https://e-dmj.org/journal/view.php?doi=10.4093/dmj.2026.0248
chatGPT:
Summary
This is a review article on metformin, arguing that it should no longer be understood simply as a glucose-lowering drug for type 2 diabetes, but as a multi-organ metabolic modulator with effects on mitochondria, redox state, gut biology, ageing pathways, neurodegeneration and cancer.
The paper’s central claim is that metformin’s actions arise from several interacting mechanisms:
-
Classical hepatic AMPK pathway
Metformin can inhibit mitochondrial complex I, increasing AMP/ATP and ADP/ATP ratios, which activates AMPK. This then suppresses hepatic gluconeogenesis, inhibits mTORC1, reduces lipogenesis and increases metabolic flexibility. -
AMPK-independent mitochondrial mechanisms
The paper emphasizes that AMPK is not always required for metformin’s glucose-lowering effects. It discusses direct effects on mitochondrial respiration, complex I, ROS production, Nrf2 activation and mitochondrial glycerophosphate dehydrogenase/GPD2 inhibition. The GPD2 mechanism is presented as a way to alter the cytosolic NADH/NAD⁺ balance and selectively reduce gluconeogenesis from lactate and glycerol. -
Gut-centred pharmacology
A major theme is that oral metformin accumulates at very high levels in the intestine, far above plasma levels. The authors argue that intestinal actions are not secondary but central to the drug’s effects. These include altered glucose handling, lactate production, bile acid metabolism, GLP-1 release, microbiome changes and gut–brain–liver signalling. -
Microbiome and bile acid effects
Metformin is described as increasing organisms such as Akkermansia muciniphila and some short-chain-fatty-acid-producing taxa, although the paper acknowledges taxon-specific and inconsistent findings. It also discusses the microbiome–bile acid–FXR pathway and SCFA-mediated effects on gut–brain signalling. -
“Intestinal glucotonic effect”
This is one of the paper’s more distinctive sections. The authors describe a proposed process in which metformin promotes movement of glucose from the blood into intestinal epithelial cells and then into the gut lumen. Mechanistically, they link this to complex I inhibition, reduced ROS, lower TXNIP, and GLUT1 membrane translocation. This would turn the intestine into an active glucose-disposal organ rather than merely an absorptive surface. -
Ageing and neurodegeneration
The paper argues that metformin may influence ageing hallmarks through AMPK–mTORC1, SIRT1, NF-κB, mitochondrial quality control, autophagy, antioxidant signalling and microbiome-derived metabolites. It discusses observational links with lower dementia risk and possible effects in Alzheimer’s and Parkinson’s disease, while noting that RCT evidence is still limited. -
Cancer and immunometabolism
The review is cautious about earlier cancer claims. It notes that some observational benefits may have been overstated due to immortal time bias. The authors shift the emphasis from metformin as a direct anticancer drug to metformin as a possible immunometabolic adjuvant, affecting tumour metabolism, CD8 T cells, NK cells, macrophages and the tumour immune microenvironment.
What is novel or interesting?
The paper is not presenting new experimental data; its novelty is mainly conceptual synthesis. The most novel aspects are:
-
Reframing metformin as intestine-first rather than liver-first
The review gives strong emphasis to gut sequestration, intestinal glucose uptake, microbiome changes and gut–brain–liver signalling. This is a useful correction to the older model in which metformin’s main action was simply hepatic AMPK activation. -
The “intestinal glucotonic effect”
This is the most distinctive mechanistic idea in the paper. The concept that metformin may actively divert circulating glucose into the gut lumen, potentially feeding microbiota and lowering systemic glucose, is a more sophisticated model than simply “metformin reduces hepatic glucose production.” -
Integration of redox biology with glucose flux
The ROS–TXNIP–GLUT1 axis in enterocytes links mitochondrial redox state to intestinal glucose transport. That is mechanistically interesting because it connects mitochondrial inhibition, antioxidant signalling and systemic glucose disposal. -
A precision-medicine framing
The paper repeatedly argues that response to metformin depends on transporter genetics, tissue exposure, microbiome composition, age, metabolic phenotype and context. This is a useful move away from treating metformin as a uniform intervention. -
Balanced discussion of geroprotection
The paper does not simply promote metformin as an anti-ageing drug. It notes possible negative interactions, including attenuation of exercise adaptation and context-dependent effects in older organisms.
