Increasing evidence suggests that microbiota composition is not only associated with several human diseases such as autoimmune disease and cancer (Kostic et al., 2013; Scher et al., 2013; Huttenhower et al., 2014; Matson et al., 2021; Connell et al., 2022), but is also responsible for therapeutic efficacy (Allen-Vercoe and Coburn, 2020; Finlay et al., 2020; Huang et al., 2020; Ansaldo and Belkaid, 2021; Pham et al., 2021; Singh et al., 2021), drug resistance (Baruch et al., 2021a; Goc and Sonnenberg, 2022), and adverse effects of multiple therapies (Pham et al., 2021; Seton-Rogers, 2021). Small molecule drugs can be used to ameliorate the altered microbiota composition of cancer patients (Vezza et al., 2019; Huang et al., 2022). More crucially, emerging evidence highlights that microbiota is an invaluable resource for discovering small molecule drugs and microbiota-derived medications have wide applications in anticancer therapy.

Microbiota affects the host in several ways, including the release of metabolic products, cellular components, and secreted proteins that can activate various host receptors to modulate immune responses. Especially, microbiota-derived metabolites have obvious safety advantages over microbiota intervention, without the risk of systemic infection. Microbial metabolites, such as short-chain fatty acids (SCFAs), bile acid, lactic acid, spermidine, indole, and retinoic acid, have been demonstrated to link the intestinal microbiome to systemic immunity (Levy et al., 2017), indicating a future potential perspective in the treatment of tumors (Zhou and Fang, 2018). Furthermore, microbiota and their metabolites also affect the efficacy and toxicity of immunotherapy by modulating host immune responses via different regulatory mechanisms (Inamura, 2020; Zhou et al., 2021). Hence, it is essential to exploit the microbiota and their metabolites to develop novel therapeutic strategies as well as to identify druggable targets that can assist cancer immunotherapy. Here, we highlight recently gained insights into the immunomodulatory functions of microbiota and their metabolites and further demonstrate the unique impacts of microbiota-based therapies on immunotherapy efficacy in patients with cancer, by providing evidence from clinical trials (Figure 1).

A significant metabolic function of the intestinal microbiome is the anaerobic fermentation of ingested dietary fiber and mucosal glycans to produce SCFAs, such as acetate, propionate, and butyrate (Kim et al., 2017). SCFAs have extensive effects on the physiology of the host, especially immunological regulation (Buck et al., 2017). By inhibiting histone deacetylase or combining G-protein-coupled receptors, SCFAs regulate the release of cytokines (Li et al., 2018), as well as the activity of B cells, T cells, Tregs, and innate immune cells (Trompette et al., 2018; Kim, 2021). Specifically, butyrate can increase the clearance of activated T cells by upregulating the Fas cell surface death receptor (Furusawa et al., 2013). In a clinical study of solid tumor patients treated with anti-CTLA-4 antibody, low levels of serum butyrate and propionate were observed to be associated with prolonged progression-free survival (PFS) in the pooled cohort (n = 85), and there was a correlation between gut bacteria and systemic SCFA concentrations (Coutzac et al., 2020). The findings from the other two clinical studies, however, are in contrast with those shown above. Higher fecal SCFA concentrations were substantially related to prolonged PFS in a study of 52 patients with solid tumors under anti-PD-1 treatment (Nomura et al., 2020). Another study indicated that fecal SCFAs, particularly propionate, were associated with improved long-term responses to ICIs in non-small cell lung cancer (NSCLC) patients treated with anti-PD-1 antibody (Botticelli et al., 2020).

Bile acids (BAs) are another type of metabolite biotransformed by microbiota that exert immunomodulatory properties in the intestine of the host (Campbell et al., 2020; Li et al., 2021). Crosstalk between intestinal microbiota, BAs, and the host influences immune functions, metabolic phenotypes, and risk factors for various cancers (Schroeder and Bäckhed, 2016; Sun et al., 2021). Several intestinal strains of the genus Clostridium have been discovered to produce secondary BAs such as deoxycholic acid (DCA) and lithocholic acid (LCA) by removing the 7α/β-hydroxy group from primary BAs (Ridlon et al., 2016). One of the BAs called isodeoxycholic acid produced by gut bacteria through the epimerization of DCA (Devlin and Fischbach, 2015) increases the production of peripherally generated Treg cells by suppressing the immunostimulatory capacity of dendritic cells (Campbell et al., 2020). Moreover, a preclinical study showed that the LCA derivative 3-oxoLCA inhibited T_H17 cell differentiation, whereas another derivative of LCA called isoalloLCA boosted Treg cell expansion (Hang et al., 2019).

As a necessary amino acid for humans, tryptophan and its metabolites are known to be bioactive substances that can affect immune cell differentiation by acting as arylhydrocarbon receptor (AhR) ligands (Sun et al., 2020) and thus have significant effects on the regulation of the immune system and the development of cancer. Most of the dietary tryptophan is absorbed in the small intestine, while a minor amount reaches the colon and is metabolized by the intestinal microbiome (Cryan et al., 2019). Kynurenine produced from tryptophan was the most remarkably elevated serum metabolite in response to anti-PD-1 antibody, and an increased serum kynurenine/tryptophan ratio was related to shorter overall survival in melanoma and renal cell carcinoma (RCC) patients (Li et al., 2019). A clinical study that evaluated plasma tryptophan metabolites in 19 NSCLC patients treated with ICIs discovered that low levels of 3-hydrozyanthranilic acid were substantially correlated with prolonged median PFS (Karayama et al., 2021).

Inosine derived from gut microbes has also been demonstrated to be associated with ICI therapy responses. He et al. (2017) found that the gut microbiota regulated the concentration of inosine, which inhibited Th1/Th2 cell differentiation by combining with adenosine A2A receptor. Another study demonstrated that inosine improved anti-tumor responses of immunotherapy in mouse models of various cancer types (Mager et al., 2020). Furthermore, the combination of inosine and a PD-L1 inhibitor resulted in delayed tumor growth and prolonged survival time in a melanoma mouse model (Wang et al., 2020). However, some cancer cells compete with T cells for inosine as an alternative energy source when glucose is lacking, dampening the efficacy of inosine supplementation in combination with anti-PD-L1 therapy. Conversely, several microbial metabolites, such as polyamines and lipoteichoic acid, contribute to the shaping of the tumor-promoting microenvironment (Loo et al., 2017; Proietti et al., 2020). With a growing number of microbiota-derived mediators being discovered, those that cause favorable immune responses in the host are being turned into possible therapeutic medications.

Dietary habits play a pivotal role in shaping microbiome variance among people, and their alterations can modify the gut microbiome, but not necessarily permanently. Dietary intervention targeting the gut microbiome has emerged as an appealing approach due to its safety and convenience, although it may have fewer striking effects than FMT or probiotics. Growing evidence supports the use of the ketogenic diet as an adjunctive strategy to the antitumor effects of ICI. In a murine melanoma model, the ketogenic diet and its main ketone body, three hydroxybutyrate (3HB), not only pronouncedly shifted the composition of the gut microbiota but also restored the effectiveness of ICI by inhibiting PD-L1 expression and promoting the expansion of CXCR3⁺ T cells (Ferrere et al., 2021). A clinical trial evaluating the potential effect of the ketogenic diet on immunotherapy outcomes is currently underway (NCT04316520).

The fasting mimicking diet (FMD), a plant-based, calorie-restricted, low-carbohydrate, low-protein diet, has been proposed as a potential anticancer dietary intervention to modulate gut microbiota composition and immune cell profiles. Recently, a clinical study focusing on the association of FMD intervention with ICIs in patients with NSCLC is ongoing (NCT03700437). Another phase II trial using metformin plus/minus cyclic FMD is currently being conducted in advanced LKB1-inactive lung cancer to increase the effectiveness of first line chemoimmunotherapy (NCT03709147). Patients with head and neck cancer are being studied to determine whether they will respond better to ICIs if they restrict their eating to an 8–10-h window each day and then fast for a more extended period each night (NCT05083416). Moreover, several clinical studies are evaluating the feasibility of high-fiber or low-protein diet and resistant starch in cancer patients who are receiving ICI treatment (Table 1). As mentioned above, there are some ongoing efforts to combine ICI with dietary intervention, but no results have been obtained thus far. It is necessary to further explore the potential mechanisms explaining the immunomodulatory effects of dietary intervention. Moreover, considering the limited influence of diets on gut microbiota, combining dietary intervention with FMT may help to expand the benefits of immunotherapy.

Dietary interventions may appear to be easy to implement, but their impacts on the gut microbiota are often limited, and patients’ compliance is challenging to enforce and monitor. Direct administration of specific probiotics or their metabolites could be a superior choice due to their high efficiency but few side effects. Currently, numerous clinical studies are being conducted to further investigate the therapeutic potential of prebiotics in combination with immunotherapy due to their immunomodulatory effects on the host immune system (Table 1).

Various single strains of bacteria have been used as probiotics to improve clinical outcomes through modulation of the immune response. A preclinical study in syngeneic mouse models of breast and lung cancer demonstrated that the E. gallinarum strain MRX0518 could inhibit tumor growth in conjunction with an elevation in the CD8⁺ T cell: Treg ratio (Stevenson et al., 2018). Preliminary data from a small cohort of multiple cancer types demonstrated that oral live biotherapeutic MRx0518 enhanced antitumor activity, manifested by significant increases in genes and metagenes involved in antigen presentation, innate immune processes, interferon response, Th1 cells, and CD8⁺ cells (NCT03934827) (Lythgoe et al., 2021). The clinical studies of MRx0518 as a cotherapy with PD-1/PD-L1 inhibitors to treat various tumors are ongoing (NCT03637803, NCT05107427).

More interestingly, an increasing number of preclinical studies and clinical trials have reported that Bifidobacteria and Lactic acid bacteria can improve the curative effect of cancer immunotherapy. A single strain of Bifidobacterium animals lactis called EDP1503 has been shown to induce systemic antitumor immunity by elevating the production of cytokines (such as IFN- and CXCL10), activating NK cells and CD8⁺ T cells and triggering a proinflammatory signature within the TME (Gardner et al., 2019). The safety, tolerability, and efficacy of EDP1503 in combination with ICI have been evaluated in multiple cancer types (NCT03775850). EDP1503 administered with pembrolizumab was safe and well tolerated in patients with microsatellite stable colorectal cancer, without significant serious therapeutic toxicity (McHale et al., 2020). Again, the satisfactory safety and tolerability of EDP1503 and pembrolizumab combination therapy were proven in Triple-negative breast cancer (TNBC) patients, who experienced clinical benefit with an ORR of 18% and a disease control rate (DCR) of 27% in evaluable patients (n = 11) (Yuan et al., 2021). A phase 2 trial (NCT03595683) is presently recruiting patients with advanced melanoma to evaluate the efficacy of EDP1503 in enhancing the response to conventional immunotherapy.

A specific strain of Lactobacillus lactis known as GEN-001 can generate immunogenic metabolites, thus activating multiple immune cells, including T cells, DCs, and macrophages, potentially affecting the efficacy of immunotherapy (Chen et al., 2022). A phase I clinical study to investigate the safety, tolerability, and biological and clinical activities of GEN-001 in combination with avelumab in advanced solid tumor patients who have progressed during or after receiving ICI therapy is underway (NCT04601402). To explore the oral probiotics V9 (Lactobacillus bifidobacterium) and M9 (Lactobacillus rhamnosus) combined with PD-1 inhibitors for patients with NSCLC and liver cancer, two clinical studies were conducted (NCT05094167, NCT05032014).

CMB588 contains Clostridium butyricum, which belongs to SCFA-producing probiotics and has immunomodulatory activity. The addition of CBM588 prolonged median PFS (12.7 versus 2.5 months, hazard ratio (HR) 0.15, 95% CI 0.05–0.47, p < .001) and ORR (58% versus 20%) in patients with metastatic RCC receiving nivolumab plus ipilimumab. Furthermore, patients receiving CBM588 had alterations in their gut microbiota, including upregulation of rhamnose synthesis and increased SCFA propionate production (NCT03829111) (Dizman et al., 2022). Currently, many other clinical trials are underway to evaluate the safety and anticancer effect of commensal bacteria and immunotherapy combinations in various cancer patients (NCT04699721, NCT04167137, NCT04009122, NCT03817125).

Compared to FMT and single-strain probiotics, microbial ecosystem therapeutics (METs) are more effective and safer since they are composed of deliberately engineered bacterial communities derived from the intestinal microorganisms isolated from the feces of a healthy donor. Initial results from an early phase 1 study have thus far supported the safety of oral administration of MET-4 in combination with an anti-PD-1/PD-L1 inhibitor in cancer patients (NCT03686202) (Araujo et al., 2020). A consortium of 11 commensal strains isolated from the feces of healthy donors was found to improve the therapeutic efficacy of ICIs in mouse models by activating interferon-γ-producing CD8⁺ T cells in the intestine (Tanoue et al., 2019). This consortium, which is now being produced as a probiotic capsule called VE800, is undergoing Phase 1 clinical testing as an orally delivered therapy in combination with the anti-PD-1 inhibitor nivolumab (NCT04208958). MaaT013 is regarded as the next-generation FMT product with high standardization, produced from pooled healthy donors, and characterized by a highly consistent richness of 455 intestinal microbiome species. A prospective randomized clinical trial (NCT04988841) assessing the tolerance and clinical benefit of MaaT013 (administered via enema) in melanoma patients treated with CTLA-4 and PD1 inhibitors has begun. A phase 4 trial is now being conducted to assess the clinical efficacy of combination probiotics (Bifidobacterium, Lactobacillus, and Enterococcus capsules) in immunotherapy for patients with bladder urothelial cancer (NCT05220124).

Microbiota represents a rich source of small molecule drug discovery. However, emerging endeavors in microbiota mainly focus on manipulating microbiota directly to enhance immunotherapy efficacy. Small molecules derived from microbiota are rarely developed into drugs specific to immunotherapeutic effects. Identifying responsible effector molecules from microorganisms that affect host immunity and deciphering the corresponding regulatory mechanisms is a crucial step for the development of microbiota-based therapeutics. Moreover, multicenter trials are warranted to further evaluate the feasibility and safety of microbiota-based therapies in patients with cancer. As an increasing number of immunomodulators are discovered from microbiota, microbiota-derived medications will be developed and offer better options to improve the benefit of checkpoint immunotherapy.