Present address: Institute of Gastroenterology, Tokyo Women's Medical University, Tokyo, Japan.
Role of Shiga toxin versus H7 flagellin in enterohaemorrhagic Escherichia coli signalling of human colon epithelium in vivo
Version of Record online: 13 JAN 2006
Volume 8, Issue 5, pages 869–879, May 2006
How to Cite
Miyamoto, Y., Iimura, M., Kaper, J. B., Torres, A. G. and Kagnoff, M. F. (2006), Role of Shiga toxin versus H7 flagellin in enterohaemorrhagic Escherichia coli signalling of human colon epithelium in vivo. Cellular Microbiology, 8: 869–879. doi: 10.1111/j.1462-5822.2005.00673.x
- Issue online: 13 JAN 2006
- Version of Record online: 13 JAN 2006
- Received 20 September, 2005; revised 26 November, 2005; accepted 29 November, 2005.
Enterohaemorrhagic Escherichia coli O157:H7 (EHEC) is a clinically important foodborne pathogen that colonizes human colon epithelium and induces acute colonic inflammation, but does not invade the epithelial cells. Whereas Shiga toxin (Stx) and bacterial flagellin have been studied for their ability to upregulate the production of proinflammatory chemokines by cultured human colon cancer cell lines, the relevance of studies in colon cancer cell lines to the production of proinflammatory signals by normal epithelial cells in EHEC-infected human colon is not known. We show herein that Stx does not bind to human colon epithelium in vivo. Moreover, globotriaosylceramide (Gb3/CD77) synthase, the enzyme required for synthesis of the Gb3/CD77 receptor for Stx, was not expressed by normal or inflamed human colon epithelium in vivo. In contrast, Toll-like receptor (TLR) 5, the receptor for bacterial flagellin, was expressed by normal human colon epithelium and by colon epithelium in human intestinal xenografts. EHEC H7 flagellin instilled in the lumen of human colon xenografts that contain an intact human epithelium upregulated the expression of epithelial cell proinflammatory chemokines, which was accompanied by a subepithelial influx of neutrophils. Isogenic mutants of EHEC that lacked flagellin did not significantly upregulate prototypic neutrophil and dendritic cell chemoattractants by model human colon epithelia, irrespective of Stx production. We conclude that EHEC H7 flagellin and not Stx is the major EHEC factor that directly upregulates proinflammatory chemokine production by human colon epithelium in vivo.
The foodborne pathogen enterohaemorrhagic Escherichia coli O157:H7 (EHEC) adheres to the apical membrane of colon epithelial cells, but is not invasive for those cells. Nonetheless, infection is characterized by acute inflammation of the colonic mucosa (Besser et al., 1999). EHEC is thought to signal colon epithelial cells to produce proinflammatory chemokines that can chemoattract and activate leukocytes, resulting in acute mucosal inflammation (Thorpe et al., 1999; 2001).
Enterohaemorrhagic E. coli produces several proteins that are candidates for signalling the upregulated production of epithelial chemokines. These include the EHEC surface protein intimin encoded by eae, which is important for development of the characteristic attaching and effacing (A/E) lesion on the apical cell membrane of colon epithelial cells (Yu and Kaper, 1992). Nonetheless, EHEC isogenic mutants with an inactivated eae gene are as effective as wild-type EHEC at inducing proinflammatory chemokine production by human colon epithelial cell lines (Berin et al., 2002). This is also the case for another A/E lesion producing pathogen, enteropathogenic E. coli (Zhou et al., 2003).
Shiga toxin (Stx) produced by EHEC is another candidate for stimulating upregulated epithelial chemokine production. Moreover, Stx has been reported to stimulate the production by some cultured colon cancer cell lines of proinflammatory chemokines, which hypothetically could chemoattract neutrophils and other leukocytes (Thorpe et al., 1999; 2001; Yamasaki et al., 1999). Stx's produced by EHEC are AB5 holotoxins whose B subunits interact with the glycolipid receptor globotriaosylceramide (Gb3/CD77) that is expressed by some long-term colon cancer cell lines (Nakao and Takeda, 2000) and whose synthesis is regulated by Gb3/CD77 synthase (α1,4-galactosyltransferase) (Jacewicz et al., 1995; Keusch et al., 1995; Kojima et al., 2000). Nonetheless, the ability of Stx to act as a potent stimulus for the production of proinflammatory mediators by human colon cancer cell lines was recently called into question (Berin et al., 2002; Rogers et al., 2003).
