Cellular senescence

Senescent cells stop dividing but refuse to die, leaking inflammatory signals that poison the tissue around them; clearing them rejuvenates mice, which has launched a gold rush of "senolytic" drugs and supplements. In humans those drugs remain unproven and are not risk-free — the best-supported options today are diet and sleep, with exercise and stress control worth doing for many other reasons.

Cellular senescence is the eighth of the twelve hallmarks of aging, and one of the strangest. A senescent cell is one that has permanently stopped dividing in response to stress — DNA damage, the gradual erosion of telomeres (the protective caps on chromosome ends), an activated cancer gene, or mitochondrial breakdown — but has not died. Instead it enters a durable, metabolically active state, resisting the normal signal to self-destruct and broadcasting a continuous stream of inflammatory molecules.[1] In youth this is protective; with age these cells accumulate, and their accumulation is a genuine driver of tissue decline, not just a symptom of it.[2]

How confident should you be

Descriptive — Strong. Intervention in humans — Weak, with an adverse biomarker signal that warrants Caution about self-administration.

That senescent cells accumulate and secrete a damaging mixture is as well established as any hallmark on this site, and in mice the transplant experiments below show they can cause decline, not merely accompany it, which most of the twelve cannot claim. In people the causal link is still indirect: in 1,678 adults aged 70 to 79 followed for about 11 years, higher blood levels of the proteins senescent cells release went with a higher risk of death, loss of mobility and heart failure, but other cells release the same proteins too.[3] The intervention side is the widest mouse-to-human gap in geroscience. The flagship human trial in knee osteoarthritis missed, and so did a randomised dasatinib-plus-quercetin trial for bone in 60 postmenopausal women. The first study to show senescent cells being cleared in people had nine people and no control group; the cognitive pilot had twelve and no control group. The one study to look at what senolytics do to biological-age markers in healthy adults found several clocks moving the wrong way — though the two clocks best validated against real health outcomes did not move, and most of the effect had reversed by six months. Nothing here is ready for use outside a trial, and this is the one hallmark page where the honest advice includes don't.

One note on the numbers below. Where a result is followed by a range in parentheses — a 95% confidence interval — that range is where the true effect most plausibly lies; when it spans zero, the result is compatible with no effect at all.

Why the body makes zombie cells in the first place

Strong. Settled cell biology, worked out across development, wound healing and tumour suppression.

It is tempting to read "cells that won't die" as pure malfunction, but senescence is a deliberate safety mechanism that evolution built for good reasons. When a cell sustains damage that could turn it cancerous — a string of mutations, an activated oncogene — pushing it into permanent arrest is one of the body's strongest defences against tumours. A cell that cannot divide cannot become a tumour.[4] The arrest is enforced by two molecular brakes that, once engaged, are very hard to release: one that halts division in response to acute damage, and the p16 pathway, which locks the arrest in place for the long term. Together they hold the cell's division switch in its "off" position.[5]

Senescence also does useful work beyond cancer prevention. During embryonic development, programmed senescence helps sculpt tissues and is then cleared by the immune system. In wound healing, a transient wave of senescent cells limits scarring by stopping repair cells from over-proliferating, after which immune cells mop them up.[6] The common thread is that healthy senescence is usually temporary: the cell does its job and is then removed. Not always, though. In mice, some long-lived senescent cells do lasting work: many line the liver's small blood vessels, and wiping them out broke down the barrier between blood and tissue, caused scarring in the liver and around blood vessels, and worsened overall health.[7] That is one more reason why clearing as many senescent cells as possible may not be the right goal. The problem of aging is not that senescent cells exist. It is that, as we get older, the immune cells that patrol for damaged cells and remove them lose their edge, so senescent cells begin to pile up and linger — the failing surveillance covered under immunosenescence.[8]

How one senescent cell poisons its neighbourhood

Strong. The mechanism is well characterised, and the transplant experiment below is the closest thing this framework has to a causal proof.

