Objective Personalised healthcare is contingent on the identification of biomarkers that represent disease relevant pathways and predict drug response. The authors aimed to develop a gene expression signature in synovial tissue that could enrich clinical response of rheumatoid arthritis (RA) patients to rituximab.

Methods The authors studied synovial gene expression using high-throughput quantitative real-time-PCR in 20 RA patients who underwent arthroscopy before and after treatment with rituximab. Several objective approaches were used to explore patterns in the data and to find genes associated with changes in disease activity due to treatment.

Results This analysis revealed two patient populations associated with distinct clinical, laboratory and histological features and, importantly, showed enrichment for response (60% non-responders vs 90% responders). A composite baseline gene score (GS) correlated with change in disease activity score (ΔDAS) between baseline and month 3 (r=0.74, p=0.0002), but also with ΔDAS at later time-points (month 9, r=0.54, p=0.016; month 15, r=0.45, p=0.06; month 21, r=0.72, p=0.003). Notably, the GS significantly correlated with baseline erythrocyte sedimentation rate (r=0.69, p=0.0008), but not with other DAS components. The GS genes represented T cell, macrophage, remodelling and interferon-α biology. Responders demonstrated higher expression of macrophage and T cell genes, while non-responders showed higher expression of interferon-α and remodelling genes.

Conclusions This study reveals a baseline synovial GS that correlates with early and late clinical responses to rituximab. The GS biology suggests that T cells and macrophages are important for response to B cell depleting therapy, while expression of remodelling and interferon-α genes correlates with poor response.