Background Inhibition of IL-6 receptor with tocilizumab (TCZ) is an effective strategy for treating rheumatoid arthritis (RA) patients. However, the effect of TCZ on lymphocytes remains poorly studied. Regulatory T cells (Tregs) and NK cytotoxicity are known to be impaired in RA.

Objectives The objective of the present study was to analyze the effect of TCZ on B, T, NK and NKT cells and to find predictive factors of response.

Methods To be included, patients should meet ACR/EULAR criteria, require the introduction of TCZ, have corticosteroid doses below 10 mg per day, stable for more than a week and should not have received rituximab in the previous year. Extra-cellular staining with antibodies directed against CD19, CD24, CD27, CD38, IgD, CD3, CD4, CD8, CD25, CD127, CD56, CD16, g9d2 and intra-cellular staining of Granzyme B and Perforine A were performed after ficoll at baseline, 3 and 6 months. IL-10 was measured in sera by ELISA.

Results Twenty-three patients requiring a treatment with TCZ as well as 31 healthy controls were included. Patients with RA had significantly lower percentage of Tregs as defined by CD4+CD25HiCD127lo (3.8 [2.7-4.9] vs 4.8 [3.5-6.0]% of CD4+ lymphocytes; respectively, p=0.007). After 3 months of TCZ, Tregs significantly increased (+1.27±1.00% of CD4+ cells; p<0.001), meeting normal values. Tregs and IL-10 were correlated at baseline (r=0.52; p=0.01) and variations of Tregs were correlated with variations of IL-10 after 3 months of TCZ (r=0.69; p<0.01). No change in percentage of B, NK and NKT cells subsets was observed. Patients in EULAR remission at 3 months had higher percentage of CD3-CD56+ NK cells at baseline (10.4±7.1 vs 3.6±3.1% of lymphocytes; p=0.01), especially CD56BriCD16- NK cells (0.88±0.68 vs 0.28±0.28% of CD3- cells; p=0.04) than patients with a DAS28 at 3 months above 2.6. NK cells were more frequently positive for perforine A and tended to be more frequently positive for granzyme B after TCZ treatment (+6.8 ±2.9% of CD3-CD56+ cells; p=0.02 and +7.1 ±3.8% of CD3-CD56+ cells; p=0.06, respectively).

Conclusions TCZ increased the number of phenotypic Tregs. Significant correlation between variation of IL-10 and variations of Tregs suggest than these Tregs are probably functional, at least in terms of IL-10 production. TCZ also seem to restore NK cells cytotoxicity, known to be impaired in RA. Values of NK cells at baseline could help to predict remission after 3 months of treatment.

Acknowledgements To the French Society of Rheumatology and Roche-Chugaï (France) for their financial support

Disclosure of Interest: None Declared