Autoimmunity to stromal-derived autoantigens in rheumatoid ectopic germinal centers exacerbates arthritis and affects clinical response
Elisa Corsiero, Mattia Caliste, Lucas Jagemann, Liliane Fossati-Jimack, Katriona Goldmann, Cankut Cubuk, Giulia M. Ghirardi, Edoardo Prediletto, Felice Rivellese, Cristiano Alessandri, Mark Hopkinson, Behzad Javaheri, Andrew A. Pitsillides, Myles J. Lewis, Costantino Pitzalis, Michele Bombardieri
Elisa Corsiero, Mattia Caliste, Lucas Jagemann, Liliane Fossati-Jimack, Katriona Goldmann, Cankut Cubuk, Giulia M. Ghirardi, Edoardo Prediletto, Felice Rivellese, Cristiano Alessandri, Mark Hopkinson, Behzad Javaheri, Andrew A. Pitsillides, Myles J. Lewis, Costantino Pitzalis, Michele Bombardieri
View: Text | PDF
Research Article Autoimmunity

Autoimmunity to stromal-derived autoantigens in rheumatoid ectopic germinal centers exacerbates arthritis and affects clinical response

Abstract

Ectopic lymphoid structures (ELSs) in the rheumatoid synovial joints sustain autoreactivity against locally expressed autoantigens. We recently identified recombinant monoclonal antibodies (RA-rmAbs) derived from single, locally differentiated rheumatoid arthritis (RA) synovial B cells, which specifically recognize fibroblast-like synoviocytes (FLSs). Here, we aimed to identify the specificity of FLS-derived autoantigens fueling local autoimmunity and the functional role of anti-FLS antibodies in promoting chronic inflammation. A subset of anti-FLS RA-rmAbs reacting with a 60 kDa band from FLS extracts demonstrated specificity for HSP60 and partial cross-reactivity to other stromal autoantigens (i.e., calreticulin/vimentin) but not to citrullinated fibrinogen. Anti-FLS RA-rmAbs, but not anti–neutrophil extracellular traps rmAbs, exhibited pathogenic properties in a mouse model of collagen-induced arthritis. In patients, anti-HSP60 antibodies were preferentially detected in RA versus osteoarthritis (OA) synovial fluid. Synovial HSPD1 and CALR gene expression analyzed using bulk RNA-Seq and GeoMx-DSP closely correlated with the lympho-myeloid RA pathotype, and HSP60 protein expression was predominantly observed around ELS. Moreover, we observed a significant reduction in synovial HSP60 gene expression followed B cell depletion with rituximab that was strongly associated with the treatment response. Overall, we report that synovial stromal-derived autoantigens are targeted by pathogenic autoantibodies and are associated with specific RA pathotypes, with potential value for patient stratification and as predictors of the response to B cell–depleting therapies.

Authors

Elisa Corsiero, Mattia Caliste, Lucas Jagemann, Liliane Fossati-Jimack, Katriona Goldmann, Cankut Cubuk, Giulia M. Ghirardi, Edoardo Prediletto, Felice Rivellese, Cristiano Alessandri, Mark Hopkinson, Behzad Javaheri, Andrew A. Pitsillides, Myles J. Lewis, Costantino Pitzalis, Michele Bombardieri

×

Figure 6

HSP60 analysis of paired pre- and post-rituximab treatment synovial biopsies.

Options: View larger image (or click on image) Download as PowerPoint
HSP60 analysis of paired pre- and post-rituximab treatment synovial biop...
(A) Schematic representation of the R4RA clinical trial (27, 28). At baseline, patients with RA underwent a synovial biopsy of a clinically active joint for histological evaluation. Following the synovial biopsy, patients were randomized 1:1 to either the rituximab or tocilizumab treatment group. An optional repeated synovial biopsy of the same joint previously sampled was performed at 16  weeks, followed by histological and transcriptomic evaluation. A CDAI of 50% was used as a primary endpoint to define responders and nonresponders to treatment at 16 weeks. Patients were initially followed up every 4 weeks, with the end of the trial at 48 weeks. Numbers in parentheses represent patients with available RNA-Seq data. (B) Synovium HSPD1 variance-stabilizing transformation (VST) gene expression in patients stratified by histology into B cell–poor (B poor) and B cell–rich (B rich) categories. (C) HSPD1-normalized synovial gene expression levels assessed at baseline in CDAI nonresponders versus responders and DAS28-CRP moderate/nonresponders versus responders to rituximab treatment. (D) HSPD1 RNA-Seq counts were assessed at baseline and 16 weeks following rituximab and tocilizumab treatment in paired synovial biopsies (rituximab, n = 29 individuals, n = 58 samples; tocilizumab, n = 15 individuals, n = 30 samples). (E and F) HSPD1 counts were assessed at baseline and 16 weeks following rituximab treatment in paired synovial biopsies. In D–F, gray data points represent RNA-Seq counts between paired samples for individuals. Statistical analysis was performed using a negative binomial mixed-effects model on RNA-Seq counts. Overlaid green/purple data points show estimated marginal means of the fitted mixed-effects model, with error bars showing 95% CIs for the fixed effects of the mixed-effects model. To assess the clinical response, the following parameters were used: CDAI 50% improvement and DAS28-CRP European Alliance of Associations for Rheumatology (EULAR) response. RTX, rituximab.