The zoonotic pathogen Campylobacter jejuni is a leading cause of bacterial foodborne enterocolitis in humans worldwide. The understanding of immunopathology underlying human campylobacteriosis is hampered by the fact that mice display strong colonization resistance against the pathogen due to their host specific gut microbiota composition.
Since the microbiota composition changes significantly during intestinal inflammation we dissected factors contributing to colonization resistance against C. jejuni in murine ileitis, colitis and in infant mice. In contrast to healthy animals C. jejuni could stably colonize mice suffering from intestinal inflammation. Strikingly, in mice with Toxoplasma gondii-induced acute ileitis, C. jejuni disseminated to mesenteric lymphnodes, spleen, liver, kidney, and blood. In infant mice C. jejuni infection induced enterocolitis. Mice suffering from intestinal inflammation and C. jejuni susceptible infant mice displayed characteristical microbiota shifts dominated by increased numbers of commensal Escherichia coli. To further dissect the pivotal role of those distinct microbiota shifts in abrogating colonization resistance, we investigated C. jejuni infection in healthy adult mice in which the microbiota was artificially modified by feeding live commensal E. coli. Strikingly, in animals harboring supra-physiological intestinal E. coli loads, colonization resistance was significantly diminished and C. jejuni infection induced enterocolitis mimicking key features of human campylobacteriosis.
Murine colonization resistance against C. jejuni is abrogated by changes in the microbiota composition towards elevated E. coli loads during intestinal inflammation as well as in infant mice. Intestinal inflammation and microbiota shifts thus represent potential risk factors for C. jejuni infection. Corresponding interplays between C. jejuni and microbiota might occur in human campylobacteriosis. Murine models introduced here mimick key features of human campylobacteriosis and allow for further analysis of immunological and molecular mechanisms of C. jejuni – host interactions.
Citation: Haag L-M, Fischer A, Otto B, Plickert R, Kühl AA, Göbel UB, et al. (2012) Intestinal Microbiota Shifts towards Elevated Commensal Escherichia coli Loads Abrogate Colonization Resistance against Campylobacter jejuni in Mice. PLoS ONE 7(5): e35988. doi:10.1371/journal.pone.0035988
Editor: Georgina L. Hold, University of Aberdeen, United Kingdom
Received: February 3, 2012; Accepted: March 28, 2012; Published: May 1, 2012
Copyright: © 2012 Haag et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by grants from the 15 German Research Foundation (DFG) to UBG (GO363/12-1, CampyGerm; SFB633, TP A7), SB (SFB633, TP A7), AK (SFB633, TP Z1), MMH (SFB633, TP B6), LMH and BO (SFB633, Immuco), and from the German Federal Ministery of Education and Research (BMBF) to SB (“Lab in a hanky” projects TP1.1 and TP 8.2). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: Co-authors Stefan Bereswill and Markus M. Heimesaat are PLoS ONE Editorial Board members. This does not alter the authors' adherence to all the PLoS ONE policies on sharing data and materials.
Campylobacter (C.) jejuni continues to be the most frequently reported causative agent of bacterial gastroenteritis in humans worldwide , . The gram-negative bacteria form part of the commensal flora in many wild and domestic animals. Zoonotic transmission from livestock animals via consumption of contaminated meat products is the most common source for human C. jejuni infections , , . Infected patients present a broad range of symptoms including abdominal pain, watery or bloody diarrhea, and fever. Histological examination of affected intestinal tissues commonly reveals crypt abscesses, ulcerations, and elevated numbers of immune cells in the colon in situ , , , . Although the infection is self-limiting in most cases, long-term sequelae such as reactive arthritis, gastrointestinal and neurological disorders have been reported in susceptible hosts , . C. jejuni infection correlates with reactivation of inflammatory bowel diseases (IBD), Miller-Fisher-Syndrome and Guillian-Barré-Syndrome , . The high prevalence of campylobacteriosis poses a significant health burden and accentuates the need for further unravelling the underlying pathogenesis of C. jejuni infections. The lack of understanding mainly stems from the scarcity of suitable experimental in vivo models of human infection . Although chickens allow for the investigation of colonization and transmission, intestinal disease manifestations such as enteritis and diarrhea are absent , . Murine models of C. jejuni infection have the disadvantage of sporadic colonization, absence of clinical signs of disease and/or lack of intestinal immunopathology , . This is in part due to the colonization resistance against C. jejuni exhibited by adult mice older than two months (commonly used in experimental settings) harboring a conventional intestinal microbiota . Results from the 1980s demonstrated that colonization resistance was abrogated in isolator raised germfree (GF) mice. In these animals C. jejuni colonized the entire gastrointestinal (GI) tract and induced clinical signs of enterocolitis including granulocyte infiltrates, bloody diarrhea, and humoral immune responses , , . More recent investigations showed that colonization resistance is mainly caused by the host-specific intestinal microbiota composition , . Previous work in our group revealed that colonization resistance of mice against C. jejuni is abrogated in gnotobiotic (GB) mice generated by quintuple antibiotic treatment as well as in GB mice replenished with human flora (hfa) . In contrast to isolator raised germfree mice, however, GB and hfa animals display a fully established immune system. The finding that GB mice replenished with murine flora were completely resistant to C. jejuni proved that colonization resistance is caused by the host-specific microflora composition . Most strikingly, C. jejuni infected GB and hfa mice displayed characteristic inflammatory responses seen in human campylobacteriosis . The major role of the microflora composition in colonization resistance against C. jejuni was further supported by the recent finding that obesity caused by both feeding human “Western style” cafeteria diet (CAF) as well as leptin deficiency in ob/ob−/− mice rendered mice susceptible to C. jejuni infection . Interestingly, quantitative analysis of main gut bacterial groups revealed 2–3 orders of magnitude higher loads of commensal E. coli in C. jejuni susceptible animals in all these models as a common feature as compared to respective resistant controls , . Recent studies on Salmonella enterica also point towards a pivotal role of microbiota composition in host susceptibility to enteropathogenic bacteria , . Mice displaying high commensal E. coli densities were more susceptible to S. enterica induced gut inflammation , , . We have previously shown that murine models of acute small as well as large intestinal inflammation were associated with tremendous changes of the intestinal microbiota with decreasing species diversity and lactobacilli loads and an intestinal overgrowth with commensal E. coli and Bacteroides/Prevotella spp. , , , . We could further show that TLR-4-signalling of E. coli LPS initiated and perpetuated acute ileitis and colitis in mice , . Thus, these experimental models were mimicking immunopathological key features of human IBD Crohn’s disease (CD) and ulcerative colitis (UC). The exact etiology of IBD remains unknown and the pathogenesis seems to be multifactorial, involving the complex interaction between the intestinal microflora, host immune and genetic factors as well as environmental stimuli , , Gram-negative bacteria such as E. coli and Bacteroides spp. have been reported to accumulate at inflamed tissue sites in IBD patients with active intestinal inflammation, and further potentiate immunopathology by translocation via microlesions and ulcerations , . It is under current debate whether IBD patients might be more susceptible to intestinal pathogens such as Campylobacter or Salmonella spp. and whether acute bacterial enterocolitis is one of the factors inducing or exacerbating IBD in susceptible individuals , , . Results from population-based studies point towards a potential role of enteric Campylobacter or Salmonella infections in IBD development , , .
In the present study, we investigated colonization resistance in mice displaying an altered gut microbiota composition under conditions of acute or chronic intestinal inflammation as well as in infant mice after weaning (lacking signs of intestinal inflammation) having all elevated intestinal E. coli loads in common. Finally we asked the question whether feeding healthy adult mice with live E. coli and thus artificially elevating the enterobacteria population within the commensal gut microbiota towards supra-physiological loads might overcome colonization resistance and renders the animals susceptible to C. jejuni. The results clearly demonstrate that colonization resistance is completely abrogated under conditions of elevated E. coli loads within the commensal intestinal microbiota. Following infection, C. jejuni induced pro-inflammatory immune responses mimicking key features of human campylobacteriosis. Taken together, the murine C. jejuni infection models presented here proved to be suitable for reproducible investigations of factors that predispose C. jejuni colonization and trigger the onset and progress of infection and immunopathology.
T. gondii-induced Acute Ileitis Abrogates Colonization Resistance of Mice Against C. jejuni Infection
Colonization resistance against C. jejuni exerted by mice harbouring a conventional intestinal microbiota can be overcome by modification of the gut microbiota composition. We therefore investigated C. jejuni infection capacities in mice with acute or chronic intestinal inflammation displaying distinct microbiota shifts towards higher intestinal enterobacteria loads. Three-months-old adult C57BL/6 mice harboring a conventional gut microbiota were infected perorally with T. gondii in order to induce acute ileitis. Four days thereafter, the time point when initial histopathological changes in the ileum mucosa can be observed, mice were orally subjected to 109 CFU C. jejuni ATCC 43431 by gavage. At day 7 following T. gondii infection, mice suffered from acute ileitis and were prone to death. Interestingly, by that time (i.e. 3 days following C. jejuni ATCC 43431 infection) C. jejuni ATCC 43431 could be cultured from the entire GI tract of diseased, but not healthy mice with high loads of 109 CFU/g in duodenum, ileum and colon (Fig. 1A). Notably, the E. coli concentrations in the ileum lumen had increased by more than 6 orders of magnitude up to more than 1010 CFU/g during ileitis development (Fig. 1B). Live C. jejuni ATCC 43431 and commensal E. coli translocated into MLNs draining the small intestine (100±0% and 78.6±7.1%, respectively), spleen (57.1±1.49% and 50.0±0.68%, respectively) and liver (57.1±0% and 42.9±0%, respectively) (Fig. 1C). Strikingly, in 92.9±0.68%, but only in 7.1±0.7% of mice with acute ileitis following T. gondii infection, live C. jejuni ATCC 43431 and E. coli, respectively, could be cultured from cardiac blood indicating C. jejuni bacteremia (Fig. 1C). A molecular survey of the ileum microbiota composition revealed that concentrations of total eubacteria, enterobacteria, enterococci and Bacteroides/Prevotella spp. were increased in T. gondii-induced acute ileitis, whereas the lactobacilli population was lower as compared to healthy mice (Fig. 1D). Thus, during acute small intestinal inflammation accompanied by distinct intestinal microbiota shifts towards E. coli overgrowth in the GI tract colonization resistance against C. jejuni is abrogated with hazardous consequences for the host, as indicated by translocation and systemic spread of the pathogen.
Acute ileitis was induced in conventionally colonized 3-months-old C57BL/6 mice by oral infection with 100 cysts of Toxoplasma gondii ME49 strain. Four days following ileitis induction, mice were orally infected with C. jejuni strain ATCC 43431 as described in Methods. (A) At day 7 p.i. when T. gondii infected mice display severe ileitis (ILE) C. jejuni densities in distinct compartments of the gastrointestinal tract were determined by quantification of live C. jejuni in luminal samples taken from stomach, duodenum, ileum, and colon by cultural analysis (CFU, colony forming units; solid circles) and compared to control mice without ileitis (Con). (B) E. coli loads in luminal ileum samples of mice with T. gondii-induced acute ileitis (day 7 p.i.; open circles) were compared to naïve controls (open diamonds). (C) Relative translocation frequencies (%) of live C. jejuni 43431 ATCC (Cj, black bars) and E. coli (Ec, open bars) were determined in ex vivo biopsies of mesenteric lymphnodes (MLN), spleen, liver, and blood derived from mice with T. gondii-induced acute ileitis at day 7 p.i. by culture in enrichment broths. (D) A complex molecular analysis of the ileal microbiota composition was assessed in mice with T. gondii-induced acute ileitis at day 7 p.i. (ILE; solid circles) and compared to naïve control animals (N; open circles). Main gut bacterial groups were quantified by Real-Time-PCR analyses amplifying bacterial 16S rRNA variable regions and 16S rRNA gene numbers/ng DNA derived from luminal ileal samples: Enterobacteriaceae (EB), Enterococci (EC), Lactic acid bacteria (LB), Bifidobacteria (BIF), Bacteroides/Prevotella spp. (BP), Clostridium leptum group (CLEP), Clostridium coccoides group (CLOCC), Mouse intestinal bacteroidetes (MIB), and total eubacterial load (TL). Numbers of animals harboring C. jejuni or E. coli out of the total number of analyzed animals are given in parentheses. Medians (black bars), standard deviations and significance levels (P-values) determined by Mann-Whitney-U test are indicated. Data shown were pooled from at least three independent experiments.
IL-10 Deficient Mice with Chronic Colitis are Susceptible to C. jejuni Colonization
Next, we were interested whether colonization resistance against C. jejuni is also abrogated in mice with chronic intestinal inflammation accompanied with gut microbiota shifts towards higher commensal enterobacteria loads . To address this question we infected 6-months-old IL-10−/− mice with chronic colitis harboring a conventional gut microbiota with C. jejuni ATCC 43431. Twelve days following oral infection, C. jejuni could be detected in the ileum and colon (in 55.2% and 69.0% of infected mice, respectively), but not in the stomach or duodenum of IL-10−/− mice with colitis, whereas healthy wildtype control animals remained free of C. jejuni (Fig. 2A). Furthermore, the median C. jejuni loads in infected IL-10−/− mice were more than 4 orders of magnitude lower (105 CFU/g luminal content) when compared to mice with T. gondii-induced acute ileitis (Fig. 1A, 2A). Of note, uninfected IL-10−/− mice with colitis harbored 4 orders of magnitude higher E. coli counts in their large intestines when compared to healthy wildtype mice (Fig. 2B). Interestingly, live C. jejuni could be cultured only from less than 5% (3 out of 67) of MLNs derived from infected IL-10−/− mice, whereas no C. jejuni at all translocated into spleen, liver, kidney or cardiac blood at day 12 p.i. (not shown). A molecular survey of the main bacterial groups within the colonic lumen revealed that IL-10−/− mice with colitis harbored significantly higher numbers of enterobacteria, Bacteroides/Prevotella spp. and bifidobacteria, whereas lactobacilli, Clostridium leptum group and mouse intestinal bacteroidetes rRNA were lower as compared to wildtype controls, (Fig. 2C). Thus, colonization resistance against C. jejuni is abrogated by chronic large intestinal inflammation in IL-10−/− mice accompanied by distinct changes in colonic microbiota composition.
Conventionally colonized 6-months-old IL-10−/− mice with chronic colitis were orally infected with C. jejuni strain ATCC 43431 as described in Methods. (A) At day 14 p.i. C. jejuni densities in distinct compartments of the gastrointestinal tract were determined by quantification of live C. jejuni in luminal samples (CFU, colony forming units; solid circles) taken from stomach, duodenum, ileum, and colon by cultural analysis and compared to wildtype controls (WT, open circles). (B) Furthermore, E. coli loads in luminal colonic samples of C. jejuni infected IL-10−/− mice (IL-10, open circles) were compared to healthy wildtype control animals (WT, open diamonds). (C) A complex molecular analysis of the colonic microbiota composition was assessed in IL-10−/− mice with chronic colitis (IL-10; solid circles) and compared to wildtype control animals (WT; open circles). Main gut bacterial groups were quantified by Real-Time-PCR analyses amplifying bacterial 16S rRNA variable regions and 16S rRNA gene numbers/ng DNA derived from luminal colonic samples: Enterobacteriaceae (EB), Enterococci (EC), Lactic acid bacteria (LB), Bifidobacteria (BIF), Bacteroides/Prevotella spp. (BP), Clostridium leptum group (CLEP), Clostridium coccoides group (CLOCC), Mouse intestinal bacteroidetes (MIB), and total eubacterial load (TL). Numbers of animals harboring C. jejuni or E. coli out of the total number of analyzed animals are given in parentheses. Medians (black bars) and significance levels (P-values) determined by Mann-Whitney-U test are indicated. Data shown were pooled from at least three independent experiments.
C. jejuni Infection of Infant Mice Harboring a Conventional Intestinal Microbiota
In earlier studies weaned mice harboring a conventional gut microbiota were shown to be susceptible to C. jejuni infection , , , , . Very recently, we could demonstrate that compared to adult animals, infant mice harbored significantly higher concentrations of E. coli, Bacteroides/Prevotella spp. and Clostridium coccoides whereas lactobacilli, bifidobacteria and Clostridium leptum numbers were lower . To further unravel immunopathological consequences of abrogated colonization resistance in infant mice in more detail, we infected 3-weeks-old animals right after weaning with C. jejuni strains ATCC 43431 or B2. Adult mice 3 months of age with a common commensal gut flora served as controls. Kinetic analyses of fecal pathogen burden revealed that the infant mice harbored higher C. jejuni ATCC 43431 and B2 loads until days 5 and 6 p.i., respectively, when compared to adult animals (Fig. 3A, B). Interestingly, C. jejuni strain B2, but not strain ATCC 43431, could stably infect infant mice until day 6 p.i., as indicated by 66.7% and 30.0% of culture-positive cases, respectively, with a median load of approximately 105 CFU C. jejuni B2/g feces (Fig. 3A, B). Feces samples taken from naïve infant mice contained significantly higher numbers of E. coli (median 106 CFU/g), as compared to uninfected adult animals (Fig. 3C). Thus, these data provide evidence that E. coli might be causatively involved in the abrogation of murine colonization resistance against C. jejuni.
The analysis of the clinical outcome revealed that acute campylobacteriosis was self-limiting in infant mice. A cumulative score addressing the clinical condition and occurrence of bloody diarrhea in infected animals revealed that either C. jejuni strain induced enterocolitis within 24 hours p.i. remaining on a comparable level until day 6 p.i. (Fig. 3D). C. jejuni infected infant mice were not severely compromised clinically but developed mild diarrhea with haemoccult-positive feces irrespective of the C. jejuni strain analyzed. Furthermore, no live C. jejuni could be cultured from MLNs, spleen, liver, kidney or cardiac blood at day 6 following infection with either strain (not shown).
Conventionally colonized 3-weeks-old infant (3W; open circles) and 3-months-old adult (3M; solid circles) mice were orally infected with C. jejuni strain ATCC 43431 (A) or B2 (B) as described in Methods. Kinetic analyses of pathogen densities were performed by quantification of live C. jejuni in fecal samples by cultural analysis (CFU, colony forming units) at day (d) 6 p.i. ** indicates significant differences (p<0.01) of C. jejuni 43431 loads in infected infant mice at d1 versus d3, d5, and d6 p.i. (C) E. coli loads in colonic samples taken from naïve 3-weeks-old infant mice (3W, open diamonds) were compared to conventionally colonized 3-months-old adult mice (3M, open circles). (D) Kinetic analyses of disease activity of conventionally colonized 3-weeks-old infant mice following infection with C. jejuni strain ATCC 43431 (43431, open circles, n = 10) or B2 (solid circles, n = 12) were performed applying a standardized clinical score (see Methods). Numbers of animals harboring C. jejuni or E. coli out of the total number of analyzed animals are given in parentheses. Days past infection, medians (black bars) and significance levels (P-values) determined by Mann-Whitney-U test are indicated. Data shown were pooled from at least three independent experiments.
Pro-inflammatory Immune Responses Following C. jejuni Infection in Infant Mice
Since human campylobacteriosis is accompanied by an influx of immune cells to the inflamed colon , , , , we investigated numbers of apoptotic cells and immune cells within the colon mucosa of C. jejuni infected infant mice in situ by immunohistochemistry. Naive infant mice served as controls. At day 6 p.i., irrespective of the C. jejuni strain used, infected infant mice displayed a two-fold increase of apoptotic cells, paralleled by a significant increase of neutrophilic granulocytes and macrophages as compared to naive controls (Fig. 4A–C). In addition, both C. jejuni strains induced multifold increases in T-lymphocytes, regulatory T-cells (Tregs) as well as B-lymphocytes in infected versus naïve infant mice (Fig. 4D–F). The immune cell responses were accompanied by increased expression of pro-inflammatory cytokines and mediators in the colon. The levels of nitric oxide, TNF-α, IL-6, and MCP-1 were significantly higher in colon biopsies derived from C. jejuni ATCC 43431 and B2 infected mice at day 6 p.i. as compared to naive controls (Fig. 5A–D). For IFN-γ a significant increase was measured in C. jejuni B2 infected mice while C. jejuni ATCC 43431 infected mice displayed a trend towards elevated concentrations (Fig. 5E). Of note, following infection with either C. jejuni strain, adult mice harboring a conventional gut microbiota displayed neither clinical signs of enterocolitis nor relevant immune cell or pro-inflammatory responses in their colon (also refer to ). Taken together, C. jejuni infection of infant mice resulted in distinct pro-inflammatory responses and apoptotic changes in the colonic mucosa.
Conventionally colonized 3-weeks-old infant mice were orally infected with C. jejuni strain ATCC 43431 (crossed circles) or B2 (solid circles) as described in Methods. The average numbers of apoptotic cells (positive for caspase-3, panel A), neutrophilic granulocytes (neutrophils, positive for MPO-7, panel B), macrophages (positive for F4/80, panel C), T-lymphocytes (positive for CD3, panel D), regulatory T-cells (Tregs, positive for FOXP3, panel E) and B-lymphocytes (positive for B220, panel F) from at least six high power fields (HPF, ×400 magnification) per animal were determined microscopically in immunohistochemically stained colon sections at day 6 p.i. Uninfected mice (Naïve, open circles) served as negative controls. Numbers of analyzed animals are given in parentheses. Means (black bars) and levels of significance (P-values) determined by the Student’s t-test are indicated. Data shown are representative for three independent experiments.
Conventionally colonized 3-weeks-old infant mice were orally infected with C. jejuni strain ATCC 43431 (43431, crossed circles) or B2 (solid circles) as described in Methods. Concentrations (pg per mg colon) of (A) nitric oxide, (B) TNF-α, (C) IL-6, (D) MCP-1, and (E) IFN-γ were determined in supernatants of ex vivo colon cultures at day 6 p.i. by Griess reaction or CBA-ELISA, respectively. Uninfected mice (Naïve, open circles) served as negative controls. Numbers of analyzed animals are given in parentheses. Means (black bars) and levels of significance (P-values) as compared to the respective group (determined by Student’s t-test) are indicated. Data shown are representative for three independent experiments.
Treatment with Live E. coli Abrogates Murine Colonization Resistance Against C. jejuni
In order to dissect whether colonization resistance is abrogated by shifts in the microbiota composition or by conditions of inflammation we increased intestinal E. coli loads artificially by treating healthy adult mice with live E. coli in the drinking water (108 CFU/mL) and investigated the susceptibility to C. jejuni infection. Analysis of E. coli in feces samples revealed that the intestinal E. coli loads increased by more than 3 orders of magnitude (from median 105 CFU/g to 108 CFU/g feces) and remained stable as long as the E. coli suspension was kept (Fig. 6). Within 24 hours after withdrawl of the E. coli suspension, the intestinal E. coli loads decreased back to basal levels (Fig. 6). Oral infection of E. coli-treated mice and non-treated controls with C. jejuni strains ATCC 43431 or B2 revealed that E. coli did significantly increase the colonization capacity of either strain. (Fig. 7A and B). C. jejuni B2 could be cultured from the entire GI tract of E. coli treated animals at day 12 p.i. with highest counts in distal parts such as ileum and colon (Fig. 7C). In non-treated control animals, however, pathogen burdens were significantly lower as compared to mice with elevated intestinal E. coli load (Fig. 7).
In conventionally colonized 3-months-old adult mice (n = 6) intestinal E. coli loads were raised by adding live E. coli into the drinking water (+ E. coli, until day 8) as described in Methods. At day 8, the E. coli suspension was withdrawn and replaced by regular tap water. Kinetic analyses of E. coli densities were performed by quantification of live bacteria in fecal samples by cultural analysis (CFU, colony forming units) until day 13 p.i. Medians (black bars) and levels of significance (P-values) as compared to the indicated groups (determined by Mann-Whitney-U test) are indicated. Data shown are representative for three independent experiments.
In conventionally colonized 3-months-old adult mice intestinal E. coli loads were raised up to 108 bacteria per gram feces by adding live E. coli into the drinking water (+ E. coli) as described in Methods. Mice were orally infected with C. jejuni strain ATCC 43431 (A) or B2 (B). Kinetic analyses of pathogen densities were performed by quantification of live C. jejuni in fecal samples derived from mice with elevated (Ec; solid circles) or normal (N; open circles) E. coli loads by cultural analysis (CFU, colony forming units) until day 12 p.i. (C) C. jejuni B2 densities in distinct compartments of the gastrointestinal tract were determined in luminal samples taken from stomach, duodenum, ileum, and colon of mice with elevated (Ec; solid circles) or normal (N; open circles) E. coli loads at day 12 p.i. Numbers of animals harboring C. jejuni out of the total number of analyzed animals are given in parentheses. Days post infection, medians (black bars) and levels of significance (P-values) as compared to the respective group (determined by Mann-Whitney-U test) are indicated. Data shown are pooled from at least three independent experiments.
Pro-inflammatory Immune Responses in C. jejuni Infected Adult Mice with Elevated E. coli Loads
To further elucidate whether C. jejuni infection of adult mice with high intestinal E. coli loads was accompanied by pro-inflammatory immune responses, we analyzed apoptosis and inflammatory cells in the colon mucosa by in situ immunohistochemistry. Irrespective of the C. jejuni strain analyzed, infected mice displayed significantly higher numbers of apoptotic cells, neutrophils, macrophages, T- and B-lymphocytes as well as regulatory T-cells in the colon in situ at day 12 p.i. as compared to non-infected E. coli-treated control animals (Fig. 8).
In conventionally colonized 3-months-old adult mice intestinal E. coli loads were raised up to 108 bacteria per gram feces by adding live E. coli into the drinking water (+ E. coli) as described in Methods. Then, mice were orally infected with C. jejuni strain ATCC 43431 (crossed circles) or B2 (solid circles). The average numbers of apoptotic cells (positive for caspase-3, panel A), neutrophilic granulocytes (Neutrophils, positive for MPO-7, panel B), macrophages (positive for F4/80, panel C), T-lymphocytes (positive for CD3, panel D), regulatory T-cells (Tregs, positive for FOXP3, panel E) and B-lymphocytes (positive for B220, panel F) from at least six high power fields (HPF, 400×magnification) per animal were determined microscopically in immunohistochemically stained colon sections at day 12 p.i. Uninfected mice (Naïve, open circles) served as negative controls. Numbers of analyzed animals are given in parentheses. Means (black bars) and levels of significance (P-values) as compared to uninfected (Naïve) control mice (determined by the Student’s t-test) are indicated. Data shown are pooled from at least three independent experiments.
Secretion of nitric oxide, IL-6, MCP-1, IFN-γ and TNF-α in ex vivo colon biopsies were significantly increased at day 12 following infection with C. jejuni ATCC 43431 or B2 versus naïve animals with supra-physiological intestinal E. coli loads (Fig. 9). Thus, C. jejuni infection of adult mice treated with supra-physiological E. coli loads is accompanied by pro-inflammatory immunopathology of the colon. Of note, elevation of intestinal E. coli loads alone did not affect pro-inflammatory immune responses (data not shown). This was even true in our recent study when gnotobiotic mice generated by antibiotic treatment were replenished with a complex human microbiota .
In conventionally colonized 3-months-old adult mice intestinal E. coli loads were raised up to 108 bacteria per gram feces by adding live E. coli into the drinking water (+ E. coli) as described in Methods. Then, mice were orally infected with C. jejuni strain ATCC 43431 (crossed circles) or B2 (solid circles). Concentrations (pg per mg colon) of (A) nitric oxide, (B) IL-6, (C) MCP-1, (D) IFN-γ, and (E) TNF-α were determined in supernatants of ex vivo colon cultures at day 12 p.i. by Griess reaction or CBA-ELISA, respectively. Uninfected mice (Naïve, open circles) served as negative controls. Numbers of analyzed animals are given in parentheses. Means (black bars) and levels of significance (P-values) as compared to the uninfected (Naïve) group determined by Student’s t-test are indicated. Data shown were pooled from at least three independent experiments.
Taken together, physiological (infancy) and pathological conditions (e.g. acute and chronic intestinal inflammation in mice and men) associated with shifts of the intestinal microbiota composition such as elevated commensal E. coli loads facilitate C. jejuni infection inducing or perpetuating immunopathology. Most strikingly, the physiological CR against C. jejuni can be abrogated by elevating a single bacterial commensal gram-negative species such as E. coli within the complex commensal intestinal microbiota towards supra-physiological levels. These results point towards a pivotal role of E. coli in facilitating C. jejuni infection.
Investigations of the molecular mechanisms underlying C. jejuni induced immunpathology in vivo are hampered by colonization resistance of mice due to their host-specific intestinal microbiota composition. In the present study we investigated the susceptibility of an initially resistant host to C. jejuni during acute and chronic intestinal inflammation. In our acute T. gondii-induced ileitis model colonization resistance against C. jejuni was completely abolished and the pathogen colonized the entire GI tract with high counts. Moreover, C. jejuni and commensal E. coli translocated into MLNs, spleen, and liver. Strikingly, C. jejuni could be cultured from cardiac blood samples in >90% of infected animals displaying severe ileitis indicating bacteremia and subsequent sepsis. The impact of intestinal inflammation in overcoming colonization resistance was further supported by the finding that IL-10−/− mice suffering from chronic colitis were susceptible to C. jejuni infection. This is well in line with previous findings of Mansfield et al. showing that C. jejuni induced significant clinical signs of enterocolitis following infection of IL-10−/− mice . However, in the IL-10−/− mice analyzed here the clinical signs of enterocolitis following C. jejuni infection remained rather subtle. The reason for this discrepancy in colitis severity is most likely attributed to differences in the genomic contents of the C. jejuni strains used.
The intestinal inflammation models exerting a compromised colonization resistance against C. jejuni had all in common significantly higher E. coli loads at inflamed tissue site, for instance, when compared to healthy controls with an intact colonization resistance, irrespective whether the inflammatory condition was acute or chronic, located in the small or large intestine. These findings are well in line with results from previous studies pointing towards a weakness of the physiological barrier during intestinal inflammation , , , . Bacteremia and sepsis display hazardous sequelae of C. jejuni infection in hosts with a compromized epithelial barrier function (e.g. Crohn’s disease patients with acute episodes). A high prevalence of C. concisus was recently detected in paediatric CD as well as adult UC patients , . Although distinct from IBD, irritable bowel syndrome was described to occur as a potential post-infectious event following C. jejuni infection . Taken together, these findings once more point towards a correlation between inflammation, intestinal microbiota and susceptibility to C. jejuni infection.
Given that colonization of the intestinal tract starts right after birth and complexity increases thereafter, the question arises whether infant mice already display a fully established colonization resistance. Neonatal mice had been used in Campylobacter research already three decades ago, investigating transmissibility from infant to dam, virulence factors, vaccines and the efficacy of bismuth subsalicylate in preventing the growth and colonization of the intestinal tract , , , , . Remarkably, infection of 3-weeks-old mice right after weaning in our study revealed that C. jejuni B2, and to a less distinct extent strain ATCC 43431, could stably colonize the GI tract of mice. Moreover, infant mice displayed approximately 2–3 orders of magnitude higher E. coli counts in their feces as compared to 3-months-old adult animals. Thus, there is evidence that an altered microbiota towards higher enterobacteria numbers (such as E. coli) might facilitate C. jejuni infection. When assessing C. jejuni induced clinical pathology due to blood-positive stool samples and mild diarrhea in the first week following infection, the clinical outcomes were similar in infant mice infected with C. jejuni strains B2 or ATCC 43431. Both strains originate from humans with severe enteritis, are invasive in vitro and induced distinct pro-inflammatory immune responses within the colon mucosa of C. jejuni infected infant mice. The elevated intestinal E. coli loads in mice suffering from intestinal inflammation and in infant mice provide evidence that E. coli is essentially involved in abrogation of CR. This is further underlined by a substantially increased colonization capacity of C. jejuni in adult mice in which the commensal intestinal microbiota was artificially modified towards supra-physiological intestinal E. coli loads. This finding demonstrates again that CR against C. jejuni is mediated by the microbiota and that shifts in the flora represent a risk for C. jejuni infection.
We conclude that the murine models of C. jejuni infection presented here shed further lights into the mechanisms underlying CR against C. jejuni offering promising tools to further unravel the interplay between intestinal pathogens, gut microbiota and the innate and adaptive immune system.
Materials and Methods
Animal experiments were conducted according to the European Guidelines for animal welfare (2010/63/EU) with approval of the commission for animal experiments headed by the “Landesamt für Gesundheit und Soziales” (LaGeSo, Berlin, Germany; Registration numbers: G0173/07 and G135/10). Animal welfare was monitored twice daily by assessment of clinical conditions.
Mice were bred and maintained in the facilities of the “Forschungsinstitut für Experimentelle Medizin” (FEM, Charité - Universitätsmedizin, Berlin, Germany), under specific pathogen-free (SPF) conditions. C57BL/6 mice were used in the experiments. Age matched female mice 3 weeks (right after weaning; infant mice) and 3 months of age (adult mice) as well as 6 months old IL-10−/− mice (C57BL/6 background) with overt chronic colitis were used.
To assess clinical/macroscopic signs of C. jejuni induced infection on a daily basis, a standardized cumulative clinical score (maximum 12 points; modified according to  addressing the occurrence of blood in feces (0 points: no blood; 2 points: microscopic detection of blood by the Guajac method using Haemoccult, Beckman Coulter/PCD, Krefeld, Germany; 4 points: macroscopic blood visible), diarrhea (0: formed feces; 2: pasty feces; 4: liquid feces), and the clinical aspect (0: normal; 2: ruffled fur, less locomotion; 4: isolation, severely compromized locomotion, pre-final aspect) was used.
Induction of Acute Ileitis
For induction of ileitis, C57BL/6 mice were infected orally with 100 Toxoplasma (T.) gondii cysts (ME49 strain) from homogenized brains of intraperitoneally infected NMRI mice in a volume of 0.3 ml phosphate-buffered saline (PBS, pH 7.4) by gavage, as described previously , , , .
C. jejuni Infection of Mice
Mice were infected with approximately 109 viable CFU of C. jejuni strains ATCC 43431 or B2 by gavage in a total volume of 0.3 mL PBS on two consecutive days as described earlier in detail . C. jejuni strain B2 was kindly provided by Prof. Dr. Uwe Groß, University of Göttingen, Germany.
Elevation of Intestinal E. coli Loads
In order to raise intestinal E. coli loads in 3-months-old adult mice harboring a conventional gut microbiota to supra-physiologic levels, live E. coli were isolated from naïve, healthy 3 months old C57BL/6 mice by culture on MacConkey media. This strain was shown to be a common commensal in healthy mice. PCR-based detection revealed that this strain does not contain known virulence factors of pathogenic E. coli such as stx 1 and 2, catA, hlyA, cspA, katP, and astA. This commensal E. coli strain was subcultivated on blood agar and added to the drinking water at a final concentration of 108 CFU per mL suspension starting at day 0 in naïve animals. The E. coli suspension was renewed daily. For kinetic experiments, the E. coli suspension was withdrawn after 8 days and replaced by regular tap water thereafter. For C. jejuni infection studies, the application of E. coli in drinking water was started 8 days prior the first oral C. jejuni infection and continued until the end of the experiment (12 days following the latest C. jejuni gavage).
Sampling Procedures, Determination of Colon Length and Histopathology
Mice were sacrificed by isofluran treatment (Abbott, Germany). From each mouse samples derived from the GI tract (stomach, duodenum, ileum, colon) were collected in parallel for histological, microbiological, immunobiological and molecular analyses. Cardiac blood and tissue samples from mesenteric lymphnodes, spleen, liver and the GI tract were asserved under sterile conditions. Histopathological changes were determined in colon samples immediately fixed in 5% formalin and embedded in paraffin. Sections (5 µm) were stained with respective antibodies for immunohistochemistry.
In situ immunohistochemical analysis of colonic paraffin sections were performed as described previously , , . Primary antibodies against CD3 (#N1580, Dako, Denmark, dilution 1∶10), FOXP-3 (FJK-16s, eBioscience, 1∶100), B220 (eBioscience, San Diego, CA, USA, 1∶200), myeloperoxidase-7 (MPO-7, # A0398, Dako, 1∶400), F4/80 (#14–4801, clone BM8, eBioscience, 1∶50), and cleaved caspase-3 (Asp175, Cell Signaling, USA, 1∶200) were used. For each animal, the average number of positive stained cells within at least six independent high power fields (HPF, 400×magnification) were determined microscopically and subjected to statistical analysis as indicated.
Cytokine Detection in Colon Culture Supernatants
Colon biopsies were cut longitudinally, washed in PBS and strips of 1 cm2 placed in 24-flat-bottom well culture plates (Nunc, Wiesbaden, Germany) containing 500 µl serum-free RPMI 1640 medium supplemented with penicillin (100 U/ml) and streptomycin (100 ug/ml; PAA Laboratories). After 18 h at 37°C, culture supernatants were tested for TNF-α, IL-6, MCP-1 and IFN-γ by the Mouse Inflammation Cytometric Bead Assay (CBA; BD Biosciences) on a BD FACSCanto II flow cytometer (BD Biosciences). Nitric oxide (NO) was determined by Griess-reaction as described earlier .
Analysis of the Intestinal Microflora
Cultural analyses, biochemical identification and molecular detection of luminal bacterial communities from stomach, duodenum, ileum, and colon as well as feces were performed as previously described , , , .
Assessment of Bacterial Translocation Frequencies
For bacterial translocation experiments, mesenteric lymphnodes (MLNs) draining the small intestine, spleen, liver (1 cm2) and cardiac blood (1 mL) were transferred into thioglycollate enrichment broths under sterile conditions and cultivated for 7 days. Turbid broths were streaked onto sheep blood, MacConkey as well as Karmali agar in order to detect translocated live E. coli and C. jejuni followed by biochemical species identification (API 20E and API Campy, respectively, Biomérieux). Given that bacterial growth in enrichment broths cannot be quantitated, the relative rates of translocating C. jejuni or E. coli into the respective extra-intestinal compartment are expressed in % (number of broths culture-positive for either E. coli or C. jejuni multiplied with 100 and divided by the total number of broths inoculated with the respective organ or blood).
Mean values, medians, standard deviations, and levels of significance were determined using appropriate tests as indicated (two-tailed Student’s t-Test, Mann-Whitney-U Test). Two-sided probability (P) values ≤0.05 were considered significant. All experiments were repeated at least twice.
We thank Michaela Wattrodt, Ursula Rüschendorf, Gernot Reifenberger, Uwe Lohmann, Silvia Boelke, and the staff of the animal research facility for excellent technical assistance and animal breeding. We are grateful to Simone Spieckermann for immunohistochemistry staining of colon sections.
Conceived and designed the experiments: SB AF UBG LMH MMH. Performed the experiments: AF LMH BO MMH RP. Analyzed the data: AF LMH BO AAK MMH RP. Contributed reagents/materials/analysis tools: AAK. Wrote the paper: LMH SB AF AAK MMH.
- 1. Hermans D, Pasmans F, Messens W, Martel A, Van Immerseel F, et al. (2012) Poultry as a Host for the Zoonotic Pathogen Campylobacter jejuni. Vector Borne Zoonotic Dis 12: 89–98.
- 2. Man SM (2011) The clinical importance of emerging Campylobacter species. Nat Rev Gastroenterol Hepatol 8: 669–685.
- 3. Lane JA, Mehra RK, Carrington SD, Hickey RM (2010) The food glycome: a source of protection against pathogen colonization in the gastrointestinal tract. Int J Food Microbiol 142: 1–13.
- 4. Guerry P, Szymanski CM (2008) Campylobacter sugars sticking out. Trends Microbiol 16: 428–435.
- 5. Young KT, Davis LM, Dirita VJ (2007) Campylobacter jejuni: molecular biology and pathogenesis. Nat Rev Microbiol 5: 665–679.
- 6. Kist M, Bereswill S (2001) Campylobacter jejuni. Contrib Microbiol 8: 150–165.
- 7. van Spreeuwel JP, Duursma GC, Meijer CJ, Bax R, Rosekrans PC, et al. (1985) Campylobacter colitis: histological immunohistochemical and ultrastructural findings. Gut 26: 945–951.
- 8. Walker RI, Caldwell MB, Lee EC, Guerry P, Trust TJ, et al. (1986) Pathophysiology of Campylobacter enteritis. Microbiol Rev 50: 81–94.
- 9. Ketley JM (1997) Pathogenesis of enteric infection by Campylobacter. Microbiology 143 (Pt 1): 5–21.
- 10. Allos BM (1997) Association between Campylobacter infection and Guillain-Barre syndrome. J Infect Dis 176: Suppl 2S125–128.
- 11. Butzler JP (2004) Campylobacter, from obscurity to celebrity. Clin Microbiol Infect 10: 868–876.
- 12. Dorrell N, Wren BW (2007) The second century of Campylobacter research: recent advances, new opportunities and old problems. Curr Opin Infect Dis 20: 514–518.
- 13. Dasti JI, Tareen AM, Lugert R, Zautner AE, Gross U (2010) Campylobacter jejuni: a brief overview on pathogenicity-associated factors and disease-mediating mechanisms. Int J Med Microbiol 300: 205–211.
- 14. Yrios JW, Balish E (1986) Immune response of athymic and euthymic germfree mice to Campylobacter spp. Infect Immun 54: 339–346.
- 15. Yrios JW, Balish E (1986) Colonization and infection of athymic and euthymic germfree mice by Campylobacter jejuni and Campylobacter fetus subsp. fetus. Infect Immun 53: 378–383.
- 16. Yrios JW, Balish E (1986) Pathogenesis of Campylobacter spp. in athymic and euthymic germfree mice. Infect Immun 53: 384–392.
- 17. Bereswill S, Fischer A, Plickert R, Haag LM, Otto B, et al. (2011) Novel Murine Infection Models Provide Deep Insights into the “Menage a Trois” of Campylobacter jejuni, Microbiota and Host Innate Immunity. PLoS One 6: e20953.
- 18. Bereswill S, Plickert R, Fischer A, Kühl AA, Loddenkemper C, et al. (2011) What you eat is what you get: Novel Campylobacter models in the quadrangle relationship between nutrition, obesity, microbiota and susceptibiliy to infection. Eur J Microbiol Immunol 1: 237–248.
- 19. Stecher B, Hardt WD (2011) Mechanisms controlling pathogen colonization of the gut. Curr Opin Microbiol 14: 82–91.
- 20. Stecher B, Robbiani R, Walker AW, Westendorf AM, Barthel M, et al. (2007) Salmonella enterica serovar typhimurium exploits inflammation to compete with the intestinal microbiota. PLoS Biol 5: 2177–2189.
- 21. Stecher B, Chaffron S, Kappeli R, Hapfelmeier S, Freedrich S, et al. (2010) Like will to like: abundances of closely related species can predict susceptibility to intestinal colonization by pathogenic and commensal bacteria. PLoS Pathog 6: e1000711.
- 22. Heimesaat MM, Bereswill S, Fischer A, Fuchs D, Struck D, et al. (2006) Gram-negative bacteria aggravate murine small intestinal Th1-type immunopathology following oral infection with Toxoplasma gondii. J Immunol 177: 8785–8795.
- 23. Heimesaat MM, Fischer A, Jahn HK, Niebergall J, Freudenberg M, et al. (2007) Exacerbation of murine ileitis by Toll-like receptor 4 mediated sensing of lipopolysaccharide from commensal Escherichia coli. Gut 56: 941–948.
