A large genetic study has uncovered tentative evidence that specific gut bacteria could influence the risk of multiple sclerosis (MS) through the activation of particular immune cells, offering a fresh — though still highly preliminary — glimpse into the hidden communication channels that may connect the microbes living in our intestines with the destruction of myelin in the brain and spinal cord.

The new research, published in the journal Immunity, Inflammation and Disease, used a sophisticated statistical technique known as Mendelian randomization (MR) to probe whether the gut microbiota might exert causal effects on MS risk, and whether immune cell traits stand in the middle of that relationship. The team, led by Pingping Ning and Rui Li, drew on genetic data from tens of thousands of people to build what they describe as an exploratory “gut microbiota–immune traits–MS axis.” Their headline finding: two well-known gut bacterial genera, Roseburia and Ruminococcus2, appear to influence MS risk at least partly by modulating markers on the surface of CD4+ and CD8+ T cells — the very soldiers of the adaptive immune system that are widely suspected of turning against the nervous system in people with MS.

Multiple sclerosis is one of the most common causes of non-traumatic disability in young adults, affecting an estimated 2.8 million people worldwide. The disease is characterized by the immune system’s assault on the myelin sheaths that insulate nerve fibers in the central nervous system, producing symptoms ranging from motor weakness and sensory loss to vision impairment and cognitive decline. Its incidence has climbed sharply in recent years — with roughly 500,000 new cases added globally since 2013 — and women are about three times more likely than men to be diagnosed. The annual economic burden of MS in the United States alone is estimated at $85.4 billion. Yet despite decades of research, the precise triggers of the disease remain stubbornly unclear, with evidence pointing to an intricate interplay between genetics, environmental exposures, and lifestyle factors.

That is where the gut microbiome enters the picture. The intestine is far more than a digestive organ; it is a major hub of immune regulation, and the trillions of microbes it harbors are known to shape immune function in profound ways. When the microbial community becomes imbalanced — a state known as dysbiosis — the production of microbial metabolites can shift, altering the behavior of antigen-presenting cells, T cells, and B cells. This, in turn, may set the stage for autoimmune diseases. Previous studies have found altered gut bacterial composition and reduced levels of short-chain fatty acids in people with MS, but whether these microbial changes actually cause the disease — and through what biological intermediaries — has remained an open question.

Mendelian randomization offered the researchers a way to address that question with observational data. The technique exploits the random allocation of genetic variants at conception as a natural experiment: because single nucleotide polymorphisms (SNPs) are inherited randomly and are fixed at conception, they are largely immune to the confounding and reverse-causation biases that plague traditional observational studies. In essence, MR uses genetic variants as proxies — instrumental variables — to ask whether an exposure genuinely influences an outcome, in a manner loosely analogous to a randomized controlled trial.

The data underpinning the study were formidable. For gut microbiota, the team used summary statistics from the MiBioGen consortium, covering 18,340 participants from 24 cohorts, most of European ancestry, spanning 211 microbial taxa across phyla, classes, orders, families, and genera. For immune traits, they turned to a dataset of 731 immune cell phenotypes measured by flow cytometry in 3,757 individuals from a Sardinian founder population — a rich catalog capturing absolute cell counts, surface antigen expression levels, morphological parameters, and ratios between cell populations. For MS outcomes, they accessed genome-wide association data from the International Multiple Sclerosis Genetics Consortium comprising 47,429 MS cases and 68,374 controls of European ancestry.

The analysis proceeded in stages. First, the researchers screened all 211 microbial taxa for genetic associations with MS and identified six taxa with nominally significant associations — four apparently protective, two apparently risk-increasing. Notably, the genus Ruminiclostridium5 showed the strongest nominal protective signal (odds ratio 0.695, 95% confidence interval 0.554–0.871; p = 0.0016). Next, they tested the 731 immune traits for causal links to MS, finding 52 with nominal associations. Finally, they searched for the crucial middle links: microbial taxa and immune traits that formed a statistically coherent triangle, in which the microbe was associated with both the immune trait and MS, and the immune trait was associated with MS, with effect directions consistent with a genuine mediation pathway.

Only two such pathways survived this filtering. The genus Roseburia appeared to influence MS through the expression of CD28 on CD28+ CD4+ T cells, with an estimated 8.97% of the taxon’s nominal effect on MS mediated through this immune trait. The genus Ruminococcus2 appeared to act through CD45 expression on HLA-DR+ CD8+ T cells, mediating an estimated 12.18% of its nominal effect. Both mediation effects had confidence intervals that did not cross zero, lending them a degree of statistical credibility despite the overall exploratory nature of the analysis.

