ABSTRACT
Helicobacter pylori stimulates secretion of monocyte chemoattractant protein 1 (MCP-1) from gastric epithelial cells. Secretion of this chemokine may be instrumental in monocyte infiltration of the gastric epithelium that characterizes H. pylori gastritis. The aim of this study was to identify the mechanism by which H. pylori induces MCP-1 production. Induction of MCP-1 mRNA was assessed by reverse transcription-PCR. We used luciferase reporter assays to monitor activation of the MCP-1 gene promoter and electrophoretic mobility shift assays to explore binding of transcription factors to this promoter. H. pyloriinfection increased MCP-1 mRNA expression from gastric epithelial cells. Induction of MCP-1 mRNA relies on an intact cagpathogenicity island. We identified two closely spaced NF-κB-binding sites within the MCP-1 distal enhancer as required for H. pylori-induced MCP-1 gene transcription. H. pyloriinfection led to the specific activation of NF-κB complexes containing p50 and p65. Kinase-deficient mutants of NF-κB-inducing kinase (NIK) and IκB kinases (IKK) caused suppression of MCP-1 distal enhancer-dependent reporter activity following H. pyloriinfection. H. pylori infection induces the activation of NF-κB via the NIK-IKK signaling complex, leading to MCP-1 gene transcription in gastric epithelial cells.
Helicobacter pylori is the major cause of chronic active gastritis and is strongly associated with both duodenal and gastric ulceration (8, 10), as well as gastric cancer (19, 42) and mucosa-associated lymphoid tissue lymphoma (24). H. pylori-infected gastric mucosa is frequently infiltrated with inflammatory cells including neutrophils, monocytes, and lymphocytes (18). However, H. pylori is minimally invasive, which is why most investigators have focused on host or bacterial soluble factors as potential mediators of inflammatory cell recruitment.
Chemoattractant cytokines (chemokines) form a superfamily of closely related secreted proteins which specialize in mobilizing leukocytes to areas of immune challenge (4, 6, 43). These inducible proinflammatory peptides potently stimulate leukocyte migration along a chemotactic gradient. They also modulate leukocyte adhesion, activate signal transduction cascades leading to novel gene expression programs, and mediate other leukocyte functions necessary for leukocytes to leave the circulation and infiltrate tissues. Thus, increased chemokine production and release is an important mechanism for leukocyte recruitment in response to injury or infection. Chemokines are divided into groups or families that are defined by characteristic cysteine motifs. Four families of chemokines—CXC, CC, C, and CX3C (C is a conserved cysteine residue and X is any other amino acid)—have been described (4, 6, 43). Among these chemokines, interleukin-8 (IL-8), a prototype CXC chemokine, seems to play an important role in recruiting and activating neutrophils (2, 14, 15, 20, 31, 34) in the gastric mucosa. Several reports suggest that gastric epithelial cells represent an important source of IL-8 (14, 23, 39). Infection with H. pylori strains expressing cytotoxin-associated antigen (cagA) has been associated with increased gastric mucosal IL-8 production in vivo (35, 50) and with induction of IL-8 in gastric epithelial cell lines in vitro (13, 16, 39). The cagA gene is part of the cag pathogenicity island (PAI). ThecagA gene product itself is not the direct inducer of IL-8 (13, 39). Multiple genes in the cag PAI participate in the induction of epithelial IL-8 (11).
Since chronic H. pylori-associated gastritis is accompanied by monocyte and lymphocyte infiltration, in addition to neutrophil infiltration, we were interested in determining whether members of the CC subfamily of chemokines, which recruit monocytes and lymphocytes, participated in H. pylori-associated pathogenesis. Monocyte chemoattractant protein 1 (MCP-1) is a CC chemokine that stimulates mononuclear leukocytes. Like IL-8 induction, the in vivo expression of MCP-1 is elevated in the gastric mucosa following H. pyloriinfection (40, 48). H. pylori infection of gastric epithelial cell lines also stimulates MCP-1 expression (25, 48). The purpose of this study was to explore the molecular mechanism responsible for increased MCP-1 expression by gastric epithelial cells in response to H. pylori infection. We demonstrate that two NF-κB binding sites located approximately 2.6 kb from the transcription initiation site are essential for activation of MCP-1 gene expression by H. pylori infection. Following exposure to cag PAI+H. pyloristrains, gastric epithelial cells exhibit increased NF-κB activity, which subsequently elevates MCP-1 gene expression. Thus, the heightened inflammatory response observed in patients infected with pathogenic (PAI+) strains of H. pylori is due in part to the ability of these pathogenic bacteria to induce NF-κB, a key regulation of proinflammatory chemokines such as MCP-1.
MATERIALS AND METHODS
Bacterial strains. H. pylori (ATCC 49503; American Type Culture Collection, Manassas, Va.) was used in most of this study. Other clinical strains (OHPC0001, OHPC0002, and OHPC0003), isolated from chronic gastritis patients, were kind gifts from T. Kitahora (Ohkura Hospital, Tokyo, Japan). The presence of cag PAI andvacA in these strains was determined by PCR using specific sets of primers (1, 47). H. pylori strains were recovered from frozen stocks by seeding them on a blood agar plate (Mueller Hinton II agar with 7% horse blood) at 37°C for 3 days under microaerophilic conditions (10% O2 and 10% CO2) generated with Anaeropack Campylo (Mitsubishi Gas Chemicals Co., Tokyo, Japan). Bacteria harvested from the plates, using cotton swabs, were suspended in 200 ml of brain heart infusion broth containing 10% fetal calf serum and were then liquid cultured at 37°C for 2 days with vigorous shaking in a controlled microaerophilic atmosphere. Bacteria were harvested from a broth culture by centrifugation and were resuspended at the indicated concentrations in antibiotic-free medium. At this time, bacteria reached a concentration of 4 × 108 CFU/ml. All procedures were performed with the approval of the appropriate institutional biosafety review committees and in compliance with their guidelines for biohazards.
Cell culture.A panel of epithelial cell lines was cultured in RPMI 1640 or Ham F-12 medium supplemented with 10% fetal bovine serum (GIBCO Laboratories, Grand Island, N.Y.) and antibiotics (penicillin [50 U/ml] and streptomycin [50 μg/ml]). In these experiments, the gastric epithelial cell lines MKN45, MKN28, MKN74, AGS, and Kato III were used (32). MKN45 and AGS cells are derived from human gastric adenocarcinomas. MKN28 and MKN74 cells are derived from human gastric tubular adenocarcinomas. Kato III cells are derived from a human gastric signet ring cell carcinoma.
RT-PCR for MCP-1 mRNA.Gastric epithelial cells were cocultured with H. pylori for the indicated time intervals. Total cellular RNA was extracted from the cells with Trizol (GIBCO-BRL, Gaithersburg, Md.) according to the protocol provided by the manufacturer, and the amount of total RNA was determined by measuring the absorbance at 260 nm. First-strand cDNA was synthesized in a 20-μl reaction volume using an RNA PCR kit (Takara Shuzo, Kyoto, Japan) with random primers. Thereafter, cDNA was amplified for 35 and 28 cycles for MCP-1 and β-actin, respectively. The sense and antisense oligonucleotide primers used were 5′-TCGCTCAGCCAGATGCAATCAATGC-3′ and 5′-CCCAGGGGTAGAACTGTGGTTCAA-3′ (26), respectively, for MCP-1 and 5′-GTGGGGCGCCCCAGGCACCA-3′, and 5′-CTCCTTAATGTCACGCACGATTTC-3′, respectively, for β-actin. Product sizes were 479 bp for MCP-1 and 548 bp for β-actin. Cycling conditions were as follows: denaturing at 94°C (30 s), annealing at 60°C (30 s), and extension at 72°C (60 s for MCP-1 and 90 s for β-actin). The PCR products were fractionated on 2% agarose gels and visualized by ethidium bromide staining. This reverse transcription-PCR (RT-PCR) is not semiquantitative.