Flagellin signalling through Toll-like receptor (TLR)5 activates NF-κB and MAP kinases in colon cancer cell lines (Steiner et al., 2000; Gewirtz et al., 2001; Hayashi et al., 2001; Berin et al., 2002). Moreover, T84, a long-term human crypt-like human colon cancer cell line, was shown to express TLR5 on its basolateral membrane. Consistent with that distribution, basolateral but not apical stimulation of polarized monolayers of those cells with Salmonella flagellin induced basolateral secretion of CXCL8 (Gewirtz et al., 2001). However, studies with other human colon epithelial cell lines, also modelled as polarized epithelia (e.g. HCA-7, Caco-2), suggested that flagellin activation of epithelial chemokine production was not simply restricted to flagellin encounter at the basolateral membrane (Sierro et al., 2001; Berin et al., 2002). Nonetheless, those studies did not directly address whether flagellin signalling occurred through TLR5, and the possibility remained that another cell surface molecule (e.g. asialoGM1) that was reported to mediate flagellin signalling in airway epithelial cells (DiMango et al., 1995; Adamo et al., 2003), might explain the reported differences. Importantly, the relevance of any of the studies using colon cancer cell lines to colon epithelium in vivo has not been addressed. In the present study we show that EHEC, an epithelial adherent but non-invasive pathogen for colon epithelial cells, stimulates normal human colon epithelial cells to produce prototypic proinflammatory chemokines by flagellin signalling through TLR5 and not by Stx signalling through Gb3.
Human colonic epithelium in vivo does not bind Stx or express Gb3/CD77 synthase
To assess whether Stx binds to human colon epithelial cells in vivo, as is the case for some colon cancer cell lines (Jacewicz et al., 1995), normal and inflamed human colon were stained with Cy3-conjugated Stx2. Whereas endothelial cells in the vessels of normal (Fig. 1) and inflamed colon mucosa were brightly stained, there was no staining of colon epithelial cells. Importantly, there also was no Stx staining of epithelial cells in sections of human colon cancers from nine different individuals, although Stx bound to the vasculature in those cancers (not shown). Consistent with the lack of Stx2 staining of human colon epithelium in vivo, mRNA transcripts for Gb3/CD77 synthase, the enzyme required for synthesis of the Stx receptor Gb3/CD77, were not detected in freshly isolated epithelial cells from normal (Fig. 2A) or inflamed human colon biopsies or human colon cancer tissue. Although some long-term colon cancer epithelial cell lines (e.g. Caco-2, HCA-7) expressed Gb3/CD77 synthase and stained with Cy3-conjugated Stx2 (Fig. 2B), others expressed little (HT-29) or no (T84) Gb3/CD77 synthase and this was paralleled by a lack of staining with Cy3-conjugated Stx2. Notably, the epithelium of colon cancers in vivo appears to more closely resemble that of normal colon epithelium with respect to Gb3/CD77 synthase expression and Stx binding, and differs from that of some long-term cultured human colon cancer cell lines. Moreover, only those human colon cancer epithelial cell lines that stained positively for Gb3/CD77 produced CXCL8 in response to Stx, albeit at low levels (Berin et al., 2002).
Butyrate, a short chain fatty acid generated in the colon by the bacterial digestion of complex carbohydrates, can upregulate Gb3/CD77 expression by some long-term colon cancer epithelial cell lines (Jacewicz et al., 1995; Keusch et al., 1996). To determine whether this was also the case for in vivo intestinal epithelium, we used a human intestinal xenograft model, because human colon xenografts contain an intact intestinal epithelium that is strictly of human origin (Savidge et al., 1995) and are devoid of intraluminal bacteria (Hase et al., 2002). Instillation of sodium butyrate (NaB) into the lumen of xenografts at a concentration present in the normal colon lumen (10 mM), and which we found to upregulate Gb3/CD77 synthase expression and Gb3/CD77 production by some human colon cancer cell lines, did not induce the expression of Gb3/CD77 synthase by epithelium in the xenografts or result in Stx binding to xenograft epithelial cells (not shown).
Toll-like receptor 5 expression, and EHEC H7 flagellin induced chemokine production, by human colon epithelium in vivo
Bacterial flagellin is known to upregulate chemokine production by human colon cancer cell lines in vitro. The receptor for bacterial flagellin is TLR5. As shown in Fig. 3, human colon epithelium in vivo expresses TLR5 mRNA and protein, although the distribution of TLR5 in colon epithelium in vivo was predominately intracellular (Fig. 3A), which contrasts with the selective basolateral expression of TLR5 reported for a human colon cancer cell line, T84 (Gewirtz et al., 2001).
We next asked whether EHEC H7 flagellin increased chemokine production by human colon epithelial cells in vivo. We again used the human colon xenograft model, because xenografts are well-suited for studies of early events that ensue after agonist stimulation of human intestinal epithelium in vivo (Eckmann et al., 1997; O’Neil et al., 1999; Berin et al., 2001; Dwinell et al., 2001; Izadpanah et al., 2001; Maaser et al., 2004; Ogawa et al., 2004). The lack of a commensal bacterial flora in human colon xenografts (Hase et al., 2002) was a further advantage allowing us to assess flagellin stimulation of intact human colon epithelium in vivo in the absence of other microbial products. Colon epithelial cells in the xenografts expressed TLR5 intracellularly, with some staining also being apically and basolaterally distributed (Fig. 4A). EHEC H7 flagellin, instilled into the lumen of xenografts, upregulated mRNA expression of the prototypic human neutrophil chemoattractant CXCL8 (Fig. 4B) and CCL20, the chemoattractant for CCR6 expressing dendritic cells, as well as production of CCL20 (Fig. 4C). This was accompanied by a subepithelial infiltrate consisting mostly of neutrophils (Fig. 4D). Based on nuclear morphology, the infiltrating neutrophils were of mouse origin. Consistent with this, human IL-8 has previously been shown to chemoattract mouse neutrophils (Laurent et al., 1997).
Upregulation of CXCL8 production in response to isogenic mutants of EHEC
To complement the in vivo studies, we used Caco-2 human epithelial colon cancer cells as an in vitro model system, because those cells express TLR5 mRNA and protein and purified EHEC H7 flagellin upregulates CXCL8 secretion by Caco-2 cells (Figs 3B and 5A). Importantly, irrespective of whether isogenic mutants of EHEC expressed both flagellin and Stx2, or only flagellin, they upregulated CXCL8 production by Caco-2 cells to a similar magnitude (Fig. 5B). Conversely, EHEC lacking flagellin stimulated little if any CXCL8 production irrespective of whether they produced Stx2 (Fig. 5B).
Epithelial cell encounter with EHEC flagellin at the apical surface upregulates CXCL8 by signalling through TLR5 and not an alternative flagellin receptor
Enterohaemorrhagic E. coli colonize the apical membrane of colon epithelial cells, form A/E lesions, and signal colon epithelial cells through a Type III secretion system (DeVinney et al., 1999). However, whether or not TLR5 is apically expressed by polarized colon epithelial cell lines, and whether apically encountered bacterial flagellin in cultures of polarized colon epithelial cells signals through TLR5 has been controversial (Gewirtz et al., 2001; Sierro et al., 2001). As shown by confocal microscopy, Caco-2 cells, like human epithelium in colon xenografts in vivo, expressed TLR5 both intracellularly, apically and basolaterally (Fig. 6A). Moreover, the magnitude of basolateral CXCL8 secretion was similar irrespective of whether flagellin was added to the apical or basal chambers of polarized cultures of Caco-2 cells (Fig. 6B).
To show that the CXCL8 response in these studies was specific for flagellin and not due to potentially contaminating TLR ligands in the flagellin preparations, Caco-2 cells were stimulated with other TLR ligands. Neither adding titrated doses of LPS to EHEC H7 flagellin preparations nor adding polymyxin B to preparations of EHEC H7 flagellin, to inactivate potentially contaminating bacterial LPS, altered CXCL8 production by Caco-2 cells in response to EHEC H7 flagellin (data not shown). Furthermore, NF-κB was not activated in cells expressing TLR2 that were stimulated with EHEC H7 flagellin, whereas NF-κB was activated in those cells in response to stimulation with the TLR2 ligand Pam3Cys (10 ng ml−1) (data not shown). Finally, to show a specific requirement for bacterial flagellin, we used the fliC–/stx+ strain to prepare a flagellin negative reagent that would contain the same potential contaminants as the wild-type flagellin preparation. The flagellin negative preparation did not activate NF-κB or stimulate CXCL8 production by Caco-2 cells.
Flagellin from Pseudomonas aeruginosa was reported to activate NF-κB production in airway epithelial cells by signalling through either TLR5 or the ganglioside asialoGM1 (DiMango et al., 1995; Feldman et al., 1998; Adamo et al., 2003), raising the possibility that flagellin induction of the NF-κB target gene CXCL8 in colon cancer cell lines might take place through an alternative receptor. However, this was not the case. Stimulation of Caco-2 cells with IL-1α or EHEC H7 flagellin, but not anti-asialoGM1 antibody at concentrations able to ligate asialoGM1 on the cell surface of Caco-2 cells, activated NF-κB, as assessed by the translocation of RelA/p65 from the cytoplasm to the nucleus (Fig. 7). However, NF-κB was activated by anti-asialoGM1 antibody in BEAS-2B airway epithelial cells (not shown).
To further show that H7 flagellin activates colon epithelial cells solely through TLR5, Caco-2 cells were transiently transfected with a plasmid expressing dominant negative (DN) TLR5 or a control empty vector together with a 3 × NF-κB-luciferase reporter plasmid. Cells were left unstimulated or stimulated with EHEC H7 flagellin (100 ng ml−1) and luciferase activity was determined 6 h later. DN TLR but not the empty vector significantly blocked flagellin stimulated increases in luciferase activity (fold increase in flagellin stimulated luciferase activity: empty vector 2.7 ± 0.3; DN TLR5 1.2 ± 0.4, P < 0.01). As a control, IL-α stimulation of luciferase activity in those cells was not decreased by DN TLR5.
Enterohaemorrhagic E. coli colonizes human colon epithelial cells and induces an A/E lesion on their apical membrane, but does not invade those cells. Nonetheless, colonization of the apical epithelium induces acute colonic inflammation. This is thought to be initiated, at least in part, by the release from the colon epithelium of cytokines that chemoattract and activate mucosal inflammatory cells. Prior reports that used long-term cultured human colon cancer cell lines known to express Gb3/CD77, the receptor for Stx (Thorpe et al., 1999; 2001; Yamasaki et al., 1999), suggested that Stx was the major EHEC factor causing upregulated epithelial cell chemokine production. However, different human colon cancer epithelial cell lines vary markedly as to whether or not they express Gb3/CD77, and those that express Gb3/CD77 vary in its level of expression. We found that cell lines lacking Gb3/CD77 synthase and Gb3/CD77 do not produce proinflammatory chemokines in response to stimulation with Stx (M. Kagnoff et al., unpubl. data). Furthermore, we found that relative to proinflammatory cytokines like IL-1α or TNFα, Stx2 was a weak inducer of CXCL8 in colon cancer cell lines that express Gb3/CD77 synthase and Gb3/CD77 (Eckmann et al., 1993a; Yang et al., 1997; Berin et al., 2002).