The mechanistic story behind these findings is biologically plausible. Both Roseburia and Ruminococcus2 are short-chain fatty acid (SCFA)-producing bacteria — fermenters of dietary fiber that generate acetate, propionate, and butyrate. SCFAs nourish colonic epithelial cells and regulate inflammatory responses throughout the body, and they play a key role in gut-brain communication. In MS patients, gut bacterial composition and SCFA levels are often reduced, which is thought to impair regulatory T cell function. Intriguingly, oral propionate supplementation has been shown to increase propionate availability in cerebrospinal fluid and improve clinical outcomes in MS patients, while butyrate ester treatment in mouse models of MS alleviated symptoms, reduced pro-inflammatory Th1 and Th17 cells, and boosted regulatory T cell proportions and IL-10 secretion.

The immune mediators identified in the study are equally compelling. CD28 is a co-stimulatory receptor on T cells that binds CD80 and CD86 on antigen-presenting cells, amplifying T cell activation when the T cell receptor recognizes an antigen. In MS patients, CD28+ CD4+ T cell activation is typically enhanced, and over-activated CD28+ T cells secrete interferon-γ and tumor necrosis factor-α, recruit B cells and macrophages, and contribute to the demyelination process. Excessive CD28 signaling may drive T cells to mount responses against neuroantigens, making CD28-targeting inhibitors a potential future therapeutic strategy. On the CD8+ side, HLA-DR expression marks activated cytotoxic T cells, and HLA-DR+ CD8+ T cells have been implicated in myelin destruction and neurodegeneration — they can recognize central nervous system antigens, exert cytotoxic effects, and release pro-inflammatory cytokines that worsen neuronal damage. CD45, meanwhile, is a tyrosine phosphatase that fine-tunes T cell receptor signaling, making its expression level a meaningful indicator of T cell responsiveness.

The authors are emphatic that these results must be interpreted with caution. All associations reported in the study were nominally significant only — with uncorrected p-values below 0.05 — and none survived false discovery rate (FDR) correction. In other words, the findings are hypothesis-generating clues, not proven causal relationships. The team itself attributes this partly to the low heritability of gut microbial taxa, which limits the statistical power of MR analyses of the microbiome, and to the fact that gut microbiota are heavily shaped by diet and environment, factors that genetic data alone cannot fully capture.

Several other limitations temper the conclusions. The immune trait data came from a Sardinian founder population, which may not represent immune trait distributions elsewhere in the world. The researchers did not perform MR-PRESSO outlier detection due to the computational burden of 58 exposure factors, and reverse MR analysis was not possible because no valid instrumental variables existed for MS as an exposure — meaning reverse causality cannot be ruled out. Multivariable MR and genetic colocalization analyses, which could help disentangle confounding and verify shared genetic signals, were likewise beyond the scope of the study and are flagged as priorities for future well-powered investigations.

Still, the study represents a meaningful step toward mapping the causal architecture of the microbiota-immune-disease axis in MS. By combining two-stage MR with formal mediation analysis, the researchers constructed a coherent, testable framework: fiber-fermenting gut bacteria produce SCFAs, which modulate the activation state of CD4+ and CD8+ T cells, which in turn influence the risk and progression of multiple sclerosis. If confirmed in independent populations and validated in laboratory experiments, this framework could point toward novel biomarkers for predicting disease progression — and toward microbiome-targeted or immune-cell-targeted therapies that intervene early in the disease process. For now, the message from this study is one of cautious optimism: the gut may indeed be whispering to the immune system about MS, and scientists are finally beginning to hear what it says.

The compound at the center of the study is 3,4-dihydroxybenzoic acid, or DHB, a metabolite generated by members of the gut microbiome. The researchers found that DHB can influence the behavior of CD8+ T cells, immune cells responsible for identifying and destroying infected or malignant cells. In cancer, however, these cells can gradually become dysfunctional or “exhausted,” losing their ability to proliferate, survive, and maintain sustained attacks against tumors. This exhaustion is one reason why therapies such as immune checkpoint inhibitors, including anti-PD-1 drugs, work effectively in only a subset of patients.

According to the study, DHB promotes the development of antigen-specific CD8+ T cells with stem-like properties. These cells are not fully differentiated into short-lived effector cells; instead, they retain the capacity for self-renewal and can generate new waves of tumor-directed immune cells. This stem-like state is considered important for durable antitumor immunity because it provides a continuing source of T cells capable of responding to cancer over extended periods. The researchers reported that DHB helped preserve this functional state while enhancing the ability of the cells to recognize tumor-associated antigens.

