The plasma metabolome of juvenile idiopathic arthritis varies according to subtype and underlying inflammatory status

Background: Juvenile idiopathic arthritis (JIA) is challenging to classify and effectively monitor due to the lack of disease- and subtype-specific biomarkers. A robust molecular signature that tracks with specific JIA features over time is urgently required, and targeted plasma metabolomics may reveal such a signature. The primary aim of this study was to characterise the differences in the plasma metabolome between JIA patients and non-JIA controls and identify specific markers of JIA subtype. We also assessed the extent to which these signatures are due to underlying inflammation as assessed by glycoprotein acetyls (GlycA) and high-sensitivity C-Reactive Protein (hsCRP) levels.

Methods: Targeted nuclear magnetic resonance (NMR) metabolomic profiles of plasma of 72 children with JIA and 18 controls were assessed cross-sectionally. Associations between 71 metabolomic biomarkers and JIA, JIA subtype, disease activity status, and inflammation markers (GlycA and hsCRP) were assessed using multivariable linear regression models.

Results: JIA was associated with higher GlycA (mean difference = 0.93 standard deviations, 95% confidence interval = [0.370, 1.494], P_adj = 0.039) and docosahexaenoic acid (1.06, [0.51, 1.60], P_adj = 0.021), and lower acetate (-0.92, [-1.43, -0.41], P_adj = 0.024) relative to controls. This variation was largely driven by systemic JIA (sJIA), with 24 of 71 total biomarkers significantly different (P_adj <0.05) relative to controls. There were no specific differences identified in oligoarticular (oJIA) or polyarticular (rheumatoid factor positive or negative) JIA relative to controls. Despite being generally highly correlated with hsCRP (r > 0.70), GlycA, but not hsCRP, was positively associated with active disease in sJIA (0.22, [-0.40, -0.04], P_adj = 0.018), and 6 of 24 sJIA-associated markers were associated with GlycA levels. Only 1 sJIA-associated biomarker, histidine, was associated with hsCRP levels.

Conclusion: Differences in the plasma NMR metabolomic profiles are apparent in children with sJIA, but not other JIA subtypes, relative to non-JIA controls. These findings suggest a potential utility for classifying and monitoring JIA through metabolomic profiling, with chronic inflammation, measured by GlycA, potentially playing a role in at least some of these metabolomic differences.