- 24. Heimesaat MM, Fischer A, Siegmund B, Kupz A, Niebergall J, et al. (2007) Shift towards pro-inflammatory intestinal bacteria aggravates acute murine colitis via Toll-like receptors 2 and 4. PLoS One 2: e662.
- 25. Erridge C, Duncan SH, Bereswill S, Heimesaat MM (2010) The induction of colitis and ileitis in mice is associated with marked increases in intestinal concentrations of stimulants of TLRs 2, 4, and 5. PLoS One 5: e9125.
- 26. Kalischuk LD, Buret AG (2010) A role for Campylobacter jejuni-induced enteritis in inflammatory bowel disease? Am J Physiol Gastrointest Liver Physiol 298: G1–9.
- 27. Jess T, Simonsen J, Nielsen NM, Jorgensen KT, Bager P, et al. (2011) Enteric Salmonella or Campylobacter infections and the risk of inflammatory bowel disease. Gut 60: 318–324.
- 28. Swidsinski A, Ladhoff A, Pernthaler A, Swidsinski S, Loening-Baucke V, et al. (2002) Mucosal flora in inflammatory bowel disease. Gastroenterology 122: 44–54.
- 29. Swidsinski A, Weber J, Loening-Baucke V, Hale LP, Lochs H (2005) Spatial organization and composition of the mucosal flora in patients with inflammatory bowel disease. J Clin Microbiol 43: 3380–3389.
- 30. Man SM, Kaakoush NO, Mitchell HM (2011) The role of bacteria and pattern-recognition receptors in Crohn’s disease. Nat Rev Gastroenterol Hepatol 8: 152–168.
- 31. Gradel KO, Nielsen HL, Schonheyder HC, Ejlertsen T, Kristensen B, et al. (2009) Increased short- and long-term risk of inflammatory bowel disease after salmonella or campylobacter gastroenteritis. Gastroenterology 137: 495–501.
- 32. Helms M, Simonsen J, Molbak K (2006) Foodborne bacterial infection and hospitalization: a registry-based study. Clin Infect Dis 42: 498–506.
- 33. Garcia Rodriguez LA, Ruigomez A, Panes J (2006) Acute gastroenteritis is followed by an increased risk of inflammatory bowel disease. Gastroenterology 130: 1588–1594.
- 34. Wohlgemuth S, Keller S, Kertscher R, Stadion M, Haller D, et al. (2011) Intestinal steroid profiles and microbiota composition in colitic mice. Gut Microbes 2: 159–166.
- 35. Diker KS, Hascelik G, Diker S (1992) Colonization of infant mice with flagellar variants of Campylobacter jejuni. Acta Microbiol Hung 39: 133–136.
- 36. Hanninen ML (1989) Effect of endotoxin on colonisation of Campylobacter jejuni in infant mice. J Med Microbiol 30: 199–206.
- 37. Hanninen ML (1990) Bismuth subsalicylate in the prevention of colonization of infant mice with Campylobacter jejuni. Epidemiol Infect 104: 397–404.
- 38. Abimiku AG, Dolby JM (1987) The mechanism of protection of infant mice from intestinal colonisation with Campylobacter jejuni. J Med Microbiol 23: 339–344.
- 39. Field LH, Headley VL, Payne SM, Berry LJ (1986) Influence of iron on growth, morphology, outer membrane protein composition, and synthesis of siderophores in Campylobacter jejuni. Infect Immun 54: 126–132.
- 40. Haag LM, Fischer A, Otto B, Grundmann U, Kühl AA, et al. (2012) Campylobacter jejuni infection of infant mice: acute enterocolitis is followed by asymptomatic intestinal and extra-intestinal immune response. Eur J Microbiol Immunol 2: 2–11.
- 41. Mansfield LS, Bell JA, Wilson DL, Murphy AJ, Elsheikha HM, et al. (2007) C57BL/6 and congenic interleukin-10-deficient mice can serve as models of Campylobacter jejuni colonization and enteritis. Infect Immun 75: 1099–1115.
- 42. Lupp C, Robertson ML, Wickham ME, Sekirov I, Champion OL, et al. (2007) Host-mediated inflammation disrupts the intestinal microbiota and promotes the overgrowth of Enterobacteriaceae. Cell Host Microbe 2: 204.
- 43. Man SM, Zhang L, Day AS, Leach ST, Lemberg DA, et al. (2010) Campylobacter concisus and other Campylobacter species in children with newly diagnosed Crohn’s disease. Inflamm Bowel Dis 16: 1008–1016.
- 44. Mahendran V, Riordan SM, Grimm MC, Tran TA, Major J, et al. (2011) Prevalence of Campylobacter species in adult Crohn’s disease and the preferential colonization sites of Campylobacter species in the human intestine. PLoS One 6: e25417.
- 45. Marshall JK, Thabane M, Garg AX, Clark WF, Salvadori M, et al. (2006) Incidence and epidemiology of irritable bowel syndrome after a large waterborne outbreak of bacterial dysentery. Gastroenterology 131: 445–450.
- 46. Field LH, Underwood JL, Pope LM, Berry LJ (1981) Intestinal colonization of neonatal animals by Campylobacter fetus subsp. jejuni. Infect Immun 33: 884–892.
- 47. Siegmund B, Fantuzzi G, Rieder F, Gamboni-Robertson F, Lehr HA, et al. (2001) Neutralization of interleukin-18 reduces severity in murine colitis and intestinal IFN-gamma and TNF-alpha production. Am J Physiol Regul Integr Comp Physiol 281: R1264–1273.
- 48. Bereswill S, Munoz M, Fischer A, Plickert R, Haag LM, et al. (2010) Anti-inflammatory effects of resveratrol, curcumin and simvastatin in acute small intestinal inflammation. PLoS One 5: e15099.
- 49. Munoz M, Heimesaat MM, Danker K, Struck D, Lohmann U, et al. (2009) Interleukin (IL)-23 mediates Toxoplasma gondii-induced immunopathology in the gut via matrixmetalloproteinase-2 and IL-22 but independent of IL-17. J Exp Med 206: 3047–3059.
- 50. Heimesaat MM, Nogai A, Bereswill S, Plickert R, Fischer A, et al. (2010) MyD88/TLR9 mediated immunopathology and gut microbiota dynamics in a novel murine model of intestinal graft-versus-host disease. Gut 59: 1079–1087.