Toll-Like Receptor-4 Is Involved in Mediating Intestinal and Extra-Intestinal Inflammation in Campylobacter coli-Infected Secondary Abiotic IL-10−/− Mice

Human Campylobacter infections are emerging worldwide and constitute significant health burdens. We recently showed that the immunopathological sequelae in Campylobacter jejuni-infected mice were due to Toll-like receptor (TLR)-4 dependent immune responses induced by bacterial lipooligosaccharide (LOS). Information regarding the molecular mechanisms underlying Campylobacter coli-host interactions are scarce, however. Therefore, we analyzed C. coli-induced campylobacteriosis in secondary abiotic IL-10−/− mice with and without TLR4. Mice were infected perorally with a human C. coli isolate or with a murine commensal Escherichia coli as apathogenic, non-invasive control. Independent from TLR4, C. coli and E. coli stably colonized the gastrointestinal tract, but only C. coli induced clinical signs of campylobacteriosis. TLR4−/− IL-10−/− mice, however, displayed less frequently fecal blood and less distinct histopathological and apoptotic sequelae in the colon versus IL-10−/− counterparts on day 28 following C. coli infection. Furthermore, C. coli-induced colonic immune cell responses were less pronounced in TLR4−/− IL-10−/− as compared to IL-10−/− mice and accompanied by lower pro-inflammatory mediator concentrations in the intestines and the liver of the former versus the latter. In conclusion, our study provides evidence that TLR4 is involved in mediating C. coli-LOS-induced immune responses in intestinal and extra-intestinal compartments during murine campylobacteriosis.


Introduction
Campylobacter infections are among the most prevalent causes of bacterial infectious gastroenteritis worldwide [1,2]. In most cases, C. jejuni and, less frequently, C. coli induce the diarrheal disease complex campylobacteriosis in infected human patients. The acute phase of campylobacteriosis is characterized by diarrhea and abdominal cramps, commonly accompanied by fever and bloody stools [3,4]. The disease is mostly self-limiting, and antimicrobial therapy is, therefore, only needed in severe incidents, particularly in infected immunocompromised individuals. In rare cases, however, campylobacteriosis is associated with long-term post-infectious sequelae, such as the autoimmune diseases Guillain-Barré syndrome (GBS), reactive arthritis, or intestinal inflammatory morbidities, IL-10 −/− mice carrying a human gut microbiota displayed less distinct immune responses upon peroral C. coli infection as compared to IL-10 −/− counterparts [36].
In order to investigate C. coli-induced campylobacteriosis and to determine the immunopathogenic role of C. coli LOS in the absence of any commensal gut microbiota, we here assessed the gastrointestinal colonization properties and the clinical, macroscopic, and microscopic inflammatory sequelae in intestinal and extra-intestinal compartments following peroral challenge of secondary abiotic TLR4-deficient IL-10 −/− mice.

Ethics Statement
The mouse studies were approved by the local commission for animal experiments ("Landesamt für Gesundheit und Soziales", LaGeSo, Berlin, Germany, registration numbers G0172/16 and G0247/16) and performed according to the European Guidelines for animal welfare (2010/63/EU). Throughout the experiment, the clinical conditions of mice were evaluated twice daily.

Bacterial Challenge and Gastrointestinal Colonization and Translocation
The used C. coli strain had originally been isolated from the stool of a patient suffering from bloody diarrhea and was kindly provided by Dr. Torsten Semmler (Robert-Koch-Institute, Berlin, Germany). A commensal intestinal Escherichia coli isolate from a healthy wildtype mouse served as a Gram-negative rod control [38,39]. The absence of commonly known virulence factors of pathogenic E. coli, such as stx 1 and 2, catA, hlyA, cspA, katP, and astA, was verified by PCR analysis [39].
On two consecutive days (i.e., days 0 and 1), sex-and age-matched secondary abiotic mice (three months of age) were perorally challenged with 10 9 colony forming units (CFU) of the C. coli patient isolate or the commensal E. coli strain by gavage. Throughout the experiment, mice were maintained in a sterile environment (autoclaved food and drinking water) and handled under strict aseptic conditions to prevent contamination.
To monitor the colonization properties over a 28-day-period, C. coli and E. coli loads were enumerated in fecal samples collected on selected days post-infection (p.i.) and in luminal samples derived from distinct parts of the gastrointestinal tract (i.e., from the stomach, duodenum, ileum, and colon) upon necropsy by culture, as stated elsewhere [16,40]. In order to quantity C. coli burdens, samples were homogenized with a sterile pistil and serial dilutions plated onto Columbia agar plates containing 5% sheep blood and, additionally, onto selective Karmali plates (both from Oxoid, Wesel, Germany). The inoculated plates were incubated in a jar under microaerophilic conditions for 48 h at 37 • C. E. coli loads were enumerated in serial dilutions of homogenized fecal samples plated onto Columbia agar plates containing 5% sheep blood and selective MacConkey agar plates (Oxoid, Wesel, Germany) following incubation for 48 h at 37 • C in aerobic atmosphere. Bacterial translocation of the respective bacterial strains was assessed in homogenized ex vivo biopsies obtained from mesenteric lymph nodes (MLN), spleen, kidneys, liver, and lungs and from cardiac blood, as described elsewhere [16,40]. The detection limit of viable pathogens was ≈ 100 CFU per g (CFU/g).

Clinical Conditions
Before and after bacterial application, the clinical conditions of mice were evaluated by daily applying a standardized clinical score (maximum 12 points). The score addressed the clinical aspect/wasting (0: normal; 1: ruffled fur; 2: less locomotion; 3: isolation; 4: severely compromised locomotion, pre-final aspect), the abundance of blood in feces (0: no blood; 2: microscopic detection of blood by the Guajac method using Hemoccult, Beckman Coulter, Krefeld, Germany; 4: macroscopic blood visible), and diarrhea (0: formed feces; 2: pasty feces; 4: liquid feces), as described earlier [33]. Fecal blood positivity rates were determined by the ratio of overt and occult fecal blood positive mice to the total number of analyzed animals.

Sampling Procedures
On day 28 p.i., mice were sacrificed by CO 2 asphyxiation. Immediately after, ex vivo biopsies from spleen, kidneys, liver, lungs, and MLN, as well as luminal gastrointestinal samples (from the stomach, duodenum, ileum, and colon), were isolated under sterile conditions. Cardiac blood was obtained for serum cytokine measurements. Colon, MLN, and liver tissue samples were collected from each mouse for subsequent immunological, histological, and microbiological analyses.

Histopathology
Histological analyses were performed in colonic ex vivo biopsies following immediate fixation in 5% formalin and embedding in paraffin. Sections (5 µm) were stained with hematoxylin and eosin (H&E), examined by light microscopy (100× magnification), and histopathological changes in the large intestines were quantitatively assessed by applying an established histopathological scoring system, as reported previously [41]: Score 1: minimal inflammatory cell infiltrates in the mucosa with intact epithelium. Score 2: mild inflammatory cell infiltrates in the mucosa and submucosa with mild hyperplasia and mild goblet cell loss. Score 3: moderate inflammatory cell infiltrates in the mucosa with moderate goblet cell loss. Score 4: marked inflammatory cell infiltration into the mucosa and submucosa with marked goblet cell loss, multiple crypt abscesses, and crypt loss.

Role of TLR4 in
Intestinal Colonization of C. coli and E. coli in Secondary Abiotic IL-10−/− Mice We first surveyed the intestinal colonization properties of pathogenic C. coli and commensal E. coli following oral challenge of secondary abiotic IL-10 −/− mice with or without TLR4. Our cultural analyses revealed stable intestinal colonization efficiencies of either strain that was TLR4-independent, as indicated by high median fecal loads of more than 10 8 viable bacterial cells per g feces in both IL-10 −/− mice lacking TLR4 and IL-10 −/− counterparts ( Figure 1). From day 2 until day 28 post-challenge, fecal C. coli numbers marginally decreased by approximately 1.0-1.5 orders in mice of either genotype (p < 0.01-0.001; Figure 1C,D).
Upon necropsy, comparable E. coli numbers could be cultivated from luminal samples taken from stomach, duodenum, ileum, and colon of TLR4 −/− IL-10 −/− as compared to IL-10 −/− mice (not significant (n.s.); Figure 2A), which also held true for C. coli, except for gastric luminal C. coli loads that were approximately two orders of magnitude higher in the former as compared to the latter at day 28 post-challenge (p < 0.001; Figure 2B). Furthermore, no viable C. coli or E. coli could be isolated from extra-intestinal tissue sites, such as liver, kidneys, lungs, and cardiac blood (data not shown). Hence, upon peroral application, C. coli and E. coli stably colonized the murine gastrointestinal tract at high loads in a TLR4-independent fashion.

Role of TLR4 in Intestinal Colonization of C. coli and E. coli in Secondary Abiotic IL-10-/-Mice
We first surveyed the intestinal colonization properties of pathogenic C. coli and commensal E. coli following oral challenge of secondary abiotic IL-10 -/-mice with or without TLR4. Our cultural analyses revealed stable intestinal colonization efficiencies of either strain that was TLR4independent, as indicated by high median fecal loads of more than 10 8 viable bacterial cells per g feces in both IL-10 -/-mice lacking TLR4 and IL-10 -/-counterparts ( Figure 1). From day 2 until day 28 post-challenge, fecal C. coli numbers marginally decreased by approximately 1.0-1.5 orders in mice of either genotype (p < 0.01-0.001; Figure 1C,D). Upon necropsy, comparable E. coli numbers could be cultivated from luminal samples taken from stomach, duodenum, ileum, and colon of TLR4 -/-IL-10 -/-as compared to IL-10 -/-mice (not significant (n.s.); Figure 2A), which also held true for C. coli, except for gastric luminal C. coli loads that were approximately two orders of magnitude higher in the former as compared to the latter at day 28 post-challenge (p < 0.001; Figure 2B). Furthermore, no viable C. coli or E. coli could be isolated from extra-intestinal tissue sites, such as liver, kidneys, lungs, and cardiac blood (data not shown). Hence, upon peroral application, C. coli and E. coli stably colonized the murine gastrointestinal tract at high loads in a TLR4-independent fashion.

Kinetic Survey of Clinical Signs Following Peroral C. coli or E. coli Application to Secondary Abiotic IL-10−/− Mice Lacking TLR4
We further quantitatively surveyed clinical signs following peroral bacterial challenge over time by applying a standardized scoring system assessing key features of severe human campylobacteriosis, such as wasting and bloody diarrhea. Whereas mice of either genotype were virtually uncompromised following E. coli challenge (Supplementary Figure S1A,B), clinical scores were significantly higher in IL-10 −/− mice at days 7 and 28 post-C. coli challenge as compared to unchallenged conditions (p < 0.001 and p < 0.05, respectively), which also held true for day 7 in TLR4 −/− IL-10 −/− mice (p < 0.001; Supplementary Figure  commensal murine E. coli strain ((A) squares) or a C. coli patient isolate ((B) circles) on days 0 and 1. On day 28 post-challenge, luminal bacterial loads were quantitatively surveyed in distinct gastrointestinal compartments by culture (in colony-forming units per g; CFU/g). Medians (black bars), levels of significance (p-values) as determined by the Mann-Whitney U test, and numbers of culture-positive mice out of the total number of analyzed animals (in parentheses) are given. Data were pooled from three independent experiments.

Kinetic Survey of Clinical Signs Following Peroral C. coli or E. coli Application to Secondary Abiotic IL-10-/-Mice Lacking TLR4
We further quantitatively surveyed clinical signs following peroral bacterial challenge over time by applying a standardized scoring system assessing key features of severe human campylobacteriosis, such as wasting and bloody diarrhea. Whereas mice of either genotype were virtually uncompromised following E. coli challenge (Supplementary Figure S1A,B), clinical scores were significantly higher in IL-10 -/-mice at days 7 and 28 post-C. coli challenge as compared to unchallenged conditions (p < 0.001 and p < 0.05, respectively), which also held true for day 7 in TLR4 -/-IL-10 -/-mice (p < 0.001; Supplementary Figure S1C,D). Notably, none of the C. coli-infected TLR4 -/-IL-10 -/-mice and only 18.8% of the IL-10 -/-counterparts were suffering from acute campylobacteriosis (as indicated by clinical scores of at ≥10), whereas 81.8% and 56.3% of infected TLR4 -/-IL-10 -/-and IL-10 -/-mice remained clinically unaffected, respectively (Supplementary Figure S1C,D). When focusing on the abundance of fecal blood, however (Figure 3), a maximum of 93.8% of IL-10 -/-mice were fecal blood-positive as early as 5 days post-challenge ( Figure 3C), which was the case for 95.5% of TLR4 -/-IL-10 -/-mice ( Figure 3D), whereas fecal blood was less frequently detected at day 14 and thereafter in the latter as compared to the former ( Figure 3C,D). At the end of the observation period, the fecal blood positivity rates were lower in IL-10 -/-mice lacking TLR4 as compared to IL-10 -/-counterparts (18.2% versus 43.8%; Figure 3C,D). Hence, pathogenic C. coli but not commensal E. coli caused differential kinetics in fecal bleeding in IL-10 -/-and TLR4 -/-IL-10 -/-mice, indicating that TLR4 was involved in mediating the pathogen-induced signs of intestinal disease.

Role of TLR4 in C. coli-Mediated Colonic Histopathology and Apoptosis
We next assessed potential TLR4-dependent inflammatory responses affecting the large intestines upon C. coli versus E. coli challenge of secondary abiotic IL-10 −/− mice. Therefore, the histopathological changes within the colon were quantified by applying an established histopathological scoring system [41]. On day 28 following C. coli infection, higher histopathological scores were assessed in IL-10 −/− when compared to TLR4 −/− IL-10 −/− mice (p < 0.001) and also when compared to E. coli challenged IL-10 −/− control mice (p < 0.001; Figure 4A; Supplementary Figure S2A).
Since apoptosis is a hallmark of C. jejuni-induced murine campylobacteriosis [16], we furthermore enumerated apoptotic colonic epithelial cells following in situ immunohistochemistry upon necropsy. In line with the obtained histopathological results, a multifold higher number of caspase3 + apoptotic cells were found in colonic epithelia at day 28 following C. coli versus E. coli challenge of IL-10 -/-mice (p < 0.05-0.01; Figure 4B; Supplementary Figure S2B). These increases were, however, far less pronounced in IL-10 -/-mice lacking TLR4 as compared to IL-10 -/-controls (p < 0.001; Figure 4B; Supplementary Figure S2B). Hence, TLR4 signaling was involved in mediating C. coliinduced colonic histopathological and epithelial cell apoptosis upon peroral infection of secondary abiotic IL-10 -/-mice.  Since apoptosis is a hallmark of C. jejuni-induced murine campylobacteriosis [16], we furthermore enumerated apoptotic colonic epithelial cells following in situ immunohistochemistry upon necropsy. In line with the obtained histopathological results, a multifold higher number of caspase3 + apoptotic cells were found in colonic epithelia at day 28 following C. coli versus E. coli challenge of IL-10 −/− mice (p < 0.05-0.01; Figure 4B; Supplementary Figure S2B). These increases were, however, far less pronounced in IL-10 −/− mice lacking TLR4 as compared to IL-10 −/− controls (p < 0.001; Figure 4B; Supplementary Figure S2B). Hence, TLR4 signaling was involved in mediating C. coli-induced colonic histopathological and epithelial cell apoptosis upon peroral infection of secondary abiotic IL-10 −/− mice.

TLR4-Dependent Intestinal Immune Cell Responses Induced by C. coli Infection
We next surveyed TLR4-dependent innate and adaptive immune responses in the large intestines upon the bacterial challenge of secondary abiotic IL-10 −/− mice by applying quantitative in situ immunohistochemistry. On day 28, the numbers of innate immune cell populations, such as F4/80 + Microorganisms 2020, 8, 1882 9 of 17 cells macrophages and monocytes, as well as adaptive immune cell subsets, including CD3 + T lymphocytes, FOXP3 + regulatory T cells, and B220 + B lymphocytes, were higher in the colonic mucosa and lamina propria of C. coli as compared to E. coli challenged mice of either genotype (p < 0.01-0.001; Figure 5; Supplementary Figure S2C-F). Upon either bacterial application, colonic numbers of macrophages and monocytes and of T lymphocytes were lower in TLR4 −/− IL-10 −/− as compared to respective IL-10 −/− counterparts (p < 0.01-0.001; Figure 5A,B; Supplementary Figure  S2C,D), which also held true for large intestinal B lymphocyte counts at day 28 following C. coli infection (p < 0.05; Figure 5D; Supplementary Figure S2F). Hence, C. coli induced colonic immune cell responses in a TLR4-dependent manner. independent experiments.

