Background and purpose Seropositivity for Rheumatoid factors and for antibodies to citrullinated (Cit) protein antigens (ACPA) is the strongest predictor for clinical response to Rituximab (rituximab) in rheumatoid arthritis (RA). The aim of this study was to follow fluctuations of IgG- and IgA-ACPA to multiple individual citrullinated protein epitopes in relation to clinical improvement and resumption of symptoms after rituximab.
Methods 16 patients with severe, active RA (DAS28 ≥ 5.1) undergoing initial cycles of rituximab were included. All achieved peripheral B-cell depletion (<5 CD19+cells/μl) and clinically responded (ΔDAS28 ≥ 1.2) within 5 months. Follow-up to disease flare (Relapse) was available in 11/16 patients. IgG- and IgA-ACPAbinding to a custom, bead-based, antigen array comprising 30 Cit-antigens and 22 corresponding native antigenswas assessed using mean fluorescence intensity (MFI). Data was Z-normalised to compensate for variability in number of autoantibody binding sites per antigen.
Results At Baseline, MFI of IgG- and of IgA-Cit antibodies were strongly correlated (R2 = 0.75; p < 0.0001). Between Baseline and Depletion phase (median 3 months), IgG- and IgA-ACPA showed mean% decreases of -21.4% and -30.2% respectively. In 11 patients followed up to Relapse, ACPA fluctuated with different patterns. Of ACPAinitially decreasing following rituximab, 30/52 IgG- and 31/48 IgA-ACPA then rose before Relapse with reactivity largelyshared by IgA- and IgG-ACPA. 24% (16/68) of IgG-Cit and 6% (3/51) IgA-Cit, unexpectedly showed a transient increase following rituximab, subsequently decreasing to pre-treatment levels at Relapse. ACPA in different sera recognising the same epitopes were not linked to particular patterns of response. Putative ‘new’ specificities (Z-scores: Baseline <1 and Relapse ≥1)were rare, with most detectable at low levels pre-rituximab.
Conclusion Dissection of the ACPA response after rituximab using multiplex bead array revealed diverse kinetics of autoantibody production and followed several distinct patterns, largely un-related to Cit-epitope specificity. At Relapse, rising levels of ACPA with the same specificities as those at baseline suggested re-expansion of residual Cit-specific memory B cells. As B cell return is mandatory for relapse, and with no marked presence of new autoantibody specificities, potential interactions between newly generated and residual B cells could be exploited for more efficient B cell directed therapies.