FRI0247 Serum level of interleukin 33, a novel independent predictive biomarker of clinical response to rituximab in rheumatoid arthritis: results from the smart trial

Background Interleukin-33 (IL33) is a recently identified member of the interleukin-1 family that binds to its ST2 receptor expressed on various cells including mast cells, T-cells and also B1-cells in mice ([1][1]). IL-33 may activate B-cell, immunoglobulin secretion (IgA and IgG) and induce immunoglobulin class switching. We previously demonstrated in a whole-blood transcriptomic analysis performed in 69 patients (pts) with rheumatoid arthritis (RA) included in the SMART trial that baseline IL-33 mRNA upregulation was associated with subsequent clinical response to rituximab (RTX) ([2][2]). Objectives To investigate whether baseline serum level of IL-33 may predict response to RTX in RA. Methods Before treatment, serum level of IL-33 using ELISA and of B-cell activation biomarkers (rheumatoid factor [RF], anti-CCP antibodies, free light chains, IgG, IgA, IgM, and BAFF levels) were assessed in 205 RA pts participating to the SMART study and treated by a first course of RTX (1g day 1 and 15). Association between IL-33 level and RA features were determined. Results are presented in median and interquartiles. Uni and multivariate (clinical features, auto-antibodies status and IL33 level) analyses were performed to identify factors associated with a EULAR response at 24 weeks. Results Serum IL-33 level was higher in pts with anti-CCP (anti-CCP+: 359 [87; 1619] vs anti-CCP-: 102 [39; 368]; p=0.001) and in those with RF (RF+: 365 [95; 1677] vs RF-: 131 [38; 602]; p=0.001). IL-33 level was not modulated by anti-TNF prior exposure (n=140) or current steroid use (n=161) (anti-TNF +: 216 [73; 978] vs anti-TNF -: 359 [78; 1874] pg/mL; NS) and steroid +: 254 [70; 1348] vs steroid -: 258 [79; 1345] pg/mL; NS). Serum IL-33 level was not correlated to DAS28-CRP (r=-0.04; NS). There were 146 (71%) responders (i.e. 44 good and 102 moderate). IL-33 level was associated with EULAR response (IL33 level: 361 [87; 1654] vs 131 [46; 475] pg/mL in responders and non responders to RTX, respectively; p=0.006). After adjustment on baseline DAS28-CRP, multivariate analysis indicated that IL33 level was independently associated with subsequent response to RTX (OR [95% CI]: 1.3 [1.02; 1.6]), as well as the presence of anti-CCP antibodies or RF (OR [95% CI]: 2.7 [1.2; 6.0]). The combination of the 2 parameters further increased the likelihood of response. Conclusions Serum IL33 level is increased in RA pts with anti-CCP and/or RF. Moreover, we confirm our transcriptomic analysis and demonstrate for the first time that serum IL-33 level represents a novel simple useful biomarker predicting clinical response to RTX, independently of auto-antibodies status, which usually displays a strong impact on rituximab response. The role of the IL-33/ST2 axis in B-cell immunopathology in RA needs to be further addressed. References 1. Komai-Koma M et al. J Immunol 2011;186:2584-91 2. Sellam et al ACR congress 2011 Disclosure of Interest : J. Sellam Grant/research support from: Roche, S. Rouanet Employee of: Roche, H. Hendel-Chavez: None Declared, S. Marion-Thore: None Declared, C. Miceli-Richard: None Declared, B. Combe Grant/research support from: Roche, J. Sibilia Grant/research support from: Roche, X. Le Loet Grant/research support from: Roche, J. Tebib Grant/research support from: Roche, G. Chiocchia: None Declared, R. Jourdan Employee of: Roche, M. Dougados Grant/research support from: Roche, Y. Taoufik: None Declared, X. Mariette Grant/research support from: Roche [1]: #p-6 [2]: #p-7