enterohaemorrhagic Escherichia coli
mesenteric lymph node
natural killer cell
pattern recognition receptor
segmented filamentous bacteria
small intestinal lymphoid tissue
The consortia of microorganisms inhabiting the length of the gastrointestinal tract, the gastrointestinal microbiota, are vital to many aspects of normal host physiology. In addition, they are an active participant in the progression of many diseases, among them enteric infections. Healthy intestinal microbiota contribute to host resistance to infection through their involvement in the development of the host immune system and provision of colonization resistance. It is not surprising then that disruptions of the microbial community translate into alterations of host susceptibility to infection. Additionally, the process of the infection itself results in a disturbance to the microbiota. This disturbance is often mediated by the host inflammatory response, allowing the pathogen to benefit from the inflammation at the intestinal mucosa. Uncovering the mechanisms underlying the host–pathogen-microbiota interactions will facilitate our understanding of the infection process and promote design of more effective and focused prophylactic and therapeutic strategies.
Humans and other multicellular eukaryotic organisms are made up of more than just their own cells – they exist in a tight relationship with myriad assorted microorganisms that inhabit numerous sites on their bodies. This entourage of microorganisms constitutes the normal flora, or microbiota; colonization by the microbiota is a very dynamic process that begins at birth and continues throughout life. Although every surface that is exposed to the environment harbours some number and variety of microbial tenants, in animals the gastrointestinal tract contains by far the most numerous and diverse microbial community (Salzman et al. 2007). The intestinal microbiota are notable for their many roles in the proper development, health and disease of the host, contributing to numerous aspects of the development of the gastrointestinal tract itself, and also impacting on many other organ systems.
The intestinal microbiota comprise eukaryotes (Scupham et al. 2006), viruses (Furuse et al. 1983) and bacteria. The bacterial microbiota constituents are incredibly diverse and numerous, outnumbering the host cells by at least one order of magnitude (Savage, 1977). Microbial colonization in humans varies along the length of the gastrointestinal tract, with a low of 101–103 bacteria ml−1 in the stomach and the duodenum, progressing to 104–107 bacteria ml−1 in the jejunum and ileum and culminating in 1011–1012 bacteria ml−1 in the colon (O’Hara & Shanahan, 2006). Obligate anaerobes make up most of the colonic microbiota, with facultative anaerobes about ∼1000-fold fewer. Firmicutes and Bacteroidetes phyla dominate both the human and the murine intestinal microbiota. Proteobacteria, Actinobacteria, Verrucomicrobia, Cyanobacteria and Deferribacteres have also been detected in humans and mice, and Fusobacteria in humans only (Backhed et al. 2005; Eckburg et al. 2005; Stecher & Hardt, 2008).
Intestinal immune system
In addition to providing a home to numerous microbial inhabitants, the intestinal tract is also an active immunological organ, with more resident immune cells than anywhere else in the body.
A great variety of immune cells are found in the intestine. They are organized in lymphoid structures called Peyer's patches (PP) and small intestinal lymphoid tissue (SILT), as well as residing in the intestinal mesenteric lymph nodes (MLN) (Mowat, 2003; Herbrand et al. 2008). Macrophages, dendritic cells (DC), various subsets of T cells, B cells and the secretory IgA they produce all contribute to the generation of a proper immune response to invading pathogens, while keeping the resident microbial community in check without generating an overt inflammatory response to it (Sansonetti, 2008; Hooper, 2009).
In addition to the immune cells, the intestinal epithelial cells also contribute to the mucosal immunity. Enterocytes and colonocytes have been shown to express pattern recognition receptors (PRR), both extracellular Toll-like receptors (TLR) and intracellulalr NOD-like receptors (NLR) that sense conserved bacterial products. While in a normal state, TLRs and NLRs remain relatively unresponsive to the myriad bacteria overlying the mucosa. In a state of infection, injury, or another assault they initiate a cascade of events that contributes to the induction of inflammatory host response (Reaves et al. 2005; Gribar et al. 2008; Rakoff-Nahoum & Medzhitov, 2008).
Another component of the intestinal mucosal defences is the antimicrobial proteins secreted by intestinal epithelial cells, both enterocytes and Paneth cells (specialized secretory cells found in the small intestine). They belong to several different families, such as defensins, cathelicidins and C-type lectins, and act by disrupting bacterial surface structures (Salzman et al. 2007; Hooper, 2009).
