Listen to a Summary of This Article
Metformin, the generic drug sold in Britain and parts of Europe under brands including Glucophage, and in the United States under names including Glumetza and Fortamet, has spent decades doing an ordinary job remarkably well: helping people with type 2 diabetes control their blood sugar. Yet it has also acquired a second, more speculative life as one of the medicines most closely watched by scientists trying to understand ageing.
There is still no convincing evidence that healthy people can take metformin and expect to live longer. A major review published last year concluded that some of the early enthusiasm had outrun the evidence, with weaknesses emerging in animal studies, observational research and trials involving people without diabetes. What has endured is the biological puzzle behind the drug: metformin interferes with several of the same systems that respond to exercise, fasting and shortages of cellular energy.
One of those systems is AMPK, an enzyme that acts as an ancient fuel sensor. When a cell has plenty of energy, it can afford to build proteins, store fat and grow. When energy falls, AMPK helps change the budget, restraining some expensive activities while encouraging the cell to release or generate fuel. Exercise and fasting can stimulate the system, while metformin can activate it indirectly through changes in cellular metabolism.
For years this presented researchers with a problem. Metformin does too many things at once to reveal whether AMPK itself has much to do with longevity. Now an international group of scientists has tried a cleaner experiment: leave metformin aside and switch on AMPK directly.
The compound they used, known simply as 991, extended lifespan in three organisms separated by vast stretches of evolution: fission yeast, the nematode worm C. elegans and the fruit fly Drosophila. In some experiments the increase was greater than 25 per cent. The findings were published in Aging Cell.
Reaching past metformin
The distinction between the two drugs is important. Metformin alters metabolism in ways that can eventually activate AMPK. Compound 991 binds directly to the enzyme, allowing researchers to ask a more precise question than previous metformin experiments had permitted: if AMPK itself is pharmacologically activated, does anything happen to lifespan?
It did.
About 150 female fruit flies were used in each condition in the principal survival experiment, while the worm experiments involved roughly 110 animals per condition. Lifespan increased significantly in flies, worms and yeast. The researchers also checked whether the flies were simply eating less because the treated food was unpleasant. Feeding did not fall significantly, and their reproductive output was not significantly affected.
Then they performed what is probably the most important part of the study. In worms and yeast, the scientists genetically disabled the relevant AMPK machinery and repeated the experiment. The extension of life disappeared. In the mutant yeast, treatment with 991 was actually detrimental.
That makes it much harder to explain the results as some unrelated property of the chemical. Its beneficial effect depended on the cellular system it had been designed to activate.
Helena Cochemé, who leads the Redox Metabolism Group at the MRC Laboratory of Medical Sciences and is a senior author of the paper, described it as the first demonstration that directly targeting AMPK with a drug can have longevity benefits in living organisms.
There was also an important limit. Increasing the dose did not simply increase the benefit. A high dose shortened life in the fruit flies, and another high dose shortened survival in yeast. AMPK therefore looks less like a longevity button than a regulator whose setting matters. A system designed to help an organism cope with scarcity may be useful when appropriately activated and harmful when driven too hard.
The mouse problem
Yeast, worms and flies are useful because they live quickly enough for ageing experiments to be completed in reasonable time. A laboratory worm lives for roughly three weeks and a fruit fly for about three months, while a mouse may live for around three years. But that convenience also marks the great weakness of experiments like this one: humans are not large worms.
The researchers therefore conducted a preliminary experiment in mice, packaging 991 inside biodegradable nanoparticles and treating the animals for three weeks. Analysis of liver proteins showed the biochemical response they were looking for: AMPK activation, increased processes associated with ATP production and mitochondrial biogenesis, and suppression of mTOR, another nutrient sensitive pathway with a long history in ageing research.
But only three normal mice were analysed in each group for that part of the study, and they were not followed through old age. Nothing in this experiment shows that 991 makes mice live longer.
The authors say that longer term mouse studies are now required and will probably use newer AMPK activators with better pharmaceutical properties than 991. One of them, PXL770, has already been tested in a randomised human trial for fatty liver disease. That study concerned metabolic disease, not ageing, but it showed that direct AMPK activation in humans is no longer merely a laboratory idea.
The leap to longevity remains much larger.
That caution matters because metformin itself has accumulated a mythology that the clinical evidence has not yet justified. The 2025 review in Ageing Research Reviews found increasing uncertainty around claims that the drug slows ageing in people, noting that many of the anticipated benefits have not appeared consistently in trials involving people without diabetes.
The new study does not rescue those claims, nor does it show that people should take metformin to live longer. What it does is answer a narrower and perhaps more interesting question.
Metformin has spent years pointing researchers towards AMPK without revealing how important that connection really was. By reaching past the diabetes drug and manipulating the cellular mechanism directly, scientists have now shown that AMPK activation can extend life across three very different branches of the evolutionary tree, and that when the mechanism is removed, the advantage disappears.
The experiment has not found an anti ageing pill. It has done something more useful at this stage of the science: it has taken one plausible piece of the machinery of ageing and made the case for it considerably stronger. Metformin may have been leaving that clue in plain sight for years.