Background In a transcriptomic study, we found that whole blood interleukin-33 (IL-33) mRNA expression is increased in patients with rheumatoid arthritis (RA) responding to rituximab (RTX) (1). Using one ELISA IL-33 kit (DuoSet, R&D) for protein assessment, we previously reported an association between serum IL-33 level and RTX response in a single cohort, but additional experiments have recently raised caution about accuracy of this kit for sera measurements (2). Recommendations from the manufacturer advice to use for human sera another assay (Quantikine, R&D).
Objectives To replicate in different cohorts and with a more accurate ELISA assay the possible association between detectable serum level of IL-33 and response to RTX in RA patients.
Methods Serum IL-33 level and B-cell biomarkers known to be associated with RTX response (rheumatoid factor [RF], anti-citrullinated cyclic peptide [anti-CCP] antibodies, high serum IgG level above upper limit normal >12.7 g/L) were assessed in 111 RA patients receiving a first course of 2 grams RTX in the SMART cohort (85% of patients with RF or anti-CCP) as well as, for replication, in 74 RA patients (including 32 patients from Leeds, UK and 42 from Clermont-Ferrand, France) treated with the same schedule in routine care (97% of patients with RF or anti-CCP). We performed uni- and multivariate analyzes to identify factors associated with a EULAR response at 24 weeks.
Results At week 24, 84 (76%) and 54 (73%) patients reached EULAR response in SMART and in replication cohort, respectively. Serum IL-33 was detectable in 27/111 (24%) and 35/74 (47%) patients from SMART and replication cohort, respectively. In SMART, multivariate analysis indicated that auto-antibodies status (positive RF or anti-CCP) were associated with RTX response (OR 3.47 95%CI 1.04–11.60; p=0.04). The same trend was observed for high serum IgG (OR 2.82 95%CI 0.97–8.14; p=0.056) but not for serum IL-33. In the replication cohort, detectable serum IL-33 was independently associated with RTX response (OR 3.73, 95%CI [1.12–12.38]; p=0.03) with the same trend for auto-antibodies status (OR 32, 95%CI [0.69–999]; p=0.08), but without significant result for high serum IgG. When we combined both cohorts (n=185), the 3 parameters were associated with RTX response in multivariate analysis: presence of RF or anti-CCP (OR 3.27 95%CI 1.13–9.46; p=0.03), high serum IgG (OR 2.32, 95%CI 1.01–5.33; p=0.048) and detectable IL-33 (OR 2.40 95%CI 1.01–5.72; p=0.047). There was synergy between the 3 predictive factors since detection of IL-33 in patients having positive auto-antibodies and high IgG level improved likelihood of response (100% EULAR responders) compared with patients having only positive auto-antibodies and high IgG level (77% responders) or compared with those without any of these factors (55% responders) (Figure 1).
Conclusions Using an accurate ELISA kit, serum IL-33 may predict clinical response to RTX, independently of and synergistically with auto-antibodies and serum IgG level
Sellam J A Rheum 2014
Rivière E Ann Rheum Dis 2015
Disclosure of Interest None declared