Integrated serum proteomics and autoantibody analyses reveal a biomarker signature predictive of flare during biologic tapering in rheumatoid arthritis

BackgroundRheumatoid arthritis (RA) is a chronic immune-mediated inflammatory disease characterized by a heterogeneous clinical course with periods of remission and flare. Although biologic DMARDs (bDMARDs) have revolutionized RA treatment by enabling sustained disease control, their long-term use is associated with adverse effects and high costs, making dose tapering an attractive but clinically challenging strategy. The lack of reliable biomarkers to predict flare risk limits safe implementation of treatment de-escalation. This study aimed to identify novel circulating protein biomarkers associated with flare risk in RA patients undergoing bDMARDs tapering, useful to enable biomarker-guided treatment optimization strategies. MethodsA discovery proteomic analysis using mass spectrometry was performed on baseline serum samples from a subset of the OPTIBIO clinical trial (n=44), followed by validation in the full cohort (n=194) using ELISA. Functional pathway analysis explored biological processes associated with candidate biomarkers. In parallel, anti-cytokine autoantibodies were profiled using multiplex immunoassays. Logistic and Cox regression models were used to assess associations with flare risk. Predictive models integrating biomarkers and clinical variables were evaluated using receiver operating characteristic (ROC) analysis, sensitivity and specificity metrics, and decision curve analysis to assess clinical utility. ResultsMass spectrometry identified 806 proteins, of which 87 were differentially expressed at baseline between patients who flared and those who maintained remission during follow-up within the intervention (tapering) arm. Functional enrichment analysis highlighted immune-regulatory and innate immune pathways. Among the candidates, V-set immunoglobulin-domain-containing 4 (VSIG4) was validated as a biomarker associated with increased flare risk. Anti-interferon-{gamma} (anti-IFN{gamma}) autoantibodies were also associated with flare. A combined model including VSIG4, anti-IFN{gamma}, and the clinical variable DAS28-CRP improved predictive performance compared with clinical variables alone (AUC 0.76 vs 0.66), achieving significantly higher sensitivity. Decision curve analysis demonstrated higher net benefit of the combined model, indicating improved clinical decision-making. In a secondary analysis focused on patients with prolonged remission, representing the most suitable candidates for safe treatment tapering, the model performance further improved (AUC 0.84). ConclusionIntegration of novel serum proteomic and autoantibody biomarkers with clinical parameters improves prediction of flare during biologic tapering in RA and provides clinically relevant benefit for patient stratification. These findings support further development of biomarker-driven approaches for personalized treatment optimization strategies.