Integrated Bulk and Single-Cell Transcriptomic Analysis Identifies a Reproducible SASP-Related Three-Gene Panel and...

Rheumatoid arthritis (RA) is a chronic inflammatory autoimmune disease in which senescence-associated secretory phenotype (SASP)-related transcriptional programs may contribute to synovial inflammation and fibroblast activation. This study aimed to integrate bulk transcriptomic cohorts with the gene-expression component of a single-nucleus multimodal dataset to identify reproducible SASP-related candidate biomarkers and prioritize fibroblast-associated signals in RA.

SASP-score correlation analysis, differential expression analysis, cross-cohort evaluation, logistic regression, gene set enrichment analysis, single-nucleus transcriptomic characterization, predicted regulatory network analysis, and RT-qPCR assessment were performed. GSE89408 was used as the discovery bulk transcriptomic cohort, GSE77298 as the external evaluation cohort, and GSE243917 as the single-nucleus gene-expression dataset.

Fibroblasts showed relatively high SASP-related transcriptional scores within the RA single-nucleus dataset. The cross-cohort evaluation retained RCAN1, IRF1, CFB, and TIMP1 as reproducibly elevated candidates. Among all 15 non-empty panels derived from these four genes, the RCAN1/IRF1/CFB panel achieved the highest five-fold cross-validated AUC in GSE89408 and tied for the highest external AUC in GSE77298. Adding TIMP1 did not improve the external AUC or the 95% confidence interval. Among the retained genes, CFB showed the strongest fibroblast-associated expression pattern and SASP-score correlation. The RT-qPCR analysis provided preliminary mRNA-level support for the increased expression of RCAN1, IRF1, and CFB in the RA-related cell samples.

RCAN1, IRF1, and CFB form a reproducible SASP-related candidate panel in RA, with CFB prioritized as a fibroblast-associated marker. Further protein-level, functional, and controlled single-cell validation studies are required.