Probiotics ease migraine-like responses in mice with genetic gut-brain support

A compound probiotic supplement reduced migraine-like pain and suppressed neuronal activation in a mouse model of migraine, according to a study that combined human genetic data with laboratory experiments in male mice. The work began with two-sample Mendelian randomization analyses of gut microbial traits, migraine and migraine subtypes, followed by mediation analyses of brain MRI phenotypes. In the animal experiment, eight-week-old male C57BL/6J mice received preventive oral supplementation with a formulation containing six bacterial species for 21 days before repeated nitroglycerin injections used to induce migraine-like pain.

The findings support a possible role for the microbiota–gut–brain axis in migraine susceptibility, but they do not establish that probiotics can prevent migraine in people. The human genetic component used a gut microbiome genome-wide association study involving 8,956 people from Germany. Its 430 microbial features ranged from broad taxonomic groups to sequence-defined clusters, providing researchers with genetic instruments for testing whether inherited variation in microbial abundance could influence migraine risk.

Migraine summary statistics came from an international dataset of 1.3 million individuals, including 74,495 people with migraine, 16,603 with migraine with aura and 11,718 with migraine without aura. The researchers used single-nucleotide polymorphisms as instrumental variables, applied linkage disequilibrium pruning, removed variants associated with smoking and drinking, and evaluated results with several Mendelian randomization estimators. A causal association was considered supported when inverse-variance weighted estimates passed a Bonferroni-corrected threshold and estimates from additional methods pointed in the same direction.

Eight gut microbial features were causally associated with migraine or one of its clinical subtypes in the genetic analysis. By contrast, higher genetically predicted Paraprevotella abundance was associated with lower migraine odds, with an odds ratio of 0.95, a 95% confidence interval of 0.94–0.97 and a p-value of 2.69 × 10−6. The subtype results were more heterogeneous: several sequence-defined taxa were associated with lower risk of migraine without aura, whereas another Prevotella feature and a Parasutterella feature were associated with higher risk.

The authors noted that different sequence-defined features within the same broad family need not have identical relationships with disease, making the taxonomy-specific findings important. The researchers then screened 2,347 brain MRI phenotypes and identified 30 causal associations with migraine or its subtypes that survived multiple-testing correction. These included regional measures of cortical thickness, surface area, volume, folding, curvature, fractional anisotropy, mean diffusivity, intracellular volume fraction, isotropic volume fraction and orientation dispersion.

Among the findings, genetically predicted mean curvature in the anterior cingulate and medial prefrontal cortex had the largest reported association with migraine, corresponding to an odds ratio of 2.30 and a 95% confidence interval of 1.76–3.01. Intracellular volume fraction in posterior insular area 2 was associated with higher migraine odds, whereas the folding index in the same region was associated with lower odds. Such estimates concern genetically instrumented differences in brain structure; they do not mean that changing an MRI measure would directly change a person’s risk of developing migraine.

Mediation analysis narrowed attention to three microbial–brain pathways that passed correction. Holdemanella OTU99_54 was positively associated with cortical thickness in posterior cingulate area 31a, Parasutterella OTU97_15 was positively associated with intracellular volume fraction in the orbital frontal complex, and Paraprevotella was negatively associated with intracellular volume fraction in the posterior insular cortex. The latter pathway accounted for an estimated 19.01% of the genetically predicted association between Paraprevotella and migraine, with a 95% confidence interval of 4.62–33.28% and a p-value of 0.0019.

Multivariable analysis indicated that the Paraprevotella association remained after adjustment for the insular measure, while the adjusted insular effect was not statistically significant, suggesting that the proposed mediator explained only part of the relationship. Reverse Mendelian randomization found no consistent evidence that migraine caused the tested changes in Paraprevotella or posterior insular intracellular volume fraction. The laboratory component tested whether preventive probiotic supplementation could reproduce aspects of the genetic signal.

