Proefschrift de Lange by Nicole Nijhuis

Synovial Inflammation In Knee Osteoarthritis

Synovial Inflammation In Knee Osteoarthritis Histological and Imaging Studies

Uitnodiging Voor het bijwonen van de openbare verdediging van het proefschrift

Synovial Inflammation In Knee Osteoarthritis Histological and Imaging studies

De openbare verdediging zal plaatsvinden op dinsdag 27-10-2015 om 13:45 in het Academiegebouw, Rapenburg 73, te Leiden Aansluitend aan de promotie bent u van harte welkom op de receptie in het Academiegebouw. Badelog Jeanine Elise de Lange-Brokaar Vuurbloem 8 2317 LP Leiden badelog@hotmail.com

Paranimfen: Rosaline van den Berg r.van_den_Berg@lumc.nl Marieke Twickler mtwickler@hotmail.com

B.J.E. de Lange-Brokaar

Badelog Jeanine Elise de Lange-Brokaar

SYNOVIAL INFLAMMATION IN KNEE OSTEOARTHRITIS Histological and Imaging Studies

Badelog Jeanine Elise de Lange-Brokaar

COLOPHON ISBN: 978-94-6233-098-6 Cover Design: Badelog de Lange-Brokaar Thesis lay-out: Nicole Nijhuis, Gildeprint Printing: Gildeprint, Enschede ÂŠ B.J.E. de Lange-Brokaar, 2015 badelog@hotmail.com All rights reserved. No part of this book may be reproduced in any form without written permission from the author or, when appropriate, of the publishers of the publications. The printing of this thesis was kindly supported by: Pfizer, ABN-AMRO, AbbVie B.V.

SYNOVIAL INFLAMMATION IN KNEE OSTEOARTHRITIS Histological and Imaging Studies

Proefschrift

ter verkrijging van de graad van Doctor aan de Universiteit Leiden, op gezag van de Rector Magnificus prof. mr. C.J.J.M. Stolker, volgens het besluit van het College voor Promoties te verdedigen op dinsdag 27 oktober 2015 klokke 13:45 uur door

Badelog Jeanine Elise de Lange-Brokaar geboren in Zwijndrecht in 1980

Promotor: Co-promotor: Promotiecommissie:

Prof. dr. M. Kloppenburg Dr. A. Ioan-Facsinay Prof. dr. T.W.J. Huizinga Prof. dr. W.F. Lems (VuMC te Amsterdam) Prof. dr. F.P.J.G. Lafeber (UMC te Utrecht) Prof. dr. J.L. Bloem Prof. dr. T.P.M. Vliet Vlieland

TABLE OF CONTENTS CHAPTER 1

Introduction

PART I THE NATURE OF SYNOVIAL INFLAMMATION IN KNEE OA CHAPTER 2 CHAPTER 3 CHAPTER 4 CHAPTER 5

Synovial inflammation, immune cells and their cytokines in osteoarthritis: a review Inflammatory cells in end stage knee osteoarthritis patients: a comparison between the synovium and the infrapatellar fat pad (IFP) Characterization of synovial mast cells in knee osteoarthritis: association with clinical parameters Degree of synovitis on MRI by comprehensive whole knee semi-quantitative scoring method correlates with histologic and macroscopic features of synovial tissues inflammation in knee osteoarthritis

55 73

PART II THE ROLE OF SYNOVITIS IN THE CLINICAL BURDEN OF OA CHAPTER 6 CHAPTER 7 CHAPTER 8 CHAPTER 9 CHAPTER 10

Association of pain in knee osteoarthritis with distinct patterns of synovitis Radiographic progression of knee osteoarthritis is associated with MRI abnormalities in both the patellofemoral and the tibiofemoral joint Evolution of synovitis in osteoarthritic knees and its association with clinical features Summary, general discussion and future perspectives Nederlandse samenvatting, discussie en toekomstperspectieven

109

125 143 159 175

APPENDIX Artrosegroep reumatologie valt in de prijzen Publicatielijst Curriculum Vitae Dankwoord

193 199 201 203

CHAPTER 1 INTRODUCTION

Chapter 1

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 8

Introduction

INTRODUCTION Osteoarthritis (OA) is a heterogeneous disorder primarily resulting in joint destruction with high clinical burden that can affect all joints in the body, but is especially prevalent in the knee joint1. Knee OA is characterized by knee pain and stiffness, and is accompanied by cartilage loss and abnormalities in the subchondral bone. For scientific purposes, criteria sets for classification of knee OA were developed though a multicentre study group by Altman et al in 1986. A prerequisite for all criteria sets was knee pain (table 1)2. Table 1: American college of Rheumatology (ACR) criteria sets for Knee Osteoarthritis (OA), Modified after Altman et al, 1986 2 Clinical and laboratory Knee pain +

Clinical and radiographic Knee pain +

At least 5 of 9: At least 1 of 3:  Age > 50  Age > 50  Stiffness < 30 minutes  Stiffness < 30 minutes  Crepitus  Crepitus  Bone tenderness  Bone enlargement + Osteophytes  No palpable warmth  ESR < 40 mm/hour  RF < 1:40  SF OA

Clinical Knee pain + At least 3 of the 6:  Age > 50  Stiffness < 30 minutes  Crepitus  Bone tenderness  Bone enlargement  No palpable warmth

* ESR: erythrocyte sedimentation rate, RF: rheumatoid factor, SF OA: synovial fluid signs of OA (clear, viscous, or white blood cell count < 2.000/mm2).

Epidemiology and risk factors Not only is OA one of the most common rheumatic disorders, it is also in the top 25 most prevalent diseases in the population worldwide as was reported by the World Health Organization (3.64% of the population in both sexes (2.56% in males and 4.74% in females) 3 4 . There were approximately 1.189.000 patients (444.000 male and 745.000 women) with OA (prevalence: 53,8 per 1.000 males and 88,5 per 1.000 women) in the Netherlands in 2011 (Source: www.volksgezondheidenzorg.info/onderwerp/artrose). OA is most prevalent in the knee joint (prevalence 227.000 in males and 367.000 in women). The clinical burden of OA is high, which is underscored by the fact that OA is responsible for the largest number of years lived in disability (YLD) in the elderly female Dutch population5. Therefore, OA is becoming a significant medical and financial burden in a world whose population is aging. OA is considered a multifactorial disease, in which factors determine the susceptibility for OA. These factors, together with local biomechanical factors, determine the localization and severity of OA in a certain joint. The most important risk factors are female gender and age.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 1

High body mass index (BMI) also shows an increased risk of developing OA. BMI acts both via systemic and local pathways (table 2)1. Table 2: Risk factors for the occurrence and progression of knee Osteoarthritis (OA) â&#x20AC;&#x201C; modified after Bijlsma et al, 2011 1 Risk factors for knee OA Occurrence - Deleterious

- Protective Protective/deleterious - Progression - Deleterious

Protective

age, gender, physical activity, body-mass index, intense sport activities, quadriceps strength, bone density, previous injury, vitamin D, malalignment (including varus and valgus), genetics Hormone replacement therapy Smoking Age, body-mass index (including obesity), vitamin D, malalignment (varus and valgus), chronic joint effusion, synovitis, intense sport activities, subchondral bone oedema on MRI Hormone replacement therapy

Pathogenesis OA is characterized by focal lesions of the articular cartilage, combined with a hypertrophic reaction (sclerosis) in the subchondral bone and new bone formation (osteophytes) at the joint margins. Although for a long time OA was considered to be a degenerative process, a combination of complex degenerative and repair processes leads to changes observed in OA. Several factors are involved in these processes, such as mechanical stress, biochemical (such as inflammatory mediators) and genetic factors1. Recently, OA has been relabelled as a whole organ disease as not only abnormalities in the cartilage but also in subchondral bone and synovial tissue are involved. Moreover, pathologic abnormalities such as periarticular muscle weakness, lax ligaments, meniscal degeneration and neurosensory system alteration are frequently present in these patients6. - Cartilage abnormalities in OA Articular cartilage is composed of extracellular matrix (ECM) which contains collagen (mainly type II fibrils with both collagen IX and XI integrated) and proteoglycans. Chondrocytes are the only cells found in the ECM and are responsible for production, maintenance and destruction of the cartilaginous matrix. Healthy articular chondrocytes maintain a stable phenotype and do not show proliferation and differentiation. In OA, chondrocytes develop terminal differentiation and hypertrophy leading to disruption of the cartilage. Moreover, chondrocytes respond to injuries by producing degrading enzymes and by developing inappropriate repair responses1,7. - Subchondral bone abnormalities in OA The subchondral bone is a global term that includes the subchondral bone plate, the underlying trabecular bone and bone marrow. An important feature of OA pathophysiology

Introduction

is subchondral bone remodelling, which is characterized by increased subchondral bone thickness (sclerosis), formation of new bone at the joint margins (osteophytes) and subchondral bone cysts development7,8. Another feature in the subchondral bone that received a lot of attention is bone marrow lesions (BML), which can be visualized on MRI. The nature of BMLs is not fully elucidated, but seems to reflect local regions of sclerosis with increased bone volume fraction and increased trabecular thickness. However, OA subchondral bone also contains lower trabecular spacing. These features are thought to represent trabecular remodelling or compression9,10. Both osteoblasts as well as increased expression of insulin-like growth factor 1 (IGF-1) and transforming growth factor - β (TGF-β) by subchondral bone are thought to play a role in the abnormal bone remodelling in OA11 7,8. - Synovial tissue abnormalities in OA For a long time, OA was considered a non-inflammatory condition. More recently, however, it became evident that synovial inflammation could play an important role in the pathophysiology of OA 3,12-16 as it is a predictor of cartilage destruction 17,18 and a determinant of pain19,20. Although the biological processes underlying the appearance of synovial inflammation are poorly understood, it has been suggested that cartilage breakdown products could lead to activation of immune cells and production of pro- and anti-inflammatory mediators, which in turn could stimulate further cartilage breakdown, creating a negative feedback loop in OA (Figure 1) 16. Several immune cells have been found in OA synovial tissue, such as macrophages, T cells, mast cells, B cells and others. Furthermore, several cytokines such as tumor necrosis factor- α (TNF-α), interleukin 1- β (IL1-β) and others have been implicated in OA. The role of inflammation, immune cells and cytokines in OA, however, is less clear and pathophysiologic mechanisms are still hypothetical as these are not linked to clinical parameters. Likewise, it is unclear how synovial inflammation evolves during the disease course, since studies investigating this topic were conflicting 12,14,15,21. Therefore, it still seems unclear whether synovitis causes pathological changes or whether OA causes synovitis.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Figure 1 Involvement of the synovium in OA pathophysiology

Chapter 1

Figure 1. Involvement of synovial tissue in Osteoarthritis (OA) pathophysiology, from Sellam & Berenbaum, 2010. 16 * ADAMTS: a disintegrin and metalloproteinase with thrombospondin motifs, BMP: bone morphogenetic protein, Sellam, J. & Berenbaum, F. (2010) The role of synovitis in pathophysiology and clinical symptoms of CCL2: CC motif chemokine ligand 2, CXCL13: C-X-C motif chemokine 13 ligand 13, EGF: endothelial growth factor, osteoarthritis GM‑CSF: granulocyte‑macrophage colony stimulating factor, IL: interleukin, IL‑1Ra: IL‑1 receptor antagonist, Nat. Rev. Rheumatol. doi:10.1038/nrrheum.2010.159 LIF: leukaemia inhibitory factor, LTB4: leukotriene B4, MMP: matrix metalloproteinase, NAMPT: nicotinamide phosphoribosyl transferase (also called visfatin), NO: nitric oxide, NGF: nerve growth factor, PGE2: prostaglandin E2, TIMP: tissue inhibitor of metalloproteinase, TNF: tumor necrosis factor, vCAM-1: vascular cell adhesion molecule 1, vEGF: vascular endothelial growth factor.

- The role of infrapatellar fat pad in OA The infrapatellar fat pad (IFP), also known as Hoffa’s fat pad, is an intracapsular structure that fills the anterior knee compartment. Due to its location, adjacent to the synovial tissue, and due to its inflammatory nature, it is conceivable that the infrapatellar fat pad has a role in pathophysiology of OA, although its role in OA is largely unknown22,23. Clinical aspects of knee OA Knee OA is characterized by pain, stiffness and reduced function, and consequently, by decreased mobility and participation. Knee OA can involve two joints in the knee; the tibiofemoral joint, encompassing both medial and lateral tibiofemoral compartment, and the patellofemoral joint. Pain experienced during walking on level ground is usually a symptom observed in patients with tibiofemoral OA, while pain during climbing or ascending stairs is more often observed in patients with patellofemoral OA, although OA is usually found in both joints simultaneously creating a possible overlap of symptoms. Physical examination of the knee reveals bony enlargements (osteophytes), crepitus, bone tenderness, restricted joint movement and effusion. In clinical practise plain radiographs of the knee can be used to confirm the diagnosis, by showing structural abnormalities such as osteophytes and joint space narrowing. Joint space narrowing reflects a loss of articular cartilage (Figure 2). However, the role of radiographs in research of the knee is much debated as OA is now

Introduction

thought of as a whole organ disease including subchondral bone and synovial tissue, which cannot be visualized on radiographs. Therefore, MRI plays an upcoming role in OA as it enables visualisation of all structures of importance in the knee1,24.

Figure 2. Radiological views of knee osteoarthritis (OA). (A) Medial tibiofemoral knee OA on a weight-bearing fixed flexion posteroanterior view. (B) Patellofemoral OA on a lateral view.

Natural history OA Knee OA is a chronic disorder but its natural history can vary greatly. OA generally develops progressively over several years although may remain relatively stable for prolonged periods25. In patients at risk, local mechanical factors such as malalignment, muscle weakness or alterations in the structural integrity of the joint environment (for instance abnormalities of the meniscus or ligaments or articular abnormalities caused by previous trauma) facilitate the progression of the disease, especially in knee OA26. The correlation between clinical outcome and radiographic course is relatively poor at the individual level: whereas symptoms can improve, the radiographic picture rarely does27. Progression of OA can result in disability, pain and joint destruction, requiring a total joint replacement28. Therapy Unfortunately, no disease modifying therapies exist for OA and current therapies rely on symptom relief such as pain medication, NSAIDs and physical therapy. In end-stage knee OA, an arthroplasty is the only effective therapy.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 1

AIM OF THESIS Synovial inflammation is present in knee OA and seems to be of importance in pain perception and disease course. However, its role in the pathophysiology of OA is largely unclear. Therefore, in this thesis we aimed to investigate: 1. The nature of synovial inflammation in knee OA 2. The role of synovitis in the clinical burden of knee OA Insight in synovial inflammation could ultimately lead to better understanding of the pathophysiology of OA and therefore might provide clues for developing disease modifying drugs in OA.

GEMSTOAN STUDY In the current thesis patients have been included, which participated in the geMstoan study. The geMstoan study (GEneration of Models, Mechanism & Markers for STratification of OsteoArthritis patieNts), an observational study in established and end-stage knee OA patients to find new biomarkers for OA progression. Between 2008 and 2013, patients with symptomatic radiographic primary knee OA, following the American College of Rheumatology (ACR) classification criteria 2 and who were attending the rheumatology or orthopaedic department of the LUMC or orthopaedic department of the Diaconessenhuis, Leiden, were included. The geMstoan study comprises two groups of patients: one group of patients with end-stage disease that were planned to receive an arthroplasty and another group with mild to established OA that had no indication for an arthroplasty. Patients with mild to established disease received an arthroscopy and were followed for 2 years. Synovial tissues and infra patellar fat pad were obtained during arthroscopy or arthroplasty which enabled us to investigate the nature of synovitis. Furthermore, as baseline and 2 year follow-up radiographs and MR images were made in every patient it was possible to investigate in detail MRI abnormalities and especially. This study has been approved by the ethics committee of the Leiden University Medical Center (LUMC). All patients provided written informed consent.

OUTLINE THESIS Part I The nature of synovial inflammation in knee OA There is increasing evidence that inflammation is present in synovial tissue of OA patients. However the role of synovial inflammation, including immune cells and their cytokines,

Introduction

are not fully understood. Since OA synovial tissues are frequently used as control tissue in histological studies in rheumatoid arthritis (RA), a considerable amount of knowledge is readily available in scientific literature. Therefore, we capitalized on earlier histological studies using OA synovial tissue and performed an extensive narrative review (chapter 2), summarizing all current knowledge of synovial inflammation. In this review, we focused on two aspects of inflammation: the degree of inflammation as can be evaluated by H&E staining and secondly, on specific immune cells and their cytokines as these might provide important insights in the underlying mechanisms of synovitis and pain in osteoarthritis. Although it is shown that synovitis is associated with pain, we learned from chapter 2 that it is unclear which immune cells and their mediators play a role in pain experience. Therefore, in chapter 3 we extensively characterized immune cells in both synovial tissue and infrapatellar fat pad of end-stage knee OA patients by flow cytometry analysis and investigated relation of immune cells with pain. To further investigate the role of immune cells in OA different subtypes and activation states of immune cells and their intracellular cytokines were investigated. Furthermore, to investigate underlying mechanisms in pain, associations between immune cells and pain were investigated. As observations from chapter 2 suggested that mast cells could play an important role in OA, in chapter 4 number and granulation state was investigated by immunofluorescence stainings in both mild to established and in end-stage knee OA patients. The association of mast cells with clinical parameters was also investigated. Furthermore, we compared our findings to RA synovial tissue samples. Although histological assessment of synovial biopsies is currently the golden standard for evaluating synovial inflammation in knee OA, acquisition of synovial biopsies is technically difficult and patient unfriendly as it involves an invasive procedure like arthroscopy. Therefore, a non-invasive method, such as contrast-enhanced (CE) MR imaging constitutes an attractive alternative for visualizing synovial tissue inflammation 29-31. As the anatomical distribution of synovitis on CE-MRI is patchy and heterogeneous 32, a MRI scoring method should encompass a sufficient number of compartments. A recently developed method by Guermazi et al. is a semi-quantitative method that scores synovitis on CE-MRI at 11 different sites throughout the knee 33 and constitutes a comprehensible and practical method for assessing synovitis in the whole joint. In chapter 5 we used this scoring system to investigate whether the degree of synovitis on CE-MRI correlates with microscopic and macroscopic features of inflammation in knee OA patients. Furthermore, we aimed to investigate whether the degree of inflammation differs in patients with different stages of knee OA. Part II The role of synovitis in the clinical burden of OA Although synovitis is known to be heterogeneous and therefore current scorings systems should investigated synovitis at different anatomical sites, it was not known whether

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 1

synovitis at different sites occurs independently or whether they form patterns. Therefore, in chapter 6 we aimed to investigate whether patterns of inflammation exist in an unbiased manner in knee OA and whether they associate with pain. Several MRI features are known to be related to radiographic progression. However, these MRI features are known to be highly correlated. Therefore, to increase our insight into the interaction between MRI abnormalities in the different joint tissues and the interaction between the patellofemoral joint (PFJ) and tibiofemoral joint (TFJ) in relation to radiological progression, we aimed to investigate patterns of different tissue abnormalities as assessed with MRI of both the PFJ and TFJ in an unbiased manner in Chapter 7. Furthermore, to understand the role of synovitis in the disease course of OA and its relation to disease progression and pain, in chapter 8 we investigated changes in synovitis over a 2-year period in knee OA patients. Finally, in chapter 9, we provide an overview of the thesis and discuss our findings.

Introduction

REFERENCES

1 Bijlsma JW, Berenbaum F, Lafeber FP. Osteoarthritis: an update with relevance for clinical practice. Lancet 2011;377:2115-26. 2 Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K et al. Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arthritis Rheum 1986;29:1039-49. 3 Felson DT. Clinical practice. Osteoarthritis of the knee. N Engl J Med 2006;354:841-8. 4 Vos T, Flaxman AD, Naghavi M, Lozano R, Michaud C, Ezzati M et al. Years lived with disability (YLDs) for 1160 sequelae of 289 diseases and injuries 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012;380:2163-96. 5 Zantinge EM, van der Wilk EA, van Wieren S. Gezond ouder worden, Rijksinstituut voor Volksgezondheid en Milieu (RIVM). 2011;270462001: 6 Loeser RF, Goldring SR, Scanzello CR, Goldring MB. Osteoarthritis: a disease of the joint as an organ. Arthritis Rheum 2012;64:1697-707. 7 Sharma L, Chmiel JS, Almagor O, Felson D, Guermazi A, Roemer F et al. The role of varus and valgus alignment in the initial development of knee cartilage damage by MRI: the MOST study. Ann Rheum Dis 2013;72:235-40. 8 Henrotin Y, Pesesse L, Sanchez C. Subchondral bone and osteoarthritis: biological and cellular aspects. Osteoporos Int 2012;23 Suppl 8:S847-S851. 9 Driban JB, Tassinari A, Lo GH, Price LL, Schneider E, Lynch JA et al. Bone marrow lesions are associated with altered trabecular morphometry. Osteoarthritis Cartilage 2012;20:1519-26. 10 Hunter DJ, Gerstenfeld L, Bishop G, Davis AD, Mason ZD, Einhorn TA et al. Bone marrow lesions from osteoarthritis knees are characterized by sclerotic bone that is less well mineralized. Arthritis Res Ther 2009;11:R11. 11 Hilal G, Martel-Pelletier J, Pelletier JP, Ranger P, Lajeunesse D. Osteoblast-like cells from human subchondral osteoarthritic bone demonstrate an altered phenotype in vitro: possible role in subchondral bone sclerosis. Arthritis Rheum 1998;41:891-9.

12 Loeuille D, Chary-Valckenaere I, Champigneulle J, Rat AC, Toussaint F, Pinzano-Watrin A et al. Macroscopic and microscopic features of synovial membrane inflammation in the osteoarthritic knee: correlating magnetic resonance imaging findings with disease severity. Arthritis Rheum 2005;52:3492-501. 13 Loeuille D, Rat AC, Goebel JC, Champigneulle J, Blum A, Netter P et al. Magnetic resonance imaging in osteoarthritis: which method best reflects synovial membrane inflammation? Correlations with clinical, macroscopic and microscopic features. Osteoarthritis Cartilage 2009;17:1186-92. 14 Pearle AD, Scanzello CR, George S, Mandl LA, DiCarlo EF, Peterson M et al. Elevated high-sensitivity C-reactive protein levels are associated with local inflammatory findings in patients with osteoarthritis. Osteoarthritis Cartilage 2007;15:516-23. 15 Benito MJ, Veale DJ, FitzGerald O, van den Berg WB, Bresnihan B. Synovial tissue inflammation in early and late osteoarthritis. Ann Rheum Dis 2005;64:1263-7. 16 Sellam J and Berenbaum F. The role of synovitis in pathophysiology and clinical symptoms of osteoarthritis. Nat Rev Rheumatol 2010;6:62535. 17 Roemer FW, Guermazi A, Felson DT, Niu J, Nevitt MC, Crema MD et al. Presence of MRI-detected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30-month follow-up: the MOST study. Ann Rheum Dis 2011;70:1804-9. 18 Roemer FW, Zhang Y, Niu J, Lynch JA, Crema MD, Marra MD et al. Tibiofemoral joint osteoarthritis: risk factors for MR-depicted fast cartilage loss over a 30-month period in the multicenter osteoarthritis study. Radiology 2009;252:772-80. 19 Yusuf E, Kortekaas MC, Watt I, Huizinga TW, Kloppenburg M. Do knee abnormalities visualised on MRI explain knee pain in knee osteoarthritis? A systematic review. Ann Rheum Dis 2011;70:60-7. 20 Baker K, Grainger A, Niu J, Clancy M, Guermazi A, Crema M et al. Relation of synovitis to knee pain using contrast-enhanced MRIs. Ann Rheum Dis 2010;69:1779-83. 21 Smith MD, Triantafillou S, Parker A, Youssef PP, Coleman M. Synovial membrane inflammation

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 1

and cytokine production in patients with early osteoarthritis. J Rheumatol 1997;24:365-71. Klein-Wieringa IR, Kloppenburg M, BastiaansenJenniskens YM, Yusuf E, Kwekkeboom JC, El-Bannoudi H et al. The infrapatellar fat pad of patients with osteoarthritis has an inflammatory phenotype. Ann Rheum Dis 2011;70:851-7. Ioan-Facsinay A and Kloppenburg M. An emerging player in knee osteoarthritis: the infrapatellar fat pad. Arthritis Res Ther 2013;15:225. Hayashi D, Roemer FW, Guermazi A. Osteoarthritis year 2011 in review: imaging in OA--a radiologists’ perspective. Osteoarthritis Cartilage 2012;20:207-14. Botha-Scheepers S, Riyazi N, Watt I, Rosendaal FR, Slagboom E, Bellamy N et al. Progression of hand osteoarthritis over 2 years: a clinical and radiological follow-up study. Ann Rheum Dis 2009;68:1260-4. Hunter DJ and Felson DT. Osteoarthritis. BMJ 2006;332:639-42. Dieppe PA, Cushnaghan J, Shepstone L. The Bristol ‘OA500’ study: progression of osteoarthritis (OA) over 3 years and the relationship between clinical and radiographic changes at the knee joint. Osteoarthritis Cartilage 1997;5:87-97. Yusuf E, Bijsterbosch J, Slagboom PE, Kroon HM, Rosendaal FR, Huizinga TW et al. Association between several clinical and radiological determinants with long-term clinical progression and good prognosis of lower limb osteoarthritis. PLoS One 2011;6:e25426. Guermazi A, Roemer FW, Hayashi D, Crema MD, Niu J, Zhang Y et al. Assessment of synovitis with contrast-enhanced MRI using a wholejoint semiquantitative scoring system in people with, or at high risk of, knee osteoarthritis: the MOST study. Ann Rheum Dis 2011;70:805-11. Loeuille D, Sauliere N, Champigneulle J, Rat AC, Blum A, Chary-Valckenaere I. Comparing non-enhanced and enhanced sequences in the assessment of effusion and synovitis in knee OA: associations with clinical, macroscopic and microscopic features. Osteoarthritis Cartilage 2011;19:1433-9. Hayashi D, Roemer FW, Katur A, Felson DT, Yang SO, Alomran F et al. Imaging of synovitis in osteoarthritis: current status and outlook. Semin Arthritis Rheum 2011;41:116-30.

32 Roemer FW, Kassim JM, Guermazi A, Thomas M, Kiran A, Keen R et al. Anatomical distribution of synovitis in knee osteoarthritis and its association with joint effusion assessed on non-enhanced and contrast-enhanced MRI. Osteoarthritis Cartilage 2010;18:1269-74. 33 Guermazi A, Roemer FW, Hayashi D, Crema MD, Niu J, Zhang Y et al. Assessment of synovitis with contrast-enhanced MRI using a wholejoint semiquantitative scoring system in people with, or at high risk of, knee osteoarthritis: the MOST study. Ann Rheum Dis 2011;70:805-11.

Introduction

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

PART I THE NATURE OF SYNOVIAL INFLAMMATION IN KNEE OA

CHAPTER 2 SYNOVIAL INFLAMMATION, IMMUNE CELLS AND THEIR CYTOKINES IN OSTEOARTHRITIS: A REVIEW

de Lange-Brokaar BJE, Ioan-Facsinay A, van Osch GJVM, Zuurmond A-M, Schoones J, Toes REM, Huizinga TWJ, Kloppenburg M

Osteoarthritis Cartilage. 2012 Dec;20(12):1484-99.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 2

ABSTRACT Objective: Although osteoarthritis (OA) is considered a non-inflammatory condition, it is widely accepted that synovial inflammation is a feature of OA. However, the role of immune cells and their cytokines in OA is largely unknown. This narrative systematic review summarizes the knowledge of inflammatory properties, immune cells and their cytokines in synovial tissues (STs) of OA patients. Methods: Broad literature search in different databases was performed which resulted in 100 articles. Results: Of 100 articles 33 solely investigated inflammation in OA ST with or without comparison with normal samples; the remaining primarily focussed on rheumatoid arthritis (RA) ST. Studies investigating different severity stages or cellular source of cytokines were sparse. OA ST displayed mild/moderate grade inflammation when investigated by means of haematoxylin and eosin (H&E) staining. Most frequently found cells types were macrophages, T cells and mast cells (MCs). Overall the number of cells was lower than in RA, although the number of MCs was as high as or sometimes even higher than in RA ST. Cytokines related to T cell or macrophage function were found in OA ST. Their expression was overall higher than in normal ST, but lower than in RA ST. Their cellular source remains largely unknown in OA ST. Conclusion: Inflammation is common in OA ST and characterized by immune cell infiltration and cytokine secretion. This inflammation seems quantitatively and qualitatively different from inflammation in RA. Further research is needed to clarify the role of inflammation, immune cells and their cytokines in the pathogenesis of OA.

Synovial inflammation, immune cells and their cytokines in OA

INTRODUCTION Osteoarthritis (OA) is one of the most common rheumatic disorders. A majority of the elderly have radiographic or clinical evidence of OA1. For a long time, OA was considered a non-inflammatory condition and OA synovial tissue (ST) samples were used as controls for comparative purposes in studies that focused on rheumatoid arthritis (RA) pathogenesis. However, increasing evidence that inflammation is present in ST of OA patients 1-21 has brought to light the possibility that synovitis and the immune system could be active players in OA development and progression. Because immune cells and their cytokines may play an important role in the pathogenesis of OA and because a better understanding of the biological mechanisms involved in this process may lead to better therapies for OA patients, we performed a systematic narrative review to summarize the data published thus far regarding inflammation, immune cells and their cytokines in ST of OA patients.

METHODS In cooperation with a trained librarian, a broad search strategy in the following databases: Pubmed (1946–September 2011), Embase (Ovid-version 1980–September 2011) and Web of science (1945–September 2011) was composed (Fig. 1). Two different search strategies were used. The first consisted of the AND combination of following search terms: “osteoarthritis” and “synovium” as a major topic (major subject, title words). The Second strategy consisted of the AND combination of three concepts; “osteoarthritis”, “synovium” and “inflammation” (major or minor topic). All relevant keywords variations were used and search strategy was optimized for all consulted databases. See Appendix 1 for literature search details. Because the goal of this review was to summarize the existent knowledge about immune cells and their cytokines in OA ST, a broad search was initiated in which all studies were included that reported data regarding immune cells or their cytokines in OA ST, either in text or figures/tables. Moreover, studies that investigated at least five OA patients were included. 32 articles that investigated ST samples of less than five OA patients were excluded. Only studies investigating cytokines that are commonly associated with immune cells were included (e.g., interleukins (ILs)). Animal studies, experiments with non-immune cells and genetic studies were excluded. Studies concerning temporomandibular joints were excluded. After screening full text articles and after additional hand search of references and screening of related articles, a total of 100 articles remained.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 2

Figure 1. Flowchart of first and second (Italic) search. WoS (Web of Science).

RESULTS 100 articles are included in this narrative review. 35 articles investigated OA ST by standard haematoxylin and eosin (H&E) staining (Table I). In 55 the focus was on different cell types (Table II), and in 37 articles the focus was on cytokines (Table III). Only scarce data are available regarding the cellular source of cytokines in OA ST. Only results relevant for OA are described. Thirty-three articles investigated exclusively OA ST, while the majority of the articles used OA samples for comparative purposes. Moreover, only a few studies compared samples of patients undergoing arthroscopy with patients undergoing arthroplasty in OA.

Synovial inflammation, immune cells and their cytokines in OA

Synovitis in OA Houli et al. were among the first to describe features of inflammation (31% samples showed cellular infiltration) in ST from OA patients22. Later, several studies confirmed the presence of inflammation in OA ST2,5-7,9,13,16,18,19,23-43. Histological features were hyperplasia, with an increased number of lining cells (LCs) and a mixed cellular infiltrate2,6,7,9,18. Overall, inflammation of OA ST was found to be less pronounced than in RA19,23,25-27,30-34,36,37,39, but higher than in healthy controls13,18,19,25,32-34,37,40,41,43,44. Synovitis score Different scoring systems for synovitis were used throughout the years, some designed specifically for OA16, others with focus on RA31,45. An early scoring system of OA was designed by Salvati et al. for ST inflammatory responses in coxarthrosis7,23. In 2002 Krenn et al. developed a scoring system for the entire spectrum of rheumatic diseases including OA46. This system was based on the following morphologic alterations: hyperplasia/enlargement of synovial LCs, activation of resident cells/synovial stroma and inflammatory infiltration46. It is a validated scoring system that can accurately discriminate between high and low grade synovitis32 and is frequently used21,25,34,41,44. With this scoring system a mild to moderate synovitis was found in OA (mean score around 2)25,32,34,44,46 vs 1.4 for healthy controls and 5.7 for RA (scale 0–9). This mild to moderate synovitis was confirmed in other studies using author developed scoring systems2,5-7,13,16,18,24,2628,39. Additional histological studies revealed that also fibrosis and detritus (non-living organic material) can be seen in OA ST9,23,33,35. Synovitis in different severity stages in OA Tissues obtained from arthroscopy, reflecting early to established disease, were usually compared with arthroplasty, reflecting end-stage disease. Results vary, as one study reports a significantly increased mononuclear cell (MNC) infiltration in the arthroscopy samples17, whereas others find these markers to be increased in the arthroplasty samples2,3,18. Pearle et al. did not find difference16. One study has extensively characterized the different inflammatory patterns and showed that hyperplasia and infiltrate are characteristic for the arthroscopy group, while fibrotic and detritus rich domains together with infiltrate are typical for arthroplasty group33.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

28 Knee biopsies Arthroplasty

OA (17)

OA (15), N (4)

OA (9)

OA (104)

OA (16)

OA (13)

Goldenberg, 19785

Goldenberg, 19826

Haynes, 200227

Haywood, 200328

Houli, 195922

Huss, 201029

Knee or hip arthroscopy (7) or arthroplasty (97)

Knee or hip arthroplasty Knee arthroplasty

Biopsies

Hip arthroplasty

OA (19)

Gedikoglu, 19867

Knee or hip arthroplasty

OA (5), RA (5), N (5)

FuruzawaCarballeda, 199919

Histopathological: cellular infiltration, fibrosis, etc Histology: fibrin and aggregates

Grading: hyperplasia, cellular infiltration and specific changes Grading: hyperplasia, cellular infiltration and changes Histology: inflammatory infiltrate Inflammation grading

Grading by Salvati et al., 1977

Histology: degree of inflammation

Dense fibrous bands, with aggregates with macrophages and lymphocytes

5/16 cellular infiltration, 11/16 fibrosis, oedema 11/16

All samples had features of inflammation: minimal inflammation 32%, moderate 53% and advanced 16% Mostly quite normal in appearance, few moderate synovitis. 35% MNC infiltrate and 70% SLC hyperplasia 87% samples abnormal; synovitis: 40% mild, 20% moderate, 27% markedly inflamed 67% evidence inflammation, 44% large perivascular accumulations 7% no inflammation, mild inflammation 28%, moderate 35% and severe 31%

No difference between infiltrate type (perivascular and interstitial) or degree (varied moderate – abundant) between OA and RA

Table I. Studies that investigated ST in OA patients by means of H&E staining (phosphotungstic acid hematotoxylin (PTAH) and berlin blue) Author Arthropaties OA phenotype Methods Results/conclusion OA ST OA (24̂∗), RA Hip arthroplasty Grading by Salvati et al., 1977 Inflammation moderate. Two histological types: Proliferation Arnoldi, 198023 (12), N (8) and fibrous form Da, 200724 OA (41) Knee needle Synovial lining thickness, Mild to moderate inflammatory inflammation arthroscopy vascularity and infiltration Knee biopsy or Synovitis score Krenn et al., 2002 Intermediate inflammation scores Diaz-Torne, 200725 OA (12), RA (10), N (9) arthroplasty OA (8), RA (11) Knee blind biopsy or Inflammation score: SLL Inflammation score 3–8, RA 2–9 (score 0–15) Dolhain, 199626 arthroplasty hyperplasia and infiltration Knee arthroplasty Histology Mild chronic inflammation all cases Fernandez-Madrid, OA (9) 200511 Mean global score (0–60) 11.1, RA 19.8 OA (16), RA (26) Arthroscopy Grading: LL hyperplasia, vessel Fonseca, 200040 infiltrate + density, fibrosis, lymphocyte clusters + %

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 Chapter 2

Knee arthroscopy Arthroscopy (29) and arthroplasty (37)

OA (212), RA (247), N (49)

OA (39)

OA (15)

OA (29), N (14)

OA (66), RA (22), N (8)

OA (54)

OA (8)

OA (25), RA (28), N (10)

Korkusuz, 200513

Krenn, 200632

Loeuille, 20053

Loeuille, 20094

Myers, 19902

Oehler, 200233

Pearle, 200716

Pelletier, 1989113

Pessler, 2008, second34 Knee needle biopsy or arthroplasty

Arthroplasty

Knee or hip arthroscopy or arthroplasty

Knee arthroscopy

Hip or knee arthroplasty Knee synovectomy

OA (10), N (6)

Koizumi, 199931

Arthropaties OA (16), RA (6)

OA phenotype Knee or hip arthroplasty OA (22), RA (40) Knee arthroplasty

Author Johnell, 198530

Synovitis score Krenn et al., 2002, mean (SD)

Histology: morphological changes

Histology: grading synovial inflammation

Synovitis: MNC infiltration graded 0–3 Histology: degree infiltration (0–4) and hyperplasia

Grading: SLL cells, infiltration, surface fibrin, features vessels, fibrosis, perivascular oedema

Score synovitis OA 91% <10 pts, all RA >11 pts (range 0—20 pts)

Also PTAH, berlin blue staining. Scoring: proliferation synovial cells and stroma (infiltration, etc) Histology: hyperplasia and hypertrophy SLL, fibrosis Synovitis score by Krenn et al., 2002

Mean synovitis score 2.23 (low-grade), normal 1.38 (no synovitis) and RA 5.74 (high-grade)

Significant amount chroming inflammation, morphological changes were hyperplasia and hypertrophia SLC, MNC infiltrate and proliferation blood vessels

50% Low-grade and 7% high-grade synovitis. No differences grade between arthroscopy and arthroplasty group

Four patterns: hyperplastic (72% arthroscopy group), lymphocytic infiltrates (28% arthroscopy, arthroplasty 38%), fibrotic (arthroplasty 32%), detritus rich (arthroplasty 24%)

55% Moderate or severe synovitis. Normal ST 50% synovitis

Increased number of synovial LCs, fibrosis, oedema and infiltration

Hyperplasia and hypertrophy synovial LL suggesting an immunogenetic role Mean values synovitis score: control 1.0 (no synovitis), OA 2.0 (low-grade), RA 5.0 (high-grade). Significant differences between diagnoses and between high- and low-grade synovitis Severe OA showed more fibrin, infiltrate and higher total score than mild OA

Results/conclusion OA ST Inflammatory cells around small vessels and capillaries

Methods Histology: inflammatory cells

Synovial inflammation, immune cells and their cytokines in OA

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

OA (221) RA (341) OA (8), RA (16), N (5) OA (40), N (23)

OA (20)

OA (19)

OA (74), RA (127)

OA (9), RA (10)

Slansky, 201041

Scanzello, 201142

Soren, 197838

Soren, 198843

Thurkow, 199739

N, normal. ∗ Number of patients.

Smith, 199718

Smith, 199237

OA (30), RA (10) Knee or hip arthroplasty

Sakkas, 199836

Knee needle biopsy

Arthroplasty and arthroscopy Knee, hip, ankle synovectomy (8), arthroplasty (10), arthrodesis (1) Knee, hip arthrotomy

Knee synovectomy or arthroplasty Knee arthroscopy or arthroplasty Knee arthroscopy (13) and arthroplasty (27)

Knee arthroplasty

OA (19)

Saito, 200235

OA phenotype Knee or hip arthroscopy (3) or arthroplasty (35)

Arthropaties OA (38)

Author Revell, 19889

All samples infiltrates and vascular proliferation. 45% SLL thickening Histology: various degree of MNC infiltration, mild to moderate synovial hyperplasia. Inflammatory changes less prominent than in RA

Histology: inflammatory features, hyperaemia, oedema, haemosiderin, fibrosis, other features Histology: degree inflammation and infiltration

Inflammation score: perivascular MNC infiltration (0–3) Histology: hyperplasia and infiltration

Inflammation 4.7 vs 11.4 RA (range 0–15), little infiltration

Presence synovial inflammation features in lesser incidence and intensity than RA

25% No inflammation, 40% mild, 35% moderate and no severe inflammation Some focal inflammatory infiltrates and moderate hyperplasia

Synovitis score Krenn et al., 2002 Median synovitis score +/− 2 (figure), RA 5.8 (figure), normal 0.6 (figure) Histology: hyperplasia and Variable LL hyperplasia as in RA. SLL: acellular till infiltrate LL and SLL predominantly macrophage, lymphocyte or mixed infiltrates Low-grade synovitis, arthroplasty group highest degree of Histology: hyperplasia, cellularity, vascularity and inflammatory infiltrate vascularity, cellularity and infiltrate

Histology: hyperplasia, infiltrate and vascular proliferation Histology: hyperplasia, infiltrate

Methods Results/conclusion OA ST Histology: hyperplasia, lymphoid Significantly more fibrosis and lymphoid aggregates than aggregates and fibrosis samples with machanical or traumatic background.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 Chapter 2

Synovial inflammation, immune cells and their cytokines in OA

Cell types found in OA ST (Table II, Fig. 2) Many studies investigated inflammatory cell types in OA ST (Table II) 8-12,17,20,21,24,25,27,28,30,3337,44,45,47-77 . Several authors reported that macrophages and T cells are the most abundant cells 8,9,11,17,20,21,25,34,35, 37,45,47,48,51,56, 57,61,62,69,78 . Mast cells (MCs)49,50,54,55,58-60,64,67,70,79, as well as B cells and plasma cells were found, although the last two in lower amounts than other ce lls9,11,21,24,25,33-35,44,45,57,69,73. Natural killer cells were detected in OA ST65 and several authors describe the presence of low numbers of dendritic cells (DCs)8,53,66,69. Neutrophils were almost never found21,34,35,44,48,62. Overall OA ST contained fewer number of inflammatory cells than RA ST25,30,36,37,44,45,56,57,63,65,68,69,71,75,78, but more than normal control ST25,34,44,45,50,51,55,57,59,6163,65,67,70,71,75,78 . One extensive study by Pessler et al. also investigated the relative abundance of immune cells in infiltrates and found that macrophages represented approximately 65%, T cells 22%, B cells 5% and plasma cells <1% of the infiltrate. MCs were not analysed in this study. Relatively more B cells were found in RA44. Macrophages Macrophages were mostly distributed in the lining layer (LL)20,25,28,33,37,44,48,56,61,62,78. With increasing histological infiltration grade, both macrophage fraction areas as intensity of macrophage infiltration increased17,28. Subsets of macrophages, “classically activated” (M1) and “alternatively activated” (M2)30,68, have not been investigated in OA ST. Although the importance of macrophages in OA ST is hypothesized21,33,35,44,48, their role in human OA ST is still largely unknown. T cells T cells were predominantly detected in the sublining layer (SLL) and to some extent in the deep layer (DL)9,12,37,44,56,57,63. T cells expressing activation antigens were found by different authors27,36,72. Sakkas et al. found that the MNC infiltrate consisted of T cells expressing early (cluster of differation (CD)69), intermediate (CD25 and CD38) and late (CD45RO human leukocyte antigen (HLA) class II) activation antigens in OA ST36. A later study by the same author showed a decrease in CD3ζ protein in OA ST, which is suggestive for a chronic T cell stimulation72. An altered ratio of CD4/CD8 T cells was found, showing a relative enrichment of CD4+ T cells9,11,27,35,44,45,68,74. Saito et al. reported a CD4+/CD8+ ratio in OA ST of 5:1, compared to normal ST, where the ratio is 2:135. Steiner et al. found that the CD4+/CD8+ ratio in OA ST was comparable to RA ST74. Benito et al. reported that the abundance of CD4+ T cells was significantly greater in OA ST of the arthroscopy group compared to the arthroplasty group17. Several populations of T helper (Th) cells can be distinguished: Th1, Th2 and Th17 as well as regulatory cells (Th3 and Tr1) with unique function and unique cytokine patterns (Fig. 2)8083 . Different types of Th cells have been identified in OA ST, based on their cytokine profile 31

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 2

upon in vitro activation. The Th1/Th2 cell ratio in OA reported by Yudoh et al. was 1.5 compared with 6.1 in RA77. Th1 cell was seen more frequently than Th17 cells84. One study by Yudoh et al. has investigated the presence of Tr1, a regulatory T cell that inhibits immune responses, in OA. The authors describe a higher abundance of Tr1 cells in OA ST compared to RA ST77. Moreover, several authors favour the concept that T cells play an important part in pathogenesis of OA8,12,18,36,63,72,85.

Figure 2. T cells and T cell subsets in OA, with most important cytokines. Th cells (Th1, Th2, and Th17), T regulatory (Treg) cells and induced T regulatory (iTreg) cells (Th3 and Tr1).

MCs Different authors detected MCs in OA ST35,49,50,54,55,58-60,64,67,70. Bridges et al. reported a percentage of 1.6 in OA ST49. Buckley et al. found a percentage of 2.4 compared to 1.1% post-mortem and 1.3% in amputation controls50. Overall MC numbers were as high49,50,55,79 or higher58,70 compared to RA and higher than in controls50,55,59,67,70,79, although one study reported lower numbers of MCs in OA compared to RA59. The highest abundance of MCs was found within the SLL54,55,58,59 and around blood vessels54. The number of degranulated MCs was highest in superficial layers of OA ST, indicating their active state. In RA ST the number of MCs was highest in the capsule55. Despite a common origin, similar granulated morphology and functions, MCs are a heterogeneous group of cells. MCs can be subdivided into MCs containing only tryptase (MCT) and in MC containing both tryptase and chymase (MCTC) as defined by Irani et al.86. Buckley et al. investigated the MC subpopulations in OA ST vs control subjects. They found a striking shift in the relative proportions of MCs with a MCTC phenotype to MCs with a MCT phenotype. The number of MCs with a MCT phenotype was higher in OA ST (median 53 MCT/mm2) than in post-mortem control ST (7.5 MCT/mm2) or amputation controls (12 MCT/

Synovial inflammation, immune cells and their cytokines in OA

mm2)50. Gotis-Graham et al. reported a lower ratio of MCTC/MCT in OA than in RA ST (3:4 vs 3:2)59. Similarly, a higher tryptase activity of MCs in OA ST was found by Nakano et al.67. Furthermore the histamine content of MCs was the same or higher in OA ST than in RA ST49 whereas a lower histamine release is reported60,64. B cells Although present in low numbers, B cells and plasma cells were detected in OA ST8,9,11,29,34,35,44. However, their relative abundance is lower than in RA44. Da et al. reported that OA ST with increased inflammatory infiltrate contained relatively more B cells. Moreover, infiltrated B cells in ST of patients with OA were shown to be oligo-clonal, suggesting an antigen-driven expansion73. Supporting this observation sequencing of the complementarity determing region (CDR) regions of these cells indicated that B cells have been clonally expanded. Therefore, a role for these cells during the course of OA cannot be excluded24. Other cells Natural killer cells were not frequently studied but have been detected in OA ST65. Additionally very few DCs were seen in OA ST8,53,66,69, although both plasmacytoid DC and myeloid DC subsets were found66.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

34 IHC Ab CD1a, CD1b, CD1c IHC c-kit and SCF. Toluidine blue staining

Arthroplasty Arthroplasty, arthroscopy Knee needle arthroscopy Knee or hip arthroplasty Knee biopsy or arthroplasty

OA (8), RA (8) OA (12), RA 10)

OA (41)

OA (10)

OA (12), RA (10), N (9)

Cauli, 200053 Ceponis, 199879

Da, 200724

Damsgaard, 199954

Diaz-Torne, 200725

Giesma staining. Stereological microscopy MCs ICH Ab CD3, CD68, CD20, CD38

IHC Ab CD3, CD20, CD138

IHC Ab cells: 27E10, CD14, 25F9, RM3/1. Ab cytokines: IL-1α, IL-1β, IL-1Ra Double staining for macrophages and cytokines

OA (10), RA (10) Unknown

Cauli, 199752

Morphology: MCs found in vicinity blood vessels, not in LL. MCs 0.8% total number of cell profiles present in ST Significantly more macrophages (LL and subintima (SI)) and T cells than gulf war veterans illness samples. B cells and plasma cells were sparse. Lower number inflammatory cells than in RA ST

No B lymph infiltration in 54%, mild in 32%, moderate/ strong in 15% of the samples. Plasma cells 17%. No follicular DCs

Few CD1b and CD1c, no CD1a, similar to RA Increased densities c-kit+ synovial MCs. Mean number c-kit+ MC/mm2 (135 ± 26) not significantly different from RA (85 ± 16)

No difference subset macrophages between LL and SLL. Higher % mature macrophages, IL-1Ra+ than IL-1α+ in SLL. % 25F9+, IL-1+ α cells OA <RA. % 25F9+, IL-1Ra+ and RM3/1+, IL-1Ra+ similar OA and RA

T cells isolated assay hprt-mutant Isolated MNCs were predominantly CD4+ T cells. T cells, Mab CD3, CD8, TCR α/β, Number of hprt-mutant T cells lower than RA TCR γ/δ, CD29, flow cytometry

Higher number MCs than control ST, distributed throughout ST. OA ST shift MCTC to MCT (77%) phenotype, normal ST MCT (42%)

OA (8), RA (93), Arthroplasty N (19)

IHC Ab chymase spec, tryptase. Double labelling procedure

Cannons, 199851

Knee arthroplasty

OA (14), N (14)

Significantly more T cells and Macrophages in arthroscopy group compared with arthroplasty group Mainly fibroblast like synoviocytes. 2–7% macrophages, <0.5% neutrophils, <0.1% T Cells In digest 1.6% MCs, not significantly different from RA 1.3%. Histamine content comparable in OA and RA

Results/conclusion OA ST No γ/δ T cells expansion

Buckley, 199850

Table II. Studies that investigated different cell types in ST in OA patients Author Arthropaties OA phenotype Methods OA (5∗), RA (7) Arthroplasty MNCs isolated. Pan-γ/δ T cell Andreu, 199147 Mab TCR δ1 OA (25) Knee arthroscopy (10) IHC Ab CD4, CD68 Benito, 200517 vs arthroplasty (15) 48 OA (19) Knee or hip Cell suspension ST. Flow Bondenson, 2006 arthroplasty cytometry analysis OA (42), RA (48) Knee, hip or wrist Digest ST. Alcian blue staining. Bridges, 199149 arthroplasty or Histamine release assay synovectomy

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 Chapter 2

IHC Ab chymase and tryptase

OA (30), RA (90) Unknown

Arthroscopy or arthroplasty Unknown Arthroplasty

Knee or hip arthroscopy (7) or arthroscopy (97)

OA (18), RA (16), N (15)

OA (7), RA (8) OA (9), RA (6)

OA (104)

OA (25), RA (62) Arthroplasty

OA (8), RA (9)

Fritz, 198458

Gotis-Graham, 199759

Gruber, 198660 Haynes, 200227

Haywood, 200328

Helbig, 198861

Hogg, 198562

Arthroplasty, arthroscopy

Avidin-peroxidase vs toluidine Distribution MCs in SLL. Number of MCs significantly blue and Giesma staining of MCs. higher than in RA ST in sub synovial layer Double staining method

Arthroplasty

OA (6), RA (6), N (3)

Fonseca, 200257

Mab Ki-M6, Ki-M8, anti-HLA-DR, OKT 9 Mab 24, UCHM1, 44, 24, DA2, 5.5, 28

IHC Ab CD14, CD31. Mean macrophage fractional area

Histamine release studies IHC Ab CD4, CD8, CD28, CD69, CD40L

IHC CD2, CD4, CD8, CD19, I3, RM3/1 IHC and IF Ab CD68, CD163, CD14, CD19, CD45, CD4, CD3, CD8, double staining: IFN-γ

Arthroplasty

OA (9)

Fernandez-madrid11

Mean HLA-DR+ similar RA. OKT9+ and ki-M8+ cells (macrophages) lower than in RA Less mature monocyte in intima, macrophages scattered intima, HLA-D-positive cells were synovial monocytes. Subset synovial cells was densly stained by monoclonal antibody 5.5 No neutrophils

Macrophages (LL) evident in inflamed ST samples. Mean macrophage fractional area 9.1%. Fractional area increases with increased inflammation

Significantly lower histamine content than in RA ST T cells major constituents of infiltrates. Aggregates: CD4+ >CD8+ cells. Some CD4+ cells were CD3−. Cellular activation marker (CD69) expressed by aggregated cells

Number MCs higher in superficial layers than in deeper layers. Number of MCs significantly lower than in RA ST, higher than normal ST. Ratio MCTC:MCT 3:4, RA 3:2

MNC infiltrate composed of B and T cells, predominantly T cells (CD4 >CD8) Macrophages found in I, SI and around lymphocytic clusters. T cells in SI. Fewer numbers of cells than RA. Some T cells stained for IFN-γ. In normal ST no cells stained for IFN-γ

Macrophages predominantly in LL, T cells mostly perivascular and in connective tissue layer not LL. Both T cells and macrophages in lower numbers than in RA

IHC Ab CD3, CD4, CD8, CD68, CD14, 27E10, RM3/1

OA (10), RA (10) Arthroplasty

Farahat, 199356

OA phenotype Methods Results/conclusion OA ST Knee arthroscopy and Toluidine blue staining and alcian MCs most numerous in SI, number of deregulated arthroplasty blue safranin staining MCs MCs was greater in the superficial layer. MC numbers comparable to RA ST, higher than in control ST

Arthropaties OA (36), RA (21), N (8)

Author Dean, 199355

Synovial inflammation, immune cells and their cytokines in OA

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

36 Knee arthroscopy

Knee arthroplasty Arthroplasty

OA (10)

OA (5)

OA (10), RA (14), N (6) OA (5)

Lebre, 200866

Lindblad, 19878

Lindblad, 198910

Mitchell, 200878

Oehler, 200233

Nakano, 200767

OA (12), RA (14), N (4) OA (66), RA (22), N (8)

IHC Ab CD3+

OA (10), RA (20) Arthroscopy Knee arthroscopy

OA (5), RA (5)

Kummer, 199465

Nakamura, 199912

IHC CD68+

OA (22), RA (19) Arthroplasty

Kopicky-Burd, 198864

Arthroscopy (29) and arthroplasty (37)

Knee arthroplasty

Knee Arthroplasty and arthroscopy

OA (17), RA (36), N (5)

Kraan, 199945

Supernatants ST. Western blot and IHC Ab tryptase ICH Ab CD3, CD4, CD8, CD20, CD68 and CD138

IHC Ab αLeu-1, αLeu-4, αLeu-2a, αLeu-3a, OKM1. Double staining with HLA-DR

Histology MCs, histamine content perchloric acid extraction Mab granzymes A and B. IHC double staining UCHT1, MT310, DK25, CD3, CD4, CD8, CD16, CD56, CD15 IHC Ab CD1c, CD304, CD3, CD8, CD11c, CD123 IHC Ab αLeu-1, αLeu-4, αLeu-2a, αLeu-3a, OKM1. Double staining with HLA-DR

IHC Ab CD3, CD4, CD8, CD68, CD22, CD38, CD55

IHC Ab leu-1, 2a, 3a cells

Knee or Hip arthroplasty Knee needle biopsy

OA (16), RA (6)

Johnell, 198530

Methods IHC Ab CD3, CD4, CD8

OA phenotype Arthroplasty

Arthropaties OA (10), RA (10), N (10)

Author Ishii, 200263

Lower number macrophages than in RA ST, more than in normal ST Mild or moderate infiltration of CD3+ T Cells observed in all samples in mainly perivascular areas MC tryptase activity similar in OA and RA, sign. higher than in control ST In ST with histological inflammatory features more B cells, T cells and plasma cells, compared with other histological patterns. More macrophages in lining LL than in RA ST or normal ST

Local inflammation with T cells, B cells, plasma cells and HLA-DR-expressing DCs adjacent to CD4+ T-helper cells

Both myeloid DCs and plasmacytoid DCs were observed in low numbers, mostly in SLL All cell types, but mostly macrophages and T cells (predominantly T-helper phenotype) were found. HLADR+ cells near αLeu-3a

MCs in SI often in perivascular locations. Histamine content not significantly different (OA 5.4, RA 3.7) 2/5 samples OA granzyme A and B, 3/5 in RA, granzyme A and B positive cells mostly CD16+, CD56+ NK cells

Leu-1 T cells and Leu-3a, leu-2a sparse. Leu-3a near HLA-DR+ cells. T cells RA ST >OA ST Mostly T cells (CD4+ >CD8+) and macrophages, few B cells and plasma cells. Lower number of cells than RA ST, more than normal ST

Results/conclusion OA ST More CD3+ and CD4+ positive T cells than normal ST, but less than in RA ST. T cells found in SLL and DL, RA in all ST layers

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 Chapter 2

Knee needle biopsies or arthroplasty

OA (25), RA (25), N (15)

Pessler, 2008, first44

Pessler, 2008, second34 OA (25), RA (28), N (10)

Knee arthroplasty

OA (19)

Saito, 200235

IHC Ab CD68, CD2, CD4, CD8, CD15, CD19, CD25, CD1a, toluidine blue for MC (n = 11)

Flow cytometry CD5+, CD69+ IHC (OA n = 9, RA n = 6) Ab CD5, CD69

Tendency to higher expression lymphocytes than normal ST, lower than RA ST. Lymphocyte recent activation phenotype comparable RA ST and significantly higher than normal ST Mostly T cells (CD4+> CD8+) and macrophages. 45% samples MCs in vicinity blood vessels. Macrophage/ helper T cell interaction might be involved in synovitis in OA

Knee arthroplasty

OA (19), RA (11), N (3)

Rollin, 200871

Revell, 19889

MCs found near small blood vessels. MC count significantly higher than in RA ST and normal ST Knee, hip arthroscopy IHC Ab 24, 44, OKM1, 10.1, OKT4, Macrophages found in most samples at different (3), arthroplasty (35) OKT8, Pan B, 52, 29 locations, T cells 50% samples in SLL (OKT4+ >OKT8+), few pan B cells

Prussian blue staining MCs

Knee arthroplasty

OA (13), RA (17), N (3) OA (38)

Pu, 199870

IHC Ab C-19, (Rel-B), CD-20, CD3, Perivascular MNC aggregates were small and few. CD68, HLA-DR Some had nRelB+ differentiated DCs, B cells, T cells and macrophages, 84% lacked nRelB+ diff DCs

Knee arthroplasty

QA: macrophages and T cells most common cells. Numbers lower than RA and higher than normal. Little B cells and plasma cells in OA ST

Results/conclusion OA ST Mostly T cells and macrophages, in LL. Little B cells and plasma cells and neutrophils. Intermediate cell densities in OA ST. CD68 distinguishes OA from RA and normal Lower number T cells than in RA ST, CD4+ >CD8+ cells. Relatively more CD28+ cells in older people, not sign Significant amount chronic inflammation. Significant amount IL-1 in MNC infiltrate SLL and at LL level Mostly macrophages in LL and SI (RC ± 65%) and T cells in SI (RC ± 22% (40% CD8+)), no/few neutrophils and B cells and plasma cells in SI (RC ± 5% B cells, plasma cells <1%). RC similar normal ST, RA more CD20, CD38+, less CD68+

OA (10), RA (18),

ICH Ab CD15, CD3, CD8, CD20, CD38, CD68. Quantitative assessments (QA)

T cell population. FACS. Ab CD3, CD4. CD8, CD28 Immunostaining studies. PAb human IL-1 (α and β) ICH Ab, CD3, CD20, CD38, CD68. Quantitative assessment (QA) and relative composition infiltrate (RC)

Methods IHC Ab CD3, CD15, CD38, CD20, CD68

Pettit, 200169

Pelletier, 1989113

OA (11), RA (11) Knee synovectomy and hip arthroplasty OA (8) Arthroplasty

Pawlowska, 200968

OA phenotype Knee needle biopsy, arthroscopy and arthroplasty

Arthropaties OA (31), RA (28), N (22)

Author Ogdie, 201021

Synovial inflammation, immune cells and their cytokines in OA

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

OA (19), RA (9), Arthroplasty N (9)

OA (11)

OA (6)

OA (8), RA (16), Knee arthroscopy or N (5) arthroplasty

OA (5), RA (8)

OA (7), RA (8)

OA (8), RA (9)

OA (12), RA (18) Arthroplasty

OA (18), RA (25) Arthroplasty or synovectomy

Sakkas, 200472

Scanzello, 200920

Shiokawa, 200173

Smith, 199237

Steiner, 199974

Warren, 199175

Weidler, 200476

Yamada, 201184

Yudoh, 200077

IHC Ab CD3, CD4, CD8, CD20, CD45RA, CD45RO, CD68, IL-2, IL-4, IL-6, IFN-γ, TNF-α, double staining IHC CD3, CD4, CD8, CD45, CD45RA, CD45RO mAB CD3, CD163, CD1a, CD1b, CD1c Isolation MNC, FACS IL-17A, CD57, CD45RO, CD28, IFN-γ, CD4, HLA-DR, CD69, CCR5 and intracellular staining Lymphocytes isolated. Intracellular staining. FACS. Ab IL-4, IFN-γ, IL-2, IL-10, subtyping T cells by production cytokines

Reverse transcriptase-PCR B-cell clonotypes DNA-cloning and sequencing IHC Ab CD3, CD8, CD4, IL-2r, CD25, CD5, CD14, CD64, CD11c

Digested ST. Flow cytometry, IF, FACS and IHC. Ab anti-CD3ζ, anti-CD3ε IHC Ab CD3, CD8, CD68

Methods IHC Ab CD3, CD69, CD25, CD45RO, HLAII, CD38, CD43

N, normal; diff, diffuse; IF, immunofluorescence; Ab, anti-bodies.∗ Number of patients.

Arthroplasty or synovectomy Arthroplasty

Knee arthroplasty

Knee arthroscopy (4) and arthroplasty (7)

Arthropaties OA phenotype OA (30), RA (13) Knee or hip arthroplasty

Author Sakkas, 199836

CD 4 T Cells expressed activation markers higher level than RA T cells. Th1 (IFN-γ) cells predominate in both OA as RA ST. Th17 (double producers IFN-γ and IL-17) were scarcely found TH1 (IFN-γ, no IL-4) /TH2 (IL-4, no IFN-γ) ratio 1.5, 6.1 in RA. Tr1 (IL-10, no IL2 and no IL-4), thought to inhibit Th1-respons, higher than in RA ST. Other subsets CD4+ comparable RA

Lower number of all investigated cells than in RA ST

Little lymphoid infiltrate

Results/conclusion OA ST 65% samples lymphoid cells aggregates, containing predominantly CD3+ T cells. Expression of activation antigens, although in fewer numbers than in RA ST Decreased CD3ζ protein relative to CD3ε. Suggests chronic T cells stimulation and confirms T cell involvement in OA Macrophages two distributions: in LL and scattered throughout SLL (similar in both groups) and lymphocytic accumulations (all arthroscopy samples, 57% arthroplasty samples) Infiltrating B-cells are oligoclonal (antigen-driven immune response may play a part in disease progress OA) Mostly T cells (SLL) and macrophages (LL). IHC changes in similar as in RA ST, suggesting quantitative rather than qualitative differences 80% samples only a few T cells, 20% had more T cells. CD4:CD8 ratio comparable to RA ST. No T cell cytokine expression in OA could be found

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 Chapter 2

Synovial inflammation, immune cells and their cytokines in OA

Cytokines associated with immune cells found in OA ST (Table III) Although cytokines were extensively studied in animal and experimental studies, and in synovial fluid from OA patients, studies that primary focus on cytokines in OA ST are less abundant17,18,20,87-91. Various cytokines associated with immune cells have been detected in OA ST17-20,26,39,52,56,57,63,75,85,88-105, and only a few studies reported no cytokine expression at all74,106. Cytokines detected in OA ST IL-1β and tumour necrosis factor alpha (TNF-α) are both pro-inflammatory cytokines and are most frequently studied and detected17-20,52,56,87,89-91,97,102,107. These cytokines were most frequently seen in LL52,102,107 and to a lesser amount in SLL52,90,102. Moreover, the presence of IL-1β-converting enzyme (ICE) indicates that IL-1β is not only present in OA ST, but that it can also be activated108. However, their cellular source is still largely unclear. Only one study used a double staining method for IL-1 and macrophage markers. In this study, macrophages were found positive for IL-1 and IL-1Ra. Likewise, several other pro- and anti-inflammatory cytokines have been investigated in OA19,20,26,39,56,57,63,74,75,77,85,87-90,92-95,97-101,103-106. Among these, interferon gamma (IFN-γ), IL-2, IL-4, IL-6, IL-8, IL-18, transforming growth factor (TGF)-β and IL-10 have received most attention. IFN-γ was detected by means of immunohistochemistry (IHC)26,57,63. The cellular origin of IFN-γ remains unclear. Although Yudoh et al. showed that ST T cells in OA can produce IFN-γ when stimulated ex vivo77. IHC studies using counterstaining for CD3 showed only very little57 or no IFNγ-positive T cells in situ74. These data indicate that IFNγ-producing T cells might not be activated in vivo and that other cells could be the main source of this cytokine in OA. IL-10 and IL-4 were also found in OA ST19,63,77,89. Yudoh et al. showed that ST T cells in OA can produce IL-10 and IL-4 when stimulated ex vivo77. Similarly to IFNγ, however, IHC using counterstaining for CD3 did not detect T cells positive for IL-4, indicating that T cells might not be the source of IL-4 in OA ST74. The cellular source of IL-10 has not been investigated yet. Ex vivo stimulated T cells from OA ST were also found to produce IL-277, although this cytokine could not be detected by means of IHC in ST74,97. In conclusion, the cytokines secreted by ST T cells in OA are still unknown. IL-6 has been detected in OA ST19,56,88. In the largest study by Doss et al., IL-6 was detected in LL88. Stainings for various cellular markers indicated that plasma cells are the main source of IL-688. Another cytokine associated with innate immunity, IL-8 was found to be produced in OA ST by means of IHC and enzyme linked immuno sorbent assay (ELISA) by several authors19,89,93. Deleuran et al. found IL-8 in OA samples predominantly in deeper layers and around vessels93. However, the cellular source of IL-8 remains unclear.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

OA (10), RA (10)

OA (8), RA (18), N (7)

OA (8), RA (18)

OA (11), RA (13), N (6) Arthroplasty

OA (8), RA (11)

OA (49)

OA (10), RA (10)

Cauli, 199752

Chu, 199199

Deleuran, 1992107

Deleuran, 199493

Dolhain, 199626

Doss, 200788

Farahat, 199356

Fonseca, 200257 OA (6), RA (6), N (3)

OA (41)

Brenner, 200487

IHC Ab IL-23 subsets p19 and p40. Total RNA extraction. RT-PCR IL23p19 and IL-23p40 IHC Ab TNF-α, IL-1β

Arthroplasty

Knee or hip arthroplasty Arthroplasty

Knee blind biopsy or arthroplasty

Knee arthroplasty and arthroscopy

IHC, IF Ab CD68, CD163, CD14, CD19, CD45, CD4, CD3, CD8, double staining: IFN-γ

IHC Ab IL-6 counterstaining for different cells IHC Ab IL-1α, IL-1β, IL-6, TNF-α, GM-CSF

IHC Ab IFN-γ and IFN-γR

IHC and IF Ab IL-8

IHC IL-1α, IL-1R1, IL-1Ra

Knee arthroscopy (10) vs arthroplasty (15) Arthroscopy RNA ST isolation, RT-PCR IL-6, IL1β, IL-1α, TNF-α Unknown Ab cytokines: IL-1α, IL-1β, IL-1Ra Double staining for macrophages and cytokines Arthroscopy IHC Ab TGF-β1

OA (25)

Benito, 200517

Synovectomy or arthroplasty

OA (8), RA (9)

Brentano, 200998

Fewer IL-1α and IL-1R1 in interstitium compare with RA. Similar percentage IL-1α in LL. Staining intensity IL-1α lower than in RA. IL-1Ra mostly LL as in RA IL-8 LL and in deeper layers around vessels. Numbers significantly higher than normal ST, lower than RA ST (not significant) IFN-γ+ cells in 88% samples and IFN-γR in 38% samples, mainly around blood vessels and SL. Number cells lower than in RA Plenty IL-6 cells in synovial LL. Counterstaining revealed IL-6 production by plasma cells Expression IL-1α, IL-1β, IL-6, TNF-α, GM-CSF. Intensities expressions significantly lower than in RA ST. Differences in cytokine production OA and RA quantitative, not qualitative Some T cells stained for IFN-γ. In normal ST no cells stained for IFN-γ

No TNF-α or IL-1α. IL-1β in 14 of 16 investigated samples and IL-6 in all 17 samples % 25F9+, IL-1Ra+ higher than IL-1α+ in SLL., IL-1+ α macrophages OA <RA. % 25F9+, IL-1Ra+ and RM3/1+, IL-1Ra+ similar OA and RA TGF-β1 detected, although in reduced quantities than in RA

TNF-α and IL-1β expression significantly increased in arthroscopy group compared with arthroplasty group

No/little IL-23 subsets. Subunits IL-23p19 and IL-23p40 not or weakly

Table III. Studies that investigated cytokines in ST in OA patients Author Arthropaties OA phenotype Methods Results/conclusion OA ST Arthroplasty Total RNA extraction, RT-PCR IL-10, IL-10 and IL-26 expression similar as in RA, IL-19 lower than in Alanärä, 201092 OA (10∗), RA (10) IL-19, IL-20, IL-22, IL-24, IL-26 RA ST. no IL-20 and IL-22 in OA

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 Chapter 2

Arthroplasty

Hip arthroplasty Arthroplasty

Knee arthroscopy Knee replacement surgery (19), arthroscopy (4) Knee arthroplasty

Heinhuis, 201194 OA (9), RA (15)

Hulejova, 200789 OA (55), N (10)

OA (10), RA (10), N (10)

OA (6), RA (12)

OA (10), RA (11) OA (5), RA (8)

OA (18), RA (6)

OA (23)

OA (6), N (7)

OA (16), RA (10)

Ishii, 200263

Jungel, 2004106

Kohno, 200895 Kragstrup, 2008101

Melchiorri, 1998102 Ning, 201190

Saha, 1999108

Sakkas, 199885 Arthroplasty or synovectomy

Arthroplasty Arthroplasty

Synovectomy or arthroplasty

Arthroplasty

OA (7), RA (18)

Gracie, 1999100

OA phenotype Knee or hip arthroplasty

Arthropaties OA (5), RA (5), N (5)

Author FuruzawaCarballeda, 199919

All isoforms detected IL-32 in OA. IL-32 γ significantly lower than in RA ST

Little IL-18 compared to RA ST. IL-18 detected in OA ST by RT-PCR

Results/conclusion OA ST Little IL-8 and IL-10 expression, lower than RA ST, higher than normal ST. RT-PCR IL-1β, TNF-α, IL-6, IL-8, IL-10, and TGF-β1

IHC Ab IL-12p40, IL-12p70, counterstaining Ab CD68, CD20. RNA extraction. RT-PCR IL-12p40

In situ hybridization and IHC ICE

IHC Ab IL1β, TGF-β. Comparison K/L 2 or 3 vs K/L 4

cDNA samples RT-PCR IL-17 IHC Ab IL-20 and IL-24. Total RNA extraction, RT-PCR IL-20, IL-24 and TNF-α IHC Ab IL-1β, TNF-α

IL-17 gene not different from RA Only few IL-20 positive cells. IL-24 staining present but weaker than in RA ST, with discreet staining of endothelial cells. IL-20 and IL-24 expression PRC same RA ST IL-1β and TNF-α expression mostly in LL and less in SLL. Expression both cytokines lower than in RA ST IL-1β expression in both LL and SLL, higher expression in patients with K/L 2or 3 than in patients with K/L 4, TGF-β higher in K/L 4 than in patients with K/L 2 or 3 Expression IL-1β-converting enzyme (ICE) was detected in OA ST 31% samples IL-12p70+ cells (synovial LCs and monocyte/ macrophages), no IL-12p40+. Presence IL-12p70 suggests an immunoregulatory role for IL-12. IL-12p40 transcripts detected, lower than in RA ST

IL-8 and IL-10 significantly higher than normal ST, TNF-α and IL-1α comparable to normal ST IFN-γ+ five-fold higher than IL-4+ cells. Number both cells ± three-fold lower than in RA. Normal ST no IFN-γ and IL-4 staining IHC Ab IL-21R and IL-21 total RNA Weak IL-21R expression 33% samples. No IL21 in RA and OA extraction. RT-PCR IL-21 and IL-21R ST. No IL-21 and IL-21R expression by RT-PCR

Total mRNA extraction. RT-PCR isoforms IL-32α, IL-32β, IL-32γ and IL-32δ Tissue extracts OA ST. ELISA. Ab IL-1α, IL-10, IL-8, TNFα IHC Ab IL-4 and IFN-γ

IHC Ab IL-18. Total RNA extraction OA (10), RA (20), RT-PCR IL-18

Methods IHC Ab IL-8, IL-10. Total extracted RNA. RT-PCR IL-1β, TNF-α, IL-4, IL6, IL-8, IL-10, IL-13, and TGF-β1

Synovial inflammation, immune cells and their cytokines in OA

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

OA (40), N (13)

OA (5), RA (8)

Smith, 199718

Steiner, 199974

Knee arthroscopy (13) and arthroplasty (27) Knee arthroplasty

Methods IHC Ab rec-IL-15. Total RNA extraction. RT-PCR IL-1β, TNF-α, IL-2, IL-6 IL-15 and IL-21 Western blot IL-18, IL-18R, IL18BP. Total RNA extraction. RT-PCR IL-18, IL-18R and IL-18BP IHC Ab IL-1α, IL-1β, IL-1Ra, TNF-α

N, normal; IF, immunofluorescence; Ab, anti-bodies. ∗ Number of patients.

Yudoh, 200077

Wassilew, 201091 Yamada, 201184

Warren, 199175

Thurkow, 199739 Wagner, 199797

Suzuki, 200696 Szekanecz, 1995104 Tanaka, 2001105

OA (8), RA (12), N (7)

Shao, 2009103

OA phenotype Knee arthroscopy (4) and arthroplasty (7) Unknown

IHC: Ab IL-2, IL-4, IL-6, IFN-γ, double staining method OA (12), RA (21) Arthroplasty Total RNA extraction. RT-PCR TNF-α OA (10), RA (10), N (4) Arthroplasty IHC Ab TGFβ1, CD11c, f VIII H&E staining combined with IHC OA (12), RA (24) Arthroplasty IHC IL-18, western blot IL-18 (RA = 10, OA = 6) total RNA extraction. RT-PCR IL-18, IL-18R, IL-18Rβ OA (9), RA (10) Knee needle biopsy IHC IL-15 Ab OA (5), RA (14) Arthroplasty, Total RNA extraction, RT-PCR IL-1β, synovectomy TNF-α, IL-2, IL-4, IL-5, IL-6 and IL-10 OA (7), RA (8) Arthroplasty or IHC IL-2. Total RNA extraction, PRC synovectomy IL-2 OA (12) Arthroplasty RNA extraction, RT-PCR IL-1β, TNF-α OA (12), RA (18) Arthroplasty Isolation MNC, FACS IL-17A, CD57, CD45RO, CD28, IFN-γ, CD4, HLADR, CD69, CCR5 and intracellular staining OA (18), RA (25) Arthroplasty or Lymphocytes isolated. Intracellular synovectomy staining. FACS. Ab IL-4, IFN-\γ, IL-2, IL-10

Arthropaties OA (11)

Author Scanzello, 200920

TH1 (IFN-γ, no IL-4) /TH2 (IL-4, no IFN-γ) ratio 1.5, 6.1 in RA. Tr1 (IL-10, no IL2 and no IL-4) higher than in RA ST

Little IL-15 expression compared with RA Expression IL-1β, TNF-α, IL-6 and IL-10 detected, expression IL-2, IL-4 and IL-5 not detected IHC: no IL-2 staining. mRNA IL-2 not increased above background Both IL-1β and TNF-α found in ST. No differences between OA and traumatic joint disorder samples CD 4 T Cells expressed activation markers higher level than RA T cells. Th1 (IFN-γ) cells predominate in both OA as RA ST. Th17 (double producers IFN-γ and IL-17) were scarcely found

TNF-α gene significantly higher in OA compared with RA Little TGF-β1 expression by LCs, macrophages and endothelial cells, less than RA ST, more than normal ST Small number IL-18 positive cells SLL lower than RA. All samples pro-IL-18, 67% no/weakly mature IL-18. IL-18 mRNA two of four samples. No IL-18R and IL-18Rβ.

Results/conclusion OA ST Rec-IL-15 staining in LL and endothelium. IL-1β, TNF-α, IL-6 and IL-15 similar groups. IL-21 sign higher in arthroplasty group. IL-2 detectable over half patients IL-18 and IL-18R expression higher than control, lower than RA ST. Lower IL-18 and IL-18R than RA ST, higher N ST. Higher IL-18BP than RA ST, lower than N ST In arthroplasty group higher production IL-1α, IL-1β and TNF-α than arthroscopy group. Similar IL-1Ra expression between OA groups and normal ST No T cell cytokine expression could be found

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 Chapter 2

Synovial inflammation, immune cells and their cytokines in OA

IL-18 and its precursor forms were also detected in several studies100,103,105 and IL-18producing cells were found predominantly in the SLL105. Finally, several studies found the presence of TGF-β in OA ST90,99,104. By means of H&E staining combined with IHC, it was found that macrophages were the primary immune cells expressing TGF-β104. For most of the cytokines described above, messenger Ribonucleic acid (mRNA) transcripts were also detected19,20,75,85,87,91,92,94,96-98,100,101,103,105,106. Additionally, some cytokines have only been investigated by means of mRNA analysis. Of those, IL-5, IL-13, IL-17, IL-19, IL-21, IL-26 and IL-3219,20,92,94,95,97 could be detected, whereas IL-22 could not be detected in OA ST92. Differences in cytokine profiles between arthropaties The overall conclusion emerging from the above-mentioned studies is that differences in cytokine expression between OA and RA are mainly quantitative, not qualitative19,56. In general, the number of cytokine-positive cells was lower in OA ST than in RA19,26,39,56,63,74,89,93,98105 and higher than in normal ST18,19,51,57,63,93,103,104, indicating that OA is characterized by less inflammation than RA. Only one study found similar TNF-α and IL-1α expression in OA ST compared with normal ST89. Cytokine profiles in different severity stages in OA Inflammation is believed to play a role in OA severity and progression. However, only very few studies exist that compare inflammation with different severity stages in OA. Moreover, only few cytokines were investigated in these studies. Moreover, conflicting findings exist in expression of IL-1 and TNF-α between ST of patients undergoing arthroscopy and patients undergoing arthroplasty. Benito et al. found a higher expression of both cytokines in the arthroscopy group17, which is supported by results by Ning et al. The latter reported a higher IL-1β expression in patients with less severe disease (Kellgren–Lawrence score 2 and 3 vs 4). In the same study TGF-β expression was found to be lower in patients with less severe disease90. In contrast Smith et al. reported higher expression of IL-1α, IL-1β and TNF-α in the arthroplasty group18, suggesting that additional studies are necessary for a firm conclusion.

DISCUSSION In this review we aimed at summarizing the published data regarding the presence and phenotype of immune cells and their cytokines in ST of OA patients. Our analyses revealed that synovitis is a common feature of OA and is usually characterized by the presence of infiltrating immune cells, such as macrophages, T cells and MCs. Likewise, among the cytokines investigated, the pro-inflammatory cytokines TNFα and IL-1β were most frequently

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 2

detected in OA ST. Based on our analyses, we will discuss the limitations of our study, as well as possible research directions that could facilitate understanding the role of the immune system in OA. Our search revealed that over 100 studies investigated and reported data on ST in relation with inflammatory markers. Because we intended to give a comprehensive overview of the literature, we included all studies in which more than five OA patients participated, regardless of the disease they focus on. This low number of patients, however, implies that some of the presented data needs confirmation in additional studies or in larger cohorts. A meta-analysis of the studies was unfortunately not feasible, due to differences in methods and outcome measurements (e.g., different semi-quantitative systems and quantitative measures (IHC cells/mm2 or cells/hpf or % infiltrate, % total ST)). Development of standardized methods for evaluating synovial inflammation could therefore be beneficial in the future. Another limitation of the studies published thus far is that, while the knee is the most investigated joint (68% of articles that reported joint site), several studies exist that investigated both knee and hip but combined the data instead of presenting them separately. This complicates the interpretation of the results, as it is currently unclear how similar STs from different anatomical positions are in OA. It is likely that differences will be present, due to different influences from neighbouring organs, such as the infrapatellar fat pad (IFP) in the knee, which is lacking in the hip. We and others have previously shown that the IFP is a source of inflammatory mediators and could influence the pathophysiological processes in the knee joint109-111. Comparative studies of different ST would be of great interest for our understanding of the disease. One of the most important conclusions of our review is that inflammation and synovitis are present in OA ST. More importantly, the features found by IHC seem to correlate with the inflammation observed by magnetic resonance imaging (MRI). Several histological features of ST (except oedema) and total composition score were significantly correlated (littlemoderate) with MRI synovitis grade. Much less is known, however, on the correlation between inflammation and clinical characteristics. In a study in 39 OA patients both function and pain were not associated with macroscopic and microscopic parameters of ST, visualized by H&E staining3. Because this is the only study that investigated correlation between clinical symptoms and ST, it remains largely unknown how features of OA ST translate to signs and symptom of OA in patients. Addressing this question in future studies will likely constitute a considerable step forwards to a better understanding and potentially treatment of OA patients. The role of the immune system and of cytokines in OA is still poorly understood. Although several studies have shown the presence of different immune cells in ST in patients with OA, only few have attempted to further characterize these cells phenotypically and functionally. For example, it is still unclear which cytokines are secreted by macrophages, T cells and

Synovial inflammation, immune cells and their cytokines in OA

MCs, the most abundant immune cell populations in OA ST. Moreover, it is unclear how their phenotype relates to clinical symptoms or radiological features. Pain is one of the most important features of clinical OA. Although the biological mechanisms involved in pain are still largely unclear, it has been suggested that local inflammation could play an important role16. This hypothesis is also supported by the finding that TNF-α concentration in synovial fluid was found to correlate with pain in knee OA patients112. In conclusion, more detailed knowledge of the immune cells and their cytokines in synovitis is important for a better understanding of clinical features in OA, including pain. Comparison between OA and other arthritides or between different clinical phases of OA (early and late) could also offer insight into the role of synovial inflammation in disease progression. Several studies summarized in this analysis have addressed this topic. Unfortunately, the data are contradictory probably due to different definitions used for “early” or “late” OA; some authors found synovitis and cytokine expression more pronounced in patients with “late” OA undergoing arthroplasty2,3,18, others declared the opposite17,90 or did not find difference16. Longitudinal studies in different OA populations might offer a more definite conclusion. Studies comparing OA with RA point largely to the same direction; the number of infiltrating immune cells as well as the expression of cytokines were higher in RA than in OA ST and both were higher than in normal ST. This is in line with the clinical observations that OA is less inflammatory than RA. However, there is one population of immune cells that is enriched in OA compared to RA, namely MCs. This indicates that the inflammation in OA may be less than in RA, but could be qualitatively different. This intriguing finding was confirmed by several studies and point to MCs as potentially important players in synovitis in OA. Future studies are needed to elucidate the role of MCs in OA. In conclusion, our study indicates that inflammation is commonly detectable in OA ST and is characterized by immune cell infiltration and cytokine secretion. This inflammation seems in some aspects qualitatively different from the inflammation in RA. Future studies are needed to elucidate the role of different immune cells types and their cytokines in the development and progression of OA and their association with the different clinical features of the disease. Based on our results we suggest the research agenda depicted in Table IV.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 2

Table IV. Research agenda for future investigations of ST in OA patients Research agenda •To investigate severity and histological features at different OA severity stages •To correlate histological synovial inflammation severity and histological features with clinical parameters (pain/function) •To investigate course/persistence of synovial inflammation during OA disease course in longitudinal studies •To investigate the role of synovial inflammation in disease progression in longitudinal studies •To further investigate cellular source of cytokines in ST of (different stages) OA patients •To further investigate different subtypes of cells in ST of (different stages) OA patients •To investigate presence of cells and cytokines in relation to disease severity and clinical features (pain/ function) in (different stages) OA patients

Author contributions Following authors participated in design of the study: de Lange-Brokaar BJE, Ioan-Facsinay A, van Osch GJVM, Zuurmond A-M, Schoones J, Toes REM, Huizinga TWJ, Kloppenburg M. Interpretation of data was done by following authors: de Lange-Brokaar BJE, Ioan-Facsinay A, Kloppenburg M. Drafting of manuscript was done by de Lange-Brokaar BJE, Ioan-Facsinay A, van Osch GJVM, Zuurmond A-M, Schoones J, Toes REM, Huizinga TWJ, Kloppenburg M. Final approval of manuscript was provided by following authors: de Lange-Brokaar BJE, Ioan-Facsinay A, van Osch GJVM, Zuurmond A-M, Schoones J, Toes REM, Huizinga TWJ, Kloppenburg M. Role of funding source Financial support was obtained from TI Pharma, however TI Pharma did not contribute to design, interpretation of data, drafting and final approval of the manuscript. Conflict of interests None. Acknowledgements The study was sponsored by TI Pharma, but TI Pharma did not contribute to design, interpretation of data, drafting and final approval of the manuscript.

Synovial inflammation, immune cells and their cytokines in OA

Appendix 1. Literature search details Database Strategies

Number Number of unique of references references

PubMed

1. (“osteoarthritis”[Majr] OR osteoarthritis[ti] OR osteoarthritic[ti] OR Osteoarthritides[ti] OR Osteoarthrosis[ti] OR Osteoarthroses[ti] OR “Degenerative Arthritides”[ti] OR “Degenerative Arthritis”[ti] OR (oa[ti] AND knee) OR arthrosis[ti] OR “degenerative joint disease”[ti] OR “degenerative joint diseases”[ti]) AND (“Synovial Membrane”[Majr:noexp] OR synovium[ti] OR “Synovial Membrane”[ti] OR “Synovial Membranes”[ti] OR “Synovial tissue”[ti] OR “synovial tissues”[ti] OR Synovialis[ti]) 2. (“osteoarthritis”[Majr] OR osteoarthritis[tiab] OR 1460 osteoarthritic[tiab] OR Osteoarthritides[tiab] OR Osteoarthrosis[tiab] OR Osteoarthroses[tiab] OR “Degenerative Arthritides”[tiab] OR “Degenerative Arthritis”[tiab] OR (oa[tiab] AND knee) OR arthrosis[tiab]) AND (“Synovial Membrane”[Majr:noexp] OR synovium[tiab] OR “Synovial Membrane”[tiab] OR “Synovial Membranes”[tiab] OR “Synovial tissue”[tiab] OR “synovial tissues”[tiab] OR Synovialis[tiab]) AND (inflammatory[tw] OR “inflammation”[mesh] OR inflammation[tw] OR “Synovitis”[mesh] OR synovitis[tw])

1460

EMBASE (OVIDversion)

1. (exp *osteoarthritis/OR (osteoarthritis OR osteoarthritic OR Osteoarthritides OR Osteoarthrosis OR Osteoarthroses OR “Degenerative Arthritides” OR “Degenerative Arthritis” OR arthrosis OR “degenerative joint disease” OR “degenerative joint diseases”).ti OR (oa.ti AND knee.mp)) AND (*synovium/ OR (“Synovial Membrane” OR synovium OR “Synovial Membrane” OR “Synovial Membranes” OR “Synovial tissue” OR “synovial tissues” OR Synovialis).ti) 2. (exp *osteoarthritis/OR (osteoarthritis OR osteoarthritic 2.089 OR Osteoarthritides OR Osteoarthrosis OR Osteoarthroses OR “Degenerative Arthritides” OR “Degenerative Arthritis” OR arthrosis OR “degenerative joint disease” OR “degenerative joint diseases”). ti,ab OR (oa.ti AND knee.ti,ab)) AND (*synovium/ OR (“Synovial Membrane” OR synovium OR “Synovial Membrane” OR “Synovial Membranes” OR “Synovial tissue” OR “synovial tissues” OR Synovialis).ti,ab) AND (exp *inflammation/ OR inflammation.ti OR inflammatory.ti OR exp *Synovitis/ OR synovitis.ti)

1018

Web of Science

1. TI=((osteoarthriti* OR Osteoarthritides OR Osteoarthrosis OR Osteoarthroses OR “Degenerative Arthritides” OR “Degenerative Arthritis” OR arthrosis OR (oa AND knee)) AND (synovium OR “Synovial Membrane” OR synovium OR “Synovial Membrane” OR “Synovial Membranes” OR “Synovial tissue” OR “synovial tissues” OR Synovialis)) 2. TS=((osteoarthriti* OR Osteoarthritides OR Osteoarthrosis OR Osteoarthroses OR “Degenerative Arthritides” OR “Degenerative Arthritis” OR arthrosis OR (oa AND knee)) AND (synovium OR “Synovial Membrane” OR synovium OR “Synovial Membrane” OR “Synovial Membranes” OR “Synovial tissue” OR “synovial tissues” OR Synovialis) AND (inflammat* OR synovitis))

1.271

382

4820

2.860

Total

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 2

REFERENCES

1 Felson DT. Clinical practice. Osteoarthritis of the knee. N Engl J Med 2006;354:841-8. 2 Myers SL, Brandt KD, Ehlich JW, Braunstein EM, Shelbourne KD, Heck DA, et al. Synovial inflammation in patients with early osteoarthritis of the knee. J Rheumatol 1990;17:1662-9. 3 Loeuille D, Chary-Valckenaere I, Champigneulle J, Rat AC, Toussaint F, Pinzano-Watrin A, et al. Macroscopic and microscopic features of synovial membrane inflammation in the osteoarthritic knee: correlating magnetic resonance imaging findings with disease severity. Arthritis Rheum 2005;52:3492-501. 4 Loeuille D, Rat AC, Goebel JC, Champigneulle J, Blum A, Netter P, et al. Magnetic resonance imaging in osteoarthritis: which method best reflects synovial membrane inflammation? Correlations with clinical, macroscopic and microscopic features. Osteoarthritis Cartilage 2009;17:1186-92. 5 Goldenberg DL, Cohen AS. Synovial membrane histopathology in the differential diagnosis of rheumatoid arthritis, gout, pseudogout, systemic lupus erythematosus, infectious arthritis and degenerative joint disease. Medicine (Baltimore) 1978;57:239-52. 6 Goldenberg DL, Egan MS, Cohen AS. Inflammatory synovitis in degenerative joint disease. J Rheumatol 1982;9:204-9. 7 Gedikoglu O, Bayliss MT, Ali SY, Tuncer I. Biochemical and histological changes in osteoarthritic synovial membrane. Ann Rheum Dis 1986;45:289-92. 8 Lindblad S, Hedfors E. Arthroscopic and immunohistologic characterization of knee joint synovitis in osteoarthritis. Arthritis Rheum 1987;30:1081-8. 9 Revell PA, Mayston V, Lalor P, Mapp P. The synovial membrane in osteoarthritis: a histological study including the characterisation of the cellular infiltrate present in inflammatory osteoarthritis using monoclonal antibodies. Ann Rheum Dis 1988;47:300-7. 10 Lindblad S. Arthroscopic and synovial correlates of pain in osteoarthritis. Semin Arthritis Rheum 1989;18:91-3. 11 Fernandez-Madrid F, Karvonen RL, Teitge RA, Miller PR, An T, Negendank WG. Synovial thickening detected by MR imaging in osteoarthritis of the knee confirmed by biopsy

as synovitis. Magn Reson Imaging 1995;13:17783. 12 Nakamura H, Yoshino S, Kato T, Tsuruha J, Nishioka K. T-cell mediated inflammatory pathway in osteoarthritis. Osteoarthritis Cartilage 1999;7:401-2. 13 Korkusuz P, Dagdeviren A, Eksioglu F, Ors U. Immunohistological analysis of normal and osteoarthritic human synovial tissue. Bull Hosp Jt Dis 2005;63:63-9. 14 Ayral X, Pickering EH, Woodworth TG, Mackillop N, Dougados M. Synovitis: a potential predictive factor of structural progression of medial tibiofemoral knee osteoarthritis Ââ&#x2C6;&#x2019; results of a 1 year longitudinal arthroscopic study in 422 patients. Osteoarthritis Cartilage 2005;13:361-7. 15 Walsh DA, Bonnet CS, Turner EL, Wilson D, Situ M, McWilliams DF. Angiogenesis in the synovium and at the osteochondral junction in osteoarthritis. Osteoarthritis Cartilage 2007;15:743-51. 16 Pearle AD, Scanzello CR, George S, Mandl LA, DiCarlo EF, Peterson M, et al. Elevated high-sensitivity C-reactive protein levels are associated with local inflammatory findings in patients with osteoarthritis. Osteoarthritis Cartilage 2007;15:516-23. 17 Benito MJ, Veale DJ, FitzGerald O, van den Berg WB, Bresnihan B. Synovial tissue inflammation in early and late osteoarthritis. Ann Rheum Dis 2005;64:1263-7. 18 Smith MD, Triantafillou S, Parker A, Youssef PP, Coleman M. Synovial membrane inflammation and cytokine production in patients with early osteoarthritis. J Rheumatol 1997;24:365-71. 19 Furuzawa-Carballeda J, Alcocer-Varela J. Interleukin-8, interleukin-10, intercellular adhesion molecule-1 and vascular cell adhesion molecule-1 expression levels are higher in synovial tissue from patients with rheumatoid arthritis than in osteoarthritis. Scand J Immunol 1999;50:215-22. 20 Scanzello CR, Umoh E, Pessler F, Diaz-Torne C, Miles T, Dicarlo E, et al. Local cytokine profiles in knee osteoarthritis: elevated synovial fluid interleukin-15 differentiates early from end-stage disease. Osteoarthritis Cartilage 2009;17:1040-8. 21 Ogdie A, Li J, Dai L, Paessler ME, Yu X, Diaz-Torne C, et al. Identification of broadly discriminatory

Synovial inflammation, immune cells and their cytokines in OA

tissue biomarkers of synovitis with binary and multicategory receiver operating characteristic analysis. Biomarkers 2010;15:183-90. Houli J, Roimicher S, Paciornik I, De PD. Synovial tissue in osteo-arthritis of the knee. Acta Rheumatol Scand 1959;5: 122-35. Arnoldi CC, Reimann I, Bretlau P. The synovial membrane in human coxathrosis: light and electron microscopic studies. Clin Orthop Relat Res 1980:213-20. Da RR, Qin Y, Baeten D, Zhang Y. B cell clonal expansion and somatic hypermutation of Ig variable heavy chain genes in the synovial membrane of patients with osteoarthritis. J Immunol 2007;178:557-65. Diaz-Torne C, Schumacher HR, Yu X, GomezVaquero C, Dai L, Chen LX, et al. Absence of histologic evidence of synovitis in patients with Gulf War veterans’ illness with joint pain. Arthritis Rheum 2007;57:1316-23. Dolhain RJ, ter Haar NT, Hoefakker S, Tak PP, de LM, Claassen E, et al. Increased expression of interferon (IFN)-gamma together with IFNgamma receptor in the rheumatoid synovial membrane compared with synovium of patients with osteoarthritis. Br J Rheumatol 1996;35:24-32. Haynes MK, Hume EL, Smith JB. Phenotypic characterization of inflammatory cells from osteoarthritic synovium and synovial fluids. Clin Immunol 2002;105:315-25. Haywood L, McWilliams DF, Pearson CI, Gill SE, Ganesan A, Wilson D, et al. Inflammation and angiogenesis in osteoarthritis. Arthritis Rheum 2003;48:2173-7. Huss RS, Huddleston JI, Goodman SB, Butcher EC, Zabel BA. Synovial tissue-infiltrating natural killer cells in osteoarthritis and peri-prosthetic inflammation. Arthritis Rheum 2010;62: 3799805. Johnell O, Hulth A, Henricson A. T-lymphocyte subsets and HLA-DR-expressing cells in the osteoarthritic synovialis. Scand J Rheumatol 1985;14:259-64. Koizumi F, Matsuno H, Wakaki K, Ishii Y, Kurashige Y, Nakamura H. Synovitis in rheumatoid arthritis: scoring of characteristic histopathological features. Pathol Int 1999;49:298-304. Krenn V, Morawietz L, Burmester GR, Kinne RW, Mueller-Ladner U, Muller B, et al. Synovitis score: discrimination between chronic lowgrade and high-grade synovitis. Histopathology 2006;49:358-64.

33 Oehler S, Neureiter D, Meyer-Scholten C, Aigner T. Subtyping of osteoarthritic synoviopathy. Clin Exp Rheumatol 2002;20: 633-40. 34 Pessler F, Dai L, Diaz-Torne C, GomezVaquero C, Paessler ME, Zheng DH, et al. The synovitis of “non-inflammatory” orthopaedic arthropathies: a quantitative histological and immunohistochemical analysis. Ann Rheum Dis 2008;67:1184-7. 35 Saito I, Koshino T, Nakashima K, Uesugi M, Saito T. Increased cellular infiltrate in inflammatory synovia of osteoarthritic knees. Osteoarthritis Cartilage 2002;10:156-62. 36 Sakkas LI, Scanzello C, Johanson N, Burkholder J, Mitra A, Salgame P, et al. T cells and T-cell cytokine transcripts in the synovial membrane in patients with osteoarthritis. Clin Diagn Lab Immunol 1998;5:430-7. 37 Smith MD, O’Donnell J, Highton J, Palmer DG, Rozenbilds M, Roberts-Thomson PJ. Immunohistochemical analysis of synovial membranes from inflammatory and noninflammatory arthritides: scarcity of CD5 positive B cells and IL2 receptor bearing T cells. Pathology 1992;24:19-26. 38 Soren A, Klein W, Huth F. The synovial changes in posttraumatic synovitis and osteoarthritis. Rheumatol Rehabil 1978;17:38-45. 39 Thurkow EW, van der Heijden IM, Breedveld FC, Smeets TJ, Daha MR, Kluin PM, et al. Increased expression of IL-15 in the synovium of patients with rheumatoid arthritis compared with patients with Yersinia-induced arthritis and osteoarthritis. J Pathol 1997;181:444-50. 40 Fonseca JE, Canhao H, Resende C, Saraiva F, da Costa JC, Pimentao JB, et al. Histology of the synovial tissue: value of semiquantitative analysis for the prediction of joint erosions in rheumatoid arthritis. Clin Exp Rheumatol 2000;18:559-64. 41 Slansky E, Li J, Haupl T, Morawietz L, Krenn V, Pessler F. Quantitative determination of the diagnostic accuracy of the synovitis score and its components. Histopathology 2010;57:43643. 42 Scanzello CR, McKeon B, Swaim BH, Dicarlo E, Asomugha EU, Kanda V, et al. Synovial inflammation in patients undergoing arthroscopic meniscectomy: molecular characterization and relationship to symptoms. Arthritis Rheum 2011;63:391-400. 43 Soren A, Cooper NS, Waugh TR. The nature and designation of osteoarthritis determined by its histopathology. Clin Exp Rheumatol 1988;6:416.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 2

44 Pessler F, Chen LX, Dai L, Gomez-Vaquero C, Diaz-Torne C, Paessler ME, et al. A histomorphometric analysis of synovial biopsies from individuals with Gulf War Veterans’ illness and joint pain compared to normal and osteoarthritis synovium. Clin Rheumatol 2008;27:1127-34. 45 Kraan MC, Haringman JJ, Post WJ, Versendaal J, Breedveld FC, Tak PP. Immunohistological analysis of synovial tissue for differential diagnosis in early arthritis. Rheumatology 1999;38:1074-80. 46 Krenn V, Morawietz L, Haupl T, Neidel J, Petersen I, Konig A. Grading of chronic synovitis − a histopathological grading system for molecular and diagnostic pathology. Pathol Res Pract 2002;198:317-25. 47 Andreu JL, Trujillo A, Alonso JM, Mulero J, Martinez AC. Selective expansion of T cells bearing the/receptor and expressing an unusual repertoire in the synovial membrane of patients with rheumatoid arthritis. Arthritis Rheum 1991;34:808-14. 48 Bondeson J, Wainwright SD, Lauder S, Amos N, Hughes CE. The role of synovial macrophages and macrophage-produced cytokines in driving aggrecanases, matrix metalloproteinases, and other destructive and inflammatory responses in osteoarthritis. Arthritis Res Ther 2006;8:R187. 49 Bridges AJ, Malone DG, Jicinsky J, Chen M, Ory P, Engber W, et al. Human synovial mast cell involvement in rheumatoid arthritis and osteoarthritis. Relationship to disease type, clinical activity, and antirheumatic therapy. Arthritis Rheum 1991;34:1116-24. 50 Buckley MG, Gallagher PJ, Walls AF. Mast cell subpopulations in the synovial tissue of patients with osteoarthritis: selective increase in numbers of tryptase-positive, chymasenegative mast cells. J Pathol 1998;186:67-74. 51 Cannons JL, Karsh J, Birnboim HC, Goldstein R. hprt-mutant T cells in the peripheral blood and synovial tissue of patients with rheumatoid arthritis. Arthritis Rheum 1998;41:1772-82. 52 Cauli A, Yanni G, Panayi GS. Interleukin-1, interleukin-1 receptor antagonist and macrophage populations in rheumatoid arthritis synovial membrane. Br J Rheumatol 1997;36:935-40. 53 Cauli A, Pitzalis C, Yanni G, Awad M, Panayi GS. CD1 expression in psoriatic and rheumatoid arthritis. Rheumatology (Oxford) 2000;39:66673.

54 Damsgaard TE, Sorensen FB, Herlin T, Schiotz PO. Stereological quantification of mast cells in human synovium. APMIS 1999;107:311-7. 55 Dean G, Hoyland JA, Denton J, Donn RP, Freemont AJ. Mast cells in the synovium and synovial fluid in osteoarthritis. Br J Rheumatol 1993;32:671-5. 56 Farahat MN, Yanni G, Poston R, Panayi GS. Cytokine expression in synovial membranes of patients with rheumatoid arthritis and osteoarthritis. Ann Rheum Dis 1993; 52:870-5. 57. Fonseca JE, Edwards JCW, Blades S, Goulding NJ. Macrophage subpopulations in rheumatoid synovium: reduced CD163 expression in CD4þ T lymphocyte-rich microenvironments. Arthritis Rheum 2002;46:1210-6. 58 Fritz P, Reiser H, Saal JG. Analysis of mast cells in rheumatoid arthritis and osteoarthritis by an avidin-peroxidase staining. Virchows Arch B Cell Pathol Incl Mol Pathol 1984;47:35-45. 59 Gotis-Graham I, McNeil HP. Mast cell responses in rheumatoid synovium. Association of the MCTC subset with matrix turnover and clinical progression. Arthritis Rheum 1997;40:479-89. 60 Gruber B, Poznansky M, Boss E. Characterization and functional studies of rheumatoid synovial mast cells. Activation by secretagogues, antiIgE, and a histamine-releasing lymphokine. Arthritis Rheum 1986;29:944-55. 61 Helbig B, Gross WL, Borisch B, Starz H, MullerHermelink HK. Characterization of synovial macrophages by monoclonal antibodies in rheumatoid arthritis and osteoarthritis. Scand J Rheumatol Suppl 1988;76:61-6. 62 Hogg N, Palmer DG, Revell PA. Mononuclear phagocytes of normal and rheumatoid synovial membrane identified by monoclonal antibodies. Immunology 1985;56:673-81. 63 Ishii H, Tanaka H, Katoh K, Nakamura H, Nagashima M, Yoshino S. Characterization of infiltrating T cells and Th1/Th2-type cytokines in the synovium of patients with osteoarthritis. Osteoarthritis Cartilage 2002;10:277-81. 64 Kopicky-Burd JA, Kagey-Sobotka A, Peters SP, Dvorak AM, Lennox DW, Lichtenstein LM, et al. Characterization of human synovial mast cells. J Rheumatol 1988;15:1326-33. 65 Kummer JA, Tak PP, Brinkman BM, van Tilborg AA, Kamp AM, Verweij CL, et al. Expression of granzymes A and B in synovial tissue from patients with rheumatoid arthritis and osteoarthritis. Clin Immunol Immunopathol 1994;73:88-95.

Synovial inflammation, immune cells and their cytokines in OA

66 Lebre MC, Jongbloed SL, Tas SW, Smeets TJ, McInnes IB, Tak PP. Rheumatoid arthritis synovium contains two subsets of CD83-DC. Am J Pathol 2008;172:940-50. 67 Nakano S, Mishiro T, Takahara S, Yokoi H, Hamada D, Yukata K, et al. Distinct expression of mast cell tryptase and protease activated receptor-2 in synovia of rheumatoid arthritis and osteoarthritis. Clin Rheumatol 2007;26:1284-92. 68 Pawlowska J, Mikosik A, Soroczynska-Cybula M, Jozwik A, Luczkiewicz P, Mazurkiewicz S, et al. Different distribution of CD4 and CD8 T cells in synovial membrane and peripheral blood of rheumatoid arthritis and osteoarthritis patients. Folia Histochem Cytobiol 2009;47:627-32. 69 Pettit AR, Ahern MJ, Zehntner S, Smith MD, Thomas R. Comparison of differentiated dendritic cell infiltration of autoimmune and osteoarthritis synovial tissue. Arthritis Rheum 2001;44:105-10. 70 Pu J, Nishida K, Inoue H, Asahara H, Ohtsuka A, Murakami T. Mast cells in osteoarthritic and rheumatoid arthritic synovial tissues of the human knee. Acta Med Okayama 1998;52:359. 71 Rollin R, Marco F, Jover JA, Garcia-Asenjo JA, Rodriguez L, Lopez-Duran L, et al. Early lymphocyte activation in the synovial microenvironment in patients with osteoarthritis: comparison with rheumatoid arthritis patients and healthy controls. Rheumatol Int 2008;28:757-64. 72 Sakkas LI, Koussidis G, Avgerinos E, Gaughan J, Platsoucas CD. Decreased expression of the CD3zeta chain in T cells infiltrating the synovial membrane of patients with osteoarthritis. Clin Diagn Lab Immunol 2004;11:195-202. 73 Shiokawa S, Matsumoto N, Nishimura J. Clonal analysis of B cells in the osteoarthritis synovium. Ann Rheum Dis 2001;60:802-5. 74 Steiner G, Tohidast-Akrad M, Witzmann G, Vesely M,Studnicka-Benke A, Gal A, et al. Cytokine production by synovial T cells in rheumatoid arthritis. Rheumatology 1999;38:202-13. 75 Warren CJ, Howell WM, Bhambhani M, Cawley MID, Smith JL. An investigation of T-cell subset phenotype and function in the rheumatoid synovium using in situ hybridization for IL-2 mRNA. Immunology 1991;72:250-5. 76 Weidler C, Kroll R, Miller LE, Scholmerich J, Grifka J, Straub RH. Low density of CD1Ăž cells in

the synovial tissue of patients with rheumatoid arthritis. Clin Exp Rheumatol 2004;22:433-40. Yudoh K, Matsuno H, Nakazawa F, Yonezawa T, Kimura T. Reduced expression of the regulatory CD4Ăž T cell subset is related to Th1/Th2 balance and disease severity in rheumatoid arthritis. Arthritis Rheum 2000;43:617-27. Mitchell A, Rentero C, Endoh Y, Hsu K, Gaus K, Geczy C, et al. LILRA5 is expressed by synovial tissue macrophages in rheumatoid arthritis, selectively induces pro-inflammatory cytokines and IL-10 and is regulated by TNF-alpha, IL-10 and IFN-gamma. Eur J Immunol 2008;38:345973. Ceponis A, Konttinen YT, Takagi M, Xu JW, Sorsa T, Matucci-Cerinic M, et al. Expression of stem cell factor (SCF) and SCF receptor (c-kit) in synovial membrane in arthritis: correlation with synovial mast cell hyperplasia and inflammation. J Rheumatol 1998;25:2304-14. Schulze-Koops H, Kalden JR. The balance of Th1/Th2 cytokines in rheumatoid arthritis. Best Pract Res Clin Rheumatol 2001;15:677-91. Packard KA, Khan MM. Effects of histamine on Th1/Th2 cytokine balance. Int Immunopharmacol 2003;3:909e20. 82. Battaglia M, Gregori S, Bacchetta R, Roncarolo MG. Tr1 cells: from discovery to their clinical application. Semin Immunol 2006;18:120-7. Harrington LE, Mangan PR, Weaver CT. Expanding the effector CD4 T-cell repertoire: the Th17 lineage. Curr Opin Immunol 2006;18:349-56. Yamada H, Nakashima Y, Okazaki K, Mawatari T, Fukushi J, Oyamada A, et al. Preferential accumulation of activated Th1 cells not only in rheumatoid arthritis but also in osteoarthritis joints. J Rheumatol 2011;38:1569-75. Sakkas LI, Johanson NA, Scanzello CR, Platsoucas CD. Interleukin-12 is expressed by infiltrating macrophages and synovial lining cells in rheumatoid arthritis and osteoarthritis. Cell Immunol 1998;188:105-10. Irani AA, Schechter NM, Craig SS, DeBlois G, Schwartz LB. Two types of human mast cells that have distinct neutral protease compositions. Proc Natl Acad Sci U S A 1986;83:4464-8. Brenner SS, Klotz U, Alscher DM, Mais A, Lauer G, Schweer H, et al. Osteoarthritis of the knee - clinical assessments and inflammatory markers. Osteoarthritis Cartilage 2004;12:46975. Doss F, Menard J, Hauschild M, Kreutzer HJ, Mittlmeier T, Muller-Steinhardt M, et al.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 2

Elevated IL-6 levels in the synovial fluid of osteoarthritis patients stem from plasma cells. Scand J Rheumatol 2007;36:136-9. 89 Hulejova H, Baresova V, Klezl Z, Polanska M, Adam M, Senolt L. Increased level of cytokines and matrix metalloproteinases in osteoarthritic subchondral bone. Cytokine 2007;38:151-6. 90 Ning L, Ishijima M, Kaneko H, Kurihara H, Arikawa-Hirasawa E, Kubota M, et al. Correlations between both the expression levels of inflammatory mediators and growth factor in medial perimeniscal synovial tissue and the severity of medial knee osteoarthritis. Int Orthop 2011;35:831-8. 91 Wassilew GI, Lehnigk U, Duda GN, Taylor WR, Matziolis G, Dynybil C. The expression of proinflammatory cytokines and matrix metalloproteinases in the synovial membranes of patients with osteoarthritis compared with traumatic knee disorders. Arthroscopy 2010;26:1096-104. 92 Alanara T, Karstila K, Moilanen T, Silvennoinen O, Isomaki P. Expression of IL-10 family cytokines in rheumatoid arthritis: elevated levels of IL-19 in the joints. Scand J Rheumatol 2010;39:118-26. 93 Deleuran B, Lemche P, Kristensen M, Chu CQ, Field M, Jensen J, et al. Localisation of interleukin 8 in the synovial membrane, cartilageepannus junction and chondrocytes in rheumatoid arthritis. Scand J Rheumatol 1994;23:2-7. 94 Heinhuis B, Koenders MI, van de Loo FA, Netea MG, van den Berg WB, Joosten LA. Inflammation-dependent secretion and splicing of IL-32{gamma} in rheumatoid arthritis. Proc Natl Acad Sci U S A 2011;108:4962-7. 95 Kohno M, Tsutsumi A, Matsui H, Sugihara M, Suzuki T, Mamura M, et al. Interleukin-17 gene expression in patients with rheumatoid arthritis. Mod Rheumatol 2008;18:15-22. 96 Suzuki E, Tsutsumi A, Sugihara M, Mamura M, Goto D, Matsumoto I, et al. Expression of TNF-alpha, tristetraprolin, T-cell intracellular antigen-1 and Hu antigen R genes in synovium of patients with rheumatoid arthritis. Int J Mol Med 2006;18:273-8. 97 Wagner S, Fritz P, Einsele H, Sell S, Saal JG. Evaluation of synovial cytokine patterns in rheumatoid arthritis and osteoarthritis by quantitative reverse transcription polymerase chain reaction. Rheumatol Int 1997;16:191-6. 98 Brentano F, Ospelt C, Stanczyk J, Gay RE, Gay S, Kyburz D. Abundant expression of the

100

101

102

103

104

105

106

107

interleukin (IL)23 subunit p19, but low levels of bioactive IL23 in the rheumatoid synovium: differential expression and Toll-like receptor(TLR) dependent regulation of the IL23 subunits, p19 and p40, in rheumatoid arthritis. Ann Rheum Dis 2009;68:143-50. Chu CQ, Field M, Abney E, Zheng RQH, Allard S, Feldmann M, et al. Transforming growth factorbeta1 in rheumatoid synovial membrane and cartilage/pannus junction. Clin Exp Immunol 1991;86:380-6. Gracie JA, Forsey RJ, Chan WL, Gilmour A, Leung BP, Greer MR, et al. A proinflammatory role for IL-18 in rheumatoid arthritis. J Clin Invest 1999;104:1393-401. Kragstrup TW, Otkjaer K, Holm C, Jorgensen A, Hokland M, Iversen L, et al. The expression of IL-20 and IL-24 and their shared receptors are increased in rheumatoid arthritis and spondyloarthropathy. Cytokine 2008;41:16-23. Melchiorri C, Meliconi R, Frizziero L, Silvestri T, Pulsatelli L, Mazzetti I, et al. Enhanced and coordinated in vivo expression of inflammatory cytokines and nitric oxide synthase by chondrocytes from patients with osteoarthritis. Arthritis Rheum 1998;41:2165-74. Shao XT, Feng L, Gu LJ, Wu LJ, Feng TT, Yang YM, et al.Expression of interleukin-18, IL-18BP, and IL-18R in serum, synovial fluid, and synovial tissue in patients with rheumatoid arthritis. Clin Exp Med 2009;9:215-21. Szekanecz Z, Haines GK, Harlow LA, Shah MR, Fong TW, Fu R, et al. Increased synovial expression of transforming growth factor (TGF)-beta receptor endoglin and TGF-beta 1 in rheumatoid arthritis: possible interactions in the pathogenesis of the disease. Clin Immunol Immunopathol 1995;76:187-94. Tanaka M, Harigai M, Kawaguchi Y, Ohta S, Sugiura T, Takagi K, et al. Mature form of interleukin 18 is expressed in rheumatoid arthritis synovial tissue and contributes to interferon production by synovial T cells. J Rheumatol 2001;28:1779-87. Jungel A, Distler JHW, Kurowska-Stolarska M, Seemayer CA, Seibl R, Forster A, et al. Expression of interleukin-21 receptor, but not interleukin-21, in synovial fibroblasts and synovial macrophages of patients with rheumatoid arthritis. Arthritis Rheum 2004;50:1468-76. Deleuran BW, Chu CQ, Field M, Brennan FM, Katsikis P, Feldmann M, et al. Localization of interleukin-1alpha, type 1 interleukin-1

Synovial inflammation, immune cells and their cytokines in OA

108

109

110

111

112

113

receptor and interleukin-1 receptor antagonist in the synovial membrane and cartilage/ pannus junction in rheumatoid arthritis. Br J Rheumatol 1992;31:801-9. Saha N, Moldovan F, Tardif G, Pelletier JP, Cloutier JM, Martel-Pelletier J. Interleukin1beta-converting enzyme/caspase-1 in human osteoarthritic tissues: localization and role in the maturation of interleukin-1beta and interleukin-18. Arthritis Rheum 1999;42:157787. Clockaerts S, Bastiaansen-Jenniskens YM, Runhaar J, van Osch GJ, Van Offel JF, Verhaar JA, et al. The infrapatellar fat pad should be considered as an active osteoarthritic joint tissue: a narrative review. Osteoarthritis Cartilage 2010;18:876-82. Klein-Wieringa IR, Kloppenburg M, BastiaansenJenniskens YM, Yusuf E, Kwekkeboom JC, El-Bannoudi H, et al. The infrapatellar fat pad of patients with osteoarthritis has an inflammatory phenotype. Ann Rheum Dis 2011;70:851-7. Bastiaansen-Jenniskens YM, Clockaerts S, Feijt C, Zuurmond AM, Stojanovic-Susulic V, Bridts C, et al. Infrapatellar fat pad of patients with end-stage osteoarthritis inhibits catabolic mediators in cartilage. Ann Rheum Dis 2012;71:288-94. Orita S, Ishikawa T, Miyagi M, Ochiai N, Inoue G, Eguchi Y, et al. Pain-related sensory innervation in monoiodoacetateinduced osteoarthritis in rat knees that gradually develops neuronal injury in addition to inflammatory pain. BMC Musculoskelet Disord 2011;12:134. Pelletier JP, Martel-Pelletier J. Evidence for the involvement of interleukin 1 in human osteoarthritic cartilage degradation: protective effect of NSAID. J Rheumatol Suppl1989;18:19-27.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

CHAPTER 3 INFLAMMATORY CELLS IN END STAGE KNEE OSTEOARTHRITIS PATIENTS: A COMPARISON BETWEEN THE SYNOVIUM AND THE INFRAPATELLAR FAT PAD (IFP)

de Lange-Brokaar BJE*, Klein-Wieringa IR*, Yusuf E, Andersen SN, Kwekkeboom JC, Kroon HM, van Osch GJVM, Zuurmond A-M, Stojanovic-Susulic V, Nelissen RGHH, Toes REM, Kloppenburg M, Ioan-Facsinay A

Submitted * These authors contributed equally to the manuscript

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 3

ABSTRACT Introduction: Synovitis has been associated with radiographic progression of osteoarthritis (OA) and pain, suggesting its involvement in disease pathophysiology. To get a better understanding of inflammatory pathways active in the OA joint, we characterized and compared inflammatory cells in synovium and infrapatellar fat pad (IFP) of knee OA patients. Methods: Infiltrating immune cells were characterized by flow cytometry in 76 patients with knee OA from whom synovial tissue (n = 34) and IFP (n = 59) samples were obtained. Pain and Kellgren-Lawrence scores were determined using validated methods. Results: Macrophages and T cells, followed by mast cells, were the most predominant immune cells in the inflamed synovia and IFP and they were equally abundant in these tissues. Interestingly, both macrophages and T cells secreted mostly pro-inflammatory cytokines even without additional stimulation, indicating their activated state. In line with this, most CD4+ T cells had a memory phenotype and contained a significant population of cells expressing activation markers (CD25+, CD69+). Interestingly, synovial CD4+ T cells had a more activated phenotype than IFP T cells. Preliminary analyses indicated that the number of synovial CD4+ T cells were associated with visual analog scale (VAS) pain (0.55 (0.09â&#x20AC;&#x201C;1.02), p = 0.02). Conclusions: Our data suggest that the immune cell composition of synovium and IFP is similar and includes activated cells that could contribute to inflammation through secretion of pro-inflammatory cytokines. Moreover, preliminary analyses indicate that synovial CD4+ T cells might associate with VAS pain in patients with end-stage OA of the knee.

Inflammatory cells in end stage knee OA patients

INTRODUCTION Knee Osteoarthritis (OA) is a joint disease characterized by radiographic damage and pain. During the last decade, it became clear that synovial inflammation is present in a significant number of OA patients. In several studies, inflammation has been associated with pain13 and radiographic progression4,5 of the disease, indicating it as an important player in diseases pathogenesis. The inflamed synovium of OA patients is infiltrated with immune cells, which can secrete various inflammatory mediators (reviewed in6) that could mediate the association between inflammation and pain/radiographic progression. Another source of immune cells and inflammatory mediators in the joint is the infrapatellar fat pad (IFP)7-9. Since the IFP is situated adjacent to synovium, it is conceivable that a cross-talk exists between these tissues and that IFP could contribute directly or indirectly to disease pathogenesis. The association between joint inflammation and pain/radiographic progression is primarily established by imaging using (contrast-enhanced) MRI. Signal alterations have been reported both at synovium and IFP level and have been reported to be associated with radiographic damage10,11 and pain1-3,12. However, it is still unclear how these signal alterations relate to histological changes and cellular infiltrate in these two tissues and whether inflammatory changes in synovium and IFP are related. Moreover, the relative contribution of these two tissues to clinical parameters such as radiographic damage and pain in OA remains to be established. To get a better insight into inflammatory changes in these tissues, we compared the abundance and phenotype of the main immune cells populations in synovium and IFP. Furthermore, we used this knowledge to get insight into the association of different immune cell populations with radiographic damage and pain in a hypotheses generating setting using a well-defined cohort of knee OA patients.

METHODS Patients and study design Patients with osteoarthritis (OA) (n = 76), undergoing joint replacement surgery were recruited into the study and synovial and IFP samples were obtained and immediately processed. Forty two of the patients included in the present study are part of the ongoing geMstoan study (GEneration of Models, Mechanism & Markers for STratification of OsteoArthritis patieNts), an observational study in knee OA patients to find new biomarkers for OA progression. The geMstoan patients were included between 2008 and 2013, had symptomatic radiographic

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 3

primary knee OA, following the ACR criteria13, and attended the orthopaedic department of the LUMC or the orthopaedic department of the Diaconessenhuis in Leiden. In this study, patients with other rheumatic diseases, using immunosuppressive drugs or having knee injections (corticosteroids, etc) in the past 3 months were excluded. Written informed consent is available from all geMstoan patients. From the remaining 34 patients, leftover synovial tissues and IFP samples were obtained during arthroplasty, thus no additional clinical data were available, besides diagnosis, age, gender and BMI. Both studies were approved by the ethical committee of the Leiden University Medical Center (LUMC). Isolation of synovial cells and the IFP-derived stromal vascular fraction (SVF) fraction. Synovial tissue was cut and subsequently digested with collagenase type II (Sigma, Germany) for 90 min 37°C under continuous rolling. The digested tissue was filtered through a cell strainer with a pore size of 70μm with and the cell suspension was stained with the appropriate antibodies. Isolation of SVF cells from IFP was performed as previously described 7. Flow cytometric analysis For surface staining, 100,000 SVF or synovial cells were stained with mixes of the following antibodies (Abs): phycoerythrin (PE)-conjugated CD3; fluorescein isothiocyanate (FITC)conjugated CD45, CD8 and CD45RA; PacificBlue (PB)-conjugated-CD4; Allophycocyanin (APC)-conjugated CD117 and CD8; and phycoerythrin-Cy-7 (PE-CY7)-conjugated CD14, CD25 (all Abs were from BD Biosciences, The Netherlands). All incubations were performed at 4°C, for 30 min. For intracellular cytokine staining, SVF or synovial cells were plated o/n in medium containing 50 IU/ml IL-2 in the presence or absence of 3 μg/ml brefeldin A. The next day, cells cultured in the absence of brefeldin A were activated with 20ng/ml PMA and 200ng/ml ionomycin for 5 hours and 10 μg/ml brefeldin A was added the last 4 hours of culture. Cells cultured in the presence of brefeldin A were stained directly the next day. Approximately 400,000 SVF or synovial cells were stained using the BD intracellular cytokine fixation/permeabilization solution kit (BD Biosciences, The Netherlands), according to the manufacturer’s instructions. The following antibodies were used: AlexaFluor700 (AF700)conjugated CD3; PB-conjugated CD4; APC-conjugated CD8; FITC-conjugated CD45RA; PE-CY-7-conjugated CD14 and IL-4; PE-conjugated IFNγ, TNFα, IL-10 and IL-6 (all from BD Biosciences, The Netherlands). Cells were fixed with 1% paraformaldehyde and analyzed with a LSR II flow cytometer using Diva 6 software (BD Biosciences). Dead cells were excluded based both on FSC/SSC and staining with the Dead cell discrimination kit (Miltenyi Biotec, Germany), as indicated. Single cells were selected based on FSC/SSC analyses, as described in the gating strategy.

Inflammatory cells in end stage knee OA patients

Scoring knee radiographs In the geMstoan patients X-rays of the affected knees (posterior anterior (PA) fixed flexion) were obtained and scored, blinded for patient characteristics, by an experienced musculoskeletal radiologist (HK) according the Kellgren- Lawrence (KL) scale14. The intraclass correlation coefficient (ICC, with 95% confidence interval) was based on a randomly selected sample of 36 radiographs (18 right and 18 left knees) and was 0.99 (0.98-0.99). The knees with KL < 2 were rescored in consensus between HK and an experienced rheumatologic reader (MK). Clinical data In geMstoan patients, demographics and disease characteristic were collected via standard questionnaires. Measurement of pain was performed preoperatively using 3 different questionnaires investigating various dimensions of pain for the affected knee. First, general assessment of self-reported pain was assessed by the visual analogue scale (VAS, 0-100) for the affected knee, a one-dimensional measure of pain intensity, in which a score of 100mm represents worst possible pain.. Second, the measure of Intermittent and Constant OsteoArthritis Pain (ICOAP) 15,16 was used. Higher scores indicate worst pain experience. Thirdly, the Knee injury and Osteoarthritis Outcome Score (KOOS subscale pain, 0-100) 17,18 was used. In contrast to VAS and ICOAP, a KOOS score of 0 equals the worst pain experience. Statistics Associations between percentages of immune cells and pain scales were determined by linear regression analyses and were adjusted for age, gender and BMI when necessary. Log transformations were used when appropriate and log transformed variables are indicated. Correlations were determined by Spearmanâ&#x20AC;&#x2122;s rank test. A correlation < 0.3 was considered as weak, 0.3-0.7 as moderate and > 0.7 as strong. A p-value <0.05 was considered significant. The statistical package for the social sciences (SPSS) version 20.0 was used (SPSS, Chicago, IL- USA) for analyses.

RESULTS Immune cell populations in synovium and IFP First, we characterized the immune cell population in synovium and the stromal vascular fraction (SVF) of the infrapatellar fat pad (IFP) in detail. Due to the limited amount of tissue, not all experiments could be performed with each tissue sample. Therefore, variable numbers of samples were included in different experiments.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 3

Flow cytometric analysis of the SVF revealed a heterogeneous forward/side scatter (FSC/ SSC). In supplementary Figure 1A we describe the performed gating strategies. To determine the abundance of various cell types we used specific cell-surface markers. Macrophages, mast cells and T cells were readily detectable (Figure 1A, B, F, G), while B cells (CD19+ cells) were virtually absent (less than 1%) in both tissues (data not shown). Within the CD3+ T cell population, the ratio between CD4+ and CD8+ T cells was similar to the one usually found in blood for both tissues (Fig. 1C, H). Further phenotypic characterization of CD4+ and CD8+ T cells in IFP and synovium revealed that these populations consisted mainly of memory cells (Fig. 1D, I). Interestingly, a substantial percentage of both CD4+ and CD8+ T cells had an activated phenotype and expressed CD25 and CD69 (Fig. 1E, J). To directly compare the abundance of different cell populations in synovium and IFP to detect possible tissue specific effects, we have selected and compared paired synovium-IFP samples (Fig. 2). We found that T cells (CD3+ cells) were more abundant in IFP compared to synovium (Fig 2A) However, this difference disappeared when percentages of CD3+ T cells within the CD45+ cell population were compared, indicating that the differences were probably due to the presence of tissue-resident cells (fibroblasts) in the synovial digest (Fig. 2B). Moreover, no differences were found in percentages of macrophages (Fig. 2C), mast cells (Fig. 2D), memory T cells or T cells bearing the late activation marker CD25 (data not shown) between tissues. Interestingly, percentages of T cells positive for the early activation marker CD69 were significantly higher in synovium-derived CD4+ T cells compared to IFPderived T cells and a trend was present for synovium-derived CD8+ T cells (Fig. 2E).

Inflammatory cells in end stage knee OA patients

100

40 20

Macrophages (CD14+ )

20 10

CD45RA+

40 20 CD25+

CD69+

CD25+

60 40 20

100 80

15 10 5

Macrophages (CD14+ )

CD4 +

J % of lindicated population

40 30 20 10

CD45RA+

T cells (CD3+ )

(CD117+)

CD8 +

Mast cells

CD69+

25 20

% of live cells

% of CD45 +cells

CD8+

+ CD8

CD45RA+

100

% of indicated population

CD4+

+ CD4

100

T cells (CD3+ )

(CD117 + )

CD8+

Mast cells

% of live cells

% of indicated population

CD4+

% of CD3+ cells

100

% of live cells

% of CD45 +cells

% of CD3+ cells

CD4+

CD8+

100 80 60 40 20 0

CD25+

CD69+

CD25+

CD69+

Figure 1. Immune cell characterization 1 of the synovium (A – E) and the stromal vascular fraction of the infrapatellar fat pad (IFP) (F – J) of all end stage OA patients (n = 76) by flow cytometry. Gating strategies were performed as described in suppl. Fig1. Percentages of macrophages (CD45+CD14+), mast cells (CD45+CD117+) (A, F) and T cells (CD3+) (B, G) are depicted. Phenotypic characterization of T cells by flow cytometry; (C, H) Percentages of CD4+ and CD8+ T cells; (D, I) CD45RA expression within the CD4+ and CD8+ T cell populations; (E, J) CD25 and CD69 expression within the CD4+ and CD8+ T cell populations. Indicated are mean percentage positive cells ± SEM. Each dot represents one patient.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

% indicated population

% of live cells

25 20 15 10 5 0

SYN

CD3+ CD45+ cells p = 0.85

CD3+ p = 0.002

100

IFP

SYN

p = 0.30

p = 0.15 25

% indicated population

100 80 60 40 20 0

IFP

CD117+CD45+

CD14+CD45+

SYN

20 15 10 5 0

IFP

CD69+ p = 0.02

SYN

IFP

CD69+ p = 0.07

% indicated population

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 3

60 40 20 0

SYN

IFP

CD4+

SYN

IFP

CD8+

Figure 2. Paired synovium-IFP samples analysis of different cells population for end stage knee OA patients. Paired analysis of T cells (% of CD3+ T cells of live population (A) and % CD3+ T cells of CD45+ cells (B)), macrophages (CD114+CD45+) (C) and mast cells (CD117+CD45+) (D) are depicted. Furthermore, percentages of T cells (CD4+ and CD8+ T cells) positive for the early activation marker CD69 are shown (E).

T cell and macrophage cytokine production Macrophages and T cells are the most abundant immune cell types present in synovium and in IFP. To assess their possible contribution to the release of inflammatory mediators, we investigated the polarization state of these cells by studying their cytokine secretion.

Inflammatory cells in end stage knee OA patients

Cytokines typically associated with different polarization states were investigated. Gating strategies are described in supplementary figure 1B. Intracellular cytokine staining showed that both CD4+ and CD8+ T cell subsets produced mainly IL-6 and IL-4 (Fig. 3A, B) in the absence of extra stimulation. Upon in vitro stimulation, these cells were able to produce IFNγ, TNFα, IL-6, IL-4 and very little, if any, IL-10 (Fig. 3C, D). In both tissues, macrophages secreted IL-6 and TNFα, but also some IL-10 ex vivo (Fig. 3E, F). Synovium A

IFP 40

% of positive cells

0 IFN-γ

TNFα

IL6

IL10

IL4

IFN-γ

TNFα

CD4+

IL6

IL10

IFN γ

IL4

PMA/ionomycin

IL10

IL4

IFN γ

TNF α

IL6

IL10

IL4

CD8+ PMA/ionomycin

% of positive cells

IL6

CD4+

0 IFNγ

TNFα

IL6

IL10

IL4

IFNγ

TNFα

CD4+

IL6

CD8+

IL10

IFNγ

IL4

TNFα

IL6

IL10

IL4

IFNγ

TNFα

CD4+

IL6

IL10

IL4

CD8+

cells

% of CD14

TNF α

CD8+

% of CD14

unstimulated

0 IL6

IL10

TNFα

40 20 0

IL6

IL10

TNFα

Figure 3. T cell and macrophage cytokine production in synovium (A, C, E) and the stromal vascular fraction of the infrapatellar fat (IFP) (B, D, F). Gating strategies were performed as described in suppl. Fig1. Spontaneous(A, B), or PMA/ionomycin-induced cytokine production (C, D) of CD4+ and CD8+ T cells are depicted. Spontaneous cytokine production by CD14+ macrophages in synovium (E) or IFP (F). Indicated are mean percentage cytokine positive cells ± SEM. Each dot represents one patient.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 3

Association of immune cells with knee pain To get first insight into the possible contribution of immune cells to clinical characteristics of OA such as radiographic damage and pain, we next investigated whether the synovial- or IFP-derived immune cell populations were associated with Kellgren-Lawrence (KL) scores or self-reported knee pain. For this, we made use of a clinically well-characterized population of patients, participating in the geMstoan study (n=42). No significant differences in immune cell populations were observed between the geMstoan and non-geMstoan samples (data not shown). Furthermore, there were no significant differences in age, gender and BMI between geMstoan patients (n=42) and total group of patients (n=76), indicating that no bias was present in further analysis (data not shown). Patient characteristics are shown in Table 1. Table 1. geMstoan patient characteristics (N = 42) Age Female, No (%) BMI, median (IQR) Fat % Fat mass, median (IQR) VAS pain (0-100) KOOS, subscale pain (0-100) ICOAP (0-100), median (IQR) KL grade 1, Nr (%) KL grade 2, Nr (%) KL grade 3, Nr (%) KL grade 4, Nr (%) KL total, median (IQR)

63.3 (8.4) 22.0 (52.4%) 28.9 (25.7 â&#x20AC;&#x201C; 31.6) 35.98 (11.36) 29.74 (21.3 - 40.5) 65.19 (16.14) 41.14 (18.49) 47.73 (31.82 â&#x20AC;&#x201C; 63.07) 1 (2.4%) 2 (4.8%) 17 (40.5%) 21 (50.0%) 4.00 (3.00-4.00)

Unless specified otherwise, depicted are means (SD); BMI = Body Mass Index; KL = Kellgren and Lawrence; IQR = interquartile range; the visual analogue scale (VAS) for pain; the Knee Injury Osteoarthritis Outcomes Score (KOOS); the Intermittent and Constant Osteoarthritis Pain score (ICOAP).

No associations between immune cells and KL scores could be detected (data not shown). Linear regression analyses, adjusted for age and gender, indicated that the percentage of CD4+ T cells in synovium associated with VAS pain (Table 2; Fig. 4),while no such association was found for other immune cells. This association was not present for IFP-derived CD4+ T cells. Because this part of the study was meant as hypotheses-generating study, we did not adjust for multiple testing.

Inflammatory cells in end stage knee OA patients

Table 2. Association of immune cells with knee pain. VAS pain (0-100) Synovium

IFP

KOOS (0-100) Synovium

IFP

Log ICOAP (0-100) Syn

IFP

Nr. patients

Adjusted B (95% confidence interval)

CD4+ T cells CD8+ T cells macrophages mast cells CD4+ T cells CD8+ T cells macrophages mast cells

22 22 14 13 24 24 20 18

0.55 (0.09 – 1.02)* 0.39 (-0.50 – 1.29) -0.26 (-0.60 – 0.08) -0.49 (-1.71 – 0.74) -0.12 (-0.80 – 0.56) 0.17 (-0.53 – 0.86) 0.07 (-0.38 – 0.52) -0.05 (-1.89 – 1.79)

CD4+ T cells CD8+ T cells macrophages mast cells CD4+ T cells CD8+ T cells macrophages mast cells

22 22 14 13 24 24 20 18

-0.22 (-0.90 – 0.46) -0.87 (-1.96 – 0.21) 0.09 (-0.49 – 0.67) 0.29 (-1.67 – 2.25) 0.56 (-0.30 – 1.41) -0.68 (-1.54 – 0.18) 0.14 (-0.49 – 0.76) 1.41 (-0.72 – 3.54)

CD4+ T cells CD8+ T cells macrophages mast cells CD4+ T cells CD8+ T cells macrophages mast cells

21 21 14 13 23 23 19 17

-0.002 (-0.04 – 0.03) 0.04 (-0.02- 0.92) -0.014 (-0.03 – 0.004) 0.005 (-0.06 – 0.07) -0.02 (-0.07 – 0.03) 0.03 (-0.02 – 0.08) 0.03 (-0.006 – 0.06) -0.006 (-0.07 – 0.06)

Linear regression analysis of T cells, macrophages and mast cells and the visual analogue scale (VAS) for pain, the Knee Injury Osteoarthritis Outcomes Score (KOOS), and the Intermittent and Constant Osteoarthritis Pain score (ICOAP). Corrections were made for age and gender. *P-value < 0.05

Interestingly, the percentage of CD4+ T cells in synovium correlated significantly with BMI (r = 0.45; p = 0.03) in our cohort. Since VAS pain also correlated with BMI (r = 0.34; p = 0.04), we next investigated whether the association of CD4+ T cells with pain is, in statistical terms, mediated by BMI. These analyses revealed that the association of CD4+ T cells with VAS pain is independent of BMI (Adjusted B (95% CI) = 0.55 (0.02-1.08). None of the other immune cells from either synovial tissue or IFP were significantly correlated with BMI.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Synovium

CD4+ T cells

100

r = 0.53 p = 0.01

80 60 40 20 20

VAS

100

IFP 100

CD4+ T cells

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 3

r = 0.21 p = 0.33

80 60 40 20 20

VAS

100

Figure 4. Spearmanâ&#x20AC;&#x2122;s rank correlations between synovial (SYN) and Infrapatellar fat pad (IFP)-derived CD4+ T cells and VAS. A P-value â&#x2030;¤ 0.05 was considered significant.

In an attempt to get more insight into the mechanisms underlying this association, we investigated the relationship between different T cell characteristics and pain. Although the association between different cytokine producing T cells and pain could not be studied due to limited availability of cytokine data in de geMstoan population, we did investigate whether phenotypically distinct CD4+ populations correlated with pain. No correlations could be found between CD4+ T cells expressing CD45RA+, CD25+ or CD69+ with VAS (data not shown).

DISCUSSION In this study, we investigated in detail the immune cells populations present in synovium and IFP of OA patients. Our data indicate similarities both in composition and phenotype of different immune cells populations in these tissues. Moreover, our exploratory studies investigating the association of different cell types with radiographic damage and pain

Inflammatory cells in end stage knee OA patients

suggest that synovial CD4+ T cells might be related to pain perception in patients with end stage knee OA. Macrophages and T cells are the most predominant immune cell types present in synovial tissue and IFP of osteoarthritis patients, which is in line with earlier observations (7 and 6 ). A direct comparison between synovial cells and IFP has never been performed before. Interestingly, we found that the immune cell composition of these tissues was very similar, with no significant differences in the abundance of different cell populations at the population level. In contrast, we were able to detect differences in T cells and mast cells between paired IFP-subcutaneous adipose tissue samples from OA patients (Klein-Wieringa ARD) in an earlier study. Together, these data could indicate that infiltration of immune cells into tissues might be more affected by the disease process (i.e. knee OA) than by tissuespecific signals. Moreover, these data could indicate, as suggested by other studies 9,19, that a cross talk between synovial tissue and IFP exists, leading to a common inflammatory state. Our analyses further indicate that both tissues are an important source of inflammatory mediators and therefore could both have an important role in the pathophysiology of OA. Indeed, both macrophages and T cells could produce cytokines directly ex vivo, which suggests that they are in an activated state in the tissues. Moreover, they secreted predominantly pro-inflammatory and little or no anti-inflammatory cytokines, which is in agreement with the diseased state and with our earlier finding that haematoxylin and eosin staining of these tissue samples revealed the presence of at least one feature of synovitis (i.e. lining cell layer thickening, stromal activation, presence of cellular infiltrate) in each patient. This was in contrast to synovial samples from OA patient undergoing arthroscopy, in which several samples had no detectable signs of inflammation 20. In addition to the ex vivo cytokine secretion, CD4+ T cells in both tissues had an activated phenotype, indicated by the expression of CD69, an early activation marker on T cells. Interestingly, percentages of T cells positive for the early activation marker CD69 were significantly lower in IFP-derived CD4+ T cells and a trend was present for IFP-derived CD8+ T cells indicating that T cells in the synovial tissue are more frequently activated than in the IFP. As it is believed that T cells stop secreting cytokines when their cognate antigens is removed and as the expression of CD69 is transient and relatively short-lived, it is tempting to speculate that the activated T cells recognize antigens locally, in the tissues from which they originate. These antigens remain, however, to be determined. Previous studies have indicated the presence of Th1 cells in synovial tissue. However, IFNÎł, the typical Th1 cytokine, proved difficult to detect by immunohistochemistry (reviewed in O&C de Lange), which might indicate that these cells are present in this tissue, but not in an activated state. Our data support indeed this hypothesis, as we could clearly identify IFNÎłsecreting Th1 cells upon ex vivo activation, but not in the absence of activation.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 3

Although the number of available samples allows only for a hypotheses-generating study regarding the association of various infiltrating cell types with pain, we believe that the found association of synovial CD4+ T cells with pain is of interest and plausible in view of the literature. Several studies have implicated a potential role for T cells in nociception. T cells have not only been shown to infiltrate the site of nerve injury, elimination of T cells in nude rats attenuated hyperalgesia and allodynia in a model of neuropathic pain 21. In addition, this effect appeared to be facilitated by T-helper 1 (Th1) cells through the release of IL-2 and IFNÎł21, cytokines that were also secreted by synovial CD4+ T cells in our study. This indicates a possible mechanism by which synovial T cells could be associated with pain in OA. Moreover, in our study, synovial CD4+ T cells secreted also TNFÎą and IL-6. Interestingly, these cytokines have also been shown to directly affect sensory fiber function 22-25, which further supports the hypothesis that CD4+ T cells could be associated with pain. Therefore, future studies replicating our preliminary findings are certainly of great interest. Previous studies have shown infiltration of (CD4+) T cells into the OA synovium (summarized in 6). We confirmed and expanded these findings, by observing BMI-dependent infiltration of CD4+ T cells in synovium (r = 0.45; p = 0.03), for the first time revealing an associative link between obesity and synovial inflammation. Whether adipose tissue secreted factors or other obesity related changes underlie this association needs to be established. A limitation of our study was the preferential inclusion of patients who underwent arthroplasty, as the only possible source of IFP and of sufficient synovial tissue for exhaustive analyses. Therefore, all patients included in the present study had synovial inflammation and experienced pain in the presence of advanced radiographic knee OA, which could bias our data. Although no patients free of pain were included, the variation in all pain scores (ranges of pain measures; VAS 32-95, KOOS subscale pain 2.78-86.11 and ICOAP 0-72.73) were sufficient to enable us to study associations between immune cells and pain. Other limitations include the lack of healthy control subjects, due to technical difficulties, and the low number of samples available for cytokine analysis which precluded the study of the association between cytokines and pain measures. In summary, our study indicates that the inflammatory cell composition and phenotype is similar between synovium and IFP of OA patients and that part of the infiltrating cells have an activated phenotype and secrete pro-inflammatory cytokines. Moreover, while our data awaits replication, this study indicates that synovial CD4+ T cells could contribute to the perception of knee pain in patients with end-stage knee OA offering a cellular bases for the association between synovial inflammation and knee pain in OA patients. Conflict of interests: the authors have no conflicts of interest

Inflammatory cells in end stage knee OA patients

Author contributions: Study conception and design: MK, AIF, GJEO, AMZ, VSS, RGHHN, REMT; Acquisition of data: BJELB, IRKW, EY, SNA, JCK, HMK; Analysis and interpretation of data: BJELB, IRKW, AIF, SNA, JCK, MK, REMT. All of the authors were involved in drafting the article or revising it critically for important intellectual content, and all of the authors approved the final version to be submitted for publication. Dr. AIF had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Acknowledgements This work was performed within the framework of Dutch Top Institute Pharma, project “Generation of models, mechanisms and markers for stratification of osteoarthritis patients” (project nr. T1-213). The authors would like to acknowledge support of the cooperating hospital Diaconessenhuis, Leiden and orthopaedic surgeons and nurse practitioners. This work was also financially supported by The Dutch Arthritis Association, EU FP6 program Autocure, FP7 program Masterswitch, a grant from Centre for Medical Systems Biology (CMSB) within the framework of the Netherlands Genomics Initiative (NGI) and The Netherlands Organization of Health Research and Development.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 3

REFERENCES

1 Kornaat PR, Bloem JL, Ceulemans RY, Riyazi N, Rosendaal FR, Nelissen RG et al. Osteoarthritis of the knee: association between clinical features and MR imaging findings. Radiology 2006;239:811-7. 2 Yusuf E, Kortekaas MC, Watt I, Huizinga TW, Kloppenburg M. Do knee abnormalities visualised on MRI explain knee pain in knee osteoarthritis? A systematic review. Ann Rheum Dis 2011;70:60-7. 3 Baker K, Grainger A, Niu J, Clancy M, Guermazi A, Crema M et al. Relation of synovitis to knee pain using contrast-enhanced MRIs. Ann Rheum Dis 2010;69:1779-83. 4 Roemer FW, Guermazi A, Felson DT, Niu J, Nevitt MC, Crema MD et al. Presence of MRI-detected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30-month follow-up: the MOST study. Ann Rheum Dis 2011;70:1804-9. 5 Roemer FW, Zhang Y, Niu J, Lynch JA, Crema MD, Marra MD et al. Tibiofemoral joint osteoarthritis: risk factors for MR-depicted fast cartilage loss over a 30-month period in the multicenter osteoarthritis study. Radiology 2009;252:772-80. 6 de Lange-Brokaar BJ, Ioan-Facsinay A, Van Osch GJ, Zuurmond AM, Schoones J, Toes RE et al. Synovial inflammation, immune cells and their cytokines in osteoarthritis: a review. Osteoarthritis Cartilage 2012;20:1484-99. 7 Klein-Wieringa IR, Kloppenburg M, BastiaansenJenniskens YM, Yusuf E, Kwekkeboom JC, El-Bannoudi H et al. The infrapatellar fat pad of patients with osteoarthritis has an inflammatory phenotype. Ann Rheum Dis 2011;70:851-7. 8 Bastiaansen-Jenniskens YM, Clockaerts S, Feijt C, Zuurmond AM, Stojanovic-Susulic V, Bridts C et al. Infrapatellar fat pad of patients with endstage osteoarthritis inhibits catabolic mediators in cartilage. Ann Rheum Dis 2012;71:288-94. 9 Ushiyama T, Chano T, Inoue K, Matsusue Y. Cytokine production in the infrapatellar fat pad: another source of cytokines in knee synovial fluids. Ann Rheum Dis 2003;62:10812. 10 Roemer FW, Guermazi A, Zhang Y, Yang M, Hunter DJ, Crema MD et al. Hoffaâ&#x20AC;&#x2122;s Fat Pad: Evaluation on Unenhanced MR Images as a Measure of Patellofemoral Synovitis

in Osteoarthritis. AJR Am J Roentgenol 2009;192:1696-700. 11 de Lange-Brokaar BJ, Ioan-Facsinay A, Yusuf E, Visser AW, Kroon HM, Andersen SN et al. Degree of synovitis on MRI by comprehensive whole knee semi-quantitative scoring method correlates with histologic and macroscopic features of synovial tissue inflammation in knee osteoarthritis. Osteoarthritis Cartilage 2014;22:1606-13. 12 Hill CL, Hunter DJ, Niu J, Clancy M, Guermazi A, Genant H et al. Synovitis detected on magnetic resonance imaging and its relation to pain and cartilage loss in knee osteoarthritis. Ann Rheum Dis 2007;66:1599-603. 13 Altman R, Alarcon G, Appelrouth D, Bloch D, Borenstein D, Brandt K et al. The American College of Rheumatology criteria for the classification and reporting of osteoarthritis of the hand. Arthritis Rheum 1990;33:1601-10. 14 The Atlas of Standard Radiographs of Arthritis. Rheumatology 2005;44:iv43-iv72. 15 Hawker GA, Davis AM, French MR, Cibere J, Jordan JM, March L et al. Development and preliminary psychometric testing of a new OA pain measure--an OARSI/OMERACT initiative. Osteoarthritis Cartilage 2008;16:409-14. 16 Kessler S, Grammozis A, Gunther KP, Kirschner S. [The intermittent and constant pain score (ICOAP) - a questionnaire to assess pain in patients with gonarthritis]. Z Orthop Unfall 2011;149:22-6. 17 Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD. Knee Injury and Osteoarthritis Outcome Score (KOOS)--development of a self-administered outcome measure. J Orthop Sports Phys Ther 1998;28:88-96. 18 Roos EM and Toksvig-Larsen S. Knee injury and Osteoarthritis Outcome Score (KOOS) validation and comparison to the WOMAC in total knee replacement. Health Qual Life Outcomes 2003;1:17. 19 Clockaerts S, Bastiaansen-Jenniskens YM, Feijt C, De CL, Verhaar JA, Zuurmond AM et al. Cytokine production by infrapatellar fat pad can be stimulated by interleukin 1beta and inhibited by peroxisome proliferator activated receptor alpha agonist. Ann Rheum Dis 2012;71:1012-8. 20 de Lange-Brokaar BJ, Ioan-Facsinay A, Yusuf E, Visser AW, Kroon HM, Andersen SN et al.

Inflammatory cells in end stage knee OA patients

Degree of synovitis on MRI by comprehensive whole knee semi-quantitative scoring method correlates with histologic and macroscopic features of synovial tissue inflammation in knee osteoarthritis. Osteoarthritis Cartilage 2013; Moalem G, Xu K, Yu L. T lymphocytes play a role in neuropathic pain following peripheral nerve injury in rats. Neuroscience 2004;129:767-77. Ren K and Dubner R. Interactions between the immune and nervous systems in pain. Nat Med 2010;16:1267-76. Marchand F, Perretti M, McMahon SB. Role of the immune system in chronic pain. Nat Rev Neurosci 2005;6:521-32. Richter F, Natura G, Loser S, Schmidt K, Viisanen H, Schaible HG. Tumor necrosis factor causes persistent sensitization of joint nociceptors to mechanical stimuli in rats. Arthritis Rheum 2010;62:3806-14. von Banchet GS, Richter J, Huckel M, Rose C, Brauer R, Schaible HG. Fibroblast-like synovial cells from normal and inflamed knee joints differently affect the expression of pain-related receptors in sensory neurones: a co-culture study. Arthritis Res Ther 2007;9:R6.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 3

SUPPORTING INFORMATION

Supplementary Figure 1. (A) Gating strategy for Fig 1 and 2. A live gate based on FSC-A/SSC-A was first set on stromal vascular fraction cells (SVF), followed by two gates to exclude double cells. (B) The intersection of this gate was used for gating the lymphocytes positive for CD3 and the hematopoietic cells positive for CD45. Lymphocytes positive for CD3 were selected and, within this population, the CD4+/CD8+ cells were further analyzed as depicted in Fig.1. Hematopoietic cells were selected and within this population the CD117+ and the CD14+ populations were analyzed. (B) Gating strategy used for Fig 2. A live gate based on cell FSC-A/SSC-A was set, followed by two gates to exclude double cells (as depicted in A). Within the intersection of these gates, cells positive for dead cell discriminator were excluded as described in â&#x20AC;&#x153;Materials and Methodsâ&#x20AC;?. CD14+ cells in this live gate were further analyzed as depicted in Fig. 2. The lymphocyte population was gated based on FCS-A/SSC-A. Lymphocytes positive for CD3 and negative for CD14 were selected, the CD4+/CD8+ cells were further analyzed as depicted in Fig. 2.

CHAPTER 4 CHARACTERISATION OF SYNOVIAL MAST CELLS IN KNEE OSTEOARTHRITIS: ASSOCIATION WITH CLINICAL PARAMETERS

de Lange-Brokaar BJE, Kloppenburg M, Andersen SN, Dorjée AL, Yusuf E, Herb-van Toorn L, Kroon HM, Zuurmond A-M, Stojanovic-Susulic V, Bloem JL, Nelissen RGHH, Toes REM, Ioan-Facsinay A

submitted

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 4

ABSTRACT Objective: To investigate the presence of mast cells in the osteoarthritic (OA) synovium and their association with clinical parameters in comparison with rheumatoid arthritis (RA) samples. Methods: Synovial tissues of 56 symptomatic OA and 49 RA patients were obtained. Two to three paraffin slides were used to quantify inflammation using haematoxylin and eosin staining (synovitis score 0-9), and numbers of mast cells (per 10 high-power fields) using double immunofluorescence for CD117 and tryptase. Average scores per patient were used for analysis. Knee radiographs of OA patients were scored according to the Kellgren and Lawrence (KL) system and pain was determined in OA patients at baseline by visual analogue scale. Results: Median (range) of mast cells was significantly higher in OA samples 45 (1-168) compared to RA samples 4 (1-47) (p-value < 0.001), despite a lower median (range) synovitis score in OA (2.5 (0-6.0)) compared to 4.6 (0-8.0) in RA samples. The synovitis score was significantly correlated with the number of mast cells (in OA Spearmanâ&#x20AC;&#x2122;s rho (p-value) 0.3 (0.023) and RA 0.5 (p-value < 0.001)). Clinically, the number of mast cells was associated with an increased KL-grade (p-value 0.05) in OA patients, but not with self-reported pain. Conclusion: Prevalence of mast cells in OA synovial tissue is relatively high and associates with structural damage in OA patients, suggesting a role of mast cells in this disease.

Characterization of synovial mast cells in knee OA

INTRODUCTION Osteoarthritis (OA) is a common rheumatic disorder that has been considered to be a noninflammatory condition1. Synovitis, however, could play a role in the pathophysiology of OA1-6 as it is a predictor of cartilage destruction7,8 and a determinant of pain9,10. Although the biological processes underlying these associations are poorly understood, it has been suggested that immune cells infiltrating the synovial tissue could be important determinants of synovial inflammation6,11. Most frequent types of immune cells found in OA are macrophages, T cells and mast cells11. In contrast to macrophages and T cells, mast cells have been less frequently investigated in OA12-22. Interestingly however, it has been reported in the 1980s and 90s that mast cells were the only type of immune cells found in equal numbers14,15,17,23 or higher12,18 in OA compared to rheumatoid arthritis (RA) and healthy controls15-18,20,23. These observations suggest a role for these cells in the pathophysiology of OA, although little additional studies have been performed since then. Moreover, no studies investigated the relationship between mast cells and clinical symptoms and signs of OA. Several mechanisms have been put forward through which mast cells could contribute to the disease process in RA and these could also apply in OA. For example, mast cells are thought to attract other immune cells through cytokine release, chemokine release or direct cell to cell contact, which leads to activation of the synovium and ultimately to inflammation and bone destruction. Other possible mechanisms include the activation of fibroblasts and other synovial cells, stimulation of angiogenesis, upregulation of adhesion molecules on endothelium and others24. Furthermore, mast cells could contribute to pain in OA, as they have been implicated in pain perception in several disorders25. How mast cells exert these effects is also poorly understood, but release of soluble mediators and enzymes have been suggested as possible mechanisms. The contribution of mast cells to radiographic changes and pain is still unknown. To get insight into the potential role of mast cells in OA, we investigated the difference in mast cell number or degranulation status in OA compared to RA and at different stages of disease. Furthermore, we investigated the association of the number of mast cells with the grade of synovitis, structural damage and pain in OA.

METHODS Osteoarthritis (OA) Patients The OA patients participated in the geMstoan study (GEneration of Models, Mechanism & Markers for Stratification of OsteoArthritis patieNts), an observational study in established and end-stage knee OA patients26. This study has been approved by the ethics committee

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 4

of the Leiden University Medical Center (LUMC) and the Diaconessenhuis Leiden. All patients provided written informed consent. Inclusion and exclusion criteria are described elsewhere26. Two groups of patients with symptomatic radiographic primary knee OA, following the ACR criteria,27 have been included in the geMstoan study: one group with endstage disease that were planned to receive an arthroplasty (“late OA”) and another group with mild to established OA that had no indication for an arthroplasty (“early OA”). Rheumatoid Arthritis (RA) patients Tissue samples were obtained as leftover material from patients who underwent arthroplasty (“late RA”). Arthroscopy samples (“early RA”) were collected within a study described in 28, including patients with RA; excluded were those on oral prednisolon > 10 mg/day, recent (less than six weeks) change of disease modifying anti-rheumatic drugs (DMARDs) or recent joint injection. The study was approved by the Medical Ethical Committee of the LUMC, Leiden. Synovial tissue collection Synovial samples were either collected during arthroplasty or arthroscopy. Synovial samples were collected from patients admitted for arthroplasty for OA or RA to the orthopaedic department of the LUMC or Diaconessenhuis Leiden. Arthroscopy was performed at the department of Rheumatology for research purposes only, using a small-bore 2.7 mm arthroscope (Storz, Turrlingen, Germany) with sterile technique, as described previously26,28,29. Synovial tissues obtained during arthroscopy and arthroplasty were fixed in formalin for approximately 24 hours and then transferred to 70% ethanol where they were stored until embedded in paraffin. Histological staining of biopsies Four micrometer thick sections were stained with haematoxylin and eosin (H&E) after deparaffinization. For each patient, 2-6 H&E stained synovial tissue samples were scored for 3 features, according to Krenn et al.30: lining cell layer, synovial stroma and inflammatory infiltrate. The grading system of Krenn et al. was modified as described earlier26. All synovial tissue samples were scored by 2-3 independent observers (BDL, SNA and ALD for OA, BDL and ALD for RA) who were blinded to imaging clinical data; an average score per feature was calculated (0-3) for each patient. Furthermore, a total synovitis score was calculated (sum of averaged scores of all 3 features (0-9)). Only for the OA samples, in case the scoring of one observer was evidently different from the other two, one rescoring was performed by that observer.

Characterization of synovial mast cells in knee OA

Mast cell staining and scoring Four micrometer sections of synovial tissue from patients with OA were deparaffinized, rehydrated and treated with DAKOpen (DAKO, Glostrup, Denmark). After antigen retrieval by heating 30 min in prewarmed EDTA pH9 in 96ºC, the sections were blocked with blocking solution (10% Normal Donkey Serum (DS) in 1%BSA/PBS) (Sigma Aldrich, Saint Louis, Missouri (MO), USA) for 30 min. at RT. Sections were incubated first with 1:100 polyclonal rabbit anti-human CD117 (DAKO, Glostrup, Denmark) and 0.125 µg/ml monoclonal mouse anti-human tryptase antibody (Merck Millipore, Darmstadt, Germany) in 1%DS/1%BSA/PBS for 1 hour at RT, followed by wash with PBS and incubation with 1:1000 polyclonal donkey anti-rabbit IgG Alexa Fluor 488 for mast cells and 1:1000 polyclonal donkey anti-mouse IgG Alexa Fluor 568 for tryptase (both Invitrogen, Carlsbad, California, USA) in 1%DS/1%BSA/ PBS., for 1 hour at RT. Control rabbit Ig (Antibodies online, Aachen, Germany) and control mouse IgG1 (DAKO, Glostrup, Denmark) were included as isotype controls. Slides were covered with Vectashield Hard Set mounting medium with DAPI (Vector laboratories, Burlingame, California (CA), USA) and analyzed on a fluorescence imaging microscope (Axio Scope A1, Zeiss, Oberkochen, Germany) coupled to a MRc5 camera using AxioVision 4.9.1. software. Degranulated mast cells were defined as CD117+ or CD117-tryptase+ cells for which at least 2 tryptase granules next to the cells were readily detectable. All other tryptase+ cells were defined as non-degranulated mast cells. Degranulated, non-degranulated and total number of mast cells (degranulated + non-degranulated), were determined in 10 adjacent highpower fields in the middle of the tissue section, starting on the left side of the section at the lining layer. Examples of mast cell stainings are presented in Figure 1. Direction of sub sequential high power fields was from left to right and from top to bottom. Mast cells in OA tissues were counted manually by three independent blinded observers (BDL, SNA and ALD), while two blinded observers counted the mast cells in RA tissues (BDL, ALD). Average (between all observers) total number of mast cells per 10 high power fields and percentage of degranulated cells were used for analyses.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 4

Figure 1. Paraffin sections were stained for DAPI, CD117 and tryptase as described in Materials and Methods. Individual stainings and the overlay (right column) are presented, as well as examples of degranulated and nondegranulated mast cells, isotype stainings and overview image on a lower magnification.

Scoring knee radiographs Radiographs (posterior anterior (PA) fixed flexion) were obtained of all patients in the geMstoan study. Radiographs were scored, blinded for patient characteristics, by an experienced musculoskeletal radiologist (HK) according the Kellgren- Lawrence (KL) scale31. Reproducibility was as earlier described26. Clinical data In the geMstoan patients, demographics and disease characteristic were collected via standard questionnaires. Pain was measured by the visual analogue scale (VAS, 0-100), in whichparticipants were asked to place an X at a 100 mm line that represented the general pain intensity for that knee. A score of 100 represents worst possible pain intensity. Statistics For comparison between groups, unpaired t-test, Mann-Whitney-U test or Kuskal-Wallis test were used as indicated. Correlations were investigated using Spearmanâ&#x20AC;&#x2122;s rank correlation test. Linear regression was used to investigate association between mast cells (variable) and

Characterization of synovial mast cells in knee OA

inflammation (outcome). Statistical analyses were calculated performed using SPSS 22.0 and GraphPad Prism version 6.02.

RESULTS Patient characteristics: Fifty-six patients with symptomatic knee OA and 49 RA patients were included. Their characteristics are summarized in Table 1. Twenty-two patients with “early” OA and 23 patients with “early” RA underwent knee arthroscopy and synovial biopsies were obtained. Thirty-four patients with “late” knee OA and 26 “late” RA patients were included and synovial biopsies of the knee were obtained. Table 1. Characteristics of 56 patients with knee osteoarthritis (OA) and 49 patients with rheumatoid arthritis (RA). OA, nr. of patients Age (years), mean (sd) Female, n (%) BMI (Kg/m2) VAS pain (0-100) KL grade (0-4) KL 1 (n= 4 (7.3%) KL 2 (n=13,23.6%) KL 3 (n=21, 38.2%) KL 4 (n= 17,30.9%) Total synovitis score (0-9) Lining layer Stromal activation Infiltrate RA, nr. of patients Age (years), mean (sd) Female, n (%) Total synovitis score (0-9) Lining layer Stromal activation Infiltrate

All 56 61.9 (8.1) 36 (64.3) 29.2 (20.5-49.9) 64.5 (4-95) 3.00 (1-4) 4 (7.3) 13 (23.6) 21 (38.2) 17 (30.9) 2.5 (0-6.0) 0.8 (0-2.8) 0.9 (0-2.7) 0.5 (0-3.0) 49 62.3 (11.9) 32 (68.1) 4.6 (0-8.0) 1.5 (0-3.0) 2.0 (0-3.0) 1.6 (0-3.0)

early 22 60.4 (7.3) 16 (72.7) 29.2 (24.1-43.9) 42.5 (4-84) 2.0 (1-4) 3 (13.6) 11 (50.0) 6 (27.3) 2 (9.1) 1.8 (0-4.8) 0.5 (0-2.8) 0.7 (0-2.7) 0.2 (0-1.8) 23 64.0 (14.2) 15 (68.2) 4.5 (0-7.0) 1.0 (0-3.0) 2.0 (0-3.0) 1.1 (0-3.0)

late 34 62.8 (8.6) 20 (58.8) 29.2 (20.5-49.9) 68.0 (32-95) 3.0 (1-4) 1 (3.0) 2 (6.1) 15 (45.5) 15 (45.5) 2.9 (0-5.6) 1 (0-2.5) 1 (0-2.3) 0.9 (0-3) 26 63.0 (9.7) 17 (68.0) 5.1 (0-8.0) 1.5 (0-2.3) 1.5 (0-3.0) 2.0 (0-3.0)

Numbers represent median (range) unless otherwise specified. Abbreviations: BMI= body mass index; KL= KellgrenLawrence

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 4

Mast cells in OA vs RA samples First, we have quantified mast cells in a set of synovial tissues obtained from both “early” and “late” OA and RA patients. Mast cells were mainly located in the sub-lining layer in both OA and RA samples (data not shown). Median (range) of mast cells per patient per 10 HPF was 45 (1-168) in the OA samples and this was significantly higher than 14 (1-47) in the RA samples (p-value < 0.0001, Fig. 2a). Similarly, mast cells were more abundant in both the early and late OA samples than in their RA counterparts (Fig. 2b). A relatively small proportion of the mast cells were degranulated in both OA and RA samples (Fig. 2c) and no significant differences could be detected between any of the groups (Fig. 2d), indicating that the abundance rather than the degranulation state of the mast cells is associated with having OA.

Figure 2. Total mast cell numbers in 10 high power fields in A) total osteoarthritis (OA) and Rheumatoid Arthritis (RA) population, B) early and late OA and RA populations are depicted. The percent degranulated mast cells in the total OA and RA population (C) or early and late OA and RA populations (D) is depicted. Each dot represents one patient. The median per group is indicated. P-values for differences between groups (as indicated) tested by MannWhitney (A-C) or Kruskal-Wallis (D) are depicted.

Characterization of synovial mast cells in knee OA

Association with histological synovial inflammation Patients with OA had lower histological synovitis scores than patients with RA; this was observed for the three separate features (lining layer, stromal activation and infiltrate), as is depicted in Table 1. Moreover, a higher total synovitis score could be found in the RA compared to the OA samples (Fig. 3a and Table 1), despite a lower number of mast cells, and these differences were still present when arthroscopy or arthroplasty samples from OA and RA were compared (Fig. 3b). Next, we studied whether the higher abundance of mast cells could be related to the presence of higher synovitis in general26.

Figure 3. Synovitis scores determined by H&E staining (as described in Materials and Methods) in A) total osteoarthritis (OA) and Rheumatoid Arthritis (RA) population or B) early and late OA and RA populations are depicted. Each dot represents one patient. The median per group is indicated. P-values for differences between groups (as indicated) tested by student’s t-test are depicted.

Mast cell numbers associated with the synovitis score both in the OA and the RA samples (Fig. 4), although the effect size of this association was lower in OA than in RA (beta (95% CI) = 0.48 (1.35 – 4.52) for RA, (beta(95%CI) = 0.35 (1.82 – 14.68 for OA), as shown by univariate linear regression analyses. Interestingly, as shown in fig.4, there is an increased number of mast cells in OA synovium compared to RA synovium, even in the absence of significant inflammation (synovitis score ≤ 1). Moreover, mast cells explained only 12% of the variance in total synovitis score in OA, while this was 23% in RA, in univariate linear regression analyses.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 4

Figure 4. The correlation between total number of mast cells and synovitis scores (determined by H&E staining) is presented in osteoarthritis (OA, closed circles) and Rheumatoid Arthritis (RA, open squares) patients. Each dot represents one individual. Spearmanâ&#x20AC;&#x2122;s correlation coefficient and p-values are presented for each group.

Association of mast cells with clinical features in OA To understand the possible contribution of mast cells to the disease process in OA, we further investigated the association between mast cell numbers and clinical disease parameters. To ensure that enough variation is present in VAS pain values and radiographic severity scores, we performed these analyses in the whole OA group. As depicted in Fig.5a, there is no significant association between the number of mast cells and VAS pain. Because a trend for higher number of mast cells in females than in males was observed (median (range) 50 (1-169) for females, 39 (2-67) for males, p = 0.060 (Mann-Whitney-U test)), we have also stratified our analyses for gender and found that there is no correlation between mast cells abundance and VAS pain in any of the groups (data not shown). To evaluate the association with radiographic damage, we studied the amount of mast cells present in patients with a KL score of 1/2 vs a KL score of 3/4. As shown in Fig. 5b, there is an increased number of mast cells in the KL3/4 (median (range), 49 (2-169) compared to KL1/2 group (median (range) 26 (1-82) (p = 0.05 (Mann-Whitney-U test)). This association remains in stratified analyses in females (Fig. 5c), and in males, although it does not reach statistical significance in males (Fig. 5d).

Characterization of synovial mast cells in knee OA

Figure 5: A) The correlation between number of mast cells and visual analog scale (VAS) pain is depicted. Differences in mast cell numbers between patients with low Kellgren-Lawrence (KL) scores (KL1 and KL2) and high KL scores (KL3 and KL4) were evaluated in the total osteoarthritis (OA) population (B), as well as in females (C) and (D) males. P-values were determined by Mann-Whitney.

DISCUSSION In this study, we investigated the presence and degranulation state of mast cells in OA and RA samples at different disease stages. Our data indicate that there is an increase in mast cell number in OA compared to RA synovium, while there is no difference in mast cell number or degranulation state between end-stage disease (late) compared to mild disease (early) in either disease, confirming and expanding previous findings. Furthermore, we found that higher numbers of mast cells are associated with more synovitis and more structural damage in OA patients. To our knowledge, this is the first published report to find an association between mast cells synovial infiltration and clinical parameters in OA, suggesting a role of these cells in disease pathogenesis.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 4

Mast cells were readily detected in OA synovial tissue in present study. By staining for tryptase and CD117, we detected a higher number of mast cells in the OA compared to RA synovium. Previous studies comparing OA and RA synovial mast cells reported controversial results. Although the underlying reason for this is unclear, most previous studies have used toluidine blue to detect mast cells and this staining is possibly less sensitive than the tryptase staining employed by us and others32. Moreover, in contrast to previous studies that also used tryptase staining, we have counterstained the tissue sections with CD117, allowing also detection of mast cells that have partially or totally degranulated. Using this technique, we could assess the number of degranulated mast cells in situ, while most previous studies have investigated degranulation in synovial fluid33 or rather the capacity of synovial mast cells to degranulate upon ex vivo stimulation13,14. Our data indicate no differences in mast cell number or degranulation between “early” and “late” disease in either group of patients. Moreover, percentage of degranulated mast cells was around 20% in both diseases and independently of disease stage, which is in line with an earlier study15. This indicates that, independently of disease severity, a certain percentage of mast cells is degranulated in the synovium. Moreover, neither the number nor the percentage of degranulated mast cells was associated with inflammation, pain or radiographic damage (data not shown), while the number of mast cells was. Mast cells can be activated by several stimuli and degranulation is only one of the possible outcomes of mast cell activation. Our data suggest that other mechanisms than degranulation probably contribute to synovitis or radiographic damage in OA. Both in OA and RA, mast cell numbers were correlated with the degree of synovitis. This indicates that mast cells could be recruited or stimulated to proliferate in the presence of synovitis or that a higher number of resident mast cells leads to recruitment of inflammatory cells into synovium. Indeed, previous studies have suggested that stem cell factor (SCF) secreted by synovial cells could lead to recruitment and hyperplasia of mast cells into synovium and this could be associated with histologic inflammation23,34. Whether this mechanism plays a role in our patient cohort, remains to be elucidated. Interestingly, OA patients have consistently higher numbers of mast cells than RA patients, independently of synovitis score and even in the absence of significant inflammation. Moreover, mast cells seem to contribute less to synovitis in OA than in RA. Taken together, these data suggest that the higher number of mast cells in OA than in RA is to a certain extent mediated by mechanisms unrelated to synovitis. This is not surprising, considering that the mechanisms involved in mast cells activation in OA and RA could be very different. For example, it is conceivable that in RA, autoantibodies could contribute to mast cells activation, while in OA other factors, such as matrix degradation products could be of importance. Indeed, low molecular weight hyaluronic acid, fibronectin or fibronectin fragments are potentially able to act as danger signals for cells expressing TLR2,-4 or CD44 and some of these receptors have been previously shown to be expressed on mast cells35. 84

Characterization of synovial mast cells in knee OA

Our data suggest that mast cells are associated with the degree of radiographic damage in OA patients. Although the mechanisms involved in this association are unclear, it is likely that synovitis is not the main mechanisms mediating this association, as synovitis is not associated with radiographic damage in our population (data not shown) and is only weakly associated with mast cell numbers. Mast cells could contribute to cartilage damage and/or osteophyte formation through various mechanisms. For instance, tryptase released by mast cells could contribute to cartilage damage through activation of protease-activated receptor 2 (PAR2), which has been identified in human OA and RA synovium20 and cartilage36 and has been shown to be critical for the induction of OA in a mouse model37. Similarly, mast cells have been previously associated with osteopenia, as a result of activation of osteoclasts, indicating that they can influence bone turnover (reviewed in24) Moreover, previous reports have suggested that increased numbers of mast cells could lead not only to increased osteoclast activation, but also to increased osteoblast numbers38. Possible mechanisms include release of tryptase and activation of PAR-2 on osteoblasts39 and release of oncostatin M40 and possibly other mechanisms with direct or indirect effects on osteoblasts. The present study has some limitations. Firstly, only a relatively low number of patients were included in the present study, most of whom were females. This could explain the fact that most significant associations found were only detected in females. Furthermore, the present study is cross-sectional, precluding establishment of cause and consequence. Similar studies in a longitudinal manner could be very informative in assessing whether high mast cells numbers are a cause of inflammation/radiographic damage or a consequence of active disease in arthritis patients. Moreover, both RA and OA patients were taking medication at the time of the study. Although it is nowadays known that some of the anti-rheumatic drugs, such as corticosteroids and cyclosporine could affect mast cell function41-44, the effect of most drugs taken by the patients in our cohort on mast cells is still unknown. Finally, the present study lacks comparison with normal controls, associated with difficulties of obtaining synovial tissues from healthy subjects age-matched with our geMstoan population. Taken together, our data show for the first time an association between mast cell numbers and radiographic damage in OA patients, suggesting a possible role of mast cells in this disease. Acknowledgments This work was performed within the framework of Dutch Top Institute Pharma, project â&#x20AC;&#x153;Generation of models, mechanisms and markers for stratification of osteoarthritis patientsâ&#x20AC;?Â (project nr. T1-213) and was co-funded by the Dutch Arthritis Foundation, the

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 4

Dutch Science Foundation (NWO) and by the IMI JU funded project BeTheCure, contract no 115142-2. Author Stojanovic-Susulic is an employee of Janssen Research & Development LLC, a subsidiary of Johnson & Johnson. The authors would like to acknowledge the support of the cooperating hospital (Diaconessenhuis, Leiden, The Netherlands), including nurse practitioner C. E. Jonxis and orthopaedic surgeons Drs. R. Krips, J. B. Mullers, and H. M. Schuller. We also thank the referring rheumatologists, orthopaedic surgeons, and nurse practitioners.

Characterization of synovial mast cells in knee OA

REFERENCES

1 Felson DT. Clinical practice. Osteoarthritis of the knee. N Engl J Med 2006;354:841-8. 2 Loeuille D, Chary-Valckenaere I, Champigneulle J, Rat AC, Toussaint F, Pinzano-Watrin A et al. Macroscopic and microscopic features of synovial membrane inflammation in the osteoarthritic knee: correlating magnetic resonance imaging findings with disease severity. Arthritis Rheum 2005;52:3492-501. 3 Loeuille D, Rat AC, Goebel JC, Champigneulle J, Blum A, Netter P et al. Magnetic resonance imaging in osteoarthritis: which method best reflects synovial membrane inflammation? Correlations with clinical, macroscopic and microscopic features. Osteoarthritis Cartilage 2009;17:1186-92. 4 Pearle AD, Scanzello CR, George S, Mandl LA, DiCarlo EF, Peterson M et al. Elevated high-sensitivity C-reactive protein levels are associated with local inflammatory findings in patients with osteoarthritis. Osteoarthritis Cartilage 2007;15:516-23. 5 Benito MJ, Veale DJ, FitzGerald O, van den Berg WB, Bresnihan B. Synovial tissue inflammation in early and late osteoarthritis. Ann Rheum Dis 2005;64:1263-7. 6 Sellam J and Berenbaum F. The role of synovitis in pathophysiology and clinical symptoms of osteoarthritis. Nat Rev Rheumatol 2010;6:62535. 7 Roemer FW, Guermazi A, Felson DT, Niu J, Nevitt MC, Crema MD et al. Presence of MRI-detected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30-month follow-up: the MOST study. Ann Rheum Dis 2011;70:1804-9. 8 Roemer FW, Zhang Y, Niu J, Lynch JA, Crema MD, Marra MD et al. Tibiofemoral joint osteoarthritis: risk factors for MR-depicted fast cartilage loss over a 30-month period in the multicenter osteoarthritis study. Radiology 2009;252:772-80. 9 Yusuf E, Kortekaas MC, Watt I, Huizinga TW, Kloppenburg M. Do knee abnormalities visualised on MRI explain knee pain in knee osteoarthritis? A systematic review. Ann Rheum Dis 2011;70:60-7. 10 Baker K, Grainger A, Niu J, Clancy M, Guermazi A, Crema M et al. Relation of synovitis to knee pain using contrast-enhanced MRIs. Ann Rheum Dis 2010;69:1779-83.

11 de Lange-Brokaar BJ, Ioan-Facsinay A, Van Osch GJ, Zuurmond AM, Schoones J, Toes RE et al. Synovial inflammation, immune cells and their cytokines in osteoarthritis: a review. Osteoarthritis Cartilage 2012;20:1484-99. 12 Fritz P, Reiser H, Saal JG. Analysis of mast cells in rheumatoid arthritis and osteoarthritis by an avidin-peroxidase staining. Virchows Archiv Abteilung B Cell Pathology 1984;47:35-45. 13 Gruber B, Poznansky M, Boss E. Characterization and functional studies of rheumatoid synovial mast cells. Activation by secretagogues, antiIgE, and a histamine-releasing lymphokine. Arthritis and rheumatism 1986;29:944-55. 14 Bridges AJ, Malone DG, Jicinsky J, Chen M, Ory P, Engber W et al. Human synovial mast cell involvement in rheumatoid arthritis and osteoarthritis. Relationship to disease type, clinical activity, and antirheumatic therapy. Arthritis Rheum 1991;34:1116-24. 15 Dean G, Hoyland JA, Denton J, Donn RP, Freemont AJ. Mast cells in the synovium and synovial fluid in osteoarthritis. Br J Rheumatol 1993;32:671-5. 16 Gotis-Graham I and McNeil HP. Mast cell responses in rheumatoid synovium. Association of the MCTC subset with matrix turnover and clinical progression. Arthritis Rheum 1997;40:479-89. 17 Buckley MG, Gallagher PJ, Walls AF. Mast cell subpopulations in the synovial tissue of patients with osteoarthritis: selective increase in numbers of tryptase-positive, chymasenegative mast cells. J Pathol 1998;186:67-74. 18 Pu J, Nishida K, Inoue H, Asahara H, Ohtsuka A, Murakami T. Mast cells in osteoarthritic and rheumatoid arthritic synovial tissues of the human knee. Acta Med Okayama 1998;52:359. 19 Damsgaard TE, Sorensen FB, Herlin T, Schiotz PO. Stereological quantification of mast cells in human synovium. APMIS 1999;107:311-7. 20 Nakano S, Mishiro T, Takahara S, Yokoi H, Hamada D, Yukata K et al. Distinct expression of mast cell tryptase and protease activated receptor-2 in synovia of rheumatoid arthritis and osteoarthritis. Clin Rheumatol 2007;26:1284-92. 21 Saito I, Koshino T, Nakashima K, Uesugi M, Saito T. Increased cellular infiltrate in inflammatory synovia of osteoarthritic knees. Osteoarthritis Cartilage 2002;10:156-62.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 4

22 Kopicky-Burd JA, Kagey-Sobotka A, Peters SP, Dvorak AM, Lennox DW, Lichtenstein LM et al. Characterization of human synovial mast cells. J Rheumatol 1988;15:1326-33. 23 Ceponis A, Konttinen YT, Takagi M, Xu JW, Sorsa T, Matucci-Cerinic M et al. Expression of stem cell factor (SCF) and SCF receptor (c-kit) in synovial membrane in arthritis: correlation with synovial mast cell hyperplasia and inflammation. J Rheumatol 1998;25:2304-14. 24 Nigrovic PA and Lee DM. Synovial mast cells: role in acute and chronic arthritis. Immunol Rev 2007;217:19-37. 25 Chatterjea D and Martinov T. Mast cells: Versatile gatekeepers of pain. Mol Immunol 2015;63:38-44. 26 de Lange-Brokaar BJ, Ioan-Facsinay A, Yusuf E, Visser AW, Kroon HM, Andersen SN et al. Degree of synovitis on MRI by comprehensive whole knee semi-quantitative scoring method correlates with histologic and macroscopic features of synovial tissue inflammation in knee osteoarthritis. Osteoarthritis Cartilage 2014;22:1606-13. 27 Altman R, Alarcon G, Appelrouth D, Bloch D, Borenstein D, Brandt K et al. The American College of Rheumatology criteria for the classification and reporting of osteoarthritis of the hand. Arthritis Rheum 1990;33:1601-10. 28 van OM, Sont JK, Bajema IM, Breedveld FC, van Laar JM. Comparison of efficacy of arthroscopic lavage plus administration of corticosteroids, arthroscopic lavage plus administration of placebo, and joint aspiration plus administration of corticosteroids in arthritis of the knee: A randomized controlled trial. Arthritis Rheum 2006;55:964-70. 29 van OM, Sont JK, van Laar JM. Superior effect of arthroscopic lavage compared with needle aspiration in the treatment of inflammatory arthritis of the knee. Rheumatology (Oxford) 2003;42:102-7. 30 Krenn V, Morawietz L, Haupl T, Neidel J, Petersen I, Konig A. Grading of chronic synovitis--a histopathological grading system for molecular and diagnostic pathology. Pathol Res Pract 2002;198:317-25. 31 The Atlas of Standard Radiographs of Arthritis. Rheumatology 2005;44:iv43-iv72. 32 Ehara T and Shigematsu H. Contribution of mast cells to the tubulointerstitial lesions in IgA nephritis. Kidney Int 1998;54:1675-83. 33 Buckley MG, Walters C, Wong WM, Cawley MI, Ren S, Schwartz LB et al. Mast cell activation in arthritis: detection of alpha- and beta-tryptase,

histamine and eosinophil cationic protein in synovial fluid. Clin Sci (Lond) 1997;93:363-70. Kiener HP, Hofbauer R, Tohidast-Akrad M, Walchshofer S, Redlich K, Bitzan P et al. Tumor necrosis factor alpha promotes the expression of stem cell factor in synovial fibroblasts and their capacity to induce mast cell chemotaxis. Arthritis Rheum 2000;43:164-74. Suurmond J, Rivellese F, Dorjee AL, Bakker AM, Rombouts YJ, Rispens T et al. Toll-like receptor triggering augments activation of human mast cells by anti-citrullinated protein antibodies. Ann Rheum Dis 2014; Xiang Y, Masuko-Hongo K, Sekine T, Nakamura H, Yudoh K, Nishioka K et al. Expression of proteinase-activated receptors (PAR)-2 in articular chondrocytes is modulated by IL1beta, TNF-alpha and TGF-beta. Osteoarthritis Cartilage 2006;14:1163-73. Ferrell WR, Kelso EB, Lockhart JC, Plevin R, McInnes IB. Protease-activated receptor 2: a novel pathogenic pathway in a murine model of osteoarthritis. Ann Rheum Dis 2010;69:2051-4. Seitz S, Barvencik F, Koehne T, Priemel M, Pogoda P, Semler J et al. Increased osteoblast and osteoclast indices in individuals with systemic mastocytosis. Osteoporos Int 2013;24:2325-34. Smith R, Ransjo M, Tatarczuch L, Song SJ, Pagel C, Morrison JR et al. Activation of proteaseactivated receptor-2 leads to inhibition of osteoclast differentiation. J Bone Miner Res 2004;19:507-16. Hoermann G, Cerny-Reiterer S, Perne A, Klauser M, Hoetzenecker K, Klein K et al. Identification of oncostatin M as a STAT5-dependent mediator of bone marrow remodeling in KIT D816V-positive systemic mastocytosis. Am J Pathol 2011;178:2344-56. Cirillo R, de PA, Ciccarelli A, Triggiani M, Marone G. Ciclosporin A inhibits mediator release from human Fc epsilon RI+ cells by interacting with cyclophilin. Int Arch Allergy Appl Immunol 1991;94:76-7. Eklund KK. Mast cells in the pathogenesis of rheumatic diseases and as potential targets for anti-rheumatic therapy. Immunol Rev 2007;217:38-52. Frenzel L and Hermine O. Mast cells and inflammation. Joint Bone Spine 2013;80:141-5. de PA, Ciccarelli A, Marino I, de CG, Marino D, Marone G. Human synovial mast cells. II. Heterogeneity of the pharmacologic effects of antiinflammatory and immunosuppressive drugs. Arthritis Rheum 1997;40:469-78.

CHAPTER 5 DEGREE OF SYNOVITIS ON MRI BY COMPREHENSIVE WHOLE KNEE SEMI-QUANTITATIVE SCORING METHOD CORRELATES WITH HISTOLOGIC AND MACROSCOPIC FEATURES OF SYNOVIAL INFLAMMATION IN KNEE OSTEOARTHRITIS

de Lange-Brokaar BJE, Ioan-Facsinay A, Yusuf E, Visser AW, Kroon HM, Andersen SN, Herb-van Toorn L, van Osch GJVM, Zuurmond A-M, Stojanovic-Susulic V, Bloem JL, Nelissen RGHH, Huizinga TWJ, Kloppenburg M

Osteoarthritis Cartilage. 2014 Oct: 22 (10): 1606-13.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 5

ABSTRACT Objective: To evaluate the association between synovitis on contrast enhanced (CE) MRI with microscopic and macroscopic features of synovial tissue inflammation. Methods: Forty-one patients (mean age 60 years, 61% women) with symptomatic radiographic knee OA were studied: twenty underwent arthroscopy (macroscopic features were scored (0–4), synovial biopsies obtained), twenty-one underwent arthroplasty (synovial tissues were collected). After haematoxylin and eosin staining, the lining cell layer, synovial stroma and inflammatory infiltrate of synovial tissues were scored (0–3). T1-weighted CE-MRI’s (3 T) were used to semi-quantitatively score synovitis at 11 sites (0–22) according to Guermazi et al. Spearman’s rank correlations were calculated. Results: The mean (SD) MRI synovitis score was 8.0 (3.7) and the total histology grade was 2.5 (1.6). Median (range) scores of macroscopic features were 2 (1–3) for neovascularization, 1 (0–3) for hyperplasia, 2 (0–4) for villi and 2 (0–3) for fibrin deposits. The MRI synovitis score was significantly correlated with total histology grade [r = 0.6], as well as with lining cell layer [r = 0.4], stroma [r = 0.3] and inflammatory infiltrate [r = 0.5] grades. Moreover, MRI synovitis score was also significantly correlated with macroscopic neovascularization [r = 0.6], hyperplasia [r = 0.6] and villi [r = 0.6], but not with fibrin [r = 0.3]. Conclusion: Synovitis severity on CE-MRI assessed by a new whole knee scoring system by Guermazi et al. is a valid, non-invasive method to determine synovitis as it is significantly correlated with both macroscopic and microscopic features of synovitis in knee OA patients.

Synovitis on MRI by new scoring method correlates with synovial tissue inflammation in knee OA

INTRODUCTION For a long time, OA was considered a non-inflammatory condition. More recently, however, it became evident that synovial inflammation could play an important role in the pathophysiology of OA1-6 as it is a predictor of cartilage destruction7,8 and a determinant of pain9,10. Although the biological processes underlying the appearance of synovial inflammation are poorly understood, it has been suggested that cartilage breakdown products could lead to activation of immune cells and production of pro- and anti-inflammatory mediators, which in turn could stimulate further cartilage breakdown, creating a negative feedback loop in OA6. Likewise, it is unclear how synovial inflammation evolves during the disease course, since studies investigating this topic were conflicting2,4,5,11. Therefore, more studies are necessary to elucidate the evolution of synovial inflammation during the disease course. Histological assessment of synovial biopsies is currently the golden standard for evaluating synovitis in knee OA. However, acquisition of synovial biopsies is technically difficult and patient unfriendly as it involves an invasive procedure like arthroscopy. Similar difficulties are encountered with another method used for assessing synovitis, the macroscopic scoring of the synovial tissue during arthroscopy. Moreover, no validated scoring system exists for the latter method. Therefore, a non-invasive method, such as contrast enhanced (CE) MR imaging constitutes an attractive alternative for visualizing synovial tissue inflammation12-14. Synovial inflammation in OA has proven to be different from rheumatoid arthritis15, therefore a specific scoring method for OA is warranted. Several scoring methods for synovitis on CEMRI in OA exist2,16,17, but only a limited number of studies focused on validating MRI scoring methods by comparing with histological scores2,3,13. Furthermore, the scoring method should encompass a sufficient number of compartments, as the anatomical distribution is patchy and heterogeneous18. A recently developed method by Guermazi etÂ al. is a semiquantitative method that scores synovitis on CE-MRI at 11 different sites throughout the knee16 and constitutes a comprehensible and practical method for assessing inflammation in the whole joint. This method, however, has not yet been validated with histology samples. In the present study, we aimed at validating the assessment of synovitis on CE-MRI with a new comprehensive whole knee synovitis scoring system in knee OA by comparison with histologic and macroscopic features of synovial tissue inflammation and by correlation with pain severity. Furthermore, we evaluated synovitis at different stages of OA to elucidate the role of synovial inflammation in different clinical stages of OA.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 5

METHODS Study design This study is part of the ongoing GEneration of Models, Mechanism & Markers for STratification of OsteoArthritis patieNts study (geMstoan), an observational study in established and end-stage knee OA patients to find new biomarkers for OA progression. This study has been approved by the ethics committee of the Leiden University Medical Center (LUMC). All patients provided written informed consent. Patients Between 2008 and 2012, patients with symptomatic radiographic primary knee OA, following the clinical and radiographic ACR criteria set19, attending the rheumatology or orthopaedic department of the LUMC or orthopaedic department of the Diaconessenhuis, Leiden, were included. The geMstoan study comprises of two groups of patients: one group of patients with end-stage disease that were planned to receive an arthroplasty and another group with mild to established OA that had no indication for an arthroplasty. Patients with mild to established disease received an arthroscopy. Patients with other rheumatic diseases, using immunosuppressive drugs or having knee injections (corticosteroids, etc) in the past 3 months were excluded. Patients with renal insufficiency (Cockroft-Gault < 60 mL/min) did not undergo CE-MRI. Patients using anticoagulants did not undergo arthroscopy. Patients having both synovial tissue samples and CE-MRI images were included in the present study. Arthroplasty Synovial tissue samples were collected from patients admitted for arthroplasty for OA to the orthopaedic departments. One sample of 3 × 3 × 3 mm was obtained from a random location form the knee. Arthroscopy and macroscopic scoring of synovial tissue Arthroscopy was performed only for the purpose of this research study, using a smallbore 2.7 mm arthroscope (Storz, Turrlingen, Germany) with sterile technique, as described previously20,21. After maximal needle aspiration of synovial fluid, intra-articular and local skin anaesthesia was achieved by a lidocaine injection (1%). The skin inferolateral to the patella was also injected with lidocaine 0.5%. Two small skin incisions were made to introduce two portals into the joint. The lower portal was used for introduction of the arthroscope and instillation of saline. The upper portal was used for collecting 15–20 blind synovial tissue biopsies and draining of the saline. Biopsies were taken from the patellar regions of the medial capsule using 2.0 mm forceps. All biopsies were physically combined to create one

Synovitis on MRI by new scoring method correlates with synovial tissue inflammation in knee OA

tissue block. Arthroscopic exploration was combined with joint lavage with at least 1000 ml of saline. At the end of the procedure 6 ml of 0.5% marcaïne was administered. During the procedure macroscopic features (neovascularization, villi, fibrin deposits and hyperplasia) were visualized by using a 30° flex scope (magnification 40×) and the general appearance of synovial tissue was semi-quantitatively scored from 0 to 4 (0 = absent, 1 = little, 2 = moderate, 3 = much/many, 4 = very much present/very many) by one physician according to a non-validated scoring method, which has been used for over 10 years in our centre. The physician was blinded for imaging data during evaluation. Synovial tissue samples handling and staining All synovial tissue samples were fixated in formalin for approximately 24 h and then transferred to 70% ethanol in which they were stored until they were imbedded in paraffin. Fifty-five μm thick coupes were cut. From 50 coupes of tissue, three coupes in the beginning of the tissue block and three coupes from the middle part of tissue block were selected and stained with haematoxylin and eosin (H&E) after deparaffinization. From these 2–6 coupes for each patient had a high enough quality to be scored. Scoring synovial tissue samples The samples were microscopically scored for three features, according to Krenn et al.22 lining cell layer, synovial stroma and inflammatory infiltrate. The grading system of Krenn et al. 22 was modified in the present study as follows: for each feature the samples with the least severe and the most severe score were determined and they provided the basis for developing a scoring system for each feature ranging from 0 (least severe) to 3 (most severe) (Fig. 1). Subsequently, all synovial tissue samples were scored by three independent observers who were blinded to MRI and clinical data; for each feature an average grade was calculated (0–3) for each patient. Furthermore, a total grade was calculated (sum of averaged grades of all three features (0–9)). In case the scoring of one observer was evidently different from the other two, one rescoring was performed by that observer. MRI acquisition Patients underwent MRI of the index knee less than 7 days and more than 36 h before the arthroscopy. MR scanning was performed on a 3 T MR system (Philips Achieva, Philips Healthcare, Best, The Netherlands) using an eight-channel dedicated knee coil. Gadolinium (Gd) contrast agent (Gd-DOTA, Dotarem, Guerbet, 0.2 ml/kg @ 2 ml/s followed by a saline flush 40 ml @ 2 ml/s) was injected in the cubital vein for visualizing synovitis. Both axial and sagittal CE, T1-weighted, turbo spin echo (TSE), spectral presaturation with inversion recovery (SPIR) sequences were used for scoring synovitis. Scan parameters were (for axial and sagittal): multi-slice, spin echo, TSE factor 6, TR 655 ms, TE 20 ms, FA 90, FOV

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 5

160 × 160 mm, pixel size 0.75 × 0.75 mm, slice thickness 2.5 mm, slice gap 0.8 mm, 24 slices. Sequences were obtained between 8 and 10 min after contrast injection.

Figure 1. Scoring system for scoring histology features based on own samples. Representative pictures of synovial tissue samples with H&E staining of all three features: lining layer, stroma and infiltrate. Grade 0 (normal) and Grade 3 (most severe) sample for each feature are shown.

MRI scoring Sagittal and axial T1-weighted CE-MR images (3 T) were used to semi-quantitatively score synovitis at 11 different sites according to Guermazi et al. Ref. 16. Synovial thickness was measured and scored as followed: 0, when synovial thickness was less than 2 mm, 1 when thickness was between 2 and 4 mm and 2 when synovial thickness was above 4 mm. The total sum score of 11 sites was used for the analysis (range 0–22). A total score of 0–4 was considered normal (no synovitis); 5–8 represents a mild, 9–12 a moderate and above 13 a severe synovitis 16. All MR images were analysed by means of consensus between two readers both experienced in reading MR images of the knee (BDL and WV). Both BDL and WV (over 1000 MRI’s scored) have more than 3 years of experience in scoring knee MR images.

Synovitis on MRI by new scoring method correlates with synovial tissue inflammation in knee OA

Scoring was done after extensive learning sessions and under supervision of experienced musculoskeletal radiologist (JB). During the assessment, the readers were blinded to radiographic results and patient data. Intraclass correlation (ICC, with 95% confidence interval (CI)) was based on random sample of 14 CE-MRI’s and was 0.84 (0.58–0.95). Scoring knee radiographs Radiographs (posterior anterior (PA) fixed flexion) were obtained for all patients. Radiographs were scored, blinded for patient characteristics, by an experienced musculoskeletal radiologist (HK), with 30 years of experience in scoring musculoskeletal radiographs, according to the Kellgren–Lawrence (KL) scale23. The ICC, with 95% CI was based on a randomly selected sample of 36 radiographs (17 right and 17 left knees) and was 0.99 (0.98–0.99). The knees with KL < 2 were rescored in consensus between HK and an experienced rheumatologic reader (MK). Clinical data Self-reported pain of the index knee was assessed by the visual analogue scale (VAS, 0–100) within 2 weeks of MRI acquisition. Statistics Parameters normally distributed are described as means (SD), otherwise medians (ranges) are given. Comparison between groups was calculated with independent t-test for MRI synovitis score, total histology grade, age, fat percentage and VAS, and Mann–Whitney U test for histology features, KL grade and BMI. Chi-squared test was used for comparison of proportions for gender and index knee, Chi-squared test for trend was used for comparison of BMI groups, KL grade groups and MRI synovitis groups. Spearman’s rank correlations for all patients were used for correlation between total MRI synovitis score and total histology synovitis grade, histology features, macroscopy features and VAS. A correlation <0.3 was considered as weak, 0.3–0.7 as moderate and >0.7 as strong. SPSS 20.0 was used for statistical analyses.

RESULTS Patient characteristics Of 95 patients included in the geMstoan study, 42 patients had both CE-MRI and synovial tissue available. Those 42 patients did not differ for age, sex, BMI, fat percentage, KL grade or VAS pain from original 95 patient samples (data not shown). One patient developed

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 5

after 1 year a CCP positive, rheumatoid factor positive oligoarthritis and was excluded, resulting in 41 patients for present study. Contrast administration was well tolerated by all patients. The patients receiving arthroscopy had a median KL score of 2 (1–4) and patients receiving arthroplasty had a median KL score on radiographs of 3 (2–4), difference was significant between groups. There were no significant differences in gender, age, BMI and fat percentage between the groups. As expected, mean (SD) VAS pain was significantly lower in the arthroscopy (42.9 (23.6), range (4–84)) than in the arthroplasty group (61.4 (14.5), range (32–79)); mean difference was −18.5 (95% CI −31.0 to −5.9). Patient characteristics are displayed in Table I. Table I. Patients characteristics of all (n = 41) patients and different groups: patients receiving arthroscopy (n = 20) and patients receiving arthroplasty (n = 21) Parameters

All

Arthroscopy group

Arthroplasty group

Diff (95% CI)/P-value

Age, yrs, mean (SD) Female sex, no. (%) BMI, kg/m2 - Total, median (range) - Groups, no. (%) <25 25–30 30–35 35–40 >40 Fat percentage, mean (SD) Right knee, no. (%)

60.2 (9.0) 25 (61.0)

60.8 (7.2) 14 (70.0)

59.7 (10.7) 11 (52.4)

1.0 (−4.7 to 6.8)* P = 0.248***

29.1 (22.2–43.9) 29.2 (24.4–43.9) 29.1 (22.2–40.9) P = 0.696** 4 (9.8) 23 (56.1) 9 (22.0) 2 (4.9) 3 (7.3) 36.6 (8.4) 25 (61.0)

1 (5.0) 13 (65.0) 3 (15.0) 1 (5.0) 2 (10.0) 38.5 (8.1) 11 (55.0)

3 (14.3) 10 (47.6) 6 (28.6) 1 (4.8) 1 (4.8) 34.8 (8.4) 14 (66.7)

P = 0.704****

3.0 (1–4)

2.0 (1–4)

3.0 (2–4)

P < 0.001**

2 (4.9%) 12 (29.3%) 16 (39.0%) 11 (26.8%) 52.4 (21.4)

2 (10.0%) 11 (55.0%) 6 (30.0%) 1 (5.0%) 42.9 (23.6)

0 1 (4.8%) 10 (47.6%) 10 (47.6%) 61.4 (14.5)

P < 0.001****

3.7 (−1.5 to 9.0)* P = 0.444***

KL grade - Total, median (range) - Groups, no. (%) 1 2 3 4 VAS pain knee, mm, mean (SD)

−18.5 (−31.0 to −5.9)*

* Independent t-test, ** Mann–Whitney U test, ***Chi-squared test and **** Chi-squared trend test.

Synovitis on CE-MR images (Table II) The mean (SD) synovitis score on MRI for all patients was 8.0 (3.7), representing a mild synovitis, and was significantly lower in patients in the arthroscopy (6.1 (2.6)) than in the arthroplasty group (9.7 (3.8)); mean difference −3.6 (95% CI −5.7 to −1.6). Representative examples of CE-MR images of patient from both groups are displayed in Fig.2.

Synovitis on MRI by new scoring method correlates with synovial tissue inflammation in knee OA

Table II. MRI synovitis total scores and distribution of severity of synovitis score at MRI in patients with knee OA MRI synovitis score Total (0–22), mean (SD) Groups, no. (%) Normal (0–4) Mild (5–8) Moderate (9–12) Severe (>13)

All patients (n = 41) 8.0 (3.7)

Arthroscopy group Arthroplasty group (n = 20) (n = 21) 6.1 (2.6) 9.7 (3.8)

Mean difference (95% CI)/P-value −3.6 (−5.7 to −1.6)*

6 (14.6) 21 (51.2) 9 (22.0) 5 (12.2)

4 (20.0) 13 (65.0) 3 (15.0) 0

P = 0.009****

2 (9.5) 8 (38.1) 6 (28.6) 5 (23.8)

* Independent t-test, **** Chi-squared trend test.

When all 11 sites on CE-MRI were investigated and compared between the arthroscopy and the arthroplasty group, no patients in the arthroscopy group had a total MRI score representing a severe synovitis, whereas five patients (24%) in the arthroplasty group had. Overall, the arthroplasty group showed increased synovial thickness at all 11 different sites, while the arthroscopy group showed less synovial thickness at all sites.

Figure 2. Examples of histology features (H&E staining) and CE-MRI from a patient from the arthroscopy group and a patient from the arthroplasty group. Patient from the arthroscopy group displaying a mild synovitis and patient from the arthroplasty displaying a severe synovitis. Both histology features and CE-MR images are shown. Open black arrow = hyperplasia lining layer, Black arrow = infiltrate, open white arrow = Bakers Cyst with synovial thickening, white arrow = synovial inflammation.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 5

Histology of synovial tissue of OA patients (Table III) A mean total histology grade of 2.5 was observed for all patients. Medians in all patients were 0.8 for the lining layer, 0.9 for the stroma and 0.5 for the infiltrate. Overall, 60% of patients showed an inflammatory infiltrate. More patients in the arthroplasty group showed infiltrates (69%) compared to the arthroscopy group (50%). Table III. Histological total synovitis grade and distribution of histological features in patients with knee OA Histological synovitis grade

All patients (n = 41)

Total (0–9), mean (SD) 2.5 (1.6) Features, median (range) Lining layer (0–3) 0.8 (0–2.8) Stroma (0–3) 0.9 (0–2.7) Infiltrate (0–3) 0.5 (0–3)

Arthroscopy group (n = 20)

Arthroplasty group (n = 21)

Mean difference (95% CI)

P-value

2.0 (1.4)

3.0 (1.7)

−1.0 (−2.0 to −0.07)*

0.037*

0.5 (0–2.8) 0.8 (0–2.7) 0.2 (0–1.8)

1 (0–2.5) 1 (0–2.3) 0.9 (0–3)

NA NA NA

0.027** 0.601** 0.072**

*Independent t-test, ** Mann–Whitney U test.

The mean (SD) total histology grade differed significantly between patients with mild/ established and patients with end-stage knee OA (2.0 (1.4) vs 3.0 (1.7), respectively). The grades for both the lining layer and inflammatory infiltrate were significantly higher in the arthroplasty group. Median (range) for the lining layer was 0.5 (0–2.8) for the arthroscopy group vs 1 (0–2.5) for the arthroplasty group. For the infiltrate median (range) was 0.2 (0– 1.8) vs 0.9 (0–3). For the stroma no significant differences were seen between the groups. Representative examples of histology features of a patient in both groups are displayed in Fig 2. Macroscopy features assessed during arthroscopy In 22 patients undergoing arthroscopy, the median (range) score for macroscopic features was 2 (1–3) for neovascularization, 1 (0–3) for hyperplasia, 2 (0–4) for villi and 2 (0–3) for fibrin. Correlations between synovitis assessed by CE-MRI, histology and macroscopy and association with pain As summarized in Table IV, there was a moderate significant correlation between the total synovitis grade on MRI and the total histology grade, lining layer, stroma cells and inflammatory infiltrate. Furthermore, the MRI synovitis grade was significantly correlated with the macroscopic features: neovascularization, hyperplasia and villi, but not with fibrin. A significant correlation between total histology grade and the macroscopic neovascularization was seen as well. Moreover, a moderate, but statistically significant correlation was seen between the MRI synovitis score and the VAS pain level. In arthroscopy group correlation of

Synovitis on MRI by new scoring method correlates with synovial tissue inflammation in knee OA

MRI synovitis score with total histology grade was a little bit higher 0.58 (P = 0.007), while in the arthroplasty groups the correlation was lower 0.45 (P = 0.041). Table IV. Correlations between macroscopic features at arthroscopy, histological synovitis grade and pain score with total MRI synovitis score in patients with knee OA Parameters Macroscopy (n = 20)

Histology (n = 41)

VAS pain (n = 41)

Neovascularization (0–4) Hyperplasia (0–4) Villi (0–4) Fibrin (0–4) Total synovitis grade (0–9) Lining layer (0–3) Stroma (0–3) Infiltrate (0–3)

Spearman’s rank correlations 0.64 0.64 0.61 0.34 0.57 0.38 0.31 0.47 0.32

P-value 0.002 0.002 0.004 0.149 <0.001 0.015 0.049 0.002 0.041

DISCUSSION To our knowledge we are the first to validate a new comprehensive and practical synovitis scoring method for assessing degree of synovitis on CE-MRI in the whole knee. We found a significant correlation of total synovitis score on MRI with both macroscopic and microscopic features of synovitis, which indicates that the method by Guermazi et al.16 is a valid method to assess degree of inflammation in knee OA patients. A significant correlation between VAS pain scores and synovial inflammation on CE-MRI was also seen. Furthermore we found that the total histology grade and the MRI synovitis score was less in patients undergoing an arthroscopy, representing a mild to established knee OA patient group, than in patients receiving an arthroplasty, representing end-stage knee OA. Therefore, it seems that synovial inflammation is a more pronounced feature of end-stage knee OA patients. In the present study we found a moderate significant correlation of synovitis severity on MR images with both histological and macroscopic features of synovitis and with pain scores. Our correlation with pain was as expected moderate (0.3), while pain is multidimensional and other dimensions of pain (depression, personality, etc.) were not included in our study. Our dedicated set-up enabled us to find these correlations. Firstly, synovial inflammation was scored on contrast enhanced (CE) MR images, as this is the best way to distinguish synovial inflammation from synovial fluid and fat tissues12,13. Studies that investigated the correlation cross-sectionally between severity of only synovitis (not the combination of synovitis and effusion) in osteoarthritis assessed on non-CE MR images and VAS pain, no significant correlation was found24,25. Secondly, we used a 3 T scanner, instead of a 1.5 T scanner, which is more often used, but have a lower signal to noise ratio. Finally, in the

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 5

present study extensive synovitis scoring was performed addressing 11 sites in the whole knee. These advantages may explain the differences with other studies performed in the same research field. Loeuille et al. assessed synovitis on MR images and performed a validation with clinical symptoms, histology samples and macroscopic features2,3,13. Although contrast enhancement was used, no correlation was found between the MRI synovitis score and VAS pain. Furthermore, a lower correlation with histology was found (r = 0.41 vs 0.6 in the present study). The lower correlations found could be explained by the use of a scoring method for synovitis severity on MR images addressing only five sites in the knee. The present extensive scoring would have allowed more comprehensive evaluation. The aim of present study was to investigate whether synovitis on CE-MRI as assessed in all compartments of the knee could be used as surrogate for synovial tissue inflammation. Therefore, we compared total MRI score with the total histology grade. In arthroplasty patients we investigated a random sample of synovial tissue from the knee, while in the arthroscopy group we evaluated the medial capsule, since this is the only site that can be reached safely. It could have been possible that the correlation of the total score on MRI with the total histology score would have been higher in arthroscopy patients if we had biopsies from the whole knee instead of only at a selected site of all our patients or if we had performed visualized sampling of the tissue to assess the exact location of the biopsies enabling more direct comparisons. However, this was not observed since the correlation between the MRI score of the two medial sites (creating a range 0–4) with the total histology grade in the arthroscopy patient group (n = 20), was the same (0.58 (P = 0.008)) as the correlation between the total MRI score and total histology grade in the arthroscopy group (0.58 (P = 0.007)). This observation suggests that sampling of the medial site during arthroscopy is a good representation of the synovitis in the whole knee, but these results should be interpreted with caution. New in this study is the extensive investigation of synovitis in different stages of OA. Not only were both MRI and histology used to compare mild to established OA to end-stage disease, but we were also able to study real mild to established knee OA since we investigated patients with knee OA without a clinical indication for arthroscopy. MRI and histology data, although scored independently of each other, are supporting each other; We found that the patients with end-stage OA, had a higher MRI synovitis score, a higher total histology score of the synovial biopsies and reported more pain, compared to patients with mild to established OA. In the literature results regarding the degree of synovial inflammation at different stages of OA are conflicting and vary from no difference to more inflammation in patient with earlier than in end-stage OA or to more inflammation in patients with endstage than in earlier OA2,4,5,11. Differences in results between the studies could be explained by dissimilarity in defining the stage of OA in the cases under study, and by the methods used to score the histological samples and MR images. In a study by Pearle et al. in primary

100

Synovitis on MRI by new scoring method correlates with synovial tissue inflammation in knee OA

OA patients no differences were found in the prevalence of synovial infiltrates between early and advanced OA. Early OA was defined as KL grade ≤ 2 and no intraoperative evidence of full-thickness chondral wear, while advanced was defined as KL > 2 and intraoperative evidence of full-thickness chondral loss. In this study no MR imaging was performed and the sample size of the patients in the early group was only 9, vs 43 in the advanced group Pearle et al.4. In a study by Benito et al. more pronounced inflammation was seen in early OA patient group than in end-stage OA. Early OA was defined when patients had clinical features of OA and cartilage degeneration during arthroscopy, but no radiographic OA signs5, which is different from the present study where all patients had radiographic knee OA. In the study by Benito et al. also no MR images were available. Two studies found a higher synovitis severity score in late-stage OA than in earlier phases of the disease as in the present study. Loeuille et al. found microscopically a significantly lower mean total composite score and infiltration score for the early OA (mild chondral lesions at arthroscopy) group2 than for the end-stage (severe cartilage lesions with exposure of subchondral bone at arthroscopy) group, but no difference in MRI synovitis score between the groups. Smith et al. found in a histological study more infiltrate and a trend towards increased synovial lining layer in endstage OA, when compared to early OA11. The present study has some limitations. A relatively small number of patients were included in the present study (n = 41), therefore findings should be interpreted with caution and should be replicated in larger samples. Furthermore, in the present study we scored macroscopic features according to a nonvalidated method. Moreover, reproducibility could not be done, because repetitive procedures are not ethical. However the person performing the arthroscopy was blinded for CE-MRI data, therefore misclassification is random. But, taken together data concerning macroscopic features should be interpreted with caution. The time from Gd injection to acquisition of T1-W images was 8–10 min due to the fact that our protocol included several dynamic sequences (that are not part of the current analysis). Although, some controversy exists concerning the optimal time after Gd injection17,18,26 we feel this could have potentially have led to washout of the contrast into the cavity and might have led to increased measurements of synovial thickness. However, because the aim of present study was to compare findings between patients and in all patients the time after injection was comparable, the actual measurements of synovial membrane were of lesser importance and this is less of a problem. Finally, we modified the validated histological synovitis scoring system by Krenn et al. Therefore our mean total histology scores cannot be compared to results that use the synovitis score by Krenn et al. 22. However this modification made our score more sensitive to discriminate between groups.

101

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 5

In present study we validated a new whole knee scoring system on CE-MRI by Guermazi et al. by comparing synovial inflammation on CE-MRI with histological and macroscopic features of the synovial tissue. Furthermore, we have shown that synovial inflammation was more prevalent and severe in end-stage knee OA and that synovial inflammation is of clinical importance as it is associated with pain. Further research is necessary to elucidate the role of synovitis in the pathophysiology in OA. Author contributions Authors made substantial contributions to the following: (1a) conception and design of the study: BDL, GVO, AMZ, VSS, TH, MK; (1b) acquisition of data: BDL, AIF, EY, AWV, HK, SA, LVT, JB, RN, MK; (1c) interpretation of data BDL, AIF, GVO, AMZ, VSS, JB, TH, MK; (2) drafting or critical revision of manuscript: BDL, AIF, EY, AWV, HK, SA, LVT, GVO, AMZ, VSS, JB, RN, TH, MK; (3) final approval of manuscript BDL, AIF, EY, AWV, HK, SA, LVT, GVO, AMZ, VSS, JB, RN, TH, MK. Role of funding source Financial support was obtained from TI Pharma, however TI Pharma did not contribute to design, interpretation of data, drafting and final approval of the manuscript. Conflict of interests None. Acknowledgement This work was performed within the framework of Dutch Top Institute Pharma, project “Generation of models, mechanisms and markers for stratification of osteoarthritis patients” (project nr. T1-213). The authors would like to acknowledge support of the cooperating hospital Diaconessenhuis, Leiden and referring Rheumatologists, Orthopaedic Surgeons and Nurse practitioners.

102

Synovitis on MRI by new scoring method correlates with synovial tissue inflammation in knee OA

REFERENCES

1 Felson DT. Clinical practice. Osteoarthritis of the knee. N Engl J Med 2006;354:841-8. 2 Loeuille D, Chary-Valckenaere I, Champigneulle J, Rat AC, Toussaint F, Pinzano-Watrin A, et al. Macroscopic and microscopic features of synovial membrane inflammation in the osteoarthritic knee: correlating magnetic resonance imaging findings with disease severity. Arthritis Rheum 2005;52: 3492-501. 3 Loeuille D, Rat AC, Goebel JC, Champigneulle J, Blum A, Netter P, et al. Magnetic resonance imaging in osteoarthritis: which method best reflects synovial membrane inflammation? Correlations with clinical, macroscopic and microscopic features. Osteoarthritis Cartilage 2009;17:1186-92. 4 Pearle AD, Scanzello CR, George S, Mandl LA, DiCarlo EF, Peterson M, et al. Elevated high-sensitivity C-reactive protein levels are associated with local inflammatory findings in patients with osteoarthritis. Osteoarthritis Cartilage 2007;15:516-23. 5 Benito MJ, Veale DJ, FitzGerald O, van den Berg WB, Bresnihan B. Synovial tissue inflammation in early and late osteoarthritis. Ann Rheum Dis 2005;64:1263-7. 6 Sellam J, Berenbaum F. The role of synovitis in pathophysiology and clinical symptoms of osteoarthritis. Nat Rev Rheumatol 2010;6:62535. 7 Roemer FW, Guermazi A, Felson DT, Niu J, Nevitt MC, Crema MD, et al. Presence of MRIdetected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30-month follow-up: the MOST study. Ann Rheum Dis 2011;70:1804-9. 8 Roemer FW, Zhang Y, Niu J, Lynch JA, Crema MD, Marra MD, et al. Tibiofemoral joint osteoarthritis: risk factors for Mrdepicted fast cartilage loss over a 30-month period in the multicenter osteoarthritis study. Radiology 2009;252:772-80. 9 Yusuf E, Kortekaas MC, Watt I, Huizinga TW, Kloppenburg M. Do knee abnormalities visualised on MRI explain knee pain in knee osteoarthritis? A systematic review. Ann Rheum Dis 2011;70:60-7. 10 Baker K, Grainger A, Niu J, Clancy M, Guermazi A, Crema M, et al. Relation of synovitis to knee pain using contrastenhanced MRIs. Ann Rheum Dis 2010;69:1779-83.

11 Smith MD, Triantafillou S, Parker A, Youssef PP, Coleman M. Synovial membrane inflammation and cytokine production in patients with early osteoarthritis. J Rheumatol 1997;24:365-71. 12 Guermazi A, Roemer FW, Hayashi D. Imaging of osteoarthritis: update from a radiological perspective. Curr Opin Rheumatol 2011;23:484-91. 13 Loeuille D, Sauliere N, Champigneulle J, Rat AC, Blum A, Chary-Valckenaere I. Comparing non-enhanced and enhanced sequences in the assessment of effusion and synovitis in knee OA: associations with clinical, macroscopic and microscopic features. Osteoarthritis Cartilage 2011;19:1433-9. 14 Hayashi D, Roemer FW, Katur A, Felson DT, Yang SO, Alomran F, et al. Imaging of synovitis in osteoarthritis: current status and outlook. Semin Arthritis Rheum 2011;41:116-30. 15 Ayral X. Diagnostic and quantitative arthroscopy: quantitative arthroscopy. Baillieres Clin Rheumatol 1996;10:477-94. 16 Guermazi A, Roemer FW, Hayashi D, Crema MD, Niu J, Zhang Y, et al. Assessment of synovitis with contrast-enhanced MRI using a wholejoint semiquantitative scoring system in people with, or at high risk of, knee osteoarthritis: the MOST study. Ann Rheum Dis 2011;70:805-11. 17 Rhodes LA, Grainger AJ, Keenan AM, Thomas C, Emery P, Conaghan PG. The validation of simple scoring methods for evaluating compartment-specific synovitis detected by MRI in knee osteoarthritis. Rheumatology (Oxford) 2005;44:1569-73. 18 Roemer FW, Kassim JM, Guermazi A, Thomas M, Kiran A, Keen R, et al. Anatomical distribution of synovitis in knee osteoarthritis and its association with joint effusion assessed on non-enhanced and contrast-enhanced MRI. Osteoarthritis Cartilage 2010;18:1269-74. 19 Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K, et al. Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arthritis Rheum 1986;29:1039-49. 20 van Oosterhout M, Sont JK, Bajema IM, Breedveld FC, van Laar JM. Comparison of efficacy of arthroscopic lavage plus administration of corticosteroids, arthroscopic

103

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 5

104

lavage plus administration of placebo, and joint aspiration plus administration of corticosteroids in arthritis of the knee: a randomized controlled trial. Arthritis Rheum 2006;55:964-70. van Oosterhout M, Sont JK, van Laar JM. Superior effect of arthroscopic lavage compared with needle aspiration in the treatment of inflammatory arthritis of the knee. Rheumatology (Oxford) 2003;42:102-7. Krenn V, Morawietz L, Haupl T, Neidel J, Petersen I, Konig A. Grading of chronic synovitis e a histopathological grading system for molecular and diagnostic pathology. Pathol Res Pract 2002;198:317-25. The atlas of standard radiographs of arthritis. Rheumatology 2005;44:iv43-72. Pelletier JP, Raynauld JP, Abram F, Haraoui B, Choquette D, Martel-Pelletier J. A new noninvasive method to assess synovitis severity in relation to symptoms and cartilage volume loss in knee osteoarthritis patients using MRI. Osteoarthritis Cartilage 2008;16(Suppl 3):S8S13. Hill CL, Hunter DJ, Niu J, Clancy M, Guermazi A, Genant H, et al. Synovitis detected on magnetic resonance imaging and its relation to pain and cartilage loss in knee osteoarthritis. Ann Rheum Dis 2007;66:1599-603. Ostergaard M, Klarlund M. Importance of timing of postcontrast MRI in rheumatoid arthritis: what happens during the first 60 minutes after IV gadolinium-DTPA? Ann Rheum Dis 2001;60:1050-4.

Synovitis on MRI by new scoring method correlates with synovial tissue inflammation in knee OA

105

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

PART II THE ROLE OF SYNOVITIS IN THE CLINICAL BURDEN OF OA

CHAPTER 6 ASSOCIATION OF PAIN IN KNEE OSTEOARTHRITIS WITH DISTINCT PATTERNS OF SYNOVITIS

de Lange-Brokaar BJE, Ioan-Facsinay A, Yusuf E, Visser AW, Kroon HM, van Osch GJVM, Zuurmond A-M, Stojanovic-Susulic V, Bloem JL, Nelissen RGHH, Huizinga TW, Kloppenburg M

Arthritis Rheumatol. 2015 Mar;67(3):733-40.

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 6

ABSTRACT Objective: To determine possible patterns of synovitis on contrast-enhanced magnetic resonance imaging (CE-MRI) and its relation to pain and severity in patients with radiographic knee osteoarthritis (OA). Methods: In total, 86 patients (mean age 62 years, 66% women, median body mass index 29 kg/m2) with symptomatic knee OA (Kellgren/Lawrence radiographic score 3) were included. T1-weighted, gadolinium-chelate–enhanced MRI with fat suppression was used to semiquantitatively score the extent of synovitis at 11 knee sites (total score range 0–22). Self-reported pain was assessed with 3 standardized questionnaires. Principal components analysis (PCA) was used to investigate patterns (the location and severity) of synovitis. Subsequently, these patterns were assessed for associations with pain measures and radiographic severity in adjusted logistic regression models. Results: Synovitis was observed in 86 patients and was found to be generally mild on CEMRI (median total synovitis score 7, range 0–16). The median pain scores were 53 (range 0–96) on the visual analog scale for pain, 51.4 (range 2.8–97.2) on the Knee Injury and Osteoarthritis Outcome Score (KOOS) for pain, 35 (range 0–75) on the Intermittent and Constant Osteoarthritis Pain (ICOAP) score for constant pain, and 40.6 (range 0–87.5) on the ICOAP score for intermittent pain. PCA resulted in extraction of 3 components, explaining 53.4% of the variance. Component 1 was characterized by synovitis at 7 sites (mainly medial parapatellar involvement) and was associated with scores on the KOOS pain subscale and the ICOAP constant pain subscale. Component 2 was characterized by synovitis at 4 sites (mainly the site adjacent to the anterior cruciate ligament), but was not associated with pain measures or with radiographic severity. Component 3, characterized by synovitis at 3 sites (mainly at the loose body site), was associated with radiographic severity. Conclusion: Different patterns of synovitis in knee OA were observed. The pattern that included several patellar sites was associated with pain, whereas other patterns showed no association, suggesting that pain perception in patients with knee OA is a localized response.

110

Association of pain in knee OA with distinct patterns of synovitis

INTRODUCTION Although one of the major outcomes of knee osteoarthritis (OA) is pain, the pathophysiologic and pain-causing mechanisms in OA remain largely unknown1. Studies in past years have elucidated the role of synovitis as one of the key players in pain perception, since it has been associated with pain in knee OA2,3. For a long time, OA was considered a noninflammatory condition. However, recently it became evident that synovial inflammation is prevalent in OA4 and could play an important role in the pathophysiology of OA5. Although histologic assessment of synovitis in human synovial biopsy tissue is currently the gold standard for defining synovial inflammation, contrast-enhanced magnetic resonance imaging (CE-MRI) has proven to be a good surrogate in evaluating synovitis in patients with knee OA6-11. Because the anatomic distribution of synovitis on CE-MRI is patchy and heterogeneous12, the optimal MRI scoring method should encompass a sufficient number of compartments. A method recently developed by Guermazi et al is a semiquantitative method that scores the extent of synovitis on CE-MRI at 11 different sites throughout the knee13and constitutes a comprehensible and practical method for assessing synovitis in the whole joint. Synovitis on CE-MRI as assessed by this scoring method compares well with synovial inflammation identified histologically in synovial biopsy tissue from patients with knee OA9. Only a small number of studies have investigated synovitis on CE-MRI, and even fewer studies have used a scoring system that encompasses a sufficient number of sites throughout the whole knee12,13. These studies have found that some sites of the knee display more synovitis than others. Whether synovitis at different sites within the knee joint occurs independently or whether distinct patterns of synovitis may form is currently unknown, since patterns of synovitis have not been investigated in an unbiased manner. Moreover, investigation of potential patterns of synovitis could contribute to the understanding of disease mechanisms in OA, as it might help to unravel both the underlying mechanisms of synovitis and the pain mechanisms involved in OA. In the present study, we aimed to investigate whether different patterns of synovitis, as assessed by its location and extent on gadolinium-chelateâ&#x20AC;&#x201C;enhanced MRI of the whole knee joint, exist in patients with symptomatic primary knee OA, and if these patterns exist, whether they are associated with the extent of pain and severity of radiographic knee OA.

111

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 6

METHODS Study design This study is part of the ongoing geMstoan Study (Generation of Models, Mechanism & Markers for Stratification of Osteoarthritis Patients), an observational study in patients with established or end-stage knee OA that is aimed at finding new biomarkers for OA progression. This study has been approved by the ethics committee of the Leiden University Medical Center (LUMC). All patients provided written informed consent. Patients Study patients comprised individuals with symptomatic radiographic primary knee OA who were attending the rheumatology or orthopedic department of the LUMC or orthopedic department of the Diaconessenhuis Hospital (Leiden, The Netherlands) between 2008 and 2013. Symptomatic radiographic knee OA was diagnosed according to the American College of Rheumatology classification criteria for knee OA14. The geMstoan Study involves 2 groups of patients stratified according to a clinical end point: one group of patients with end-stage knee OA who had received a total knee arthroplasty, and another group of patients with mild to established OA who had no indication for an arthroplasty. Patients with other rheumatic diseases who had received immunosuppressive drugs or knee injections (i.e., corticosteroids) in the past 3 months were excluded. Patients with renal insufficiency (Cockcroft-Gault glomerular filtration rate <60 ml/minute) did not undergo gadolinium-chelate–enhanced MRI. MRI acquisition MR scanning was performed on a 3T Philips Achieva MR system (Philips Healthcare) using an 8-channel dedicated knee coil. To visualize synovitis, gadoterate meglumine (0.2 ml/kg) (Dotarem; Guerbet) was injected in the cubital vein at a rate of 2 ml/second followed by a 40-ml saline flush at a rate of 2 ml/second. CE-MRI with T1-weighted, turbo spin-echo, spectral presaturation with inversion recovery sequences and fat suppression in both the axial and sagittal planes were used for scoring the extent of synovitis. Scan parameters (for both the axial and sagittal planes) were as follows: multislice spin-echo sequence, echo train length 6 msec, time to recovery 655 msec, time to echo 20 msec, field of view 160 × 160 mm, pixel size 0.75 × 0.75 mm, slice thickness 2.5 mm, slice gap 0.8 mm, and 24 slices. Sequences were obtained between 8 and 10 minutes after injection of the contrast agent.

112

Association of pain in knee OA with distinct patterns of synovitis

MRI scoring Sagittal and axial T1-weighted, gadolinium-chelate–enhanced MRI (3T Philips Achieva MR) was used to semiquantitatively score the extent of synovitis at 11 different sites within the whole knee joint, using the scoring system of Guermazi et al13. Synovial thickness was measured and scored as follows: 0 = synovial thickness <2 mm, 1 = synovial thickness between 2 and 4 mm, and 2 = synovial thickness >4 mm (Figure 1). The total summed score of the 11 sites was calculated (total score range 0–22). A total MRI synovitis score of 0–4 was considered normal (no synovitis), a total score of 5–8 represented mild synovitis, a total score of 9–12 represented moderate synovitis, and a total score of >13 represented severe synovitis13.

6 Figure 1. Scoring of synovitis on gadolinium-chelate contrast–enhanced magnetic resonance imaging (CE-MRI), using fat-suppressed T1-weighted MR images of the axial and sagittal planes. A, Sagittal CE-MRI at the level of the posterior cruciate ligament, showing synovitis and a large loose body (arrow). B, Sagittal CE-MRI at the level of the anterior cruciate ligament (ACL), showing synovitis adjacent to the ACL (#) and at the suprapatellar (white arrow), infrapatellar (*), and intercondylar (black arrow) sites. C, Axial CE-MRI, showing synovitis at both the medial (black arrow) and the lateral (white arrow) parapatellar sites and a large medial Baker’s cyst with surrounding white peripheral rim indicating synovitis (*). D, Sagittal CE-MRI at the level of the tibiofibular joint, showing synovitis at the posterior horn of the lateral meniscus (arrow). E, Sagittal CE-MRI, showing synovitis at the posterior horn of the medial meniscus (arrow) and a Baker’s cyst. F–J, CE-MRI of different knees at the same locations as in A–E, showing almost no synovitis. Please note the small Baker’s cyst (*) in H.

Bone marrow lesions (BMLs) and effusion were scored on axial and coronal proton density– weighted images, using the Knee Osteoarthritis Scoring System in 9 compartments, as described elsewhere15. BMLs were defined as an ill-defined area in the subchondral bone extending from the articular surface, with a grade range from 0 to 3 as follows: 0 = absent, 1 = minimal lesion (<5 mm), 2 = moderate lesion (5–20 mm), and 3 = severe lesion (≥20 mm). For the purposes of the present analysis, BML grades were recoded into a binary variable (0 = not present, 1 = present). Effusion was scored from 0 to 3, as follows: 0 = no joint effusion,

113

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 6

with only a small, physiologic sliver of synovial fluid, 1 = small effusion, with 1 or 2 distended joint recesses (suprapatellar pouch, medial or lateral patellar recess, dorsal femorotibial joint space, popliteal tendon sheath, recesses surrounding the cruciate ligaments, meniscosynovial recesses), 2 = moderate effusion, with more than 2 joint recesses partially distended, 3 = massive effusion, with full, marked distention of all of the joint recesses. For the purposes of the present analysis, effusion scores were recoded into a binary variable (0 = not present, 1 = present). All MR images were analyzed by 2 readers (BDL and AWV) by means of consensus. Both readers have >3 years of experience in scoring knee MR images (more than 1,000 MR images scored). Scoring was done after extensive learning sessions and under the supervision of an experienced musculoskeletal radiologist (JLB). During the assessment, the readers were blinded with regard to the radiographic findings and patient data. The intraclass correlation coefficient (ICC) was 0.84 (95% confidence interval [95% CI] 0.58–0.95), based on a random sample of 14 gadolinium-chelate–enhanced MR images. Knee radiograph scoring Radiographs of the knees of all patients (posteroanterior fixed-flexion view) were obtained. The radiographs were scored in a blinded manner by an experienced musculoskeletal radiologist (HMK) with 30 years of experience in scoring musculoskeletal radiographs. The Kellgren/Lawrence (K/L) scale for scoring radiographic damage16 was used. The ICC was 0.99 (95% CI 0.98–0.99), based on a randomly selected sample of 36 radiographs (18 right knees and 18 left knees). Those knees assigned a K/L grade <2 were rescored in consensus between HMK and an experienced rheumatologist (MK). Clinical data In the geMstoan Study, patient demographic features and disease characteristics were collected from the patients via standard questionnaires. Measurement of pain was achieved using different questionnaires, each of which investigates different dimensions of pain. All questionnaires were filled in with reference to the imaged knee. Three pain questionnaires were used. First, general self-reported pain was assessed on a visual analog scale (VAS) (scale 0–100), a one-dimensional measure of pain intensity. Participants were asked to place an “X” on a 100-mm line to represent the intensity of general pain in the reference knee. A score of 100 represents worst possible pain intensity. Second, the constant pain and intermittent pain subscales of the Intermittent and Constant Osteoarthritis Pain (ICOAP) scoring system (scale 0–100)17 were used. Higher scores indicate worse pain experience. Third, the pain subscale of the Knee Injury and Osteoarthritis Outcome Score (KOOS) (scale 0–100)18 was used. In contrast to all other scales, a score of 0 represents worst possible pain. Patients were asked to fill in both the KOOS and the ICOAP questionnaires to indicate the level of pain intensity experienced in the last 7 days. 114

Association of pain in knee OA with distinct patterns of synovitis

Statistical analysis For comparisons of age between groups, an independent t-test was used. The MannWhitney U test was used for comparisons of body mass index (BMI), K/L grade, MRI total synovitis score, and all pain scales. The chi-square test was used for comparisons of sex distribution and frequency of affected right knees. Principal components analysis (PCA) was used to investigate possible patterns of synovitis. All 11 items in the synovitis scoring system of Guermazi et al13 were subjected to PCA. Prior to performing PCA, the suitability of the data was assessed. For this purpose, the correlation matrix was inspected, which revealed correlation coefficients of ≥0.3. Furthermore, the Kaiser-Meyer-Oklin value of sampling adequacy was 0.738, which exceeded the recommended value of 0.619,20. Finally, the Bartlett’s test of sphericity21 was calculated, yielding significant values (P < 0.001) and thereby supporting the factorability of the correlation matrix. Using eigenvalues of >1, the number of components was determined. A varimax rotation with Kaiser normalization was done to help with the interpretation of the results. In the present study, a site of synovitis was said to load significantly on a component if the factor loading was at least 0.422. To investigate the association between synovitis components and pain measures and radiographic severity in patients with knee OA, regression factor scores that represented location on a factor for each patient were calculated for all components23. Spearman’s rank correlation analysis was used to investigate the unadjusted correlations between regression factor scores and the MRI total score and clinical data. Subsequently, regression factor scores were transformed into binary variables, with 0 representing negative location on a component and 1 representing positive location on a component. A logistic regression adjusted for age, sex, and BMI, and additionally for BMLs and effusion, was used to investigate the association between components and pain and radiographic severity scores. Statistical analyses were performed using SPSS version 20.0.

RESULTS Patient characteristics Of the 101 patients included in the geMstoan Study (mean ± SD age 62 ± 7.5 years, 68% women, median BMI 29 kg/m2 [range 21–49], median K/L grade 3 [range 1–4]), 87 patients had undergone gadolinium-chelate–enhanced MRI. One patient developed anti–cyclic citrullinated peptide–positive, rheumatoid factor–positive oligoarthritis after 1 year, and was therefore excluded, resulting in a total of 86 patients for analysis in the present study. These 86 patients (Table 1) did not differ significantly from the original 101 patients (results not shown) in terms of age, sex, BMI, or K/L grade. Gadolinium-chelate administration was well tolerated by all patients. 115

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 6

Table 1. Characteristics of the study patients (n = 86) Age, mean ± SD years Sex, no. (%) female BMI, median (range) kg/m2 Right knee affected, no. (%) K/L grade, median (range) MRI total synovitis score, median (range) ICOAP score, median (range) Constant pain Intermittent pain KOOS pain score, median (range) VAS knee pain, median (range) mm

62.3 ± 7.4 57 (66) 28.6 (22.0–47.8) 45 (52) 3 (1–4) 7 (0–16) 35 (0–75) 40.6 (0–87.5) 51.4 (2.8–97.2) 53 (0–96)

BMI = body mass index; K/L = Kellgren/Lawrence; MRI = magnetic resonance imaging; ICOAP = Intermittent and Constant Osteoarthritis Pain; KOOS = Knee Injury and Osteoarthritis Outcome Score; VAS = visual analog scale.

The patients with mild to established OA had a median K/L grade of 2 (range 1–4), and patients with end-stage disease requiring arthroplasty had a median K/L grade of 4 (range 1–4); the difference in radiographic damage scores between the groups was significant (P < 0.001). The percentage of female subjects was significantly lower in the end-stage OA group than in the mild to established OA group (P = 0.007). Moreover, as expected, all pain scale scores were significantly higher in the patients with end-stage knee OA. Synovitis on gadolinium-chelate–enhanced MRI The MRI synovitis scores for the affected knees of OA patients are shown in Figure 2. The median total MRI synovitis score was 7 (range 0–16) and was significantly different between the groups. The median score was 6 (range 0–14) in the mild to established OA group compared to a median score of 8 (range 1–16) in the end-stage OA group (P = 0.005). Synovitis was most frequently present at the medial parapatellar site (n = 77), the site adjacent to the posterior cruciate ligament (PCL) (n = 74), the suprapatellar site (n = 65), and the lateral parameniscal site (n = 70). The site adjacent to the PCL was most frequently scored the maximal MRI synovitis score of 2 (n = 31). Loose body synovitis was found in only 6 patients. Overall, the pattern of synovitis in the mild to established OA group resembled the pattern in the end-stage OA group, although synovitis was more extensive and severe in end-stage disease (Figure 2). BMLs were seen in 76 patients (88%), whereas effusion was seen in 80 patients (93%).

116

Association of pain in knee OA with distinct patterns of synovitis

6 Figure 2. Total magnetic resonance imaging scores of synovitis at 11 sites of the whole knee joint in patients with mild to established knee osteoarthritis (OA) and those with end-stage knee OA requiring an arthroplasty. Each row represents 1 patient. Columns represent the 11 different sites. The score range was 0â&#x20AC;&#x201C;2: 0 = white, 1 = light gray, 2 = dark gray. ACL = anterior cruciate ligament; PCL = posterior cruciate ligament.

Synovial patterns based on PCA analysis Analysis of the MRI synovitis findings by PCA resulted in extraction of 3 components, which together explained 53.4% of the variance. After rotation, the loading factors for each anatomic site of synovitis for all 3 components were calculated (Table 2). Component 1 was characterized by the presence of synovitis at 7 sites, with mainly medial parapatellar involvement. Component 2 was characterized by the presence of synovitis at the site adjacent to the anterior cruciate ligament (ACL) and at the medial parameniscal, intercondylar, and suprapatellar sites. Component 3 was characterized by the presence of synovitis at 3 sites (mainly at the loose body site).

117

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 6

Table 2. Principal components analysis with varimax rotation of a 3-component solution for 11 different knee sites on gadolinium-chelate–enhanced magnetic resonance imaging in 86 patients with symptomatic knee osteoarthritis 1 Anatomic site of synovitis Medial parapatellar Lateral parapatellar Adjacent to posterior cruciate ligament Lateral parameniscal Baker’s cyst Infrapatellar Adjacent to anterior cruciate ligament Medial parameniscal Intercondylar (at surface of Hoffa’s fat pad) Loose body Suprapatellar % variance explained

Component 2

0.808 0.693 0.538 0.534 0.530 0.488

0.506

0.854 0.622 0.533 0.421 24.4

0.404 15.6

0.840 0.598 13.4

*Only factor loadings >0.4 are displayed.

Correlation of synovitis sites with clinical outcomes To investigate the association between the synovitis components and the pain scores and severity of radiographic damage, regression factor scores were calculated for all 3 components. Statistically significant correlations were seen between the regression factor score for component 1 and the KOOS pain subscale score (r = −0.24; P = 0.03) and the ICOAP score for constant pain (r = 0.22; P = 0.05), but not with the other pain measures. The regression factor scores for the other 2 synovitis components and the total MRI score did not correlate with any of the pain measures. Moreover, the total MRI synovitis score did not correlate with any of the pain measures. Statistically significant correlations were observed between radiographic severity and the regression factor score for component 1 (r = 0.267; P = 0.013), the regression factor score for component 3 (r = 0.312; P = 0.004), and the total MRI score (r = 0.411; P < 0.001). No significant correlations between BMI and the regression factors were found. Subsequently, the association between synovitis components and the pain and radiographic severity measures was investigated in a logistic regression model, with regression factor scores transformed into binary variables. Logistic regression adjusted for age, sex, and BMI revealed a statistically significant association between component 1 and the KOOS pain subscale score. A trend toward a statistically significant association was observed between component 1 and the ICOAP constant pain score (P = 0.07). Further adjustment for BMLs and effusion did not change the association between component 1 and the KOOS pain subscale score (odds ratio [OR] 0.8, 95% CI 0.6–0.998) and the ICOAP constant pain score (OR 1.3, 95% CI 1.001–1.6). Components 2 and 3 were not associated with any of the pain measures.

118

Association of pain in knee OA with distinct patterns of synovitis

Furthermore, radiographic severity, as measured by the K/L grade, was significantly associated with component 3, and component 1 showed a trend toward a significant association with radiographic severity (P = 0.05) (Table 3). Sensitivity analyses restricted to only knees with a K/L grade of at least 2 decreased the association of component 1 with radiographic severity (OR 2.5, 95% CI 0.8–7.5; adjusted for BMLs and effusion, OR 1.8, 95% CI 0.6–5.8), but increased the association between component 3 and radiographic severity (OR 3.8, 95% CI 1.1–12.4; adjusted for BMLs and effusion, OR 3.4, 95% CI 1.0–11.5). Table 3. Association of the synovitis components with pain and radiographic severity* VAS paina KOOS paina ICOAP intermittent paina ICOAP constant paina K/L grades 3 and 4

Component 1 1.2 (1.0–1.4) 0.8 (0.6–0.954) 1.2 (1.0–1.5) 1.2 (1.0–1.5) 2.7 (1.0–7.4)

Component 2 1.1 (0.9–1.3) 1.0 (0.8–1.2) 1.0 (0.8–1.2) 0.9 (0.8–1.2) 0.9 (0.3–2.2)

Component 3 1.0 (0.9–1.3) 0.9 (0.7–1.1) 1.0 (0.8–1.2) 1.0 (0.8–1.2) 3.2 (1.1–9.0)

* For each component (as displayed in Table 2), regression factor scores were calculated and transformed into binary variables, with 0 representing negative location and 1 representing positive location. Values are the odds ratio (95% confidence interval) adjusted for age, sex, and body mass index. VAS = visual analog scale; KOOS = Knee Injury and Osteoarthritis Outcome Score; ICOAP = Intermittent and Constant Osteoarthritis Pain (score); K/L = Kellgren/Lawrence.aOdds ratio shown per 10 units on pain scale.

6 DISCUSSION In the present study, we found distinct patterns of synovitis as assessed on gadoliniumchelate–enhanced MRI in patients with knee OA. Furthermore, these different patterns of synovitis seemed to be of clinical relevance, since only one pattern was associated with pain. Moreover, we demonstrated that the patterns at different stages of the disease are roughly the same, although synovitis in end-stage disease is more severe. To our knowledge, this is the first study that has investigated patterns of synovitis on gadolinium-chelate–enhanced MRI with the use of PCA. Although synovitis in OA is known to be patchy and heterogeneous, it is not known whether the distribution of synovitis is similar in all patients or whether it differs between patients. Synovitis could differ between patients as a result of a localized response to triggers, such as a microtrauma, cartilage breakdown products, or mechanical loading. Therefore, we chose to study the 11 sites of synovitis by PCA without the inclusion of patient characteristics and without assumptions based on anatomic location. Investigation by PCA resulted in extraction of 3 components. Component 1 explained the largest portion of the total variance and consisted of 7 sites of synovitis with mainly medial parapatellar involvement. In the present study, the results of logistic regression revealed

119

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 6

that only component 1 was associated with pain. Several mechanisms could underlie this observation. Four of the 7 sites of component 1 were in the vicinity of the patella. Earlier studies on the innervation and pain sensation of the knee showed that the patellar region is richly innervated and that anterior synovial tissue in the vicinity of the patella is very sensitive to pain stimulation, which could explain the association between component 1 and pain24,25. The association between synovitis in the parapatellar subregion and pain is in accordance with earlier observations of an association between this region and the pain score on the Western Ontario and McMaster Universities OA index13,26. Although component 1 explained the largest portion of the variance and included sites that displayed the most severe synovitis (medial parapatellar site, suprapatellar site, lateral meniscal site, and site at the PCL), we do not believe that the association found could be a reflection of synovitis in the whole knee with pain. The total MRI score was not correlated with pain as assessed with the different questionnaires, thereby implicating a different mechanism underlying the association of component 1 with pain. These data suggest that each synovitis location has its location-specific properties, and further investigation is needed to unravel the underlying mechanisms. In our study, component 2 was characterized by synovial inflammation at the ACL, medial parameniscal site, suprapatellar site, and intercondylar site. We did not find an association of this component with pain. The infrapatellar fat pad is located at the intercondylar site. Previous studies that investigated an association between synovitis at the infrapatellar fat pad on nonâ&#x20AC;&#x201C;gadolinium-chelateâ&#x20AC;&#x201C;enhanced MRI and pain did not find a significant association with pain, which is in accordance with the present finding26,27. We found that synovitis was most frequently present at 4 sites throughout the whole knee: the site adjacent to the PCL, the medial parapatellar site, the suprapatellar site, and the lateral parameniscal site. Of these, the site adjacent to the PCL, the medial parapatellar site, and the suprapatellar site were, in a study by Roemer et al, also reported to be among the top 4 most frequent sites affected with synovitis in an American population of patients with knee OA, supporting the robustness of the present findings12. Remarkably, the most frequent sites of synovitis were not localized at one specific subregion of the knee, but were found throughout the whole knee, thus emphasizing that use of a system for scoring the whole knee, instead of predefined subregions, is the best way to study synovitis in patients with knee OA. This also stresses the importance of an inflammatory component of the knee joint as a whole and the purely mechanical, one-compartment origin of OA. Loose bodies were seen in the knee joints of 7 patients. Although not frequently observed, they usually display synovitis when present, which is in accordance with other studies12,13. In the present study, investigation of the raw data revealed different distributions of synovitis in these 7 patients, and therefore we chose to include the loose body site in our PCA. This

120

Association of pain in knee OA with distinct patterns of synovitis

resulted in a separate component, underscoring the fact that patients who have synovial inflammation at the loose body site are indeed different from other OA patients. These results, however, should be interpreted with caution and should be replicated in larger cohorts. In the present study, we aimed to investigate different stages of OA in symptomatic patients. Therefore, we chose to divide our patients based on a hard clinical end point: those who did and those who did not receive an arthroplasty. Remarkably, a radiographic K/L grade of 4 was observed in the mild to established OA group, whereas a K/L grade of 1 was observed in the arthroplasty group. These findings underscore the discrepancy between radiographic features of OA and clinical symptoms and emphasize the fact that radiographs do not accurately define the clinical severity of OA. Sensitivity analysis in knees with a K/L grade â&#x2030;Ľ2 revealed no effect on the associations with component 1; however, the association between radiographic severity and component 3 became stronger. The present study also has some limitations. First, only 86 patients were included in the present study, which might explain why some of the pain measures failed to reach significance in their association with component 1. Nevertheless, because the effect sizes and direction (e.g., more severe pain) were quite similar in the different pain scales, we do think our conclusions are valid. The effect sizes, however, are small, reflecting the multidimensionality of pain, which is also influenced by genetic predisposition and psychosocial factors28-30. Only the KOOS pain subscale score and the ICOAP constant pain score showed a statistically significant association with component 1, which confirms the notion that different pain scales measure different pain dimensions. Furthermore, the time from gadolinium-chelate injection to acquisition of the T1-weighted images was 8â&#x20AC;&#x201C;10 minutes, due to the fact that our protocol included several dynamic sequences (that are not part of the current analysis). Although some controversy exists concerning the optimal time for imaging after contrast injection 12,31,32, we believe that this could potentially have led to washout of the contrast into the cavity and might have led to increased measurements of synovial thickness. However, because the aim of the present study was to compare findings between patients, and in all patients the time after injection was comparable, the actual measurements of the synovial membrane were of lesser importance, and therefore this is less of a problem. Finally, we adjusted our logistic regression model for age, sex, and BMI and additionally for BMLs and effusion. Although BMI was significantly correlated with all of the pain measures, it was not correlated with the components of synovitis, and therefore BMI is, by definition, not a confounder. When the models were adjusted for age and sex, both the KOOS pain subscale score and the ICOAP constant pain score were significantly associated with component 1 (results not shown), and adjustments for BMI, BMLs, and effusion did not change the effect sizes.

121

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 6

In conclusion, the present study confirms that the distribution of synovitis on gadoliniumchelateâ&#x20AC;&#x201C;enhanced MRI is patchy in patients with knee OA. This distribution is not random, because distinct patterns of synovitis on gadolinium-chelateâ&#x20AC;&#x201C;enhanced MRI can be identified. Furthermore, these patterns are of clinical relevance, since different patterns were associated differentially with pain. The next research step is to understand what the underlying mechanisms are for these patterns of synovitis in patients with knee OA. Author contributions All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be published. Dr. Kloppenburg had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study conception and design. de Lange-Brokaar, Yusuf, van Osch, Stojanovic-Susulic, Bloem, Kloppenburg. Acquisition of data. de Lange-Brokaar, Yusuf, Visser, Kroon, Bloem, Nelissen, Kloppenburg. Analysis and interpretation of data. de Lange-Brokaar, Ioan-Facsinay, Zuurmond, Huizinga, Kloppenburg. Additional disclosures Author Stojanovic-Susulic is an employee of Janssen Research & Development LLC, a subsidiary of Johnson & Johnson. Acknowledgments The authors would like to acknowledge the support of the cooperating hospital (Diaconessenhuis, Leiden, The Netherlands), including nurse practitioner C. E. Jonxis and orthopedic surgeons Drs. R. Krips, J. B. Mullers, and H. M. Schuller. We also thank the referring rheumatologists, orthopedic surgeons, and nurse practitioner.

122

Association of pain in knee OA with distinct patterns of synovitis

REFERENCES

1 Felson DT. Clinical practice. Osteoarthritis of the knee. N Engl J Med 2006;354:841-8. 2 Yusuf E, Kortekaas MC, Watt I, Huizinga TW, Kloppenburg M. Do knee abnormalities visualised on MRI explain knee pain in knee osteoarthritis? A systematic review. Ann Rheum Dis 2011;70:60-7. 3 Baker K, Grainger A, Niu J, Clancy M, Guermazi A, Crema M et al. Relation of synovitis to knee pain using contrast-enhanced MRIs. Ann Rheum Dis 2010;69:1779-83. 4 de Lange-Brokaar BJ, Ioan-Facsinay A, Van Osch GJ, Zuurmond AM, Schoones J, Toes RE et al. Synovial inflammation, immune cells and their cytokines in osteoarthritis: a review. Osteoarthritis Cartilage 2012;20:1484-99. 5 Sellam J and Berenbaum F. The role of synovitis in pathophysiology and clinical symptoms of osteoarthritis. Nat Rev Rheumatol 2010;6:62535. 6 Loeuille D, Chary-Valckenaere I, Champigneulle J, Rat AC, Toussaint F, Pinzano-Watrin A et al. Macroscopic and microscopic features of synovial membrane inflammation in the osteoarthritic knee: correlating magnetic resonance imaging findings with disease severity. Arthritis Rheum 2005;52:3492-501. 7 Loeuille D, Rat AC, Goebel JC, Champigneulle J, Blum A, Netter P et al. Magnetic resonance imaging in osteoarthritis: which method best reflects synovial membrane inflammation? Correlations with clinical, macroscopic and microscopic features. Osteoarthritis Cartilage 2009;17:1186-92. 8 Loeuille D, Sauliere N, Champigneulle J, Rat AC, Blum A, Chary-Valckenaere I. Comparing non-enhanced and enhanced sequences in the assessment of effusion and synovitis in knee OA: associations with clinical, macroscopic and microscopic features. Osteoarthritis Cartilage 2011;19:1433-9. 9 de Lange-Brokaar BJ, Ioan-Facsinay A, Yusuf E, Visser AW, Kroon HM, Andersen SN et al. Degree of synovitis on MRI by comprehensive whole knee semi-quantitative scoring method correlates with histologic and macroscopic features of synovial tissue inflammation in knee osteoarthritis. Osteoarthritis Cartilage 2014;22:1606-13. 10 Roemer FW, Felson DT, Yang T, Niu J, Crema MD, Englund M et al. The association between

meniscal damage of the posterior horns and localized posterior synovitis detected on T1weighted contrast-enhanced MRI--the MOST study. Semin Arthritis Rheum 2013;42:573-81. 11 Guermazi A, Hayashi D, Roemer FW, Zhu Y, Niu J, Crema MD et al. Synovitis in Knee Osteoarthritis Assessed by Contrast-enhanced Magnetic Resonance Imaging (MRI) is Associated with Radiographic Tibiofemoral Osteoarthritis and MRI-detected Widespread Cartilage Damage: The MOST Study. J Rheumatol 2014;41:501-8. 12 Roemer FW, Kassim JM, Guermazi A, Thomas M, Kiran A, Keen R et al. Anatomical distribution of synovitis in knee osteoarthritis and its association with joint effusion assessed on non-enhanced and contrast-enhanced MRI. Osteoarthritis Cartilage 2010;18:1269-74. 13 Guermazi A, Roemer FW, Hayashi D, Crema MD, Niu J, Zhang Y et al. Assessment of synovitis with contrast-enhanced MRI using a wholejoint semiquantitative scoring system in people with, or at high risk of, knee osteoarthritis: the MOST study. Ann Rheum Dis 2011;70:805-11. 14 Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K et al. Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arthritis Rheum 1986;29:1039-49. 15 Kornaat PR, Ceulemans RY, Kroon HM, Riyazi N, Kloppenburg M, Carter WO et al. MRI assessment of knee osteoarthritis: Knee Osteoarthritis Scoring System (KOSS)--interobserver and intra-observer reproducibility of a compartment-based scoring system. Skeletal Radiol 2005;34:95-102. 16 Kellgren JH and Lawrence JS. Radiological assessment of osteo-arthrosis. Ann Rheum Dis 1957;16:494-502. 17 Hawker GA, Davis AM, French MR, Cibere J, Jordan JM, March L et al. Development and preliminary psychometric testing of a new OA pain measure--an OARSI/OMERACT initiative. Osteoarthritis Cartilage 2008;16:409-14. 18 Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD. Knee Injury and Osteoarthritis Outcome Score (KOOS)--development of a self-administered outcome measure. J Orthop Sports Phys Ther 1998;28:88-96.

123

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 6

19 Kaiser H. A second generation little jiffy. Psychometrika 1970;35:401-15. 20 Kaiser H. An index of factorial simplicity. Psychometrika 1974;39:31-6. 21 Bartlett M. A note on the multiplying factors for various chi-square approximations. J.R. Stat Soc Series B Stat Methodol 1954; 16:296-8. 22 Stevens J. Applied multivariate statistics for the social sciences. Hillsdale (NJ): L. Erlbaum Associates; 1986. 23 DiStefano C, Zhu M, Mindrila D. Understanding and using factor scores: Considerations for the applied researcher. Practical Assessment, Research & Evaluation 2009;14:1-11. 24 Dye SF, Vaupel GL, Dye CC. Conscious neurosensory mapping of the internal structures of the human knee without intraarticular anesthesia. Am J Sports Med 1998;26:773-7. 25 Buma P. Innervation of the patella. An immunohistochemical study in mice. Acta Orthop Scand 1994;65:80-6. 26 Crema MD, Felson DT, Roemer FW, Niu J, Marra MD, Zhang Y et al. Peripatellar synovitis: comparison between non-contrast-enhanced and contrast-enhanced MRI and association with pain. The MOST study. Osteoarthritis Cartilage 2013;21:413-8. 27 Hill CL, Hunter DJ, Niu J, Clancy M, Guermazi A, Genant H et al. Synovitis detected on magnetic resonance imaging and its relation to pain and cartilage loss in knee osteoarthritis. Ann Rheum Dis 2007;66:1599-603. 28 Mogil JS. The genetic mediation of individual differences in sensitivity to pain and its inhibition. Proc Natl Acad Sci U S A 1999;96:7744-51. 29 Colloca L and Benedetti F. How prior experience shapes placebo analgesia. Pain 2006;124:12633. 30 Wager TD. Expectations and anxiety as mediators of placebo effects in pain. Pain 2005;115:225-6. 31 Rhodes LA, Grainger AJ, Keenan AM, Thomas C, Emery P, Conaghan PG. The validation of simple scoring methods for evaluating compartment-specific synovitis detected by MRI in knee osteoarthritis. Rheumatology (Oxford) 2005;44:1569-73. 32 Ostergaard M and Klarlund M. Importance of timing of post-contrast MRI in rheumatoid arthritis: what happens during the first 60 minutes after IV gadolinium-DTPA? Ann Rheum Dis 2001;60:1050-4.

124

CHAPTER 7 RADIOGRAPHIC PROGRESSION OF KNEE OSTEOARTHRITIS IS ASSOCIATED WITH MRI ABNORMALITIES IN BOTH PATELLOFEMORAL AND TIBIOFEMORAL JOINT

de Lange-Brokaar BJE, Bijsterbosch J, Kornaat PR, Yusuf E, Ioan-Facsinay A, Zuurmond A-M, Kroon HM, Meulenbelt I, Bloem JL, Kloppenburg M

submitted

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 7

ABSTRACT Objective: To investigate patterns of MRI abnormalities in the patellofemoral (PFJ) and tibiofemoral joint (TFJ) and their association with radiographic progression, using hypothesis free analyses. Methods: 205 patients from the GARP study with symptomatic OA at multiple sites (mean age 60 years, 80% woman, median BMI 26 kg/m2), underwent knee MRI at baseline. Cartilage damage, osteophytes, cysts, bone marrow lesions (BMLs) and effusion/synovitis were scored according to a validated scoring method. Baseline and 6-year TFJ and PFJ radiographs were scored (0-3) for JSN and osteophytes according to OARSI and Burnett atlases, respectively; progression was defined as â&#x2030;Ľ1 point increase. Baseline patterns of MRI abnormalities derived from principal component analysis (PCA) were associated with progression using adjusted generalized estimating equations. Results: PCA resulted in extraction of 6 components, explaining 69% of variance. In 29% and 29% of 133 patients with follow-up the TFJ progressed, whereas in 15% and 9% the PFJ progressed for osteophytes and JSN, respectively. Component 1 (cartilage damage of the PFJ and osteophytes of both joints) was statistically significant associated with TFJ JSN progression and PFJ osteophyte progression. Component 2 (all lateral PFJ abnormalities except osteophytes) was associated with JSN/osteophyte progression in the PFJ alone, whereas component 3 (all medial TFJ abnormalities except osteophytes) was associated with JSN and osteophyte progression in both PFJ and TFJ. Conclusion: Baseline structural damage and bone turnover activity, as reflected by BMLs, seem to be involved in knee OA progression. Moreover, progression in PFJ and TFJ seems to be related.

126

Radiographic progression of knee OA is associated with MRI abnormalities

INTRODUCTION Knee OA is the most frequent subtype of OA. It can result in end-stage knee OA with cartilage damage and abnormalities in the subchondral bone with pain and functional loss1,2. In some patients rapid progression can be observed, while others progress very slowly3. Unfortunately, due to insufficient knowledge concerning pathophysiological mechanisms of OA, we cannot identify patients at risk for rapid progression. This lack of knowledge also limits the development of structure modifying drugs for OA. Degenerative and repair mechanisms in cartilage, subchondral bone and synovium are involved in OA development and progression. Processes, such as increased bone turnover as assessed by bone marrow lesions (BMLs) and cysts, and inflammation as assessed by synovitis and effusion, have been shown to associate with radiographic progression represented by loss of joint space4-6. Radiographic progression was especially related to ipsilateral MRI abnormalities, suggesting effects of local processes7. But also the presence of radiographic features, such as osteophytes and joint space narrowing (JSN), are associated with radiographic progression8. Since all these structural abnormalities are known to be highly correlated investigation of the role of the individual abnormalities is difficult. Knee OA comprises two joints: the patellofemoral joint (PFJ) and the tibiofemoral joint (TFJ). Although most studies investigate the TFJ, the PFJ might also be of clinical importance as the exclusive presence of PFJ OA is known to lead to functional limitation, pain and stiffness9,10. Several studies have observed that interaction of the TFJ and PFJ could be linked to OA development in those joints and some have suggested that PFJ OA precedes TFJ OA11F. Earlier studies have shown that MRI abnormalities in knee OA have a tendency to follow a medialized or lateralized pattern in both PFJ and TFJ 12, supporting that similar underlying mechanical processes play a role and that these ipsilateral compartments are part of one single joint surface. On the other hand the two joints have different weight bearing properties due to their anatomical location, which could lead to different OA phenotypes. Currently, the relationship between the TFJ and PFJ remains largely unclear. The aim of this study is to investigate patterns of different OA tissue abnormalities as assessed with MRI of both the PFJ and TFJ by using principal component analysis allowing for an objective analysis without assumptions concerning underlying relationships. Subsequently, the association between these patterns with radiological progression over 6 years has been investigated.

127

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 7

PATIENTS AND METHODS Study design This study is part of the Genetics, Osteoarthritis, and Progression (GARP) study, an observational cohort aimed at identifying determinants of OA susceptibility and progression13. This study has been approved by the ethics committee of the Leiden University Medical Center (LUMC). All patients provided written informed consent. Patients Probands, between 40 and 70 years of age, and their siblings, both with symptomatic OA at multiple sites, were included between 2000 and 200314.Patients with secondary OA were excluded. Sib pairs with at least one subject with symptomatic hip or knee OA (but not in a radiographic end-stage) were eligible for the MRI sub-study15,16. Patients were followed for 6 years (For details see17).Demographics and disease characteristic were collected via standard questionnaires. 105 sibpairs with at least one subject with symptomatic hip or knee OA (but not in a radiographic end-stage) were eligible for the MRI sub-study15,16. In 5 out of 210 patients no MRI (claustrophobia (N=1), not fitting into the knee-coil (N=1)) or an MRI of insufficient quality (motionartefacts (n=3)) was available, leaving 205 patients with an MRI for present study. As purpose of the study was to assess progression of OA, no MR images were made of a knee that already had an end-stage (KL score 4)18. In 133/205(65%) patients TFJ radiographic follow-up was available and in 130/205(63%) patients for the PFJ (of 3 patients no lateral knee radiographs were available) Radiographic scoring and definition of progression Semi-flexed posterior-anterior and lateral knee radiographs were obtained at baseline and follow-up. Radiographic severity at baseline was assessed by a single radiologist (HMK) according to the KL atlas13. Progression was scored (0-3) for both osteophytes and JSN at both TFJ and PFJ according to the Osteoarthritis Research Society International (OARSI) atlas 19 and Burnett atlas 20, respectively. Radiographs were scored paired in chronological order blinded for patient characteristics by consensus opinion of two experienced readers (JB, EY). Intra-observer reproducibility was good, as reflected by the intra-class correlation coefficient (ICC, (95% CI)) (One-way random models, single measures). The ICC calculated for the sum scores based on a randomly selected sample of 48 radiographs (both baseline and 6 year followup) of 24 patients ranged from 0.93 (0.89-0.96) to 1.00 (1.00-1.00). The mean ICCs for TFJ and PFJ were 0.97 and 0.99, respectively. Radiographic progression was defined as an increase of 1 point in JSN which was calculated to be the smallest detectable change (SDC). The SDC indicated changes on summarized score above the measurement error21. 128

Radiographic progression of knee OA is associated with MRI abnormalities

MRI acquisition and scoring MR scanning was performed at baseline using a 1.5 T superconducting magnet (Philips Medical Systems, Best, The Netherlands) using a 8-channel dedicated knee coil as described elsewhere18,22. The following images were obtained: coronal proton density- and T2-weighted dual spin echo (SE) images without fat-suppression (with repetition time (TR) of 2,200; echo time (TE) of 20/80; 5 mm slice thickness; 0.5 mm intersection gap;16 cm field of view; 206 x 256 acquisition matrix); sagittal proton density- and T2-weighted dual SE images without fat-suppression(TR 2,200; TE 20/80; 4 mm slice thickness; 0.4 mm intersection gap;16 cm field of view; 205 x 256 acquisition matrix); sagittal three-dimensional (3D) T1-weighted spoiled gradient echo (GE) frequency selective fat-suppressed images (TR 46; TE 2,5; flip angle 40°; 3.0 mm slice thickness; slice overlap 1.5 mm; no gap; 18 cm field of view; 205 x 256 acquisition matrix); and axial proton density- and T2-weighted turbo spin echo (TSE) fat-suppressed images (TR 2,500; TE 7.1/40; echo train length 6,2 mm slice thickness; no gap;18 cm field of view; 205 x 256 acquisition matrix). Total acquisition time (including the initial survey sequence) was 30 min. Cartilage damage (thinning and focal lesions), osteophytes (central and marginal), cysts, bone marrow lesions (BMLs) and effusion (reflecting a combined feature that incorporates both effusion and synovitis) were scored according the Knee Osteoarthritis Scoring System (KOSS) score for presence or absence in 9 compartments, including 5 compartments of the patellofemoral joint (PFJ) and 4 compartments of the tibiofemoral joint (TFJ)20,22. Cartilage defects (diffuse and focal) were scored from 0-3 for depth extent ( 0= absent, 1 < 50% reduction, 2 ≥ 50% reduction, 3 = (near) full-thickness cartilage loss), osteophytes were scored from 0-3 ( 0= no osteophyte, 1= minimal osteophytes < 3 mm from base to tip, 2= moderate osteophyte 3-5 mm, 3= severe osteophyte ≥ 5 mm), subchondral cysts ( 0= absent, 1 minimal < 3 mm greatest dimension measures, 2 moderate 3-5 mm, 3= severe ≥ 5 mm) and BML were defined as an ill-defined area in the subchondral bone extending from the articular surface and also graded from 0-3 (0 = absent, 1= minimal < 5 mm, 2= moderate 5-20 mm, 3= severe ≥ 20 mm). Presence of knee effusion was scored from 0-3 (0= physiological shiver of synovial fluid, 1= small amount of fluid distended one or two joint recesses, 2= >two joint recesses partially distended, 3= full distension of all joint recesses; evaluated were lateral, medial and suprapatellar joint recesses). MRIs were scored in consensus by three readers (PK, RC, JLB) with 3, 15 and 25 years of experience in scoring MRIs (blinded to radiographic results and patient data)16. In an earlier study a good to very good inter- and intra-reader reproducibility was observed using Kappa, weighted kappa and intraclass correlation coefficient (ICC). Intra-observer reliability was assessed by two readers (PK,JLB) using at least a 2 week interval between the randomized readings23. To investigate a single joint surface, compartments were combined to create 4

129

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 7

locations: medial PFJ (containing patellar medial facet and trochlea medial facet), lateral PFJ (containing patellar lateral facet and trochlea lateral facet), medial TFJ (medial femoral condyle and tibia plateau) and lateral TFJ (lateral femoral condyle and tibia plateau). We aimed to investigate local longitudinal associations, as was done before by Felson et al24. We therefore combined different compartments of the KOSS to create the following compartments: medial PFJ (patellar medial facet and trochlea medial facet), lateral PFJ (patellar lateral facet and trochlea lateral facet), medial TFJ (medial femoral condyle and tibia plateau) and lateral TFJ (lateral femoral condyle and tibia plateau). As the patellar crest is in continuum with both the medial and the lateral patellar facet and articulates with both medial and lateral trochlear articular facet, it cannot be medialized/lateralized and we therefore excluded the patellar crest from the analysis15. To create a uniform score the maximum defect score was used to create a score for all variables from 0-3 (0=absent;1=minimal;2=moderate;3=severe). For cartilagedefects a maximum depth of either diffuse or focal cartilagedefect was used. For example if an osteophyte of medial femoral condyle had a score of 2 and the osteophyte at the tibia plateau was scored a 3, the score for medial TFJ osteophyte was a 3. Statistics Normal distributed variables are displayed as mean (SD), otherwise median (range) is given. To investigate whether patients with complete follow-up differed from the total study population the following tests were used: Students-t test (age), Mann-Witney-U test (BMI), Chi-squared test (gender) and Kruskal-Wallis test (KL grade). Principal component analysis (PCA) was used to investigate patterns of MRI features (osteophytes, cartilage defects, subchondral cysts, bone marrow lesions and effusion) in the different joint sites. PCA is a statistical method that determines groups or patterns (named components) based on correlation of features with each other. This method allows for an objective analysis taking in account all variables, without inclusion of assumptions related to possible mechanisms or anatomical sites, and interrelationships can be investigated. For PCA analysis ordinal variables were used and all medial and lateral MRI features of both the PFJ and TFJ (0-3) were subjected to PCA. Prior to performing PCA, suitability of data was assessed. For this purpose the correlation matrix was inspected and revealed the presence of correlation coefficients of 0.3 and above. Furthermore, the Kaiser-Meyer-Oklin value was 0.728, which exceeded the recommended value of 0.625,26. Finally, the Barlettâ&#x20AC;&#x2122;s Test of Sphericity 27 was calculated, which turned out to be significant (p < 0.001), supporting the factorability of the correlation matrix. Using Eigen values >1, the number of components was determined. A Varimax rotation with Kaiser Normalization was done to help with the interpretation of the results. In the present study all variables were said to load significantly on a component if the factor loading was at least 0.428. 130

Radiographic progression of knee OA is associated with MRI abnormalities

To investigate the association between components and radiologic progression at follow-up, regression factor scores that represented location on a factor for each patient, were calculated for all components29. Subsequently, the association of patterns of MRI abnormalities with radiographic progression (outcome, 0 = no progression, 1 = progression) was calculated using logistic regression analysis with generalized estimating equations (GEE) to correct for possible family effects, since observations were obtained from probands and their siblings. The results were presented as odds ratio (OR) with 95% confidence intervals (CI). The GEE was performed with adjustment for age, gender and BMI. Statistics were calculated using SPSS 20.0 (Amonk, NY: IBM Corp.).

RESULTS Patients In the present study 205 patients (mean age (SD) 60 (7) years, 79.5% woman, median BMI (range) 26 (20-40), median (range) KL score 1 (0-3) were included. Patients with KL 0, 1, 2 and 3 were included in the present study. In this study 68% (139/205) of the patients had a Kellgren-Lawrence (KL) score ≥1 in their imaged knee at baseline (32% had a KL score of 0). More than half (55%) of patients had a JSN score ≥ 1 and 47% osteophyte score ≥ 1 in PFJ or TFJ. The prevalence of MRI abnormalities at baseline is displayed in Table 1. Cartilage damage and osteophytes were the most frequent MRI abnormalities, both in the PFJ and TFJ. Principal component analysis of MRI features PCA of all 205 patients resulted in extraction of six components (Eigen value > 1), explaining 69% of the variance (table 1). Component 1 was characterized by medial and lateral cartilage damage and osteophytes of the PFJ and medial and lateral osteophytes of the TFJ; hence included both abnormalities of the PFJ and of the TFJ. The other five components included either abnormalities of the PFJ or of the TFJ. In all compartments, except for the lateral TFJ, cartilage damage was incorporated in the same component as BMLs and cysts. Interestingly effusion/synovitis was not incorporated in any of the components.

131

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 7

Table 1. Loading table of MRI abnormalities of the knee in 205 patients with osteoarthritis at multiple sites; only loadings â&#x2030;Ľ 0.4 are displayed. MRI feature PFJ medial cartilage damage PFJ medial OP PFJ medial cyst PFJ medial BML PFJ lateral cartilage damage PFJ lateral OP PFJ lateral cyst PFJ lateral BML TFJ medial cartilage damage TFJ medial OP TFJ medial cyst TFJ medial BML TFJ lateral cartilage damage TFJ lateral OP TFJ lateral cyst TFJ lateral BML Effusion joint Explained variance 68.6%

Component Prevalence of MRI 1 2 feature (%) 61 48 15 19 44 46 16 17 59 81 16 13 42 61 4 5 55

0.723 0.755

0.414 0.869 0.838

0.577 0.761

0.500 0.902 0.909 0.546

0.407 0.401

0.585 0.738 0.792

0.830 0.536

0.454 0.827 0.787 16.9

11.9

11.2

10.0

9.3

9.1

Abbreviations: PFJ=patellofemoral joint, TFJ= tibiofemoral joint, OP=osteophytes, BML=bone marrow lesion. For description of variables see Patients and Methods.

Correlation of patterns of MRI abnormalities with radiographic progression The demographic characteristics and KL scores of the patients available for follow-up (n=133) did not differ statistically significant from the total study population (data not shown). Progression of JSN and osteophytes in the TFJ were seen in 38/133 (28.6%) and 39/133 (29.3%) of patients, respectively. Progression of JSN and osteophytes in the PFJ were less frequently being observed in 12/130 (9.2%) and 20/130 (15.4%) of patients, respectively. Radiographic JSN or osteophyte progression in the PFJ and TFJ were not related; the Spearman rank correlation was 0.095 (p-value 0.296) for JSN and 0.136 (p-value 0.133) for osteophytes. The associations of the components with radiographic progression in JSN are shown in Table 2. JSN progression in the PFJ was associated with component 2 (incorporating cartilage damage, BML and cysts of the lateral PFJ); the OR (95%CI) of this association was 2.0 (1.23.2). JSN progression was also associated with component 3 (incorporating cartilage damage, BML and cysts of the medial TF); OR (95%CI) 2.9 (1.5-5.7)). All associations were adjusted for age, sex and BMI. No independent associations were seen with component 1, 4, 5 and 6. JSN progression in the TFJ was associated with component 1 (incorporating cartilage damage

132

Radiographic progression of knee OA is associated with MRI abnormalities

of the PFJ and osteophytes of both PFJ and TFJ), with an OR (95%CI) of 1.6 (1.1-2.5), and with component 3 (incorporating cartilage damage, BML and cysts of the medial TF with an OR (95%CI) of 1.4 (1.002-2.0), all after adjustment for age, sex and BMI. No association were seen with components 2, 4, 5 and 6. Associations of the components with osteophyte progression are depicted in Table 3. Osteophyte progression in both the PFJ and TFJ was associated with component 3 (incorporating cartilage damage, BMLs and cysts of the medial TFJ). The OR (95%CI) for PFJ was 1.9 (1.3-2.9), whereas the OR (95%CI) for TFJ was 1.8 (1.2-2.6). In addition, associations were observed between osteophyte progression in the PFJ and component 1 (incorporating cartilage damage of the PFJ and osteophytes of both PFJ and TFJ) with an OR (95%CI) of 1.7 (1.039-2.9), and component 2 (incorporating cartilage damage, BMLs and cysts of the lateral PFJ) with an OR (95%CI) of 1.6 (1.031-2.4). Table 2. Association of components with radiographic joint space narrowing progression over 6 years, analysed for all patients. Joint space narrowing PFJ (n=130)

Joint space narrowing TFJ (n =133)

Component

Crude OR (95%CI)

Adjusted* OR (95%CI)

Crude OR (95%CI)

Adjusted* OR (95%CI)

Cart/OP med/lat PFJ,OP med/lat TFJ

1.8 (1.01-3.2)

1.6 (0.8-3.2)

1.8 (1.2-2.7)

1.6 (1.1-2.5)

2 3 4 5 6

Cart, cyst, BML lat PFJ Cart, cyst, BML med TFJ Cart, cyst, BML med PFJ Cyst, BML lat TFJ Cart/OP med/lat TFJ

1.5 (1.03-2.3) 2.1 (1.3-3.1) 0.9 (0.5-1.7) 1.1 (0.6-1.9) 1.4 (0.8-2.3)

2.0 (1.2-3.2) 2.9 (1.5-5.7) 1.0 (0.5-2.2) 1.0 (0.5-1.7) 1.6 (0.8-2.9)

0.8 (0.6-1.2) 1.3 (0.9-1.8) 0.9 (0.6-1.3) 1.0 (0.6-1.5) 1.2 (0.8-1.7)

0.8 (0.5-1.3) 1.4 (1.002-2.0) 0.9 (0.6-1.3) 1.0 (0.6-1.6) 1.2 (0.9-1.7)

*Adjusted OR: GEE model adjusted for age, gender, BMI and family effect. Outcome is progression in joint space narrowing (0 = no, 1 = yes). Abbreviations: Cart =cartilage damage, OP = osteophytes, PFJ = patellofemoral joint, med =medial, lat = lateral, TFJ =tibiofemoral joint, cyst = subchondral cyst, BML = bone marrow lesion, OR =odds ratio

133

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 7

Table 3. Association of components with radiographic osteophyte progression over 6 years, analysed for all patients. Component

OP PFJ (n =130) Crude OR (95%CI)

Adjusted* OR (95%CI)

OP TFJ (n = 133) Crude OR (95%CI)

Adjusted* OR (95%CI)

Cart/OP med/lat PFJ, OP med/lat TFJ

1.8 (1.1-3.0)

1.7 (1.04-2.9)

1.4 (0.9-2.0)

1.4 (0.9-2.1)

2 3 4 5 6

Cart, cyst, BML lat PFJ Cart, cyst, BML med TFJ Cart, cyst, BML med PFJ Cyst, BML lat TFJ Cart/OP med/lat TFJ

1.4 (0.9-1.9) 1.8 (1.2-2.6) 1.1 (0.7-1.6) 1.1 (0.6-1.7) 1.4 (0.9-2.1)

1.6 (1.03-2.4) 1.9 (1.3-2.9) 1.2 (0.7-1.9) 1.0 (0.7-1.5) 1.6 (0.9-2.6)

0.8 (0.5-1.2) 1.7 (1.2-2.4) 0.8 (0.5-1.2) 1.4 (0.9-2.1) 1.4 (0.97-2.0)

0.8 (0.5-1.2) 1.8 (1.2-2.6) 0.8 (0.5-1.3) 1.4 (0.98-1.9) 1.5 (0.97-2.2)

* Adjusted OR: GEE model adjusted for age, gender, BMI and family effect. Outcome is progression in joint space narrowing (0 = no, 1 = yes). Abbreviations: Cart =cartilage damage, OP = osteophytes, PFJ = patellofemoral joint, med =medial, lat = lateral, TFJ =tibiofemoral joint, cyst = subchondral cyst, BML = bone marrow lesion, OR =odds ratio

Associations of components at baseline with both medial and lateral progression of radiographic osteophytes and JSN of the TFJ are depicted in Table 4. Component 1 (incorporating cartilage damage of the PFJ and osteophytes of both PFJ and TFJ) was associated with JSN at the medial TFJ with an OR (95%CI) of 1.9 (1.2-2.9). Component 3 (including cartilage, cysts and BML of the medial TFJ) was associated with medial JSN progression with an OR (95%CI) of 1.6 (1.1-2.3), with medial osteophyte progression with an OR (95%CI) of 2.2 (1.4-3.4) and interestingly also with lateral osteophyte progression with an OR (95%CI) of 1.7 (1.2-2.5). Furthermore, a statistical significant association between lateral osteophyte progression with both component 5 (including lateral MRI features of the TFJ) with an OR (95%CI) of 1.4 (1.01-2.0) and component 6 (including both cartilage and osteophyte features of the TFJ) with an OR (95%CI) of 1.6 (1.1-2.5) was observed.

134

Radiographic progression of knee OA is associated with MRI abnormalities

Table 4. Association of components with radiographic joint space narrowing and osteophyte progression over 6 years, analysed for all patients.

Component All patients, N = 133 1 Cart/OP med/lat PFJ, OP med/lat TFJ 2 Cart, cyst, BML lat PFJ 3 Cart, cyst, BML med TFJ 4 Cart, cyst, BML med PFJ 5 Cyst, BML lat TFJ 6 Cart/OP med/lat TFJ

JSN progression JSN TFJ med JSN TFJ lat Adjusted* OR Adjusted* OR (95%CI) (95%CI)

OP progression OP TFJ med Adjusted* OR (95%CI)

OP TFJ lat Adjusted* OR (95%CI)

1.9 (1.2-2.9)

0.9 (0.4-2.0)

1.4 (0.9-2.3)

1.1 (0.7-1.8)

0.6 (0.3-1.2) 1.6 (1.1-2.3) 0.9 (0.6-1.4) 0.6 (0.3-1.5) 1.2 (0.9-1.8)

1.0 (0.7-1.7) 2.1 (0.01-448,3) 1.1 (0.7-1.7) 1.4 (0.9-2.4) 1.2 (0.7-2.0)

0.6 (0.2-1.6) 2.2 (1.4-3.4) 0.7 (0.4-1.2) 0.9 (0.6-1.5) 0.9 (0.6-1.6)

0.8 (0.5-1.2) 1.7 (1.2-2.5) 0.9 (0.5-1.4) 1.4 (1.01-2.0) 1.6 (1.1-2.5)

* Adjusted OR: GEE model adjusted for age, gender, BMI and family effect. Outcome is progression in osteophyte or joint space narrowing in the tibiofemoral joint (0 = no, 1 = yes). Abbreviations: Cart =cartilage damage, OP = osteophytes, PFJ = patellofemoral joint, med =medial, lat = lateral, TFJ =tibiofemoral joint, cyst = subchondral cyst, BML = bone marrow lesion, OR =odds ratio

DISCUSSION To our best knowledge this is the first study that investigates MRI abnormalities and their association with radiographic knee OA progression without inclusion of assumptions related to possible mechanisms or anatomical sites. Clustering in MRI abnormalities of the PFJ and TFJ were observed at baseline, since the component that explained most of the variance incorporated both abnormalities seen in the PFJ and the TFJ. Interestingly, in some components cartilage damage clustered with osteophytes, whereas in other components clustering of cartilage damage with BMLs and cysts was seen. Longitudinal analyses showed that radiological progression over 6 years in the PFJ was associated with components including MRI abnormalities of both the TFJ and the PFJ. Likewise, radiological progression in the TFJ was associated with components including MRI abnormalities of both the PFJ and TFJ. These results suggest that underlying processes in PFJ and TFJ as visualized on MRI at baseline are related with respect to radiological progression. Furthermore, radiographic progression was associated with components including cartilage damage and osteophytes, or with components including cartilage damage and BMLs and cysts in the medial TFJ or lateral PFJ. These observations suggest that not only osteoarthritic structural damage enhance further progression, but that also processes reflecting increased bone turnover result in progression.

135

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 7

The component incorporating medial and lateral cartilage damage and osteophytes of the PFJ and medial and lateral osteophytes of the TFJ (no.1) was significantly associated with JSN progression of the TFJ and osteophyte progression in the PFJ. This component explains the largest proportion of variance (26.3% unrotated and 17% after rotation) and probably reflects the largest proportion of people in our population. This observation suggests that patients with evident OA involvement of both compartments of the knee are at risk for ongoing OA progression, as a vicious circle, which involves the whole joint. An alternative explanation could be the presence of effusion. Since we set our cut-off value for significant loading to 0.4 28, effusion was not incorporated in any of the components, although effusion was loaded on both component 1 (loading 0.326) and 3 (loading 0.350). Component 1 and 3 both were associated with radiographic progression. Several studies in the past have found that effusion significantly correlates with cartilage damage over time, which are in accordance with our results4. The role of osteophytes in OA is much debated as some argue that osteophytes are related to normal aging rather than to the presence of OA30. This argument is supported by studies that have found that osteophytes do not always relate to cartilage damage31,32. However, in the present study using PCA osteophytes, which were highly prevalent, clustered with cartilage damage, supporting that osteophytes are part of the OA process, for instance as compensatory mechanism to stabilize the joint33. Although osteophytes did not cluster with BMLs and cysts cross-sectionally, component 3 (including cartilage damage, cysts and BMLs in the medial TFJ) is associated with osteophyte progression, which further supports that osteophytes are a result of the OA process. In the present study osteophytes did cluster with local cartilage damage, but also with osteophytes at other locations, suggesting also a non-location specific mechanism for formation osteophytes. The association between component 3 and osteophyte progression in the lateral TFJ suggest additional systemic mechanism in the formation of osteophytes. These findings encourage further studies looking into inflammatory and metabolic factors in the formation of osteophytes and development and progression of OA34. The present study, using PCA to investigate association with MRI abnormalities in an unbiased way, reveals that BMLs and cysts were associated with cartilage damage in almost all compartments. Furthermore, two of the three components that were associated with radiographic progression in longitudinal analyses incorporated BMLs and cysts. These observations are in line with earlier studies that showed that BMLs are associated with cartilage damage cross-sectionally and predict cartilage loss over tďťżime4,6,35,35,36,36. The nature of BMLs is not fully elucidated, but seems to represent local regions of trabecular remodelling or compression37,38. Both osteoblasts as well as increased expression of insulinlike growth factor 1 (IGF-1) and transforming growth factor - Î˛ (TGF-Î˛) by subchondral bone

136

Radiographic progression of knee OA is associated with MRI abnormalities

are thought to play a role in the abnormal bone remodelling in OA39 40,41. In line with potential local processes we report an associated between BMLs and cysts medially with JSN and OP progression medially in the TFJ, and BMLs and cysts laterally with OP progression laterally in the TFJ. In addition we also found BMLs and cysts in the medial TFJ to be associated with OP progression in the lateral TFJ, suggestion a reflection of a systemic effect. The latter effect encourages further study. The most extensive component for radiographic progression in the knee joint appeared to be component 3. BMLs and cysts together with cartilage damage located at the medial TFJ were associated positively with progression of JSN and osteophytes in both joints. This could be explained by varus malalignment, which has a significant influence on the biomechanical properties of both the TFJ and the PFJ. Varus malalignment has been shown to be a risk factor for the development of BMLs in the medial TFJ35,42,43 and leads to progression of OA in the medial compartment43,44. Furthermore, in patients with a varus malalignment the q-angle of the patella increases resulting in an increasing medial patellar force and increased load on the medial compartment of the patella and subsequently in PFJ progression45-47. Unfortunately, only lateral view, not skyline view, radiographs were available in the present study and therefore we could not confirm that medial TFJ abnormalities lead to medial PFJ JSN progression. The majority of scientific literature suggests that PFJ OA is more common in the lateral patella compared to the medial patella in OA patients, due to a naturally occurring lateral reaction force vector46,47. Our study does not support this notion, as at baseline medial PFJ MRI abnormalities were more common. This is in concordance with a study by Gross et al. that found a higher prevalence of medial PFJ OA compared to lateral PFJ OA48. However, our longitudinal data suggest that lateral PFJ processes are of higher impact than medial PFJ processes since radiological progression in PFJ is especially associated with lateral PFJ involvement. The present study has some limitations. First, the study population is relatively small and PFJ progression was seen only in a limited number of patients. Therefore, findings should be interpreted with caution and should be replicated in larger samples. Second, our results suggest that MRI abnormalities in the TFJ can influence progression of the PFJ. Since knees with a KL score of 4 were excluded the role of severe structural damage of the TFJ on progression in the PFJ could not be studied. Third, our study did not include synovitis in PCA. Since synovitis has been implicated in radiographic progression, inclusion of synovitis in the PCA would have been preferable. Yet, we did include effusion in our analysis, which can be considered as surrogate of synovial inflammation. Fourth, only lateral view radiographs

137

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 7

were available for scoring PFJ progression, not skyline view. Therefore we were unable to investigate medial and lateral PFJ progression, which could have led to an underestimation of PFJ progression. Finally, in present study we only included depth of cartilage defect (either diffuse or focal) in our analysis, while the KOSS scoring system scores the surface of the defect. Therefore small full-thickness cartilage defects were treated the same as large full-thickness cartilage defects and could have biased our results. Furthermore, the patellar crest was excluded from our analysis as it could not be medialized or lateralized. This could have biased our results. In conclusion our findings, showing that both osteophytes as well as BMLs and subchondral cysts are clustered with cartilage defects and are associated with radiographic progression, in a localized way, but also in the whole knee joint, suggest that the OA process and progression is a complicated interplay between different underlying pathogenetic processes, where likely biomechanics, but also systemic factors play a role. Therefore these factors should be investigated as a whole. Author contributions Authors made substantial contributions to the following: (1a) conception and design of the study: BDL, JLB, MK; (1b) acquisition of data: JB, PK, EY, MK; (1c) interpretation of data BDL, JB, PK, EY, AIF, AMZ, HK, IM, JLB, MK; (2) drafting or critical revision of manuscript: BDL, JB, PK, EY, AIF, AMZ, HK, IM, JLB, MK; (3) final approval of manuscript BDL, JB, PK, EY, AIF, AMZ, HK, IM, JLB, MK Role of funding source Financial support was obtained from the Dutch Arthritis Association and Pfizer Groton Inc.. However they did not contribute to design, interpretation of data, drafting and final approval of the manuscript. Conflict of interests: none

138

Radiographic progression of knee OA is associated with MRI abnormalities

REFERENCES

1 Felson DT. Clinical practice. Osteoarthritis of the knee. N Engl J Med 2006;354:841-8. 2 Vos T, Flaxman AD, Naghavi M, Lozano R, Michaud C, Ezzati M et al. Years lived with disability (YLDs) for 1160 sequelae of 289 diseases and injuries 1990-2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet 2012;380:2163-96. 3 Bijlsma JW, Berenbaum F, Lafeber FP. Osteoarthritis: an update with relevance for clinical practice. Lancet 2011;377:2115-26. 4 Roemer FW, Guermazi A, Felson DT, Niu J, Nevitt MC, Crema MD et al. Presence of MRI-detected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30-month follow-up: the MOST study. Ann Rheum Dis 2011;70:1804-9. 5 Crema MD, Felson DT, Roemer FW, Wang K, Marra MD, Nevitt MC et al. Prevalent cartilage damage and cartilage loss over time are associated with incident bone marrow lesions in the tibiofemoral compartments: the MOST study. Osteoarthritis Cartilage 2013;21:306-13. 6 Roemer FW, Guermazi A, Javaid MK, Lynch JA, Niu J, Zhang Y et al. Change in MRIdetected subchondral bone marrow lesions is associated with cartilage loss: the MOST Study. A longitudinal multicentre study of knee osteoarthritis. Ann Rheum Dis 2009;68:1461-5. 7 Felson DT, McLaughlin S, Goggins J, LaValley MP, Gale ME, Totterman S et al. Bone marrow edema and its relation to progression of knee osteoarthritis. Ann Intern Med 2003;139:3306. 8 Chapple CM, Nicholson H, Baxter GD, Abbott JH. Patient characteristics that predict progression of knee osteoarthritis: a systematic review of prognostic studies 1. Arthritis Care Res (Hoboken) 2011;63:1115-25. 9 Duncan R, Peat G, Thomas E, Wood L, Hay E, Croft P. Does isolated patellofemoral osteoarthritis matter? Osteoarthritis Cartilage 2009;17:1151-5. 10 Peat G, Duncan RC, Wood LR, Thomas E, Muller S. Clinical features of symptomatic patellofemoral joint osteoarthritis. Arthritis Res Ther 2012;14:R63. 11 Duncan R, Peat G, Thomas E, Hay EM, Croft P. Incidence, progression and sequence of development of radiographic knee osteoarthritis in a symptomatic population. Ann Rheum Dis 2011;70:1944-8.

12 Kornaat PR, Watt I, Riyazi N, Kloppenburg M, Bloem JL. The relationship between the MRI features of mild osteoarthritis in the patellofemoral and tibiofemoral compartments of the knee. Eur Radiol 2005;15:1538-43. 13 Riyazi N, Meulenbelt I, Kroon HM, Ronday KH, Hellio le Graverand MP, Rosendaal FR et al. Evidence for familial aggregation of hand, hip, and spine but not knee osteoarthritis in siblings with multiple joint involvement: the GARP study. Ann Rheum Dis 2005;64:438-43. 14 Riyazi N, Meulenbelt I, Kroon HM, Ronday KH, Hellio Le Graverand MP, Rosendaal FR et al. Evidence for familial aggregation of hand, hip, and spine but not knee osteoarthritis in siblings with multiple joint involvement: the GARP study. Ann Rheum Dis 2005;64:438-43. 15 Kornaat PR, Watt I, Riyazi N, Kloppenburg M, Bloem JL. The relationship between the MRI features of mild osteoarthritis in the patellofemoral and tibiofemoral compartments of the knee. Eur Radiol 2005;15:1538-43. 16 Kornaat PR, Bloem JL, Ceulemans RY, Riyazi N, Rosendaal FR, Nelissen RG et al. Osteoarthritis of the knee: association between clinical features and MR imaging findings. Radiology 2006;239:811-7. 17 Bijsterbosch J, Meulenbelt I, Watt I, Rosendaal FR, Huizinga TW, Kloppenburg M. Clustering of hand osteoarthritis progression and its relationship to progression of osteoarthritis at the knee. Ann Rheum Dis 2014;73:567-72. 18 Kornaat PR, Bloem JL, Ceulemans RY, Riyazi N, Rosendaal FR, Nelissen RG et al. Osteoarthritis of the knee: association between clinical features and MR imaging findings. Radiology 2006;239:811-7. 19 Altman RD, Hochberg M, Murphy WA, Jr., Wolfe F, Lequesne M. Atlas of individual radiographic features in osteoarthritis. Osteoarthritis Cartilage 1995;3 Suppl A:3-70. 20 Burnett S Hart DJ CCST. A radiographic atlas of osteoarthritis. 1994; 21 Bruynesteyn K, Boers M, Kostense P, van der Linden S, van der Heijde D. Deciding on progression of joint damage in paired films of individual patients: smallest detectable difference or change 4. Ann Rheum Dis 2005;64:179-82. 22 Kornaat PR, Ceulemans RY, Kroon HM, Riyazi N, Kloppenburg M, Carter WO et al. MRI assessment of knee osteoarthritis: Knee

139

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 7

140

Osteoarthritis Scoring System (KOSS)--interobserver and intra-observer reproducibility of a compartment-based scoring system. Skeletal Radiol 2005;34:95-102. Kornaat PR, Ceulemans RY, Kroon HM, Riyazi N, Kloppenburg M, Carter WO et al. MRI assessment of knee osteoarthritis: Knee Osteoarthritis Scoring System (KOSS)--interobserver and intra-observer reproducibility of a compartment-based scoring system. Skeletal Radiol 2005;34:95-102. Felson DT, McLaughlin S, Goggins J, Lavalley MP, Gale ME, Totterman S et al. Bone marrow edema and its relation to progression of knee osteoarthritis. Ann Intern Med 2003;139:3306. Kaiser HF. Second Generation Little Jiffy. Psychometrika 1970;35:401-&. Kaiser HF. Index of Factorial Simplicity. Psychometrika 1974;39:31-6. Bartlett MS. A Note on the Multiplying Factors for Various Chi-2 Approximations. Journal of the Royal Statistical Society Series B-Statistical Methodology 1954;16:296-8. Stevens J. Applied multivariate statistics for the social sciences. 1986; DiStefano C, Zhu M, Mindrila D. Understanding and Using Factor Scores:Considerations for the Applied Researcher. Practical Assessment, Research & Evaluation 2009;14:1-11. Hernborg J and Nilsson BE. The relationship between osteophytes in the knee joint, osteoarthritis and aging. Acta Orthop Scand 1973;44:69-74. Brandt KD, Fife RS, Braunstein EM, Katz B. Radiographic grading of the severity of knee osteoarthritis: relation of the Kellgren and Lawrence grade to a grade based on joint space narrowing, and correlation with arthroscopic evidence of articular cartilage degeneration. Arthritis Rheum 1991;34:1381-6. Boegard T, Rudling O, Petersson IF, Jonsson K. Correlation between radiographically diagnosed osteophytes and magnetic resonance detected cartilage defects in the tibiofemoral joint. Ann Rheum Dis 1998;57:401-7. Pottenger LA, Phillips FM, Draganich LF. The effect of marginal osteophytes on reduction of varus-valgus instability in osteoarthritic knees. Arthritis Rheum 1990;33:853-8. Thijssen E, van CA, van der Kraan PM. Obesity and osteoarthritis, more than just wear and tear: pivotal roles for inflamed adipose tissue and dyslipidaemia in obesity-induced

osteoarthritis. Rheumatology (Oxford) 2015;54:588-600. Segal NA, Kern AM, Anderson DD, Niu J, Lynch J, Guermazi A et al. Elevated tibiofemoral articular contact stress predicts risk for bone marrow lesions and cartilage damage at 30 months. Osteoarthritis Cartilage 2012;20:11206. Hunter DJ, Zhang Y, Niu J, Goggins J, Amin S, LaValley MP et al. Increase in bone marrow lesions associated with cartilage loss: a longitudinal magnetic resonance imaging study of knee osteoarthritis. Arthritis Rheum 2006;54:1529-35. Driban JB, Tassinari A, Lo GH, Price LL, Schneider E, Lynch JA et al. Bone marrow lesions are associated with altered trabecular morphometry. Osteoarthritis Cartilage 2012;20:1519-26. Hunter DJ, Gerstenfeld L, Bishop G, Davis AD, Mason ZD, Einhorn TA et al. Bone marrow lesions from osteoarthritis knees are characterized by sclerotic bone that is less well mineralized. Arthritis Res Ther 2009;11:R11. Hilal G, Martel-Pelletier J, Pelletier JP, Ranger P, Lajeunesse D. Osteoblast-like cells from human subchondral osteoarthritic bone demonstrate an altered phenotype in vitro: possible role in subchondral bone sclerosis. Arthritis Rheum 1998;41:891-9. Henrotin Y, Pesesse L, Sanchez C. Subchondral bone and osteoarthritis: biological and cellular aspects. Osteoporos Int 2012;23 Suppl 8:S847-S851. Sharma L, Chmiel JS, Almagor O, Felson D, Guermazi A, Roemer F et al. The role of varus and valgus alignment in the initial development of knee cartilage damage by MRI: the MOST study. Ann Rheum Dis 2013;72:235-40. Hayashi D, Englund M, Roemer FW, Niu J, Sharma L, Felson DT et al. Knee malalignment is associated with an increased risk for incident and enlarging bone marrow lesions in the more loaded compartments: the MOST study. Osteoarthritis Cartilage 2012;20:1227-33. Sharma L, Chmiel JS, Almagor O, Felson D, Guermazi A, Roemer F et al. The role of varus and valgus alignment in the initial development of knee cartilage damage by MRI: the MOST study. Ann Rheum Dis 2013;72:235-40. Walker EA, Davis D, Mosher TJ. Rapidly progressive osteoarthritis: biomechanical considerations. Magn Reson Imaging Clin N Am 2011;19:283-94.

Radiographic progression of knee OA is associated with MRI abnormalities

45 Huberti HH and Hayes WC. Patellofemoral contact pressures. The influence of q-angle and tendofemoral contact. J Bone Joint Surg Am 1984;66:715-24. 46 Elahi S, Cahue S, Felson DT, Engelman L, Sharma L. The association between varus-valgus alignment and patellofemoral osteoarthritis. Arthritis Rheum 2000;43:1874-80. 47 Cahue S, Dunlop D, Hayes K, Song J, Torres L, Sharma L. Varus-valgus alignment in the progression of patellofemoral osteoarthritis. Arthritis Rheum 2004;50:2184-90. 48 Gross KD, Niu J, Stefanik JJ, Guermazi A, Roemer FW, Sharma L et al. Breaking the Law of Valgus: the surprising and unexplained prevalence of medial patellofemoral cartilage damage. Ann Rheum Dis 2012;71:1827-32.

141

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

CHAPTER 8 EVOLUTION OF SYNOVITIS IN OSTEOARTHRITIC KNEES AND ITS ASSOCIATION WITH CLINICAL FEATURES

de Lange-Brokaar BJE, Ioan-Facsinay A, Yusuf E, Kroon HM, Zuurmond A-M, Stojanovic-Susulic V, Nelissen RGHH, Bloem JL, Kloppenburg M

submitted

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 8

ABSTRACT Objective: To investigate the course of synovitis on contrast-enhanced MRI In osteoarthritic knees over 2 years, and its association with pain and cartilage deterioration. Methods: Consecutive patients (n=39, mean age 61 years, 79% woman, median (range) BMI 29 (24-48) kg/mm2) with clinical OA were included. Baseline and follow-up contrastenhanced MR images (3T) were scored paired in chronological order for synovitis (semiquantitatively at 11 sites (range 0-22)), cartilage deterioration and bone marrow lesions (BML) (semi-quantitatively according to KOSS). Changes in sum scores were calculated. Cartilage deterioration was defined as change of â&#x2030;Ľ 2 above the smallest detectable change (SDC). Pain was assessed by standardized questionnaires. Logistic and linear regression models were used to investigate association between synovitis change and cartilage deterioration and between synovitis change or cartilage deterioration and change in pain. Results: The total synovitis score did not change over 2 years (mean change 0.2 (SD 3.2), although changes in individual patients were observed. Cartilage deterioration was observed in 51% of patients. Increase in synovitis score was significantly associated with cartilage deterioration, independently of BML change (adjusted OR (95% CI) 1.3 (1.0041.8). Change in synovitis was not associated with change in pain, whereas cartilage deterioration was associated with change in ICOAP constant pain in adjusted models (OR (95%CI) 2.8 (0.4-5.3). Conclusion: In individual patients synovitis fluctuates during disease course. Increase in synovitis is associated with cartilage deterioration, suggesting a role for synovitis as a target for disease-modifying treatment.

144

Evolution of synovitis in osteoarthritic knees

INTRODUCTION The disease course of osteoarthritis (OA) in the knee is known to be variable; some patients are known to progress rapidly while others remain stable over a long time1,2. However, which processes underlie these differences in disease course remain unknown. Although OA is considered a non-inflammatory condition, synovial inflammation is prevalent3 and could play an important role in the pathophysiology of the disease4. However, to investigate synovitis in knee OA patients a valid method to assess synovitis is necessary. Currently, the gold standard for assessing synovial inflammation is based on histological analysis of synovial biopsy samples, a methodology that is not patient friendly and technically difficult. Alternatively synovitis can be assessed by contrast enhanced (CE) MRI. It has been proven to be a practical and reliable alternative in evaluating synovitis in knee OA patients5-12. As synovitis is known to be patchy and heterogeneous11, a synovitis scoring method on MRI should encompass a sufficient number of compartments. The synovitis scoring method developed by Guermazi et al.13 is a comprehensible and practical method, which meets these requirements and compares well with synovial inflammation in tissue biopsies of knee OA patients8. The importance of synovitis in knee OA has been supported by several MRI studies that showed an association of synovitis with cartilage deterioration14-17 and pain12,18. Nevertheless, very few studies investigated changes in synovitis over time19,20 and these indicated an association with pain, but not with cartilage deterioration. However, in these studies no contrast enhancement was used, which precludes a precise determination of synovitis. Therefore, the evolution of synovitis during the disease course and its role in cartilage progression and change of pain in knee OA patients remains unclear. Therefore, in the present study, we aimed to investigate the change of synovitis on contrast Gd-chelate-enhanced MRI over 2 years and its association with cartilage deterioration and change in pain in knee OA patients.

METHODS Study design This study is part of the ongoing geMstoan study (GEneration of Models, Mechanism & Markers for STratification of OsteoArthritis patieNts)8, a cohort study in established and endstage knee OA patients to find new biomarkers for OA progression. This study has been approved by the ethics committee of the Leiden University Medical Center (LUMC). All patients provided written informed consent.

145

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 8

Patients Between 2008 and 2013, patients with symptomatic radiographic primary knee OA, according to American College of Rheumatology (ACR) criteria,21 attending the rheumatology or orthopaedic department of the LUMC or orthopaedic department of the Diaconessenhuis, Leiden, were included. The geMstoan study comprises two groups of patients based on their clinical status; one with end-stage disease who received a total knee arthroplasty and the other group with a mild to established OA, with no indication for an arthroplasty. For the current analysis patients with mild to established OA were investigated. Patients with other rheumatic diseases, using immunosuppressive drugs or having knee injections (i.e. corticosteroids) in the past 3 months were excluded. Patients with chronic renal insufficiency (Cockroft-Gault < 60 mL/min) did not undergo Gd-chelate enhanced MRI. MRI acquisition We used a 3T Philips Achieva MR system (Philips Healthcare, Best the Netherlands) with an 8-channel dedicated knee coil. Coronal and sagittal proton density (PD) fast spin-echo (FSE) driven equilibrium images were obtained with a field of view (FOV) of 150X 150 mm, an acquisition matrix of 304X 240, and slice thickness of 3 mm., repetition time (TR) was 3000 ms; echo time (TE) 34 ms. Axial and coronal frequency selective fat suppressed PD FSE images were obtained with the same geometric parameters, and TR of 2675, TE 24. A T1-weighted axial sequence with TR 581 ms, and TE 20 ms was obtained with slice thickness of 3.3 mm, FOV 160. The sixth sequence was a sagittal three-dimensional (3D) T1-weighted spoiled gradient echo (GE) frequency selective fat-suppressed sequence with TR 16,3; TE 9.2; flip angle 35째; 1.5 mm slice thickness; FOV 150x150; 304x304 acquisition matrix. Finally contrast enhanced (CE)-MR images were obtained following injection of gadoterate meglumine (0.2 ml/kg) (Dotarem; Guerbet) in the cubital vein using a power injector (Medrad) with a rate of 2 ml/second followed by a 40-ml saline flush also at a rate of 2 ml/second. We subsequently obtained frequency selective fat suppressed T1-weighted, FSE with TR of 655 ms, and TE of 20 ms, in both the axial and sagittal planes MRI scoring Cartilage was assessed on PD-FSE and GE images. BML were scored using the fat suppressed PD-FSE and GE images. Synovial tissue was assessed on the Gd-chelate enhanced images8,22,23. All MRIs were scored in paired samples in a chronological order. Synovitis was scored in a semi-quantitative way at 11 different sites according to Guermazi et al.13. Synovial thickness was measured and scored as followed: 0, when synovial thickness was less than 2 mm, 1 when thickness was between 2-4 mm and 2 when synovial thickness was above 4 mm. The total synovitis score of 11 sites was calculated (range 0-22). A total score of 0-4 was considered normal (no synovitis); 5-8 represents a mild, 9-12 a moderate and above 13 a

146

Evolution of synovitis in osteoarthritic knees

severe synovitis13. Intra-class correlation (ICC) was based on a random sample of 10% Gdchelate enhanced MR images and was 0.93 for synovitis and 0.90 for synovitis change. Cartilage damage and bone marrow lesions (BMLs) were scored according the Knee Osteoarthritis Scoring System (KOSS) score in 9 compartments, as described elsewhere22. In short, cartilage damage was defined as a combination of diffuse and focal cartilage defect (0 = absent (no abnormality in signal intensity or morphology), 1 = less than 50% reduction of cartilage thickness, 2= 50% or greater reduction of cartilage thickness, grade 3= full-thickness or near-full-thickness cartilage defect). To investigate cartilage damage throughout the whole knee diffuse defects (0-27) and focal defects (0-27) were summarized, creating a total cartilage damage score (possible range 0-54). Subsequently, change in summarized cartilage damage scores between two time points was defined as cartilage deterioration. ICC for total cartilage score was 0.96 and ICC for cartilage deterioration was 0.73. Cartilage deterioration was defined based on the smallest detectable change (SDC), being measurement error; a change in the summarized score of cartilage defects â&#x2030;Ľ 2 was used to define cartilage deterioration. Cartilage deterioration was used as dichotomous variable. BMLs were defined as an ill-defined area in the subchondral bone extending from the articular surface and were graded from 0-3 (0= absent, 1= minimal < 5 mm, 2 = moderate 5-20mm, 3= severe â&#x2030;Ľ 20mm). BML scores were summarized (range 0-27) to reflect BMLs throughout the knee. Subsequently the change in total BML scores between time points were calculated and defined as change in total BML score. ICC was 0.98 for total BML scores and 0.57 for change in total BML scores. All MRI were analyzed by one experienced reader (BdL). Scoring was done after extensive learning sessions and under supervision of an experienced musculoskeletal radiologist (JB). During the assessment, the reader was blinded to radiographic results and patient data. Scoring knee radiographs: Baseline radiographs (posterior anterior (PA) fixed flexion) were obtained of all patients. Radiographs were scored, blinded for patient characteristics, by an experienced musculoskeletal radiologist (HK), with 30 years of experience in scoring musculoskeletal radiographs, according to the Kellgren- Lawrence (KL) scale24. Reproducibility was good as described elsewhere8. Clinical data In the geMstoan patients demographics and disease characteristic were collected via standard questionnaires. Measurement of pain in the imaged knee was achieved by using three questionnaires that each investigates different dimensions of pain. General assessment of self-reported pain was assessed by the visual analogue scale (VAS, 0-100). The VAS is a one-dimensional measure of pain intensity. A score of 100 represents worst

147

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 8

possible pain intensity. The measure of Intermittent and Constant OsteoArthritis Pain (ICOAP)25 was filled in to assess constant pain and intermittent pain. Higher scores indicate worse pain experience. The Knee injury and Osteoarthritis Outcome Score (KOOS subscale pain, 0-100)26 was used. In contrast to all other scales a score of 0 represents worst possible pain. Patients were asked to fill in both KOOS and ICOAP questionnaire for pain experienced in the last 7 days. For analysis change in pain was used. Statistics Normal distributed variables are depicted as mean (standard deviation), otherwise as median (range). For comparison between time points, paired sample t-test was used for all variables except for cartilage deterioration for which Wilcoxon signed rank test was used and comparison of NSAID use for which Chi-squared test was used. Depending on normal distribution of the data, comparison of the sample of patients used in the follow-up analysis with the original baseline patient population, independent t-test was used for age, Mann-Whitney U tests were used for BMI and KL grade and Chi-squared test was used for gender. To investigate the association of synovitis change with cartilage deterioration both unadjusted as well as adjusted logistic regression models were performed. To investigate the association of synovitis change with pain both unadjusted and adjusted linear regression models were performed. Statistics were calculated by SPSS 20.0 (IBM, Armonk, NY).

RESULTS Patient characteristics Of 62 patients at baseline, one patient developed after one year an Anti-cyclic Citrullinated Peptide (anti-CCP)-antibody positive, rheumatoid factor positive oligoarthritis and was excluded from the study, resulting in 61 patients at baseline (mean (SD) age 61.5 (6.9) years, 79.5% woman, BMI median (range) 28.6 (22.8-47.8) kg/mm2, median (range) KL score 2 (0-4)) included in the geMstoan study. Thirty nine out of these 61 patients had Gdchelate enhanced MR images at both baseline and follow-up that has been used for current analysis (Figure 1). These 39 patients did not significantly differ from the original 61 patients at baseline in age, gender, BMI or KL grade (data not shown). Patient characteristics and MRI features are displayed in Table 1. Gd-chelate administration was well tolerated by all patients.

148

Evolution of synovitis in osteoarthritic knees

8 Figure 1. Flowchart of patients available for current analysis. Abbreviations: RF = rheumatoid factor, OA = osteoarthritis, pt =patients, N = number, MRI = magnetic resonance imaging, CE = contrast-enhancement

149

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 8

Table 1 Patient characteristic 39 patients Age Gender, n (%) BMI* KL grade* NSAID, n (%) ICOAP constant pain (0-100)* ICOAP intermittent pain (0-100) KOOS pain (0-100) VAS pain* (0-100) Total synovitis score on MRI (0-22) Cartilage deterioration on MRI (0-54)* Total BML score on MRI (0-27)*

Baseline 61.1 (6.7) 31/39 (80%) 28.8 (23.3-47-8) 2 (0-4) 17/38 (45%) 25 (0-75) 35.9 (24.9) 61.4 (23.2) 42 (0-96) 5.59 (2.3) 16 (4-37) 4 (0-11)

2 years follow-up 12/38 (32%) 15 (0-75) 28.2 (20.2) 63.2 (21.7) 29 (0-87) 5.74 (2.9) 20 (4-37) 5 (0-12)

Difference - 5 (13%) -8.1 (21.0) -7.6 (20.2) 1.7 (14.0) -6.2 (20.6) 0.15 (3.2) 2 (-1 -11) 0.3 (1.8)

p-value 0.065 0.028 0.030 0.465 0.069 0.764 < 0.001 0.299

Mean (SD) are given except for variables that were not normal distributed (median (range)), indicated with a *, for gender and NSAID use number (%) is given. Abbreviations: n = number, BMI = body mass index, KL = kellgren and Lawrence, ICOAP = measure of Intermittent and Constant OsteoArthritis Pain, KOOS = Knee injury and Osteoarthritis Outcome Score,VAS = visual analogue scale, CE-MRI = contrast enhanced Magnetic resonance images, BML = bone marrow lesions

Course of synovitis, cartilage damage and BMLs over a 2-year period The mean follow-up time was 2 years and 3 months. A mild synovitis in the knees was observed that did not change over time: the mean (SD) total synovitis score at baseline was 5.6 (2.3) and was similar at follow-up (5.6 (2.9)) (Table 1). However, changes were seen on individual levels, illustrated by Figure 2 and Figure 3A. Figure 2 shows synovitis scores per anatomical site at baseline and follow-up. Synovitis (score 1 or 2) was most frequently present at the medial parapatellar site (in 31 and 33 patients at baseline and follow-up visits respectively), adjacent to the posterior cruciate ligament (PCL) (in 30 patients at baseline and in 28 patients at follow-up) and at the suprapatellar site (in 22 and 24 patients at baseline and follow-up respectively). The site adjacent to the PCL was most frequently displaying the maximal score of 2 at both baseline (n= 8) and follow-up (n=9). Loose bodies, and therefore synovitis surrounding loose bodies, were not seen in any patient at both baseline and follow-up. When the 11 sites were investigated separately over time, the sites that most frequently displayed synovitis at baseline were also the sites that most frequently showed changes over time (39 % of patients in both medial and suprapatellar site and 44 % at the site adjacent to PLC). In contrast, patients who displayed no synovitis at a particular site at baseline often did not display synovitis at 2-year follow-up either (69% of patients at the Bakersâ&#x20AC;&#x2122; cyst site and 72% of patients at the intercondylar site did not show synovitis at both baseline and follow-up). Subsequently, we investigated the change in total synovitis scores on individual patient level. Figure 3A shows distribution plots for change in total synovitis score over a 2-year period. This figure clearly shows that the change in total synovitis score varies between patients, which is also reflected by its range (-7 till 9). 150

Evolution of synovitis in osteoarthritic knees

Figure 2. Total magnetic resonance imaging scores of synovitis at 11 sites of the whole knee joint in knee osteoarthritis patients at baseline and follow-up visit. Each row represents 1 patient; the baseline patient row corresponds to the follow-up patient row. Columns represent the 11 different synovitis sites. The score range was 0â&#x20AC;&#x201C;2: 0 =white, 1 = light gray, 2 = dark gray. Abbreviations: ACL = anterior cruciate ligament, PCL = posterior cruciate ligament.

Figure 3. Cumulative probability plots of change in total synovitis score, change in total BML score, cartilage deterioration and for change in total synovitis score stratified for cartilage deterioration progression in all knee osteoarthritis (OA) patients over 2-year period. A. Change in total synovitis score. B. Cartilage deterioration. The dotted line represents the smallest detectable change (SDC). Patients above the dotted line were classified as having cartilage deterioration progression. C. Change in total BML score. D. Change in total synovitis score stratified for cartilage deterioration progression.

151

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 8

Relation of synovitis change and change of BMLs with cartilage deterioration Cartilage deterioration significantly increased over time (Table 1 and Figure 3B). Figure 3B shows that 20 patients (51%) had cartilage deterioration over 2 years (the dotted line represents the SDC of 1.95). Cartilage deterioration was seen in 17 (43.6%) patients in the patellofemoral joint and in 17 (43.6%) patients in the tibiofemoral joint, and was more frequent in the medial compartment compared to the lateral compartment of the tibiofemoral joint (35.9% versus 20.5%). The fluctuating nature of BMLs over time is underscored by the change in total BML scores that ranges from -3 till 5. The large variety in change in BML scores between patients is illustrated by Figure 3C. Figure 3D shows that patients with cartilage deterioration had on average an increase in total synovitis score over time (positive mean (SD) total score 1.3 (3.1)), while patients without cartilage deterioration had on average a decrease in total synovitis score over time (negative mean (SD) -1.1 (2.9)). The difference in change in total synovitis scores between patients without cartilage deterioration and with cartilage deterioration was statistically significant (difference (95% CI) (-2.4 (-4.3 till -0.4)). To further investigate the association between change in total synovitis score and cartilage deterioration we adjusted for other variables in logistic regression analyses. A statistically significant association was observed between change in the total synovitis score and cartilage deterioration, taking into account the change in the total BML score (OR (95% CI) 1.3 (1.004-1.772). The effect size of the association increased when correcting for baseline characteristics, although the statistical significance was lost (Table 2). Table 2: The association between change in total synovitis score and change in total bone marrow lesion (BML) score over 2-year period with cartilage progression in 39 patients with knee osteoarthritis. Synovitis change BML change Synovitis adjusted for BML Synovitis adjusted for age, gender and BMI, BML

OR (95% CI) 1.3 (1.022-1.774) 1.3 (0.865-1.816) 1.3 (1.004-1.772) 1.4 (0.981-1.858)

p-value 0.035 0.232 0.047 0.066

Abbreviations: BML = bone marrow lesions, BMI = body mass index, OR = odds ratio, CI = confidence interval. P- value < 0.05 was considered significant (bold), P-value < 0.10 was considered a trend (bold and italic). Maximum synovitis score at 11 anatomical areas is 22

Relation of synovitis change, cartilage deterioration and change in BMLs with change in pain At follow-up, significantly less pain was reported for both subscales of the ICOAP and a trend for less pain was reported for the VAS pain (Table 1). To investigate the associations of several MRI features with pain linear regression models were used (Table 3). Although in univariate analyses associations were suggested for change in the total synovitis score or change in the total BML score with change in pain over the 2-year period, these

152

Evolution of synovitis in osteoarthritic knees

associations were lost when adjusted models were used. Cartilage deterioration was found to be associated with change in pain and the association with change in ICOAP constant pain remained statistically significant even in adjusted models (B (95%CI) 2.8 (0.4-5.3)). Additional adjustment for baseline variables (age, gender and BMI), showed similar results; only cartilage deterioration was significantly associated with change in ICOAP constant pain over time (B (95%CI) 3.1 (0.5-5.7)). Table 3: The association between change in total synovitis score, change in total bone marrow lesion score and cartilage deterioration over 2-year period with change in pain in knee osteoarthritis patients.

