Present address: Division of Pediatric Allergy and Immunology, Mount Sinai School of Medicine, One Gustave L. Levy Place, New York, NY 10029, USA;
Role of EHEC O157:H7 virulence factors in the activation of intestinal epithelial cell NF-κB and MAP kinase pathways and the upregulated expression of interleukin 8
Article first published online: 4 OCT 2002
Volume 4, Issue 10, pages 635–648, October 2002
How to Cite
Berin, M. C., Darfeuille-Michaud, A., Egan, L. J., Miyamoto, Y. and Kagnoff, M. F. (2002), Role of EHEC O157:H7 virulence factors in the activation of intestinal epithelial cell NF-κB and MAP kinase pathways and the upregulated expression of interleukin 8. Cellular Microbiology, 4: 635–648. doi: 10.1046/j.1462-5822.2002.00218.x
- Issue published online: 4 OCT 2002
- Article first published online: 4 OCT 2002
- Received 26 March, 2002; revised 24 June, 2002; accepted 24 June, 2002.
Enterohaemorrhagic Escherichia coli O157:H7 (EHEC) is a gastrointestinal pathogen that is generally non-invasive for intestinal epithelial cells, yet causes acute intestinal inflammation, diarrhoea, haemorrhagic colitis and haemolytic uraemic syndrome. To study signal transduction pathways activated in human intestinal epithelial cells by EHEC, we took advantage of EHEC O157:H7 and isogenic mutants deficient in the major EHEC virulence factors, intimin (eae–) and Shiga toxin (stx–). Infection with wild-type EHEC activated p38 and ERK MAP kinases and the nuclear translocation of the transcription factor NF-κB. Downstream, this was accompanied by increased expression of mRNA and protein for the neutrophil chemoattractant IL-8. Isogenic eae– and stx– mutants also activated p38 and ERK MAP kinases, and NF-κB and stimulated increases in IL-8 protein secretion similar to those of wild-type EHEC. Further, inhibition of either p38, ERK or NF-κB activation abrogated the IL-8 response induced by wild-type EHEC and the mutants. Epithelial cell MAP kinase and NF-κB pathways leading to IL-8 secretion were also activated by isolated EHEC H7 flagellin, which was active when added to either the apical or basolateral surface of polarized human intestinal epithelial cells. We conclude that EHEC interacting with intestinal epithelial cells activates intracellular signalling pathways and an epithelial cell proinflammatory response independent of either Shiga toxin or intimin, two of the major known virulence factors of EHEC. The activation of proinflammatory signals in human colon epithelial cells in response to this non-invasive pathogen appears to depend to a significant extent on H7 flagellin.
Epithelial cells that separate the host from intestinal luminal contents can function as sentries for enteropathogenic bacteria and produce signals that alert immune and inflammatory cells in the underlying mucosa to the presence of bacterial infection (Kagnoff and Eckmann, 1997). Enterohaemorrhagic Escherichia coli (EHEC) O157:H7 is a pathogen that causes diarrhoea, and can result in haemorrhagic colitis and the potentially fatal haemolytic uraemic syndrome (HUS) (Besser et al., 1999). In contrast to other adherent strains of E. coli, EHEC can produce Shiga toxins (Stxs), which are thought to be important in the pathogenesis of EHEC-induced renal disease and vascular complications of haemorrhagic colitis by direct cytotoxic effects on microvascular cells of the kidney and colonic mucosa (Obrig et al., 1988; Tesh et al., 1991).
Stxs are AB5 toxins that contain an enzymatically active A subunit combined with a pentamer of B subunits. The majority of EHEC O157:H7 isolates produce Stx-2 and some produce Stx-1 in addition to Stx-2 (Law, 2000). Although the role of Stxs in the development of EHEC-induced colonic inflammation has not been defined, purified Stxs are known to stimulate low level production of the neutrophil chemoattractant IL-8 by human intestinal epithelial cells (Thorpe et al., 1999; 2001). However, it is not known whether Stx is required for the activation of IL-8 by epithelial cells during infection with live EHEC.
EHEC expresses the bacterial surface protein intimin, encoded by the eae gene (Beebakhee et al., 1992). Intimin is a key virulence factor for EHEC, and plays an essential role in formation of the characteristic EHEC attaching and effacing (A/E) lesion (Yu and Kaper, 1992) and in establishing a pathway through which EHEC can signal into target cells after binding to its receptor Tir, which EHEC inserts into the target cell membrane (DeVinney et al., 1999). Although EHEC are related to enteropathogenic E. coli (EPEC) that also induce A/E lesions but do not produce Stxs, recent studies indicate that signal transduction pathways initiated by EPEC and the O157:H7 strain of EHEC diverge downstream of Tir (DeVinney et al., 2001).
The interaction of EHEC with the apical surface of intestinal epithelial cells in human colon is an important step in the pathogenesis of infection. EHEC generally does not appear to be invasive for human intestinal epithelium, yet initial signals between the pathogen and the underlying mucosa are putatively relayed through the epithelium. It is now known that a number of enteric bacterial pathogens that have varying degrees of invasiveness (e.g. Salmonella, Shigella, Listeria, EPEC, Helicobacter pylori) can signal human intestinal or gastric epithelial cells to produce chemokines and proinflammatory mediators thought to be important in initiating the host mucosal inflammatory response (Eckmann et al., 1993; Crabtree et al., 1994; Jung et al., 1995; Savkovic et al., 1996).
Many of the chemokines, cytokines, and innate defence molecules produced by intestinal epithelial cells in re-sponse to bacterial infection are target genes of the transcription factor, nuclear factor (NF)-κB (Hobbie et al., 1997; Savkovic et al., 1997; Elewaut et al., 1999; O’Neil et al., 1999; Philpott et al., 2000). In this regard, NF-κB appears to be a central regulator of the epithelial cell innate immune response to infection by a spectrum of bacteria that use different strategies to adhere to and invade intestinal epithelial cells (Elewaut et al., 1999). Bacteria and bacterial products may signal the activation of NF-κB through pathways that include extracellular receptors [e.g. Toll-like receptors (TLRs)] or after bacterial invasion, possibly by activating intracellular receptors (e.g. Nod1/CARD4) that signal the activation of the inhibitory κB kinase (IKK) complex (Chow et al., 1999; Schwandner et al., 1999; Girardin et al., 2001; Hayashi et al., 2001; Inohara et al., 2001). In addition, mitogen activated protein (MAP) kinases, including p38 MAP kinases, extracellular signal-regulated kinases (ERK), and c-Jun N-terminal kinases (JNK) have been noted to play a role in the upregulated expression of IL-8 in epithelial cells in response to infection with bacterial pathogens (Tang et al., 1994; Hobbie et al., 1997; Keates et al., 1999). Moreover, activation of the same receptors that lead to NF-κB activation in response to bacterial infection has also been shown to result in MAP kinase activation (Akira et al., 2001; Girardin et al., 2001).
In the present study we report that EHEC infection of human intestinal epithelial cells activates MAP kinase (p38 and ERK 1/2) and NF-κB signalling pathways, both of which are required for the induction of an intestinal epithelial cell proinflammatory response. Moreover, we demonstrate that the major EHEC virulence factors Stx and intimin are not required for activation of those signalling pathways and the epithelial cell proinflammatory response in the context of infection with EHEC. In contrast, H7 flagellin appears to be a key virulence factor for the upregulated expression and production of IL-8 by human intestinal epithelial cells infected with EHEC.
EHEC infection increases IL-8 mRNA expression and IL-8 secretion by Caco-2 cells independent of Stx or intimin
We first determined the requirement for Stx in EHEC-induced upregulation of IL-8 production by human intestinal epithelial cells. Caco-2 cells were infected with either wild-type EHEC 86–24 (producing Stx-2), an isogenic stx– mutant that does not produce Stx, or were stimulated with purified Stx-2. Wild-type EHEC induced a greater than 250-fold increase in IL-8 mRNA after 3 h of infection (Fig. 1A) whereas the isogenic stx– EHEC induced a 100–150-fold increase in IL-8 mRNA. As the response to the Stx-2 null mutant was less than that induced by the wild-type EHEC (P < 0.05), we assessed whether purified Stx-2 at concentrations spanning those produced by wild-type EHEC could explain the difference. As shown in Fig. 1A, purified Stx at a dose of 10 µg ml−1 (100-fold higher than that released by wild-type EHEC at an MOI of 10) induced only a 6.7-fold increase in IL-8 mRNA levels (Fig. 1A), while lower concentrations of Stx between 0.01 and 1 µg ml−1 induced lower levels of IL-8 mRNA (data not shown). Despite the lower magnitude of IL-8 mRNA upregulation in response to the stx– EHEC mutant, IL-8 protein levels did not differ significantly between wild-type and stx– EHEC infected cells (Fig. 1B). Thus, the majority of the increase in IL-8 after infection with wild-type EHEC is dependent on the presence of the bacteria or bacterial products other than Stx that are produced by EHEC.
Infection of intestinal epithelial cells with wild-type EHEC results in a characteristic A/E lesion that is dependent on the eae gene and its product, intimin. To determine if the increase in IL-8 expression after EHEC infection required intimin, Caco-2 cells were infected with an intimin-deficient (eae–) isogenic mutant of 86–24. Infection with either the eae– mutant or wild-type EHEC resulted in an increase in IL-8 mRNA of a similar magnitude (Fig. 1A), indicating that intimin was not required for the upregulated expression of IL-8 mRNA. Further, IL-8 protein secretion was similar in cells infected with wild-type EHEC or the eae– mutant (Fig. 1B). In contrast to EHEC and its stx– and intimin-negative mutants, non-pathogenic DH5αE. coli upregulated IL-8 mRNA levels only ninefold over control. Similar data for IL-8 protein and mRNA responses to infection with EHEC and the EHEC mutants were obtained after infection of HT-29, another human colon epithelial cell line (data not shown). Thus, intestinal epithelial cells respond to EHEC infection with a rapid upregulation of IL-8 mRNA and release of IL-8, which does not require Stx or intimin.
EHEC activates NF-κB in human intestinal epithelial cells
NF-κB has a central role in the activation of several proinflammatory genes, including IL-8, in human intestinal epithelial cells in response to bacterial infection (Keates et al., 1997; Savkovic et al., 1997; Elewaut et al., 1999). To determine if EHEC infection of human colon epithelial cells activates NF-κΒ, Caco-2 cells were infected with wild-type EHEC for up to 3 h and NF-κΒ activation was assessed by electrophoretic mobility shift assay (EMSA) (Fig. 2). As shown in Fig. 2A, NF-κB binding activity increased by 1 h after infection and, in the continued presence of bacteria, remained increased at 3 h. However, the activation of NF-κB was transient when bacteria were not continuously present. As shown in Fig. 2B, removal of infecting bacteria after 2 h of infection resulted in a return of binding activity to baseline within an additional 2 h period, whereas in the continued presence of bacteria, binding activity remained persistently increased (Fig. 2B). Infection with the stx– or eae– isogenic mutants resulted in NF-κB activation similar to that seen with the wild-type EHEC indicating that neither Stx nor intimin were required for the activation of NF-κB by EHEC (Fig. 2C). In contrast, non-pathogenic E. coli DH5α resulted in little NF-κB activation (Fig. 2C). Supershift assays were done to determine which NF-κB subunits were activated after EHEC infection. As shown in Fig. 2D, incubation with an antibody against the p50 subunit induced a shift in both bands, whereas incubation with the anti-p65 antibody preferentially induced a shift in the upper band, indicating that the NF-κB binding complex induced by EHEC likely contains both p50 homodimers and p50/p65 heterodimers.
EHEC activates MAP kinase signalling pathways in human intestinal epithelial cells
S. typhimurium, some strains of H. pylori, and EPEC have been reported to increase IL-8 production by epithelial cells by activating one or more of the MAP kinase signalling pathways (Hobbie et al., 1997; Keates et al., 1999; Savkovic et al., 2001). To determine if those pathways are activated by EHEC infection, Caco-2 cells were infected with EHEC for up to 3 h after which p38 and ERK 1/2 phosphorylation was assessed. As shown in Fig. 3, increased phosphorylation of p38 MAP kinase occurred within 1–2 h of infection with wild-type EHEC. As Stx-2 had been reported to activate p38 (Thorpe et al., 1999), we asked whether p38 phosphorylation required Stx-2. As shown in Fig. 3, p38 phosphorylation also occurred in Caco-2 cells infected with the stx– mutant, and followed an identical time course and magnitude as that seen after infection with wild-type EHEC. Moreover, p38 phosphorylation also was independent of the bacterial expression of intimin, as similar p38 phosphorylation was seen in response to infection with wild-type and eae– EHEC.
Also shown in Fig. 3, EHEC infection resulted in the phosphorylation of ERK 1/2. As with p38 phosphorylation, the phosphorylation of ERK 1/2 after EHEC infection was independent of stx or eae gene expression. However, in contrast to p38, phosphorylated ERK1/2 levels decreased markedly by 3 h post infection.
Both MAP kinases and NF-κB are required for EHEC-induced IL-8 secretion
The phosphorylation of p38 and ERK 1/2 MAP kinases in response to EHEC infection suggested those kinases might be important for the upregulation of IL-8 expression in Caco-2 cells. To test this possibility, Caco-2 cells were pretreated with specific inhibitors of p38 (SB203580, 10 µM) or MEK, the upstream kinase that activates ERK 1/2 (PD98059, 50 µM), before EHEC infection. Inhibition of p38 with SB203580 resulted in a significantly diminished IL-8 response to EHEC infection, which was observed with both wild-type EHEC (Fig. 4) and the stx– and eae– mutants (data not shown). In addition, MEK inhibition inhibited IL-8 secretion to a similar extent to that seen after p38 inhibition. Inhibition of both p38 and MEK resulted in more marked inhibition of the IL-8 response than that noted with each inhibitor used separately.
To determine if NF-κB activation also was essential for the upregulated expression and production of IL-8 in response to EHEC infection, activation of NF-κB in Caco-2 cells was inhibited using a recombinant adenovirus that expresses a mutant IκBα protein, which acts as a super-repressor of NF-κB activation. Caco-2 cells were infected with the recombinant adenovirus that expresses the super-repressor (Ad5IκB-A32/36) or, as a control, with an adenovirus expressing β-galactosidase (Ad5LacZ) prior to EHEC infection. Three hours after EHEC infection, RNA was isolated for quantification of IL-8 mRNA. As shown in Fig. 5, in Caco-2 cells infected with the control Ad5LacZ, EHEC infection resulted in a 180-fold increase in IL-8 mRNA. In contrast, IL-8 mRNA levels were significantly, but not completely, inhibited in cells infected with the Ad5IκB-A32/36 super-repressor before stimulation with wild-type EHEC, or the stx– and eae– mutants. Because we determined that only 50% of Caco-2 cells were infected with the adenovirus, parallel experiments were performed using HT-29 cells that were>80% infected with the adenovirus under similar conditions. Control HT-29 cells also responded to EHEC infection with an upregulation of IL-8 mRNA (20-fold increase), and this was completely abolished in cells infected with the Ad5IκB-A32/36 super-repressor. These results indicate that NF-κB is essential for EHEC-induced upregulation of intestinal epithelial cell IL-8.
Because blocking either the MAP kinases or NF-κB activation inhibited the EHEC-induced IL-8 response, we sought to determine if blocking the MAP kinases prevented NF-κB activation. In this regard, blocking MEK had been recently reported to abrogate IκBα degradation in cells infected with a closely related bacterial pathogen, EPEC (Savkovic et al., 2001). We assessed if the p38 and MEK MAP kinase inhibitors blocked the activation of NF-κB, as assessed both by EMSA and IκBα degradation. As shown in Fig. 6A and B, neither NF-κB binding activity nor IκBα degradation in response to EHEC infection was affected by MAP kinase inhibition. Thus, NF-κB activation in response to EHEC occurs independent of the activation of the MAP kinases that we show also play a significant role in regulating intestinal epithelial cell IL-8 after EHEC infection.
EHEC releases a factor that activates intestinal epithelial cells
We next sought to determine if direct EHEC-intestinal epithelial cell contact is required for EHEC to activate NF-κB and MAP kinase pathways and the IL-8 response, or if EHEC releases a factor that can activate MAP kinases and NF-κB. For these studies, Caco-2 cells were incubated for 3 h with supernatants derived from bacterial cultures of wild-type and stx– EHEC. As shown in Fig. 7A, incubation of Caco-2 cells with a 1:10 dilution of EHEC supernatant prepared from an overnight bacterial culture of either wild-type or stx– EHEC, but not non-pathogenic E. coli DH5α, induced a significant increase in IL-8 secretion. To determine if the bacterial culture supernatant activated the same signalling pathways shown to be activated by EHEC infection, NF-κB and MAP kinase activation were assessed in supernatant-treated Caco-2 cells. The supernatants resulted in an increase in NF-κB binding as shown by EMSA (Fig. 7B). Moreover, NF-κB activation was similar in cells incubated with supernatants from the wild-type or stx– EHEC. In addition, treatment of Caco-2 cells with supernatant from either wild-type or stx– EHEC, but not supernatant from E. coli DH5α, resulted in both p38 and ERK1/2 phosphorylation (Fig. 7C). These results demonstrated that a soluble factor(s) produced and released by EHEC can upregulate IL-8 production by human intestinal epithelial cells. Moreover the soluble factor(s) uses the same pathways we showed to be used by EHEC to upregulate IL-8 production.
H7 flagellin is the soluble factor in EHEC supernatants that upregulates epithelial cell IL-8 production
We previously showed that bacterial LPS does not induce IL-8 production in Caco-2 cells (Yang et al., 1997). However, flagellin from enteroaggregative E. coli (EAEC) and from S. typhimurium was recently reported to induce IL-8 secretion by intestinal epithelial cells (Steiner et al., 2000; Eaves-Pyles et al., 2001; Gewirtz et al., 2001a). To test the possibility that EHEC H7 flagellin was the significant IL-8-inducing factor in EHEC culture supernatants, we initially demonstrated the presence of H7 flagellin in the supernatants by immunoblotting. As shown in Fig. 8A, EHEC culture supernatants contain H7 flagellin, with the expected apparent molecular weight of approximately 66 kDa (Sherman et al., 1988). Smaller bands were also present, presumably due to proteolytic cleavage of the flagellin (Sherman et al., 1988). To determine if H7 flagellin can induce IL-8 expression by intestinal epithelial cells, Caco-2 cells were stimulated with flagellin isolated from the EHEC stx– mutant. As shown in Fig. 8B, flagellin induced a dose-dependent increase in IL-8 secretion. Moreover, removal of H7 flagellin from EHEC culture supernatants by immunodepletion almost completely inhibited the increase in IL-8 secretion in response to those supernatants (Fig. 8C). Consistent with this, an isogenic Stx–, fliC– mutant of EHEC 86–24 that lacks both Stx2 and flagellin completely failed to stimulate IL-8 secretion (IL-8 in the presence and absence of the mutant was 0.29 ng ml−1 and 0.24 ng ml−1 respectively).
Polarized intestinal epithelial cells produce IL-8 in response to apical EHEC infection or apical flagellin
EHEC is generally not invasive for intestinal epithelial cells, and therefore predictably would have its major interaction with intestinal epithelial cells through their apical, rather than basolateral membrane. To determine if the major IL-8 inducing activity of EHEC was mediated from the apical epithelial cell surface, HCA-7 human colonic epithelial cells were grown as a polarized monolayer on transwells and infected apically with wild-type EHEC or the stx– and eae– mutants. We used HCA-7 cells (Kirkland, 1985) for these studies as HCA-7 cells in our laboratory consistently form functional polarized monolayers with a trans-epithelial resistance that is greater than that of Caco-2 cells (i.e. approximately 500 Ω·cm2), do not respond to bacterial LPS stimulation, and like other polarized model intestinal epithelia (e.g. T84 cells) secrete IL-8 basolaterally (Kirkland, 1985 and Kagnoff, M.F., unpublished data). As shown in Fig. 9A, the wild-type and mutant EHEC strains induced IL-8 when added to the apical compartment of the polarized cells. Moreover, in each case, greater than 90% of the IL-8 was secreted into the basolateral media. Whereas the wild-type EHEC and the eae– mutant induced similar levels of IL-8, the stx– mutant induced even greater levels of IL-8, perhaps caused by the known ability of Stx to inhibit protein synthesis (Sandvig and van Deurs, 1996). The non-pathogenic E. coli O111:B4 that has a similar growth rate to the EHEC strains used herein, but lacks flagella, did not induce IL-8 secretion at any time tested after infection.
To determine if polarized intestinal epithelial cells could respond to soluble EHEC H7 flagellin presented to the apical (luminal) side (i.e. the normal site of epithelial cell contact with EHEC), H7 flagellin was added to the apical or, alternatively to the basolateral chamber, of HCA-7 cells grown as polarized monolayers. As shown in Fig. 9B, maximal stimulatory doses of flagellin added to either the apical or basolateral side of HCA-7 cells induced similar levels of basolateral IL-8 secretion. To exclude the possibility that some apically added flagellin was simply gaining access to the basolateral chamber through leaky tight junctions between epithelial cells and then stimulating IL-8 from the basolateral surface, we added submaximal stimulating doses of flagellin (3 and 10 µl) to either the apical or basolateral transwell chambers and assayed IL-8 production at earlier times (i.e. 2 and 4 h later). For each dose and duration of culture, flagellin added to the apical chamber stimulated two–threefold or more basolateral IL-8 production than the equivalent amount of flagellin added to the basolateral chamber (data not shown).
EHEC O157:H7 is not invasive for human colon epithelial cells, yet causes acute intestinal inflammation and haemorrhagic colitis. Although EHEC is a significant human enteropathogen, there is a paucity of information on intestinal epithelial cell proinflammatory responses to infection with EHEC O157:H7. The studies herein investigated epithelial cell signalling pathways and EHEC O157:H7 virulence factors that contribute to the intestinal epithelial cell proinflammatory response to this bacterial pathogen. EHEC O157:H7 is shown to activate MAP kinase signalling pathways and the transcription factor NF-κB, and both are required for the upregulated mRNA expression and secretion of the neutrophil chemoattractant IL-8 by human intestinal epithelial cells.
These studies focused on major known E. coli O157:H7 virulence factors. Stx is a virulence factor that distinguishes EHECs from other enteroadherent disease-causing E. coli (e.g. EPEC, EAEC). However, E. coli O157:H7 lacking Stx was a potent activator of p38 and ERK1/2 MAP kinases and NF-κB, and markedly upregulated epithelial cell IL-8 production. We observed, as others have reported (Thorpe et al., 1999), that purified Stx alone can induce low levels of IL-8 production by human intestinal epithelial cells. However, our studies clearly demonstrate for the first time that Stx is not required for a robust epithelial IL-8 response to infection with E. coli O157:H7. Previous in vivo studies have demonstrated that rabbits infected with O157:H7 develop diarrhoea and changes in colonic solute transport that are dependent on neutrophil recruitment (Elliott et al., 1994), and furthermore, these changes are independent of Stx expression (Li et al., 1993). These findings could potentially be explained by our results, which demonstrate that the neutrophil chemoattractant IL-8 is upregulated in epithelial cells by EHEC-derived factors other than Stx.
The A/E lesion is a characteristic feature of EHEC infection and is associated with rearrangements of the actin cytoskeleton resulting in the formation of a pedestal-like structure that supports the bacteria. Formation of this lesion requires the expression of bacterial intimin, which is encoded by the eae gene (Yu and Kaper, 1992). Intimin binds to its receptor, Tir, which is translocated from the bacteria to the host membrane and interacts with host cell proteins forming a pathway for the transduction of signals from the bacteria to the epithelial cell (reviewed in Donnenberg and Whittam, 2001). We used an isogenic eae– mutant of E. coli O157:H7 to determine if intimate bacterial attachment was required for activating the p38 and ERK 1/2 MAP kinase pathways, NF-κB, and the ensuing production of IL-8 in human colon epithelial cells. Infection with the eae– mutant resulted in approximately the same level of IL-8 expression and secretion as infection with the wild-type strain, indicating that intimin and the formation of the A/E lesion are not required for inducing the expression and secretion of IL-8 by colon epithelial cells. Others recently reported that shiga toxin producing E. coli (STEC) that do not induce A/E lesions can induce IL-8 production from intestinal epithelial cells (Hurley et al., 2001). However, they did not directly address the role of intimin in the intestinal epithelial cell proinflammatory response to EHEC infection through the use of isogenic mutants. Consistent with our results, enteroaggregative E. coli (EAEC), which also causes intestinal inflammation and diarrhoea in humans but does not induce an attaching and effacing lesion, also can upregulate epithelial cell IL-8 production (Steiner et al., 1998).
Epithelial cells can respond to invasive bacterial infection with the expression of an inflammatory phenotype that includes the production of chemokines (Yang et al., 1997; Philpott et al., 2000; Izadpanah et al., 2001), expression of adhesion molecules (Huang et al., 1996), nitric oxide production (Witthoft et al., 1998), and the release of antimicrobial peptides (O’Neil et al., 1999). The transcription factor NF-κB has been shown to be a central regulator of this inflammatory phenotype (Elewaut et al., 1999; Philpott et al., 2000). To determine if EHEC, a non-invasive pathogen, also induced NF-κB activation, we examined NF-κB nuclear binding activity by EMSA after EHEC infection. In concordance with our observations of IL-8 secretion, NF-κB was activated by the wild-type as well as the Stx– and eae– isogenic mutants. Moreover, inhibition of NF-κB activation abrogated EHEC-induced IL-8 expression indicating that NF-κB has a central role in the epithelial cell response to EHEC infection.
Infection with wild-type EHEC and the Stx– and eae– isogenic mutants also activated signalling in human intestinal epithelial cells through two MAP kinase pathways, p38 and ERK 1/2. Furthermore, specific pharmacological inhibitors of p38 and MEK (the upstream activator of ERK 1/2) inhibited IL-8 production in response to infection with EHEC. Our results indicate that EHEC has in common with several other enteric bacterial pathogens, including Helicobacter pylori, Salmonella typhimurium, Listeria monocytogenes and EPEC, the ability to activate MAP kinase signalling pathways and epithelial cell proinflammatory mediators such as IL-8, independent of invasion of epithelial cells. Nonetheless, the specific mechanisms and pathways used by these bacteria to activate these MAP kinases may differ (Tang et al., 1994; Hobbie et al., 1997; Savkovic et al., 2001).
The IL-8 response to EHEC infection was inhibited either by inhibition of the MAP kinase pathways or by inhibition of NF-κB activation. Others recently reported that MEK is important for activation of NF-κB in response to EPEC infection, as pharmacologic inhibition of MEK inhibited the degradation of IκBα after EPEC infection (Savkovic et al., 2001). In our studies, and those of others studying H. pylori infection of gastric epithelium (Keates et al., 1999), inhibitors of p38 and MEK did not prevent IκBα degradation or inhibit NF-κB DNA binding as assessed by gel shift mobility assays. Thus, the NF-κB and MAP kinase signalling pathways that result in upregulated IL-8 mRNA and protein in EHEC infected cells appear to function independently. It is possible that NF-κB activates IL-8 gene transcription, whereas MAP kinase activation increases production by stabilization of IL-8 mRNA (Holtmann et al., 1999).
H7 flagellin in supernatants generated from wild-type and the stx– mutant EHEC were potent inducers of epithelial cell IL-8 production. Bacterial supernatants were able to reproduce the findings with intact bacteria with respect to activation of the MAP kinase pathways and NF-κB, depletion of flagellin from bacterial supernatants completely abolished their IL-8 stimulating activity, and a double stx–, fliC- mutant completely failed to stimulate IL-8 production. Consistent with our findings, flagellin recently was reported to be the major IL-8 inducing factor produced by EAEC (Steiner et al., 2000). Moreover, previous studies have indicated the importance of flagellin for the upregulated production of IL-8 by bronchial epithelial cells in response to infection with Pseudomonas aeruginosa (DiMango et al., 1995) and flagellin was recently shown to be key for the activation of NF-κB and the upregulated production of IL-8 in T84 human colon carcinoma cells infected with S. typhimurium (Gewirtz et al., 2001a).
Using polarized HCA-7 cells as a model epithelium, we found that the addition of H7 flagellin to either the apical or basolateral epithelial compartments resulted in the polarized basolateral secretion of IL-8, and that addition of submaximal doses of H7 flagellin to the apical surface resulted in significantly greater IL-8 responses than addition of the same dose to the basolateral surface. These results are in accord with studies showing that flagellin from Salmonella can activate Caco-2 cells when added apically to polarized monolayers (Sierro et al., 2001), but differ from those reporting that flagellin from Salmonella activated IL-8 expression in T84 human colon epithelial cells only when added to the basolateral compartment (Gewirtz et al., 2001b). As flagellin can activate NF-κB through TLR5 (Hayashi et al., 2001), this suggests both an apical and basolateral distribution of TLR5 on the cell membrane in these models or additional signalling receptors that respond to bacterial flagellin, as has been proposed for flagellin from Pseudomonas (McNamara et al., 2001). It is not known if flagellin can activate epithelial cells in vivo when present in the intestinal lumen. Further, it will be important in future studies to determine whether or not other bacterial virulence factor(s) contribute to the intestinal epithelial cell proinflammatory response to EHEC infection in vivo.
In summary, EHEC infection of human intestinal epithelial cells activates MAP kinase signalling pathways and the transcription factor NF-κΒ, leading to secretion of the neutrophil chemoattractant IL-8 from the basolateral surface of colonic epithelial cells. Activation of these MAP kinase signalling pathways and NF-κΒ does not require the EHEC virulence factors intimin or Stx and appears to depend to a significant extent on an interaction between the epithelial cells and H7 flagellin.
Caco-2 (American type culture collection; ATCC HTB 37), HT-29 (ATCC HTB 38) and HCA-7 human colon carcinoma cells (a gift from S. C. Kirkland) were grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum (FCS), 2 mM glutamine, and 25 mM HEPES (all from Life Technologies, Gaithersburg, MD). For polarized monolayers, HCA-7 cells were seeded on tissue culture-treated transwell filters (0.4 µm pore size, 1 cm2 surface area; Costar, Cambridge, MA) and allowed to grow for ∼7 days, at which time a resistance of 500–700 Ω·cm2 was established.
Bacterial strains and shiga toxin
The bacterial strains used in this study were wild-type EHEC strain 86–24 (serotype O157:H7) that produces Stx-2 and UMD619 which is an isogenic mutant of 86–24 with an inactivated eae gene, referred to herein as eae– (both provided by M. S. Donnenberg, University of Maryland, MD), TUV86-2, an isogenic mutant of 86–24 that does not produce Stx-2 (referred to as stx–) (provided by J. Hacker, Würzburg University, Germany) and 86–24 DM1, an isogenic fliC- mutant of TUV86-2 that does not produce Stx-2 or flagellin (generated by A.D-M, Université d’Auvergne, using a kanamycin cassette). Non-pathogenic E. coli DH5α was from Life Technologies, and non-pathogenic E. coli O111:B4 was provided by D. Guiney (University of California, San Diego, CA) (Heffernan et al., 1992). Purified Stx-2 was from Toxin Technology (Sarasota, FL).
Bacterial infection and preparation of bacterial supernatants
For infection studies, bacteria were grown overnight without agitation at 37°C in Luria–Bertani (LB) broth, and centrifuged and resuspended in complete DMEM prior to addition to epithelial cells. For infection, Caco-2 cells grown to confluence in 6 or 12-well plates were incubated with bacteria at a multiplicity of infection (MOI) of 10 for up to 3 h. For longer incubations, bacteria were removed by washing after 3 h, and the cells were incubated for the additional period in the presence of 50 µg ml−1 gentamicin to kill any remaining bacteria. In some experiments, confluent monolayers of Caco-2 cells were pretreated for 30 min with the p38 MAP kinase inhibitor SB203580 (10 µM, Calbiochem) or the MEK-1 inhibitor PD98059 (50 µM, Calbiochem) before bacterial infection. For the preparation of bacterial supernatants, bacteria were grown overnight at 37°C with shaking. Bacterial cultures were centrifuged and culture media was sterile-filtered (0.22 µm) before addition to the cells at a 1:10 dilution in complete DMEM.
Isolation of flagellin from EHEC
Flagellin was isolated by acid depolymerization. Briefly, EHEC were grown with shaking overnight at 37°C after which bacteria were pelleted and resuspended in 10 mM HCl, 150 mM NaCl, and incubated with rocking for 30 min at RT. After initial centrifugation to remove bacteria, supernatants were clarified by ultracentrifugation at 100 000 g for 90 min and the pH was adjusted to 8.0 with 10 mM NaOH, 50 mM Tris. Supernatants were stored at −80°C prior to use.
Caco-2 cells grown overnight in serum-free DMEM were infected and then washed with ice-cold PBS containing 20 mM β-glycerophosphate and 20 mM NaF, and lysed at 4°C with lysis buffer [50 mM Tris HCl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Ipegal CA-630, 0.25% sodium deoxycholate, plus phosphatase inhibitors (10 mM NaF, 10 mM Na3VO4, 20 mM β-glycerophosphate), and protease inhibitors (protease inhibitor cocktail III, Calbiochem, La Jolla, CA)]. Total protein in the cell extracts was determined by Bradford assay (Bio-Rad, Hercules, CA). Samples containing 40 µg total protein were electrophoresed on 10% SDS-PAGE gels (Bio-Rad) and then transferred to nitrocellulose membranes (Hybond, Amersham Pharmacia, Arlington Heights, IL) after which membranes were blocked with a 5% (w/v) solution of non-fat milk powder in Tris-buffered saline/0.1% Tween-20 (TBS/T) for 1 h at RT. Membranes were probed with polyclonal rabbit antibodies specific for phosphorylated p38, phosphorylated ERK 1/2, total ERK 1/2, or IκBα (Cell Signalling Technology, Beverly, MA). Detection was performed using donkey anti-rabbit antibody (Amersham Life Science) followed by an ECL detection kit (Amersham) and exposure to X-ray film (Biomax, Eastman Kodak, Rochester, NY).
Nuclear extracts and EMSA for NF-κB activation
Cells in 10 cm dishes were washed with ice-cold PBS, after which cells were scraped into cold microfuge tubes and pelleted. Cell pellets were resuspended in 1 ml of hypotonic buffer (10 mM HEPES pH 7.9, 1.5 mM MgCl2, 10 mM NaCl) containing 0.1% Nonidet P-40 with freshly added protease inhibitor cocktail III and 1 mM DTT and incubated on ice for 10 min. Samples were pelleted for 3 min at 8000 r.p.m. at 4°C and the pellets were washed in hypotonic buffer lacking Nonidet P-40 and repelleted. The resulting nuclear pellets were resuspended in high-salt buffer (20 mM HEPES pH 7.9, 0.4 M NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 20% v/v glycerol) with freshly added protease inhibitors and 1 mM DTT and incubated on ice for 10 min with intermittent vortexing. Samples were then spun at 12 000 r.p.m. for 20 min and the resulting supernatant was used as the nuclear extract. A double-stranded oligonucleotide corresponding to an NF-κB consensus sequence (DiDonato et al., 1995) was end-labelled with γ33P-ATP (Amersham) using T4 polynucleotide kinase (Stratagene, San Diego, CA). Five µg nuclear extract were incubated for 30 min at RT with ∼0.3 ng of labelled oligonucleotide in buffer containing 10 mM Tris pH 7.6, 50 mM NaCl, 1 mM EDTA, 5 mM DTT, 5% glycerol, 0.1 mg ml−1 BSA and 2 µg of poly(dIdC)·poly(dIdC). (Pharmacia). Free and bound probe were separated in 5% native polyacrylamide gels.
Recombinant adenovirus containing an IκBα-AA super-repressor (Ad5IκB-A32/36) or the E. coliβ-galactosidase gene (Ad5LacZ) was as described before (Elewaut et al., 1999). Briefly, Ad5IkB-A32/36 contains alanine for serine substitutions at residues 32 and 36, cannot be phosphorylated at those positions, and functions as a super-repressor of NF-κB activation. Caco-2 cells at approximately 50% confluence were infected with Ad5IκB-A32/36 or Ad5LacZ (both at an MOI of 100) for 12 h in serum-free media (Optim-MEM, Life Technologies). After infection, adenovirus was removed by washing, fresh medium containing serum was added, and cells were incubated for an additional 12 h before bacterial infection.
Reverse transcription and real-time RT-PCR
Total RNA was extracted from cells (RNeasy, Qiagen, Valencia, CA) and treated with DNase (Qiagen) to remove any contaminating genomic DNA. One µg of total RNA was reverse transcribed and amplified using primers specific for IL-8 and β-actin as described before (Elewaut et al., 1999). One microlitre of cDNA was amplified using an ABI Prism 7700 Sequence Detection System (PE Applied Biosystems, Foster City, CA) with 2X SYBR Green master mix (Applied Biosystems) as described before (Berin et al., 2001). Amplification of the single expected PCR product was confirmed by electrophoresis on agarose gels and ethidium bromide staining.
IL-8 in cell culture supernatants was measured by ELISA, using a monoclonal mouse anti-human IL-8 antibody for capture and a biotinylated polyclonal anti-human IL-8 antibody for detection (both from R and D Systems, Minneapolis, MN). The ELISA was sensitive to 3 pg IL-8 ml−1.
Immunoblotting for H7 flagellin
Bacterial supernatants or isolated surface flagella were electrophoresed on 10% SDS-PAGE gels and transferred to nitrocellulose membranes. Membranes were blocked for 60 min at RT in TBS/T with 5% non-fat milk powder. Anti-H7 flagellin serum (Lee Laboratories, Grayson, GA) was diluted 1:1000 in TBS/T with 5% milk powder, and incubated with the membrane for 60 min at RT. HRP-conjugated donkey anti-rabbit Ig (Amersham Life Science) was used at a 1:2000 dilution for detection, followed by incubation with ECL reagents and exposure to X-ray film.
Immunoprecipitation of H7 Flagellin
Rabbit anti-H7 serum (Lee Laboratories) was added to supernatants from EHEC (stx–) cultures at a 1:50 dilution and incubated for 60 min with rotating. One hundred microlitres of a 1:1 slurry of protein A-sepharose (Sigma) in PBS was added to 500 µl of bacterial supernatant and incubated for 60 min with rotating, after which the Sepharose beads were pelleted and the supernatant was harvested. Depletion of flagellin from the supernatants was confirmed by immunoblotting. As a control, bacterial culture supernatants were incubated with normal rabbit serum, rather than rabbit anti-H7 serum.
This work was supported by National Institutes of Health grants DK58960 and DK35180. M. C. Berin was supported in part by a fellowship from the Canadian Institute of Health Research. A. Darfeuille-Michaud was supported in part by the Philippe Foundation.
- 2001) Toll-like receptors: critical proteins linking innate and acquired immunity. Nat Immunol 2: 675–680. , , and (
- 1992) Cloning and nucleotide sequence of the eae gene homologue from enterohemorrhagic Escherichia coli serotype O157: H7. FEMS Microbiol Lett 70: 63–68. , , , and (
- 2001) Production of MDC/CCL22 by human intestinal epithelial cells. Am J Physiol Gastrointest Liver Physiol 280: G1217–G1226. , , , and (
- 1999) Escherichia coli O157: H7 gastroenteritis and the hemolytic uremic syndrome: an emerging infectious disease. Annu Rev Med 50: 355–367. , , and (
- 1999) Toll-like receptor-4 mediates lipopolysaccharide-induced signal transduction. J Biol Chem 274: 10689–10692. , , , , and (
- 1994) CagA/cytotoxic strains of Helicobacter pylori and interleukin-8 in gastric epithelial cell lines. J Clin Pathol 47: 945–950. , , , , , and (
- 1999) Enterohemorrhagic Escherichia coli O157: H7 produces Tir, which is translocated to the host cell membrane but is not tyrosine phosphorylated. Infect Immun 67: 2389–2398. , , , , , and (
- 2001) Enterohaemorrhagic and enteropathogenic Escherichia coli use a different Tir-based mechanism for pedestal formation. Mol Microbiol 41: 1445–1458. , , , , and (
- 1995) Phosphorylation of I kappa B alpha precedes but is not sufficient for its dissociation from NF-kappa B. Mol Cell Biol 15: 1302–1311. , , and (
- 1995) Diverse Pseudomonas aeruginosa gene products stimulate respiratory epithelial cells to produce interleukin-8. J Clin Invest 96: 2204–2210. , , , and (
- 2001) Pathogenesis and evolution of virulence in enteropathogenic and enterohemorrhagic Escherichia coli. J Clin Invest 107: 539–548. , and (
- 2001) Flagellin, a novel mediator of Salmonella-induced epithelial activation and systemic inflammation: I kappa B alpha degradation, induction of nitric oxide synthase, induction of proinflammatory mediators, and cardiovascular dysfunction. J Immunol 166: 1248–1260. , , , , , et al. (
- 1993) Epithelial cells secrete the chemokine interleukin-8 in response to bacterial entry. Infect Immun 61: 4569–4574. , , and (
- 1999) NF-kappa B is a central regulator of the intestinal epithelial cell innate immune response induced by infection with enteroinvasive bacteria. J Immunol 163: 1457–1466. , , , , , and (
- 1994) Modulation of host response to Escherichia coli O157: H7 infection by anti-CD18 antibody in rabbits. Gastroenterology 106: 1554–1561. , , , , , , et al. (
- 2001a) Salmonella typhimurium translocates flagellin across intestinal epithelia, inducing a proinflammatory response. J Clin Invest 107: 99–109. , , , , , , et al. (
- 2001b) Cutting edge: bacterial flagellin activates basolaterally expressed TLR5 to induce epithelial proinflammatory gene expression. J Immunol 167: 1882–1885. , , , , and (
- 2001) CARD4/Nod1 mediates NF-kappaB and JNK activation by invasive Shigella flexneri. EMBO Rep 2: 736–742. , , , , , , et al. (
- 2001) The innate immune response to bacterial flagellin is mediated by Toll- like receptor 5. Nature 410: 1099–1103. , , , , , et al. (
