Background/Purpose: Systemic onset juvenile idiopathic arthritis (sJIA) is an autoinflammatory disease, characterized by fever, rash and arthritis. The IL-1 and IL-6 pathway are crucial in sJIA, exemplified by excellent responses to IL-1 and IL-6 blocking therapy. However, so far it has not been feasible to base treatment choice on individual inflammatory characteristics of patients. Moreover, we cannot predict which patients can safely stop therapy in inactive disease. Here, we aim to identify biomarkers that predict which sJIA patients will respond to IL-1 blockade and which patients will sustain clinical remission when IL-1 blockade is stopped in clinically inactive disease.

Methods:

All patients were treated with recombinant human IL-1 receptor antagonist (rhIL1RA) as first-line therapy. If the response to rhIL1RA was insufficient, either low dose corticosteroids were added or the patients were switched to other biologicals. If patients acheived clinically inactive disease at time point 3 months after start of rhIL1RA, we attempt to taper and stop rhIL1RA. Biomarker discovery was performed in the serum of patients at onset (before start of rhIL1RA) and at time point 3 months in patients with clinically inactive disease on rhIL1RA, using cytokine analysis (Luminex multiplex assay), determination of S100A12 and Mrp8/14 (ELISA) and miRNAs (qPCR). We used Mann-Whitney U tests, with false discovery rate (FDR) correction by Benjamini-Hochberg methodology.

Results: In total, serum of 19 patients at disease onset and 20 patients at inactive disease was available. Fifty-five proteins and 10 miRNAs were significantly different between patients at onset and at inactive disease after FDR-correction. Proteins included known biomarkers as S100A12, Mrp8/14, IL-6 and IL-18, but also novel markers including metalloproteinases and angiogenesis-related proteins.

Within the onset patients, thirteen patients had a complete response with rhIL1RA monotherapy (CR), while five patients needed additional therapy/switch of therapy (NR) because of persistent fever (n=2), arthritis (n=2) or both fever and arthritis (n=1). One patient had a partial response on rhIL1RA monotherapy. On univariate (uncorrected) testing, 3 proteins and 35 miRNAs were significantly different between CR and NR. After FDR-correction none remained significant.

In the stop-prediction part of this study, ten patients flared during tapering or stopping therapy, while nine patients remained in remission. The PCA of the Luminex data showed that patients that will flare cluster separately from patients that will maintain clinical remission. On univariate analysis, 16 proteins and 2 miRNAs were significantly different, however none were significant after FDR correction.

Conclusion:  This study aimed to identify novel biomarkers for the prediction of treatment response to rhIL1RA in new-onset sJIA patients. Although in our small cohort no markers tested significantly after FDR correction, our results provide potentially interesting novel biomarkers that need further modelling and validation.We recently started a prospective study optimising our stop-strategy of rhIL1RA, providing the opportunity to validate the predictive value of suggested biomarkers.