TLR4-Dependent Intestinal Immune Cell Responses Induced by C. coli Infection
We next surveyed TLR4-dependent innate and adaptive immune responses in the large intestines upon the bacterial challenge of secondary abiotic IL-10 -/-mice by applying quantitative in situ immunohistochemistry. On day 28, the numbers of innate immune cell populations, such as F4/80 + cells macrophages and monocytes, as well as adaptive immune cell subsets, including CD3 + T lymphocytes, FOXP3 + regulatory T cells, and B220 + B lymphocytes, were higher in the colonic mucosa and lamina propria of C. coli as compared to E. coli challenged mice of either genotype (p < 0.01-0.001; Figure 5; Supplementary Figure S2C-F). Upon either bacterial application, colonic numbers of macrophages and monocytes and of T lymphocytes were lower in TLR4 -/-IL-10 -/-as compared to respective IL-10 -/-counterparts (p < 0.01-0.001; Figure 5A,B; Supplementary Figure S2C,D), which also held true for large intestinal B lymphocyte counts at day 28 following C. coli infection (p < 0.05; Figure  5D; Supplementary Figure S2F). Hence, C. coli induced colonic immune cell responses in a TLR4dependent manner.
(CD3 + ), (C) regulatory T cells (FOXP3 + ), and (D) B lymphocytes (B220 + ) were microscopically surveyed from six high power fields (HPF, 400× magnification) per animal in immunohistochemically-stained colonic paraffin sections. Medians (black bars), levels of significance (p-values) as determined by (A,B) the one-way ANOVA and Tukey's post-correction and (C,D) the Kruskal-Wallis test and Dunn's post-correction, and the numbers of analyzed animals (in parentheses) are indicated. Data were pooled from three independent experiments.

Discussion
C. coli constitutes the second most prevalent causative agent of human Campylobacter infections after C. jejuni and was recently reported to account for approximately 10% of the confirmed campylobacteriosis cases in the European Union [8]. Yet, little is known about the crosstalk between C. coli and the immune system of the vertebrate host, and the lack of molecular data concerning bacterial factors triggering campylobacteriosis underlines the urgent need for a convenient murine model of C. coli infection. In the present study, we, therefore, analyzed the pathogenic potential of C. coli in the secondary abiotic IL-10 -/-murine infection model, which was successfully established and further optimized as one of the valid clinical murine models for acute C. jejuni infection, mimicking key features symptoms of severe campylobacteriosis in humans. This acute infection and inflammation model has unraveled the major role of LOS in C. jejuni-induced immunopathologies and is nowadays used for the evaluation and validation of novel therapeutic intervention strategies against campylobacteriosis, as shown in recent studies [46][47][48][49]. Whereas several in vitro investigations have proven that C. jejuni LOS activates TLR4 in different avian, murine, and human cell lines [19,50,51], only one study to the best of our knowledge revealed that a C. coli chicken isolate could activate TLR4 in vitro [20].
In the present study, we provided evidence for TLR4-dependent C. coli vertebrate-host interactions in vivo and determined the pathogenic potential of C. coli by comparison of data obtained from colonization of mice with an apathogenic commensal E. coli isolate. Results indicated that both C. coli and E. coli stably colonized the small and large intestines of IL-10 -/-mice with high loads up to four weeks post-challenge in a TLR4-independent manner. Stable bacterial establishment within the intestinal lumen was accompanied by an abundance of fecal blood in mice of either genotype that had been challenged with C. coli as opposed to E. coli, which was most prominent within the first week. The finding that C. coli but not E. coli induced signs of intestinal inflammation proved the pathogenic potential of C. coli, which was mediated by the barrier-breaking properties, such as

