Scientists discover metformin's brain mechanism after 60 years of use

For more than six decades, metformin has been the standard first treatment for type 2 diabetes. Doctors prescribed it by the millions. Patients took it daily. And for all that time, nobody fully understood how it actually worked. The assumption was straightforward enough: the drug suppressed glucose production in the liver or altered how the gut processed food. A team at Baylor College of Medicine has now upended that understanding. Their research, published in Science Advances, reveals that metformin's primary effect happens not in the liver or digestive tract but in the brain itself—a discovery that reframes how scientists think about one of medicine's most common medications.

The brain, it turns out, is far more central to controlling blood sugar than previous research had suggested. The Baylor team investigated whether the brain played any role in metformin's ability to lower glucose levels. What they found was a specific protein called Rap1, located in a region of the brain called the ventromedial hypothalamus, or VMH. When metformin is present at clinically relevant doses, it suppresses this protein's activity. That suppression is what actually drives the drug's blood-sugar-lowering effect.

To test this hypothesis, the researchers used genetically modified mice that lacked the Rap1 protein in their VMH. They fed these mice a high-fat diet to induce a diabetic state. The results were striking: low-dose metformin treatment had no effect on these animals. Their blood sugar remained elevated even after receiving the medication. Other diabetes drugs—insulin and GLP-1 receptor agonists—still worked in these mice, which meant the problem was specific to metformin's mechanism, not a general failure of glucose control. The absence of Rap1 essentially rendered metformin useless.

What happened next demonstrated just how sensitive the brain is to metformin's action. The researchers injected a tiny amount of the drug directly into the brains of diabetic mice—a dose far smaller than what would normally be given by mouth. The effect was dramatic: blood sugar dropped significantly. This suggested that the brain's responsiveness to metformin far exceeds what happens elsewhere in the body. The team also identified another key player: SF1 neurons in the VMH. These neurons showed enhanced electrical activity when metformin was present, but only if Rap1 was also there. Without Rap1, the neurons did not respond.

The implications are substantial. For the first time, researchers have mapped a direct pathway by which one of the world's most widely used diabetes medications actually operates. This opens a door to designing new drugs that could target the same brain mechanism with greater precision and potentially greater effectiveness. The findings suggest that future diabetes treatments might work by directly modulating brain activity rather than relying on liver or gut effects alone.

Still, caution is warranted. These are animal studies. The mice were genetically engineered in specific ways, and their physiology does not perfectly mirror human biology. Before any new drug based on this mechanism reaches patients, the findings will need to be confirmed in human trials. The researchers themselves acknowledge this: the results are preliminary and require validation in people. But they have provided a foundation. After sixty years of prescribing metformin without fully understanding its mechanism, science has finally caught up with clinical practice. What comes next is the harder work of translating this knowledge into better treatments.