The mechanism appears to involve the regulation of cellular metabolism. Activated T cells commonly increase glycolysis, a rapid pathway for converting glucose into energy and producing metabolic intermediates needed for growth. Although glycolysis is useful during an acute immune response, excessive or prolonged reliance on it can contribute to T-cell dysfunction in the tumor microenvironment. The study indicates that DHB suppresses this metabolic shift and modulates the Akt-mTORC1-Myc signaling pathway, a network that coordinates nutrient sensing, cell growth, protein production, and metabolic activity. By restraining glycolytic overactivation, DHB may help T cells avoid exhaustion and maintain a more resilient, stem-like identity.

The researchers faced a significant pharmacological challenge when attempting to use DHB as an oral treatment. In its natural form, the metabolite is rapidly degraded after ingestion and may be removed from the body within minutes, limiting the amount that reaches the circulation. To address this problem, the team chemically converted DHB into a more stable prodrug. This modified form was then incorporated into an oleic-acid-based oral nano-emulsion known as Prodrug 201. The formulation was designed to protect the active compound during gastrointestinal processing and improve its absorption into the bloodstream.

Experiments showed that the nano-emulsion substantially increased the compound’s exposure in the body. The researchers reported a 14.3-fold improvement in bioavailability compared with unformulated DHB. In practical terms, this means that a greater proportion of the administered material reached the circulation and remained available to influence immune cells. Oral delivery could also offer advantages over repeated injections, potentially making microbiome-based immunomodulation easier to administer and more compatible with combination treatment strategies.

The team tested the formulation in animal models of colorectal cancer, melanoma, and breast cancer. In these models, treatment promoted the accumulation of stem-like, tumor-reactive T cells at tumor sites and was associated with marked tumor regression. The strongest effects were observed when the oral nanomedicine was combined with immune checkpoint blockade. In the reported test models, the combination eliminated tumors and generated immune memory capable of protecting against subsequent tumor recurrence. These findings suggest that DHB may help overcome one of the limitations of checkpoint therapy: the need for a sufficiently large and functionally competent pool of T cells that can be reinvigorated once inhibitory signals are removed.

The results also highlight the broader therapeutic potential of metabolites produced by the microbiome. Gut bacteria are known to influence immune development and cancer treatment responses, but translating these biological associations into medicines has been difficult. Live bacterial therapies can be complex to manufacture, standardize, and administer, while many natural microbial metabolites have short half-lives or poor absorption. By stabilizing a defined metabolite and packaging it in an oral delivery system, the researchers sought to transform an otherwise fleeting microbial signal into a reproducible pharmacological intervention. The approach could eventually be adapted to other microbiome-derived molecules with immunological activity.

The study remains preclinical, and its results in animal models do not establish whether the treatment will be safe or effective in people. Human tumors, immune systems, microbiomes, and drug responses can differ substantially from those observed in laboratory models. Future studies will need to determine the formulation’s toxicity profile, optimal dosing, interactions with existing immunotherapies, and effects across different cancer types and patient populations. Nevertheless, the work provides a technically distinct route toward cancer immunotherapy: rather than directly activating T cells through a conventional immune stimulant, it modifies the metabolic conditions that help them survive and retain their antitumor potential. If validated in clinical trials, the oral DHB nano-emulsion could represent a new class of microbiome-inspired medicines aimed at making immune checkpoint therapy more durable and broadly effective.

Subject of Research: Gut microbial metabolite DHB, CD8+ T-cell stemness, cancer immunotherapy, and oral nanomedicine.

Article Title: Oral nano-delivery of a gut microbial metabolite enhances T cell stemness for cancer immunotherapy.

Web References: https://doi.org/10.1038/s41565-026-02235-9

References: Han, K., Cho, Y. S., Takahashi, M. et al. “Oral nano-delivery of a gut microbial metabolite enhances T cell stemness for cancer immunotherapy.” Nature Nanotechnology. DOI: 10.1038/s41565-026-02235-9.

Image Credits: Han, K., Cho, Y. S., Takahashi, M. et al.

Keywords: DHB; 3,4-dihydroxybenzoic acid; gut microbiome; CD8+ T cells; T-cell stemness; cancer immunotherapy; immune checkpoint inhibitors; nanomedicine; oral drug delivery; Akt-mTORC1-Myc pathway; glycolysis; tumor immunity.