Intestinal Host-Microbiome Interactions

A human body contains at least tenfold more bacteria cells than human cells and the most abundant and diverse microbial community (also known as microbiota or microbiome) resides in the large intestine (colon). It is estimated that this colonic microbiome is composed of ∼1014 bacterial cells, comprising >103 species (Dethlefsen et al., 2006; Qin et al., 2010). Intestinal microbiomes differ from individual to individual but remain relatively stable during adult life (Green et al., 2006; Arumugam et al., 2011). The resident microbiome provides the host with core functions that are essential for digestion of food and control of intestinal epithelial homeostasis. Conversely, an increasing body of evidence supports a relationship between infective agents and human colorectal cancer (CRC) by production of DNA damaging metabolites or toxins, and the induction of cell proliferation and procarcinogenesis pathways by a subpopulation of the intestinal microbiota. It could be speculated that the intrinsic intestinal microbiome of a certain individual may contain an unfavorable number of disease-inducing bacteria. On the long term, their activities may override the health-promoting activities of the commensal bacterial population. On the other hand, the dramatic physiological alterations that result from colon carcinogenesis itself (Hirayama et al., 2009) disturbs the local intestinal microenvironment and causes (local) shifts in the microbiota composition and provides a portal of infection for certain opportunistic pathogens. The latter phenomenon could explain why some uncommon bacterial infections are often associated with CRC. In this chapter we will discuss the mechanisms by which intestinal bacteria may drive the initiation and progression of sporadic CRC, but also the driving forces of intestinal carcinogenesis on local microbial dysbiosis and the consequences thereof will be reviewed.


Introduction
A human body contains at least tenfold more bacteria cells than human cells and the most abundant and diverse microbial community (also known as microbiota or microbiome) resides in the large intestine (colon).It is estimated that this colonic microbiome is composed of ∼10 14 bacterial cells, comprising >10 3 species (Dethlefsen et al., 2006;Qin et al., 2010).Intestinal microbiomes differ from individual to individual but remain relatively stable during adult life (Green et al., 2006;Arumugam et al., 2011).The resident microbiome provides the host with core functions that are essential for digestion of food and control of intestinal epithelial homeostasis.Conversely, an increasing body of evidence supports a relationship between infective agents and human colorectal cancer (CRC) by production of DNA damaging metabolites or toxins, and the induction of cell proliferation and procarcinogenesis pathways by a subpopulation of the intestinal microbiota.It could be speculated that the intrinsic intestinal microbiome of a certain individual may contain an unfavorable number of disease-inducing bacteria.On the long term, their activities may override the health-promoting activities of the commensal bacterial population.On the other hand, the dramatic physiological alterations that result from colon carcinogenesis itself (Hirayama et al., 2009) disturbs the local intestinal microenvironment and causes (local) shifts in the microbiota composition and provides a portal of infection for certain opportunistic pathogens.The latter phenomenon could explain why some uncommon bacterial infections are often associated with CRC.In this chapter we will discuss the mechanisms by which intestinal bacteria may drive the initiation and progression of sporadic CRC, but also the driving forces of intestinal carcinogenesis on local microbial dysbiosis and the consequences thereof will be reviewed.

Intestinal microbiome
The colonic epithelium is the first line of defense against enteric antigens and bacteria.In a healthy colon, the epithelial barrier regulates uptake of nutrients and limits uptake of potential toxic substances and infectious agents (Chichlowski & Hale, 2008).Goblet cells are specialized epithelial cells within the mucosa that produce a viscous mucus layer that covers the intestinal epithelium (Heazlewood et al., 2008).This mucus layer is thick and consists of an inner firmly attached layer, that excludes bacteria from direct contact with the (NTBF) showed no increased tumor formation compared to conventional mice (Housseau & Sears, 2010).These data clearly show that the intestinal microbial population has a strong promoting effect on tumor progression in mice that have a genetic predisposition for developing intestinal adenomas and that certain species within the intestinal microbiota contribute more than average to this process.

Stimulation of TLR signaling
A balanced immune stimulation to commensal and pathogenic bacteria is crucial for a healthy intestinal tract.Toll-like receptors (TLRs) are proteins that activate immune responses towards potentially harmful pathogens upon sensing of pathogenic substances, such as cell wall components.However, chronic overstimulation of these responses may be detrimental by leading to the initiation and progression of CRC (Fukata & Abreu, 2007).A direct impact of bacteria on the development of CRC through the TLR5/MyD88 pathway was demonstrated in germ-free and gnotobiotic mice.These animal experiments revealed that MyD88 -/-knock-out mice that were treated with the carcinogen azoxymethane (AOM) failed to develop colorectal tumors when these mice were subjected to bacteria.In contrast, control mice rapidly developed CRC upon bacterial colonization of their intestinal tract.These results implicate that TLR/MyD88 signaling is a prerequisite for the development of CRC (Uronis et al., 2009).In addition, it was shown that tumors in Apc min/+ MyD88 -/-mice were significantly smaller than those found in Apc min/+ mice (Rakoff-Nahoum & Medzhitov, 2007).Another study showed that TLR4 -/-mice were partly protected against the development of neoplasia by tumor-inducing chemical agents (Killeen et al., 2009).Additional evidence was presented that TLR4 signaling can promote colon carcinogenesis by stimulating tumor infiltration of Th17 cells (T-helper cell subset that produces IL-17) through the increased production of pro-inflammatory signals (Su et al., 2010).It can be envisaged that bacterial TLR4 ligands, such as LPS, play an important role in this increased chemotactic activity of tumor cells (Scanlan et al., 2008).Importantly, Th17 cells have directly been implicated in the pathogenesis of Enterotoxigenic Bacteroides fragilis-induced CRC (Housseau & Sears, 2010;Wu et al., 2009).Thus, although TLR signaling is important for the effective clearance of harmful pathogens and can mediate anti-tumor cell responses, chronic TLR activation may tip the delicate balance towards tumor-promoting activities (Rakoff-Nahoum & Medzhitov, 2009).Altogether, the above mentioned studies indicate that chronic bacterial stimulation of inflammatory pathways at malignant sites promotes, and may even be a prerequisite for, intestinal tumor development.

Upregulation of COX-2
Cyclooxygenase-2 (COX-2) is one of the key players in the progression of CRC.The expression of COX-2 is highly elevated in colonic tumors and correlated with disease stage and stimulates cell proliferation and pro-inflammatory pathways by the production of prostaglandins (Menter et al., 2010).Human intervention studies have clearly shown that the usage of Non-Steroidal Anti-Inflammatory Drugs (NSAIDS) can reduce CRC risk by as much as 75% (Eaden et al., 2000;Labayle et al., 1991;Thun et al., 1991).Evidence for bacterial involvement in the upregulation of COX-2 during CRC development was gained through animal and in vitro studies.First, superoxide radicals produced by Enterococcus faecalis were shown to upregulate the expression of COX-2 in hybrid hamster cells containing human chromosomes, as well as in macrophages (Wang & Huycke, 2007).Furthermore, macrophages that were pre-treated with a COX-2 inhibitor and subsequently exposed to E. faecalis totally inhibited the induction of chromosome instability (CIN) in these hybrid hamster cells.Second, an animal study published by Ellmerich et al. (2000b) indicated that Streptococcus bovis biotype II.1 (Streptococcus infantarius) could also play a role in the progression of CRC through induction of the COX-2 pathway.These investigators employed a rat model in which pre-treatment with azoxymethane (AOM) induced preneoplastic aberrant crypt foci (ACF).When such rats were co-exposed to S. infantarius or cell wall antigens from this bacterium, the number of ACF increased drastically and also adenomas were found, whereas the latter were totally absent in the control mice treated with AOM alone.In addition, the production of the pro-inflammatory cytokine IL-8 in the mucosa of rats exposed to S. infantarius was increased.This finding is in accordance with in vitro studies on epithelial Caco-2 cells that release both IL-8 and PGE2 upon incubation with S. infantarius (Biarc et al., 2004).Moreover, Abdulamir et al. (2010) have recently shown that increased COX-2 and IL-8 expression was associated with the presence of Streptococcus gallolyticus (S. bovis biotype I) in human colon tumor tissue.However, IL-8 expression was not increased in non-malignant tissue that contained S. gallolyticus.Together these studies indicate that COX-2 induction is associated with both tumor development and exposure to bacterial stimulants.

Toxin-induced promotion of cell proliferation
Enterotoxigenic Bacteroides fragilis (ETBF) has been implicated in the promotion of CRC through inflammatory pathways.B. fragilis is a normal inhabitant of the gastrointestinal tract, but its enterotoxigenic form is only present in approximately 20% of the healthy population (Sears, 2009).ETBF produces the B. fragilis toxin that degrades E-cadherin in epithelial cells, which causes β-catenin to migrate towards the nucleus where it can activate cell proliferation pathways (Wu et al., 2003).Consequently, APC min/+ mice colonized with ETBF were shown to suffer from increased tumor burden compared to control mice colonized with non-toxigenic B. fragilis (NTBF) strains (Housseau & Sears, 2010;Wu et al., 2009).Importantly, Wu et al (2009).showed that this increased tumor burden was mediated through the increased expression of STAT3 that leads to a Th17 response.Importantly, increased tumor formation could be blocked by anti-IL17 therapy.These experiments clearly show that induction of a STAT3/Th17-dependent pathway for inflammation, leads to inflammation-induced cancer by ETBF in a mouse model.Since ETBF is a quite common bacterium in the gastro-intestinal tract, this finding could have major implications for the role of these bacteria in the development of CRC in the human population.This idea is further corroborated by the fact that patients with CRC have indeed increased carriage rates of ETBF compared to NTBF (Toprak et al., 2006).It should be realized that this mechanism of tumor induction could also be associated with other toxigenic intestinal bacterial strains.

Toxin-induced DNA damage
Certain E. coli strains can induce increased mutation rates in eukaryotic cells as demonstrated by Cuevas-Ramos and colleagues (2010).Their experiments showed that E. coli strains harboring the pks island caused DNA damage in human epithelial cells and in an ex vivo mouse intestinal model by the induction of single strand breaks and activation of DNA damage signaling pathways.The pks gene cluster codes for nonribosomal peptide synthetases and polyketide synthetases (pks) that synthesize a genotoxin named Colibactin.The pks island is commonly present in about 34% of commensal E. coli isolates.Upon infection of epithelial cells with physiological concentrations of pks + strains, initial DNA damage occurred.Furthermore, it was shown that cells continued to proliferate in the presence of DNA damage after E. coli infection, resulting in an increased mutation frequency (Cuevas-Ramos et al., 2010).These studies suggest that pks + strains of E. coli could be involved in the initiation and progression of CRC.As, E. coli is generally regarded as a normal commensal inhabitant of the gastro-intestinal tract, Bronowski and co-workers investigated the differences between E. coli strains collected from healthy individuals and CRC patients (Bronowski et al., 2008).These experiments showed that a subset of E. coli strains recovered from CRC tissue shared pathogenicity islands, encoding an alfa haemolysin and a cytotoxic necrotizing factor, with uropathogenic E. coli strains.This suggests that besides Colibactin production, other virulence characteristics may also mediate the tumor promoting capacity of E. coli pks + strains.

Metabolite-induced DNA damage
Sulfate reducing bacteria use sulfate as energy source by converting it to sulfide and hydrogen sulfide (H 2 S) in the human colon.The genotoxic potential of H 2 S is in part mediated by oxidative free radicals, which results in increased levels of DNA damage in cultured epithelial cells (Attene-Ramos et al., 2006;Attene-Ramos et al., 2007;Attene-Ramos et al., 2010).Furthermore, exposure to H 2 S may disrupt the balance between apoptosis, proliferation and differentiation (Cai et al., 2010;Deplancke & Gaskins, 2003).Interestingly, also COX-2 was shown to be upregulated in epithelial cells after H 2 S treatment at physiological concentrations, probably through generation of reactive oxygen species (Attene-Ramos et al., 2010).Increased fecal H 2 S concentration was implicated as a risk factor for the development of colonic neoplasia in a clinical study (Kanazawa et al., 1996).Whether these increased H 2 S levels originates from increased activity of sulfate reducing bacteria and/or reduced epithelial capacity to degrade H 2 S remains to be investigated.E. faecalis was also found to produce extracellular superoxide in colonic tissue of rats, which is the result of dysfunctional microbial respiration (Huycke et al., 2002).These rats produced up to 25-fold increased concentrations of hydroxylated aromatic metabolites in urine than rats colonized with a closely-related strain.Importantly, superoxide can be converted to hydrogen peroxide, which has the potential to diffuse into epithelial cells and cause DNA damage.In an in vitro setup, it was shown that the formation of DNA adducts by E. faecalis was mediated by activated COX-2 expression in macrophages that in turn promoted DNA damage in epithelial target cells (Wang & Huycke, 2007;Wang et al., 2008).Since COX-2 induction has a clear clinical association with CRC, this might indicate that superoxideproducing bacteria have a contributing role in disease development.This notion is further underscored by the finding that E. faecalis fecal carriage was increased in CRC patients, whereas the number of butyrate producing bacteria was decreased (Balamurugan et al., 2008).However, no clinical evidence has been presented that associates superoxide producing enterococci with adenomas or CRC (Winters et al., 1998).This clearly indicates that, although the in vitro data and animal studies strongly suggest that oxygen radicals from bacterial origin could play an important role in CRC initiation or progression, the clinical impact of these findings remains to be properly examined in well-designed clinical studies (Huycke & Gaskins, 2004).Bacteroides species produce fecapentaenes that are potent mutagens that have been shown to alkylate DNA, which leads to mutagenic adducts.Some evidence points towards a mechanism in which oxygen radicals cause oxidative damage to DNA (Hinzman et al., 1987;Povey et al., 1991;Shioya et al., 1989).Fecapentaenes appear in relatively high concentrations in human feces, however, no significant differences in fecapentaene levels were found in feces from CRC patients and controls (Schiffman et al., 1989).In view of their mutagenic potential, however, fecapentaenes should be regarded as possible bacterial inducers of CRC (de Kok & van Maanen, 2000).For instance, their detrimental effects may locally contribute to the accumulation of mutations in epithelial cells, which is not directly reflected by the increased levels in fecal material.

Induction of pro-carcinogenic pathways
Some evidence exists that certain intestinal bacteria can also directly induce host epithelial pathways that make cells more susceptible to DNA damage by carcinogenic substances.Maddocks et al. (2009) have shown that enteropathogenic E. coli can down-regulate mismatch repair genes in colon epithelial cells.It may be envisaged that this impaired expression can lead to a net increased mutation rate upon co-exposure to genotoxic dietary compounds.This study accentuates that bacteria can directly interfere with gene expression in epithelial cells which, under certain conditions, may lead to increased carcinogenesis rates.

CRC microbiome
The preceding paragraphs describe the potential mechanisms by which bacteria can play a role in the initiation and progression of CRC.In the following paragraphs, the effects of colonic malignancies on the (local) microbial composition are discussed.It is evident that the dramatic physiological and metabolic alterations that result from colon carcinogenesis itself (Hirayama et al., 2009) will locally disturb the intestinal environment.Consequently, this will cause (local) shifts in microbiota composition as the altered tumor metabolites and intestinal physiology will recruit a bacterial population with a competitive advantage in this specific microenvironment.This is exemplified by the fact that infections with certain opportunistic intestinal pathogens have been associated with CRC for many years (see Section 5).Thus pre-malignant sites seem to constitute a preferred niche for a subset of intestinal bacteria and facilitate their outgrowth and eventually entry into the human body.Importantly, local outgrowth of harmful bacteria could also accelerate tumor progression after disease has been initiated by other factors.The effect of colonic tumors on the microbiome composition has been investigated by several studies.First, Scanlan et al. (2008) investigated the bacterial diversity in healthy, polypectomized patients with increased risk for CRC and CRC patients.These studies showed a significant increased diversity of the Clostridium leptum and coccoides subgroups in the CRC patients compared to a healthy control group.Importantly, metabonomic faecal water analysis was able to distinguish CRC and polypectomized patients from healthy individuals, which is indicative for an altered metabolic activity of the intestinal microbiota in these patients.In another study by Maddocks et al. (2009) it was shown that the mucosa of adenomas and carcinomas contained increased numbers of E. coli compared to colonic mucosa from healthy controls.It was speculated that certain surface antigens on tumor cells, which display homology to surface antigens of fetal origin, may be responsible for the binding of E. coli and thus local recruitment of these bacterial strains (Martin et al., 2004;Maddocks et al., 2009;Swidsinski et al., 1998).A similar relation has been described for the opportunistic pathogen Streptococcus bovis.This bacterium is thought to selectively colonize malignant and pre-malignant colonic sites by which it can cause systemic infections in susceptible individuals (see Section 5).Some contradicting results on actual S. bovis colonization of tumor tissue have, however, been reported.Conventional culturing techniques to determine the carriage rate of S. bovis in adenoma, carcinoma and healthy biopsies did not provide clear evidence for the selective colonization of adenomas or carcinomas by this bacterium (Norfleet & Mitchell, 1993;Potter et al., 1998).More recently, Abdulamir and co-workers showed the presence of Streptococcus gallolyticus (S. bovis biotype I) DNA in carcinoma and adenoma tissue via polymerase chain reaction (PCR)-based techniques, which are more sensitive than conventional culturing techniques.DNA from S. gallolyticus was detected in about 50% of the tumor biopsies and in 35% of off-tumor tissue samples from the same patients.Strikingly, however, S. gallolyticus DNA was only found in <5% of the colonic tissue samples of healthy control subjects (Abdulamir et al., 2010).More recently, several studies have assessed the bacterial communities in healthy, adenoma and CRC tissue by deep 16S ribosomal DNA sequencing approaches.Shen and colleagues compared the bacterial composition in normal tissue samples from adenoma patients and from individuals without colon abnormalities.The data showed increased levels of proteobacteria and decreased bacteroidetes species in off-tumor tissue samples from adenoma patients (Shen et al., 2010).Interestingly, Sobhani et al. (2011) reported that the abundance of Bacteroides was significantly increased in tumor and normal tissue of cancer patients compared to healthy controls.More importantly, the abundance of Bacteroides was higher in tumor tissue of cancer patients than adjacent off-tumor tissue, which was paralleled by an increased IL-17/CD3 immune cell infiltration in the malignant tissues.Another recent study by Marchesi et al. (2011), compared differences in healthy and cancerous tissue within cancer patients and found that tumor tissue was overrepresented by species of the genera Coriobacteridae, Roseburia, Fusobacterium and Faecalibacterium that are generally regarded as gut commensals with probiotic features.On the contrary, this study found decreased colonization of Enterobacteriaceae, such as Citrobacter, Shigella, Cronobacter, and Salmonella in adjacent off-tumor mucosa from the same investigated patients.The development of colorectal tumors is schematically depicted from left to right.Initiation of carcinogenesis is a process in which many factors are involved.As discussed in this Chapter, certain bacterial pathogens, bacterial toxins, or bacterial toxic metabolites (1) may contribute to the initiation and progression of CRC by causing DNA damage, induction of COX-2/IL-8, TLR signalling and/or cell proliferation pathways (2).Consequently, the altered metabolic profile of colon tumor cells and/or differentially expression of bacterial receptor molecules on tumor cells (3) creates a new niche that recruits a different bacterial population (4) of which certain opportunistic pathogens can eventually breach the bowel wall and cause a systemic bacterial infection (5).The latter group of bacteria may play an important signalling function for the early detection of CRC by serological assays.Mason, 1951).In the 1970's this association was re-discovered by Hoppes and Lerner, who reported that 64% of the S. bovis endocarditis cases had gastrointestinal disease (Hoppes & Lerner, 1974).A few years later, Klein et al. (1977) reported an increased incidence of CRC in patients with S. bovis endocarditis.These investigators additionally discovered that fecal carriage of S. bovis in CRC patients was increased about 5-fold compared to healthy controls.At the time, these findings led to the recommendation to perform colonic evaluation in patients that were diagnosed with an S. bovis infection.Over the years, many studies have confirmed the association between S. bovis infection and CRC.In these studies, the prevalence of S. bovis infection with underlying CRC ranged from 10 -100% (median 60%) for patients that underwent colonic evaluation (Boleij et al., 2011b).

Streptococcus bovis biotypes
Based on phenotypic diversity, S. bovis was previously divided into three biotypes I, II.1 and II.2.Of these biotypes, biotype I is most often associated with endocarditis, while biotype II is mostly found in cases of bacteremia or liver disease.Strikingly, the association between S. bovis biotype I infection and CRC (21-71%) is much higher then that of S. bovis biotype II (11-30%) (Corredoira et al., 2008;Corredoira et al., 2005;Giannitsioti et al., 2007;Herrero et al., 2002;Jean et al., 2004;Lee et al., 2003;Ruoff et al., 1989;Vaska & Faoagali, 2009)(Beck et al., 2008;Tripodi et al., 2004).In fact, the reported incidences of carcinomas and adenomas in S. bovis biotype II infected patients are within the range for the normal asymptomatic population (0.3% for carcinomas / 10-25% for adenomas), whereas the rates for S. bovis biotype I were significantly increased (Lieberman & Smith, 1991;Lieberman et al., 2000;Spier et al., 2010).The distinct association of these different S. bovis biotypes with CRC may have accounted for the wide range of association percentages that have been reported over the years in literature.More importantly, because most studies have not discriminated between S. bovis biotypes the association between S. bovis biotype I and CRC may have structurally been underestimated.It is important to note that Schlegel et al. (2003) suggested renaming S. bovis biotype I into S. gallolyticus subsp.gallolyticus, S. bovis biotype II/1 into S. infantarius subsp.coli or S. infantarius subsp.infantarius and to rename S. bovis biotype II/2 into S. gallolyticus subsp.pasteurianus.This new nomenclature should be used to better discriminate between the different S. bovis subspecies of which S. gallolyticus is the only species with an unambiguous association with CRC (Boleij et al., 2011b).

Streptococcus gallolyticus
Recently, some striking differences between S. bovis biotypes were revealed that could explain their different association rates with CRC.First of all, S. gallolyticus seems to contain distinguished mechanisms to adherence to extracellular matrix (ECM) structures like collagen and fibrinogen (Ellmerich et al., 2000a;Sillanpaa et al., 2008;Sillanpaa et al., 2009).Interestingly, (pre-)malignant colonic sites are characterized by displaced collagen of the lamina propria (Galbavy et al., 2002;Yantiss et al., 2001), through which specifically S. gallolyticus may colonize these sites.Besides the ECM components, also other structures at the epithelial surface may play a role in the initial adhesion to enterocytes.For example, Henry-Stanley et al. (2003) reported binding of S. bovis strains to heparan sulfate proteoglycans, which may be mediated by surface-associated HlpA (Boleij et al., 2009).In an in vitro trans-well model containing a differentiated intestinal monolayer, the paracellular translocation efficiency of S. gallolyticus was shown to be significantly higher than that of other S. bovis biotypes.This could mean that this bacterium has an advantage over other S. bovis subspecies to cross an intestinal epithelium, which possibly only occurs at (pre-)malignant sites with reduced barrier function (Boleij et al., 2011a).Recent data suggested that S. gallolyticus does not induce a strong pro-inflammatory IL-8 response in epithelial cells in contrast to other S. bovis strains, which may be a possibly mechanism by which S. gallolyticus stays rather invisible for macrophages in the lamina propria.Furthermore, Hirota et al. (1995) discovered that S. gallolyticus isolates from endocarditis patients, express human sialyl Lewis x antigens on their cell surface unlike other fecal isolates.Mimicking human sialyl antigens, which are naturally present on monocytes and granulocutes, could therefore be a second mechanism of S. gallolyticus to remain unnoticed by the human innate immune system.Moreover, sialyl Lewis x antigens could make these bacteria more efficient in binding to endothelial cells and invasion into the circulatory system (Hirota et al., 1996).Finally, S. gallolyticus was shown to have superior efficiency to form biofilms on collagen I and IV surfaces (Boleij et al., 2011a;Sillanpaa et al., 2008).The latter finding could explain the increased incidence of S. gallolyticus as causative agent in infective endocarditis.Based on the current state-of-the-literature (July 2011), the following events in CRC-associated S. gallolyticus endocarditis can be envisaged i) S. gallolyticus specifically adheres to (pre-)malignant colonic sites for instance via binding to displaced collagen of the lamina propria or other tumor cell specific adherence factors; ii) S. gallolyticus may promote tumor progression by induction of the COX-2 pathway; iii) S. gallolyticus takes advantage of the distorted structure of the colonic epithelium at (pre-)malignant sites to pass the colonic wall; iv) S. gallolyticus stays relatively invisible for the innate immune system and can reach the blood stream; v) S. gallolyticus can cause a secondary infection at sites with high exposure of collagens, such as present at damaged heart values.It should be noted, however, that many of these data were obtained by in vitro studies and that it remains to be determined how this relates to the in vivo situation.

Clostridium septicum
In addition to S. gallolyticus endocarditis, also Clostridium septicum infections have been clinically associated with sporadic CRC (Chew & Lubowski, 2001;Mirza et al., 2009).C. septicum is not considered to be part of the normal intestinal microbiota and is a rare cause of bacteremia (<1% of all cases).Hermsen et al. (2008) investigated 320 cases of C. septicum infections, 42% of which had a gastrointestinal origin.Malignant disease was present in 30-50% of these cases.The underlying mechanism of this association is not known, but it has been speculated that the hypoxic and acidic environment of the tumor specifically favor germination of C. septicum spores that enter the gastrointestinal tract via contaminated food (Dylewski & Luterman).A direct involvement of C. septicum in the development of CRC has thus far not been investigated, but it is hypothesized that C. septicum infections are primarily a consequence of CRC itself.Also Clostridium perfringens and Clostridium butyricum have been described in relation with CRC (Cabrera et al., 1965;Rathbun, 1968).However, these strains are much less virulent than C. septicum and their association with CRC is less evident.Although infections with C. septicum are rare, underlying malignancy should be suspected and also in these cases full bowel examination could eventually save patients' lives.

Helicobacter pylori
Helicobacter pylori has been classified as gastric cancer-causing infective agent by the International Agency for Research on Cancer (IARC) in 1994.Most H. pylori strains, however, are non-invasive organism and exist in a non-adherent extracellular mucous environment.A small number of strains adheres to gastric epithelial cells, which most likely involves a number of different surface receptors (Wilkinson et al., 1998).The presence of the pathogenicity island, expressing the cytotoxins VacA and CagA, is an important virulence determinant in these strains (Ekstrom et al., 2001;Huang et al., 2003;Crabtree et al., 1994;Kuipers et al., 1995).It is thought that long-term exposure to these toxins induces gastric inflammation that can eventually lead to gastric carcinomas (Higashi et al., 2002;Fox, 2002).A meta-analysis conducted in 2006 by Zumkeller et al. indicated also a slightly increased risk for CRC (factor 1.4) in individuals with a H. pylori infection (Zumkeller et al., 2006).Another study showed that CagA status was associated with a significantly increased risk (factor >10) for CRC among hospitalized patients that were H. pylori seropositive (Shmuely et al., 2001).Notably, this study again underscores the importance of proper microbiological classification and characterization of cancer-associated infectious agents, since not all Helicobacter strains may be associated with CRC.Like has been the case for S. bovis, lack of proper distinction between H. pylori subspecies could have biased or even underestimated a possible association of this bacterium with this disease (Erdman et al., 2003a,b).

