Lessons learnt from the synovial tissue response to anti-rheumatic treatment

There has been increased interest in studies on synovial tissue in recent years. This can be explained by technical developments and by the recognition of the synovium as the primary site of inflammation in patients with rheumatoid arthritis (RA) and other arthritides. Obviously, examination of peripheral blood and synovial fluid may provide insight into the production of soluble mediators as well as the dynamics of migration of inflammatory cells into different compartments. However, it has become clear that such studies yield only indirect information about the events in the synovium. Hence, there has been an upsurge in investigations of the pathological changes taking place in the rheumatoid synovium [1]. These studies have been used to provide more insight into the pathogenesis and to identify features which might be used for diagnostic purposes and to provide a prognosis [2]. The relevance of examination of the synovium has been underscored by the observation that clinical arthritis activity is accompanied by histological signs of synovitis after treatment of RA patients with the monoclonal antibody Campath‐1H, despite profound depletion of circulating lymphocytes [3]. Conversely, our unpublished observations have shown that the changes occurring after treatment with an effective anti‐inflammatory drug, such as prednisone, are more pronounced in the synovium compared with peripheral blood and synovial fluid. Studies of the changes in synovial biopsy specimens after specific therapeutic interventions may reveal information about the mechanism of action of treatment. By the analysis of synovial tissue, factors might also be identified that are associated with clinical signs and symptoms of arthritis. However, several methodological problems must be addressed first. Some of these will be reviewed in this editorial. Many investigators have used blind needle biopsy to obtain synovial tissue samples. This is a safe, well‐tolerated, cheap and technically easy method that allows synovial tissue to be acquired in a high percentage of RA patients [4]. Comparison of measures of inflammation in synovial tissue samples obtained by blind needle biopsy with samples selected under vision at arthroscopy revealed that the results were generally similar [5]. In addition, it was shown that it is not possible to predict the microscopic features based on the macroscopic appearance of the synovium. Therefore, both blind needle biopsy and arthroscopy, which is similarly safe and well tolerated [6], can be used in cross‐sectional studies. Arthroscopy has, however, several advantages for studies of serial biopsies, in spite of the fact that it is more complicated and expensive. A major advantage is that it is always possible to obtain tissue in adequate amounts, even when the synovial tissue volume has decreased substantially as a result of effective treatment. Arthroscopy allows access not only to the large joints, but also to small joints such as the wrist, the metacarpophalangeal joints and the proximal interphalangeal joints [7, 8]. Furthermore, arthroscopy makes it possible to evaluate the effects of treatment on articular cartilage [9]. Several studies have suggested that there is morphological heterogeneity in the synovium from a single joint [10, 11]. Other studies have indicated that, despite histological heterogeneity, markers of synovial inflammation can be practically and reliably quantified and that sampling error can be overcome [12, 13]. This raises the question of how many biopsy samples should be examined to get a representative impression of the microscopic features of the synovium. A study which focused on variation in the presence of T cells and signs of cell activation within the joint revealed that examination of slides derived from at least six biopsy specimens results in variance of less than 10% [14]. Thus, acceptable results can be obtained when at least six biopsy samples are used for immunohistological analysis. Presumably, this applies also to most other cell markers. An extensive quality control system is required for reliable immunohistological analysis. Protocols need to be optimized for (new batches of) various antibodies and chemicals. For some markers, such as p53, results obtained by immunohistology alone must be interpreted with caution, and it is essential to confirm the observations using other methods, for example Western blot analysis [15]. Since it is likely that studies of synovial biopsy samples will increase studies need to be initiated comparing the results obtained in various laboratories. Standardization of the methods will be necessary. There are essentially three methods of quantifying the features of synovial inflammation: semiquantitative analysis, conventional counting of cells, and computerized image analysis [5, 16]. Counting cells and computerized image analysis appear to be better able to detect small changes after treatment, whereas semi‐quantitative analysis and computerized analysis are generally more time‐efficient than conventional counting of cells. Recent sophisticated computer systems allow reliable and efficient evaluation of the cell infiltrate and the expression of cytokines in the synovium [16], and computerized image analysis will therefore become increasingly important. However, further development