Shiga Toxin 2 and Flagellin from Shiga-Toxigenic Escherichia coli Superinduce Interleukin-8 through Synergistic Effects on Host Stress-Activated Protein Kinase Activation

Abstract

Shiga toxins expressed in the intestinal lumen during infection with Shiga-toxigenic Escherichia coli must translocate across the epithelium and enter the systemic circulation to cause systemic (pathological) effects, including hemolytic uremic syndrome. The transepithelial migration of polymorphonuclear leukocytes in response to chemokine expression by intestinal epithelial cells is thought to promote uptake of Stx from the intestinal lumen by compromising the epithelial barrier. In the present study, we investigated the hypothesis that flagellin acts in conjunction with Shiga toxin to augment this chemokine expression. We investigated the relative contributions of nuclear factor κB (NF-κB) and mitogen-activated protein kinase (MAPK) signaling to transcription and translation of interleukin-8. Using reporter gene constructs, we showed that flagellin-mediated interleukin-8 gene transcription is heavily dependent on both NF-κB and extracellular signal-regulated kinase 1 and 2 (ERK-1/2) activation. In contrast, inhibition of p38 has no detectable effect on interleukin-8 gene transcription, even though flagellin-mediated activation of host p38 is critical for maximal interleukin-8 protein expression. Inhibition of MAPK-interacting kinase 1 suggests that p38 signaling affects the posttranscriptional regulation of interleukin-8 protein expression induced by flagellin. Cotreatment with Stx2 and flagellin results in a synergistic upregulation of c-Jun N-terminal protein kinases (JNKs), p38 activation, and a superinduction of interleukin-8 mRNA. This synergism was also evident at the protein level, with increased interleukin-8 protein detectable following cotreatment with flagellin and Stx2. We propose that flagellin, in conjunction with Shiga toxin, synergistically upregulates stress-activated protein kinases, resulting in superinduction of interleukin-8 and, ultimately, absorption of Stx into the systemic circulation.

Shiga-toxigenic Escherichia coli (STEC) strains are a major cause of severe gastrointestinal disease in humans, as well as of hemolytic uremic syndrome (HUS), a life-threatening sequela characterized by a clinical triad of acute renal failure, microangiopathic hemolytic anemia, and thrombocytopoenia (34, 51). Many STEC strains associated with serious pathology in humans form attaching and effacing (A/E) lesions on enterocytes, a property mediated by the locus of enterocyte effacement (LEE) pathogenicity island (7). It is thought that these lesions contribute significantly to disease (44, 45, 64). However, some LEE-negative strains are also responsible for cases of severe disease, including HUS (50, 51). Aside from Shiga toxin (Stx) production, the virulence factors important in pathogenicity of LEE-negative strains are undefined.

STEC strains are generally noninvasive pathogens. After ingestion and establishment of intestinal colonization, they release Stx into the gut lumen. When the toxin crosses the gut epithelium and enters the circulation, it targets host cells that express the glycolipid receptor globotriaosylceramide (Gb₃) or globotetraosylceramide (Gb₄) (16, 31, 37, 38, 57, 70). The penetration of Stx into underlying tissues is a crucial step for the development of HUS, and polymorphonuclear leukocytes (PMNs) are thought to be important in this process (28, 65, 66). A high peripheral PMN count during STEC infection in humans is a predictor of severe disease, including the progression to HUS and death (8, 30, 71). CXC chemokines are potent PMN attractants and may therefore be important in PMN migration during human infection. In a previous study, we have shown that STEC induces high levels of CXC chemokines, including interleukin-8 (IL-8) and macrophage inflammatory protein-2α (MIP-2α or GRO-β), in an intestinal epithelial cell (IEC) line (54). This response was more intense and rapid when using STEC strains lacking LEE, and it was mediated largely by a flagellin (FliC) of the H21 serotype.

Flagellin is recognized by Toll-like receptor 5 (TLR5) (26). TLR signaling through the adaptor protein MyD88 can result in the activation of both the mitogen-activated protein kinases (MAPKs) and NF-κB (1-4) and may also lead to the upregulation of receptors on endothelial cells (42) that are important for the recruitment and migration of PMNs (48). In addition, flagellins from LEE-negative STEC, including that of the strain, 98NK2, used in this study, have been shown to promote STEC invasion of tissue culture cells (40, 41, 55). The MAPKs include extracellular signal-regulated kinase 1 (ERK-1), ERK-2, and the stress-activated protein kinases (SAPKs) p38, c-Jun N-terminal protein kinase 1 (JNK-1), and JNK-2. Several bacterial pathogens, including STEC, enteroinvasive E. coli (EIEC), and enteropathogenic E. coli (EPEC), have been shown to activate MAPKs and/or the transcription factors NF-κB and activator protein 1 (AP-1) in vitro (14, 15, 19, 58). Flagellins purified from EPEC and Salmonella spp. have been shown to activate MAPK and NF-κB pathways (17, 18, 23, 74-76). Therefore, in inflammatory diarrheal infections, flagellins probably play an important role in the upregulation of chemokines via these pathways.

Stx1 and Stx2 are also capable of inducing MAPK activation in vitro, but a consensus regarding the role of NF-κB in Stx-induced inflammation has not been reached (5, 9, 21, 29, 56, 63, 67, 77). Even though Stxs are potent protein synthesis inhibitors, both Stx1 and Stx2 have been shown to upregulate IL-8 protein expression in IECs via the ribotoxic stress response, a signal transduction event resulting from specific damage to the 28S rRNA that causes activation of host SAPKs (63). Stx1 has also been shown to superinduce the mRNAs for other CXC chemokines, including IL-8, epithelium-derived neutrophil-activating peptide 78 (ENA-78), and growth-regulated oncogene alpha (GRO-α) (67, 68). Expression of these chemokines is ERK and p38 dependent (unpublished data) and is mediated in part through enhanced mRNA stability (67, 68).

Humans infected with STEC show fecal leukocyte counts similar to those observed in invasive bacterial diarrheas, suggesting that Stxs may augment the proinflammatory response in the gut induced by the bacteria during human infection (27, 62). Furthermore, animal models demonstrate that during oral challenge with Stx-producing organisms, Stx production contributes to the intestinal immune responses (53, 61). While two or more different stimuli could result in additive or synergistic signaling, previous studies have not assessed the contribution of Stx2 to the flagellin-induced immune response in the intestine. In this work, we investigated the proinflammatory role of the recombinant flagellin from the STEC strain 98NK2 (rFliC_H21). Whereas previous studies (74) have indicated the relative importance of posttranscriptional mechanisms to flagellin-induced IL-8, we now provide data describing the relative contributions of MAPK and NF-κB activation to flagellin-induced IL-8 expression and thereby provide mechanistic insight into the relative contributions of these signaling pathways to transcriptional and translational activation of IL-8. We also evaluate for the first time the combined effects of Stx2 and flagellin on signal transduction and IL-8 production, thereby demonstrating how protein synthesis inhibitors such as Stx2 may act as immunomodulators. Finally, we compare flagellins of serotypes H21, H7, H8, H9, and H11 to determine if they differ in their ability to stimulate IL-8 secretion.

MATERIALS AND METHODS

Bacterial strains and reagents.

STEC strain 98NK2 (O113:H21) was isolated from a patient with HUS at the Women's and Children's Hospital, South Australia, as previously described (50). A flagellin deletion mutant of 98NK2 (98NK2 ΔfliC) has also been described (54, 55). E. coli strains were routinely cultured in Luria-Bertani broth with appropriate antibiotics as needed. Both p38 and JNK immunoprecipitation kinase assay kits were purchased from Cell Signaling (Beverly, MA). Dithiothreitol (DTT), β-glycerol phosphate, leupeptin, phenylmethylsulfonyl fluoride (PMSF), and sodium orthovanadate were obtained from Sigma Chemical Co. (St. Louis, MO.). The p38 inhibitor SB203580 (SB) and the MEK1 and -2 inhibitor PD98059 (PD) were obtained from Calbiochem (La Jolla, CA). The JNK-1 and -2 inhibitor SP600125 (SP) was obtained from A. G. Scientific, Inc. (San Diego, CA). The Mnk1 inhibitor CGP57380 was the kind gift of Hermann Gram at Novartis Pharmaceutical (Switzerland).

Cloning and expression of H7, H8, H9, H11, and H21 flagellins.

Purification of the recombinant flagellins was carried out using the QIAexpressionist His₆ fusion protein system (Qiagen GmbH, Germany). In order to construct an N-terminal six-histidine-tagged flagellin fusion protein, the genes encoding flagellin (fliC in E. coli) were amplified by high-fidelity PCR (Roche Diagnostics Australia Pty. Ltd., Sydney, New South Wales, Australia) from chromosomal DNA using the primers TR45 (5′-ACATGCGCATGCATGGCACAAGTCATTAATACCCAAC-3′) and TR46 (5′-CATCGGTCGACCTCTAACCCTGCAGCAG AGACA-3′). The PCR primers incorporated the SphI and SalI sites (underlined), enabling the resultant product to be cloned between the SphI and SalI sites of pQE30 (Qiagen GmbH). Correct insertion of the fliC fragment was confirmed by sequence analysis. The recombinant plasmid was transformed into E. coli M15(pREP4). For purification of recombinant FliC, fliC gene expression was induced by the addition of 2 mM IPTG (isopropyl-β-d-thiogalactopyranoside) to log-phase cultures grown in Terrific broth. Following 3 h of induction, cells were washed in phosphate-buffered saline (PBS), resuspended in 10 ml lysis buffer (50 mM sodium phosphate, 300 mM NaCl, 10 mM imidazole, pH 8.0), and lysed using a French pressure cell (SLM Instruments) operated at 12,000 lb/in². The lysate was centrifuged at 8,000 × g for 15 min at 4°C to remove cellular debris. The supernatant was centrifuged at 40,000 × g for 1 h at 4°C and the soluble supernatant used for purification by nickel-nitrilotriacetic acid (Ni-NTA) agarose purification. A 2-ml Ni-NTA column was equilibrated with 20 ml lysis buffer. M15(pQE-30/fliC) lysate supernatant was treated with 5 μg/ml DNase and 10 μg/ml RNase for 10 min on ice and was loaded onto the column at a rate of 15 ml/h. After the entire supernatant had been loaded, the column was washed with 15 ml wash buffer (50 mM sodium phosphate, 300 mM NaCl, 10% [vol/vol] glycerol, pH 6.0). Bound FliC was eluted from the column with a 0 to 500 mM imidazole gradient in wash buffer (50 mM sodium phosphate, 300 mM NaCl, 10% [vol/vol] glycerol, pH 6.0). Fractions containing FliC were dialyzed overnight against PBS and stored in 50% (vol/vol) glycerol at −20°C or at 4°C (if for immediate use). Protein concentrations were determined using the method of Lowry et al. (39).

Cell culture media and conditions for HCT-8 cells.

All tissue culture media and reagents were obtained from Gibco BRL-Life Technologies (Grand Island, NY). Unless otherwise indicated, HCT-8 cells were maintained at 37°C in an atmosphere of 5% CO₂ in RPMI 1640 medium supplemented with 10 mM HEPES, 1 mM sodium pyruvate, 10% heat-inactivated fetal calf serum (FCS), and 50 IU of penicillin and 50 μg of streptomycin per ml and used at 90 to 100% confluence.

Transfection of HCT-8 cells and assays for luciferase activity.

The DNA constructs used in the luciferase expression assays (−133-luc, −50-luc, and NF-κB mut-luc) were previously reported (47, 49). Briefly, −133-luc carries the IL-8 promoter (−133 to +44) linked to the firefly luciferase reporter gene, −50-luc carries the enhancerless IL-8 core promoter (−50 to +44) linked to the firefly luciferase reporter; and NF-κB mut-luc is the same as −133-luc except for three transitions in the NF-κB binding site (−80 to −71) that prevent binding of NF-κB. These constructs were the kind gift of Naofumi Mukaida (Kanazawa University, Japan).

Transfection of HCT-8 cells was performed using the FuGENE 6 transfection reagent (Roche). Cells were plated in a 24-well culture plate at a density of 0.5 × 10⁵ to 1 × 10⁵ cells/ml and incubated overnight under the standard culture conditions described above. Plasmids were transfected together with the control plasmid phRL-TK, which constitutively expresses Renilla luciferase under the control of the herpes simplex virus (HSV) thymidine kinase (TK) promoter (Promega, Madison, WI). Following overnight incubation, transiently transfected cells were treated with either interleukin-1β (IL-1β) at 5 ng/ml, rFliC_H21 at 100 ng/ml, heat-inactivated rFliC_H21 (HI FliC) at 100 ng/ml, or no additives (control) and incubated for 4 h. Cells were lysed and assayed for luciferase activity using the Dual-Glo luciferase assay system (Promega, Madison, WI) and a Wallac Victor₂ luminometer (Perkin-Elmer, Waltham, MA). Firefly luciferase luminescence was normalized to Renilla luciferase luminescence.

To study the effects of MAPK inhibitors on IL-8 transcription, PD98059, SB203580, or SP600125 was dissolved in dimethyl sulfoxide (DMSO) and added separately to cells 30 min prior to addition of 100 ng/ml of rFliC_H21. Cells were incubated for 4 h. A relative response ratio (RRR) was used to compare samples treated with MAPK inhibitors: RRR = (normalized experimental sample − normalized negative control)/(normalized positive control − normalized negative control), where the positive and negative controls were cells pretreated for 30 min with DMSO (at a volume equal to the maximum volume used for inhibitor treatment) followed by incubation for 4 h either untreated (negative control) or with 100 ng/ml rFliC_H21.

RNA extraction and real-time RT-PCR.

RNA was extracted using TRIzol reagent (Life Technologies, Grand Island, NY). RNA was precipitated at −80°C overnight in 1/10 volume of sodium acetate (pH 4.8) and 2 volumes of 100% ethanol, as previously described (54). RNasin RNase inhibitor (Promega, Madison, WI) was added to samples. Contaminating DNA was digested with RQ1 RNase-free DNase followed by DNase stop solution, according to the manufacturer's instructions (Promega, Madison, WI). The absence of DNA contamination in all RNA preparations was confirmed by reverse transcription-PCR (RT-PCR) analysis using primers specific for the gene encoding the housekeeping enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH). For the experiments shown in Fig. 4B and C, RNA was purified from treated cells using the RNeasy mini kit (Qiagen, Valencia, CA). DNA was removed from the RNA preparations using the DNA-free kit (Ambion, Austin, TX), and cDNA was made using the ImProm-II reverse transcription system (Promega, Madison, WI). Levels of chemokine mRNA produced by HCT-8 cells after stimulation with E. coli strains or proteins were determined by quantitative real-time RT-PCR using a one-step Access RT-PCR system (Promega, Madison, WI) with a 1/20,000 dilution of SYBR green I nucleic acid stain (Invitrogen, Carlsbad, CA) and 20 nmol of each oligonucleotide. The quantitative RT-PCR was performed on a Rotorgene RG-2000 cycler (Corbett Research, Mortlake, New South Wales, Australia). Each RNA sample was assayed in triplicate using primers specific for the various chemokine mRNAs (or mRNA for GAPDH as an internal control) as previously described (54). Results were calculated using the comparative cycle threshold (2^−ΔΔCT) method, as previously described (54).

FIG. 4.

Effect of Mnk1 inhibition on eIF4E phosphorylation, IL-8 mRNA, and IL-8 protein induction by flagellin. HCT-8 cells were preincubated with the Mnk1 inhibitor CGP57380 for 30 min or with the CGP57380 diluent DMSO prior to stimulation with 100 ng/ml rFliC_H21 for 1 and 4 h. (A) Whole-cell extracts were prepared and used to perform Western blotting for phosphorylated eIF4E. (B) Northern blot for IL-8 mRNA from cells treated as described above. (C) Real-time PCR for IL-8 mRNA was performed in triplicate using GAPDH as an internal mRNA control. (D) From cells treated as described above, supernatants were collected and assayed for IL-8 by ELISA in duplicate. Differences in IL-8 levels were analyzed by one-way ANOVA, and Bonferroni's multiple-comparison test for CGP57380-treated control cells versus diluent-treated control cells or for rFliC_H21-plus-CGP57380-treated cells versus rFliC_H21-plus-diluent-treated cells was made. *, P < 0.05; NS, not significant.