Determination of MCP-1 secretion in gastric epithelial cells.MKN45 cells were cultured in complete medium in 24-well plates. After the cells reached subconfluency, H. pylori was added to a final concentration of 108 CFU/ml. After 24 h, the cell supernatants were collected after centrifugation to remove bacteria and stored at −80°C until measurement. MCP-1 concentrations were measured in the culture supernatants using a commercially available enzyme-linked immunosorbent assay (ELISA) kit (Biosource International, Inc., Camarillo, Calif.). ELISA sensitivity was 20 pg/ml.
Luciferase assay.The luciferase reporter constructs contain the proximal promoter region and distal enhancer region of the human MCP-1 gene and have recently been described in detail (45). IκB-αΔN (9) and IκB-βΔN (kindly provided by D. W. Ballard, Vanderbilt University School of Medicine, Nashville, Tenn.) (28) are deletion mutants of IκB-α and IκB-β lacking the NH2-terminal 36 amino acids and 23 amino acids, respectively. The IκB kinase alpha (IKK-α) (pEV-IKK-α-T7), IKK-β (pcDNA-IKK-β-FLAG), and NF-κB-inducing kinase (NIK) (pRK-Myc-NIK) expression constructs and the kinase-deficient versions of these kinases, K44M IKK-α, K44A IKK-β, and KK429/430AA NIK, have been described previously (21). For the luciferase assay, 7 × 105MKN45 cells were transfected with 1 μg of each luciferase reporter construct along with 2 μg of pRL-TK, an internal controlRenilla luciferase expression vector (Toyo Ink Co., Tokyo, Japan), using a calcium phosphate mammalian cell transfection kit (5 Prime→3 Prime, Inc., Boulder, Colo.) according to the manufacturer's recommendations. After 24 h, the transfected cells were divided into two parts, and H. pylori (final concentrations, 108 CFU/ml) was added separately to each part. After an additional 6 h of incubation, cell lysates were prepared using PicaGene Dual (Toyo Ink Co.). Twenty microliters of the cell lysates was assayed for both firefly and Renilla luciferase activities using the dual-reporter assay system in a Lumat model LB9505C luminometer (Berthold, Bad Wildbad, Germany). The firefly luciferase activity was normalized to the Renilla luciferase activity, and fold stimulation was calculated as the ratio of the value for H. pylori-infected cells to that for uninfected cells.
Preparation of nuclear extracts and electrophoretic mobility shift assay (EMSA).Nuclear proteins were extracted from epithelial cells incubated in the presence or absence of H. pylori (108 CFU/ml) as described by Antalis et al. (3), with modifications. Five-microgram aliquots of nuclear proteins were incubated for 15 min at room temperature with 1 ng of [α-32P]dCTP- and [α-32P]dATP-labeled oligonucleotide probe and 1 μg of poly(dI-dC) (Pharmacia, Piscataway, N.J.) in 20 μl of a solution containing 10 mM Tris-HCl (pH 7.5), 50 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, and 5% glycerol. In competition studies, a 100-fold excess of the unlabeled oligonucleotides was included in the reaction mixture along with the radiolabeled probe. The probes or competitors used were prepared by annealing the sense and antisense synthetic oligonucleotides as follows: NF-κB element A1 in the MCP-1 gene, 5′-gatcGATCTGGGAACTTCCAAAGC-3′; A1 mutant (MA1), 5′-gatcGATCTaGaAACTTCCAAAGC-3′; NF-κB element A2 in the MCP-1 gene, 5′-gatcAGAGTGGGAATTTCCACTCA-3′; A2 mutant (MA2), 5′-gatcAGAGTGGGAATTcggACTCA-3′; and a typical NF-κB element from the IL-2 receptor alpha (IL-2Rα) gene, 5′-gatcCGGCAGGGGAATCTCCCTCTC-3′. Underlined sequences represent the NF-κB binding site, and mutations are indicated in lowercase. In some experiments, nuclear proteins were preincubated with 2 μg of antibodies to p65, p50, p52, or c-Rel (Santa Cruz Biotechnology Inc., Santa Cruz, Calif.) for 45 min at room temperature before addition of the labeled probe. The mixtures were loaded onto a 4% polyacrylamide gel with 0.25× Tris-borate buffer. After electrophoresis, gels were dried and analyzed by autoradiography.
RESULTS
H. pylori infection increases steady-state MCP-1 mRNA levels and secretion of MCP-1 in gastric epithelial cells.Using RT-PCR, we first examined whether coculture of several gastric epithelial cell lines with H. pylori led to the induction of MCP-1 mRNA. Coculture with H. pylori significantly enhanced steady-state levels of MCP-1 mRNA in all gastric epithelial cell lines examined (Fig. 1A). MCP-1 mRNA levels clearly increased 2 h after addition of H. pylori to the cells and remained elevated at least 24 h following cocultivation (Fig. 1B). Supernatants derived from H. pyloricultures failed to induce MCP-1 mRNA expression in MKN45 cells (Fig.2A). Moreover, neither heat-killed bacteria nor live bacteria separated by a permeable membrane induced MCP-1 mRNA expression in MKN45 cells (data not shown). These results suggest that the interaction with live H. pylori itself, rather than products secreted by these bacteria, upregulates the steady-state levels of MCP-1 mRNA. We examined the capacity of MKN45 cells to secrete MCP-1 upon coculture with H. pylori. Although MCP-1 secretion by MKN45 cells was below the sensitivity of ELISA in the control condition, MCP-1 secretion was found to increase after coculture with H. pylori (40 pg/ml/24 h).
H. pylori-induced MCP-1 mRNA expression in gastric epithelial cell lines. (A) MCP-1 mRNA profiles in various gastric epithelial cell lines (MKN45, Kato III, AGS, MKN28, MKN74) infected with H. pylori. Total RNA was isolated from uninfected (−) gastric epithelial cells or from H. pylori(ATCC 49503)-infected (+) cells and subjected to RT-PCR to evaluate steady-state MCP-1 and β-actin mRNA levels. (B) Time course ofH. pylori-induced MCP-1 mRNA expression. Total RNA was extracted from MKN45 cells infected with H. pylori for the indicated time intervals and used for RT-PCR.
The cag PAI of H. pylori is required for induction of MCP-1 expression in MKN45 cells. (A)cag PAI+H. pylori strains (ATCC 49503, OHPC0001, and OHPC0003) are capable of inducing MCP-1 mRNA expression in MKN45 cells compared with a cagPAI−H. pylori strain (OHPC0002). Total RNA was extracted from the cells infected with H. pylori for 6 h and used for RT-PCR. (B) Role of cag PAI in H. pylori-induced MCP-1 promoter and enhancer activity. MKN45 cells were transfected with 1 μg of either pGLM-PRM or pGLM-ENH firefly luciferase reporter and 2 μg of the pRL-TK Renillaluciferase reporter as an internal control. Twenty-four hours later, cells were treated with medium alone (control), with medium containingH. pylori, or with medium containing H. pyloriculture supernatants. MKN45 cell lysates were prepared 6 h after stimulation and analyzed for luciferase activity. Data represent means ± standard deviations from three independent experiments and are expressed as fold induction relative to the basal level measured in cells treated with medium alone (control).
H. pylori strains differ in the ability to induce MCP-1 mRNA expression.Since recent studies indicated that expression of multiple genes in the cag PAI is necessary for cytokine production by gastric epithelial cells in vitro (11), we examined the ability of several H. pylori strains, possessing or lacking the cag PAI, to induce MCP-1 mRNA expression. Infection with H. pylori strains ATCC 49503, OHPC0001, and OHPC0003, which contain the entire cag PAI (47), led to increased MCP-1 mRNA levels in MKN45 cells (Fig. 2A). Strain OHPC0002, which lacked the cag PAI (47), failed to induce MCP-1 mRNA levels (Fig. 2A). The ability of H. pylori to induce MCP-1 mRNA expression was independent of the vacA locus, as all bacterial isolates possessed this gene (1) (Fig. 2A). These results suggest that the H. pylori cag PAI may play an important role in the induction of MCP-1 mRNA expression.
