Campylobacter jejuni Disrupts Protective Toll-Like Receptor 9 Signaling in Colonic Epithelial Cells and Increases the Severity of Dextran Sulfate Sodium-Induced Colitis in Mice

Abstract

Inflammatory bowel disease (IBD) is characterized by chronic intestinal inflammation associated with a dysregulated immune response to commensal bacteria in susceptible individuals. The relapse of IBD may occur following an infection with Campylobacter jejuni. Apical epithelial Toll-like receptor 9 (TLR9) activation by bacterial DNA is reported to maintain colonic homeostasis. We investigated whether a prior C. jejuni infection disrupts epithelial TLR9 signaling and increases the severity of disease in a model of mild dextran sulfate sodium (DSS) colitis in mice. In a further attempt to identify mechanisms, T84 monolayers were treated with C. jejuni followed by a TLR9 agonist. Transepithelial resistance (TER) and dextran flux across confluent monolayers were monitored. Immunohistochemistry, Western blotting, and flow cytometry were used to examine TLR9 expression. Mice colonized by C. jejuni lacked any detectable pathology; however, in response to low levels of DSS, mice previously exposed to C. jejuni exhibited significantly reduced weight gain and increased occult blood and histological damage scores. Infected mice treated with DSS also demonstrated a significant reduction in levels of the anti-inflammatory cytokine interleukin-25. In vitro studies indicated that apical application of a TLR9 agonist enhances intestinal epithelial barrier function and that this response is lost in C. jejuni-infected monolayers. Furthermore, infected cells secreted significantly more CXCL8 following the basolateral application of a TLR9 agonist. Surface TLR9 expression was reduced in C. jejuni-infected monolayers subsequently exposed to a TLR9 agonist. In conclusion, infection by C. jejuni disrupts TLR9-induced reinforcement of the intestinal epithelial barrier, and colonization by C. jejuni increases the severity of mild DSS colitis.

INTRODUCTION

Inflammatory bowel disease (IBD), including Crohn's disease and ulcerative colitis, are chronic, relapsing inflammatory disorders of the gastrointestinal tract (47). Symptoms associated with IBD include diarrhea, bloody stools, abdominal pain, and weight loss. The pathogenesis of this debilitating disease involves the interaction of multiple factors, including host immune response, genetics, and environmental triggers. A widely held hypothesis is that an inappropriate immune response to the resident microflora is generated in genetically susceptible individuals (31, 47). Although a specific etiological agent has yet to be identified, several reports suggest that acute bacterial enteritis may initiate or exacerbate disease (1, 9, 12, 13, 30, 46, 49, 51).

Campylobacter species, including C. jejuni, are the leading cause of acute bacterial enteritis in the developed world (20). Infection can cause inflammatory diarrhea, fever, and abdominal pain (54). While most cases are self-limiting, postinfection complications, including Guillain-Barré syndrome and irritable bowel syndrome, can arise (29, 34). Campylobacter species are also commonly isolated from patients with colitis (28, 55), and an acute enteric infection can induce the relapse of IBD (35, 36, 53). Moreover, several recent clinical studies indicate that an acute infection with Campylobacter is a risk factor for the subsequent development of IBD (12, 13, 51). Despite significant advances, the mechanisms by which these events induce the onset or relapse of disease symptoms remain elusive.

Bacterial pathogens such as C. jejuni contain pathogen-associated molecular patterns that are recognized by Toll-like receptors (TLRs). Intestinal epithelial cells are the first line of defense against bacterial pathogens and express several TLRs (25). Activation of TLRs is typically associated with an inflammatory immune response that is required to clear the invading pathogen. Interestingly, TLRs also recognize the same conserved molecular patterns found in the resident microflora. However, there is a striking contrast in the response elicited by commensal bacteria. Not only does the healthy intestinal epithelium tolerate the presence of the resident microflora, but it also appears to require commensal recognition of bacteria via TLRs to maintain intestinal homeostasis (4, 26, 44).

Prior studies have demonstrated that bacterial DNA or its synthetic oligodeoxynucleotide (ODN) analogues (immunostimulatory sequence [ISS] ODN or CpG ODN) can effectively ameliorate the severity of colitis in various animal models (2, 38, 42). Bacterial DNA and its synthetic analogues are recognized by TLR9. Importantly, mice deficient in TLR9 are more susceptible to colitis (26) and TLR9 genetic polymorphisms are associated with an increased risk of IBD (11, 52), suggesting a critical role for abnormal bacterial DNA sensing in the development of IBD. Whether and how C. jejuni-induced TLR9 disruptions may increase the severity of colitis remains obscure.