Northern blotting.

Cells were disrupted using Qiagen shredders, and total RNA was prepared using RNA binding columns (Qiagen, Valencia, CA). For Northern blotting, 10 to 30 μg of total RNA was separated on glyoxyl agarose gels and transferred to nylon membranes (Ambion, Austin, TX). IL-8 and GAPDH probes were used as previously described (67). DNA probes were synthesized by random priming and labeled with [α-³²P]dCTP. Blots were hybridized overnight at 65°C, washed, and detected using phosphate-based hybridization and wash solutions (11).

Indirect immunofluorescence for detection of RelA.

HCT-8 cells were plated in collagen-coated eight-chamber slides at a density of approximately 1 × 10⁴ to 3 × 10⁴ cells/well and allowed to adhere overnight at 37°C in 5% CO₂. The following day, FliC or heat-inactivated FliC was added to a final concentration of 100 ng/ml and incubated for 0, 5, 10, 15, 30, 45, and 60 min. Some chambers had no additives (control). Slides were then washed twice with PBS and fixed by incubating in 4% paraformaldehyde in PBS for 10 min at room temperature. Slides were washed twice more with PBS, extracted with acetone at approximately −20°C for 3 to 5 min, washed twice more with PBS, blocked in 1.5% normal goat serum (NGS) in PBS, and labeled with Oregon Green 488 phalloidin (Molecular Probes, Eugene, OR). The slides were washed twice with PBS, incubated for 1 h with anti-RelA antibody diluted 1:200 in blocking buffer (sc-372; Santa Cruz Biotechnology, Santa Cruz, CA), washed three times in PBS, and then incubated for 1 h with a goat anti-rabbit secondary antibody conjugated to Alexa Fluor 594 (Molecular Probes) diluted 1:1000 in blocking buffer. The slides were then washed three times in PBS, allowed to dry, and then mounted using Vectashield Hard Set (Vector Laboratories, Inc. Burlingame, CA). For DAPI (4′,6′-diamidino-2-phenylindole) staining, mounting was performed in Vectashield mounting medium (Vector Labs, Burlingame, CA). Images were obtained using a Zeiss AxioImager Z.1 microscope (Carl Zeiss Microscopy, Jena, Germany). Images were captured with an IEEE1394 digital charge-coupled-device (CCD) camera (Hamamatsu, Hamamatsu-City, Japan). Pictures showing two different fluorescent labels were created using Volocity software (Improvision Inc., Lexington, MA).

MAPK protein assays.

Cells were plated in 100- by 20-mm tissue culture dishes at approximately 1 × 10⁷ to 1.5 × 10⁷ cells per dish and allowed to adhere overnight. Stimuli were added without changing the medium, and the cells were then incubated at 37°C for various amounts of time as noted. Following treatment and removal of cell culture medium, cell culture plates were transferred to ice, and the plates were washed twice with cold phosphate-buffered saline (PBS). Cold PBS containing 10 μg/ml leupeptin, 1 mM PMSF, and 0.5 mM DTT was added to the cells, which were subsequently scraped off the plates. Cells were pelleted by centrifugation at 4°C, the supernatant was removed, and pelleted cells were resuspended in Triton lysis buffer, consisting of 25 mM HEPES (pH 7.5), 300 mM NaCl, 1.5 mM MgCl₂, 2 mM EDTA, 0.05% Triton X-100, 0.1 mM Na₃VO₄, 20 mM β-glycerol phosphate, 10 μg/ml leupeptin, 1 mM PMSF, and 0.5 mM DTT. The resuspended cells were gently rocked at 4°C for 30 min; cellular debris was removed by centrifugation at 4°C. The supernatants were collected and frozen at −80°C. The concentrations of the cell extracts were determined using the Bio-Rad protein assay reagent according to manufacturer's instructions (Bio-Rad, Hercules, CA).

All protein kinase assays were performed on cell extracts of equal protein concentrations. p38 kinase activity was assessed using the p38 MAPK assay kit from Cell Signaling according to the manufacturer's instructions. Briefly, 200 μg of protein extract was mixed with monoclonal antibody specific for phosphorylated p38 (Thr180/Tyr182) immobilized on Sepharose beads. These mixtures were incubated overnight at 4°C with gentle rocking. Immunoprecipitates were washed twice with the lysis buffer supplied in the kit to which 1 mM PMSF had been added, followed by washing with kinase buffer also supplied in the kit. Immunoprecipitates were resuspended in 50 μl of kinase buffer that was supplemented with 200 μM ATP and 1 μg ATF-2 fusion protein. Kinase reaction mixtures were incubated for 30 min at 30°C, followed by addition of SDS-PAGE sample buffer to terminate the reaction. The samples were analyzed by SDS-PAGE and Western blotting using antibody specific for ATF-2 phosphorylated at Thr 71.

SAPK/JNK activity was assessed using the SAPK/JNK assay kit from Cell Signaling according to the manufacturer's instructions. Procedures were similar to that outlined for p38 activity above, except JNK/SAPK was precipitated from 250 μg of cell extract by adding c-Jun fusion protein immobilized on Sepharose beads. Kinase reactions were performed, followed by Western blotting using antibody specific for Jun phosphorylated at Ser 63. Jun fusion protein substrate can undergo phosphorylation at Ser 63 or at both Ser 63 and Ser 73, yielding two bands of 33 and 35 kDa, respectively. Densitometric analysis of Western blot data was performed using ImageJ 1.38x software (52).

Detection of phosphorylated eIF4E.

Cell extracts were resolved by SDS-PAGE, transferred to membranes by electroblotting, incubated overnight with antibody specific for the phosphorylated form of eukaryotic initiation factor 4E (eIF4E) (Ser 209), and then detected using horseradish peroxidase (HRP)-linked secondary antibody and chemiluminescent substrate. Subsequently, membranes were stripped and reprobed with appropriate antibody to detect total eIF4E. (Both phospho-eIF4E [Ser 209] and total eIF4E antibodies were obtained from Cell Signaling, Beverly, MA.)

IL-8 ELISA.

The levels of IL-8 in culture supernatants (cotreatment studies) were assayed using a commercial sandwich enzyme-linked immunosorbent assay (ELISA) (R&D Systems, Minneapolis, MN) as previously described (54). Briefly, ELISAs were performed using monoclonal mouse anti-human IL-8 antibody at 2 μg/ml for capture and a biotinylated polyclonal anti-human IL-8 antibody at 20 ng/ml for detection in 96-well trays (Maxisorp Nunc-Immuno plates; Nunc, Roskilde, Denmark). The assay was calibrated using recombinant human IL-8 (R&D Systems). The sensitivity limit of the ELISA was 31.25 pg/ml.

Statistical analysis.

Statistical analyses were performed using Prism 3.03 software (GraphPad Software, San Diego, CA). For Fig. 1, significance was compared using a one-way analysis of variance (ANOVA) with Tukey's multiple-comparison test. For Fig. 1, 2, 3, 7, and 8, comparisons between two values were analyzed using Student's t test. Differences were considered significant at P values of <0.05. For Fig. 4, one-way ANOVA with Bonferroni's multiple-comparison test was used because of the low number of replicates (51).

FIG. 1.

Inhibition of the MAPKs (p38, JNK-1 and -2, and ERK-1 and -2) reduces IL-8 secretion by HCT-8 cells in response to rFliC_H21. HCT-8 cells were preincubated for 60 min with one or more of 10 μM SB203580 (SB) (p38 inhibitor), 50 μM PD98059 (PD) (ERK-1 and -2 inhibitor), 10 μM SP600125 (SP) (JNK-1 and -2 inhibitor), or just DMSO (control) before the addition of 100 ng/ml rFliC_H21. Control monolayers were incubated with all three inhibitors but without rFliC_H21 (negative control). After 4 h, supernatants were collected and assayed for IL-8 by ELISA. Data are the means ± standard deviations (SD) from two experiments. Significant differences relative to DMSO- and rFliC_H21-treated cells are indicated as follows: *, P < 0.05; ** P < 0.005; ***, P < 0.001.

FIG. 2.

FliC induces rapid nuclear-to-cytoplasmic translocation of RelA, and FliC-induced NF-κB binding is required for transcription of the IL-8 promoter. (A) HCT-8 cells were stimulated for 30 min with either rFliC_H21 or heat-inactivated rFliC_H21 (HI). An untreated control counterstained with DAPI containing mounting medium to show nuclei is also shown (bottom panel). For each of these, NF-κB was detected using indirect immunofluorescence. Oregon Green-labeled phalloidin was used to better visualize individual cells. (B) HCT-8 cells were transiently cotransfected using one of three IL-8 promoter-firefly luciferase gene constructs as designated, together with a plasmid constitutively expressing Renilla luciferase to control for transfection efficiency. Transiently cotransfected cells were then treated with IL-1β, rFliC_H21 (FliC), or heat-inactivated rFliC_H21 (HI) or were left untreated (control). Cell lysates were harvested after 4 h, and both firefly luciferase and Renilla luciferase levels were measured. Firefly/Renilla luciferase ratios for each sample were obtained, following which the mean and SD (n = 2) of the firefly/Renilla luciferase ratio for each condition was calculated. (C) Transiently cotransfected cell monolayers were pretreated for 30 min with DMSO (positive and negative controls), 10 mM SB203580 (p38 inhibitor), 10 mM SP600125 (JNK inhibitor), or 20 mM PD98059 (ERK inhibitor). Cells were then left untreated or treated with 100 ng/ml rFliC_H21 for 4 h. Data shown are the relative response ratios ± SD (n = 3 for all samples, including controls).

FIG. 3.

Effect of p38 MAPK inhibition on IL-8 mRNA and protein induction by flagellin. HCT-8 cells were stimulated with 100 ng/ml rFliC_H21 in the presence of either 10 μM SB203580 (SB) (p38 inhibitor) or an equal volume of vehicle (DMSO). (A) At 1 or 4 h, total RNA was extracted from cells and IL-8 mRNA was quantitated by real-time RT-PCR. Results are expressed as the fold increase in [IL-8 mRNA] relative to levels at 0 h, and data are shown as the means ± SD for triplicate assays. (B) At 4 h, supernatants were collected and assayed for IL-8 by ELISA. Data shown are the means ± SD from two experiments. Significant differences relative to DMSO- and rFliC_H21-treated cells are indicated as follows: **, P < 0.015.

FIG. 7.

DHP-2 does not inhibit FliC activation of p38 and JNKs. HCT-8 cells were incubated for 1 h with 200 nM DHP-2 (+) or an equal volume of DMSO (−), followed by treatment with 100 ng/ml rFliC_H21 or no treatment for 1 h. (A) Western blot of phospho-ATF-2 (indicative of p38 activity). (B) Western blot of phospho-c-jun (indicative of JNK activity). The relative density of each band was calculated by dividing its density by that of the DMSO/untreated control.

FIG. 8.

Cotreatment with rFliC_H21 and Stx2 causes an increase in IL-8 protein. HCT-8 cells were incubated overnight with 100 ng/ml rFliC_H21, 1 μg/ml Stx2, or both rFliC_H21 and Stx2. IL-8 protein in cell supernatants was measured by ELISA. The results represent the combined data from four experiments. Error bars represent SD. *, P < 0.05.

RESULTS

ERKs, JNKs, and p38 contribute to FliC-induced IL-8 secretion by HCT-8 cells.

Harrison et al. (25) have shown that Vibrio cholerae FlaA induction of IL-8 in the T84 intestinal epithelial cell line requires ERKs, JNKs, and p38 activity. However, a separate study found that IL-8 induction by Salmonella flagellin did not require JNK-1 and -2 signaling (74). Therefore, we investigated the roles of all three of the MAPK pathways in the rFliC_H21-induced IL-8 production by using specific inhibitors to each pathway. Stimulation of HCT-8 cells with rFliC_H21 (Fig. 1) in the presence of MAPK inhibitors resulted in a reduction in induced IL-8, ranging from a 40 to 61% reduction with PD98059 and SB203580, respectively. The combination of any of two inhibitors further decreased the amount of IL-8 secreted, while the combination of inhibitors of all three MAPKs had the greatest effect, resulting in only 2% of the IL-8 secreted by DMSO- and rFliC_H21-treated cells.

NF-κB mediates the FliC upregulation of IL-8 at the transcriptional level.

To demonstrate NF-κB activation by rFliC_H21, we determined the location of RelA in HCT-8 cells at 0, 5, 10, 15, 30, 45, and 60 min after treatment with rFliC_H21. While initially most of the RelA is cytoplasmic, by 30 min posttreatment, RelA has translocated to the nucleus (Fig. 2 A). Heat-inactivated rFliC_H21 does not cause such translocation. To confirm that the FliC-induced NF-κB activation has a direct effect on IL-8 transcription, we transfected HCT-8 cells with plasmids containing luciferase constructs under the regulation of the wild-type IL-8 promoter (−133-luc), the IL-8 promoter with mutations in the NF-κB binding site (NF-κB-mut), and a truncated nonexpressing IL-8 promoter (−50-luc) and measured luciferase activity 4 h after treatment with rFliC_H21 (Fig. 2B). Impairment of NF-κB binding abrogated luciferase activity. These data suggest that activation/translocation of NF-κB is essential to rFliC_H21-induced IL-8 gene transcription.

MAPK pathways contribute to FliC-mediated IL-8 mRNA transcription, but the effect of p38 activation appears to be predominantly posttranscriptional.

Luciferase activity was measured in cells transiently transfected with the −133-luc plasmid and treated with inhibitor PD98059, SB203580, or SP600125 in order to determine the contribution of the MAPK pathways to transcriptional activation of the IL-8 gene in response to rFliC_H21. While inhibition of ERKs and JNKs decreased rFliC_H21-mediated transcription from the −133-luc construct, inhibition of p38 had minimal, if any, impact (Fig. 2C).

Upregulation of chemokines by p38 occurs via a posttranscriptional mechanism, possibly involving the poly(A) tail of the transcript (12, 33). While there is only a modest difference between the levels of IL-8 mRNA in HCT-8 cells stimulated with rFliC_H21 in the presence of the p38 inhibitor SB203580 versus the carrier DMSO (Fig. 3 A), when IL-8 protein was measured (Fig. 3B), inhibition of p38 resulted in a 57% reduction in IL-8 secretion (P < 0.015). We therefore hypothesize that the primary influence of p38 signaling on IL-8 production is posttranscriptional.

FliC-induced IL-8 protein expression but not IL-8 mRNA expression occurs in an MnkI-dependent manner.

Nuclear transcribed mRNAs bear a 5′ cap structure that is recognized by the translation initiation factor eIF4E, which recruits the mRNA to the translation apparatus, allowing potential control of translation by mRNA selection (24). Because eIF4E is present in limited quantities in eukaryotic cells, eIF4E activity can regulate the global translation initiation rate; this activity is tightly controlled by the MAPK signal-integrating kinases Mnk1 and Mnk2 (22, 73). Mnk1 can be phosphorylated, and thus activated, by the MAPK family members ERK-1/2 and p38. Mnk1 in turn phosphorylates Ser 209 of eIF4E, resulting in enhanced translation of 5′-capped mRNAs (59). Since this canonical pathway is active in HCT-8 cells (13), we used the specific Mnk1 inhibitor CGP57380 to assess the role of Mnk1 in rFliC_H21-induced eIF4E phosphorylation. rFliC_H21 (Fig. 4 A) treatment resulted in increased phosphorylation of eIF4E at 1 and 4 h, consistent with our observations that rFliC_H21 induces both p38 and ERKs in these cells. Pretreatment of HCT-8 cells with CGP57380 resulted in decreased phosphorylation of eIF4E at Ser 209 in the presence and absence of rFliC_H21.

We then assessed the effects of CGP57380 on rFliC_H21 IL-8 mRNA and protein expression. Inhibition of Mnk1 with CGP57380 had little or no effect on IL-8 mRNA induced by rFliC_H21 (Fig. 4B and C). Pretreatment of cells with CGP57380 significantly decreased rFliC_H21-induced IL-8 protein levels, while in the absence of any stimulus, CGP57380 had no effect on basal IL-8 secretion by HCT-8 cells (Fig. 4D). These data suggest that a translational activation mechanism involving eIF4E phosphorylation is likely involved in the maximal expression of IL-8 protein following exposure of intestinal epithelial cells to flagellin.