of software might be required for the quantification of blood vessels of variable size [17]. Various factors related to the biopsy procedure itself could theoretically influence the pathological changes in the synovium. Obviously, this could be a source of bias in studies of serial biopsies. Adrenalin will influence the macroscopic effect of the synovium, but its use is generally not necessary for needle arthroscopy. It is, however, essential to use adequate anaesthesia. Of importance, it has recently been shown that intra‐articular injection of bupivacaine and adrenaline does not affect the cell infiltrate [18]. Joint lavage during arthroscopy has a beneficial effect on the clinical signs of arthritis [6, 19] and presumably also on the characteristics of the synovium. Therefore, only limited use of lavage should be made in studies evaluating serial synovial biopsy samples. When needle arthroscopy was used without significant lavage, persistent disease activity was shown to be associated with unaltered synovial inflammation in serial biopsy samples after placebo treatment [20]. What lessons have been learned from the response of synovial tissue to therapeutic interventions? Several studies have described changes in the synovium after antirheumatic treatment (Table 1). The description of changes after specific interventions has provided insight into the mechanism of action of the treatment as well as into the role of certain cells and molecules in the pathogenesis of RA. Histological examination revealed a significant reduction in the cell infiltrate after treatment with the depleting anti‐CD4 antibodies (cM‐T412) [21]. There was not only a decrease in the number of CD4+ T cells, but also reduced infiltration by other inflammatory cells, such as macrophages, plasma cells and fibroblast‐like synoviocytes. The clear drop in the number of inflammatory cells in the synovium after depletion of CD4+ T cells supports the view that CD4+ T cells orchestrate the local cell infiltrate. There was, however, no clinical improvement after treatment. This might be explained by the fact that pre‐activated Th1‐like memory cells were relatively spared [22, 23] and by the persistent production of macrophage‐derived cytokines in the synovium [21]. Based on these observations, it is conceivable that higher doses or prolonged treatment might have resulted in clinical improvement. After treatment with anti‐tumour necrosis factor‐α (TNF‐α) antibodies (infliximab), we observed a significant reduction in infiltration by T cells [24]. More detailed immunohistological examination in five RA patients revealed a decrease in macrophages, plasma cells and fibroblast‐like synoviocytes as well [25]. The significantly diminished expression of adhesion molecules, such as E‐selectin and VCAM‐1, in the synovium [24] was associated with a reduction in serum concentrations of soluble adhesion molecules [26]. Accordingly, it has been suggested that the decrease in cellular infiltration resulted from reduced migration of inflammatory cells into the joint secondary to inhibition of the expression of cytokine‐induced vascular adhesion molecules [24] and chemokines [27]. This hypothesis was supported by the demonstration of reduced migration of 111In‐labelled neutrophil migration into the joints after anti‐TNF‐α treatment [27]. Thus, analysis of serial synovial samples has helped to elucidate one of the mechanisms of action of anti‐TNF‐α therapy. The beneficial clinical response was also associated with a decrease in the expression of macrophage‐derived cytokines in the synovium [25]. Taken together, these studies suggest that especially the number of macrophages and the expression of macrophage‐derived cytokines in the synovium are associated with the clinical expression of the disease. This was recently confirmed by two cross‐sectional studies [28, 29]. It can be expected that examination of serial synovial samples will also be used to predict the effects of therapeutic interventions. The feasibility of this approach is supported by the highly significant correlation between measures of macrophage infiltration and expression of their mediators on the one hand and clinical signs and symptoms of arthritis on the other [28]. Successful treatment with disease‐modifying anti‐rheumatic drugs (Table 1), such as gold, methotrexate and leflunomide, inhibits cytokine production by macrophages. In fact there are no examples of successful anti‐rheumatic therapies where macrophage infiltration in the synovium was shown to be unaffected. As illustrated by observations made after anti‐CD4 antibody therapy [21], the analysis of serial synovial biopsies appears to be more sensitive to change than, for example, the American College of Rheumatology 20% criteria for improvement. Another study showed changes in the synovium in patients with only a modest decrease in serum levels of acute phase reactants [30]. This suggests that this approach can be used as a screening method to test new compounds requiring relatively small numbers of patients. Conceivably, the expression of mediators of joint destruction, such as matrix metalloproteinases, will be used as surrogate markers for joint destruction, although their value remains to be shown. Studies using serial synovial biopsy for evaluation of anti‐rheumatic therapy Studies using serial synovial biopsy for evaluation of anti‐rheumatic therapy