Discussion
C. coli constitutes the second most prevalent causative agent of human Campylobacter infections after C. jejuni and was recently reported to account for approximately 10% of the confirmed campylobacteriosis cases in the European Union [8]. Yet, little is known about the crosstalk between C. coli and the immune system of the vertebrate host, and the lack of molecular data concerning bacterial factors triggering campylobacteriosis underlines the urgent need for a convenient murine model of C. coli infection. In the present study, we, therefore, analyzed the pathogenic potential of C. coli in the secondary abiotic IL-10 −/− murine infection model, which was successfully established and further optimized as one of the valid clinical murine models for acute C. jejuni infection, mimicking key features symptoms of severe campylobacteriosis in humans. This acute infection and inflammation model has unraveled the major role of LOS in C. jejuni-induced immunopathologies and is nowadays used for the evaluation and validation of novel therapeutic intervention strategies against campylobacteriosis, as shown in recent studies [46][47][48][49]. Whereas several in vitro investigations have proven that C. jejuni LOS activates TLR4 in different avian, murine, and human cell lines [19,50,51], only one study to the best of our knowledge revealed that a C. coli chicken isolate could activate TLR4 in vitro [20].
In the present study, we provided evidence for TLR4-dependent C. coli vertebrate-host interactions in vivo and determined the pathogenic potential of C. coli by comparison of data obtained from colonization of mice with an apathogenic commensal E. coli isolate. Results indicated that both C. coli and E. coli stably colonized the small and large intestines of IL-10 −/− mice with high loads up to four weeks post-challenge in a TLR4-independent manner. Stable bacterial establishment within the intestinal lumen was accompanied by an abundance of fecal blood in mice of either genotype that had been challenged with C. coli as opposed to E. coli, which was most prominent within the first week. The finding that C. coli but not E. coli induced signs of intestinal inflammation proved the pathogenic potential of C. coli, which was mediated by the barrier-breaking properties, such as motility, adhesion, and invasion, as demonstrated earlier for C. jejuni in the same murine model of infection [52,53].
In TLR4-deficient IL-10 −/− mice, however, the maximum fecal blood positivity rates occurred two days later than in infected IL-10 −/− counterparts (day 5 versus day 7 p.i.), whereas in the later stage, i.e., from week two to four of C. coli infection, fecal blood could less frequently be detected in the former versus the latter. The role of TLR4 in C. coli-induced disease outcome is supported by less distinct histopathological changes in the large intestinal tract and by lower numbers of apoptotic epithelial cells in the colon of TLR4 −/− IL-10 −/− mice versus IL-10 −/− mice at day 28 post-C. coli challenge. Given that TLR4 can potently induce cell apoptosis [16,17,54], our finding provided evidence that C. coli LOS was involved in triggering the inflammatory scenario in the intestines upon infection.
The role of LOS as a virulence factor was further confirmed by the findings that C. coli induced both innate and adaptive immune cell responses in a TLR4-dependent manner. Compared to the E. coli controls, C. coli-infected IL-10 −/− mice displayed elevated levels of innate immune cells, such as macrophages and monocytes, which also held true for adaptive immune cells populations, including T and B lymphocytes, which were lower in the TLR4 −/− IL-10 −/− cohort at day 28 p.i. These results are supported by our previous studies showing TLR4-dependent immune cell responses upon infection with C. jejuni 81-176 in the same murine model of infection [17], as well as in secondary abiotic wildtype mice [16,17]. This confirms that C. coli induces campylobacteriosis similar to C. jejuni in a TLR4-dependent manner and that LOS plays a major role in C. coli immunopathogenesis.
Furthermore, in line with results derived from C. jejuni-infected IL-10 −/− mice suffering from acute enterocolitis within the first week of infection [17], pro-inflammatory mediator secretion was less pronounced in intestinal ex vivo biopsies derived from TLR4-deficient IL-10 −/− mice as compared to IL-10 −/− controls at day 28 post-C. coli infection. Remarkably, TLR4-dependent immune responses were not restricted to the intestinal tract but could also be observed in extra-intestinal compartments, as indicated by less hepatic IFN-γ secretion in C. coli-infected TLR4-deficient IL-10 −/− mice as compared to IL-10 −/− mice. Interestingly, no viable C. coli that might have translocated from the lumen of the inflamed intestines could be isolated from extra-intestinal tissue sites including the blood stream at all. We could not exclude, however, that translocated bacteria would most likely have been cleared by the immune system as late as day 28 p.i. One needs to take further into consideration that soluble C. coli constituents, including LOS, might exert potent pro-inflammatory effects at any extra-intestinal, including systemic tissue sites [55].