- 1992) Mechanism of resistance to complement-mediated killing of bacteria encoded by the Salmonella typhimurium virulence plasmid gene rck. J Clin Invest 90: 953–964. , , , , , and (
- 1997) Involvement of mitogen-activated protein kinase pathways in the nuclear responses and cytokine production induced by Salmonella typhimurium. cultured intestinal epithelial cells. J Immunol 159: 5550–5559. , , , and (
- 1999) Induction of interleukin-8 synthesis integrates effects on transcription and mRNA degradation from at least three different cytokine- or stress-activated signal transduction pathways. Mol Cell Biol 19: 6742–6753. , , , , , , et al. (
- 1996) Infection of human intestinal epithelial cells with invasive bacteria upregulates apical intercellular adhesion molecule-1 (ICAM) -1) expression and neutrophil adhesion. J Clin Invest 98: 572–583. , , , and (
- 2001) Shiga Toxin Translocation across Intestinal Epithelial Cells Is Enhanced by Neutrophil Transmigration. Infect Immun 69: 6148–6155. , , and (
- 2001) Human Nod1 confers responsiveness to bacterial lipopolysaccharides. J Biol Chem 276: 2551–2554. , , , , and (
- 2001) Regulated MIP-3alpha/CCL20 production by human intestinal epithelium: mechanism for modulating mucosal immunity. Am J Physiol Gastrointest Liver Physiol 280: G710–G719. , , , , and (
- 1995) A distinct array of proinflammatory cytokines is expressed in human colon epithelial cells in response to bacterial invasion. J Clin Invest 95: 55–65. , , , , , , and (
- 1997) Epithelial cells as sensors for microbial infection. J Clin Invest 100: 6–10. , and (
- 1997) Helicobacter pylori infection activates NF-kappa B in gastric epithelial cells. Gastroenterology 113: 1099–1109. , , , and (
- 1999) Differential activation of mitogen-activated protein kinases in AGS gastric epithelial cells by cag+ and cag- Helicobacter pylori. J Immunol 163: 5552–5559. , , , , Jr, , and (
- 1985) Dome formation by a human colonic adenocarcinoma cell line (HCA-7). Cancer Res 45: 3790–3795. (
- 2000) Virulence factors of Escherichia coli O157 and other Shiga toxin-producing E. coli. J Appl Microbiol 88: 729–745. (
- 1993) The effect of enterohemorrhagic Escherichia coli 0157: H7 on intestinal structure and solute transport in rabbits. Gastroenterology 104: 467–474. , , , , , and (
- 2001) ATP transduces signals from ASGM1, a glycolipid that functions as a bacterial receptor. Proc Natl Acad Sci USA 98: 9086–9091. , , , , , , and (
- 1988) Direct cytotoxic action of Shiga toxin on human vascular endothelial cells. Infect Immun 56: 2373–2378. , , , , , , and (
- 1999) Expression and regulation of the human beta-defensins hBD-1 and hBD-2 in intestinal epithelium. J Immunol 163: 6718–6724. , , , , , , and (
- 2000) Invasive Shigella flexneri activates NF-kappa B through a lipopolysaccharide-dependent innate intracellular response and leads to IL-8 expression in epithelial cells. J Immunol 165: 903–914. , , , and (
- 1996) Endocytosis, intracellular transport, and cytotoxic action of Shiga toxin and ricin. Physiol Rev 76: 949–966. , and (
- 1996) Attachment of a noninvasive enteric pathogen, enteropathogenic Escherichia coli, to cultured human intestinal epithelial monolayers induces transmigration of neutrophils. Infect Immun 64: 4480–4487. , , and (
- 1997) Activation of NF-kappaB in intestinal epithelial cells by enteropathogenic Escherichia coli. Am J Physiol 273: C1160–C1167. , , and (
- 2001) EPEC-activated ERK1/2 participate in inflammatory response but not tight junction barrier disruption. Am J Physiol Gastrointest Liver Physiol 281: G890–G898. , , , and (
- 1999) Peptidoglycan- and lipoteichoic acid-induced cell activation is mediated by toll-like receptor 2. J Biol Chem 274: 17406–17409. , , , , and (
- 1988) Characterization of flagella purified from enterohemorrhagic, vero-cytotoxin-producing Escherichia coli serotype O157: H7. J Clin Microbiol 26: 1367–1372. , , and (
- 2001) Flagellin stimulation of intestinal epithelial cells triggers CCL20- mediated migration of dendritic cells. Proc Natl Acad Sci USA 98: 13722–13727. , , , , , and (
- 1998) Enteroaggregative Escherichia coli produce intestinal inflammation and growth impairment and cause interleukin-8 release from intestinal epithelial cells. J Infect Dis 177: 88–96. , , , and (
- 2000) Enteroaggregative Escherichia coli expresses a novel flagellin that causes IL-8 release from intestinal epithelial cells. J Clin Invest 105: 1769–1777. , , , , and (
- 1994) Listeria monocytogenes, an invasive bacterium, stimulates MAP kinase upon attachment to epithelial cells. Mol Biol Cell 5: 455–464. , , and (
- 1991) Evaluation of the role of Shiga and Shiga-like toxins in mediating direct damage to human vascular endothelial cells. J Infect Dis 164: 344–352. , , , , and (
- 1999) Shiga toxins stimulate secretion of interleukin-8 from intestinal epithelial cells. Infect Immun 67: 5985–5993. , , , , , and (
- 2001) Shiga Toxins Induce, Superinduce, and Stabilize a Variety of C-X-C Chemokine mRNAs in Intestinal Epithelial Cells, Resulting in Increased Chemokine Expression. Infect Immun 69: 6140–6147. , , , and (
- 1998) Enteroinvasive bacteria directly activate expression of iNOS and NO production in human colon epithelial cells. Am J Physiol 275: G564–G571. , , , and (
- 1997) Differential and regulated expression of C-X-C, C-C, and C-chemokines by human colon epithelial cells. Gastroenterology 113: 1214–1223. , , , and (
- 1992) Cloning and characterization of the eae gene of enterohaemorrhagic Escherichia coli O157: H7. Mol Microbiol 6: 411–417. , and (