MEK Is a Key Modulator for TLR5-induced Interleukin-8 and MIP3 (cid:1) Gene Expression in Non-transformed Human Colonic Epithelial Cells*

Flagellin, a specific ligand for Toll-like receptor 5 (TLR5), is a molecular pattern associated with several bacterial species. Recently, TLR signaling has been in-tensively studied. However, TLR5-associated signaling in non-transformed colonocytes has not been investi-gated. Here we studied the expression of cytokines induced by flagellin in non-transformed human colonic NCM460 cells and the signaling mechanisms mediating these responses. Cytokine expression array experiments showed that exposure of the cells to flagellin (100 ng/ml) for 12 h increased the expression of interleukin (IL)-8 and macrophage-inflammatory protein 3 (cid:1) (MIP3 (cid:1) ) in a TLR5-specific manner. Flagellin also activated MAP kinases (ERK1/2, JNK, and p38) and degraded I (cid:2) B (cid:1) . Dominant negative MEK1 (a kinase that activates ERK1/2) blocked flagellin-stimulated IL-8 and MIP3 (cid:1) transcriptional activity, while the MEK-specific inhibitors PD98059 and U0126 reduced protein production of these cytokines. Conversely, transfection with a constitutively active MEK1 increased IL-8 and MIP3 (cid:1) transcriptional activity in a NF (cid:2)

main and an intracellular Toll/IL-1R (TIR) domain (1). TLRs recognize several pathogen-associated molecular patterns such as Gram-negative bacterial lipopolysaccharide, lipoprotein from Gram-positive bacteria, double-stranded RNA, bacterial hypomethylated DNA, or flagellin and then triggers innate and adaptive immune responses against these pathogens (2). Studies on TLR signaling provided information on several adaptor molecules of TLR such as MyD88 (myeloid differentiation factor 88), TIRAP/Mal (TIR domain-containing adapter protein/ MyD88-adapter-like), TRIF (TIR domain-containing adapter inducing interferon ␤), and TRAM (TRIF-related adaptor molecule) (3,4). These adaptor molecules mediate NFB and MAP kinase activation, leading to proinflammatory cytokine gene expression and immune cell maturation. Among the various pathogen-associated molecular patterns, flagellin, the major component of the flagellar filament, is produced from Gramnegative and Gram-positive bacterial species and stimulate TLR5 leading to cytokine gene expression (5,6). So far MyD88 is the only known immediate downstream adaptor molecule of TLR5, mediating TLR5-associated signaling.
The human gut harbors a large collection of commensal microbes (ϳ500 -1000 bacterial species) at the concentration of ϳ10 11 organisms/ml of proximal colonic contents (7,8). These commensal microbiota are compartmentalized in the intestinal lumen by the intestinal epithelium that serves as an interface between the host and the bacterial milieu. Several clinical observations and animal experiments suggest that intestinal bacteria play a major role in the pathogenesis of chronic bowel inflammation (8 -11). In addition, enteroinvasive pathogens such as Salmonella, Shigella, Yersinia, and Listeria can invade the epithelium and provoke inflammatory responses characterized by secretion of proinflammatory cytokines (12)(13)(14). TLR5 is highly expressed in intestinal epithelial cells lining the gastrointestinal tract (15). Thus TLR5 engagement by flagellin may activate various signal transduction pathways leading to proinflammatory responses.
Here, we used non-transformed human colonic NCM460 cells (19,20) to investigate which cytokines are induced by TLR5 and studied the signaling pathways involved in cytokine expression. We found that exposure of NCM460 cells to flagellin specifically increased IL-8 and MIP3␣ expression, which was blocked by a TLR5-dominant negative mutant. We also present novel evidence indicating that MEK (mitogen-activated protein kinase/extracellular signal-regulated kinase kinase), a kinase involved in ERK1/2 activation, mediates IL-8 and MIP3␣ expression via NFB-independent pathway(s) in flagellin/TLR5induced signal transduction pathways.
