All disease begins in the gut

The gut microbiome is an active ecosystem that produces many metabolic by-products with immunomodulatory functions. Dysbiosis disrupts or alters the metabolic output from gut commensals, leading to reduced T-cell effector function and permitting immunosuppressive cells to migrate from the gut to distant tumor sites (figure 1). Short-chain fatty acids (SCFAs) and bile acid (BA) derivatives are important metabolic compounds that directly affect immune responses and lymphocyte migration. SCFAs like butyrate, propionate, and acetate are fermentation products from dietary fiber that promote T-cell activation and effector function. For example, butyrate is produced by many commensal species, such as Faecalibacterium prausnitzii and Clostridium butyricum, and can remodel T cell and dendritic cell activity through epigenetic mechanisms via histone deacetylase inhibition.8 Indeed, patients with melanoma with a higher relative abundance of F. prausnitzii have significantly longer progression-free survival following anti-programmed cell death protein 1 (PD-1) therapy.1 F. prausnitzii has also been associated with improved CAR-T cell therapy responses, as has Bifidobacterium longum—another SCFA-producing gut commensal.3 4 Butyrate’s immunological influences also extend to regulatory T-cell induction, mucosal integrity, and blockade of proinflammatory cytokines,8 thus giving it a possible role in preventing or reducing the rate of irAEs following treatment.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Dysbiosis versus eubiosis and their impact on immune responses during cancer therapy. The figure compares the gut mucosal environment and its effect on tumor immunity under conditions of dysbiosis (left) and eubiosis (right). In dysbiosis, low microbial richness and diversity, altered bile acid (BA) metabolism, and the resulting immunosuppressive microenvironment facilitate the trafficking of RORγt+regulatory Th17 (Tr17) cells from the gut to distant tumor sites. This migration contributes to reduced local antitumor immunity and diminished immunotherapy efficacy. Conversely, in a healthy eubiotic state, a microbiome characterized by high richness and diversity supports the production of short-chain fatty acids (SCFAs), improved T-cell activation, and maintenance of gut-imprinting addressins (eg, mucosal addressin cell adhesion molecule 1 [MAdCAM-1]). These conditions collectively promote potent antitumor immune responses and enhance immunotherapy success. Symbols are provided to indicate key cell types, including activated T-cells, macrophages, dendritic cells, and cancer cells, as well as representative bacterial species (eg, Enterocloster spp). Figure created in BioRender.com.

BAs are synthesized in the liver and can be metabolized into secondary BAs by gut bacteria, which influence systemic immunity by regulating the expression of adhesion molecules and affecting T-cell migration. Our laboratory recently found mucosal addressin cell adhesion molecule 1 (MAdCAM-1), an addressin that is highly expressed on the luminal side of high endothelial venules in the gastrointestinal tract, governs the retention of α4β7+RORγt+regulatory Th17 (Tr17) cells within the gut mucosa.9 Importantly, the blooming of Enterocloster spp following antibiotic cessation directly led to disruptions in BA metabolism—specifically lithocholic acid and ursodeoxycholic acid—which reduced enteric MAdCAM-1 levels and soluble MAdCAM-1 (sMAdCAM-1) in the serum. By tracking immune cells in the mesenteric lymph node with carboxyfluorescein succinimidyl ester labeling and transgenic photoconvertible Kaede mouse models, our laboratory demonstrated that the downregulation of MAdCAM-1 allowed the exodus of immunosuppressive Tr17 cells out of the gut and permitted their accumulation in distant tumor sites, where they foster an immunosuppressive microenvironment.9 In both mice and patients with non-small cell lung cancer, increased circulation of Tr17 cells and decreased concentrations of sMAdCAM-1 in sera correlated with dampened local antitumor responses and diminished effectiveness of anti-PD-1 therapies.

SCFAs and BAs are well-known examples of metabolites that shape immune function, yet they are not the only ones. Others, such as inosine, L-arginine, or tryptophan derivatives, can boost effector T-cell functions.5 The next important research challenge will be to unravel the effector functions of other microbiota-derived metabolites.

Restoring balance

Dysbiosis is a treatable condition, and MCIs are being investigated to treat or ameliorate this imbalance and boost immunotherapy success. Examples of MCIs include concomitant changes in diet with therapy, probiotics and prebiotics to encourage growth or actively incorporate select commensals, metabolites known to have immunomodulatory effects (eg, SCFAs), and fecal microbiota transplantation (FMT).

Dietary changes may be the most straightforward and less invasive MCI, as increased fiber intake, reduced consumption of processed sugars and fats, and moderate alcohol use can encourage the outgrowth of SCFA-producing bacteria.10 While dietary interventions are appealing due to their simplicity and non-invasive nature, they rely heavily on patient compliance and may require highly individualized nutritional plans. Transforming nutrition into a science will be challenging, as will determining optimal dietary patterns and frequencies for improving immunotherapy outcomes.

Prebiotics, such as dietary fibers, 3-hydroxybutyrate, or castalagin (a polyphenol found in the Camu-camu berry (Myrciaria dubia)), are “food supplements” that selectively feed beneficial bacteria and have shown potential in improving the efficacy of immunotherapy in preclinical models.10 Live biotherapeutic bacterial strains such as C. butyricum MIYAIRI 588 (CBM588), commonly used to treat gastrointestinal disorders in Japan and China, appeared to reduce the deleterious effects of antibiotics and proton pump inhibitors in retrospective studies of patients with ICI-treated cancer.11 CBM588 has additionally shown effectiveness in boosting the immunostimulatory effects of anti-cytotoxic T lymphocyte-associated protein 4 or tyrosine kinase inhibitors combined with anti-PD-1 antibodies in randomized trials involving patients with first-line metastatic kidney cancer.12 13

The most direct intervention is FMT, the transfer of stool (and thus, the entire gut ecology) from a healthy donor or treatment-responsive patient to a recipient’s gastrointestinal tract. Several proof-of-concept clinical studies have demonstrated that FMT can significantly improve responses to anti-PD-1 therapy in patients with advanced melanoma14–16 and in a randomized trial in patients with metastatic kidney cancer.17 Although these early short-term results are encouraging, challenges remain in standardizing donor selection, ensuring product safety, and maintaining consistent efficacy in the long term. Moreover, individual patient factors and cancer characteristics may influence FMT outcomes, necessitating more sophisticated stratification tools to identify which patients are most likely to benefit from this MCI.

Conclusion and future directions

Investigating methods to leverage the gut microbiome to improve immunotherapy success is an exciting area of research. However, alongside those previously mentioned, several challenges need to be addressed. Critically, we still lack a standardized definition for dysbiosis. The current broad definition of a ‘misbalanced gut microbiome’ is very likely an oversimplification and evokes the idea of ‘misbalanced humors’ used in antiquity. Is dysbiosis defined strictly by diversity metrics, or should we focus on specific taxa or functions? Which biomarkers are the most predictive and reliable, and can these be harmonized into a single diagnostic test? Should dysbiosis be considered a transition between health and disease? We know that at least two types of dysbiosis exist wholly based on the prevalence and relative abundance of Akkermansia muciniphila: its absence or overabundance is associated with significantly worse clinical outcomes.18

Given the complex relationships between gut bacteria consortia within the holo-ecosystem, metagenomics data could be further analyzed using network interactions evolving in a seesaw manner. The TOPOSCORE (‘topological score’) is one such approach developed by our laboratory as a personalized metric of enteric health that could predict treatment response across various malignancies.18 In parallel to determining dysbiosis from fecal materials, identifying robust plasma biomarkers that represent proxies for gut inflammation will have great potential in oncological practice. As described above, serum sMAdCAM-1 levels may be one of many potential key surrogates for intestinal dysbiosis that could be used to assess the need for MCIs and as a pharmacodynamic tool to monitor their efficacy.9 Soon, machine learning may identify functional interactions between metagenomics species and metabolomics features that will pave the way to biomarker discovery.

Bacterial metabolites are not the only means by which the gut microbiome influences systemic immunity. For example, it is also a vast reservoir of antigens that can potentially prime lymphocytes that are capable of cross-reactivity with bacterial and cancer-derived antigens. Pre-clinical studies have demonstrated cross-reactivity between bacterial species and murine tumor models,19 20 and recent clinical studies have extended this finding to glioblastoma21 and bladder cancer.22 However, the clinical relevance of this cross-reactivity remains unclear.23 Additionally, emerging research indicates non-mucosal tumors harbor intrinsic microbiomes, challenging previous assumptions that cancers are sterile.24 Moreover, they may impact anti-tumor immune responses, as immunogenic bacterial peptides have been identified on melanoma cells presented via human leukocyte antigen molecules.25 However, the bacterial biomass in non-mucosal tumors is very low (approximately 1–10 bacteria per 10⁵ tumor cells).26 Consequently, the tumor microbiome’s significance may be far greater at mucosal sites where microbes naturally reside, whereas its presence in non-mucosal tissues could reflect systemic or local immunosuppression. Thus, the clinical impact of these sparse ‘passenger microbes’ remains an important area for further investigation.

As our understanding of the relationships within the holobiont deepens, MCIs can be better utilized to enhance the efficacy of immunotherapies by preserving or restoring a metabolically supportive microbiome. We are actively bridging the gap between complex microbial data and clinical decision-making and stand poised to open a new era of precision medicine in oncology—one in which the gut microbiome becomes as integral to treatment planning as tumor staging and molecular profiling.