Helicobacter pylori Induces IκB Kinase α Nuclear Translocation and Chemokine Production in Gastric Epithelial Cells

ABSTRACT NF-κB is an important transcriptional factor that is involved in multiple cellular responses, such as inflammation and antiapoptosis. IκB kinase α (IKKα) and IKKβ, which are critical regulators of NF-κB activity, possess various mechanisms for NF-κB activation. This variability in NF-κB signaling may be associated with distinct inflammatory responses in specific cell types. The gastric pathogen Helicobacter pylori is known to activate NF-κB. However, the role of IKK in H. pylori infection remains unclear. In this report, we show that H. pylori activates both IKKα and IKKβ in gastric cancer cells and enhances NF-κB signaling in distinct manners. We found that IKKβ acted as an IκBα kinase during H. pylori infection, whereas IKKα did not. H. pylori induced IKKα nuclear translocation in time-, multiplicity of infection-, and cag pathogenicity island-dependent manners. In contrast, p100 processing, which is a known IKKα activity induced by several cytokines, was not induced by H. pylori. Both IKKs were responsible for chemokine secretion by infected cells. However, the antiapoptotic effect of H. pylori was merely transduced by IKKβ. Microarray analysis and real-time PCR indicated that both IKKs were involved in the transcriptional activation of genes associated with inflammation, antiapoptosis, and signal transduction. Our results indicate that H. pylori activates NF-κB via both IKKα and IKKβ using distinct mechanisms. IKKα nuclear translocation induced by H. pylori is indispensable for appropriate inflammatory responses but not for antiapoptosis, which suggests a critical role for IKKα in gastritis development.

Helicobacter pylori is a pathogen that causes human gastric disease. About half of the world population is infected with this bacterium, although only a relatively small proportion of infected patients develop symptomatic disease, such as gastroduodenal ulcer, gastric cancer, and mucosa-associated lymphoid tissue lymphoma. Bacterial, environmental, and host genetic factors may affect the progress and outcome of gastric disease. One such factor that is responsible for severe disease is the bacterial virulence factor cag pathogenicity island (PAI) (reviewed in references 6, 32, and 36). H. pylori strains that carry cag PAI genes, called type I strains, are highly prevalent in patients with gastroduodenal ulcer and gastric cancer (2,4,8). Previous studies have revealed that type I H. pylori strains are capable of activating multiple intracellular signaling pathways in infected epithelial cells (22,26). The inflammatory, proliferative, and antiapoptotic responses observed in H. pylori-infected cells in culture and in gastric tissues are possibly mediated by the activation of intracellular signaling pathways, such as those for NF-B and mitogen-activated protein kinase.
NF-B is an important transcriptional factor that controls various biological processes, such as inflammation, cell survival or death, and cell cycle (reviewed in references 5, 14, 18, 20, and 23). The mechanism of NF-B activation by a variety of extracellular stimuli is unique in that it is induced rapidly and does not require de novo protein synthesis, thereby allowing the cells to respond quickly to emergent situations, such as bacterial infection (49). Most forms of NF-B, especially the most common form of the p50-RelA dimer, are rendered inactive through binding of the inhibitory protein IBs. Phosphorylation-induced ubiquitination of IBs promotes its degradation, which in turn liberates NF-B dimers as their active forms (reviewed in references 5, 14, 18, and 23). The IB kinase (IKK) complex is a protein complex that phosphorylates IBs in response to upstream stimuli, and it is considered to be a critical regulator of NF-B activity (14,20,49).