H. pylori regulates MCP-1 gene transcription.To examine whether H. pylori-mediated upregulation of MCP-1 mRNA levels was due to the enhanced activity of the MCP-1 promoter and enhancer, we transfected luciferase reporters under the control of the MCP-1 5′ regulatory sequences. Structural features of the human MCP-1 promoter have recently been described (36, 44, 45). While basal promoter activity is dependent on a proximal promoter region containing an SP-1 site, cytokine-inducible promoter activity is mainly mediated by a distal enhancer region (Fig.3). Two NF-κB binding sites (A1 and A2) which are crucial for enhancer activity have been identified in the distal enhancer region (36, 44, 45). Therefore, we used an MCP-1 promoter/enhancer driven luciferase construct, pGLM-ENH, containing the enhancer (situated between nucleotides −2742 and −2513 in the original MCP-1 gene) and promoter (between nucleotides −107 and +60) regions of the human MCP-1 promoter to test for H. pylori responsiveness (Fig. 2B and Fig. 3). As shown in Fig. 2B,H. pylori strain ATCC 49503 caused a sixfold stimulation of the pGLM-ENH reporter construct in MKN45 cells. Two other H. pylori strains, OHPC0001 and OHPC0003, also induced luciferase activity from this reporter to similar levels. However, strain OHPC0002, devoid of the cag PAI, and H. pylori (ATCC 49503) culture supernatants were incapable of increasing luciferase activity from the pGLM-ENH reporter. These findings parallel the MCP-1 mRNA responses shown in Fig. 2A. Because the pGLM-PRM reporter, containing only the proximal MCP-1 promoter region, was not responsive to any of the H. pylori strains tested, we conclude that the H. pylori-responsive sites within the 5′ regulatory sequences of the MCP-1 gene are localized in the distal enhancer.
Cooperation between the A1 and A2 sites of the MCP-1 distal enhancer in H. pylori-induced MCP-1 gene transcription. (A) Schematic representation of the MCP-1 reporter constructs. The proximal promoter region and distal enhancer region of the MCP-1 gene are indicated by closed and open boxes, respectively. The A1 and A2 sites of the MCP-1 distal enhancer region were mutated (indicated by ×) in some of the constructs. These constructs were transfected into MKN45 cells with pRL-TK, and the cells were subsequently infected with H. pylori for 6 h. (B) Normalized relative luciferase activities corresponding to each construct expressed in untreated MKN45 cells or in cells induced withH. pylori. Data represent means ± standard deviations from three independent experiments.
H. pylori-induced activation of the MCP-1 distal enhancer can be blocked by dominant interfering signaling components of the NF-κB pathway.NF-κB activity is normally regulated through its cytoplasmic retention by specific inhibitors, including IκB-α and related proteins (5). Activation of NF-κB by cytokines is mediated by signal transduction cascades, leading to activation of IKK-α and IKK-β (17, 29, 37, 49, 51). These kinases phosphorylate the IκBs, a modification that triggers their ubiquitination and proteolysis, allowing the released NF-κB dimers to enter the nucleus and activate target genes (41). Recent studies also suggested that a kinase of the mitogen-activated protein (MAP) kinase kinase kinase family termed NIK participates in NF-κB activation induced by cytokines (27). NIK was suggested to physically interact with and activate the IKKs (37, 49). Through the use of dominant interfering mutants of IκB-α and IκB-β and kinase-deficient mutants of IKK-α, IKK-β, and NIK, we examined whether these signaling intermediates participated in H. pylori-induced activation of NF-κB and the MCP-1 distal enhancer. The constitutive repressor mutants of IκB-α and IκB-β (IκB-αΔN and IκB-βΔN) effectively blocked H. pylori induction of MCP-1 distal enhancer-driven luciferase activity (Fig. 4B). These findings suggest that the activation of the MCP-1 distal enhancer byH. pylori involves NF-κB. Since the wild-type IKKs and NIK were included as controls for their kinase-deficient mutant counterparts, the functional effects of wild-type IKK-α, IKK-β, and NIK on H. pylori-induced MCP-1 enhancer activation were first evaluated (Fig. 4A). IKK-α and IKK-β produced only a 1.7-fold increase in pGLM-ENH luciferase activity in MKN45 gastric epithelial cells, while NIK mediated a 4.8-fold increase in luciferase activity. The addition of the wild-type IKK-α, IKK-β, and NIK expression vectors produced an amplification of the H. pylori response (Fig. 4A). Finally, the dominant interfering mutants were cotransfected along with the pGLM-ENH prior to addition of H. pylori. As shown in Fig. 4B, the elevated MCP-1 enhancer-dependent luciferase activity in response to H. pylori was markedly suppressed by cotransfection with these mutants. These results confirmed that the NF-κB pathway via the NIK-IKK signaling components was involved in H. pylori-induced activation of the MCP-1 distal enhancer.
Functional analysis of the roles of IKK-α, IKK-β, and NIK in H. pylori-induced activation of the MCP-1 distal enhancer. (A) Functional effects of wild-type IKK-α, IKK-β, and NIK on H. pylori-induced activation of the MCP-1 distal enhancer. MKN45 cells were transfected with 1 μg of pGLM-ENH and 5 μg of the wild-type IKK-α, IKK-β, or NIK expression vector. Each transfection also contained 2 μg of pRL-TK and was supplemented to 5 μg with pCMV4 vector. Twenty-four hours after transfection, half of the transfectants were cultured with H. pylori for 6 h, while the other half were left untreated. The cells were lysed and assayed for luciferase activity. Luciferase activity is presented as fold induction relative to the basal level measured in cells transfected with pCMV4 without further treatment. (B) Functional effects of IκB-α and IκB-β dominant interfering mutants and kinase-deficient IKK-α, IKK-β, and NIK mutants on H. pylori-induced activation of the MCP-1 distal enhancer. MKN45 cells were transfected with 1 μg of pGLM-ENH and 5 μg of the IκB-α mutant IκB-αΔN, IκB-β mutant IκB-βΔN, the K44M mutant of IKK-α, the K44A mutant of IKK-β, or the kinase-deficient KK429/430AA mutant of NIK and then infected withH. pylori for 6 h. All values were first calculated as fold induction relative to cells transfected with pCMV4 without further treatment. These values were then expressed as percentages of the response obtained with H. pylori infection. Data represent means ± standard deviations from three independent experiments.
Cooperation between the A1 and A2 sites of the MCP-1 distal enhancer in H. pylori-induced activation.To investigate the role of the A1 NF-κB site and a possible cooperation with the A2 site in H. pylori-induced activation, the sequences of the A1 and A2 sites were mutated (Fig. 3). The A1 and A2 mutant reporter constructs were transfected into MKN45 cells prior to addition of H. pylori. As shown in Fig. 3, H. pylori-induced luciferase activity was significantly reduced by mutation in either the A1 or A2 sequence, indicating that H. pylori-induced activation of the MCP-1 distal enhancer involves both A1 and A2 sites. These results suggest that both the A1 and A2 sites of the distal enhancer are important for H. pylori-induced MCP-1 gene transcription.
H. pylori infection of gastric epithelial cells induces binding of NF-κB/Rel family proteins to the A1 and A2 κB elements of the MCP-1 distal enhancer.We tested by EMSA whether H. pylori infection induced binding of NF-κB/Rel family members to the putative A1 and A2 κB elements of the MCP-1 distal enhancer. Synthetic oligonucleotides containing the A1 site (nucleotide positions −2640 to −2632 relative to the MCP-1 transcription start site) and the A2 site (nucleotide positions −2612 to −2603) of the MCP-1 distal enhancer were used as probes. MKN45 cells were infected with H. pylori, and at different time points postchallenge, nuclear protein extracts were prepared and analyzed for NF-κB DNA binding activity. As shown in Fig. 5B, when the A1 site was used as a probe, a complex was induced in MKN45 cells within 60 min after infection with H. pylori (lanes 1 to 5). Infection of other gastric epithelial cell lines with H. pylori resulted in the induction of a DNA binding activity similar to that observed in MKN45 cells at 60 min after H. pyloriinfection (Fig. 5A, lanes 1 to 10). This complex was specific to the A1 fragment because it was competed by addition of excess cold probe. This complex most likely contains members of the NF-κB/Rel family, as it was also competed by the typical NF-κB sequence of the IL-2Rα enhancer. The A2 oligonucleotide was also an effective competitor of the A1 bound complex, but a mutant sequence of the A1 site (MA1) failed to compete the complex (Fig. 6A, lanes 2 to 6). The A2 probe essentially showed the same pattern of complex formation as probe A1 (Fig. 5A, lanes 11 to 20; Fig. 5B, lanes 6 to 10; Fig. 6A, lanes 8 to 12). To identify which NF-κB/Rel family members were binding to the A1 and A2 elements of the MCP-1 distal enhancer, we performed EMSA using antibodies specific for members of the NF-κB/Rel family. Supershifts were seen with anti-p65 and anti-p50 antibodies in complexes formed with both probes A1 and A2, illustrating that these complexes contain the p50 and p65 subunits of NF-κB (Fig. 6B).
