Background and Objectives Rituximab induces clinical remission in a majority of seropositive patients with Rheumatoid arthritis (RA), however, all patients eventually relapse. Although B cell return precedes clinical relapse absolute B cell numbers are not predictive of relapse. The aim of this study was to determine whether there were any consistent dynamic changes in combinations of serological parameters following up to 3 cycles of B cell depletion therapy based on rituximab (RTX).

Materials and Methods We included 23 RA patients (all clinical responders in each cycle) over 1, 21 over 2, and 15 over 3 cycles of RTX followed for up to 108 months and analysed in relation to B cell kinetics at 4 key points within each cycle of therapy. Serum analytes including Rheumatoid factors (RhF) and anti-citrullinated protein/peptide antibodies (ACPA), BAFF, serum free light chains (SFLC), soluble CD23 (sCD23 - cleaved from naïve B cells coincident with CD27 expression), antibodies to tetanus toxoid (TT) and pneumococcal capsular polysaccharide (PCP) were measured by ELISA at 4 key points in each cycle: Baseline (pre-RTX in each cycle); when B-cell depleted (CD19⁺ B-cells<5/ml); at B-cell return (CD19⁺ B-cells ≥5/ml); and at clinical relapse (ΔDAS28>1.2).

Results Comparing levels of analytes at relapse after 3 cycles with baseline values before 1^st Cycle, BAFF levels were higher, anti-microbial remained relatively unchanged and IgM-RhF and IgM-CCP and IgG-CCP autoantibodies had decreased (p<0.05). The most consistent changes between baseline to B cell depletion in each cycle, over all 3 cycles, were in lSFLC, sCD23 and IgM-RhF which fell and BAFF levels which rose (p<0.05). Incremental rises in sCD23 levels in Cycles 2 and 3 were associated with time to relapse. In patients with relapse >5 months after B-cell return, significant rises in IgM-RhF, lSFLC and sCD23, and falls in BAFF, occurred between B cell return until clinical relapse.

Conclusions Dynamic changes in the serum analytes IgM-RhF, λSFLC, sCD23 and BAFF were consistent over 3 cycles of rituximab and closely associated with clinical response/relapse. This suggests that the clinical response to RTX involves interuption of the genesis of short-lived immunoglobulin-secreting cells. Resumption of this process at relapse involved maturation of B-cell clones associated with sCD23 cleavage and of immunoglobulin-secreting cells accompanied by lSFLC release and IgM-RhF specificity. In patients with the longer periods of remission after B cell return the ‘delay’ in re-starting the inflammatory process may be due to a slower rate of selection and/or maturation of autoreactive B cell clones, despite raised serum BAFF.