Gut microbiota–derived metabolites in immunomodulation and gastrointestinal cancer immunotherapy

Gut microbiota derived metabolites constitute a diverse group of low molecular weight compounds, primarily including SCFAs, bile acid derivatives, and tryptophan derivatives, as well as emerging molecules such as inosine, trimethylamine N oxide (TMAO), and urolithin A (UroA). Acting as key mediators of metabolic reprogramming, epigenetic regulation, and immune cell modulation, these metabolites collectively shape the tumor microenvironment and influence responses to immunotherapy. Table 1 summarizes the major classes, microbial sources, metabolic pathways, and immunoregulatory functions of gut microbiota derived metabolites in GI cancers.

The metabolic activity of the gut microbiota extends beyond the intestinal milieu and influences systemic immunity through two complementary routes: a chemical route and a cellular route. In the chemical route, small microbial metabolites are absorbed through the intestinal epithelium into the portal circulation (80), processed in the liver, and subsequently spill over into the systemic bloodstream (81, 82). In parallel, a portion of these molecules is carried by afferent lymph to the draining mesenteric lymph nodes (MLNs) (83, 84). This dual vascular–lymphatic conduit enables microbial metabolic signals to reach peripheral immune sites such as the liver, spleen, and MLNs. During development, maternally derived microbiota-dependent metabolites can traverse the placenta and modulate fetal immune organogenesis, thereby establishing a metabolic foundation for thymic maturation and immune tolerance (85, 86).

In the cellular route, dendritic cells and T cells conditioned by microbial metabolites egress via lymphatic vessels and home to the MLNs under the guidance of the C-C chemokine receptor 7 (CCR7)–C-C motif chemokine ligands 19 and 21 (CCL19/CCL21) axis (87, 88). Within these lymphoid niches, they prime naïve T cells and imprint tissue-homing programs (89). The resulting gut-conditioned effector and regulatory T cells subsequently disseminate through the circulation to peripheral organs, translating local metabolic cues into systemic immune effects (90). In adults, there is currently no conclusive evidence that gut-derived metabolites can cross the blood–thymus barrier and directly act on thymic tissue. This is likely related to the selective permeability of the cortical barrier region, which restricts molecular access under steady-state conditions (91). Instead, a more widely recognized mechanism involves cell-mediated communication, whereby dendritic cells acquire microbial or peripheral antigens from the circulation and deliver them to the thymus, enabling intestinal signals to indirectly influence thymic selection and the maintenance of immune homeostasis (92, 93).

Collectively, the diffusion of metabolites and trafficking of immune cells form a bidirectional communication axis through which microbial metabolic activity in the gut becomes systemically perceptible, fostering dynamic crosstalk and coordinated regulation between the intestinal and peripheral immune systems.

In addition to SCFAs, bile acids, and tryptophan derivatives, recent metabolomic and functional studies have identified a variety of emerging microbiota-derived metabolites (114). These molecules not only participate in maintaining host metabolic homeostasis but also modulate immune cell function and reshape the tumor microenvironment, thereby exerting significant and complex effects on GI tumorigenesis and responses to immunotherapy (11).

The underlying mechanisms by which these metabolites exert their systemic and immunological effects involve both chemical and cellular routes. One class of molecules such as inosine, TMAO and UroA can be detected in peripheral circulation (139–141), indicating their systemic exposure via intestinal absorption, portal venous transport, hepatic processing, and recirculation through the bloodstream and lymphatics (142). These metabolites have been reported to modulate immune microenvironments across organs and to influence antitumor immune responses (143). Other emerging metabolites primarily condition intestinal immune cells locally; these gut-imprinted cells then migrate through MLNs to disseminate their effects systemically via a cellular route (83, 144). In addition, vitamin-like or small molecules that are highly sensitive to dietary or therapeutic interventions show markedly context-dependent levels of peripheral exposure and immunological outcomes (145–147).

ICIs, particularly those targeting the PD-1/PD-L1 axis, have profoundly reshaped the therapeutic landscape of multiple cancers (153). However, their overall response rate remains limited, with challenges including resistance and immune-related adverse events. Increasing evidence suggests that gut microbiota diversity and community composition influence ICI efficacy, clinical studies and reviews consistently report an association between higher microbial diversity and improved clinical outcomes (11).In HCC, cohort findings remain inconsistent, responders have been associated with enrichment of Lachnoclostridium, Lachnospiraceae, Veillonella, and higher fecal UDCA, while retrospective studies also link Akkermansia, Ruminococcaceae, and Bifidobacterium to favorable prognosis. These heterogeneous results require validation in larger, standardized cohorts (154).

Akkermansia muciniphila (Akk), regarded as a “next-generation probiotic,” can significantly enhance the efficacy of PD-1/PD-L1 blockade by reshaping the immune ecosystem and boosting CD8⁺ T-cell function. Key mediators include the outer membrane protein Amuc_1100 and the protease Amuc_1434, which also regulate immune responses through modulation of SCFA and bile acid metabolism (155). In HCC cohorts, its abundance correlates with elevated UDCA/TUDCA levels, suggesting that Akk may improve ICI efficacy through bile acid metabolism. Beyond functional bacterial strains, dietary components can also shape the microbiota and metabolic profiles. For instance, Chinese yam polysaccharide (CYP) has been shown to synergize with PD-1 therapy in CRC mice by improving microbial diversity and reducing deoxyguanosine (dG) levels, thus preventing M2 polarization and reversing immunosuppression (156).

Building on microbial diversity, microbiota-derived metabolites are regarded as functional mediators that directly account for differences in ICI efficacy. Distinct classes of metabolites can exert either positive or negative effects on immunotherapy through signaling pathways, metabolic reprogramming, and immune cell regulation.

As key metabolites of the gut microbiota, SCFAs have attracted increasing attention for their immunosensitizing effects in ICI therapy. Multiple preclinical studies suggest that SCFAs enhance antigen presentation and promote CD8⁺ T-cell effector functions, thereby augmenting the antitumor efficacy of PD-1/PD-L1 blockade (157, 158).

As key metabolites of the gut microbiota, SCFAs have recently gained significant attention for their immunosensitizing effects in ICI therapy. Multiple preclinical studies suggest that SCFAs can enhance antigen presentation and promote CD8⁺ T-cell effector functions, thereby augmenting the antitumor efficacy of PD-1/PD-L1 blockade. Among them, butyrate is particularly representative, as it enhances CD8⁺ T-cell activity through epigenetic mechanisms such as HDACs inhibition, generating a synergistic effect with PD-1 blockade (159, 160). In gastric cancer models, microbiota-derived butyrate downregulates PD-L1 and IL-10 expression in TAMs, thereby alleviating immunosuppression and reinforcing antitumor responses (161). Evidence at the bacterial strain level further supports this conclusion. For example, Faecalibacterium prausnitzii (EXL01) restores immune function after antibiotic-induced damage and enhances PD-1/PD-L1 efficacy (162). Similarly, the butyrate-producing Clostridium butyricum strengthens antitumor responses to PD-1 therapy in colorectal cancer models by suppressing IL-6–mediated immunosuppression. Notably, the clinically promising CBM588 strain has been shown to promote CD8⁺ T-cell infiltration and functional activation, thereby prolonging survival. However, under inflammatory conditions, CBM588 may also exacerbate inflammation, highlighting the need for context-specific application (163). These findings suggest that butyrate and its producing bacteria act not only as metabolite-level immunosensitizers but also hold translational potential. However, their effects remain context-dependent: while enhancing PD-1 blockade, SCFAs may attenuate CTLA-4 inhibition by promoting Treg expansion, underscoring the heterogeneity of their immunoregulatory roles across therapeutic settings (155, 164).

Moreover, the ability of SCFAs to potentiate PD-1 blockade efficacy is highly dependent on molecular and immunological context. Studies have shown that butyrate and propionate can induce the DNA damage response (DDR), upregulate antigen-presenting molecules (MHC-I) and chemokines, thereby enhancing CD8⁺ T-cell infiltration and effector function, ultimately improving the antitumor efficacy of PD-1 blockade (160). In microsatellite instability–high (MSI-H) colorectal cancer (CRC), tumors already exhibit a high mutational burden and abundant neoantigens, resulting in relatively high response rates to PD-1 blockade. In this context, SCFAs mainly act to further enhance effector T-cell function and sustain immune-inflammatory activity. In contrast, microsatellite-stable (MSS) CRC is typically resistant to PD-1 therapy. However, preclinical studies demonstrate that propionate, in synergy with Bacteroides fragilis, can reshape the immunosuppressive tumor microenvironment and markedly enhance PD-1 efficacy. This offers a new therapeutic entry point for MSS CRC, which has traditionally been unresponsive to ICIs (165, 166). Both clinical and mechanistic evidence are steadily accumulating: in solid tumor patients treated with PD-1 inhibitors, higher SCFA levels in blood or feces are often associated with longer PFS. Functional studies further reveal that metabolites such as isobutyric acid and butyrate can enhance CD8⁺ T-cell cytotoxic activity by modulating TCR signaling, thereby providing immunosensitizing support for ICIs at the metabolite level (159, 167).

Bile acids and their metabolites regulate the immune microenvironment through receptors such as FXR and TGR5, and are closely associated with the efficacy of ICIs. Multiple studies have suggested that secondary bile acid profiles correlate with PD-1 responsiveness. In HCC cohorts, responders exhibited relatively higher fecal levels of UDCA and TUDCA, which were associated with PD-L1 downregulation and enhanced CD8⁺ T-cell function (168). In preclinical models, UDCA worked with anti–PD-1 therapy to augment CD8⁺ T-cell effector functions (169). Mechanisms included carboxy terminus of Hsc70-interacting protein (CHIP)–mediated TGF-β degradation, Treg suppression, and EMT downregulation through the TGR5–Yes-associated protein (YAP) axis, thereby relieving the immunosuppressive TME. Other studies have shown that Akkermansia muciniphila can increase TUDCA levels in parallel with tumor PD-L1 downregulation, suggesting that a “bile acid–PD-1/PD-L1–T-cell” axis may represent a novel target for immunotherapy (170). More recently, Jin et al. reported that specific secondary bile acids antagonize the androgen receptor (AR), activate CD8⁺ T cells, and enhance anti–PD-1 efficacy, providing new molecular evidence for the role of bile acids in sensitizing tumors to immunotherapy (171).

In GI cancers, the IDO/TDO–Kyn–AhR axis has been well established as a key driver of immunosuppression. By activating AhR, Kyn induces Treg differentiation, suppresses CD8⁺ T-cell function, and promotes immune checkpoint upregulation, thereby attenuating the efficacy of immunotherapy (172). In colorectal and hepatocellular carcinoma models, blockade of Kyn synthesis or transport has been shown to restore T-cell function and act synergistically with PD-1 blockade (173, 174).

Purine metabolism likewise plays a pivotal role in immune evasion in GI cancers. Among its pathways, the adenosine–A2A receptor axis has been extensively demonstrated to suppress effector T-cell and NK-cell activity while promoting the accumulation of immunosuppressive cell populations, constituting a major factor underlying immunotherapy resistance in colorectal and gastric cancers (175). Blocking adenosine signaling can remodel the tumor microenvironment and enhance the efficacy of PD-1/PD-L1 blockade (176).

Indole derivatives have been shown to activate the AhR–IL-22 pathway to improve mucosal barrier homeostasis, while also alleviating Kyn–AhR–mediated T-cell dysfunction in multiple models, thereby promoting responsiveness to ICIs. Further studies revealed that IPA upregulates the TCF7 super-enhancer, expands PD-1⁺/TCF1⁺ stem-like CD8⁺ T cells, and enhances the efficacy of anti–PD-1 therapy (177). IAld, via the AhR–cAMP response element-binding protein (CREB) axis, reinforces effector programming and sensitizes tumors to PD-(L)1 therapy (178). Meanwhile, IAA exhibits a typical dose- and context-dependent pattern. At moderate levels, IAA promotes mucin sulfation and maintains barrier and immune homeostasis through the AhR–3′-phosphoadenosine 5′-phosphosulfate synthase 2 (Papss2)–solute carrier family 35 member B3 (Slc35b3) pathway. However, at high concentrations or within specific tumor microenvironments, IAA may excessively activate AhR, skewing toward immunosuppressive networks and attenuating CD8⁺ T-cell effector function, ultimately diminishing the efficacy of PD-1/PD-L1 blockade (179). The functional direction of IAA also varies across GI cancers: in colorectal cancer, it predominantly exerts barrier-protective and antiproliferative effects, whereas in pancreatic cancer, it may instead promote immune evasion (48, 172).

Inosine, a purine metabolite, has recently attracted attention for its immunosensitizing effects. Studies indicate that inosine, acting as a ligand of the A2A receptor, activates the cAMP–PKA signaling pathway to enhance CD8⁺ T-cell metabolic adaptability and effector function (141). In GI tumor models, exogenous inosine supplementation synergizes with PD-1 or CTLA-4 blockade to improve antitumor immunity (150), suggesting its potential translational value in optimizing immunotherapy efficacy.

TMAO has been shown in pancreatic cancer models to drive TAMs toward an immune-activating phenotype and to enhance the effector functions of intratumoral IFNγ⁺TNFα⁺ T cells. When combined with ICB, TMAO significantly reduces tumor burden and prolongs survival, an effect dependent on type I interferon (IFN-I) signaling (67).

Exercise-induced microbiota-derived metabolite formate has recently been identified as a potential sensitizer for ICIs. Mechanistic studies demonstrate that formate activates CD8⁺ T-cell effector functions via the nuclear factor erythroid 2–related factor 2 (Nrf2) signaling pathway, enhancing IFN-γ and granzyme B secretion, thereby amplifying the antitumor immune response to PD-1/PD-L1 blockade. Fecal microbiota transplantation and human cohort analyses further indicate that a high formate-producing microbiota is positively correlated with improved ICI responsiveness (71), providing new therapeutic avenues for combining exercise or dietary interventions with ICIs.

Diet directly shapes the gut microbiota and its metabolites, influencing ICI efficacy through the “diet–microbiota–metabolite” axis. High-salt diets promote pro-inflammatory metabolites like skatole, driving immunosuppression, whereas indoles such as IPA and IAA improve the tumor immune microenvironment and T-cell function. Thus, dietary modulation of tryptophan metabolism can critically affect immunotherapy outcomes, underscoring the potential of diet as an adjunct strategy (180). Multiple studies have demonstrated that a high-fiber diet significantly elevates SCFA levels, thereby enhancing dendritic cell antigen presentation and effector T-cell function, ultimately improving responses to ICIs (181). High-fiber intake and Mediterranean diets are associated with improved ICI outcomes through enhanced microbial diversity and butyrate production, whereas antibiotics, PPIs, and corticosteroids reduce efficacy. These findings highlight diet–microbiota interactions as both predictive markers and modifiable strategies to optimize immunotherapy (181). The underlying mechanisms through which various dietary patterns enhance immune checkpoint inhibitor (ICI) efficacy are summarized in Table 2.

However, a recent study across five murine tumor models indicated that simply increasing dietary fiber does not consistently potentiate ICI efficacy; rather, baseline dietary composition and non-fiber components exert stronger effects on gut microbiota composition, metabolite profiles, and tumor immunity (182). Building on this, other dietary patterns are gaining attention. For instance, ketogenic diets and choline supplementation may enhance immune activation via TMAO metabolism (183, 184), while exercise and intermittent fasting promote the abundance of butyrate-producing bacteria and increase immune sensitivity (185, 186). Dietary fibers and polyphenols enhance mucosal barrier function by driving beneficial indole metabolites and can potentiate immunotherapy through the “gut microbiota–metabolite” axis. In preclinical models, pectin boosts anti–PD-1 efficacy via butyrate-producing bacteria and CD8⁺ T-cell infiltration, while inulin increases probiotic abundance, SCFAs, and immune metabolites to improve PD-1 blockade (187, 188).

Similarly, components such as Lycium barbarum polysaccharides, anthocyanins, Ganoderma lucidum polysaccharides, flavonoids, and Inonotus hispidus have been reported to exert comparable effects, mainly by facilitating probiotic colonization and enhancing SCFA or L-arginine metabolism to improve antitumor immunity (189). Ginseng polysaccharides (GPS), in particular, have been found to increase valerate levels and decrease kynurenine, thereby suppressing Treg activity and promoting effector T-cell differentiation. In CRC models, GPS enhances CD8⁺ T-cell function and suppresses MDSCs by promoting Lachnospiraceae colonization and SCFA production. Clinical sample analyses further confirmed that Lachnospiraceae abundance correlates positively with CD8⁺ T cells and negatively with MDSCs, underscoring its potential as an immunotherapy sensitizer in CRC (190). Plant-derived nanoparticles (PNPs) show promise as immunotherapy sensitizers. Ginger-derived exosome-like nanoparticles (GELNs) reshape microbial metabolism via miR-159a, enriching DHA and downregulating PD-L1 to enhance anti–PD-L1 efficacy, with clinical data linking higher DHA to lower PD-L1 expression (191). Sea cucumber polysaccharides (SCPs) synergize with PD-1 blockade in CRC models by enhancing CD8⁺ T-cell function, reducing Tregs, and remodeling the microbiota to elevate indole-3-carboxylic acid. Fecal microbiota transplantation (FMT) confirmed this microbiota-dependent effect, highlighting SCPs as promising dietary adjuvants in immunotherapy (192). Huaier polysaccharides (HP) show strong antitumor effects in HCC models by remodeling bile acid metabolism, notably reducing TCDCA. This shift suppresses M2 macrophage polarization, increases the M1/M2 ratio, and boosts pro-inflammatory cytokines, thereby reversing immunosuppression (193).

Diet shapes the “microbiota–metabolite–immunity” axis, where balanced nutrition enhances antitumor immunity, while unhealthy patterns foster pro-inflammatory metabolites and weaken immunotherapy efficacy.

The butyrate-producing probiotic Clostridium butyricum (CBM588) enhances antitumor immunity by expanding IL-17A⁺ T-cell subsets and modulating gut microbiota (194). Preclinical and early clinical studies show that CBM588 combined with PD-1/PD-L1 or PD-1/CTLA-4 therapy improves response rates and PFS, underscoring the clinical potential of functional probiotics (195).

In colorectal cancer liver metastasis models, Kang’ai Jinfang (KAJF) reshapes the immune microenvironment by enriching SCFA-producing bacteria and elevating butyrate, which improves barrier function and suppresses M2 macrophages. KAJF also reduces MDSCs and enhances CD8⁺ T-cell infiltration, thereby inhibiting metastatic growth (196). In addition to butyrate-producing bacteria, other probiotics such as Akkermansia muciniphila and Bifidobacterium have also been shown to enhance PD-1/PD-L1 blockade efficacy by reshaping metabolic profiles and optimizing the tumor immune microenvironment. For example, Bifidobacterium-derived inosine can augment CD8⁺ T-cell effector functions in the context of PD-1/CTLA-4 blockade (197). Further studies revealed that Bifidobacterium activates the cyclic GMP–AMP synthase–stimulator of interferon genes (cGAS–STING) pathway, enhancing dendritic cell function and CD8⁺ T-cell responses, thereby improving ICB efficacy (143).