Costimulation with flagellin and Stx2 synergistically induces IL-8 and MIP-2α mRNA accumulation and SAPK activation.

HCT-8 cells were incubated with either 100 ng/ml of rFliC_H21, 1 μg/ml Stx2, both rFliCH₂₁ and Stx2, or heat-inactivated (HI) rFliC_H21 for 1 or 4 h. By Northern blotting, superinduction was observed at 4 h following cotreatment with Stx2 and rFliC_H21 (Fig. 5 A). Using real-time RT-PCR, no differences in IL-8 mRNA levels were found after 1 h in either the presence or absence of Stx2 alone (Fig. 5B). At the same time point, rFliC_H21 induced a 345-fold upregulation of IL-8, with no significant effect caused by the addition of Stx2. However, superinduction of IL-8 mRNA was observed at 4 h in cells treated with both rFliC_H21 andStx2. At this time point, Stx2 alone induced a modest 4-fold upregulation of IL-8 mRNA, and rFliC_H21 induction had decreased to 26-fold upregulation. Strikingly, IL-8 mRNA was upregulated by greater than 3,000-fold following cotreatment with rFliC_H21 and Stx2 (P < 0.0001 compared to the rFliC_H21 alone). Similar results were observed for MIP-2α mRNA, with superinduction also occurring at 4 h with Stx2 and rFliC_H21 (Fig. 5C). Therefore, Stx2 is able to induce and superinduce IL-8 and MIP-2α mRNAs in the presence of flagellin.

FIG. 5.

Superinduction of flagellin-mediated IL-8 mRNA in HCT-8 cells by Stx2. HCT-8 cells were treated with 100 ng/ml rFliC_H21 (FliC) or HI rFliC_H21 (HI FliC), with or without 1 μg/ml Stx2, for 1 or 4 h, and then total RNA was extracted and IL-8 mRNA observed by Northern blotting (A), real time RT-PCR using GAPDH as an internal mRNA control (B), or real-time RT-PCR for MIP-2α (C). The densities of the IL-8 bands were normalized to the respective GAPDH loading controls. The real-time RT-PCR results are expressed as the fold increase in [mRNA] relative to levels at 0 h, and data shown are the means ± SD for triplicate assays. No cell death is observed by 4 h following treatment of HCT-8 cells with 1 μg/ml Stx2.

Costimulation with flagellin and Stx2 synergistically activates SAPKs p38 and JNKs, and only Stx2-induced SAPK activation is on ZAK.

FliC and Stx2 both activate p38 and JNKs, although there are qualitative differences in activation, with FliC-induced SAPK activation being earlier and transient and Stx2-induced activation more delayed but more sustained. As there was a synergistic activation of IL-8 message by the addition of both rFliC_H21 and Stx2, we evaluated the effect of costimulation with rFliC_H21 and Stx2 on MAPK activation and found synergistic activation of p38 and JNKs (Fig. 6A and B). Activation of both SAPKs occurs early, as with rFliC_H21 alone, and is sustained, as with Stx2 alone. However, the level of sustained activation at 4 to 6 h is larger in magnitude following costimulation than that observed with either stimulus alone. We then assessed the effect of 7-[3-fluoro-4-aminophenyl-(4-(2-pyridin-2-yl-5,6-dihydro-4H-pyrrolo[1,2-b]pyrazol-3-yl))]-quinoline (DHP-2), an inhibitor of the ribotoxic stress response MAP3K ZAK (32), on SAPK activation resulting from costimulation with rFliC_H21 and Stx2 (Fig. 6C and D). As expected, DHP-2 almost completely abrogates Stx2-induced SAPK activation and significantly but not completely diminishes the synergistic activation observed with costimulation. Interestingly, DHP-2 does not diminish FliC-induced p38 or JNK activation but instead appears to have the effect of slightly increasing it (Fig. 7).

FIG. 6.

Synergistic upregulation of p38 and JNKs following costimulation with rFliC_H21 and Stx2 that can be partially blocked with the ZAK inhibitor DHP-2. HCT-8 cells were incubated with 100 ng/ml rFliC_H21, 100 ng/ml HI rFliC_H21, 1 μg/ml Stx2, or both 100 ng/ml rFliC_H21 and 1 μg/ml Stx2 or were not treated. (A and C) Western blots of phospho-ATF-2 (indicative of p38 activity). (B and D) Western blots of phospho-c-Jun (indicative of JNKs activity). (A and B) HCT-8 cells were incubated for 30 min, 2 h, 4 h, or 6 h following no treatment or treatment as described above. *, lanes showing synergistic upregulation of JNKs or p38. (C and D) HCT-8 cells were pretreated with 200 nM DHP-2 (+) or equal volume of DMSO (−) for 1 h, following which they were either not treated or treated for 4 h with FliC, Stx2, or Stx2 and FliC.

Costimulation with flagellin and Stx2 results in an increased IL-8 response.

Since cotreatment with rFliC_H21 and Stx2 resulted in synergistic activation of both MAPK signaling and IL-8 message, we investigated whether cotreatment would also result in an increased IL-8 response. Indeed, cotreatment for 24 h (Fig. 8) resulted in a significant increase in IL-8 protein over treatment with either stimulus alone.

Comparison of rFliC_H21 with other flagellin serotypes.

