Sensitive B-cell receptor repertoire analysis shows repopulation correlates with clinical response to rituximab in rheumatoid arthritis

Abstract

Background: Although B-cell depleting therapy in rheumatoid arthritis (RA) is clearly effective, response is variable and does not correlate with B cell depletion itself.

Methods: The B-cell receptor (BCR) repertoire was prospectively analyzed in peripheral blood samples of twenty-eight RA patients undergoing rituximab therapy. Timepoints of achieved BCR-depletion and -repopulation were defined based on the percentage of unmutated BCRs in the repertoire. The predictive value of early BCR-depletion (within one-month post-treatment) and early BCR-repopulation (within 6 months post-treatment) on clinical response was assessed.

Results: We observed changes in the peripheral blood BCR repertoire after rituximab treatment, i.e., increased clonal expansion, decreased clonal diversification and increased mutation load which persisted up to 12 months after treatment, but started to revert at month 6. Early BCR depletion was not associated with early clinical response but late depleters did show early response. Patients with early repopulation with unmutated BCRs showed a significant decrease in disease activity in the interval 6 to 12 months. Development of anti-drug antibodies non-significantly correlated with more BCR repopulation.

Conclusion: Our findings indicate that rather than BCR-depletion it is repopulation with unmutated BCRs, possibly from naïve B cells, which induces remission. This suggests that (pre-existing) differences in B-cell turnover between patients explain the interindividual differences in early clinical effect.

Keywords: AIRR-seq; B cells; B-cell receptor repertoire; Rheumatoid arthritis; Rituximab.

Fig. 1

Peripheral blood BCR repertoire in patients undergoing rituximab treatment. Boxplots showing (A) clonal expansion: Gini index, (B) clonal diversity: Shannon index and (C) the average IGHV gene mutation load in samples obtained before (M0), and at one (M1), three (M3), six (M6) and twelve (M12) months after treatment with rituximab. Boxplots show the median, 25th and 75th interquartile, error bars show the range, and single data points are depicted in grey (*p ≤ 0.05, ***p ≤ 0.001, using one-way ANOVA)

Fig. 2

Monitoring of changes in the peripheral blood BCR repertoire during treatment with rituximab. A Barplot showing the percentage of unmutated clonotypes in samples obtained before (M0), and at one (M1), three (M3), six (M6) and twelve (M12) months after treatment with rituximab. Bars height shows the median, error bars show the range. Single data points are depicted in grey (**p ≤ 0.01, ***p ≤ 0.001, using one-way ANOVA). B Disease activity score (DAS28: red dots and line, red scale on the right Y-axis) and percentage of unmutated clonotypes (white bars, scale on the left Y-axis) in the individual patients. Grey arrows indicate the first post-BCR depletion timepoint, while black arrows indicate the first post-BCR repopulation timepoint. Green triangles indicate the timepoint at which each patient received rituximab treatment. C Percentage of patients that reached the first post-BCR depletion (grey) and first post-BCR repopulation timepoint (black) at the timepoints analyzed

Fig. 3

Timing of BCR depletion does not predict long-term clinical outcome but correlates with short-term disease activity. A-B Boxplot showing the changes in DAS28-score between month 1 (A) or month 3 (B) post-treatment and baseline in early or late depleting patients. Boxplots show the median and 25th and 75th interquartile, error bars show the range, and single data points are depicted in grey (*p ≤ 0.05, using Mann-Whitney test in A and unpaired t-test in B)

Fig. 4

BCR repopulation is associated with improvement of disease activity shortly after and possibly with anti-drug antibodies development. A Boxplot showing the changes in DAS28-score between the month 6 and month 12 after treatment in early or late  BCR repopulating patients. Boxplots show the median and 25th and 75th interquartile, error bars show the range, and single data points are depicted in grey (** p ≤ 0.01, using unpaired t-test). B Barplot showing the correlation between anti-drugs antibodies (ADA) development and repopulation within 12 months of treatment