What makes a lingering senescent cell harmful is not that it has stopped working, but what it secretes. Senescent cells pour out a complex mixture of inflammatory signalling proteins, tissue-dissolving enzymes, and growth factors collectively called the senescence-associated secretory phenotype (SASP).[9] The SASP is driven largely by a master inflammatory switch, and it has two corrosive effects. First, it degrades the surrounding tissue and recruits inflammatory immune cells, creating a self-sustaining inflamed micro-environment. Second — and this is the part that turns a local problem systemic — the SASP can convert nearby healthy cells into senescent ones, spreading the state by paracrine (cell-to-cell) signalling.[10]

Multiplied across millions of cells and decades of life, this is one of the main engines of inflammaging — the chronic, sterile, low-grade inflammation that pervades aging tissue and feeds nearly every age-related disease. The most direct evidence that senescent cells cause aging rather than merely accompanying it comes from a striking mouse experiment. Injecting about half a million senescent cells into young mice weakened their grip and slowed their walking, and the slowdown lasted up to six months. The same dose given to older mice (17 months old) also shortened their lives: over the following year they had about five times the risk of dying of mice given normal cells.[11] A small population of misbehaving cells, in other words, can age a whole animal.

Where it shows up: muscle, joints, vessels, and brain

Moderate as description. Each link below is well supported mechanistically and in animal or ex-vivo human tissue; none rests on a human outcome trial, and the brain evidence is animal-only.

The damage from accumulated senescent cells is not evenly spread; it concentrates in the tissues where decline is most visible with age.

Muscle. Senescent cells build up across the muscle's repair niche — including the stem cells (satellite cells) that rebuild fibres — and disrupt the local signalling that muscle regeneration depends on. This impaired repair is a contributor to sarcopenia, the progressive loss of muscle mass and strength that accelerates after midlife.[12] It is one reason resistance training earns its place as a core longevity intervention.

Joints. In cartilage, stress pushes the resident cells (chondrocytes) into senescence, and their SASP both degrades the cartilage matrix and spreads the state to neighbouring cells — a local "inflammaging" that drives osteoarthritis.[13] In lab experiments on cartilage cells from older donors and on tissue pieces from knees with early osteoarthritis, dasatinib plus quercetin cleared senescent chondrocytes and switched the remaining cells back toward making cartilage; the rebuilding effect came mainly from dasatinib, and a different senolytic, navitoclax, made the cells worse at building cartilage.[14]

Blood vessels. The cells lining arteries face circulating damage plus constant mechanical shear from blood flow, which makes them highly prone to senescence; their accumulation degrades the vessel wall, reduces the nitric oxide that vessels use to relax and widen, and stiffens the arteries — the vascular route from a microscopic problem to a hard clinical outcome, covered in more depth under genomic instability.[15]

Brain. In the brain's immune cells (microglia), aging mitochondria leak their DNA into the cell interior, where a sensor that evolved to detect viral DNA reads it as an infection and switches on a lasting inflammatory program linked to the loss of nerve cells. In old mice, a drug that blocks this sensor pathway calmed brain inflammation and improved memory in maze and fear-learning tests.[16] This is animal evidence only.

Senolytics vs. senomorphics: the two drug strategies

Weak in humans; the mouse record is consistent and large. No human trial of a senolytic has yet shown a clinical benefit (one small randomised fatty-liver trial met a biopsy goal and awaits repetition) — the closest thing to a human success on this page, a phase 2 immune result for mTOR inhibitors further down, did not hold up in phase 3.

The discovery that clearing senescent cells reverses frailty, bone loss, and cardiovascular dysfunction in mice created the most active drug-development race in geroscience. The field has split into two approaches.[17]

Senolytics selectively kill senescent cells. Because these cells survive by switching on anti-death (anti-apoptotic) defences, senolytics work by briefly disabling those defences so the cell's own self-destruct machinery can finish the job. The best-known combination pairs dasatinib (a leukaemia drug that blocks pro-survival signals) with quercetin (a plant flavonoid), abbreviated D+Q; the two are paired because they work best on different kinds of senescent cells: in the original laboratory study, dasatinib cleared senescent human fat-cell precursors while quercetin worked better against senescent cells from blood-vessel linings, and the combination lowered the senescent-cell load in aged mice.[18] Other senolytics, such as navitoclax, which blocks a family of survival proteins, are limited by toxicity: platelets and some white blood cells rely on the same proteins, so continuous navitoclax treatment can cause a sharp drop in platelets, with a risk of internal bleeding, and in infection-fighting white cells.[19]

Senomorphics take a gentler tack: instead of killing senescent cells, they quiet the harmful SASP. Both rapamycin (a direct inhibitor of the growth-signalling mTOR pathway) and metformin fall here, suppressing the inflammatory secretome without forcing cell death.[20] The site covers both, alongside the D+Q and fisetin senolytics, under geroprotectors.