In contrast to some of the human colon cancer cell lines, human colon epithelium in vivo did not express Gb3/CD77 synthase, the enzyme required for synthesis of the Gb3/CD77 receptor for Stx. Consistent with this, Stx did not bind to the epithelium in normal or inflamed human colon in vivo or to the epithelium in human colon cancers, although Stx bound to endothelial cells in normal and inflamed colon and in colon cancers. The absence of Gb3 synthase and Stx binding to human colon epithelium was further shown in human colon xenografts that contain an epithelium that is strictly of human origin. In addition, stimulation with Na butyrate, which upregulated the expression of Gb3 synthase and Gb/CD77 expression by some human colon cancer cell lines that constitutively express Gb3 synthase (M. Kagnoff et al., unpubl. data), did not induce Gb3 synthase expression or Stx binding in colon epithelial cells in human intestinal xenografts. These results are also supported also by biochemical and immunochemical studies that have failed to detect Gb3/CD77 in human intestinal epithelium (Bjork et al., 1987; Holgersson et al., 1991; Schuller et al., 2004). Of note, these studies do not exclude the possibility that Stx, which gains access to endothelial cells in the subepithelial region of the colon mucosa, has a role in the pathogenesis of mucosal inflammation. They also do not exclude the possibility that mediators released by endothelial cells in response to Stx might act, either directly or indirectly, on the epithelium and alter epithelial cell function.
Exposure of human intestinal epithelial cells to EHEC flagellin instilled into the lumen of human colon xenografts resulted in the upregulated production in vivo of CXCL8 and CCL20 by human colon epithelium. Moreover, this was accompanied by a marked influx of neutrophils into the subepithelial region of the xenograft mucosa. Stimulation of human colon cancer cell lines with bacterial flagellin has been reported to upregulate the expression of CXCL8 and/or CCL20 (Gewirtz et al., 2001; Sierro et al., 2001; Berin et al., 2002; Bambou et al., 2004; Tallant et al., 2004). Moreover, at least some of the cell lines used in those prior studies have been reported to express the flagellin receptor, TLR5. Although TLR5 expression was basolaterally restricted in the T84 cell line (Gewirtz et al., 2001), which led to the notion that bacterial flagellin signals through TLR5 only when it can access the epithelial cell basolateral membrane (Gewirtz et al., 2001), this did not appear to be the case for other colon cancer cell lines which, as shown herein for Caco-2 cells, express TLR5 apically, basolaterally and intracellularly. Consistent with that, as shown herein and by others (Sierro et al., 2001; Berin et al., 2002; Bambou et al., 2004), bacterial flagellin was equally effective at inducing CXCL8 or CCL20 production regardless of whether it was encountered apically or basolaterally by human colon epithelial cells. As with Caco-2 cells, TLR5 was not restricted to the basolateral domain of normal colon epithelium in vivo or colon epithelium in colon xenografts. This suggested Caco-2 cells as a suitable in vitro model for studying flagellin signalling. Using that model, flagellin signalling was shown to be mediated strictly through TLR5, and not an alternate receptor such as the ganglioside asialoGM1, which had been reported by others to function as a alternative receptor for bacterial flagellin on airway epithelial cell lines (DiMango et al., 1995; Feldman et al., 1998; Adamo et al., 2003; Ogushi et al., 2004; West et al., 2005). Studies in this model further showed that EHEC flagellin was both necessary and sufficient for upregulating CXCL8 in the context of EHEC infection.
Finally, we note that the distribution of TLR5 in human colon xenografts differed somewhat from that of epithelium in normal colon, with more pronounced apical and basolateral surface expression of TLR5 in the xenografts. This could reflect greater recycling of TLR5 to the intracellular compartment in normal colon epithelium due to TLR5 contact with flagellin released by the commensal colon bacteria, whereas the lumen of the xenografts lacks a commensal bacterial flora. We envision that this, or flagellin-induced self-tolerance (Mizel and Snipes, 2002), is more important than the previously proposed selective basolateral distribution of TLR5 for regulating TLR5 activation in the colon epithelium in vivo. The notion that differences in the activation of mucosal inflammation by commensal bacteria and pathogens can be explained simply by the polarized basolateral expression of TLRs by intestinal epithelial cells seems particularly unlikely, as it is clear that both commensals and pathogens express TLR ligands and both signal the intestinal epithelium (Rakoff-Nahoum et al., 2004).
Goat anti-human TLR5, rabbit anti-human RelA/p65 antibody (sc-372) and TLR5 blocking peptide (sc-8695P) were from Santa Cruz Biotechnology (Santa Cruz, CA). Cy-3-conjugated donkey anti-goat IgG, Cy-3-conjugated goat anti-mouse IgG, Cy-3-conjugated anti-rabbit IgG, and normal goat IgG were from Jackson ImmunoResearch Laboratories (West Grove, PA). Purified Stx2 was from Toxin Technology (Sarasota, FL). Recombinant human (rh)IL-1α was from Pepro Tech (Rocky Hill, NJ). Mouse anti-human IL-8 (CXCL8) monoclonal antibody (MAb), biotinylated goat anti-human IL-8 (CXCL8) antibody and mouse anti-human MIP3α (CCL20) MAb were from R and D systems (Minneapolis, MN). Rabbit anti-asialoGM1 antibody was from Wako Chemical (Richmond, VA). Alexa 488-conjugated phalloidin was from Molecular Probes (Eugene, OR). Hoechst 33258 and sodium butryate (NaB) were from Sigma-Aldrich (St Louis, MO).
Caco-2 cells (ATCC HTB 37; American Type Culture Collection, Manassas, VA) and HCA-7 cells (Berin et al., 2002) were grown in plastic tissue culture dishes in Dulbecco's modified Eagle's medium (DMEM) and BEAS-2B cells (ATCC) were cultured in DMEM/F12 medium. Media were supplemented with 10% heat-inactivated fetal bovine serum (FBS) and 2 mM l-glutamine (Life Technologies, Gaithersburg, MD). Cells were cultured in an atmosphere of 95% air/5% CO2 at 37°C. To obtain polarized epithelial monolayers, Caco-2 cells were grown on polyester membranes in Transwells (1 cm2, 0.4 µm pore size; Costar, Cambridge, MA) for 10–12 days, at which time cells expressed increased levels of alkaline phosphatase as a marker of differentiation and a transepithelial resistance of > 400 Ω cm2 was established.
The following bacterial strains were used: wild-type motile EHEC strain 86–24 of serogroup O157:H7 that produces Stx2 (referred to as 86–24:fliC+/stx+), a member of the Stx family that is frequently associated with severe disease (gift of M. Donnenberg, University of Maryland) (Donnenberg et al., 1993; Boerlin et al., 1999); TUV86-2, an isogenic motile mutant of 86–24 that has a deletion of the entire region encoding the A and B subunits of Stx2 and does not produce Stx2 (referred to as 86–24:fliC+/stx–) (gift from J. Hacker, Würzburg University) (Gunzer et al., 1998), AGT06, a non-motile isogenic fliC– mutant of 86–24 (referred to as 86–24:fliC–/stx+), and AGT05, a non-motile isogenic fliC– and stx– mutant of 86–24 that does not produce Stx2 and lacks flagella (referred to as 86–24:fliC–/stx–). The fliC mutants were constructed by allelic exchange as previously described (Torres et al., 2002).
The elisa for IL-8 used mouse anti-human IL-8 MAb as the capture antibody and biotinylated polyclonal anti-human IL-8 antibody for detection as described before (Eckmann et al., 1993b). The IL-8 elisa was sensitive to 3 pg ml−1.
Motility assay and staining of flagella
Motility of wild-type EHEC and isogenic EHEC mutants was assessed 48 h after incubation of bacteria on 0.3% Luria–Bertani (LB) agar plates at 25°C. The presence of flagella in wild-type and mutant EHEC 86–24 strains was further tested using flagella staining solution according to the manufacture's instructions (Hardy Diagnostics, Santa Maria, CA).
Epithelial cells grown to confluence in 12-well plates were infected with EHEC or isogenic EHEC mutant bacteria at a multiplicity of infection (moi) of 100. Cells were incubated with bacteria for 3 h after which bacteria were removed by washing. To kill remaining extracellular bacteria, the cells were further incubated for additional time periods in the presence of 50 µg ml−1 gentamicin.
Isolation of EHEC H7 flagellin
Enterohaemorrhagic E. coli flagellin was prepared as described before (Steiner et al., 2000; Berin et al., 2002). Briefly, bacteria grown in LB medium for 48 h at 25°C were centrifuged. Bacterial pellets were resuspended in 500 mM Tris (pH 8.0), sheared (Polytron PT10-35, Brinkmann Instruments, Westbury, NY) for 60 s, and centrifuged at 8000 g for 15 min after which supernatants were filtered. Flagella were centrifuged at 100 000 g for 90 min at 20°C and resuspended in PBS. Purity and molecular mass of the flagellin was verified by SDS-PAGE and Coomassie blue staining. To prepare a control an H7 flagellin negative reagent, strain 86–24:fliC–/stx+ was subjected to the identical protocol in parallel.
RelA/p65 immunostaining was performed as described before (Egan et al., 2003). Briefly, cells grown on chamber slides (Nunc, Naperville, IL) were fixed with 4% paraformaldehyde, permeabilized with 0.5% Triton X-100, and blocked with PBS containing 1% BSA and 5% goat serum, after which polyclonal rabbit anti-p65 IgG in blocking buffer was added and cells were further incubated at 4°C overnight. After washing, Cy3-conjugated goat anti-rabbit Ig in blocking buffer was added as secondary antibody. Cells in which nuclear fluorescence was the same or greater than cytoplasmic intensity were considered positive for RelA/p65 translocation.
Plasmids, transfections and luciferase assay
A plasmid expressing DN TLR5 (TLR-5 ΔTIR) in pZERO and control pZERO empty vector were from InvivoGen, San Diego, CA. A plasmid expressing Flag-tagged human TLR2 in pCMV and empty vector were provided by J. Lee (Lee et al., 2003). Rous sarcoma virus β-galactosidase (RSV-β-gal) and 3 × NF-κB-luciferase transcriptional reporter constructs were as described before (Elewaut et al., 1999). Cells cultured in 24-well dishes were transfected with plasmid DNA using Lipofectamine 2000 (Invitrogen, Carlsbad, CA) according to the manufacturer's instructions. Luciferase activity was assayed and normalized to β-gal activity (Roebuck et al., 1993; Eckmann et al., 1995).
Reverse transcription, real-time reverse transcription polymerase chain reaction (RT-PCR) and qualitative RT-PCR
Total RNA was extracted from isolated epithelium of human colon biopsies (Iimura et al., 2000) using RNeasy (Qiagen, Valencia, CA). One µg of isolated total cellular RNA was reverse transcribed. CXCL8 and β-actin mRNA were amplified from cDNA by real-time PCR using specific primers for CXCL8 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 2 × SYBR Green master mix (Applied Biosystems) as described before (Berin et al., 2002; Hase et al., 2002). Data analysis used sequence detection system software provided by the manufacturer. For qualitative RT-PCR, the following primers were used to amplify Gb3/CD77 synthase cDNA: sense primer 5′-CACCT CTCTGCAATGGGCTGC-3′; anti-sense primer 5′-CGTGAACTT GAAGCCGATGATG-3′ and TLR5 sense primer 5′-TCAAACCC CTTCAGAGAATCCC; anti-sense primer 5′-TTGGAGTTGAGG CTTAGTCCCC-3′. The amplification profile for Gb3/CD77 synthase was 15 s denaturation at 95°C, 1 min annealing 63°C and for TLR5 was 15 s denaturation at 95°C, 1 min annealing at 60°C. The primers and amplification profile for GAPDH were as described before (Iimura et al., 2000). Amplification of the expected single PCR product was confirmed on 1% agarose gels stained with ethidium bromide.
Colon biopsies were obtained from 10 different normal individuals and seven individuals with marked colon inflammation, three due to ulcerative colitis, two Crohn's disease, and two diverticulitis. In addition, surgical specimens and mucosal biopsies were obtained from nine different individuals with adenocarcinoma of the colon. Tissue was embedded in OCT and snap frozen in isopentane-dry ice (Dwinell et al., 1999; O’Neil et al., 1999) after which cryostat sections (5 µm) were prepared. Caco-2 cells were detached with 0.25% trypsin/EDTA, washed with PBS, centrifuged and resuspended in PBS. Aliquots were centrifuged onto positively charged glass slides (Fisher Scientific, Houston, TX) using a cytospin centrifuge (Shandon, Pittsburgh, PA). For TLR5 immunostaining, sections were fixed with 4% paraformaldehyde for 20 min and blocked in PBS containing 0.5 M NaCl, pH 8.6 and normal donkey serum (1 mg ml−1) for 1 h at room temperature (RT). Sections were incubated for 1 h with goat anti-human TLR5 (5 mg ml−1) or control goat IgG (5 mg ml−1) as primary antibody and Cy3-conjugated donkey anti-goat IgG antibody as secondary antibody at RT for 60 min. As a further specificity control, anti-human TLR5 was absorbed at 4°C with a peptide of TLR5 (TLR5 blocking peptide; sc-8695P, Santa Cruz) at a ratio 1:10 (Bambou et al., 2004). For counter-staining nuclei, Hoechst 33258 (1 µg ml−1) was added for 3 min at RT. Conditions for CCL20 immunostaining were as described before (Izadpanah et al., 2001). For Gb3/CD77 staining, sections or cell samples were blocked in PBS (pH 7.4)-0.2% BSA for 1 h at RT followed by incubation with Cy3-conjugated Stx2 (10 mg ml−1) (kindly provided by L. Johannes, Institut Curie) for 1 h at RT.
Caco-2 cell monolayers on cell insert membranes were fixed for 10 min with 4% paraformaldehyde (pH 7.0) and incubated overnight at 4°C with goat anti-TLR5 or control goat IgG. Membranes were washed and incubated for 60 min at RT with Cy3-conjugated donkey anti-goat antibody together with Alexa-488-coupled phalloidin for F-actin staining. Monolayers were removed from the surrounding plastic frame, mounted on glass slides and examined by laser scanning confocal microscopy (MRC Bio-Rad 1024, Bio-Rad Laboratories).
Human intestinal xenografts
Human fetal colon of gestational age 12–18 weeks (Advanced Biosciences Resources, Alameda, CA), was transplanted subcutaneously onto the backs of C57BL/6 severe combined immunodeficiency (SCID) mice as described before (Huang et al., 1996; Eckmann et al., 1997; Laurent et al., 1997). Xenografts were allowed to develop for 14–18 weeks after implantation, at which time the epithelium, which is strictly of human origin, is fully differentiated (Savidge et al., 1995). Mature intestinal xenografts were injected intraluminally with either l00 µl of H7 flagellin in PBS (10 µg ml−1) or PBS alone, and harvested 6 h later. Epithelium from the xenografts was isolated as described before (Iimura et al., 2000; Hase et al., 2003) and used for RNA isolation. In other experiments, xenografts were injected intraluminally with 200 µl of NaB (10 mM) for 5 or 24 h prior harvesting the xenografts. Segments of xenografts were embedded in optimum cutting temperature compound (OCT, TissueTek, Torrance, CA) and frozen in isopentane-dry ice for immunohistochemical analysis. These studies were approved by the University of California, San Diego Human and Animal Subjects Committees.
