Background RA is a very heterogeneous disease, recently highlighted by distinct differences in the genetic contribution to APCA(+) and ACAP(–) disease suggesting two divergent pathogenic models, with different rates of progression and response to treatment. Synovial tissue studies showed that the cellular composition of synovial membrane is relatively consistent however the size of the inflammatory cell infiltrate is variable in direct relation to local levels of inflammation and sometimes highly organised with germinal centre like structures (GCL). The role of IL-7 as an essential factor in tissue organisation has recently been recognised and we hypothesize that the synovial tissue architecture (TA) may be driven by IL-7 expression in RA.
Methods Synovial tissue biopsies were obtained during arthroscopy (n = 93, distributed between studies). Histology, IHC and two TaqMan gene-arrays were used. Tissues were classified as diffuse infiltration, aggregates and GCL-structures (defined B and T-cell zones). Clustering was used to analyse gene expression data.
Results IL-7 was detected with large variation between samples. There was association between% of IL-7(+) cells and local inflammation (n = 30, rho = 0.620, p<0.0001), synovitis score (rho = 0.556, p = 0.001) but only in the sublining layer, as well as with increased TA-complexity (p = 0.001), CD3(+) cells (rho = 0.599, p = 0.002) and CD20(+) (rho = 0.605, p = 0.001) but not with CD68(+) cells. IL-7 gene expression was measured in 29 biopsies. IL-7 mRNA expression was significantly higher in tissue presenting aggregates over diffuse infiltration (p = 0.019) and GCL structure (p = 0.037). A first TaqMan-arrays (48-genes) of 44 RA biopsies generated a 2 groups samples hierarchy clearly associated with TA (p = 0.002) separating genes associated with B-cell biology from those associated with stroma, T cells and macrophages. IL-7 surprisingly clustered with the B-cell-group however its expression was highly correlated with CD4 (rho = 0.520, p<0.0001). A second array (96-genes) of 29 biopsies also showed association with TA (p = 0.032) and clustered B-cell maturation genes (CD80, CD86, CD40L, RAG-1, BLIMP, AIOLOS, XBP-1, all Ig-genes) but not CD19/CD20 therefore suggesting a plasma cell specific signature. CD68 and CD4 were excluded from this gene signature. IL-7 was included in the plasma-cell cluster and was closely associated with CD38 (rho = 0.381, p = 0.042) and CD31 (Rho = 0.381, p = 0.042). A regulatory loop between IL-6/IL-7 was also highlighted by the gene dendrogram.
Conclusion Our data suggest that IL-7 could orchestrate the synovial tissue architecture providing a niche for B-cell maturation. Blocking IL-7 may therefore be of therapeutic value as was recently demonstrated in several animal models of auto-immune diseases.