Mechanisms of relapse in rheumatoid arthritis following B-lymphocyte depletion

An open label trial of B-lymphocyte depletion therapy (BLDT) based on rituximab in 5 patients with rheumatoid arthritis, initiated in 1998, had shown that this therapy was associated with major clinical improvement. B-lymphocyte repopulation had occurred before relapse in all patients and, interestingly, it had not been always associated with relapse. This thesis starts with an extension of the initial open label trial to 22 patients with results suggesting a dose response to rituximab. Further follow up of these and an additional 18 patients revealed that, although all patients eventually relapsed, as in the previous small study, relapse could occur either at the time of B-lymphocyte repopulation of the peripheral blood or at a variable time after this time point, up to almost 3 years. Re-treatment with BLDT was found to be well tolerated and effective. Upon re-treatment, patients usually repeated the same pattern of relapse. Disease-associated autoantibodies, rheumatoid factor of immunoglobulin A, G and M classes and antibodies to cyclic citrullinated peptides, fell in all patients following BLDT but only decreased significantly in those who responded to treatment. Comparison with changes observed in total immunoglobulin levels and anti-microbial antibodies, suggested a specific effect on autoantibodies. Relapse was almost always preceded or associated with a rise in autoantibody levels, particularly rheumatoid factor of immunoglobulin M isotype, and was more closely associated with a rise in autoantibody levels than with the presence of B lymphocytes per se. B-lymphocyte stimulator (BLyS) serum levels increased following BLDT. Levels decreased following B-lymphocyte repopulation but tended to be lower at the time of repopulation in the group of patients who relapsed at this time point. Immunophenotyping of peripheral blood B lymphocytes showed that depletion was major but not complete and that it involved all B-lymphocyte subpopulations. It suggested that a quantitative threshold of depletion needed to be reached for patients to respond to therapy. B-lymphocyte repopulation occurred mainly from naive cells with immature B lymphocytes present in increased numbers and frequency. Patients who relapsed at the time of B-lymphocyte repopulation tended to show increased frequency and increased numbers of circulating memory B lymphocytes at the time of repopulation. Immunophenotyping studies of bone marrow aspirates in 6 patients, 3 months after treatment with BLDT, showed that the frequency of cells of B-lymphocyte lineage at this time point varied between patients. The variable relationships between the frequencies of more immature and more mature B-lymphocyte precursors in the samples suggested different degrees of depletion. Samples from patients who relapsed the earliest showed a pattern suggesting less efficient depletion. In conclusion, the results presented here suggest that relapse following BLDT in patients with rheumatoid arthritis can be attributed to incomplete depletion of pathogenic B-lymphocyte clones, persistence of long-lived plasma cells producing pathogenic autoantibodies or to a combination of both. This will contribute to hypothesis-based development of B-lymphocyte targeting therapies.