- The study is real and well run. A research compound called 991, which switches on the energy-sensing enzyme AMPK directly, extended lifespan in fission yeast, roundworms and fruit flies, "in some cases by more than 25%" (MRC Laboratory of Medical Sciences release on the Aging Cell paper, 30 September 2026)
- No mouse lifespan test yet. Mice were dosed only for a protein-profile analysis; a mouse lifespan study is the stated next step. Write-ups saying the drug extended lifespan in mice, or naming zebrafish, are wrong
- The worm-to-mouse record is poor. In the NIA's Interventions Testing Program, resveratrol, green tea extract, curcumin, fish oil and metformin alone all failed to extend mouse lifespan. Rapamycin and acarbose did, by up to 26% and 22% of median lifespan
- You already own an AMPK activator. Exercise activates AMPK in human muscle. Metformin does so indirectly, and its longevity trial has never launched; calorie restriction improved risk factors in a two-year human trial but has no lifespan data
- Don't buy anything off the back of this. Compound 991 has never been given to a human. One direct AMPK activator, MK-8722, caused heart-muscle thickening in rodents and monkeys. No supplement sold as an "AMPK activator" has been shown to extend human life
You may have seen the headline this week: scientists extend lifespan by a quarter with a single drug. The study behind it is genuine, it comes from a serious lab, and the result is interesting. It's also in yeast, worms and flies. That's not a put-down. It's the whole story, and it's worth understanding why.
Here's what the paper found, why "it worked in flies" has so often meant "it didn't work in mice", what already activates the same enzyme in your own body, and what to do with the news.
Yeast, Worms, Flies, and (So Far) No Mouse Lifespan
Evidence: Strong in model organisms — no mammalian lifespan data yetThe paper, "Direct Pharmacological Activation of AMPK Extends Lifespan in Yeast, Worms and Flies", was published in Aging Cell on 30 September 2026 (press release date) by a team led from the MRC Laboratory of Medical Sciences and Imperial College London, with the University of Cologne and others (Dos Santos et al., Aging Cell 2026; UKRI press release).
AMPK, AMP-activated protein kinase, is an enzyme in our cells that tracks how much energy the cell has. When energy runs low it shuts down expensive building projects like protein synthesis and fat storage and turns on the breakdown of fat and sugar for fuel. In the words of Dr Helena Cochemé, one of the lead authors: "AMPK is effectively the body's equivalent to the 'energy saving mode' on a mobile phone."
The idea that switching on AMPK might slow ageing isn't new. What's new here is how the switch was flipped. Metformin and semaglutide activate AMPK indirectly, as a side-effect of what they do elsewhere, which makes it hard to say whether AMPK is responsible for any benefit. Compound 991 acts on AMPK directly. The team gave it to three distantly related organisms: the fission yeast S. pombe, the roundworm C. elegans and the fruit fly Drosophila. Lifespan went up in all three, "in some cases by more than 25%". Cochemé's summary: "The fact that we can extend lifespan in yeast, worms and flies is very exciting."
Two things the headlines tend to drop. First, the per-species numbers: the press release gives only "more than 25% in some cases", and the abstract gives no per-species figures and the full text was not accessible to us, so we won't invent precision the release doesn't offer. Second, the species list. If the version you read said it extended lifespan in mice, or mentioned zebrafish, it was wrong. The paper's title names yeast, worms and flies; mice were treated only for a protein-profile analysis, and a mouse lifespan study is what comes next. Professor David Carling, a co-author: "Now that we have very solid, convincing evidence from the laboratory model organisms, the next step will be to show that we can also improve health and extend lifespan in mice."
That last quote deserves its full context. Professor Filipe Cabreiro, another of the senior authors: "The field is still a long way from anti-ageing clinical trials in humans." The researchers are being careful. The question is whether the rest of us will be.
From Worm to Mouse, Most Drugs Fail
Evidence: Strong — multi-site randomised mouse trialsWhy does "it worked in flies" carry so little weight on its own? Because we have an unusually good scorecard. The US National Institute on Aging runs the Interventions Testing Program (ITP), which feeds candidate compounds to genetically diverse mice at three independent labs, with enough animals in each group to detect modest effects. The same fly-and-worm favourites that fill supplement shelves have been through it.
The results are sobering. Resveratrol, green tea extract and curcumin, each fed from four months of age, produced no significant change in the lifespan of male or female mice (Strong et al., J Gerontol A 2013). Fish oil didn't extend lifespan either, and nor did metformin on its own at 0.1% of the diet (Strong et al., Aging Cell 2016). Metformin is worth pausing on, because it extends lifespan in worms and had the same "AMPK activator" billing. In C. elegans it works through the worm's bacterial food: metformin "increases or decreases worm lifespan, depending on E. coli strain metformin sensitivity and glucose concentration" (Cabreiro et al., Cell 2013). Cabreiro, incidentally, is a senior author on the new paper, which is one reason to expect the mouse study will actually be done.