Human Cytokine Expression Array Assay-The human cytokine array 5.1 was purchased from Ray Biotech (Norcross, GA) and used following the manufacturer's instructions (27). Briefly, NCM460 cells were plated 1 day prior to the experiment. The cells then were stimulated with flagellin (100 ng/ml) for 12 h followed by washing with 1ϫ phosphate-buffered saline. Cell lysates were then prepared with lysis buffer, and the protein concentration was measured by the Bradford assay system from Bio-Rad. Membranes, immobilized with 79 different capture antibodies, were incubated with the blocking buffer and then incubated with 300 g of total protein for 2 h followed by washing with washing buffer. The membranes were then incubated with biotin-conjugated antibodies for 2 h. After incubation, membranes were washed and then incubated with horseradish peroxidase-conjugated streptavidin for 30 min. Unbound reagents were removed by washing, and the bound antibodies on the membranes were visualized using the enhanced chemiluminescence system (Pierce).

IL-8 and MIP3␣
Measurements-An enzyme-linked imunosorbent assay (ELISA) was performed to measure the level of human IL-8 and MIP3␣ using the appropriate kits from R&D Systems (Minneapolis, MN) and BIOSOURCE International (Camarillo, CA), respectively. Briefly, NCM460 cells were plated 1 day prior to the experiment and then stimulated with flagellin (100 ng/ml) for 12 h. The cultured supernatants in turn were used for ELISA following the manufacturer's protocol. All assays were performed in triplicate, and a single representative experiment is shown. Data are expressed as mean values ϩ S.E.
Immunoblot Analysis-Cells were harvested and washed once with ice-cold 1ϫ phosphate-buffered saline (pH 7.5) and then lysed for 60 min on ice in lysis buffer (150 mM NaCl, 50 mM Tris-Cl (pH 8.0), 5 mM EDTA, 1% Nonidet P-40) with a protease inhibitor mixture (Roche Applied Science) and a phosphatase inhibitor mixture (Sigma). Cell lysates were clarified by centrifugation at 4°C for 10 min at 12,000 ϫ g. Protein concentrations of the lysates were measured by the Bradford method (Bio-Rad), and equal amounts of total protein were fractionated on 10% SDS-polyacrylamide gels. Gels were transferred to polyvinylidene difluoride membranes. The membranes were blocked with Trisbuffered saline containing 0.05% Tween 20 and 5% nonfat dry milk and then incubated with the indicated antibodies and an appropriate horseradish peroxidase-conjugated secondary antibody as described elsewhere (25,28). Bound antibodies were visualized using the enhanced chemiluminescence system (Pierce).
Transfection and Luciferase Reporter Assays-NCM460 cells were plated in 6-well plates (0.4 ϫ 10 6 cells/well) and transfected with the appropriate plasmid DNA, including a ␤-galactosidase expression plasmid (HSP70-␤-gal) as an internal control, using TransIT-LT1 from Mirus (Madision, WI) according to the manufacturer's instructions. One day after transfection, cells were stimulated with flagellin (100 ng/ml) for 6 h, and the relative luciferase activity was determined by normalization with ␤-galactosidase activity as described previously (25,28). The total amount of plasmid DNA was kept consistent by adding the empty vector for each transfection. All assays were performed in triplicate, and a single representative experiment is shown. Data are expressed as mean values ϩ S.E.
Immunoprecipitation-NCM460 cells (2.5 ϫ 10 6) were plated in 60-mm dishes and stabilized overnight. Cells were stimulated with flagellin (100 ng/ml) for various time points, and cell lysates were prepared in lysis buffer as described above. Equal amounts of protein were mixed with the IRAK antibody, incubated at 4°C for 3-4 h, and protein A beads (Pierce) were added and incubated overnight. The precipitated mixture was washed three times with lysis buffer; proteins were fractionated on 10% SDS-polyacrylamide gel, followed by immunoblotting as described above.

