Gut & MicrobiomeReview ArticleOpen Access

Gut Microbes May Unlock Exercise's Health Benefits as Drug-Like Molecules

Researchers propose microbial-derived exerkines — gut metabolites shaped by exercise — as a new drug discovery framework targeting aging, cancer, and more.

Thursday, October 1, 2026 3 views
Published in Am J Physiol Cell Physiol
A split image: on one side, a person jogging on a trail; on the other, a close-up of labeled glass vials containing gut bacterial cultures in a research laboratory

Summary

Exercise changes the bacteria in your gut, and those bacteria release chemicals that may carry some of exercise's health benefits to the rest of the body. This review calls these chemicals microbial-derived exerkines, or MDEs. Examples include short-chain fatty acids, pipecolic acid, succinate, indole-3-propionic acid, two cardioprotective acids (3-HPA and 4-HBA), and lithocholic acid. In animal studies, giving these compounds directly protected muscle, improved exercise performance, shielded the heart, and reduced Alzheimer's-type pathology. Because no approved drug exists for general muscle wasting, the authors argue that mining the exercise-microbiome link could open a new route to drug discovery. Most of the evidence comes from mice, though, so human trials are still needed before anyone can swallow these benefits in a pill.

Detailed Summary

Exercise is one of the most powerful therapies we have, yet many people who would benefit most, such as those with cancer cachexia or advanced heart failure, cannot do it. This review by Burke, Valentino, Ismaeel, Wen and McCarthy asks whether the gut microbiome can deliver some of exercise's benefits in a different form. It proposes a new class of molecules, microbial-derived exerkines (MDEs), defined by three criteria: gut microbial production, exercise responsiveness, and a role in mediating exercise benefits.

The authors first lay out the two-way relationship between exercise and gut bacteria. Wheel running in rodents changed cecal microbiota and raised butyrate. In humans, athletes carry different microbial communities than untrained people, and cardiorespiratory fitness independently predicts microbial diversity. Supervised training shifts composition in ways that reverse with detraining. In the other direction, germ-free mice show impaired exercise performance and muscle growth, and recolonization or metabolite delivery can rescue these deficits.

The review then catalogs established and candidate MDEs. Short-chain fatty acids (acetate, butyrate, propionate) were produced preferentially in the microbiomes of prediabetic people who responded to exercise. In obese mice colonized with non-responder microbiota, butyrate and propionate partially rescued glucose homeostasis. Formate enabled exercise-enhanced CD8+ T cell antitumor immunity in melanoma models. The authors' own work found pipecolic acid and succinate elevated in exercised mice, and together they prevented disuse atrophy and improved strength and mitochondrial respiration. Indole-3-propionic acid, 3-hydroxyphenylacetic acid, 4-hydroxybenzoic acid and lithocholic acid (a candidate MDE, since some reports show it falling with training) round out the list.

Disease applications are broad. In colitis, SCFAs, IPA and lithocholic acid were protective, but succinate accumulates in inflamed bowel and Sucnr1-deficient mice were protected, so effects are organ-specific. In aging, IPA extended Drosophila lifespan and improved muscle function in aged mice, and acetate blunted age-related muscle changes in rats. In heart disease, 3-HPA and 4-HBA reproduced exercise's cardioprotection through NRF2 even during antibiotic depletion. IPA was low in humans and mice with HFpEF and improved diastolic dysfunction in mice. In Alzheimer's models, IPA and butyrate reduced amyloid pathology and improved cognition. Higher IPA tracked with lower type 2 diabetes incidence in Finnish cohorts.

The pitch for drug discovery is strongest for muscle loss. No FDA-approved drug exists for general skeletal muscle atrophy, and target-based candidates have repeatedly failed. An exercise-guided, target-agnostic screen of microbial metabolites offers an alternative. The authors also note possible synergy between MDEs, which is untested.

The caveats are substantial. Much causal evidence comes from germ-free or antibiotic-treated rodents, which are immunologically and metabolically atypical, and antibiotics themselves harm host tissue. Mouse and human gut microbiomes share only about 4% of genes. Effects are context-dependent: formate helps antitumor immunity in melanoma yet promotes colorectal cancer. Human dose-finding, target engagement and randomized trials with functional endpoints are needed.

Key Findings

  • Butyrate and propionate partially rescued glucose homeostasis and insulin sensitivity in obese mice given post-training microbiota from exercise non-responders with prediabetes
  • Pipecolic acid plus succinate together prevented disuse muscle atrophy and improved strength, exercise performance and mitochondrial respiration in mice
  • 3-HPA and 4-HBA, enriched in feces of exercised mice, each reproduced exercise's cardioprotection (less hypertrophy, fibrosis, apoptosis) even during antibiotic-induced microbiome depletion, via NRF2
  • Indole-3-propionic acid attenuated amyloid-β pathology and improved cognition in APP/PS1 mice, extended Drosophila lifespan, and improved muscle function in aged mice
  • Higher serum IPA was associated with lower type 2 diabetes incidence and better insulin secretion in the Finnish Diabetes Prevention Study (associative only)
  • Exercise-induced microbial formate was both sufficient and required for CD8+ T cell antitumor immunity and checkpoint inhibitor efficacy in preclinical melanoma, yet formate promotes colorectal cancer progression
  • Mouse and human gut microbiomes share only about 4% of microbial genes, and no FDA-approved drug exists for general skeletal muscle atrophy

Methodology

This is a narrative review, not a systematic review or meta-analysis, so no formal search strategy, pooled sample size or statistical synthesis is reported. The authors propose a three-criteria definition for MDEs (microbial production, exercise responsiveness, mediation of benefit) and map the literature onto it in two tables. Evidence is drawn mostly from rodent studies, including germ-free, antibiotic-depleted and fecal transplant models, with supporting human observational and intervention data. Candidate versus established MDEs are distinguished by how many criteria are met.

Study Limitations

The authors acknowledge that most causal evidence comes from germ-free or antibiotic-treated rodents, which are atypical and may exaggerate metabolite effects, and that human-mouse microbiome differences limit translation. Several candidates, such as lithocholic acid, have conflicting exercise data, MDE synergy is untested, and effects are disease- and organ-specific. As a narrative review by authors who have their own MDE research on pipecolic acid and succinate, selection bias is possible; no specific conflict-of-interest statement was in the supplied text.

Enjoyed this summary?

Get the latest longevity research delivered to your inbox every week.

Enter your email to subscribe: