Microbial-derived metabolism is intimately connected with our own dietary intake. For example, riboflavin (vitamin B2) —found in dairy, meats, cereals and grains, nuts, and green vegetables (37, 38) —is metabolized by the GMB to produce derivatives that can promote expansion of mucosal-associated invariant T cells, which are seen to be activated to a greater extent in responders to PD-1 ICB (39). However, while several microbial metabolites have been linked to improvements in immunity and immunotherapy response, the same metabolites are often also implicated in cancer formation. For instance, microbial metabolism of primary bile acids from the consumption of animal fat/proteins produces secondary bile acids (SBAs), which have been implicated in hepatocellular carcinoma carcinogenesis. Despite this, the SBA ursodeoxycholic acid has been found to reduce cancer immunosuppression by inducing TGF-β degradation (40).

The role of microbial tryptophan metabolism is also controversial, with some studies suggesting indoles can act as aryl hydrocarbon receptor (AhR) agonists, promoting immune evasion in the tumor microenvironment (TME), and others showing these metabolites can promote anti-tumor immunity (41–44). One study observed that selective antibiotic targeting of Lactobacilli, which are indole-producers, reduced pancreatic tumor growth and increased T cell infiltration. Conversely, provision of L. reuteri resulted in increased tumor growth and reduction in intratumoral CD8⁺ T cells (45). On the other hand, other studies have found that microbial tryptophan metabolites can promote chemotherapy and ICB efficacy (43, 44). Recently, Bender et al. demonstrated that L. reuteri was able to translocate from the intestine to the tumor bed and produce indole-3-aldehyde in situ which promoted antitumor immunity and ICB response in an AhR-dependent, Tc1-cell-inducing manner (43, 44). Notably, cruciferous vegetables like broccoli, cabbage, and brussels sprouts possess dietary indole compound precursors that can activate AhR yet demonstrate reduced cancer risk in a potentially dose- and context-dependent manner (46–51).

However, the GMB metabolites that have garnered the most attention as potential mediators of ICB response are short chain fatty acids (SCFAs), which will be discussed in detail below. Many of the bacteria associated with ICB efficacy and response across studies are involved in fiber fermentation and the production of SFCAs and/or support other fiber-fermenting bacteria (2, 3, 6, 7, 52, 53). Ultimately, further investigation into the function and metabolic output of the GMB rather than its taxonomical constituents may overcome some of the lack of concordance observed across studies of microbiome and ICB response given functional redundancy of the GMB (6, 7).

A common feature shared by plant-based diets is the possession of high amounts of dietary fiber, which is defined as nondigestible carbohydrates and lignin that are intrinsic and intact in plants (80–82). In preclinical models, fiber deprivation lowers microbial diversity (83) and decreases the abundance of beneficial fiber-fermenting bacteria such as F. prausnitzii (bacteria associated with response to ICB). Insufficient fiber can also cause some bacterial taxa to begin degrading host mucin glycans, thus thinning the mucus layer; promoting inflammation, bacterial translocation, and oncogenic signaling pathways; and inducing greater tumor development as a result (83, 84).

SCFAs, i.e., butyrate, acetate, and propionate, are the product of dietary fiber fermentation by gut microbiota and a major mediator of the effects of plant-based diets. SCFAs serve as the main energy source for intestinal epithelial cells (IECs) (85), ensuring mucosal barrier integrity and preventing bacterial translocation (86–89), and also have both indirect and direct effects on immunity. In vitro administration of SCFAs to IECs activates G protein coupled receptors (GPCRs), increases the production of anti-inflammatory cytokines like IL-10 (85), and can also abrogate neutrophil recruitment, regulating the production of inflammatory mediators like TNF-a and IL-17 and thus playing an important role in recruiting leukocytes and activating T lymphocytes (88). SCFAs can also induce plasma cell production of IgA, blocking virulence and bacterial adherence to epithelial cells (90, 91). Perhaps most relevant for ICB response is that SCFAs can also enhance T cell activity. For example, acetate has been shown to enhance IFN-γ production from CD8+ and intratumoral cytotoxic T cells (CTLs) via the mammalian target of rapamycin pathway (92) as well as memory CD8+ T cell responses via GAPDH enzyme activity and glycolysis (93). Meanwhile, the role of the SCFA butyrate in suppressing carcinogenic pathways, such as the Wnt signaling pathway, has been well-studied in colonocytes (94, 95). More recent studies have demonstrated that butyrate can also impact immune cell function via histone deacetylation inhibition (88), which upregulates IFN-γ and granzyme B expression while suppressing IL-17A production in CTLs, thus facilitating an increase in the switch of Tc17 cells to a CTL phenotype (92).

However, although multiple studies have demonstrated the potential benefits of fiber and SCFAs for cancer immunity, there is also evidence for their negative impact. For instance, one study found that incorporating the fiber inulin into mice’s diets induced hepatocellular carcinoma via a microbiota-dependent mechanism (96). Another experiment detected an increase in mouse colonic tumor load when the amount of fiber—consisting of either inulin, cellulose, or brewers spent grain—was increased in their diets (97). Other studies, both human and preclinical, have also found an association between fiber/SCFAs and aggravation of factors promoting cancer risk/development, such as greater production of cancer-associated cytokines (88, 98–101). These seemingly discrepant results may be driven by the specific sources of dietary fiber incorporated, types of fiber, and differences in study populations, including cancer type (102). Isolated fiber may also have different effects than whole-food-derived fiber, i.e., it is possible that the combination of fiber and the micronutrients found in whole-food/natural dietary fiber sources is required in order for patients to reap the potential cancer-inhibiting benefits of a high-fiber diet. Furthermore, specific types of fiber may elicit differential biological effects (97, 103, 104), with one murine study observing a significantly lower tumor load in inulin-fed mice than in mice fed cellulose or brewers spent grain (97). It thus remains to be seen how we may optimize fiber-related dietary interventions targeted at improving cancer immunity in a microbiota-dependent manner. However, given the enormous amount of data supporting the health benefits of a whole-food, high-fiber diet and the context-dependent factors that may have led to the negative effects observed in other studies, it is reasonable to recommend increasing fiber consumption through natural sources while continuing to research its impact on cancer immunity.

Given the existing literature on the immunomodulatory role of SCFAs and the known role in fiber fermentation of many pro-ICB-response bacteria, there was a strong rationale to examine the potential link between dietary fiber consumption and immunotherapy response. Indeed, a 2021 study by Spencer et al. incorporated dietary screener questionnaires into their observational microbiome cohort and found that habitual higher dietary fiber consumption was associated with significantly improved progression-free survival in melanoma patients undergoing ICB (105). Similarly, a 2022 multicenter cohort study observed that a Mediterranean diet pattern—which features high intake of plant-based fiber alongside omega-3 fatty acids and various micronutrients—was positively associated with ICB response, progression-free survival, and less immunotoxicity among patients with advanced melanoma (106). As expected, dietary fiber intake is positively correlated with pro-ICB-response, fiber-fermenting bacteria (58). These findings were recapitulated in murine melanoma models, where high-fiber diet fed mice displayed improved treatment response to anti-PD-1 ICB—with increased CD4+ T cell infiltration in tumors, and greater expression of genes linked to T-cell activation and interferon response in CD45+ TILs —compared to anti-PD1 low-fiber fed mice (105).

Rapid shifts in the GMB and lower stool concentrations of the SCFA propionate are also observed with fiber deprivation in preclinical models. Importantly, synergy between ICB and dietary fiber were not seen in germ-free mice, supporting that the effects were indeed microbiome-mediated. In another preclinical study, a high-fiber diet modulated the GMB in a manner that triggered this intratumoral IFN-1/natural killer cell (NKC)/DC axis, improving ICB efficacy and programming TME mononuclear phagocytes toward immunostimulatory phenotypes (107). Given these observational and preclinical data supporting the role of fiber in modulating the microbiome and immunity, human interventional studies (NCT03950635 and NCT04645680) are now ongoing that investigate high-fiber diet as a GMB-focused strategy to improve anti-cancer immunity. Be that as it may, future microbiome-centered dietary interventions will likely require tailoring to individual patients. Recent findings suggest that GMB compositions associated with high fiber intake—namely, high-diversity GMBs with high relative Ruminococcaceae, butryogen, and methanogen abundance—are associated with ICB response, but that modulating taxa composition alone may not be enough to improve outcome. For instance, having higher relative abundance of merely one putatively beneficial taxon, such as F. prausnitzii, may be insufficient to induce response, and was in fact observed to pose no further advantage for ICB outcome in patients already possessing high-fiber-associated GMBs prior to ICB (58). Employing high-fiber diets in the context of immunotherapy may thus be an intervention better suited for patients with a GMB composition shaped by habitual low-fiber intake [i.e., low-diversity, Bacteroides-dominated microbiomes (58)], but even then, tumor-intrinsic, systemic, and other factors also need to be considered.

Probiotics are live microorganisms with putative health benefits when consumed. They are found in fermented foods such as kimchi, kefir, sauerkraut, etc., which are “produced through controlled microbial growth … and the conversion of food components through enzymatic action” (108). As such, probiotic foods contain both live microorganisms and food sources for said microorganisms, such as non-digestible carbohydrates like inulin-type fructans and galacto-oligosaccharides (109, 110). This allows probiotic foods to support a beneficial GMB composition by encouraging the growth of bacteria like F. prausnitzii while pulling resources from and thus suppressing the growth of pathogenic bacteria.

These naturally occurring dietary probiotics, however, stand in contrast to commercial probiotic supplements, which are generally not targeted/rationally designed, but instead frequently composed of only aerobic, easy-to-culture bacteria in order to minimize production costs (111, 112). Commercial probiotics also lack the beneficial nutrients inherent to probiotic foods, such as essential amino acids, polyphenols, different types of fibers, and other fermentable bioactive compounds (113, 114). In fact, over-the-counter probiotic supplements have been shown to lower overall GMB diversity and worsen ICB response in preclinical models and were also associated with worse outcomes with ICB in an observational study (105). The negative impacts of unselected probiotics on GMB homeostasis has also been demonstrated after antibiotic use, wherein probiotics inhibited the restoration of normal gut ecology after antibiotics (105, 111). Thus, the use of probiotics outside the context of a clinical trial should be discouraged at this time.

Conversely, observational data suggests a beneficial impact of dietary probiotics on the GMB and immunity. Natural yogurt consumers have been found to display healthier metabolic profiles, with lower levels of inflammation and serum lipid peroxidation and higher fecal concentrations of Akkermansia (115, 116). In a preclinical study by Woo et al., mice with DSS-induced colitis were fed with a fermented barley and soybean mixture which led to increased intestinal barrier function and decreased levels of inflammatory cytokines in colonic tissue. The fermented mixture was also found to suppress bacterial translocation to mLNs and promote growth of the anti-inflammatory bacterial genera Lactobacilli and Bacteroides, the latter of which has been demonstrated to be an important contributor to anti-CTLA-4 ICB efficacy (9, 117).

A recent human interventional study substantiated these health-promoting effects of dietary probiotics. Wastyk et al. conducted a high-fermented-food diet intervention in healthy individuals targeting consumption of 6 servings of fermented foods daily. Post-intervention, participants exhibited increased GMB diversity; importantly, beneficial taxa that were enriched were in large part not derived from the microbes found in the fermented foods consumed, suggesting that dietary probiotics can influence the GMB composition both directly and indirectly and seem to support the overall ecosystem. They further examined the effects of this diet on circulating immunocytes, finding increased effector memory CD4+ T cells with decreased non-classical monocytes and circulating inflammatory cytokines in the post-intervention specimens (21).

Ketogenic Diets (KD), characterized by low carbohydrate and high fat content, aim to reduce carbohydrate intake to the point of ketosis, where ketone bodies are produced and metabolized as an energy source rather than glucose, the usual main source of energy. The premise of using the ketogenic diet in cancer is based on the “Warburg effect”, an observation that cancer cells rely primarily on aerobic glycolysis rather than mitochondrial oxidative phosphorylation (118). However, it has subsequently been shown that cancer cells are much more metabolically plastic than originally thought (119). Nonetheless, the ketogenic diet has been shown to limit cancer growth in many preclinical models via lowering of insulin and IGF-1, and intriguingly has shown synergy in combination with cancer therapies that disrupt the insulin/IGF-1/PI3K axis in host organs (120).

Recent studies have examined the impact of the KD on anti-tumor immunity and on the GMB. In murine models, Ferrere et al. found that KD inhibited tumor growth and cancer progression and promoted ICB efficacy, at least in part via the ketone body 3-hydroxybutyrate (3HB) in a T-cell dependent manner (121). Notably, systemic administration of 3HB mimicked the antitumor effects of KD, enhancing ICB efficacy and promoting CXCR3+ T cell expansion. In terms of the effects on the microbiome, KD led to an increased abundance of A. muciniphila and other immunogenic bacteria. The tumor-suppressive effects of KD and 3HB were attenuated by antibiotics, indicating that the microbiome may mediate the KD’s anti-cancer effects (121). Other murine studies have similarly found that KD can increase potentially beneficial bacteria like A. muciniphila and Lactobacillus while decreasing putative inflammatory taxa, such as Desulfovibrio and Turicibacter (122).

However, implementation of KD in humans, especially cancer patients, requires close supervision by registered dietitians, as hypoglycemia, weight loss, and electrolyte imbalances may occur (123). Further, KD limits fruit, vegetables, whole grains, and fiber ingestion, and thus the effects of this diet on the human microbiome need to be carefully considered. Even so, clinical data has shown some promise. In a human diet intervention study by Link et al, participants sequentially consumed KD and vegan diets for 2 weeks each and in randomized order. Both KD and vegan diet induced differential changes on host immunity and GMB composition: the KD increased activated T_reg and CD16⁺ NKC abundance as well as T cell activation; meanwhile, the vegan diet increased activated T helper and NKCs and upregulated innate-immunity-associated pathways (124). Further investigation is needed to elucidate the potential implications for anti-cancer immunity.

Calorie restriction (CR) is defined as consuming 50-90% of usual ad libitum calorie intake without going to the extent of malnutrition. On the other hand, a fasting-mimicking-diet (FMD) reduces calorie intake to a lesser extent and is specifically low in protein and sugar yet high in unsaturated fat content (125–129). Multiple preclinical studies have demonstrated that CR and/or fasting-like conditions reduce cancer risk, incidence, progression, and mortality, though data in humans remains limited (107, 130–133). Like the KD, a major target of CR and FMD is insulin and IGF-1 levels. High serum IGF-1 has been associated with increased risk of multiple cancers (134). In rodents, long-term CR reduces serum IGF-1 concentration by 30-40%, and fasting in humans markedly decreases serum IGF-1 concentration provided that protein intake is not excessively high (135). In terms of immune effects, preclinical studies have found that CR alters NKC maturation and function; meanwhile, fasting-like conditions have been shown in pre-clinical models to increase cytotoxic CD8 T cell frequency and reduce T_reg tumors in the TME while also impacting NKC maturation (107, 136–138).

FMD has been shown to enhance chemotherapy efficacy in murine breast cancer models by promoting T-cell-mediated tumor immunogenicity and cytotoxicity (107, 136–138). More specifically, FMD administration induced greater cytotoxic CD8+ T cell recruitment, greater levels of common lymphoid progenitor cells and circulating/tumor-infiltrating CD8+ lymphocytes, and decreased activation and frequency of T_reg cells. These effects were mediated at least in part by a decrease in expression of the heme oxygenase-1 (HO-1) enzyme in cancer cells. This downregulation of HO-1 expression served to selectively sensitize cancer cells to chemotherapy while bypassing normal cells, whose HO-1 expression levels were not impacted (137, 139, 140).

Similarly, preclinical studies have shown that FMD elevates IFN-γ levels, increases CTLs and NKCs, promotes switching of CD8+ T cells to an activated/memory phenotype, and decreases immunosuppressive myeloid-derived suppressor cells (MDSCs) and Tregs (107). A periodic FMD has also been shown sensitize cancer cells to ICB, increasing anti-PD-L1 and anti-OX40 ICB efficacy against triple-negative breast tumors. In these murine subjects, FMD cycles were found to reactivate and expand early exhausted effector T cells and induce cancer cells to switch from respiratory to glycolytic metabolism (141). Meanwhile, human studies found that fasted cancer patients who adhered to a 5-day FMD (700 kcal consumed on day 1; 300 kcal consumed days 2-5) every 21-28 days displayed greater amounts of activated T cells and cytolytic NKCs in peripheral blood (142). In another study of heterogeneously treated (not ICB) cancer patients, FMD augmented intratumoral Th1/cytotoxic responses and reduced immunosuppressive MDSCs and Tregs. The study also noted greater amounts of CTLs and NKCs in patient blood; increased intratumoral populations of CD8+ T cells, DCs, NKCs, and effector memory T lymphocytes; a heightened switch of CD8+ cells toward a memory/activated phenotype; and increased levels of T cell stimulating cytokines (107). Another clinical trial observed that administration of hydroxycitrate, a CR mimetic, lead to depletion of Tregs and thus heightened anticancer immunosurveillance and decreased tumor burden (143).

In terms of the GMB, murine studies on CR/FMD have observed an increase in populations of putative protective/beneficial bacteria, such as SCFA-producing Bifidobacterium and Faecalibaculum as well as immunogenic A. muciniphila (144–147). In the same vein, preclinical models have noted a decrease in reputed inflammatory taxa, such as Helicobacter, Streptococcaceae, and Desulfovibrionaceae (144, 145, 148). Therefore, FMD and CR lead to marked systemic changes in the GMB, metabolism, and immunity in ways that substantiate the further investigation of FMD/CR. However, these types of diet interventions in patients with cancer should only be utilized under physician supervision and in the context of a clinical trial.