What the New Study Found
Metformin has been a first-line treatment for type 2 diabetes for more than six decades, and researchers have spent most of that time arguing about how it works. A 2026 study suggests the answer may involve a different organ than the textbooks say.
The standard account has been that metformin suppresses glucose production in the liver. A Northwestern University study in mice, published in Nature Metabolism, found the drug's primary target appears to be the gut instead.
In that model, metformin slows mitochondrial energy production in the cells lining the intestine. Those cells respond by consuming more glucose themselves. The practical effect is that the intestine functions as a glucose sink, pulling sugar out of circulation before it can raise blood levels.
The specific target is mitochondrial complex I, an enzyme in the cellular respiration machinery. Prior work from the same laboratory had established that metformin blocks complex I. What this study adds is where that blockade matters most.
How They Tested It
The experimental design is what makes the claim credible rather than suggestive, and it is straightforward enough to describe.
The researchers engineered mice to produce a yeast enzyme called NDI1 specifically in intestinal cells. NDI1 does the same job as mitochondrial complex I but cannot be blocked by metformin. In effect, they gave the gut cells a workaround the drug could not shut down.
In those mice, metformin lost much of its ability to lower blood sugar. If the drug's glucose-lowering worked primarily through the liver, disabling its effect in the intestine should not have mattered much. It mattered a great deal.
One finding from the study has consumer relevance beyond diabetes care. Berberine, a plant-derived over-the-counter supplement that has circulated on social media as a natural alternative to prescription drugs, appears to engage the same intestinal pathway. Northwestern researchers noted the parallel and cautioned that evidence for berberine remains limited and that it should not be used as a substitute for approved medications. A shared pathway is not equivalent to efficacy, and supplements are not held to the manufacturing or evidentiary standards that prescription drugs are.
Two Other 2026 Findings Complicate the Picture
This was not the only mechanism paper published this year, and the others point elsewhere, which is worth knowing before treating any single result as settled.
Researchers at Baylor College of Medicine and international collaborators reported a brain pathway, finding that metformin acts in the hypothalamus through a protein called Rap1. In mice lacking Rap1 in those neurons, the drug did not activate the cells or reduce blood sugar through that route. The Baylor group also reported that the brain responds to much lower drug concentrations than the liver or intestine require.
Separately, a team at Université de Montréal reported in eLife in April that metformin directly targets ATP5I, a small subunit of the enzyme that produces ATP, the molecule cells use to carry energy. Senior author Gerardo Ferbeyre described the finding as raising more questions than it answered.
These results are not necessarily in conflict. A drug taken daily for decades by hundreds of millions of people may well act in several tissues at once, with the relative contribution of each depending on dose, formulation and where the drug concentrates. What the collective picture suggests is that the single-organ explanation most patients were given is incomplete.
What Does Not Change for Patients
This is the section that matters most for the very large number of people reading this while holding a prescription bottle.
Nothing about metformin dosing changes. Nothing about timing changes. Nothing about laboratory monitoring changes. The drug's indications, its established benefits, its side effect profile and its place as a first-line therapy in diabetes guidelines are all unaffected by mechanism research conducted in mice.
Mechanism studies explain why an established drug works. They do not re-establish whether it works. That question was settled by decades of clinical outcome data, and this research does not revisit it.
If you take metformin, keep taking it as prescribed. Do not adjust the dose, change the timing relative to meals, or stop because a headline suggested scientists were wrong about something. If gastrointestinal side effects are a problem, that is a longstanding and manageable issue worth raising with a clinician, who may consider an extended-release formulation or a slower titration.
Do not substitute berberine or any supplement for prescribed metformin. Sharing a pathway in a mouse experiment is not a basis for swapping a medication with decades of outcome data for one without it.
Why Mechanism Research Matters Anyway
If nothing changes clinically, the reasonable question is why any of this is worth publishing, and the answer is about what comes next rather than what exists now.
Knowing which tissue matters allows drug designers to build compounds that concentrate there. If glucose control depends substantially on the intestine, a drug engineered to stay in the gut could deliver the benefit with fewer systemic effects. Northwestern researchers made exactly that point, suggesting that directing drugs or supplements to the gut could be an effective strategy.
The brain finding carries a similar implication. Few diabetes drugs are known to act centrally, and if metformin has been doing so all along at low concentrations, that opens a design space that was not previously recognized.
There is also a broader stake. Metformin is under study for effects beyond diabetes, including on aging and cancer risk, and those investigations have been hampered by not knowing the mechanism well enough to predict where benefit would come from.
All of this work was conducted in animal models and cells. Human confirmation is the necessary next step. MedicalDaily will report human mechanistic studies and any resulting change to clinical guidance.
Frequently Asked Questions
What did the new study find? That in mice, metformin lowers blood sugar mainly by acting on mitochondrial complex I in intestinal cells, turning the gut into a glucose sink, rather than working primarily through the liver.
How confident should I be in that? It is a well-designed animal study with a strong confirmatory experiment. It has not been demonstrated in humans.
Do I need to change my dose or timing? No. Nothing about dosing, timing or monitoring changes based on this research.
Were there other findings this year? Yes. A Baylor team reported a hypothalamic pathway involving a protein called Rap1, and a Montreal team identified ATP5I as a direct target. The drug may act in more than one place.
Can I take berberine instead? No. Berberine appears to engage the same intestinal pathway in this research, but evidence for it is limited, and it is not a substitute for approved medication.
Why does mechanism research matter if nothing changes? It guides the design of future drugs. If the gut is the key tissue, compounds can be engineered to concentrate there.
Does this mean metformin is less effective than we thought? No. Its effectiveness rests on decades of clinical outcome data, which this research does not revisit.
