Gut Microbiota and Colorectal Cancer Development: A Closer Look to the Adenoma-Carcinoma Sequence

There is wide evidence that CRC could be prevented by regular physical activity, keeping a healthy body weight, and following a healthy and balanced diet. Many sporadic CRCs develop via the traditional adenoma-carcinoma pathway, starting as premalignant lesions represented by conventional, tubular or tubulovillous adenomas. The gut bacteria play a crucial role in regulating the host metabolism and also contribute to preserve intestinal barrier function and an effective immune response against pathogen colonization. The microbiota composition is different among people, and is conditioned by many environmental factors, such as diet, chemical exposure, and the use of antibiotic or other medication. The gut microbiota could be directly involved in the development of colorectal adenomas and the subsequent progression to CRC. Specific gut bacteria, such as Fusobacterium nucleatum, Escherichia coli, and enterotoxigenic Bacteroides fragilis, could be involved in colorectal carcinogenesis. Potential mechanisms of CRC progression may include DNA damage, promotion of chronic inflammation, and release of bioactive carcinogenic metabolites. The aim of this review was to summarize the current knowledge on the role of the gut microbiota in the development of CRC, and discuss major mechanisms of microbiota-related progression of the adenoma-carcinoma sequence.


Introduction
Colorectal cancer (CRC) is a leading cause of cancer mortality worldwide with approximately 900,000 deaths every year, and the increasing age-standardized incidence rate of CRC in most countries represents an important public health challenge [1]. Indeed, the global incidence of CRC was 1.8 million (95% UI 1.8-1.9) in 2017, with an age-standardized incidence rate of 23.2 per 100,000 person-years that raised by 9.5% (4.5-13.5) between 1990 and 2017 [2]. There is wide evidence that CRC risk is highly modifiable through diet and lifestyle [3]. Several studies suggested that a significant number of CRC cases could be prevented by regular physical activity, keeping a healthy body weight, and following a healthy and balanced diet [4][5][6].
Around 60-90% of sporadic CRCs arise via the traditional adenoma-carcinoma pathway, starting as premalignant lesions represented by conventional, tubular, or tubulovillous adenomas [7]. Cancers that derive from this pathway are frequently associated with male sex, and located in the distal colon. These tumors are characterized by chromosomal instability (CIN), inactivating mutations or losses in the adenomatous polyposis coli (APC) tumor suppressor gene, and in some cases mutations in the KRAS oncogene, SMAD4, PIK3CA, and TP53 genes [8,9].
The term "gut microbiota" indicates the collection of microorganisms (bacteria, archaea and eukarya) colonizing the human gastrointestinal tract. Overall, the number of these microorganisms has been calculated to exceed 10 14 , with a ratio of human:bacterial cells closer to 1:1 [10,11]. The gut bacteria play a crucial role in regulating host metabolism (i.e., absorption of indigestible The adenoma-carcinoma progression may occur because of the genomic instability caused by alterations in the gut microbiota. These changes may be supported by diet and lifestyle, which promote dysbiosis, inflammatory state and epithelial DNA damage, thus contributing to CRC development. The carcinogenesis leads to gut niche changes, which may favor the proliferation of opportunistic pathogens.

Specific Bacteria Associated with Colorectal Adenoma and Cancer Development
Numerous studies have identified tumour-specific bacteria present in colorectal mucosal and/or faecal samples, and not detectable in healthy controls or tumour tissue versus the bordering healthy mucosa [90] (Table 1). A metagenome-wide association study (MGWAS) on stools from advanced adenoma and CRC patients and from healthy individuals, detected microbial genes, strains and functions enriched in each group. High consumption of red meat relative to fruits and vegetables seems to be associated with development of specific bacteria that could contribute to a more hostile intestinal milieu [91]. In general, microbial species associated with CRC development are represented by specific strains of Escherichia coli, Streptococcus gallolyticus, Bacteroides fragilis, Fusobacterium nucleatum, and Enterococcus faecalis among others [16]. Dysbiosis and other factors contributing to the adenoma-carcinoma progression. The adenoma-carcinoma progression may occur because of the genomic instability caused by alterations in the gut microbiota. These changes may be supported by diet and lifestyle, which promote dysbiosis, inflammatory state and epithelial DNA damage, thus contributing to CRC development. The carcinogenesis leads to gut niche changes, which may favor the proliferation of opportunistic pathogens.
Changes of the microbiota profile in adenomas could enhance the production of primary and secondary bile acids, as well as sucrose, lipid, starch, and phenylpropanoid metabolism, thus supporting an intestinal environment that favors the growth of bile-resistant and sulfidogenic microorganisms including Desulfovibrio and Bilophilia [81,82].
It is well recognized that hydrogen sulfide (H 2 S) generated by bacteria in the gut is related to adenoma development and eventually CRC [83]. Many anaerobic bacterial strains such as Salmonella enterica, Clostridia, Escherichia coli, and Enterobacter aerogenes are able to convert cysteine to H 2 S, ammonia and pyruvate by cysteine desulfhydrase; moreover some gut bacteria (i.e., Escherichia coli, Salmonella, Enterobacter, Staphylococcus, Bacillus, Klebsiella, Corynebacterium, and Rhodococcus) may generate H 2 S by sulfite reduction [84]. H 2 S modulates inflammation, ischemia and/or perfusion injury and motility, and exerts a toxic activity on the colonic epithelium [85]. Phenolic substances such as amines, N-nitroso compounds (NOCs) found in processed meat, may also exert toxic activities favoring carcinogenesis [86,87].
Colibactin is a genotoxin produced by certain strains of bacteria, such as B2 phylogroup E. coli strains that colonize the human gut [88]. The synthesis of colibactin by the polyketide synthetase (pks) genomic island, especially in members of the family Enterobacteriaceae, may lead to chromosomal instability and DNA damage in eukaryotic cells, apoptosis of immune cells, and in turn the development of CRC [89].

Specific Bacteria Associated with Colorectal Adenoma and Cancer Development
Numerous studies have identified tumour-specific bacteria present in colorectal mucosal and/or faecal samples, and not detectable in healthy controls or tumour tissue versus the bordering healthy mucosa [90] (Table 1). A metagenome-wide association study (MGWAS) on stools from advanced adenoma and CRC patients and from healthy individuals, detected microbial genes, strains and functions enriched in each group. High consumption of red meat relative to fruits and vegetables seems to be associated with development of specific bacteria that could contribute to a more hostile intestinal milieu [91]. In general, microbial species associated with CRC development are represented by specific strains of Escherichia coli, Streptococcus gallolyticus, Bacteroides fragilis, Fusobacterium nucleatum, and Enterococcus faecalis among others [16].
Hale et al. observed significant abundances of multiple taxa in subjects with adenomas, such as Bilophila, Desulfovibrio, pro-inflammatory bacteria in the genus Mogibacterium, and Bacteroidetes spp. On the other hand, Veillonella, Firmicutes (class Clostridia), and Actinobacteria (family Bifidobacteriales) were more represented in patients without adenomas [81].
A study by Peters et al. analyzed for the first time the link between the gut microbiota and specific colorectal polyp types in 540 subjects, and showed that conventional adenomas (CA) cases had lower species diversity in faeces compared to controls (p = 0.03), especially with regard to advanced CA cases (p = 0.004). Only subjects with distal or advanced CA showed significant differences in general microbiota composition compared to controls (p = 0.02 and p = 0.002). Faeces of CA cases were characterized by the reduction in Clostridia from families Ruminococcaceae, Clostridiaceae, and Lachnospiraceae, and the increase in the classes Gammaproteobacteria and Bacilli, order Enterobacteriales, and genera Streptococcus and Actinomyces. There were not significant differences between sessile serrated adenoma (SSA) and hyperplastic polyps (HP) cases in diversity or composition compared to controls [92].
Feng et al. detected a great amount of Bacteroides and Parabacteroides, together with Bilophila wadsworthia, Lachnospiraceae bacterium, Alistipes putredinis, and Escherichia coli in CRC compared with both healthy and advanced adenoma. Also, gut commensals such as Bifidobactium animalis and Streptococcus thermophilus, were diminished in stools from adenoma or CRC patients, thus highlighting a divergence from healthy microbiota. Patients with advanced adenoma or CRC seem to be lacking in lactic acid-producing commensals such as Bifidobacterium that could facilitate epithelium regeneration and inhibition of opportunistic pathogens [91]. Association between S. bovis bacteremia and CRC risk.

