What Is the Mechanism of Action of Metformin?

Metformin lowers blood sugar primarily by reducing the amount of glucose the liver produces, and it does this through several overlapping molecular pathways rather than a single clean switch. Despite being prescribed to hundreds of millions of people worldwide and studied for decades, the full picture of how metformin works has only come into focus recently, with key discoveries as late as 2022 revealing binding targets that nobody expected. The drug’s reach extends well beyond glucose control, touching energy sensing, gut bacteria, immune regulation, and even cellular aging.

Where It All Starts in the Cell

The first thing metformin does once it enters a cell is interfere with the energy-producing machinery inside mitochondria. Specifically, it inhibits Complex I of the mitochondrial respiratory chain, one of the protein assemblies responsible for generating the cell’s main energy currency, ATP.¹ When Complex I is partially blocked, cells produce less ATP and accumulate more AMP, its low-energy counterpart. That shift in the energy balance is the trigger for most of what follows.

The rise in AMP relative to ATP activates a sensor called AMPK (AMP-activated protein kinase), which functions as a cellular fuel gauge. When AMPK detects that energy is running low, it flips the cell into a conservation mode: it dials down processes that consume energy (like making new glucose or new fat) and ramps up processes that generate energy (like burning stored fat). AMPK activation by metformin involves an upstream enzyme called LKB1 that tags AMPK’s catalytic subunit, switching it on.² Low concentrations of metformin can also promote the physical assembly of AMPK’s component parts, making the enzyme easier to activate.³

A Surprising New Binding Target on Lysosomes

For years, the Complex I story was treated as essentially the whole upstream explanation. Then in 2022, a research team identified something unexpected: metformin physically binds to a protein called PEN2, a subunit of an enzyme complex called γ-secretase, which sits on the surface of lysosomes (the cell’s recycling compartments).⁴ When metformin locks onto PEN2, PEN2 pairs up with ATP6AP1, a component of a proton pump on the lysosome. That pairing shuts the pump down and recruits the LKB1 machinery to the lysosome surface, activating AMPK right there without any change in the cell’s overall AMP levels.⁵

This discovery matters because it means metformin can activate AMPK at low doses, potentially even at the concentrations that actually circulate in a patient’s blood, through a route that does not require the broader energy stress of Complex I inhibition. It also plugs into the lysosomal glucose-sensing pathway, a nutrient-detection system the cell uses independently of mitochondrial energy status. In practical terms, the finding helps explain something clinicians have observed for a long time: metformin works at doses that seem too low to substantially block Complex I.

How Liver Glucose Production Gets Dialed Down

The liver is constantly making glucose, a process called gluconeogenesis. In people with type 2 diabetes, this process runs too hot, flooding the bloodstream with sugar even when food is not being absorbed from the gut. Metformin’s most clinically important action is dampening that overproduction, and it does so through at least two distinct biochemical routes.

The first involves the enzyme fructose-1,6-bisphosphatase (FBP1), which controls a rate-limiting step in the glucose-making pathway. AMP, the low-energy molecule that builds up when metformin inhibits Complex I, directly blocks FBP1. Research has identified FBP1 as a major contributor to metformin’s glucose-lowering effect, essentially acting as a brake pedal on gluconeogenesis.⁶

The second route works through the cell’s internal balance of oxidized and reduced molecules, its redox state. Metformin inhibits an enzyme called mitochondrial glycerophosphate dehydrogenase (mGPD), which is part of a shuttle system that moves electrons between different compartments of the cell.⁷ When this shuttle is disrupted, the liver cell’s internal chemistry shifts in a way that makes it harder to convert raw materials like lactate and glycerol into glucose. Studies in animal models have confirmed this: when the mGPD enzyme is genetically removed, metformin loses its ability to lower glucose production and plasma glucose levels.⁸

What Metformin Does in the Gut

The gastrointestinal tract is not just a transit point for metformin; it is an active site of the drug’s effects. That is why delayed-release formulations designed to keep metformin in the gut still lower blood sugar effectively, and why the drug’s most common side effects are GI-related. In the intestine, metformin increases glucose uptake by the cells lining the gut wall, redirecting blood sugar into the tissue where it gets converted to lactate rather than circulating further.⁹ This effect is robust enough to show up on PET-CT imaging scans, where the gut of a metformin-treated patient lights up from increased glucose-analogue uptake.

Metformin also boosts levels of GLP-1, a hormone released from specialized cells in the lower intestine that stimulates insulin secretion and suppresses appetite. The drug appears to raise GLP-1 through both direct stimulation of the cells that produce it and indirect effects involving bile acid metabolism.¹⁰ This overlap with the GLP-1 pathway is interesting given that newer diabetes drugs (GLP-1 receptor agonists) have become blockbusters partly by targeting that same hormonal signal.

Then there is the gut microbiome. A randomized trial found that metformin reshapes the bacterial community in the intestine, increasing species that produce short-chain fatty acids like butyrate and acetate.¹¹ People taking metformin have been found to carry higher levels of Akkermansia muciniphila, a bacterium associated with healthy gut lining, along with several other short-chain fatty acid producers.¹² Short-chain fatty acids have wide-ranging metabolic benefits, from improving insulin sensitivity to strengthening the gut barrier. These microbiome changes may also explain some of metformin’s GI side effects, since rapid shifts in bacterial populations can produce bloating, gas, and diarrhea.

A Gut-Brain-Liver Communication Loop

One of the more surprising findings is that metformin in the upper intestine can send a signal to the liver via the nervous system. In rat models of insulin resistance, a brief infusion of metformin directly into the duodenum activated AMPK in the gut lining, which triggered local GLP-1 receptor signaling. That signal traveled through a neural pathway connecting the gut to the brain and then to the liver, ultimately reducing hepatic glucose production.¹³ When the duodenal AMPK was blocked, metformin lost this liver-directed effect entirely. This gut-brain-liver axis adds another layer to the drug’s action that is independent of metformin concentrations reaching the liver directly through the bloodstream.

How Metformin Gets into Cells

Metformin is a small, positively charged molecule that cannot easily cross cell membranes on its own. Instead, it hitches a ride on organic cation transporters, a family of membrane proteins that ferry charged molecules into and out of cells. The two most relevant are OCT1, which is heavily expressed in the liver and gut, and OCT2, which dominates in the kidneys.¹⁴ Metformin also uses OCT3, which is present in muscle and other tissues.¹⁵

This transporter dependence has real clinical implications. Genetic variants in the OCT1 gene are common and can reduce how much metformin the liver absorbs, which may partly explain why some patients respond less well to the drug. Certain medications can also compete for these same transporters. Proton pump inhibitors, for instance, have been shown in laboratory studies to inhibit metformin uptake through OCTs.¹⁶ Since proton pump inhibitors are among the most widely used drugs in the world, this interaction could be quietly affecting metformin’s effectiveness in many patients, though the clinical significance in practice is still debated.

Beyond Blood Sugar

Metformin’s mechanism of action extends into territory that has nothing obvious to do with diabetes. One important downstream effect is suppression of mTOR, a protein complex that acts as a master switch for cell growth and protein production. AMPK activation by metformin feeds into the mTOR pathway through an intermediary complex called TSC, putting the brakes on mTOR signaling. In liver cells, this suppression reduced active protein synthesis by roughly 75% in lab conditions.¹⁷ Separately, metformin can also inhibit mTOR through a protein called REDD1, a route that operates independently of AMPK.¹⁸

The mTOR connection is what drives interest in metformin as a potential anti-cancer agent. Because mTOR promotes cell growth and proliferation, suppressing it could theoretically slow tumor development. Population studies have suggested that metformin users may have lower cancer rates or better outcomes, and laboratory research supports an anti-cancer mechanism through this pathway.¹⁹ The evidence is still being tested in clinical trials, and it is worth noting that lab concentrations of metformin often far exceed what circulates in a patient’s blood. Still, the biological plausibility is there, and a drug that hundreds of millions of people already take safely would be a remarkably convenient addition to cancer prevention if the trials bear out.

Effects on Inflammation and Immune Cells

Metformin shifts immune cells toward a less inflammatory profile. In particular, it pushes macrophages (the immune system’s all-purpose cleanup crew) from a pro-inflammatory state toward an anti-inflammatory one, a shift characterized by increased production of anti-inflammatory signals like IL-10 and decreased production of inflammatory mediators like TNF-α and IL-6.²⁰ This macrophage reprogramming involves the same AMPK/mTOR axis and also reduces activity of the NLRP3 inflammasome, a molecular alarm system that drives sterile inflammation in many chronic diseases.²¹

The drug also appears to reshape the adaptive immune system, increasing the differentiation of T cells into regulatory and memory subtypes while reducing the capacity of neutrophils to form inflammatory traps.²² These immunomodulatory effects have generated interest in using metformin for autoimmune conditions and chronic inflammatory diseases, though clinical evidence in those areas is still limited.