Binding of nuclear proteins from H. pylori-infected gastric epithelial cell lines to the A1 and A2 probes derived from the MCP-1 distal enhancer. (A) NF-κB activation in several gastric epithelial cell lines treated for 1 h withH. pylori (ATCC 49503). Nuclear extracts were prepared from the indicated cell lines and incubated with radiolabeled A1 (lanes 1 to 10) and A2 (lanes 11 to 20) probes. (B) Time course of NF-κB activation in MKN45 cells infected with H. pylori, evaluated by EMSA. (C) cag PAI products of H. pylori are required for induction of NF-κB binding activity in MKN45 cells. Nuclear extracts from unstimulated MKN45 cells (lanes 1 and 6) and cells cocultured with the indicated H. pylori strains (lanes 2 to 5 and 7 to 10) were analyzed for NF-κB.
Induction of specific NF-κB complex formation by coculture of gastric epithelial cells with H. pylori. (A) Sequence specificity of NF-κB binding activity in MKN45 cells. The radiolabeled A1 (lanes 1 to 6) and A2 (lanes 7 to 12) probes derived from the MCP-1 distal enhancer were incubated with nuclear extracts from unstimulated (lanes 1 and 7) or H. pylori-stimulated (lanes 2 to 6 and 8 to 12) MKN45 cells without (lanes 2 and 8) or with (lanes 3 to 6 and 9 to 12) a 100-fold excess amount of each specific competitor oligonucleotide. (B) Characterization of NF-κB/Rel proteins that bound to the A1 and A2 κB sites of the MCP-1 gene. Nuclear extracts from MKN45 cells that were unstimulated (lanes 1 and 7) or cocultured with H. pylori (lanes 2 to 6 and 8 to 12) were preincubated with antibodies against the NF-κB/Rel family member p50, p65, c-Rel, or p52 before addition of radiolabeled A1 (lanes 1 to 6) and A2 (lanes 7 to 12) probes.
Since we have shown that cag PAI+ strains induce significantly more MCP-1 mRNA than do cag PAI−H. pylori strains, we sought to determine whethercag PAI+H. pylori strains better induced NF-κB (Fig. 5C). Markedly increased NF-κB DNA binding activity was induced by cag PAI+H. pylori strains (lanes 2, 3, 5, 7, 8, and 10) compared with thecag PAI− strain (lanes 4 and 9). These results indicate that the increased ability of cag PAI+strains to activate NF-κB binding may play an essential role in the observed activation of the MCP-1 distal enhancer by these bacterial strains. Thus, H. pylori infection induces MCP-1 gene expression, at least in part, through the induced binding of the p50 and p65 NF-κB/Rel family members to the A1 and A2 elements of the MCP-1 distal enhancer, and this effect is dependent on thecag PAI products.
DISCUSSION
There has been recent interest in the role of chemokines inH. pylori-associated pathogenesis. Although early studies showed increased IL-8 production in gastric mucosa infected withH. pylori (14, 15), the involvement of other chemokines in gastritis has not been fully investigated. Locally produced CC chemokines that target monocytes and lymphocytes may play important roles in chronic inflammation associated with H. pylori infection. Consistent with this notion, the in vivo expression of CC chemokines, such as MCP-1 or RANTES (regulated on activation normal T-cell expressed and secreted), is increased inH. pylori positive mucosa compared with uninfected mucosa (40). A significant correlation is seen only between MCP-1 mRNA expression and mononuclear cell infiltration (40). The present results showed that H. pylori can induce MCP-1 mRNA expression and MCP-1 secretion in gastric epithelial cells. Although MCP-1 is known to be produced by many cell types, such as monocytes/macrophages or fibroblasts (6, 38), we have shown that gastric epithelial cells may be an important source of MCP-1 production in the stomach.