Akkermansia muciniphila provides another representative case. Its metabolite c-di-AMP activates the STING–type I interferon pathway, thereby strengthening antitumor immunity and responsiveness to immune checkpoint therapy (198). In metabolic dysfunction–associated fatty liver disease (MAFLD)–associated HCC models, Akkermansia abundance declines during disease progression. In contrast, supplementation significantly improves the local immune landscape. It restores intestinal barrier integrity, reduces serum lipopolysaccharide (LPS) and bile acid levels, alleviates hepatic inflammation, diminishes infiltration of monocytic MDSCs and M2 macrophages, and promotes CD4⁺/CD8⁺ T-cell recruitment and activation, thereby reversing immunosuppression (27). In addition, Bifidobacterium and Lactobacillus species can also produce and release c-di-AMP, which acts as a STING agonist to trigger cGAS–STING–IFN-β signaling in dendritic cells. This strengthens type 1 conventional dendritic cell (cDC1)–mediated cross-priming and expands tumor-specific CD8⁺ T cells, creating synergy with PD-1 blockade and improving antitumor efficacy (199–201). Specific Lactobacillus species enhance ICB efficacy via tryptophan metabolism, L. johnsonii and L. gallinarum modulate CD8⁺ T-cell stemness and Treg differentiation, while L. reuteri generates indole-3-aldehyde to activate AhR signaling and boost IFN-γ⁺ CD8⁺ T-cell function (177, 202).

Faecalibaculum has been reported to enhance immunotherapy efficacy through inosine production, a process that depends on ICB-induced disruption of the intestinal barrier, which permits systemic translocation of inosine and subsequent activation of antitumor T cells (203). Furthermore, engineered microbial strategies have demonstrated even greater translational potential. Xu et al. reported that engineered Lactobacillus reuteri (LR-S-CD/CpG@LNPs), when administered orally in combination with near-infrared irradiation, induced immunogenic cell death in CRC tissues and promoted dendritic cell maturation as well as CD8⁺ T-cell infiltration. Meanwhile, it enhanced I3A production via tryptophan metabolism to activate AhR signaling, thereby markedly strengthening antitumor immunity (204). Similarly, oral delivery of the oncolytic magnetotactic bacterium (MSR-CPT/APP@LPs) has been shown to not only induce immunogenic cell death and activate the cGAS–STING pathway within tumors, but also to reprogram metabolism and optimize the gut microbiota–metabolite axis. This is characterized by elevated SCFA levels that enhance effector T-cell function, reduced kynurenine pathway metabolites that alleviate Treg induction, and decreased glutamine consumption that suppresses tumor energy metabolism (205). The mechanisms by which different bacterial strains promote ICI responsiveness are summarized in Table 3.

These findings suggest that microbial therapies exert dual effects—activating immunity and relieving immunosuppression—through the microbiota–metabolite axis, highlighting the pivotal role of metabolites in reshaping the TME and enhancing immunotherapy efficacy.

Despite the important contributions of conventional in vitro systems and mouse models to our understanding of gut microbiota–immune crosstalk, both approaches have notable limitations. In vitro platforms often lack a complete immune compartment or must rely on peripheral immune cells of heterogeneous origin, and it is technically challenging to maintain steady-state coexistence of an oxygen-requiring epithelium with strict anaerobic commensals within the same system (206). Mouse models, in turn, differ substantially from humans in anatomy, physiology, immune composition, and microbiome structure, which limits the clinical translatability of their findings. Moreover, widely used cancer cell lines are closer to a small-intestinal phenotype and therefore fail to recapitulate the physiological immune responses of colonic epithelium (207). Traditional three-dimensional intestinal organoids can support replication of enteric viruses, but because they inherently lack immune cell components, they cannot reproduce the complex dynamics of epithelial and immune interactions and inflammatory responses. The recently proposed MaugOs (macrophage-augmented gut organoids) address this gap by integrating human macrophages, enabling parallel analysis of viral infection, immune signaling, and inflammatory readouts on a single platform and thereby improving immunophysiological relevance (208). Comprehensive reviews likewise emphasize that organoids can closely reproduce the structure and function of native intestinal epithelium and have become powerful tools for studying TME interactions; however, their capacity to model systemic immune crosstalk and dynamic shifts in the microbiota remains limited (209).

With advances in three-dimensional organoid culture and strategies for integrating immune cells, immune–organoid co-culture systems are emerging as more physiologically relevant experimental platforms. In these systems, intestinal organoids derived from pluripotent or adult stem cells are co-cultured with dendritic cells, macrophages, or T lymphocytes to reconstruct the epithelial–immune interface under controlled conditions, while also sustaining long-term co-culture with strict anaerobic commensals. This configuration more accurately models the human gut immune microenvironment. For example, Recaldin and colleagues established human intestinal immuno-organoids (IIOs) by integrating autologous tissue-resident memory T cells (TRM) into epithelial organoids, thereby creating a self-organizing immune compartment that recapitulates intraepithelial T-cell migration, adhesion, and cytokine signaling (210). In parallel, the MaugOs (Macrophage-augmented Intestinal Organoids) platform demonstrates the functional advantages of immune-enhanced organoids by enabling concurrent analysis of viral infection, epithelial interferon-mediated antiviral pathways, and macrophage-driven inflammatory responses, and by distinguishing virus-specific immune profiles (208).Reviews centered on the gut microbiota–tumor microenvironment axis further emphasize that, although organoids capture tumor heterogeneity and architecture, they remain limited in reproducing global immune crosstalk. Consequently, incorporating immune cells into organoid systems to build immune–organoid co-cultures has become a key direction for improving physiological relevance and translational utility (209).

These systems also offer new strategies to dissect how gut microbial metabolites regulate the epithelial barrier and immune cell differentiation. In the GuMI-APC (Gut–Microbe Interface with Antigen-Presenting Cells) platform, the strict anaerobic commensal Faecalibacterium prausnitzii induces upregulation of epithelial genes such as TLR1 and interferon alpha 1 (IFNA1), while CD4⁺ T-cell–mediated feedback suppresses transcription of inflammation-related genes. These findings indicate that the composition and activation state of immune cells can rewire microbe–epithelium signaling and the associated cytokine networks (211). Further studies highlight SCFAs as the most well-substantiated beneficial metabolites. However, conventional in vitro experiments are limited by physiological constraints and absorption kinetics. Immune-competent organoid co-cultures and multi-organ microphysiological systems, including gut–liver and gut–liver–brain chips, are therefore required to resolve absorption, distribution, metabolism, and excretion and to localize sites of action in the host (212). Within the MaugOs model, acetate markedly suppresses inflammation triggered by LPS and, through inhibition of NLRP3 inflammasome, coordinates inflammatory and stress pathways across macrophages and epithelium. This reveals core mechanisms at the interface of metabolites and inflammation (208, 213). In both human and mouse systems, butyrate inhibits epithelial HDAC3 and, via the HDAC3–sprouty RTK signaling antagonist 2 (SPRY2) pathway, restricts tuft-cell differentiation, which in turn dampens the IL-13–tuft-cell–ILC2 axis and downregulates type-2 immunity (214). Type-2 responses in tumors often promote M2-like macrophage polarization, matrix remodeling, and immunosuppression, which correlate with poor prognosis and reduced sensitivity to ICB. Consequently, attenuation of type-2 immunity through the butyrate–HDAC3 pathway may alleviate such immunosuppression, although the net effect is context dependent and should be validated in immune–organoid co-culture platforms (215).

A recent study reported the PD-IO platform, which integrates patient-derived epithelial organoids, autologous T cells, and myeloid-derived suppressor cell (MDSC)–conditioned medium, with optional anti-PD-1 treatment using pembrolizumab, to reconstruct tumor–immune interactions and ICI response or resistance in vitro. This platform distinguished MSI from MSS responses and, using caspase-3/7 and granzyme B as functional readouts, identified regenerating family member 4 (REG4) and mucin family members such as MUC1 and MUC13 as markers associated with ICI resistance; targeted knockout restored sensitivity to PD-1 blockade (216).Taken together with the mechanistic insights from the butyrate–HDAC3–SPRY2 axis, these observations support evaluating butyrate-producing microbes or butyrate itself in combination with ICI within immune–organoid co-culture systems to accelerate personalized translation. Looking ahead, immune–organoid co-cultures are poised to become important preclinical platforms for testing metabolite-based immunologic adjuvants, probiotic interventions, and ICB combination strategies. Incorporating such systems into studies of gut microbiota–immune crosstalk will not only clarify metabolite-mediated mechanisms of immune regulation but also advance microbiota-informed precision immunotherapy toward clinical application.