Rogers et al. identified H21 flagellin as a major inflammatory mediator in a strain of E. coli responsible for an outbreak of HUS in 1998 (54). To determine whether or not this H21 flagellin was unique in its ability to act as a proinflammatory agonist, we compared its ability to induce IL-8 with those of four other flagellins of serotypes H7, H8, H9, and H11. No difference between the different flagellins was detectable (Fig. 9 A). Although the protein concentration was predetermined using the method of Lowry et al., we performed SDS-PAGE to ensure that the concentrations of flagellins being used were similar (Fig. 9B). These data suggest that H21 flagellins are not unique in comparison to other E. coli flagellins in their ability to promote IL-8 production.

FIG. 9.

Comparison of recombinant flagellins of the H7, H8, H9, H11, and H21 serotypes. (A) HCT-8 cells were treated with recombinant H7, H8, H9, H11, or H21 flagellins for 4 h. Cell supernatants were assayed for IL-8 by ELISA. No significant difference in IL-8 production was detected between samples. The data are representative of five separate experiments. Error bars indicate SD. (B) SDS-polyacrylamide gel showing 5 μg of each of the recombinant flagellins as determined by the Lowry assay.

DISCUSSION

This is the first time a flagellin has been shown to act with Shiga toxin to synergistically promote inflammatory signaling. Our data provides mechanistic insight showing that the synergistic effects of combined FliC and Shiga toxin treatment occur at the level of the MAPKs, specifically the stress-activated MAPKs (SAPKs) JNKs and p38. As the ZAK-specific inhibitor DHP-2 was unable to block FliC-induced SAPK activation, we conclude that at least two separate pathways converge in our model. Interestingly, inhibition of ZAK appears to slightly potentiate FliC-induced SAPK activation (Fig. 7), and this appears to allow for sustained SAPK activation at 4 h (Fig. 6C and D). As can be seen in Fig. 6A and B, FliC-induced SAPK activation is normally diminished by 4 h. It is therefore reasonable to speculate that there exists some cross talk between ZAK and the SAPK-destined pathway employed by TLR5 signaling.

Previous studies have shown that NF-κB and MAPK signaling are required components of FliC-induced IL-8 production (6, 25, 36, 74, 76). The lack of consensus on the role of each of the three MAPKs (i.e., ERKs, JNKs, and p38) is possibly complicated by variations between different model cell systems. In our system, NF-κB, ERKs, and JNKs appeared to be required for complete FliC-induced transcription of IL-8 message, while the role of p38 was posttranscriptional. Work by Yu et al. also suggested that p38 acted posttranscriptionally in FliC induction of IL-8 (74). Activation of p38 and ERKs promotes phosphorylation of eIF4E via activation of the MAPK signal-integrating kinase Mnk1 (20, 69, 72, 73), and it has been hypothesized that eIF4E phosphorylation results in enhanced initiation of 5′-capped mRNAs (59). We determined not only that H21 flagellin induces phosphorylation of eIF4E but that Mnk1 inhibition decreases rFliC_H21-induced phosphorylation of eIF4E and production of IL-8 protein without influencing the level of rFliC_H21-induced IL-8 mRNA transcript. The Mnk1 pathway may promote preferential translation initiation of flagellin-induced IL-8 mRNA, which probably accounts for some of the observed differences between the effects of SB203580 on flagellin-induced IL-8 mRNA versus IL-8 protein.

Rogers et al. 2003 have shown that flagellins purified from STEC strains of different H serotypes differed in their ability to induce IL-8 message or protein suggesting that these flagellins vary in their ability to act as a TLR5 agonist (54). We compared flagellins of serotypes H7, H8, H9, H11, and H21 all produced in an isogenic background and found no differences in the induction of IL-8 protein. Although far from being a comprehensive comparison of FliC serotypes, these data suggest that a relationship between serotype and virulence may not be attributed to flagellin's ability to act as a TLR5 agonist as much as some other factor.

There is currently some contention concerning the role of Shiga toxins as mediators of intestinal inflammation. Work by Miyamoto et al. (46) has shown that human colon epithelium does not produce Gb3 synthase, the enzyme that makes the Stx receptor Gb3. In addition they have shown that Stxs do not bind to human colon epithelium. In contrast, Malyukova et al. (43) have shown using tissue from patients infected with STEC that E. coli O157:H7 largely colonizes the area comprising the ileocecal valve versus the colon and that crypt and surface intestinal epithelial cells, as well as epithelial cells that have been sloughed into the lumen, contain large amounts of Stx1 and Stx2. Finally, while studies by Miyamoto et al. (46) using Caco-2 cells suggest that there is no epithelial inflammatory response to Stx in the gut, Bellmeyer et al. (5), using the Gb3 negative intestinal cell line T84, have shown that the genes encoding one or more Shiga toxins were required for a full IL-8 response to EHEC. Our data using the HCT-8 model support a role for Shiga toxin as a mediator of intestinal inflammation. Since ZAK is required for transduction of the ribotoxic stress response, it would be interesting to determine whether ZAK is activated in enterocytes from patients infected with STEC.

In conclusion, simultaneous activation of the TLR5 pathway and the ribotoxic stress response results in a synergistic upregulation of MAPKs and IL-8 signaling in HCT-8 cells. Another TLR ligand, lipopolysaccharide (LPS), has been shown to act synergistically with Stx1 to activate p38 and JNKs in the macrophage-like cell line THP-1 (10). In addition, some animal models of HUS require coadministration of LPS and Stx in order to cause or promote illness (35, 60). Since increased inflammation may promote increased absorption of Stx from the intestinal lumen, it is tempting to speculate that synergistic upregulation of inflammatory pathways by one or more TLR ligands together with Stx promotes the onset of systemic disease during STEC infection.