Unveiling the gastric microbiota: implications for gastric carcinogenesis, immune responses, and clinical prospects

Recent studies have investigated the variances in gastric microbial diversity across Correa's cascade, ranging from healthy controls (HC) to SG, AG, IM, and ultimately GC [21, 22, 38, 49‐52]. Among these investigations, two studies did not observe significant differences in gastric microbial diversity among these subgroups [22, 51]. Other larger-scale studies reported significantly reduced microbial richness and diversity in GC when compared with SG or HC groups. Whereas the discrimination of gastric flora diversity between GC and precancerous lesions (IM and dysplasia) did not achieve statistical significance in these studies [53, 54]. Microbial diversity may undergo a substantial decrease between the AG and IM stages [67]. As shown in previous studies, H. pylori eradication significantly decreased the risk of GC whereas its impact is constrained in patients with IM and dysplasia [68, 69]. This indicated that the optimal therapeutic window for antibiotic intervention along Correa's cascade lies before the transition from AG to IM and this timepoint may be correlated with the observed changes in the microbial composition and ecosystem dynamics within the gastric environment.

In parallel, the variations in the composition of the microbiota along the spectrum spanning from healthy gastric mucosa to GC were explored in several studies (Table 2). Through inter-comparison between the gastric microbiome of GC, pre-cancerous stages and normal tissue, these investigations unveiled the distinctive dysbiosis that characterizes gastric carcinogenic progress. A noteworthy observation is that gastric microbial composition is significantly influenced by hypochlorhydria, a main pathological manifestation of atrophy. This alteration in gastric acidity may result in increased vulnerability to colonization by microorganisms from other origins, primarily the oral cavity and intestine [33, 43, 48, 50, 53, 54].

For instance, a prospective cohort study undertaken by Sung et al. recruited 587 H. pylori-positive patients. They reported the presence of a specific assembly of oral bacteria, including Prevotella, Rothia, Peptostreptococcus, Parvimonas and Streptococcus, which were implicated in the occurrence and prolonged course of GA and IM [70]. A similar enrichment of oral microbes was found in a case–control study performed on 89 IM patients, where the abundance of Johnsonella ignava, Peptostreptococcus stomatis, Neisseria elongata, and Neisseria flavescens was found to be significantly higher in IM cases [56]. Moreover, an insightful study was conducted on a well-defined cohort, recruiting 34 cases of SC, 20 cases of AG, 5 cases of gastric intraepithelial neoplasia(GIN), and 15 cases of intestinal-type GC. Remarkably, they revealed a noticeable increase in the abundance of specific oral bacteria, including Slackia, Selenomonas, Bergeyella, and Capnocytophaga, which continuously increased from SG to GC. Concurrently, intestinal bacteria including Romboutsia, Fusicatenibacter, Prevotellaceae-Ga6A1-group and Intestinimonas were observed to be overabundant in GIN patients [22].

The predominance of these “non-indigenous colonizers” persists in GC which may significantly characterize malignant transformation. Coker et al. detected distinct Operational Taxonomic Units (OTUs) that were associated with five oral-derived species, including Peptostreptococcus stomatis, Streptococcus anginosus, Parvimonas micra, Lackia exigua, and Dialister pneumosintes, which can significantly distinguish GC from SG [53]. In another investigation where 54 GC patients and 81 chronic gastritis patients were enrolled, GC-representative genera exhibited a remarkable overabundance of intestinal commensals such as Citrobacter, Clostridium, Lactobacillus, Achromobacter and Rhodococcus [57]. Even by autologous comparison, tumor microhabitats were characterized by the discernible increase of oral pathogen, Prevotella melaninogenica, Streptococcus anginosus and skin pathogen Propionibacterium acnes (P.acnes), setting them apart from both paired paracancerous and normal tissues [58].

The recognition of microbial features during gastric carcinogenesis instigates exploration into the correlation between gastric microbiota and the multifaceted aspects of GC, such as risk, progression and prognosis. A Korean study, comprising a cohort of 268 GC patients and 288 controls, identified H. pylori, P.acnes and Prevotella copri as significant risk factors for GC, as indicated by odds ratios (ORs) of 1.86, 4.7 and 2.54 respectively [9]. Concurrently, another case–control study employed Whole metagenomic shotgun sequencing (WMS) on the gastric biopsy samples, which were obtained from a cohort of Colombian individuals, including both high-risk (n = 10) and low-risk (n = 10) populations for GC. The results revealed that the typical soil bacteria including Bacillus, Actinomyces and Arthrobacter spp., and Keratinibaculum spp. that was over representative in males, as well as skin commensal Staphylococcus and oral microbe Streptococcus, associated with an increased risk of GC [59].

In addition, a prospective longitudinal study was conducted to investigate the predictive value of gastric microbiota in GC progression. This comprehensive investigation involved 43 participants who underwent initial gastroscopy, followed by regular monitoring through 1–2 yearly gastroscopies over a minimum duration of 5 years. The results showed that patients who progressed from IM to early gastric neoplasia (EGN) during the study period had significantly higher relative abundance of Moryella genus and Vibro genus at baseline, compared to those who did not exhibit such progression [51].

Regarding the prognostic assessment of GC based on the gastric microbiome, multiple retrospective studies delved into distinctions in gastric microbial composition between GC patients with favorable and unfavorable prognoses. Several investigations highlighted the association between heightened levels of Fusobacterium nucleatum (F. nucleatum) in tumor samples and unfavorable prognostic outcomes among GC patients [60, 71, 72]. Notably, in Lauren’s diffuse-type GC, F. nucleatum positivity was found to be associated with markedly diminished overall survival rates. The abundance of F. nucleatum demonstrated a positive correlation with patient age, although it exhibited no significant associations with gender, H. pylori infection status, tumor stage, or tumor location [71]. Survival analysis from an additional cohort study has indicated that colonization of F. nucleatum species correlates with poorer prognosis among late-stage GC patients with H. pylori positivity [72]. These findings were further corroborated by an additional study showing that the presence of F. nucleatum and Prevotella species in tumor specimens has most significant impact on prognosis [60]. Furthermore, emerging evidence has also suggested potential correlations between specific bacterial species and favorable prognoses in GC. Using weighted gene co-expression network analysis (WGCNA), a recent study unveiled that GC patients exhibiting higher microbial diversity experienced poorer outcomes compared to those with lower microbial diversity. However, Tissierella enrichment was frequently observed in TP53 wild-type tumors and was associated with a favorable prognosis in various tumor types including GC, breast cancer and lung adenocarcinoma [73]. Another study revealed that GC patients with favorable prognoses displayed elevated H. pylori abundance alongside a decrease in Halomonas and Shewanella abundance [74]. Moreover, the presence of Asinibacterium in non-tumor adjacent tissues correlated with improved overall survival in GC patients, suggesting a potential influence of the gastric microenvironment on the pathophysiology of GC [60].

H. pylori, a known Class I carcinogen in the development of GC, exerts a significant predisposing effect on carcinogenesis by employing various virulence factors [75]. Its remarkable feature of urease synthesis enables it to thrive within the acidic environment of the stomach. The primary mechanism underlying H. pylori-induced carcinogenesis involves the expression of two oncogenic effector proteins: vacuolating cytotoxin A (VacA) and cytotoxin-associated gene A(CagA) protein [76] (Fig. 2).

VacA, a potent vacuolating cytotoxin present in all H. pylori strains, is capable of inducing vacuolation in host cells. VacA-forming pores can target mitochondria, causing depolarization of the mitochondrial transmembrane potential and subsequent mitochondrial dysfunction. This event triggers the release of cytochrome c from the mitochondria, which ultimately leads to the initiation of the apoptotic pathway [77]. The cag pathogenicity island (cagPAI) contains genes responsible for the expression of a type IV secretion system (T4SS) and the oncogenic protein CagA [78]. Upon the action of T4SS, H. pylori translocates CagA into the cytosol of host cells, where it serves as a hub protein to disrupt multiple cellular signaling pathways. CagA, when phosphorylated by Src family kinases, can engage with SH2 domain-containing proteins, which are mostly effectors in the mitogenic signaling pathway [79, 80]. Concurrently, unphosphorylated CagA interacts with E-cadherin to disassociate the E-cadherin/β-catenin complex and activate β-catenin-dependent expression of pro-oncogenic genes [81]. In addition to triggering oncogenic signals, CagA also elicits genetic instability, thereby exacerbating the potential for oncogenic transformation. The CagA-mediated sequestration of the polarity-regulating kinase PAR1b, initially known for its disruption of cellular polarity via tight junction perturbation [82], has recently been elucidated to play a role in inhibiting BRCA1 phosphorylation. This inhibition results in the confinement of BRCA1 to the cytoplasm, called BRCAness, where it fails to perform DNA repair in the nucleus, thereby inducing DNA double-strand breaks (DSBs) and genetic instability. The PAR1b kinase inhibited by CagA also activates Hippo signaling, thereby providing DNA-damaged cells with a means to escape apoptosis and facilitating the repair of DSBs via mutagenic repair mechanisms. As the cells with BRCAness expand, the pro-oncogenic role of CagA is eventually replaced by genomic instability which gives rise to cells with cancer predisposition phenotype independent of CagA. Consequently, the neoplastic phenotype no longer relies on the presence of CagA [83].

Moreover, H. pylori seems to assume a pioneering role by establishing a low-acid environment conducive to the colonization of potential pathogenic microorganisms in the stomach. For instance, a longitudinal study showed that patients with successful H. pylori eradication but still with persistent inflammation exhibit an increase in oral bacteria and pathogenic bacterial load [70]. The mechanisms underlying H. pylori-induced hypochlorhydria were partially unveiled. During acute H. pylori infection, T4SS protein CagL plays a dual role: it facilitates the intracellular delivery of CagA by binding to α5β1 integrins and also influences the dissociation of the metalloenzyme ADAM17 from α5β1 integrins. CagL stimulates ADAM17 activation, initiating ADAM17-triggered NF-κB-mediated suppression of H,K-ATPase α subunit (HKα) [84]. Besides, pro-inflammatory cytokines IL-1ß and TNF-α released by immune cells elicited by H. pylori infection are also robust inhibitors on the gastric acid secretion of parietal cells [85].

The “hit and run” theory was initially introduced by Skinner et al. to elucidate how viruses contribute to carcinogenesis by promoting the accumulation of mutations and inducing genomic instability until the virus is no longer necessary for tumor maintenance [86]. Consistently, genetic instability and neutralized PH in the stomach caused by H. pylori infection induces the accumulation of oncogenic insults including driver mutations acquisition in cancer-predisposing cells and pathogenic bacteria colonization in the stomach, which collectively contribute to the establishment and persistence of a neoplastic phenotype. And the actions of H. pylori successively fade in malignancy maintenance [87]. Aforementioned longitudinal study reported that eradicating H. pylori is insufficient to completely reverse the progression of IM [70]. Moreover, a study incorporating 9 gastritis, 7 IM and 11 GC subjects showed a decrease in H. pylori abundance in GC while no significant differences in the abundance of other dominant species between cancer and non-cancer group [55]. Similar conclusions were found in a Portuguese study (54 GC and 81 CG) in which H. pylori load reduced with increased abundance of intestinal commensal in GC versus chronic gastritis [57]. Based on a pathology review conducted by Stewart et al., the prevalence of H. pylori was found to be as high as 90% in cases of active gastritis, but it decreased to a range of 30% to 72% in individuals with atrophic gastritis and to approximately 30% to 35% in those with intestinal metaplasia (IM) and dysplasia. Notably, only 24.6% of GC patients had detectable H. pylori [3]. The “hit and run” model of H. pylori was initially introduced and documented by Hatakeyama. On the one hand, it highlights the action of CagA in cell reprogramming by facilitating genetic and epigenetic alternation, on the other hand, elevated genetic instability evokes host cell protection mechanisms such as cell senescence and apoptosis, impeding the role of CagA in transformed cells, ultimately resulting in the diminishing influence of CagA over time. Likewise, in the context of reduced gastric acid secretion and the absence of H. pylori, invasive microbiota successfully colonizes the stomach, initiating a progressive replacement of H. pylori by other microbial species that eventually come to dominate the gastric environment [87].

The presence of H. pylori influences the microecological landscape in the stomach. In turn, gastric microbiota can also modulate the outcome of H. pylori infection. This interplay between H. pylori and the microbiota throughout Correa's cascade is potentially involved in GC development.

H. pylori demonstrates a strong competitive advantage, rendering it the predominant microorganism in the gastric microbiota, constituting 40%–90% of the total gastric microbiota composition [30, 38, 57, 88]. The presence of H. pylori is inversely correlated with the alpha diversity of gastric bacteria [58, 88‐90]. While the impact of H. pylori on alpha diversity can be reversible after the removal of H. pylori [70, 91]. Evidence from beta diversity analysis suggested that the presence of H. pylori infection leads to significant modifications to microbial composition in the stomach [89, 92]. Individuals infected with H. pylori exhibit a predominant composition of the same phyla as those without H. pylori, but with low levels of Actinobacteria, Bacteroidetes, and the overrepresentation of Proteobacteria, probably attributed to the influence of H. pylori [93]. Co-occurring and co-excluding networks showed the potential interplay between H. pylori and other gastric microbes. In SG, H. pylori manifests a pattern of co-exclusion with Methylobacillus and co-occurrence with Arthrobacter. Within the stomach of subjects with IM, H. pylori demonstrates co-exclusion with various members of the Firmicutes phylum, such as Ruminococcus, Bacillales, SMB53 and Lactobacillus, while concurrently displaying co-occurrence with Prevotella, Moryella and H. ganmani [53]. Among GC patients, a notable negative correlation persists between Lactobacillus and Helicobacter [23].

A study employing germ-free insulin-gastrin (INS-GAS) transgenic mice illuminated the crucial role of gastric microbiota in H. pylori-driven Correa's cascade progression. H. pylori infection in germ-free mice induced milder lesions and delayed progression to gastric intraepithelial neoplasia (GIN), in comparison to H. pylori-infected specific pathogen-free (SPF) INS-GAS mice [94]. Remarkably, it has been documented that gastric colonization involving H. pylori, along with a consortium of commensal microbiota encompassing species such as Clostridium spp., Bacteroides spp., and Lactobacillus murinus, led to the development of GC in INS-GAS mice to a degree on par with that observed in mice harboring a diverse microbiota. Moreover, the colonization of the stomachs of INS-GAS mice with complex microbiota even reduced the level of gastric H. pylori colonization. These findings indicate that the effective colonization with specific commensals seems to exert a greater influence than the overall microbial diversity in the context of H. pylori-associated GC [94, 95].

Several gastric microbiota with potential modulatory effects on the severity and outcomes of H. pylori infection have been identified [96‐101]. For instance, Staphylococcus epidermidis and Streptococcus salivarius, both urease-positive bacteria, were isolated from individuals in low- and high-cancer-risk regions of Colombia respectively. Co-infection experiments in germ-free mice showed that neither bacterial strain significantly altered H. pylori abundance. However, a more severe pathological lesion was found in H. pylori-S. salivarius dual infection while H. pylori-S. epidermidis co-infection showed alleviation of tissue injury and inhibited gene expression of pro-inflammatory cytokines including IL-22, IL-17A, IL-1β, IFN-γ and TNF-α in contrast to H. pylori mono-infection [96].

Moreover, Weizmannia coagulans (BCF-01), a strain isolated from the healthy gastric mucosa, displayed strong anti-H. pylori activity. This strain could significantly alleviate H. pylori-triggered gastric dysbiosis and reduce inflammation following H. pylori infection. Mechanistically, it enhanced the expression of mucosal barrier proteins and inhibited the TLR4-NFκB-mediated pyroptotic pathway in macrophages [102]. Similarly, Lactobacillus gasseri (Kx110A1), another strain isolated from gastric mucosa, effectively suppressed the synthesis of TNF and IL-6 in macrophages upon H. pylori infection through downregulation of ADAM17 [98]. Also Lactobacillus acidophilus and Lactiplantibacillus plantarum, both indigenous gastric strains, attenuated mucosal inflammation in murine gastritis afflicted with H. pylori [99]. Furthermore, Streptococcus mitis, commonly found in the human stomach, can inhibit the growth of H. pylori and prompt its transformation into a coccoid form when co-cultivated with H. pylori [100, 101]. Variations in the prevalence of these bacteria interacting with H. pylori within the gastric microbial community may underlie the disparate susceptibilities of individuals to H. pylori infection and the development of H. pylori-propelled GC.

A recent study by Kwon et al. demonstrated the successful reproduction of premalignant lesions in the stomach by transplanting the gastric microbiota from IM and GC patients to germ-free mice, even in the absence of H. pylori. This discovery underscores the contribution of gastric dysbiosis beyond the involvement of H. pylori to the induction of GC [103]. While a definitive mechanistic elucidation remains insufficient, numerous investigations have indicated the potential implication of specific bacterial species in GC development.

To begin with, Streptococcus anginosus (S. anginosus),a species originating from the oral cavity and known for its marked resilience in acidic conditions (pH 3–5), has shown significant overabundance in the gastric mucosa of GC patients across various human cohorts [53, 58, 104]. Administration of S. anginosus via oral gavage in conventional and germ-free mice induced progressive precancerous lesions, including gastritis, parietal cell atrophy, mucinous metaplasia, and dysplasia. Further, S. anginosus in carcinogen-induced or allograft GC models expedited tumor progression by disrupting gastric barrier function, stimulating cell proliferation and preventing apoptosis. Notably, the influence of S. anginosus on gastric carcinogenesis appeared to be species-autonomous, which relied on the interaction between a bacterial lipoprotein, TMPC, and Annexin A2 (ANXA2) receptor on gastric epithelial cells. This interaction instigated oncogenic MAPK signaling cascades, as evidenced by the phosphorylation of ERK and JNK, as well as activation of downstream oncogenic targets within gastric epithelial cells following S. anginosus infection [105].

Another frequently cited candidate associated with GC, showing significant overabundance in the gastric microbiota of GC patients, is Lactobacillus species [8, 22, 23, 54, 55, 57, 61, 62]. Some hypotheses suggested that Lactobacillus spp. may contribute to GC development by producing lactic acid, which could serve as an energy source for tumor cells, promote reactive oxygen species (ROS) production, and contribute to tumor angiogenesis [67, 106, 107]. Nevertheless, the current body of proofs falls short of providing definitive support for these hypotheses. A converse viewpoint has been proposed by several studies, suggesting that Lactobacillus spp. could play an anti-tumor role in GC. This includes their potential to activate the intrinsic apoptotic pathway in cancer cells [108], prevent H. pylori infection [109, 110], and regulate gastric dysbiosis [111]. Furthermore, recent observations presented an alternative perspective indicating that the overgrowth of Lactobacillus spp. may be a consequence rather than a cause of gastric carcinogenesis [112, 113]. Jin et al. conducted long-term treatment of INS-GAS mice with deoxycholic acid (DCA), a secondary bile acid that increases from chronic gastritis to intestinal metaplasia. DCA treatment accelerated the progression to IM and enriched the population of the Lactobacillus genus [113]. In light of these complexities, further scrutiny is indispensable to elucidate the multifaceted role of Lactobacillus spp. in GC.

Additionally, lipopolysaccharide (LPS)-producing bacteria were found to be present at the high relative abundance in the gastric microenvironment of patients suffering from bile reflux gastritis (BRG) and GC. Among these bacteria, an oral-derived pathogen, Prevotella melaninogenica (P. melaninogenica) was found to be prominently elevated [58, 114, 115]. The abundance of LPS-producing bacteria exhibited a positive correlation with taurodeoxycholic acid (TDCA), a secondary bile acid that can stimulate the growth of gastric epithelial cells by triggering a pro-inflammatory IL-6/JAK1/STAT3 signaling cascade. The gavage instillation of TDCA, LPS, and P. melaninogenica in mice induced gastric inflammation and pre-neoplastic lesions, indicating the potential involvement of LPS-producing microbes in the pathogenesis of GC [114]. Therefore, targeting these microbes through alternative interventions may hold promise as a preventive measure against BRG-associated GC.

Microbial species participating in the metabolism of nitrate and nitrite have been implicated in GC. Phylogenetic Investigation of Communities by Reconstruction of Unobserved States (PICRUSt) analysis from multiple studies unveiled distinctive microbial functional characteristics in GC that are linked to heightened reduction reactions, converting nitrate to nitrite, a precursor of N-nitroso compounds (NOC) [57, 61, 63]. Indeed, bacteria possessing nitrate reductase enzymes such as Clostridium, Lactobacillus, and Neisseria were found to be significantly overrepresented in GC patients [23, 55, 57, 91]. Conversely, Nitrospira, a bacterium that oxidizes nitrite to nitrate, were found to progressively decrease in individuals with intraepithelial neoplasia (IN), IM and GC. This decrease may indicate the accumulation of nitrite during the progression of neoplastic lesions [52]. The buildup of intragastric nitrite and NOC facilitates malignant transformation by initiating DNA damage through the methylation of purine and guanine with electrophilic species, thereby increasing mutagenicity [116, 117].

A novel investigation unveiled distinctive mucin-microbiome profiles in association with clinical outcomes in GC. The results indicated that tumors displaying an intestinal mucin phenotype with upregulated MUC13 expression are associated with an unfavorable prognosis, whereas tumors characterized by dominant expression of MUC5AC or MUC6 exhibit a more encouraging outcome. The presence of specific oral or intestinal microbes was discovered to be dependent on the mucin phenotype. Notably, oral taxa such as Neisseria, Prevotella, and Veillonella showed a significant association with MUC13 overexpression [118]. In prior studies, MUC13 overexpression was documented in GC [119], and this dysregulated MUC13 signaling was shown to confer protection against apoptosis in colorectal cancer (CRC) cells through the activation of the NF-κB pathway [120], as well as to facilitate the progression of intrahepatic cholangiocarcinoma by activating the EGFR/PI3K/AKT pathways [121]. Nevertheless, its specific involvement in GC has yet to be thoroughly investigated. Whether such interactions between mucin and the microbiome causally contribute to gastric carcinogenesis remains an intriguing area for further exploration.

Microbiota-mediated epigenetic regulation represents a novel mechanism by which gastric bacteria may contribute to carcinogenesis [122]. For instance, tumors harboring F. nucleatum exhibited a trend towards decreased DNA methylation of the long interspersed nuclear elements 1 (LINE1) [71]. Kytococcus sedentarius, Actinomyces oris and Staphylococcus saccharolyticus were implicated in regulating DNA methylation of immune-related genes in gastric adenocarcinoma, which consequently promoted distal metastasis in GC. In an in vitro experiment, the co-culture of GC cells with Staphylococcus saccharolyticus was observed to enhance the proliferation and clonogenicity of GC cells by inducing loss of ZNF215 methylation imprinting [123].

Several microorganisms implicated in other cancers were observed to be also enriched in GC [23, 55, 60, 91]. For instance, F. nucleatum, a periodontal commensal, was reported to be involved in the carcinogenesis of CRC [124]. Two vitro studies provided preliminary insights into the potential role of F. nucleatum in the development of GC. Hsieh et al. revealed that F. nucleatum colonization results in the dysregulation of actin cytoskeletal dynamics, which in turn is likely to alter the motility of GC cells [72]. In another study, it was shown that F. nucleatum is capable of inducing the overexpression of exosomal HOTTIP from GC cells, which subsequently accelerates the progression of GC through the activation of the miR-885-3p/EphB2/PI3K/AKT signaling cascade [125]. However, the impact of F. nucleatum on GC lacks validation from in vivo animal studies, and the molecular mechanisms underlying this effect are not fully understood. In CRC, the carcinogenic effect is primarily mediated by its virulence factors [8, 64, 126, 127]. Fap2 lectin ensures the localization and enrichment of F.nucleatum in cancerous sites [128, 129]. And the interaction of FadA adhesin and E-cadherin of cancer cells over-activate the β-catenin/Wnt signaling pathway [130]. Additionally, LPS of F.nucleatum recognized by the Toll-Like Receptor 4/Myeloid Differentiation Primary Response 88 (TLR4/MYD88) system induces NF-κB signaling pathway activation [131]. Notably, sustained activation and abnormal modulation of the NF-κB signaling pathway play a crucial role in the maintenance of stem-cell-like prosperity of GC cells [131‐137]. Except for F.nucleatum, H. pylori has been demonstrated to stimulate NF-κB signaling pathway activation via its distinctive effector molecule, ADP-glycerol-β-D-mannoheptulose [138, 139], Peptostreptococcus anaerobius and Bacteroides fragilis which were observed to increase in GC, can activate NF-κB signaling pathway via inducing intracellular synthesis of ROS and IL-17 respectively in CRC [140, 141]. However, it remains unclear whether these mechanisms of the bacteria are replicated in the acidic environment of the stomach.