Annals ofthe Rheumatic Diseases 1995; 54: 511-515 51

Synovial membrane cellularity and vascularity Ann Rheum Dis: first published as 10.1136/ard.54.6.511 on 1 June 1995. Downloaded from

Oliver FitzGerald, Barry Bresnihan

Studies of the in vascularity is seen, depending on the type of inflammation, over many years, have proven of membrane surface. mixed benefit: while it is rarely of diagnostic The -pannus junction, important in value, clues to understanding the mechanisms the development of joint erosion, has also been involved both in the initiation and main- studied carefully in normal samples. Allard tenance of the inflammatory response have et al analysed the cellular composition both of been uncovered. While the path to under- the cells adjacent to and of the cells standing is fraught with difficulty, progress is contained in the layer of tissue which extends slowly being made and, it is hoped, will lead over and merges with the cartilage, lining the in turn to novel and better treatment joint cavity.7 Adjacent to bone, the cells stain approaches. This review of synovial membrane predominantly with the monocyte/ cellularity and vascularity will describe our marker 63D3; in contrast, those cells lining the current understanding of the pathways to joint joint cavity express the monoclonal antibody damage. 67, thought to be a marker for material surrounding type B synoviocytes.8 Normal celiularity and vascularity Normal synovial tissue is characterised by a Synovial membrane in early rheumatoid surface layer of cells, or intima, below which is a loose which contains fibro- Most of the studies of blasts, , adipocytes, mast cells, (RA) synovial membrane have been performed nerve fibres, vascular endothelial cells, and on samples obtained from patients with long occasional polymorphs and .' The established disease, largely at arthroplasty. intima contains two or possibly three types. There have only been a few studies in early Type A cells are of bone marrow origin and disease, when samples are more likely to reflect conform to the ultrastructural criteria for factors important in disease initiation rather

monocyte derived macrophages. They express than those contributing to chronicity. http://ard.bmj.com/ macrophage markers including CD68 and Schumacher and Kitridou described 24 CD 14 and also stain with non-specific patients with of recent onset, six of esterase.2 Type B cells are fibroblastic in type whom developed RA.9 They described variable and can be distinguished cytochemically and (< 10 cells in depth) lining cell proliferation immunochemically from deeper in with largely a perivascular lymphocytic the tissues. Being involved in collagen infiltration. There were prominent vascular the synthesis, they express ,B subunit of the changes, but no lymphoid follicles and only on September 27, 2021 by guest. Protected copyright. enzyme prolyl hydroxylase.3 They are distinct occasional polymorphs were observed. The in terms of their ability to express uridine intensity of the inflammation did not appear to diphosphoglucose dehydrogenase (UDPGD) correlate with the disease duration. Lindblad which is an important enzyme involved in et al also studied arthroscopic biopsies from a hyaluronan synthesis-a specialised synovial small number of RA patients with disease membrane function.3 The constitutive duration between nine months and five years.'0 expression of vascular cell molecule Again, lining layer thickening was observed, (VCAM)-1 by type B synoviocytes is also an with cells positive for both OKM1 and OKT9, important distinguishing feature not shared by suggestive ofmacrophage infiltration and acute fibroblastic cells elsewhere in the synovial proliferation. The sublining cells were largely membrane.4 The exact function of this of the helper T phenotype and VCAM- 1 expression remains a matter of focal aggregates of B cells were also seen. speculation, but it suggests a role in trapping More recent studies from our own group of cells within the lining layer. The third type of 11 RA patients with a mean disease duration intimal cell is dendritic in type, accounts for of 22 months have quantified the cellular Department of Rheumatology, < 1% of the total, is CD68 negative but infiltrate." Again, variable lining layer thick- St Vincent's Hospital expresses the class II related marker, RFDIL. ening was seen (mean 4 cells in depth; and University Studies on synovial vascularity date back to range 2-32). In the sublining, 10 of the 11 College, Dublin 1743, when Hunter described the plexus of patients had a diffuse mononuclear infiltrate 0 FitzGerald vessels which supplied the synovial membrane comprising predominantly CD 14 macro- B Bresnihan and which he called the circulus articuli phages and CD5 T lymphocytes, with roughly Correspondence to: vasculosus. A more recent detailed study of equal numbers of CD4 and CD8 cells. Small Dr 0 FitzGerald, Department of normal synovial membrane has highlighted the numbers of B cells and plasma cells were also Rheumatology, high frequency of capillaries, with a peak seen. Thus lining layer thickening, vascular St Vincent's Hospital, Elm Park, density just below the lining layer between 6 proliferation and a diffuse sublining infiltrate Dublin 4, Ireland. and 11 ,um deep.6 Considerable variation in made up of macrophages, T lymphocytes, 512 FitzGerald, Bresnihan

more CD4 than CD8, and a small number of with a shorter duration of disease, it may be B cells, appear to be typical features of the that E-selectin expression is more upregulated involved synovial membrane in early RA. in early disease, ICAM-1 then correlating with Ann Rheum Dis: first published as 10.1136/ard.54.6.511 on 1 June 1995. Downloaded from The examination of synovial membrane the intensity of the inflammatory response. obtained from clinically uninvolved is another approach to exploring changes early in the disease. The clinically uninvolved joint is Synovial membrane in established RA histologically involved, though there is less Most studies of RA synovial membrane have lining layer hyperplasia and fewer mononuclear been performed on samples obtained at joint cells infiltrate the sublining compared with the arthroplasty. While the immunohistological involved joint.'2 In a further study, a vascular findings in established disease are not dis- proliferation index and a high endothelial similar from those seen earlier on, the mono- venule (HEV)-index were calculated both in nuclear infiltrate is focal in about 70%, the clinically involved and clinically uninvolved compared with about 10% ofbiopsy samples.20 RA samples, in addition to postmortem Duke et al described four different patterns of controls.'3 Of interest, vascular proliferation infiltrate: scattered/diffuse, perivascular, lym- was not a feature of the uninvolved joint, and phoid follicles and germinal centres.2' HEV-type vessels, readily seen in actively Kurosaka and Ziff went on to study the infil- inflamed joints, were also not observed. This trates in more detail, and described lympho- suggests that vascular changes, which may cyte rich areas where most cells stained positive determine the rate of mononuclear trafficking, for OKT4, and more peripheral transitional are important in the clinical expression of joint areas containing plasma cells, macrophages, inflammation in early RA. and both OKT4 positive and OKT8 positive lymphocytes.22 More recently, Yanni et al analysed lymphoid infiltrates of different sizes Adhesion molecule expression in RA and found that the composition of each aggre- The receptor-ligand pairs important in the gate depended on the total number of mono- control of mononuclear trafficking into the nuclear cells it contained:23 large aggregates synovial membrane have been the subject of a contained a substantial number of B cells and number of recent studies.'1'6 The important cells bearing the CD45RA phenotype; in adhesion molecules expressed on the endo- contrast, small aggregates contained few B cells thelium include E-selectin, which plays a and relatively larger numbers ofT cells bearing role in the initial margination and rolling of CD8 and CD45RO. leucocytes, and both intracellular adhesion Lining layer thickening is also a prominent molecule-i (ICAM-1) and VCAM-1, which feature of established RA. Considerable are involved in leucocyte adhesion and controversy exists as to whether this thickening

penetration through the endothelial junctions. is a result of macrophage recruitment or of http://ard.bmj.com/ Studies ofRA synovial membrane by our group synoviocyte proliferation. In favour of recruit- and others have shown E-selectin expression to ment, mitotic cells are rarely seen, and staining be vessel specific, the majority of vessels for the proliferation marker Ki-67 is not staining positive, particularly in the superficial observed in RA synovial membrane.24 In synovial membrane.'7 ICAM-1 expression addition, lining layer cells express a wide range is not vessel specific, cells in the lining and of macrophage antigens, and after lethal