CRC Microbiome-based Immunoassays
The occurrence of specific CRC-associated bacterial infections, as discussed in the previous section, paves the way for the development of novel diagnostic tools.In this respect, it is important to realize that S. gallolyticus infections occur without clinical symptoms due to its mild virulence (Haimowitz et al., 2005).Clinical manifestation of S. gallolyticus infections in otherwise compromised patients (e.g.damaged heart valves), may very well only represent the tip of the iceberg of all infections with this bacterium in individuals with (pre-)malignant colonic lesions.This notion has been the incentive to investigate whether a humoral immune response to sub-clinical S. gallolyticus infections could aid in the early detection of CRC.Notably, as infectious agents in general induce a more pronounced immune response compared to tumor "self" antigens, CRC-associated bacterial antigens could be instrumental in the immunodiagnosis of this disease (Tjalsma, 2011).Furthermore, several features of circulating antibodies make these attractive targets in diagnostic medicine: i) they reflect a molecular imprint of disease-related antigens from all around the human body, ii) although an antigen may be present only briefly, the corresponding antibody response is likely to be persistent, iii) the half-life of antibodies is about 15 days which minimizes daily fluctuations, iv) antibodies are highly stable compared to many other serum proteins making serumhandling protocols less stringent, v) the amplification cascade governed by the humoral immune system causes a surplus of circulating antibodies after appearance of the cognate (low-abundance) antigen.Several studies have shown that serum antibody levels against S. bovis/S.gallolyticus antigens could discriminate CRC cases from healthy controls (Abdulamir et al., 2009;Darjee & Gibb, 1993;Tjalsma et al., 2006).Interestingly, the humoral immune response to ribosomal protein (Rp) L7/L12 from S. gallolyticus was found to be higher in early CRC compared to late CRC stages, whereas this was not paralleled by increased antibody production to endotoxin, an intrinsic cell wall component of the majority of intestinal bacteria (Boleij et al., 2010).This implies that the immune response to RpL7/L12 is not a general phenomenon induced by the loss of colonic barrier function.Furthermore, this observation could point to a temporal relationship between S. gallolyticus and CRC, suggesting that late stage tumors may change in such a way that bacterial survival in the tumor microenvironment is diminished.The possibility that disease progression may drive bacteria out of the cancerous tissue is similar to what has been reported for H. pylori during gastric cancer progression (Corfield et al., 2000;Kang et al., 2006).A relationship of S. bovis with early stages of CRC is underscored by a vast amount of case studies showing that its infection was associated with pre-malignant adenomas.These cases would have remained undiscovered if these patients did not present with an active S. bovis infection.Future research should be aimed at development of more specific S. gallolyticus-based serological assays to investigate the clinical utility of such tests for the early detection of CRC (Tjalsma et al., 2006(Tjalsma et al., , 2008;;Tjalsma, 2010).Furthermore, as CRC is a highly heterogeneous disease that is probably accompanied by even more heterogeneous microbiome shifts, accurate diagnosis based on biomarkers from a single bacterial species on the population level is highly unlikely.Therefore, future research should also be aimed at the identification of additional tumor-associated intestinal bacteria that may never have been found to cause clinical infections but do induce a humoral immune response.Furthermore, as discussed in Section 3 of this Chapter, certain mucosa-associated bacteria may be involved in CRC initiation or progression.Invasiveness of these pathogens or exposure to their antigens may elicit IgG responses that are valuable for CRC risk assessment.These individuals may not directly need bowel examination, but could be enrolled in a more strict monitoring program.

Conclusions
The development of CRC is a multistep process that may take over 20 years to progress from an adenoma into an advanced carcinoma.The fact that the intestinal microbiome plays an important role in this process is clearly shown by the inflammatory effects of intestinal www.intechopen.combacteria, which are essential to develop disease in animal models.Furthermore, accumulating evidence suggests that bacterial production of toxins, toxic metabolites and the direct influences on pro-carcinogenic pathways in host epithelial cells are contributing factors that promote the accumulation of mutations that may eventually lead to carcinomas.However, still many questions remain to be answered.For example, our knowledge on the on the impact of CRC on the local intestinal microbiota and vice versa, is still in its infancy.Future research should focus on the detailed mapping of the microbiota in close proximity of early adenomas and carcinomas.These local changes in microbiota may for instance provide clues in the understanding why only 10% of the adenomas progress into carcinomas.Such knowledge could give us new leads for cancer diagnosis, for example by using signaling bacteria, such as S. gallolyticus that benefit from the altered tumor environment, as diagnostic targets.Furthermore, this knowledge could provide leads for the selective removal of high-risk bacterial populations by health promoting species, as a new strategy in CRC prevention.Altogether, this Chapter points out that the colonic microbiota should be regarded as an important factor in intestinal carcinogenesis.Further research in this field is crucial to fully understand the etiology of CRC and has a high potential to lead to new diagnostic tools and therapeutic interventions.