The here presented results are well in line with our very recent study, where we addressed the role of TLR4 in C. coli-infected mice harboring a complex human gut microbiota [36]. The pathogen could establish within the gastrointestinal tract of both IL-10 −/− and TLR4-deficient IL-10 −/− mice until day 21 p.i., but induced less pronounced immunopathological sequelae in the latter versus the former upon oral challenge. This study provided the first evidence for the pivotal role of TLR4-dependent C. coli-host responses in concert with the human gut microbiota [36]. Our actual study involving microbiota-depleted mice, however, further underlined the important immunopathological role of TLR4 in mediating pathogenic LOS responses in the absence of any other TLR4 ligands derived from the commensal gut microbiota.
The auxiliary finding that TLR4-deficient IL-10 −/− mice harbored approximately two orders of magnitude higher C. coli bacterial numbers in the stomach as compared to their IL-10 −/− counterparts provides evidence that TLR4 triggers immune defense mechanisms in the stomach to protect from pathogens and to maintain intestinal homeostasis [56]. The basolateral expression has been shown for several TLRs, resulting in less frequent activation by stimuli derived from luminal bacteria [57]. One study revealed that TLR4 is mainly localized at the basolateral surface of the colonic epithelium in a healthy murine intestine, whereas another study showed that TLR4 is expressed both on the apical and basolateral surface of the gastric epithelium in mice infected with Helicobacter pylori, which is closely related to Campylobacter [58,59]. In this scenario, the basolateral expression of TLR4 could possibly allow C. coli to colonize at higher levels. However, it cannot be excluded that apical TLR4 expression is induced upon an increase in LPS or LOS from Gram-negative pathogens.
In summary, our results demonstrate that TLR4 signaling is involved in mediating the inflammatory responses upon C. coli infection in the vertebrate host. However, we cannot exclude that also other TLRs are participating in the recognition of this pathogen, such as TLR2, TLR5, and TLR9, recognizing bacterial lipoprotein, flagella, and CpG-DNA, respectively [60]. The C. coli chicken isolate investigated by Zoete et al. could activate TLR2 and additionally TLR21, which is the receptor recognizing CpG-DNA in poultry but not TLR5 and TLR9 [20]. Moreover, our group showed earlier that besides TLR4, also TLR2 and TLR9 play a role in mediating C. jejuni-induced immunopathology [16,17]. In addition, secondary abiotic IL-10 −/− mice deficient in the innate receptor nucleotide-oligonucleotide-domain 2 (Nod2) developed less severe enterocolitis upon peroral C. jejuni infection [35]. Although C. jejuni is able to evade recognition by TLR5 [61], we do not know whether this also may apply to C. coli. Furthermore, in order to unravel potential strain-specific host interactions, it is necessary to include C. coli strains isolated from different sources. One study, for instance, assessed the colonization abilities of several C. coli strains isolated from three different sources, namely symptomatic carriers, asymptomatic carriers, and chicken carcasses, and revealed marked functional differences between the strains, which were most likely source-dependent [62]. It would, therefore, be of high interest to unravel whether these differences in colonization capacities can be attributed to LOS diversity, affecting host TLR4 recognition.

Conclusions
We further conclude that the here applied infection model of LOS-sensitized secondary abiotic IL-10 −/− mice, which had initially been established to unravel C. jejuni-host interactions, mimicking severe human campylobacteriosis, can also be used to study the molecular mechanism underlying C. coli-induced immunopathogenesis and, hence, paves the way to develop therapeutic and prophylactic options to combat campylobacteriosis as well as post-infectious sequelae.
Author Contributions: S.K.: Performed experiments, analyzed data, co-wrote paper; C.G.: Performed experiments, analyzed data, co-edited paper; D.W.: Performed experiments, analyzed data, co-edited paper; S.M.: Performed experiments, analyzed data, co-edited paper; S.B.: Provided advice in experimental design, critically discussed results, co-edited paper; M.M.H.: Designed and performed experiments, analyzed data, co-wrote paper. All authors have read and agreed to the published version of the manuscript.
Funding: This work was supported from the German Federal Ministries of Education and Research (BMBF) in the frame of the zoonoses research consortium PAC-Campylobacter to SM, SB and MMH (IP7/01KI1725D) as part of the Research Network Zoonotic Infectious Diseases and from the German Federal Ministries of Economy and Energy to SB and MMH (ZIM; ZF4117904 AJ8). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.