The frontier is precision: because first-generation senolytics act bluntly and risk off-target harm, researchers are developing antibody-guided drugs and engineered immune cells that home in on surface markers unique to senescent cells.[21] None of this is ready for use outside trials. The most advanced attempt to sidestep navitoclax-style toxicity by local delivery has stalled. Foselutoclax blocks the same family of survival proteins but is injected straight into the eye. In 65 people with diabetic swelling of the retina, one injection caused no clear safety problem, and after 48 weeks vision was about 5.6 letters better on an eye chart than after a sham injection, a result also compatible with no benefit at all.[22] A follow-up trial in 52 people, designed to show it was no worse than a standard eye injection (aflibercept), missed that main goal at weeks 20 to 24, though it kept pace with aflibercept at most other time points,[23] and the company, the field's best-known senolytic developer, dissolved in September 2025.[24]

The human reality check

Caution — not because harm is proven, but because the only study to measure biological-age markers in healthy adults taking these drugs found several moving the wrong way, at exactly the doses people self-administer.

The gap between the mouse data and the human data is the single most important thing to understand about this field. In mice, senolytics look close to miraculous. In humans, the results so far are modest, mixed, and occasionally cautionary — and the studies are tiny. Most of them enrolled fewer than twenty people and several had no control group; the two randomised trials large enough to test a benefit both missed their main goal.

  • The foundational human result — which itself needed correcting. The landmark study — the first to show senescent cells being cleared in people — was an open-label Phase 1 pilot in nine patients with diabetic kidney disease, given three days of D+Q; it reported fewer senescent cells in fat and skin afterwards, the first evidence that the drugs do in people what they do in mice.[25] The next year the authors published a corrigendum that re-analysed the raw data and revised some of its conclusions — a reminder that even this field's cornerstone human finding did not survive scrutiny unchanged.[26]
  • A joint-disease failure. UBX0101, a senolytic that works by disrupting the p53/MDM2 interaction, cleared senescent cells and regrew cartilage in rodents. It then went into a randomised, placebo-controlled Phase II trial of a single knee injection in 183 people with knee osteoarthritis, and it failed outright: after 12 weeks, knee pain had improved no more on any of the three doses than on placebo, and the long-term follow-up study was stopped early.[27] Some reviews still describe this trial as having reduced pain; the results posted on the trial registry show it did not.[28] This is the flagship human failure of the field, and it is worth being precise about the drug: UBX0101 is not one of the Bcl-2-family senolytics such as navitoclax, which fail for a different reason — toxicity.
  • A randomised D+Q trial for bone — also a miss. In a randomised phase 2 trial, 60 postmenopausal women took dasatinib plus quercetin intermittently for 20 weeks or received no treatment. The main goal, slowing bone breakdown, was not met: a blood marker of bone breakdown did not differ from the untreated group. A bone-building marker rose by about 16% in the first month, then fell back to the untreated group's level by week 20. There were no serious side effects. In an exploratory look at the third of women with the highest levels of the senescence marker p16 in a type of immune cell (T cells), forearm bone density rose by about 3%, a lead worth testing, not a result.[29]
  • A small randomised liver trial met its goal, for now. In 31 adults with scarring fatty-liver disease (fibrotic MASH, metabolic dysfunction-associated steatohepatitis), about half of those given intermittent D+Q for 21 weeks showed at least one stage less liver scarring on biopsy, against one in fourteen on placebo. Side effects were more common on D+Q, though all passed on their own. The result rested on few responders and lost significance under the most pessimistic handling of missing data, so the authors treat it as preliminary until a larger trial repeats it; two of them have a financial interest in Mayo Clinic senolytic patents.[30]
  • A cognitive pilot that cannot carry much weight. A twelve-person pilot (STAMINA) gave intermittent D+Q to older adults at risk of Alzheimer's disease over twelve weeks. It was well tolerated with no serious adverse events related to the intervention — which is what a safety pilot is for. Beyond that, little can be concluded, because the study was single-arm and open-label with no control group, a limitation the authors state outright. Cognitive scores rose by 1.0 point on a 30-point screening test (the Montreal Cognitive Assessment), a change well within the range expected from chance — the data are compatible with anything from a 0.7-point fall to a 2.7-point rise. The inflammatory marker they tracked, tumour necrosis factor-alpha, fell by 3%, but there too the range compatible with the data runs from a 13% fall to a 7% rise. The one finding the authors lead on is a correlation: participants whose inflammatory marker fell the most tended to be the ones whose cognitive scores rose the most.[31] Hypothesis-generating in twelve uncontrolled people; nothing more. An earlier five-person pilot in people with mild Alzheimer's disease, also with no control group, checked whether the drugs reach the brain at all: dasatinib turned up in the fluid around the brain in four of the five, at under 1% of its blood level, and quercetin was not detected. Thinking tests and brain scans did not change, and two markers of brain inflammation and injury rose while some other inflammatory signals trended down, a mixed picture that five people cannot resolve.[32]
  • When a placebo group was added, the benefit did not show up. An early study in people with lung scarring (idiopathic pulmonary fibrosis) had no comparison group and reported better walking and physical function after D+Q.[33] The same team repeated it in 12 patients, half given placebo. Walking, lung and frailty measures did not appear to differ between the groups, though the study was too small to be sure, and sleep problems and anxiety showed up in four of the six people on D+Q and in none of the six on placebo.[34]
  • A genuine warning for healthy people — and one worth reading carefully. A Phase I study followed 19 healthy adults aged 43 to 87 through six months of D+Q, measuring a panel of DNA-methylation "clocks." All three first-generation clocks showed accelerated biological aging at three months, and the methylation estimate of telomere length fell — the opposite of the intended effect.[35] Four qualifications belong with that, and the paper's tables support them better than its summary does. GrimAge and DunedinPACE — the clocks best validated against actual health outcomes — did not move at all. The acceleration on two of the accelerated clocks had returned to baseline by six months, though two others (Hannum and PhenoAge) had not, and the telomere shortening did not reverse. A second-generation clock, PhenoAge, also moved, so this cannot be dismissed as a first-generation-only artefact. And the follow-on arm that added fisetin was a different, only partly overlapping group of participants — its apparent "mitigating" effect is a comparison across two small studies, not within the same people, and on the clock that drove the original finding the fisetin arm's age estimate rose significantly, meaning more apparent aging, not less. None of this is evidence that senolytics "age you." All of it is a concrete reason healthy, non-frail adults should not self-administer these drugs: a tool designed to clear sick cells can stress healthy ones, and nineteen people cannot tell you how much.

A related laboratory finding sharpens the point. When healthy young vascular smooth-muscle cells are exposed to D+Q, they briefly take on the nuclear and chromatin features of senescent cells, recovering only after the drug is withdrawn.[36] And in rodents with folic-acid-induced acute kidney injury, the D+Q combination worsened the damage rather than helping.[37] The lesson running through all of it: clearing senescent cells is context-dependent, and "more" is not safely "better."

Natural compounds and the hormetic paradox

Weak. The fisetin case rests on one lab's mouse lifespan result, which a larger multi-site test did not reproduce, and on small human trials, of which the only one with results so far, a placebo-controlled trial in knee osteoarthritis, found no benefit; the curcumin dose-flip is a cell-culture finding, not a demonstrated harm in people.

The plant compounds at the centre of the senolytic story — the flavonoids quercetin and fisetin, and curcumin, the polyphenol found in turmeric — are sold as supplements, which makes their real evidence worth examining carefully.

Fisetin is the most talked-about of them as a senolytic. In one lab's study it reduced senescence in some cell types in mouse fat and in samples of human fat tissue treated in the lab, and given to mice late in life it extended both median and maximum lifespan.[38] But when the US National Institute on Aging's Interventions Testing Program, which runs lifespan tests at three sites in genetically mixed mice, tried fisetin, it did not extend lifespan in either sex.[39] The human results so far are no better. In 74 people with knee osteoarthritis, two short courses of high-dose fisetin (about 20 mg per kilogram of body weight for two days, repeated a month later) caused no more side effects than placebo, the trial's main goal, but also did no better than placebo on knee pain over 12 weeks or on walking tests, cartilage on MRI scans or blood markers of inflammation over a year, according to results posted on the trial registry.[40] Other trials are still running: on blood-vessel function in adults aged 65 and over (NCT06133634),[41] on low-dose fisetin and inflammation in healthy adults aged 50 and over (NCT07195318),[42] and on physical function in breast-cancer survivors.[43] No human result yet supports taking fisetin to slow aging.

Curcumin illustrates why "natural antioxidant" is not the same as "safe at any dose." Its effect on blood vessels appears biphasic (a hormetic, U-shaped curve): at low, dietary doses it lowers inflammation and supports vascular health, but in cultured vascular cells high concentrations flip it into a pro-oxidant that pushes those cells into senescence — potentially accelerating the very aging it is taken to prevent. Whether any dose people actually swallow reaches that range is unknown, which is exactly why this belongs in the "watch it" column rather than the "proven harm" one.[44][45] The site's dedicated curcumin page covers dosing and the broader evidence.

There is also a practical wrinkle that undercuts much of the supplement marketing: these compounds are poorly absorbed. They are water-insoluble, unstable in the gut, and rapidly broken down by the liver, so swallowing unformulated powder produces negligible blood levels.[46] Piperine, a black-pepper compound that slows the breakdown of curcumin in the gut wall and liver, is often added to raise absorption. The headline figure, a roughly 20-fold increase, comes from a single small 1998 study in healthy volunteers who took 2 g of curcumin with 20 mg of piperine; one of its authors was from a company that sells the ingredient.[47] Newer work has not confirmed it. A 2025 independent crossover study in 9 healthy men found that adding piperine did not raise curcumin levels in the blood,[48] and a 2024 laboratory study by a manufacturer found that curcumin from a curcumin-piperine product crossed a layer of gut cells less well than curcumin from ground turmeric root, though it did not compare curcumin with and without piperine.[49]

Rapamycin and the immune angle

Moderate — the strongest human result on this page, and still a six-week trial judged on one year of infections.

A different strategy sidesteps the killing of cells entirely: slow the formation of senescent cells and quiet their secretions by inhibiting mTOR, the master nutrient-sensing kinase that, when chronically active, suppresses the cell's self-cleaning autophagy and tilts cells toward a pro-senescent state.[50] Rapamycin does this, and it remains the most consistently life-extending drug in animal studies. The human evidence in healthy adults is still early and revolves around low, intermittent (weekly) dosing intended to capture the benefit while avoiding the immune suppression seen with the daily doses used in transplant patients.[51] A year-long placebo-controlled trial in healthy adults, PEARL, of 5 or 10 mg rapamycin once a week, found the drug reasonably safe but did not reduce deep belly (visceral) fat, its main goal; small gains in lean mass and less pain appeared only in women on the higher dose, and several of its authors were employees and shareholders of the company AgelessRx.[52] Other trials are covered under geroprotectors and deregulated nutrient sensing.

The most clinically compelling human result in this area is about the immune system. A randomised trial gave 264 older adults six weeks of low-dose mTOR inhibitors, including the rapamycin relative everolimus, and found significantly fewer infections reported over the following year along with an improved antibody response to influenza vaccination — a sign that quieting mTOR may partly reverse immunosenescence.[53] It did not hold up at scale: a 1,000-person phase 3 trial of one of those drugs found no reduction in respiratory illness (26% of people on the drug vs 25% on placebo).[54] Why such low doses? The daily doses given to transplant patients suppress the immune system, so the trial designers deliberately stayed low. Reviewers have suggested there may be a dose threshold above which mTOR inhibition stops boosting immunity and starts suppressing it, but no trial in older adults has shown that switch.[55] The trials do not point to a clear best dose either. In the phase 2b trial that preceded the phase 3, the experimental mTOR inhibitor RTB101 at 5 mg a day did no better than placebo, and neither did the higher-dose arms (10 mg twice a day, or 10 mg plus everolimus); only 10 mg once a day reduced laboratory-confirmed respiratory infections (19% of people vs 28% on placebo), and only in a second analysis that pooled both parts of the trial.[56] Getting the dose right is an open question across this field, not a solved one.

What actually helps healthy adults

Diet — Moderate (two randomised trials, judged on blood markers). Exercise — Weak in people (consistent in animals; the one positive human study had no comparison group, and randomised trials found no change in blood markers). Sleep — Weak (one single-night experiment and one randomised trial of insomnia treatment, judged on a blood marker). Stress — Weak (mechanistic). None has been tested against a hard outcome in this context.

Everything in this section works through one idea. Senescent cells pile up with age for two reasons: the body makes more of them as damage accumulates, and it gets much slower at clearing them. In mice, a senescent cell is removed within days in youth but lingers for weeks in old age, partly because the senescent cells themselves slow their own removal.[57] Weakening immune surveillance is one reason for that slowdown,[58] so the levers that help are the ones that limit damage and restore clearance, rather than ones that try to do the immune system's job with a drug.

Caloric restriction and fasting. Chronic overeating keeps the growth-and-storage pathways (insulin/insulin-like growth factor 1 and mTOR) switched on, suppressing autophagy and favouring senescence; eating less, or eating within a compressed daily window, reverses that signalling and lets cells clear damaged components before they tip into senescence.[59] In a randomised trial, 58 middle-aged and older adults with obesity and prediabetes were assigned to 18 weeks of moderate calorie restriction or a control diet; the restricted group lost about 11% of body weight. Senescence markers in their immune cells did not change — but a composite of specific circulating SASP proteins moved significantly. The authors write "altered" rather than "lowered," because the composite did not shift uniformly in one direction, and the size of the shift tracked only loosely with how much weight each person lost, explaining roughly a tenth of the variation between people.[60] In other words it behaved like a senomorphic, calming some of the harmful secretions rather than clearing the cells. The longest test comes from CALERIE 2, which randomly assigned healthy young and middle-aged adults, of normal weight or moderately overweight, to two years of calorie restriction or their usual diet. Blood samples from 199 participants showed that, compared with controls, the people who cut calories had lower levels of nine senescence-linked proteins after one year, and four of them stayed lower at two years. They managed this by eating only about 12% less, half the 25% target.[61] None of these blood markers is specific to senescent cells, so this shows that the secretions calmed down, not that the cells were cleared. The dietary detail lives under fasting and time-restricted eating.

Exercise — and not too much. In animals, regular exercise reduces the number of senescent cells in the heart, kidney, liver, muscle and fat, and it sharpens the immune cells that patrol for damaged cells — natural-killer cells and cytotoxic T-cells — so they find and kill senescent ones more readily.[62][63] The evidence in people is thinner. Older adults who did 12 weeks of twice-weekly strength and endurance training had lower levels of p16 and other senescence markers in their T-cells, and several senescence-related proteins in their blood fell modestly, but the study had no comparison group.[64] A 2026 review of randomised trials in people over 60 found only three; none measured p16, and in the two that could be pooled, blood levels of inflammatory proteins such as interleukin-6 did not differ between exercisers and controls.[65] The dose-response may also be hormetic, though the evidence is from animals: in rats, swimming to exhaustion pushed cells in the brain's memory centre into senescence and impaired memory.[66] The practical reading is the same one the site reaches everywhere: train consistently across zone 2 and high-intensity intervals, and recover properly.

Sleep. In 29 adults aged 61 to 86, one night of short sleep (allowed to sleep only from 3 to 7 a.m.) switched on SASP genes in their blood immune cells; a DNA-damage signal rose only borderline, and an early rise in the senescence marker p16 did not hold up on a second, more precise test.[67] Better evidence comes from treating insomnia. In a randomised trial of 231 adults over 60 with insomnia, p16 levels in blood immune cells rose over two years in people given sleep education but stayed flat in those given cognitive behavioural therapy for insomnia (CBT-I); among CBT-I patients whose insomnia stayed in remission p16 actually fell, though that comparison was not randomised.[68] Chronic sleep-disordered breathing compounds this: the repeated oxygen dips of obstructive sleep apnea drive oxidative stress and premature senescence, which is one reason treating it matters beyond daytime alertness.

Stress. Chronic psychological stress is not merely a mood problem at the cellular level. Sustained activation of the stress-hormone axis floods cells with cortisol and adrenaline-type hormones whose breakdown generates reactive oxygen species; this oxidative damage preferentially attacks the repetitive DNA at the ends of chromosomes, which is unusually vulnerable to it, accelerating telomere erosion and driving cells toward premature senescence.[69] The stress and recovery page covers what reliably lowers that load.

What this does and doesn't tell you

What it tells you: cellular senescence can cause decline in mice — the transplant experiments show that, and in people high levels of the proteins senescent cells release go with death and disability — and the SASP is the mechanism that turns a per-cell problem into the body-wide inflammation behind so many age-related diseases. Clearing or quieting senescent cells is therefore a legitimate target, and the levers with some human support today are diet and sleep, with exercise well supported in animals; none has yet been shown to change a health outcome through this route.

What it doesn't tell you: that any senolytic drug or supplement is established for healthy adults. The mouse-to-human gap is wide; the flagship osteoarthritis trial failed; a randomised D+Q trial for bone loss missed its main goal; the cognitive study was an uncontrolled twelve-person pilot; and a nineteen-person Phase I study found the popular D+Q combination accelerated several epigenetic-aging measures in healthy adults and shortened a methylation estimate of telomere length that had not recovered by six months. Add the poor absorption of these supplement compounds, curcumin's dose-dependent flip into a senescence inducer, and the still-unsettled dosing of rapamycin-class drugs, and the picture is clear: this is one of the most exciting areas of aging biology and one of the least ready for self-experimentation. The discipline is to treat senolytics as a research frontier, not a supplement-shelf decision.

Further reading

  • Li W, et al. Cellular senescence: from homeostasis to pathological implications and therapeutic strategies. Front Immunol 2025.[70]
  • Fu TE, Zhou Z. Senescent cells as a target for anti-aging interventions: from senolytics to immune therapies. J Transl Int Med 2025.[71]
  • Alum EU, et al. Targeting cellular senescence for healthy aging: advances in senolytics and senomorphics. Drug Des Devel Ther 2025.[72]
  • Lelarge V, et al. Senolytics: from pharmacological inhibitors to immunotherapies, a promising future for patients' treatment. NPJ Aging 2024.[73]
  • Martel J, et al. Lifestyle interventions to delay senescence. Biomed J 2024.[74]
  • Justice JN, Leng XI, LeBrasseur NK, et al. Caloric restriction intervention alters specific circulating biomarkers of the senescence-associated secretome in middle-aged and older adults with obesity and prediabetes in an 18-week randomised controlled trial. J Gerontol A Biol Sci Med Sci 2024.[75]
  • Lee E, Carreras-Gallo N, Lopez L, et al. Exploring the effects of dasatinib, quercetin, and fisetin on DNA methylation clocks: a longitudinal study on senolytic interventions. Aging (Albany NY) 2024 — n=19 Phase I; read the tables, not the abstract.[76]
  • Millar CL, et al. A pilot study of senolytics to improve cognition and mobility in older adults at risk for Alzheimer's disease. EBioMedicine 2025 — STAMINA; single-arm, open-label, n=12.[77]
  • Yousefzadeh MJ, et al. Fisetin is a senotherapeutic that extends health and lifespan. EBioMedicine 2018.[78]
  • Picos A, et al. Vascular senescence and aging: mechanisms, clinical implications, and therapeutic prospects. Biogerontology 2025.[79]
  • Hickson LJ, et al. Senolytics decrease senescent cells in humans: preliminary report from a clinical trial of dasatinib plus quercetin in individuals with diabetic kidney disease. EBioMedicine 2019 — open-label Phase 1, n=9.[80]
  • Corrigendum to the above. EBioMedicine 2020 — the re-analysis that revised some of its conclusions.[81]
  • Maurer S, et al. Senolytic therapy combining dasatinib and quercetin restores the chondrogenic phenotype of human osteoarthritic chondrocytes by the release of pro-anabolic mediators. Aging Cell 2025.[82]
  • Mannick JB, et al. TORC1 inhibition enhances immune function and reduces infections in the elderly. Sci Transl Med 2018.[83]
  • Hands JM, et al. What is the clinical evidence to support off-label rapamycin therapy in healthy adults? Aging (Albany NY) 2025.[84]

— § —