Identification of synovial biomarkers of response to experimental treatment in early-phase clinical trials in spondylarthritis

Objective. To identify biomarkers for effective treatment in early-phase clinical trials of spondyl-arthritis (SpA), by analyzing which synovial features can be reliably identified in patients with SpA. Methods. Synovial biopsies were performed at weeks 0 and 12 in 20 SpA patients treated with infliximab, 20 treated with etanercept, and 12 who were not treated. Primary clinical outcome measures were patient and physician global assessment of disease activity. Extensive histologic evaluation included assessment of lining layer hyperplasia, vascularity, markers of cellular infiltration, and metalloproteinases (MMPs) in the lining and sublining layers. Results. Changes in levels of CD163 (resident tissue macrophages) in the lining, and CD163, MMP-3, and myeloid-related protein 14 (MRP-14; infiltrating myeloid cells) in the sublining correlated significantly with changes in the primary clinical outcomes. Comparison between responders (n = 35) and nonresponders (n = 17) showed differences in the degree of change in the levels of CD163 in the lining and CD163, MMP-3, and CD3 in the sublining, whereas trends in change in the levels of MRP-8 and MRP-14 in the lining and sublining were similar in the 2 groups. Accordingly, the highest differences in standardized response means (SRMs) between the 2 groups were found for CD163 in the lining, MMP-3, CD163, CD3, and MRP-8 in the sublining, and the level of polymorphonuclear cells (PMNs). When comparing treated and untreated patients, high differences in SRMs were again found for CD163 in the lining, MMP-3, CD163, and MRP-8 in the sublining, and PMNs. These parameters performed prognostically as well as the erythrocyte sedimentation rate and better than the C-reactive protein level. Class prediction analysis yielded a 90% correct prediction using 8 synovial parameters, as follows: lining and sublining CD163, MRP-8, and MRP-14, sublining MMP-3, and PMNs. In validation analyses with independent samples, effective treatment was correctly predicted in 24 of 30 SpA patients and in 2 of 2 placebo-treated patients. Conclusion. Changes in synovial macrophage subsets, PMN levels, and MMP-3 expression reflect response to treatment in SpA. The ability of these parameters to correctly identify effective therapy makes them interesting biomarkers for use in early-phase clinical trials in SpA.