MCP-1 gene transcription requires the activation of NF-κB, AP-1, SP-1, or STAT, depending on the types of cells and stimuli (36, 44-46). The H. pylori-responsive elements within the 5′ regulatory sequences of the MCP-1 gene are localized in the MCP-1 distal enhancer. Mutation of the A1 or A2 NF-κB site in this region resulted in loss of H. pylori responsiveness, indicating that MCP-1 regulation by H. pylori is mediated by NF-κB sites. Because neither supernatants of H. pylori cultures nor H. pylori separated by a permeable membrane induces MCP-1 expression, components of the H. pylori bacterium most likely act to trigger the induction of MCP-1 in gastric cells. Since acag PAI− strain of H. pylori is incapable of inducing MCP-1 expression, it is likely that the gene products of the cag PAI are involved in the induction of MCP-1 gene expression. The cag proteins form a multimeric structure on the H. pylori surface, and this structure seems to be capable of eliciting intracellular signaling in target cells (12). In this study we analyzed the capacities of different H. pylori strains to induce the chemokine MCP-1 and identified the signaling components NIK and IKKs as likely participants in H. pylori-mediated NF-κB activation. Compared with the cag PAI−H. pyloristrain, the more virulent cag PAI+ strains showed an enhanced ability to induce MCP-1 promoter activity and NF-κB binding activity. This was consistent with our observations that increased MCP-1 induction in MKN45 cells was associated withcag PAI+ strains. These results demonstrate that the ability of cag PAI+H. pyloristrains to activate MCP-1 expression is dependent on prior activation of the NF-κB p50 and p65 subunits.
The mammalian signaling pathways triggered by H. pyloriremain largely unknown. In this study, we have identified the cellular kinases NIK and IKKs as participants in NF-κB-dependent MCP-1 induction by H. pylori in gastric epithelial cells. In addition to NIK and the IKKs, the MAP kinase p38 has been reported to be involved in NF-κB-dependent gene expression (7). Goebeler et al. (22) report a crucial role for p38 in MCP-1 gene expression following tumor necrosis factor alpha stimulation of endothelial cells. However, Naumann et al. (33) report that p38 was not activated by H. pylori in gastric epithelial cells. Consistent with their observations, blocking the p38 pathway by the p38 inhibitor SB203580 or by dominant negative mutants of p38 did not inhibit H. pylori-induced expression of MCP-1 (data not shown). Thus, the p38 MAP kinase does not seem to contribute to H. pylori-induced expression of the MCP-1 gene in our experimental system.
H. pylori-mediated NF-κB activation via the NIK-IKK signaling complex is likely to induce other genes involved in the inflammatory responses occurring during H. pylori-associated gastritis. For instance, in addition to activating MCP-1 as reported here and other proinflammatory cytokines following infection of gastric epithelial cells with H. pylori, NIK/IKK-mediated activation of NF-κB may also induce the expression of cell adhesion molecules (30). The identification of NF-κB-mediated induction of MCP-1 following H. pylori infection of gastric epithelial cells provides a better understanding of the pathogenic mechanism leading to the recruitment of leukocytes in H. pylori-associated gastritis.
ACKNOWLEDGMENTS
We thank Mika Yamamoto and Masako Sasaki for skilled technical assistance, T. Kitahora for gifts of clinical isolates of H. pylori, K. Oishi for gifts of the oligonucleotide primers for MCP-1, and D. W. Ballard for gifts of plasmids.
This work was supported in part by a Grant-in-Aid for Encouragement of Young Scientists from the Ministry of Education, Science, Sports and Culture of Japan.
Notes
Editor: R. N. Moore
FOOTNOTES
- Received 26 June 2000.
- Returned for modification 14 September 2000.
- Accepted 8 December 2000.
- Copyright © 2001 American Society for Microbiology
