The human gastrointestinal microbiota comprises the bacteria, archaea and eukarya species colonizing the human gut. Moreover, the total amount of genetic material exceeds that of the genetic material of the human host [1]. There are more than 1000 presently known species, distributed mainly among the phyla Bacteroidetes and Firmicutes (around 90%) and the rest among Proteobacteria, Verrucomicrobia and Actinobacteria [2]. However, the human gut microbiota demonstrates high geographical, interpersonal and age variability ^[1][2][1,2]. Moreover, its genetic and metabolomic abundance comprises a more extensive set of genes than the entire human genome [3]. The latter justifies the numerous metabolic and immunological interactions between the gut microbiota and the host, with variations in plasma metabolites and the host immune system activity correlating to the gut microbiota signature ^[4][5][6][4,5,6]. These host–microbiota interactions have local and systemic impacts, including on the pathological processes of carcinogenesis.

It has been demonstrated that carcinogenesis could be promoted or suppressed by the interaction between the microbiome and the host. For example, a considerable amount of evidence for the role of the gut microbiota and intestinal dysbiosis in the development of colorectal cancer exists ^[7][8][7,8]. There is also mounting evidence for the significance of gut microbiota for other non-intestinal locations of carcinogenesis, such as breast and lung cancer ^[9][10][9,10].

While the potential gut microbiota signatures are endless and numerous correlations can be made, in our opinion, it is convenient to approach the complex interactions between gut microbiota and cancer from the viewpoint of the major oncogenic and tumor-suppressive pathways involved.

2. Local Oncogenic Effects of Gut Microbiota

The gut microbiota can promote local colonic oncogenesis through the production of cancerogenic metabolites, oncogenic exotoxins and the axis of chronic inflammation, including biofilm production, pathogenic adhesins stimulation and local immune system mediation ^[11][12]. For example, the enterotoxigenic Bacteroides fragilis strains (EBFT) produce one of the three isotypes of a 20 kDa zinc metalloprotease (BFT-1, BFT-2 and BFT-3). BFT modify the permeability of the colonic epithelium, inducing colitis, and promote cell proliferation through the NF-kB and the MAPK pathways ^[12][13]. In addition, it was revealed in a mouse model that BFT enhance the oncogenesis of colorectal cancer (CRC) through the Th17 pathway, and EBFT has been linked to an increased risk of CRC in human patients ^[11][12][12,13]. Recently, one team demonstrated that upregulated IL-6 is crucial for both inflammatory bowel diseases and CRC development, whereas Th17/T regulatory (Treg) cells and related genes are activated primarily in CRC ^[13][14]. It has been suggested that chronic inflammation mediated by Th17 cells and related cytokines is associated with an increased risk of malignant transformation. Another prominent group of microbiota toxins interferes with the DNA and cell cycle of the intestinal epithelium. Colibactin is a byproduct of the pks genetic island in some Enterobacteriaceae, particularly Escherichia coli. It is an unstable compound, which, when introduced directly on the mucosal surfaces or intracellularly, causes DNA alkylation, interstrand crosslinking and double-strand breaks ^[14][15][15,16]. Then, DNA damage translates into mutagenesis and carcinogenesis potential, which has been demonstrated on epithelial cell lines and linked to colon cancer risk ^[16][17][17,18]. A Swedish study discovered colibactin-producing bacteria in 56% of the CRC samples versus 19% of control samples ^[18][19]. The cytolethal distending toxin (cdt) is another potent bacterial toxin isolated initially from E. coli and consequently described in Shigella, Campylobacter and other Gram-negative bacteria. It can be coded both on the bacterial chromosome, but also on plasmid vectors. Cdt is a virulence factor that enhances the invasive properties of its carriers through damage to the epithelial layer of the intestinal mucosa. Cdt also extensively suppresses lymphocytes and macrophages ^[19][20][21][20,21,22]. The mechanism of action of cdt relies on its structural similarity to the human DNAse I protein family that leads to cell cycle damage through the infliction of double-strand breaks. The described processes are often observed in mutagenesis and carcinogenesis ^[22][23]. It has been demonstrated in animal models that mice bearing cdt-positive Campylobacter jejuni are more likely to develop CRC and larger tumors ^[23][24]. The typhoid toxin (TT) found in Salmonella also induces double strand breaks in a similar manner, as the active part of TT actually comprises the cdt B subunit. However, TT is unique in being directly linked to an increased risk of developing CRC in humans. A recent Dutch study demonstrated a statistically significant increase in standardized incidence risk for CRC development after salmonellosis infection. The overall risk increased by 1.54 times for the age group above 20 and below 60 years, while for the age groups 20–39 and 40–49, the increases were 2.55 and 1.62 times, respectively ^[24][25][25,26]. Gut microbiota can also promote oncogenesis in a non-DNA-related way; intestinal commensals can promote inflammation-related epithelial cell proliferation in an IL-6 STAT3-dependent way, in IL-17C dependent way or through the PI3K-Akt-axis ^[26][27]. These proinflammatory pathways relate to specific virulence factors enhancing biofilm production, specific adhesins and immune cell recruitment in the tumor stroma. For example, Fusobacterium nucleatum was related to IL-17A-related inflammation in human CRC, and Peptostreptococcus anaerobius was described to induce a proinflammatory response in the tumor stroma, contributing to tumor progression via the recruitment of tumor-infiltrating immune cells ^[26][27][27,28]. The above mechanisms are further implied by novel sequencing techniques suggesting there are specific gut microbiota signatures in CRC patients. A few studies have demonstrated the loss of alpha diversity of the gut microbiota of CRC patients ^[28][29][30][29,30,31]. A large meta-analysis of 386 samples from CRC patients and 392 tumor-free controls revealed heterogenic data on the alpha diversity but demonstrated the prevalence of 29 bacterial species in the CRC samples, among which were the already discussed Fusobacterium and Peptostreptococcus ^[31][32]. The same study also demonstrated significant enrichment in the pks gene already discussed above and also of the fadA gene (and adhesin and virulence factor of F. nucleatum). Another large study demonstrated a correlation between a western style diet and high levels of pks positive E. coli and CRC ^[32][33]. Additionally, while some of the largest of these studies demonstrate significant correlations in samples from cancer patients versus healthy controls, further prospective long-term studies will be needed to further elaborate on the role of metagenomic profiles and their contribution to the development of CRC and to exclude the potential reciprocal relationship where CRC impacts the local flora.