The Role of IL-23, IL-22, and IL-18 in Campylobacter Jejuni Infection of Conventional Infant Mice.

We have recently shown that, within 1 week following peroral Campylobacter jejuni infection, conventional infant mice develop self-limiting enteritis. We here investigated the role of IL-23, IL-22, and IL-18 during C. jejuni strain 81-176 infection of infant mice. The pathogen efficiently colonized the intestines of IL-18(-/-) mice only, but did not translocate to extra-intestinal compartments. At day 13 postinfection (p.i.), IL-22(-/-) mice displayed lower colonic epithelial apoptotic cell numbers as compared to wildtype mice, whereas, conversely, colonic proliferating cells increased in infected IL-22(-/-) and IL-18(-/-) mice. At day 6 p.i., increases in neutrophils, T and B lymphocytes were less pronounced in gene-deficient mice, whereas regulatory T cell numbers were lower in IL-23p19(-/-) and IL-22(-/-) as compared to wildtype mice, which was accompanied by increased colonic IL-10 levels in the latter. Until then, colonic pro-inflammatory cytokines including TNF, IFN-γ, IL-6, and MCP-1 increased in IL-23p19(-/-) mice, whereas IL-18(-/-) mice exhibited decreased cytokine levels and lower colonic numbers of T and B cell as well as of neutrophils, macrophages, and monocytes as compared to wildtype controls. In conclusion, IL-23, IL-22, and IL-18 are differentially involved in mediating C. jejuni-induced immunopathology of conventional infant mice.

Introduction

Human gastroenteritis cases caused by the zoonotic gram-negative bacteria Campylobacter jejuni are emerging worldwide [1, 2]. As part of the commensal gut microbiota in a plethora of wild and domestic animal species, transmission to humans occurs from livestock animals via consumption of contaminated meat products or water for instance [3, 4]. Infected patients present with symptoms of considerable variability ranging from mild, non-inflammatory, watery diarrhea to severe, inflammatory, bloody diarrhea associated with abdominal pain that might last for a few weeks, but mostly resolve spontaneously. In rare cases, however, infected patients develop post-infectious sequelae including reactive arthritis and peripheral neuropathies such as Guillain–Barré and Miller–Fisher syndromes later on [5, 6]. Histological changes such as apoptosis, crypt abscesses, ulcerations, and pronounced influx of pro-inflammatory immune cell populations including lymphocytes and neutrophils into the intestinal mucosa and lamina propria can be observed in intestinal tissues derived from infected patients [7, 8]. Despite the global importance of human C. jejuni infection, our understanding of the molecular mechanisms underlying campylobacteriosis is limited due to the scarcity of appropriate in vivo models. Whereas newborn piglets, weanling ferrets, chicken, gnotobiotic canine pups, and primates have been more or less successfully used for studying C. jejuni–host interactions [6], our group has recently shown that, upon peroral C. jejuni infection immediately after weaning, 3-week-old conventional infant mice develop acute enteritis within 1 week that resolves thereafter [9-11]. C. jejuni-induced immune responses were characterized by increased colonic abundances of effector cells and innate as well as adaptive immune cell subsets that were accompanied by increased colonic secretion of pro-inflammatory mediators including TNF, IFN-γ, IL-6, MCP-1, and nitric oxide [9-13]. Interestingly, as compared to adult mice, infant mice harbored higher intestinal loads of commensal enterobacteria such as Escherichia coli in their intestines facilitating C. jejuni colonization [9-11]. Overall, the infant mouse model displayed key features of human campylobacteriosis and can be regarded as well suitable in order to investigate Campylobacter–host interactions in more detail [6, 14]. Very recently, we were able to show that IL-23p19, IL-22, and IL-18 were upregulated in the large intestines of not only C. jejuni-infected conventional infant [13] but also of gnotobiotic (i.e., secondary abiotic) mice generated by broad-spectrum antibiotic treatment [15]. Moreover, IL-22 and IL-17 were upregulated in C. jejuni-infected IL-10-deficient mice [16]. IL-22 is a cytokine of the IL-10 family and well known not only for its antimicrobial and tissue-protective but also pro-inflammatory properties [17, 18]. Particularly in the intestinal tract, IL-22 exerts its dichotomous actions in a tissue-dependent fashion. Whereas in the large intestinal tract, IL-22 has been shown to act as an anti-inflammatory mediator [18], we have recently shown its pro-inflammatory properties within the small intestines. In acute Toxoplasma gondii-induced ileitis, IL-23p19-dependent IL-22 induction resulted in small intestinal necrosis [19-21]. In addition, IL-22 induced the expression of IL-18 mRNA in intestinal epithelial cells following T. gondii infection, whereas, conversely, IL-18 amplified IL-22 production from innate lymphoid cells (ILCs) and T helper (Th) -1 cell mediated intestinal inflammation [21].

In the present study, we aimed to shed further light onto the impact of cytokines belonging to the IL-23/IL-22/IL-18 axis during C. jejuni infection. To address this, we infected 3-week-old conventional infant IL-23p19–/–, IL-22–/–, IL-18–/–, and corresponding wildtype (WT) mice perorally with C. jejuni strain 81-176 immediately after weaning and investigated 1) the gastrointestinal colonization and translocation properties of C. jejuni as well as of commensal E. coli facilitating pathogenic infection, 2) the clinical outcome of infection, 3) the histopathological changes in the colon including apoptosis, 4) the abundances of distinct immune cell populations in the colonic mucosa and lamina propria, and, furthermore, 5) the large intestinal expression of pro- and anti-inflammatory cytokines.

Methods

Mice and C. jejuni infection

Female IL-23p19–/–, IL-22–/–, and IL-18–/– mice (all in C57BL/6j background) as well as age- and sex-matched C57BL/6j wildtype (WT) control mice were bred and maintained within the same specific pathogen-free (SPF) unit in the Forschungseinrichtungen für Experimentelle Medizin (FEM, Charité – University Medicine Berlin). In order to confirm absence of IL-23p19, IL-22, or IL-18 gene expression, genomic DNA was isolated and disruption of either gene was confirmed by polymerase chain reaction (PCR) [19]. Immediately after weaning, 3-week-old conventional infant mice were perorally infected with 109 colony forming units (CFU) of viable C. jejuni strain 81-176 in a volume of 0.3 ml phosphate buffered saline (PBS) on two consecutive days (day 0 and day 1) by gavage as described earlier [22].

Clinical score

To assess clinical signs of C. jejuni-induced infection on a daily basis, a standardized cumulative clinical score (maximum 12 points), addressing the occurrence of blood in feces (0: no blood; 2: microscopic detection of blood by the Guaiac method using Haemoccult, Beckman Coulter/PCD, Krefeld, Germany; 4: 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 compromised locomotion, prefinal aspect) was used [23, 24].

Sampling procedures

Mice were sacrificed at day 6 or day 13 p.i. by isoflurane treatment (Abbott, Greifswald, Germany). Cardiac blood and tissue samples from the gastrointestinal tract (i.e., stomach, duodenum, terminal ileum, and colon), mesenteric lymphnodes (MLN), spleen, liver, and kidney were asserved under sterile conditions. Colonic ex vivo biopsies were collected in parallel for immunohistochemical, microbiological, and immunological analyses. Immunohistopathological changes were assessed in colonic samples that were immediately fixed in 5% formalin and embedded in paraffin. Sections (5 μm) were stained with hematoxylin and eosin (H&E) or respective antibodies for in situ immunohistochemistry as described earlier [13, 25].

Histopathological grading of large intestinal lesions

Histopathological changes were quantitatively assessed in H&E-stained large intestinal paraffin sections, applying a histopathological scoring system by two independent double-blinded investigators as described previously [26]. In brief:

Colonic histopathology (max. 4 points; according to ref. [27]): 0: no inflammation; 1: single isolated cell infiltrates within the mucosa; no epithelial hyperplasia; 2: mild scattered to diffuse cell infiltrates within the mucosa and submucosa; mild epithelial hyperplasia; starting loss of goblet cells; 3: cell infiltrates within mucosa, submucosa, and sometimes transmural; epithelial hyperplasia; loss of goblet cells; 4: cell infiltrates within mucosa, submucosa, and transmural; severe inflammation; loss of goblet cells, loss of crypts; ulcerations; severe epithelial hyperplasia.

Immunohistochemistry

In situ immunohistochemical analysis of colonic paraffin sections was performed as described previously [26]. Primary antibodies against cleaved caspase-3 (Asp175, Cell Signaling, Beverly, MA, USA, 1:200), Ki67 (TEC3, Dako, Denmark, 1:100), myeloperoxidase (MPO-7, no. A0398, Dako, 1:500), F4/80 (no. 14-4801, clone BM8, eBioscience, San Diego, CA, USA, 1:50), CD3 (no. N1580, Dako, 1:10), FOXP3 (FJK-16s, eBioscience, 1:100), and B220 (eBioscience, 1:200) were used. For each animal, the average number of positively stained cells within at least six high power fields (HPF, 0.287 mm[2], 400× magnification) were determined microscopically by a double-blinded investigator.

Quantitative analysis of bacterial colonization and translocation

Viable C. jejuni were detected in feces over time p.i. or at time of necropsy (day 6 or day 13 p.i.) in luminal samples taken from stomach, duodenum, terminal ileum, and colon, by culture of serial dilutions in PBS on Karmali- and Columbia-Agar supplemented with 5% sheep blood (Oxoid) for 2 days at 37 °C under microaerobic conditions using CampyGen gas packs (Oxoid). To quantify bacterial translocation, ex vivo biopsies derived from MLN, spleen, liver, and kidney were homogenized in 1 ml sterile PBS, whereas cardiac blood (≈100 μL) was directly streaked onto Karmali-Agar and Columbia-Agar supplemented with 5% sheep blood and cultivated accordingly. Numbers of viable E. coli were quantitatively assessed by culture as described earlier [28]. The respective weights of fecal or tissue samples were determined by the difference of the sample weights before and after asservation. The detection limit of viable C. jejuni by direct plating was 100 CFU per gram of sample.

Cytokine detection in supernatants of colonic ex vivo biopsies

Colonic ex vivo biopsies were cut longitudinally and washed in PBS. Strips of approximately 1 cm[2] intestinal tissue were placed in 24-well flat-bottom culture plates (Nunc, Wiesbaden, Germany) containing 500 μL serum-free RPMI 1640 medium (Gibco, Life Technologies, Paisley, UK) supplemented with penicillin (100 U/ml) and streptomycin (100 μg/ml; PAA Laboratories). After 18 h at 37 °C, culture supernatants or serum samples were tested for TNF, IFN-γ, IL-6, MCP-1, and IL-10 by the Mouse Inflammation Cytometric Bead Assay (CBA; BD Biosciences) on a BD FACSCanto II flow cytometer (BD Biosciences).

Statistical analysis

Medians and levels of significance were determined using Mann–Whitney U test (GraphPad Prism v6.05, La Jolla, CA, USA) as indicated. Two-sided probability (p) values of <0.05 were considered significant.

Ethics statement

All 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, registration number G0135/10). Animal welfare was monitored twice daily by assessment of clinical conditions.

Results

Gastrointestinal colonization and translocation of C. jejuni in infant mice lacking IL-23p19, IL-22, or IL-18 upon peroral infection

In the present study, we aimed to dissect the impact of the cytokines IL-23p19, IL-22, and IL-18 in murine campylobacteriosis and applied the infant mouse model of C. jejuni infection. Immediately after weaning, 3-week-old IL-23p19–/–, IL-22–/–, IL-18–/–, and corresponding WT mice were perorally infected with 109 CFU C. jejuni strain 81-176 on two consecutive days (namely, day 0 and day 1) by gavage. We then performed a kinetic survey of colonization densities in individual fecal samples. At day 2 p.i. (i.e., as early as 24 h after the latest infection), 72.7%, 81.0%, 78.9%, and 78.3% of infected infant WT, IL-23p19–/–, IL-22–/–, and IL-18–/– mice, respectively, harboured C. jejuni with a median load of approximately 104 CFU per gram fecal samples, whereas mice of the former three genotypes successively expelled the pathogen from their intestinal tract ( In fact, fecal C. jejuni could be isolated in two thirds of infected IL-18–/– mice, but only in 18.2%, 23.8%, and 27.8% of WT, IL-23p19–/–, and IL-22–/– mice, respectively, at day 6 p.i. ( Interestingly, median pathogenic loads in fecal samples taken from IL-18–/– mice were even one order of magnitude higher at day 6 as compared to days 2 or 3 p.i. (p < 0.05; ). Mice were then sacrificed at two different time points, namely, day 6 and day 13 p.i., and the C. jejuni infection efficiencies were determined alongside the gastrointestinal tract. At day 6 p.i., IL-18–/– mice exhibited higher pathogenic loads in their stomach (p < 0.05; ) and colonic lumen (p < 0.05–0.001; ) as compared to mice of the remaining genotypes. Also the small intestinal tract of IL-18–/– mice could be colonized by C. jejuni, with higher loads in the duodenum as compared to IL-23p19–/– and IL-22–/– mice (p < 0.01; ) and higher ileal pathogenic numbers versus WT and IL-22–/– animals (p < 0.05; ). Until day 13 p.i., however, conventional infant mice had completely expelled the pathogen from their gastrointestinal tract, except for single IL-18–/– animals ( Hence, the gastrointestinal tract of infant IL-18–/– contrary to WT, IL-23p19–/–, and IL-22–/– mice could be efficiently colonized by C. jejuni following peroral infection. We further investigated whether C. jejuni was able to translocate from the intestinal to extra-intestinal tissue sites. Whereas viable C. jejuni could be isolated from MLN in single cases only (i.e., 5.9% of WT mice at day 6 p.i. and 20% of IL-18–/– mice at day 13 p.i.; ), homogenates of spleen, liver, and kidney as well as blood samples were all C. jejuni negative as determined by culture (

Commensal gastrointestinal E. coli loads and bacterial translocation in infant mice lacking IL-23p19, IL-22, or IL-18 upon peroral C. jejuni infection

We next addressed the question whether observed differences in pathogenic colonization efficiencies might be due to different gastrointestinal loads of commensal E. coli known to facilitate murine C. jejuni infection [9]. Surprisingly, before infection, naive IL-22–/–, but not IL-18–/–, mice exhibited the highest fecal E. coli densities with median loads of more than 109 CFU per g feces (p < 0.01–0.001 vs. remaining groups of mice; ). Furthermore, naive IL-23p19–/–, but not IL-18–/–, infant mice exhibited higher fecal E. coli numbers as compared to WT controls (p < 0.01; ). We further assessed E. coli colonization properties in the gastrointestinal tract at days of necropsy. At days 6 and 13 p.i., IL-22–/– mice exhibited higher E. coli loads in the duodenum, ileum, and colon (p < 0.001 and p < 0.01, respectively), but not stomach as compared to respective WT mice ( In IL-23p19–/– mice, E. coli numbers were higher in the duodenum and colon at day 6 p.i. (p < 0.001; and D) and at either time point in the ileum as compared to WT controls (p < 0.05–0.01; ). In MLN, E. coli could be isolated in 12.5% and 25.0% of IL-23p19–/– and 61.5% and 20.0% of IL-22–/– mice at day 6 and day 13 p.i., respectively ( At day 6 p.i., E. coli loads in MLN derived from IL-22–/– mice were higher as compared to the remaining genotypes of mice (p < 0.05–0.001; ). Hence, IL-23p19–/– and IL-22–/–, but not IL-18–/–, infant mice exerted the highest small and large intestinal E. coli loads, even though the latter, but not the former two, were stably infected by C. jejuni. As for C. jejuni, no viable commensal E. coli could be isolated from extra-intestinal tissue sites such as spleen, liver, kidney, and cardiac blood (

Macroscopic and microscopic aspects of campylobacteriosis in C. jejuni-infected infant mice lacking IL-23p19, IL-22, or IL-18