Mice were assigned to a saline-treated control group, a vehicle-treated nitroglycerin model group or a probiotic-treated nitroglycerin group. The supplement contained Lactobacillus acidophilus, Lactobacillus casei, Lactobacillus rhamnosus, Bifidobacterium lactis, Bifidobacterium adolescentis and Bifidobacterium longum. The source did not specify the individual strains, which were administered by oral gavage for 21 consecutive days.

Repeated intraperitoneal nitroglycerin lowered mechanical pain thresholds in the periorbital and hind-paw regions, indicating increased sensitivity. Compared with untreated model animals, probiotic-pretreated mice had higher thresholds after nitroglycerin exposure. Probiotics did not significantly change baseline or post-injection thresholds in saline-treated mice.

These observations concern a preventive regimen in an induced mouse model; they do not show that the supplement treats established migraine attacks or that the same dose and formulation would affect human symptoms. Microbiome sequencing indicated that nitroglycerin-treated mice had reduced gut microbial richness and a community structure distinct from that of controls. The probiotic intervention significantly restored microbial richness.

An analysis of community dissimilarities found differences among the three groups, with a PERMANOVA p-value of 0.003, but the test did not specifically establish that probiotics changed beta diversity relative to the model group. Paraprevotella abundance declined after nitroglycerin treatment and rose again after probiotic supplementation. Because its sequence reads were relatively sparse, the researchers also used taxon-specific quantitative PCR, which confirmed lower Paraprevotella abundance in the model group than in controls and a significant increase with probiotics.

No significant between-group differences were detected for Lachnospiraceae. The experiment therefore connected the supplement with preservation of a microbial feature highlighted by the human genetic analysis, but it did not establish that Paraprevotella alone caused the behavioral or neural effects. Immunofluorescence imaging extended the behavioral findings to brain regions implicated in nociception.

In the spinal trigeminal nucleus caudalis, nitroglycerin increased both c-Fos expression and calcitonin gene-related peptide immunoreactivity. Probiotics reduced the number of activated neurons and lowered CGRP fluorescence intensity relative to the model group. The supplement also diminished c-Fos activation in the paraventricular thalamus and medial prefrontal cortex.

These markers indicate changes in neuronal activity and peptide expression after treatment, rather than direct evidence of restored normal brain structure or the elimination of central sensitization. The investigators collected brain tissue within 30 minutes of behavioral testing to reduce the possibility that the testing itself caused c-Fos expression. The study has important limitations.

Its genetic datasets were restricted to people of European ancestry, so the results may not generalize to other populations. The mouse experiments included only males, limiting conclusions about sex-specific effects despite the greater prevalence of migraine among women. Probiotic responses may also vary with strain, dose, duration, formulation, baseline microbiota and disease phenotype, and the specific contribution of Paraprevotella remains unresolved.

Brain microstructure was not directly measured in the animals, so the connection between the human mediation analysis and the mouse neural findings remains inferential. Although earlier clinical studies discussed by the authors reported benefits from other probiotic or synbiotic formulations, the present experiment did not test migraine prevention in patients. The combined results identify testable links among microbial composition, brain traits and experimental migraine-like responses, while controlled human trials would be needed to determine whether targeted microbiome intervention has clinical value.

Subject of Research: Genetic and mouse-model evidence linking gut microbiota, brain microstructure and migraine-like responses

Article Title: Preventive effects of compound probiotics supplementation on migraine: Mendelian randomization analysis and experimental evidence

Article References: Bai, L., Zhang, Q., Zhu, J., Zhang, W., Cui, Z., Jia, T., Li, C., & Guo, Z. (2026). Preventive effects of compound probiotics supplementation on migraine: Mendelian randomization analysis and experimental evidence. Frontiers in Microbiology, 17, Article 1939532. https://doi.org/10.3389/fmicb.2026.1939532

Image Credits: AI Generated

DOI: 10.3389/fmicb.2026.1939532

Keywords: migraine, gut microbiome, probiotics, Mendelian randomization, microbiota–gut–brain axis, brain MRI, Paraprevotella, mouse model, CGRP, neuronal activation, Preventive, effects