Cardiovascular and Endothelial Mechanisms

Blood vessels are lined by a thin layer of endothelial cells, and when those cells stop functioning properly, it sets the stage for high blood pressure, atherosclerosis, and cardiovascular events. Metformin improves endothelial function through several routes. In animal models of diabetes, the drug reduced blood pressure and restored the ability of blood vessels to relax, partly by increasing nitric oxide availability and partly through a separate relaxation pathway involving endothelium-derived hyperpolarizing factor, an effect that was independent of blood sugar control.²³ A review of the evidence identified multiple molecular targets through which metformin protects the endothelium, including AMPK-dependent and AMPK-independent pathways involving NF-κB, eNOS, and SIRT1, among others.²⁴

Epigenetic Fingerprints

Metformin does not just change which enzymes are active right now; it can alter how genes are read over longer time scales. Through AMPK, metformin influences the activity of enzymes that add or remove chemical tags on DNA and on the histone proteins that DNA wraps around, including histone acetyltransferases, histone deacetylases, and DNA methyltransferases.²⁵ Metformin also changes the levels of numerous microRNAs, small molecules that fine-tune gene expression after the fact. These epigenetic effects may help explain some of metformin’s longer-term benefits, particularly in cancer prevention, because abnormal epigenetic patterns are a hallmark of tumor cells.

Metformin and Cellular Aging

The interest in metformin as a potential anti-aging drug stems directly from its mechanism of action. By modulating AMPK, mTOR, and cellular energy sensing, the drug touches several of the biological hallmarks of aging: mitochondrial dysfunction, chronic low-grade inflammation, loss of protein quality control, and cellular senescence.²⁶ In laboratory studies on human fat-tissue stem cells from older donors, metformin reversed markers of stress-induced senescence, restoring cell proliferation and reducing oxidative stress to levels resembling younger cells.²⁷ Whether these cellular-level effects translate to meaningful lifespan extension in humans is the question behind the ongoing TAME (Targeting Aging with Metformin) clinical trial, which aims to test whether metformin delays the onset of age-related diseases in older adults who do not have diabetes.

Metformin in Polycystic Ovary Syndrome

Polycystic ovary syndrome is one of metformin’s most established off-label uses, and the mechanism ties back to insulin. PCOS is driven in part by insulin resistance, which leads to excess androgen production from the ovaries. By improving insulin sensitivity through AMPK activation and the metabolic pathways already described, metformin reduces circulating androgen levels, restores more regular menstrual cycles, and can trigger ovulation.²⁸ Interestingly, the drug’s effect on ovulation may not rely entirely on those systemic improvements. Research has found evidence that the ovulatory response to metformin is probably also due to a direct action on the ovary itself, with the response depending on local sensitivity to the drug.²⁹

Safety Concerns Tied to the Mechanism

The side effects people worry most about with metformin connect directly to its mechanism. Lactic acidosis, the rare but serious complication, is essentially the drug’s intended effect taken too far. When metformin inhibits Complex I and shifts the redox balance in the liver, one byproduct is increased lactate. Under normal circumstances the body clears this easily. But when metformin accumulates to very high levels, particularly in people whose kidneys are failing, the combination of excessive lactate production and impaired clearance can tip blood pH dangerously low. In a case series of patients with metformin-associated lactic acidosis, the average blood metformin concentration on admission was about 30 mg/L with average lactate around 13 mmol/L and average pH around 7.0, all in the context of acute kidney injury, vomiting, diarrhea, or tissue hypoxia.³⁰ The condition is genuinely dangerous but also genuinely rare, and in almost all reported cases it occurred alongside another major insult to the body rather than from metformin alone.

The other mechanistic side effect worth knowing about is vitamin B12 deficiency. Metformin interferes with B12 absorption in the terminal ileum, likely by disrupting a calcium-dependent step in which B12 bound to intrinsic factor attaches to receptors on the gut wall.³¹ Over years, this can deplete B12 stores enough to cause neuropathy, which is particularly problematic because diabetic neuropathy can mask or mimic B12-related nerve damage, making it easy to miss. The encouraging finding is that supplemental calcium appears to reverse this malabsorption, restoring B12 uptake even while metformin use continues.³² Periodic B12 monitoring in long-term metformin users is a sensible precaution that is underutilized in routine practice.

From Goat’s Rue to Modern Medicine

Metformin descends from a compound found in Galega officinalis, a plant commonly known as goat’s rue that was used in traditional European herbal medicine. The plant is rich in guanidine, which was shown to lower blood glucose back in 1918.³³ Guanidine itself was too toxic for clinical use, but chemists developed related biguanide compounds, of which metformin turned out to have the best safety profile. It was introduced clinically in France in the late 1950s but did not gain FDA approval in the United States until 1995, decades behind most of the world. The irony is that this extraordinarily well-studied drug’s molecular mechanism is still being actively revised, with the PEN2 discovery in 2022 showing that the field had missed a direct binding target for over sixty years. If anything, metformin is a reminder that widespread clinical use and deep mechanistic understanding do not always advance on the same timeline.