Critique
The review is useful, but there are several weaknesses.
1. It sometimes over-integrates mechanisms that are not equally proven
The paper presents a highly connected model: complex I inhibition, AMPK, GPD2, Nrf2, GLUT1, GDF15, microbiome, GLP-1, bile acids, FXR, ageing and cancer immunity. This is intellectually attractive, but the evidence strength varies greatly across these claims. Some mechanisms are well established; others remain early, model-dependent or inferential.
The risk is that the reader comes away with a unified mechanistic map that looks more settled than the field really is.
2. The intestinal glucotonic effect is interesting but still not fully established
The proposed blood-to-lumen glucose flux is one of the most interesting ideas, but it remains relatively new. Increased intestinal FDG uptake on PET does not automatically prove luminal glucose excretion, and the quantitative importance of this pathway in ordinary clinical metformin use remains uncertain.
A key missing question is: how much of metformin’s glucose-lowering effect in humans is actually explained by this pathway compared with hepatic gluconeogenesis, appetite reduction, GLP-1, weight loss and microbiome effects?
3. The ageing section risks extrapolating beyond the clinical evidence
The ageing biology is plausible, especially via mTOR, AMPK, inflammation, mitochondrial stress and redox pathways. But the clinical geroprotection case remains unproven. Observational data are confounded by diabetes status, health-system contact, prescribing bias and comparator choice.
The paper does mention TAME and the need for long-term RCTs, but the mechanistic discussion may still give a stronger impression than the clinical evidence warrants.
4. Neurodegeneration claims need caution
The paper notes associations between metformin use and lower dementia risk, but this is a difficult area because diabetes itself affects dementia risk, and metformin users may differ systematically from non-users. There are also potential countervailing concerns, such as vitamin B12 depletion, frailty, renal impairment and heterogeneity by age and metabolic state.
The review would be stronger if it separated more sharply:
- metformin’s effects in diabetic patients,
- possible effects in non-diabetic ageing,
- and possible disease-modifying effects in established neurodegenerative disease.
5. Cancer section is appropriately cautious, but still speculative
The authors rightly acknowledge that earlier observational oncology findings may have been inflated by immortal time bias and that large trials have often been disappointing. Their pivot to “immunometabolic adjuvant” is plausible, but this remains more of a research programme than a clinical conclusion.
The paper could have been clearer that metformin should not currently be considered a general anticancer therapy outside specific trial contexts.
6. Little emphasis on dose, tissue concentration and tolerability
The review discusses pharmacokinetics well, but many mechanistic studies use concentrations that may not correspond to human tissue exposures. Metformin’s intestinal concentrations may be high, but plasma and tissue levels are much lower. This matters especially for claims about complex I inhibition, cancer cell metabolism and anti-ageing effects.
It also could say more about practical constraints: gastrointestinal intolerance, B12 deficiency, renal function, frailty, lactic acidosis risk in vulnerable patients and possible blunting of exercise adaptation.
Overall assessment
This is a good, modern review of metformin biology. Its main value is in shifting the conceptual frame from “metformin activates AMPK in the liver” to a broader model involving intestinal drug accumulation, mitochondrial redox reprogramming, microbiome signalling, gut–brain–liver communication and context-dependent ageing biology.
Its weakness is that it sometimes compresses very different levels of evidence into one coherent diagram. The best-supported areas are metformin’s gut accumulation, transporter-dependent pharmacokinetics, hepatic and intestinal glucose effects, microbiome changes, and AMPK-independent mitochondrial/redox mechanisms. The more speculative areas are broad geroprotection, neuroprotection and oncology adjuvant effects.
In short: strong as a mechanistic synthesis, weaker as evidence for broad clinical repurposing.