within the mononuclear infiltrates also irradiation and heterologous bone marrow on September 27, 2021 by guest. Protected copyright. staining positive. Perhaps surprisingly, vascular transplantation, the synovial macrophages are VCAM- 1 expression is either minimal or replaced by cells of bone marrow donor absent, depending on the monoclonal antibody strain. 2 25 In favour of proliferation, some used. VCAM-1 is found in the lining, as it is 3H-thymidine incorporation of synovial lining in normal individuals, and occasional dendritic cells in RA is seen, and lining layer thickening type cells within the infiltrates also stain without mononuclear infiltration occasionally positive. observed.26 In a recent study, Qu et al further In a recent study of biopsies taken from addressed this question by examining synovial patients with actively inflamed joints (nine membrane samples for three markers of seropositive RA, three , one cell proliferation: proliferating cell nuclear ankylosing spondylitis), correlations were antigen, C-myc proto-oncogene, and nucleolar sought between the expression of adhesion organiser regions.26 All three markers indicated molecules in the synovial membrane, measures that there was active proliferation in the of disease activity, and concentrations of synovial lining ofRA patients. Double labelling soluble adhesion molecules in both sera and experiments indicated that the proliferating .'8 19 A consistent positive cells were fibroblastic in type, as they were correlation was identified between ICAM- 1 type I procollagen positive and CD68 negative. concentrations in serum, synovial fluid, Thus lining layer thickening is likely to be the synovial membrane, and measures of disease result of a combination of macrophage activity (Ritchie articular index, erythrocyte recruitment and like synoviocyte sedimentation rate (ESR)), but this was proliferation. not seen with VCAM-1 or E-selectin.'8 How- The central role ofthe T cell in the initiation ever, a consistent negative relationship was and maintenance ofthe inflammatory response identified between ICAM-1 and E-selectin in RA has often been stressed.2" It is ofinterest, expression in these patients.'9 As serum levels though, that T cell products such as interferon of soluble E-selectin were greater in patients gamma and interleukin-2 (IL-2) are difficult to Synovial membrane ceMlularity and vascularity 513

find in RA synovial fluid and membrane, and cartilage-pannus junction and elsewhere in the that gene expression for these lymphokines is synovial membrane, and it appears to be the also sparse. In contrast, monocyte/macrophage dominant cytokine in samples containing a Ann Rheum Dis: first published as 10.1136/ard.54.6.511 on 1 June 1995. Downloaded from derived products are plentiful in RA synovial transitional fibroblastic zone. Finally, IL-ira, membrane with cells, mostly of macrophage seen in 35% of lining cells in RA, is seen in origin, staining positive with antibodies against < 10% of cells at the cartilage-pannus IL-la, tumour necrosis factor (TNF) a, junction.3' Particularly in samples of invasive TNF,B, and IL-6, in addition to both the type cartilage-pannus junction, there appears to be 1 IL-1 receptor (IL-iRI) and the TNF recep- a predominance of proinflammatory cytokines tors (p55TNF-R and p75TNF-R).2"32 While capable of promoting release of factors IL-1 receptor antagonist (IL-Ira) expression is involved in joint destruction. also seen, it has been suggested that insufficient amounts are produced by RA synovial membrane to neutralise IL-I activity.3" Finally, Follow up studies ofRA synovial in vitro production ofcytokines IL- 1 [ and IL-6 membrane and effects oftreatment by RA synovial membrane explants was It has long been recognised that certain joints compared in two groups of arthroplasty may deteriorate radiologically even in the samples, those with focal and those with diffuse absence of clinical inflammation. Follow up infiltrates.20 Focal infiltrates occur more studies in RA have confirmed that radiological frequently in arthroplasty samples, and both deterioration occurs despite significant IL-i1[ and IL-6 production was significantly improvement in parameters of disease enhanced in those with focal disease. This activity. 36 In a study of patients for 40 months suggests that lymphoid follicle formation is after total lymphoid irradiation, a significant associated with enhanced IL- 1 [ and IL-6 and sustained decrease in peripheral blood T production, which in turn leads to greater cell numbers was observed, accompanied by a synthesis of the proteases involved in joint continued, gradual radiological deterioration.37 destruction. Rooney et al compared synovial membrane immunohistological features with the clinical outcome of disease over a one year period.38 In Cartilage-pannus junction in established those patients in whom there was a measurable RA clinical improvement, there was a reduction in Two main histological patterns at the cartilage- spontaneous in vitro IgM rheumatoid factor pannus junction have been recognised in production, a decrease in the intensity of the established RA.33 These patterns are based on T cell infiltrate, and reductions in the T helper the presence or absence of a transitional fibro- cell to T suppressor cell ratio and the number blastic zone overlying the cartilage and of samples that contained B cells. These obser-