Given the importance of bacterial DNA sensing by TLR9 in maintaining intestinal homeostasis, we hypothesized that polarized and protective epithelial TLR9 signaling is disrupted following an acute infection with C. jejuni and that this effect may induce a predisposition to heightened intestinal inflammation. We first developed an animal model to examine whether an acute infection with C. jejuni increases severity in a model of mild experimental colitis. We also utilized in vitro cell culture techniques to examine the effects of an infection on TLR9 signaling in intestinal epithelial cells. Our results reveal that an infection with C. jejuni disrupts protective apical TLR9 signaling and sensitizes proinflammatory basolateral TLR9 signaling. Furthermore, colonization by C. jejuni, in the absence of an inflammatory response, increases the severity of mild DSS-induced colitis in mice.

MATERIALS AND METHODS

Reagents, inhibitors, and antibodies.

Type B CpG oligodeoxynucleotide (ODN 2006), a human TLR9 ligand, in addition to ODN 2006 control oligodeoxynucleotide, was purchased from Invivogen (San Diego, CA). Bay 11-7085, apigenin, and LY294002 were purchased from CalBiochem (Gibbstown, NJ), Calbiochem, and Cell Signaling Technology (Danvers, MA), respectively.

The following antibodies were used in this study: mouse monoclonal anti-TLR9 (IMG-305A; Imgenex, San Diego, CA) for immunocytochemistry and flow cytometry; rabbit polyclonal anti-TLR9 (IMG-431; Imgenex) for immunohistochemistry; and rabbit polyclonal anti-TLR9 (GTX111547; GeneTex, Irvine, CA) and mouse anti-GAPDH (anti-glyceraldehyde-3-phosphate dehydrogenase) (sc-59540; Santa Cruz Biotechnology, Santa Cruz, CA) for Western blotting. Mouse Alexa Fluor 488- and 555-conjugated secondary antibodies (Invitrogen, Carlsbad, CA) were used for immunofluorescence and flow cytometry. Horseradish peroxidase (HRP)-conjugated secondary antibodies (Cell Signaling Technology, Danvers, MA) were used for Western blotting.

Mice.

Wild-type C57BL/6 mice (Charles River; Montreal, Quebec, Canada), 4 to 6 weeks of age, were housed in a temperature-controlled room. Mice were maintained on a normal 12-h/12-h light-dark cycle and allowed free access to standard laboratory chow and water. All methods used in this study were approved by the University of Calgary Animal Care Committee and were carried out in accordance with the guidelines of the Canadian Council on Animal Care.

Cell culture.

T84 human colonic epithelial cells (American Type Culture Collection, Manassas, VA) were cultured in Dulbecco's modified Eagle's medium (DMEM)–F-12 (Sigma, St. Louis, MO), supplemented with 10% heat-inactivated fetal bovine serum (FBS), 100 μg ml⁻¹ streptomycin, 100 U ml⁻¹ penicillin, 80 μg ml⁻¹ tylosin, and 200 mM l-glutamine (all from Sigma). Cells were kept at 37°C in 5% CO₂ at 96% humidity. Culture medium was replenished every 2 to 3 days, and cells were subjected to passage with 2× trypsin-EDTA (Sigma). Cells were grown to confluence in 6-well plates for immunoblotting and flow cytometry or on Laboratory-Tek chamber slides (Nalge Nunc International, Naperville, IL) for immunocytochemistry. Cells were also grown on Transwell filter units containing a 1.13-cm² or 5-cm² semipermeable filter membrane (Costar, Cambridge, MA) (0.4-μm pore size) for determination of epithelial permeability, CXCL8 secretion, or flow cytometry. T84 cells were used at passages 78 to 83. All experiments were performed using serum and antibiotic-free media (infection media).

Bacteria and infection of mice.

Mice were inoculated with 0.1 ml of Casamino Acids-yeast extract (CYE) broth containing approximately 10⁸ CFU C. jejuni 81-176 plus 2% NaHCO₃ by oral gavage on days 1 and 2, as previously described (17). Sham-treated controls were challenged with equal volumes of sterile CYE broth plus 2% NaHCO₃. To assess fecal shedding of C. jejuni, fecal pellets were homogenized in phosphate-buffered saline (PBS) and plated on Karmali agar containing selective SR167 supplement (Oxoid; Nepean, Ontario, Canada). Karmali agar dishes were examined for the presence of C. jejuni growth and scored as positive or negative.

Bacteria and infection of cells.

For all experiments, C. jejuni 81-176 was grown from frozen glycerol stocks on Karmali agar plates for 48 h at 37°C in gas jars (Oxoid) under microaerobic conditions. Inoculum was prepared from the agar plates by suspending Campylobacter bacteria in CYE broth for 14 to 16 h (37°C, 100 rpm, microaerobic conditions). Log-phase bacteria were centrifuged at 2,500 × g for 10 min and resuspended in infection media to achieve a multiplicity of infection of 100.

Induction of colitis.

At 7 days postinfection, colitis was induced by addition of dextran sulfate sodium (DSS; MP Biomedicals, Solon, OH) (2% wt/vol) to the drinking water. The mean level of DSS-water consumption was monitored for each group. Mice were weighed daily and monitored for fecal blood and changes in stool consistency. Fecal pellets were collected and examined for the presence of gross blood and tested for occult blood by the use of a Hemoccult Sensa kit according to the manufacturer's instructions (Beckman Coulter, Fullerton, CA). An occult blood score was assigned according to the following criteria: no blood, score = 0; hemoccult positive, score = 2; gross visible blood, score = 3.

Tissue collection and histology.

Mice were euthanized by cervical dislocation 12 days following the start of DSS administration. The colon was removed, and segments were frozen in liquid nitrogen for enzyme-linked immunosorbent assay (ELISA) or fixed in 10% buffered formalin for histology. Paraffin sections (8 μm) were cut and stained with hematoxylin and eosin (H&E). The severity of colonic inflammation was analyzed in a blinded fashion according to previously described criteria (7), with some modification. Briefly, the severity of inflammation was assigned a histological damage score based on the presence of ulceration (0 to 1), inflammatory infiltrate (0 to 3), or edema (0 to 1) and on general tissue architecture, including colonic wall thickening (0 to 3).

Analysis of cytokine levels.

Interleukin-17E (IL-17E; IL-25) and IL-17 cytokine levels in colonic homogenates were analyzed by ELISA. Briefly, segments of distal colon from each group were homogenized in ice-cold Tris-HCl buffer containing protease inhibitors (Complete-Mini; Roche Diagnostics, Laval, Quebec, Canada). Samples were centrifuged at 10,000 × g for 30 min, and the supernatants were collected. Aliquots were stored at −70°C until assay. The concentrations of IL-25 and IL-17 were analyzed using an enzyme immunoassay according to the instructions of the manufacturers (mouse IL-25 kit [Biolegend, San Diego, CA] and mouse IL-17A kit [eBioscience, San Diego, CA]).

CXCL8 ELISA.

Control and infected monolayers were treated with apical or basolateral ODN 2006 as described above. After incubation, cell supernatants were collected from the basolateral compartment, snap-frozen in liquid nitrogen, and stored at −80°C. Concentrations of CXCL8 secreted from polarized T84 cells were analyzed with an enzyme-linked immunoabsorbant assay according to the manufacturer's instructions (R&D Systems, Minneapolis, MN).

In vitro epithelial permeability.

Transepithelial resistance (TER) was monitored using an electrovoltohmmeter (World Precision Instruments, Sarasota, FL). Confluent monolayers (TER > 1,000 Ω/cm²) were washed twice with sterile Hanks' balanced salt solution (HBSS), followed by the addition of either sterile infection media or infection media containing C. jejuni for 24 h. In all cases, C. jejuni was added to the apical compartment. Monolayers were washed twice with sterile HBSS and treated either apically or basolaterally with a TLR9 agonist (ODN 2006 or control ODN) (5 μg/ml) for an additional 24 h. For some experiments, monolayers were treated apically for 1 h prior to administration of the TLR9 agonist with one of the following inhibitors: Bay 11-7085 (NF-κB inhibitor) (10 μmol/liter), apigenin (mitogen-activated protein kinase [MAPK] inhibitor) (20 μmol/liter), or LY294002 (phosphatidylinositol 3-kinase [PI3K] inhibitor) (10 μmol/liter).

Paracellular permeability was assessed using a nonabsorbable fluorescein isothiocyanate (FITC)-conjugated 3-kDa dextran probe. Briefly, the apical and basolateral compartments were washed twice with sterile bicarbonate-buffered Ringer's solution. The FITC-dextran probe (500 μl; 100 μM in Ringer's solution) was added to the apical compartment, and Ringer's solution (1 ml) was added to the basal chamber. After a 3-h incubation (37°C, 5% CO₂, 96% humidity), samples were collected from the basal compartment and relative fluorescence units were calculated with a microplate fluorometer (Spectra Max Gemini; Molecular Devices, Sunnyvale, CA).

Immunohistochemistry.

Segments of colon were fixed in 4% paraformaldehyde overnight at 4°C. Samples were washed, transferred to 20% sucrose in PBS overnight at 4°C, and embedded in OCT compound (Miles, Elkhardt, IN). Sections of colon (8 μm) were cut on a cryostat and mounted onto poly-d-lysine coated slides. The slides were then washed in PBS–0.1% Triton X-100 (three times at 10 min each time), blocked in 2% normal goat serum for 1 h, and incubated with rabbit anti-TLR9 (1/100) for 48 h at 4°C. Sections were washed again and incubated with secondary antiserum for 2 h at room temperature. Slides were examined with a Leica DMR fluorescence microscope. Photos were taken at the same exposure and magnification (×40) using a Retiga 2000x camera (Q Imaging, Surrey, BC).

Immunocytochemistry.

Cells were washed 3× with sterile PBS, followed by fixation and permeabilization in Cytofix/Cytoperm (BD Biosciences) for 20 min at 4°C. Following three washes in Permwash (BD Biosciences), cells were blocked with PBS containing 2% BSA for 1 h at room temperature and then stained with mouse anti-TLR9 antibody (1/200) overnight at 4°C. Cells were washed with Permwash and incubated with secondary antibody (1/1,000) for 1 h at 37°C in the dark. Slides were mounted with Aqua PolyMount (Polysciences, Warrington, PA) and visualized using a Leica DMR fluorescence microscope. Micrographs were taken using a Retiga 2000x camera.

Western blot analysis.

T84 cells were washed twice with HBSS and lysed in RIPA buffer (1× PBS, 1% Igepal CA-630, 0.5% sodium deoxycholate, 0.1% sodium dodecyl sulfate [SDS]; all from Sigma) containing a protease inhibitor tablet (Complete Mini; Roche Diagnostics, Laval, Quebec, Canada) for 30 min at 4°C. Lysates were sonicated and centrifuged at 10,000 × g for 10 min. Supernatants were collected, and protein levels were normalized using a Bradford assay (Bio-Rad, Hercules, CA). Samples were diluted in SDS sample buffer at a 1:1 ratio, boiled for 5 min, and stored at −20°C.

Protein samples were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) (7 to 10%) and transferred to nitrocellulose membranes (Whatman, Buckinghamshire, England). Membranes were blocked for 1 h in Tris-buffered saline (TBS) plus 0.1% Tween (TBS-T) containing 5% nonfat skim milk, followed by incubation with primary antibodies overnight at 4°C. Membranes were washed with TBS-T and incubated with HRP-conjugated secondary antibodies (1:1,000) for 1 h at room temperature. Bands were visualized using an ECL Plus chemiluminescence detection system (GE Healthcare, Pittsburgh, PA), and band density was determined using a Canon CanoScan 4400F scanner and Image J densitometry software. To confirm equal loading, membranes were stripped in 0.5 M acetic acid for 1 h and reprobed for GAPDH.

Flow cytometry.

For evaluation of surface TLR9 expression, T84 cells were detached with 0.2% EDTA–PBS and washed with PBS containing 3% FBS. Live cells were stained with mouse anti-TLR9 antibody (1/100) for 30 min at 4°C, washed twice with PBS, and incubated with Alexafluor 488-goat anti-mouse secondary antibody (1/1,000) for 30 min at 4°C in the dark. Expression of TLR9 was measured with a BD LSR II flow cytometer.

Statistical analysis.

Results are expressed as means ± standard errors of the means (SEM), and statistical comparisons were conducted with GraphPad Prism software (GraphPad Software, San Diego, CA). Comparisons of three or more groups were analyzed by one-way analysis of variance (ANOVA) followed by Tukey's multiple-comparison test. Data with a nonparametric distribution were analyzed by a Kruskal-Wallis test followed by Dunn's multiple-comparison test. Differences of P < 0.05 were considered statistically significant.

RESULTS

An acute infection with C. jejuni increases the severity of DSS-induced colitis in mice.

Wild-type mice are transiently infected by C. jejuni, with no underlying pathology or detectable clinical response (54). In the present study, C. jejuni was detected in the feces of most mice (80%) 1 week after inoculation. By the end of the study, detection of C. jejuni decreased notably and bacteria could be detected in 56.7% of infected mice (data not shown). The absence of detectable pathology or clinical response was confirmed in our study, as evidenced by the similar weight gain, disease activity index, and histology scores measured in sham-treated mice and in mice treated with C. jejuni alone (Fig. 1 and 2). Mice treated with low levels of DSS had a pattern of weight gain similar to that of control mice (Fig. 1). However, mice previously exposed to C. jejuni demonstrated a significant drop in weight 3 days following the start of DSS treatment. After 12 days of DSS treatment, the infected mice had gained significantly less weight than all other groups (Fig. 1).

Fig 1.

C. jejuni infection increases clinical disease activity of mild DSS colitis. (A) Percent change in body weight in C57BL/6 mice. Three days after the start of DSS, infected mice exhibited a significant drop in weight compared to control mice, whereas mice treated with DSS or C. jejuni alone had weight gain similar to that of control mice. By 12 days of DSS treatment, infected mice had gained significantly less weight than all other groups (n = 7 or 8 mice/group; mean ± SEM). *, P < 0.05 compared to controls; σ, P < 0.05 compared to DSS-treated mice. (B) C. jejuni-infected mice showed an increased incidence of occult blood following DSS treatment compared to all other groups (n = 15 to 16 mice/group; mean ± SEM). *, P < 0.05.

Fig 2.

C. jejuni infection increases the severity of histological inflammation in a model of mild DSS colitis. (A) Representative micrographs of H&E-stained colonic sections. Colons from control (upper left panel) and C. jejuni-treated (upper right panel) mice appeared microscopically similar. Control mice treated with low levels of DSS (lower left panel) displayed mild levels of inflammation. In contrast, infected mice treated with a low level of DSS exhibited extensive inflammation, as evidenced by increased mucosal thickening, inflammatory infiltrate, and ulceration. Photos were taken at the same magnification. (B) Histological damage scores in control and C. jejuni-, DSS-, and C. jejuni-plus-DSS-treated mice. Infected mice treated with DSS demonstrated a significantly increased histopathological score compared to control mice (n = 5 to 8 mice/group; mean ± SEM). *, P < 0.05.

C. jejuni-infected mice treated with DSS also exhibited significantly worse fecal blood scores than mice treated with DSS alone (Fig. 1B). Seven days after the start of DSS treatment, 81.3% of infected mice tested positive for occult blood compared with 46.7% of noninfected mice treated with DSS. Furthermore, representative micrographs (Fig. 2A) illustrate that noninfected mice treated with low levels of DSS exhibited very mild levels of inflammation (Fig. 2B). In contrast, DSS treatment resulted in significant colonic thickening and inflammatory infiltrate and extensive alteration to the tissue architecture in mice exposed to C. jejuni (Fig. 2).

Effect of a prior C. jejuni infection on cytokine levels in DSS colitis.

To determine whether the increased severity of intestinal inflammation was accompanied by altered cytokine production, we measured IL-25 and IL-17 levels in colonic homogenates. Our results show similar levels of colonic IL-25 in sham-, C. jejuni-, and DSS-treated mice, whereas there was a significant reduction of IL-25 in C. jejuni-infected mice treated with DSS (Fig. 3A). In addition, infected mice treated with DSS exhibited lower levels of IL-17 than sham-treated mice (Fig. 3B).

Fig 3.

Modulation of Th17 cytokine levels following DSS treatment in C. jejuni-infected mice. IL-25 (A) and IL-17 (B) levels were significantly reduced in colonic homogenates from C. jejuni-infected mice treated with DSS (n = 7 or 8 mice/group; mean ± SEM). ***, P < 0.0001; **, P < 0.01; *, P < 0.05.

TLR9 immunofluorescence is reduced in the colons of C. jejuni-infected mice.

To evaluate whether TLR9 expression is disrupted in C. jejuni-infected mice, TLR9 protein expression was examined using immunohistochemistry. In sham-treated mice, TLR9 fluorescence was observed throughout the colonic mucosa and was particularly concentrated in the epithelial cells lining the lumen of the gut (Fig. 4). Epithelial TLR9 fluorescence was notably reduced in mice exposed to C. jejuni compared to the sham-treated control results (Fig. 4).

Fig 4.

Representative micrographs comparing TLR9 immunofluorescence results in the colons of sham- and C. jejuni-treated mice. Magnifications (×40) and exposure times are the same. TLR9 fluorescence is notably reduced in colonic epithelial cells following infection with C. jejuni (right panel) compared to sham-treated controls (left panel).

TLR9 ligand increases intestinal epithelial barrier function in a polarized manner.

In vitro experiments were used to assess the mechanisms by which a prior infection with C. jejuni increases the sensitivity to gut injury. We examined the effect of TLR9 stimulation on epithelial barrier function in polarized epithelial monolayers. TER and the apical-to-basolateral flux of a FITC-labeled marker were monitored following administration of ODN 2006 to the apical or basolateral chambers of transwell units. Following the apical administration of ODN 2006, TER was increased by more than 50% (Fig. 5A). This was accompanied by a significant reduction in solute flux across the monolayer (Fig. 5B), indicating a tightening of the epithelial barrier. When the TLR9 agonist was added to the basolateral compartment, no effect on paracellular permeability was observed. The administration of control ODN 2006 had no effect on TER (data not shown).

Fig 5.

C. jejuni attenuates TLR9-induced increase in epithelial barrier function. (A) Transepithelial resistance of T84 monolayers was significantly increased in response to the administration of ODN 2006 for 24 h to the apical compartment of transwell units. Basolateral application of ODN 2006 failed to elicit a change in TER (n = 18; mean ± SEM). ***, P < 0.0001. (B) The apical-to-basolateral flux of a FITC-dextran probe was significantly lower in cells treated with apical ODN 2006 (n = 15 to 16/group; mean ± SEM). ***, P < 0.0001; **, P < 0.001. (C and D) T84 cells pretreated with C. jejuni for 24 h did not exhibit a significant change in TER (C) or apical-to-basolateral flux of a FITC-dextran probe following the administration of ODN 2006 for an additional 24 h to the apical or basolateral chambers (D) (n = 15 to 18/group; mean ± SEM).

Effect of C. jejuni pretreatment on TLR9-induced increase in epithelial barrier function.

To determine whether a prior infection with C. jejuni could disrupt the TLR9-induced increase in TER, monolayers were pretreated with C. jejuni for 24 h, followed by the administration of ODN 2006, as described above. Monolayers pretreated with C. jejuni failed to exhibit the TLR9-induced increase in TER (Fig. 5C), and infected monolayers did not exhibit a reduction in solute flux in response to apically applied ODN 2006 (Fig. 5D).

C. jejuni infection alters the distribution and expression of TLR9 in epithelial cells.

TLR9 appears as continuous and uniform staining at the cell periphery in control monolayers and ODN 2006-treated monolayers (Fig. 6A). After 24 h of incubation with C. jejuni, there was some redistribution of TLR9 staining to the cytoplasmic compartment. The redistribution of TLR9 was enhanced when the infected monolayers were subsequently treated with ODN 2006. Western blot analyses of whole-cell lysates indicate that infected cells treated with the TLR9 agonist expressed significantly less total epithelial TLR9 (Fig. 6B). To further examine the expression pattern of TLR9, live T84 cells were stained and analyzed by flow cytometry. Our results confirm previous studies demonstrating surface expression of TLR9 in control intestinal epithelial cells (Fig. 7A), and administration of ODN 2006 to C. jejuni-infected cells resulted in a significant reduction in surface TLR9 fluorescence (Fig. 7).

Fig 6.

C. jejuni infection modulates TLR9 protein expression in epithelial cell monolayers. (A) Representative micrographs of TLR9 immunofluorescence, demonstrating the redistribution of TLR9 away from the cell membrane in C. jejuni-infected monolayers (lower left panel). Infected monolayers treated with ODN 2006 (lower right panel) exhibit further redistribution. (B) Loss of TLR9 protein levels as determined by immunoblotting. Densitometry data are expressed as percent GAPDH (n = 9/group; mean ± SEM). ***, P < 0.0001; **, P < 00.1.

Fig 7.

Surface TLR9 expression is lost in C. jejuni-infected intestinal epithelial cells following ODN 2006 treatment. Live T84 cells were labeled with an anti-hTLR9 antibody and analyzed by flow cytometry for surface detection. (A) Representative fluorescence-activated cell sorter histograms comparing surface expression of TLR9 in control (upper left panel) and ODN 2006 (upper right panel)-, C. jejuni (lower left panel)-, and C. jejuni-plus-ODN 2006-treated intestinal epithelial cells (lower right panel). The green histogram represents the level of surface TLR9 expression. The blue-filled histogram represents the unlabeled cells. (B) The surface expression of TLR9 is significantly reduced following ODN 2006 treatment in C. jejuni-infected intestinal epithelial cells (n = 6/group; mean ± SEM). *, P < 0.05.

Effect of C. jejuni on TLR9-induced CXCL8 secretion.

To investigate whether the C. jejuni-induced disruption of TLR9 signaling had an effect on CXCL8, a proinflammatory chemokine upregulated in tissues from IBD patients (37), we examined CXCL8 secretion from T84 cells in response to ODN 2006. We observed increased CXCL8 secretion from control monolayers when a TLR9 agonist was added to the basolateral chamber of the transwell unit. Moreover, significantly more CXCL8 was secreted from infected monolayers in response to basolateral activation of TLR9 compared to control monolayers (Fig. 8).

Fig 8.

C. jejuni infection potentiates ODN 2006-induced CXCL8 secretion. Control and C. jejuni-infected T84 monolayers were treated apically or basolaterally with ODN 2006 for 24 h, and CXCL8 secretion was measured by ELISA. Basolateral stimulation with ODN 2006 induced CXCL8 secretion from control cells. Cells previously infected with C. jejuni secreted significantly more CXCL8 in response to basolateral stimulation with ODN 2006 (n = 24/group; mean ± SEM). *, P < 0.05.

TLR9-induced increase in TER is independent of NF-κB, MAPK, and PI3K signaling pathways.

We then investigated possible signaling pathways that are triggered by apical TLR9 to induce the increase in TER. Two common signaling pathways associated with TLR-induced responses are those of NF-κB and the mitogen-activated protein kinases (MAPKs) (50). Previous studies have also demonstrated that epithelial TLR2 stimulation preserves intestinal epithelial barrier integrity via a PI3K/Akt-dependent pathway (4). However, pretreatment with an NF-κB (Fig. 2A), MAPK (Fig. 9B), or PI3K (Fig. 9C) inhibitor did not significantly affect baseline TER and these inhibitors did not attenuate the TLR9-induced increase in TER.

Fig 9.

TLR9-induced increases in TER are independent of NF-κB, MAPK, and PI3K signaling. Pretreatment of T84 cell monolayers with an NF-κB (Bay 11-7085; 10 μmol/liter) (A), MAPK (apigenin; 20 μmol/liter) (B), or PI3K (LY294002; 10 μmol/liter) (C) inhibitor for 1 h did not attenuate the increase in TER induced by the administration of ODN 2006 to the apical compartment (n = 12 to 16/group; mean ± SEM). *, P < 0.05.

DISCUSSION

In this study, we demonstrated that apical activation of TLR9 increases TER and decreases solute flux across intestinal epithelial cell monolayers. Infection with C. jejuni attenuates the TLR9-induced increase in epithelial barrier function, and this appears to occur via the loss of surface TLR9. C. jejuni also sensitizes intestinal epithelial cells to a subsequent proinflammatory stimulus, as infected monolayers secrete significantly more CXCL8 in response to basolateral activation of TLR9. Furthermore, mice infected with C. jejuni, in which epithelial TLR9 levels were reduced, were more sensitive to a mild chemical insult, exhibiting significantly worse signs of inflammation. We postulate that these mechanisms may, in part, contribute to the C. jejuni-induced exacerbation of symptoms in IBD patients after a bout of acute bacterial gastroenteritis.

Campylobacter jejuni is the leading cause of gastroenteritis worldwide, and increasing evidence suggests that an acute infection with this common pathogen can initiate the relapse or onset of IBD symptoms (12, 13, 35, 51). Despite its prevalence and links to IBD, relatively little is understood regarding the pathogenesis of C. jejuni infection and the mechanisms by which this pathogen triggers postinfection complications. This is, in part, due to the lack of a suitable murine model that reliably reproduces C. jejuni pathogenesis (54). C. jejuni typically colonizes the murine intestine without inducing an overt inflammatory response (54). The present findings demonstrate an increase in the severity of mild DSS colitis after an acute infection with C. jejuni. Importantly, the low level of DSS used in this study did not induce significant histological damage or weight loss on its own. This suggests that an infection with C. jejuni may not only exacerbate but also initiate the onset of inflammation in response to an otherwise innocuous stimulus.

Increased susceptibility to intestinal inflammation following infection with C. jejuni may be due to a dysregulation in the production of pro- and anti-inflammatory mediators. Members of the Th17 family of cytokines have been linked to the pathogenesis of IBD (10, 21, 45). IL-25 is a member of the IL-17 cytokine family that is reduced in the colonic mucosa of patients with IBD (6). Moreover, treatment with recombinant IL-25 attenuates experimental colitis in mice (6, 32). In the present study, colonic IL-25 levels were significantly lower in DSS-treated mice previously exposed to C. jejuni. Defective bacterial DNA sensing leads to an imbalance in T cell subsets, including a reduction in the IL-17-producing subset of T cells (14). It is possible that C. jejuni infection interferes with commensal bacterial DNA sensing and, by this means, leads to a downregulation in IL-25 expression. Indeed, we observed a notable reduction in epithelial TLR9 fluorescence in C. jejuni-infected mice.

We also observed a reduction in colonic IL-17 levels in the infected mice treated with DSS. Although IL-17 is reported to be upregulated in patients with IBD (10, 45), the exact role of IL-17 in the pathogenesis of IBD remains unclear. For instance, O'Connor and colleagues demonstrated a protective effect for IL-17A in T cell-driven intestinal inflammation (39), and neutralization of IL-17 has been reported to exacerbate DSS-induced colitis in mice (40). IL-17 can also induce the development of intestinal epithelial barrier function in T84 monolayers via an upregulation of claudin-1 and -2 expression (19). More research is needed to determine whether our observation is related to a more acute response that compromises intestinal barrier function, thereby contributing to the increased severity of colitis.

Within a healthy host, recognition of commensal flora by TLRs is essential for protection against gut injury and maintenance of intestinal homeostasis (44). This phenomenon was illustrated in a study demonstrating increased susceptibility to DSS colitis in mice deficient in MyD88, an adaptor molecule crucial for TLR-mediated signaling (44). Previous experiments have demonstrated that TLR9-deficient mice are more susceptible to the development of colitis (26) and that administration of CpG ODN significantly ameliorates chemically induced colonic inflammation in wild-type mice (2, 38, 42, 43).

Significant advances have been made in understanding the pathway by which TLR9 activation limits intestinal inflammation. The polarized expression of TLR9 on the surface of intestinal epithelial cells appears to have an important role in maintaining intestinal homeostasis. TLR9 is expressed on the surface of epithelial cells (8, 26, 48), including both the apical and basolateral membranes of intestinal epithelial cells (26). Interestingly, basolateral, but not apical, activation of TLR9 induces the secretion of CXCL8 (26). We have further characterized the domain-specific response of intestinal epithelial TLR9 and demonstrated that apical activation of this receptor induces an increase in epithelial barrier function whereas basolateral activation had no effect.

An intact intestinal epithelial barrier is required to restrict the translocation of luminal bacteria to the subepithelial compartment (16). Compromised intestinal epithelial barrier function and the translocation of luminal antigens are key features of IBD (16, 33). Therefore, mechanisms that disrupt protective apical TLR9 signaling may destabilize this barrier function and prime the intestine for the development of inflammation. Results from the present study demonstrate that an infection with C. jejuni disrupts TLR9-induced increases in epithelial barrier function. This could then facilitate the translocation of resident bacteria to the subepithelial compartment. Indeed, infection with C. jejuni is reported to increase the translocation of noninvasive bacteria through both the paracellular and transcellular pathways (17, 24).

The attenuation of apical TLR9-induced increases in epithelial barrier function following an infection with C. jejuni appears to be due to a loss of surface TLR9 expression. In the present study, exposure to CpG ODN or C. jejuni alone did not significantly alter TLR9 expression in intestinal epithelial cells. Instead, C. jejuni appears to sensitize intestinal epithelial cells, such that a significant reduction in total and surface TLR9 expression is observed upon the subsequent exposure of the infected cells to a TLR9 agonist. A downregulation of TLR9 expression and function has also been reported in keratinocytes and cervical cancer-derived epithelial cell lines following viral infection (15). Intriguingly, enteric viral infections are also associated with postinfection onset of symptoms in IBD patients (18, 41).

The strengthening of intestinal epithelial barrier integrity may be a fundamental characteristic of epithelial TLR signaling in response to commensal bacteria (3–5). Activation of TLR2 signaling also exhibits an epithelial barrier-protective function in a PI3K/Akt-dependent fashion (4). In contrast to TLR2, the present results indicate that the barrier-protective property of TLR9 is independent of the PI3K/Akt pathway. We also examined the roles of NF-κB and MAPK, which are activated in response to TLR interaction with their respective bacterial ligands. Nevertheless, our results indicate that the TLR9-induced increase in TER does not involve the NF-κB or MAPK signaling pathway.

It is also known that CpG ODN can elicit an antiapoptotic effect associated with a TLR9-induced upregulation of heat shock proteins (22, 23). Heat shock proteins have a cytoprotective role in several cell types, including intestinal epithelial cells (44), and can protect against intestinal epithelial barrier dysfunction (27). Importantly, mice deficient in MyD88 have significantly reduced levels of heat shock proteins within the colonic epithelium, suggesting that the expression of these cytoprotective proteins may be induced directly by TLR signaling in the epithelial cells (44). Whether CpG ODN-TLR9 interaction in intestinal cells induces the upregulation of heat shock proteins, thereby regulating the protection of epithelial barrier integrity, remains to be determined.

Our data also suggest that C. jejuni may sensitize the proinflammatory basolateral TLR9 pathway, which is associated with secretion of CXCL8 (26). Thus, C. jejuni not only attenuates apical TLR9 signaling but also appears to enhance the inflammatory response elicited by the basolateral TLR9 pathway. Although we observed a reduction in surface TLR9, these findings indicate that the loss of TLR9 protein and function may be exclusive to the apical membrane.

In summary, our results show that apical epithelial TLR9 plays an active role in protecting epithelial barrier function. This protective response is lost following exposure to C. jejuni. Furthermore, C. jejuni appears to potentiate the proinflammatory CXCL8 response elicited by basolateral activation of TLR9. Taken together, these responses indicate that a prior infection with C. jejuni disrupts bacterial DNA sensing via TLR9, which has previously been shown to be essential for tolerance to commensal bacteria and the maintenance of intestinal homeostasis. Importantly, we have developed a murine model that can be used to more closely examine mechanisms by which colonization by C. jejuni contributes to the increased risk of exacerbating or initiating symptoms associated with IBD in susceptible individuals.