Change in total synovitis score, crude Change in total synovitis score, adjusted* Cartilage deterioration, crude Cartilage deterioration, adjusted * Change in total BML score, crude Change in total BML score, adjusted*

ICOAPc B (95% CI) 1.7 (-0.6-3.9) 0.3 (-2.2-2.8) 3.0 (0.9-5.1) 2.8 (0.4-5.3) 0.8 (-3.2-4.8) 0.1 (-3.8-3.9)

ICOAPi B (95% CI) 2.0 (-0.1-4.1) 1.1 (-1.4-3.6) 2.3 (0.2-4.5) 1.8 (-0.7-4.2) 1.4 (-2.5-5.2) 0.5 (-3.3-4.3)

KOOSp B (95% CI) -1.1 (-2.6-0.4) -0.3 (-2.1-1.4) -1.6 (-3.1- -0.1) -1.4 (-3.2-0.3) -0.9 (-3.6-1.8) -0.4 (-3.2-2.3)

VAS B (95% CI) 2.0 (-0.02-4.1) 0.9 (-1.5-3.2) 2.1(0.005-4.2) 1.5 (-0.9-3.8) 3.7 (0.1-7.3) 3.0 (-0.7-6.6)

* adjusted for other two change parameters. Abbreviations: ICOAP = measure of Intermittent and Constant OsteoArthritis Pain, ICOAPc = subscale constant pain, ICOAPi = subscale intermittent pain, KOOSp = Knee injury and Osteoarthritis Outcome Score subscale pain, VAS = visual analogue scale for pain, BML = bone marrow lesions. Pvalue < 0.05 was considered significant (bold), P-value < 0.01 was considered a trend (bold and italic)

DISCUSSION To our knowledge this is the first study that investigates synovitis change over 2 years using CE-MRI. We found that synovitis was most frequently seen at the medial parapatellar and the suprapatellar sites and at the site adjacent to posterior cruciate ligament (PCL). Over a 2-year period the total amount of synovitis in the osteoarthritic knees was relatively stable on group level, although fluctuations were seen on individual level. An increase in synovitis severity was significantly associated with cartilage deterioration over time. Although in univariate analyses an association of change in total amount of synovitis in the osteoarthritic knee with a change in pain was seen, this association was lost after adjustment. Also a change in the total amount or size of BMLs in the osteoarthritic knee was not associated anymore with change in pain after adjustment. Cartilage deterioration over 2 years was associated with change in pain, also taking in account changes in the total amount of synovitis and BMLs in the osteoarthritic knee. In the present study increase in synovitis in the whole knee was associated with cartilage deterioration and only in unadjusted models with increase in pain. This is in contrast with two studies that suggested a role for synovitis in the increase in knee pain, not in

153

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 8

cartilage progression19,20. This discrepancy could be explained by a number of reasons. First, both previously reported studies used signal changes in Hoffaâ&#x20AC;&#x2122;s fat pad on non-CE MRI as surrogate for whole knee synovitis, while in the present study Gd-chelate enhanced images were used. Past studies have shown that signal changes in Hoffaâ&#x20AC;&#x2122;s fat pad detected on non-CE MRI are not specific for synovitis in the knee. Therefore, findings reported in these studies could be a reflection of other processes in the knee instead of synovitis9,27. Second, both studies used a synovitis sum score that was composed of a limited number of anatomical sites in the knee, while in the present study the total synovitis score according to Guermazi et al.13, encompassing 11 different sites throughout the whole knee, was used. In the present study the total synovitis score was especially influenced by the suprapatellar site (synovitis change frequently observed) and not by the two other sites (the infrapatellar site and intercondylar site), investigated by the studies by Hill et al. and Zhang et al.. In accordance with this explanation Hill et al. did not find an association of synovitis at the suprapatellar site with VAS pain. Finally, both studies adjusted for baseline cartilage scores but not for cartilage deterioration in their investigation with pain. In our study cartilage deterioration was significantly associated with pain and adjusting for cartilage deterioration led to the loss of association of synovitis change with change in pain. Therefore, cartilage deterioration could serve as mediator in the association between synovitis and pain, explaining differences with earlier literature. In our study the mean total synovitis score at group level was almost the same between baseline and follow-up, while overall pain was less frequently reported at the follow-up visit compared to the pain score at baseline. These observations were also described by Kortekaas et al. who investigated synovitis change using ultrasound in patients with hand OA28. Decrease in pain could be explained by the fact that at time of inclusion patients consulted their physician because they had a health problem, which is most frequently due to pain. After 2 years these patients are more likely to have accepted diagnosis, have adapted expectations and have received treatment, resulting in a subjective pain reduction29,30. An alternative explanation is that the decrease in pain is a result of a statistical phenomenon called regression to the mean. Future longitudinal studies investigating pain over longer periods of time are needed to elucidate the underlying mechanism behind this pain decrease. In the present study the total synovitis scores on group level over 2-year period did not change, which seems to be in contrast with our previous study in which we found that patients with end-stage OA (with an indication for a knee arthroplasty) displayed more severe synovitis compared to patients who did not have an indication for a knee arthroplasty8. The discrepancy could be explained by the fact that a 2-year interval is too short for most patients

154

Evolution of synovitis in osteoarthritic knees

to progress to an end-stage disease with a significant increase in synovitis. In accordance in the present study only 4 of 61 patients progressed from baseline to an end-stage disease. Unfortunately, those patients that did progress to an end-stage were excluded because no follow-up MRI could be made. In the present study cartilage deterioration was seen in 43.6% of patients in the tibiofemoral joint (TFJ) and in 43.6% of patients in the patellofemoral joint (PFJ) over 2 years, which is higher than described in a previous study investigating cartilage progression on MRI (27% in the TFJ, 24% PFJ)14. The difference in progression could be explained by the difference in study population: symptomatic patients in secondary care in present study versus a general study population in the other study. Therefore, in the present study patients were more likely to progress. The present study has several limitations. Firstly, only 39 patients form the original 61 baseline patients were available for Gd-enhanced MRIs after 2-year follow-up. Eleven patients were not included because no CE-MRI was performed (e.g. discontinuation of MRI, contra-indications for MRI, claustrophobia), which underscores the difficulty for acquiring longitudinal data on synovitis using CE-MRI. Due to the low number of patients we could only adjust for a limited number of variables. Since our aim was to investigate synovitis change, we chose to adjust for variables that changed over time, not for baseline characteristics; baseline characteristics should not influence the results as in-patient relationships were investigated. To test this notion we ran analysis with additional adjustments for age, gender and BMI, resulting in the improvement of the effect size thus validating the found association. However, these results should be interpreted with caution and should be confirmed in larger studies. Another limitation of our study is that patients that progressed to an end-stage knee OA stage were excluded from our analysis as no follow-up MRI was available. The exclusion of these subjects could have potentially biased the study results considering that these patients often display more severe synovitis, and their inclusion could hypothetically have led to increase of mean synovitis score over time. In conclusion our data suggest that on an individual level increase in synovitis is associated with cartilage deterioration not with pain, suggesting a role for synovitis as a target for disease-modifying treatment. Acknowledgement The authors would like to acknowledge support of the cooperating hospital Diaconessenhuis, Leiden and referring rheumatologists, orthopaedic surgeons and nurse practitioners.

155

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 8

Author contributions Authors made substantial contributions to the following: (1a) conception and design of the study: BDL, AIF, AMZ, VSS, MK, JB; (1b) acquisition of data: BDL, AIF, EY, HK, JB, RN, MK; (1c) interpretation of data BDL, AIF, EY, HK, AMZ, VSS, RN, JB, MK; (2) drafting or critical revision of manuscript: BDL, AIF, EY, HK, AMZ, VSS, RN, JB, MK; (3) final approval of manuscript BDL, AIF, EY, HK, AMZ, VSS, RN, JB, MK. Role of funding source Financial support was obtained from TI Pharma, however TI Pharma did not contribute to design, interpretation of data, drafting and final approval of the manuscript. Conflict of interests: none

156

Evolution of synovitis in osteoarthritic knees

REFERENCES

1 Dieppe PA, Cushnaghan J, Shepstone L. The Bristol ‘OA500’ study: progression of osteoarthritis (OA) over 3 years and the relationship between clinical and radiographic changes at the knee joint. Osteoarthritis Cartilage 1997;5:87-97. 2 Yusuf E, Bijsterbosch J, Slagboom PE, Kroon HM, Rosendaal FR, Huizinga TW et al. Association between several clinical and radiological determinants with long-term clinical progression and good prognosis of lower limb osteoarthritis. PLoS One 2011;6:e25426. 3 de Lange-Brokaar BJ, Ioan-Facsinay A, Van Osch GJ, Zuurmond AM, Schoones J, Toes RE et al. Synovial inflammation, immune cells and their cytokines in osteoarthritis: a review. Osteoarthritis Cartilage 2012;20:1484-99. 4 Sellam J and Berenbaum F. The role of synovitis in pathophysiology and clinical symptoms of osteoarthritis. Nat Rev Rheumatol 2010;6:62535. 5 Loeuille D, Chary-Valckenaere I, Champigneulle J, Rat AC, Toussaint F, Pinzano-Watrin A et al. Macroscopic and microscopic features of synovial membrane inflammation in the osteoarthritic knee: correlating magnetic resonance imaging findings with disease severity. Arthritis Rheum 2005;52:3492-501. 6 Loeuille D, Rat AC, Goebel JC, Champigneulle J, Blum A, Netter P et al. Magnetic resonance imaging in osteoarthritis: which method best reflects synovial membrane inflammation? Correlations with clinical, macroscopic and microscopic features. Osteoarthritis Cartilage 2009;17:1186-92. 7 Loeuille D, Sauliere N, Champigneulle J, Rat AC, Blum A, Chary-Valckenaere I. Comparing non-enhanced and enhanced sequences in the assessment of effusion and synovitis in knee OA: associations with clinical, macroscopic and microscopic features. Osteoarthritis Cartilage 2011;19:1433-9. 8 de Lange-Brokaar BJ, Ioan-Facsinay A, Yusuf E, Visser AW, Kroon HM, Andersen SN et al. Degree of synovitis on MRI by comprehensive whole knee semi-quantitative scoring method correlates with histologic and macroscopic features of synovial tissue inflammation in knee osteoarthritis. Osteoarthritis Cartilage 2014;22:1606-13.

9 Roemer FW, Guermazi A, Zhang Y, Yang M, Hunter DJ, Crema MD et al. Hoffa’s Fat Pad: Evaluation on Unenhanced MR Images as a Measure of Patellofemoral Synovitis in Osteoarthritis. AJR Am J Roentgenol 2009;192:1696-700. 10 Hayashi D, Roemer FW, Guermazi A. Osteoarthritis year 2011 in review: imaging in OA--a radiologists’ perspective. Osteoarthritis Cartilage 2012;20:207-14. 11 Roemer FW, Kassim JM, Guermazi A, Thomas M, Kiran A, Keen R et al. Anatomical distribution of synovitis in knee osteoarthritis and its association with joint effusion assessed on non-enhanced and contrast-enhanced MRI. Osteoarthritis Cartilage 2010;18:1269-74. 12 Baker K, Grainger A, Niu J, Clancy M, Guermazi A, Crema M et al. Relation of synovitis to knee pain using contrast-enhanced MRIs. Ann Rheum Dis 2010;69:1779-83. 13 Guermazi A, Roemer FW, Hayashi D, Crema MD, Niu J, Zhang Y et al. Assessment of synovitis with contrast-enhanced MRI using a wholejoint semiquantitative scoring system in people with, or at high risk of, knee osteoarthritis: the MOST study. Ann Rheum Dis 2011;70:805-11. 14 Roemer FW, Guermazi A, Felson DT, Niu J, Nevitt MC, Crema MD et al. Presence of MRI-detected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30-month follow-up: the MOST study. Ann Rheum Dis 2011;70:1804-9. 15 Roemer FW, Zhang Y, Niu J, Lynch JA, Crema MD, Marra MD et al. Tibiofemoral joint osteoarthritis: risk factors for MR-depicted fast cartilage loss over a 30-month period in the multicenter osteoarthritis study. Radiology 2009;252:772-80. 16 Guermazi A, Hayashi D, Roemer FW, Zhu Y, Niu J, Crema MD et al. Synovitis in Knee Osteoarthritis Assessed by Contrast-enhanced Magnetic Resonance Imaging (MRI) is Associated with Radiographic Tibiofemoral Osteoarthritis and MRI-detected Widespread Cartilage Damage: The MOST Study. J Rheumatol 2014;41:501-8. 17 Krasnokutsky S, Belitskaya-Levy I, Bencardino J, Samuels J, Attur M, Regatte R et al. Quantitative magnetic resonance imaging evidence of synovial proliferation is associated with radiographic severity of knee osteoarthritis. Arthritis Rheum 2011;63:2983-91.

157

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 8

18 Yusuf E, Kortekaas MC, Watt I, Huizinga TW, Kloppenburg M. Do knee abnormalities visualised on MRI explain knee pain in knee osteoarthritis? A systematic review. Ann Rheum Dis 2011;70:60-7. 19 Hill CL, Hunter DJ, Niu J, Clancy M, Guermazi A, Genant H et al. Synovitis detected on magnetic resonance imaging and its relation to pain and cartilage loss in knee osteoarthritis. Ann Rheum Dis 2007;66:1599-603. 20 Zhang Y, Nevitt M, Niu J, Lewis C, Torner J, Guermazi A et al. Fluctuation of knee pain and changes in bone marrow lesions, effusions, and synovitis on magnetic resonance imaging. Arthritis Rheum 2011;63:691-9. 21 Altman R, Asch E, Bloch D, Bole G, Borenstein D, Brandt K et al. Development of criteria for the classification and reporting of osteoarthritis. Classification of osteoarthritis of the knee. Diagnostic and Therapeutic Criteria Committee of the American Rheumatism Association. Arthritis Rheum 1986;29:1039-49. 22 Kornaat PR, Ceulemans RY, Kroon HM, Riyazi N, Kloppenburg M, Carter WO et al. MRI assessment of knee osteoarthritis: Knee Osteoarthritis Scoring System (KOSS)--interobserver and intra-observer reproducibility of a compartment-based scoring system. Skeletal Radiol 2005;34:95-102. 23 de Lange-Brokaar BJ, Ioan-Facsinay A, Yusuf E, Visser AW, Kroon HM, vanOsch GJ et al. Pain in knee osteoarthritis patients associates with distinct patterns of synovitis. Arthritis Rheumatol 2015;67:733-740. 24 The Atlas of Standard Radiographs of Arthritis. Rheumatology 2005;44:iv43-iv72. 25 Hawker GA, Davis AM, French MR, Cibere J, Jordan JM, March L et al. Development and preliminary psychometric testing of a new OA pain measure--an OARSI/OMERACT initiative. Osteoarthritis Cartilage 2008;16:409-14. 26 Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD. Knee Injury and Osteoarthritis Outcome Score (KOOS)--development of a self-administered outcome measure. J Orthop Sports Phys Ther 1998;28:88-96. 27 Crema MD, Felson DT, Roemer FW, Niu J, Marra MD, Zhang Y et al. Peripatellar synovitis: comparison between non-contrast-enhanced and contrast-enhanced MRI and association with pain. The MOST study. Osteoarthritis Cartilage 2013;21:413-8.

158

28 Kortekaas MC, Kwok WY, Reijnierse M, Huizinga TW, Kloppenburg M. Follow-up study of inflammatory ultrasound features in hand osteoarthritis over a period of 3 months: variable as well as constant. Osteoarthritis Cartilage 2014;22:40-3. 29 Colloca L and Benedetti F. How prior experience shapes placebo analgesia. Pain 2006;124:12633. 30 Wager TD. Expectations and anxiety as mediators of placebo effects in pain. Pain 2005;115:225-6. 31 Rhodes LA, Keenan AM, Grainger AJ, Emery P, McGonagle D, Conaghan PG. The relationship between limited MRI section analyses and volumetric assessment of synovitis in knee osteoarthritis. Clin Radiol 2005;60:1295-9. 32 Ostergaard M and Klarlund M. Importance of timing of post-contrast MRI in rheumatoid arthritis: what happens during the first 60 minutes after IV gadolinium-DTPA? Ann Rheum Dis 2001;60:1050-4.

CHAPTER 9 SUMMARY, GENERAL DISCUSSION & FUTURE PERSPECTIVES

Chapter 9

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 160

Summary, general discussion & future perspectives

SUMMARY Osteoarthritis (OA) is a heterogeneous disorder, which can affect all joints in the body, but is especially prevalent in the knee joint1. Eventually, the disease progresses and leads to joint destruction with symptoms of pain and functional impairments. Currently, no disease modifying drugs are available and therapy is focused on symptom relief. OA is becoming a significant medical and financial burden in a world whose population is aging. Although knee OA is a very prevalent disease, underlying pathological and pain mechanisms remain unknown. Therefore, research investigating underlying mechanisms is of utmost importance. For a long time, OA was considered a non-inflammatory condition. More recently, however, it became evident that synovial inflammation could play an important role in the pathophysiology of OA 2-7 as it is a predictor of cartilage destruction 8,9 and a determinant of pain10,11. The role of synovitis in knee OA, however, is still largely unknown. Therefore, in this thesis we investigated the nature of synovial inflammation in knee OA and its possible contribution to the clinical manifestations of knee OA. Results presented in this thesis provide insight into different aspects of synovial inflammation aimed at increasing our understanding of the pathophysiology of OA and aiding to the development of disease modifying drugs in OA. Part I The nature of synovial inflammation in knee OA As first step in the analysis of the cellular and molecular nature of synovial inflammation, we performed in chapter 2 a narrative systematic review to summarize the current knowledge of inflammatory properties, immune cells and their cytokines in synovial tissues of OA patients. Our hypothesis that knowledge was readily available was underscored by the fact that we extracted 100 articles. In this literature overview, we found that synovial inflammation histologically similar to the one observed in rheumatoid arthritis (RA) was commonly described in OA and that the most frequently detected cell types were macrophages, T cells and mast cells, while B cells were almost never found. Cytokines related to T cell or macrophage function were also described in OA synovial tissues, although their cellular source was only scarcely investigated. Overall, the conclusion was that inflammation is in general lower in OA than in RA and this reflects also in the abundance of most infiltrating immune cells. Strikingly, however, the number of mast cells was as high as, or sometimes even higher than in RA synovial tissue. Based on the results of this extensive review a research agenda was made, which served as a foundation for the follow-up studies described in this thesis. - Analysis of immune cells and their activation state in synovial tissues of knee OA patients in relation to clinical disease characteristics.

161

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 9

- -

In-depth analysis of mast cells and their activation state in synovial tissues in knee OA patients at various disease stages and their relationship to clinical disease characteristics. Investigation of synovitis throughout the knee OA disease course. Association of synovitis with clinical parameters as pain, structural damage and progression.

FACS analysis, as described in chapter 3, of synovial tissues and of infrapatellar fat pad (IFP), an articular tissue in which immune cells are abundantly present, of patients with knee OA revealed that macrophages and T cells, followed by mast cells, were the most prominent immune cells, and that subpopulations of both T cells and macrophages were in an activated state. Furthermore, we found that CD4+ T cells were associated with pain, offering a cellular basis for the long-known association between synovitis and pain in knee OA. As we learned from chapter 2, mast cells could be important in OA. Therefore, in chapter 4, we showed that synovial tissue of OA patients had a significantly higher number of mast cells than synovial tissue of RA patients, although RA synovial tissues displayed more severe inflammation grade by H&E staining compared to OA synovial tissue. The clinical relevance of mast cells in OA synovial tissues was suggested by the association between the number of mast cells and structural damage. No association with pain was observed. The synovitis scoring method developed by Guermazi et al. 12 enables the scoring of synovitis throughout the whole knee on contrast-enhanced (CE)-MRI. In chapter 5, we validated this scoring method by comparing the scores with microscopic and macroscopic features of inflammation in OA synovium. From our results we concluded that the synovitis scoring method by Guermazi et al. is a comprehensive and valid non-invasive method to investigate the degree of synovitis in the whole knee in knee OA patients and we therefore used this synovitis scoring method to investigate the role of synovitis in the clinical burden of knee OA (chapter 6 and 8) and the course of synovitis over time (chapter 8). Furthermore, our dedicated setup enabled us to compare synovial inflammation in different OA severity stages. We showed that synovial inflammation was more severe in end-stage knee OA patients compared to patients with mild to established knee OA. Part II The role of synovitis in the clinical burden of OA To gain a better understanding of the role of synovitis in the OA disease process, we studied the association between the presence of synovial inflammation and clinical manifestations of OA. An earlier study showed that OA synovitis on CE-MRI is patchy and heterogeneous throughout the knee 13. However, it was unknown whether synovitis at different sites occurs independently or forms patterns. In chapter 6, we used principal component analysis (PCA)

162

Summary, general discussion & future perspectives

to determined patterns of synovitis on CE-MRI. PCA is a statistical method that determines groups or patterns (named components) based on correlation of features with each other, without including assumptions related to possible mechanisms or anatomical sites. We observed three distinct patterns (Figure 1). Further analyses showed that the pattern that included several patellar sites as well as the site adjacent to the posterior cruciate ligament (PCL) was associated with pain, whereas the other two patterns did not associate with pain, suggesting that pain perception is a localized response. Furthermore, the pattern that included several patellar sites and the site adjacent to the PCL, as well as the pattern that included synovitis at site of a loose body, was associated with radiographic damage in crosssectional analysis. As we (in the chapters 5 and 6) and others have shown, synovitis might play a role in structural damage in OA. Other studies have also shown that other MRI abnormalities besides synovitis could be implicated in radiological progression in knee OA patients, as well. However, many of these features are known to be correlated to each other. Moreover, the knee consists of two joints: the patellofemoral joint (PFJ) and the tibiofemoral joint, and each has different weight bearing properties. Therefore, in chapter 7 we investigated patterns of different tissue abnormalities of both joint using PCA in relation to radiographic progression over a 5-year period. Results suggested that, one the one hand, there seems to be a local response to triggers (clustering of MRI features at same anatomical site), while on the other hand, there is also a non-location specific mechanism for formation of osteophytes. With respect to radiographic progression, results suggest that the PFJ and the TFJ are related and that not only existing structural damage enhance further progression, but also processes reflecting increased bone turnover, such as bone marrow lesions, result in progression. Interestingly, effusion was not incorporated in any of the components, although this was probable due to used cut-off value. Synovitis can only be properly visualized using CE-MRI. Therefore, to understand the development of synovitis throughout the disease course and the effect of change of synovitis over time on OA disease progression and in pain, in chapter 8 we investigated synovitis change on CE-MRI over a 2-year period. Results showed that changes in mean total synovitis scores were not significant on group level although the synovitis score changed during the disease course in individual patients. Increase of synovitis over time was associated with cartilage deterioration, suggesting a role for synovitis as a target for disease-modifying treatment. Change in synovitis was not associated with change in pain, whereas cartilage deterioration was, suggesting that synovitis is of lesser importance in change of subjective pain or cartilage deterioration serves as a mediator in the association between synovitis and pain.

163

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 9

Figure 1. Gadolinium(Gd)-chelate enhanced T1-weighted images showing thee synovitis patterns observed in knee osteoarthritis (OA) patients. Pattern 1 (A-C), A. Axial image showing synovitis at the medial parapatellar site, lateral parapatellar site and Baker cyst, B. Sagittal image with synovitis at the suprapatellar site, infrapatellar site and site adjacent to the posterior cruciate ligament (PCL), C. Sagittal image showing synovitis at the lateral meniscal site. Pattern 2 (D and E), D Sagittal image showing synovitis at the suprapatellar site, intercondylar site and site adjacent to anterior cruciate ligament (ACL), E. Sagittal image showing synovitis at the medial meniscal site. Pattern 3 (F), F. Sagittal image showing synovitis at the suprapatellar site, at the site adjacent to PCL and surrounding loose body (green arrowhead).

164

Summary, general discussion & future perspectives

GENERAL DISCUSSION Immune cells in OA Our aim was to investigate which cells are present in the inflamed synovium of OA patients and which could potentially play a role in OA. In chapter 3 we detected roughly the same cells as described in our systematic review (chapter 2). Interestingly, we found that several immune cells have been activated in the inflamed synovium: T-cells in chapter 3, but also macrophages and mast cells in chapters 3 and 4. These observations suggest that these cells could potentially play a role in the disease. Results from chapter 3 show that in the absence of further activation, T cells mainly produce IL-6 and IL-4 ex vivo, suggesting that they might have been activated in vivo. The possible role of this subset of T cells in disease progression remains to be investigated. Intriguingly, T cells are able to produce a variety of cytokines (including IFNγ and TNFa), upon stimulation, suggesting that IFNγ producing T cells exist in the inflamed OA synovium, but are not activated and therefore other cells might be responsible for producing this cytokine in OA. This is in line with previous investigations, in which IFNγ could also not be detected in CD3-positive cells in OA synovium, by means of immunohistochemistry14,15. Macrophages are the most abundant immune cells in OA (chapter 2 and chapter 3), however their role in OA remains largely unclear. Therefore, in chapter 3 we investigated for the first time, different subsets of activated macrophages (“classically activated” (M1) and “alternatively activated” (M2)) in OA synovium. Interestingly, macrophages isolated from OA synovium produce mainly M1 cytokines and very few M2 cytokines (chapter 3), suggesting a predominantly pro-inflammatory response. Moreover, production of these cytokines in the absence of additional activation indicates that these cells might have an activated state in the synovium. Future studies investigating the localization of these activated cells and their potential association with clinical disease characteristics will offer more insight into the mechanisms involved in the activation of these cells and in their potential involvement in disease pathogenesis. With respect to the activation of mast cells in chapter 4, we observed degranulation of mast cells in OA synovium. However, only a small percentage of mast cells was degranulated and this was not different at different stages of the disease, nor between RA and OA. Furthermore, although degranulation of mast cells was not associated with pain or radiographic damage, the number of mast cells did show an association with structural damage. This observation suggests that mechanisms other than degranulation could be of importance in OA. Observations from this thesis suggest that several synovial immune cells could be of clinical importance in OA; T cells could be of importance in pain perception (chapter 3) and mast cells seem to have a role in structural damage (chapter 4). However, the numbers of patients investigated were small and therefore findings should be replicated in larger

165

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 9

cohorts. Furthermore, it remains unknown which immune cells are associated with cartilage destruction and it is currently unclear which cells infiltrate or become activated in the early stages of the disease course and how these immune cells act throughout the disease course. Inflammation in knee OA not only quantitatively but also qualitatively different from RA In both chapters 2 and 6, differences in the nature of synovial inflammation between OA and RA were addressed. Although for a long time it was assumed that the degree of synovial inflammation is less severe in OA compared with RA, there was a general assumption that the nature of inflammation (e.g. composition of immune cells) was the same. RA, however, is a systemic autoimmune disease, leading to prominent inflammation and joint destruction, whereas OA is not considered to be an auto-immune disease. Therefore, it seems unlikely that the nature of synovial inflammation, such as ratios of immune cells, would be exactly the same in the two different diseases. We hypothesized that inflammation in OA is not only quantitatively different, but also qualitatively different compared to synovial inflammation in RA. In this thesis we found evidence for this hypothesis, since B cells are virtually absent in OA (chapter 2 and 3) and since numbers of mast cells are the same (chapter 2) or even higher (chapter 2 and 4) in synovial tissues of OA patients compared to synovial tissues of RA patients. Another observation that supports the hypothesis that synovial inflammation in OA is different from synovial inflammation in RA is that on CE-MRI synovitis in OA is heterogeneous 13 and patterns of synovitis seems to occur (chapter 4), suggesting that synovitis in knee OA is a localized, rather than a general systemic response observed in RA (personal observation). However, since no valid scoring system exists for CE-MRI in RA it is currently unknown whether a homogeneous nature exists. Based on lack of B cells, higher number of mast cells and the patchy nature one could speculate that inflammation in OA seems to have a more innate immunity profile, whereas RA is more profoundly an adaptive immunity disease. Mast cells in OA and other diseases Mast cells are thought to play an important role in the pathophysiology of RA and are known to have destructive effects on the knee joint though cytokine and chemokine release, direct cell to cell contact or by tryptase release (degranulation)16. Unfortunately, the role of mast cells in OA remains largely unknown as only a small number of studies focused on mast cells in OA (chapter 2). Observations in this thesis suggest that mast cells could be of importance in OA, as we found that the numbers of mast cells are as high or even higher in OA synovial tissues compared to RA synovial tissues (chapter 2 and 4) and mast cells seem to have a role in structural damage (chapter 4). Interestingly, observations from chapter 4 resemble results from a recent study comparing mast cells in synovial tissues of spondylarthropathy (SpA) patients with synovial tissues of RA patients. This study described that mast cells in

166

Summary, general discussion & future perspectives

synovial tissues of patients with SpA were more numerous compared to synovial tissues of RA patients, despite a less degree of inflammation. 17 Although inflammation in SpA and OA are thought to be different, some similarities are seen. For instance, in SpA formation of bony enlargements called syndesmofytes are observed, which are comparable although not similar to osteophytes observed in OA. In SpA the only disease modifying drug with proven efficacy is sulfasalazine, which has been shown to inhibit both degranulation and TNF secretion by mast cells 18,19 and this drug has also been found to have an protective role on chondrocytes in an experimental study20. Therefore, targeting mast cells in OA (for instance via sulfasalazine) could be potentially beneficial in OA patients. However, first further research is needed to elucidate the role of mast cells in OA. Role of imaging in knee OA - MR abnormalities in knee OA As OA is considered to be a whole organ disease and therefore MRI is widely used in OA research to investigate OA processes that might be of importance in pathophysiology of OA. However, the question arises whether MRI is too sensitive in detecting abnormalities and whether these abnormalities are of clinical relevance. This is underscored by a population based study by Guermazi et al. that found an overall prevalence of any MR abnormalities to be 89% in patients without radiographic OA (KL 0) over 50 years, which is extremely high. Interestingly, the prevalence of MR abnormality was high in both patients with painful knees (91%) and without painful knees (88%) 21. - Synovitis and contrast enhancement in knee OA The importance of contrast enhancement in evaluating synovitis on MRI is well recognized in distinguishing synovitis from effusion 11,13,22,23 and is illustrated by Figure 2 It is extremely difficult to assess synovitis on non-CE-MR images. Therefore, synovial inflammation was scored on CE-MR images in most of our studies. Due to possible side effects (mostly concerning the kidney) and associated costs, contrast enhancement has not been widely used in large OA cohorts 24. However, the contrast agent used in our study Gd-DOTA (Dotarem,Guerbet) produces very little side effects.

167

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 9

Figure 2. Axial proton density weighted (PDW) image (left) and Gadolinium (Gd)-chelate enhanced T1-weighted image (right) of same OA patients showing more severe synovitis on the medial site (arrow) of the knee on Gdchelate enhanced image, whereas on the PDW image it is suggested that synovitis is more pronounced at the lateral side of the knee (left image) (adopted from NTvR,2013(4),45-47).

Synovitis throughout the disease course in knee OA Only a few studies investigated synovial inflammation in synovial tissues in different severity stages of knee OA and the results were conflicting, as some authors found synovitis and cytokine expression more pronounced in patients with â&#x20AC;&#x153;lateâ&#x20AC;? OA undergoing arthroplasty3,25,26 whereas others reported the opposite 6,27 or did not find differences28. However, to fully understand the role of inflammation in the disease course of OA, we believe it is of importance to understand whether synovial inflammation precedes structural damage or whether it is a response triggered by the OA process. Further investigation of synovial cells at different disease stages of OA could provide insight into this question, as it could indicate which immune cells infiltrate the synovium in early stages of the disease and at which stage immune cells become activated. Synovitis in knee OA is not a homogeneous process but affects some anatomical sites in the knee more frequently than others (chapter 6). This suggests that synovitis seems to be a response to local triggers rather than a generalized inflammation response. We hypothesized that synovitis could either be triggered by cartilage breakdown products, trauma of structures of the knee or enhanced loading properties. We found that synovial inflammation was most severe in end-stage knee OA in histological samples as well as assessed on CE-MRI (chapter 5, chapter 6) However, when synovitis was investigated over 2-year period, synovitis was found to have a fluctuating nature and did not significantly increase over time (chapter 8). These results suggest that degree of inflammation

168

Summary, general discussion & future perspectives

does not follow a linear increase over time, but tend to worsen in patients with end-stage knee OA. Suggesting that synovitis is a process marker and perhaps a catalyser rather than a causative agent of cartilage deterioration. Which seems in contrast with previous literature suggesting that synovitis precedes cartilage damage 6. Differences could in part be explained by the fact that there is no consensus on the definition of early OA. The role of synovitis in structural damage in OA An important focus in OA research is the development and progression of OA structural damage. Therefore we investigated especially the relationship between synovitis and radiographic damage in cross-sectional analysis (chapter 4-6) or between synovitis and structural OA progression in longitudinal analysis (chapter 7 and 8). As we learned from chapter 5 and 6, synovitis was associated with radiographic damage, although effusion/ synovitis seemed to be of lesser importance in radiographic progression when we took also other joint abnormalities in account in a principal component analysis (chapter 7). However, the lack of association between effusion/synovitis and radiographic progression found in chapter 7 could be explained by the fact that no contrast enhancement was used and by the fact that effusion/synovitis did load on components that were associated with progression. Interestingly, in chapter 8 we found that synovitis change was associated with cartilage deterioration over time. However, it remains unknown whether synovitis acts as an amplifier in the OA process or precedes structural damage. The results in the current thesis suggest that synovitis is a part of the pathophysiological mechanism that leads to structural damage in knee OA and therefore suggests a role for synovitis as a target for disease-modifying treatment. Malalignment in knee OA Malalignment of the knee (e.g. varus or valgus deformity) has been suggested in various OA processes. In the current thesis association of patterns of MR abnormalities with radiographic progression in chapter 7 could in part be explained by varus malalignment, which is often observed in OA. Varus alignment has been shown to be a risk factor for the development of BMLs in the medial TFJ 29-31 and leads to progression of OA in the medial compartment31,32. Furthermore, in patients with a varus malalignment the q-angle of the patella increases resulting in an increasing medial patellar force and increased load on the medial compartment of the patella and subsequently in PFJ progression33-35. Valgus malalignment has been shown to be a risk factor for development and progression of the lateral knee compartment36. In conclusion, malalignment could prove to be of importance in understanding the relationship between joints in the knee and progression in knee OA. It seems conceivable that malalignment would alter the weight baring properties in the knee leading to progression of structural damage and consequently to increased synovitis perhaps also explaining pattern of synovial inflammation observed in chapter 6. 169

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 9

Synovitis as determinant of pain Previous literature has suggested that synovitis is an important determinant of pain. Therefore, a large part of this thesis was aimed to investigate the association between pain and synovitis (chapter 3-6 and chapter 8). Interestingly, synovitis at especially the parapatellar sites was more prone to pain experience than synovitis at other sites (chapter 6). Earlier studies on the innervation and pain sensation of the knee showed that the patellar region is richly innervated and that anterior synovial tissue in the vicinity of the patella is very sensitive to pain stimulation37,38. These findings suggest the location specific role of synovitis on subjective pain and could provide an opportunity for local treatment of synovitis and pain. Although, we did find an association with pain (chapter 3, 5 and 6) overall the effect sizes were not high. In chapter 8 synovitis change was only associated with change of pain in an unadjusted model, however cartilage deterioration was associated with change in pain. This could suggest that cartilage deterioration is a mediator in the association with pain. In conclusion synovitis therefore seems to contribute, but seems not to be the only factor in underlying biological mechanism in experiencing pain in knee OA patients. Another explanation for the low effect sizes lies in the fact that subjective pain and especially chronic pain is not only determined by biological processes, but is determined by several biological, psychological and social processes interacting creating a biopsychosocial pain model 39-41.

FUTURE PERSPECTIVES We started this thesis in chapter 2 with a research agenda for the investigation of synovitis. We conclude this thesis by introducing topics for a new research agenda for future investigations, aiming to fill the gaps of current knowledge concerning the role of synovitis in knee OA that will lead to development of disease modifying drugs. - Consensus in defining OA severity stage. To further understand whether synovitis precedes knee OA or whether synovitis is a process marker in OA, it is imperative to have a clear definition of OA. Especially, consensus should be reached on defining different severity stages in knee OA. Currently intensive work is performed to reach a consensus for the definition of â&#x20AC;&#x153;early osteoarthritisâ&#x20AC;?. - Understanding synovitis throughout the disease course of knee OA To understand the role of inflammation in the disease course of OA, more longitudinal data of larger groups over longer periods of time are necessary. Furthermore, to fully understand the role of immune cells throughout the disease course it would be interesting to investigate synovial biopsies at several time points throughout the disease course. This would provide insight into the kinetic of cell infiltration and activation during the disease, which could lead to a better understanding of the cellular mechanisms involved in disease progression. 170

Summary, general discussion & future perspectives

- Further investigation of the role of CD4+ T cells in pain To further investigate the role of T cells in pain perception it would be interesting to see whether T cells accumulate in the vicinity of nerve endings. Interestingly, in chapter 3 we found an association of CD4+ T cells with pain. Furthermore, we found that histological inflammation is correlated with inflammation seen on MRI (chapter 5). Therefore, it would be interesting to see whether T cells accumulate at the anatomical sites that formed a pattern that was associated with pain in chapter 6. - Further investigation of the role of mast cells as possible target for therapy in patients with OA The data presented in this thesis suggest a role for mast cells in the disease process in OA. Future studies could focus on a better characterization of mast cells in OA, for example through large-scale mRNA sequencing, which could provide insight into the possible effector molecules expressed by mast cells, as well as the possible mechanism of activation of mast cells in OA. Furthermore, investigating soluble mediators released upon mast cell activation and their effect on bone/cartilage could contribute to our understanding of the role of mast cells in the pathogenesis of knee OA. Finally, the possibility of using disease modifying drugs that target mast cells should be further explored in pre-clinical and clinical studies of OA. - Synovitis on CE-MRI in other inflammatory diseases Although CE-MRI compares well with histological signs of inflammation in knee OA patients, it is unknown whether CE-MRI could be of use in the differential diagnosis of inflammatory diseases in research or clinical practice. Therefore, it would be interesting to see whether the synovitis scoring system by Guermazi et al. could be applied in RA or other inflammatory arthritis diseases. - Alternatives for visualization of synovitis Although CE-MRI constitutes a non-invasive method to investigate synovitis in geMstoan study a large proportion of patients could not undergo MRI due to contra-indications for MRI. As OA is a disease of the aging population this is more of a problem in OA compared to other inflammatory arthritis diseases such as RA or SpA. Therefore, other non-invasive methods should be explored to investigate synovial inflammation in knee OA patients. A promising alternative is ultrasound of the knee. Ultrasound is not only is able to visualize synovitis but also has the ability to visualize osteophytes and cartilage damage in multiple planes. Only recently an ultrasound scoring method has been developed, that compares well with radiographic features of OA 42. It would be interesting to see how well this score compares with abnormalities as seen on MRI, especially with synovitis on contrast enhanced MRI. When ultrasound proofs to be a valid alternative for contrast enhanced MRI it will offer an opportunity to follow synovitis in clinical practice for experienced users of ultrasound. - To understand the role of biomechanics in knee OA

171

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 9

Especially in the knee OA biomechanics seem to be of importance in developing structural damage and could explain why structural abnormalities in the PFJ and TFJ are related. Investigation of association of patterns of MRI abnormalities with malalignment of the knee could provide more insight. Furthermore, varus or valgus malalignment could also explain the different patterns of synovitis and therefore it would be interesting to whether patterns of synovitis could be explained by either varus or valgus abnormalities.

172

Summary, general discussion & future perspectives

REFERENCES

1 Bijlsma JW, Berenbaum F, Lafeber FP. Osteoarthritis: an update with relevance for clinical practice. Lancet 2011;377:2115-26. 2 Felson DT. Clinical practice. Osteoarthritis of the knee. N Engl J Med 2006;354:841-8. 3 Loeuille D, Chary-Valckenaere I, Champigneulle J, Rat AC, Toussaint F, Pinzano-Watrin A et al. Macroscopic and microscopic features of synovial membrane inflammation in the osteoarthritic knee: correlating magnetic resonance imaging findings with disease severity. Arthritis Rheum 2005;52:3492-501. 4 Loeuille D, Rat AC, Goebel JC, Champigneulle J, Blum A, Netter P et al. Magnetic resonance imaging in osteoarthritis: which method best reflects synovial membrane inflammation? Correlations with clinical, macroscopic and microscopic features. Osteoarthritis Cartilage 2009;17:1186-92. 5 Pearle AD, Scanzello CR, George S, Mandl LA, DiCarlo EF, Peterson M et al. Elevated high-sensitivity C-reactive protein levels are associated with local inflammatory findings in patients with osteoarthritis. Osteoarthritis Cartilage 2007;15:516-23. 6 Benito MJ, Veale DJ, FitzGerald O, van den Berg WB, Bresnihan B. Synovial tissue inflammation in early and late osteoarthritis. Ann Rheum Dis 2005;64:1263-7. 7 Sellam J and Berenbaum F. The role of synovitis in pathophysiology and clinical symptoms of osteoarthritis. Nat Rev Rheumatol 2010;6:62535. 8 Roemer FW, Guermazi A, Felson DT, Niu J, Nevitt MC, Crema MD et al. Presence of MRI-detected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30-month follow-up: the MOST study. Ann Rheum Dis 2011;70:1804-9. 9 Roemer FW, Zhang Y, Niu J, Lynch JA, Crema MD, Marra MD et al. Tibiofemoral joint osteoarthritis: risk factors for MR-depicted fast cartilage loss over a 30-month period in the multicenter osteoarthritis study. Radiology 2009;252:772-80. 10 Yusuf E, Kortekaas MC, Watt I, Huizinga TW, Kloppenburg M. Do knee abnormalities visualised on MRI explain knee pain in knee osteoarthritis? A systematic review. Ann Rheum Dis 2011;70:60-7.

11 Baker K, Grainger A, Niu J, Clancy M, Guermazi A, Crema M et al. Relation of synovitis to knee pain using contrast-enhanced MRIs. Ann Rheum Dis 2010;69:1779-83. 12 Guermazi A, Roemer FW, Hayashi D, Crema MD, Niu J, Zhang Y et al. Assessment of synovitis with contrast-enhanced MRI using a wholejoint semiquantitative scoring system in people with, or at high risk of, knee osteoarthritis: the MOST study. Ann Rheum Dis 2011;70:805-11. 13 Roemer FW, Kassim JM, Guermazi A, Thomas M, Kiran A, Keen R et al. Anatomical distribution of synovitis in knee osteoarthritis and its association with joint effusion assessed on non-enhanced and contrast-enhanced MRI. Osteoarthritis Cartilage 2010;18:1269-74. 14 Steiner G, Tohidast-Akrad M, Witzmann G, Vesely M, Studnicka-Benke A, Gal A et al. Cytokine production by synovial T cells in rheumatoid arthritis. Rheumatology 1999;38:202-13. 15 Fonseca JE, Edwards JCW, Blades S, Goulding NJ. Macrophage subpopulations in rheumatoid synovium: Reduced CD163 expression in CD4+ T lymphocyte-rich microenvironments. Arthritis and rheumatism 2002;46:1210-6. 16 Frenzel L and Hermine O. Mast cells and inflammation. Joint Bone Spine 2013;80:141-5. 17 Noordenbos T, Yeremenko N, Gofita I, van de Sande M, Tak PP, Canete JD et al. Interleukin17-positive mast cells contribute to synovial inflammation in spondylarthritis. Arthritis Rheum 2012;64:99-109. 18 Barrett KE, Tashof TL, Metcalfe DD. Inhibition of IgE-mediated mast cell degranulation by sulphasalazine. Eur J Pharmacol 1985;107:27981. 19 Bissonnette EY, Enciso JA, Befus AD. Inhibitory effects of sulfasalazine and its metabolites on histamine release and TNF-alpha production by mast cells. J Immunol 1996;156:218-23. 20 Lakey RL and Cawston TE. Sulfasalazine blocks the release of proteoglycan and collagen from cytokine stimulated cartilage and downregulates metalloproteinases. Rheumatology (Oxford) 2009;48:1208-12. 21 Guermazi A, Niu J, Hayashi D, Roemer FW, Englund M, Neogi T et al. Prevalence of abnormalities in knees detected by MRI in adults without knee osteoarthritis: population based observational study (Framingham Osteoarthritis Study). BMJ 2012;345:e5339.

173

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 9

22 Roemer FW, Guermazi A, Zhang Y, Yang M, Hunter DJ, Crema MD et al. Hoffaâ&#x20AC;&#x2122;s Fat Pad: Evaluation on Unenhanced MR Images as a Measure of Patellofemoral Synovitis in Osteoarthritis. AJR Am J Roentgenol 2009;192:1696-700. 23 Hayashi D, Roemer FW, Guermazi A. Osteoarthritis year 2011 in review: imaging in OA--a radiologistsâ&#x20AC;&#x2122; perspective. Osteoarthritis Cartilage 2012;20:207-14. 24 Collidge TA, Thomson PC, Mark PB, Traynor JP, Jardine AG, Morris ST et al. Gadoliniumenhanced MR imaging and nephrogenic systemic fibrosis: retrospective study of a renal replacement therapy cohort. Radiology 2007;245:168-75. 25 Smith MD, Triantafillou S, Parker A, Youssef PP, Coleman M. Synovial membrane inflammation and cytokine production in patients with early osteoarthritis. J Rheumatol 1997;24:365-71. 26 Myers SL, Brandt KD, Ehlich JW, Braunstein EM, Shelbourne KD, Heck DA et al. Synovial inflammation in patients with early osteoarthritis of the knee. J Rheumatol 1990;17:1662-9. 27 Ning L, Ishijima M, Kaneko H, Kurihara H, Arikawa-Hirasawa E, Kubota M et al. Correlations between both the expression levels of inflammatory mediators and growth factor in medial perimeniscal synovial tissue and the severity of medial knee osteoarthritis. Int Orthop 2010; 28 Pearle AD, Scanzello CR, George S, Mandl LA, DiCarlo EF, Peterson M et al. Elevated high-sensitivity C-reactive protein levels are associated with local inflammatory findings in patients with osteoarthritis. Osteoarthritis Cartilage 2007;15:516-23. 29 Segal NA, Kern AM, Anderson DD, Niu J, Lynch J, Guermazi A et al. Elevated tibiofemoral articular contact stress predicts risk for bone marrow lesions and cartilage damage at 30 months. Osteoarthritis Cartilage 2012;20:11206. 30 Hayashi D, Englund M, Roemer FW, Niu J, Sharma L, Felson DT et al. Knee malalignment is associated with an increased risk for incident and enlarging bone marrow lesions in the more loaded compartments: the MOST study. Osteoarthritis Cartilage 2012;20:1227-33. 31 Sharma L, Chmiel JS, Almagor O, Felson D, Guermazi A, Roemer F et al. The role of varus and valgus alignment in the initial development of knee cartilage damage by MRI: the MOST study. Ann Rheum Dis 2013;72:235-40.

174

32 Walker EA, Davis D, Mosher TJ. Rapidly progressive osteoarthritis: biomechanical considerations. Magn Reson Imaging Clin N Am 2011;19:283-94. 33 Huberti HH and Hayes WC. Patellofemoral contact pressures. The influence of q-angle and tendofemoral contact. J Bone Joint Surg Am 1984;66:715-24. 34 Elahi S, Cahue S, Felson DT, Engelman L, Sharma L. The association between varus-valgus alignment and patellofemoral osteoarthritis. Arthritis Rheum 2000;43:1874-80. 35 Cahue S, Dunlop D, Hayes K, Song J, Torres L, Sharma L. Varus-valgus alignment in the progression of patellofemoral osteoarthritis. Arthritis Rheum 2004;50:2184-90. 36 Felson DT, Niu J, Gross KD, Englund M, Sharma L, Cooke TD et al. Valgus malalignment is a risk factor for lateral knee osteoarthritis incidence and progression: findings from the Multicenter Osteoarthritis Study and the Osteoarthritis Initiative. Arthritis Rheum 2013;65:355-62. 37 Dye SF, Vaupel GL, Dye CC. Conscious neurosensory mapping of the internal structures of the human knee without intraarticular anesthesia. Am J Sports Med 1998;26:773-7. 38 Buma P. Innervation of the patella. An immunohistochemical study in mice. Acta Orthop Scand 1994;65:80-6. 39 Gatchel RJ, Peng YB, Peters ML, Fuchs PN, Turk DC. The biopsychosocial approach to chronic pain: scientific advances and future directions. Psychol Bull 2007;133:581-624. 40 Colloca L and Benedetti F. How prior experience shapes placebo analgesia. Pain 2006;124:12633. 41 Wager TD. Expectations and anxiety as mediators of placebo effects in pain. Pain 2005;115:225-6. 42 Riecke BF, Christensen R, Torp-Pedersen S, Boesen M, Gudbergsen H, Bliddal H. An ultrasound score for knee osteoarthritis: a cross-sectional validation study. Osteoarthritis Cartilage 2014;22:1675-91.

CHAPTER 10 NEDERLANDSE SAMENVATTING, DISCUSSIE & TOEKOMSTPERSPECTIEVEN

Chapter 10

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 176

Nederlandse samenvatting

SAMENVATTING Artrose is een heterogene aandoening die alle gewrichten in het lichaam kan treffen, maar vooral voorkomt in het kniegewricht1. Het is een progressieve ziekte, die uiteindelijk leidt tot gewrichtsdestructie met als klinische symptomen pijn en functionele beperkingen. Helaas, zijn er voor artrose op dit moment nog geen ziektemodificerende geneesmiddelen. De huidige therapie richt zich dan ook voornamelijk op symptoomverlichting. Aangezien artrose toeneemt met de leeftijd is artrose een groot medisch, sociaal en economisch probleem in onze vergrijzende samenleving. Alhoewel, knieartrose een zeer veel voorkomende ziekte is, zijn de onderliggende pathofysiologische en pijn mechanismen nog grotendeels onbekend. Daarom is wetenschappelijk onderzoek naar deze onderliggende mechanismen van groot belang. Lange tijd werd artrose beschouwd als een niet inflammatoire ziekte. Recentelijk is echter gebleken dat synoviale ontsteking mogelijk wel een belangrijke rol speelt in de pathofysiologie van artrose 2-7, omdat het kraakbeenschade op de lange termijn voorspelt8,9 en het een determinant van pain is10,11. Echter, welke rol synovitis precies speelt in de pathofysiologie van artrose is nog grotendeels onbekend. Daarom onderzochten we in dit proefschrift de aard van de synoviale ontsteking in knieartrose en wat de mogelijke bijdrage van synovitis is aan de klinische verschijnselen van de knieartrose. De resultaten gepresenteerd in dit proefschrift geven inzicht in de verschillende aspecten van synoviale ontsteking in knieartrose gericht op het vergroten van onze kennis over de pathofysiologie van artrose die uiteindelijke zou kunnen leiden tot de ontwikkeling van ziektemodificerende medicijnen voor artrose. Deel I De aard van de synoviale ontsteking in de knie OA Als eerste stap in ons onderzoek naar de cellulaire en moleculaire aard van de synoviale ontsteking, hebben we in hoofdstuk 2 middels een systematische review de huidige kennis ten aanzien van inflammatoire eigenschappen, immuuncellen en hun cytokines in het synovium van artrose patiĂŤnten samengevat. Onze hypothese dat al veel kennis direct voorhanden is, werd bevestigd door het feit dat we 100 artikelen vonden over het onderwerp. Uit dit literatuuroverzicht werd duidelijk dat de histologie van de synoviale ontsteking bij artrose vergelijkbaar is met de ontsteking die wordt gezien bij reumatoĂŻde artritis (RA) patiĂŤnten. De meest frequent waargenomen celtypen zijn macrofagen, T-cellen en mestcellen, terwijl B-cellen bijna nooit werden gevonden. Bij artrose zijn voornamelijk cytokines gerelateerd aan T-cellen en macrofagen functie beschreven en onderzocht. Echter over hun daadwerkelijke cellulaire oorsprong is weinig bekend. Over het algemeen was de hoeveelheid ontsteking (een afspiegeling van hoeveelheid van de (meeste) infiltrerende immuuncellen) in artrose lager dan in RA weefsels. Het was daarom dan ook opvallend dat er

177

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 10

evenveel of in sommige studies zelf meer mestcellen in het synovium van artrosepatiënten werden gevonden dan in het synovium van RA patiënten. Gebaseerd op de resultaten van onze review werd eerder wetenschappelijke onderzoeksagenda gemaakt, die als uitgangspunt diende voor de verdere studies beschreven in dit proefschrift: - Analyse van de immuuncellen en hun activatie in het synovium van knieartrose patiënten en hun relatie tot klinische ziekte kenmerken. - Een diepgaande analyse van de mestcellen en hun activatie in synovium van knieartrose patiënten in verschillende ziekte stadia en hun relatie met de klinische ziekte kenmerken. - Onderzoek van synovitis gedurende het ziektebeloop van knie artrose. - Associatie van synovitis met klinische parameters zoals pijn, structurele schade en progressie. In hoofdstuk 3 onderzochten we door middel van FACS-analyse de verschillende immuuncellen in het synovium en in het vetlichaam van Hoffa, een weefsel waarin immuuncellen in overvloede aanwezig zijn, van knieartrose patiënten. We vonden dat macrofagen en T-cellen, gevolgd door mestcellen, de meest voorkomende immuuncellen waren en dat subpopulaties van zowel T-cellen en macrofagen zich in een geactiveerde toestand bevonden. De geobserveerde associatie van CD4 + T-cellen met pijn, geeft mogelijk voor het eerst een cellulaire verklaring voor de al lang bekende associatie tussen synovitis en pijn in knieartrose patiënten. Aangezien we van hoofdstuk 2 leerden dat mestcellen een belangrijke rol zouden kunnen spelen in artrose, hebben we, in hoofdstuk 4, mestcellen in synovium van knieartrose patiënten onderzocht. De resultaten laten zien dat in het synovium van knieartrose patiënten een significant hoger aantal mestcellen aanwezig is dan in het synovium van RA patiënten, alhoewel er bij RA juist meer ontsteking (H&E kleuring) werd gezien. De gevonden associatie van aantal mestcellen en structurele schade, suggereert dat mestcellen tevens van klinisch belang zijn. Er werd geen associatie van aantal mestcellen met pijn gevonden. De synovitis scoringsmethode, ontwikkeld door Guermazi et al.12, maakt het scoren van synovitis door de hele knie op MRI met contrast mogelijk. In hoofdstuk 5, hebben we deze scoringsmethode gevalideerd door de synovitis scores op MRI te vergelijken met microscopische (gouden standaard) en macroscopische kenmerken van synoviale inflammatie. Uit onze resultaten concludeerde we dat de synovitis scoringsmethode door Guermazi et al. een makkelijk hanteerbare, valide en niet invasieve methode is om synovitis door de hele knie te beoordelen. Daarom, hebben we deze scoringsmethode gebruikt om

178

Nederlandse samenvatting

de rol van synovitis in de ziektelast van knieartrose patiënten (hoofdstuk 6 en 8) en het verloop van synovitis over de tijd verder te onderzoeken (hoofdstuk 8). Bovendien, hadden wij de mogelijkheid om de hoeveelheid ontsteking in het synovium van patiënten met verschillende ziektestadia met elkaar te vergelijken. We toonden aan dat er meer synoviale ontsteking was bij patiënten met een eindstadium van knieartrose dan bij patiënten met een milde tot matige vorm van knieartrose. Deel II De rol van synovitis in de klinische ziektelast van artrose Om de rol van synovitis in het ziekteproces van artrose beter te begrijpen, bestudeerden we het verband tussen de aanwezigheid van synoviale ontsteking en de klinische manifestaties van artrose. Vanuit eerder artroseonderzoek was bekend dat synovitis op MRI met contrast niet op alle plekken in de knie evenveel aanwezig is, maar dat er op sommige plekken in de knie meer ontsteking zit dan op andere plekken13. Het was echter niet bekend of synovitis op verschillende locaties onafhankelijk van elkaar voorkomen (locaal proces), of dat er patronen waren van synovitis (bepaalde plekken van synovitis in de knie die samengaan). In hoofdstuk 6, hebben we gebruik gemaakt van een principal component analyse (PCA) om patronen van synovitis op MRI met contrast aan te tonen. PCA is een statistische methode die groepen of patronen (genoemd componenten) vormt op basis van correlatie die de variabelen met elkaar hebben, zonder dat hierbij gebruik wordt maakt van aannames met betrekking tot mogelijke mechanismen of anatomische locatie. We vonden met deze methode drie verschillende patronen (figuur 1, hoofdstuk 9). Verder vonden we dat het patroon dat verschillende locaties in de buurt van de patella en de locatie bij de achterste kruisband omvatte met pijn geassocieerd was, terwijl de andere twee patronen niet met pijn associeerde. Dit suggereert dat pijnperceptie locatie specifiek is. Bovendien vonden we in deze cross-sectionele analyse dat dit patroon en ook het patroon dat waarin loose body omringd met synovitis zat, associeerde met structurele schade op een röntgenfoto. Zoals wij (in de hoofdstukken 5 en 6) en eerdere wetenschappelijke onderzoeken hebben aangetoond, zou synovitis een rol kunnen spelen in de structurele schade bij artrose. Echter, eerder onderzoek leert ons ook dat andere met artrose gerelateerde afwijkingen in de knie (te visualiseren op MRI) een relatie hebben met radiografische progressie. Het feit dat al deze kenmerken met elkaar correleren, maakt onderzoek hiernaar lastig. Bovendien, bestaat de knie uit twee gewrichten: het patellofemoraal gewricht (PFJ) en het tibiofemoraal gewricht (TFJ), die beide verschillende gewichtsdragende eigenschappen hebben. Daarom onderzochten we in hoofdstuk 7 met behulp van PCA patronen van verschillende weefsel abnormaliteiten op MRI van beide gewrichten. Vervolgens onderzochten we de relatie van deze patronen met radiografische progressie over een periode van 5 jaar. Resultaten suggereerden dat er aan de ene kant een lokale reactie op triggers lijkt te zijn (clustering

179

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 10

van MRI kenmerken op dezelfde anatomische locatie), terwijl er anderzijds ook een nietlocatie specifieke mechanisme lijkt te zijn voor de vorming van osteofyten. Wat betreft radiografische progressie bleken beide gewrichten van invloed op elkaar te zijn en dat niet alleen bestaande structurele schade, maar ook processen die een rol spelen bij het botmetabolisme, zoals beenmerg laesies, leiden tot progressie. Een interessante observatie was dat effusie niet in de patronen voorkwam, hoewel dit waarschijnlijk aan het gebruikte afkappunt te wijten is. Omdat op een MRI synovitis het betrouwbaarste kan worden gevisualiseerd wanneer er gebruik gemaakt wordt van contrast, hebben we in hoofdstuk 8 van deze methode gebruik gemaakt om de ontwikkeling van synovitis gedurende het ziektebeloop van artrose (2 jaar) en om het effect van verandering op progressie en pijn te onderzoeken. De resultaten toonden aan dat op groepniveau er geen significante verschillen waren na 2 jaar in de gemiddelde totale synovitis score, alhoewel bij individuele patiënten de synovitis score wel fluctueerde over de tijd. Verergering van synovitis over de tijd was geassocieerd met kraakbeen schade op MRI na 2 jaar. Dit suggereert dat synovitis een mogelijk aangrijpingspunt is voor ziekte modificerende medicatie. Verandering in synovitis over de tijd was niet geassocieerd met veranderingen van pijn, terwijl verergering van kraakbeenschade wel een associatie met pijn toonde. Deze resultaten suggeren dat ofwel synovitis een minder grote rol speelt bij pijnperceptie dan eerder werd gedacht of dat kraakbeenschade fungeert als een mediator in de associatie tussen synovitis en pijn.

ALGEMENE DISCUSSIE Immuuncellen in artrose Één van de doelen van dit proefschrift was om te onderzoeken welke immuuncellen, die mogelijk een rol bij de pathofysiologie van artrose kunnen spelen, aanwezig zijn in het ontstoken synovium van artrose patiënten. In hoofdstuk 3 observeerde we ongeveer dezelfde soort immuuncellen als eerdere studies, zoals beschreven in onze systematische review (hoofdstuk 2). Opvallend was dat we zagen dat verschillende immuuncellen geactiveerd waren in het ontstoken synovium: T-cellen in hoofdstuk 3, maar ook macrofagen en mestcellen in de hoofdstukken 3 en 4. Hun geactiveerde staat suggereert dat deze immuuncellen mogelijk een actieve rol spelen in het artrose ziekteproces. Uit de resultaten uit hoofdstuk 3 blijkt dat T-cellen bij afwezigheid van verdere stimulatie tot activatie, voornamelijk IL-6 en IL-4 produceren (ex vivo). Dit suggereert dat deze cellen mogelijk al eerder (in vivo) werden geactiveerd. Wat de precieze rol van deze subset van T-cellen in de progressie van artrose is, moet nog worden onderzocht.

180

Nederlandse samenvatting

Het is intrigerend dat T-cellen wanneer ze verder worden gestimuleerd (ex vivo) ook een aantal andere cytokines (zoals IFNy en TNFa) kunnen produceren. Dit suggereert dat alhoewel er wel IFNy producerende T-cellen aanwezig zijn in het ontstoken synovium, ze niet in vivo worden geactiveerd. Dit zou betekenen dat andere cellen dan IFNy producerende T-cellen verantwoordelijk zouden zijn voor de productie van IFNy in artrose. Deze observatie geeft een mogelijke verklaring voor het feit dat in eerdere onderzoeken IFNy (met behulp van immunohistochemie) niet kon worden gedetecteerd in CD3-positieve cellen in synovium van artrose patiënten14,15. Ondanks dat macrofagen de meest voorkomende soort immuuncellen zijn in artrose (hoofdstuk 2 en 3), blijft hun rol in artrose nog grotendeels onduidelijk. Daarom onderzochten we (als eerste) in hoofdstuk 3 de verschillende subsets van geactiveerde macrofagen ((M1) “klassiek geactiveerd” en “alternatief geactiveerd” (M2)) in het synovium van artrose patiënten. We vonden dat uit het synovium geïsoleerde macrofagen voornamelijk M1 cytokines en zeer weinig M2 cytokines produceren, wat op een overwegend proinflammatoire respons duidt. Bovendien geeft het feit dat de macrofagen deze cytokinen produceren zonder dat zij werden gestimuleerd tot activatie, aan dat deze cellen mogelijk al een geactiveerde toestand hadden in het synovium. Toekomstige studies naar de lokalisatie van deze geactiveerde cellen en onderzoek naar hun mogelijke associaties met klinische ziektekenmerken, zal meer inzicht bieden in de activatie mechanismen van deze cellen en hun potentiële betrokkenheid bij de pathogenese van artrose. In hoofdstuk 4 observeerde we degranulatie van mestcellen, hetgeen activatie van de mestcellen in synovium betekend. Slecht een klein percentage van de mestcellen was gedegranuleerd en dit percentage verschilde niet tussen patiënten met verschillende ziektestadia van artrose, noch tussen verschillende ziektebeelden (artrose en RA). Alhoewel degranulatie geen rol leek te spelen bij pijn of radiografische schade, was het aantal mestcellen wel geassocieerd met structurele schade. Deze observatie suggereert dat mechanismen anders dan degranulatie van belang zijn in artrose. Verschillende observaties uit dit proefschrift suggeren dat verschillende synoviale immuuncellen van klinisch belang in artrose kunnen zijn; zo kunnen T-cellen mogelijk van belang zijn bij pijnperceptie (hoofdstuk 3) en lijken mestcellen een rol te spelen bij structurele schade (hoofdstuk 4). Echter, aangezien het aantal onderzochte patiënten in onze studies klein was, zullen deze bevindingen gerepliceerd moeten worden in grotere cohorten. Bovendien blijft het op dit moment nog onduidelijk welke immuuncellen zorgen voor kraakbeen destructie, is het momenteel nog onduidelijk welke cellen infiltreren of geactiveerd worden in de vroege stadia van artrose en hoe deze immuuncellen zich gedragen gedurende het ziekteverloop.

181

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 10

Ontsteking in de knieartose is niet alleen kwantitatief, maar ook kwalitatief verschillend van RA In hoofdstukken 2 en 6, werden de verschillen in de aard van synoviale ontsteking tussen artrose en RA patiënten onderzocht. Lange tijd werd aangenomen dat ondanks dat de mate van synoviale inflammatie beduidend minder ernstig is bij artrosepatiënten in vergelijking met RA patiënten, de aard van ontsteking (bijvoorbeeld de verhoudingen van immuuncellen in het weefsel) hetzelfde was. RA is echter een systemische auto-immuunziekte met ernstige gewrichtsontsteking met gewrichtsschade tot gevolg, terwijl artrose niet als een autoimmune ziekte wordt beschouwd. Het lijkt daarom uitermate onwaarschijnlijk dat de aard van synoviale inflammatie van deze twee verschillende ziektes wel precies hetzelfde is. Onze hypothese op voorhand was dan ook dat de aard van synoviale ontsteking in artrose niet alleen kwantitatief, maar ook kwalitatief verschilt van de ontsteking gezien bij RA patiënten. De resultaten uit dit proefschrift bevestigen deze hypothese aangezien B-cellen vrijwel afwezig zijn bij artrose (hoofdstuk 2 en 3) en aangezien het aantal mestcellen hetzelfde (hoofdstuk 2) of zelfs hoger (hoofdstuk 2 en 4) is in de synoviale weefsels van OA patiënten vergeleken met synoviale weefsels van RA patiënten. Een andere observatie die de hypothese ondersteund is dat synovitis op MRI met contrast bij artrose een heterogeen aspect heeft 13 en dat er patronen van synovitis worden gevormd (hoofdstuk 4). Deze observaties op MRI suggereren dat synovitis bij knieartrose vooral lokaal voorkomt in tegenstelling tot de gegeneraliseerde ontsteking van de knie op basis van een systemische reactie die wordt waargenomen bij RA patiënten (persoonlijke waarneming). Echter, aangezien er nog geen valide scoringsmethode voor het scoren van synovitis op MRI met contrast voor de knie in RA patiënten is, blijft dit een persoonlijke observatie en moet toekomstig onderzoek uitwijzen of er inderdaad sprake is van homogene synovitis bij RA patiënten. Op basis van het vrijwel afwezig zijn van B-cellen, het hoger aantal mestcellen en het lokale karakter van synovitis bij artrose, zouden we kunnen speculeren dat er bij artrose spraken is van een overwegend aspecifieke immuunrespons, terwijl er bij patiënten met RA spraken is van een overwegend adaptieve immuunrespons. Mestcellen bij artrose en andere ziekten Van mestcellen wordt gedacht dat ze een belangrijke rol in de pathofysiologie van RA spelen. In RA staan mestcellen bekend om hun destructieve effecten op het kniegewricht doormiddel van cytokine en chemokine release, direct cel tot cel contact of door hun tryptase afgifte (degranulatie)16. De rol van mestcellen in artrose is echter nog grotendeels onbekend aangezien slechts een handvol studies mestcellen in het synovium van artrsoepatienten heeft onderzocht (hoofdstuk 2). Dit proefschrift toont dat mestcellen van belang kunnen zijn bij artrose door aan te tonen dat er evenveel (hoofdstuk 2) of soms meer mestcellen (hoofdstuk 2 en 4) gezien worden bij artrose dan in het synovium van

182

Nederlandse samenvatting

RA patiënten (hoofdstuk 2 en 4) en door aan te tonen dat mestcellen associëren met structurele schade (hoofdstuk 4). Opvallend is dat onze observaties uit hoofdstuk 4 een grote gelijkenis hebben met de resultaten van een recente studie, die het aantal mestcellen in het synovium van patiënten met spondylarthropathy (SpA) patiënten vergeleek met die van RA patiënten. Deze studie beschreef een hoger aantal mestcellen in het synovium van SpA patiënten dan in het synovium van RA patiënten, ondanks dat er minder ontsteking bij SpA patiënten werd gezien17. Alhoewel er wordt verondersteld dat er verschillen zijn tussen de inflammatie gezien bij SpA en artrose patiënten, zijn er ook een aantal gelijkenissen. Zo worden bij SpA bijvoorbeeld ook botuitstulpingen (syndesmofyten) waargenomen, die vergelijkbaar (alhoewel niet geheel in overeenstemming met) zijn met de osteofyten bij artrose. Bij patiënten met SpA is sulfazalazine het enige ziektemodificerende medicijn met bewezen werkzaamheid. Van sulfasalazine is aangetoond dat het zowel de TNFα secretie als de degranulatie van mestcellen remt18,19. Opvallend is dat een recente experimentele studie aantoonde dat sulfazalazine ook beschermend werkt op chondrocyten, welke van belang zijn bij artrose 20. Daarom zouden medicijnen die zich richten op de remming van mestcellen (bijvoorbeeld via sulfasalazine) mogelijk een gunstig effect kunnen hebben bij artrose patiënten. Echter is het eerst van belang om meer wetenschappelijk onderzoek te verrichten gericht op het ophelderen van de precieze rol van mestcellen in het artrose proces. De rol van beeldvorming in knieartrose - MR afwijkingen in knieartrose Tegenwoordig weten we dat artrose niet alleen maar een ziekte van het kraakbeen en het bot is, maar dat ook andere structuren in de knie een rol spelen. Daarom wordt er, om de onderliggende processen die van belang kunnen zijn in de pathofysiologie van knieartrose in beeld te krijgen, bij wetenschappelijk onderzoek naar artrose steeds vaker gebruik gemaakt van MRI. Echter bij de interpretatie van de afwijkingen op MRI rijst de vraag of MRI misschien niet te gevoelig is in het detecteren van afwijkingen en of de op MRI geobserveerde afwijkingen wel klinisch relevant zijn. Zo werd recentelijk in een groot populatieonderzoek gevonden dat 89% van de patiënten boven de 50 zonder radiografische artrose (KL 0) afwijkingen samenhangend met artrose op de MRI hadden. Deze afwijkingen werden niet alleen gezien bij de patiënten met pijnlijke knieën (91%) maar ook bij de patiënten zonder pijnlijke knieën (88%)21. - Synovitis en gebruik van contrast bij artrose Dat het gebruik van contrast van essentieel belang is bij het onderscheiden van synovitis van effusie op MRI is algemeen bekend11,13,22,23 en wordt geïllustreerd in figuur 2 van hoofdstuk 9 waar we zien dat synovitis uitermate lastig te beoordelen is wanneer er geen contrast wordt gebruikt. Daarom hebben we bij het merendeel van de studies beschreven in dit proefschrift

183

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 10

contrast gebruikt. Er kleven wel een aantal nadelen aan het gebruik van contrast zoals het gevaar op complicaties bij gebruik van contrast (voornamelijk gerelateerd aan de nieren) en de bijkomende kosten. Dit is de reden waarom contrast eerder niet op grote schaal werd gebruikt in artrose cohorten 24. Echter het contrastmiddel dat wij gebruikten bij de studies beschreven in dit proefschrift zijn (Gd-DOTA (Dotarem, Guerbet)) is relatief veilig en kent weinig bijwerkingen, waardoor wij de mogelijkheid hadden om uitgebreid onderzoek naar synovitis te doen. Synovitis in het artrose ziektebeloop Slechts een aantal studies hebben synoviale ontsteking in het synovium van patiënten met verschillende artrose ziektestadia vergeleken en hun resultaten zijn tegenstrijdig; zo vonden sommige studies (en wij ook in hoofdstuk 5) dat synoviale ontsteking en cytokine-expressie meer uitgesproken was bij patiënten met een laat stadium, die in aanmerking kwamen voor een knievervangende operatie3,25,26, sommige beweerden het tegenovergestelde6,27 en weer anderen konden geen verschil vinden28. Echter om volledig de rol van synovitis in het ziekteverloop van artrose te kunnen begrijpen, is het van belang om te weten of synovitis voorafgaat aan de structure afwijkingen bij artrose of dat synovitis een reactie is dat wordt veroorzaakt door het artrose proces. Toekomstig onderzoek naar de immuuncellen in de verschillende ziektestadia van artrose zou hierbij van toegevoegde waarde kunnen zijn, aangezien hiermee duidelijk zou worden welke immuuncellen het synovium in een vroeg stadium van de ziekte infiltreren en in welk stadium van de ziekte ze worden geactiveerd. Synovitis in de knie gezien bij artrose patiënten is geen homogeen proces, maar treft sommige anatomische locaties frequenter dan andere anatomische locaties (hoofdstuk 6). Dit suggereert dat synovitis bij artrose een reactie is op locale triggers en geen gegeneraliseerde ontstekingsreactie betreft. Mogelijke locale triggers zouden kraakbeen afbraakproducten, trauma van verschillende structuren in de knie of vergrote belasting van de knie kunnen zijn. In dit proefschrift vonden we dat synovitis het meest ernstig was in het eindstadium van knieartrose zowel als we keken naar histologische ontsteking in biopten als wanneer we synovitis observeerde op MRI met contrast (hoofdstuk 5, hoofdstuk 6). Echter, wanneer we het synovitis beloop over 2 jaar onderzochten, vonden we dat synovitis over de tijd fluctueerde en niet geleidelijk aan toeneemt over de tijd (hoofdstuk 8). Deze resultaten suggereren dat de mate van ontsteking niet lineair stijgende lijn volgt, maar vaker aanwezig is bij patiënten met een eindstadium van artrose. Een mogelijke verklaring hiervoor is dat synovitis een procesmarker is, die werkt als een katalysator in het artrose proces in plaats van dat het de oorzaak is van structurele schade. Dit is in tegenstelling met eerdere literatuur die suggereert dat synovitis vooraf gaat aan kraakbeenschade6. Het verschil met onze bevindingen kan deels verklaard worden door het feit dat er geen consensus is over de definitie van vroege artrose. 184

Nederlandse samenvatting

De rol van synovitis in structurele schade bij artrose Een belangrijk aandachtspunt in het wetenschappelijk artroseonderzoek is het ontstaan en de progressie van structurele schade. Daarom hebben we in dit proefschrift extra aandacht besteed aan de relatie tussen synovitis en radiologische schade in cross-sectionele analyse (hoofdstuk 4-6) en aan de relatie tussen synovitis en structurele artrose progressie in longitudinaal onderzoek (hoofdstuk 7 en 8). Zoals we leerden van hoofdstuk 5 en 6, is synovitis geassocieerd met radiografische schade, alhoewel effusie / synovitis van minder groot belang leek te zijn bij radiografische progressie als we ook rekening hielden met andere afwijkingen in de knie middels een PCA (hoofdstuk 7). Het ontbreken van een relatie tussen effusie / synovitis en radiografische progressie in hoofdstuk 7 kan deels verklaard worden door het feit dat hier geen contrast werd gebruikt en deels doordat synovitis/effusie wel onderdeel was van de componenten die werden geassocieerd met progressie, maar net niet de waarde van het afkap punt haalde. Een interessante observatie was dat we in hoofdstuk 8 vonden dat verergering van synovitis over de tijd geassocieerd was met verergering van kraakbeenschade over de tijd. Het blijft echter nog steeds onbekend of synovitis hierin fungeert als een versterker van het artrose proces of vooraf gaat aan structurele schade. De resultaten in dit proefschrift suggereren wel dat synovitis een onderdeel is van de pathofysiologische mechanismen dat leidt tot structurele schade in knieartrose en dat synovitis daarom een aangrijpingspunt kan zijn voor ziekte-modificerende medicatie in de toekomst. Malalignment van de knie in knieartrose Malalignment van de knie (bijvoorbeeld varus of valgus deformiteit) wordt frequent als oorzaak van verschillende knieartrose processen gesuggereerd. De in dit proefschrift gevonden associatie van verschillende patronen van MR afwijkingen met radiografische progressie in hoofdstuk 7 kunnen deels verklaard worden door varus deformiteit, die frequent bij artrose voorkomt. Van bestaande varus deformatie is aangetoond dat het een risicofactor is voor de ontwikkeling van BMLs in de mediale tibiofemorale gewricht29-31 en leidt tot artrose progressie in het mediale compartiment van het tibiofemorale gewricht31,32. Verder zorgt een varus deformatie voor een toename van de Q-hoek van de patella. Deze toename resulteert in een mediaal werkende kracht en verhoogde belasting op het mediale compartiment van de patella met als uiteindelijk resultaat progressie van structurele schade in het patellofemoraal gewricht33-35. Van een bestaande valgus deformatie is aangetoond dat het een risicofactor is voor de ontwikkeling en progressie artrose in het laterale compartiment van het tibiofemorale gewricht36. Concluderend lijkt malalignment derhalve van belang te zijn bij het begrijpen van de relatie tussen verschillende gewrichten in de knie (tibiofemoraal gewricht en patellofemoraal gewricht) en het begrijpen van progressie in knieartrose. Het lijkt denkbaar dat malalignment de belasting in de knie veranderd wat

185

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 10

leidt tot progressie van structurele schade en/of toename van synovitis, wat mogelijk de gevonden patronen in hoofdstuk 6 verklaard. Synovitis als determinant van pijn Eerdere literatuur suggereert dat synovitis een belangrijke determinant van pijn is. Daarom is een groot deel van dit proefschrift gericht op het onderzoeken van de associatie tussen pijn en synovitis (hoofdstuk 3-6 en hoofdstuk 8). Een interessante bevinding was dat synovitis bij de patella meer subjectieve pijn gaf dan synovitis op andere locaties (hoofdstuk 6). Uit eerdere studies naar de innervatie en pijnsensatie van de knie bleek dat het gebied van de patella rijkelijk is geïnnerveerd en anterieur synoviaal weefsel in de nabijheid van de patella zeer gevoelig is voor pijn stimulatie37,38. Deze bevindingen suggereren een locatiespecifieke rol van synovitis op subjectieve pijnbeleving en hiermee ook mogelijk een rol voor locale behandeling van synovitis en pijn. Alhoewel wij in dit proefschrift associaties van synovitis met pijn vonden (hoofdstuk 3, 5 en 6), waren de effect-sizes over het algemeen niet erg hoog. Verder vonden we in hoofdstuk 8 dat verandering van synovitis alleen geassocieerd was met verandering van de pijn in het niet-gecorrigeerde model, terwijl kraakbeen achteruitgang wel was geassocieerd met een verandering in pijn. Dit laatste zou eventueel verklaard kunnen worden door het feit dat kraakbeen achteruitgang een mediator is in de associatie van synovitis met pijn. Concluderend, lijkt synovitis bij te dragen, maar niet de enige factor te zijn in het onderliggende biologische mechanisme van pijnbeleving in knieartrose patiënten. Een andere verklaring voor de lage effect-sizes ligt in het feit dat subjectieve pijn en in het bijzonder chronische pijn niet alleen wordt bepaald door biologische processen, maar wordt ook door psychologische en sociale processen, die allen tezamen een biopsychosociaal pijnmodel vormen39-41.

TOEKOMSTPERSPECTIEVEN We begonnen dit proefschrift in hoofdstuk 2 met een onderzoeksagenda voor het onderzoeken van synovitis. We sluiten dit proefschrift af met een suggestie voor een nieuwe onderzoeksagenda voor toekomstig onderzoek gericht op het aanvullen van de bestaande lacunes in de huidige kennis over de rol die synovitis speelt bij knieartrose, die uiteindelijke zullen leiden tot de ontwikkeling van ziektemodificerende medicatie. - Consensus over de definitie van ziektestadia bij artrose Om betrouwbaar te kunnen onderzoeken of synovitis voorafgaat aan knieartrose of dat synovitis een procesmarker is van het ziekteproces, is het noodzakelijk om een eenduidige definitie van artrose te hebben. Het bereiken van een consensus over de definitie van verschillende ziektestadia bij knieartrose is hierbij van groot belang, zodat toekomstige

186

Nederlandse samenvatting

wetenschappelijke resultaten beter kunnen worden vergeleken. Op dit moment wordt er hard gewerkt om een consensus voor de definitie van “vroege artrose” te bereiken. - Inzicht in de evolutie van synovitis gedurende het artrose ziektebeloop Om de rol van ontsteking in het ziekteverloop van artrose beter te begrijpen, zijn meer longitudinale gegevens van meer patiënten over langere tijd noodzakelijk. Bovendien, zou het interessant zijn om de immuuncellen in synoviale biopten van dezelfde patiënten op verschillende tijdstippen in het artrose ziektebeloop te onderzoeken. Dit zou inzicht geven in de immuuncel infiltratie en hun activatiepatroon gedurende de ziekte, wat kan leiden tot een beter begrip van de cellulaire mechanismen die betrokken zijn bij ziekteprogressie. - Verder onderzoek naar de rol van CD4 + T-cellen pijn In hoofdstuk 3 vonden we een mogelijke associatie van CD4 + T-cellen met pijn. Daarom zou het om de rol van T-cellen in pijnperceptie verder te onderzoeken, interessant zijn om te onderzoeken of de T-cellen zich ophopen in de nabijheid van de zenuwuiteinden. Aangezien we ook vonden dat histologische synoviale ontsteking samenhangt met ontsteking gezien op MRI (hoofdstuk 5), zou het tevens interessant zijn om te zien of de T-cellen zich vooral ophopen op de anatomische plaatsen die het patroon vormen dat in hoofdstuk 6 geassocieerd is met pijn. - Verder onderzoek naar de rol van mestcellen als mogelijk aangrijpingspunt voor therapie bij artrose De in dit proefschrift gepresenteerde data suggereren een rol voor mestcellen in het artrose ziekteproces. Toekomstig onderzoek zou zich kunnen richten op een betere karakterisering van mestcellen bij artrosepatiënten, bijvoorbeeld door middel van een grootschalige mRNA sequentie. Dit zou namelijk meer inzicht kunnen bieden in de mogelijke effector moleculen die mestcellen tot expressie brengen en in de mogelijke activatie mechanismen van mestcellen in artrose. Verder onderzoek naar de oplosbare mediators die worden uitgescheiden door geactiveerde mestcellen en hun effect op het bot en kraakbeen, kan bijdragen aan ons begrip van de rol van mestcellen in de pathogenese van knieartrose. Ten slotte moet het gebruik van ziektemodificerende medicatie die zich richt op mestcellen verder worden onderzocht in preklinische en klinische artrosestudies. - Synovitis op MRI met contrast bij andere inflammatoire ziekten Alhoewel synovitis op MRI met contrast goed overeen komt met de histologische ontsteking bij knieartrose patiënten, is het nog niet bekend of MRI van contrast van toepassing zou kunnen zijn bij de differentiële diagnostiek van inflammatoire ziekten in de klinische praktijk. Daarom zou het interessant zijn om te zien of het synovitis scoresysteem door Guermazi et al. zou kunnen worden toegepast in RA of andere inflammatoire ziekten. - Alternatieven voor de visualisatie van synovitis Alhoewel MRI met contrast een niet-invasieve methode is waarmee we synovitis in geMstoan studie hebben onderzocht, kon een deel van de patiënten geen MRI ondergaan in

187

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 10

verband met contra-indicaties voor de MRI. Aangezien artrose meer voorkomt naarmate de leeftijd vordert is de kans op het bestaan van contra-indicaties in deze populatie groter dan in andere gewrichtsziekten zoals RA of SpA. Daarom is het van belang om ook andere nietinvasieve methoden te onderzoeken die synovitis betrouwbaar zichtbaar kunnen maken. Een veelbelovend alternatief is echografie van de knie. Echo onderzoek kan niet alleen synovitis in beeld brengen, maar heeft ook de mogelijkheid om andere afwijkingen zoals osteofyten en kraakbeenschade in meerdere vlakken in beeld te brengen. Pas onlangs werd er een echo scoringsmethode ontwikkeld, die goed overeenkwam met de artroseafwijkingen die op een rĂśntgenfoto werden gezien42. Het zou interessant zijn om te zien hoe goed deze echoscore associeert met artroseafwijkingen op MRI en dan met name hoe synovitis op de echo associeert met synovitis op MRI met contrast. Als de echo een valide alternatief blijkt te zijn voor MRI met contrast zal dit een mogelijkheid bieden voor ervaren echografisten om synovitis in de klinische praktijk synovitis te vervolgen. â&#x20AC;&#x2039; - Beter inzicht in de rol van biomechanica bij knieartrose Biomechanica lijkt vooral in de knie van belang te zijn bij de ontwikkeling van structurele schade en geeft een verklaring voor het feit dat structurele schade in de verschillende gewrichten in de knie (het tibiofemoraal gewricht en het patellofemoraal gewricht) zijn gerelateerd. Onderzoek van de associatie van patronen van MRI afwijkingen met malalignment van de knie, kan hierin meer inzicht geven. Bovendien zou varus of valgus deformiteit mogelijk ook de gevonden patronen van synovitis kunnen verklaren. Verder onderzoek naar de rol biomechanica bij het ontstaan van structurele afwijkingen en synovitis zou dus interessant zijn.

188

Nederlandse samenvatting

REFERENCES

189

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Chapter 10

190

APPENDIX ARTROSEGROEP REUMATOLOGIE VALT IN DE PRIJZEN PUBLICATIELIJST

CURRICULUM VITAE

DANKWOORD

Appendix

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 192

Artrosegroep reumatologie valt in de prijzen

ATROSEGROEP REUMATOLOGIE VALT IN DE PRIJZEN de Lange-Brokaar BJE Nederlands Tijdschrift voor Reumatologie. 2013 (4), 45-47

& 193

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Appendix

Drs. Badelog de Lange-Brokaar kreeg in juni 2013 in Madrid een EULAR Abstract Award in Clinical Science uitgereikt, voor haar abstract: Different patterns of synovitis present in osteoarthritis patients associate differentially with pain. Dit is al de tweede keer dit jaar dat zij in de prijzen viel, want in april kreeg zij voor haar werk naar synoviitis bij artrose de Young Investigators Award in Philadelphia, Verenigde Staten op het jaarlijkse Osteoarthritis Research International congres (figuur 1).

Figuur 1. Badelog de Lange (OARSI 2013): voorste rij, tweede van rechts.

Synoviitis bij artrose Hoewel lange tijd werd gedacht dat artrose een puur degeneratieve ziekte is zonder synoviitis, blijkt uit recent onderzoek dat synoviitis wel van belang is bij mensen met artrose. Zo zijn er aanwijzingen dat het een rol speelt bij de pathofysiologie van artrose, voorspelt synoviitis kraakbeenschade op de lange termijn en is het geassocieerdmet pijn1-4. De gouden standaard voor het vaststellen van synoviitis is nog steeds het beoordelen van tekenen van ontsteking op weefselniveau (histologie) door middel van het nemen van biopten of het zien van macroscopische tekenen van ontsteking tijdens een artroscopie. Omdat we natuurlijk niet iedereen kunnen scopieĂŤren om te kijken of er sprake is van synoviitis hebben we in eerder onderzoek gekeken of synoviitis met MRI met contrast accuraat kan worden vastgesteld. We hebben gekozen voor MRI met contrast omdat hiermee een duidelijk onderscheid gemaakt kan worden tussen effusie en synoviitis (figuur

194

Artrosegroep reumatologie valt in de prijzen

2). Het vergelijken van de hoeveelheid synoviitis op MRI met contrast met kenmerken van ontsteking op weefselniveau (histologie) en macroscopische kenmerken (geobserveerd tijdens artroscopie) hebben laten zien dat synoviitis gemeten op MRI met contrast een goede maat is voor synoviitis5.

Figuur 2. Patiënt met patellofemorale osteoartrose, met ernstig kraakbeenverlies aan de laterale zijde, Osteofyten en zowel effusie als synoviitis, waarbij het verschil tussen de twee goed te zien is bij T1W gadoliniumopname.

Synoviitis in osteoartrose onregelmatig aspect Als we synoviitis bekijken op MRI met contrast zien we dat de het niet op alle plekken in de knie evenveel aanwezig is, maar dat er op sommige plekken in de knie meer ontsteking zit dan op andere plekken. Daarom is het van belang een scoringsmethode te hebben die op verschillende plekken in de knie de synoviitis scoort. Zo’n methode is de scoringsmethode van Guermazi et al die synoviitis scoort op basis van dikte van het door contrast opgelichte deel van het synovium van 0-2 op 11 verschillende plekken door de hele knie6. Dat synoviitis bij artrosepatiënten een onregelmatig aspect heeft was bekend. Het is echter minder duidelijk of dit een lokaal proces is waarbij ontsteking in de knie onafhankelijk van de locatie voorkomt, of dat er patronen van ontsteking (bepaalde ontstekingsplekken in de knie die samen gaan) zijn te ontdekken. Daarom hebben wij gekeken of er sprake is van patronen van synoviitis bij patiënten met knieartrose en als deze bestaan of die patronen klinisch relevant zijn (correlatie van patronen met pijn en schade).

& 195

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Appendix

Onderzoek naar patronen van synoviitis in de knie Om dit goed te kunnen onderzoeken hebben we bij patiĂŤnten met symptomatische knieartrose T1-gewogen MRI-opnames met contrast gemaakt. Vervolgens hebben we in de hele groep gekeken of er sprake was van patronen van synoviitis door alle plekken met synoviitis te analyseren met een principal component analysis (PCA). PCA is een methode van analyseren die verschillende groepen (componenten) probeert te maken op basis van correlaties die bestaan tussen de verschillende synoviitisplekken in de knie. Vervolgens hebben we bekeken of deze patronen associeerden met uitkomstmaten als pijn en radiografische schade. Uit onze analyses bleek dat er sprake was van drie verschillende synoviitispatronen die in onze populatie met knieartrose konden worden gevonden (figuur 3). Het eerste patroon bestond uit zeven verschillende plekken die vooral door synoviitis aan de mediale zijde van de patella werd gedefinieerd. Dit patroon associeerde met pijn. Het tweede patroon bestond uit synoviitis op vier verschillende plekken en werd met name gedefinieerd door synoviitis bij de voorste kruisband en synoviitis ter plaatse van de mediale meniscus. Er werden geen associaties met klinische uitkomsten gezien. Het derde patroon bestond uit drie plekken van synoviitis waarbij synoviitis bij loose body (botfragmenten) het grootste aandeel had. Deze laatste component associeerde met radiografische schade, maar niet met pijn. Met dit onderzoek hebben we aangetoond dat er verschillende patronen zijn van ontsteking en dat deze patronen verschillende klinische relevanties hebben. Hoe deze patronen ontstaan en wat het mechanisme is achter de gevonden associatie met pijn wordt op dit moment onderzocht.

196

Artrosegroep reumatologie valt in de prijzen

Figuur 3. Patronen van synoviitis. Patroon 1: synoviitis bij mediale zijde patella, laterale zijde patella, suprapatellaire locatie, infrapatellaire locatie, ter plaatste van laterale meniscus, synoviitis bij een bakercyste en ter plaatste van de achterste kruisband. Patroon 2: synoviitis ter plaatste van de voorste kruisband, ter plaatste van de mediale meniscus, bij suprapatellaire locatie en ter plaatse van het hoffavetlichaam. Patroon 3: synoviitis bij â&#x20AC;&#x2DC;loose bodyâ&#x20AC;&#x2122; (botfragment), bij suprapatellaire locatie en synoviitis ter plaatste van de achterste kruisband.

& 197

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Appendix

REFERENCES

198

1 Sellam J and Berenbaum F. The role of synovitis in pathophysiology and clinical symptoms of osteoarthritis. Nat Rev Rheumatol 2010;6:62535. 2 Crema MD, Felson DT, Roemer FW, Niu J, Marra MD, Zhang Y et al. Peripatellar synovitis: comparison between non-contrast-enhanced and contrast-enhanced MRI and association with pain. The MOST study. Osteoarthritis Cartilage 2013;21:413-8. 3 Yusuf E, Kortekaas MC, Watt I, Huizinga TW, Kloppenburg M. Do knee abnormalities visualised on MRI explain knee pain in knee osteoarthritis? A systematic review. Ann Rheum Dis 2011;70:60-7. 4 Roemer FW, Guermazi A, Felson DT, Niu J, Nevitt MC, Crema MD et al. Presence of MRI-detected joint effusion and synovitis increases the risk of cartilage loss in knees without osteoarthritis at 30-month follow-up: the MOST study. Ann Rheum Dis 2011;70:1804-9. 5 de Lange-Brokaar BJ, Ioan-Facsinay A, Yusuf E, Visser AW, Kroon HM, Andersen SN et al. Degree of synovitis on MRI by comprehensive whole knee semi-quantitative scoring method correlates with histologic and macroscopic features of synovial tissue inflammation in knee osteoarthritis. Osteoarthritis Cartilage 2014;22:1606-13. 6 Guermazi A, Roemer FW, Hayashi D, Crema MD, Niu J, Zhang Y et al. Assessment of synovitis with contrast-enhanced MRI using a wholejoint semiquantitative scoring system in people with, or at high risk of, knee osteoarthritis: the MOST study. Ann Rheum Dis.2011;70:805-11.

Publicatielijst

PUBLICATION LIST de Lange-Brokaar BJE, Ioan-Facsinay A, van Osch GJVM, Zuurmond AM, Schoones J, Toes REM, Huizinga TWJ, Kloppenburg M. Synovial inflammation, immune cells and their cytokines in osteoarthritis: a review. Osteoarthritis Cartilage 2012;20:1484-99. Klein-Wieringa IR, Andersen SN, Kwekkeboom JC, Giera M, de Lange-Brokaar BJE, van Osch GJVM, Zuurmond AM, Stojanovic-Susulic V, Nelissen RG, Pijl H, Huizinga TWJ, Kloppenburg M, Toes REM, Ioan-Facsinay A. Adipocytes modulate the phenotype of human macrophages through secreted lipids. J Immunol 2013 ;191:1356-63. de Lange-Brokaar BJE, Ioan-Facsinay A, Yusuf E, Visser AW, Kroon HM, Andersen SN, Herbvan Toorn L, van Osch GJVM, Zuurmond AM, Stojanovic-Susulic V, Bloem JL, Nelissen RGHH, Huizinga TWJ, Kloppenburg M. Degree of synovitis on MRI by comprehensive whole knee semi-quantitative scoring method correlates with histologic and macroscopic features of synovial tissues inflammation in knee osteoarthritis. Osteoarthritis Cartilage 2014;22:160613. de Lange-Brokaar BJE, Ioan-Facsinay A, Yusuf A, Visser AW, Kroon HM, van Osch GJVM, Zuurmond AM, Stojanovic-Susulic V, Bloem JL, Nelissen RGHH, Huizinga TWJ, Kloppenburg M. Association of pain in knee osteoarthritis with distinct patterns of synovitis. Arthritis Rheumatol 2015;6:733-40. de Lange-Brokaar BJE. Artrosegroep reumatologie valt in de prijzen. Nederlands Tijdschrift voor Reumatologie 2013;4:45-47. Klein-Wieringa IR*, de Lange-Brokaar BJE*, Yusuf E, Andersen SN, Kwekkeboom JC, Kroon HM, van Osch GJVM, Zuurmond A-M, Stojanovic-Susulic V, Nelissen RGHH, Toes REM, Kloppenburg M, Ioan-Facsinay A, Inflammatory cells in end stage knee osteoarthritis patients: a comparison between the synovium and the infrapatellar fat pad (IFP). Submitted. * These authors contributed equally to the manuscript de Lange-Brokaar BJE, Kloppenburg M, Andersen SN, DorjĂŠe AL, Yusuf E, Herb-van Toorn L, Kroon HM, Zuurmond AM, Stojanovic-Susulic V, Bloem JL, Nelissen RGHH, Toes REM, IoanFacsinay A. Characterisation of synovial mast cells in knee osteoarthritis: association with clinical parameters. Submitted.

199

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Appendix

de Lange-Brokaar BJE, Bijsterbosch J, Kornaat PR, Yusuf E, Ioan-Facsinay A, Zuurmond A-M, Kroon HM, Meulenbelt I, Bloem JL, Kloppenburg M. Radiographic progression of knee osteoarthritis is associated with MRI abnormalities in both the patellofemoral and the tibiofemoral joint. Submitted. de Lange-Brokaar BJE, Ioan-Facsinay A, Yusuf E, Kroon HM, Zuurmond AM, StojanovicSusulic V, Nelissen RGHH, Bloem JL, Kloppenburg M. Evolution of synovitis in osteoarthritic knees and its association with clinical features. Submitted.

200

Curriculum Vitae

CURRICULUM VITAE Badelog de Lange-Brokaar werd geboren op 27 maart 1980 te Zwijndrecht. Na haar eindexamen VWO startte zij in 2001 met haar opleiding geneeskunde aan de Universiteit Utrecht. Naast haar studie was zij vooral actief in het beoefenen van Acrobatische Rock & Roll en in 2005 werd zij samen met haar partner Nederlands Kampioen. Na haar artsenexamen in 2009 heeft zij één jaar gewerkt als arts-assistent revalidatiegeneeskunde (ANIOS & AIOS) in het VU Medisch Centrum te Amsterdam. In 2010 startte zij haar promotietraject op de afdeling reumatologie van het Leids Universitair Medisch Centrum (LUMC) onder begeleiding van prof. dr. Margreet Kloppenburg. Tijdens haar onderzoeksperiode won zij de ‘EULAR Abstract Award in Clinical Science 2013’, de ‘Young Investigator Award OARSI 2013’ en meerdere reisbeurzen. Tevens werd zij in 2013 voor haar wetenschappelijke prestaties genomineerd voor ‘VIVA vrouw 400, knappe koppen’. In 2014 volgde zij 6 maanden de vooropleiding interne geneeskunde in het kader van haar opleiding tot reumatoloog in het Bronovo Ziekenhuis te Den Haag. Op dit moment werkt zij in het Spine and Joint Centre, een poliklinisch revalidatiecentrum voor patiënten met nek-, rug- en bekkenklachten in Rotterdam.

& 201

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Appendix

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39 202

Dankwoord

DANKWOORD Hierbij wil ik iedereen bedanken die heeft bijgedragen aan de tot standkoming van dit proefschrift. Ik wilde een aantal mensen in het bijzonder bedanken. Allereerst wil ik alle deelnemers van de geMstoan studie bedanken voor hun deelname. Jullie ervaringsdeskundigheid is met geen pen te beschrijven. Graag wilde ik mijn promotor en copromotor bedanken. Geachte prof.dr Kloppenburg, beste Margreet, jij liet mijn wetenschappelijke interesse verder opbloeien. Ik heb veel geleerd van jouw artrose expertise. Geachte dr. Ioan- Facsinay, beste Andreea, bedankt voor je begeleiding wat betreft de lab georiënteerde stukken en je poëtische zinnen. Natuurlijk wil ik ook de overige leden van de “artrose groep” bedanken voor de nuttige en gezellige werkbesprekingen en congressen: Wing-Yee, Marion, Anja, Wendy, Rani, Willemien, Angga, Linda en Inge. Inge, bedankt voor de fijne samenwerking tijdens ons artikel. Rani, jouw humoristische kijk op de wetenschap en het leven verveelt nooit. Willemien, bedankt voor al het nuttige overleg over knieartrose en MRI. I would also like to thank the other members of TI-Parma consortium (Gerjo, Annemarie, Yvonne, Rene, Lobke, Patricia, Angela, Vedrana, Stefan, Wu) for the many constructive meetings (live and teleconference) and for the useful discussions. Geachte prof.dr. Huizinga, beste Tom, bedankt dat ik op jouw afdeling mocht promoveren. Alle reumatologen, arts-assistenten en onderzoeksverpleegkundigen van de afdeling reumatologie en orthopedie van het LUMC bedank ik voor hun hulp bij de inclusie van patienten in de geMstoan studie. In het bijzonder wil ik prof.dr. Nelissen bedanken voor zijn nuttige commentaar op vele artikelen beschreven in dit proefschrift. Verder wil ik Emilie Jonxis en de orthopedische chirurgen van het Diaconessenhuis bedanken voor de prettige samenwerking. Mijn nog niet genoemde kamergenootjes van C1-45 en C1-46 wil ik alle bedanken voor nuttige discussies en gezelligheid. Marianne, een avontuurlijkere overbuurvrouw kon ik mij niet wensen. Lotte, dank je wel voor de gezelligheid zowel op als naast het werk. Wouter, bedankt voor het mij attent maken op de “1 hour loop” (aanrader voor lange scoorsessies en afronding van het proefschrift). Onze data managers, Jozé en Cedric, wil ik bedanken voor de goede zorg voor de geMstoan data en de hulp bij niet functionerende computers. De dames van het secretariaat wil ik allen bedanken voor de ondersteuning in bredere zin. Hughine, dank je voor het behapbaar maken van de promotieregelementen. Het laboratorium van de afdeling reumatologie wil ik bedanken voor de hulp bij het vinden van mijn weg in het Lab. In het bijzonder wil ik Linda, Stefan, Joanneke en Annemarie bedanken voor hun hulp bij de laboratorische experimenten beschreven in dit proefschrift.

203

R1 R2 R3 R4 R5 R6 R7 R8 R9 R10 R11 R12 R13 R14 R15 R16 R17 R18 R19 R20 R21 R22 R23 R24 R25 R26 R27 R28 R29 R30 R31 R32 R33 R34 R35 R36 R37 R38 R39

Appendix

Stefan en Annemarie, ook bedankt voor jullie hulp en inzet bij het mestcellen project. Annemarie, dankjewel voor je energie zowel op als naast het werk. Verder wil ik alle niet genoemde medewerkers bedanken van het zowel het stafcentrum en van de polikliniek reumatologie. In het bijzonder wil ik de secretaresses van de poli bedanken voor het altijd vinden van een kamertje. De afdeling radiologie wilde ik bedanken voor hun ondersteuning bij het radiologische deel van dit proefschrift. Een aantal mensen wil ik in het bijzonder bedanken: Geachte prof.dr. Bloem, bedankt dat u altijd mijn vragen over de MRI scanparameters wilde beantwoorden. Beste Addy, bedankt dat je al die jaren de geMstoan röntgenfoto’s wilde maken. Geachte dr. Kroon, beste Herman, bedankt voor het scoren van de gemaakte röntgenfoto’s op de vroege ochtend. Lieve Tina, bedankt voor de gezellige scanavonden, je hulp bij het halen van mijn scanbrevet en je geduld bij het controleren van mijn scaninstellingen. Paul, bedankt voor je hulp bij alle werkgerelateerde en niet werkgerelateerde technische zaken. In het bijzonder bedank ik mijn paranimfen, Rosaline en Marieke. Rosaline, bedankt voor de tips en checklijsten voor de laatste loodjes van de promotie en de “groeiende” vriendschap. Marieke, dank je wel voor je steun en vele gezellige etentjes waarin “carriére maken” regelmatig op het menu staat. Ik dank mijn familie en vrienden, in het bijzonder mijn (schoon)ouders en oma Bep, voor hun interesse en begrip voor mijn afwezigheid tijdens vele sociale gelegenheden. Dear Allan, you may not know this, but you inspired me to become a doctor. Tenslotte wil ik mijn gezin bedanken voor hun liefde en geduld tijdens eindeloze avonden en weekenden “werken op zolder”. Janneck, dank wel dat je mij accepteert zoals ik ben en dat je mij steunt in het volgen van mijn dromen.

204

Synovial Inflammation In Knee Osteoarthritis

Synovial Inflammation In Knee Osteoarthritis Histological and Imaging Studies

Uitnodiging Voor het bijwonen van de openbare verdediging van het proefschrift

Synovial Inflammation In Knee Osteoarthritis Histological and Imaging studies

Paranimfen: Rosaline van den Berg r.van_den_Berg@lumc.nl Marieke Twickler mtwickler@hotmail.com

B.J.E. de Lange-Brokaar

Badelog Jeanine Elise de Lange-Brokaar