Background/Purpose:

Synovial membrane inflammation is common in osteoarthritis (OA). Inflammation may be present at all disease stages, increases the risk of cartilage injury, and is heterogeneous histologically, with variable synovial hyperplasia, inflammatory infiltrates, and angiogenesis. Our objective in this study was to develop an approach to characterize this variability in inflammation using flow cytometry and multiplex assays. We focused particularly on IL-6, as this inflammatory cytokine independently associates with poor OA outcomes.

Methods:

Cell composition and soluble protein release was measured in synovium collected from thirty-five knee OA patients undergoing joint replacement surgery. Correlation-based clustering identified patient-specific inflammatory networks.

Results:

We first measured release of thirteen inflammatory cytokines and adipokines by intact synovium. These proteins separated into at least two expression groups: (1) strong correlation between IL-6 and IL-8, linked through IL-8 to correlations with adipsin, CCL2, CXCL10, and adiponectin (potential fibroblast/macrophage pattern); and (2) correlations between IL-10, IFN-γ, resistin, TNFα and IL-1β (potential T cell pattern). We next enzymatically disaggregated synovium and found that identifying mesenchymal and hematopoietic cell populations required different digestion conditions. Collagenase with higher dispase concentrations markedly improved total cell yield and mesenchymal marker staining. However, it negatively impacted T and natural killer (NK) cell staining. Intracellular IL-6 expression was detected in many cells, but was highest in mesenchymal cells, especially those expressing both CD90 and CD34. IL-6⁺ synovial cell numbers correlated strongly with IL-6 tissue release, validating our flow cytometry assay. IL-6 release did not correlate with leptin or body mass index, suggesting adipocytes produce little IL-6. OA synovial T cell infiltration was highly variable, with increased percentage of T cells correlating with CD8+ cytotoxic T cell influx. CD8⁺ T cells showed variable CD45RA and CD45RO expression and their presence correlated positively the soluble T cell pattern above. Finally, combining flow cytometric and multiplex data identified at least five possible OA synovial inflammation patterns, expanding on the networks identified by soluble protein analysis alone.

Conclusion:

We have developed a novel approach to analyze OA inflammation that has identified patient-specific inflammatory clusters, including an unexpected involvement of CD8+ T cells. This study argues that identifying synovial inflammatory will provide new insights into OA patient heterogeneity and biomarker development.

« Back to 2018 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/synovial-inflammation-identifies-patients-clusters-in-osteoarthritis/