Longevity & AgingResearch PaperOpen Access

New Fisetin-Derived Senolytics Outperform the Original and Extend Healthspan in Mice

Two novel flavonoid compounds derived from fisetin clear senescent cells more effectively and improve multiple healthspan markers in aging mice.

Wednesday, October 7, 2026 3 views
Published in Aging Cell
A cluster of ripe red strawberries beside a small glass vial of amber liquid and a molecular model, arranged on a white lab bench under bright overhead lighting

Summary

Researchers at the University of Minnesota designed two fisetin-derived flavonoid compounds — SR29384 and SR31133 — using phenotypic drug discovery and structure-activity relationship optimization. In senescent cell cultures, both compounds killed senescent cells more efficiently and selectively than fisetin. In aged and progeroid mouse models, they reduced tissue senescence markers and improved physical function and healthspan parameters. Mechanistic analysis using transcriptomics, machine learning, and molecular docking pointed to PARP1, CDK2, and possibly BCL-xL as key targets driving selective apoptosis in senescent cells. The compounds represent an advance over fisetin's poor bioavailability and weak potency, offering translational promise for treating age-related diseases.

0:00--:--

Detailed Summary

Cellular senescence — the state in which damaged cells arrest permanently but remain metabolically active and secrete pro-inflammatory signals (the SASP) — is a well-established driver of aging and age-related disease. While senolytics that selectively eliminate senescent cells have proven effective in animal models, the best-studied natural senolytic, fisetin, suffers from poor bioavailability and modest potency. This study set out to engineer superior fisetin analogs using a phenotypic drug discovery framework, bypassing the need to predefine a single molecular target — an important design choice given fisetin's multi-pathway promiscuity.

The team at the University of Minnesota screened a focused library of natural flavonoids in senescent Ercc1−/− mouse embryonic fibroblasts (MEFs), a well-validated progeroid model. A selectivity index (SI) — the ratio of EC50 in non-senescent cells to EC50 in senescent cells — guided compound prioritization. From preliminary SAR analyses, several structural features emerged as critical: hydroxyl group positioning, the degree of conjugation, and specific substitutions on the B-ring. These insights drove iterative chemical optimization, ultimately producing SR29384 and SR31133, which consistently outperformed fisetin on both potency and selectivity across multiple cell types.

In head-to-head cell culture comparisons, SR29384 and SR31133 demonstrated significantly lower EC50 values in senescent MEFs than fisetin while maintaining higher selectivity indices, meaning they killed senescent cells at concentrations that spared normal proliferating cells. The compounds were tested across a panel of senescent cell types induced by different stressors (replicative, DNA-damage-induced, and oncogene-induced), confirming broad-spectrum senolytic activity rather than efficacy limited to a single senescence model.

In vivo, both compounds were tested in aged (24-month-old) wild-type mice and in Ercc1−/Δ progeroid mice. Treated animals showed reduced expression of senescence markers including p16Ink4a and p21Cip1 in multiple tissues, as well as lower levels of SASP components. Importantly, treatment was associated with improvements in physical performance metrics — including grip strength and rotarod performance — and other healthspan-related parameters compared to vehicle-treated controls, demonstrating functional as well as molecular benefits.

Mechanistic investigation combined RNA-seq transcriptomics of treated versus untreated senescent cells with machine learning-based target prediction, network pharmacology, and molecular docking. This multi-pronged approach converged on PARP1 and CDK2 as likely primary targets, with BCL-xL as a possible secondary target contributing to senescent cell apoptosis. These targets align with known pro-survival pathways upregulated in senescent cells. The authors note that while these findings are compelling, definitive target engagement in vivo and precise binding affinities await further validation. Pharmacokinetic data suggesting improved solubility and absorption over fisetin were also reported, though full PK profiling in humans remains to be done. Collectively, SR29384 and SR31133 represent well-characterized, translatable senolytic candidates poised for further preclinical development and eventual clinical testing.

Key Findings

  • SR29384 and SR31133 showed lower EC50 values in senescent Ercc1−/− MEFs than fisetin, with higher selectivity indices indicating superior selective killing of senescent over normal cells
  • Both compounds demonstrated broad-spectrum senolytic activity across multiple senescence models including replicative, DNA-damage-induced, and oncogene-induced senescent cells
  • In aged (24-month-old) wild-type mice, treatment reduced tissue expression of p16Ink4a and p21Cip1 senescence markers across multiple organs vs. vehicle controls
  • Ercc1−/Δ progeroid mice treated with SR29384 or SR31133 showed measurable improvements in grip strength and rotarod performance compared to untreated progeroid controls
  • SASP component levels (pro-inflammatory cytokines and chemokines) were reduced in tissues of treated aging mice relative to controls
  • Multi-modal mechanistic analysis (transcriptomics, machine learning, network pharmacology, molecular docking) identified PARP1 and CDK2 as primary targets and BCL-xL as a likely secondary target mediating selective apoptosis in senescent cells
  • Structure-activity relationship analysis identified specific hydroxyl group positioning and B-ring substitution patterns as the key determinants of enhanced senolytic potency in this flavonoid scaffold

Methodology

The study used phenotypic drug discovery with a focused natural flavonoid library screened in Ercc1−/− senescent MEFs using SA-β-gal activity measured by high-content fluorescent imaging (Cytation 1) with C12FDG and Hoechst 33342 staining. SAR-guided iterative chemical optimization produced SR29384 and SR31133, which were then tested in multiple senescent cell types in vitro and in vivo in both aged (24-month-old) wild-type and Ercc1−/Δ progeroid mice with vehicle controls. Mechanistic studies integrated RNA-seq transcriptomics, machine learning target prediction, network pharmacology, and molecular docking to identify candidate molecular targets.

Study Limitations

All in vivo testing was conducted exclusively in mouse models (aged wild-type and Ercc1−/Δ progeroid mice), and translation to human pharmacology remains unconfirmed. Target engagement data are primarily computational (molecular docking and network pharmacology), with definitive biochemical target validation in intact cells and in vivo pending. The authors do not report conflicts of interest related to commercial interests in these compounds, but the research was conducted at the institution developing them, and full human pharmacokinetic profiling has not been performed.

Enjoyed this summary?

Get the latest longevity research delivered to your inbox every week.

Enter your email to subscribe: