- Split View
-
Views
-
Cite
Cite
Anthony Lopez, Franck Hansmannel, Tunay Kokten, Jean-Pierre Bronowicki, Hassan Melhem, Harry Sokol, Laurent Peyrin-Biroulet, Microbiota in digestive cancers: our new partner?, Carcinogenesis, Volume 38, Issue 12, December 2017, Pages 1157–1166, https://doi.org/10.1093/carcin/bgx087
- Share Icon Share
Abstract
Evolution led to an essential symbiotic relationship between the host and commensal microbiota, regulating physiological functions including inflammation and immunity. This equilibrium can be disturbed by environmental factors such as lifestyle, diet or antibiotic pressure, contributing to create a dysbiosis. There is much evidence about the gut microbiota’s contribution to carcinogenesis, involving pro-inflammatory and immunosuppressive signals. At the same time, it seems to be increasingly clear that commensal microbes can modulate cancer therapy efficacy and safety, in particular, innovating treatments as immune checkpoint inhibitors. In this review, we discuss how the microbiota can promote digestive tract carcinogenesis, responsiveness to cancer therapeutics and cancer-associated complications.
Introduction
Human body contains ten times more commensal microorganisms than cells (1), with a symbiotic interaction between the host and the components of the microbiota, contributing to maintain the gut homeostasis (2). Lifestyle, diet and host genotype are the main determinants of the gut microbiota’s composition, but it is roughly consistent over time in the same person. Microbiota’s genome, also called microbiome, was characterized by metagenomic analyses using shotgun sequences (3). Approximatively 3.3 million of non-redundant microbial genes were reported, which are 150 times larger than the human genome. The microbiome is a key to immune development and protection of the host from invading pathogens (4). Germ-free (GF) mice, which have not been colonized by microorganisms, are highly susceptible to primary infection with impaired activation and accumulation of phagocytes to the site of infection (5). Inversely, the reconstitution of a gut microbiota in GF mice is sufficient to restore the mucosal immune system. If the modifications in the microbial composition exceed the resilience capacity of the microbiota, alterations become permanent, resulting in a dysbiosis, which seem to be involved in many diseases such as inflammatory bowel diseases (IBDs), asthma and even mental disorders. Dysbiosis is also implied in carcinogenesis by initiating pro-inflammatory or immunosuppressive signals (6). Nod2-deficient mice are more susceptible to colitis-associated colorectal cancer (caCRC) in a microbiota-dependent way (7). Indeed, higher susceptibility of caCRC in Nod2-deficient mice can be transferred to wild-type mice via microbiota transplantation (7). According to the seminal works of Heinrich H.R. Koch, we accepted that pathogenicity was an intrinsic characteristic of a microbial species or strain. For example, Helicobacter pylori was classified 20 years ago in the Group 1 of human carcinogens by the International Agency for Research on Cancer (IARC), explaining the majority of gastric cancers (8). However, this ‘one microbe-one disease’ paradigm is rarely true, contrary to global shifts in the microbiome composition promoting carcinogenesis. The ways in which the microbiota contributes to carcinogenesis encompass direct pathogenic mechanisms, impairment of host’s inflammation, cell proliferation and death pathways, immunosuppression and modification of host’s metabolism (6). Recently, data about the impact of gut microbiota in eliciting innate and adaptive immune responses in the context of cancer therapy were reported (4), representing a paradigm shift of the host–microbe mutualism, in which microbiota is no longer a potential pathogen but a possible therapeutic partner.
Role of microbiota in cancer
The infectious hypothesis
In 2012, approximately 15% of new cancer cases worldwide were attributable to carcinogenic infections, with significant differences between Occidental countries (<5% of cases) and some sub-Saharan African countries (>50% of cases) (9). Ten microorganisms, including H. pylori, human papillomavirus, hepatitis B and C viruses or Epstein–Barr virus, are considered as carcinogenic to humans by the IARC (6). Immunoproliferative small intestinal disease (also known as alpha chain disease) is a subtype of lymphoma affecting small intestinal mucosa-associated lymphoid tissue, which responds to antibiotics in case of early onset disease, suggesting a role of some bacteria in carcinogenesis (10). Campylobacter jejuni seems to be involved in this mechanism, by secreting a toxin (CdtB) leading to DNA damage and by inducing a persistent immune stimulation which could eventually lead to the selection of a clone secreting α-heavy chains (11). Helicobacter pylori is directly and strongly linked to gastric cancer, with several mechanisms involved, including H.pylori strain-specific virulence factors, the host genotype, environmental factors such as diet and alterations in stem cell populations and the microbiome (12). In case of H.pylori infection, neutrophils are activated and produce reactive oxygen species (ROS) and reactive nitrogen species (RNS), involved in inflammation and carcinogenesis due to DNA damage (12). But H.pylori induces more complex pathways, via its oncoprotein CagA, able to reprogram epithelial cells and activate the Wnt/β-catenin signaling pathway, which upregulates all central processes involved in carcinogenesis as cellular proliferation, survival, migration and angiogenesis (12). Several authors reported overabundance of Fusobacterium nucleatum in CRC tissues (13,14). Colorectal carcinogenesis could be driven in these cases by expression of FadA, a bacterial cell surface adhesion component that binds host E-cadherin, leading to β-catenin activation (15). In the same way, Salmonella typhi, through its enteric bacterial protein AvrA, could also be a player in oncogenesis (16) (Figure 1).
Dysbiosis, inflammation and cancer
Mucosal surface barriers permit host-microbial symbiosis, but infections, traumatisms or dietary factors can induce a persistent barrier breach leading to an impaired host and microbial resiliency with an increase in pro-inflammatory signals. In mice with intestinal dysbiosis characterized by a proportional expansion of the bacterial phyla Bacteroidetes, production of the pro-inflammatory cytokine IL-18, implied in mediation of mucosal protective mechanisms, is decreased, leading to elevated susceptibility to chemically induced colon carcinogenesis (17). NF-κB is a transcription factor considered as a master regulator of cancer-associated inflammation, by activating survival genes within neoplastic cells and inflammation-promoting genes in components of the tumor microenvironment (18). In the ApcMin/+ mouse model of intestinal tumorigenesis, F.nucleatum generates a pro-inflammatory microenvironment by recruiting tumor-infiltrating myeloid cells, leading to CRC progression (19). This process requires the activation of NF-κB. Several mice models with deficiency in nucleotide-binding oligomerization domain-like receptor (NLR) family members suggest that pro-inflammatory responses can lead to carcinogenesis. For example, NLRP6 is implicated in inflammasome signaling to activate caspase-1, which is essential for the production of mature pro-inflammatory cytokines, IL-1β and IL-18 (20). After dextran sulfate sodium treatment, NLRP6-deficient mice have an impaired ability to resolve inflammation and to repair damaged epithelium compared with wild-type mice, increasing their risk of caCRC (20). Interestingly, microbiota transfer from caCRC mice models to wild-type mice enhances the susceptibility of colonic tumor in these latter (7,20). These data underline the relationship between microbiota and digestive cancer.
Microbiota, immunosuppression and tumor evasion
Microorganisms not only stimulate pro-inflammatory pathways but they can also induce immunosuppressive responses. In addition, microorganisms can also protect tumors from host defense systems. TIGIT (T-cell immunoglobulin and ITIM domain) are inhibitory receptors present on all human natural killer (NK) cells and on various T cells (21). The Fap2 protein of F.nucleatum directly interacts with TIGIT, leading to the inhibition of NK cell cytotoxicity (22). It illustrates how microbes can contribute to tumor-immune evasion mechanisms. The way the microbiota is involved in immunotherapeutic resistance in cancer remains to be investigated, especially as the interaction between microorganisms and tumor cells is complex and ambivalent. For example, several studies support the role of Bacteroides fragilis toxin (BFT) in colon carcinogenesis. However, oral administration with lower-dose biologically active recombinant BFT-2 inhibited colorectal tumorigenesis in mice (23). This effect could be partially explained by an increased expression of caspase-3 involved in apoptosis. A similar paradox exists concerning regulatory T cells (Treg cells) and cancer. FOXP3+ CD4+ CD25+/high Treg cells are dominant cellular elements of the professional suppressor arm of the immune system and are important for orchestrating the control of peripheral immunological tolerance (24). Numerous studies showed that high densities of tumor-associated cells expressing Treg markers including FOXP3 were associated with poorer outcomes in patients with digestive cancers (25,26), but opposite results were also reported (27,28). More than CD8+ or FOXP3+ cells alone, the strongest prognostic factor seems to be the CD8+/FOXP3+ ratio. Analyzing 426 archival tumor tissue samples from patients surgically resected for CRC, Ling et al. demonstrated that a high intraepithelial CD8 expression was associated with a better patient outcome, independent of FOXP3 infiltration (29). But in the same way, in groups of low intraepithelial CD8 expression, a high infiltration rate of FOXP3+ cells at the tumor invasive front significantly improved prognosis. In mice, oral inoculation of a defined mix of Clostridium strains promotes Treg cells accumulation in the colonic mucosa (30), affecting the tumor microenvironment which induces remission of already established intestinal cancers (31). Complex interaction between gut microbiota, immune system and tumor cells led us into a new era in which microorganisms can be either our foes or friends.
Relationship between diet, microbiota, bile acids and immunity
High-saturated fat intake significantly increases the risk of CRC, but multiple actors are involved, with interconnection between obesity, inflammation, bile acids and the microbiome. Obesity is now considered as an inflammatory state, associated with an increased risk of several cancers (32). Moreover, the gut microbiota is altered in obese patients, and some microbes have been shown to impact insulin resistance, inflammation and adiposity via interactions with both epithelial and endocrine cells (32). High-fat diet (HFD) fed mice have altered commensal communities leading to a pro-inflammatory Th1 immune response and an increased incidence of colitis (33). Interestingly, dysbiosis seems to be the central element of intestinal carcinogenesis in case of HFD. In genetically susceptible mice for intestinal tumors (with k-ras mutation), HFD consumption mediated a shift in the composition of the gut microbiota, associated with a decrease in Paneth-cell-mediated antimicrobial host defense that compromised dendritic cell recruitment and antigen presentation in the gut-associated lymphoid tissues (34). As a result, tumor growth and spread were amplified, in the absence of obesity or the development of a robust pro-inflammatory response. Moreover, fecal transplantation from HFD-fed mice with intestinal tumors to healthy adult k-ras mutant mice was sufficient to transmit disease independently of obesity, highlighting the key role of dysbiosis in this model. The gut microbiota is essential for bile acid metabolism. In the enterohepatic circulation, primary bile acids are produced in the liver from cholesterol and are then transformed into secondary bile acids (SBAs) and deconjugated by the intestinal microbiota. Deoxycholic and lithocholic acids are SBA potentially involved in colon carcinogenesis. In vitro, they can induce oxidative DNA damage and decreased apoptosis. SBA also regulated carcinogenesis pathways as EGFR/PKC/Ras/ERK/CREB and PI3K/Akt/IκB/NF-κB (35). In HFD-fed animals, deconjugation of primary bile acids is increased, resulting in a Bilophila wadsworthia growth, which induces colitis (36). Intestinal fermentation of dietary fiber by the colonic microbiota results in the generation of several short-chain fatty acids (SCFAs), such as acetic, propionic and butyric acids. SCFAs are implied in several aspects of the immune response, by recruiting and facilitating differentiation of Treg cells in the colonic environment (37). However, SCFA could have opposite effects. In mice feed with specialized diets and defined microbes, treated with azoxymethane and dextran sodium sulfate for inducing neoplastic colonocytes, the glycolytic metabolism of cancer cells resulted in reduced metabolism of butyrate and enhanced butyrate nuclear accumulation (38). High intranuclear butyrate levels increased histone acetylation and led to increased apoptosis and reduced cellular proliferation. In another CRC mouse model with mutations in both the Apc gene and the mismatch repair gene Msh2, the microbiota and butyrate had tumor-promoting effects (39). These results underline that the relationship between the gut microbiota and the host’s immune system is highly contextual. In other words, some microbes can shift from mutualist to commensal to pathogen according to the state of activation of the host, co-infection or localization. The microbiota can directly inhibit exogenous pathogens, but it is able to stimulate and regulate innate and adaptive immunities.
Microbiota and digestive cancers
Colorectal cancer
Some indirect arguments suggest a potential role of the gut microbiota in colorectal carcinogenesis. Bacteria levels in the colon are approximately one million-fold higher than those in the small intestine, and incidence of CRC is 12-fold higher than that of small bowel adenocarcinoma (40). CRC is essentially a genetic disease, but the microbiota has the potential to explain host gene-environmental interactions in carcinogenesis. However, cancer remains a multifactorial disease, and establishing a robust causality between microbes and CRC is difficult, with regard to several limitations of available colon tumor microbiome studies (e.g. small sample sizes, undefined tissue sampling sites, limited or absence of control samples, no consideration of confounding factors and poorly described analyses) (41). In colon carcinogenesis, few bacteria met usual criteria of causality, such as epidemiologic factors, measurable immunological responses, experimental disease reproduction, biological plausibility and prevention of the disease by elimination or modification of the agent. However, some bacterial species seem to be frequently associated with CRC: Streptococcus gallolyticus (42,43), Enterococcus faecalis (44,45), enterotoxigenic B.fragilis (ETBF) (46–48), Escherichia coli (49–54), and F.nucleatum (13,14,19,55–57) (Table I).
Author . | Year . | Methods . | Cases/ controls . | Variation in CRC . | Bacteria . | Ref. . |
---|---|---|---|---|---|---|
Swidsinski et al. | 1998 | Colonic mucosa, PCR | 31/34 | Escherichia coli | (50) | |
Martin et al. | 2004 | Colonic mucosa, PCR | 21/21 | E.coli | (49) | |
Toprak et al. | 2006 | Stool, PCR | 73/59 | Enterotoxigenic Bacteroides fragilis | (46) | |
Balamurugan et al. | 2008 | Stool, PCR, 16S rDNA sequencing | 20/17 | Enterococcus faecalis Eubacterium rectal Faecalibacterium prausnitzii | (45) | |
Maddocks et al. | 2009 | Colonic mucosa, FISH, PCR | 20/20 | E.coli | (53) | |
Sobhani et al. | 2011 | Stool, pyrosequencing | 60/119 | Bacteroides Prevotella | (47) | |
Marchesi et al. | 2011 | Colonic mucosa, 16S rRNA sequencing | 6/6 | Fusobacterium Roseburia Slackia | (55) | |
Wang et al. | 2012 | Stool, 16S rRNA sequencing | 46/56 | B.fragilis Porphyromonas Escherichia Shigella Enterococcus Roseburia | (48) | |
Kostic et al. | 2012 | Colonic mucosa, FISH, 16S rDNA sequencing | 95/95 | Fusobacterium Bacteroidetes Firmicutes | (13) | |
Castellarin et al. | 2012 | Colonic mucosa, PCR, 16S rRNA sequencing | 11/11 | Fusobacterium | (14) | |
Wu et al. | 2013 | Stool, 16S rRNA sequencing | 19/20 | Fusobacterium Bacteroides Faecalibacterium Roseburia | (56) | |
Kostic et al. | 2013 | Colonic mucosa, stool, FISH, 16S rRNA sequencing | 27/30 | Fusobacterium nucleatum | (19) | |
Bonnet et al. | 2014 | Colonic mucosa, PCR | 50/33 | E.coli | (52) | |
Viljoen et al. | 2015 | Colonic mucosa, PCR, 16S rRNA/DNA sequencing | 55/55 | Fusobacterium | (57) | |
Mira-Pascual et al. CRC, colorectal cancer. | 2015 | Colonic mucosa, stool, PCR, 16S rRNA sequencing | 7/10 | F.nucleatum Enterobacteriaceae | (44) |
Author . | Year . | Methods . | Cases/ controls . | Variation in CRC . | Bacteria . | Ref. . |
---|---|---|---|---|---|---|
Swidsinski et al. | 1998 | Colonic mucosa, PCR | 31/34 | Escherichia coli | (50) | |
Martin et al. | 2004 | Colonic mucosa, PCR | 21/21 | E.coli | (49) | |
Toprak et al. | 2006 | Stool, PCR | 73/59 | Enterotoxigenic Bacteroides fragilis | (46) | |
Balamurugan et al. | 2008 | Stool, PCR, 16S rDNA sequencing | 20/17 | Enterococcus faecalis Eubacterium rectal Faecalibacterium prausnitzii | (45) | |
Maddocks et al. | 2009 | Colonic mucosa, FISH, PCR | 20/20 | E.coli | (53) | |
Sobhani et al. | 2011 | Stool, pyrosequencing | 60/119 | Bacteroides Prevotella | (47) | |
Marchesi et al. | 2011 | Colonic mucosa, 16S rRNA sequencing | 6/6 | Fusobacterium Roseburia Slackia | (55) | |
Wang et al. | 2012 | Stool, 16S rRNA sequencing | 46/56 | B.fragilis Porphyromonas Escherichia Shigella Enterococcus Roseburia | (48) | |
Kostic et al. | 2012 | Colonic mucosa, FISH, 16S rDNA sequencing | 95/95 | Fusobacterium Bacteroidetes Firmicutes | (13) | |
Castellarin et al. | 2012 | Colonic mucosa, PCR, 16S rRNA sequencing | 11/11 | Fusobacterium | (14) | |
Wu et al. | 2013 | Stool, 16S rRNA sequencing | 19/20 | Fusobacterium Bacteroides Faecalibacterium Roseburia | (56) | |
Kostic et al. | 2013 | Colonic mucosa, stool, FISH, 16S rRNA sequencing | 27/30 | Fusobacterium nucleatum | (19) | |
Bonnet et al. | 2014 | Colonic mucosa, PCR | 50/33 | E.coli | (52) | |
Viljoen et al. | 2015 | Colonic mucosa, PCR, 16S rRNA/DNA sequencing | 55/55 | Fusobacterium | (57) | |
Mira-Pascual et al. CRC, colorectal cancer. | 2015 | Colonic mucosa, stool, PCR, 16S rRNA sequencing | 7/10 | F.nucleatum Enterobacteriaceae | (44) |
Author . | Year . | Methods . | Cases/ controls . | Variation in CRC . | Bacteria . | Ref. . |
---|---|---|---|---|---|---|
Swidsinski et al. | 1998 | Colonic mucosa, PCR | 31/34 | Escherichia coli | (50) | |
Martin et al. | 2004 | Colonic mucosa, PCR | 21/21 | E.coli | (49) | |
Toprak et al. | 2006 | Stool, PCR | 73/59 | Enterotoxigenic Bacteroides fragilis | (46) | |
Balamurugan et al. | 2008 | Stool, PCR, 16S rDNA sequencing | 20/17 | Enterococcus faecalis Eubacterium rectal Faecalibacterium prausnitzii | (45) | |
Maddocks et al. | 2009 | Colonic mucosa, FISH, PCR | 20/20 | E.coli | (53) | |
Sobhani et al. | 2011 | Stool, pyrosequencing | 60/119 | Bacteroides Prevotella | (47) | |
Marchesi et al. | 2011 | Colonic mucosa, 16S rRNA sequencing | 6/6 | Fusobacterium Roseburia Slackia | (55) | |
Wang et al. | 2012 | Stool, 16S rRNA sequencing | 46/56 | B.fragilis Porphyromonas Escherichia Shigella Enterococcus Roseburia | (48) | |
Kostic et al. | 2012 | Colonic mucosa, FISH, 16S rDNA sequencing | 95/95 | Fusobacterium Bacteroidetes Firmicutes | (13) | |
Castellarin et al. | 2012 | Colonic mucosa, PCR, 16S rRNA sequencing | 11/11 | Fusobacterium | (14) | |
Wu et al. | 2013 | Stool, 16S rRNA sequencing | 19/20 | Fusobacterium Bacteroides Faecalibacterium Roseburia | (56) | |
Kostic et al. | 2013 | Colonic mucosa, stool, FISH, 16S rRNA sequencing | 27/30 | Fusobacterium nucleatum | (19) | |
Bonnet et al. | 2014 | Colonic mucosa, PCR | 50/33 | E.coli | (52) | |
Viljoen et al. | 2015 | Colonic mucosa, PCR, 16S rRNA/DNA sequencing | 55/55 | Fusobacterium | (57) | |
Mira-Pascual et al. CRC, colorectal cancer. | 2015 | Colonic mucosa, stool, PCR, 16S rRNA sequencing | 7/10 | F.nucleatum Enterobacteriaceae | (44) |
Author . | Year . | Methods . | Cases/ controls . | Variation in CRC . | Bacteria . | Ref. . |
---|---|---|---|---|---|---|
Swidsinski et al. | 1998 | Colonic mucosa, PCR | 31/34 | Escherichia coli | (50) | |
Martin et al. | 2004 | Colonic mucosa, PCR | 21/21 | E.coli | (49) | |
Toprak et al. | 2006 | Stool, PCR | 73/59 | Enterotoxigenic Bacteroides fragilis | (46) | |
Balamurugan et al. | 2008 | Stool, PCR, 16S rDNA sequencing | 20/17 | Enterococcus faecalis Eubacterium rectal Faecalibacterium prausnitzii | (45) | |
Maddocks et al. | 2009 | Colonic mucosa, FISH, PCR | 20/20 | E.coli | (53) | |
Sobhani et al. | 2011 | Stool, pyrosequencing | 60/119 | Bacteroides Prevotella | (47) | |
Marchesi et al. | 2011 | Colonic mucosa, 16S rRNA sequencing | 6/6 | Fusobacterium Roseburia Slackia | (55) | |
Wang et al. | 2012 | Stool, 16S rRNA sequencing | 46/56 | B.fragilis Porphyromonas Escherichia Shigella Enterococcus Roseburia | (48) | |
Kostic et al. | 2012 | Colonic mucosa, FISH, 16S rDNA sequencing | 95/95 | Fusobacterium Bacteroidetes Firmicutes | (13) | |
Castellarin et al. | 2012 | Colonic mucosa, PCR, 16S rRNA sequencing | 11/11 | Fusobacterium | (14) | |
Wu et al. | 2013 | Stool, 16S rRNA sequencing | 19/20 | Fusobacterium Bacteroides Faecalibacterium Roseburia | (56) | |
Kostic et al. | 2013 | Colonic mucosa, stool, FISH, 16S rRNA sequencing | 27/30 | Fusobacterium nucleatum | (19) | |
Bonnet et al. | 2014 | Colonic mucosa, PCR | 50/33 | E.coli | (52) | |
Viljoen et al. | 2015 | Colonic mucosa, PCR, 16S rRNA/DNA sequencing | 55/55 | Fusobacterium | (57) | |
Mira-Pascual et al. CRC, colorectal cancer. | 2015 | Colonic mucosa, stool, PCR, 16S rRNA sequencing | 7/10 | F.nucleatum Enterobacteriaceae | (44) |
S.gallolyticus
In a meta-analysis published in 2011, patients with S.gallolyticus (the former S.bovis biotype I) infection had a strongly increased risk of having CRC (pooled odds ratio [OR], 7.26; 95% confidence interval [CI], 3.94–13.36) compared with S.bovis biotype II-infected patients (42). The positive detection of S.gallolyticus DNA in tumorous and non-tumorous tissues of CRC patients varied between 23% and 49%, compared with <5% in control tissues (P < 0.05) (43). It remains uncertain whether S.gallolyticus is an opportunist invader through a loss in the mucosal barrier integrity or if it contributes to colon carcinogenesis due to specific virulence mechanisms. However, S.gallolyticus possesses a pilus protein associated with enhanced inflammatory signals including COX-2 (58). Other mechanisms were reported, as adhesion to both normal epithelial and neoplastic cells, or a competitive growth advantage in a tumor microenvironment (59).
E.faecalis
Human data are scarce (44,45), but in a study comparing fecal bacterial composition between 20 CRC patients and 17 healthy volunteers, E.faecalis was significantly higher in case of CRC (P = 0.0294) (44). E.faecalis is able to produce ROS (inducing DNA damage and genomic instability) and extracellular superoxide anions (potential initiators of CRC) (41). IL-10−/− mice colonized with superoxide-producing E.faecalis develop inflammation and CRC, whereas colonization with a superoxide-deficient strain results in inflammation but not cancer (60).
Enterotoxigenic B.fragilis
ETBF is often acquired during childhood, leading to diarrheal illnesses (61), but asymptomatic adult colonization with ETBF is also common, occurring in up to 40% of individuals (62). ETBF was detected significantly higher in the stools of CRC patients compared with controls in some studies (46–48), whereas several evidences of its potential role in colon carcinogenesis are available in the literature. Virulence of ETBF is essentially due to the BFT, which cleaves the E-cadherin tumor suppressor protein, resulting in enhanced nuclear Wnt/β-catenin signaling that yields increased colonic carcinoma cell proliferation (41). BFT also triggers the NF-κB-carcinogenesis pathway (61). Finally, specific STAT3 activation with induction of a mucosal IL17 response is also associated with ETBF colon carcinogenesis (63).
E.coli
E.coli was suspected as a pro-carcinogenic factor almost 20 years ago (50). Several human studies reported an association between E.coli and CRC (49–54). For example, Martin et al. reported that 71% of mucosa samples from patients with CRC were colonized by E.coli compared with 42% in controls (49). Genotoxic E.coli strains, and especially B2 colibactin-producing group, can induce double-strand DNA breaks and genomic instability through the polyketide synthase (pks) island containing the genotoxin colibactin (64). In IL10-−/− mice treated with azoxymethane, deletion of the pks island reduces DNA damage, tumor numbers and bacterial invasion, but not inflammation (51). Moreover, possible interactions have been described between another subgroup of E.coli (enteropathogenic E.coli) and DNA mismatch repair system (65).
F.nucleatum
This phylum probably provided the strongest evidence about the role of microbes in colon carcinogenesis. Using metagenomic methods (16S rDNA and RNA-sequence analyses), several studies found an enrichment of Fusobacterium spp. sequences, especially F.nucleatum, associated with tumor samples relative to the normal colon tissue from the same cancer-bearing host (13,14,19,55–57). Interestingly, rectal adenoma subjects seem to have significantly higher abundance of Fusobacterium species compared with controls (P = 0.01) (66). In the same study, subjects with high abundance of Fusobacterium spp. were significantly more likely to have adenomas (OR, 3.66; 95% CI, 1.37–9.74; p = 0.005) compared with the lowest tertile (66). These data suggest that F.nucleatum acts at the early steps of colorectal carcinogenesis promotion. As described above, carcinogenic properties of F.nucleatum could be due to a FadA expression, activating the β-catenin/Wnt pathway (15). NF-κB activation is also involved, inducing an expansion of myeloid-derived immune cells in the tumor microenvironment (19). Finally, F.nucleatum could be a prognosis marker. Recently, high-level colonization by Fusobacterium was significantly associated with stages III and IV CRC, as well as with microsatellite instability phenotype (57).
Fungi
In addition to bacteria, the microbiota is composed of archaea, fungi, viruses and bacteriophages, and that dysbiosis is most often associated with changes in the reciprocal composition of the different members of the microbiota. Metagenomic analyses reported that only 0.03 to 0.1% of the genes in the fecal material were of eukaryotic or viral origin (67), explaining the few number of studies assessing the role of fungi in human cancers. Indirect arguments suggest that the mycobiome could interact with the immune system and enhance inflammatory pathways, leading to an increased risk of malignancy. For example, shifts of gut fungal microbiota composition may be associated with mucosal inflammation and disease activity of Crohn’s disease (68). In animal models, antibiotic treatment results in the overgrowth of a commensal fungal Candida species in the gut and increased plasma concentrations of prostaglandin E2, inducing an inflammatory response (69). Fungi are recognized by a number of innate immune receptors among which Dectin-1 has emerged as key for phagocytosis and killing by myeloid phagocytes. Mice lacking Dectin-1 exhibit increased susceptibility to chemically induced colitis, which is the result of altered responses to indigenous fungi (70). Using a deep sequencing technology, Luan et al. studied the fungal microbiota of adenomas and adjacent tissues from 27 subjects (71). Although Ascomycota, Glomeromycota and Basidiomycota were identified as the dominant phyla in both adenomas and adjacent tissues, two opportunistic pathogenetic fungi genera, Phoma and Candida, were abundantly present (45%). At the operational taxonomic unit (OTU) level, a decreased diversity in adenomas was observed, and three OTUs differed significantly from the adjacent tissues (OTU 144089 assigned to phylum Basidiomycota, OTU 196869 assigned to phylum Glomeromycota and OTU 697566 assigned to phylum Chytridiomycota). This study thus identified fungal microbiota profiles potentially involved in colon carcinogenesis, but further investigation is needed to better understand the impact of the mycobiome on human digestive cancers.
Gastric cancer
According to gastric acidity, the bacterial density appears five times lower in the stomach than in the colon. In H.pylori-negative subjects, the most abundant phylum is Proteobacteria, followed by Firmicutes, Bacteroidetes, Actinobacteria and Fusobacteria, whereas Streptococcus and Prevotella are the most abundant genus (72). Prevalence of H.pylori infection considerably varies worldwide, from 45% to 84% in recent studies (73). Although 74% of the global gastric cancer burden worldwide is attributable to H.pylori (9), this latter may change the composition and biodiversity of the gastric microbiota, promoting gastric carcinogenesis. Using 16S rRNA gene microarray, a small study (n = 12) found that the relative abundance of Proteobacteria and Acidobacteria was higher in H.pylori-infected patients, and a greater relative abundance of Actinobacteria and Firmicutes in H.pylori-negative patients (74). Similar data were reported in some mice models (75). However, conflicting findings are available, suggesting that the presence of H.pylori in the gastric mucosa does not affect the composition of gastric community (72,76). Yang et al. recently compared the gastric microbiota of two groups of subjects issued from two Colombian towns with dramatically different gastric cancer incidences, despite similar H.pylori prevalences (77). Leptotrichia wadei and Veillonella spp. were significantly more abundant in the high incidence group, whereas Staphylococcus spp. was significantly more abundant in the low incidence group. Few studies compared the constitution of gastric microbiota between gastric cancer patients and controls with conflicting results, possibly due to different analysis methods. However, the most recent studies reported an increase in Firmicutes (e.g. Lachnospiraceae, Bacilli and Streptococacceae) and a decrease in Proteobacteria (e.g. Neisseria spp., Haemophilus spp. and Bergeriella denitrificans) and Bacteroidetes (e.g. Porphyromonas spp. and Prevotella pallens) in case of gastric cancer (78–80). Studies on animal models strongly support the fundamental role of microbiota in the development of gastric cancer. Transgenic insulin–gastrin (INS–GAS) mice over-expressing human gastrin may spontaneously develop intramucosal gastric carcinoma (81). This process is accelerated in GF INS–GAS mice after gastric microbiota colonization when compared with those infected by H.pylori alone (81). In contrast, antibiotherapy delayed significantly the development of gastric cancer (81). However, disturbance of the gastric bacterial community due to antibiotics allows colonization by C.albicans, promiting C.albicans-induced gastritis (82) and potentially neoplasia. The way the microbiota could influence gastric carcinogenesis is not well documented, but it could involve amino-nitroso compounds. These latter are significantly associated with gastric cancer, and nitrite is a known precursor of the endogenous amino-nitroso compounds. In the case of gastric dysbiosis, nitrate-reducing bacteria can be over-represented, as Veillonella parvula, leading to a nitrite accumulation in the gastric juice (83). Finally, H.pylori could act as a trigger for the development of chronic atrophic gastritis, associated with a decreased acidity and changes in the constitution of the microbiota that could play important roles in the later stages of gastric carcinogenesis (84).
Esophageal cancer (EC)
Comparing CCR with gastric cancer, data about microbiota and EC are poor. In 2009, Yang et al. analyzed 6,800 16S rRNA gene clones from 34 subjects, isolating two microbiome profiles: type I, mainly associated with normal esophagus, was predominated by bacteria from the Firmicutes phylum, of which Streptococcus was the most dominant genus; type II had a greater proportion of bacteria from Bacteroidetes, Proteobacteria, Fusobacteria and Spirochaetes phyla and was correlated with reflux esophagitis (RE) (OR, 15.4) and Barrett’s esophagus (BE) (OR, 16.5) (85). These results were subsequently refined by Liu et al. who reported that Veillonella, Prevotella, Neisseria and Fusobacterium were prevalent in patients with RE and BE but were not detected in controls (86). Data about modification of esophageal microbiota in case of tumor are scarce. In a British study comparing 30 cases of EC with 39 controls, no statistical difference in specific taxa was reported (87). Results were similar in a recent study analyzing five EC and eight healthy subjects (88). The way the microbiota could play a role in esophageal carcinogenesis remains uncertain, but toll-like receptors (TLRs) expressed in the microenvironment of the esophageal mucosa are suggested as potential mediators of the progression from reflux disorders to cancer, mediating the interaction between the immune system and the microbiome (89).
Hepatocellular carcinoma
The liver is constantly exposed to microbial products from the enteric microbiota via the portal vein flow. In mice models, gut sterilization can protect from development of liver cancer (90). In the same way, an antibiotic treatment leads to a reduction of tumor number and size, whereas GF mice develop fewer and smaller tumors (90). Promotion of hepatocellular carcinoma (HCC) development in these mice could be due to the activation of TLR4 by lipopolysaccharide from gut some bacteria (90). The link between microbiota and HCC seems to be closely related to obesity. In HFD mice, the percentage of gut Gram-positive bacteria is increased, potentially the source of DNA damage in hepatic stellate cells in obese mice (91). SBA could also play a role in liver carcinogenesis. SBAs are increased in HFD mice promoting HCC development through activation of a mitogenic and pro-inflammatory response program in hepatic stellate cells (91). Although these data highlight an impact of intestinal microbiota and gut homeostasis on liver carcinogenesis, human longitudinal studies are needed to confirm this relationship.
Pancreatic cancer
Pancreatic cells can be in touch with microbes through intestinal and oral microflora translocation. Several studies evaluating the role of microbiota in pancreatic cancer (PC) did it through the oral microbiome (92,93). Farrell et al. compared the saliva microbiota of patients with PC (n = 38) to healthy controls (n = 38), finding that Neisseria elongata and S.mitis were decreased in case of tumor, whereas Granulicatella adiacens levels were increased (92). In a large case-controls study (n = 821), high levels of antibodies against Porphyromonas gingivalis (pathogenic periodontal bacteria) were associated with a higher risk of PC (OR, 2.14; 95% CI, 1.05–4.36) (93). Controversial results are available concerning H.pylori and PC, but a meta-analysis of nine studies involving 3033 subjects found a modest increased risk of pancreatic neoplasia in case of H.pylori infection (OR, 1.47; 95% CI, 1.22–1.77) (94). This effect could be explained by an upregulation of the NF-κB pathway, but the role of global shifts in the microbiome composition has not been evaluated in the context of pancreatic carcinogenesis. However, chronic activation of the immune system and perpetuation of tumor-associated inflammation, involving upregulation of TLRs, seems to be a plausible mechanism (95). In conclusion, the microbiome holds promise as a biomarker for early detection of PC, but further studies are warranted to better explore the relationship between PC and microbes. Main studies reporting microbiota variations in digestive neoplasms are summarized in Table 2.
Author . | Year . | Methods . | Cases/ controls . | Variation . | Bacteria . | Ref. . |
---|---|---|---|---|---|---|
Gastric cancer | ||||||
Eun et al. | 2014 | Gastric mucosa, PCR, 16S rRNA gene sequencing | 11/20 | Bacilli Streptococacceae Epsilonproteobacteria Helicobacteriaceae | (78) | |
Aviles-Jimenez et al. | 2014 | Gastric mucosa, PCR, 16S rRNA microarray | 5/10 | Lachnospiraceae Lactobacillus coleohominis Neisseria spp. Haemophilus spp. Haemophilus spp. Haemophilus spp. Porphyromonas spp. Prevotella pallens | (79) | |
Khosravi et al. | 2014 | Gastric mucosa, 16S rRNA gene sequencing | 8/207 | Klebsiella pneumonia Acinetobacter baumanii | (80) | |
Pancreatic cancer | ||||||
Farrell et al. | 2012 | Saliva, 16S rRNA gene sequencing | 38/38 | Granulicatella adiacens Neisseria elongate S.mitis | (92) | |
Michaud et al. | 2013 | Blood, antibodies detection, immunoblot array | 405/416 | Porphyromonas gingivalis | (93) | |
Xiao et al.a | 2013 | Blood, antibodies detection, ELISA or WB | 1083/1950 | Helicobacter pylori | (94) |
Author . | Year . | Methods . | Cases/ controls . | Variation . | Bacteria . | Ref. . |
---|---|---|---|---|---|---|
Gastric cancer | ||||||
Eun et al. | 2014 | Gastric mucosa, PCR, 16S rRNA gene sequencing | 11/20 | Bacilli Streptococacceae Epsilonproteobacteria Helicobacteriaceae | (78) | |
Aviles-Jimenez et al. | 2014 | Gastric mucosa, PCR, 16S rRNA microarray | 5/10 | Lachnospiraceae Lactobacillus coleohominis Neisseria spp. Haemophilus spp. Haemophilus spp. Haemophilus spp. Porphyromonas spp. Prevotella pallens | (79) | |
Khosravi et al. | 2014 | Gastric mucosa, 16S rRNA gene sequencing | 8/207 | Klebsiella pneumonia Acinetobacter baumanii | (80) | |
Pancreatic cancer | ||||||
Farrell et al. | 2012 | Saliva, 16S rRNA gene sequencing | 38/38 | Granulicatella adiacens Neisseria elongate S.mitis | (92) | |
Michaud et al. | 2013 | Blood, antibodies detection, immunoblot array | 405/416 | Porphyromonas gingivalis | (93) | |
Xiao et al.a | 2013 | Blood, antibodies detection, ELISA or WB | 1083/1950 | Helicobacter pylori | (94) |
ameta-analysis; WB, western blot.
Author . | Year . | Methods . | Cases/ controls . | Variation . | Bacteria . | Ref. . |
---|---|---|---|---|---|---|
Gastric cancer | ||||||
Eun et al. | 2014 | Gastric mucosa, PCR, 16S rRNA gene sequencing | 11/20 | Bacilli Streptococacceae Epsilonproteobacteria Helicobacteriaceae | (78) | |
Aviles-Jimenez et al. | 2014 | Gastric mucosa, PCR, 16S rRNA microarray | 5/10 | Lachnospiraceae Lactobacillus coleohominis Neisseria spp. Haemophilus spp. Haemophilus spp. Haemophilus spp. Porphyromonas spp. Prevotella pallens | (79) | |
Khosravi et al. | 2014 | Gastric mucosa, 16S rRNA gene sequencing | 8/207 | Klebsiella pneumonia Acinetobacter baumanii | (80) | |
Pancreatic cancer | ||||||
Farrell et al. | 2012 | Saliva, 16S rRNA gene sequencing | 38/38 | Granulicatella adiacens Neisseria elongate S.mitis | (92) | |
Michaud et al. | 2013 | Blood, antibodies detection, immunoblot array | 405/416 | Porphyromonas gingivalis | (93) | |
Xiao et al.a | 2013 | Blood, antibodies detection, ELISA or WB | 1083/1950 | Helicobacter pylori | (94) |
Author . | Year . | Methods . | Cases/ controls . | Variation . | Bacteria . | Ref. . |
---|---|---|---|---|---|---|
Gastric cancer | ||||||
Eun et al. | 2014 | Gastric mucosa, PCR, 16S rRNA gene sequencing | 11/20 | Bacilli Streptococacceae Epsilonproteobacteria Helicobacteriaceae | (78) | |
Aviles-Jimenez et al. | 2014 | Gastric mucosa, PCR, 16S rRNA microarray | 5/10 | Lachnospiraceae Lactobacillus coleohominis Neisseria spp. Haemophilus spp. Haemophilus spp. Haemophilus spp. Porphyromonas spp. Prevotella pallens | (79) | |
Khosravi et al. | 2014 | Gastric mucosa, 16S rRNA gene sequencing | 8/207 | Klebsiella pneumonia Acinetobacter baumanii | (80) | |
Pancreatic cancer | ||||||
Farrell et al. | 2012 | Saliva, 16S rRNA gene sequencing | 38/38 | Granulicatella adiacens Neisseria elongate S.mitis | (92) | |
Michaud et al. | 2013 | Blood, antibodies detection, immunoblot array | 405/416 | Porphyromonas gingivalis | (93) | |
Xiao et al.a | 2013 | Blood, antibodies detection, ELISA or WB | 1083/1950 | Helicobacter pylori | (94) |
ameta-analysis; WB, western blot.
Role of microbiota in cancer therapy
Recently, a new landscape research emerged where microbiota members would not be the etiology of cancer, but rather a way to control it (Figure 2). The role of intestinal microbiota in the development and severity of chemotherapy-induced mucositis has also been proposed. In 2013, Viaud et al. first described the impact of gut microbiota on anticancer therapy’s efficacy (96). They showed that cyclophosphamide altered the composition of microbiota in mice and induced the translocation of some Gram-positive bacteria into secondary lymphoid organs, where they stimulated the generation of Th17 cells and memory Th1 immune responses. In GF mice, there was a reduction in Th17 response, associated with a resistance to cyclophosphamide. As described previously, TLRs are pattern-recognition receptors used by the innate immune system to detect microbial signals, as CpG oligonucleotides (CpG-ODN). Interaction between CpG-ODN and TLR9 expressed by dendritic cells and B-cells can enhance antitumor immunity (97). Intralesional injection of CpG-ODN with IL-10 antibodies in mice induces a tumor regression, but this treatment is not effective in GF or antibiotic-treated animals, due to a decreased cytokines production as IL-12 or TNF (98). In these mice, gut microbiota composition impacts the anti-tumor response, which is increased in presence of Alistipes shahii and decreased with Lactobacillus fermentum (98). Similar to the findings with CpG-ODN treatment, Iida et al. demonstrated that antibiotics reduce the therapeutic efficacy of oxaliplatin against subcutaneous transplanted tumors, due to a lower release of ROS from myeloid cells (98). Recently, Vetizou et al. showed that immune checkpoints inhibitors, as monoclonal antibodies against lymphocyte-associated antigen 4 (CLTA-4), were not effective for treating subcutaneous tumors in GF mice or in mice treated with antibiotics (99). Interestingly, this defect was overcome by gavage with B.fragilis, by immunization with B.fragilis polysaccharides or by adaptive transfer of B.fragilis-specific T cells (99). Authors suggested that anti-CTLA-4 efficacy depended on the abundance in the intestinal microbiota of Bacteriodes species (e.g. B.thetaiotaomicron and B.fragilis) and the proteobacteria Burkholderia cepacia, by inducing activation of IL-12 producing dendritic cells and activation of Th1 cells (99). Although microbiota did not seem so essential for anti-PD-L1 efficacy, Bifibacterium spp. induced an activation of CD11c+ dendritic cells leading to an elevated anti-tumor response and a slower tumor growth in murine models (100). Finally, the gut microbiota could be implied in the pathogenesis of chemotherapy-induced gastrointestinal mucositis. In a recent meta-analysis, patients receiving chemotherapy showed a decrease in Bifidobacterium, Clostridium cluster XIVa and Faecalibacterium prausnitzii, and an increase in Enterobacteriaceae and Bacteroides (101). These modifications may contribute to the development of mucositis, particularly diarrhea and bacteremia. Incidence of chemotherapy-induced diarrhea could be reduced with probiotic. In a randomized controlled trial, 150 CRC patients treated with adjuvant 5-fluorouracil chemotherapy received L.rhamnosus or placebo (102). Grade 3/4 diarrhea was significantly lower in case of probiotic administration compared with placebo (37% versus 22%, P = 0.027).
Conclusion
Over the past few years, we saw significant advances concerning knowledge of the gut microbiota, thanks to spreading of metagenomic analyses and the development of bioinformatic algorithms. There is body of evidence that the commensal microbiota has an impact on carcinogenesis, tumor progression and the response to therapy. Dysbiosis is highly implied in cancer-associated inflammation, by activating survival genes within neoplastic cells and inflammation-promoting genes in the tumor microenvironment. However, microbiota and cancer maintain a complex relationship, with multiple interactions between the diet, bile acids and the immune system. Recent data showed that the gut microbiota is involved in the reprogramming of intratumoral myeloid cells, contributing to pro-inflammatory antitumor mechanisms as ROS production or modulation of adaptive immunity. Naturally, modification of microbiota emerged as a potential therapeutic weapon, but results are contrasting. Although fecal transplantation is highly effective in case of Clostridium difficile infection, no randomized controlled trial is available in Crohn’s disease patients, but some open studies suggested that fecal transplantation could be a promising therapy in IBD. In the near future, the microbiota could be a biomarker for CRC diagnosis as well as a predictive factor for chemotherapy side effects and efficacy. In the therapeutic field, probiotics and prebiotics could improve the safety profile of chemotherapy, but also be prescribed as adjuvants for cancer treatment. Further, robust studies are needed to find the right place for the gut microbiota in the cancer armamentarium.
Funding
None.
Conflicts of Interest Statement; AL: board for Amgen, lecture fees from Vifor Pharma, research grants from Roche; HS received consultancy fees from Enterome, Maat Pharma, Astellas, Danone, MSD, Ferring and Roche, and speaker’s fees from Takeda, Abbvie, Astellas and Biocodex; LPB: consulting fees from Merck, Abbvie, Janssen, Genentech, Mitsubishi, Ferring, Norgine, Tillots, Vifor, Therakos, Pharmacosmos, Pilège, BMS, UCB-pharma, Hospira, Celltrion, Takeda, Biogaran, Boerhinger-Ingelheim, Lilly, Pfizer, HAC-Pharma, Index Pharmaceuticals, Amgen, Sandoz, Forward Pharma GmbH and Celgene. Lecture fees from Merck, Abbvie, Takeda, Janssen, Takeda, Ferring, Norgine, Tillots, Vifor, Therakos, Mitsubishi and HAC-pharma. FH, TK and HM have no conflict of interest.
Abbreviations
- BE
Barrett’s esophagus
- BFT
Bacteroides fragilis toxin
- caCRC
colitis-associated colorectal cancer
- CLTA-4
lymphocyte-associated antigen-4
- CpG-ODN
CpG oligonucleotide
- CRC
colorectal cancer
- EC
esophageal cancer
- ETBF
enterotoxigenic B.fragilis
- GAS
gastrin
- GF
germ-free
- HCC
hepatocellular carcinoma
- HFD
high-fat diet
- IARC
International Agency for Research on Cancer
- IBD
inflammatory bowel disease
- INS
insulin
- NK
natural killer
- NLR
domain-like receptor
- OTU
operational taxonomic unit
- PC
pancreatic cancer
- pks
polyketide synthase
- RE
reflux esophagitis
- RNS
reactive nitrogen species
- ROS
reactive oxygen species
- SBA
secondary bile acid
- SCFA
short-chain fatty acid
- TLR
toll-like receptor
References