Worms and flies vs mice vs humans: what's actually been shown
| Intervention | Worms / flies | Mice | Humans |
|---|---|---|---|
| Compound 991 direct AMPK activator | Lifespan up in yeast, worms and flies, >25% in some cases Aging Cell 2026 | Dosed for proteomics only; 'pro-longevity' protein profile, no lifespan test stated next step | Never given to a person |
| Metformin indirect AMPK activator | Extends worm lifespan, but depends on the worms' bacteria and glucose Cell 2013 | No significant extension alone at 0.1% of diet ITP, Aging Cell 2016 | No lifespan trial. TAME planned (3,000 people, 65–79, six years) but not launched; earlier "outlive non-diabetics" finding reversed in Danish registers Int J Epidemiol 2022 |
| Rapamycin mTOR inhibitor | — | Median lifespan +23% males, +26% females at 42 ppm from 9 months; a lower dose works even when started at 20 months ITP, Aging Cell 2014; Nature 2009 | No lifespan data |
| Resveratrol | — | No significant effect in either sex ITP, J Gerontol A 2013 | No lifespan data |
| Calorie restriction | — | — | Two years at ~12% below usual intake: 7.5 kg lost, LDL, blood pressure, CRP and insulin sensitivity improved. No lifespan data CALERIE-2, Lancet Diabetes Endocrinol 2019 |
| Exercise | — | — | Activates AMPK in human skeletal muscle Biochem J 2009; the only AMPK route with decades of human outcome data |
"—" means we didn't verify a primary source for that cell, not that no study exists. ITP = NIA Interventions Testing Program, which tests compounds in genetically heterogeneous mice at three sites. Sources numbered in the list at the end.
It's not all failure, and the exceptions matter because they show what a real hit looks like. Rapamycin, which blocks the mTOR pathway, is the ITP's clearest success: fed from 600 days of age, about 20 months, it raised the age at which 90% of mice had died by 14% in females and 9% in males (Harrison et al., Nature 2009). At a dose three times higher, median lifespan rose by 23% in males and 26% in females (Miller et al., Aging Cell 2014). Acarbose, a diabetes drug that slows starch digestion, increased male median lifespan by 22% but female median lifespan by only 5% (Harrison et al., Aging Cell 2014). 17α-estradiol, a weak oestrogen, added 12% to male median lifespan at the first dose tested and did nothing for females; at three times the dose it "robustly extended both median and maximal lifespan, but still only in males" (Strong 2016).
Change in median lifespan of mice, by compound and sex
Genetically heterogeneous (UM-HET3) mice, pooled across three sites. Rapamycin: Miller et al. 2014. Acarbose and 17α-estradiol: Harrison et al. 2014 (acarbose females +5%, P = 0.01). Metformin: Strong et al. 2016. Resveratrol: Strong et al. 2013. Green tea extract, curcumin and fish oil also showed no significant lifespan effect in the same programme and are omitted for space.
Notice the pattern in the hits. Effects differ by sex, sometimes sharply, and depend on dose. And the compounds that worked in mice weren't the ones with the loudest fly-and-worm stories. None of this says compound 991 will fail in mice. It says the mouse study is the test, not a formality.
You Already Have an AMPK Activator
Evidence: Strong for exercise · Moderate for calorie restriction · Weak for metforminThe odd thing about the excitement is that AMPK is not an exotic target waiting for a drug. Three things people already do switch it on.
Exercise. AMPK responds to the ratio of spent to available energy in a cell, and nothing changes that ratio in muscle like contraction. A review in the Biochemical Journal puts it plainly: during exercise AMPK "is activated in skeletal muscle in humans" (Richter & Ruderman, Biochem J 2009). This is the one AMPK activator with a long human track record attached, not because anyone ran an AMPK trial but because exercise has been studied for mortality for decades. If a reader wants a dose, our guide to the Norwegian 4x4 covers the protocol with the best VO2 max evidence.
Not overeating. Calorie restriction is the oldest lifespan intervention in lab animals and lowers the energy supply AMPK monitors. In humans we have one serious trial: CALERIE-2, which randomised 218 healthy non-obese adults aged 21 to 50 to two years of a 25% calorie cut or their usual diet. The restriction group managed an average cut of 11.9%, not 25%, and lost a sustained 7.5 kg. Their LDL cholesterol and blood pressure fell significantly from baseline, and C-reactive protein, insulin sensitivity and metabolic-syndrome score improved relative to controls (Kraus et al., Lancet Diabetes Endocrinol 2019). That's cardiometabolic risk, not lifespan; a two-year trial in 40-year-olds can't measure lifespan. But it's human, randomised and consistent.
