Background Rituximab (RTX) has shown clinical efficacy but up to 40% of RTX treated rheumatoid arthritis (RA) patients are poor responders (ref:1) and the commonly used RA biomarkers (RF/ACPA) are poor predictors for therapy response. Others reported associations between BAFF/BLyS levels, Fcg RIII and IL-6 genotype, interferon type I signatures and Epstein-Barr virus genome in bone marrow and clinical outcome (ref:2).
Objectives Analysis of the autoantibody repertoire from RA patients under RTX treatment on human genome protein macroarrays and it's correlation to clinical DAS28 response.
Methods Screening of RA sera was conducted on 37.830 unique human expression clones with sera taken before and 24 weeks after treatment. The autoantibody response of different immunoglobulin classes IgD, IgA, and IgG was recorded and bioinformatically evaluated. Response was determined according to DAS28 criteria. DAS 28 scores in the responder group before treatment was from 5.4 – 7.8 and in the non-responder group 5,6 – 6,8. We analyzed 26 RA patient sera (9 responder, 7 non-responder and 10 patients with blinded response classification) investigated the data of found autoantigens in silico and by hierarchical clustering.
Results In the cohort of 26 patients 1292 different autoantigens (IgD,IgA,IgG) were detected. Using protein array we investigated clusters of autoantigen responses that disappeared or developed during RTX treatment of RA patients. Post treatment developing responses against new autoantigens can be correlated to mRNA tissue expression data and were interestingly found organ-specifically expressed. RA autoantigenic patterns before and 6 month after RTX treatment were patient-specific and no relevant autoantigenic cluster was found that was shared between patients or associated with response. However, RTX reduced the repertoire of autoantibodies after 24 weeks of treatment in the tested RA patient cohort on average by 60%. RA patients which do not respond are generating on average 63% new autoantibodies. In good responders to RTX only 5,5% (±3%) new autoantibodies can be detected. The IgA and IgG autoantibody repertoire in the serum after 24 weeks of RTX treatment is reduced (IgA: 41%, IgG: 31%) in good responders whereas it is increased (IgA: 1,3%, IgG: 24%) in non responders to RTX.
Conclusions RA patients present highly individual autoantibody profiles before and after RTX treatment, which do not reveal antigen-specific patterns for outcome prediction. However, after 6 month of RTX treatment the autoantibody repertoire in all good responding RA patients is reduced and non responders to RTX change their autoantibody repertoire directed against new but patient specific antigens. The fast rebuilding of functional B cells might be a underlying mechanistic difference between responders and non-responders to rituximab.
Ann Rheum Dis. 2005 Feb;64(2):246–52.
Arthritis Res Ther. 2012 Apr 27;14(2):R95
Disclosure of Interest None declared