Background
Campylobacter jejuni are spiral-shaped, highly motile, Gram-negative bacteria that frequenly asymptomatically colonize birds, including poultry. In humans the bacteria cause campylobacteriosis, the most prevalent cause for enteric bacterial infections [
1‐
4]. Human
C. jejuni infections are predominantly caused by consumption of contaminated animal products and surface water [
5]. Campylobacteriosis is accompanied with clinical manifestations such as abdominal pain, fever, and watery or bloody diarrhea that are mostly self-limiting [
1,
6,
7]. In a minority of cases, severe post-infectious sequelae such as Guillain-Barré syndrome or reactive arthritis can occur [
7,
8].
The exact molecular mechanisms underlying the development of acute and invasive enterocolitis that is typical for campylobacteriosis are unclear, but the immunopathological nature of the disease has been recognized for decades [
6]. We and others have shown that
C. jejuni interact with pattern recognition receptors such as Toll-like receptor 4 (TLR-4) [
9] and nucleotide-oligomerization-domain-2 (Nod2) [
10,
11], and interfere with signaling pathways dependent on MAPK/ERK (mitogen-activated protein kinases/extracellular signal-regulated kinases) and NF-κB (nuclear factor kappa-light-chain-enhancer of activated B cells) [
12]. Activation of those signaling cascades induces the expression of a variety of immune response genes [
13,
14]. As a result, an inflammation response is triggered, characterized by the recruitment of immune cells to the site of infection and up-regulation of cytokine production [
14].
As a prerequisite for induced immunopathology,
C. jejuni needs to adhere to and invade into epithelial host cells. Amongst a number of other factors, the flagellar filaments consisting of FlaA and FlaB, and the major adhesin CadF (
Campylobacter adhesin to fibronectin) are considered to be major players in these processes [
15]. To adhere to intestinal host cells, the bacteria need to cross the overlying mucus layer by flagella-generated motility [
16]. Moreover, the flagellum can secrete molecules that promote
C. jejuni adhesion to and invasion into host cells [
17‐
20]. The adhesin CadF permits host cell adhesion by binding to the extracellular matrix protein fibronectin, which enables the interaction with integrin receptors and results in bacterial internalization into host cells [
19,
21,
22].
The dependence of adherence and invasion on flagella has been demonstrated in vitro and in vivo by gene knockout experiments [
23,
24]. It was also shown that knockout of
cadF resulted in reduced adhesion and invasion of
C. jejuni into host cells in vitro [
21,
25] and abolished colonisation in the chicken host [
26]. Both the flagellum and CadF also activate a signaling cascade in cultured INT-407 cells and other cell lines that results in the activation of the small Rho GTPase Rac1, which in turn leads to actin and/or microtubule rearrangements that trigger internalization of
C. jejuni [
27].
In order to study pathogenesis, treatment and prophylaxis of campylobacteriosis in vertebrate hosts in more detail, we have established a murine
C. jejuni infection model based on secondary abiotic IL-10
−/− mice that not only allows for investigation of colonisation properties, but also reproducibly displays clinical symptoms resembling those of the compromized infected human host [
28‐
30]. Applying this clinical infection model we have recently shown, for instance, that
C. jejuni lipooligosaccharide (LOS) is essential for the induction of campylobacteriosis and this pathogen surface molecule thus represent an important
C. jejuni pathogenicity factor [
28,
31,
32].
To further validate this murine infection model for the study of C. jejuni virulence factors for induction and progression of acute disease, we here addressed whether bacterial flagella and the major adhesin CadF are pivotal prerequisites for inducing enteric disease in the murine host. To this aim, we infected secondary abiotic IL-10−/− mice with C. jejuni strain 81-176, its isogenic non-motile mutant ∆flaA/B and its CadF-deficient ∆cadF mutant. The colonization capacities of these isogenic strains were compared, while clinical outcome as well as intestinal, extraintestinal and systemic immunopathologocal responses were monitored and bacterial translocation to extra-intestinal organs was determined.