Metformin. The diabetes drug activates AMPK indirectly, which is exactly the problem the new study was designed to get around. Its longevity reputation rests on a widely quoted observational finding that people with diabetes on metformin outlived people without diabetes. A 2022 Danish register study set out to reproduce that and found the reverse: starting metformin alone was associated with higher all-cause mortality than people with no diabetes diagnosis, with an incidence rate ratio of 1.32 (95% CI 1.16–1.50) after adjusting for comorbidities and social factors; in twin pairs discordant for metformin the adjusted ratio was 1.64, and the authors concluded that metformin "was not associated with survival equal or superior to that of the general population without diabetes" (Keys et al., Int J Epidemiol 2022). The trial meant to settle the question, TAME (Targeting Aging with Metformin), would enrol 3,000 people aged 65 to 79 for six years at 14 sites. As of this writing, the American Federation for Aging Research's own page still describes it as a trial to be launched once donors are found (AFAR). It has not reported anything because it hasn't started.
The Drugs That Are Already in Humans
Evidence: Early — short phase 2 trials, no long-term safety dataCarling notes that "Direct AMPK activators have already been used safely in clinical trials for the treatment of specific metabolic diseases." That's true, and it's worth seeing exactly what it rests on, because it's thinner than it sounds.
Three compounds, three very different results
| Compound | Tested in | Finding | Safety signal |
|---|---|---|---|
| O304 | Mice, rats, dogs; 65 people with type 2 diabetes on metformin, 28 days phase 2a | Fasting glucose −0.60 mmol/L vs −0.10 on placebo in a post-hoc subgroup of 49; systolic BP −5.8 mmHg, diastolic −3.8 | No increase in heart weight in mice or rats; no ECG abnormality or change in heart muscle mass in the 28-day trial JCI Insight 2018 |
| PXL770 | 121 people with fatty liver disease, 12 weeks phase 2, NCT03763877 | Liver fat −14.3% (250 mg twice daily) and −14.7% (500 mg daily) vs −1.1% on placebo; neither reached P < 0.05 | 3 serious adverse events in the two higher-dose arms, 0 on placebo; no long-term data ClinicalTrials.gov |
| MK-8722 | Rodents and rhesus monkeys, including diabetic monkeys | Robust, insulin-independent glucose uptake in muscle; better glycaemia with no hypoglycaemia | Cardiac hypertrophy and increased heart glycogen, "without apparent functional sequelae" Science 2017 |
| Compound 991 | Yeast, worms, flies; mice (proteomics only) | Lifespan extended, >25% in some cases | Not reported in the press release; no mammalian safety or lifespan data |
O304: Steneberg et al. 2018. PXL770: ClinicalTrials.gov results posted October 2021, least-squares mean relative change in liver fat by MRI-PDFF. MK-8722: Myers et al. 2017. None of these trials lasted longer than 12 weeks, and none measured lifespan.
So the human experience with direct AMPK activators amounts to a 28-day trial in 65 people and a 12-week trial in 121 people, in disease, with reassuring but short-term safety. Set against that is MK-8722, a direct activator of all twelve mammalian AMPK complexes, which did what it was meant to do for blood sugar in monkeys and also thickened their heart muscle (Myers et al., Science 2017). The findings aren't contradictory; different compounds behave differently. But they're a reminder that AMPK has many jobs, the heart's included, and that "safe for 28 days" and "safe to take for decades to age slower" are not the same claim. Nobody has dosed healthy people with a direct activator for years, and it isn't in the plan.
What to Do With This News
Here's the honest ledger. A direct AMPK activator extends lifespan in three short-lived laboratory species. The right next experiment is in mice, which the authors plan to run. The historical odds for a compound at this stage are not good, the human safety record for this class is short, and compound 991 has never been in a person. None of that should change what you do this week, except to notice that the enzyme in the headline is one you can already switch on.
- Exercise is the AMPK activator with human outcome data. It activates AMPK in your muscle, and unlike every compound in this article it has been studied against mortality in people for decades. Interval work like the Norwegian 4x4 is a reasonable dose; so is anything you'll keep doing.
- Don't overeat. Even a 12% calorie cut, sustained for two years, moved blood pressure, LDL, inflammation and insulin sensitivity in healthy adults. You don't need to starve; you need to stop the slow creep.