Fusobacterium nucleatum
F. nucleatum is an oral symbiont, and opportunistic pathogen that has been detected in intestinal cancers [93,94]. F. nucleatum may enhance CRC carcinogensis by stimulating the production of interleukin (IL)-17F/21/22/23/31/cluster of differentiation (CD)40L and protein expression of phospho-STAT3 (p-STAT3), p-STAT5, and phospho-extracellular regulated protein kinases (p-ERK)1/2 [95]. A great amount of Fusobacteria has been observed in SSA [108,109]; a study by Yu et al. reported that the prevalence of invasive Fusobacteria within proximal SSAs (78.8%) and HPs (65.7%) was significantly more elevated than that of proximal and distal traditional adenomas (28.9% and 24.4% respectively; p < 0.05) [96]. The presence of F. nucleatum has been associated with poor prognosis in CRC patients and development of chemoresistance [97,98]. F. nucleatum binds E-cadherin in the clonic epithelium and stimulates colorectal carcinogenesis through the fusobacterial adhesin FadA [110,111]. The interplay between Gal-GalNAc, a host polysaccharide, with fusobacterial lectin (Fap2) may promote the increase of F. nucleatum in colorectal adenoma and cancer [112]. A study by Mima et al. showed that multivariable hazard ratios (HRs) for CRC-specific mortalityin F. nucleatum-low subjects and F. nucleatum-high subjects, compared with F. nucleatum-negative subjects, were 1.25 (95% C.I. 0.82 to 1.92) and 1.58 (95% C.I. 1.04 to 2.39), respectively (p for trend = 0.020). The quantity of F. nucleatum was correlated with microsatellite instability (MSI)-high (multivariable odd ratio (OR), 5.22; 95% CI 2.86 to 9.55) independent of the presence of CIMP and BRAF mutation. A significant association between CIMP and BRAF mutation with F. nucleatum was observed only in univariate analyses (p < 0.001) but not in multivariate analysis that adjusted for MSI status [97].
Yang et al. observed that an infection of CRC cells lines (HCT116, HT29, LoVo, and SW480) with F. nucleatum increased cell growth, invasiveness, and capability to form xenograft cancers in mice. F. nucleatum promoted Toll-like receptor 4 (TLR4) signaling to myeloid differentiation factor 88 (MYD88), activating NFκB signaling pathways and increasing the expression of microRNA-21 (miR21), which reduced the levels of the RAS GTPase p21 protein activator 1 (RASA1). Shorter survival times were observed for tumors with high amounts of F. nucleatum DNA and miR21 [113].
It has been also observed that F. nucleatum may promote LC3-II protein expression, autophagy pathway, and autophagosome production in CRC cells. F. nucleatum may favor the release of the autophagy-related proteins, pULK1, ULK1, and ATG7, contributing to the resistance to oxaliplatin and 5-fluorouracil regimens in CRC cells [98].
A study by Bullman et al. showed the persistance of F. nucleatum also in distal metastatic lesions of CRC patients. Administration of metronidazole in mice bearing a colon cancer xenograft decreased F. nucleatum load, tumor cell proliferation, and overall cancer development, thus suggesting that specific antibiotics could potentially be used to treat patients with Fusobacterium-associated CRC [114].

Streptococcus gallolyticus (Formerly S. bovis)
Streptococcus gallolyticus subsp. gallolyticus (SGG), formerly known as S. bovis biotype I, represents a common causative agent for bacteremia and endocarditis in older adults. Gut colonization by SGG is strongly correlated with the development of CRC [99,115]. Indeed, both American and European guidelines recommended colonoscopy in patients with SGG bacteremia [116,117].
A large epidemiological study by Butt et al. showed for the first time a statistically significant association between exposure to SGG antigens and CRC, and pointed out that the risk for CRC was stronger among subjects younger than 65 years [101].
Aymeric et al. observed that CRC-specific conditions may favor SGG colonization of the gut at the expense of commensal enterococci. Indeed, gut colonization by SGG is promoted by a bacteriocin called "gallocin", which is enhanced by bile acids and may exert toxic activity to enterococci. Also, the stimulation of the Wnt pathway, and the reduced expression of the bile acid apical transporter gene Slc10A2, may act on the APC founding mutation, supporting the gut colonization by SGG [115].

Enterotoxigenic Bacteroides fragilis (ETBF)
Enterotoxigenic B. fragilis (ETBF) may support colorectal carcinogenesis by the production of pro-inflammatory cytokines and the stimulation of Wnt signaling. Expression of B. fragilis toxin (BFT), a 20 kDa metalloprotease produced by ETBF, is able to promote persistent colitis in mice, damage E-cadherin junctions, as well as stimulate B-catenin signaling and IL-8 production in colonic epithelial cells [118].
A study by Purcell et al. underlined the key role of ETBF in the development of colorectal low-grade dysplasia, tubular adenomas, and serrated polyps (p-values of 0.007, 0.027 and 0.007, respectively) [102]. Similar findings were reported in a study of patients with colonic adenomas that presented higher expression of the B. fragilis toxin gene (bft) associated with adenoma tissue compared to normal healthy mucosa [103].
Zamani et al. reported an increased positivity of ETBF in patients with precancerous and cancerous lesions compared to healthy controls. Higher ORs of ETBF were significantly associated with serrated lesions and adenoma with low-grade dysplasia. The most common subtype of bft gene was the bft1 gene, followed by the bft2 gene. An assessment of ETBF could represent a marker of CRC prognosis, especially in the precancerous lesions, and could be used for the screening of these conditions [104].

Enterococcus faecalis
E. faecalis is a Gram-positive commensal bacterium, that may be responsible for human disease through translocation from intestinal wall, oral cavity, and genito-urinary mucosa, leading to a systemic infection [119]. E. faecalis represents one of the most frequent causes of infection in older adults, and some studies underlined its importance for the development of cancer [120]. It has also been reported an association between enterococcal endocarditis and hidden CRC [119,121]. On the other hand, E. faecalis showed anti-inflammatory properties and probiotic activity, and is frequently administered in subjects with chronic sinusitis and bronchitis or in infant acute diarrhea [122].
Actually, there is no consensus on the role of E. faecalis in CRC: some studies highlighted its protective role or no role in CRC, whereas others reported potential pro-carcinogenic effects [123].
A study by Viljoen et al. carried out on 55 patients, did not highlight any significant clinical association between E. faecalis and CRC. However, the same study showed a relevant association bewteen clinicopathological features of CRC and Fusobacterium spp. and ETBF [105]. Miyamoto et al. observed that heat-killed E. faecalis strain EC-12 could suppress intestinal polyp development in Apc mutant Min mice. Administration of heat-killed EC-12 reduced the levels of c-Myc and cyclin D1 mRNA expression in intestinal polyps, by blocking the transcriptional activity of the T-cell factor/lymphoid enhancer factor [124].
E. faecalis could play a role in inducing CRC by activation of Wnt/β-catenin signaling and induction of pluripotent transcription factors linked to dedifferentiation. Indeed, exposure of murine primary colon epithelial cells to E. faecalis-infected macrophages contributed to CRC initiation through gene mutation, chromosomal instability, and endogenous cell transformation, which involved the transcription factors c-Myc, Klf4, Oct4, and Sox2i [125].
Perhaps, these controversial data could be explained taking into account the different geographical origin of the isolated strain, and dysbiosis due to the use of antibiotics or changes in diet [126,127].

