Gut microbiota-mediated gut-liver axis: a breakthrough point for understanding and treating liver cancer

Origin of gut dysbiosis in CLDs

Large population-based studies have shown that the composition of human gut microbiota varies among individuals and is predominantly influenced by environmental factors [8]. Diet is considered a dominant determinant of interindividual microbiome variation. Certain dietary patterns can modulate the composition and metabolism of gut microbiota, which in turn affect human health (Fig. 1A) [9]. Preclinical studies have demonstrated that a high fat diet (HFD) induced dysbiosis, which impacts the lipid metabolism in the liver and promotes the development of MASLD [10]. An increase in LPS-producing bacteria, such as Enterobacteriaceae and Desulfovibrionaceae and a decrease in the production of short-chain fatty acid (SCFA)-producing bacteria including Bacteroides, Ruminococcaceae and Lachnospiriaceae were commonly observed in HFD-induced MASLD mice [11]. Moreover, a western diet characterized by low-choline, high fat and high sugar leads to an increased abundance of Blautia producta and elevated level of its outcome metabolites 2-oleoylglycerol, thereby contributing to the pathogenesis of metabolic dysfunction-associated steatohepatitis (MASH) [12]. Others like high dietary fructose or cholesterol were also shown to alter the gut microbiota, whereby they were implicated in the development of MASLD and HCC [13-15]. Alterations in several Clostridium species and disrupted bile acid metabolism were identified as key mediators in fructose-induced MASLD [13]. In mice fed a high-fat/high-cholesterol diet, Bifidobacterium and Bacteroides were reduced, a finding also corroborated in hypercholesterolemia patients [15]. Additionally, alcoholic consumption has a profound impact on gut microbiota, potentiating the pathogenesis of alcohol-associated liver diseases [16]. Intestinal microbiota dysbiosis and bacterial overgrowth have been observed in both acute and chronic alcohol exposure [16,17]. Loss of butyrate-producing bacteria belonging to Lachnospiraceae, Ruminococcaceae, and Clostridiaceae families, along with downregulation of butyrate synthesis pathways, has been suggested as key pathogenic features of ethanol-induced dysbiosis [18]. A recent study has corroborated that ethanol-induced gut microbiota alterations are a result of elevated acetate levels, rather than direct ethanol metabolism by gut microbiota [19]. Collectively, despite the demonstration of diet's ability to remodel gut microbiota in preclinical animal models, the significance of this cause-and-effect relationship in human and the extent to which it contributes to the development and progression of liver disease remain to be investigated.

Figure 1.

Origin and outcomes of gut dysbiosis in the gut-liver axis. (A) Environmental factors such as the Western diet, alcohol consumption, drugs, and genetic predispositions are primary contributors to gut dysbiosis in the context of liver diseases. (B) This dysbiosis disrupts microbial metabolism, leading to intestinal barrier dysfunction and subsequent translocation of microorganisms and microbial products to the liver. These events typically occur early in the pathogenic process.

Antibiotics are known to have a dramatic impact on the gut microbiota communities due to their direct antibacterial activity. Long-term, low-dose penicillin exposure, particularly during early life, has been shown to exacerbate metabolic disturbances and hepatic steatosis in HFD-fed mice by inducing gut dysbiosis, characterized by a significant reduction in genera Lactobacillus, Allobaculum, Rikenellaceae, and Candidatus Arthromitus [20,21]. A human study also demonstrated that prior antibiotic use, including fluoroquinolones, cephalosporins, macrolides, penicillins and tetracyclines, was associated with an increased risk of future MASLD, with the highest risk linked to fluoroquinolones [22]. Furthermore, antibiotic exposure around the time of treatment initiation was found to correlate with a worse prognosis in HCC patients receiving targeted or immune-based therapy, suggesting impaired treatment efficacy preceded by antibiotic-induced dysbiosis [23]. In addition to the effects of antibiotics, the long-term consumption of several non-antibiotic drugs like proton pump inhibitors (PPI), also influences the gut microbiome function and composition (Fig. 1A), consequently leading to hepatic inflammation in the liver [24]. Consumption of PPI can cause an oralization of the gut microbiome due to the reduction of gastric acidity [24]. It has been demonstrated that expansion of intestinal Enterococcus faecalis induced by PPI promoted progression of alcoholic liver disease (ALD) and MASLD in preclinical models [25]. Prospective cohort studies have also observed the positive correlation between use of PPI and readmission risk in patients with cirrhosis [26].

While host genetics are known to regulate the gut microbiome, as demonstrated in general population studies, there is currently a lack of research specifically investigating the influence of host genetics on gut microbiome in patients with liver diseases (Fig. 1A) [27,28]. Most of the present evidence comes from experimental mouse models. E74-like ETS transcription factor 4 (Elf4) was identified as a host protective gene that maintains the integrity of intestinal barrier and gut homeostasis. Ablation of Elf4 in mice induced gut dysbiosis characterized by an increase in Helicobacter and Escherichia genera within the Proteobacteria phylum and exaggerated alcohol-induced hepatic steatosis and liver injury [29]. Another host gene, G protein-coupled receptor 35 (Gpr35), was also shown to modulate hepatic steatosis by regulating gut microbial ecology [30]. Expansion of pro-inflammatory bacteria from Bacteroides and Ruminococcus genera and reduction in probiotics such as Akkermansia muciniphila were observed in Gpr35^KO mice. Additionally, hepatocyte-specific deletion of myeloid differentiation primary-response gene 88 (Myd88) induced enrichment of specific bacterial genera including Sutterella, Allobaculum, Ruminococcus, and Oscillospira in mice fed with normal diet [31]. This effect may be related to the alterations of hepatic bile acid metabolism induced by Myd88 deficiency, suggesting the potential liver-to-gut communication [31]. In addition to protective genes, a gene encoding miRNA-21 was found to promote liver injury in cholestasis by modulating small intestinal Lactobacillus in mice [32].

Conversely, conditions such as liver cirrhosis-induced small bowel dysmotility, reduced bile acids and impaired intestinal immunity also contribute to gut dysbiosis, indicating a reciprocal effect in the advanced stages of liver disease.

Aberrant gut-liver communicating axis owing to dysbiosis

Among the intestinal microorganisms, the bacteria take the main responsibility for microbial metabolic processes, encompassing the transformation of bile acids, SCFAs via fermentation of indigestible carbohydrates, and the synthesis of essential amino acids [33]. These microbial metabolites serve as essential messengers and regulators of the gut-liver communicating axis, maintaining the gut homeostasis in situ and influencing the distal liver function after translocating to the liver through the portal vein [1,33]. A direct consequence of disturbances in the microbial composition community is the alteration of microbiome-derived or microbiome-modified metabolites, as evidenced by fecal metabolomics studies in patients with various CLDs [34-38]. Disruption of metabolic processes within the microbiome can impair the integrity of intestinal epithelial, immune and vascular barriers (Fig. 1B and Table 1), which is a prerequisite for microbiota-associated hepatic pathogenesis [39]. Unconjugated bile acids such as deoxycholic acid (DCA) and chenodeoxycholic acid (CDCA) have been shown to increase intestinal permeability via epithelial growth factor receptor activation [40], while conjugated bile acids prevent the unconjugated bile acids-mediated intestinal and hepatic damage by forming micelles to sequester them [41]. SCFAs like propionate and butyrate have been demonstrated to protect the integrity of intestinal mucus and epithelial barriers [42-44]. Butyrate, in particular, enhances the tight junction of intestinal epithelial cells by promoting epithelial O₂ consumption and stabilizing hypoxiainducible factor, a transcription factor that coordinates barrier protection [43]. Microbial tryptophan metabolites, especially indoles, which are reduced in patients with alcoholic hepatitis also play a crucial role in maintaining the intestinal immune and epithelial barrier by activating the aryl hydrocarbon receptor (AhR) signaling pathway [45,46]. Indole-3-acetic acid, for instance, can enhance the antibacterial function of the intestinal mucus layer by stimulating interleukin-22 (IL-22) production in type 3 innate lymphoid cells (ILC3s), thereby inducing the expression of antibacterial proteins such as regenerating islet-derived 3 gamma [45]. In addition to mucus layer and epithelial barrier, a diet induced-dysbiosis also leads to the disruption of gut vascular barrier (GVB) and the farnesoid X receptor (FXR) agonist obeticholic acid protects against the GVB damage by driving β-catenin activation in endothelial cells, although by which microbial metabolites this is mediated is still unascertained [39].

Table 1.

Microbial metabolites involved in the gut-liver communicating axis during the pathogenesis of CLDs

Beyond the intestine, specific microbial metabolites also directly influence hepatic lipid metabolism and inflammation through interacting with hepatocytes and macrophages after they are absorbed by the gut and transported to the liver (Table 1) [40-54]. Propionate can mitigate alcohol-induced liver injury by activating major urinary protein 1 and suppressing endoplasmic reticulum stress in hepatocytes [47]. Indoles and tryptamine inhibit proinflammatory activation and migration of macrophages through direct activation of the AhR signaling pathway. Additionally, N,N,N-trimethyl-5-aminovaleric acid, a microbial metabolite derived from trimethyllysine, has been shown to reduce carnitine synthesis and decrease fatty acid oxidation in hepatocytes, thereby promoting hepatic steatosis.

Gut barrier dysfunction facilitates the translocation of microorganisms and microbial products, such as lipopolysaccharide (LPS) and exotoxins, into the liver through the portal circulation, thereby triggering a continuous inflammatory response in the liver [55]. Human studies have demonstrated that patients with various liver diseases, including MASLD, ALD, cirrhosis, and PSC, present increased intestinal permeability and concomitant increase in endotoxin levels in both the serum and the liver [5]. LPS were found to be more frequently present in the liver of patients with MASLD compared to those without MASLD, and these LPS particles predominantly localize in the portal tracts, suggesting an intestinal origin [56]. These microbial components can activate pathogen recognition receptors (PRR) such as Toll-like receptors (TLRs) on Kupffer cells in the liver, resulting in increased hepatic inflammation, hepatocyte damage and liver fibrosis [57]. Administration of dextran sodium sulfate, which is known to compromise the integrity of gut barrier, has been shown to facilitate hepatic tumorigenesis in both choline-deficient high-fat diet MASH mouse model [58] and diethylnitrosamine (DEN)-induced HCC model [59]. Moreover, E. faecalis, that is abundant in human CLD and HCC feces, could increase gut permeability conferred by GelE, a metallopeptidase, and thereby promote the formation of HCC in collaboration with other bacteria in a TLR4-dependent manner [6].

In addition to microbial products, the intrahepatic microbiota has recently received increasing attention, with respect to its physiologic roles and pathophysiologic contributions to CLDs [60]. The liver microbiome, which is suggested to originate from the gut, exhibits distinct characteristics from the gut microbiome, possibly as a result of the selective filtering by the gut-liver barrier [61]. The bacterial signature within the liver has been identified in patients suffering from MASLD and showed diverse patterns associated with obesity [56]. Findings from the preclinical models have corroborated that translocated pathobionts such as E. faecalis on the premise of impaired gut barrier could exacerbate ALD via secretion of cytolysin that is toxic to hepatocytes, and/or activation of TLR2 on Kupffer cells [25,62].

It is extensively accepted that gut dysbiosis predisposes individuals to liver diseases. However, hepatic abnormalities can also have a reciprocal effect on the gut microbiome (Fig. 1), creating a cyclical relationship where it seems difficult to discern which factor comes first–much like the classic conundrum of the chicken and the egg. Indeed, diet or alcohol-induced gut dysbiosis can determine at least partially the individual’s susceptibility to MASLD and ALD, which tends to be an early event in the pathogenic process [10,63,64]. Nevertheless, as diseases progress to advanced stages like cirrhosis, they can in turn lead to changes in the microbiota composition due to reduced small bowel motility, abnormal bile acid levels and impaired intestinal immunity [65]. The biliary system serves as the central mediator of liver-to-gut communication by transporting bile, a solution rich in bile acids, immunoglobulin A (IgA) and other antimicrobial molecules, to the intestinal lumen. Liver cirrhosis is characterized by impaired bile transport and reduced bile acid levels in the gut, resulting from extensive destruction of liver tissue structure and hepatocyte dysfunction [66]. Bile acids, particularly unconjugated bile acids, can directly disrupt bacterial membranes due to their detergent properties [67,68]. Notably, different gut microorganisms exhibit varying sensitivity to bile acid toxicity [68]. Gram-negative bacteria tend to be more resistant than Gram-positive bacteria, possibly due to their differences in cell wall structure [68]. Some probiotic bacteria, such as Lactobacillus and Bifidobacterium, also show greater resistance, which may be linked to the activation of glycolysis [68]. In addition to their direct bactericidal effects, bile acids also stimulate the production of antimicrobial molecules such as angiogenin, inducible NO synthase and IL-18 in the intestine through activation of FXR [69]. These mechanisms may collaboratively contribute to the connection between disrupted bile acid metabolism and microbial dysbiosis in cirrhosis. Additionally, diminished production of antimicrobial peptides by Paneth cells, such as α-defensins 5 and 7, was found in experimental decompensated cirrhosis rats and contributed to intestinal bacterial overgrowth and translocation [70]. Others like IgA antibodies in the gut lumen were decreased in cirrhotic patients, suggesting a reduction of antimicrobial immunity [71].

Hepatocellular carcinoma

Clinical studies have delineated the dynamics of gut microbiota during the progression from CLDs to HCC, shedding light on potential anti-tumorigenic probiotics and protumorigenic pathobionts (Fig. 2 and Table 2) [6,7,83-105]. In MASLD patients with HCC, the stool microbiota exhibited an increased abundance of SCFA-producing bacteria such as Bacteroides caecimuris and Veillonella parvula, and a corresponding functional enrichment in SCFA synthesis compared to those with MASLD-cirrhosis. In this study, the MASLD-HCC microbiota was confirmed to elicit an immunosuppressive phenotype in the peripheral blood from non-MASLD controls, which is specific to HCC rather than cirrhosis [7]. In addition, an integrated analysis of gut microbiome and tumor transcriptome has revealed a correlation of intestinal microbiota dysbiosis with the tumor immune microenvironment and bile acids metabolism in patients with hepatitis B virus (HBV)-related HCC, indicating potential mechanistic relationships [83]. Another study has characterized the gut microbial profiles in patients with HCV-related HCC, revealing a notable rise in the abundance of E. faecalis and several species belonging to genera Streptococcus and Lactobacillus [6]. This study further validated that the colonization of E. faecalis, linked to decreased intestinal concentrations of DCA, could promote the liver carcinogenesis in mice [6]. Additionally, studies have also compared the differences in intestinal bacteria composition between viral and non-viral HCC, consistently uncovering an increase in Faecalibacterium in viral HCC [84,85]. Numerous studies have examined the compositional shifts of gut microbiota during cirrhosis-associated hepatocarcinogenesis (Table 2). However, the taxonomic changes in gut microbiota were not aligned across these investigations, potentially due to the diverse etiologies, varying study designs, and differences in confounder controls. Notably, cirrhotic patients with HCC also exhibited a gut fungal dysbiosis characterized by an increase in opportunistic pathogenic fungi such as Malassezia sp. and Candida albicans [86]. Characterization of gut microbial profiles in prospectively recruited patients with cirrhosis has revealed a correlation between specific microbes and future HCC risk, such as Alloprevotella, whose reduction predicts a higher risk for developing HCC [87]. Intestinal permeability was also increased in patients with cirrhosis and HCC, as suggested by elevated levels of colonic tight junction protein zonulin-1 and microbial components LPS in the serum, as well as calprotectin, a surrogate marker of intestinal inflammation, in the feces [88]. Overall, these clinical studies substantiate the correlation between the aberrant gut-liver axis and the development of HCC, highlighting selective expansion or depletion of specific microorganisms during hepatocarcinogenesis.

Figure 2.

Gut microbiome evolves with liver carcinogenesis from different etiologies. Hepatocellular carcinoma developing from different etiologies exhibits diverse fecal microbiome profiles. Similarly, iCCA is associated with specific changes in gut bacteria and metabolites. Blue boxes represent bacterial genus, yellow boxes species, red boxes fungi and green boxes metabolites. CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; GUDCA, glycoursodeoxycholic acid; HCC, hepatocellular carcinoma; iCCA, intrahepatic cholangiocarcinoma; MASLD, metabolic dysfunction-associated steatotic liver disease; PSC, primary sclerosing cholangitis; SCFAs, short-chain fatty acids; TUDCA, tauroursodeoxycholic acid.

Table 2.

Investigations into the gut microbiota composition and function in HCC and CCA patients

Experimental studies have confirmed the casual relationship between pre-existing gut microbiota dysbiosis in CLDs and carcinogenesis within the liver. Transplantation of fecal microbiota from patients with HCV-related CLD or HCC, but not from healthy individuals, into microbiota-depleted mice could increase the intestinal permeability and promote tumorigenesis induced by DEN and carbon tetrachloride (CCl4) administration [6]. This carcinogenic effect was also verified in a streptozotocin and HFD-induced MASH mouse model [6]. In addition, the pro-tumorigenic role of gut microbiota has been confirmed in other spontaneous HCC models, such as models induced by high fat and high cholesterol (HFHC) diet, combination of HFD with 7,12-dimethylbenz [a]anthracene (DMBA), and several genetically engineered mouse models (Table 3) [15,57-59,106,109-114]. The involvement of intestinal microbiota in hepatocarcinogenesis was initially observed in DEN and CCl4-induced HCC mouse models, in which LPS derived from the intestinal microbiota collaborated with the chemical carcinogen-induced chronic liver injury to promote HCC formation by activating TLR4 pathway [57]. Gut sterilization efficiently prevents this malignant progression [57]. Furthermore, depletion of gut commensal microbiota using an antibiotic cocktail, consisting of vancomycin, neomycin and primaxin, also showed a protective effect on HCC development in MYC transgenic mice [107]. Mechanistically, gut microbiota was demonstrated to suppress liver antitumor immunity via modulating NKT cells accumulation through regulating primary-to-secondary bile acid conversion, in which Clostridium sp. were suggested to play key roles [107]. These results indicated that the gut microbiota, when considered as an integral unit, may constitutively shape the hepatic microenvironment and therefore amplify carcinogenic effects in combination with other carcinogens. Experimental studies also verified the hypothesis that the gut microbiota determines the susceptibility to HCC. It is demonstrated that a dysbiotic gut microbiota induced by an inulin-containing diet (ICD) predispose TLR5-deficient mice to HCC [106]. Moreover, supplementing this with a high fat diet, known for inducing dysbiosis, further exacerbated the ICD-induced formation of HCC. A dysregulated microbial metabolism of fermentable fiber, leading to the subsequent overproduction of SCFAs such as butyrate, was confirmed to mediate this process [106]. Vancomycin treatment prevented this carcinogenic effect by depleting the SCFAs and major secondary bile acid-producing bacteria [108].

Table 3.

Mouse models for HCC and CCA involving the gut microbiota

As microbiome sequencing technologies advance, recent studies have been increasingly dedicated to precisely pinpointing specific microorganisms and metabolites that are crucial in either promoting or impeding liver carcinogenesis. The probiotics A. muciniphila and Bifidobacterium pseudolongum, whose reduction has been consistently observed in many human studies [36,77,88,95,115], were confirmed to protect against MASLD-HCC [109,110]. A. muciniphila suppressed HCC formation in the STAM model by promoting the accumulation of NKT cells [110]. Supplementation of B. pseudolongum restored a healthy gut microbiome composition, improved gut barrier function and prevented the development of HCC in a MASLD-HCC mouse model induced by DEN and HFHC diet [109]. Acetate generated from B. pseudolongum was identified as the functional mediator. It can not only restore gut homeostasis in situ, but also bind to GPR43 on hepatocytes via the gut-liver axis and inhibit the oncogenic IL-6/JAK1/STAT3 signaling pathway, thereby suppressing cell proliferation and inducing apoptosis [109]. Another SCFA, valeric acid generated by lactobacillus acidophilus, also exhibits anti-tumorigenic effects through binding to GPR41/43 and inhibiting oncogenic Rho-GTPase signaling pathway [116]. D-lactate, a microbiome small-molecule metabolite, has recently been identified to promote the transformation of M2 tumor-associated macrophages to M1 via inhibition of PI3K/Akt pathway and activation of NF-_κB pathway, and remodel immunosuppressive microenvironment for HCC [117]. Additionally, gut microbiota-associated bile acids also profoundly affect the development of HCC. Primary bile acids such as CDCA and taurocholic acid facilitated NKT cells accumulation in the liver, while secondary bile acids lithocholic acid (LCA) or ω-muricholic acid showed an opposite effect, and therefore suppressed the hepatic antitumor immunity [107]. Moreover, increased taurocholic acid and depleted indole-3-propionic acid were found in excess dietary cholesterol-induced MASLD-HCC mouse model, functionally aggravating hepatic lipid accumulation and promoting hepatocyte proliferation [15]. In a DMBA and HFD-induced obesity-associated HCC mice model, an increase in DCA resulted from gut dysbiosis provoked the senescence-associated secretory phenotype (SASP) in hepatic stellate cells (HSCs), which in turn secreted multiple tumor-promoting factors and eventuated in the promotion of HCC [111].

Cholangiocarcinoma

CCA is known to be associated with liver fluke infection and chronic biliary diseases such as choledocholithiasis and PSC [3]. Liver cirrhosis, viral hepatitis, alcohol consumption and MASLD also represent strong risk factors for CCA, especially intrahepatic CCA. Studies on gut microbiota have provided new insights into the pathogenesis of CCA. Compositional alterations of gut microbiome have been observed in patients with CCA (Table 2). Changes of specific genera in CCA such as increase in Muribaculacea [98,105] and Klebsiella [98,104], and decrease in Ruminococcus [103,105] were concordantly seen in different studies. Examination of stool microbiota and serum bile acids in patients with intrahepatic CCA revealed a positive correlation between the genera Lactobacillus and Alloscardovia and the plasma-stool ratio of tauroursodeoxycholic acid, indicating potential interacting mechanisms [102]. The gut leakiness preceded by microbial dysbiosis and subsequent translocation of microbial components and microorganisms was also shown to facilitate the development of CCA in a bile duct ligation-induced PSC mouse model [114]. Although solid evidence of gut dysbiosis in CCA has been attained in clinical studies and the role of aberrant gut-liver axis in cholangiocarcinogenesis was also verified in preclinical models, the processes leading to gut dysbiosis, the distinctions among CCA subtypes, and the mechanisms underlying microbiota-associated malignant transformation of cholangiocytes still remain largely unknown and need to be further explored.

Triggering persistent inflammation

Under physiological conditions, the liver continuously receives moderate inflammatory stimuli from gut-derived microbial pathogens, achieving a balance between immune elimination and immunotolerance [119]. However, prolonged exposure to pathogen-associated molecular patterns (PAMPs), microbial exotoxins, and live microorganisms as a result of a compromised intestinal barrier, disturbs this delicate immunological balance in the liver, leading to tissue damage and sustained inflammation that promotes carcinogenesis (Fig. 3A). Cytolysin, a microbial exotoxin secreted by E. faecalis, has been identified as a direct cause of hepatocyte death and liver injury during the progression of ALD [62]. Furthermore, accumulation of PAMPs activates multiple pro-inflammatory signaling pathways via PRRs, resulting in the production of tumor-promoting cytokines and the malignant transformation of hepatic parenchymal cells. LPS, one of the most well-known PAMPs, was shown to activate TLR4 signaling in multiple hepatic cells, including Kupffer cells, HSCs, endothelial cells and hepatocytes [120]. It has been demonstrated that activation of LPS-TLR4 pathway in Kupffer cells enhanced the production of TNF and IL-6, which are known to be crucial mediators in hepatic carcinogenesis [121,122]. Moreover, TLR4 activation also promotes hepatocytes proliferation by upregulating expression of epiregulin in HSCs and prevents cell apoptosis through NF-_κB signaling [57]. Another influential PAMP, LTA from Gram-positive bacteria, was shown to enhance the SASP of HSCs in collaboration with DCA-induced DNA damage through activation of TLR2 [111,123]. Consequentially, pro-inflammatory cytokines and chemokines such as IL-1β, Groα, and IL-6, termed SASP factors, were secreted by senescent HSCs and contributed to hepatocarcinogenesis [123]. In addition to TLR, nucleotide-binding oligomerization domain 2, an intracellular PRR that recognizes muramyl dipeptide from the gut bacteria, has also been demonstrated to promote hepatic tumorigenesis via receptor-interacting-serine/threonine-protein kinase 2-mediated pro-inflammatory response and nuclear autophagy-mediated DNA damage in hepatocytes [124]. Intriguingly, intestinal inflammatory cytokines owing to dysbiosis may also participate in the development of HCC. IL-25, generated from colonic epithelial tuft cells, was found to stimulate hepatic M2 macrophages polarization and secretion of CXCL10, which fostered the epithelial-to-mesenchymal transition of HCC cells [125]. Additionally, it is noteworthy that the intrahepatic colonization of bacteria such as Stenotrophomonas maltophilia also functions as an oncogenic factor by promoting hepatic fibrosis and senescent HSC-associated inflammation through activating TLR4 [126]. However, the relationship between this process and bacterial translocation resulting from gut permeability was not investigated in this study.

Promoting metabolic reprogramming

Metabolic rearrangement is a critical feature during the development of liver cancers, particularly on a MASLD background, where metabolic dysfunction resulting from long-term dietary challenges leads to the sequential progression of steatosis, fibrosis, and ultimately HCC [127,128]. Gut microbiota is involved in this process by promoting de novo lipogenesis and glutamine synthesis under stimulation of high dietary fructose (Fig. 3B) [14,129]. Excessive fructose intake causes gut dysbiosis and intestinal barrier deterioration in mice, resulting in the translocation of endotoxins such as LPS to the liver [129]. In addition to the aforementioned proinflammatory effects of LPS-TLR4 pathway, the production of TNF triggered by TLR4 signaling also induces de novo lipogenesis in hepatocytes [130]. Mechanistically, TNF signaling activates sterol regulatory element-binding transcription factor 1 (SREBP1), a key regulator of lipogenic genes, by upregulating the expression of caspase-2, which triggers proteolytic activation of SREBP1 by cleaving site 1 protease [130]. This de novo lipogenesis promoted by fructose-associated inflammation facilitated the progression from MASLD to MASH and the spontaneous development of HCC [129]. Furthermore, gut microbiota-mediated conversion from fructose to acetate has recently been demonstrated to promote HCC progression by influencing glutamine synthesis [14]. Microbiota-derived acetate, as an additional carbon source, supports glutamine synthesis via the tricarboxylic acid cycle and subsequently upregulates uridine diphospho-N-acetylglucosamine levels through the hexosamine biosynthetic pathway, ultimately enhancing O-GlcNAcylation of downstream proteins and promoting HCC cell proliferation [14]. Taking into consideration the intricate metabolic networks and numerous dysregulated metabolic pathways in liver cancers, investigations into the role of gut microbiota and microbial metabo-lites in reprogramming the tumor metabolic microenvironment remain relatively limited. Future studies are anticipated to profile the relationship between the gut microbiota and the tumor metabolic microenvironment in the liver, as well as to unravel the underlying interactional mechanisms.

Targeting the gut microbiota for liver cancer prevention

On the basis of ample preclinical evidence and the emergence of innovative therapeutic strategies, significant advancements have been made in the clinical translation of manipulating the gut microbiota for treating liver diseases, particularly MASLD and cirrhosis [78]. Therapeutic approaches including probiotic supplementation, rifaximin, and FMT from healthy donors have shown success in ameliorating hepatic steatosis in MASLD and reducing complications and hospitalizations among cirrhotic patients, as demonstrated by numerous human clinical trials [78]. While direct clinical trials investigating the preventive effects of microbiota-based therapies on the development of HCC are yet to be completed (Table 4), improvements in microbial composition, intestinal permeability, and anti-inflammatory effects, all pivotal factors in the malignant progression of CLDs, have been observed in many clinical studies. For example, rifaximin, a non-absorbable broad-spectrum antibiotic, was shown to reduce circulating neutrophil TLR4 expression and plasma TNF-α levels in the RIFSYS trial [156]. Concurrently, it fostered an intestinal environment resistant to pathobionts and conducive to gut barrier repair, as indicated by elevated fecal levels of TNF-α and IL-17. Moreover, rifaximin suppressed oralization of the gut microbiome by reducing mucin-degrading species rich in sialidase, such as Streptococcus spp., Veillonella atypica and V. parvula [156]. In addition, probiotic supplementation has also exhibited moderate efficacy in gut-derived inflammation and intestinal permeability. Treatment with probiotic Lactobacillus casei Shirota reduced plasma IL-1β in alcoholic cirrhosis and IL-17A in non-alcoholic cirrhosis [157]. Another clinical trial revealed that a probiotic blend containing six Lactobacillus and Bifidobacterium species helped sustain the expression of CD8⁺ T cells and zonulin-1 in the duodenal mucosa of MASLD patients [158]. Surprisingly, a retrospective study that included 1,267 patients with HBV-related cirrhosis found that probiotic treatment was associated with a lower risk of HCC, exhibiting a dose-response pattern [159]. In addition to gut microbiota, targeting gut-liver barrier may also offer a promising approach to prevent HCC development. Non-selective β-blockers, utilized for preventing variceal bleeding and managing portal hypertension in cirrhotic patients, can additionally reduce intestinal transit time and mitigate gut bacterial overgrowth, intestinal permeability, and bacterial translocation [160]. A recent meta-analysis has revealed that nadolol and carvedilol could decrease the risk of HCC in cirrhotic patients by 26% and 38%, respectively [161]. Others like bile acids, an important regulator of the intestinal barrier through activation of FXR, have been demonstrated to prevent the progression of MASLD and PSC in multiple clinical trials [78,162]. Nevertheless, the preventive effect of FXR agonist on HCC has yet to be directly verified in human studies.

Table 4.

Clinical trials of microbiota-based strategies for liver cancer prevention and therapy

Monitoring the gut microbiota for liver cancer diagnosis and prognosis

The feasibility of using gut microbiome profiling as a diagnostic tool for early screening of HCC has been validated by a cross-regional population study, in which the prediction model consisting of 30 microbial markers achieved an area under the curve (AUC) of 80.64% between 75 early HCC and 105 non-HCC samples [92]. Other studies have further strengthened this diagnostic potential in HCC and extended its application to CCA, demonstrating superior diagnostic performance of gut microbiota-based signatures compared to the classic markers alpha-fetoprotein and carbohydrate antigen 19-9 [98,105,152]. Fungal biomarkers such as increased abundance of C. albicans and depletion of Saccharomyces cerevisiae have also been associated with the progression from cirrhosis to HCC [163].

In addition to its diagnostic role, the gut microbiome also shows predictive potential for the progression and prognosis of HCC. Integrating Veillonella and Streptococcus pneumoniae with TNM stages and aspartate transaminase levels achieved an AUC of 78.0% for predicting early recurrence of HCC [164]. Another study identified six microbial markers through integrated analysis of the microbiome and host transcriptome demonstrating their potential to predict clinical outcomes in HCC patients, with an AUC of 81% [83]. A recent study has also validated the predictive ability of a specific gut microbiome signature, consisting of nine microbial markers, for microvascular invasion in HCC patients, achieving an AUC of 79.80% [165]. Nonetheless, the current diagnostic and prognosis models based on selected gut microbial markers are limited to HBV-related HCC and are restricted to the genus-level analysis of gut microbiome. Future studies involving HCC patients with diverse etiologies, along with more in-depth microbiome profiling, are essential for advancing this field.

