Integrated Transcriptomic Analysis Identifies a Ferroptosis-Related Gene Signature Underlying Macrophage Heterogeneity and Immune Dysregulation in Rheumatoid Arthritis Synovium

Ferroptosis contributes to rheumatoid arthritis (RA), but the key regulators and their cell-specific functions in the synovial microenvironment remain incompletely understood. This study aimed to identify robust ferroptosis-associated biomarkers and define their cellular context through an integrated transcriptomic framework. Bulk synovial transcriptomic datasets were integrated using differential expression, weighted gene co-expression network analysis, and a curated ferroptosis gene set. Candidate genes were prioritized using three feature-selection algorithms and validated across twelve models. Immune infiltration, single-cell RNA sequencing, pseudotime, CellChat, and virtual gene perturbation analyses characterized their cellular and functional context. Six ferroptosis-related genes (AIM2, HCP5, CAPG, ALOX5, TYMS, and RUNX3) formed a diagnostic signature discriminating RA from healthy synovium (mean leakage-free test AUC 0.977 across five training-only re-derivations; 0.9375 in the primary split), with consistent leave-one-dataset-out and osteoarthritis-control validation. These genes linked to adaptive immune and ferroptosis-associated transcriptional programs. Single-cell analysis localized this transcriptional score to macrophages, where ferroptosis-signature-high macrophages occupied a central communication-network position and showed collagen-pathway enrichment; because RA and healthy samples were not fully independent of dataset origin, this finding is hypothesis-generating. Pseudotime analysis suggested progressive induction of these genes toward an inflammatory NF-κB/TNF/IL-17 terminal state, and virtual perturbation suggested convergence on mitochondrial oxidative phosphorylation networks that persisted after removing sex-chromosome genes, though residual sex-associated effects cannot be fully excluded. This study identifies a biologically interpretable ferroptosis-related gene signature and suggests ferroptosis-associated macrophage reprogramming may contribute to immune dysregulation in RA synovium. The single-cell findings are hypothesis-generating given the cohort design, while the diagnostic signature was supported by leakage-free, cross-cohort, and disease-specificity analyses, providing candidate biomarkers and therapeutic targets for RA.