Statistics from Altmetric.com
Rituximab treats rheumatoid arthritis (RA) by temporarily depleting B cells. Initial clinical response and time to relapse after B cell repopulation are variable. Using highly sensitive flow cytometry we have shown that good clinical response is predicted by rapid complete B cell depletion (below 0.0001 cells × 109/l) and lower levels of memory and preplasma cells at baseline. There remains a need for more practically applicable biomarkers predicting response. B cell activating factor (BAFF) promotes the differentiation, activation and survival of B cells, is increased in RA serum and has been shown to inhibit the killing of B cells by rituximab in animal models.
To determine whether: (1) BAFF levels in RA can be used to predict clinical response; (2) BAFF levels affects B cell depletion; (3) BAFF is associated with earlier B cell repopulation.
30 RA patients treated with rituximab. Clinical response was assessed by EULAR criteria. BAFF levels were measured by ELISA (R&D Systems).
Prediction of response: Patients with non-response had significantly lower serum BAFF at baseline (p=0.15). Baseline: In RA compared to controls, BAFF levels were high (1.48 [0.62–4.43] vs 0.78 [0.32–5.67] pg/l) while total B cell counts were low (median [range]: 0.1228 [0.0261–0.3294] vs 0.1686 [0.1316–0.3088] cells × 109/l). Levels of BAFF negatively correlated with memory B cell numbers (r=0.570, p=0.1). Depletion: Contrary to expectations, higher baseline BAFF levels were not associated with incomplete depletion. Indeed, a trend for lower baseline BAFF levels in patients with incomplete depletion after rituximab was observed (p=0.07). After treatment, BAFF levels significantly increased. This related to the baseline level (r=631, p<0.01) rather than the extent of B cell depletion. Repopulation: There was no relationship between BAFF levels at 0, 3 or 6 months and rate of B cell repopulation or time to relapse.
Patients that respond best (and deplete most efficiently) after rituximab are those in whom baseline BAFF is high and memory and plasmablastic cells are low. In this study BAFF levels did not affect B cell depletion nor determine the rate of B cell repopulation. Instead, high levels of BAFF at baseline related to low numbers of memory cells, which were themselves predictors of clinical response. It must be noted that these relationships may differ in the synovium and future work will address this. Although it appears secondary to B cell numbers, baseline BAFF level may be a more convenient tool to predict clinical response to rituximab.