Fisetin and senolytics: what the 2025 trials say about clearing senescent cells.

What senescent cells are — and why they matter

Every cell in your body has a limited capacity to divide. Under stress — DNA damage, oxidative insult, oncogene activation, telomere shortening — cells hit a point where they stop replicating permanently. This is cellular senescence (CS): a state of essentially irreversible proliferative arrest. The signal it pulls is protective in the short term. Stopping division prevents a damaged cell from multiplying into a tumor. In the context of wound healing and development, transient senescence is necessary and cleared efficiently by the immune system.^[1]

The problem is chronic accumulation. Senescent cells are not silent. They actively secrete a cocktail of pro-inflammatory cytokines, proteases, growth factors, and chemokines collectively called the senescence-associated secretory phenotype (SASP). The SASP signal spreads dysfunction through surrounding tissue — it can induce senescence in neighboring cells, degrade the extracellular matrix, impair stem cell niches, and sustain a low-grade chronic inflammatory state sometimes called "inflammaging." The landmark 2010 review by Campisi and colleagues described SASP as "the dark side of tumor suppression" — the very mechanism that prevents cancer in the short term contributes to tissue degeneration over decades.^[2]

The causal link between senescent cell burden and physical dysfunction was established definitively by Baker et al. in 2011. Using transgenic mice with an inducible system — INK-ATTAC — that allowed targeted elimination of p16^Ink4a-positive senescent cells on demand, the researchers demonstrated that lifelong clearance of these cells delayed the onset of multiple age-related disorders including cataracts, sarcopenia, and loss of adipose tissue. Critically, clearance started late in life still slowed progression of already-established pathology.^[3] This was the first causal evidence that senescent cells drive aging phenotypes — not just correlate with them.

That finding launched a field. If you could clear senescent cells pharmacologically — without genetic engineering — the implications for aging and age-related disease were significant. Parallel longevity-pathway work, such as mTOR inhibition with rapamycin, was converging on similar questions about slowing hallmarks of cellular aging. The compounds developed to do exactly that are called senolytics (from the Greek lysis, dissolving). This is the landscape those compounds currently occupy.

The two-tier landscape: pharmaceutical vs. accessible senolytics

Senolytics fall into two broadly different categories, defined not by mechanism but by regulatory status, evidence depth, and access.

The pharmaceutical tier is anchored by dasatinib (D), a BCR-ABL tyrosine kinase inhibitor originally approved for leukemia, and quercetin (Q), a widely available flavonoid. Used together as D+Q, they represented the first clinical application of senolytics in humans — a 2019 open-label pilot in patients with idiopathic pulmonary fibrosis (IPF) that demonstrated reduced senescent cell burden and improved physical function.^[4] Navitoclax (ABT-263), a Bcl-2/Bcl-xL inhibitor with strong preclinical data, belongs here too — but its thrombocytopenia profile has constrained clinical development. Pharmaceutical-tier senolytics carry both the strongest mechanism data and the most substantial side-effect considerations.

The accessible tier includes naturally occurring compounds — primarily fisetin, standalone quercetin, and piperlongumine — that show senolytic activity in preclinical models and are obtainable as dietary supplements without a prescription. The evidence base here is less mature than D+Q, but it is growing fast. Fisetin in particular has moved from preclinical observations to registered multicenter trials over the past three years, and 2025 produced the most substantive human-adjacent data to date.

What matters for understanding either tier: senolytics are not continuous supplements. The working model is "hit-and-run" — intermittent pulsed dosing clears the senescent cell burden, after which the drug is discontinued until burden accumulates again. This matters for both efficacy and safety calculations. Kirkland and Tchkonia's 2020 review in Journal of Internal Medicine remains the clearest articulation of why intermittent dosing is central to the senolytic strategy, not a limitation of it.^[5]

Fisetin has moved from preclinical observations to registered multicenter trials in under five years. The 2025 data is the most substantive human-adjacent read the field has produced.

Fisetin's mechanism: Bcl-2 inhibition and CNS penetration

Fisetin (3,3′,4′,7-tetrahydroxyflavone) is a flavonol found at highest concentrations in strawberries, with smaller amounts in apples, persimmons, and onions. It was first identified as having senolytic properties by Zhu et al. in a 2017 Aging (Albany NY) paper that screened novel agents against senescent human endothelial cells. The researchers showed that fisetin "selectively induces apoptosis in senescent but not proliferating" cells — the definitional property of a senolytic.^[6]

The mechanism runs through the PI3K–Akt–Bcl-2/Bcl-xL axis. Senescent cells resist apoptosis in part because they upregulate anti-apoptotic proteins — chiefly Bcl-2 and Bcl-xL — that function as survival signals, preventing the cell from completing the death program even when it has stopped dividing. Fisetin inhibits this pathway, withdrawing the survival signal and allowing apoptosis to proceed selectively in senescent cells. A 2025 study in Phytotherapy Research confirmed this pathway in the context of diabetic vascular aging, showing fisetin reduced aortic senescent cell burden through PI3K–Akt–Bcl-2/Bcl-xL inhibition while simultaneously suppressing the SASP inflammatory output.^[7]

Compared to quercetin, fisetin has two structural advantages worth noting. First, CNS penetration: fisetin crosses the blood-brain barrier more readily than quercetin, a property that makes it relevant to neurological aging research and creates a mechanistic rationale for the cognitive outcome measures appearing in newer trial protocols. Second, it was found to be the most potent senolytic among ten flavonoids screened in the Yousefzadeh et al. 2018 EBioMedicine study — the paper that established fisetin's preclinical legitimacy as a lifespan-extending compound in mice, including when given late in life.^[8]

That "late in life" finding matters. The original Baker 2011 genetic clearance work showed senolysis remains effective even when started after pathology is already present. Yousefzadeh demonstrated the pharmacological analog: fisetin administered late-life in mice "restored tissue homeostasis, reduced age-related pathology, and extended median and maximum lifespan." The implication for human translation is that the intervention window may be broader than typical preventive medicine.

The 2025 Aging Cell benchmark: fisetin vs. genetic clearance

The most significant piece of fisetin research published in 2025 is the study by Murray, Mahoney, Ludwig, and colleagues at the University of Colorado Boulder, published in Aging Cell (PMC 12341784).^[9] Its title is precise: "Intermittent Supplementation with Fisetin Improves Physical Function and Decreases Cellular Senescence in Skeletal Muscle with Aging: A Comparison to Genetic Clearance of Senescent Cells and Synthetic Senolytic Approaches."

The comparison structure is what makes this study important. Rather than simply asking whether fisetin works, the researchers asked whether fisetin works as well as the gold-standard methods — genetic senescent cell clearance (analogous to the Baker INK-ATTAC system) and synthetic pharmacological senolytics (ABT-263, navitoclax). That benchmark framing is rare in the flavonoid literature, where most papers compare treatment to untreated controls and call it a day.

The key findings:

• Oral intermittent fisetin supplementation in aged mice produced a 15% reduction in frailty index scores compared to untreated aged controls.
• Grip strength improved by approximately 14%.
• RNA sequencing showed fisetin reduced Cdkn1a (p21) expression by roughly 46% in skeletal muscle — a direct marker of senescence load reduction.
• These results were comparable to genetic clearance and ABT-263 on the primary physical function endpoints.

The comparable-to-genetic-clearance finding is the benchmark that gives this paper its weight. The INK-ATTAC system is considered the definitional proof of concept for senolysis — it is the cleanest possible demonstration of what removing senescent cells does. Finding that an oral flavonoid at intermittent doses matches that outcome in skeletal muscle is a strong signal, even with the obvious caveat that this is animal data. The researchers were appropriately measured in framing the translational implications, noting it "suggests potential for clinical translation in healthy aging interventions."

What the study does not establish: the human dose equivalent, whether skeletal muscle findings translate to other tissues, or whether these effects hold in the presence of human metabolic variability and co-medications. These remain open questions, and they are exactly the questions that registered human trials need to answer.

Endothelial and vascular findings: the CXCL12 signal

The second major 2025 fisetin paper (PMC 12393244) focuses on a completely different tissue and asks a more specific mechanistic question: which SASP factor is primarily responsible for age-related vascular dysfunction, and does fisetin reverse it?^[10] (Note: this work was posted as a bioRxiv preprint in August 2025 and has since been published in Aging Cell; the findings described here are from the peer-reviewed version.)

The researchers — Mahoney, Mazan-Mamczarz, Tsitsipatis, VanDongen, and colleagues at the University of Colorado and the National Institute on Aging — used single-cell RNA sequencing to profile senescent endothelial cells in aged mice, with validation in cultured human aortic endothelial cells. Their finding: CXCL12 (C-X-C motif chemokine ligand 12, also called SDF-1) was the most highly upregulated SASP factor in senescent endothelial cells with aging.

The signal it pulls from CXCL12 is damaging in several directions simultaneously:

• Reduced nitric oxide availability — impairing endothelial-dependent vasodilation, the mechanism by which arteries dilate in response to increased blood flow demand.
• Increased mitochondrial oxidative stress — feeding a vicious cycle of further senescence induction in nearby endothelial cells.
• Promotion of endothelial-to-mesenchymal transition (EndMT) — a process associated with arterial stiffening and fibrosis.

The team then applied the SASP-transfer assay: they took circulating factors from aged mice and exposed young arteries to them. Vascular function deteriorated — and adding CXCL12 back to fisetin-treated plasma recapitulated that deterioration, establishing CXCL12 as causal rather than merely correlational. Fisetin treatment reduced circulating CXCL12 and reversed the associated functional impairments.

Why does this matter beyond the mechanism? Because it names a specific, measurable biomarker. Circulating CXCL12 becomes a potential surrogate endpoint for future fisetin cardiovascular trials — a way to know whether the drug is working before you have years of hard cardiovascular outcome data. That is the kind of translational detail that accelerates trial design. It also adds specificity to what senolytic treatment does that biomarker-agnostic studies cannot provide.

CXCL12 is now the named SASP factor mediating fisetin's vascular benefit. That names a specific, measurable endpoint future cardiovascular trials can be powered against.

STOP-Sepsis: fisetin in elderly ICU patients

The STOP-Sepsis trial (PMC protocol, PMID 39434114, NCT05758246) occupies a different corner of the senolytic space than the longevity work above — but the implications run in the same direction.^[11]

The trial's hypothesis: elderly patients with sepsis carry a higher burden of senescent immune cells. These cells exhibit "altered gene expression and resistance to apoptosis" — they do not respond normally to the inflammatory signals that should coordinate and then resolve the septic response. The result is exaggerated and dysregulated immunopathology. Clearing senescent immune cells with fisetin during the acute phase, the researchers reason, could modulate that response and reduce organ failure progression.

The design: 220 patients aged 65 and older with sepsis are randomized to fisetin (single dose), fisetin (two doses, one day apart), or placebo. The primary outcome is change in organ failure score at 7 days. Secondary endpoints include senescent cell biomarkers, organ failure-free days, ICU duration, and 28-day mortality. The trial is phase 2, multicenter, and adaptive in allocation — results inform a subsequent phase 3 design.

The context here is important to read correctly. This is not a longevity trial — it is an acute intervention trial in a critically ill population. The safety and tolerability data it will generate, however, speaks directly to the question that makes some clinicians hesitant about fisetin in any population: is it safe enough to deploy at therapeutic doses? Elderly ICU patients are the most pharmacologically vulnerable population imaginable. If fisetin clears that bar, the implications for lower-risk longevity-oriented applications become substantially clearer.

The STOP-Sepsis population also provides a unique natural experiment: elderly patients represent the highest senescent cell burdens, making them both the most likely responders to senolytic intervention and the population where SASP-mediated dysregulation is most acute. A positive result here would be one of the strongest human-context validations the accessible senolytic space has produced.

D+Q status: where the pharmaceutical-grade research stands

The dasatinib + quercetin combination sits at a different tier of the evidence hierarchy. Dasatinib is a prescription drug with established pharmacokinetics, known drug-drug interactions, and real toxicity considerations. The combination requires physician oversight in a way that fisetin — available OTC — does not. That is not a weakness of D+Q; it is the appropriate match between evidence strength and clinical caution.

The most recent addition to the D+Q trial registry is a 2025 pilot protocol published in F1000Research (PMID 40443429): "Protocol for a pilot clinical trial of the senolytic drug combination Dasatinib Plus Quercetin to mitigate age-related health and cognitive decline in mental disorders."^[12] Led by Schweiger, Diniz, Nicol, and Lenze at Washington University and collaborating institutions, this trial targets adults aged 50+ with schizophrenia or 60+ with treatment-resistant depression.

The dosing protocol is consistent with prior D+Q work: 100 mg dasatinib plus 1,250 mg quercetin for two consecutive days per week, repeated over four cycles (two days on, five days off per cycle). Outcomes are measured at multiple timepoints over one year, covering cognitive function, physical performance, brain imaging, and blood biomarkers of cellular aging.

The targeting of serious mental illness reflects a specific mechanistic rationale: senescent cells accumulate in psychiatric conditions accelerated by chronic stress and inflammatory load, and the cognitive trajectory in these populations tracks senescent burden in ways that make D+Q a logical intervention candidate. This is not a general aging trial — it is a precision application to a population where the senescence-disease link has mechanistic grounding.

The broader D+Q clinical program spans IPF, Alzheimer's disease, frailty, diabetic kidney disease, and now psychiatric aging. The STAMINA trial (Harvard-Mayo, 2025) added the cognitive-aging angle: D+Q reduced TNF-α and produced memory improvements that correlated with that inflammatory reduction. In the 2020 review by Kirkland and Tchkonia, the preclinical applications of senolytics numbered over 40 conditions. D+Q remains the furthest advanced in human evidence — but it is a prescription intervention, not a supplement optimization.

D+Q has the deepest human evidence in senolytics. It is also a prescription intervention that requires physician management. The tiers are different by design — different evidence strength, different access, different risk profile.

What to make of it in 2026

The senolytic field has changed materially over the past three years. In 2022, fisetin's clinical case rested almost entirely on Yousefzadeh 2018 and a handful of small preclinical studies. In 2026, it rests on a growing registered trial infrastructure, two major mechanistic papers, a comparison-to-genetic-clearance benchmark in skeletal muscle, and a named cardiovascular pathway.

That is a meaningful shift. It does not mean the case is closed. Here is what remains genuinely uncertain:

• Human dose: Fisetin's bioavailability in humans is variable and poorly characterized. The doses used in rodent studies do not translate directly. No published dose-finding trial in healthy adults has established the therapeutic window for senolytic effects. This is the most important gap in the literature.
• Tissue specificity: The 2025 Aging Cell paper establishes skeletal muscle data convincingly; the vascular paper establishes endothelial biology. What happens in visceral adipose, bone marrow, and the CNS in humans under accessible-tier dosing is not yet established.
• Long-term safety: Fisetin's favorable preclinical safety profile — including the Yousefzadeh data — is reassuring. But systematic long-term safety data in humans at therapeutic senolytic doses does not yet exist. STOP-Sepsis will provide the most rigorous single-dose safety dataset for the elderly.
• Hard outcome endpoints: Physical function, frailty indices, and biomarker surrogates are all legitimate intermediate endpoints. They are not mortality, hospitalization, or major cardiovascular event data. Those trials haven't been powered and run yet.

What would change the calculus decisively: a completed phase 2/3 trial in a human aging population showing durable physical function improvement or reduced all-cause hospitalization. The TROFFi trial (PMID 41835341) — a Phase II randomized study of fisetin in breast cancer survivors with chemotherapy-induced physical decline, run by investigators at UCLA, City of Hope, and Duke — represents the most structurally rigorous fisetin efficacy trial currently underway.^[13] Its primary endpoint is six-minute walk distance change after four 14-day treatment cycles. A positive result there would be the strongest human efficacy evidence the accessible senolytic space has produced.

The Biogerontology review by Li et al. synthesizes the broader clinical translation picture for senolytics across 2023–2024 and makes the point plainly: the evidence is strong enough to justify rapid clinical expansion, but not strong enough yet to recommend outside clinical protocols.^[14] That framing remains accurate in 2026.

For anyone reading the 2025 data and thinking about what it means practically: the biology is real, the signal is accumulating, and the trial infrastructure has moved from speculative to operational. The gap between "interesting preclinical compound" and "established clinical tool" is closing faster for fisetin than for almost any other compound in the accessible longevity space — including urolithin A, taurine, and NAD+ precursors, all of which share the "strong preclinical, emerging human" evidence profile. It hasn't closed yet.