During the course of C. jejuni infection, we surveyed clinical conditions of mice, applying a standardized scoring system. At days 6 and 13 p.i., mice of either genotype (except for IL-22–/– mice during late course of infection) displayed higher clinical scores as compared to day 0 (i.e., immediately before infection) (, but were suffering from only minor C. jejuni-induced clinical sequelae as ruffling fur and/or microscopic detection of occult blood in fecal samples. Infant IL-23p19–/– mice, however, displayed slightly higher clinical scores as compared to WT and IL-18–/– mice at day 6 p.i. (p < 0.05 and p < 0.01, respectively; ). In addition to macroscopic aspects of infected mice, we assessed C. jejuni-induced microscopic (i.e., histopathological) changes of H&E-stained colonic paraffin section, applying an established histopathological scoring system. Irrespective of the genotype, histopathological scores were higher in infected as compared to naive mice at either time point and indicative of comparable moderate colonic inflammation. In IL-18–/– mice, however, histopathological sequelae were even more severe at day 13 as compared to day 6 p.i. (p < 0.01; ).

Since apoptosis is a commonly used diagnostic marker for histopathological grading of intestinal inflammation [22], we further stained colonic paraffin sections against caspase-3. Six days following C. jejuni infection, WT, IL-23p19–/–, and IL-18–/– mice displayed higher apoptotic cell numbers in the colonic epithelial layer as compared to naive controls (p < 0.05–0.001; ). In IL-22–/– mice, however, apoptotic cell numbers tended to increase from day 0 until day 6 p.i. (n.s.), but were lower at days 13 p.i. than 7 days before (p < 0.05; ). At the later time point of necropsy, colonic apoptotic cell numbers were lower in IL-22–/– as compared to WT and IL-23p19–/– mice (p < 0.05; ). In IL-18–/– mice, a trend towards lower apoptotic cell numbers in the colonic epithelium could be observed at both days 6 and 13 p.i. as compared to respective WT mice, but did not reach statistical significance due to high standard deviations within the respective groups (n.s.; ). Given that Ki67 is a well-known nuclear protein necessary for cellular proliferation [29], we stained colonic paraffin sections against Ki67 to assess proliferative measures of the colonic epithelium counteracting apoptosis following C. jejuni infection. Upon C. jejuni infection and conversely to apoptotic cell numbers, Ki67-positive cell numbers increased in the colonic epithelial layer of infant IL-18–/– mice (p < 0.05–0.01, ), and were higher than in WT controls at days 6 and 13 p.i. (p < 0.05 and p < 0.01, respectively; ). An increase of colonic proliferating cells was also determined in infant IL-22–/– mice at day 13 p.i. versus WT animals (p < 0.01; ). Hence, despite stable and highest C. jejuni colonization densities in IL-18–/– mice, overall, rather subtle macroscopic and moderate microscopic C. jejuni-induced sequelae could be observed which did not significantly differ between genotypes of infected mice.

Immune cell responses in C. jejuni-infected infant mice lacking IL-23p19, IL-22, or IL-18

Recruitment of pro-inflammatory immune cells to sites of inflammation is a key event in intestinal pathogenic infection including campylobacteriosis [22]. We therefore quantitatively assessed effector cell as well as innate and adaptive immune cell numbers within the large intestinal mucosa and lamina propria of infected mice by in situ immunohistochemical staining of colonic paraffin sections. Six days following C. jejuni infection, colonic MPO7-positive neutrophilic granulocyte numbers increased in WT and IL-23p19–/– (p < 0.001 and p < 0.01, respectively), but neither in IL-22–/– nor IL-18–/– mice, and were significantly lower in respective gene-deficient animals as compared to WT controls (p < 0.01-0.001; ). Until day 13 p.i., neutrophil numbers decreased back to naive levels in IL-23p19–/– mice (

We next stained another subset of effector cells, namely, F4/80-positive macrophages and monocytes. Colonic F4/80-positive cells increased 6 days upon C. jejuni infection in WT, IL-23p19–/–, and IL-22–/– (p < 0.01–0.001), but not IL-18–/–, mice ( Seven days later, however, F4/80-positive cell numbers were higher in colons of infant mice, irrespective of their genotype, as compared to respective naive controls (p < 0.01–0.001; ). Moreover, IL-18–/– mice exhibited lower numbers of colonic macrophages and monocytes at both day 6 and day 13 p.i. as compared to WT mice (p < 0.05 and 0.01, respectively; ).

We next investigated C. jejuni-induced changes in intestinal adaptive immune cell numbers, namely, T and B lymphocytes as well as Tregs, by staining colonic ex vivo biopsies with antibodies directed against CD3, B220, and FOXP3, respectively. Whereas colonic T lymphocytes increased in WT, IL-23p19–/–, and IL-18–/– until day 6 p.i., elevated CD3+ cell counts could be observed in IL-22–/– mice later on at day 13 p.i. (p < 0.001; ). Interestingly, already in the naive state, IL-22–/– infant mice displayed higher T cell numbers in their large intestines than WT controls (p < 0.05; ). At day 6 p.i., however, mice of either genotype exhibited lower colonic T as well as B lymphocyte numbers as compared to WT animals (p < 0.01–0.001; and C), whereas, at day 13 p.i., B cells were lower in the colonic mucosa and lamina propria of IL-18–/– than WT controls (p < 0.05; ). Interestingly, FOXP3+ Treg numbers increased successively in the course of C. jejuni infection in WT mice (p < 0.05 vs. naive mice; ), whereas this increase was rather delayed in IL-22–/– and IL-18–/– mice as indicated by higher colonic FOXP3+ cell numbers at day 13, but not day 6 p.i. (p < 0.05 and p < 0.01, respectively vs. naive controls; ). Furthermore, at day 6 p.i., Treg counts were lower in the large intestines of IL-23p19–/– and IL-22–/– as compared to WT mice (p < 0.05; ), whereas, at the same time point, colonic B lymphocytes were lower in IL-18–/– versus WT mice (p < 0.05; ).

In the following, we measured pro- and anti-inflammatory cytokine secretion in colonic ex vivo biopsies taken from C. jejuni-infected infant mice. Until day 6 p.i., TNF, IFN-γ, IL-6, and MCP-1 concentrations increased in large intestines of IL-23p19–/– mice only (p < 0.05–0.01; ), whereas respective pro-inflammatory cytokines were significantly lower in colons of IL-18–/– as compared to WT mice at day 6 p.i. (p < 0.05–0.01; ). Notably, IL-23p19–/– and IL-22–/– mice exhibited even higher colonic IFN-γ concentrations as compared to respective WT controls at both day 6 and day 13 p.i. (p < 0.05–0.01; ). Furthermore, only in IL-22–/– mice, colonic secretion of the anti-inflammatory cytokine IL-10 was elevated at either time point following C. jejuni infection and higher at day 6 p.i. when compared to WT mice, whereas basal IL-10 expression, however, was lower in naive IL-22–/– vs. WT animals (all p < 0.05; ). Taken together, during the early stage of C. jejuni infection of infant mice, pro-inflammatory cytokines such as TNF, IFN-γ, IL-6, and MCP-1 were increased in large intestines of IL-23p19–/– mice only, whereas IL-18–/– mice, however, exhibited lower colonic pro-inflammatory cytokine levels at day 6 p.i., which is well in line with observed less distinct increases in colonic macrophages and monocytes.

Discussion

In the present study, we investigated the role of mediators belonging to the IL-23/IL-22/IL-18 axis during C. jejuni strain 81-176 infection of conventional infant mice that were gene-deficient for the respective regulatory and inflammatory cytokines. Infant mice displayed rather subtle clinical sequelae of infection, whereas moderate histopathological changes of the colonic mucosa and lamina propria could be observed, but irrespective of the murine genotype. Overall, differences in distinct C. jejuni-induced pro-inflammatory immune responses in infant mice were inconsistent between genotypes. For instance, in IL-22–/– mice, lower numbers of colonic epithelial apoptotic cells, but increased levels of the anti-inflammatory cytokine IL-10, could be observed upon C. jejuni infection as compared to WT controls. Infected IL-18–/– mice exhibited a trend towards less distinct large intestinal apoptosis, but significantly lower colonic numbers of innate immune cells such as macrophages and monocytes that were accompanied by lower colonic pro-inflammatory cytokines including TNF, IFN-γ, IL-6, and MCP-1 as compared to WT controls. Furthermore, higher numbers of regenerating/proliferating cells within the colonic epithelium counteracting potential C. jejuni-induced epithelial damage could be observed in IL-18–/– than in WT mice. Immune cell populations such as T and B cells, and also neutrophilic granulocytes, were less abundant in the colonic mucosa and lamina propria of IL-23p19–/–, IL-22–/–, and IL-18–/– as compared to WT mice, whereas regulatory T cells were decreased in IL-23p19–/– and IL-22–/– mice.

Recently, IL-23 was highlighted as a master regulator of mucosal immune responses upon intestinal infection and inflammation [30]. Furthermore, IL-22 was shown to exert effective antimicrobial defence mechanisms against C. jejuni including enhanced β-defensin production [31] and was upregulated following C. jejuni infection of human intestinal ex vivo biopsies [32]. These ex vivo results were supported by increased IL-22 concentrations in large intestines and MLNs following C. jejuni infection of IL-10–/– mice [16]. To date, however, further data regarding the role of IL-18 in C. jejuni-host interaction are lacking. C. jejuni infection of three different cell lines (derived from premalignant Barrett’s esophagus) resulted in upregulated IL-18 gene expression [33]. Moreover, transcriptomic and proteomic analyses revealed that genes encoding IL-23 and IL-18, but not IL-22, were regulated in differentiated THP-1 macrophages following infection with an adherent and invasive strain of Campylobacter concisus [34]. Our data are further supported by results from our previous infection studies in gnotobiotic IL-10–/– mice with a different gram-negative bacterial species, namely, Arcobacter butzleri, sharing taxonomic relationship with C. jejuni. Results revealed that, in the colon, IL-18 was upregulated upon A butzleri infection during both the early and late phase of infection, whereas colonic IL-22 mRNA increased until day 6 p.i. [35].

In the present study, C. jejuni strain 81-176 was able to readily colonize the intestines of infant IL-18–/– mice only, whereas the pathogen was virtually expelled from the intestinal tract of IL-23p19–/–, IL-22–/–, and WT mice within the first 4 days postinfection. In a previous infection study where we had infected infant mice with a different pathogenic strain, namely, C. jejuni strain B2, the pathogen was also cleared rather early in the course of infection [13]. Nevertheless, infected mice exerted infection-induced phenotypes that were depending on the respective genotype of infant mice. The pathogenic colonization kinetics observed in IL-23p19–/–, IL-22–/–, and WT mice here were virtually comparable. Hence, genotype-dependent differences in immune responses cannot be attributed to differences in intestinal C. jejuni densities. It is rather the initial hit of infection that tips the balance towards immunopathological responses [13]. C. jejuni colonization is facilitated under conditions of elevated intestinal enterobacterial (i.e., E. coli) loads as shown not only in conventional infant mice [10, 11], but also in conventional mice fed viable E. coli or a Western diet [9, 36], in mice with a human microbiota [23], and in conventional mice suffering from T. gondii-induced acute ileitis [9, 37] or from chronic IL-10–/– colitis [38]. Unexpectedly, not infant IL-18–/– but IL-22–/– mice exhibited the highest gastrointestinal E. coli loads. This is well in line with a previous study showing that IL-22–/– mice harboured an altered colonic microbiota towards a higher abundance of the phylum Proteobacteria such as commensal gram-negative bacterial species including E. coli [39]. The altered micobiota composition that predisposed IL-22–/– mice to enhanced colitis susceptibility was attributed to the lacking regulatory properties of IL-22 including expression of antimicrobial peptides that are key components for epithelial barrier maintenance [39]. We are finally unable to explain why particularly infant IL-18–/–, but not mice of the remaining genotypes, were readily colonized by C. jejuni strain 81-176 in our present study. It is, however, highly likely that, so far, unidentified host-related factors might predispose infant IL-18–/– for C. jejuni infection.

In summary, our study indicates that cytokines belonging to the IL-23/IL-22/IL-18 axis are differentially involved in mediating and orchestrating pro- and anti-inflammatory immune responses in the large intestinal tract of C. jejuni-infected infant mice. We conclude that the regulatory pathways of IL-23, IL-22, and IL-18 following C. jejuni infection need to be further unravelled in future studies in order to improve our understanding of the distinct molecular mechanisms underlying campylobacteriosis.