Distinct mechanism of Helicobacter pylori-mediated NF-kappa B activation between gastric cancer cells and monocytic cells.

NF-kappaB is a critical regulator of genes involved in inflammation. Gastric epithelial cells and macrophages are considered the main sources of pro-inflammatory cytokines. We investigated NF-kappaB activation by Helicobacter pylori in MKN45 gastric epithelial cells and THP-1 monocytic cells. Although, cag pathogenicity island (PAI)-positive H. pylori (wild type) activated NF-kappaB in both cells, isogenic mutant of cagE (DeltacagE) activated it only in THP-1 cells. Supernatant from the wild type culture could activate NF-kappaB in THP-1 cells but not in MKN45 cells. High density cDNA array analysis revealed that mRNA expression of NF-kappaB-regulated genes such as interleukin (IL)-8, tumor necrosis factor-alpha (TNFalpha), and IL-1beta was significantly up-regulated by the wild type in both cells, whereas it was up-regulated by DeltacagE only in THP-1 cells. Experiments using CD14-neutralizing antibody and IL-1 receptor-associated kinase (IRAK) assay showed that both wild type and DeltacagE H. pylori activated NF-kappaB through CD14 and IRAK in THP-1 cells but not in MKN45 cells. Macrophages from C3H/HeJ mice carrying point mutation in the Toll-like receptor 4 (TLR4) gene showed decreased NF-kappaB activation and TNFalpha secretion compared with C3H/HeN mouse macrophage when treated with H. pylori. In conclusion, H. pylori-induced NF-kappaB activation in epithelial cells is dependent on cag PAI and contact but does not involve CD14 and IRAK, whereas in macrophage/monocytic cells it is independent of cag PAI or contact but involves CD14 and TLR4.

Infection with Helicobacter pylori is strongly associated with gastroduodenal diseases such as chronic active gastritis, gastroduodenal ulcers, and gastric malignancies (1)(2)(3). Although these organisms do not appear to invade the gastric epithelium, the infection induces chronic inflammation in the lamina propria. Although the pathogenesis of gastroduodenal diseases caused by this bacterium is not well understood, some immunologic mechanisms are considered to be involved in gastric inflammation. It has been demonstrated recently that H. pylori affects intracellular signal conduction in the cells, leading to the activation of transcriptional factors (4 -8) and the induction of pro-inflammatory cytokines (9 -11).
NF-B is a critical regulator of genes involved in inflammation, cell proliferation, and apoptosis (12,13). In most resting cells, NF-B is bound to the inhibitory IB proteins (IB␣ and IB␤) and is present in the cytoplasm (14). However, NF-B enters the nucleus in response to certain stimuli, including pro-inflammatory cytokines such as tumor necrosis factor-␣ (TNF␣), 1 interleukin-1 (IL-1), and bacterial lipopolysaccharide (LPS) (15). The activation of NF-B requires phosphorylation of IBs, which results in their ubiquitin-dependent degradation. The released NF-B dimers enter the nucleus and activate the genes (15). Two upstream kinase, IB kinase ␣ (IKK␣) and ␤ (IKK␤), phosphorylate IB␣ in response to the pro-inflammatory cytokines. NF-B-inducing kinase, NIK, is a member of the MEKK (mitogen-activated protein kinase/extracellular signal-regulated kinase kinase kinase) family and has been shown to activate both IKK␣ and IKK␤ through interaction with the adapter proteins associated with the receptors for TNF␣ and IL-1 (16 -19). The adapter proteins, TRAF2 and TRAF6, belong to the TNF receptor-associated factor (TRAF) family and act as effectors for TNF␣ and IL-1/Toll-like receptors, respectively (20,21). The potential pathogenic bacteria are able to deposit their toxic and pro-inflammatory constituents such as LPS derived from the outer membrane of pathogenic Gram-negative bacteria. Recently, several Toll-like receptors (TLR) have been identified in blood macrophages and monocytes based on homology to the Drosophila protein (22,23). One of them, TLR4, acts as trans-membrane coreceptor to CD14 in the cellular response to LPS (24,25). Positional cloning of the Lpsd locus, which is responsible for LPS hyposensitivity in C3H/HeJ mice, mapped it to the tlr4 gene, which encodes a TLR4 protein with a single point mutation (P712H) (26). TLR4 has a leucine-rich ectodomain and a cytoplasmic domain with sequence homology to several mammalian proteins such as IL-1 receptor and IL-18 receptor. It associates with CD14 on the cell surface and induces NF-B activation thorough IL-1 receptor signaling molecules, namely MyD88, IRAK, and TRAF6 (27,28). We have recently shown that H. pylori activates NF-B through a signaling pathway involving IKKs, NIK, TRAF2, and TRAF6 in gastric cancer cells. However, the products of H. pylori, which lead to the recruitment of the adapter proteins TRAF2/TRAF6, remain to be identified (8).
The cag pathogenicity island (PAI), an ϳ40-kilobase region of possibly extraneous origin, is present in about 50 -60% of H. pylori isolates in Western countries and in more than 90% in Japan (29 -32). The proteins encoded by genes located in cag PAI are suggested to function as secretion machinery (type IV)-exporting molecules possibly involved in H. pylori-host cell interaction. Recently, it has been reported that CagA protein, present in PAI, can be tyrosine-phosphorylated in the gastric epithelial cells (33,34). Isogenic mutant studies have demonstrated that some of proteins encoded by cag PAI genes are essential for NF-B activation in the gastric epithelial cells (35). Recently, we have demonstrated that cag PAI is responsible for various gastric lesions in the gerbil model (36). These observations suggest that cag PAI may play an important role in the pathogenesis of gastric diseases related to H. pylori infection.
The proinflammatory cytokines are considered to be secreted from both gastric epithelial cells and infiltrating cells such as macrophages and lymphocytes (37,38). It is not known whether the mechanism of NF-B activation elucidated by us in gastric epithelial cells also functions in other cell types.
In this study, we analyzed the molecular mechanism of H. pylori-mediated NF-B activation in both a gastric epithelial cell line, MKN45, and in a monocytic cell line, THP-1. Our results indicate that the mechanism of NF-B activation is different in the two cells and that H. pylori-mediated NF-B activation in monocytic/macrophage cells involves TLR4 but is independent of cag PAI.

EXPERIMENTAL PROCEDURES
Bacterial Strains-The H. pylori isolates used in this study are as follows. TN2, a strain positive for CagA and VacA (vacuolating cytotoxin), was generously given by Dr. Nakao (Takeda Chemical Industries, Ltd., Osaka, Japan). Southern blot analysis confirmed the presence of at least 13 cag PAI genes. Infection with this strain induced gastric cancer in Mongolian gerbils (39). Isogenic cagE-negative mutants (TN2-⌬cagE), PAI totally deleted mutants (TN2-⌬PAI), vacA-negative mutants (TN2⌬vacA), and HP0638-negative mutants (TN2⌬0638) were prepared by the insertion of a kanamycin-or chloramphenicol-resistant gene into a locus of the TN2 genome as described previously (40).
H. pylori strains were cultured on Columbia agar with 5% (v/v) horse blood and Dent antibiotic supplement (Oxoid, Basingstoke, UK) at 37°C for 3 days under microaerobic conditions (Campy-Pak Systems, Becton Dickinson, BBL, Cockeysville, MD). The isolates were kept at Ϫ80°C in Brucella broth with 5% (v/v) fetal bovine serum (FBS) containing 16% (v/v) glycerol. In co-culture experiments, H. pylori was co-cultured with cells in Brucella broth containing 7.5% FBS for 24 h, centrifuged, resuspended in cell culture medium containing 10% FBS, and immediately thereafter applied to assays. The range of the bacteria: cancer cell ratio was 50:1 to 75:1 in all of the experiments. Bacterial supernatant was prepared by the culture of H. pylori (10 7 /ml) at 37°C in cell culture medium with 10% FBS under microaerobic conditions for 24 h. The culture media were centrifuged at 10,000 ϫ g for 20 min, and the cell-free supernatants were sterilized by passage through a 0.2-mpore size filter (Millipore, Tokyo).

FIG. 1. Specific protein binding activities of NF-B sequences (EMSA).
A, the nuclear extracts were prepared from MKN45 and THP-1 cells. Cells were untreated or treated with H. pylori (TN2 and ⌬cagE) or 10 g /ml LPS for 90 min. Four g of nuclear extracts were incubated with 32 P-labeled oligonucleotide for 30 min. DNA-protein complexes were separated on 4% polyacrylamide gel. The migration of the DNA-protein complex containing NF-B is indicated. This complex was found to be specific as judged by super-shifting antibodies against p50 and p65 and unlabeled NF-B competitor oligonucleotides (data not shown). B, the nuclear extracts were prepared from TMK-1 and J774A.1 cells. Cells were treated, and EMSA was performed as described under "Experimental Procedures." C, H. pylori culture supernatant was prepared as described under "Experimental Procedures." Cells were treated with live H. pylori and culture supernatant for 90 min. EMSA was performed as described under "Experimental Procedures." D, culture supernatants from various H. pylori knockout mutants (⌬cagE, ⌬PAI, ⌬vacA, ⌬HP638) were prepared, and NF-B activation was analyzed by EMSA. chased from Monosan (Uden, Netherlands), rabbit anti-IRAK polyclonal antibody from Upstate Biotechnology (Lake Placid, NY), polyclonal anti-IB␣, p50, and p65 antibodies from Santa Cruz Biotechnology (Santa Cruz, CA), and myelin basic protein and lipopolysaccharide (Escherichia coli serotype 0111:B4) from Sigma.
Cell Lines-Human gastric cancer cell lines MKN45 and TMK-1 were maintained in RPMI 1640 containing 10% FBS, L-glutamine, 100 units of penicillin-G, and 100 g/ml streptomycin. MKN45 was obtained from the Riken Gene Bank (Tsukuba, Japan). TMK-1 was provided by Dr. E. Tahara (Hiroshima University, Hiroshima, Japan). MKN45 and TMK-1 were established from poorly differentiated gastric adenocarcinoma. Human monocytic cell line THP-1 and mouse macrophage cell line J774A.1 were maintained in RPMI 1640 containing 10% FBS, L-glutamine, 100 units of penicillin-G, and 100 g/ml streptomycin. THP-1 was obtained from Health Science Research Resources Bank (Osaka, Japan). J774A.1 was obtained from the Riken Gene Bank (Tsukuba, Japan). THP-1 was established from a monocytic leukemia.
Preparation of Murine Macrophage-Six-week-old C3H/HeN and C3H/HeJ male mice were obtained from Japan SLC (Hamamatsu, Japan) as specific pathogen-free animals. The mice were sacrificed, and their peritoneal resident macrophages were collected by washing the peritoneal cavity with 5 ml of ice-cold saline. After washing with phosphate-buffered saline, the macrophages (1 ϫ 10 6 cells) were seeded onto plastic Petri dishes in Ham's F-12 medium (Sigma) supplemented with 10% heat-inactivated FBS. After incubation at 37°C for 1 h, nonadherent cells were removed by repeated washing. The culture medium was replaced with fresh medium either with or without H. pylori or LPS, and the cells were then incubated at 37°C for the indicated period.
Electrophoretic Mobility Shift Assay-Nuclear extracts were prepared by using a nonionic detergent method as described previously (41). NF-B was assayed with a [ 32 P]dATP-labeled oligo probe containing the NF-B recognition site purchased from Promega (Madison, WI). The DNA binding reactions were performed at room temperature for 30 min in a 10-l mixture consisting of 4% glycerol, 1 mM MgCl 2 , 0.5 mM EDTA, 0.5 mM dithiothreitol (DTT), 50 mM NaCl, and 0.5 g of poly(dI-dC). Supershift analysis was performed using antibodies against p65 and p50. DNA-protein complexes were loaded onto a chilled 4% nondenaturing acrylamide gel. Gel electrophoresis was executed in 0.5ϫ Tris borate-EDTA at 4°C. The gel was dried, and autoradiography was performed using a Fujix bio-imaging analyzer FLA3000 (Fuji Photo Film, Tokyo, Japan).
Analysis of mRNA Expression Using cDNA Arrays-Total cellular RNA was extracted using the acid guanidine thiocyanate-phenol-chloroform method according to the manufacturer's instruction (Isogen, Nippon Gene, Tokyo, Japan). Poly(A) mRNA was purified from total RNA using an Oligotex kit (Takara, Tokyo, Japan). Analysis of mRNA expression was performed by radioactively labeled cDNA on high density arrays of membrane-bound cDNA probes. We used the Atlas Cytokine/Receptor cDNA expression array from CLONTECH Laboratories (Palo Alto, CA). The preparation of radioactively labeled cDNA and the subsequent hybridization to cDNA arrays were conducted as recommended by the manufacturer's instructions. Analysis of mRNA Expression by RT-PCR-Total cellular RNA was extracted using acid guanidine thiocyanate-phenol-chloroform method according to the manufacturer's instruction (Isogen, Nippon Gene). Reverse transcription (RT) was used to generate cDNA using the Su-perScript II preamplification system (Invitrogen). The mRNA expression of MIP-1␣, MIP-1␤, TNF␣, IL-1␤, inhibin A, CD14, TLR2, or TLR4 was determined by RT-PCR using the primers described below. The PCR primers (sense (S) and antisense (AS)) used in this study are as follows: The primers for glyceraldehyde-3-phosphate dehydrogenase were obtained from Maxim Biotech, Inc. (San Francisco). The temperature and time schedule were as follows: incubation for 5 min at 96°C followed by 35 cycles at 96°C for 20 s, at 58 -62°C for 1 min, and at 72°C for 1 min. PCR products were analyzed by electrophoresis on 1.2% agarose/ethidium bromide gels.
Nuclease Protection Assay-Human chemokine mRNAs were detected with the multiprobe RNase protection assay system from PharMingen (San Diego, CA). In brief, a mixture of [ 32 P]UTP-labeled antisense riboprobes was generated from chemokine template DNAs including lymphotactin, RANTES, IP-10 (␥ interferon-inducible protein 10), MIP-1␤, MIP-1␣, MCP-1 (macrophage chemoattractant protein-1), IL-8, and I-309. The template DNAs for the human housekeeping genes encoding glyceraldehyde-3-phosphate dehydrogenase and a human ribosomal protein, L-32, were also included to confirm equal loading of total RNA. Ten g of total RNA was hybridized with antisense RNA probes overnight at 56°C, and digested with RNases A and T1. The RNA duplexes were run on 5% polyacrylamide gels and scanned autoradiographically using a FLA3000 image analyzer (Fuji Photo Film Co., Ltd., Tokyo). Immunoblot Analysis-Cells, either unstimulated or stimulated with H. pylori for various periods of time or with LPS, were suspended in 50 mM Tris-HCl (pH 7.4) buffer containing 1 mM EGTA, 2 mM DTT, 25 mM sodium ␤-glycerophosphate, 0.1 mM phenylmethylsulfonyl fluoride, and 10 g/ml aprotinin. An equal amount of protein extracts were fractionated by SDS-polyacrylamide gel electrophoresis and transferred electrophoretically to a polyvinylidene difluoride membrane (Amersham Pharmacia Biotech, Buckinghamshire, UK). The membrane was probed with antibodies for IB␣ or IRAK. An ECL detection assay (Amersham Pharmacia Biotech) was performed according to the manufacturer's instructions.
Immunoprecipitation and Kinase Assay-For the analysis of endogenous IRAK kinase activity, 3 ϫ 10 6 cells were incubated with H. pylori or LPS for various periods and lysed in 50 mM Tris-HCl (pH 7.4) buffer containing 1 mM EGTA, 2 mM DTT, 25 mM sodium ␤-glycerophosphate, 0.1 mM phenylmethylsulfonyl fluoride, and 10 g/ml aprotinin. The cell lysates were immunoprecipitated with anti-IRAK and 20 g of protein A-Sepharose. The immunoprecipitates were washed five times in lysis buffer and washed additionally with kinase buffer containing 10 mM HEPES (pH 7.5), 1 mM MgCl 2 , 10 mM ␤-glycerophosphate, 50 M DTT, and 10 mM ATP. After suspension in 20 l of kinase buffer, the immunoprecipitates were incubated using 5 Ci of [ 32 P]ATP with 2 g of myelin basic protein as an exogenous substrate for 30 min at 30°C. The reaction was terminated by adding SDS sample buffer. The samples were analyzed by SDS-polyacrylamide gel electrophoresis followed by signal detection using the FLA3000 system (Fuji Photo Film, Tokyo).
Human IL-8 and Mouse TNF␣ Measurement-The IL-8 and TNF␣ contents in the culture supernatants were measured by ELISA as FIG. 2. A, the respective effect of H. pylori on the expression of mRNA for MIP1␣, MIP1␤, TNF␣, IL-1␤, and inhibin A in MKN45 cells was analyzed by RT-PCR. Cells were co-cultured with the wild type or ⌬cagE for 1 or 3 h. Following stimulation, cells were lysed, and total RNA was prepared. GAPDH, glyceraldehyde-3-phosphate dehydrogenase. B, THP-1 cells were incubated with H. pylori (TN2, ⌬cagE) or the culture supernatant, and total RNA was extracted at 3 h. The ribonuclease protection assay was performed according to the supplier's instructions. Ltn, lymphotactin. C, MKN45 and THP-1 cells were prepared and treated with or without H. pylori (wild-type and ⌬cagE) for the indicated time periods. The supernatant was recovered, and IL-8 concentration was determined by ELISA. The data represent the means from four independent experiments. specified by the manufacturer (BIOSOURCE International). MKN45 cells were routinely maintained in RMPI 1640 supplemented with 10% FBS. Confluent monolayers of MKN45 in 24-well plates were co-cultured with H. pylori or treated with LPS for various periods of time. The supernatants were then aspirated and stored at Ϫ70°C until assayed for IL-8 by ELISA. The concentration of IL-8 was determined using a standard curve obtained with recombinant IL-8. The murine macrophages were prepared as described above. Macrophages in 24-well plates were co-cultured with H. pylori or treated with LPS for 16 h. The supernatants were then aspirated and stored at Ϫ70°C until assayed for TNF␣ by ELISA.
Statistical Analysis-Data were expressed as the mean Ϯ S.D. Statistical analysis was performed using Student t test, two-sided. Differences were considered statistically significant with p Ͻ 0.05.

NF-B Activation in Gastric Epithelial and Monocytic
Cells-We first investigated NF-B activation by H. pylori in MKN45 gastric epithelial cells and THP-1 monocytic cells. The wild type H. pylori activated NF-B in both cell lines, but cagE knockout mutant (⌬cagE) activated NF-B only in THP-1 cells. Furthermore, LPS activated NF-B only in THP-1 cells and not in MKN45 cells (Fig. 1A). These observations indicate that cag PAI is not essential for NF-B activation in monocytic cells. We further examined a gastric cancer cell line, TMK-1, and a murine macrophage cell line, J774A.1. As in MKN45 and THP-1 cells, NF-B activation was seen in co-culture with the wild type but not with ⌬cagE in TMK-1 cells, whereas it was seen in co-culture with both wild and ⌬cagE in J774A.1 cells (Fig. 1B). These results suggest that the mechanism of NF-B activation induced by H. pylori is different between gastric epithelial cells and monocytic cells. We previously reported that H. pylori-mediated NF-B activation in MKN45 cells required direct interaction between H. pylori and cells. However, the supernatant of H. pylori activated NF-B in THP-1 cells but not in MKN45 cells (Fig. 1C). This result indicates that direct interaction is not essential for NF-B activation in THP-1 cells. Furthermore, NF-B activation in THP-1 cells by culture supernatant is not strain-specific (Fig. 1D).
Analysis by High Density cDNA Arrays of mRNA Expression after H. pylori Infection-NF-B is an important transcrip-

FIG. 3. CD14 is associated with H. pylori-mediated NF-B activation in THP-1 cells but not in MKN45.
A, expression of CD14, TLR2, and TLR4 mRNA in MKN45 and THP-1 was analyzed by PCR following reverse transcription. RT-PCR analysis of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) expression was used as control. B, MKN45 and THP-1 cells were treated with or without neutralizing anti-CD14 antibody (MEM-18) for 1 h and incubated with H. pylori (TN2, ⌬cagE) or LPS for 90 min. Nuclear extracts were prepared, and NF-B activation was analyzed by EMSA. C, MKN45 and THP-1 cells were treated with or without neutralizing anti-CD14 antibody (MEM-18) for 1 h and incubated with H. pylori (TN2, ⌬cagE) or LPS for 16 h. Supernatants were collected, and IL-8 concentration was measured by ELISA. The values shown in this figure represent the mean Ϯ S.D. from four independent experiments. Significant differences as compared with TN2 or LPS are indicated as **, p Ͻ 0.05. tional factor in the production of pro-and anti-inflammatory cytokines. To characterize the mRNA expression profile of cytokines in MKN45 and THP-1 cells co-cultured with H. pylori, we used a cDNA microarray (Atlas Human Cytokine/Receptor Array, CLONTECH). We could reproductively detect mRNA expression for 22% (58 of 268) of the genes in the control samples and for 25% (66 of 268) in the wild type (TN2) H. pylori-infected samples at 3 h after infection in MKN45 cells. In THP-1 cells, mRNA expression was detected for 40% (107 of 268) of the genes in the control samples and for 44% (118 of 268) in the wild type (TN2) H. pylori-infected samples at 3 h after infection. Despite the increased dynamic range in these assays, the majority of genes showed relatively little change in mRNA expression after wild type H. pylori infection, with expression ratios between 0.5 and 2. Genes that were upregulated more than 2.0-fold by H. pylori are shown in Table I. Most of them are NF-B-regulated genes such as IL-8, IL-1␤, and TNF␣. Next, we co-cultured TN2-⌬cagE with MKN45 or THP-1 cells. The profiles of mRNA expression were completely different. No gene was found with an expression ratio above 2.5 in MKN45 cells. On the other hand, a gene expression profile similar to that obtained with the wild type H. pylori was observed in THP-1 cells (Table I). These results indicate that the alteration of gene expression induced by H. pylori is dependent on cag status in MKN45 cells but independent in THP-1 cells, as was the case for NF-B activation.
RT-PCR and RNase Protection Assay-To confirm the find-ings obtained by cDNA arrays, we assessed the levels of several mRNAs by RT-PCR. Oligonucleotide primers were designed for several genes showing up-regulation in cDNA array: MIP1␣, MIP1␤, TNF␣, IL-1␤, and inhibin A. In MKN45 cells, upregulation of the mRNAs by wild type H. pylori was confirmed in all tested genes, whereas none of the genes was up-regulated by TN2⌬cagE (Fig. 2A). Because it was difficult to compare mRNA levels by RT-PCR analysis in THP-1 cells, we assessed mRNA level by RNase protection assay. MIP-1␣, MIP-1␤, and IL-8 mRNA were up-regulated by co-culture with wild type TN2 or TN2⌬cagE. Culture supernatant from the wild type also up-regulated the mRNAs (Fig. 2B). Next, MKN45 and THP-1 cells were treated with or without H. pylori (wild-type and ⌬cagE) for the indicated periods, the supernatants were recovered, and the IL-8 concentration was determined by ELISA. IL-8 secretion was induced only by the wild type in MKN45 cells but by either of them in THP-1 cells (Fig. 2C). Involvement of CD14 Protein-In Gram-negative bacteria, LPS is reported to be the most potential inflammatory factor. NF-B activation induced by LPS involves CD14/Toll-like receptors in the signaling pathway. We investigated the mRNA expression of CD14, TLR2, and TLR4 in MKN45 and THP-1 cells. TLR2 and TLR4 were expressed in both cell types, but CD14 expression was detected only in THP-1 cells (Fig. 3A). To explore the role of CD14 in H. pylori-mediated NF-B activation, we used neutralizing CD14 antibody (MEM18). Pretreatment with the antibody suppressed NF-B activation in THP-1 FIG. 4. Involvement of IRAK in H.  pylori-mediated NF-B activation. A, MKN45 and THP-1 cells were treated with either H. pylori or LPS for 1 and 2 h. IRAK was immunoprecipitated using anti-IRAK antibody. The kinase activity of IRAK was measured by immunocomplex kinase assay with myelin basic protein (MBP). B, MKN45 and THP-1 cells were co-cultured with H. pylori (10 7 colonyforming units/ml) or LPS (10 g/ml), and whole cell extracts were prepared at the indicated times. Proteins (10 g) were separated by electrophoresis and immunoblotted with anti-IRAK and anti-IB␣. cells but not in MKN45 cells. NF-B activation by LPS was also suppressed by the treatment. These observations indicated that H. pylori-mediated NF-B activation involves CD14 in THP-1 cells but not in MKN45 (Fig. 3B). We also examined whether CD14-neutralizing antibody had any affect on IL-8 secretion from the cells. IL-8 secretion induced by H. pylori was inhibited in THP-1 cells but not in MKN45 cells (Fig. 3C).
