Metformin produces a modest and unreliable extension of life in animal models, and human data shows a small increase in life span in diabetic patients. This is thought to work as a calorie restriction mimetic drug, triggering one slice of the beneficial response to a reduced nutrient intake. Researchers here dig further in the biochemistry of the drug, and find that it reduces liver inflammation in addition to other, known effects. This is interesting, and suggestive that any benefits it produces are going to be much smaller in healthier older adults with lower levels of chronic inflammation. It doesn’t change the fact that metformin does have only a small and unreliable effect on life span per the existing data, and is thus not where we should be focusing our attention.

Researchers have known for 20 years that metformin activates a metabolic master switch, a protein called AMPK, which conserves a cell’s energy under low nutrient conditions, and which is activated naturally in the body following exercise. Twelve years ago, researchers discovered that in healthy cells, AMPK starts a cascade effect, regulating two proteins called Raptor and TSC2, which results in a block of the central pro-growth protein complex called mTORC1 (mammalian target of rapamycin complex 1). These findings helped explain the ability of metformin to inhibit the growth of tumor cells.

But in the intervening years, many additional proteins and pathways that metformin regulates have been discovered, drawing into question which of the targets of metformin are most important for different beneficial consequences of metformin treatment. Indeed, metformin is currently entering clinical trials in the United States as a general anti-aging treatment because it is effects are so well established from millions of patients and its side effects are minimal. But whether AMPK or its targets Raptor or TSC2 are important for different effects of metformin remains poorly understood.

In the new work, in mice, researchers genetically disconnected the master protein, AMPK, from the other proteins, so they could not receive signals from AMPK, but were able to otherwise function normally and receive input from other proteins. When these mice were put on a high-fat diet triggering diabetes and then treated with metformin, the drug no longer had the same effects on liver cells as it did in normally diabetic animals, suggesting that communication between AMPK and mTORC1 is crucial for metformin to work. By looking at genes regulated in the liver, the researchers found that when AMPK couldn’t communicate with Raptor or TSC2, metformin’s effect on hundreds of genes was blocked. Some of these genes were related to lipid metabolism, helping explain some of metformin’s beneficial effects. But surprisingly, many others were linked to inflammation. Metformin, the genetic data showed, normally turned on anti-inflammatory pathways and these effects required AMPK, TSC2, and Raptor.

Metformin and exercise elicit similar beneficial outcomes, and research has previously shown that AMPK helps mediate some of the positive effects of exercise on the body. “If turning on AMPK and shutting off mTORC1 are responsible for some of the systemic benefits of exercise, that means we might be able to better mimic this with new therapeutics designed to mimic some of those effects.”