Escherichia coli
Classification of the Gram negative bacterium E. coli includes 8 phylogenetic groups (A, B1, B2, C, D, E, F and clade I). Commensal strains are commonly represented by A and B1 groups, being the largest part of the fecal flora of healthy individuals. Extraintestinal pathogenic strains (ExPEC) include mainly B2 and D groups, and may be responsible for many extraintestinal infections, due to the achievement of numerous virulence factors that potentially support the colonization of extraintestinal tissues [128]. However, both commensals and ExPEC are considered as a part of the normal gut microbiota in healthy subjects [129].
There is evidence that E. coli could play a role in the development of CRC [106,130]. Indeed, some patients with CRC may show an excessive growth of E. coli strains, mainly B2, characterized by high expression of virulence genes, including those encoding toxins and effectors that may induce carcinogenesis, such as colibactin, cytolethal distending toxins, cytotoxic necrotizing factors, and cycle-inhibiting factor [131,132]. In vitro studies showed that colibactin could be involved in DNA alkylation on adenine residues, leading to double-strand breaks [133,134]. Pleguezuelos-Manzano et al. demonstrated that exposure to genotoxic pks + E. coli, could be responsible for specific mutational signature in human intestinal organoids; indeed, an identical mutational signature was observed in 5876 human cancer genomes from two independent study cohorts, mostly in CRC [135].
Ambrosi et al. analyzed 272 E. coli isolates from colonoscopy biopsies, and showed that E. coli strains colonizing adenomatous polyps were characterized by specific phenotypes compared to those from normal mucosa, which included lack of motility, moderate to strong biofilm forming activity, and poor proteolytic capability [106].
In a study by Iyadorai et al. pks + E. coli was detected more frequently in CRC patients compared to healthy subjects. In vitro assays carried out on primary colon epithelial (PCE) and CRC (HCT116) cell lines, highlighted that the cytopathic effect of pks + E. coli strains could support the initiation and development of CRC [107].