Analysis of IRAK Kinase Activity-IRAK is a downstream molecule of CD14/TLR-mediated NF-B signaling pathway. We investigated the involvement of IRAK in H. pylori-mediated NF-B activation. IRAK is a serine/threonine kinase that is activated by LPS. IRAK kinase activity was determined by in vitro kinase assay and was found to be up-regulated in H. pylori-and LPS-treated THP-1 cells but not in MKN45 cells (Fig. 4A). IRAK is known to undergo degradation after activation. We therefore examined the protein level of IRAK and found that it underwent degradation only in H. pylori-or LPStreated THP-1 cells but not in MKN45 cells (Fig. 4B). We performed an immunoblot analysis using the anti-IB␣ antibody and confirmed that H. pylori induced IB degradation in both cell types, whereas LPS induced IB degradation only in THP-1 cells (Fig. 4B). These observation suggested that NF-B activation by H. pylori is mediated by CD14 and IRAK in THP-1 cells but not in MKN45 cells.
Analysis of TLR4 Involvement-Recently, positional cloning analysis revealed that mice of the C3H/HeJ strains, which are profoundly unresponsive to LPS, have a TLR4 protein with a single point mutation (P712H). We investigated the role of TLR4 in H. pylori-mediated NF-B in macrophages from C3H/ HeN LPS-sensitive and C3H/HeJ LPS-insensitive mice. NF-B activation was observed in C3H/HeN mice treated with either H. pylori or LPS to a greater extent than in C3H/HeJ mice (Fig.   FIG. 5. Involvement of TLR4 in murine macrophages. A, peritoneal macrophages from C3H/HeN and C3H/HeJ mice were stimulated with H. pylori (TN2) or LPS (1g /ml) for 90 min. Nuclear extracts were prepared, and NF-B activation was analyzed by EMSA. B, peritoneal macrophages from HeN and HeJ mice were stimulated with H. pylori (TN2) or LPS (1g /ml) for 16 h. Supernatants were collected, and TNF␣ concentration was measured by ELISA. The values shown in this figure represent the mean Ϯ S.D. from four independent experiments. Significant differences to HeN are indicated as **, p Ͻ 0.05. 5A). Next, H. pylori were co-cultured with macrophage cells for 16 h, and TNF␣ concentration in the culture supernatant was examined. C3H/HeN macrophages treated with LPS or H. pylori produced a higher amount of TNF␣ than C3H/HeJ macrophages (Fig. 5B). These results indicated that H. pylori-mediated NF-B activation in macrophage cells involves TLR4 (see Fig. 6).

DISCUSSION
The present study has revealed that H. pylori activates the transcriptional factor NF-B in monocytic/macrophage cells even without direct interaction with the cell surface and does not require the presence of cag PAI. There is virtually no infiltration of macrophages in the mucosal layer in the absence of gastritis. However, once gastritis develops, numerous macrophages infiltrate into the mucosa. Our hypothesis is that in H. pylori infection, initially a small number of organisms comes in direct contact of epithelial cells, activates NF-B, and induces pro-inflammatory cytokines. This step requires cag PAI. However, once macrophages infiltrate into the mucosal layer, H. pylori is able to activate NF-B without direct contact, independently of cag PAI. This step amplifies the inflammatory changes in the mucosa. We previously found that the wild type-H. pylori induced severe gastritis, whereas ⌬cagE induced far milder changes in the gastric mucosa (38). These observations suggest that direct interaction between cag PAI-positive H. pylori and epithelial cells is essential for the initialization of severe gastritis.
