Integrated proteomic, genetic, and machine-learning analyses reveal a CXCL11-centered inflammatory signature associated with systemic activity in Sjögren’s disease

Objective Sjögren’s disease (SjD) exhibits substantial clinical heterogeneity, and molecular biomarkers for systemic activity stratification remain limited. This study aimed to identify inflammatory protein signatures associated with systemic activity in SjD and evaluate their potential value for disease stratification and dynamic monitoring. Methods A multi-stage biomarker discovery and validation framework integrating Olink targeted proteomics, enzyme-linked immunosorbent assay (ELISA), clinical correlation analysis, machine-learning-based feature selection, Mendelian randomization (MR), and longitudinal assessment was established. Plasma inflammatory proteins were profiled using the Olink Target 96 Inflammation panel in the discovery cohort, and differentially expressed proteins (DEPs) were identified. Candidate proteins were subsequently validated by ELISA in an independent cohort and evaluated for associations with clinical and immunological indices. LASSO, support vector machine-recursive feature elimination (SVM-RFE), and Random Forest algorithms were applied to identify core protein features. MR analysis was performed to assess the potential genetic relevance of prioritized inflammatory proteins to SjD susceptibility. Longitudinal changes in candidate biomarkers were further evaluated over a 24-week follow-up period. Results Olink proteomic profiling identified DEPs associated with systemic disease activity, with enrichment analyses highlighting cytokine–cytokine receptor interactions and chemokine signaling. ELISA validation confirmed differential expression of candidate inflammatory proteins between higher systemic activity (HSA) and low systemic activity (LSA) patients, and several proteins showed significant correlations with ESSDAI and immunological indices, including ESR, IgG, and RF. Machine-learning analyses consistently prioritized CXCL10 and CXCL11 as core protein features distinguishing HSA from LSA patients. MR analysis further provided genetic evidence supporting the potential relevance of CXCL11 to SjD susceptibility. During the 24-week longitudinal follow-up, plasma CXCL11 levels demonstrated dynamic changes and were significantly associated with changes in systemic activity-related parameters. Conclusion This study identifies a CXCL11-centered inflammatory signature associated with systemic activity in SjD. By integrating targeted proteomics, machine learning, genetic evidence, and longitudinal validation, our findings highlight CXCL11 as a promising candidate biomarker for molecular stratification and dynamic monitoring of systemic inflammation in SjD.