One of the most widely prescribed drugs on the planet has been hiding something from the scientists who have been studying it for six decades.
Metformin, the first-line medication for type 2 diabetes, taken by hundreds of millions of people worldwide, has been doing something in the brain that nobody fully knew about until now.
And the discovery, published in the journal Science Advances, doesn’t just change how we understand metformin.
It changes how we think about the relationship between the brain and blood sugar altogether.
The study comes from researchers at Baylor College of Medicine, led by pathophysiologist Makoto Fukuda, and it identifies a specific brain pathway that metformin has apparently been working through all along.
The liver. The gut. And now, the brain.
What Scientists Thought They Knew
Metformin has been prescribed for type 2 diabetes since the 1950s, formally approved in the US in 1994, and it remains the most commonly prescribed antidiabetes drug in the world.
Its defining characteristics are that it is long-lasting, relatively affordable, generally safe compared to alternatives, and extremely effective at managing blood sugar.
It works by reducing the amount of glucose the liver produces and by improving how efficiently the body uses insulin.
For most of those 60-plus years, that explanation, liver plus insulin sensitivity, was understood to be essentially the complete picture.
Later research added the gut to the story, finding that metformin also alters gut function in ways that contribute to its glucose-lowering effects.
But neither of those explanations fully accounted for everything metformin was doing, and the drug’s precise mechanism of action remained, as the researchers describe it in their paper, “controversial.”
The brain was always a candidate for investigation. It is the master regulator of metabolism across the body, it controls hunger, energy use, body temperature, hormonal output, and yes, glucose balance.
The ventromedial hypothalamus in particular, a small region deep inside the brain, has long been known to play a role in metabolic regulation.
Previous work by some members of the Baylor team had identified a protein in the brain called Rap1 as having an impact on glucose metabolism specifically through the VMH.
That earlier finding is what made this study possible.
What Have Scientists Found?
“It’s been widely accepted that metformin lowers blood glucose primarily by reducing glucose output in the liver,” Fukuda said. “Other studies have found that it acts through the gut. We looked into the brain as it is widely recognized as a key regulator of whole-body glucose metabolism. We investigated whether and how the brain contributes to the anti-diabetic effects of metformin.”
What the team found is that metformin travels to the ventromedial hypothalamus and, once there, essentially turns off Rap1.
That suppression of Rap1 activity in the VMH is not a side effect or a minor secondary pathway, it is central to how the drug works at clinically relevant doses.
The glucose-lowering effect of metformin, in other words, depends on this brain mechanism in ways the field had not previously recognized.
The team tested this with a clean experiment. They bred genetically modified mice that lacked Rap1 specifically in the VMH. When those mice, which had a diabetes-like condition, were given metformin, the drug had no effect on their blood sugar whatsoever.
None.
Meanwhile, other diabetes drugs, insulin, GLP-1 agonists, continued to work normally in the same mice.
The Rap1-deficient mice were not resistant to diabetes treatment in general.
They were specifically resistant to metformin. That points directly to Rap1 as the mechanism through which metformin does its work in the brain.
When the researchers went the other direction, forcibly activating Rap1 in the brain, blood glucose levels went up, and metformin’s ability to lower them was abolished.
Again, the evidence pointed the same direction: Rap1 is not incidental. It is indispensable.
How Much Does Your Brain Need?
One of the most striking findings in the study is how sensitive the brain is to metformin compared to the liver and intestines.
Direct brain injections of tiny amounts of metformin, doses far, far smaller than what is administered orally, produced marked reductions in blood sugar in diabetic mice.
The brain, it turns out, responds to concentrations of the drug that would be insufficient to trigger any meaningful response in peripheral tissues.
“This discovery changes how we think about metformin,” Fukuda said. “It’s not just working in the liver or the gut, it’s also acting in the brain. We found that while the liver and intestines need high concentrations of the drug to respond, the brain reacts to much lower levels.”
That discrepancy has potential implications for how the drug is dosed and how future treatments might be designed.
If the brain’s glucose-regulating pathway can be activated at much lower metformin concentrations than previously thought necessary, it raises the question of whether current dosing practices are optimized for the right target.
The Specific Neurons Involved
The researchers did not stop at identifying the VMH as the relevant brain region and Rap1 as the relevant protein.
They also identified which specific neurons within the VMH metformin was activating, a finding that opens a further, more targeted line of inquiry.