separating the cartilage from pannus which vations were consistent with the suggestion http://ard.bmj.com/ contains macrophages, fibroblasts, and endo- that some of the immunopathogenic events in thelial cells. This transitional fibroblastic zone RA may be arrested or reversed by treatment. stains positively for keratin sulphate and type A more recent study has sought correlations II collagen, suggesting that it is ofchondrocytic between clinical and radiological outcome over origin. In a further study, Allard et al examined one year and synovial membrane features the association between cartilage-pannus obtained before treatment in RA patients.'2

junction pattern, joint involved, cartilage While a significant correlation was observed on September 27, 2021 by guest. Protected copyright. degradation (proteoglycan depletion and chon- between the number of macrophages and the drocyte necrosis), and radiological appear- degree of joint erosion, there was none with ance.34 The transitional zone occurred more infiltrating T or B lymphocyte populations. In commonly in the hip and compared with addition, the number of synovial tissue macro- the metatarsophalangeal joint. When the zone phages correlated with lining layer cellularity, was absent and a distinct, well defined which in turn correlated with synovial fluid cartilage-pannus junction was present, concentrations of IL-6. These findings suggest cartilage degradation appeared to be enhanced that synovial tissue macrophages are critical in and joint erosion more common. Thus it was the pathogenesis of joint erosion, which is suggested that these different cartilage-pannus perhaps mediated through release of cytokines junction patterns may explain the predomi- such as IL-6. nance of erosive change in small joints, Mulherin et al have examined the same compared with joint space narrowing found in question in a prospective study of 58 RA large joints in RA. patients followed over a mean of 6- 1 The cytokine profile of the cartilage-pannus years.36 39 40 In that study, all clinical measures junction has also been studied in these two of disease activity improved, while the mean histological patterns.35 In those samples in Larsen score deteriorated significantly. Syno- which a distinct junction was seen, there was vial membrane was obtained at review from 28 plentiful IL-ia, IL-1,[, TNFa, and IL-6, in patients and analysis showed that sublining addition to both IL-iR1, p55TNF-R and layer CD14 counts and lining layer thickness p75TNF-R.3' 32 3 In contrast, these cytokines correlated with both the percentage change and receptors were not seen in samples which and the final Larsen scores. Additional contained the transitional fibroblastic zone. correlations were seen between lining layer Transforming growth factor [3, which is CD68 count and final Larsen score, sublining thought to inhibit the immune response and to CD 14 count and both lining layer thickness promote tissue repair, is also found at the and lining layer CD14 counts, and between 514 FitzGerald, Bresnihan

lining CD68 count and both sublining CD68 lining layer thickness in PsA explains the

count and lining layer CD14 count. While no different radiographic features observed. Ann Rheum Dis: first published as 10.1136/ard.54.6.511 on 1 June 1995. Downloaded from correlation was found between T cells or T cell subsets and radiological outcome, these con- sistent positive correlations underscore the Sunmnary important role of the macrophage in articular Both inflammation and destruction of the joint damage. are the hallmarks of RA. While the con- Current treatment strategies in RA involve ventional model of RA suggests that synovial the use of a number of drugs which are inflammation is T cell mediated and that this considered to modify the course of the disease. in turn stimulates lining layer thickening and These drugs do appear to modify indices of pannus formation leading to joint damage, disease activity, but support for their ability to Zvaifler and Firestein have recently postulated alter the radiological outcome has not been an alternative model of joint destruction.45 In convincing. In addition, only a few studies have this model they suggest that synoviocyte examined the effects of these agents on the proliferation, pannus formation and inflam- synovial membrane. Corkill et al examined mation comprise an autonomous process synovial membrane biopsy samples after two independent of T cell derived factors. In and 12 weeks of treatment with intramuscular support of their suggestion, they point out that gold.4' Interestingly, the number of vessels the RA synoviocyte expresses features of a expressing E-selectin was significantly reduced transformed phenotype: evidence of active after two and 12 weeks. As IL-1,B expression proliferation, ability to invade cartilage, the was also shown to decrease, the reduction in expression of oncogenes, and the secretion E-selectin expression could have been due to of metalloproteinases. In addition, T cell this decrease in IL-13p.42 products and lymphokine gene expression are The mechanism of action of methotrexate not a prominent feature in RA, animal models was also examined in a recent serial synovial of arthritis have shown that joint damage can membrane biopsy study of eight patients who occur in the absence of lymphocytic infil- were beginning oral therapy.43 All patients tration, and T cell targetted therapies have noted a clinical improvement, with a reduction largely been disappointing. in ESR and a variable though slight reduction Studies reported in this review lend support in synovial membrane inflammation. Given the to the suggestion that mechanisms of joint importance of macrophages and metallo- inflammation and destruction operate at least proteinases in joint destruction, collagenase in a semi-autonomous fashion. With the treat- gene expression was significantly decreased ments currently available, clinical features of after methotrexate therapy. Tissue inhibitor of joint inflammation in RA improve and yet joint metalloproteinase- 1 and stromelysin gene damage worsens. A consistent relationship is expression were not changed. As methotrexate seen between radiological deterioration and http://ard.bmj.com/ did not alter the in vitro collagenase gene synovial membrane macrophage infiltration expression by IL- 1,8 stimulated fibroblast like and lining layer thickening. In PsA, a disease synoviocytes, it was suggested that its effects with different radiological features compared in vivo are probably the result of an alteration with RA, there is a reduction in macrophage in synovial membrane cytokine profiles. numbers and in lining layer thickening. The macrophage, lining layer cells and their products clearly play a role in joint destruction, on September 27, 2021 by guest. Protected copyright. Synovial membrane in psoriatic ardritis and current treatment strategies do not appear Certain clinical features in patients with to affect this process. The treatments do psoriatic arthritis (PsA) are helpful in diag- modify inflammation which is mediated both nosis: the presence of psoriasis, an asymmetric by cells in the lining and by the sublining joint distribution, nail dystrophy, distal infiltrate, but new approaches are needed interphalangeal joint involvement and dactyl- which are aimed at altering the more itis. Radiological features are also different and destructive potential of the disease. help distinguish the disease from RA; in particular the presence of bone proliferation is common in PsA.44 The immunohistological 1 Henderson B, Edwards J C W, eds. Structure of synovial features of PsA synovial membrane have lining. In: The synovial lining in health and disease. Cambridge: Chapman and Hall Medical, 1987; 17-40. recently been described and quantified.'7 2 Athanasou N A, Quinn J. Immunocytochemical analysis of Compared with RA there was human synovial lining cells: phenotypic relation to other controls, marrow derived cells. Ann Rheum Dis 1991; 50: 311-5. significantly less lining layer hyperplasia, fewer 3 Wilkinson L S, Pitsillides A A, Worrall J G, Edwards J C W. macrophages and a greater number of blood Light microscopic characterization of the fibroblast-like synovial intimal cell (synoviocyte). Arthitis Rheum 1992; vessels in PsA synovial membrane. In addition, 35: 1179-84. E-selectin expression was less intense in PsA 4 Wikinson L S, Edwards J C W, Ponston R N, Haskard D 0. Expression of vascular cell adhesion synovial membrane, while there was no differ- molecule-1 in normal and inflamed synovium. Lab Invest ence in of ICAM-1 and 1993; 68: 82-8. expression VCAM-1. 5 Wllkinson L S, Worrall J G, Sinclair H S, Edwards J C W. Numbers of T cells, T cell subsets and B cells Immunohistochemical reassessment of accessory cell were similar in both populations in normal and diseased human synovium. Br groups. With these JRheumatol 1990; 29: 259-63. findings, it is tempting to suggest that with the 6 Knight A D, Levick J R. The density and distribution of reduction in capillaries around a synovial cavity. QJ'Exp Physiol 1983; E-selectin, fewer macrophages 68: 629-44. traffic into the synovial membrane and out to 7 Allard S A, Bayliss M T, Maini R N. The synovium- the It also be cartilage junction ofthe normal human knee. Implications lining layer. may suggested that for joint destruction and repair. Arthrtis Rheum 1990; 33: the reduction in macrophage numbers and 1170-9. Synovial membrane cellularity and vasculaity 515

8 Stevens C R, Mapp P I, Revell P A. A monoclonal antibody synoviocytes contributes to synovial hyperplasia. Arthritis (Mab 67) marks type B synoviocytes. Rheumatol Int 1990; Rheum 1994; 37: 212-20. 10: 103-6. 27 Sewell K L, Trentham D E. Pathogenesis of rheumatoid 9 Schumacher H R, Kitridou R D. Synovitis of recent onset. arthritis. Lancet 1993; 341: 283-6. Ann Rheum Dis: first published as 10.1136/ard.54.6.511 on 1 June 1995. Downloaded from A clinicopathologic study during the first month of 28 Field M, Chu C, Feldman M, Maini R N. Interleukin-6 disease. Arthritis Rheum 1972; 15: 465-85. localisation in the synovial membrane in rheumatoid 10 Lindblad S, Klareskog L, Hedfors E, Forsum U, arthritis. RheumatolInt 1991; 11: 45-50. Sundstrom C. Phenotypic characterization of synovial 29 Chu C Q, Field M, Abney E, et al. Transforming growth tissue cells in situ in different types of synovitis. Arthritis factor-P 1 in rheumatoid synovial membrane and Rheum 1983; 26: 1321-32. cartilage/pannus junction. Clin Exp Immunol 1991; 86: 11 Yanni G, Whelan A, Feighery C, Bresnihan B. Synovial 380-6. tissue macrophages and joint erosion in rheumatoid 30 Chu C Q, Field M, Feldman M, Maini R N. Localization arthritis. Ann Rheum Dis 1994; 53: 39-44. of tumor necrosis factor a in synovial tissues and at the 12 Soden M, Rooney M, Cullen A, Whelan A, Feighery C, cartilage-pannus junction in patients with rheumatoid Bresnihan B. Immunohistochemical features in the arthritis. Arthritis Rheum 1991; 34: 1125-32. synovium obtained from clinically uninvolved knee joints 31 Deleuran B W, Chu C Q, Field M, et al. Localization of of patients with rheumatoid arthritis. Br Jf Rheumatol interleukin-lt, type 1 interleukin-I receptor and inter- 1989; 28: 287-92. leukin-1 receptor antagonist in the synovial membrane 13 FitzGerald 0, Soden M, Yanni G, Robinson R, and cartilage/cannus junction in rheumatoid arthritis. Br Bresnihan B. Morphometric analysis of blood vessels in J Rheumatol 1992; 31: 801-9. synovial membranes obtained from clinically affected and 32 Deleuran B W, Chu C Q, Field M, et al. Localization of unaffected knee joints of patients with rheumatoid tumor necrosis factor receptors in the synovial tissue and arthritis. Ann Rheum Dis 1991; 50: 792-6. cartilage-pannus junction in patients with rheumatoid 14 Koch A E, Burrows J C, Haines G K, Carlos T M, arthritis. Arthritis Rheum 1992; 35: 1170-8. Harlan J M, Leibovich S J. Immunolocalization of 33 Allard S A, Muirden K D, Camplejohn K L, Maini R N. endothelial and leukocyte adhesion molecules in human Chondrocyte-derived cells and matrix at the rheumatoid rheumatoid and osteoarthritic synovial tissues. Lab Invest cartilage-pannus junction identified with monoclonal 1991; 64: 313-20. antibodies. RheumatolInt 1987; 7:153-9. 15 Gerritsen M E, Kelley K A, Ligon G, et al. Regulation of 34 Allard S A, Muirden K D, Maini R N. Correlation of the expression of intercellular adhesion molecule 1 in histopathological features of pannus with patterns of cultured human endothelial cells derived from rheuma- damage in different joints in rheumatoid arthritis. Ann toid synovium. Arthritis Rheum 1993; 36: 593-602. RheumDis 1991; 50: 278-83. 16 Szekanecz Z, Haines G K, Lin T R, et al. Differential 35 Chu C Q, Field M, Allard S, Abney E, Feldman M, distribution of intercellular adhesion molecules (ICAM-1, Maini R N. Detection of cytokines at the cartilage/ ICAM-2, and ICAM-3) and the MS-I antigen in normal pannus junction in patients with rheumatoid arthritis: and diseased human synovia. Arthritis Rheum 1994; 37: implications for the role of cytokines in cartilage 221-31. destruction and repair. BrJRheumatol 1992; 31: 653-61. 17 Veale D, Yanni G, Rogers S, Barnes L, Bresnihan B, 36 Mulherin D, FitzGerald 0, Bresnihan B. Radiologic FitzGerald 0. Reduced synovial membrane macrophage progression of rheumatoid arthritis over six years numbers, ELAM-1 expression, and lining layer hyper- correlates with macrophage numbers in the synovium. plasia in psoriatic arthritis as compared with rheumatoid Arthritis Rheum 1993; 36 (suppl): S86. arthritis. Arthritis Rheum 1993; 36: 893-900. 37 Soden M, Hassan J, Scott D L, et al. Lymphoid irradiation 18 Mulherin D, Veale D J, Belch J J F, Bresnihan B, in intractable rheumatoid arthritis: long-term follow-up of FitzGerald 0. Adhesion molecule expression in matched patients treated with 2000 rad and 750 rad. Arthritis synovial membrane, synovial fluid and serum in inflam- Rheum 1989; 32: 523-30. matory arthritis. IrJMedSci 1995; 164. In press. 38 Rooney M, Whelan A, Feighery C, Bresnihan B. Changes 19 Mulherin D, Bresnihan B, FitzGerald 0, Veale D, Belch J. in lymphocyte infiltration of the synovial membrane and Relationship of ICAM-1 with E-selectin in synovial the clinical course of rheumatoid arthritis. Arthritis Rheum membrane, fluid and serum in inflammatory arthritis. 1989; 32: 361-9. Arthritis Rheum 1994; 37 (suppl 2): S189. 39 Mulherin D, FitzGerald 0, Bresnihan B. An analysis of the 20 Yanni G, Whelan A, Feighery C, et al. Contrasting levels of discrepency between the clinical and radiologic course in in vitro cytokine production by rheumatoid synovial rheumatoid arthritis. Arthritis Rheum 1994; 37 (suppl): tissues demonstrating different patterns of mononuclear S251. cell infiltration. Clin Exp Immunol 1993; 93: 387-95. 40 Mulherin D, FitzGerald 0, Bresnihan B. The relationship 21 Duke 0, Panayi G S, Janossy G, Poulter L W. An between macrophage populations in the synovium and immunohistological analysis of lymphocyte sub- articular damage in rheumatoid arthritis. Arthritis Rheum http://ard.bmj.com/ populations and their microenvironment in the synovial 1994; 37 (suppl): S222. membranes of patients with rheumatoid arthritis using 41 Corkill M M, Kirkham B W, Haskard D 0, Barbatis C, monoclonal antibodies. Clin Exp Immunol 1982; 49: Gibson T, Panayi G S. Gold treatment of rheumatoid 22-30. arthritis decreases synovial expression of the endothelial 22 Kurosaka M, ZiffM. Immunoelectron microscopic study of leukocyte adhesion receptor ELAM-1. J Rheumatol 1991; the distribution of T cell subsets in rheumatoid synovium. 18: 1453-60. JExp Med 1983; 158: 1191-210. 42 Kirkham B W, Navarro F, Corkill M M, Gibson T, 23 Yanni G, Whelan A, Feighery C, Bresnihan B. Analysis of Panayi G S. Change in interleukin 1 beta expression in cell populations in rheumatoid arthritis synovial tissues. rheumatoid synovial membrane after treatment with gold

SeminArthritisRheum 1992; 21: 393-9. and glucocorticoid. Clin Rheumatol 1990; 9:118-9. on September 27, 2021 by guest. Protected copyright. 24 Lalor P A, Mapp P I, Hall P A, Ravell P A. Proliferative 43 Firestein G, Paine M M, Boyle D L. Mechanisms of activity of cells in the synovium as demonstrated by methotrexate action in rheumatoid arthritis. Selective monoclonal antibody Ki-67. Rheumatol Int 1987; 7: decrease in synovial collagenase gene expression. Arthritis 183-6. Rheum 1994; 37: 193-200. 25 Edwards J C, Wllloughby D A. Demonstration of bone- 44 Brewer A C. The radiographic features of psoriatic arthritis. marrow-derived cells in synovial lining by means of giant In: Gerber L H, Espinoza L R, eds. Psoriatic arthritis. intracellular granules as genetic markers. Ann Rheum Dis Orlando: Grune and Stratton, 1985: 125-46. 1991; 41: 177-82. 45 Zvaifler N J, Firestein G S. Pannus and pannocytes. 26 Qu Z, Garcia C H, O'Rourke L M, Planck S R, Kohli M, Alternative models of joint destruction in rheumatoid Rosenbaum J T. Local proliferation of fibroblast-like arthritis. Arthritis Rheum 1994; 37: 783-9.