Background We previously showed that regulatory IL-10 producing B cells (B10) were decreased in patients with rheumatoid arthritis (RA) and that B10 cells were inversely correlated with disease activity (DAS28) and with rheumatoid factor levels. We also previously showed that CD24hiCD38hi and CD24hiCD27+ B cell are the subsets that are most likely to produce IL-10 after stimulation with CpG.
Objectives In the present study, we aimed to analyze if these CD24hiCD38hi and CD24hiCD27+ B cells have altered functions in RA patients and if B10 cells decrease can be corrected by TNF inhibitors (TNFi) or tocilizumab (TCZ) therapies.
Methods CD24hiCD38hi, CD24hiCD27+ and CD24lo B cells and CD4+CD25– T cells were sorted by cytometry. B cells subsets and naïve T cells were co-cultured (1:1) with anti-CD3 in 8 controls and 6 RA patients (with less than 10 mg/d of steroids and no current biological therapy, controls being matched on age and gender). Cells were collected after 72 hours and regulatory T cells were assessed using anti-CD4, anti-CD25, anti-CD127 antibodies and DAPI. B10 cells were generated from PBMCs stimulated for 24 hours with CpG and 4 hours with PMA and ionomycine. Intra-cellular B cell IL-10 was assessed by cytometry. B10 cell measures were performed in 31 patients before introduction of therapy and 3 months after (18 patients treated with TNFi and 15 with TCZ).
Results RA patients tended to have more Tregs in response to anti-CD3 alone than controls (median [IQR25-75]: 6.5 [3.2-13.4] vs 2.9 [0.6-3.6], p=0.100). CD24hiCD27+CD19+ and CD24hiCD38hiCD19+B cells of the healthy controls induced more Tregs than CD24loCD19+ (4.1 [1.5-5.4] and 3.9 [1.4-5.0] vs 1.9 [0.9-3.7]; respectively with p=0.014 and p=0.058). Conversely, CD24hiCD27+CD19+ and CD24hiCD38hiCD19+ of RA patients failed to induce more Tregs than CD24loCD19+ (5.4 [2.5-12] and 7.6 [2.6-15.6] vs 7.6 [4.2-16.1], respectively with p=0.40 and p=0.66). Variations in the levels of Tregs induced by CD24hiCD27+CD19+ and CD24hiCD38hiCD19+ compared to those induced by CD24lo were significantly higher in controls than in RA patients (+60.7 [37.2-107.4]% vs -28 [-44.5–19.2]% for CD24hiCD27+/CD24lo in controls and in RA patients respectively, p=0.002; and +47.4 [16.2-160.4] vs -24.3 [-43.1–10.1] CD24hiCD38hi/CD24lo in controls and in RA patients respectively, p=0.01). No significant change in B10 cells was observed between baseline and after 3 months treatment (mean change ± SD 1.18±4.45%, p=0.15). The variations were not significantly different between EULAR responders and non-responders (0.95±4.59 vs 2.32±4.43%; p=0.52). No significant correlation was found between variations of DAS28 and change in B10 cells between baseline and 3 months (r=0.30; p=0.11). The variations of B10 cells were not significantly different between patients treated with TNFi and TCZ (1. 36±4.75 vs 0.94±4.16%; p=0.48).
Conclusions Conversely to healthy subjects, regulatory B cell precursors cannot convert CD4+CD25- T cells into regulatory T cells in patients with RA. Biological therapies do not restore IL-10 secretion capacities of B cells, suggesting that this impairment is constitutive in RA patients.
Acknowledgements Unrestricted grant PASSERELLE (Pfizer, France)
French Society of Rheumatology
Disclosure of Interest None declared