- Don't buy an "AMPK activator". Supplements sold under that label have not been shown to extend human lifespan, and the one with the best story, resveratrol, failed in mice. Berberine gets the same billing; we've looked at it alongside metformin here.
- Don't ask for metformin to live longer. It's a good diabetes drug. Its longevity evidence is observational, contested, and awaiting a trial that hasn't begun. If you're on it for diabetes, stay on it and talk to your doctor before changing anything.
- Watch the mouse study. If 991 or a successor extends lifespan in both sexes of ITP-style mice, that's news. Until then, this is a fly result, and fly results are where the pipeline starts, not where it ends.
And if you want to know whether any of it is working for you, biological-age tests are softer than they look; blood pressure, waist, lipids and fitness are still the numbers to track with your doctor.
Frequently Asked Questions
A good fly result is still a fly result.
Compound 991 switched on AMPK directly and three distantly related organisms lived longer, in some cases by more than a quarter. That's a real finding from a careful lab, and the mouse study it points to is worth waiting for. But the same programme that would test it has watched resveratrol, curcumin, green tea, fish oil and metformin all fail to extend mouse lifespan, and the only direct AMPK activators that have touched humans did so for weeks, not decades.
The enzyme in the headline is one you already control. Hard exercise activates it in your muscle; not overeating keeps the energy signal honest. Do those, skip the supplement aisle, and talk to your doctor before taking any drug for a disease you don't have.
Sources
- Dos Santos et al. Direct Pharmacological Activation of AMPK Extends Lifespan in Yeast, Worms and Flies. Aging Cell 2026;25(10):e70721, doi 10.1111/acel.70721 (press releases 30 September 2026). The full text was not accessible to us at the time of writing; study details are taken from the abstract (PubMed 42810392) and the institutional releases below.
- UK Research and Innovation. Experts extend lives of animals by activating 'energy saver' mode. 30 September 2026.
- MRC Laboratory of Medical Sciences. Experts significantly extend lives of animals by activating 'energy saver' mode. 30 September 2026.
- Strong R, Miller RA, Astle CM, et al. Evaluation of resveratrol, green tea extract, curcumin, oxaloacetic acid, and medium-chain triglyceride oil on life span of genetically heterogeneous mice. J Gerontol A Biol Sci Med Sci 2013;68(1):6–16.
- Strong R, Miller RA, Antebi A, et al. Longer lifespan in male mice treated with a weakly estrogenic agonist, an antioxidant, an α-glucosidase inhibitor or a Nrf2-inducer. Aging Cell 2016;15(5):872–884.
- Harrison DE, Strong R, Sharp ZD, et al. Rapamycin fed late in life extends lifespan in genetically heterogeneous mice. Nature 2009;460:392–395.
- Miller RA, Harrison DE, Astle CM, et al. Rapamycin-mediated lifespan increase in mice is dose and sex dependent and metabolically distinct from dietary restriction. Aging Cell 2014;13(3):468–477.
- Harrison DE, Strong R, Allison DB, et al. Acarbose, 17-α-estradiol, and nordihydroguaiaretic acid extend mouse lifespan preferentially in males. Aging Cell 2014;13(2):273–282.
- Cabreiro F, Au C, Leung K-Y, et al. Metformin retards aging in C. elegans by altering microbial folate and methionine metabolism. Cell 2013;153(1):228–239.
- Richter EA, Ruderman NB. AMPK and the biochemistry of exercise: implications for human health and disease. Biochem J 2009;418(2):261–275.
- Kraus WE, Bhapkar M, Huffman KM, et al. 2 years of calorie restriction and cardiometabolic risk (CALERIE): exploratory outcomes of a multicentre, phase 2, randomised controlled trial. Lancet Diabetes Endocrinol 2019;7(9):673–683.
- Keys MT, Thinggaard M, Larsen LA, et al. Reassessing the evidence of a survival advantage in Type 2 diabetes treated with metformin compared with controls without diabetes: a retrospective cohort study. Int J Epidemiol 2022;51(6):1886–1898.
- American Federation for Aging Research. The TAME Trial. Accessed 8 October 2026.
- Steneberg P, et al. PAN-AMPK activator O304 improves glucose homeostasis and microvascular perfusion in mice and type 2 diabetes patients. JCI Insight 2018;3(12):e99114.
- ClinicalTrials.gov. A Study of the Efficacy and Safety of PXL770 Versus Placebo After 12 Weeks of Treatment in Patients With NAFLD (NCT03763877). Results posted 28 October 2021.
- Myers RW, Guan H-P, Ehrhart J, et al. Systemic pan-AMPK activator MK-8722 improves glucose homeostasis but induces cardiac hypertrophy. Science 2017;357(6350):507–511.