Cytokine Microarray Assay Revealed the Selective Expression of IL-8 and MIP3␣ by TLR5 Engagement in Non-transformed
Human Colonocytes-Understanding the expression array of proinflammatory mediators by bacterial flagellin is important to study the pathophysiology of inflammatory responses in the human gut. Therefore, using a human cytokine microarray assay, we first screened the expression profile of cytokines from non-transformed human NCM460 colonocytes stimulated with flagellin. The microarray membrane was immobilized with capture antibodies against 79 different cytokines as listed in Fig.  1A. The membranes were hybridized with 300 g of total protein from flagellin-stimulated (100 ng/ml, for 12 h) or control NCM640 cells. The corresponding proteins were detected by a mixture of detection antibodies and visualized by an ECL system. Among various cytokines, the expression of IL-8 and MIP3␣ was substantially up-regulated (Fig. 1B). Subsequently, the supernatant from flagellin-stimulated cells was used for ELISA to confirm the increased expression of IL8 and MIP3␣ observed with the cytokine microarray assay. As shown in Fig.  1, C and D, IL-8 and MIP3␣ production detected by ELISA was substantially increased in response to the TLR5 agonist. These Cell lysates were prepared, and the human cytokine microarray assay was performed with 300 g of total cell lysate by following the manufacturer's protocol. The culture supernatant from the sample for B was harvested and used for ELISA to measure the level of MIP3␣ (C) and IL-8 (D) as described under "Materials and Methods." NCM460 cells were transiently co-transfected with IL-8 (E)-or MIP3␣ (F)-luciferase reporter plasmid (1 g) and the dominant negative TLR5 expression plasmid (TLR5(⌬TIR), 2-4 g) in the indicated combinations. One day after transfection, the cells were stimulated with flagellin (100 ng/ml) (E and F) for 6 h followed by a luciferase assay as described under "Materials and Methods." The empty vector was added as necessary to keep the consistent amount of total plasmid DNA for each transfection. The human cytokine array experiment was performed three times and a single representative experiment is shown. Transfections were performed in triplicate, and a single representative experiment is shown. Data are reported as mean values Ϯ S.E. (n ϭ 3). results strongly suggest that, after 12 h incubation, flagellin stimulation selectively up-regulates IL-8 and MIP3␣ in human colonocytes.
Additionally, we observed that mRNA for TLR5 is also expressed in human colonic epithelial NCM460 cells. 2 Although flagellin has been suggested as a specific ligand for TLR5, it is necessary to verify if IL-8 and MIP3␣ expression induced by flagellin is specifically mediated by TLR5 engagement and not by putative contaminants present in the purified flagellin preparation. To test this, we used a dominant negative TLR5 mutant (cytoplasmic TIR domain deleted TLR5(⌬TIR)). Cells were transiently co-transfected with IL-8 or MIP3␣ luciferase reporter constructs with or without dominant negative TLR5 mutant. As shown in Fig. 1, E and F, ⌬TIR overexpression substantially inhibited flagellin-induced IL-8 and MIP3␣ transcriptional activity. In contrast, we confirmed that overexpressing dominant negative TLR5 mutant did not alter TNF␣-mediated responses (data not shown), indicating the TLR5 specificity of this response. Together, these findings indicate that, in human colonocytes, flagellin stimulation selectively up-regulates IL-8 and MIP3␣ expression in a TLR5-specific manner.
Engagement of TLR5 Results in the Activation of MAP Kinases and NFB Signaling-Previous results in macrophages demonstrate that TLR signaling leads to the activation of MAP kinases and the transcription factor NFB (2). Subsequent studies suggested that bacterial infection with Salmonella or Escherichia coli results in NFB and p38 activation in colon adenocarcinoma HT-29 and Caco-2 cells (16,17,29). Therefore, we tested whether TLR5 engagement can give rise to a typical TLR-related signaling response. NCM460 cells were stimulated with flagellin as indicated in Fig. 2 and whole cell lysates were then prepared and analyzed by immunoblot analysis. Activation of MAP kinases was determined using specific an-tibodies against phospho-ERK1/2, phospho-JNK, and phospho-p38. As shown in Fig. 2, flagellin induced activation of ERK1/2 (A), JNK (B), and p38 (C) at early time points. Moreover, flagellin exposure triggered degradation of IB␣ (D). Thus, engagement of TLR5 leads to rapid activation of MAP kinases and NFB in non-transformed human colonocytes, indicating the presence of intact TLR5 signaling in the cells.
Blocking MEK1 Inhibited Flagellin/TLR5-induced IL-8 and MIP3␣ Expression-Based on ample evidence suggesting involvement of several cytokines in the pathogenesis of bowel inflammation (30,31), it is important to study the mechanism by which the inflammatory mediators are up-regulated in response to inflammatory stimuli, such as flagellin. Previous studies (12,13,16) showed that the expression of IL-8 and MIP3␣ by pathogenic bacteria was attenuated by the overexpression of IB␣, indicating that expression of these cytokines by enteric pathogens is NFB-dependent.
However, as shown in Fig. 2, apart from NFB activation, TLR5 signaling also leads to activation of MAP kinases. Therefore, we next determined the role of MAP kinase activation in TLR5-derived IL-8 and MIP3␣ gene expression. Since TLR5 engagement induces ERK1/2 activation, which is phosphorylated on its threonine and tyrosine residues by the MEK1/2 class of MAP kinase kinases, we stimulated NCM460 cells with flagellin in the presence or absence of the MEK-specific inhibitors U0126 or PD98059 and then measured IL-8 and MIP3␣ protein in culture supernatants by ELISA. As shown in Fig. 3, A-D, pharmacologic antagonism of MEK activity inhibited IL-8 and MIP3␣ protein production in response to flagellin in dosedependent manner.
To directly examine whether flagellin stimulation induces MEK1/2 activation in NCM460 cells we performed immunoblot analysis in cell lysates from flagellin-stimulated cells using an antibody directed against phospho-MEK1/2. As shown in Fig.  3E, flagellin stimulation induced MEK1/2 activation at early time points. Since MEK1/2 is the upstream enzyme to directly activate ERK1/2, this result is compatible to flagellin-induced ERK1/2 activation ( Fig. 2A).
To obtain further evidence for the involvement of MEK in TLR5-induced IL-8 and MIP3␣ expression, we next determined whether a dominant-negative MEK1 mutant could influence IL-8 or MIP3␣ transcriptional activity in flagellin-stimulated human colonocytes. NCM460 cells were transiently co-transfected with the IL-8-or MIP3␣-luciferase reporter plasmid with or without an expression plasmid encoding for a dominant negative MEK1 (MEK1-DN) mutant. As shown in Fig. 3, F and G, overexpression of a dominant negative MEK1 mutant in these cells resulted in the complete inhibition of IL-8 transcriptional activities, whereas MEK1-DN expression partially inhibited flagellin-induced MIP3␣ promoter activity. Together, these results suggest that flagellin-induced IL-8 and MIP3␣ expression is, at least in part, mediated by activation of MEK.
Constitutive MEK1 Activation Induces IL-8 and MIP3␣ Gene Expression-The finding that MEK activation is required for the induction of IL-8 and MIP3␣ gene expression by flagellin in NCM460 cells raises the possibility that activated MEK1 can up-regulate IL-8 and MIP3␣ gene expression in the absence of any upstream stimulation. To address this issue, a plasmid encoding for constitutively active MEK1 (MEK1-CA) was cotransfected with IL-8 or MIP3␣ luciferase reporter constructs into NCM460 cells and luciferase activity was measured. As shown in Fig. 4A, IL-8 promoter activity was substantially increased by the specific activation of MEK1 without flagellin stimulation. Surprisingly, the level of increase in IL-8 promoter activity in cells overexpressing constitutively active MEK1 alone is higher compared the level obtained following flagellin exposure (100 ng/ml for 6 h, Fig. 4A, lane 3). This finding suggests that activation of MEK1 is sufficient for IL-8 gene expression in non-transformed human colonocytes. However, compared with IL-8, MIP3␣ promoter activity was weakly induced by overexpressing the MEK1-CA construct (Fig. 4B). Moreover, we also demonstrate that overexpression of MEK1-CA alone induces endogenous IL-8 and MIP3␣ protein production (Fig. 4, C and D). Together, our findings suggest that MEK1 activation induces IL-8 and MIP3␣ expression following flagellin/TLR5 engagement.

MEK1-induced IL-8 Transcriptional Activity Is Mediated in a NFB-independent Manner-
The transcription factor NFB is known to be involved in proinflammatory cytokine gene expression in various cell types (28). Previous results indicate that overexpression of IB␣ to inhibit NFB activation inhibited IL-8 gene expression in colon carcinoma cells stimulated with flagellin-expressing enteropathogenic bacteria (12,13,16). Thus, it is necessary to determine whether MEK1-mediated IL-8 expression in response to flagellin involves NFB activation. To test this possibility, an IB␣ expression construct was transiently co-transfected with an IL-8 reporter construct in NCM460 cells, in the presence/absence of MEK1-CA. As shown in Fig. 5A, overexpressing IB␣ did not alter MEK1-derived IL-8 transcriptional activity. In contrast, flagellin-induced IL-8 transcriptional activity was abolished by IB␣ overexpression (Fig. 5B). Thus, in NCM460 colonocytes, MEK1 activation appears to induce IL-8 transcriptional activity in a NFB-independent manner. Moreover, as shown in Fig. 5C, transfection with the MEK1-CA construct does not result in NFB activation. Together, these results demonstrate that, in non-transformed human colonocytes, MEK signaling is a parallel to the NFB signaling molecular event in response to flagellin/TLR5 engagement and mediates IL-8 expression in a NFB-independent manner.

TRAF6 Is an Upstream Mediator of MEK for IL-8 and MIP3␣
Expression-TNF receptor-associated factors are identified as downstream signal transducers of the TNF receptor family. TRAF6 has been determined as a downstream signaling molecule in TLRs and IL-1R engagement, which is associated with transforming growth factor ␤-activated kinase 1 (TAK1) via the TAK1-binding protein to activate NFB-inducing kinase leading to NFB activation (32). Kopp et al. (33) identified a novel TRAF6-binding protein, ECSIT (evolutionarily conserved signaling intermediate in Toll pathways), which interacts with the MEKK1 and thereby links TRAF6 to MEKK1, leading to NFB and AP-1 activation in TLR4 response. More recently, Huang et al. (34) demonstrated that MEKK3 associates with TRAF6 to link MAP kinase (JNK and p38, but not ERK1/2) activation in response to lipopolysaccharide or IL-1. Furthermore, several studies already showed that TRAF6 associates with IRAK following the engagement of TLRs and IL-1 receptor (35)(36)(37)(38). Based on this evidence, we hypothesized that TRAF6 is involved in the flagellin/TLR5 signal transduction pathway, functions as a bridge to MEK1 activation, and mediates IL-8 and MIP3␣ gene expression following TLR5 engagement. To test this, IRAK was co-immunoprecipitated in flagellin-stimulated cell lysates, and co-precipitation of TRAF6 was determined by immunoblot analysis using a TRAF6 antibody. As shown in Fig. 6A, flagellin stimulation induced the association of IRAK and TRAF6. Additionally, we transfected the cells with a TRAF6 expression plasmid together with an IL-8 or MIP3␣ reporter construct. In some cases, MEK1-DN was included in the indicated combination. As shown in Fig. 6, B and C, overexpression of TRAF6 induced IL-8 and MIP3␣ transcriptional activity without flagellin/TLR5 stimulation. However transfection with the MEK1-DN construct completely inhibited IL-8 transcriptional activity induced by TRAF6, whereas TRAF6derived MIP3␣ promoter activity was partially inhibited by overexpressing MEK1-DN. These results indicate that TRAF6 is an upstream mediator of MEK1 mediating IL-8 and MIP3␣ expression in NCM460 cells.
Coiled-coil and TRAF-C Regions of TRAF6 Mediate TLR5induced MEK1-dependent Signaling-TRAF6 is characterized by a N-terminal Ring finger and five zinc finger domains together with a C-terminal coiled-coil (TRAF-N) and TRAF-C domains (23). TRAF6 lacking the N-terminal Ring finger and zinc finger domains acts as a dominant negative mutant and blocks NFB activation in response to lipopolysaccharide or IL-1 (23,39). We found that transfection with TRAF6-CC (TRAF6 lacking N-terminal Ring finger and zinc finger domains but harboring C-terminal coiled-coil and TRAF-C domains) inhibited flagellin-induced NFB activation (Fig. 7A), indicating that TRAF6 participates in this TLR5 signaling response. Based on this evidence, we next studied whether TRAF6 transduces the upcoming TLR5 response to MEK-mediated cytokine expression without activating the NFB pathway. To test this, cells were co-transfected with IL-8 or MIP3␣ reporter constructs, together with TRAF6-CC, and/or MEK1-DN constructs in combinations indicated in Fig. 7. Our results show that overexpressing TRAF6-CC, which is unable to mediate NFB activation, did not inhibit flagellin-induced IL-8 and MIP3␣ promoter activity. IL-8 and MIP3␣ promoter activities, however, were substantially inhibited in the presence of MEK1-DN. These data demonstrate that the C-terminal coiled-coil and TRAF-C regions of TRAF6 mediate IL-8 and MIP3␣ expression via MEK1. Taken together, our results strongly suggest that MEK mediates flagellin/TLR5-induced IL-8 and MIP3␣ expression in an NFB-independent manner. DISCUSSION An important finding of our study is the novel demonstration of specific flagellin/TLR5-induced proinflammatory responses in non-transformed human colonic epithelial cells. Among various cytokines, TLR5 engagement in human colonocytes stimulates the specific expression of the potent chemoattractants IL-8 and MIP3␣. Flagellin-stimulated TLR5 is also linked to activation of several MAP kinases, including ERK1/2, p38, and JNK, and of the transcription factor NFB. Activation of MEK, an upstream kinase of ERK1/2, is an important step for flagel-linTLR5-induced IL-8 and MIP3␣ gene expression in these cells. In contrast, flagellin/TLR5-derived IL-8 and MIP3␣ expression does not involve activation of JNK.
Our results may be pertinent to the pathophysiology of colonic inflammation. Increased MIP3␣ and IL-8 expression has been observed in colonic tissues obtained from patients with inflammatory bowel disease (30,31). Several reports underlie the importance of MIP3␣ and IL-8 in inflammatory cell recruitment during colonic inflammation. For example, MIP3␣ attracts CCR6-expressing memory T cells and immature DCs, and IL-8 is a chemoattractant for neutrophils in the subepithelial compartment (12,13,24). Thus, IL-8 and MIP3␣ participate in infiltration of immune cells to the subepithelial compartment seen during intestinal inflammatory responses.
Previous studies in colon carcinoma HT-29 and Caco-2 cells showed that overexpression of IB␣ significantly reduced IL-8 and MIP3␣ expression in response to flagellin-containing bacteria (12,13,16). Using non-transformed human colonocytes, we also found that the NFB/IB system is involved in flagel- lin-stimulated IL-8 and MIP3␣ promoter activity (see "Results"). Consistent with these findings, previous studies demonstrated the importance of NFB binding sites on the promoter regions of IL-8 and MIP3␣ for transcriptional activation of these genes (24,40,41). Thus, NFB appears to play an important role in IL-8 and MIP3␣ expression following exposure to flagellin-containing bacteria.
Using molecular as well as pharmacologic approaches we provide strong evidence that NFB-independent MEK-mediated signaling pathways are also involved in flagellin/TLR5induced IL-8 and MIP3␣ expression. MEK-associated signaling has been linked to activation of several, other than NFB, transcription factors regulating IL-8 and MIP3␣ gene expression. For example, DNA binding sites for the transcription factors AP-1 and C/EBP (NF-IL6) have been identified within the IL-8 promoter (40). Moreover, in gastric cancer cells, MEK signaling induces the transactivation of SRE and AP-1 in response to Helibacter pylori and inhibition of MEK activation reduces H. pylori-induced IL-8 secretion, indicating that MEK participates in IL-8 expression in response to this pathogen (41). MEK signaling is also related to activation of transcription factors such as Elk-1 and C/EBP in human monocytes and colonic adenocarcinoma T84 cells, respectively (18,42). In addition, in transformed endothelial cells, IL-8 gene expression induced by TNF␣ is regulated by AP-1 and C/EBP (43). Moreover, bindings sites for Sp1, the Ets nuclear factor ESE-1, and p50/p65 NFB heterodimers have been identified within the 5Ј-flanking region of MIP3␣ and substitutions in these elements alter MIP3␣ promoter activity induced by IL-1 in colon adenocarcinoma Caco-2 cells (24).
We found that TLR5-induced expression of IL-8 and MIP3␣ was not altered by exposure of cells to the specific JNK inhibitor SP600125 and by overexpressing a dominant negative JNK, 2 implying that JNK does not mediate the up-regulation of IL-8 and MIP3␣ gene expression by TLR5 engagement. Our results also indicate increased p38 activation following exposure of NCM460 cells to flagellin (Fig. 2), consistent with prior observations in colon adenocarcinoma HT-29 cells (17). Although we did not directly examine involvement of p38 in flagellin-induced IL-8 and MIP3␣ expression in NCM460 cells, Yu et al. (17) presented evidence indicating that p38 activation is functionally linked to IL-8 production in response to flagellin activation in HT-29 cells.
Several studies indicate participation of TRAF6 in other than TLR5, TLR-related signaling. Here we present direct evidence that in human colonocytes TRAF6 mediates flagellininduced TLR5 proinflammatory signaling (Figs. 6 and 7) and that the C-terminal coiled-coil and TRAF-C domains of TRAF6 (that lack the N-terminal Ring finger and zinc finger regions of TRAF6) relay the upcoming signal from TLR5 to MEK activation, leading to IL-8 and MIP3␣ gene expression (Fig. 7). Thus, in non-transformed human colonocytes, the C-terminal coiledcoil and TRAF-C domains of TRAF6 are responsible for linking TLR5 signaling to MEK1 activation and MEK-mediated IL-8 and MIP3␣ expression in a NFB-independent manner. However, the molecule(s) involved in TRAF6-MEK signaling pathway in flagellin-stimulated human colonocytes have not been examined in our study. Previous results indicate that the TRAF6-associating adaptor molecule ECSIT, which links MEKK1 to TRAF6, leads to JNK and NFB activation in response to TLR4 engagement (33). More recently, MEKK3 has been suggested to interact with TRAF6, resulting in activation FIG. 6. TRAF6 is the upstream modulator of MEK1, leading to IL-8 and MIP3␣ transcriptional activation. NCM460 cells were stimulated with flagellin (100 ng/ml) for the indicated times, and total cell lysates were prepared with lysis buffer. The same amount of total protein was mixed with IRAK antibody for 3-4 h, and then protein A beads were added for overnight incubation at 4°C. Co-precipitation of TRAF6 was determined by immunoblot analysis using a TRAF6 antibody, as described under "Materials and Methods" (A). NCM460 cells were transiently co-transfected with a luciferase reporter plasmid under the control of the IL-8 reporter plasmid (B) or MIP3␣ reporter construct (C) (1 g) and either TRAF6 expression plasmid (TRAF6 wt, 2 g) or empty vector (2 g). In some cases, a dominant negative MEK1 (MEK1-DN) expression construct (2 g) was included in the indicated combinations. The HSP70-␤-gal reporter construct (0.5 g) was included as the internal control. Transfected cells were cultivated for a day, and then cell lysates were prepared for luciferase assay. Transfections were performed in triplicate, and a single representative experiment is shown. Data are reported as mean values Ϯ S.E. (n ϭ 3). IP, immunoprecipitation; IB, immunoblot. of p38 and JNK, but not ERK1/2 in response to TLR4 activation or IL-1 exposure (34). However, both ECSIT and MEKK3 are known to directly interact with TRAF6, but not with the Nterminally truncated form of TRAF6 (33,34). Thus, it is quite likely that another molecule(s) mediates TRAF6-induced MEK activation in response to TLR5 engagement.
Our findings demonstrate that a TRAF6-mediated MEK-dependent pathway, that does not involve NFB activation, as well as TRAF6-NFB-dependent pathways are important modulators of IL-8 and MIP3␣ gene expression following TLR5 engagement in non-transformed human colonocytes. Thus, TRAF6-mediated activation of MEK and NFB-dependent pathways synergize to regulate expression of IL-8 and MIP3␣ genes in response to flagellin/TLR5 engagement in human colonocytes. We hypothesize that similar pathways may be important in the pathogenesis of several forms of colonic inflammation, including inflammatory bowel disease.