The IKK complex contains three subunits: the catalytic subunits IKK␣ and IKK␤ and the regulatory subunit IKK␥ (10,30,37,38,45,50). When overexpressed exogenously or synthesized in an in vitro system, both IKK␣ and IKK␤ phosphorylate IB proteins and activate NF-B (10,30,37,45,50). Earlier studies on IKK␤ knockout cells have indicated that IKK␤ is indispensable for IB phosphorylation, NF-B activation, and subsequent gene expression in response to proinflammatory stimuli (24,25). In contrast, IKK␣ knockout cells show normal IB␣ phosphorylation and RelA nuclear translocation in response to lipopolysaccharide or cytokines (16,43). These results have raised the question of whether IKK␣ is involved physiologically in NF-B activation. Interestingly, IKK␣-deficient mice show morphological abnormalities, indicating the specific role of IKK␣ that cannot be compensated for by IKK␤ (16,43). One of the specific activities of IKK␣ is the induction of p100 processing to p52. The phosphorylation of p100 by IKK␣ results in p100 degradation and the generation of p52, which in turn dimerizes with RelB to form the NF-B subunit (9,39). This pathway, which is called the alternative pathway, is considered to play an essential role in secondary lymphoid organ development and adaptive immunity (9,39). Another specific function of IKK␣ is to control gene expression by direct translocation into the nucleus, which appears to be important in epidermal differentiation and craniofacial morphogenesis (1,41,46). These functional diversities of IKK␣ and IKK␤, as well as the variations in extra-and intracellular signaling that lead to the activation of each IKK, may provide information on the numerous biological roles of these molecules with respect to NF-B.
As NF-B is especially important for the immune system, the constitutive activation of NF-B is associated with inflammatory diseases (20,23). Furthermore, aberrant NF-B activation leads to tumorigenesis via antiapoptotic gene expression (19,20). Thus, the detailed analysis of this signaling in disease states will be useful for therapy development. Although H. pylori persistently infects the human stomach and activates NF-B in gastric tissues, the way in which it activates NF-B is not well understood. Furthermore, the subunit of IKK complex that is involved in H. pylori-induced NF-B activation and the effects of these molecules on gastric disease remain to be resolved. To achieve a better understanding of NF-B signaling in H. pylori-related gastric disease, we examined the role of IKK in H. pylori-infected gastric cancer cells.
dominant-negative IKK vector for 24 h. Where indicated, siRNA oligonucleotides were transfected as described above 24 h before transfection of the reporter plasmids. The cells were supplemented with fresh culture medium and infected with H. pylori for 8 h. Luciferase activity was measured and calculated from cell lysates as described previously, and the results are represented as fold induction compared to the control in three independent experiments. Immunoblot analysis. For the preparation of total cell lysates, AGS cells that were treated with the indicated siRNAs and infected with H. pylori for different time periods were washed once with cold PBS and lysed in ice-cold Triton X-100 buffer (50 mM Tris/HCl [pH 7.6], 1% Triton X-100, 5 mM EDTA, 1 mM Na 3 VO 4 , 10 mM NaF, 10 g/ml leupeptin, 10 g/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride). The cell lysates were centrifuged at 10,000 ϫ g for 10 min at 4°C, and the supernatants were stored as total cell lysates. For the preparation of nuclear and cytosolic extracts, AGS cells were seeded in a 6-cm dish, transfected with the indicated siRNAs, and infected with H. pylori. The cells were washed with Tris-buffered saline (TBS), suspended in 200 l of Buffer A (10 mM HEPES [pH 7.9], 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 0.75% Nonidet P-40, 1 mM dithiothreitol, protease inhibitor cocktail [Roche Molecular Biochemicals]), incubated on ice for 3 min, and centrifuged at 1,500 ϫ g for 4 min at 4°C. The supernatant was removed and used as the cytosolic extract. The pellet was washed once with Buffer A without Nonidet P-40 and centrifuged as described above, followed by resuspension in 50 l Buffer C (20 mM HEPES [pH 7.9], 400 mM NaCl, 1 mM EDTA, 1 mM EGTA, 10 mM dithiothreitol, protease inhibitor cocktail) and incubation on ice for 10 min with frequent mixing. Finally, the suspension was centrifuged at 14,000 ϫ g for 10 min at 4°C and the supernatant was used as the nuclear extract. Equal amounts of cell lysates and extracts were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and transferred to a polyvinylidene difluoride membrane. The membrane was probed with the indicated primary antibody followed by the horseradish peroxidase-conjugated secondary antibody and developed using the ECL plus kit (Amersham, