A novel two-gene diagnostic signature for rheumatoid arthritis synovitis: Cross-platform validation of CXCL13 and IGHG1 across microarray and RNA-seq cohorts

Abstract

Rheumatoid arthritis (RA) is a persistent autoimmune disorder characterised by synovial inflammation and progressive joint deterioration. B-cells play a pivotal role in the pathophysiology of rheumatoid arthritis by producing autoantibodies and secreting cytokines. This study assessed the diagnostic efficacy of CXCL13 and IGHG1 as biomarkers for rheumatoid arthritis synovitis using a cross-platform in silico analysis. Five Affymetrix microarray datasets (GSE55235, GSE55457, GSE12021, GSE77298, GSE1919; total n=97) were subjected to pooled analysis, and one separate RNA-seq dataset (GSE89408; n=202) was used for independent validation. All datasets comprised synovial tissue samples from patients with RA and non-inflammatory controls. Differential expression analysis was conducted using t-tests with the Benjamini-Hochberg adjustment. Protein-protein interaction (PPI) networks were established using the STRING database. ClusterProfiler was utilised to conduct functional enrichment analysis. The diagnostic performance was assessed using receiver operating characteristic (ROC) curve analysis with 95% confidence intervals. Both genes exhibited considerable upregulation in rheumatoid arthritis synovium across all datasets. In a pooled analysis of five microarray datasets (n=97), CXCL13 attained an AUC of 0.899 (95% CI: 0.847-0.952), whereas IGHG1 reached an AUC of 0.663 (95% CI: 0.557-0.768). Independent validation in a substantial RNA-seq cohort (n=202) corroborated exceptional performance for CXCL13 (AUC = 0.906, 95% CI: 0.868–0.945) and enhanced performance for IGHG1 (AUC = 0.870, 95% CI: 0.819–0.921). PPI network analysis identified relationships between CXCL13 and inflammatory hubs (TNF, IL6) as well as between IGHG1 and B-cell receptor components (CD79B, MZB1, JCHAIN). Functional enrichment revealed B-cell receptor signalling (p = 8.2 × 10⁻⁴) and humoral immune response (p = 1.5 × 10⁻³) as the primary pathways. CXCL13 exhibits superior and reliable diagnostic efficacy across both microarray and RNA-seq platforms, whereas IGHG1 demonstrates robust performance, particularly in RNA-seq data. The two-gene panel (CXCL13 + IGHG1) offers a promising diagnostic signature for RA synovitis, supported by substantial cross-platform validation. Experimental validation in independent clinical cohorts is warranted.