Discussion
Both the bacterial flagella and the adhesin CadF are well-investigated pathogenicity and virulence factors of
C. jejuni, respectively, and are considered key players for colonization and subsequent host cell invasion [
2‐
4]. In vitro studies revealed that CadF-mediated invasion of intestinal epithelial cells represents a crucial prerequisite for
C. jejuni to initiate immunopathological responses via induction of cytokine responses [
19,
21,
22,
34‐
36]. In fact, these immunopathological sequelae of infection help to explain the severity of symptoms during acute campylobacteriosis which are induced by cells of the innate immune system [
37]. In our present study, we provide in vivo evidence that FlaA/B and CadF exert differential features in the interaction of
C. jejuni and the mammalian host. By means of our clinical murine infection model, we show that
C. jejuni flagellar motility but not adhesion exerted by CadF is required for induced immunopathology in the murine host. Neither inactivation of the flagellin genes nor of the
cadF gene resulted in a compromized large intestinal colonization by
C. jejuni as indicated by comparably high colonic loads of either bacterial strain. However, whereas WT bacteria and the
cadF deficient mutant could also be isolated from the stomach, duodenum and ileum upon peroral infection, these sites were only poorly colonized by the non-motile mutant strain. This pronounced phenotype provides strong evidence that motility is required to allow
C. jejuni to escape the unfavorable luminal conditions exerted by acids, bicarbonate and lytic enzymes within the lumen of the upper environmental tract, for instance. In this scenario motility allows the bacteria to reach mucus sites where the pathogen is protected from toxic influences and can adhere to epithelial cells to prevent passive transport to the colon where even the non-motile bacteria accumulate due to the low peristaltics.
It is well established that flagella-deficient
C. jejuni mutants are unable to colonize the gastrointestinal tract of infant mice [
38,
39], of young chicks [
40] and of piglets [
41]. However, a role of CadF in colonisation of the mammalian gut was mostly inferred from in vitro data, as it was shown previously that
cadF-deficient
C. jejuni mutants poorly adhered to cultured mammalian cells [
21,
25,
34], though CadF was reported to be essential for cecal colonization in chicken [
26]. Our data provide strong evidence, however, that CadF is not essential for murine colonization, and only marginally affects the outcome of infection. Thus, one or more different adhesin(s) are obviously required for
C. jejuni colonization and disease development in mice. Because this might reflect the situation in humans, it will be important to identify these factors in further screens.
The difference between mutants deficient in flagellins or
cadF extended beyond colonization capacity, given that there were also noted differences in their ability to generate macroscopic disease signs and microscopic inflammatory responses in the large intestines. Interestingly, WT and
cadF deficient bacteria were able to enhance pro-inflammatory mediator secretion in distinct compartments of the intestinal tract, which was not seen with the non-motile
C. jejuni mutant lacking flagella. The murine model applied here also allowed to investigate the capacity to generate campylobacteriosis-like symptoms in mice. Surprisingly, high numbers of
C. jejuni present in the colon were not per se responsible for triggering disease, as could be demonstrated with the Δ
flaA/B mutant that colonized the colon effectively, but did not cause enteric disease. In a recent study applying a murine
C. jejuni induced enteritis model, however, mice could not be infected by a
flaA-deficient mutant strain and did therefore not display any signs of enteritis [
42]. Notably, in this study the gut microbiota of corresponding mice was not completely eradicated and this leads to the assumption that the residual microbiota established after vancomycin treatment is responsible for the complete colonization defect of the
flaA deficient mutant. This provides evidence that motility is required to allow
C. jejuni to escape from commensal bacteria that produce harmful metabolites and thus create an unfavorable environment for the pathogen.
In our study, the presence of symptoms coincided with the ability to generate intestinal, extra-intestinal and systemic immune responses, which both WT bacteria and the Δ
cadF mutant were capable of. This observation further confirms the hypothesis of an immunopathological nature of campylobacteriosis. It has been hypothesized that bacterial invasion into host cells is regulated by pro-inflammatory mediators in the gut [
43] and this idea is in line with our observations, since non-motile mutant strains, that are impaired in their capacity to invade concurrently exhibited far weaker immune responses.
A marked difference was further observed in the ability of
C. jejuni to reach extra-intestinal sites. All three bacterial strains could be isolated from extra-intestinal organs, but the numbers of viable bacteria were much lower for the non-motile Δ
flaA/B mutant. As expected, it appeared that due to lack of flagella-dependent motility, fewer bacteria were able to reach liver, lungs, and spleen. Alternatively, it has been shown that the flagellum of
C. jejuni can act as a type III secretion apparatus for the delivery of bacterial factors such as the Cia or Fed proteins into the extracellular milieu or directly into host cells in vitro [
18‐
20]. Thus, certain exported
C. jejuni proteins may also trigger the above responses in mice. Thus, future studies should be designed to provide evidence for one or both of these options.
Methods
Ethics approval
All animal experiments were conducted in accordance with the European Guidelines for animal welfare (2010/63/EU) following approval of the protocol by the commission for animal experiments headed by the “Landesamt für Gesundheit und Soziales” (LaGeSo, Berlin, registration number G0247/16). Clinical conditions of mice were surveyed twice daily.
Generation of secondary abiotic mice and C. jejuni infection
IL-10
−/− mice of C57BL/6j background were reared and housed under specific pathogen free conditions. In order to counteract physiological colonization resistance and hence facilitate intestinal pathogenic colonization, secondary abiotic mice with a virtually depleted gut microbiota were generated upon broad-spectrum antibiotic treatment as reported earlier [
31,
33].
Sex-matched, 3 months old mice were perorally infected with either the C. jejuni parental strain 81-176 (WT), the isogenic flaA/B deletion mutant (ΔflaA/B), or the cadF deletion mutant (ΔcadF). An inoculum of 109 CFU in 0.3 mL phosphate buffered saline (PBS; Gibco, life technologies, UK) was administered on two consecutive days (i.e., days 0 and 1) by oral gavage. Mock control animals received an equal volume PBS perorally. Mice were maintained in a sterile environment and had unlimited access to autoclaved food and drinking water and were handled under strict aseptic conditions to avoid contamination.
Monitoring of clinical conditions
The clinical conditions of the mice were surveyed prior and post respective
C. jejuni infections on a daily basis by applying a standardized cumulative clinical score (maximum 12 points). These scores included the abundance of blood in feces as detected by the Guajac method using a Haemoccult, Beckman Coulter (PCD, Krefeld, Germany) (score 0: no blood; 2: microscopic detection of blood; 4: macroscopic blood visible), presence of diarrhea (score 0: formed feces; 2: pasty feces; 4: liquid feces), and by visual clinical and behavioral symptoms (score 0: normal; 2: ruffled fur and/or less locomotion; 4: isolation, severely compromised locomotion, pre-final aspect) as described earlier [
29].
Sampling procedures
At day 6 post-infection (p.i.), the animals were sacrificed upon isoflurane inhalation (Abbott, Germany). Luminal gastrointestinal samples from stomach, duodenum, ileum and colon, and ex vivo biopsies from colon, ileum, mesenteric lymph nodes (MLN), liver, kidneys, lungs, and spleen were taken under sterile conditions. Intestinal samples were collected from each mouse in parallel for microbiological, immunohistopathological and immunological analyses. The absolute colonic lengths were measured with a ruler (in cm).
Immunohistochemistry
In situ immunohistochemical analyses were performed in colonic ex vivo biopsies that had been immediately fixed in 5% formalin and embedded in paraffin as described earlier [
44‐
46]. Paraffin sections (5 μm) of ex vivo biopsies from colon, liver and kidneys were stained with primary antibodies directed against cleaved caspase 3 (Asp175, Cell Signaling, Beverly, MA, USA, 1:200) for detection of apoptotic epithelial cells; against Ki67 (TEC3, Dako, Denmark, 1:100) for detection of proliferating epithelial cells; against F4/80 (# 14-4801, clone BM8, eBioscience, San Diego, CA, USA, 1:50) for detection of macrophages/monocytes; against CD3 (#N1580, Dako, 1:10) for detection of T lymphocytes; and against B220 (No. 14-0452-81, eBioscience; 1:200) for detection of B lymphocytes. Secondary antibodies were used for detection as previously described [
31,
47]. Positively stained cells were examined by light microscopy (magnification 100× and 400×), and for each mouse the average number of respective positively stained cells was determined within at least six high power fields (HPF, 0.287 mm
2, 400× magnification) by an independent investigator using blinded samples.
Bacterial colonization
The number of viable
C. jejuni bacteria was quantitatively assessed in feces over time p.i., in homogenates of ex vivo biopsies taken MLN, spleen, liver, kidneys and lungs, and in cardiac blood at day 6 p.i. by culture as described elsewhere [
31,
47]. The detection limit of viable bacteria was ≈ 100 CFU per g.
Colonic ex vivo biopsies were cut longitudinally, washed in PBS, and strips of approximately 1 cm
2 tissue as well as ex vivo biopsies derived from MLN (3 lymph nodes), liver (approximately 1 cm
3), one kidney (cut longitudinally), and one lung were placed in 24-flat-bottom well culture plates (Nunc, Germany) containing 500 μL serum-free RPMI 1640 medium (Gibco, life technologies, UK) supplemented with penicillin (100 U/mL) and streptomycin (100 µg/mL; PAA Laboratories, Germany). After 18 h at 37 °C, culture supernatants were tested for tumor necrosis factor- (TNF-) α, interleukin (IL)-6, interferon (IFN)-γ, and monocyte chemoattractant protein (MCP)-1 by the Mouse Inflammation Cytometric Bead Array (CBA; BD Biosciences, Germany) on a BD FACSCanto II flow cytometer (BD Biosciences). Nitric oxide was measured by the Griess reaction as reported previously [
33]. Systemic pro-inflammatory mediator concentrations were assessed in serum samples.
Statistical analysis
Medians and levels of significance were determined by one-way ANOVA test followed by Tukey post-correction for multiple comparisons (GraphPad Prism v7, USA). Two-sided probability (p) values ≤ 0.05 were considered significant. Experiments were reproduced three times and pooled data are shown.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.