Background Tumor necrosis factor (TNF) inhibitors are effective treatments for rheumatoid arthritis (RA). It is known that monoclonal antibodies and soluble receptors against TNF act slightly differently. Effects of TNFi on lymphocytes are not so clear and studies are controversial. Moreover, many biologics are now available to treat patients with active RA despite methotrexate. Predictive factors of therapeutic response to these different drugs are required to help clinician to choose which treatment should be prescribed to which patient.
Objectives The present study aimed to compare the effect of monoclonal antibodies and soluble receptors TNFi on lymphocytes and to determine predictive factors of therapeutic response to TNFi based on lymphocyte phenotype at baseline.
Methods To be included, patients should met ACR/EULAR criteria, require the introduction of TNFi and have corticosteroid doses below 10 mg per day, stable for more than a week. Extra-cellular staining with antibodies directed against CD5, CD19, CD24, CD27, CD38, IgD, IgM, CD3, CD4, CD8, CD25, CD127, CD56, CD16, at baseline and 3 months. TNF intracellular staining was performed on PBMC of 10 RA patients after stimulation with CpG, ionomycine and PMA.
Results Twenty patients were included. Nine were treated by etanercept (ETN), 9 by certolizumab pegol and 2 by adalimumab. The percentage of B cells significantly increased under TNFi from (median [IQR 25-75]) 4.6 (3.5-6.7) to 7.6 (5.2-9.9) % of lymphocytes. No change was observed in the different subtypes of B cells. However, in patients treated with ETN, IgD-CD27- double negative memory B cells significantly increased from 4.6 (2.5-5.4) to 7.7 (6.2-11.0)(p=0.03). The variation of those double negative B cells were significantly different from those observed with monoclonal antibodies (+1.6 [0.0-5.4] vs 0.3 [-1.3-1.8]% of B cells, p=0.02). No change of T, NK, NKT cells was observed in either group. EULAR responders at 3 months had significantly higher percentage of CD27+ memory B cells at baseline (32.9 [25.2-40.6] vs 19.5 [12.3-19.6]% of B cells, respectively; p=0.02), especially IgD+CD27+ pre-switch memory B cells (19.3 [9.8-21.8] vs 5.9 [4.9-9.4]% of B cells, respectively; p=0.02). Since memory B cells were not decreased after treatment, the hypothesis to explain that it could predict respond was that memory B cells are an important source of TNF. To validate this theory, we analyzed the production of TNF by B cells in culture and compared memory B cells with others B cell subtypes in 10 RA patients. Indeed, memory B cells were more often positive for TNF than non-memory B cells (23[7-47] vs 19 [1-31]% of B cells; p<0.05).
Conclusions No major change in terms of lymphocytes was seen after 3 months of TNFi. Memory B cells produce more TNF than other B cell subsets and the percentage of memory B cells before starting a TNFi could help to predict response.
Acknowledgements To the french society of rheumatology, UCB (France) Sirius grant and Pfizer (France) Passerelle grant
Disclosure of Interest None Declared
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.