion: more than one mechanism, involving the HLA-BZ7. could cooperate to the pathogenesis of the disease.v' One ofthe first hypotheses that, from time to time, is revived by new evidence, considers the molecular mimicry of HLA-BZ7 self-ligands by microbial antigens as a reasonable explanation ' Corresponding Author: Rosa Sorrentino-Department of Cell Biology and Development, "Sapienza", Universita di Rorna, Vi a dei Sardi 70, 00 185 Roma, Italy. Email : rosa.sorrentin
[email protected]
Molecula r Mechanisms ofSpondyloarthropathies, edited by Carlos Lopez-Larrea and Roberto Diaz-Pena. ©Z009 Landes Bioscience and Springer Science+Business Media.
256
Molecular Mechanisms ofSpondylosrtbropatbies
for the appearance ofthe disease in genetically predisposed individuals. According to this theory, traditionally reported as "arthrirogenic" peptide hypothesis, cytotoxic CD8+ T-cell responses against one/more cross-reactive self and foreign antigens presented by the HLA-B27 alleles associated with AS, would be the primary pathogenic event in the autoimmune process. In short, an infection with a microbial agent sharing homology in its HLA-B27-restricted epitopes with sequences embedded in selfproteins, but usually ignored by the immune system, would unleash a cross-reactive autoimmune response leading to chronic inflammation and tissue damage.' In the last years, much effort has focused on elucidation ofmechanisms by which T-cells respond to cross-reactive epitopes and on correlating pathogen-specific cellular responses to the development oforgan-specific autoimmune diseases." Yet, this issue results more difficult to address in the case ofAS since the aetiology ofthe disease is unknown and gram-negative bacteria such as Chlamidia, Shigella, Salmonella and Yersinia that have a clear role in the reactive arthritis (ReA), a related inflammatory rheumatic disease, also B27-associated, do not appear to playa primary role in AS.7 The picture is complicated by the fact that more than one infectious agent could be involved in priming the auto-reactive immune response and unleash the disease. Therefore, it is not surprising that common pathogens, in combination with specific environmental and genetic factors , would set the conditions to precipitate the autoimmune disease. A further apparent complication derives from the fact that HLA-B27 is not a single allele but a wide family of41 alleles, some ofwhich have been rarely found in patients, although present in reasonable numbers in somepopulations. Besides,some allelesapparently associate with the disease in some populations but not in others. v' One ofthese alleles is the B*2709, originally described in an Italian healthy individual and subsequently found at significant frequency in Sardinia, where it accounts for 20% of the overall B27 subtypes ."!" HLA-B*2709 differs from B*2705 only at residue 116 (His to Asp) which is part of the floor of the pocket F and takes contact with the carboxy-terminal anchor of the bound peptides. But why two alleles, apparently so similar, are differently associated with AS? The reason is not yet understood but several speculations, not mutually exclusive, could be put forward. First, the explanation could be found at genetic level: they are harboured by distinct HLA haplotypes meaning that a different set ofsusceptibility or protective genes could be coinherited with B*2705 and not with B*2709 subtype orviceversa.l' 'F Second, the solution could be searched at structural! functional level: polymorphism 116 confers different properties in terms ofpeptide binding repertoire and could influence the antigenic functions oftwo B27 alleles.P:" An interesting parallel to this regard can be found in AIDS where two groups ofHLA-B35 alleles, B35-PY and B35-Px, differing at amino acid position 116, bind peptides with different C-terminus and the first but not the latter, confers resistance the AIDS. 16 In this framework, the comparative analysis ofthese closely related B27 subtypes, differently associated with AS, has revealed itselfas an useful tool to approach the molecular basis of the disease. Herein, we will summarize recent studies based on the analysisofthe antigen presenting features ofB*2705 versus B*2709molecules that have lead to the identification ofpreviously unrecognized selfepitope/s that share homology with an EBV-derived epitope and unleash auto -reactive CD8+ T-cells in patients with AS. The occurrence ofT-cell cross-reactivity and the structural molecular mimicry turn out as features strictly dependent on the B27 allele; this allows to depict a possible molecular mechanism underlying the autoimmune process in AS.
Searching for the Holy Grail: From B27-Restricted Viral Antigens to Self-Ligands Eliciting Autoreactive Cytotoxic T-Cell Responses in Patients with Ankylosing Spondylitis Supposing that HLA-B27 molecules are implicated in AS pathogenesis because ofthe presentation ofone/more "disease-inducing peptides", these ligands do not need to be selectively presented by the HLA-B27 subtypes associatedwith AS. Alternatively, they could be general ligands that bind both disease-associated and nonassociared B27 subtypes but with different affinity or structural conformations. In such a case,one could speculate that central tolerance can be influenced by the
T-Cell ResponsesAgainst Viral and Se!fEpitopes and HLA-B27
257
affinity/stability of the complex that these "disease-inducing peprides" form when bound by the different B27 alleles. How to give substance to this hypothesis? HLA-B27 seems to work very efficiently in inducing protective immunity against viral infec tions e.g., Influenza, AIDS and herpetic diseases compared to other HLA alleles.17•18This effect seems to be mediated by productive CD8+ T-cell responses directed against immunodominant epitopes. For the HLA-B'2705 in particular, several immunodominant epitopes derived from viruses mentioned above, have been identified and extensively studied." Apart from these major drivers ofviral specific T-cell reactivity, suboptimal epitopes have also been described which are likelyto evoke lessefficaciousT-cell responseswhich, for this samereason, could be not counteracted by feedback regulatory mechanisms. As a consequence, these CD8T-cells, could be maintained lifelongwithout detrimental effectsthrough stimulation by cross-reactiveself-epitopes. However, in some circumstances influenced by the genetic background and/or by infections caused by the same or a different pathogen entailing the same T-cell repertoire, these responses can be amplified and contribute to the pathology. Following this hypothesis, some years ago, we focused on a peptide, pLMP2 (residue 236-244 RRRWRRLTV) derived from the latent membrane protein 2ofEBV, that has been already described as a specific B'2704 epitope." pLMP2 possessed some peculiarities making it an excellent antigen to explore possible differences in antigen presenting features induced by the polymorphism at 116. First, pLMP2 showed a comparable and moderate binding affinity towards the two B27 molecules and, more importantly, it was able to evoke specific T-cell responses in both B'270S and B'2709 positive individuals. Second, pLMP2 came from a virus that persists in most individuals as a lifelong asymptomatic infection thus inducing long-lasting memory CTL responses . Therefore, it would be conceivable to reactivate pLMP2-specific CTLs from peripheral blood ofB27 and EBVseropositive subjects, either patients with AS or controls, by in vitro stimulation with autologous EBV-Blymphoblastoid cells.Third, it might be ofinterest to analyseB27-restricred CTL responses against a widespread virus that several studies have suggested as possibly involved in autoimmune, inflammatory diseases, among which rheumatic diseases." Studying pLMP2-responsive CTL lines generated from B'2705 patients with AS and from B'2709 healthy controls, we realized that pLMP2 and its alanine analogs did not induce the same pattern ofreactivity when associated with B*270S or 8*2709 allele." In particular, the activation ofa CTL line from a B*2709 subject showed a clear-cut difference when the peptide was presented in the context of B*270S or B*2709 molecules, displaying a high dependency on the peptide C-terminal moiety. These functional data gave a first strong indication that pLMP2 could assume two distinct conformations when associated with either B27 subtype. Later on, this hypothesis has been confirmed by X-ray crystallography showing drastically deviating binding modes of the viral peptide pLMP2 in association with B'270S (noncanonical conformation) or B*2709 subtype (canonical conformation) (Ziegler, et al, this book)." What did make this peptide unusual? At first glance, the stretch of arginines from pI to p6 represented the major peculiarity of the peptide. Therefore, by using the blastp program, we screened the public protein databases (www.ncbLnlm.nih.gov/) searching for peptide sequences related to pLMP2 and fulfilling the B27 binding constraints. Several peptides satisfying these requirements were found and seven of the best matches were chosen for further investigations. Three ofthem were human sequences derived from the vasoactive intestinal peptide Type 1 receptor (pVIPR, residues 400-408; RRKWRRWHL); from the glucagon receptor (pGR, residues 412-420; RRRWHRWRL) and from the voltage-dependent calcium channel al subunit (pCAC, residues 513-521; SRRWRRWNR) (Table 1). Other four sequences came from microbial proteins belonging to Escherichia coli, human cytomegalovirus, Salmonella typhimurium andMycobacterium smegmatis (Table 1). All but one (RQWRRLPQV from S. typhimurium) had an Arg in P2 and a hydrophobic (Leu or Val) or a positively charged (Arg) residue at the C-terminus (P8 or P9) . All ofthem shared the RXWXR motifwith the first Arg acting as primary anchor (P2) and were able to stabilize B'2705 and B'2709 molecules on T2 transfectants. Notably, the pLMP2-stimulated CTL line (MPl) from a patient with AS was responsive to pVIPR and pargA in both B'2705 and B*2709 context of presentation (Fig. 1). On the contrary, only pargA from Salmonella
2S8
Molecular Mechanisms ofSpondyloarthropathies
Table 1. Peptides sharing sequence homology with pLMP2 and possesing the 827 binding motif Sequence
Organism
Source
pVIPR
RRKWRRWHL
Homo sapiens
Vasoactiv e intestinal peptid e receptor Type 1 (VPAC1) (400-408)
pGR
RRRWHRWRL
Homo sapiens
Glucagon receptor (41 2-420)
pCAC
SRRWRRWNR
Hom o sapiens
Voltage-dependent calc ium channel a1 subunit (513-521)
pyaiP
RRWRRWIV
Esch erichia coli
Hypoth eical prot ein yaiP (246-253)
pHXLF4
RRWLRLLV
Hum an cytomegalovirus
HXLF4 prote in precur sor (2-9)
pOrgA
RQWRRLPQV
Salm onella typh imu rium
O xygen-regulated invasion protein (Org A) (77-85) Probable arabynosil transferase C (678-6 86)
pPATC
QRRWQRLLV
Mycoba cterium smegmatis
PLMP2
RRRWRRLTV
Epstein-Barr virus Latent memberane protein 2 (LMP2) (236-244)
Conventional sequence homol ogy search in the publi c protein databases (w w w.ncbLnim.nih.gov/) using t he blastp program. The selected pepti des, coming from micro bial and human proteins, share the RXWXR mot if and have mostly a hydrophobi c residue at carbo xy-te rmi nus.
typhimurium induced the effector functions of another CTL (Ter 5) originated from a B'2709 healthy subject. The occurrence of CTL cross-reactivity among a viral (pLMP2), a self (pVIPR) and a bacterial (pOrgA) peptide in a subject affected by AS, prompted us to undertake further none
TerS
MPl
MP49
pLMP2 pVIPR
~
~
pargA pyaiP
P
~
pHXLF4 pGR
~
~
~
pPATC
0
10
20
30
40
50
60
•
0
%
10 20 30 40
OT28*2709 .T28*2705 Peptides [70....]
~
pCAC
Tarvet cells:
50
60
70
0
10 20 30 40 50 60 70 80
Specific Lysis
Figure 1. CD 8+ T-c ell c ross-reac t ivity among self and nonself peptides sharing the RxWxR motif . These results are from a standard 4-hours 51-chrom ium release assay in which CTL lines obtained by pLMP2 stimulation are used as eff ec to r cells. Ter5 line or iginates from a B*2709 healthy ind iv idual wh ile MP1 and MP49 come from a B* 2705 patient with AS. T2B*2705 and T2B*2 709 stable transfectants, pre-incubated with the peptides show n on the left, have been used as target cells . Ter5 CTL line is cross-r eactive w ith pOrgA whe reas MP1 CTL line is fully ac tivated by pOrgA and pVIPR besides pLMP2.
cn
T-Cell Responses Against Viral and Self-Epitopes andHLA-B27
259
functional studies. Accordingly, pVIPR was used to stimulate CD8+ T-cells from B"270S positive patients with AS and healthy controls either B"2709 or B"270S.Such stimulation elicited relevant specific CTL responses turning out more pronounced in patients than in B"270S positive healthy subjects and virtually absent in the B"2709-positive donors." Since than, this same autoreactive T-cell response has been found to characterize many other B"270S or B"2702 positive patients with AS, giving strength to the concept that this is not a sporadic event.24 Noteworthy, the response to pVIPR persists over time since the same patients with AS maintain a reproducible pattern of self-reactivity during years (Fiorillo and Sorrentino, unpublished data). Interestingly, about 2S% ofCTL lines raised against pVIPR from B"270S positive AS patients and controls are cross-reactive with pLMP2 and both peptides are also recognized in association with the B"2709 rnolecules.P'" As for pLMP2-driven CTL linesderived from donors either B"270S or B"2709 positive, only 16% display cross-reactivity with pVIPR at high peptide concentration (70 J.lM) independently from being presented by B"270S or B"2709 molecules.F-" Therefore, the different proportion and strength of cross-reactivity could be interpret as consequence of fine-tuning regulatory mechanisms that limit the activation ofpathogen-specific T-cells by weaker, autologous ligands. 25The presence ofdouble reactive CTL lines, especially in patients, would hint a relationship among EBV infection, the expansion of a pool of pLMP2IpVIPR cross-reactive T-cells and the occurrence of autoimmunity. It is hard at the state of art to correlate these events to AS pathogenesis, all the more that, in patients, a good proportion ofpVIPR-specific responses is independent from pLMP2. There is no doubt, however, that the same autoreactive response is totally absent in B"2709 positive individuals although the B"2709 molecules can present the self peptide to T-cells and, usually, the specific TCR do not distinguish between pVIPRlB"270S and pVIPR/B "2709 complexes (Fiorillo 2000 and unpublished results). Conversely, the majority ofthe B"2709 positive healthy donors respond vigorously to pLMP2 (88%) and only few ofthese T-cells recognize pVIPR in a suboptimal way, thus indicating that in these subjects there is no reactivity driven by pVIPR itselfand the weak response to pVIPR isa side effect ofthe immunity towards EBY. How to explain the lack ofpVIPR self-reactivity in the B"2709 genetic context? B27 stabilization assays performed on T2 transfectants have clearly shown no impairment in the binding ofpVIPR to the B"2709 molecules. Moreover, CTL lines and T-cell clones, although derived from B"270S positive subjects, are triggered more efficiently by peptide/B"2709 than by peptide/B"270S com plexes.22.23These dat a were interpreted as a better capacity ofB"2709 molecules to present pVIPR perhaps due to a higher stability of pVIPR/B*2709 complex. Crystal structures ofpVIPR bound 2426 to the two B27 subt ypes have supplied fundamental pieces to compose the puzzle. . Remarkable is the finding that pVIPR, besides to bind both B27 subtypes in a conventional conformation, could adopt a second noncanonical binding mode in association with B"270S molecules." In the conventional conformation the arginine at PS is solvent exposed, while in the non conventional one the same residue points to the Boor ofbinding groove forming a salt bridge with the aspartic acid 116 present in the B"270S but not in the B"2709 allele. To the light ofthis new information, it could be speculated a cause-effect relationship between the conformational properties ofpVIPR associated to the B27 subtypes and the occurrence ofa pVIPR-driven T-cell auto-reactivity. In the B*270S groove , the pVIPR could change dynamically its conformation hindering high-affinity interaction with the TCR. On the opposite, a static conformation of the same peptide when associated with the B*2709 could favour an optimal recognition by TCR. Therefore, the differences in pVIPR specific T-cell reactivity might be seen as consequence of a thymic negative selection which is leaky in the B*270S context and efficient in the B"2709 background.
Extension ofT-Cell Cross-Reactivity to a 'Third Player: pGR Another issue that deserve further attention is degeneracy versus specificity of recognition by pVIPR-responsive CTL lines . In a previous study, numerous pVIPR specific T-cell lines from various B"270S subjects , were analysed for cross-reactivity with other peptides sharing sequence homology with pVIPR and pLMP2 (Table 1).27Apart from the recognition of pLMP2 by few CTLs, pGR (RRRWHRWRL) originated from the glucagon receptor was the only other peptide
260
Molecular Mechan isms of Spondy loarthropathies
able to induce CTL reactivity.Such a recepror is a G-coupled protein belonging, as the vasoactive intestinal peptide Type 1 recepror, ro the family of secretin receptors; this explains the very high homology between pVIPR and pGR peptides." In analogy with pVIPR, this self-peptide exhibited antigenic properties, evoking a prompt cytotoxic 'T-cell response in B*270S and B*2702 patients with AS. 27 The analysisofthe cross-recognition showed different behaviours ofthe pGR-responsive T-cell lines: the majority ofthem were specificfor pGR only, while others could be also triggered by both pVIPR and pLMP2 or one ofthe two. Interestingly, CTL specificity versus cross-recognition appeared as subject-specific with some individuals possessing cross-reactive CTLs and some others having a more restricted pattern ofrecognition. In any case, pGRsimilarly to pVIPR, could evoke a restricted T-cell reactivity apparently independent from both the anti-EBV and the pVIPR-induced responses. Work is in progress to verifywhether the self-response to pGR is a feature characterizing B*270S patients with AS or extends to B*2709 healthy subjects . X-ray crystallographic structure ofpGR/B*270S complex has been determined, showing the peptide in a dual conformation and, in both binding modes, the middle part ofthe peptide bends towards the floor of peptide binding groove." Intriguingly, the structural comparison of pGR (one of the two conformations), pVIPR (nonconvenrional conformation) and pLMP2 in the B*270S binding groove, reveals a high degree ofsimilarity that is particularly pronounced in the N-terminal part of the peptides.
T-Cell Cross-Reactivity between Selfand Non-Self: The TCR Point ofView The identification ofT-cells targeting selfpeptides only, or cross-reactivewith a sequence-related viral peptide, offers the opportunity to gain critical insights into the nature, the binding features and the topology of TCRs possibly implicated in autoimmune responses . This is of great interest since, from the few studies available so far, it has emerged that auto-reactive TCRs contact the peptide/MHC complex in unconventional ways that, lowering the binding affinity, permit auto-reactive T-ceHs to escape negative seleceion." In our case, no structure oftrimolecular TCRI peptide/HLA-B27 complexes has been determined so far. Nevertheless, some speculations can be put forward on the basisofTCR alpha and beta chain sequence analysis ofmore than forty T-cell clones originated from pVIPR- and pGR-stimulated CTLs ofdifferent B*270Spatients with AS.27 Among these clones, some were highly selective, responding to pVIPR or pGR only, while some others showed different profiles of peptide cross-reactivity among pVIPR, pGR and pLMP2. Notably, we identified a peculiar amino acid motifshared by the TCRs exhibiting a cross-reactive pattern but not by those that resulted mono-reactive to pVIPR or pGR (Table 2). This common sequence included the (D/N)RDDKIIFG motif within the CDR3 alpha loop and extended to the same] alpha segment Oa 9.4). Given the similar topology ofthe TCRs on the peptide/MHC complex shown by the different crystallographic studies, 30 it can be assumed that this CDR3 alpha loop shared by the cross-reactive TCRs interacts with the N-terminal part of the peptides where
Table 2. TCR-a chain sequences of cross-reactive T-cell clones from patients with AS
cn Line MPVPAC7"
AB4b ABSb
No. of Clones
17 2 4
AV
14-C AYS 7-C A 7-C A
l S-C A
Peptide Specificity DRDDKIIFG-9.4 VNRDDKIIFG -9 .4 VNRDDKIIFG-9 .4 ASPSSGGSYIPTFG-15 .3 C
pVIPR/pLMP2 pVIPR/pGR pVIPR/pGR /pLMP2
"Reported in refs. 22 and 27; "Reported in ref. 27; -clones from CTL line ABS have a dual TCR. The underlined CDR3a motif and the Ja 15.3 are shared with T-cell clones monospecific for pVIPR and derived from a different patient with AS (see ref. 27).
T-Cell Responses Against Viral and Self-Epitopes and HLA-B27
261
the similarity among the pLMP2, pVIPR, pGRlB'2705 structures is the highest. Moreover, it is conceivable to predict that the stretch ofpositively charged amino acid residues characterizing the N -terminal halfofthese peptides could supply a firm dockingfor Asp residues well represented in the CDR3 alpha region ofthese TCRs. An intriguing feature of several T -cell clones , either specific for pVIPR or even largely cross-reactive, was the expression of two TCR a-chains both sharing sequence homologies in their CDR3 region with other T-cell clones derived from the same or a different patient with AS. This suggested that both TCRs, having different a -chains coupled to the same 13 chain, could be involved in the recognition of these peptides. About one third ofT-cells has been reported to possess a dual TCR31 and, this condition has been for a long time suspected to be detrimental since the unselected TCR could potentially induce autoimmunity. More recently, a positive role for these "not legitimate" TCRs in expanding the repertoire for nonself antigens has been also envisaged." Conversely, no relevant similarities have been found among the TCRJ3 chains ofthese clones. irrespective of responding to one or more peptides. Nevertheless, some TCRs showed a valuable homology in the J3-chain CDR3 region with TCRs reported by others and derived from CD8+ T-cells ofHLA-B27-positive patients with ReA or AS, whose specificity is unfortunately unknown." Curiously, we have found that an unusually short CDR3J3loop characterized some TCRs turning out to be mono-specific for pVIPR or pGR. This feature makes uncertain the participation of this part of the TCR in contacting the surface ofthe peptide/B*2705 complex. In conclus ion. these data permit to depict a scenario in which the contribution of the CDR3 loops of a - and J3-chain to the cross-reactivity of these T-cell clones is substantially different: the TCR a-chain would playa pivotal role in its occurrence while the TCR J3-chain would tuning its extent.
Conclusions HLA-B27 family ofalleles, perhaps because isone ofthe most studied HLA class1allele. seems to possess so many diverse features able to justify its association with AS. The "arthrirogenic" peptide hypothesis puts forward many years ago, has undergone through alternate cycles of interest but it has never been dismissed. Our studies have focused on a model ofcross-reactivity based on self, viral and bacterial peptides that share sequence and structural similarities and trigger T-cell responses in a disease-susceptible context thus giving strong support to this hypothesis. More insights are expected from the TCR studies both at sequence and structural level. Nevertheless, it must be considered that. ifT-cell autoreactiviry is relevant, this cannot be the only explanation for AS. Probably, more than one aspect of HLA-B27 biology is implicated in AS pathogenesis. Finally,genetic studies have clearly shown that HLA-B27 is only one ofthe many players in a very complex game . The future challenge is to understand how they relate each other.
Acknowledgements The authors wish to thank Prof. A. Mathieu and Dr. A. Cauli for participating in this study and Volkswagen-Stiftung for financial support (Grant 1/79990 to R.S. and Grant 1/82735 to M.T.F).
References I. Brewerton DA , Hart FD. Nicholls A et al. Ankylosing spondyliris and HLA-B27. Lancet 1973 ;
1:904-907. 2. Schlosstein L. Terasaki PI, Bluestone Ret al. High association of an HL-A antigen. W27. with ankylosing spondylitis. N Engl] Med 1973; 288:704-706. 3. Khan MA, Mathieu A, Sorrentino R er al. The pathogenetic role of HLA-B27 and its subtypes. Autoimmun Rev 2007; 6:183-189 . 4. Lopez de Castro ]A. HLA-B27 and the pathogenesis of spondyloarthropathies. lmmunol Lett 2007 ; 108:27-33. 5. Benjamin R. Parham P. Guilt y by association: HLA-B27 and ankylosing spondylitis. lmmunol Today 1990, 11:137-142.
262
Molecular Mechanisms ofSpondyloarthropathies
6. Benoist C, Mathis D. Autoimmunity provoked by infection: How good is the case for T-cell epitope mimicry? Nature Immunol 2001: 2:797-801. 7. Hyrich KL, Inman RD. Infectious agents in chronic rheumatic diseases. Curr Opin Rheumatol 2001: 13:300-304. 8. Del Porto P, D'Amato M, Fiorillo MT et al. Identification of a novel HLA-B27 subtype by restriction analysis of a cytotoxic yf:> T-cell clone. J Immunol 1994; 153:3093-3099. 9. D'Arnato M, Fiorillo MT, Carcassi C et al. Relevance of residue 116 of HLA-B27 in determining susceptibility to ankylosing spondylitis. Eur J Immunol1995: 25:3199-3201. 10. Paladini F, Taccari E, Fiorillo MT et al. Distribution ofHLA-B27 subtypes in sardinia and continental italy and their association with spondylarthropathies. Arthritis Rheum 2005 : 52:3319-3321. I I. Fiorillo MT, Cauli A, Carcassi C et al. Two distinctive HLA haplorypes harbor the B27 alleles negatively or positively associated with ankylosing spondylitis in sardinia: implications for disease pathogenesis . Arthritis Rheum 2003; 48 :1385-1389. 12. Cascino I, Paladini F, Belfiore F er al. Identification of previously unrecognized predisposing factors for ankylosing spondylitis from analysis of HLA-B27 extended haplotypes in sardinia. Arthritis Rheum 2007; 56:2640-2651. 13. Fiorillo MT, Meadows L, D'Amato M et al. Susceptibility to ankylosing spondylitis correlates with the C-terminal residue of peptides presented by various HLA-B27 subtypes. Eur J Immunol 1997: 27 :368-373. 14. Ramos M, Paradela A, Vazquez M er al. Differential association of HLA-B'2705 and B'2709 to ankylosing spondylitis correlates with limited peptide subsets but not with altered cell surface stability. J Bioi Chern 2002; 277 :28749-28756. IS. Fiorillo MT, Greco G, Maragno M et al. The naturally occurring polymorphism Asp116-His116 that differentiates the ankylosing spondylitis associated HLA-B'2705 from the non associated HLA-B'2709 subtype influences peptide-specific CD8 T-cells recognirion. Eur J Immunol 1998: 28:2508-2516. 16. Gao X, Nelson GW; Karacki P er al. Effect of a single amino acid change in MHC class I molecules on the rate of progression to AIDS. N Engl J Med 2001: 344:1668-1675. 17. Den Uyl D, van der Horst-Bruinsma IE, van Agcmael M. Progression of HIV to AIDS: a protective role for HLA-B27? AIDS Rev 2004; 6:89-96. 18. Lopez de Castro JA. HLA-B27-portraying immunodominant viral epitopes. Eur J Immunol 2005; 35:336-340. 19. Stewart-Jones GB, di Gleria K. Kollnberger S et al. Crystal structure and KIR3DLI recognition of three immunodominant viral peptides cornplexed to HLA-B·2705. Eur J Immunol 2005; 35:341-351. 20. Brooks JM , Murray RJ, Thomas WA et al. Different HLA-B27 subtypes present the same immunodorninant Epstein-Barr virus peptide. J Exp Med 1993: 178:879-887. 21. Scorer E, Peyrat MA, Saulquin X er al. Frequent enrichment for CD8 T-cells reactive against common herpes viruses in chronic inAammatory lesions: Towards a reassessment of the physiopathological significance of T-cell clonal expansions found in autoimmune inAammatory processes. Eur J Immunol 1999: 29 :973-985. 22. Fiorillo MT, Ruckert C, Hulsmeyer M et al. Allele-dependent similarity between viral and self-peptide presentation by HLA-B27 Subtypes . J BioI Chern 2005; 280 :2962-2971. 23. Fiorillo MT. Maragno M, Butler R et al. CD8+ T-cell autoreactivity to an HLA-B27-restricted self-epitope correlates with ankylosing spondylitis. J Clin Invest 2000; 106:47-53. 24. Hulsmeyer M, Fiorillo MT, Betrosini F er al. Dual. HLA-B27 subtype-dependent conformation of a self-peptide. J Exp Med 2004: 199:271-281. 25. Grossman Z, Paul WE . Auroreactiviry, dynamic tuning and selectivity. Curr Opin Immunol 2001; 13:687-698. 26. Wucherpfennig KW. Presentation of a self-peptide in two distinct conformations by a disease-associated HLA-B27 subtype. J Exp Med 2004 : 199:151-154. 27. Ruckert C, Fiorillo MT. Loll B et al. Conformational dimorphism ofself-peprides and molecular mimicry in a disease-associated HLA-B27 subtype . J BioI Chern 2006; 281 :2306-2316. 28. Laburthe M, Couvineau A. Gaudin P et al. Receptors for VIP, PACAP, secretin, GRF, glucagon , GLP-l and other members of their new family of G protein-linked receptors: structure-function relationship with special reference to the human VIP-l receptor. Ann N Y Acad Sci 1996: 805:94-109. 29. Hahn M. Nicholson MJ, Pyrdol J et al. Unconventional topology of self peptide-major histocompatibility complex binding by a human autoimmune T-cell receptor. Nat Immunol 2005; 6:490-496. 30. Rudolph MG. Stanfield RL, Wilson IA. How TCRs bind MHCs. peptides and coreceptors. Annu Rev Immunol 2006; 24:419-466. 3 I. Padovan E, Casorati G, Dellabona P et al. Expression of two T-cell receptor a chains: Dual receptor T-cells . Science 1993: 262:422-424. 32. He X, Janeway CA, Levine M et al. Dual receptor T-cells extend the immune repertoire for foreign antigens. Nat Immuno! 2002; 3:127-134.
CHAPTER
20
Dendritic Cell: T-Cell Interactions in Spondyloarthritis J.S. Hill Gaston: Lorna B.Jarvis, Libin Zhang and Jane C. Goodall Abstract
T
he discovery ofthe association between spondyloarthritis (SpA) and HLA-B27 inevitably turned the spotlight on Tdymphocyres as the cellswhich recognize peptide antigens within the bindinggroove ofthe HLA-B27 molecule and then carry out effector functions. These include cytolysis, cytokine and chemokine production and activation ofother effector cells, such as those which could destroy joints or drive new bone formation. In this view the T -cell assumed the role of"director" ofthe immune response and therefore, in inflammatory diseases such as SpA , of immuno-pathology. The important research questions under this paradigm were the identity of the peptides recognized by T-c ells in disease, including whether they were derived from self proteins or from micro-organisms, the influence ofHLA-B27 in selecting antigenic peptides for recognition by T-cells, the T -cell receptors used in recognition and the effector programmes which the T-cells initiated. Whilst these questions continue to be explored-many have not yet been answered-attention has shifted to a new "master regulator" ofthe immune response, namely the dendritic cell and the possibility that the genetic influences which contribute to susceptibility to SpA do so at the level of the dendritic cell (DC).
Introduction There have been a number of reasons for thi s change in thinking. Firstly, as a general principle in immunology, it ha s been noted that the DC both supplies the context in which antigen is recognized and directs the differentiation of the T-cell which recognizes antigen, so as to bring about appropriate effector functions.':" Information on context is detected by pattern recognition receptors (PRR), both cell surface and intracellular, which allow DC to distinguish between a protein antigen "on its own", likely to be a self antigen and therefore innocuous and an antigen accompanied by, for example, bacterial cell wall components or viral double-stranded RNA and therefore likely to bederived from a pathogen and a potential threat to the host. DC are particularly well equipped to detect these bacterial products using the Toll-like receptors (TLR)3 and other bacterial sensors such as the NOD-like receptors (NLR) which include NOD 1 and NOD2 together with the NALP components of the inflammasome.v' The same system is also able to recognize certain selfmolecules, such as heat shock proteins and urate crystals,6.7which are used as signals of tissue damage. DC respond to activation of these recognition systems by altering their expression ofcostimulator molecules on their surface and secretion ofthose cytokines which have the ability to alter T-cell growth and differentiation. These include cytokines more usually associated with T-cells such as IFNy and IL-4, but more importantly the more recently discovered "D C-specific" cyrokines such as IL-12, IL-23 and IL-27 .8 'Corresponding Author: ).S. Hill Gaston-Department of Rheumatology, University of Cambridge, Box 157, Level 5, Addenbrooke's Hospital, Cambridge CB2 2QQ, UK. Ema il:
[email protected]. ac.uk
Molecular M echanisms ofSpondyloarthropathies, edited by Carlos Lopez-Larrea and Roberto Diaz-Pena. ©2009 Landes Bioscience and Springer Science+Business Media.
264
Molecular Me chanisms 4Spondyloarthropathies
In the context of SpA genetic findings have emphasised the significance these processes . For the SpA-associated conditions, psoriasis and inflammatory bowel disease, associations have been shown between disease and polymorphisms in the bacterial sensor NOD2 (Crohn's disease") , the DC product IL-12 (psorias is") and the receptor for another member ofthe IL-12 family of DC cytokines, IL-23 (psoriasis, Crohn's disease and ulcerative colitis 10.I 1). Polymorphisms ofthe IL-23R gene have also been shown to be associated with AS in recent whole genome studies.'? This chapter will explore two aspects ofDC function in relation to SpA. Firstly, we will discuss the phenotype ofthe CD8+ cellswhich emerge from cocultures with autologous DC and how this differs in AS patients and healthy subjects. Secondly, we present data which indicate that the Output ofDC depends not only on external signals received from PRRs, but is strongly influenced by the physiologic state ofthe cell, particularly the endoplasmic reticulum and whether it is experiencing stress related to production of large amounts of incorrectly folded proteins. In view of the demonstration that the B27 heavy chain is a molecule which can evoke ER stress through inefficient folding,13 these data assume significance in relation to our understanding of the way in which HLA-B27, together with other genes, gives rise to the inflammation characteristic ofSpA.
A Novel Subset ofCD8+ T Cells Since DC are potent antigen presenting cellsrequired for the initiation ofCD8+ T-cell responses, they can be used to expand CD8+ T-cells purified from peripheral blood (PB). In order to identify CD8+ T-cells specific for an intracellular bacterium associated with reactive arthritis, Chlamydia trachomatis (CT), we used CT-infected DC to stimulate highly purified CD8+ T-cells from CT-infected patients. These cocultures were successful in generating CT-specific CD8+ T-cells, as we have previously reported elsewhere." However, in the course ofcontrol experiments in which the same CD8+ T-cells were cocultured for two weeks with uninfected autologous DC, we noted an unusually high proportion ofIL-4 producing cells amongst the CD8+ cells. Furthermore, such IL-4+ cells were much more frequent in cultures from AS patients as compared to healthy subjects. Cloning these CD8+ T -cell lines, using a nonspecific protocol with mitogen, IL-2 and irradiated allogeneic PBMC, revealed a novel CD8+ subsen" the clones were autoreactive, restricted by HLA-class I antigens (including, but not limited to, HLA-B27) and produced "Tc2" cytokines i.e., IL-4, IL-5 and IL-B. They did not produce IFNy, but did express TGFfJ and could produce TNFa. They differed greatly from conventional cytotoxic CD8+ T-cells in that they lacked expression ofperforin or cytotoxic activity; instead they showed several characteristics normally associated with CD4+ regulatory T-cells,including constitutive high level expression ofCD25 and CTLA-4, together with intracellular expression of Foxp3. In keeping with this regulatory phenotype they showed potent, contact-dependent regulatory activity, inhibiting both proliferation and IFNy production by CD4+ T -cells.The ready isolation ofHLA-class I restricted CD8+ regulatory T -cell clones (CD8+ Treg) from AS patients raised the possibility that they were expanded in a disease with an HLA class I association perhaps in response to active inflammation. Having isolated and characterized this T -cell subset we have gone on to investigate the role of DC in both their in vitro expansion and their regulatory function.
The Role ofDendritic Cells in the Expansion ofa CDS+Regulatory T-Cell Subset We have shown that an interaction between CD8+ T-cells and DC is crucial for the expansion ofthe IL-4 producing CD8+ cells within T-cell lines. Coculture ofpurified CD8+ T-cells from AS patients with other types ofantigen presenting cell, such as EBV-transformed lymphoblastoid cells (LCL), did not promote the expansion ofIL-4+ CD8+ Tvcells, whether or not the LCL expressed HLA-B27; instead cultures were dominated by IFNy-producing cells (Fig. 1). Likewise activation and expansion ofCD8+ T-cells using anti-CD3/anti-CD28 coated magnetic beads (Dynal) resulted in the expansion of CD8+ T-cells producing predominantly IFNy. In these experiments cell contact between the CD8+ T-cells and DC was required for expansion ofIL-4+CD8+ T -cells (Fig. 2) . These results imply that a surface expressed molecule(s) specific to DC playsa critical role
42.24
I
fll~
10"
AS37 DC
10"
15.99
6.2 1
10'
i
~
te ...
'S!
-.
o
";
r " ••
..
. ~; ::" .
· '"1.: , ,
.~ ;:~.,:
....,
2.77
.t-
I
" .:
: :
'",.
•
. ....~..:',: " :.
~ ',
:-:.
.'::.
:
BLCL(B 27+)
. 10'
52.00
2.36
io'
I
~
~
~
~
:L.74
0.79
u.
IFNy
fll~
1
t. .
-:tt~" : :1"1.::
I ,2 . ~:,:~.::~ ';.;.
"
28.95 10'
I
gT,_::-::-:--...,.,----------"
Anti-CD3/CD 28
Figure 1. Effect of coculture w ith DC on outgrowth of IL-4- CD8 - T-cells Purified CD8 - T-cells were cocultured w ith either allogeneic DC from an AS patient, (left pane\), 827- allogeneic LCL (central panel) or beads coated with anti-CD3 and anti-CD28 antibodies (right panel). Following two weeks' culture cells were stimulated w ith PMA and ionomycin before analysis of cytokine production by flow cytometry. Staining for IL-4 is on the y-axis and for IFNy on the x-axis.
'S!
"
o
IL-4
~
N
e
", '
S·
;:"
s... s.. ...
~ ~ ~
i1 '"~ s: E! ii'
~
:;-
~ :::::
';-l
c~
",'
~
~,
!}
Molecular Mechanisms ofSpondyloartbropethies
266
A
B 0...
...
0
26.69
I
2.52
4.36
8.90
...
b
...
b
E
C>
'Qj
J:
~ u.
~
.....J
10' 10' Fl1 -Height
... 0...
b
FL1-Height
IFNy Figure 2. Cell contact between DC and T-cells is requ ired for outgrowth of IL-4-CD8- T-cells. Intracellular expression of IL-4 (y-axis) or IFNy (x-axis) by CD8 - T-cells cocultured for two weeks in contact with DC (A) or separated from the DC by a transwell (6) . Cells were stimulated with PMA and ionomycin before analysis of cytokine production by flow cytometry. Cells which are strongly positive for IFNy and apparently weakly so for IL-4 probably represent a compensation artefact. in the expansion of this Tvcell subset; work to identify this molecule is continuing. Whilst DC were indispensable for expansion ofIL-4+ CD8+ Tvcells, DC from healthy donors or AS patients were equally efficient for this purpose and expression of HLA-B27 was not required. Thus the difference in the frequency ofIL-4+CD8+ T-cells in cultures from AS patients and healthy donors cannot be accounted for by differences in DC derived from AS patients and healthy subjects. Expression ofHLA-B27 does appear to influence expansion of these cells, since recent data have shown higher proportions ofIL-4+ CD8+ T-cells in Tvcellslines derived from HLA-B27+ healthy subjects than from HLA -B27-negative subjects (Zhang et al, ms in preparation). The extent to which these healthy individuals may have possessed other genes associated with AS 12 is not known and requires further investigation. Although it is possiblethat DC derived signalsmight allow differentiation ofIL-4+CD8+ T-cells as well as their expansion in vitro , this does not seem to be the case. Thus cultures only contained significant numbers ofIL-4+CD8+ T-cellswhen CD45RO+ CD8+ T-cells (i.e.,memory phenotype) were cocultured with DC and not when CD8+ T-cells were depleted of CD45RO+ cells to leave a predominantly naive population. This implies that the IL-4+CD8+ T-cell subset has already differentiated in vivo; whether DC are involved in this initial differentiation is not known. In support of the differentiation ofIL-4+ CD8+ T-cells in vivo, flow cytometric analysis of PBM C stimulated immediately exvivowith PMA and ionomycin revealed the presence ofincreased numbers ofIL-4+ CD8+ T-cells in AS patients as compared to healthy donors (Fig. 3); interestingly increased numbers have also been seen in later experiments using PBMC from RA patients (Baek H et al, In press). Nevertheless, the relationships between the IL-4+CD8+ T-cells detected in ex vivo staining ofPBMC, those found in T-cell lines expanded by coculture with DC and the IL-4+CD8+ Treg obtained by cloning the T -cell lines, require further discussion . It is hypothesized that at least some of the IL-4+CD8+T-cells detected in PBMC are precursors of the IL-4+CD8+ T-cells seen in the T-cell lines and of the CD8+ Treg clones. In support ofthis, higher numbers of IL-4+CD8+ T-cells were seen in T-cell lines obtained from AS patients and CD8+ Treg clones were more readily cloned from the lines obtained from AS patients than those from healthy donors. However, recent experiment s in which Tvcell lines were stained for intracellular expression ofboth IL-4 and Foxp3 have shown that whilst the lines expanded using DC (and not those obtained using
267
Dendr iti c Cell: T-Cell Interactions in Spondy/oarthritis
A u r:..
t
<.
oC ' Q ~ o FSc
B
I
:r
I
,-
---
CD3-PerCP
Isotype controls
I
7. 1%!
IFN-y-FITC
0
~
m Q.
3.5
.5
3.0
til ~
P = 0.0007
0 0
2.5
0
o 2.0
.... +
IX)
1.5
C 1.0
0
+
0.5
-r -I 0.0 ~ 0
0
0
~ 00 0
AS
00 0
0
He
Figure 3. IL-4 production by PBMC stimulated dir ectl y ex vivo A . Four co lour intracellular cy tokine staining of PBMC from a representative AS patient fo ll owi ng stimulati on with PMA and calc ium ionophore in the presence of monensin. CDB+ CD3 + cells were gated and the percentage IL-4 and IFN-y positi ve cells within the CDB+T-cell population is shown . Quadrants wer e set by stain ing with isotype control antibodies. B. Propo rtions of IL-4 + CDB+ T-cells in PBMC of AS patients as compared to healthy individuals. The plot shows the numbers of IL-4 +CD3 + CDB + T-cells in each group and the medians (horizo ntal bar). The difference wa s significa nt, ( p < 0.001 ; Mann Whitn ey U Test).
aniCD3/CD28 beads) contain both Foxp3+and IL-4+ CD8+ T-cells, many of the IL-4+ T-cells within these T -cell lines are Foxp3-negative. In contrast, when these lines were cloned, the majority (-66%) of the clones obtained express both IL-4 and Foxp3+ and have suppressive activity. This suggests that the CD8+ Treg are preferentially selected by the cloning procedure, po ssibly because the other IL-4+T -cells represent effector cells with limited cloning potential. The "clonabiliry" of CD8+ Treg is in striking contrast to CD4+CD2S+ Tregwhich are difficult to expand in vitro and cannot therefore be cloned (to produce an analyzable T-cell done requires a rapid rate ofin vitro proliferation-a minimum of 10 divisions over two weeks). The relationship between the IL-4+CD8+ cellsseen in PBMC following stimulation with PMA and ionomycin immediately ex vivo and those obtained aslines and clone s is less clear. Foxp3+CD8+ cells are generally thought to be very rare (certainly in murine studies where the y represent < 1% Foxp3 expressingcells'<") and thus farit has not proved technicall y possible to isolate and clone the
Molecular M echanisms ofSpondyloarthropathies
268
IL-4+cells obtained by stimulation, e.g.,by using "cytokine capture" techniques" to identify them rather than the intracellular staining which destroys them. This would be required to demonstrate directly that these IL-4+CD8+ T-cells are indeed precursors ofIL-4+ CD8+ Treg.
The Role ofDendritic Cells in the Function ofa CDS+ Regulatory T-Cell Subset In addition to thei r important role in expanding IL-4+ CD8+ Tvcells, we have also shown that the mechanism by which AS derived CD8+ Treg function to suppress immune responses in vitro critically involves interactions with DC. In these studies, CD8+ Treg clones were able to suppress both proliferation and IFNy production by CD4+ T-cells stimulated with alloantigen s in a mixed lymphocyte reaction (Fig. 4, black and red bar columns 1-6). Note that in this experiment IFNy produced by CD4+ Tcells stimulated with allogeneic HLA-B44+ DC was completely suppressed by the additional of a B44-restricted CD8+ Treg clone, whereas this clone had no effect on IFNy produced in response to allogeneic HLA-B44-negative DC. This is due to the fact that the clone is not activated by the B44-negative DC in the culture. Pre-activation ofthe clone with mitogen allowed the clone to inhibit the CD4+ Tcell response to allogenic DC which it did not itself recognize. Figure 4 also demonstrates that CD8+ Treg were unable to suppress the low amounts of IFNy produced by CD4+ Tcells which were activated by sub-optimal amounts magnetic beads coated with antibodies to CD3 and CD28 ("antiCD3/CD28 beads") and in the absence ofDC. When DC were added to CD4+ T-cells cultured with antiCD3/CD28 beads, there wasa very substantial enhancement in the amounts of IFNy produced by the CD4+ T-cells . However, when a CD8+ Treg clone was added to such cultures (I.e., in the presence of DC) it was able to suppress IFNy
1000 900 800 700 ~ 600 Z LL 500 400 E Q. 300 200 100
en
o
Figure 4. CD8+ T reg mediate the ir effect through interaction with DC. IFNy production by CD4+ responders cells stimulated with HLA-B44+ or B44-ve DC, or beads coated with anti-CD3 and ant i-CD28 antibodies("dynal "), or both. The effects of the addition of an HLA-B44-restricted CD8+ Treg clone we re also measured. nd = not detected. A color version of this image is available at www.landesbioscience.com/curie.
269
Dendritic Cell: T-Cell Interactions in Spondy/oarthritis
production back to baseline levels (Fig. 4, light blue bar columns 7-10). In these cultures, since the CD8+ Treg clone could be activated by the antiCD3/CD28 beads, its ability to suppress the effect of the DC did not depend on its ability to recognize the DC, so that suppression was as effective when either HLA-B44+ or HLAB44-ve DC were used (light blue and dark blue bars respectively [columns 9-12]). In these experiments contact between the CD8+ Treg and the CD4+ Tvcell/DC cocultures was required for suppression ofIFNy secretion. since their separation by transwells abolished suppression. Together these experiments suggest that CD8+Treg bring about their effect on the output from CD4+ T-cells by interaction with DC, even though this interaction does not require recognition ofclass I HLA antigens on the DC. The interaction is likely to involve CTLA-4. expressed at high levels on the CDS+ Treg (Fig. SA) and CDSO on the DC, since suppression was partially blocked with monoclonal antibodies specific for CTLA-4 and CDSO (Fig. 5B). In addition conjugate formation between CD4 cells and DC was reduced in the presence ofCDS+ Treg (data not shown). Thus the interaction between the CDS+ Treg and the DC alters the latter in such a way as to block its ability to activate the CD4 T-cell. The exact mechanism remains to be elucidated; a key feature of the CDS+ Treg clones is
A
resting CDB Treg i-c
resting CDS • Treg surface
Act COB Treg surface
Act CD4 surface
CD4
B
4500. 4000 ~ 3500 3000 2500 .E 2000 ~ 1500 1000 500
a: =
O.
~
~
s ;.;
~
....
~
~
....
,f
~
....
~
'"
~
....
s'"
ec!l
CJ
~~
;.;
~
~
~
~
.... e
~~
~
~
~
~
J.....!!'
....
Figure 5. Involvement of CTLA-4 and COBO in COB Treg function. A) Intracellular and surface expression of CTLA-4 on a resting and activated COB' Treg clone and on activated C04' T-cells . Note the substantial surface expression of CTLA-4 on activation of the COB' Treg clone as compared to the modest levels on the activated C04+ T-cells (equivalent to the levels expressed by the resting COB' Treg clone) . B) Effects of antibodies to CTLA-4 or COBO and their isotype controls, on the ability of a COB' Treg clone to inhibit IFNy production by PBMC in response to PHA .
270
Molecular Mechanisms ofSpondyloarthropathies
their ability to quickly up-regulate the expression ofvery large amounts ofCTLA-4 on their cell surface as soon as they are activated either by recognition ofthe relevant class I HLA allele by the T-cell receptor, or following mitogen stimulation. This mobilising ofintracellular CTLA-4 to the cell surface may be under the control ofthe TRAT1 gene which CD8+ Treg express at high levels as compared to conventional CD4+ and CD8+ T-cells (Jarvis LB, unpublished data). This gene encodes TRIM (T-cell receptor interacting molecule), a cell surface molecule recently shown to bind strongly to CTLA-4 in the trans-Golgi network and to be promote its cell surface expression ." Lastly,whilst CD8+Treg clearlyexpress the transcription factor most closelyassociated with regulatory function, Foxp3, it is not known to what extent Foxp3 is essential for their regulatory function, given that certain activated human 'T-cellsare known to express Foxp3 without necessarily acquiring regulatory function." Preliminary observations suggest lower levelsofFoxp3 transcripts in CD8+ Treg than in "classical" CD4+CD25+ Treg.
What Is the Significance ofCDS+ Treg in the Pathogenesis ofSpA? At this stage it is only possible to speculate on whether CD8+ Treg have a role in vivo which relates to the pathogenesis of AS or other forms of SpA. Only transfer of such cells, or their depletion, could demonstrate unequivocally such a role and this will only really be possible in the context of an animal model such as the HLA-B27 transgenic rat. In this respect it is worth noting that a CD8+ Tregsubset has been described in the rat and has been shown to be capable of inhibiting graft-versus-host disease." It may therefore be worthwhile to seek a role for these cells in this model; interestingly recent experiments have shown that depletion ofCD8+ cells does not abolish arthritis in the B27 transgenic rat model (Taurog], data presented at the ACR meeting, Boston, November 2007), but whether it influences disease severity has not yet been conclusively demonstrated. The generation CD8+Treg might represent an attempt to modulate chronic inflammation even though this attempt fails. This kind ofsituation is not unusual-synovial macrophages make more ofthe anti-inflammatory cytokine IL-I RA than normal macrophages, even though this is insufficient to counteract their production ofIL-I and concentrations of soluble TNF receptors are elevated in active rheumatoid arthritis. Our recent suggestion that CD8+ Treg are also increased in frequency in RA would agree with the notion that they represent a general anti-inflammatory mechanism. It is also possible to postulate pathogenic effects ofthese cells. For instance, they produce TNFa, a cytokine whose over-production clearly plays a part in SpA pathogenesis, given the remarkable clinical benefits of TNFa inhibitor drugs in these conditions. Secondly, they inhibit IFNy production: it has been suggested that IFNy production is deficient in SpA and this would favour the production ofIL-I7 producing cells. The IL-I7 producing subset ofT-cells (ThI7 cells) plays a central role in several forms ofautoimmune disease (at least in murine models) and IFNy directly inhibits their differentiation. IL-23 is a crucial cytokine in the growth and function ofThI7 cells, so the association between polymorphisms in the IL-23 receptor and AS implies an important role for this subset in AS pathogenesis. Lastly, CD8+ Treg make IL-4 and IL-I3, cytokines which affect both DC growth and maturation and are able to modulate autoimmune disease.
The Influence ofthe Unfolded Protein Response on DC Function SpA is strongly associated with the human leukocyte antigen HLA-B27. The presence of HLA-B27 is not sufficient to induce disease in humans but requires other genetic or environmental factors. There is increasing evidence that infection by microorganisms may be a critical factor that can modify an individual's susceptibility to arthritis. Thi s is particularly evident in reactive arthritis where prior infection of the gastrointestinal tract or urogenital system is associated with disease. The B27 transgenic rat model ofSpA shows an important link between bacteria and arthritis since disease does not occur in germ free conditions." Although a clear role for bacterial infection in AS has not been proven, reactive arthritis patients can progress to a disease that has the same clinical features as AS. Given these findings, we suggest that HLA-B27 may confer susceptibility to SpA, by
Dendritic Cell: T-Cell Interactions in Spondyloarthritis
2 71
its influence on the quality ofthe immune response to components ofmicroorganisms.The finding that HLA-B27 induces an unfolded protein response (UPR) in cells from HLA-B27 transgenic ratsZ3 suggests that this effect of B27 expression may contribute to altered cellular physiology in cellsthat orchestrate the immune response and provide disease susceptibility. We will now describe our init ial explor ation of this hypothesis.
The Effects ofB27Expression in a ModelAntigen Presenting Cell To investigate how the presence ofHLA-B27 may influence the immune response to bacterial components, we used the myelo-monoeytic cell line U937, stably transfected with HLA-B27 , HLA-A2 or a control vecror.r' Cells were differentiated with PMA and stimulated with the TLR4 agonist, LPS. Supernatants were analysed for TNFa secretion and the gene expression profiles of unstimulated and LPS stimulated cells were analysed by microarray analysis. Although the U937 cells exhibited a stereotypical response to LPS we did not detect differences in TNFa secretion between cells transfected with B27 as compared to those transfected with HLA-A2 or a control vector. Microarray analysis ofgene expression revealed only two genes that were differentially expressed between HLA-B27 and control cellsstimulated with LPS. OAS2 was the only gene whose differential expression was confirmed by quantitative PCR. HLA-A2 expressing cells exhibited similar OAS2 gene expression to HLA-B27 U937 cells,suggesting this effect was due to presence ofadditional HLA class I molecule and not an effect specific for HLA-B27 (Fig. 6A). Microarray analysismay not detect subtle differencesin gene transcription; therefore we analysed IL-23 mRNA expression by the different cell lines using quantitative PCR but again no significant differences in gene expression were detected (Fig. 6B). Although we detected an increa sein the unfolded protein responsefollowing stimulation with LPS we were unable to show differences between HLA-A2 and HLA-B27 transfectants. These findings do not exclude the possibilit y that HLA-B27 can exhibit unusual behaviour in human cells, since B27 expression in transgeni c rats at low copy number is not sufficient to provide susceptibility to diseaseor induce an unfolded protein response, It is likely that the level of expression ofHLA-B27 in the U93 7 cells studied here did not exceed a critical threshold to reveal abnormal effects. Future studies will reveal whether high level expression can ind uce physiological effects in response to bacterial products recognized by TLR's.
The Unfolded Protein Response Influences Cytokine Production by DC Although an unfolded protein response (UPR) has been demonstrated in HLA-B27 transgenic rat myeloid cells," it is not known how this phenomenon could modify immune response s to micro-organi sms. Although we could nor induce a UPR by expression ofHLA-B27 in U937 cells, we extended our studies to investigate how a UPR induced biochemically could modulate immune responses in monocyte-derived dendritic cells. Disruption ofER homeostasis by inhibition ofthe sarcoendoplasmic reticulum Ca z. ATPase was achieved by the addition ofthapsigargin (TP). The reduction in the concentration of calcium within the ER prevents the optimal function of ER chaperones that are critical for protein folding resulting in the UPR. Therefore we were able to test whether activation ofthe UPR had the potential to modify pro-inflammatory immune responses to pattern recognition receptors that recognise molecules derived from infective agents. Monocytes isolated from the peripheral blood of healthy individuals were differentiated into DC by incubation with IL-4 and GM-CSF and stimulated with the UPR inducing agent, thapsigargin, together with agonists recognized by PRR's. Activation ofthe UPR in combination with TLR 2,3,4 and 8-specific agonists very substantially enhanced IL-23 secretion (Fig. 7) . In contrast production ofthe related cytokine IL-I2 was inhibited rather than enhanced by the same additions ofTP. Activation of the UPR still exerted an IL-23 specific effect even when optimal TLR signals for both IL-23 and IL-I2 secretion were provided con cominantly. These results suggest that TP results in expression offactors, or co-operates with factors, that are inherently IL-23 specific in their action. There is increasing evidence that the NF-KB pathway is crucial for the expression ofIL-23. Mise-Ornata et alzs have shown that the NF-kB family members RelA (p6S) and c-Rel playa critical role in the expression ofthe IL-23 p19 subunit. Furthermore pathways that negatively regulate RelA severely affect IL-23 P19 expression without affecting
2 72
A
Molecular M echanisms ofSpondyloarthropathies
A2+LP~ 1-------t ~~~
+LP~~ C C+LPS ljl
~~~~~~~~~~~==:~_.. . ,
o
0.5
1.5
2
2.5
3
OAS2 mRNA (fold upregulation from unstimulated control U937)
B
Vector
HLA-B27 HLA-A2
o
50
100
150
200
250
300
350
400
Il-23 mRNA (fold upregulation from unst imulated U937) Figure 6. HLA-B27 doe s not mod ify the response of U937 to LPS. RNA isolated fro m U93 7-B27, U937-A2 and U937-C stimulated with or without LPS for 4 hours was analysed for the expression of OAS2 (A) and IL-23 (B) using quantitative PCR. These data are presented as fol d chan ge from U937-C. These data are obtained from four independent experiments and are expressed as mean .±.SD. • is P ~ 0.05.
IL-12p70 expression ." Whether stimulating a UPR enhances the activation and expression of the se factor s remains to be determined. Even when a potent stimulus for IL-23 secretion was provided by peptidyoglycan, UPR activation was able to furth er enhance thi s response (Fig. 7 ). We and others" have shown that bacteria-derived- peptidoglycan. which stimulates TLR2 and NOD2 pathways, preferentially stimulates IL-23 pl9 and not IL-12 p70 expression. It would appe ar that this specificity is dependent on TLR2 signals since NOD2 pathways can synergise with TLR4 agonises to enhance IL-12p70 secretion." It has previously been shown IL-23 is critically regulated by PI3 kinase and its downstream pathways via TLR2 29. It is not known how TLR2/NOD2 pathways provide specificity for IL-23 activation but identification ofthe specific compon ents th at achieve this may enable us to understand how T P int eracts with thi s pathway. Evidence suggests that stimulatio n of the PRR, De ctin I . result s in IL -23 secretion in DC and th at signals from dectin one use a pathway which is distin ct from th at used by T LR's.30 To investigate if thi s pathway co uld inte grate ER stress signals to further enhance IL-23 secretion. mD C were stimulated with curdlan, a specific agonist of the dectin 1 receptor pathway. Although the addition ofcurdlan resulted in high levels ofI L-23 secretio n, no further inc rease ofIL-23 protein was detected when the cells were costimulated with TP (Fig. 7).
Dendritic Cell: T-Cell Interactions in Spondyloarthritis
273
Control TLR3
Poly I:C
TLR4
LPS
TLR8
CL-097
o Control
-+TP
LPS/CL-097
TLR4/8
TLR3/8 Poly I:C/CL-097 Deetin 1 TLR2/NOD2
Curdlan
PGN
o
4
2
8
6
10
IL-23 ngjml
Control Polyl:C
o control
-+TP
LPS CL-097 LPS/CL-097 Poly I:C/CL-097 LPS/CL-097 Poly I:C/CL-097
-
o
2
4
6
8
10
12
14
IL-12ngjml Figure 7. Stimulation of a UPR enhances IL-23 but not IL-12 secretion . Monocyte derived immature dendritic cells were stimulated with agonists for TLR3 (Poly I:C 30lAg/ml), TLR4 (ultrapure LPS Iug/ml), TLR8 (CL-097, SlAg/ml), TLR2/NOD2 (Peptldyoglycan.tuug/ml) and dectin 1 (Curdlan 1OOlAg/ml) The ERstress induci ng agent; thapsigargin (O.02SIAM) was added to cell cultures in the presence or absence of PRR stimulation and incubated for 48 hours and analysed by ELISA for IL-12p70 and IL-23 secretion . These data are expressed as the mean ± SD and are representative of four independent experiments.
The Unfolded Protein Response Influences Cell Surface Molecule Expression by DC We have also been able to show that UPR activation in immature DC increased their expression of costimulatory molecules and markers associated with DC maturation (Fig. 8). This may reflect a requirement for less complex stimulatory signals to upregulate markers of DC maturation as compared to those required to control pro-inflammatory cytokine secretion. CD83 and HLA class I molecules are already preformed within the cells and retained within intracellular compartments; thus TP induced signals must act on pathways that altering the trafficking ofthese molecules from intracellular compartments to the cell surface.
CD83
~
103
:>
103
+LPS
10'
10'
10'
~
::
CD80-
+TP~
:>
103
10 J
10'
10' 10'
+LPS
10'
t4l,A pIC'
10J
.. . . . .
HLA class I
10°
o.
N
....
01 10°
::
,....
10'
10'
Figure 8. Stimulation of a UPR induces maturation of immature monocyte derived dendritic cells. Monocyte derived immature dendritic cells were stimulated with either thapsigargin (0.05 I!M) or a crude prearation of LPS for 24 hours. Cells were removed and analysed for the cell surface expression of CD83, CD86 and HLA class I molecules.
~
~
N
~
....
S. ;S -sI:> S. ~.
s...I:>
~ ~ ~ ~
~.
~
... ~ "~
ii;'
l:!
~
~
Dendritic Cell: T-Cell Interactions in Spondyloarthritis
275
The Significance ofthe Interaction between the UPR and the Innate Immune Response Why would ER stress signals be integrated into the innate immune response mediated by TLRs? McKenzie et al2? suggest that both that microbial products and self (PGE2, ATP) signals for danger or injury can regulate the balance between IL-12 and IL-23 production. In addition it has been reported that viral infection can induce a UPR. The UPR may be usefully integrated as a danger signal which in combination with PRR activation, indicates a bacterial insult on the cell. How would the integration ofER stress signals into the IL-23 secretion pathway benefit the host? Again, McKenzie et al2? suggest that IL-23 is an important driver ofearly immune responses to pathogens. In particular, IL-23 may facilitate the rapid induction ofIL-17 secretion from tissue resident T-cells and allowing the rapid recruitment of neutrophils and thus an effective early immune response to pathogen invasion.
Conclusions DC have a pivotal role in the expansion of a novel subset of CD8+ T-cells which have both an unusual cytokine profile-Th2/Tc2 cytokines combined with TNFa and TGF~-together with a regulatory phenotype and function. IL-4+ CD8+ T regulatory subset in AS patients. The mechanism oftheir regulatory action again involvesan interaction with DC, most likely through CTLA-4 and its principal ligand , CD80. The response of DC to TLR ligands can be strongly influenced by the physiologic state of the cell's endoplasmic reticulum. Induction of an unfolded protein response pharmacologically increases DC cytokine output, with IL-23 being particularly favoured under these conditions. Expression of HLA-B27 per se was not found to either induce an unfolded protein response or to affect the response to LPS, as assessed by gene array experiments. However these experiments were carried out by transfecting HLA-B27 and other control HLA alleles into the macrophage/ monocyte cell line, U937. This may not accurately mirror the situation in vivo when HLA-B27+ DC, expressing other gene variants associated with susceptibility to AS, encounter combinations ofligands for TLR's and NLR's.
References I. Mellman I, Steinman RM . Dendritic cells: Specialized and regulated antigen proc essing mach ines. Cell 200 1; 106(3):255-8. 2. Liu Y]. Dendritic cell subsets and lineages and thei r functions in innate and adaptive immunit y. Cell 2001 ; 106(3):259-62. 3 Takeda K, Kaisho T. Akira S. Toll-like receptor s. Annu Rev Immunol 2003 ; 21 :335-76. 4. Fritz]H. Ferrero RL, Philpott D] et al. Nod-like proteins in immunit y, inflammation and disease. Nature Immunology 2006; 7(12) :1250 -7. 5. Petrill i V. Dosterr C , Muruve DA et al. The inflammasome : a danger sensing complex triggering innate immunity. Curr Opin Immunol 200 7; 19(6):615-22 . 6. Tsan ME Gao B. Endogenous ligands of Toll-like receptors.] Leukoc Bioi 2004; 76(3):514-9. 7. Martinon F, Petrilli V. Mayor A et al. Gout-associated uric acid crystals activate the NALP3 inRamma some. Nature 2006; 440(7081) :237-41. 8. Hunter CA. New IL-12-family members : IL-23 and IL-27, cytokines with divergent functions. Nat Rev Immunol 2005 ; 5(7):521-31. 9. McGovern DP, van Heel DA, Ahmad T et al. NOD2 (CARDI5), the first susceptibility gene for Crohns disease. Gut 2001 ; 49( 6 ):752-4. 10. Cargill M , Schrodi S], Chang M er al. A large-scale genetic association study confirms ILI2B and leads to the identification of IL23R as psoriasis-risk genes. Am] Hum Genet 2007; 80(2 ):273-90. 1 I. Duerr RH, Taylor KD, Brant SR ec al. A genome-wide association study identifies IL23R as an inflammatory bowel disease gene. Science 2006 ; 314(5804):1461-3 . 12. Wellcome Trust Case Control Consortium and the Ausrralo-Anglo-Arnerican Spondylitis Consortium T. Association scan of 14,500 nsSNPs in four common diseases identifies variants involved in autoimmunity. Nat Genet 2007 ; 39:1329-37. 13. Colbert RA . HLA-B27 misfolding : A solution to the spondyloarrhropathy conundrum ? Mol Med Today 2000; 6(6):224-30.
276
Molecular Mechanisms ofSpondyloarthropatbies
14. Matyszak MK . Gaston JSH. Chlamydia trachornatis- specific human CD8(-) T-cells show rwo patterns of antigen recognition. Infect Immun 2004; 72(8 ):4357-67. IS. Jarvis LB, Matyszak MK, Duggleby RC et aJ. Autorcactive human peripheral blood CD8(-) T-cells with a regulatory phenotype and function . Eur J Immunol 2005; 35(10):2896-908. 16. Fontenot JD , Rasmussen JP, Williams LM ec aJ. Regulatory T-cell lineage specification by the forkhead transcription factor foxp3. Immunity 2005 ; 22(3):329-41. 17. Wan ¥Y, Flavell RA. Identify ing Foxp3-expressing suppre ssor T-cells with a bicistronic reporter. Proc Nad Acad Sci USA 2005; 102(14):5126-31. 18. Brosterhus H . Brings S, Leyendeckers H et aJ. Enrichment and detection of live antigen-specific CD4(-) and CD8(-) T-cells based on cytokine secretion. Eur J Immunol 1999 ; 29(12):4053-9. 19. Valk E, Leung R, Kang H er a!. T-cell receptor-interacting molecule acts as a chaperone to modulate surface expression of the CTLA-4 coreceptor. Immunity 2006; 25(5) :807-21. 20. Gavin MA. Torgerson TR, Houston E et a!. Single-cell analysis of normal and FOXP3-mutant human T-cells: FOXP3 expression without regulatory T-cell development. Proc Natl Acad Sci USA. 2006; 103(17) :6659 -64. 21. Xysrrakis E, Dejean AS, Bernard I et al. Identification of a novel natural regulatory CD8 T-cell subset and analysis of its mechanism of regulation . Blood 2004 ; 104(10) :3294-301. 22. Taurog JD, Richardson JA, Crofi: JT et al, The germfree state prevents development of gut and joint inAammatory disease in HLA-B27 transgenic rats. J Exp Med 1994; 180(6):2359-64. 23. Turner MJ, Sowders DP, DeLay ML et a!. HLA-B27 misfolding in transgenic rats is associated with activation of the unfolded protein response. J Immunol 2005 ; 175(4) :24J8-48. 24. Goodall JC, Ellis L, Yeo GS et a!. Does HLA-B27 influence the monocyte inflammatory response to lipopolysaccharide? Rheumatology (Oxford) 2007 ; 46(2):232-7. 25. Mise-Ornata S, Kuroda E. NiikuraJ et al. A proximal kappaB site in the IL-23 p19 promoter is responsible for ReiA- and c-Rel-dependenr transcription. J Immunol 2007; 179(10):6596-603 . 26. Utsugi M. Dobashi K. Ishizuka T et aJ. Racl negatively regulates lipopolysaccharide-induced IL-23 p19 expression in human macrophages and dendritic cells and NF-kappaB p65 trans activation plays a novel role. J Immunol 2006; 177(7) :4550-7 . 27. McKenzie BS, Kastelein RA. Cua D]. Understanding the IL-23-IL-17 immune pathway. Trends Immunol 2006 ; 27(1) :17-23 . 28. Tada H, Aiba S. Shibata K er aJ. Synergistic effect of NodI and Nod2 agon ises with toll-like receptor agonisrs on human dendritic cells to generate int erleukin-12 and T helper type 1 cells. Infect Immun 2005 ; 73(12):7967-76. 29. Yang CS , Song CH, Lee JS er al, Intracellular network of phosphatidylinosirol 3-kinase. mammalian target of the rapamycin170 kDa ribosomal S6 kinase 1 and mitogen-activated protein kinases pathways for regulating mycobacteria-induced IL-23 expression in human macrophages. Cell Microbiol 2006 ; 8(7) :1158 -71. 30. LeibundGut Landmann S, Gross 0 , Robinson MJ er al. Syk- and CARD9-dependent coupling of innate immunity to the induction ofT helper cells that produce inrerleukin 17. Nat Immunol 2007 ; 8(6):630-8.
CHAPTER
21
The Role ofB27 Heavy Chain Dimer Immune Receptor Interactions in Spondyloarthritis Simon Kollnberger* and Paul Bowness
Abstract LA-B27 (B27) is strongly associated with spondyloarthopathy.The classical role ofB27 is to present peptides from intracellular pathogens as a heterotrimeric complex with 132 microglobulin for recognition by the T-cell receptor (TCR) of CD8 T-cells. In addition to heterotrimers, B27 can also be expressed as cell surface beta2-microglobulin (f32m)-free homodimers (B27 2 ) . In addition to the TCR, MHC class I molecules bind to immunoregulatory receptor s including members of the killer immunoglobulin-like receptor (KIR) and leukocyte immunoglobulin-like receptor (LILR) families.Rodents expressthe paired immunoglobulin receptor (PIR) family which are related to LILR . B27 2 but not 132m-associated B27 binds to KIR3DL2 and rodent PIR. NK and T-cells expressing the immune receptor KIR3DL2, which interacts with B272, are expanded in B27 AS patients. Ligation ofimmune receptors by B27 2 promotes the survival ofKIR-expressing leukocytes and modulates immune cytokine production. Upregulation ofB27 2 in spondyloarthritis and differential interaction off32m-associatedHLA-B27 and B27 2 with immune receptor s could be involved in the pathogenesis ofB27-associated spondyloarthritis (AS).
H
Introduction HLA-B27 (B27) is strongly associated with development of the spondyloarthr itides, a group ofcommon arthritic diseases including ankylosing spondylitis (AS), where 94% express B27 and reactive arthritis (Re.A).' The role of HLA-B27 in disease remains unclear.' ReA follows infection with gram-negative intracellular bacteria including Salmonella Sp.,3suggesting a causal link between infection and pathogenesis. The classicalrole ofHLA-B27 is to present peptides from intracellular pathogens as a heterotrimeric complex with 132 microglobulin (132m) for recognition by the T-cell receptor (TCR) of CD8 T-cells. Experiments in B27 transgenic rodents have shown myeloid and natural killer (NK) cells may be involved in pathogenesis and CD8 T-cells are not essential for development of the disease: suggesting a possible nonclassical role for HLA-B27 in disease pathogenesis independent of its role in ant igen presentation. In addition to binding to the TCR, HLA-B27 molecules bind other regulatory immune receptors including members ofthe killer cell immunoglobulin-like (KIR) and leukocyte immunoglobulin-like (LILR) receptor families. Interaction ofB27 with these receptors could contribute to disease pathogenesis by influencing the control ofinflammation. *CorrespondingAuthor: Simon Kollnberger-MRC Human Immunology Unit, Weatherall Institute of Molecular Medicine, John Radcliffe Hospital, Headley Way, Headington, Oxford, UK OX3 90S. Email:
[email protected]
Molecular Me chanisms ofSpondyloarthropatbies, edited by Carlos Lopez-Larrea and Roberto Diaz-Peila. ©2009 Landes Bioscience and Springer Science+Business Media.
278
Molecular Mechanisms of'Spondyloarthropathies
Leukocytes Express Cell Surface B27 Heavy Chain Homodimers All alleles ofHLA-B27 have an unpaired cysteine residue at pusition 67 (cys67) in the al domain ofthe heavy chain . In classical ~2m-associated heterotrimers ofB27 thi s amino acid is buried in the peptide binding groove. HLA-B27 can be expressed at the cell surface as cys67-disulphide bonded heavy chain homodimers and multimers.5 The monoclonal antibody H C 10 recognises a number ofHLA-A, B and C heavy chains and in addition also reacts with B27 dimers ," Bycontrast the MAb MEl reacts with ~2m-associated HLA-B27 heterotrimers. Analysis with a combination ofthese antibodies thus allows the relative proportions ofheterotrimeric B27 and B27 heavy chains expressed at the surface ofcells to be determined. Resting and lipopolysaccharide-activated monocytes from B27 spondyloarthritis patients show increased FACS staining with HClO compared with monocytes from control groups.Y'Ihe increase in H C 1O-reactivityseen on these cells could result from enhanced expression of dimeric/multimeric forms ofB27 on the surface of these cells. In western blots with H C 10 B27 dimers (B27 2) electrophorese as a 90kD band on nonreducing SDS -PAGE gels reducing to a 45kD monomeric band in the presence ofdithiothreitol (DTT).1O Although HC10 is not specific for HLA-B27, B27 heavy chain monomers and dimers can be resolved from heavy chains ofother class I molecules by nonreducing 2D isoelecrric focusing and SDS PAGE. Using this technique B27 2 can be resolved from other class I heavy chains by differences in their isoectric focusing point and molecular weight. l o •l l Cell surface labelled B27 dimers can be immunoprecipitated with HC10 from leukocytes ofB27+ patients, transgenic rodent leukocytes and transfected cells.IO•12 HLA-B27-expressing cells have both intracellular and cell surface pools ofheavy chain dimers and multirners. Cell surface B27 2 expressed by B-cells is generated in a distinct fashion from the intracellular endoplasmic reticular (ER)-associated pool ofB272. Bird et al, II showed that cell surface B27 2 is formed from unstable B27 heterotrimers recycling from the cell surface in an endosomal compartment. The cell line LBL.72 1.220 is an EBV immortalised Bscell line, lacking expression of endogeneous HLA-A and B alleles and tapasin. Tapasin plays an important role as a chaperone in MHC class I assembly by optimising the peptide cargo, ensuring expression of stable class I hererotrirners at the cell surface. Different class I molecules show varying degrees of tapasin dependence for their assembly with HLA-B27 being relatively independent. LBL.721.220 B27 transfectants lack endogeneous tapasin and express more unstable heterotrimeric B27 and B27 dimers at their cell surface. By contrast LBL.72 1.220 B27 transfectants transfected with human tapasin express less dimeric and multimeric forms of B27 heavy chains as do LBL.72 1.221 B27 transfectants as well which naturally expressfunctional tapasin.!':" Stabilisation ofB27 heterotrimers expressed by LBL.72 1.220 B27 transfectants by the addition of exogeneous peptide or disruption of endosomal acidification with bafilomcycin or hydroxychloroquine inhibits cell surface expression ofB27 2 • When B27 2 is proreolytically cleaved from the surface oftransfectanrs by papain , new dimers form from preexisting B27 heterotrimers cycling through the endosomal compartment. Other B27 expressing B-ceillines including hom-2 and C 1R transfectants also express cellsurface B27 2 •1O·11 Although the majority ofcell surface B27 multimeric heavy chain species expressed by B-cells would seem to be generated via thi s extracellular route it is possible that in other antigen presenting cells (APCs) intracellular dimers may also traffic to the cell surface. B27 transgenic rat and mice splenocytes and dendritic cells express H C 1O-reactive heavy chain dimers and multimers both at their cell surface and within the cell.' ? Stabilisation of heterotrimeric B27 by overexpression of the B27 influenza nucleoprotein 383-391 peptide epitope (flu NP) has been shown to ameliorate disease in transgenic rodents.!" Given the fact that CD8 T-cells do not seem to be essential for development of disease it is tempting to speculate that at least part ofthis effect may be mediated via inhibition ofcell surface B27 dimer formation. Transgenic B27 rats which express additional human ~2m still develop arthritls." Splenocytes from these rats express reduced levels of surface ~2m-associated B27 and reduced levels of H C 10-reactive heavy chains. Since dimers were not visualised biochemically at
TheRole tiB27 Heavy Chain D imer Immune Receptor Interactions in Spondyloarthritis
279
the plasma membrane in this study, a role for cell surface B27 dimers and multimers in transgenic rodent disease should not be excluded. In addition. these studies did not address the possibility that surface B27 dimers are expressed by a subpopulation of cells as the relative levels ofHClO expression on different populations ofleukocytes were not explored. Studies of the expression of both leukocytes and transfected cell lines suggest that, in addition to B27 dimer, B27 heavy chains may also exist as disulphide bonded multimers at the cell surface.1I •l 2 Thus a significant proportion ofB27 heavy chains immunoprecipitated with HCIO from surface-labelled transfectants or rodent splenocytes, electrophorese ashigh molecular weight aggregate by nonreducing SDS-PAGE but resolve as monomeric heavy chain under reducing conditions. Some reports have also suggested that nonclassical HLA-G can also be expressed as a cysteine 42-dependent disulphide bonded dimer and multimer at the surface of transfectants and trophoblast cells.l 6.l 7
Immune Receptors Interacting with HLA-B27 Heterotrimers and Dimers In addition to binding the TCR, HLA classI molecules bind other immune receptors, including members ofthe Killer cell Immunoglobulin-like Receptor (KIR) and the Leukocyte Ig-Like Receptor (LILR) families.P:" Rodents expresspaired immunoglobulin-like receptors (PIR) which share some homology in their bindingdomains to KIR and LILR and interact with murine H -2K, D and L class I molecules." The domain organisation. cellular expression and ligands of receptor family members whi ch have been shown to interact with B27 are summarised in Figure 1. Killer cell immunoglobulin-like receptors are expressed by NK, T and NKT cell subsets . KIR s are members ofthe immunoglobulin superfamily and possesseither 2 or 3 immunoglobulin domains. KIRs with long cytoplasmic tail regions-(denoted as KIR2/3DL) incorporating two immunoreceptor tyrosine inhibitory motifs (ITIMs) are inhibitory. I 8 Upon ligation ofinhibitory KIRs the ITIMs are phosphorylated and recruit SHP-l phosphatase which inhibits tyrosine kinase mediated cellular activation. Stimulatory KIR isoforms (denoted by KIR2/3DS) have shorter cytoplasmic tails but charged transmembrane amino acids (either arginine or lysine) , enabling interaction with stimulatory DAP12 adaptor molecules. DAP12 contains two cytoplasmic immunotyrosine activation motifs (ITAMs) that are phosphorylated by src family kinases. Three domain KIRs have a DO D 1 D2, immunoglobulin domain organisation, with the D 1 and D2 domains involved in KIR ligand recognition. NK cell activation is determined by a balance between inhibitory and stimulatory receptor ligation . Many intracellular pathogens while down-regulatingMHC classI ligands for inhibitory KIR concomitantly upregulate stimulatory ligands for stimulatory NK receptors such as NKp44 and NKp46 with the net effect thus leading to NK cell activation.2 1•22 Inhibitory KIR ligation by antibody or class l-expresslng cell lines inhibits NK IFN-y production and cytotoxic killing. By contrast stimulatory KIR isoforms have a costimulatory effect on immune function. KIRs are highly polymorphic and demonstrate allele-specific recognition of class 1. The 3 domain KIRs, KIR3DLl and KIR3DL2 have a great number of different allelic variants.P The binding specificity of these different KIR alleles for class I is still being defined. HLA-B27 hererotrimers, for example. bind the three domain KIR3DLl*OOl allele. Inhibitory KIR interaction with ligand has been shown to beimportant in the generation ofNK cells which are effective killers in the absence ofinhibitory ligand/class 1, a phenomenon known as NK licencing." The stronger the inhibitory ligand bound byKIR during NK development; the more effective are the NK cells in the absence of the ligand. In NK cells the effect of inhibitory KIR is generally considered to be dominant over stimulatory signals." They act by blocking recruitment of activating signal molecules to the NK immunological synapse. KIR expression by T-cells is seen primarily on a small population of memory effector CD8 in normal healthy individuals. 26 Although KIR-expressing CD4 T-cells are also found in healthy individuals, the proportion expressing KIR is lower than among the CD8 T-cell population.i?
280
Molecular Me chanisms ofSpondyloarthropathies
Structure
Receptor
KIR3DS 1 KIR3DS 1 KIR3DL1/ KIR3DL2 DO D1 D2
Ligand binding
KIR3DL1/2
ITIMx2
Receptor PIR-A PIR-B
Structure PIR A
PIR B
ITAM x 2
Receptor L1LR A1
KIR3DL2 : HLA-A3/A11 and B27 2 (other class I»
PIR-AlB : H-2K , D, L
ITIM x4
Structure
L1LRA L1LRB
KIR3DS1 NK, NK T , minor Bw4 lIeuBO subsets of CDB and KIR3DL 1: CD4 T cells HLA-B27 and B27 2 Bw4 (other class I)
Ligand binding
D1 D2
Ligand binding
L1LRB1 /B2
D1 D2
L1LRA1 : HLA-B27 and B272 L1LRB1 : HLA-A , B, C HLA-G
ITAM x2
ITIM x4
Expression
Expression Mouse: PIR-A in macrophages, mast cells , granulocytes DC PIR-B in B cells , macrophages mast cells , granulocytes, DC Rat: PIR-A in B cells , PIR-B in myeloid eels
Expression L1LRA1 in macrophages, monocytes, Bcells L1LRB1 in B cells, monocyteslmacrophages DC , T cells , NK cells L1LRB2 in monocytes/ macrophages, DC
L1LRB2 : HLA-A , B, C HLA-G, B272
Figure 1. Schematic structures of killer immunoglobulin-like receptors (KIR), leukocyte immunoglobulin-like receptors (L1LR) and rodent paired immunoglobulin-l ike receptors (PIR) that have been shown to bind different forms of HLA-B27. Activating immunoglobulin-like receptors have a charged amino acid in the ir transmembrane domains which is involved in their interaction w ith negativel y charged amino acid residues of FcRy homodimer (PIR and L1LR) or DAP12 (KIR3DS1) adapter molecules. By contrast, inhibitory immunoglobulin-like receptors have severallTIMs in their cytoplasmic portion, important for SHIP-1 phosphatase recruitment upon tyrosine phosphorylation . The positions of the D1 and D2 ligand binding immunoglobulin domains of the respective receptors are indicated.
TheRole ofB27Heavy Chain D imer Immune Receptor Interactions in Spondyloarthritis
281
T- cellsacquire KIR expression extrathymically and the percentage expressing KIR increases with age.28 KIR ligation on T-cells is thought to have a role in setting the TCR activation threshold. Ligation of CD8 T-cell KIR may limit damage caused by potentially autoreactive T-cells and/or may promote the survival ofthese cells.i? The presence ofKIR-expressing CD4 T-cells is more evident in individuals with chronic T-cell activation such as may occur in some autoimmune disorders.'? KIR expressing CD4 T-cells from healthy individuals have features of senescent memory effector Th l T-cells lacking CD28 and CD27 expression but expressing CDS7 and 2B4 and producing primarily IFNy.27 By contrast with NK cells, the inhibitory activity ofKIRs is incomplete in CD4 T-cells. KIR2DL2 inhibits the transcriptional activation ofsome, including some cytokine genes, but not all genes after T-cell activation.P Certain CD4 T-cell effector functions such as cytotoxicity and granule expulsion are completely unaffected. In the presence ofstimulatory KIRs, inhibitory KIRs do not inhibit T -cell receptor induced production ofIFNy. Like KIR, the LILR family of receptors exist as inhibitory ITIM-incorporating and stimulatory isoforms. By contrast, however, these receptors are much lesspolymorphic than KIR and the LILR that bind to MHC class I (LILRBI and LILRB2) bind to all the different classical class I molecules that have been rested.'? LILRBI is expressed by NK, T, B-cells and cells ofmonocyte/ macrophage/dendritic cell lineage. LILRB2 have a more restricted expression on cells of the myeloid lineage. LILR receptors have four immunoglobulin-like domains; the membrane distal D 1 and D2 doma ins ofLILRB 1 and LILRB2 are the binding domains for class I. Inhibitory LILRB 1 and LILRB2 receptors incorporate 4 ITIM motifs in their cytoplasmic tails. The role ofLILR in controlling inflammation is unclear. Tolerogenic dendritic cells (DCs) which are involved in the induction of regulatory T-cells express increased levels of LILR receptors including LILRB2. 31 Ligation ofLILRB2 expressed by DC promotes the formation ofinduced T regs, suggesting LILR ligation could have a role to play in dampening immune responses under some circumstances." Rodent PIR have five extracellular immunoglobulin-like domains and like LILR interact with ligand via membrane distal D 1 and D2 binding domains." PIR are expressed as inhibitory PIRB and stimulatory PlRA isoforms and show paired expression on myeloid and B-cells in mice and unpaired expression by myeloid cells and B-cells in rats (Fig. I) . Mice have multiple PIR-A genes but only a single PIR-B gene.
Differential Binding ofB272 and B27 Heterotrimers to KIR t LILRandPIR FACS staining with B27 dimer and B27 heterotrirner tetramers demonstrates that heterotrimeric and dirneric B27 bind to different but overlapping sets ofKIR and LILR family members .r-" Whereas both B27 dimers and heterotrimers bind to KIR3DU'OOI, LILRAI and LILRB2, only B27 heterotrimers bind to LILRBI and only B27 2 binds to KIR3DL2. LILRBI and B2 are inhibitory LILR family members whereas LILRAI is a stimulatory receptor. Consistent with the binding of B27 dimers to cells transfected with these receptors B27 dimer tetramers stain leukocytes expressing these receptors. Thus, B27 2 tetramers stain B-cells and monocytes expressing LILRAI and LILRB2 and also stain primary NK and T-cell lines that express KIR3DL2. Binding ofBz'Z, to KIR3DL2 affects NK and T-cell function. KIR3DL2-expressing NK cells secrete IFNy upon coculture with LBL.721.220 cells." Ligation ofKIR3DL2 by B27 2-expressing LBL.721.220 B27 cellsinhibits IFNy secretion by NK. Bycontrast coculture ofKIR3DL2 expressing NK with control LBL.721.220 celllines transfected with HLA-B7 and HLA-B8 has no effect on IFNy secretion. LBL.721.220 cell lines transfected with the Cys67Ser mutant ofHLA-B27 and B27 transfectants transfected with human tapasin which express reduced levelsofB27 dimers do not effect NK IFNy production. B272 and ~2m-associated class I interact with KIR dlfferenrly," Binding of KIR to beta-Z microglobulin-associated MHC class I is dependent on the sequence ofcomplexed peptide, with the sequence of amino acid residues at positions 7 and 8 being particularly important for this
282
Molecular Mechanisms ofSpondyloarthropathies
interaction. Whereas HLA-B27 andHLA-A3/AII heterotrimer binding to KIR3DU*OOI and KIR3DL2 is peptide-dependent, by contrast B27 2 binding to KIR3DL2 and KIR3DU is independent ofbound peptide sequence. Differences in how B27 2 and HLA-A3/All and other class I molecules interact with KIR might account for why B27 and KIR3DL2-expressing leukocytes are involved in spondyloarthritis whereas other KIR ligand combinations such as HLA-A3/All and KIR3DL2 are not associated with disease. B27 2 but not 132m-associated B27 heterotrimer tetramers bind to rodent PIR-A and PIR-Rl2 Moreover both PIR-Aand PIR-B can be pulled down from murine macrophage cell line lysatesby immobilised monomeric B27 2 but not other class I heterotrimers. B27 2 but not control tetramers stain PIR expressing B, NK, monocyte/macrophage and dendritic cell populations from mice and rats, consistent with the distribution of these receptors. Strikingly, B27 2 tetramers also stain a population of rat CD4+ CD8+ CDI6Ia+.CDllb/c+ Cdl72a+ expressing myelomonocytic cells which are expanded in inflammatory disease in B27 transgenic rats." HLA-G transgenic and PIR-B knockout mice havecompromised maturation offunctional DC as do HLA-B27 transgenic rats. 34-36 Moreover, in an analogous manner to LILRB2 in humans PIR-B expression is induced on tolerogenic DC in rats." Thus PIR- B272 interactions may have a role to play in disease in B27 transgenic rodents.
Involvement ofKIR3DL2-Expressing T-Cells and NK-Cells in Spondyloarthritis We have shown that KIR3DL2-expressingNK and CD4 T-cells are expanded in the peripheral blood ofpatients with B27-associated spondyloarthriris. 38 Strikingly up to 30% ofmemory effector CD4 T-cells in patients express KIR3DL2. KIR3DL2-expressing CD4 T-cells from patients have a CD28-, CCR7-, CD45RO+ memory effector phenotype and are activated, expressing CD38, suggesting encounter with antigen. In addition KIR3DL2-expressing NK cells express perforin and CD38 suggesting recent activation . Ligation ofKIR3DL2 by B27 2-expressing transfectants inhibits activation-induced celldeath (AICD) of KIR3DL2-expressing leukocytes from patients." Thus binding of B27 dimers to immune receptors could promote the survival of proinflammatory NK and Tvcell populations expressing KIR in spondyloarthritis. Other authors have shown expansion ofCD28- CD4 T-cells in spondyloarthritis some ofwhich express KIR distinct from KIR3DL2.39 However, although CD28- CD4 T-cells are expanded in AS patients, no preferential expansion ofthis subset expressing KIR apart from KIR3DL2 has been shown. In rheumatoid arthritis (RA) CD4 CD28- KIR expressing T-cells are also expanded but, by contrast with B27 + AS patients, these patients have no preferential expansion of CD4 T-cells expressing KIR3DL2. 25,38 Loss ofCD28 expression is associated with changes in the dependence ofCD4 T on professional APC for activation and a shift in the use ofcostimulatory molecules to molecules such as NKG2D and stimulatory KIR. The ligands for these molecules are expressed on somatic cells such as synoviocytes. CD28- KIR+ CD4 'T-cellsrepresent a population ofchroni cally activated T-cells found in RA and AS patients, stimulated by unknown antigen. Production of perforin and cytokines by this cell subset could contribute to inflammation in patients with spondyloarthritis.
Conclusions We hypothesise that infection stimulates the production ofcell surface B27 dimers by antigen presenting cells, either by direct mobilisation from an intracellular pool or, more probably, indirectly, from recycling unstable 132m-associated B27 hererotrimers. Binding of B27 2 to KIR and LILR receptors could then promote inflammation by enhancing the survival ofproinflammatory KIR-expressing NK and T-cells or influencing th e differentiation of LILR-expressing antigen presenting cells (Fig. 2). In addition to B27 heterotrimers, leukocytes from B27+ spondyloarthritis patients and B27 transgenic rodents expresscell surfacedisulphide-bonded B27 heavy chain homodimers. Although
The Role o/B2? Heavy Chain Dimer Immune Receptor Interactions in Spondyloarthritis
283
Ligation of KIR and LILR receptors Teen
Bacterial infection induces expression ofHLAB27 homodimers, activating Antigen receptors for presenting cell NK cells Dendritic cell/macrophageJ and bacterial peptide monocyteJ presentation toT cells
... .
IHLA·B27
B272
Hetemlrimers
f
L1LR
KIR
l
Teen
Intubition of T and NK activation induced eeD dea
A1CDvIa KJR ligation n of TandNK cen activation
Costiroo
KJR I llon . llftltlliatlon of APes, DCs Monocyteslmacrophage viaLlLR ligation
TeR
,
AdivatOlY ~ Adi:;.tory NK • receptor
Receptor ligand
Figure 2. Hypothesis for the involvement of B27 dimers in spondyloarthritis. Infection stimulates the production of cell surface B27 dimers by antigen presenting cells, either by direct mobilisation from an intracellular pool or, more probably, indirectly, from recycling unstable 132m-associatedB27 heterotrimers. Concomitantly, infection stimulates antigen presentation to T-cells and the express ion of ligands for activating NK receptors such as NKp44 and NKp46 . Binding of B27 2 to KIR and L1LR receptors co uld th en promote inflammation by enhancing the survival of proinflammator y KIR-expressing NK and T-cells or influencing the differentiation of L1 LR-expressing anti gen presenting cells .
indirect evidence and biochemical studies with cell lines indicate that B-cells and possibly other professional antigen presenting cells (monocyres, macrophagesand D C s) express dimers, the exact cell types expressing B27 2 in disease await identification. Also it remains to be determined how infection might affect expre ssion ofB27 2• B27 2 and 132m-associated B27 hererotrimers interact with different but overlapping immune receptors. Moreover, B27 2 and 132m-associated class I heterotrimers bind differently to KIR. Both differences in the quality of dimer and classical 132m-associated class I immune receptor interactions and the specific immune receptors they bind to could influence pathogenesis. B27 2 but not B27 heterotrimers bind to the NK and T-cell expressed KIR3DL2 and rodent paired immunoglobulin-like receptors. Activated NK cells and memory effector CD4 T-cells expressing KIR3DL2 are expanded in B27 + spondyloarthritis patients but the role ofthese cells in pathogenesis is at present unclear. Although KIR-B27 2 interactions promote survival ofKIR3DL2-expressing leukoc ytes in vitro how these cells survive and their role in inflammation in vivo has yet to be determined. Likewise, the role of LILR-B27 2 and PIR-B27 2 interactions in human and rodent disease is presently unclear.
Acknowledgements We thank the arthritis research campaign and medical research council (UK) for supporting thi s work.
284
Molecular Mechanisms ofSpondyloarthropathies
References 1. Brown MA . Pile KD. Kennedy LG ct al. HLA class I associations of ankylosing spondylitis in the white population in the United Kingdom . Annals Rheum Dis 1996; 55:268-270. 2. Allen RL. Bowness P. McMichael A. The role of HLA-B27 in spondyloarthritis. Immunogenetics 1999; 50:220-227. 3. Colmegna I. Cuchacovich R. Espinoza LR. HLA-B27-associated reactive arthritis : Pathogenetic and clinical considerations. Clin Microbiol Rev 2004 ; 17(2) :348-69. Review. 4. May E, Dorris ML. Saturntira N er al. CD8 «/3 T-cells are not essential to the pathogenesis of arthritis or colitis in HLA-B27 transgenic rats. J Immunol 2003; 170(2) :1099-105. 5. Allen RL, O'Callaghan CA. McMichael AJ et al. 1999. Cutting edge : HLA-B27 can form a novel beta 2-microglobulin-free heavy chain homodimer structure. J Immunol 162:5045-5048. 6. Starn NJ, Vroom TM, Peters PJ et al. HLA-A- and HLA-B-specific monoclonal antibodies reactive with free heavy chains in western blots. in formalin-fixed. paraffin-embedded tissue sections and in cryo-imrnuno-electron microscopy. Int Immunol 1990; 2(2) :113-25 . 7. Raine T. Brown D, Bowness P et al. Consistent patterns of expression of HLA class I free heavy chains in healthy individuals and raised expression in spondyloarrhropachy patients point to physiological and pathological roles. Rheumatology (Oxford) 2006; 45(11) :1338-44. 8. Tsai WC, Chen CJ. YenJ er al. Free HLA class I heavy chain-carrying monocytes-A potential role in the pathogenesis of spondyloarrhropathies, J Rheumatol 2002 ; 29:966-972. 9. Lan CC, Tsai WC. Wu CS er al. Psoriatic patients with arthropathy show significant expression of free HLA class I heavy chains on circulating monocytes : A potential role in the pathogenesis of psoriatic arthropathy. Br J Derrnarol 2004; 151(1) :24-31. 10. Kollnberger S. Bird L. Sun MY et al. Cell-surface expression and immune receptor recognition of HLA-B27 homodimers. Arthritis Rheum 2002 ; 46(11):2972-82. 11. Bird LA. Peh CA. Kollnberger S et al. Lymphoblastoid cells express HLA-B27 homodimers both intracellularly and at the cell surface following endosomal recycling. Eur J Immunol 2003 ; 33 :748-759. 12. Kollnberger S. Bird LA. Roddis M et al. HLA-B27 heavy chain homodimers are expressed in HLA-B27 transgenic rodent models of spondyloarthritis and are ligands for paired Ig-like receptors . J Immunol 2004; 173(3) :1699-710. 13. Kollnberger S, Chan A. Sun MY et al. Interaction of HLA-B27 homodimers with KIR3DLI and KIR3DL2. unlike HLA-B27 heterocrimers, is independent of the sequence of bound peptide. Eur J Immunol 2007; 37(5 ):1313-22. 14. Zhou M. Sayad A, Simmons WA et al. The specificity of peptides bound to human histocompatibility leukocyte antigen (HLA)-B27 influences the prevalence of arthritis in HLA-B27 transgenic rats. J Exp Med 1998 ; 188(5) :877-86. 15. Tran TM, Dorris ML. Satumtira N ct al. Additional human beta2-microglobulin curbs HLA-B27 misfolding and promotes arthritis and spondylitis without colitis in male HLA-B27-transgenic rats. Arthritis Rheum 2006; 54(4) :1317-27. 16. Boyson JE. Erskine R, Whitman MC et al. Disulfide bond-mediated dimerization of HLA-G on the cell surface . Proc Natl Acad Sci USA. 2002 ; 99(25) :16180-5. 17. Apps R. Gardner L. Sharkey AM et al. A homo dime ric complex ofHLA-G on normal trophoblast cells modulates antigen-presenting cells via LILRBI. Eur J Immunol 2007 ; 37(7) :1924-37 . 18. Lanier LL. Follow the leader: NK cell receptors for classical and nonclassical MHC class I. Cell 1998; 92 :705 -707 . 19. Colonna M. Nakajima H. Navarro F er al. A novel family of Ig-like receptors for HLA class I molecules that modulate function of lymphoid and myeloid cells. J Leukoc Bioi 1999; 66(3 ):375-81. Review. 20. Takai T. Paired immunoglobulin-like receptors and their MHC class I recognition. Immunology 2005 ; 115(4):433-40. Review. 21. Diefenbach A. Raulet DH. Innate immune recognition by stimulatory immunoreceptors. Curr Opin Immunol 2003 ; 15( 1):37-44. Review. 22. Vankayalapati R. Wizel B, Weis SE er aL The NKp46 receptor contributes to NK cell lysis of mononuclear phagocytes infected with an intracellular bacterium. J Immunol 2002; 168(7) :3451-7 . 23. Vilches C, Parham P. KIR : diverse. rapidly evolving receptors of innate and adaptive immunity. Annu Rev Immunol 2002 ; 20:217-51. 24. Yokoyama WM, Kim S. How do natural killer cells find self to achieve tolerance? Immunity 2006; 24(3 ):249-57. Review. 25. Goronzy JJ. Henel G. Sawai H et aL Costirnulatory pathways in rheumatoid synovitis and T-cell senescence . Ann NY Acad Sci 2005 ; 1062:182 -94. Review. 26. Young NT. Uhrberg M. KIR expression shapes cytotoxic repertoires : a developmental program of survival. Trends Immunol 2002; 23(2) :71-5. Review.
The Role ofB27Heavy Chain Dimer Immune Receptor Interactions in Spondyloartbritis
285
27. Van Bergen]. Thompson A. van der Slik A et al. Phenotypic and functional characterization of CD4 T-cells expressing killer Ig-like receptors .] Immunol 2004; 173(11) :6719-26. 28. Abedin S. Michel ]]. Lernster B ec al. Diversity of NKR expression in aging Tvcells and in T-cells of the aged: the new frontier into the exploration of protective immunity in the elderly. Mech Ageing Dev 2005; 126(6-7):722-31. 29. Ugolini S. Arpin C. Anfossi N et al. Involvement of inhibitory NKRs in the survival of a subset of memory-phenotype CD8+ T-cells. Nat Immunol 2001; 2:430-435. 30. Nelson GW. Martin MP. Gladman D et al. Cutting edge: heterozygote advantage in autoimmune disease: hierarchy of protection/susceptibility conferred by HLA and killer Ig-like receptor combinations in psoriatic arthritis. ] Immunol 2004; 173(7):4273-6. 31. Ristich V. Liang S. Zhang W et al. Tolerization of dendritic cells by HLA-G. Eur] Immunol 2005; 35(4) :1133-42. 32. Chang CC. Ciubotariu R. Manavalan ]S er al. Tolerizarion of dendritic cells by T(S) cells: the crucial role of inhibitory receptors ILT3 and ILT4. Nat Immunol 2002 ; 3(3) :215-7. 33. Allen RL. Raine T. Haude A er al. Leukocyte receptor complex-encoded immunomodulatory receptors show differing specificity for alternative HLA-B27 structures.} Immunol 2001; 167(10) :5543-7 . 34. Horuzsko A. Lenfanr F. Munn DH et al. Maturation of antigen-presenting cells is compromised in HLA-G transgenic mice. Int Immunol 2001; 13(3) :385-94 . 35. Ujike A. Takeda K. Nakamura A et al. Impaired dendritic cell maturation and increased T(H)2 responses in PIR-B(-/-) mice. Nat Immunol 2002; 3(6) :542-8. 36. Hacquard -Bouder C . Falgarone G. Bosquet A et al. Defective costimularor y function is a striking feature of antigen-presenting cells in an HLA -B27-transgenic rat model of spondylarthroparhy, Arthritis Rheum 2004 ; 50(5):1624-35. 37. Liu ], Liu Z . Witkowski P et al. Rat CD 8+ FOXP3+ T suppressor cells mediate tolerance to allogeneic heart transpl ant s. inducing PIR -B in APC and rendering the gratt invulne rable to rejection . Transpl Immunol2004; 13(4) :239-47 . 38. Chan AT. Kollnberger SD. Wedderburn LR et al. Expansion and enhanced survival of natural killer cells expressing the killer immunoglobulin-like receptor KIR3DL2 in spond ylarthritis. Arthritis Rheum 2005 ; 52(11):3586-95. 39. Dufiner C. Goldberger C. Falkenbach A et al. Prevalence. clinical relevance and characterization of circulating cytotoxic CD4+CD28- T -cells in ankylosing spondylitis. Arthritis Res Ther 2003; 5(5) :R292-300.
CHAPTER
22
KIR Genes and Their Role in Spondyloarthropathies Roberto Diaz-Pefia, Miguel Angel Blanco-Gelaz and Carlos Lopez-Larrea" Abstract
C
ellular activity ofnatural killer cells (NK cells) is defined by the balan ce bet ween activat ing and inhibitory signals com ing from their receptors. With respect to this response. killer immunoglobulin-like receptors (KIR) are unique because of their diversity and capacity to recognize specific human leukocyte antigen (HLA) class I allorypes. Up to the present few studies have experimentally been developed concerning the role ofKIR genes in spondyloarthropathies (SpA) and its clear relationship with HLA-B27. However. the role of the HLA-B27 heavy chain homodimers and their possible recognition by KIR receptors in the pathogenesis of spondylarthritides has been studied. Moreover, it has been sugge sted that NK cells and their receptors could playa role in ankylosing spondylitis (AS) development. Several association studies based on a model in which KIRs synergize with HLAs have also been published. This interactio n may generate compound genotypes which pro vide different levels of activation or inhibition. Furthermore, so me of these have been associated with certain SpA, such as ankylo sing spondylitis (A S) and psoriatic arthritis (PsA) .
Introduction N atural killer cells (N K cells) are defined as granular cytotoxic lymphocyte s th at do not express antigen receptors. Tcell receptor or B-cell receptor (TCR or BCR). but which usually express the surface markers CD 16 (FcyRIII) and CD 56 in humans. Th ey constitute a major component ofthe innate immune system and play an important role in cells infected by viruses, bacteria and parasites and in the rejection of tumors. NK cells also playa key role in regulating autoimmune responses. I Th ey can be activated through different receptors by their union to ligands present in target cells but their activit y is normally restricted by inhibitory receptors th at recognize target cell-expressed MHC class 1 mole cules. The "missing self" hypothesis proposed that, when MHC class I molecules are downregulated or absent. NK cell are discharged from the inhibitory influence ofthese receptors and kill target cells more cfficicn rly.' In this way. NK cells function is controlled by a delicate balance between activating and inhibitory signals. Thus. a revision ofthe "missing self " hypothesis proposed that NK cells che ck the abnormal cells that lack MHC class I molecules or overexpress ligands for activating NK cells receprors.i The proportion between different NK cell receptors and the number of ligands in the target cell, as well as the differences in the intensity of signal tran sduction, determine th e NK cell respon se. In humans, an important component ofNK cell target recognition depends mainly on the control of human leucocyte antigen (HLA) class I molecules by killer immunoglobulin-like *Correspo nd ing Author: Carlos l.op ez-Larrea-c-Departm ent of Immunology, Hospital Universitario Central de Asturias, C/Celestino Villamil sin, 33006-0viedo, Spa in. Email: inrnuno sehc a.es
M olecular M echanisms ofSpondyloarthropath ies, edited by Carlos Lopez-Larr ea and Roberto Diaz-Peiia. ©2009 Landes Bioscience and Springer Science+Busine ss Media.
287
KIR Genesand Their Role in Spondyloarthropathies
19
LRC Figure 1.Map of the Leukocyte Receptor Complex. The Leukocyte Receptor Complex (LRC) is formed by a cluster of genes that encode a family of proteins that contain immunoglobulin-like domains. These include the families "killer immunoglobulin-like receptors (KIR), "leukocyte immunoglobulin-like receptor" (L1LR) and " leukocyte-associ ated immunoglobulin-l ike receptor " (LAIR). The "signall ing lectins" (SIGLECs) and members of the family CD66 are found close to LCR. In humans , this complex is situated on chromosome 19.
receptors (KIR) . The presence ofdifferent combinations ofKIR genes can generate inhibitory or activatory signals to the cell and the effector function is considered to result from the balance of th ese contributing signals. Moreover, the relationship between HLA and KIR polymorphism in relation to allele frequency or level ofexpression on NK cells is very important, because NK cell ligand-receptor recognition can vary at different levels: HLA polymorphism. KIR haplotype. KIR allelic polymorphism. The cause of the diversity of KIR genes and the contribution of each one to signalling in N K and T-cells are not clear. but their relevance has been supported by different genetic studies.' : These studies show that different HLA/KIR genotypes can provide different levels of activation in the NK cell repertory and that these combinations are associated with differing susceptibility to and protection against a range of diseases in this model. The relationship between HLA-B27 and spondyloarthropathies (SpA) has been known for a long rime.?" and molecular mimicry between foreign and self-peptide has been proposed as the cause that could unleash a cytotoxic T -cell reponse leading to the autoimmune destruction of self-tissues." However, despite intensive research, the pathogenic role of the gene and its product has not been resolved as yet. It had been assumed that the association between HLA-B27 and SpA could reflect some aspect of irs principal function such as presenting antigenic peptides to CD8+ T-cells but, in addition to this classical role, a recognition of HLA-B27 by members of the KIR family (3D Ll and 3DL2) ha s been described. ' " Moreover, it has been suggested that NK cells might playa role in AS pathogenesis and implicate chemokines in the mechanism ofNK receptor expression ." Therefore, KIR receptors probably playa role in the development of SpA.
Complexity ofKIR Family KIR Structure KIRs are a diverse family of receptors expressed on human NK cells and a subset ofT-celis. The KIR gene cluster is located on human chromosome 19q13.4 within the leukocyte receptor complex (LCR) (Fig. 1) and con stitutes a multigene family whose genomic diversity is achieved through differences in gene content and allelic polymorphism. LRC also contains genes encoding other cell surface molecules with distinctive immunoglobulin-like extra-cellular domains. The se genesinclude Leukocyte Ig-like Receptors (LILR), Leukocyte-Associated Ig-like Receptors (LAIR) and the Feu Receptor. In addition the LRC contains genes encoding the CD66 family members, Sialic and binding Ig-like Lectins (SIGLEC) and the genes encoding the transmembrane adaptor molecules DAPIO and DAP12,12 The KIR gene family currently consi sts of 15 genes (KIR2DLl, KIR2DL2, KIR2DL3, KIR2DL4, KIR2DL5A, KIR2DL5B, KIR2DSl, KIR2DS2, KIR2DS3, KIR2DS4, KIR2DS5,
KIR3Ol2
KIR3DL3
-!-
-!-
I KIR3DLli
X1m M
X2mM
~
I KIR30S1 1
-!-
X 2 ITA"
0 ... 12
~
KIR30S
KIR2Ol 68
KIR20l6A
-!-
X2m "
1\'1>. A
I KIR2DUI
-!-
X1m M
UNKNOWN ADAPTOR IlOt.ECULE
1)opeU
r.---~A
KIR20L
"
I
~
~ KIR2DS5
...,
So ~.
Ol ~
S-
s;:. ...
";,
~
~
::;.
...r;;~ ~ ;,
£1
~
~
KIR2DS4
I HLA~ I
HLA-A
I HLA-Bw4 1
I HLA~2 I I HLA~1 I
HLA LIGANDS
~
KIR2DS3
KIR2DS21
I KIR20S1
-!-
X2rTAM
OAP12
~
KIR20S
IKIR2DL1 ~ IKIR2Dl2~ I IKIR2DL3~
-!-
X2m"
1\'1>< I
KIR20
Figure 2. KIR protein structures and ligands. Structural characteristics of KIR proteins . The ITAM motifs are shown in grey boxes, whereas the ITIM motifs are shown in white boxes. The adaptor molecules (DAP12) are shown in red. Individual KIR ligands are shown in small squares of different colours (KIR ligands for KIR3DL3, KIR2DL5, KIR2DS3 and KIR2DS5 are unknown) .
02
01
DO
KIR30L
r.----A~
KIR30
~
~ -~ 20
10
10
/
o
iiU
.,
•
10
'00
=~~~
olD
ID
KIR2DS2 •
~ ;.~
_
KIR2DSI •
'V
_
•
KIR3DLI
_
10
~
10
10
o
•
..
.
III
III
KIR20LI
n. ~ =:~», ~. ~:~~::~;:~
:;~~ .
.....
~2Dlt ~ n.
100
-' I!!!!!!!!!!!!!!~ ::t;;~;::====:::::1
10
. . , ~~~ • .
I!lL1 I!lL1
nl
..... --;~~
..........==-r"'====~
...
.... .. . . . . =~~
Figure 3. KIR gene distribut ion in different worldwide popul ations.
.....
~':~E .
_ _I
. . . . =~~~:========~.
..._.
n.
~~~
. . . . =~~:~:::======::J -_.
= ~!!!!!!!!!!!!!!.
_.
'C
1;;
s.. ~.
~
~
s.'"
s-
:. ~
'"
~
ii'
l\-
~
~ ::;.
:::....
:.
'"
~
~ :.
~ ~
290
Molecular Mechanisms ofSpondyloarthropathies
KIR3DU, KIR3DL2, KIR3DL3 and KIR3DSl) and two pseudogenes (KIR2DPI and KIR3DPl).The nomenclature for KIRs isaccording to whether they have two (2D) or three Ig-like domains (3D) (DO, Dl, D2) and to whether they possess a short (S) or long (L) cytoplasmic tail (Fig. 2). Additionally, the letter P denotes pseudogenes .All KIR3D molecules have the DO-D1-D2 configuration, whereas KIR2D molecules can have either D I-D2 (Type 1)or DO-D2 (Type II) configurations. Although the precise function and the ligands ofsome KIRs remain controversial they have been classified into two types. One type consists ofactivating KIRs (KIR2DS and KIR3DS), which have a short (S) cytoplamic tail with the capacity to interact with activating adaptor proteins such as DAP 12, deliveringactivating signalsthrough there immunoreceptor tyrosine-based activating (ITAM) motif. 12,131he other type consists of the inhibitory KIR (KIR2DL and KIR3DL), which has one or two immunoreceptor tyrosine-based inhibition (ITIM) motifs in their long (L) cytoplasmic tail, with the exception of2DL4. KIR2DL4 has an ITIM motif in the cytoplasmic tail and recent studies indicate that this receptor associates with the accessory protein, FCERI-y, which sends a stimulatory signal via its ITAM similar to DAP 12. 14
Haplotypic andAllelic Diversity The complexity ofKIR is compounded by susrantial allelic and haplotypic diversity. Each KIR gene has between four and nineteen alleles and individuals can have a different number of KIR genes and these can occur together in many differcnt combinations on a single chromosome to generate considerable haplotypic diversity,":" There are four KIR genes present in practically all haplorypes: KIR3DL3, 3DP1, 2DL4 and 3DL2. Based on their gene content, haplorypes have been divided into two basic groups: haplotype A and B. Haplotype A contains only one activating KIR gene, 2DS4 which frequently displays a deletion in exon 5 and does not encode a membrane protein, 16 whereas haplotype B contains various combinations ofactivating KIR genes, KIR2DS 1, 2DS2, 2DS3 , 2DS5, 3DS 1 and 2DS4 and exhibits extreme diversity both in terms ofgene content and allelic polymorphism. Several studies have focused on investigating KIR diversit y in various human popularions. F:" Different studies show a high degree of haplotype diversit y and allelic polymorphism and these result in variable activating or inhibitory signalling potentials. An example ofthe different distribution ofKIR genes is shown in Figure 3. The unusual and impressive diversity which combines HLA class I molecules and KIRs is a fertile substrate for natural selection and indicates its role in the adaptation ofthe immune system in response to a continuously challenging environment, but as yet, its functional consequences remain largely unexplored. Thus, it is important to consider ethnic differences in HLA class I as well as in KIR genes as risk factors (NT=not tested), because many effects have been reported in different diseases.P" The A and B haplorypes are almost equally distributed in Caucasians and in some Amerindian tribes ." It has been reported that, in Asian populations such as the Japanese, Koreans and Han Chinese, A haplotypes are more frequent,19.21 whereas those of B haplorypes are more frequent in Australian Aborigines.F The allelic diversity of KIR create s a further level of complexity. A good example is the KIR3DL1/ 3D S1 locus, whose allelic relationship has been recognized by the KIR Nomenclature Committee, with consecutive assignments to variations within these genes (KIR3DLI alleleshave been numbered from KIR3DU'OOI to '009, KIR3DS1'010 to '014and KIR3DU'015 to '020) (IPD-KIR Sequence Database: http:/ /www.ebLac.uk/ipd/kir/) . KIR3DU isa very polymorphic in hibitory receptor, which include an apparent activating counterpart, KIR3DS 1,with allelesthat are expressed at either high or low levelson the cell surface" and one allele, KIR3DL 1*004, that is not expressed on the cellsurface and is retained within the cell.24 Thus, the functional repercussions ofvariability at the KIR3D U locus provide different features for the alleles of this locus: alleles encoding high -expression allotypes (KIR3DU'00 1, '002, '008, '009, '015, '020), low expression allorypcs (KIR3DU*005, '006, '007) and no cell surface expression (KIR3DU·004) .23.2\
KIRLigands In addition to interaction with TCR, it has been shown that mature class I complexes bind several other immunomodulatory molecule s, including members ofthe KIR family.To date, only
IDDM
Scleroderma Behcet disease Psoriasis vulgaris
Other autoimmunes diseases
Rheumatoid arthritis
Rheum atoid vasculitis
Other rheumatic diseases
Ank ylosing spondy litis
Spo ndy larthrit ides
Psoriatic arthritis
Spondyloarthropathies
Disease
- KIR2DS2+/ KIR2DL2-A ltered KIR3D Ll expression -2DS1; H LA-Cw · 06 -2DS1; 2D L5; Haplotype B -KIR2D S2; HL A-Cl
-KI R2DS2; H LA-Cw·03 -Increased KIR2L2/2DS2 in patients with extra articular manifestation s Decreased KIR2DS1/3D51 in patient s withou t bo ne erosions -KIR2DS4; HL A-Cw 4
-2DS1/2 DS2; HLA -Cw group homoz ygosity -Increased KIR3DL 2 expression -Interactio n H LA-B27 homodimers wi th KIR3D L1IKIR3DL 2; indepe ndent of pept ide -KIR3DL1/3 DS1; HL A-B27 genotypes
Kir Association
54,55
Suscepti bility
74 75 76 -77 78
73 Susceptibi lity Susceptibilit y Asociated w ith severe eye disease Susceptibility Suscepti bility Susceptibil ity
72
Susceptibility Cli nica l manifestatio ns may have different Genetic backgro und s w ith respect to the KIR genotype
71
63,64 70 10
Reference
Susceptibility M ay co ntribute to disease pathology M ay co ntri bute to disease pathogenesis
Observation
Table 7. Associations of KIRs with spondyloarthropathies, rheumatic diseases and other autoimmunes disorders
::::
I
N
~
Q
~ .
So
I ~
...
::!. ;:,..
.... I:l
It
:ii'
~
I ~
...~ .
S:l
l:>..
;:s
I:l
~Q
::ti
292
Molecular Mechanisms ofSpondyloarthropathies
class I molecules have been identified as ligands of KIRs (Fig. 2). In this way, KIR proteins are directly involved in the activation and inhibition ofNK and, possibly, also in CD8+ T-cells. 26 Inhibitory KIR molecules bind to target cell HLA class I molecules and prevent the attack ofNK cells.i? KIR2DL2/2DL3 binds to HLA-C group 1 molecules, which have aminoacids Ser77 and Asn80, KIR2DLl binds to HLA-C group 2 molecules which have aminoacids Asn77 and Lys80 . KIR2DL4, which is an unusual KIR receptor because ofits extracellular domain organization and the possession ofonly one ITIM, interacts with HLA-G.27 KIR3DLl binds the HLA-B o l helix around residues 76-80, with specificity for all Bw4 alleles containing isoleucine at heavy chain residue 80,28 whereas the ligand for KIR3DS 1 has not been determined as ofyet." Serological Bw4 epitope is defined by a sequence rnotifin the C -terminal part ofthe HLA-B a 1helix. Serologically, heterogeneity exists in the Bw4 epitope and there is evidence that Bw4+ HLA-B allotypes differ in their capacity to engage KIR3DLl,3° In addition to the complexity of the system is the feature of polymorphism at the KIR3DLllocus to produce different levels ofcell surface expression ofthe protein." These differences will affect the strength of the inhibitory signal generated by engagement ofKIR3DLl with a Bw4+ HLA-B ligand. In contrast to the inhibitory KIR, ligands for the activating KIR are largely unknown despite possessing similar extracellular regions, although, it has been suggested that they recognize the same HLA-B or HLA-C molecules that are recognized by their related inhibitory KIRs.
KIRIHLA Peptide Dependence In general, KIR2D receptors recognize HLA-C alleles, whereas KIR3D receptors recognize HLA-A and HLA-B alleles. Our understanding of the different specificity of KIR receptors for HLA alleles derived from crystallographic studies of KIR molecules and the binding of the respective HLA ligand. The crystal structure of the KIR3DLI receptor is not known, but its interaction with B*270S can be inferred from the structures of KIR2DL2-HLA-Cw3 and KIR2DL 1-HLA-Cw4,31·32Peptide preferences have been documented in studies related to KIR2D binding to HLA-C molecules'P" and it has been shown that the KIRlHLA interaction is most sensitive to substitutions at peptide positions 7 and 8, this being in agreement with the finding that the KIR binding site is located near C-terminal residues 7 and 8 ofthe peptide. In contrast, TCRs, which exhibit much greater peptide selectivity than KIRs, generally focus on the central portion of the MHC-bound peptide, at and around the PS position.v The KIR3DLl receptor has specificity for bound peptide and shows greater sensitivity for the residue at position 8.36.37 The interactions ofdifferent peptides with the HLA-B270S groove have been analysed and it has been shown that the P8 glutamic acid residue of the EBV EBNA3C peptide prevents KIR3DLl receptor recognition of the HLA-B270S-EBV cornplex.t" Thus , peptides derived from microbial infections could be "arthritogenic" EBV epitope bound to B*270S fails to bind KIR3DLl and this suggests, that during infection, microbial peptides presented by HLA-B27 may have a pro -inflarnatory effect, through increased NK activation.
KIR and Spondyloarthropathies The spondyloarthropathies are a group ofinflammatory rheumatic diseases, typically affecting the axial skeleton. They present similarities in some of their clinical manifestations, namely, the association to HLA-B27 and certain radiological characteristics and treatments. Given their class I specificity and immunomodulatory functions, KIR receptors might play an important role in the association to disease ofHLA class I molecules. Different associations of KIRs with spondyloarthropathies and other autoimmunes disorders are shown in Table 1. A summary ofthese association studies based on the model that KIRs synergize with HLAs in order to generate compound genotypes which provide different levels ofactivation and inhibition on NK or T-cells and some of these , have been associated with some SpA , such as AS and PsA. However, few studies so far have experimentally addressed the role of leukocyte receptor recognition in spondyloarthropathies.
293
KIR Genesand Their Role in Spondyloarthropathies Psoriatic arthritis
Ankyloslng spondylills
Mcm Inh jblUon
A
Target ce ll
NK c e ll
c
T ar g et ce ll
NK c ell
D
Mcm activation
Figure 4. Possible mechanism of the pathogenic role of KIR in SpA. Ankylosing Spondylitis: A) HLA-Bw4 (B27) can be recogn ized by KIR3DLl conferring a strong inhibition and preventing the lysis of the cell B) Recognition of Bw4 (B27) by KIR3DLl can be blocked (some viral peptides may cause the blockage of the inhibitory receptor 3DLl or KIR3DL1 polymorphism can serve to alter the inhibitory response)and therefore, it is probable that a significant fraction of arthritrogenic ligands of B*27 might avoid their interaction with KIR3DLl in the course of a microbial infection and influence the activation of activating NK receptors (KIR3DS1) and the subsequent progression to AS. Psoriatic Arthritis: C) The susceptibility to PsA develop could be determined by the balance of activating and inhibitory KIR-HLA genotypes, where protective inhibitory effects against PsA appear to be restricted to KIR2DL. D) The combination of KIR2DSl and/or KIR2DS2 with their HLA-Cw ligand group homozygosity (HLA-C2 and HLA-Cl, respectively), where the inhibitory signal is missing, confers susceptib il ity to PsA. There are different intermediate combinations KIR-HLA and these genotypes cou ld be classified by their ability to confer activation (susceptibility) or inhibition (protection).
KIR andAnkylosing Spondylitis Linkage between HLA-B27 and AS remains the strongest association between an HLA class I molecule and disease,6-8as has been demonstrated worldwide and evidence for the role ofHLA-B27 in AS comes from linkage and association studies in humans and transgenic animal models.39.40 HLA-B27 represents a family of at least 39 closely related alleles: B2701 to 2739, which differ in their ethnic distribution and whose heterogeneity has been previously determined in different populations.t'<" HLA-B27 has a peptide binding cleft formed by the al and a2 domains of the heavy chain that has six side-pockets (designated by the letters A through F), which accommodate the side chains ofthe amino acids ofthe bound nonamer peptides. B27 subtypes differ from each other by one or a few amino acids and they overlap peptide repertories." Thus , differences in antigen presentation can be largely interpreted in terms ofthe effect ofpolymorphism on pocket interactions. Moreover, at least two alleles, HLA-B*2706 which is common among in Southeast Asians and HLA-B*2709 which is primarily observed in Sardinians, lack such an association with AS.44.45 The pathogenic role ofthe HLA-B27 gene and its product hasnot been resolved as yet. However, studies indicate that HLA-B27 contributes only about 16% ofthe total genetic risk for disease." There are nonB27 genes, both within and outside the HLA, that also seem to be involved in disease
294
Molecular Mechan isms ofSpondyloarthropathies
etiology. With respect to the association with the MHC region. located in the short arm ofchromo some 6. MHC class II alleles. genes that encode TNF or complement and genes which are involved in the antigenic presentation ofclass I molecules (TAP. LMP2 and LMP7) 46 were investigated. In another study. an additional effect ofHLA-DR4 in HLA-B27 positive ind ividuals in relation to the increased risk ofspondylitis ankylosing was demonstrared." Furthermore. genome-wide scans have implicated regions showing "suggestive" or stronger linkage with AS on chromosomes 2q . 6p, 6q. l Oq, Llq, 16q. 17q and 19q.48.soThese data seem to indicate that the susceptibility to develop AS is related with multigenic interactions. including HLA-B27 in a substant ial way. Given th e receptor-ligand relation ship between certain combinations ofKIR and HLA class I molecules. it is reasonable to hypothesise a synergism between th ese polymorphic loci. This is the case of the KIR3DLl /3DSllocus and HLA-B27. As far as we know. KIR3DLl is the only KIR known to recogn ise HLA-B alleles and it binds to HLA-B whi ch also contains the serological-defined epitope Bw4. 51All HLA-B27 subtypes carry the sequence s specifying the serological Bw4 epitope,with the exception ofB *2708 and other related subtypes, which carry Bw6. KIR3DS 1 have more than 97% sequence identity to KIR3DLl in the extracellular dom ains and the differences of both receptors are located in the intracytoplasmic tail. However, the ligand for KIR3DSI has not been clearly determined." although it has been shown that KIR3DS 1 in combination with HLA-B alleles that encode molecules with isoleucine at position 80 (HLA-B Bw4-801) results in delayed progression to AIDS after HIV-l infection." In another study, a diverse panel of HIV-l specific Bw4-801 HLA class I tetramers was generated and its ability to bind KIR3DLl and KIR3DSI assessed. However, tetramer binding to KIR3DS 1 was not detected.P Genetic evidence for a possible implication of the KIR3DLl/3DSI locus in AS has been reported by US. 54.55 We found that the inhibitory allele was decreased in AS patients compared to B27-po sitive healthy controls , whereas KIR3DSI was substantially increased in AS patients. The se studies were carried out in four different populations, two Caucasian and two Asian and it has been reported elsewhere that Asians differ from Caucasians in the genetic profile ofKIR and B27 alleles.20•41These studies show that the activato r effect seems to be pr evalent in all populations and the same tendencyis observed, that is,an increase ofKIR3DSI and a decrease ofKIR3DLl in AS patients. Due to polymorphism present in the KIR3DLl/Sllocus, different 3DLl allotypes can differ in their capacity to inh ibit NK cell function.P In this way, allelic polymorphism could modulate the inhibitory function ofKIR3DLl and Bw4-3DLl interaction could be a potentially important source of functional NK or T-cells diversification in all such populations. KIR3DLl polymorphism has been directly correlated with differences in cell surface phenotype as detected with the KIR3DLl-specific mAb , DX9 ,S6·57and KIR3DLl allorype s have been subdivided into three groups according to how NK or T-cells expressing them bind the DX9 Ab (Gardiner CM et al2001): alleles encoding high -expression allotypes (KIR3DLl *h), low-expression allotypes (KIR3DLl *£) and no cell surface expression (*004) . A further stu dy grouped the inhibitory KIR3DLI alleles based on distinct functional characteristics and identified differential protective effects of the se allelic groups on HIV disease.58 It has been reported that KIR3DS1, in combination with the HLA-B Bw4-80Ile. is associated with delayed progression to AIDS in individuals infected with human immunodeficiency virus Type 1 (HIV-l). Moreover. the possibility ofgreater affinity ofKIR3DL1*1 allotypes for ligands containing Bw4-80T rather th an for those containing Bw4-801 has been suggested. The role ofthe HLA-B27 heavy chain homodimers in the pathogenesis ofspondylarthritides has been investigated. Homodimer formation is dependent on the presence of Cys67 ofthe B27 molecule'? and, although its natu re and fun ction is poorly understood, cell surface expressed homodimers can be recognized for some leukocyte receptors, such as LILRA 1. LILRB 2. KIR3DLl and KIR3DL2. 60•6 1 The mechanism by which HLA-B27 confer s susceptibility to AS is poorly understood and it has been suggested that a T -cellular response originated by the binding of arthritrogenic peptides of endogenous origin and HLA-B27 molecules might be in the origin of spondiloarthroparhies, As far as we know, some KIR receptors (3D L I and 3DL2) are able to bind B27 in both the classical beta2m /heavy chain (He) and th e bet a2m / free HC homodimers
KIR Genes and Their Role in Spondyloarthropathies
295
(HC-B27) which are independent of the sequence of bound peptide.'? Moreover, recognition ofHLA-B27 bound to self or viral peptides by KIR3DLl has been demonstrated. In the case of some viral peprides, their union to HLA-B*2705 causesthe blockage ofthe NK inhibitory receptor 3DL}.38Therefore, it is probable that a significant fraction of arthritrogenic ligands ofB*27 might avoid their interaction with KIR3DLI in the course ofa microbial infection and influence the activation of activating NK receptors (3DSI and others) and the subsequent progress to AS (Fig. 4A,B). Thus, a genetic imbalance berween activating and inhibitory KIR genes might have an influence on the pathogenesis ofAS by upregulation ofactivation or by the loss ofinhibition or by a combination of both. KIR receptors might use combinations of synergistic receptors to mediate natural cytotoxicity in NK cells and act as costimulatory molecules of TCR. In this model, the presence of KIR3DS 1 or KIR3DLl in combination with HLA-B27genotypes may modulate AS development.
KIR andPsoriatic Arthritis Psoriatic arthritis (PsA) is an inflammatory skin disease associated with psoriasis and is classified among the HLA-B27-associated spondyloarthropathies. It is believed to be caused by a combination of genetic and envirommental factors. Genetic factors have been considered to be important in studies ofboth susceptibility to and the clinical evolution ofPsA and , although the inheritance of psoriasis seems to be polygenic, recent studies suggest that among such, human HLA-Cw6 itself is the true psoriasis gene.62 HLA-CI allotypes (Ser77!Asn80) are ligands for the inhibitory receptors KIR2DL2 and KIR2DL3 and the activator receptor KIR2DS2, whereas HLA-C2 allotypes (Asn77!Lys80) are ligands for KIR2DL1 and KIR2DS I. Severalstudies exist where genetic evidence ofthe influence ofKIR genes in PsA has been shown .63.64 In these studies, it has been hypothesized that susceptibility to PsA is determined by the overall balance ofactivating and inhibitory composite KIR-HLA genotypes. Firstly,it was observed that activating genes KIR2DS 1 and KIR2DS2 associatewith susceptibility to developing PsA, but only when HLA ligands for their homologous inhibitory receptors (KIR2DL1 and KIR2DL2I2DL3) are missing.Absence ofligands for inhibitory KIRs could potentially cause the lack ofan inhibitory signal and subjects who carry these variants might show a lower threshold for NK (and!or T) cell activation and consequently an increase in their cytolytic activity through activating receptors, which contributes to pathogenesis ofPsA. 63 Based on this model, the absence ofligand for inhibitory KIR (HLA-C 1 for KIR2DL2I2DL3 and HLA-C2 for KIR2DLl ) involves a lack of inhibitory signal, which confer susceptibility to PsA.64 There are different intermediate combinations and susceptibility to PsA development could be determined by the balance ofactivating and inhibitory KIR-HLA genotypes, where genotypes will be classified by their ability to confer activation (susceptibility) or inhibition (protection) (Fig. 4C,D).
Conclusions The molecular basis of the association of HLA-B27 with SpA is unknown and their linkage with AS, the strongest association between an HLA class I molecule and disease, has not been resolved as yet. Several studies indicate that HLA-B27 contributes only about 16% of the total genetic risk for AS. Therefore, other genes can also be involved in SpA susceptibility. Several genetic studies have associated KIR variation with disease. In these, the activating and inhibitory signals generated by KIR receptors are associated with susceptibility or protection against a variety ofdiseases. This model is based on KIR genotypes and KIR: HLA combinations and this is the case too of some SpA, such as AS or PsA. These genotypes provide different levels ofinhibition and activation and they are associated with protection or susceptibility with specific KIR genes, or specific KIR and HLA combinations. Another important aspect might be KIR polymorphism, because it can modulate NK cell effector function. For example, 3DLl allotypes differ in their capacity to inhibit NK cells function.
296
Molecular Mechanisms ofSpondyloartbropatbies
With respect to AS. the effects ofcognate 3DLI-B27 (Bw4) interactions could be modulated by KIR3DLI polymorphism. although there are two additional factors which can influence the 3DL1-Bw4 interaction. namely, the different variants ofBw4 epitope. Although highly conserved, there is some limited variability within the Bw4 epitope and there are four defined variants." Some ofthe polymorphic residues that define these epitopes contribute to formation ofthe F-pocket of HLA-B27 and this diversity has the potential to influence NK cell recognition through KIR3DLl and the peptide binding specificity.Severalstudies have been found where it has been demonstrated how the binding of KIR3DLito Bw4ligands depends upon the bound peptide as well as HLA and KIR3DLI polyrnorphism.w" Moreover, it has been suggested that the presence ofdifferent HLA-B alleles may also increase susceptibility to AS in B27-negative patients. 67•68 For example, B*1403, an allotype restricted to Sub-saharan populations which has been associated to AS among B27-negative individual in Togo and Zambia/" To clarify the different association ofHLA-BI4 and other HLA-B alleles to AS, it is necessary to investigate other aspects among subtypes, for example, their interaction with the Peptide loading complex (PLC), their tendency to misfolding or their capacity to form homodimers. It is known that B*14 alleles carry a B-pocket consensus motif ofB*27 that includes Cys 67. Thus, these alleles could originate beta2m/free heavy chain homodimers dependent on the presence ofCys67, which could be recognized for some leukocyte receptors, such as members ofthe KIR family.Moreover, it has been found that KIR3DL2 expression is increased in NK and CD4+ T-cells in spondylarthritis patients compared to controls.?" This receptor only recognizes HLA-B27 homodimers, suggesting an important role of homodimers in the pathogenesis to disease. Hereby, the presence ofKIR combinations in concert with HLA may modulate SpA development. SpA susceptibility could be determined by the overall balance between activating and inhibitory combined KIR- HLA genotypes , bearing in mind KIR and HLA polymorphism as well as the bound peptide.
Acknowledgments This work was supported by grant SAF2004/02669 from the Ministry of Science and Technology (Spain).
References 1. French RA. Yokoyama WM. Natural killer cells and autoimmunity. Arthritis Res Ther 2004 : 6(1):8-14. 2. Ljunggren HG, Karre K. In search of the 'missing self': MHC molecules and NK cell recognition. Imrnunol Today 1990: 11(7) :237-44. 3. Lanier LL. NK cell recognition . Annu Rev Immunol 2005; 23:225-74 . 4. Khakoo SI, Carrington M. KIR and disease: A model system or system of models? Immunol Rev 2006: 214 :186-201. 5. Williams AP, Bateman AR , Khakoo SI. Hanging in the balance. KIR and their role in disease. Mol Interv 2005: 5(4) :226-40. 6. Brewerton DA . H art FD. Nicholls A et al. Ankylosing spondylitis and HL -A 27. Lancet 1973: 1(7809) :904-7. 7. Brewerton DA. Caffrey M. Nicholls A et al. Reiter's di seas e and HL-A 27 . Lancet 1973: 2(7836) :996-8. 8. Schlosstein L. Terasaki Pl, Bluestone Ret al. High association of an HL -A ant igen. W27 , with ankylosing spondylitis. N Engl J Med 1973: 288(14) :704-6. 9. Benjamin R, Parham P. Guilt by association: HLA-B27 and ankylosing spondylitis. Immunol Today 1990; 11(4):137-42. 10. Kollnberger S. Chan A. Sun MY et al. Interaction of HLA-B27 homodimers with KIR3DLl and KIR3DL2, unlike HLA -B27 hererotrimers, is independent of the sequence of bound peptide. Eur J Immunol 2007 ; 37(5):1313-22. 11. Azuz-Lieberrnan N, Markel G, Mizrahi S et al. The involvement ofNK cells in ankylosing spondylitis. Inr Immunol 2005 : 17(7):837-45. 12. Lanier LL, Corliss BC. Wu Jet al. Immunorecepror DAP12 bearing a tyrosine-based activation motif is involved in activating NK cells. Nature 1998; 391(6668) :703-7.
KIR Genesand 1heir Role in Spondyloarthropathies
297
13. Olcese L. Cambiaggi A, Sernenzato G ec al. Human killer cell activatory receptors for MHC class I molecules are included in a multimeric complex expressed by natural killer cells. J Immunol 1997, 158(11):5083-6. 14. Kikuchi-Maki A. Carina TL. Campbell KS. Cutting edge: KIR2DL4 transduces signals into human NK cells through association with the Fc receptor gamma protein. J Immunol 2005 , 174(7):3859-63 . 15. Shilling HG. Guethlein LA, Cheng NW et al. Allelic polymorphism synergizes with variable gene content to individualize human KIR genotype. J Immunol 2002; 168(5) :2307-15. 16. Hsu KC, Chida S. Geraghty DE et al. The killer cell immunoglobulin-like receptor (KIR) genomic region: gene-order. haplotypes and allelic polymorphism . Immunol Rev 2002 ; 190:40-52 . 17. Uhrberg M, Parham P. Wernet P. Definition of gene content for nine common group B haplotypes of the caucasoid population: KIR haplotypes contain between sevenand eleven KIR genes. Immunogenetics 2002; 54(4):221 -9. 18. Norman PJ. Carrington Cv. Byng Met al. Natural killer cell immunoglobulin-like receptor (KIR) locus profiles in african and south asian populations . Genes Immun 2002; 3(2):86-95. 19. Jiang K, Zhu FM, Lv QF er al. D istribut ion of killer cell immunoglobulin-like receptor genes in the chinese han population. T issue Antigens 2005; 65(6) :556-63. 20. Parham P. MHC class 1 molecules and KIRs in human history. health and survival. Nat Rev Immunol 2005; 5(3):201 -14. 21. Yawata M, Yawata N, Moqueen KL ec al. Predominance of group A KIR haplotypes in Japanese associated with diverse NK cell reperto ires of KIR expresion, Immunogenetics 2002 , 54(8) :543-50. 22. Toneva M. Lepage V. Lafay G et al. Genomic diversity of natural killer cell receptor genes in three populations . Tissue Antigens 2001 , 57(4) :358-62. 23. Gardiner CM. Guethlein LA, Shilling HG et al. Different NK cell surface phenotypes defined by the DX9 antibody are due to KIR3DLI gene polymorphism. J Immunol 2001 ; 166(5) :2992-3001. 24. Pando MJ. Gardiner CM . Gleimer M er al. The protein made from a common allele of KIR3DLI (3DLI '004) is poorly expressed at cell surfaces due to substitution at positions 86 in Ig domain 0 and 182 in Ig domain 1. J Immunol 2003 ; 171(12) :6640-9. 25. Yawata M. Yawata N. Draghi M et al. Roles for HLA and KIR polymorphisms in natural killer cell repertoire selection and modulation of effector function . J Exp Med 2006, 203(3):633-45 . 26. Anfossi N, Doisne JM. Peyrat MA et al. Coordinated expressionof Ig-like inhibitory MHC class I receptors and acquisition of cytotoxic function in human CD8+ T-cells. J Immunol 2004, 173(12) :7223-9. 27. Rajagopalan S. Long EO. A human histocompatibility leukocyte antigen (HLA)-G-specific receptor expressed on all natural killer cells. J Exp Med 1999; 189(7):1093-100. 28. Cella M. Longo A. Ferrara GB er al. NK3 -specific natural killer cells are selectively inhibited by Bw4-positive HLA alleles with isoleucine 80. J Exp Med 1994; 180(4):1235-42. 29. Carr WH. Rosen DB, Arase H er al. Cutting edge: KIR3DSl , a gene implicated in resistance to pro gression to AIDS, encodes a DAPl2-associated receptor expressed on NK cells that triggers NK cell activation. J Immunol 2007; 178(2) :647-51. 30. Gumperz JE, Barber LD, Valiance NM et al. Conserved and variable residues within the Bw4 motif of HLA-B make separable contributions to recognition by the NKBI killer cell-inh ibitory receptor. J Immunol 1997; 158(11) :5237 -41. 31. Boyington JC , Motyka SA, Schuck P er al. Crystal Structure of an NK cell immunoglobulin-like receptor in complex with its class I MHC ligand. Nature 2000, 405(6786) :537-43. 32. Fan QR, Long EO. Wiley DC. Crystal structure of the human natural killer cell inhibitory receptor KIR2DLI -HLA-Cw4 complex. Nat Immunol 2001; 2(5):452-60. 33. Rajagopalan S. Long EO. The direct bind ing of a p58 killer cell inhibitory receptOr to human histocompatibility leukocyte antigen (HLA)-Cw4 exhibits peptide selectivity. J Exp Med 1997; 185(8) :1523-8. 34. Zappacosta F, Borrego F, Brooks AG et al. Peptides isolated hom HLA-Cw'0304 confer different degrees of protection hom natural killer cell-mediated lysis. Proc Nat! Acad Sci USA 1997; 94(12) :6313-8. 35. Rudolph MG. Luz JG. Wilson IA. Structural and thermodynamic correlates of T-cell signaling. Annu Rev Biophys Biomol Struct 2002; 31:121-49. 36. Peruzzi M. Wagtmann N. Long EO. A p70 killer cell inhibitory receptor specific for several HLA-B allorypes discriminates among peprides bound to HLA -B'2705 . J Exp Med 1996, 184(4) :1585-90. 37. Peruzzi M. Parker KC. Long EO et al. Peptide sequence requirements for the recognition of HLA -B'2 705 by specific natural killer cells. J Immunol 1996; 157(8):3350-6. 38. Stewart-Jones GB, di Gleria K, Kollnberger S et al. Crystal structures and KIR3DLI recognition of three immunodominant viral pepndes complexed to HLA-B·2705. Eur J Immunol 2005 , 35(2) :341-51. 39. Sims AM. Wordsworth BP. Brown MA. Genetic susceptibility to ankylosing spondylitis. Curr Mol Med 2004,4(1):13-20. 40. Turner MJ. Sowders DP. DeLay ML et al. HLA-B27 misfolding in transgenic rats is associated with activation of the unfolded protein response. J Immunol 2005; 175(4):2438-48.
298
Molecular Mechanisms ofSpondyloarthropathies
41. Gonzalez -Roces S. AlvarezMV, Gonzalez S er al. HLA-B27 polymorphism and worldwide susceptibility to ankylosing spondylitis. Tissue Antigens 1997; 49(2) :116-23. 42. Garda-Fernandez S. Gonzalez S, Mifia Blanco A et al. New insights regarding HLA-B27 diversit y in the asian population. Tissue Antigens 2001; 58(4) :259-62 . 43. Lopez de Castro ]A. Alvarez I. Marcilla M et al. HLA-B27: A registry of constitutive peptide ligands. Tissue Antigens 2004; 63(5) :424-45. 44. L6pez-Larrea C. Sujirachato K. Mehra NK er al. HLA-B27 subtypes in asian patients with ankylosing spondylitis. Evidence for new associations. Tissue Ant igens 1995; 45(3):169-76. 45. Paladini F. Taccari E. Fiorillo MT et al. Distribution ofHLA-B27 subtypes in sardinia and continental Italy and their association with spondylarthropathies. Arthritis Rheum 2005; 52(10):3319-21. 46. Reveille ]K. Ball E]. Khan MA. HLA-B27 and genetic predisposing factors in spondyloarthropathies. Curr Opin Rheumarol 2001; 13(4):265-72. 47. Said-Nahal R. Miceli-Richard C. Gautreau C et al. The role ofHLA genes in familial spondyloarchroparhy: a comprehensive study of 70 multiplex families. Ann Rheum Dis 2002 ; 61(3):201-6. 48. Young Ho Lee. YoungHee Rho. Seong]ae Choi et al. Ankylosing spondylitis susceptibility loci defined by genome-search meta-analysis. ] Hum Genet 2005 ; 50:453-459. 49. Zhang G. Luo ], Bruckel] et al. Genetic studies in familial ankylosing spondylitis susceptibility. Arthritis and Rheumatism 2004; 50:2246-2254. SO. Laval SH . T imms A. Edwards S er al. Whole-genome screening in ankylosing spondylitis: evidence of nonMHC genetic-susceptibility loci. Am] Hum Genet 2001b ; 68:918-926 . 51. Gumperz ]E. Litwin V, Phillips]H et al. The Bw4 public epitope of HLA-B molecules confers reactivity with natural killer cell clones that express NKB 1. a putative HLA recepcor. ] Exp Med 1995; 181(3) :1133-44. 52. Martin MP. Gao X. Lee ]H et al. Epistatic interaction between KIR3DSI and HLA-B delays the progre ssion to AIDS. Nat Genet 2002; 31(4) :429-34 . 53. Gillespie GM . Bashirova A. Dong T et al. Lack of KIR3DSI binding to MHC class I Bw4 tetramers in complex with CD8+ T-cell epitopes. AIDS Res Hum Retroviruses 2007 ; 23(3):451-5. 54. Lopez-Larrea C. Blanco-Gelaz MA. Torre-Alonso ]C et al. Contribution of KIR3DLl/3DSI to ankylosing spondylitis in human leukocyte antigen-B27 caucasian populations. Arthritis Res Ther 2006; 8(4) :RI01. 55. Dlaz-Pefia R. Blanco-Gelaz MA. Suarez-Alvarez B et al. Activating KIR genes are associated with Ankylosing spondylit is in asian populations. In press. Human Immunology. 56. Litwin V, Gurnperz ], Parham P et al. NKBl : A natural killer cell receptor involved in the recognition of polymorphic HLA-B molecules.] Exp Med 1994; 180(2):537-43. 57. Gumperz ]E. Valiance NM. Parham P et al. Heterogeneous phenotypes of expression of the NKBI natural killer cell class I receptor among individuals of different human histocompatibility leukocyte antigens types appear genetically regulated, but not linked to major hisrocomparibililty complex haplotype .] Exp Med 1996; 183(4):1817-27. 58. Martin MP. Qi Y, Gao X et al. Innate partnership of HLA-B and KIR3DLl subtypes against HIV-1. Nat Genet 2007; 39(6):733-40. 59. Appel H . Kuon W. Kuhne M et al. The solvent-inaccessible Cys67 residue ofHLA-B27 contributes to T'-cell recognition of HLA -B27/peptide complexes.] Immunol 2004; 173(11) :6564-73. 60. Allen RL. Raine T, Haude A et al. Leukocyte receptor complex-encoded immunomodulatory receptors show differing specificity for alternative HLA-B27 srructures.} Immunol 2001 ; 167(10):5543-7 . 61. Kollnberger S. Bird L. Sun MY et al. Cell-surface expression and immune receptor recognition of HLA-B27 homodimers. Arthritis Rheum 2002; 46(11) :2972 -82. 62. Nair RP, Stuart PE. Nisto r I er al. Sequence and haplotype analysis supports HLA-C as the psoriasis susceptibility 1 gene. Am] Hum Genet 2006; 78(5) :827-51. 63. Marcin MP. Nelson G, Lee ]H et al. Cutting edge: Susceptibility to psoriatic arthritis : influence of activating killer Ig-like receptor genes in the absence of specific HLA-C alleles. ] Immunol 2002 ; 169(6) :2818-22. 64. Nelson G\V, Marcin MP, Gladman D et al. Cutting edge : Heterozygote advantage in autoimmune disease: hierarchy of protection /susceptibility conferred by HLA and killer Ig-like receptor combinations in psoriatic arthritis.] Immunol 2004; 173(7) :4273-6. 65 Vilches C. de Pablo R. Herrero M] et al. HLA-B73: An atypical HLA-B molecule carrying a Bw6-epitope motif variant and a B pocket identical to HLA-B27. Immunogenetics 1994; 40(2):166. 66. Thananchai H. Gillespie G. Marcin MP et al. Cutting Edge: Allele-specific and peptide-dependent interactions between KIR3DLl and HLA-A and HLA-B .] Immunol 2007 ; 178(1):33-7. 67. Wei ]C. Tsai WC. Lin HS et al. HLA-B60 and B61 are strongly associated with ankylosing spondylitis in HLA-B27-negative Taiwan Chinese patients . Rheumatology (Oxford) 2004; 43(7) :839-42.
KIR Genes and Their Role in Spondyloarthropathies
299
68. Yamaguchi A, Tsuchiya N, Mitsui H er al. Association ofHLA-B39 with HLA-B27-negative ankylosing spondylitis and pauciarticular juvenile rheumatoid arthritis in japanese patients . Evidence for a role of the peptide-anchoring B pocket . Arthritis Rheum 1995; 38(11) :1672-7. 69. Lopez-Larrea C, Mijiyawa M, Gonzalez S et al. Association of ankylosing spondylitis with HLA-B'I403 in a west african population. Arthritis Rheum 2002; 46(11) :2968-71. 70. Chan AT, Kollnberger SD, Wedderburn LR et al. Immunoglobulin-like receptor KIR3DL2 in spondylarthritis. Arthritis Rheum 2005; 52(11):3586-95. 71. Yen]H, Moore BE, Nakajima T et al. Major histocompatibility complex class I-recognizing receptors are disease risk genes in rheumatoid artritis.] Exp Med 2001; 193(10):1159-67. 72. Majorczyk E, Pawlik A, Luszczek W et al. Associations of killer cell immunoglobulin-like receptor genes with complications of rheumatoid arthritis. Genes Immun 2007; 8(8) :678-83. 73. Yen]H, Lin CH, Tsai WC et al. Killer cell immunoglobulin-like receptor gene's repertoire in rheumatoid arthritis. Scand] Rheumarol 2006 ; 35(2) :124-7. 74. Momot T, Koch S, Hunzelmann N er al. Association of killer cell immunoglobulin-like receptors with scleroderma. Arthritis Rheum 2004; 50(5) :1561-5. 75. Takeno M, Shimoyama Y, Kashiwakura] et al. Abnormal killer inhibitory receptor expression on natural killer cells in patients with Behcer's disease. Rheurnatol Int 2004; 24(4) :212-6. 76. Luszczek W, Mariczak M, Cislo Met al. Gene for the activating natural killer cell receptor, KIR2DS1, is associated with susceptibility to psoriasis vulgaris. Hum Immunol 2004 ; 65(7) :758-66. 77. Suzuki Y, Hamamoto Y, Ogasawara Y er al. Genetic polymorphisms of killer cell immunoglobulin-like receptors are associated with susceptibility to psoriasis vulgaris.] Invest Dermatol2004; 122(5):1133-6. 78. van der Slik AR , Koeleman BP, Verduijn W et al. KIR in type 1 diabetes : Disparate distribution of activating and inhibitory natural killer cell receptors in patients versus HLA-matched control subjects. Diabetes 2003 ; 52(10) :2639-42.
CHAPTER 23
Innate Immunity ofSpondyloarthritis: The Role ofToll-Like Receptors Robert D. Inman" Abstract
A
mongst the spondyloarthroparhies (SpA), the relationship ofhost susceptibility and environmental triggers is best seen in reactive arthritis (ReA). There are several points from the clinical studies of ReA which highlight important unresolved issues in the pathogenesis. Innate and adaptive immune responses are distinct, but complementary aspects of host defense (Pacheco-Tena er al, 2002. Beutler er al, 2006) . Innate immunity is critically dependent upon non variant , genetically encoded receptors for highly conserved structures known aspathogen-associated molecular patterns (PAMP). The receptors themselves are termed pattern-recognition receptors (PRR). Among the PRR. toll-like receptors (TLR) playa central role.
Introduction Although the adaptive immune response may playa central role in ReA. to date there is no distinctive immunologic pattern that sheds light on pathogenesis. A study comparing antibody profiles in patients with uncomplicated Yersiniainfection with patients having Yersinia-associated arthritis found that serwn antibodies against the organism remained elevated 4 yearsafter the initial infection and could not differentiate against the two patient groups.' In the study ofWilkinson er al2 there was no correlation between the presence of chlamydial DNA in the joint of ReA patients and a specific cellular or humoral immune response against the organism. Secondly, clinical stu dies in ReA have not included studies in the inductive phase of the arthritis, because of the impracticability ofstudying patients in the preclinical phase ofthe disease. Yet it is this early host response to the infection, innate immunity. which may critically dictate the subsequent clinical course ofevents. Finally, while HLA-B27 does constitute a genetic risk factor. population studies have determined that at least 5 genes in addition to B27 are playing a role in SpA susceptibility. But the nature ofmost ofthese other genes and their potential role in host: pathogen interaction has not yet been defined. These important questions should be seen in the context of the imperative to explain the chronic inflammation which is the hallmark of ReA. Natural history studies have emphasized how chronic this arthritis can be. Follow-up studies have documented the significant morbidity of ReA at a 5-year interval after the inciting infection.' In th at study, 18 of the original 27 patients continued to have symptoms 5 years after onset. The early events after such an infection may well hold the key to defining these immune sequelae to the infection. Despite the fact that the B27 association is still,supported, the mechanism ofits role in ReA has proved quite resistant to a number of investigations.
"Rob ert D. Inman-University of Toronto, TorontoWestern Hospital, Fell Pavilion, 1 E-423, 399 Bathurst St, Toronto, Ontario, Canada MST258. Email: robert.inman @uhn.on.ca
Molecular Mechanisms ofSpondyloarthropathies. edited by Carlos Lopez-Larrea and Roberto Dfaz-Pefia. ©2009 Landes Bioscience and Springer Science+Business Media.
Innate Immunity ofSpondyloarthritis: The Role ofToll-Like Receptors
301
Current Concepts in Innate Immunity Innate and adaptive immune responses are distinct, but complementary aspects of host defense.4,5 Innate immunity is critically dependent upon nonvariant, genetically encoded receptors for highly conserved structures known as pathogen-associated molecular patterns (PAMP). The receptors themselves are termed pattern-recognition receptors (PRR). Adaptive immunity relies on rearrangement and clonal selection as its fundamental principles, allowing an expansion of antigen-specific effector cells. But the development of an adaptive immune response is not immediate and the processes ofaffinity maturation, clonal selection and clonal expansion take time . The ability of preformed PRR to interact with bacterial PAMP immediately allows a prompt first-line of defense that can determine the final course of the infection, not only in terms of dissemination of the pathogen but also the very survival of the host, as reflected in septic shock. But there is another way that innate immunity figures prominently in the outcome of infection, namely the generation of an effective adaptive immune response. One critical link between the two systems the up-regulation ofcostimulatory molecules occasioned by the PAMP-PRR interaction . Activation of lymphocytes entails both signal one (peptide-MHC-TCR interactions) and signal two (up-regulation of costimulatory molecules), in the absence of which anergy, rather than activation and expansion, can be the consequence ofthe antigen-presenting cell-lymphocyte encounter. Thus both infectious and immune-mediated sequelae ofa pathogen challenge involve innate immunity in a critical way. Among the PRR, toll -like receptors (TLR) playa central role.The TLR family includes at least 11 members in humans, which share a common structure with a variable number ofextra-cellular leucine-rich repeat domains paired with an intracellular signaling domain which shares a high degree ofhomology with the receptors for IL-l and IL-18. 6,7 They also share with these receptors the same signaling cascade, which includes association with MyD88,8.9activation ofthe IL-l receptor activation kinase (IRAK) and subsequent nuclear translocation ofNF-kB via the release ofIk-B. Activation of the TLR signaling pathway results in dendritic cell maturation with production of a range of cytokines.'?'!' TLR are expressed on a variety ofcell types and amongst leukocytes, predominate in immature stages of dendritic cells and monocytes. The distribution in different cells is distinctive for each TLR. It is of interest to note that the polymorphisms that exist in human populations may reflect a dynamic interaction with infectious environmental pressure." Polymorphisms in TLR4 (Asp299Gly and Thr399lle). which are known to be related to susceptibility to Gram-negative infections , have recently been shown to have unique distributions in populations from Africa, Asia and Europe. Genetic and functional studies suggest that the nonsynonymous polymorphism Asp299Gly has evolved as a protective allele against malaria, explaining its high prevalence in sub-Saharan Africa . However, the same allele could have been disadvantageous after migration of modern humans into Eurasia , putatively because ofincreased susceptibility to severe bacterial infections . In contrast, the Asp299Gly allele, when present in cosegregation with Thr399lle to form the Asp299Gly/Thr399lle haplotype, shows selective neutrality. Polymorphisms in TLR4 thus reflect an interaction between innate immunity and the infectious pressures in particular environments which may have shaped the genetic variations of the immune system during the out-of-Africa migration of modern humans . A recent study has provided important insight into the biological role ofTLR4 modulation in host immunity. 13 The activation ofTLR signaling by microbial signatures, critical to the induction of immune responses, demands tight regulation. RP 105 is a TLR homolog lacking a signaling domain and its expression directly mirrors that ofTLR4 on antigen-presenting cells. RP 105 is a specific inhibitor ofTLR4 signaling in HEK 293 cells, a function conferred by its extracellular domain. Notably, RP 105 and its helper molecule, MD-!, interact directly with the TLR4 signaling complex, inhibiting its ability to bind microbial ligand. TLR4 signaling in dendritic cells as well as endotoxin responses in vivo are regulated by RPIOS. Low-dose intraperitonealE. coli LPS produced more TNF in RPlOS -/- mice than in wild type controls. These results identify RPlOS as a physiological negati ve regulator ofTLR4 responses.
302
Molecular Mechanisms ofSpondyloarthropathies
In addition to the key roles ofaugmenting pro-inflammatory cytokine generation and ofbridging to adaptive immunity by up regulation ofcostimulatory molecules, there is recent evidence that TLR2 signaling plays an important role in modulating T regulatory cells.'? It had been observed that the CD4CD25 Tregsubset in TLR2-/- mice to be significantly reduced in number compared with WT littermate control mice, indicating a link between Tregs and TLR2. It was recentlyshown that the TLR2ligand Pam3Cys , but not LPS (TLR4ligand) or CpG (TLR9ligand), directly acts on purified Treg in a MyD88-dependent fashion. When combined with TCR stimulation, TLR2 triggering augmented Treg proliferation in vitro and in vivo and resulted in a temporal loss of the suppressive Treg phenotype in vitro by directly affecting Tregs. WT Tregs adoptively transferred into TLRZ-/- mice were neutralized by systemic administration ofTLRZligand during the acute phase of a Candida albicans infection, resulting in a IOO-fold reduced C. albicans outgrowth. This study provided evidence that in vivo TLR2 also controls the function ofTregs. This may have important implications in reactive arthritis, in which the critical variable appears to be the adequacy ofhost clearance of the pathogen and Treg modulation of this variable could playa fundamental role in this process.
Innate Immunity in Host Response to Arthritogenic Bacteria The recognition that there are critical host determinants operational as the first line ofdefense against pathogens and that these determinants precede adaptive immunity represents a major advance in our understanding ofhost-microbial interactions. Adaptive immunity is distinguished by the heterogeneity ofligand-receptor interactions and the specificity accorded by this spectrum and the memory created by it. For all its elegant specificity and recall potency, the adaptive immune response represents a temporally sluggish response to the urgent invasion of a pathogen. Such immediate defense calls for a predefined strategy that can respond immediately to the challenge. This is the essence of the innate immune response. The great advantage of this prefabricated defense strategy for the host is the immediacy of the response. As discussed above, innate immunity also becomes a critical gatekeeper for both appropriate and inappropriate immune responses by upregulation of costimulatory molecules. The pathogenesis ofspondyloarthritis has been thought to represent a dynamic interaction of genetically defined host susceptibility and an environmental trigger. ReA represents a paradigm of this construct since in this entity an infection can trigger a chronic nonseptic synovitis. A number ofinvestigators have been active in developing experimental systems with which to examine host: pathogen interactions more rigorously than is achievable in the clinical setting. This isparticularly relevant to examining innate immunity of the host, since these events which set the stage for the subsequent development of adaptive immunity occur early in the course of a challenge by an arthritogenic pathogen. One such experimental system has examined innate immune elements in Chlamydia-induced arthritis. IS In this study the role ofneutrophils was examined in experimental arthritis in mice with targeted elimination ofthe small GTPases Rac I and Rac2 as well as the role ofTLR. Arthritis was induced by intraarticular inoculation ofsynoviocyte-packaged Chldmydia tracbomatis. The microbiology, histopathology and course of this model have been shown to recapitulate the essential features of clinical Chlamydia-induced arthritis in humans." In the acute phase, wild-type mice developed more severe arthritis than did Rac-deficienr mice, with abundant infiltration of neutrophils into the joint. In the chronic phase, the Rae-deficient mice developed more severe arthritis and demonstrated defective clearance of the pathogen from the joint. In vitro stimulation of neutrophils with Cblamydi« up-regulated the expression ofTLR4, but not TLR2, in wild-type mice. However, neutrophils from Rae-deficient mice did not show this up-regulation ofTLR-4. Sustained TLR-4 expression in neutrophils was found to be dependent on the expression ofRac. Mice genetically deficient in TLR4 demonstrated more severe arthritis than did the controls. Thus, Rae expression plays a profound role in infection-triggered arthritis and demonstrates a bimodal influence on the disease process , exacerbating acute joint inflammation but controlling chronic arthritis. This was the first demonstration that the course of reactive arthritis is defined by innate immune responses to the inciting pathogen.
Innate Immunity ofSpondyloarthritis: The Role ofToll-Like Receptors
303
Experimental ReA has provided insight into the early events of host response and its impact on the course of arthritis. I? In the model of Chlamydia-induced arthritis, Lewis (LEW) rats are susceptible, whereas Brown Norway (BN) rats are relatively resistant. Significant differences in the histopathologic severity ofarthritis were originally observed at day 21 and this prompted an examination of the acute phase of the arthritis. As early as day 5 after the onset of the arthritis LEW showed more severehistopathology scores than BN . Evaluatingthe role ofcomplement with cobra venom factor treatment excluded complement asbeing the key to differential sensitivity since decornplernentarion did not eliminate the LEW-BN differencesin arthritis severity. Host clearance on the other hand demonstrated significant differences between the rat strains, with BN showing more prompt and effective clearance of the pathogen from both synovial tissues and spleen than LEW. Local cyrokine profiles demonstrated that host resistance was characterized by enhanced synovial expression ofTNF-a, IFN-y and IL-4.These studies demonstrate that cytokines thought to be pro-inflammatory in nature can play an important role in host defense in infection-triggered arthritis and serve to highlight the dynamic cytokine interactions which constitute effective host:pathogen interactions. The rapid production ofTNF-a in the host response to an infectious challenge is a product ofinnate immunity and the experimental studies provide support to further analysis of the early events in clinical ReA. An alternative model for ReA utilized mice treated with cell wall extract from Escherichia coli (ECW) by injection into the footpads." Seven days later, the mice were an intra-articular injection with LPS. ECW-primed mice manifested acute severearthritis after inrraarticular challenge with ECW or LPS, while unprimed mice exhibited modest changes after these challenges. Mutant mice lacking functional TLR-4 or myeloid differentiation factor 88 (MyD88), an adaptor molecule of TLR-4 signaling, were resistant to this arthritis. Although both TNF-a and IL-6 were equally expressed in the joint after LPS challenge, IL6-/- mice, but not TNF-/- mice, were resistant to ECW/LPS-induced arthritis. These results indicate the importance of priming with ECW and the requirement ofTLR-4/MyD88-mediated IL-6, but not TNF-a for the develop ment of ECW/LPS-induced arthritis. LPS-induced IL-6, in the absence of TNF-a, mediates LPS-induced arthritis. These findings would suggest that IL-6 could be a candidate target in novel treatment regimens in ReA. TLR may also playa keyrole in experimental arthritis that isnot initiated by an infectious trigger. Recently, treatment of murine collagen-induced arthritis using a TLR4 antagonist was shown to substantially suppressed both clinical and histologic characteristics of arthritis without influencingadaptive anti-Type II collagen immunity,' ? Treatment with TLR4 antagonist strongly reduced IL-l beta expression in articular chondrocytes and synovialtissueand also inhibited IL-l-mediated autoimmune arthritis in IL-lRa-/- mice. Thus inhibition ofTLR4 suppresses the severity of experimental arthritis and decrease IL-l expression in arthritic joints . Immune complex-mediated arthritis (ICA) may also implicate TLR Signaling as a contributory factor as demonstrated by van Lent er aFoWhen ICA was induced in knee joints of TLR2- / - and TLR4-/ - mice and their wild-Type controls it was observed that joint inflammation and cartilage destruction were not different in arthritic TLR2-/ - and wild-type mice. Bycontrast, at day 1 after ICA induction, joint swelling and proteoglycan depletion in knee joints ofTLR4-/- mice were considerably lower when compared with wild-type controls. Cytokine production (IL-l, IL-6, MIP-la) at this time point was markedly reduced in TLR4-/- mice. In the synovial tissues ofTLR4-/- mice and also after injection ofPLL lysozyme alone, mRNA levelsofFcyR and IUO were considerably lower. Stimulation ofperitoneal macrophages with PLL lysozymeup regulated mRNA levelsofFcyRand IL 10, whereas neutralization by anti -IL-l 0 antibodies largely blocked FeyR up regulation. Thus TLR4 regulates early onset ofjoint inflammation and cartilage destruction during ICA arthritis by upregulation ofFeyR expression and this is mediated primarily by IL-l O. Innate immunity may implicate receptors other than TLR. One recent experimental approach to direct host: pathogen interaction has taken advantage ofrecent genetic findings that playa role in chronic inflammatory joint disease." Blau Syndrome (BS) is a familial granulomatous disease manifested by uveitis, arthritis and skin rash. BS has recently been found to be associated with
304
Mole cular Mechanisms ofSpondyloarthropathies
a distinctive mutation in NOD2, which encodes an intracellular receptor. We have compared host cell interaction with bacterial challenge in U937 cells expressing wild type human NOD2 (NOD2w, ) ' mutant NOD2 (NOD2 B1au) , or a vector control. The cells were incubated with s. typhimurium. Invasion of target cells with S. typhimurium was diminished in the presence of NOD2B1au• Expression ofTNF-a mRNA was enhanced following bacterial invasion in all cell lines but NOD2B1au was associated with a more rapid decline in TNF-a expression. Kinetics of intracellular clearance of bacteria indicated a relative defect in NOD2 B1au compared to controls. This clearance defect maybe related to the lack ofsustained TNF-a seen in the early stages. These findings indicate that the NOD2 mutation associated with this syndrome alters host:microbial interaction and this may have relevance to triggering factors in the ocular and joint inflammation seen in BS.
Innate Immunity and Spondyloarthritis A number ofstudies haveimplicated altered TLR expression in RA synovial tissues.22•23 But the pattern ofTLR expression may be distinctive in SpA. TLR2 and TLR4 are unregulated in circulating CD163+ macrophages. In the study of Baeten et a124 higher levels ofTLR2 and TLR4 were seen in SpA synovial tissuesthan in RA synovial tissues and this upregulation corresponded to an expanded CD 163+ macrophage population in the SpA synovium . Indeed, the immunopathology ofSpA, with a predominance ofmacrophages and neutrophils contrasts with that ofRA, with the presence oflymphoid aggregates, CD4+ T-cells and anti-CCP antibody production, suggests that SpA is a disease mediated and sustained by innate immune responses, whereas RA is mediated and sustained by adaptive immune responses. Macrophages expressing CD163 are known to be increased in both the gut and the joints of SpA patients and soluble CD 163 may inhibit R cell activation in the joint.25026 While these studies have emphasized the importance of macrophage in SpA it should be recognized that macrophages are thought to be the major source ofinflammation mediators in RA. The recent study ofHuang et al2? analyzed macrophages obtained from the joints ofRA patients in order to characterize the expression ofTLR-2 and TLR-4 and the responses to TLR ligation. The expression ofTLR2 and TLR4 was increased on CD 14+ macrophages from the joints ofRA patients compared with that on control in vitro-differentiated macrophages or control PBMC. Neither TLR2 expression nor TLR4 expression differed between RA and SpA. However, PG- and LPS-induced TNF-a expression and IL-8 expression were greater with RA SF macrophages than with those obtained from SpA or with control macrophages. PG-induced TNF-a expression and IL-8 expression were highly correlated with TLR2 expression in normal macrophages, but not with that in macrophages obtained from joints ofRA or SpA patients. Thus this study contrasts with the studies cited above in that TLR2 and TLR4ligation resulted in increased activation of RA synovial macrophages compared with SpA. Factors other than the level ofTLR2 and TLR4 expression contributed to the increased activation of RA SF macrophages. These observations support the notion of a potential role for activation through TLR2 and TLR4 in the inflammation and joint destruction ofRA. A recent study examined the effect ofTNF-inhibitor treatment on these parameters. Expression ofTLR4, but not TLR2, was increased on peripheral blood mononuclear cells in SpA patients, whereas both TLR4 and TLR2 showed enhanced expression in RA patients. TLR expression was particularly noted in the CD163+ macrophages. TNF-inhibitor therapy down-regulated expression ofTLR-2 and TLR-4 monocyres and was associated with an impairment ofTNF-a generation after LPS stimulation. Expression ofCD 163 and both TLR2 and TLR4 were increased in synovial tissues to a greater extent in SpA than in RA and this local TLR expression in tissues was down-regulated with TNF-inhibitor rrearrnenr ." These alterations ofTLR expression after such treatment have suggested that TNFi treatment might impa ir innate immune responses to pathogens and result in enhanced susceptibility to infections as a result. Netea et al29 reported two patients with RA who developed Salmonella septicemia during TNFi treatment. The se investigators evaluated an additional 6 RA patients receiving TNFi, 5 control RA patients and 6
Innate Immunity ofSpondyloarthritis: The Role ofToll-Like Receptors
305
healthy volunteers. Stimulation ofPBMC with heat-killed Salmonella typhimurium or Candida albicans led to a significantly decreased production ofIFN-y, but not to a decreased production ofIL-l 0, IL-~, or IL-6, in anti-TNF-treated RA patients compared with controls. TNF-blocking treatment ex vivo significantly inhibited TLR4 expression on dendritic cells from RA patients and healthy controls. These studies suggested that the increased susceptibility to infections in patients receiving TNFi therapy might be mediated in part by interference with TLR signaling and IFN-y production, since this pathway figure prominently in host defense. In theory, this might be ofparticular concern in reactive arthritis in which an arthritogenic pathogen isimplicated. F1agget al30 recently reported on the efficacy and safety ofetanercept in 16 patients with undifferentiated or reactive arthritis. Of 16 patients, 10 completed the 6-month open-label trial. Six patients withdrew, but none had a worsening ofarthritis or infection. Ofthe 10 completers, 9 could be classified as treatment responders, despite the evidence ofbacterial organisms on PCR analysis prior to initiating etanercept in 3 patients; 2 patients became PCR negative on etanercept. Five of 6 patients with adequate synovial biopsy specimens showed improvement, but not normalization of histology. Thus ifTNFi treatment is associated with decrease TLR expression and function in the synovial tissues,it did not seem to result in any untoward events, even in patients with evidence ofmicrobial antigens persisting locally. Upregulation of TLR2 on antigen-presenting cells has recently been observed in patients with psoriatic arthritis." In this study, it was specifically the immature dendritic cell population, not mature dendritic cells or monocytes, which should altered TLR2 expresssion and this did not obtain for TLR4 expression. The cytokine profile of these cells reflected a Thl pattern, with predominance ofTNF-a, IFN-y and IL-2. These findings may provide some important links to skin inAammation in PsA. In psoriatic skin lesions there is an enhancement ofTLR2 expression on keratinocytcs in the upper epidermis whereas TLRS is down-regulated in the basal keratinocyres." TLR2 expression has also been observed on dermal dendritic cells in conjunction with TLR2 expression by epidermal and dermal dendritic cells." Recently it has been reported that CD4+CD28null T-cells in AS, PsA and RA express TLR recognizing bacterial LPS.34 In AS , TLR4 and to a smaller extent TLR-2, were expressed on CD4+CD28null T-cells, whereas expression was negligible on CD4+CD28+ and CD8+ T-cells. CD4+CD28null T -cells produced perforin upon stimulation with lipopolysaccharide and this effect was enhanced by autologous serum or recombinant soluble CD 14. Incubation of peripheral blood mononuclear cells with TNF-a led to an upregulation ofTLR4 and TLR2 on CD4+CD28null T-cells in vitro and treatment of patients with TNFi resulted in decreased expression ofTLR4 and TLR2 on CD4+CD28null T-cells in vivo. This study thus identifies a pathway for direct activation of cytotoxic CD4+ T-cells by components of infectious pathogens and supports the hypothesis that CD4+CD28null T-cells represent a link between the innate immune system and the adaptive immune system. There have been differing reports in the role ofTLR as genetic susceptibility factors. There was no association observed between the TLR4variants Asp299Gly and Thr399lle and AS in the study ofAdam et al35 nor with TLR4 A896G in the study ofvan der Paardt et al36 In contrast, a recent study from Newfoundland did observe a relationship between AS and TLR4 polymorphismsY The rationale for this study derived from two observations. Functional SNPs within the ectoplasmic domain of the TLR4 gene have been shown to result in an endotoxin hyporesponsive phenotype and aberrant signal transduction for bacterial agonises. Secondly,TLR4 is in proximity to a genome wide linkage peak in 9q32-33. Given the proposed function and location ofTLR4, the investigators set out to examine the association oftwo functional variants ofTLR4 in the Newfoundland ankylosing spondylitis (AS) population. All patients with AS met the modified New York Criteria and the controls were ethnically matched. Subjects were assessed using a standardized protocol, which included measures of disease activity (BASDI) and function (BASFI). In total, 101 AS patients and 100 ethnically matched controls were genotyped for two functional variants in the TLR4 gene: Asp299Gly (A/G polymorphism) and 1hr399lle (CIT polymorphism). The minor
306
M olecular Mechanisms ofSpondyloarthropathies
allele frequency for the Asp299Gly variant (G) was significantly higher in AS cases as compared to controls (7.5% vs 2.6 % respectively; OR = 3.1, P = 0.037). The minor allele frequency for the Thr399Ile variant (T) for the cases and controls was 7.5% vs 3.0% respectively, OR = 2.59; p = 0.07. Haplotype analysis using Haploview noted a higher proportion of GT in the cases (GT, X2 P = 0.023). Regression analyses among AS patients noted no association between age of onset, BASDAI, BASFI for the genot ype/haplotype indicator variables. Thus in this study study, a modest association was noted for the functional variants ofTLR4 and AS. G iven the critical role ofTLR4 variants in the innate immune system, larger studies are warranted to better delineate the association ofTLR4 variants in AS. These differences may reflect differing genetic heterogeneity in cases and controls, since Newfoundland represents a relatively homogeneous population genetically. Studies in clinical ReA have been hampered by the identification ofwell-defined cohorts sharing the same antigenic trigger. Most knowledge about the course ofReA is obtained through analysis ofsporadic cases in clinic populations. The occurrence ofa point source outbreak affords a unique opportunity to address genetic interaction with a common environmental trigger. Earlier studies in epidemic ReA had suggested that the distinctive immune profile ofReA was a diminished adaptive immune response to the organism." In light ofcurrent understanding ofinnate immunity playing a defining role for subsequent adaptive immunity, this earlier observation in retrospect might have reflected altered innate immune responses in ReA, but I 'LR had not been identified at that time. In a more recent outbreak of Salmonellosis, a cohort of ReA patients was identified by a screeningquestionnaire and DNA was isolated from saliva samples mailed in by the ReA patients and a large number ofcontrols. There was a significant relationship between TLR2 variants in the ReA cohort, whereas no such relationship was observed for TLR4 polymorphisrns.VThe Arg7S3Gln variant associated with ReA has recently been shown to have functional relevance in a NFkB translocation study using transfected cells.40 This candidate SNP has also shown an association with acute rheumatic fever in children." It is also of interest to note that TLR2 polymorphisms bear a relationship to serum antimicrobial profiles in Crohn's disease.P The development of antibodies reflects a loss of tolerance to intestinal bacteria that underlies Crohns disease, resulting in an exaggerated adaptive immune response to these bacteria. It has been hypothesized that the development ofantimicrobial antibodies (anti-Saccharomyces cerevisiae antibodies (gASCA) IgG, antilaminaribioside antibodies (ALCA) IgG, antichirobioside antibodies (ACCA) IgA, antimannobioside antibodies (AMCA) IgG and outer membrane porin (Omp) IgA) is influenced by the presence ofgenetic variants in pattern recognition receptor genes. When compared with Crohns disease patients without CARDIS mutations, the presence of at least one CARD IS variant in Crohn'sdisease patients more frequently led to gASCA positivity and ALCA positivity and higher gASCA titers , independent of ileal involvement. A gene do sage effect , with increasing gASCA and ALCA positivity for patients carrying none, one and two CARD 15 variants, respectively, was seen for both marker s. An opposite effect was observed for the TLR4 D299G and TLR2 P631 H variants, with a lower prevalence of ACCA antibodies and Omp antibodies. Thus variants in innate immune receptor genes were found to influence antibody formation against microbial epitopes. This suggests that integrity ofgut mucosa and its alteration by inflammation secondary to infection or inflammatory bowel di sease, may figure prominently in SpA. Another important link to the gut in the pathogenesis ofSpA derives from recent genetics studies implicating polymorphisms in the IL23R in AS. One large analysis included AS patients from the US and the UK.43This study implicated a polymorphism that had previously been implicated in CD. A Canadian study ofthree di stinct AS populations also demonstrated a disease association with the IL-23Rlocus and impli cated the same polymorphism (Arg381 Gin) associated with IBD and psorlasis." A haplotype containing the primary SNP ofinterest in IBD and psoriasis was also noted to be protective in two ofthe three Canadian AS populations. Thi s suggests that IL-23R is a pleiotropic gene that is relevant to multiple autoimmune diseases. An int eresting extension of the se genetic s studies is the analysis ofIL-23 expression in the gut. Ci ccia et al45 recently analyzed gut tissue from SpA patients without clinical gut involvement and contrast them with Crohn's d isease and normal controls. Subclinical inflammation was present in all th e SpA pat ient s. IL-23
Innate Immunity ofSpondyloartbritis: TheRole a/Toll-Like Receptors
307
was strongly over-expressed in AS, with levels in gut tissue comparable to the Crohn's patients. 3, an inhibitor ofIL-23 was only increased in CD and not in AS, suggesting a dysregulation ofIL-23. These developments in genetics and genomics add support to the longstanding notion that there is a subclinical gut inAammation in SpA patients and that gut flora and host mucosal immunology may interact to set the stage fur the development ofSpA. As the role ofTLR continues to attract attention from investigators in IBD, there no doubt will arise important new insights into the pathogenesis ofSpA.46 The other rapidly developing field is in the genetic basis of immun e response to infections"? There is increasing recognition of mutations in human genes associated with conventional primary immunodeficiencies, which confer Mendelian predisposition to multiple infectious diseases. Recently, there has been much study of monogenic traits that do not confer such a broad vulnerability. Defects in several genes confer predisposition to infection with specific bacteria and viruses in otherwise healthy individuals. Mutations in other genes even confer resistance to specific pathogens, with no detectable decrease in fitness. It is instructive to note that there are other clinical complications in common point source epidemics, as occurred when a large number ofindividuals were exposed to Legionellapneumopbila at a flowershow. Hawn et al48 examined the role ofTLR in this event . These investigators had demonstrated recently that TLRS recognizes Hagellin, a potent inflammatory stimulus present in the flagellar structure ofmany bacteria. They then showed that a common stop codon polymorphism in the ligand-binding domain ofTLRS (TLRS392STOP)was unable to mediate flagellin signaling, acts in a dominant fashion . This TLR$ mutation was associated with susceptibility to pneumonia caused by Legionellapneumopbila. They alsodemonstrated that Aagellinwas a principal stimulant ofproinAammatory cytokine production in lung epithelial cells. Together, these observations suggest that TLRS392STOP increases human susceptibility to infection through an unusual dominant mechanism that compromises TLRS's essential role as a regulator of the lung epithelial innate immune response. Such approaches will be productive in the future study ofepidemic ReA.
sacs-
Conclusions The current genome-wide scans being carried out in SpA studies are detecting new candidate genes which are highly significant statistically, but which confer a relatively low odds ratio respectively. Thus the attributable risk, beyond B27, is likely to arise from a distinct signature ofgenetic polymorphisms and genes which modulate host immune response to environmental factor s, such as TLR, will constitute an important group for attention.
References 1. Granfors K, Marilahri Palo R, Luukkainen R et al. Persistence of yersinia antigens in peripheral blood cells from patients with yersinia cntcrocolitica 0:3 infection with or without reactive arthritis. Arthritis Rheum 1998; 41 :855-862. 2. Wilkinson NZ, Kingsley GH, Sieper J er al. Lack of correlation between detection of chlamydia trachomatis DNA in synovial fluid from patients with a range of rheumatic diseases and the presence of an antichlamydial immune respon se. Arthritis Rheum 1999; 41:845-854. 3. Thomson GTD, DeRubeis DA , Hodge MA et al. Post-salmonella reactive arthritis. Late sequelae in a point source cohort. Am J Med 1995; 98:13-21. 4. Pacheco Tena C, Zhang X, Stone M et al. Innate immunity in host -microbial inte ractions : beyond B27 in the spondyloarthropathies. C urr Opin Rheurnarol 2002 ; 14:373-82. 5. Beutler B, Jiang Z, Georgel P et al. Genetic analysis of host resistance : Toll-like receptor signaling and immunity at large. Ann Rev Immunol 2006 ; 24:353-389 . 6. Kaisho T, Akira S. Dendritic cell function in toll-like receptor and MyD88-knockout mice. Trend s Immunol 2001 ; 22 :78-83. 7. Means TK, Wang S, Lien E er al. Human toll-like receptors mediate cellular activation by mycobacterium tub erculo sis. J Immunol 1999 ; 163:3920-3927. 8. Akira S, Hoshino K, Kaisho T. The role of toll-like receptors and myD88 in innate immune responses. J Endotoxin Res 2000; 6:383-387. 9. Fitzgerald KA, Paisson-McDermott EM et al. Mal (myD88 -adapter-like) is required for TLR4 signal transd uction. Nature 2001 ; 413 :78-83 .
308
Molecular M echanisms ofSpondyloarthropathies
10. Thorna-Uszynski S. Kiertscher SM. Ochoa MT et al. Activation of TLR2 on human dendritic cells triggers induction of IL-12. but not IL-I0.] Immunol 2000: 165:3804-3810. 11. Tsuji S. Matsumoto M. Takeuchi 0 et al. Maturation of human dendritic cells by cell wall skeleton of mycobacterium BCG : involvement of toll-like receptors. Infect Immun 2000; 68:6883-6890. 12. Ferwerda B. McCall MB. Alonso S et al. TLR4 polyrnorphisms, infectious diseases and evolutionary pressure during migration of modern humans. Proc Nad Acad Sci USA 2007 ; 104:16645-50. 13. Divanovic S. Trornperte A. Atabani SF et al. Negativ e regulation of toll -like recepto r 4 signaling by the toll-like receptor homolog RP105 . Nat Immunol 2005; 6:571-578. 14. Sutmuller RP, den Brok MH, Kramer M et al. Toll-like receptor 2 controls expansion and function of regulatory T-cells.] Clin Invest 2006 : 116:485-94. 15. Zhang X, Glogauer M. Zhu F et al. Innate immunity and arthritis: neutrophil rae and toll-like receptor 4 expression define outcomes in infection-triggered arthritis. Arthritis Rheum 2005 ; 52:1297-304. 16. Inman RD. Chiu B. Synoviocyte-packaged chlamydia trachomatis induces a chronic aseptic arthritis. ] Clin Invest 1998; 102:1776-1782. 17. Inman RD. Chiu B. Cytokine profiles in the joint define pathogen clearance and severity in chlamydia-induced arthritis. Arthritis Rheum 2006; 54:499-507. 18. Kyo F. Futani H. Matsui K ec al. Endogenous incerleukin-6, but not tumor necrosis factor alpha. contributes to the development of toll-like receptor 4/ myeloid differentiation factor 88-mediated acute arthritis in mice. Arthritis Rheum 2005 ; 52:2530-40. 19. Abdollahi-Roodsaz S. Joosten LA. Roelofs MF et al. Inhibition of toll-like receptor 4 breaks the inflam matory loop in autoimmune destructive arthritis. Arthritis Rheum 2007 ; 56:2957-67. 20. Van Lent PLEM. Blom AB. Grevers L er aI. Toll-like receptor 4 induced FcyR expression potentiates early onset of joint inflammation and cartilage destruction during immune complex arthritis: Toll-like receptor 4 largely regulates FcyR expression by interleukin 10. Ann Rheum Dis 2007 ; 66:334-340. 21. Kim TH. Payne U. Zhang X et al. Altered hosnparhogen interactions conferred by the blau syndrome mutation of NOD2. Rheumatol Int 2007 ; 27:257-62. 22. Radstake TR, Franke B. Hanssen S et aI. The toll-like receptor 4 Asp299Gly functional variant is associated with decreased rheumatoid arthritis disease susceptibility but does not influence disease severity and/or outcome. Arthritis Rheum 2004 ; 50:999-1001. 23. Roelofs MF, Joosten LAB. Abdollahi Roodsaz SA er al. The expression of toll-like receptors 3 and 7 in rheumatoid arthritis synovium is increased and costimulation of toll-like receptors 3. 4 and 7/8 results in synergistic cytokine production by dendritic cells. Arthritis Rheum 2005 ; 52:2313-2322. 24. Baeten D. Kru ithof E, De Rycke L et al. Infiltration of the synovial membrane with macrophage subsets and polymorphonuclear cells reflects global disease activity in spondyloarrhroparhy, Arthritis Res Ther 2005 : 7:R359-69 . 25. Baeten D, Demetter P. Cuvelier CA er al. Macrophages expressing the scavenger receptor CD163: a link between immune alterations of the gut and synovial inflammation in spondyloarchropathy. ] Pathol 2002; 196:343-50. 26. Baeten D. Meller H], Delanghe] et al. Association of CD163+ macrophages and local production of soluble CD 163 with decreased lymphocyte activation in spondylarthroparhy synovitis. Arthritis Rheum 2004; 50:1611-23. 27. Huang QQ. Ma Y. Adebayo A et al. Increased macrophage activation mediated through toll-like receptors in rheumatoid arthritis. Arthritis Rheum 2007 ; 56:2192-2201. 28. De Rycke L. Vandooren B. KruithofE et al. Tumor necrosis factor-a blockade treatment down-modulates the increased systemic and local expression of toll-like receptor 2 and toll-like receptor 4 in spondylarthropathy, Arthritis Rheum 2005 ; 52:2146-2158. 29. Netea MG . Radstake T, Joosten LA et al. Salmonella septicemia in rheumatoid arthritis patient s receiving anti-tumor necrosis factor therapy: association with decreased interferon-gamma production and toll-like receptor 4 expression. Arthritis Rheum 2003 ; 48 :1853-7. 30. Flagg SD. Meador R. Hsia E et al. Decreased pain and synovial inflammation afier etanercept therapy in patiencs with reactive and und ifferentiated arthritis: an open-label trial. Arthritis Rheum 2005; 53:613-7. 31. Candia L, Marquez ]. Hernandez C et al. Toll-like receptor-2 expression is upregulated in antigen-presenting cells from patients with psoriatic arthritis: a pathogenic role for innate immunity? ] Rheumacol 2007; 34:374-9 . 32. Baker BS. Ovigne ]M. Powles AV et al. Normal keratinocyres express toll -like receptors (TLRs) I. 2 and 5: modulation ofTLR expression in chronic plaque psoriasis . Br] Derrnarol 2003; 148:670-9 . 33. Curry j L, Qin ]Z. Bonish B et al. Innate immune -related receptors in normal and psoriatic skin. Arch Pathol Lab Med 2003; 127:178-86 . 34. Raffeiner B. Dejaco C. Dunner C et al. Between adaptive and innate immunity: Tl.Rs-mediared perforin production by cozs-« T-helpe r cells in ankylosing spondylitis. Arth Res Ther 2005; 7:1412-1420.
Innate Immunity ofSpondyloarthritis: The Role ofToil-Like Receptors
309
35. Adam R. Sturrock RD. Gracie JA . TLR4 mutations (Asp299Gly and Thr399I1e) are not associated with ankylosing spondylitis. Ann Rheum Dis 2006; 65:1099-101. 36. Van der Paardt M. Crusius JB. de Koning MH et al. No evidence for involvement of the toll-like receptor 4 (TLR4) A896G and CD14-C260T polymorphisms in susceptibility to ankylosing spondylitis. Ann Rheum Dis 2005 : 64:235-8. 37. Snelgrove T. Lim S. Greenwood C er al. Association of toll-like receptor 4 variants and ankylosing spondylitis : a case-control study. J Rheumarol 2007 : 34:368-70. 38. Inman RD. Chiu B. Johnston MEA et al: HLA class l-related impairment in IL-2 production and lymphocyte response to microbial antigens in reactive arthritis . J Immunol 1989: 142:4256-4260. 39. Inman RD. Rohekar 5, Riarrh Ret al. TLR2, but not TLR4. variants are associated with reactive arthritis. Arthritis Rheum 2007; 56:5515. 40. Merx S. Neumaier M. Wagner H et al. Characterization and investigation of single nucleotide polymorphisms and a novel TLR2 mutation in the human TLR2 gene. Hum Mol Genet 2007 ; 16:1225-32. 41. Berdeli A, Celik HA. Ozyurek Ret al. TLR-2 gene Arg753Gln polymorphism is strongly associated with acute rheumatic fever in children. J Mol Med 2005; 83:535-41. 42. Henckaerts L, Pierik M, Joossens M et al. Mutations in pattern recognition receptor genes modulate seroreactivity to microbial antigens in patients with inflammatory bowel disease. Gut 2007 ; 56:1536-42. 43. ReveilleJD, Zhou X. Ward MM et al. The IL23R Gene is a major determinant of susceptibility to but not severity of ankylosing spondylitis. Arthritis Rheum 2007; 56:5529. 44. Rahman P. Inman RD. Gladman DD et al. Association of interleukin-23R variants with ankylosing spondylitis. Arthritis Rheum 2007 ; 56:5529. 45. Ciccia F. Bombardieri M. Principato A er al. IL-23 over expression as immunologic al signature of subclinical intestinal inflammation in patients with AS. Ann Rheum Dis 2007; 66:83. 46. De Jager PL, Franchimont D. Waliszewska A er al. The role of the toll recepto r pathway in susceptibility to inflammatory bowel diseases. Genes Immun 2007; 8:387-97. 47. Picard C. Casanova JL, Abel L. Mendelian trait s that confer predisposition or resistance to specific infection s in humans . Curr Opin Immunol 2006 ; 18:383-90. 48. Hawn TR. Verbon A. Lerringa KD er al. A common dominant TLR5 Stop codon polymorphism abolishes flagellin signaling and is associated with susceptibility to legionnaires' disease. J Exp Med 2003 ; 198:1563-72.
INDEX
A
C
Activating transcription factor-6(ATF6) 224,226 Animal model 87,89,90,95,116,218,223, 226.245.249-251,270,293 Ankylosingspondylitis 1-4.6-10,18-20,37. 38,41,42,45,62,67.85,86,89,91 ,92, 114,115,119,122,125-128,133-138, 148,150,151 ,153.159,167,177,178, 196,210,217,237,245,246,255,256. 277.286,291 ,293.305 Antagonist ofnuclear factor-lCb(RANK) 74, 85.86,88-94,100,101,103-110 Antigen binding cleft 167 Anti-TNF 95, 139, 142 Arthritogenic bacteria 302 Arthritogenic peptide hypothesis 79. 177, 178.187,193 Arthropathy 37,38,40,85,100,109,110. 114,115,193,229.248 ARTS1 gene 193 Assessments in ankylosing spondylitis (ASA) 2-9,11.26,129,134-136,138.139,141. 142 Autoimmunity 63,196,204.237.259,261 Autoreacriviry 255.261
Calcification 22,39,40-42.45-49, 116,245, 250 Calnexin 211 Calorimetry 190,191 Calreticulin 205.211,212,220 Cartilage oligomeric matrix protein (COMP) 74,127 Chin to chest 6 Chlamydia trachomatis 170.204,237,264. 302 Chondrocalcinosis 37.39.40,42,44,247 Chronic inflammation 71.76,85,114,120, 178.226,256,270.300 Computed tomography 17.21 C-propeptides of collagen type 2 (CPU) 75. 122.123.124.127.128 C-reactive protein (CRP) 75,122.138,142
B Bath ankylosing spondylitis disease activity index (BASDI) 2,4.9-11, 122. 124-129,135-137,139,141 -143.148. 306 Bath ankylosing spondylitis functional index (BASFI) 2.7.8, 11,45, 126, 127. 136. 137.139,141-143,148.305,306 Bath ankylosing spondylitis metrology index (BASMI) 4-7.11.137,141 Bath ankylosing spondyliti s radiographic index (BASRI) 8,18,19,137.138.148 Biomarker 20,74-76.122-129,226 BiP 211.217.222-225.229.239
D Damage associated molecular pattern (DAMPs) 67 DAP 12 adaptor molecule 279 Dendritic cell 77.78,89. 103,213.214,225, 236.249,263.264,268,271 .273,274, 278,281,282,301 ,305 Diffuse idiopathic skeletal hyperostosis 37, 39.40 Disease association 164, 166-168, 170-172. 177.180,199.203.210,228.306
E Ectopic calcification 42,45-47 Edmonton ankylosing spondylitis metrology index (EDASMI) 6.7 Endoplasmic reticulum-associated degradation (ERAD) 220.222,236 Endoplasmicreticulum 151, 168, 171, 193, 204,210,211,217,219,235,236.237, 264,275 Endosomal recycling 206 Enthesis 4.22,33,40.47.48.57-59,61 -63, 65,66-68,72,73,85.114
312
Enchesisorgans 57,58,61,62,65,66,68 Enchesitis 1, 3,4,9,11,12,17,21,22,33, 37,39,57,58,62,63,67,71,72,74,133, 138,142,247,248,251 Erp57 205,211,212,220,223 Erythrocyte sedimentation rate (ESR) 73, 75,122,124,126-129,136,138,142
F Fibrocartilage 40,57,59,61-63,65-67,72 Finger to floor distance 5, 6 Functional encheses 66, 68
M olecular Mechanisms ofSpondyloarthropathies
K KIR3DLllocus 290,292 KIR genes 153,286,287,290,295 KIR receptors 286,287,292,294,295
L Ligand repertoire 170 LILR receptors 281-283 Lipopolysaccharide (LPS) 225-227 ,235, 238 ,240,241, 271-275 , 278, 301-305
M
G Genetic polymorphism 47,67,307
H HLA-B27 17,25,37-39,44,57,67,73,76, 77,79,122,148,151,154,159-163, 166-172,177-181,183,185-193,196, 197,199-206,210-214,217-229, 235-241, 245-251, 255-257, 260, 261, 263,264,266,270-272,275,277-282, 286,287,291-296,300,316 HLA polymorphism 287,296 Homodimer 171,178,206,212,219,224, 235.237,249,277,278,280,282,286, 291,294,296 Host response 302
I IL23 receptor (IL23R) 45,151, 152, 155, 229,306 Imaging 8,9,17,18,21,22,24-27,32,33, 58,67,86,133-136,138,143,171 Immunoreceptor tyrosine activatory motifs (ITAM) 103,107,288,290 Immunoreceptor tyrosine inhibitory motifs (ITIM) 103,107,279,281, 288, 290, 292 Immunoregulatory receptors 277 Inflammatory back pain (IBP) 2,3,18,21 , 22,25,26,33,39,50,73,133,140,143 Inflammatory response 77-79,143,237, 239,246 Intermalleolar distance 6,7, 137
Maastricht Ankylosing Spondylitis Enchesitis scale (MASES) 3.4, 12, 138 Magnetic resonance imaging 17,22,32,58, 86,138 Mander enchesitis index (MEl) 3, 138 Matrix metalloproteinases 74, 101, 122, 125, 126,128 MHC class I molecule 77,133,168,178, 179,181, 190,206,210-212,214,236, 237,245,277,286 Misfolding 171,206,217-219,220-227, 229,235,236,239,241 ,249,296 Molecular mimicry 170,171, 177, 178, 187, 188,189,192,196,203,204,206,255, 256,287
N Non-steroidal anti -inflammatory drug (NSAID) 20, 25, 139 Nuclear factor KB 77,237 Nuclear factor-xls ligand (RANKL) 74,85, 86,88-95,100,101,103-110,120 Seealso: Receptor activator Nucleotide binding oligomerization domain 2 (NOD2) 76,78,149,263,264,272, 273,304
o Occiput to wall 5,6, 142 Oligoarthritis 133,235 Ossification 13,37,39,40,43-47,50, 116, 247 Osteoarthritis 41,43,91-93,11 5, 122
Index
Osteoclast 72,74,85-95,100,101, 103-110, 115-118,120 Osteoclast differentiation 95 Osteoclastogenesis 88-91, 93, 95,101, 103-110 Osteopetrosis 89,90,100,104-106 Osteoporosis 18,39,85,86,88,91-94,96, 100,101,104,107-109,118 Osteoprotegerin (OPG) 85,86,88,89, 91-95,100,101, 103, 104, 108-110, 119, 120, 127 Outcome measure 1, 2, 4, 6, 13, 135, 136, 141 Oxidoreductase 205,211,212
p Pathogen-associated molecular pattern (PAMP) 300,301 Pattern recognition receptor (PRR) 67,225, 227,263,271-273,275,300,301,306 Peptide complex 178,180,181, 189-192, 206 Peptide loading complex (PLC) 107, 205, 211-214,220,222,296 Peptide 67,77,79,89-91,103-105,151, 152,159, 162, 1~172 , 177-181, 183, 185-193,196,199-206,210-214,217, 219,220,222-225,236,237,245,246, 249,255-261,263,277,278,281, 282, 287,291 -296,301 Plain radiography 17,18,21 -23,25,27,33 Pocket 164-169,179-181,183,187,192, 200-202,205,222,256,293,296 Population 21,37,38,40,42,44,59,72,73, 76,100,133,134,149-151, 153-155, 159,161,163,165-167,171,172,196, 199,210,212,213,218,227,228,256, 266,267,279,282,289,290,293,294, 296,300,301, 304-306 Progressive ankylosis 114, 246 Proteasome 108,204,205,222,225 Protein quality control 220 Psoriaticarthritis 1, 2, 9, 11,22,58,62,65, 66,85,89,90,93,115,140,141 ,152, 286,291 ,293,295,305
313
R Reactivearthritis 1,22,85,86,95, 148, 164, 219,235,245,246,256,264,270,277, 300, 302, 305 Receptor activator or NF-k(3 (RANK) 74, 85,86,88-94,100,101,103-110 Rheumatoid arthritits 1,2,9,22,41,75,85, 86,88,90-95,100,101,109,114,119, 126,136,140,150,151,155,245,248, 270,282,291
s Sacroiliacdisease 50 Sacroiliacjoint 17,18,21-27,32,33,37-41, 48-50,67,71 -73,86,90,91,95,118, 128,133,134,137,138,143,246 Sacroiliitis 1, 13, 18,21, 22, 24-27, 38, 39, 76,138,142,246,248,251 Schober's test 5, 6, 137 Self-peptide 183 Single nucleotide polymorphism (SNP) 45, 47,150,151,153,217,305 Spectroscopy 190-192 Spondyloarthritis 12,17,22,26,37,39, 71-73,76,77,86,90,91,122,164,167, 170,196,206,217,228,229,245,246, 249-251, 263, 277, 278, 282, 283, 300, 302,304 Spondyloarthroparhy (SpA) 1,2,4,9,17, 18,21 -24,27,28,33,37,38,57-59,62, 63,65,66,67,71-79,85,86,90-96,114, 122,125,126,128,133,139-141, 164, 166,167,172,178,193,196,214,217, 218,219,223,224,228,229,235-237, 240,241,263,264,270,286,287, 291-293,295,296,300,304,306,307 Substitution 76,153,159-161,164-168, 171, 222, 236, 292 Subtype 22,44,77,133,159-172,177,180, 181,183,185,186-190,192,193,196, 197,199-206,212,220,228,236,237, 255-257,259,293,294,296 Sulphasalazine 73 Synovialbiomarker 75 Synovialmembrane 57,61,66,67,72-76, 114,115 Synovio-enthesealcomplex (SEC) 57,66
314
T Tapasin 171,205,206,211, 212, 219, 220, 222,223,225,278,281 T cell receptor 270 T cell receptor interacting molecule (TRIM) 270 Tissue inhibitors of meralloproteinase (TIMP) 74, 126 TNF 2,6,9,11,13,22,29,30,31,44,74, 75,77,88-95,101,103,104,108,109, 114,117,118,122,125-128,133,134, 139-143,152,220,226,227,245,247, 248,250,251,270,294,301,303-305 TNF receptor superfamily 89,103 Toll-like receptor (TLR) 77,78,227,263, 271,272,275,300-307 Tragus to wall 5, 6, 137
Molecular Mechanisms ofSpondyloarthropathies
Transforming growth factor-B 88 Transporters associated with antigen processing (TAP) 205,211, 212, 220, 294 TRATI gene 270
u Ultrasonography 22, 138 Unfolded protein response (UBR) 79,171, 217,219,221,222,224-229,239,240, 249 ,270-275
V Viral peptide 171, 177,204,257,260,293, 295