The current study has shown that H. pylori-mediated NF-B activation in macrophages involves TLR4. In fact, TLR4 mRNA expression was also detected in MKN45 and other gastric cancer cell lines (AGS, TMK-1). 2 However, CD14, the co-receptor for TLR4, was detected in THP-1 but not in MKN45 cells. CD14 expression level is associated with the activity of NF-B (42)(43). These results indicate that defective CD14 expression may be responsible for low LPS sensitivity in the epithelial cells.
Reportedly, Gram-negative bacterial LPS stimulated Tolllike receptor 2 (TLR2) and initiated the NF-B signaling cascade (44). Because H. pylori is a Gram-negative bacterium, we investigated H. pylori-mediated NF-B activation using HEK293 cells, a cell line that does not express TLR2, and found that NF-B was also activated as much as TNF␣, indicating that TLR2 did not play a major role in H. pylori-mediated NF-B activation. 2 Recently, it has been reported that other TLRs are associated with bacteria-mediated NF-B activation. TLR5 and TLR9 recognize bacterial flagellin and CpG DNA, respectively. Therefore it is important to evaluate an association between these receptors and H. pylori-mediated NF-B activation (45,46).
We have considered that the H. pylori-mediated NF-B activation in gastric epithelial cells is a novel mechanism. In this study, we have revealed that CD14/TLR and IRAK are not involved in the activation. We also investigated the role of MyD88 (myeloid differentiation protein) by the expression of the dominant negative mutants and found no effect on NF-B activation by H. pylori. 2 We previously revealed that the NF-B activation by H. pylori is associated with TRAF6 which is a downstream molecule TLR receptor. We should find a molecule of H. pylori origin that directly or indirectly interacts with TRAFs.
The products of cag PAI genes are supposed to form a type IV secretion system (33)(34)47). In particular, CagE is a homologue of a transporter component in Agrobacterium tumefaciens and Bordetella pertussis, which are engaged in the transcellular transport of toxins or T-DNA (47)(48)(49). Therefore, it is possible that the putative transporter delivers active substances into human cells. Recently, it has been reported that CagA protein can be tyrosine-phosphorylated in the gastric epithelial cells (33,34). However, the expression of CagA protein in epithelial cell does not activate NF-B, and the cagA knockout mutants activate NF-B as well as the wild-type. 2 These observations suggest that certain effector proteins of H. pylori origin other than CagA that can be transported into host FIG. 6. Schematic presentation of the signaling leading to NF-B activation in response to H. pylori in gastric epithelial and macrophage cells.
In epithelial cells, only cag PAI-positive H. pylori on direct contact induces NF-B activation via TRAF2 and TRAF6, where CD14 is not involved. The upstream mechanism of TRAFs remains unknown. In macrophage cells, H. pylori activates NF-B through CD14/TLR4 and IRAK. Neither cag PAI nor direct contact is essential for activation. cells interact directly with intracellular molecules and activate NF-B.
We have previously used cDNA microarray analysis to characterize the mRNA responses of MKN45 to H. pylori. By analyzing the mRNA expression status of 2300 human genes, we showed that H. pylori induces significant changes in mRNA expression in a small portion of the tested genes (50). In contrast with the previous study, we used the array primarily designed for the detection of cytokine expression in this study and revealed that 18 of 268 genes in MKN45 and 25 of 268 in THP-1 were up-regulated by H. pylori. The pro-inflammatory cytokines such as TNF␣ and IL-1 are regulated mainly by NF-B (51). Inflammatory response to H. pylori infection is a complicated product of these cytokine networks. Among them, the up-regulation of chemokines such as MIP and IL-8 is especially important, because it may be associated with neutrophil infiltration, which is characteristic of H. pylori-infected mucosa (52).
In conclusion, we have shown that the mechanism of H. pylori-mediated NF-B activation is different in gastric epithelial cells and monocytic cells. NF-B activation is caused through a CD14/TLR4-dependent pathway in monocytic cells but through a CD14-independent mechanism in gastric epithelial cells.