Targeting the gut microbiota for liver cancer therapy

As we described above, specific microbes and their metabolites play an important role in modulating hepatic immunity, with some enhancing immune competence while others dampening immunosurveillance. An immunosuppressive tumor environment can thus be attributed to the disturbance of homeostatic gut microbiota. Multiple cohort studies have confirmed the association between the gut microbiome and the efficacy of immune checkpoint inhibitors, including tremelimumab [166], durvalumab [166], nivolumab [167,168], camrelizumab [153], and pembrolizumab [168], in HCC patients, identifying specific bacterial species and microbial metabolites that may be either beneficial or detrimental [154,169]. Reported probiotics, including A. muciniphila [153,166,167], species of Bifidobacterium [153,166], Ruminococcus [153,169], and Faecalibacterium [154,168,169], were enriched in patients with superior treatment responses, while higher abundance of Enterobacteriaceae [166,168], Bacteroides [153,167] was associated with poorer treatment outcomes. Additionally, a study investigating gut microbial metabolites revealed that the microbial metabolites ursodeoxycholic acid and ursocholic acid were more abundant in patients with objective responses [168]. These studies highlight the potential of manipulating the gut microbiota to enhance the efficacy of liver cancer immunotherapy. Several phase I/II clinical trials have already been initiated to evaluate the combined efficacy of ICB therapy and microbiota-based treatments, including vancomycin, probiotics, and FMT (Table 4). While one completed trial combining vancomycin with anti-PD-1 therapy showed no significant efficacy in 6 patients with refractory HCC (NCT03785210), the outcomes of other trials, involving larger cohorts, are still pending.

In addition to immunotherapy, the gut microbiome may also influence the efficacy of other liver cancer treatments, including the tyrosine kinase inhibitors, chemotherapy and radiotherapy [170-172]. A microbiota-derived protein, Staphylococcal superantigen-like protein 6, was found to sensitize HCC cells to sorafenib-induced apoptosis by blocking CD47 and downregulating PI3K/Akt-mediated glycolysis [170]. Moreover, the microbial metabolite butyrate can directly inhibit the proliferation and migration of HCC cells by disrupting intracellular calcium homeostasis, and synergistically improve the efficacy of sorafenib in preclinical models [173]. In HCC patients undergoing chemotherapy, depletion of intestinal anaerobic bacteria, such as Blautia, due to the overuse of antibiotics targeting anaerobes, was associated with poor prognosis [171]. The gut microbiome has also been associated with the response to radiotherapy in HCC patients, with Faecalibacterium identified as being enriched in responders [172]. Mechanistically, gut dysbiosis impaired radiotherapy-associated antitumor immune responses by suppressing stimulator of interferon genes (STING) signaling pathway [172]. Cyclic-di-AMP, a bacterium-derived molecule that is elevated in the stool of responders, may enhance the efficacy of radiotherapy as a STING agonist [172].

It is worth mentioning that encouraging success has been made in promoting liver function recovery after hepatectomy in HCC patients using a probiotic bacteria cocktail containing Bifidobacterium longum [174]. These benefits may derive from diminished liver inflammation, reduced liver fibrosis and enhanced hepatocyte regeneration facilitated by several microbial metabolites such as 5-hydroxytryptamine, secondary bile acids and SCFAs, as indicated in mouse experiments [174]. Although this study concentrated solely on postoperative liver recovery of HCC patients, its implications in improving liver function hold pivotal extending potential in various liver diseases, signifying substantial progress in clinical translation of microbiota-based strategies.

Notably, in addition to antibiotics, probiotics and FMT, novel microbiota-based therapeutic strategies have been developed for liver cancer therapy in recent years. A recent study reported a nanoparticle/engineered bacteria-based triple-strategy delivery system for HCC therapy [175]. This system utilized non-pathogenic Escherichia coli with a synchronized lysis circuit as both an immunostimulatory agent, and a source of hepatitis B surface antigen upon bacterial lysis [175]. Anchored with a nanoparticle loaded with doxorubicin and plasmid encoded human sulfatase 1, the delivery system effectively stimulated the anti-tumor immune responses and suppressed tumor growth [175]. Nano-delivery systems carrying gut microbial metabolites have also made significant advancements in HCC therapy. Based on the anti-tumor effects of butyrate and its synergistic action with sorafenib, glypican-3-targeted nanoparticles encapsulating butyrate and sorafenib were developed and shown to inhibit HCC progression [173]. Moreover, nanoparticles loaded separately with obeticholic acid, an FXR agonist, and 5β-cholanic acid 3, a G-protein-coupled bile acid receptor 1 antagonist, elicited robust anti-tumor immune responses against HCC [176,177]. Collectively, the precise delivery of gut microbiota metabolites by nano-delivery systems amplifies their therapeutic efficacy, limits side effects, and holds promise as a future direction for HCC therapy.