Fisetin Clears Senescent Cells and Restores Blood Vessel Function via CXCL12

Cardiovascular disease remains the leading cause of global mortality, and advancing age is its strongest risk factor. Central to this risk is progressive vascular endothelial dysfunction—reduced ability of blood vessels to dilate properly—which precedes clinical disease. Cellular senescence, the irreversible arrest of aged or damaged cells, drives dysfunction partly through the senescence-associated secretory phenotype (SASP), a cocktail of pro-inflammatory cytokines, chemokines, and growth factors released into circulation. Yet which specific SASP factors originate from senescent endothelial cells in vivo, and how they impair vessel function, has remained unclear.

This study combined single-cell RNA sequencing of whole aortas from young (6-month) and old (27-month) C57BL/6N mice—with and without intermittent fisetin treatment (100 mg/kg/day, one week on/two weeks off/one week on)—to map transcriptomic changes at single-cell resolution. Unsupervised clustering identified 11 distinct aortic cell types. Reactome pathway analysis of pseudo-bulk differentially abundant RNA revealed cellular senescence as the most significantly upregulated pathway with aging, followed by immune dysregulation and cell cycle arrest. Endothelial cells stood out as the cell type most susceptible to senescence with aging and most responsive to fisetin-mediated clearance. Among senescence-associated transcripts, Cxcl12 (encoding the chemokine CXCL12/SDF-1) was the most highly upregulated gene in senescent endothelial cells with aging and was significantly reduced by fisetin treatment. Circulating plasma CXCL12 protein levels mirrored these transcriptional patterns.

To test whether the aged circulating environment causally impairs endothelial function, carotid arteries from young intervention-naïve mice were cannulated ex vivo and luminally perfused for 24 hours with plasma from young, old-vehicle, or old-fisetin mice. Old-vehicle plasma significantly impaired endothelium-dependent dilation (EDD) compared to young plasma, while old-fisetin plasma partially restored EDD. Parallel experiments in cultured human aortic endothelial cells (HAECs) exposed to the same plasma sources showed that old-vehicle plasma reduced nitric oxide (NO) bioavailability, elevated mitochondrial superoxide production, increased senescence-associated β-galactosidase activity, and promoted endothelial-to-mesenchymal transition (EndoMT)—all hallmarks of vascular aging. Fisetin-treated old mouse plasma attenuated each of these effects.

Causality of CXCL12 was tested via add-back and inhibition experiments. Adding recombinant CXCL12 to young or old-fisetin plasma (to match old-vehicle concentrations of ~3,210 pg/mL) recapitulated the impairments seen with old-vehicle plasma. Conversely, pharmacological inhibition of CXCL12 signaling with LIT-927 in old-vehicle plasma partially rescued endothelial function, NO production, mitochondrial oxidative stress, and EndoMT markers. Together these experiments establish CXCL12 as a causal, circulating SASP factor mediating endothelial dysfunction with aging.

These findings have meaningful translational implications: fisetin, a naturally occurring flavonoid already in early clinical trials, reduces senescent endothelial cell burden, lowers CXCL12 secretion, and preserves endothelial homeostasis. Limitations include the exclusively mouse-based in vivo work, the use of aortic (rather than microvascular) transcriptomics, and the fact that CXCL12 inhibition only partially rescued function—indicating additional SASP factors contribute. Nonetheless, this study positions CXCL12 as a biomarker and therapeutic target for vascular aging and supports senolytic strategies as a viable approach to cardiovascular disease prevention.