“We also investigated which cells in the VMH were involved in mediating metformin’s effects,” Fukuda said. “We found that SF1 neurons are activated when metformin is introduced into the brain, suggesting they’re directly involved in the drug’s action.”
SF1 neurons, named for the protein steroidogenic factor 1 that defines them, are a specific population of cells within the VMH that are involved in metabolic regulation.
In brain tissue experiments, the team measured the electrical activity of these neurons directly.
Metformin increased the firing rate of most SF1 neurons. But critically, only when Rap1 was present. In mice that lacked Rap1 specifically in these neurons, metformin produced no change in neural activity at all.
This level of precision, identifying not just that the brain is involved, but which region, which protein, and which specific class of neurons, is what makes the study significant beyond a general headline.
It gives researchers a concrete target for follow-up work and for the development of new, brain-directed treatments that might work through this pathway more precisely than metformin does.
“These findings open the door to developing new diabetes treatments that directly target this pathway in the brain,” Fukuda said.
Why This Matters Beyond Diabetes
Metformin’s story does not end with blood sugar. For years, the drug has accumulated an evidence base suggesting it does things that go well beyond glucose management, things that have been observed consistently but not fully explained.
The Baylor discovery may help explain some of them.
Metformin is classified as a gerotherapeutic, a drug with demonstrated capacity to slow various aspects of the aging process at the biological level.
It limits DNA damage. It promotes gene activity associated with longevity. It reduces chronic low-grade inflammation, which is one of the primary drivers of age-related disease.
In animal studies, it has extended lifespan in multiple species. In human observational data, the signal keeps showing up.
A 2025 study published in the Journal of Gerontology analyzed data from over 400 postmenopausal women with type 2 diabetes drawn from the Women’s Health Initiative, a national cohort study with over 30 years of follow-up.
Half the women were taking metformin. The other half were on sulfonylurea, a different class of diabetes drug.
The researchers found that the metformin group had a 30 percent lower risk of dying before the age of 90 than the sulfonylurea group.
The benchmark for what they called “exceptional longevity” was surviving past 90. Women on metformin were significantly more likely to get there.
The study has important caveats, participants were not randomly assigned to treatments, there was no placebo group, and the sample was not large.
Observational data cannot prove causation. But the signal is consistent with a growing body of evidence pointing to metformin as something more than a blood sugar drug.
Previous research has also linked metformin to a reduction in brain wear and tear specifically, slowing the kind of cellular deterioration that accumulates in the brain over time.
And a separate line of research found that metformin may reduce the risk of developing long COVID, likely through its anti-inflammatory and immune-modulating properties.
The team at Baylor is now directly pursuing the question of whether the Rap1 brain pathway they identified is responsible for these broader effects.
“In addition, metformin is known for other health benefits, such as slowing brain aging,” Fukuda said. “We plan to investigate whether this same brain Rap1 signaling is responsible for other well-documented effects of the drug on the brain.”
What Comes Next?
All of this research was conducted in mice. That is not a dismissal, the mechanistic case built in this study is rigorous, and the specific targets identified (VMH, Rap1, SF1 neurons) are well-suited to investigation in human tissue and human trials.
But the translation from mouse biology to human biology is never guaranteed, and the VMH-Rap1-SF1 axis will need to be validated in people before any of this changes clinical practice.
The study’s authors are clear about this. The next step is human research, confirming whether the same brain pathway operates in people, whether the brain responds to similarly low concentrations of metformin, and whether targeting this pathway more precisely could produce better outcomes with fewer side effects.
The gastrointestinal problems associated with metformin, nausea, diarrhea, abdominal discomfort, affect up to 75 percent of new users, and they are primarily a consequence of high concentrations of the drug in the gut.
If the brain’s contribution to metformin’s effect can be leveraged at far lower doses, there may be a path to the same therapeutic benefit with fewer of the peripheral side effects that make the drug difficult for many patients to tolerate.
“These findings provide a previously unknown neural mechanism mediating the antidiabetic effect of metformin,” the authors write in their conclusion. The word “previously unknown” is doing a lot of work in that sentence.
Metformin has been prescribed to patients for over 60 years. It has been the subject of thousands of studies. And it was still hiding a brain mechanism that nobody had found until 2025.
The research was published in Science Advances.
If you have questions about metformin or any medication you are taking, speak with your doctor or a qualified healthcare professional.
