Gut Bacteria Secretly Destroy Metformin, But a Plant Compound Can Stop Them
A gut microbial enzyme degrades metformin before it works — and a soy-derived compound may block it, restoring drug efficacy.
Summary
Metformin is the world's most prescribed diabetes drug, yet many patients respond poorly for unknown reasons. This study from Nanjing University reveals that certain gut bacteria produce an enzyme called AdhE that chemically breaks down metformin, converting it into inactive compounds before it can act. People with a weak response to metformin harbor more of these AdhE-producing bacteria, specifically a species called Eubacterium hallii. In mice, removing or disabling AdhE restored drug levels and blood sugar control. Researchers then used AI to screen thousands of compounds and identified daidzein — a phytochemical found in soy — as a potent AdhE inhibitor that blocks metformin degradation and improves its effectiveness. This discovery reframes metformin non-response as a fixable gut microbiome problem and opens the door to microbiome-guided strategies for optimizing existing medications.
Detailed Summary
Metformin remains the cornerstone first-line therapy for type 2 diabetes and is also studied as a longevity drug, yet individual responses vary enormously. For decades, this variability was poorly understood. A new study published in Cell Metabolism offers a striking explanation: some people's gut bacteria are simply eating the drug.
Researchers at Nanjing University identified a microbial enzyme — aldehyde-alcohol dehydrogenase, or AdhE — produced by Eubacterium species that are enriched in individuals who respond poorly to metformin. AdhE converts metformin into pharmacologically inactive byproducts, reducing both intestinal absorption and systemic drug exposure before the medication can reach its metabolic targets.
To confirm causation, the team genetically deleted AdhE from the relevant bacteria and separately used targeted phage therapy to deplete AdhE-producing Eubacterium hallii MD1 in mice. Both interventions restored metformin bioavailability and improved glycemic control, demonstrating that microbial drug degradation is a direct and mechanistically tractable cause of therapeutic failure.
The researchers then leveraged artificial intelligence for structure-based screening of dietary compounds as potential AdhE inhibitors. The top hit was daidzein, an isoflavone phytochemical abundant in soy and other legumes. Daidzein blocked AdhE activity, prevented metformin degradation in the gut, and improved metformin responsiveness in animal models.
For longevity-focused clinicians and researchers, the implications extend beyond diabetes management. Metformin is under active investigation as a geroprotective agent in the TAME trial. If microbial AdhE activity limits its efficacy in a significant proportion of users, this mechanism could partly explain why some individuals see robust anti-aging benefits while others do not. Testing for AdhE-producing bacteria and co-administering daidzein or other AdhE inhibitors could eventually personalize metformin therapy. Limitations include that human clinical validation is still needed, and the summary is based on the abstract only.
Key Findings
- Gut bacteria produce AdhE enzyme that converts metformin into inactive compounds, directly reducing drug efficacy.
- Eubacterium hallii MD1 is the primary AdhE-producing species enriched in poor metformin responders.
- Genetic or phage-mediated removal of AdhE in mice restores metformin levels and blood sugar control.
- AI-guided screening identified daidzein (a soy isoflavone) as a dietary AdhE inhibitor that rescues metformin activity.
- Gut microbial drug inactivation is a modifiable, targetable factor — not a fixed pharmacological limitation.
Methodology
The study combined human microbiome profiling of metformin responders vs. non-responders, mouse genetic and phage-depletion experiments to establish causality, and an AI-guided structure-based virtual screen of dietary phytochemicals against the AdhE enzyme. Both intestinal and systemic drug exposure were measured as pharmacokinetic endpoints alongside glycemic outcomes in murine models.
Study Limitations
This summary is based on the abstract only, as the full paper was not available; specific effect sizes, patient numbers, and mechanistic details could not be assessed. All causal experiments were conducted in mice, and human clinical validation of daidzein as an AdhE inhibitor has not yet been reported. The prevalence of AdhE-producing Eubacterium species across diverse human populations remains to be characterized.
Enjoyed this summary?
Get the latest longevity research delivered to your inbox every week.
Enter your email to subscribe: