Probiotics in colorectal cancer: mechanisms, biomarkers, and adjunct strategies

Restoring the intestinal mechanical barrier

The intestinal barrier is comprised of mechanical and chemical components that work in concert to protect the host from luminal threats. Mechanical barriers physically block and remove pathogens (e.g., mucus and epithelial tight junctions), while chemical barriers (antimicrobial peptides and enzymatic mediators) kill or inhibit pathogens using specialized molecules (e.g., α/β-defensins, cathelicidin, and lysozyme). The intestinal epithelial barrier, as the primary mechanical defense, functions as a selectively permeable interface by restricting the passage of harmful substances while supporting epithelial regeneration. Bacterial endotoxins [e.g., lipopolysaccharide (LPS)], food antigens, and microbial metabolites translocate into the systemic circulation when this barrier is compromised, triggering a cascade of chronic inflammation. This persistent inflammatory state promotes local immune dysregulation and inflicts damage on epithelial cells, thereby driving malignant transformation¹⁶. Key barrier components are essential for this protective function. For example, the tight junction (TJ) protein, claudin-7, is crucial for tissue integrity. In fact, claudin-7 deficiency in mice disrupts cell junctions, exacerbates inflammation, and promotes severe colitis and colon cancer¹⁷. Similarly, mucin-2 (Muc2), the primary secreted gastrointestinal mucin, is vital for maintaining normal intestinal architecture. Mice lacking Muc2 frequently develop invasive adenocarcinomas, underscoring the critical role in suppressing CRC¹⁸.

Several probiotic strains can repair and reinforce the intestinal epithelium (Table 1). For example, Lactobacillus acidophilus (L. acidophilus) upregulates occludin²⁰, Lactobacillus plantarum (L. plantarum) upregulates ZO-1 and ZO-2¹⁴, Lactobacillus rhamnosus (L. rhamnosus) upregulates claudin-3²¹; Bifidobacterium infantis (B. infantis)upregulates claudin-4²⁴; and Escherichia coli Nissle (E. coli Nissle) 1917 upregulates ZO-2²⁶. Other strains, including Bacillus subtilis (B. subtilis) CW14 and L. plantarum ZLP001, prevent ZO-1 degradation and alleviate damage from increased intestinal permeability^19,27. Probiotics also support the epithelial barrier by increasing goblet cell populations [Faecalibacterium prausnitzii (F. prausnitzii)³⁰] and elevating mucin (MUC-2) expression [Clostridium butyricum (C. butyricum)³¹]. Mechanistically, probiotic-mediated intestinal barrier repair frequently involves activation of toll-like receptor (TLR) signaling pathways. Strains, such as Bacillus amyloliquefaciens SC06³², B. subtilis CW14²⁷, and Lactobacillus reuteri (L. reuteri)²², modulate TLR signaling pathways to improve mucosal structure and TJ integrity. For example, Bifidobacterium bifidum (B. bifidum) activates the TLR-2/TLR-6 receptor complex on intestinal epithelial cells, thereby inducing IRAK-1 phosphorylation and recruitment of the anti-inflammatory adapter protein, TOLLIP, to the apical membrane, which results in the downstream upregulation of occludin²⁵. L. acidophilus enhances barrier function through a TLR-2-dependent pathway involving PI3K-mediated inhibition of TNF-α-induced NF-κB activation, thereby protecting against inflammation-induced barrier disruption³³. Another critical pathway involves the suppression of pro-inflammatory signaling cascades. For example, B. bifidum prevents TNF-α-induced barrier disruption³⁴, while B. infantis and L. acidophilus normalize occludin and claudin-1 expression, and suppress NF-κB activation under IL-1β stimulation²³. Furthermore, L. rhamnosus CY12 can alleviate colitis by inhibiting the TLR-4/MyD88/NF-κB signaling pathway, thereby reducing pro-inflammatory cytokine (IL-6, IL-1β, and TNF-α) expression, while promoting production of the anti-inflammatory cytokine, IL-10³⁵. Suppression of NF-κB can also prevent activation of myosin light chain kinase (MLCK), which induces actomyosin contraction and TJ opening by phosphorylating the myosin II regulatory light chain (MLC2)³⁶.

In addition to direct regulation of host proteins, the protective metabolites produced by these probiotics have a critical role. For example, propionate secreted by Akkermansia muciniphila (A. muciniphila) activates the GPR43 receptor to enhance the expression of occludin, ZO-1, and mucins²⁸. In a more complex interaction, conjugated linoleic acid (CLA) produced by Bifidobacterium breve (B. breve) promotes the growth of Odoribacter splanchnicus, which in turn increases butyrate production. This cascade upregulates TJ proteins and modulates inflammatory cytokines¹². Roseburia intestinalis also directly improves intestinal permeability in Apc^Min/+ mice via butyrate production²⁹. Collectively, these mechanisms demonstrate that probiotics maintain intestinal barrier integrity through direct protein upregulation, immune modulation, and metabolite-mediated protection, thereby counteracting inflammatory and pathogenic insults and highlighting the therapeutic potential in CRC (Figure 1A).

Inhibition of pathogenic bacteria growth and colonization

CRC is marked by severe gut microbiota dysbiosis, in which pathogenic bacteria become enriched and actively contribute to disease progression⁷. Stool transplants from CRC patients can directly induce carcinogenesis in mice, leading to increased dysplasia, proliferation, and pro-tumorigenic inflammation³⁷. Pathogens, such as Fusobacterium nucleatum (F. nucleatum), Peptostreptococcus anaerobius, and Bacteroides fragilis (B. fragilis), drive tumorigenesis through multiple mechanisms, including enhanced mucosal adhesion³⁸, production of harmful metabolites³⁹, induction of damage via virulence factors⁴⁰, and conferring resistance to chemotherapy^41,42. While targeted strategies, like selective antibiotics or phage therapy, have been explored, the limitations are significant. Antibiotic overuse risks driving bacterial resistance and further disrupts microbial balance. Indeed, the elimination of a single species often fails to restore overall community homeostasis. Consequently, there is a compelling rationale for more systematic microbiota modulation, such as can be achieved with probiotic supplementation. For example, a clinical study in CRC patients demonstrated that oral administration of a probiotic mixture containing Lactobacillus lactis (L. lactis) and L. acidophilus significantly increased beneficial butyrate-producers, while reducing CRC-associated taxa, like Clostridium and Peptostreptococcus⁴³. This finding underscores the potential of probiotics to simultaneously reconfigure microbial populations and ameliorate dysbiosis.

Probiotics can significantly reshape the gut microbial community structure and exert broader modulatory effects on the intestinal ecosystem (Table 2). For example, administration of L. rhamnosus GG or B. breve CCFM683 significantly increase the abundance of key butyrate producers, such as Faecalibacterium, Roseburia, and Akkermansia^12,46,47. This shift is particularly relevant for CRC prevention because butyrate reinforces the intestinal barrier, exhibits anti-inflammatory properties, and supports colonocyte health. Another consistent pattern is the suppression of pro-inflammatory and potentially pathogenic taxa. Probiotic interventions often reduce the relative abundance of genera associated with mucosal inflammation and tumor promotion. For example, L. plantarum L168 inhibits the overgrowth of Oscillibacter and Alistipes, while simultaneously increasing the abundance of Faecalibaculum and Bifidobacterium⁵⁰. In addition to single-strain interventions, rationally designed multi-strain consortia can further enhance the precision of microbial modulation. A synthetic community (SynCom) composed of seven strains successfully eradicated F. nucleatum and suppressed F. nucleatum-associated CRC in mice⁵¹. Similarly, larger microbial consortia have been shown to enhance resistance to enteric pathogens or alleviate intestinal inflammation^52,53.

Probiotics prevent pathogenic colonization and expansion through several strategies: (1) secretion of direct antimicrobial compounds; (2) sequestration of essential nutrients; (3) blockade of mucosal adhesion; and (4) interference with pathogen quorum sensing (QS) (Figure 1B). Direct antibacterial activity could be mediated by secreted molecules. For example, nisin from L. lactis forms pores in the membrane of pathogens, particularly, gram-positive bacteria, leading to rapid bacteria death, while L. lactis self-protects from nisin by expressing NisI (a membrane-anchored lipoprotein capturing nisin before reaching the membrane) and NisFEG (an ABC transporter expelling nisin)⁵⁴. Organic acids, such as lactic and acetic acids, penetrate bacterial membranes, acidify the cytoplasm, and impair the proton motive force, thereby inhibiting growth. Engineered Saccharomyces cerevisiae with a BTS1 gene knockout produces lactic acid, lowering the environmental pH to inhibit E. coli and Staphylococcus aureus (S. aureus) growth, and disrupt biofilm formation⁵⁵. These pathogens are often more vulnerable than commensal bacteria to environmental stresses, such as acidification, metabolic strain, and membrane disruption. In addition, vitamin B6 produced by Lactobacillus spp. can compromise pathogenic cell membrane integrity, leading to leakage of intracellular contents and bacterial death⁵⁶. Competitive exclusion through nutrition limitation represents another critical mechanism. For example, B. infantis suppresses host apoptotic pathways, activates anti-inflammatory signaling, and competes with Salmonella typhimurium (S. typhimurium) for arginine, thereby limiting growth⁵⁷. Iron is another nutrient for which strains compete. The probiotic strain, E. coli Nissle 1917, produces high-affinity siderophores, such as salmochelin, enabling efficient iron acquisition and thereby outcompeting and reducing S. typhimurium colonization in mouse models of acute colitis and chronic persistent infection⁵⁸. Similarly, Bifidobacteria with strong iron-chelating capacity can inhibit the growth and adhesion of enterohemorrhagic E. coli and Salmonella in vitro⁵⁹. Blocking mucosal adhesion establishes an additional physical barrier that limits colonization by opportunistic pathogens. Probiotics facilitate this effect by competing for glycan receptors on intestinal epithelial cells, typically through surface-layer proteins or pili. For example, Bifidobacterium pseudolongum (B. pseudolongum) PV8-2 exhibits high affinity for mucins and significantly reduces S. typhimurium adhesion to epithelial cells in vitro⁵⁹. QS allows individual bacteria to coordinate colony-wide functions, making QS a promising target for intervention. Probiotics can disrupt this process by secreting enzymes or metabolites that degrade or mimic QS signaling molecules and reducing virulence without exerting direct killing. For example, Bacillus strains produce fengycins, a class of lipopeptides that inhibit the S. aureus agr QS system, reducing toxin production and biofilm formation⁶⁰. In addition, certain Lactobacillus strains degrade autoinducer-2 (AI-2), a signal molecule widely used by gram-negative bacterial pathogens to coordinate virulence⁶¹. While these mechanisms are promising, corollary human studies are needed. Ultimately, exploiting such microbial antagonism to identify targeted antimicrobial molecules could offer novel avenues for microbiota-based CRC therapies.

Probiotic metabolites target cancer signals

Probiotic-derived metabolites are central to the therapeutic benefits against CRC. These microbial molecules, encompassing carbohydrates, lipids, proteins, and small compounds, directly reprogram oncogenic signaling, modify epigenetic pathways, and enhance antitumor immunity to suppress tumor growth. Probiotic-derived metabolites could promote the accumulation of reactive oxygen species (ROS) in tumor cells to trigger mitochondrial apoptotic pathways. For example, Bifidobacterium longum D42 elevates intracellular ROS in HT-29 cells, disrupts mitochondrial membrane potential, promotes cytochrome c release, and activates downstream caspase cascades, ultimately inducing CRC cell apoptosis via the intrinsic mitochondrial pathway⁶². In contrast, probiotic metabolites could directly interfere with core oncogenic networks to halt proliferation and induce apoptosis. The well-studied metabolite, butyrate, inhibits pro-survival pathways, like NF-κB and STAT3, downregulating cyclins, and activating apoptosis genes⁶³. Moreover, butyrate has been reported to suppress vascular endothelial growth factor (VEGF) expression by blocking the transactivation of the key transcription factor [specificity protein 1 (Sp1)], thereby reducing tumor angiogenesis and migration^64,65. Riboflavin from Bacteroides cellulosilyticus binds to ceramide synthase (CERS3), blocking the synthesis of very long-chain ceramide (C26) and the interaction with epidermal growth factor receptor (EGFR), thereby attenuating proliferation⁶⁶. Conjugated linoleic acid (CLA) from B. breve activates PPAR-γ to promote tumor cell apoptosis¹². Furthermore, gut microbiota-derived N-acetylmuramic acid inhibits AKT1 phosphorylation, suppressing the AKT1-FOXO3a pathway to suppress cancer cell proliferation and induce apoptosis⁶⁷. In addition to small molecules, probiotic macromolecules exert antitumor effects. β-galactosidase from Streptococcus thermophilus generates galactose, impairing glucose utilization and activating oxidative phosphorylation to inhibit tumor growth⁴⁴. The extracellular polysaccharide, YL-11 EPS, from Lactobacillus YL-11 blocks the PI3K/AKT pathway, arresting the cell cycle at G1 phase, thereby suppressing colorectal carcinogenesis⁶⁸. In addition, extracellular vesicles from L. plantarum can be internalized by CRC cells, downregulating the desuccinylase (SIRT5) to enhance succinylation of p53, inhibiting glycolysis and proliferation while inducing apoptosis⁶⁹. Probiotic metabolites can also function as epigenetic regulators, reversibly altering chromatin states to reshape transcriptional programs. For example, butyrate acts as an endogenous histone deacetylase (HDAC) inhibitor that increases histone acetylation by suppressing HDAC activity, thereby opening chromatin structure to activate the transcription of tumor suppressor and apoptosis-related genes, while simultaneously repressing oncogene expression^70,71. Together, probiotic-derived metabolites combat CRC through diverse mechanisms, including oxidative stress induction, modulation of oncogenic signaling pathways, suppression of angiogenesis, and epigenetic reprogramming via HDAC inhibition. These findings underscore the therapeutic potential of these bioactive molecules as mechanism-based adjuncts to conventional CRC therapy.

Probiotic metabolites remodeling the tumor immune microenvironment (TIME)

While ICI therapy targeting PD-1 is approved for microsatellite instability (MSI) CRC, the efficacy remains restricted in the majority of microsatellite stable (MSS) tumors⁷². MSS CRC is typically considered an immunologically “cold” tumor that is characterized by limited infiltration of cytotoxic T cells within the TIME, and enrichment of immunosuppressive populations, such as myeloid-derived suppressor cells (MDSCs), regulatory T cells (Tregs), and M2-type tumor-associated macrophages (TAMs)⁷³. This immunosuppressive milieu leads to intrinsic resistance to ICI therapy. A key mechanism by which probiotics and metabolites may overcome this limitation is by reprogramming the immunosuppressive tumor microenvironment, converting “cold” tumors into “hot” immunologically active tumors to synergize with ICIs. For example, among patients resistant to ICIs, C. butyricum can colonize tumors and locally produce butyrate to suppress the IL-6/JAK2/STAT3 pathway, subsequently reducing immunosuppressive cell infiltration and boosting cytotoxic CD8⁺ T cell function to restore tumoral sensitivity to anti-PD-1 therapy¹³. Furthermore, L. plantarum-derived indole-3-lactic acid (ILA) orchestrates a multi-layered epigenetic attack. ILA promotes IL-12 secretion in dendritic cells (DCs) to prime T-cell immunity, while ILA remodels chromatin in CD8⁺ T cells to enhance infiltration and cytotoxic activity⁵⁰. Capsular polysaccharides derived from Bacteroides uniformis bind to Lag-3 molecules on the cell surface of natural killer (NK) cells, activating cytotoxic function and enhancing tumor cell killing⁷⁴. Importantly, this pathway synergizes with anti-PD-1/PD-L1 therapy to further strengthen immune responses against the immunologically “cold” tumors⁷⁴. Moreover, short-chain fatty acids (SCFAs), such as acetate and butyrate, have been reported to directly enhance antitumor NK cell activity in vitro. SCFAs promote the release of NK-derived extracellular vesicles, which can carry cytotoxic cargo, while reducing secretion of the immunosuppressive cytokine, IL-10⁷⁵.

Probiotic-derived metabolites can alleviate the immunosuppressive state of the TIME. For example, metabolites from L. plantarum can reprogram tumor-associated macrophages from a pro-tumor M2 phenotype to an anti-tumor M1 state by activating the NLRP3 inflammasome⁷⁶. Lysates of L. acidophilus, when combined with anti-CTLA-4 therapy, can increase the number of CD8⁺ T cells, while reducing CD4⁺CD25⁺Foxp3⁺ regulatory T cells and the M2 phenotype of macrophages within the TIME⁷⁷. Lactobacillus gallinarum secretes indole-3-carboxylic acid that diminishes recruitment of Foxp3⁺CD25⁺ regulatory T cells and enhances CD8⁺ T cell function, synergistically augmenting the inhibitory effect of anti-PD-1 antibodies on tumor growth⁴⁵. In addition, SCFAs suppress the accumulation of MDSCs in tumor tissues, which alleviates MDSC-mediated immunosuppression and promotes CD8⁺ T cell infiltration and activation⁷⁸. Notably, tyrosine kinase inhibitors can also increase the abundance of Muribaculum and the metabolite, urocanic acid (UCA), to potentiate cancer immunotherapy efficacy. Increased UCA inhibits p65 signaling in tumor endothelial cells to reduce MDSC recruitment via the CXCL1-CXCR2 axis, ultimately suppressing colorectal tumorigenesis⁷⁹. Collectively, these findings demonstrate how microbial metabolites can reshape the TIME at cellular, metabolic, and epigenetic levels to potentiate anticancer immunity (Figure 2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Regulation of the tumor microenvironment by probiotics and metabolites. Probiotics and derived metabolites suppress CRC through direct effects on tumor cells and indirect remodeling of the immune microenvironment. Direct mechanisms involve strain-specific signaling regulation. S. thermophilus produces β-galactosidase to modulate the Hippo pathway, whereas B. breve (via conjugated linoleic acid) and L. Y-11 (via YL-11 EPS) inhibit NF-κB activity, thereby reducing tumor cell survival and proliferation. Indirectly, these metabolites promote antitumor immunity by enhancing NK cell recruitment (e.g., B. uniformis secretes ZPS capsule), CD8⁺ T cell cytotoxicity (e.g., C. butyricum produces butyrate), and dendritic cell antigen presentation (e.g., L. plantarum secretes indole-3-lactic acid). B. breve, Bifidobacterium breve; B. cellulosilyticus, Bacteroides cellulosilyticus; B. uniformis, Bacteroides uniformis; C. butyricum, Clostridium butyricum; EGFR, epidermal growth factor receptor; GRP78, glucose-regulated protein 78; L. plantarum, Lactobacillus plantarum; L. Y-11, Lactobacillus. YL-11; MDSC, myeloid-derived suppressor cell; NK, natural killer; S. thermophilus, Streptococcus thermophilus. Figure created using BioRender (www.biorender.com).

Probiotic signatures as diagnostic markers

Current diagnostic research into gut microbiota for CRC and colorectal adenomas (CRAs) has predominantly focused on pathogenic bacteria (e.g., F. nucleatum), metabolites, or a combination with host genetic markers. While these microbial features are often examined as auxiliary tools, a parallel and compelling diagnostic signal lies in the progressive depletion of probiotic taxa during CRC progression. This decline is driven by multiple interrelated factors, including elevated oxidative stress within the tumor tissues that inhibits survival of strict anaerobes^80,81, tumor-associated metabolic alterations that lead to ecological niche loss⁸², and remodeling of the host immune microenvironment that excludes some probiotics and impedes stable colonization^83,84. Such antagonistic ecological dynamics suggest that the loss of beneficial microbes represents a potent yet underexploited biomarker class. The diagnostic power of this ecological shift is evident when probiotic signatures are formally quantified. For example, a simple abundance ratio of F. nucleatum-to-beneficial taxa (F. prausnitzii, Bifidobacterium, and Lactobacillus) can significantly improve diagnostic specificity over single-pathogen assays⁸⁵. More sophisticated, multi-taxa classifiers that incorporate probiotic genera have demonstrated strong performance. Specifically, a seven-genus panel, including Lactobacillus, achieved an AUC of 0.84 for CRC detection⁸⁶, while a six-taxon panel containing Roseburia inulinivorans differentiated CRA from CRC with an AUC of 0.93⁹. In specific contexts like obesity-associated CRC, the marked reduction of F. prausnitzii has been a key discriminative feature, reinforcing the broader role as a biomarker for intestinal health^87,88.

Despite this promise, probiotic signatures remain underrepresented in diagnostic models. A primary challenge is the inconsistent loss across CRC stages, complicating the development of universal single-taxon biomarkers. Furthermore, gut microbiota composition is profoundly confounded by host-specific factors, including environment, genetics, medication history, and demographic variables, like gender, as illustrated by the hormone-dependent colonization of Carnobacterium maltaromaticum⁸⁹. Another key limitation is the lack of disease specificity because the reduction of probiotic taxa is not unique to CRC. For example, depletion of F. prausnitzii is commonly observed in inflammatory bowel disease (IBD) and other metabolic disorders⁹⁰. Such overlapping dysbiosis patterns may undermine the diagnostic specificity of probiotic signatures and increase the risk of false positives. Nevertheless, CRC and IBD exhibit quantifiable differences in microbial structure^90,91, suggesting that diagnostic accuracy could be improved through multi-feature microbial profiling rather than reliance on single taxa. Therefore, these observations highlight a critical diagnostic strategy in which probiotic signatures ae integrated into composite multi-marker classifiers. Such models can account for ecologic dynamics (pathogen rise and probiotic decline), stage-specific variability, and host confounders, moving beyond auxiliary use to establish a robust, ecology-informed framework for CRC and CRA detection. From a translational perspective, routine clinical implementation of probiotic signatures faces ongoing challenges in standardization, cost-effectiveness, and analytic reproducibility, despite the encouraging feasibility demonstrated by fecal microbiome analysis using quantitative PCR and metagenomic sequencing. Future efforts should therefore focus on establishing standardized testing workflows and diagnostic thresholds, as well as validating the robustness, alone or in combination with existing biomarkers, across diverse populations through large-scale prospective studies to facilitate clinical adoption.

Probiotic signatures as predictive markers

In addition to diagnostics, fecal microbial signatures are emerging as valuable predictors of therapeutic response and clinical outcomes in CRC. The predictive relevance is particularly pronounced in the context of immunotherapy, in which the composition of the gut microbiota can critically influence efficacy. A growing body of evidence indicates that favorable responses to ICIs and prolonged survival are strongly associated with the enrichment of specific probiotic taxa and the metabolic byproducts, rather than solely the depletion of pathogens. Pan-cancer analyses have consistently shown that ICI responders harbor increased abundances of bacteria, such as A. muciniphila⁹², F. prausnitzii¹¹, and L. reuteri⁹³, as well as fungi with probiotic-like functions (e.g., Schizosaccharomyces⁹⁴ and Nemania serpens¹¹). In parallel, elevated levels of related metabolites, such as indole-3-aldehyde (I3A)⁹³, galanthamine ketone⁹⁵, and hippuric acid⁷⁹ are significantly associated with improved prognosis. Butyrate-producing strains (Agathobacter M104/1 and Blautia SR1/5) have been identified as potential prognostic indicators for combined cetuximab and avelumab therapy in CRC⁹⁶. These findings highlight a key insight, i.e., therapeutic success may depend more on the active presence of beneficial microbes than on the absence of harmful microbes.

Probiotic signatures also hold promise in predicting susceptibility to immune-related adverse events (irAEs) during immunotherapy⁹⁷. In patients with metastatic melanoma receiving anti-CTLA-4 therapy, those with higher baseline abundances of Bacteroidetes exhibited greater resistance to ICI-induced colitis⁹⁸. In contrast, microbial profiles enriched in Firmicutes were associated with treatment-related colitis⁹⁹. Therefore, observations suggest that incorporating microbial signatures into risk stratification models may enable early detection and preventive interventions, thereby expanding the therapeutic window of ICIs.

Outside of immunotherapy, the direct predictive role of probiotics is less clear. Probiotics are primarily documented for their role in mitigating treatment-related side effects in chemotherapy¹⁰⁰ with limited evidence supporting a substantial direct impact on chemotherapeutic efficacy or overall survival. Nonetheless, certain probiotic signatures may hold broader prognostic value, as illustrated by the association between higher abundance of Carnobacterium maltaromaticum and improved outcomes, particularly in female patients⁸⁹. Overall, probiotic signatures and the metabolic products represent a promising, yet underdeveloped, component of predictive oncology. Integrating these features into models of therapeutic responsiveness, especially for immunotherapy, offers a feasible strategy to refine patient stratification and personalize treatment. However, as with diagnostic applications, research in this area remains nascent. Translating these insights into clinical practice will require more comprehensive, longitudinally designed studies and multi-omics integration to validate and standardize microbial predictors across diverse patient populations and treatment regimens.

Modulating microbiomes to enhance immunotherapy

The gut microbiota is a critical determinant of immunotherapy efficacy in CRC. A compelling illustration of this is the finding that fecal microbiota transplantation (FMT) from healthy donors restores anti-PD-1 responsiveness in murine CRC models, whereas FMT from non-responding CRC patients fails to do so¹⁰¹. This finding highlights the therapeutic potential of modulating the microbiome with probiotic supplementation emerging as a promising strategy to overcome resistance to ICIs (Table 3). Probiotics exert direct antitumor effects and reprogram the TIME. For example, Lactobacillus paracasei sh2020, which has been isolated from healthy human feces, upregulates the chemokine, CXCL10, within tumors and enhances recruitment of cytotoxic CD8⁺ T cells and synergizing with anti-PD-1 therapy¹⁰¹. Similarly, the surface protein, SecD, of C. butyricum binds to glucose-regulated protein 78 (GRP78) on CRC cells, inactivating the downstream PI3K-AKT-NF-κB pathway¹³. This interaction suppresses the production of the immunosuppressive cytokine, IL-6, which alleviates the functional inhibition of CD8⁺ T cells and reduces the infiltration of TAMs, thereby creating a more permissive environment for immunotherapy¹³.

In addition to direct cellular interactions, probiotic-derived metabolites systemically potentiate antitumor immunity. SCFAs, like butyrate, produced by genera (Roseburia and Clostridium) directly enhance the cytotoxic function of CD8⁺ T cells, thereby synergizing with anti-PD-1 therapy^13,29. Mechanistically, butyrate modulates T-cell receptor (TCR) signaling and engages Toll-like receptor 5 (TLR-5) to activate NF-κB, amplifying T-cell effector capacity^29,107. Other microbial metabolites also fine-tune immune responses. Specifically, indole derivatives (e.g., indole-3-propionic acid and indole-3-carboxylic acid) influence T-cell stemness and differentiation^45,102,103, while B. pseudolongum utilizes the L-arginine pathway to drive the generation of tissue-resident memory CD8⁺ T cells, boosting anti-CTLA-4 efficacy¹⁰⁸. Notably, SCFAs, like butyrate and valerate esters, have also been shown in vitro to enhance the antitumor activity of engineered cell therapies, including chimeric antigen receptor (CAR)-T cells¹⁰⁹.

Probiotics further bridge innate and adaptive immunity by activating antigen-presenting cells. For example, the membrane protein, LIPOAF, from Alistipes fingoldii activates DCs via the TLR-2-NF-κB pathway, stimulating CXCL16 secretion and promoting CD8⁺ T-cell tumor infiltration¹⁰⁴. Similarly, outer membrane vesicles from A. muciniphila induce DC maturation, leading to more robust cytotoxic T-cell priming¹⁰⁵. At an ecologic level, probiotics can durably reshape the gut microbial community to sustain a pro-inflammatory, immune-active state. Synergistic effects are observed with combination approaches. Co-administration of C. butyricum and A. muciniphila more effectively restores a favorable microbiota composition, reduces systemic inflammation, and sensitizes tumors to anti-PD-L1 therapy than either probiotic alone¹⁰⁶.

Together, probiotics enhance ICI efficacy in CRC through the following multifaceted network of mechanisms: direct tumor targeting; metabolic reprogramming; innate immune activation; and ecologic restoration. This coordinated action helps convert immunologically “cold” tumors into “hot” tumors, offering a viable strategy to break immunotherapy resistance.

Modulating microbiomes to overcome chemoresistance

Chemotherapy, using agents, such as 5-fluorouracil (5-FU), oxaliplatin, and irinotecan, remains a cornerstone of CRC treatment. However, the long-term success is frequently compromised by intrinsic or acquired chemoresistance. Emerging research has established the gut microbiome as a critical modulator of therapeutic efficacy, influencing resistance through diverse mechanisms (Table 4). Key insights have been derived from model systems and clinical observations. Studies involving Caenorhabditis elegans fed a bacterial diet revealed that microbial genetics directly alter host responses to chemotherapeutics¹¹⁴. For example, bacterial nucleotide metabolism genes have been shown to widely modulate the efficacy of drugs, like 5-FU, with cytotoxicity surprisingly mediated through bacterial ribonucleotide metabolism rather than solely through canonical host pathways¹¹⁴. An intact microbiota is essential for optimal therapy in mammalian systems. Depletion of commensal bacteria in mouse models impairs the response to both platinum-based chemotherapy and immunotherapy¹¹⁵. This finding is linked to dysfunctional tumor-infiltrating myeloid cells, which exhibit reduced cytokine production and cytotoxic activity. Clinical data further underscore this relationship, revealing significant differences in gut microbial composition between chemotherapy responders and non-responders. Specific pathogens, such as F. nucleatum and enterotoxigenic B. fragilis (ETBF), have been directly implicated in promoting chemoresistance^41,42. Targeted modulation of the microbiome presents a promising therapeutic strategy. Phage-based targeting of ETBF successfully reverses chemoresistance in murine CRC models⁴². Furthermore, the administration of beneficial microbes, such as B. bifidum, in combination with 5-FU has been shown to enhance therapeutic efficacy significantly compared to chemotherapy alone¹¹⁰. Collectively, these results indicate that modulation of the gut microbiota represents a potential strategy to improve chemotherapy outcomes.

