Ab0062 abundance and activation of memory b-cells characterize synovitis in rheumatoid arthritis patients with early treatment response

Background A relevant percentage of patients with rheumatoid arthritis (RA) do not respond to individual (40%) or all available disease modifying anti-rheumatic drugs DMARDs (5-20%). A recent study using synovial bulk RNAseq from a phase 4 trial has shown that humoral immune response gene signatures are associated with response to rituximab and tocilizumab [1]. Other biomarkers that can predict early response to alternative DMARDs are still lacking. Objectives This study, performed in a real-life clinical care setting, aimed at characterizing synovitis of RA patients with an early drug response on a single-cell level to identify potential biomarkers and molecular signatures. Methods Synovial tissue was obtained by ultrasound-guided synovial biopsy from inflamed joints. Four patients were defined as early responders not meeting the D2T EULAR definition criteria (median (IQR) number of treatments 1.5 (1.25)) [2], but meeting 2022 ACR/EULAR remission within 3 months on the DMARD started after the synovial biopsy. Five patients met the D2T EULAR definition criteria (median (IQR) number of treatments 7 (1)). Two out of five from the D2T group went into ACR/EULAR remission 3 months after the biopsy following DMARD initiation (abatacept (n=1) and upadacitinib (n=1)). Both groups fulfilled the ACR/EULAR classification criteria for RA. 6000 cells were targeted for single-cell encapsulation (10X Genomics workflow v3.1). Libraries were sequenced on NovaSeq6000. The following R packages were used: Cell Ranger, Seurat, Harmony, PCAtools. Results Clinical characteristics (sex, age, swollen joint count, C-reactive protein, body mass index, smoking status) were comparable between the two groups, except for the presence of autoantibodies (all D2T compared to only one early responder patient were seropositive). The patients in the early responder group received the following drugs after the biopsy: methotrexate plus certolizumab (n=1), methotrexate alone (n=1) or abatacept (n=2). Patients in the D2T group had received a median of seven treatments before the biopsy. 29,408 cells were integrated for scRNA-seq analysis (mean of 3,267 cells after quality control filtering per patient). Unsupervised dimensionality reduction analysis showed 8 principal components (PCs) explaining all variations in gene expression between samples. PC6 was significantly associated with the early responder/D2T categorization, accounting for 6.72% of the sample variation. Immunoglobulin genes (IgH-, IgL-), Prostaglandin D2 Synthase (PTGDS) and B-cell antigen receptor complex-associated protein alpha chain (CD79A) were among the top 100 genes associated with early response to treatment (PC6) (Figure 1A). Analysis of the average expression of these genes in the early responder and D2T group revealed expression mostly in B cells and plasma cells (Figure 1B). Comparing the proportions of cell populations using graph-based clustering showed no significant differences between early responder and D2T patients. Further clustering of the primary cell populations also did not show changes in proportions except for B cells: the proportions of NR4A1+ memory B cells (B cells recently receiving antigen stimulation) were higher in early responder compared to D2T patients (mean 48%±10 vs 36%±10, p-value = 0.0547) (Figure 1C). Analysis of differentially expressed genes in B cells between early responder and D2T patients revealed 62 genes differentially expressed (p-value<0.05, average fold change > or < 0.3). Still, only six of these genes were among the top 100 genes associated with PC6 (Figure 1D). Conclusion Synovitis in rheumatoid arthritis patients with early treatment response was associated with activation of B cells and an abundance of NR4A1+ memory B cells. This finding in a real-life cohort is in concordance with previous results showing an association of humoral immune activation and response [1] and sets the path for future biomarkers predicting early response in RA. References [1]Rivellese et al, Nat Med 28, 1256–1268 (2022) [2]Nagy et al. ARD 80, 31-35 (2021) Acknowledgements: NIL. Disclosure of Interests Alexandra Khmelevskaya: None declared, Miranda Houtman: None declared, Kristina Buerki: None declared, Chantal Pauli: None declared, Felice Rivellese: None declared, Edoardo Prediletto: None declared, Costantino Pitzalis: None declared, Oliver Distler Consultant of: Abbvie, Adrian Ciurea: None declared, Caroline Ospelt: None declared, Raphael Micheroli: None declared.