Data were analysed using a two-tailed Student's t-test. P-values less than or equal to 0.05 were considered statistically significant.
This work was supported by NIH Grants DK58960 and DK35108 (M.F.K.), DK58957 (J.B.K.) and a grant from the Cystic Fibrosis Foundation (M.F.K.).
- 2003) P. aeruginosa flagella activate airway epithelial cells through asialoGM1 and TLR2 as well as TLR5. Am J Respir Cell Mol Biol 30: 627–634. , , , , (
- 2004) In vitro and ex vivo activation of the TLR5 signaling pathway in intestinal epithelial cells by a commensal Escherichia coli strain. J Biol Chem 279: 42984–42992. , , , , , , et al. (
- 2001) Production of MDC/CCL22 by human intestinal epithelial cells. Am J Physiol Gastrointest Liver Physiol 280: G1217–G1226. , , , (
- 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. Cell Microbiol 4: 635–648. , , , , (
- 1999) Escherichia coli O157:H7 gastroenteritis and the hemolytic uremic syndrome: an emerging infectious disease. Annu Rev Med 50: 355–367. , , (
- 1987) Structures of blood group glycosphingolipids of human small intestine. A relation between the expression of fucolipids of epithelial cells and the ABO, Le and Se phenotype of the donor. J Biol Chem 262: 6758–6765. , , , , (
- 1999) Associations between virulence factors of Shiga toxin-producing Escherichia coli and disease in humans. J Clin Microbiol 37: 497–503. , , , , , (
- 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. , , , , , (
- 1995) Diverse Pseudomonas aeruginosa gene products stimulate respiratory epithelial cells to produce interleukin-8. J Clin Invest 96: 2204–2210. , , , (
- 1993) The role of the eae gene of enterohemorrhagic Escherichia coli in intimate attachment in vitro and in a porcine model. J Clin Invest 92: 1418–1424. , , , , , (
- 1999) Chemokine receptor expression by human intestinal epithelial cells. Gastroenterology 117: 359–367. , , , , (
- 2001) Regulated production of interferon-inducible T-cell chemoattractants by human intestinal epithelial cells. Gastroenterology 120: 49–59. , , , (
- 1993a) Differential cytokine expression by human intestinal epithelial cell lines: regulated expression of interleukin 8. Gastroenterology 105: 1689–1697. , , , , , (
- 1993b) Epithelial cells secrete the chemokine interleukin-8 in response to bacterial entry. Infect Immun 61: 4569–4574. , , (
- 1995) Entamoeba histolytica trophozoites induce an inflammatory cytokine response by cultured human cells through the paracrine action of cytolytically released interleukin-1 alpha. J Clin Invest 96: 1269–1279. , , , (
- 1997) Role of intestinal epithelial cells in the host secretory response to infection by invasive bacteria. Bacterial entry induces epithelial prostaglandin H synthase-2 expression and prostaglandin E2 and F2α. J Clin Invest 100: 296–309. , , , , , , et al. (
- 2003) Nuclear factor-κB activation promotes restitution of wounded intestinal epithelial monolayers. Am J Physiol Cell Physiol 285: C1028–C1035. , , , , , (
- 1999) NF-κB is a central regulator of the intestinal epithelial cell innate immune response induced by infection with enteroinvasive bacteria. J Immunol 163: 1457–1466. , , , , , (
- 1998) Role of flagella in pathogenesis of Pseudomonas aeruginosa pulmonary infection. Infect Immun 66: 43–51. , , , , , , (
- 2001) Cutting edge: bacterial flagellin activates basolaterally expressed TLR5 to induce epithelial proinflammatory gene expression. J Immunol 167: 1882–1885. , , , , (
- 1998) Construction and characterization of an isogenic slt-ii deletion mutant of enterohemorrhagic Escherichia coli. Infect Immun 66: 2337–2341. , , , , , , (
- 2002) Cell differentiation is a key determinant of cathelicidin LL-37/human cationic antimicrobial protein 18 expression by human colon epithelium. Infect Immun 70: 953–963. , , , , (
- 2003) Expression of LL-37 by human gastric epithelial cells as a potential host defense mechanism against Helicobacter pylori. Gastroenterology 125: 1613–1625. , , , , , , et al. (
- 2001) The innate immune response to bacterial flagellin is mediated by Toll-like receptor 5. Nature 410: 1099–1103. , , , , , , et al. (
- 1991) Glycosphingolipids of human large intestine: detailed structural characterization with special reference to blood group compounds and bacterial receptor structures. J Biochem (Tokyo) 110: 120–131. , , (
- 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. , , , (
- 2000) Bax is downregulated in inflamed colonic mucosa of ulcerative colitis. Gut 47: 228–235. , , , , , , (