Despite promising data, few large-scale clinical trials have directly validated the ability of probiotics to enhance the anticancer efficacy of chemotherapy in humans, as most evidence still relies on animal models. Moreover, current research largely focuses on mitigating adverse effects rather than directly enhancing anticancer efficacy. Oral probiotics significantly reduced the risk and severity of chemotherapy-induced diarrhea and oral mucositis in a meta-analysis of 12 randomized controlled trials (1,013 patients) with benefits notably observed in Asian patients and linked to greater bacterial species diversity¹¹⁶. Another meta-analysis involving 8 studies (633 patients) demonstrated that probiotics significantly reduced chemoradiotherapy-induced diarrhea and improved pain, dyspnea, and insomnia¹¹⁷. Moreover, supplementation with L. rhamnosus GG or Lactobacillus casei (L. casei) has been shown to reduce chemotherapy-induced intestinal mucosal injury and diarrhea^111,112. Similarly, probiotic mixtures containing Lactobacillus and Bifidobacterium spp. have been reported to alleviate severe diarrhea associated with irinotecan treatment¹¹³. The protective mechanisms likely involve probiotic-derived metabolites that help strengthen the intestinal barrier and restore microbial homeostasis. Thus, clinical evidence strongly supports the use of probiotics for maintaining intestinal integrity and improving patient tolerance to chemotherapy. However, whether probiotics can directly potentiate chemotherapy and overcome tumor resistance in patients remains an open question. The molecular and immunologic mechanisms underlying such potential synergy are not well-defined and warrant focused investigation. It is worth noting that chemotherapy-induced neutropenia could substantially increase systemic infections and the risk of sepsis caused by microbial pathogens. Under such conditions, even bacteria traditionally regarded as probiotics, such as Bifidobacterium and Lactobacillus, may cause opportunistic infections in immunocompromised hosts¹¹⁸. Future efforts should therefore prioritize well-designed prospective studies to define risk stratification for live microbial therapies under varying degrees of chemotherapy intensity and neutropenia, enabling the development of clearer and safer clinical guidelines.

Dietary and lifestyle modulation of the microbiome

Unhealthy lifestyle patterns are a major source of gut microbial dysbiosis and chronic, low-grade inflammation are key drivers of CRC pathogenesis. Factors, such as high-fat diets¹¹⁹, smoking¹²⁰, and disrupted circadian rhythms¹²¹ can drive the enrichment of pathogenic bacteria, undermine intestinal barrier function, and establish an immunosuppressive tumor niche, which collectively accelerate disease progression. Conversely, positive lifestyle modifications can restore microbial homeostasis and support cancer therapy¹²². For example, regular physical activity is associated with a reduced risk of CRC, in part by enriching SCFA-producing bacteria and enhancing overall microbial diversity¹²³.

A meta-analysis of over 2.2 million individuals showed that adherence to the Mediterranean diet is associated with a 16% reduction in CRC risk¹²⁴. Similarly, high-fiber diets have shown a consistent inverse association with CRC incidence across multiple international studies¹²⁵. These protective effects are primarily attributed to increased dietary fiber intake, which promotes the production of SCFAs, such as butyrate, by the gut microbiota, thereby exerting anti-inflammatory and antitumor effects¹²⁶. Thus, targeted dietary strategies offer a more direct route to modulate the microbiome for therapeutic benefit. A ketogenic diet (KD) can suppress CRC tumorigenesis by shifting microbial communities to increase the production of stearic acid¹²⁷. This effect can be mimicked by the ketone body, β-hydroxybutyrate (BHB), which activates the receptor Hcar2 on colonic cells, induces Hopx expression, and subsequently inhibits crypt proliferation and tumor growth¹²⁸. Similarly, fasting-mimicking diets (FMDs) have emerged as a potent adjunct to immunotherapy^108,129. Combining FMD with anti-PD-1 therapy has been shown to inhibit CRC progression in preclinical models by increasing Lactobacillus abundance, reducing proliferation and angiogenesis, and boosting intratumoral CD8⁺ T-cell infiltration¹²⁹. Another mechanism involves FMD-induced enrichment of B. pseudolongum, which enhances the efficacy of anti-CTLA-4 therapy by producing the metabolite, L-arginine, and promoting tissue-resident memory CD8⁺ T cells¹⁰⁸. These findings underscore that dietary and lifestyle interventions are powerful tools for reshaping the gut microbiome (Table 5).

Carcinogens in food pose a significant health risk. Lactic acid bacteria (LAB), probiotic microorganisms that ferment lactose into lactate, have gained attention as food additives capable of counteracting carcinogenic compounds. Some LAB strains do so through direct physical binding, thereby avoiding the re-pollution associated with toxic degradation byproducts¹³¹. This binding primarily involves cell wall components with peptidoglycan serving as the main site and is strain-specific. For example, Lactobacillus and Bifidobacterium spp. show potential against common carcinogens, such as mycotoxins, polycyclic aromatic hydrocarbons, heterocyclic amines, and phthalic acid esters¹³². However, it is noteworthy that the gut microbiota can also contribute to metabolic activation of some pro-carcinogens. Roje et al.¹³³ reported that the gut microbiota can metabolize environmental nitrosamine carcinogens, such as N-butyl-N-(4-hydroxybutyl)-nitrosamine and its derivatives, via bacterial β-glucuronidase-mediated deconjugation and oxidation, converting environmental nitrosamine carcinogens into more toxic metabolites. These metabolites accumulate systemically, reach distal organs, such as the urinary bladder, and drive tumor formation. This effect is causally linked to specific bacterial strains (e.g., Lactobacillus), varies across human donor microbiomes, and extends to related nitrosamines¹³³. Together, the interactions between the gut microbiota and dietary/environmental carcinogens are complex, underscoring the importance of developing novel microbiome-targeted prophylactic approaches.

Reshaping gut microbiota dysbiosis modulated by host genetics and gene expression

The composition of the gut microbiota is shaped by a complex interplay of exogenous factors and intrinsic host genetics. Integrated analyses using metagenomic sequencing of fecal samples with host genomic and transcriptomic profiling have established significant associations between specific microbial communities, host genetic variants, and gene expression patterns^134,135. These genetic influences can define an individual’s microbial landscape with direct consequences for health and disease. Specific host genotypes are powerful drivers of microbial community structure. For example, CRCs harboring a KRAS p.G12D mutation exhibit a distinct microbial niche enriched with F. nucleatum and B. fragilis, a combination that promotes tumorigenesis in murine models^136,137. Furthermore, a host single nucleotide polymorphism (SNP rs2355016) within the KCNJ11 gene correlates strongly with the abundance of F. nucleatum¹³⁸. Carriers of the risk allele demonstrate reduced KCNJ11 expression and enhanced mucosal adhesion of F. nucleatum, accelerating CRC progression¹³⁸. In addition to oncogenes, genes regulating host defense also shape microbiota. Systemic or intestinal epithelial knockout of Mettl9 impairs secretory functions, leading to pathologic over-colonization by Candida albicans, barrier disruption, and increased mortality¹³⁹. Depending on the host context, inflammatory status, the presence of live vs. non-viable components, and strain specificity, A. muciniphila can exert protective or pathogenic effects^140,141. Loss of group 3 innate lymphoid cells (ILC3s) triggers IL-22-dependent intestinal galactosylation, fostering the expansion of A. muciniphila within the mucus layer. This effect disrupts microbial equilibrium and heightens susceptibility to S. typhimurium infection¹⁴². Similarly, functional loss of the Fut2 gene, which governs epithelial fucosylation, compromises gut homeostasis. Notably, fecal microbiota transfer from Fut2-deficient mice or inoculation with poorly fucosylated B. fragilis is sufficient to confer colitis susceptibility in recipients¹⁴³. Such genetically influenced dysbiosis is a key driver in diseases, like early-onset CRC. This finding provides a compelling rationale for combining probiotics with genotype-targeted therapies. In KRAS-mutant settings, the probiotic, Parabacteroides distasonis, competitively excludes pro-tumorigenic F. nucleatum, and oral administration attenuates CRC progression in mice¹³⁶. Advanced delivery systems are enhancing precision. Extracellular vesicles (EVs) from L. acidophilus loaded with TNF-α siRNA, enable colon-specific gene silencing, reshaping the microbiota and metabolite profile to restore homeostasis in IBD models¹⁴⁴. In another approach, engineered Pediococcus pentosaceus was developed as a safe lactic acid bacterium delivery vehicle expressing the secreted anticancer protein, P8, which has been reported to downregulate cyclin B1 and Cdk1¹⁴⁵. Oral administration of P8 suppresses tumor development in murine CRC xenografts and colitis-associated CRC and restores gut microbiota diversity¹⁴⁶. Collectively, these strategies underscore the significant translational potential of precisely modulating the microbiota based on host genetic context to improve therapeutic outcomes (Table 6). However, the genotype-specific mechanisms governing host-microbe interactions in CRC require deeper characterization. Furthermore, the combined application of probiotics and gene-targeted agents demands rigorous optimization of dosing, safety, and delivery platforms to realize the full clinical promise.