In addition to mucosal defences mediated either directly by GI tract-associated cells or by antimicrobial cellular products, there is a layer of mucus overlying the intestinal epithelium that forms a physical barrier between the mucosa and the resident microbiota (Johansson et al. 2008), minimizing both microbial translocation and excessive immune activation by the resident microbes.
Intestinal microbiota and host defence mechanisms
Intestinal microbiota contribute to many aspects of host defence against invading pathogens both through direct microbial antagonism and promotion of maturation of the intestinal immune system (Fig. 1A). Bacterial metabolism results in the production of several by-products with an antimicrobial effect, such as peroxides and various acids. Some of these substances do not just inhibit the growth of pathogenic microorganisms themselves, but might also potentiate the effectiveness of other antimicrobial substances (Alakomi et al. 2000). In addition, production of biosurfactants by the microbiota (Velraeds et al. 1996) and competition for sites of attachment and nutrients (Reid et al. 2001) prevent the pathogens from establishing themselves within the host.
The intestinal microbiota is in constant contact with a variety of host cells of the gastrointestinal tract, such as the surface epithelial cells, M cells and DCs (O’Hara & Shanahan, 2006). The commensal bacterial antigens are being constantly sampled by the host PRRs, both TLRs and NLRs. It is not surprising, therefore, that the immunomodulatory role of the intestinal microbiota is integral to the proper development of the intestinal immune system. Comparisons of germ-free (GF) and conventionally raised laboratory animals confirm a role for the indigenous microbes in cytokine production, serum immunoglobulin levels, development of intraepithelial lymphocytes and Peyer's patches (O’Hara & Shanahan, 2006), as well as a role in systemic lymphoid organogenesis (Macpherson & Harris, 2004). Depletion of intestinal microbiota constituents has demonstrated that microbiota-induced activation of TLR signalling in a healthy host is necessary to maintain proper epithelial homeostasis and promote repair following intestinal injury (Rakoff-Nahoum et al. 2004). Products of microbial metabolism, such as butyrate and lithocholic acid, have been shown to upregulate expression of antimicrobial peptides in epithelial cell lines, pointing towards a function for microbiota in the induction of this line of host defences (Schauber et al. 2003; Kida et al. 2006; Termen et al. 2008). Additional studies have implicated specific bacterial groups and species in various aspects of intestinal immunity. However, as many of these studies were done in GF animals mono-associated with a particular microbial species, they do not exclude the possibility of additional microbiota members either contributing to or being the primary inducer of each particular aspect of the mucosal immunity.
Polysaccharide antigen of Bacteroides fragilis, a prominent member of the gut microbiota, promotes the expansion of splenic CD4 T-cells and regulates the Th1/Th2 cytokine production, as well as restoring the splenic architecture to that of conventionally raised mice (Mazmanian et al. 2005). Bacteroides thetaiotaomicron, another ubiquitous microbiota species, affects host gene expression resulting in a plethora of effects in a number of organ systems (Xu & Gordon, 2003); for instance, it induces the expression of an angiogenin with bactericidal activity against intestinal microbes (Hooper et al. 2003). The presence of higher numbers of bacteria belonging to the phylum Bacteroidetes has been shown to be associated with the development of IL-17 producing T-helper cells (Ivanov et al. 2008). Different Lactobacillus species, also important members of the gut microbiota, differentially activate DCs, inducing them to produce different arrays of inflammatory cytokines, thus playing an important role in the modulation of the Th1, Th2 and Th3 balance (Christensen et al. 2002). Moreover, Lactobacillus-stimulated DCs proceed to activate natural killer (NK) cells, thus potentiating gastrointestinal immunity (Fink et al. 2007). Segmented filamentous bacteria (SFB) are implicated in the induction of the intestinal IgA (Suzuki et al. 2004) and activation of intraepithelial lymphocytes and induction of MHC class II expression on intestinal epithelial cells (Umesaki et al. 1995).
Intestinal microbiota in enteric infection
Enteric infection can either occur as a result of the host encountering an overt or an opportunistic pathogen or be due to a pathological overgrowth of an opportunistic member of the intestinal microbial community. To produce a successful infection the infecting agent has to overcome a number of host defences, such as mucosal immunity or colonization resistance posed by the normal microbiota. In recent years the host–pathogen-microbiota interplay during enteric infection has received mounting attention with novel research illuminating some interesting aspects of this relationship (Fig. 1B).
Inflammation – a double-edged sword Many facets of the pathogen–host interactions during infection have already been uncovered, but only lately the extensive effects of the pathogen on the host microbiota have begun to be characterized. Infection with Citrobacter rodentium, a murine enteric pathogen that is used to model enterohaemorrhagic Escherichia coli (EHEC) infection in a murine host, was shown to drastically reduce the total numbers of colonic microbiota and extensively alter the composition of the indigenous microbial population (Lupp et al. 2007). Interestingly, the observed microbiota perturbations were mediated by the host inflammatory response to C. rodentium, rather than by the infecting agent itself, as chemically and genetically induced intestinal inflammation produced similar alterations in the mouse colonic microbiota and facilitated colonization of the intestinal tract by Enterobacteriaceae. Mice genetically predisposed to inflammation were also shown to develop more severe histopathological changes following infection with Helicobacter trogontum, another rodent gastrointestinal pathogen (Whary et al. 2006). Additionally, infection-associated inflammation was found to be detrimental to colonization by commensal anaerobic bacteria (Whary et al. 2006; Lupp et al. 2007).
Normal mouse intestinal microbiota of commonly used C57BL/6 mice provide the host with colonization resistance to Salmonella enterica serovar Typhimurium, while a severe reduction in microbiota numbers with high doses of antibiotics allows the pathogen to successfully establish itself in the intestinal environment and generate intestinal disease manifestations (Que & Hentges, 1985; Barthel et al. 2003; Woo et al. 2008). Studies in a high dose streptomycin pre-treatment mouse model (Barthel et al. 2003) have shown that S. Typhimurium-induced intestinal inflammation allows it to outcompete the indigenous microbial community, promoting successful host colonization and alterations to microbiota composition (Stecher et al. 2007). Akin to the situation during C. rodentium infection, the inflammatory response that was detrimental to the indigenous microbiota was host-mediated, allowing a non-inflammatory S. Typhimurium mutant to establish successful host colonization in an already inflamed intestine. The ability of S. Typhimurium to utilize high-energy nutrients released in the inflamed intestine as part of the host defence was proposed to contribute to this enhanced pathogen fitness (Stecher et al. 2008). However, some degree of caution should be exercised when interpreting the results of these experiments, as the initial pre-treatment of mice with a high dose of streptomycin severely disturbs the mouse microbiota composition and numbers, potentially affecting the downstream host–microbiota, host–pathogen and microbiota–pathogen interactions.
While S. Typhimurium appears to benefit from the induced inflammatory host response, commensal bacterial strains may act to down-regulate pro-inflammatory signalling, exerting a protective effect during S. Typhimurium infection (O’Mahony et al. 2008): a probiotic preparation of B. infantis, a commensal microbe, prevented excessive NF-κB activation in vivo, reducing inflammatory host response both locally and systemically and ameliorating S. Typhimurium-induced disease severity. These findings demonstrate, yet again, the importance of a measured host response to an invading pathogen.
While induction of the immune response is necessary for the host to clear an invading pathogen, certain aspects of the host response may work to the pathogen's advantage. Normal intestinal microbiota provide the host with colonization resistance to invading pathogens and also promote proper functioning of the host immune system. It is not surprising then, that a disruption to the host's indigenous microbial population would benefit the pathogen's ability to invade and establish itself in the GI tract. It is rather ingenious, however, that at least some pathogens are able to exploit to their advantage the inflammatory response that is aimed at eradicating them, but along the way proves detrimental to the host's microbiota. The host almost appears to set itself up for a successful infection: a healthy intestinal microbial community ensures that a prompt inflammatory response would be mounted when the host immune system encounters an invading pathogen, but the inflammation at the intestinal mucosa proves detrimental to the microbiota, disrupting the colonization resistance and potentially also interfering with some aspects of the normal host homeostasis. This disruption is then exploited by the pathogen as a window of opportunity to out-compete the intestinal microbiota and successfully infect the host. If the host's mucosal immune response is sufficiently robust to ultimately control the invading pathogen, it will eventually be cleared from the system and homeostasis will resume. The intestinal microbiota will return to the pre-infection balance, once again providing the host with both colonization resistance and the ability to mount an inflammatory response to future enteric infections, yet again repeating the cycle. However, if the inflammatory response is ineffective in controlling the pathogen, then the very attempt to limit the infection would, instead, provide the enteric pathogen with an opportunity to successfully infect the host, leading to the host's demise. This cyclical system, while complex, provides many potential check-points for a successful therapeutic interference. The host inflammatory response to enteric infections needs to be appropriately balanced so as to be sufficiently robust to control the pathogen, while at the same time suitably gentle to spare the host's microbiota. To achieve this, the host's mucosal immune system could be targeted to gear the response in the appropriate direction, the host's microbiota could be modified to promote induction of a suitable response, or the pathogen's virulence factors aimed at inducing excessive inflammation could be blocked.
Microbiota perturbations – cause or consequence? The stability of the intestinal microbiota was shown to be compromised in a number of enteric infections. Since the microbiota emerge as a key component in the infection process, it is imperative to determine the relative contributions of pre- and post-infection microbiota alterations to the disease progression.
Infections with various Helicobacter species were shown to alter the composition of the host indigenous microbial community: the diversity and numbers of the microbiota in the lower intestinal tract were reduced, while the diversity of the gastric microbiota increased (Kuehl et al. 2005; Aebischer et al. 2006; Whary et al. 2006). Interestingly, in the case of H. pylori infection, the alterations in the gastric microbiota composition preceded the onset of inflammation-induced pathological changes in the intestine, indicating that the pathogen-elicited inflammatory host response is not the only factor able to effect changes in the indigenous microbial community (Aebischer et al. 2006). In fact, the authors of the study hypothesized that the metabolic by-products of the altered gastric microbial community may induce H. pylori-associated gastric pathology. In this way, H. pylori-mediated perturbations in the host microbiota would promote its pathogenicity.
Antibiotic-induced alterations in microbiota composition, without a concurrent reduction in total microbial numbers, resulted in an increased murine susceptibility to S. Typhimurium-induced intestinal disease (Sekirov et al. 2008; Croswell et al. 2009). Greater pre-infection alterations to the indigenous microbiota promoted higher colonization by S. Typhimurium and more severe inflammation, demonstrating the importance of a healthy microbial community to host resistance to infection (Sekirov et al. 2008). Additionally, host microbiota were shown to contribute to S. Typhimurium transmission, with mice whose microbiota were perturbed prior to infection shedding higher numbers of bacteria than mice whose microbiota were in a steady state prior to pathogen introduction (Lawley et al. 2008). The process of infection itself also promoted alterations in microbiota composition and numbers (Barman et al. 2008; Sekirov et al. 2008). Two pathogenicity islands of S. Typhimurium were implicated in microbiota alterations in an FvB mouse model of enteric salmonellosis (Barman et al. 2008). However, it has not been determined whether the virulence factors encoded in these pathogenicity islands interact directly with the microbiota constituents or mediate the observed alterations through induction of differential host response. It is clear, nonetheless, that perturbations to the host microbiota composition prior to infection interfere with the host's ability to control S. Typhimurium infection, and that the virulence factors of S. Typhimurium promote further alterations to the host's indigenous microbial community.
As mentioned previously, the gastrointestinal microbiota is actively interacting with the host mucosal immune system, contributing to the mounting of an adequate immune response. Commensal-derived antigens were shown to prime TLR2, resulting in an increase in RegIIIβ C-type lectin expression. This interaction was demonstrated to promote the ability of the murine host to clear Yersinia pseudotuberculosis, highlighting the importance of the microbiota in host response to this enteric pathogen (Dessein et al. 2009). When the mouse microbiota were perturbed prior to infection (via the use of an antibiotic), the mice were not able to mount a successful response to Y. pseudotuberculosis and succumbed to infection more readily than mice whose microbiota were in a steady state prior to infection (Dessein et al. 2009). Thus, similar to the case of S. Typhimurium, pre-infection perturbation to the indigenous microbial community enhanced host susceptibility to this enteric pathogen.
Antibiotic-induced alterations in the host intestinal microbiota have also been shown to interfere with innate immune defences against opportunistic pathogens, effectively creating an immune deficit and contributing to the establishment of the opportunist in the intestinal tract. Specifically, antibiotic administration to mice down-regulated intestinal expression of RegIIIγ, a C-type lectin with antimicrobial activity against Gm(+) bacteria, thereby rendering mice more susceptible to vancomycin resistant enterococci (VRE), a common complication of antimicrobial therapy in human patients (Brandl et al. 2008). Thus VRE were able to exploit immune deficits created by perturbations to the host microbial community in order to successfully colonize the host intestinal tract. Recurrent Clostridium difficile infections, another frequent complication of antibiotic therapy (Owens et al. 2008), were characterized by a decreased diversity in the fecal microbial community (Chang et al. 2008), while the main bacterial components of the fecal flora were found to have some predictive value for the future development of C. difficile-associated diarrhoea (De La Cochetiere et al. 2008). This demonstrates that a pre-infection reduction in complexity of the host microbiota prevents successful inhibition of C. difficile overgrowth by either a reduction in colonization resistance, or a failure to promote a proper host response, or both.
It appears that microbiota perturbations during enteric infections are a cyclical process. Initial perturbation to the stability of the microbial community can act as a starting point in the infection process, creating a permissive environment for pathogen colonization by either interfering with initiation of a proper immune response, or freeing up suitable niches, or both. Regardless of the mechanism, a host with a perturbed intestinal microbiota becomes essentially immunocompromised and consequently more susceptible to overt and opportunistic pathogens alike. If the pre-infection perturbation to the microbial community promotes enhanced pathogen replication and colonization, the infected host will also shed the infecting microbes in higher numbers, posing a greater health hazard to those around him. Once the process of infection is initiated, the pathogen appears to further modify the microbiota composition, possibly for the sake of generating itself a comfortable niche in the host GI tract or creating an environment that would potentiate its pathogenicity and consequently maximize its spread to other susceptible hosts. Depending on the extent of the pre-infection perturbation to microbiota stability and/or on the severity of infection and infection-associated microbiota perturbations, the indigenous microbial population will take a shorter or longer time to recover to its original steady state. If the steady state is not recovered sufficiently readily, then the infection-associated microbiota perturbations might lay a fertile ground for subsequent recurrent infections. Multiple factors might contribute to the generation of the initial disturbance in the intestinal microbiota: antimicrobial therapy, abrupt changes in diet, medical interventions, such as abdominal surgery or colonoscopy, etc. Once we improve our knowledge and understanding of the exact alterations to the indigenous microbial community resultant from these disturbances, we’d be better positioned to counteract and correct them, both preventatively and therapeutically.
Conclusions and outlook
Gastrointestinal infections are an intricate process of multifaceted interactions between three main components – the host, the pathogen and the host's intestinal bacterial community. These interactions are ongoing and continuously evolving throughout the infection process, leading to either the resolution of infection or a chronic pathogen colonization of the infected individual, or perpetuating a vicious cycle culminating in the demise of the host (Fig. 2). Resolution of details contributing to this byzantine web of interfaces is still in its infancy, but its thorough understanding is required to provide us with proper, well-targeted tools for control of gastrointestinal infections.
A number of factors from each of the three participants contribute to the eventual outcome of the gastrointestinal infection process. The host's overall fitness, genetic make-up and the choice of therapy to treat the infection will contribute to the generation of the immune response. The individual variations in microbiota composition and stability, and any pre-existing perturbations to it will contribute to colonization resistance and resilience of the intestinal mucosa, and will also affect the generation of the immune response. The virulence factors carried by the pathogen will determine its interactions with the host, the microbiota, and its ability to promote a detrimental-to-microbiota host response. The initial stability of the intestinal microbiota population and its ability to recover to steady state following an insult underlie the progression of many of the interactions happening between the players of this ménage à trois (Pedron & Sansonetti, 2008), positioning the microbiota as the cornerstone of the host's health. In view of this, future research in this area will greatly contribute to our knowledge of enteropathogenesis and provide us with tools necessary for the design of effective and specific therapeutic and prophylactic methods, tailor-made not only to target a particular infective agent, but also to account for individual variations in the indigenous microbial population.
This work was supported by operating grants to B.B.F. from the Canadian Crohn's and Colitis Foundation, the Canadian Institutes of Health Research, the Howard Hughes Medical Institute (HHMI), and Genome Canada. B.B.F. is an HHMI International Research Scholar and the University of British Columbia Peter Wall Distinguished Professor. I.S. is supported by a Michael Smith Foundation for Health Research Senior Graduate Trainee fellowship.