Future Perspectives
Modulation of the gut microbiota, aiming to reverse microbial dysbiosis, could represent a new tool for prevention and treatment of CRC. The strategies could include the use of probiotics, prebiotics, postbiotics, antibiotics, and fecal microbiota transplantation (FMT) [136][137][138][139].
Overall, the effects of microbiota modulation on CRC prevention could be due to many mechanisms, such as the suppression of inflammatory state, stimulation of apoptosis of early cancer cells, re-establishment of intestinal barrier function and correction of microbiota composition [140,141]. Also, manipulation of the gut microbiota could alleviate chemotherapy-induced side effects, such as mucositis, as confirmed by a decreased incidence of diarrhea and weight loss after the administration of several probiotics strains in animal models [142,143].
There is growing evidence that modifications of microbial abundances in some pathological conditions could affect their co-abundance interactions; indeed, Chen et al. observed specific gut microbial co-abundance networks in patients with inflammatory bowel disease (IBD) and obesity. These findings underlined the importance of microbial dysbiosis in the pathogenesis of some diseases, and suggested that even the development of CRC could share similar mechanisms [144][145][146].
Promising preclinical studies suggested that modulation of gut microbiota could increase therapeutic efficacy of anticancer drugs. There is evidence that the administration of antibiotics could lead to clinical benefits to CRC patients by gut microbiota depletion and subsequent reduction of chemotherapeutic resistance. Indeed, a study by Geller et al. observed that intratumor bacteria could favor gemcitabine resistance through enzymatic inactivation, and therefore the administration of a gemcitabine-ciprofloxacin combination therapy could enhance the efficacy of chemotherapy [147].
Some studies demonstrated that the gut microbiota is also able to affect chemotherapy and/or immunotherapy efficacy by modulating immune response [148]. Oral administration of some probiotics, such as Bifidobacterium spp. and Akkermansia muciniphila, or FMT from treatment-responsive patients, stimulated the programmed cell death protein 1 ligand 1 (PD-L1)-based immunotherapy, thus blocking cancer development through the increase of dendritic cell and T cell response [149][150][151].
There is growing evidence that microbial shift markers could be used succesfully for non-invasive early diagnosis and/or prognostic assessment of CRC and advanced adenomas [81,152]. Mangifesta et al. performed a metataxonomic analysis based on 16S rRNA gene sequencing approach, and showed that some microbial taxa such as Bacteroides, Faecalibacterium, and Romboutsia, seem to be reduced in cancerogenic mucosa and in adenomatous polyps, thus representing potential new biomarkers of early carcinogenesis. Furthermore, the detection of high amounts of F. nucleatum in polyps, underlined the key role of this microorganism as a microbial biomarker for early diagnosis of CRC [153].
A study by Hale et al. showed that the composition of the gut microbiota in subjects with adenomas is significantly different from that of healthy subjects, and is similar to the microbiota of subjects with CRC. These changes could be a consequence of the Western diet and could result in metabolic changes leading to intestinal cellular damage and mutagenesis [81,154].
The combined assessment of heterogeneous CRC cohorts detected reproducible microbiota biomarkers and disease-predictive models that could represent useful tools for clinical prognostic tests and future research. A meta-analysis of 969 stool metagenomes carried out using data from five open access datasets and two new cohorts, showed that the gut microbiota in CRC was characterized by more richness than controls (p < 0.01), partly due to the growth of some species originating from the oral cavity. The results also highlighted an association between gluconeogenesis, putrefaction and fermentation processes with CRC, while the starch and stachyose degradation were associated with controls. A significant association between microbiota choline metabolism and CRC was also observed (p = 0.001) [155]. Another meta-analysis of eight stool metagenomic studies of CRC (n = 768) from different geographical areas, reported a significant enrichment in a group of 29 species in CRC metagenomes (FDR < 1 × 10 −5 ). An elevated production of secondary bile acids from CRC metagenomes, higher expression of mucin and protein catabolism genes and reduction of carbohydrates degradation genes were observed, thus underlying a metabolic relationship between gut microbiota in CRC and a diet rich in meat and fat [156].
A study by Poore et al. carried out on The Cancer Genome Atlas (TCGA) detected specific microbial signatures in blood and tissue of different types of tumors, including CRC, which were predictive for patients with stage Ia-IIc tumor and tumors without any genomic modifications as detected by cell-free tumor DNA assessment. These findings could pave the way to a novel type of microbial-based CRC diagnostics [157].
Currently, there is a great limitation in availability of mouse models to study the interaction between gut microbiota and CRC. Zeb2 IEC-Tg/+ (intestinal epithelial cell-specific transgenic expression of the epithelial-to-mesenchymal transition regulator Zeb2) mice represented the first and only microbiota-dependent CRC mouse model available so far. Specific characteristics of Zeb 2IEC-Tg/+ mice included the presence of gut dysbiosis, and the preventive effect on carcinogenesis through the microbiota reduction by broad-spectrum antibiotics or germ-free rederivation [158].

Conclusions
In conclusion, detecting key relationships between diet, gut microbiota, and metabolites involved in the adenoma-carcinoma sequence could provide important basis for personalized medicine aimed at preventing and managing CRC. Secondary bile acids, H 2 S, and other bacterial metabolites could exert genotoxic activities and should be kept into account when investigating the adenoma and carcinoma development. Nonetheless, further studies are needed to evaluate the effects of diet, lifestyle, or medications on the gut metabolic environment and the microbiota. Finally, the identification of global microbiota signatures specific for CRC represents a promising tool in CRC diagnosis and therapy.