Integration of transcriptomes of senescent cell models with multi-tissue patient samples reveals reduced COL6A3 as an inducer of senescence

Senescent cells are a major contributor to age-dependent cardiovascular tissue dysfunction, but knowledge of their in vivo cell markers and tissue context is lacking. To reveal tissue-relevant senescence biology, we integrate the transcriptomes of 10 experimental senescence cell models with a 224 multi-tissue gene co-expression network based on RNA-seq data of seven tissues biopsies from ∼600 coronary artery disease (CAD) patients. We identify 56 senescence-associated modules, many enriched in CAD GWAS genes and correlated with cardiometabolic traits-which supports universality of senescence gene programs across tissues and in CAD. Cross-tissue network analyses reveal 86 candidate senescence-associated secretory phenotype (SASP) factors, including COL6A3. Experimental knockdown of COL6A3 induces transcriptional changes that overlap the majority of the experimental senescence models, with cell-cycle arrest linked to modulation of DREAM complex-targeted genes. We provide a transcriptomic resource for cellular senescence and identify candidate biomarkers, SASP factors, and potential drivers of senescence in human tissues.