- 2001) Regulated MIP-3α/CCL20 production by human intestinal epithelium: mechanism for modulating mucosal immunity. Am J Physiol Gastrointest Liver Physiol 280: G710–G719. , , , , (
- 1995) Maturational regulation of globotriaosylceramide, the Shiga-like toxin 1 receptor, in cultured human gut epithelial cells. J Clin Invest 96: 1328–1335. , , , , , (
- 1995) Globotriaosylceramide, Gb3, is an alternative functional receptor for Shiga-like toxin 2e. Infect Immun 63: 1138–1141. , , , , , (
- 1996) Comparison of the effects of Shiga-like toxin 1 on cytokine- and butyrate-treated human umbilical and saphenous vein endothelial cells. J Infect Dis 173: 1164–1170. , , , , (
- 2000) Molecular cloning of globotriaosylceramide/CD77 synthase, a glycosyltransferase that initiates the synthesis of globo series glycosphingolipids. J Biol Chem 275: 15152–15156. , , , , , , et al. (
- 1997) Cryptosporidium parvum infection of human intestinal epithelial cells induces the polarized secretion of C-X-C chemokines. Infect Immun 65: 5067–5073. , , , , , , (
- 2003) Molecular basis for the immunostimulatory activity of guanine nucleoside analogs: activation of Toll-like receptor 7. Proc Natl Acad Sci USA 100: 6646–6651. , , , , , , et al. (
- 2004) Expression of Epstein-Barr virus-induced gene 3 and other interleukin-12-related molecules by human intestinal epithelium. Immunology 112: 437–445. , , , , (
- 2002) Gram-negative flagellin-induced self-tolerance is associated with a block in interleukin-1 receptor-associated kinase release from toll-like receptor 5. J Biol Chem 277: 22414–22420. , (
- 2000) Escherichia coli Shiga toxin. J Nat Toxins 9: 299–313. , (
- 1999) Expression and regulation of the human beta-defensins hBD-1 and hBD-2 in intestinal epithelium. J Immunol 163: 6718–6724. , , , , , , (
- 2004) Regulated production of the chemokine CCL28 in human colon epithelium. Am J Physiol Gastrointest Liver Physiol 287: G1062–G1069. , , , (
- 2004) Gangliosides act as co-receptors for Salmonella enteritidis FliC and promote FliC induction of human β-defensin-2 expression in Caco-2 cells. J Biol Chem 279: 12213–12219. , , , , , , et al. (
- 2004) Recognition of commensal microflora by toll-like receptors is required for intestinal homeostasis. Cell 118: 229–241. , , , , (
- 1993) Identification of c-fos-responsive elements downstream of TAR in the long terminal repeat of human immunodeficiency virus type-1. J Clin Invest 92: 1336–1348. , , (
- 2003) Enhanced CXC chemokine responses of human colonic epithelial cells to locus of enterocyte effacement-negative shiga-toxigenic Escherichia coli. Infect Immun 71: 5623–5632. , , , (
- 1995) Human intestinal development in a severe-combined immunodeficient xenograft model. Differentiation 58: 361–371. , , , , , (
- 2004) Interaction of Shiga toxin from Escherichia coli with human intestinal epithelial cell lines and explants: Stx2 induces epithelial damage in organ culture. Cell Microbiol 6: 289–301. , , (
- 2001) Flagellin stimulation of intestinal epithelial cells triggers CCL20-mediated migration of dendritic cells. Proc Natl Acad Sci USA 98: 13722–13727. , , , , , (
- 2000) Enteroaggregative Escherichia coli expresses a novel flagellin that causes IL-8 release from intestinal epithelial cells. J Clin Invest 105: 1769–1777. , , , , (
- 2004) Flagellin acting via TLR5 is the major activator of key signaling pathways leading to NF-κB and proinflammatory gene program activation in intestinal epithelial cells. BMC Microbiol 4: 33. , , , , , (
- 1999) Shiga toxins stimulate secretion of interleukin-8 from intestinal epithelial cells. Infect Immun 67: 5985–5993. , , , , , (
- 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. , , , (
- 2002) Characterization of Cah, a calcium-binding and heat-extractable autotransporter protein of enterohaemorrhagic Escherichia coli. Mol Microbiol 45: 951–966. , , , , , (
- 2005) Gangliosides inhibit flagellin signaling in the absence of an effect on flagellin binding to toll-like receptor 5. J Biol Chem 280: 9482–9488. , , (
- 1999) Induction of cytokines in a human colon epithelial cell line by Shiga toxin 1 (Stx1) and Stx2 but not by non-toxic mutant Stx1 which lacks N-glycosidase activity. FEBS Lett 442: 231–234. , , , , , (
- 1997) Differential and regulated expression of C-X-C, C-C, and C-chemokines by human colon epithelial cells. Gastroenterology 113: 1214–1223. , , , (
- 1992) Cloning and characterization of the eae gene of enterohaemorrhagic Escherichia coli O157:H7. Mol Microbiol 6: 411–417. , (
- 2003) Flagellin of enteropathogenic Escherichia coli stimulates interleukin-8 production in T84 cells. Infect Immun 71: 2120–2129. , , , , , , (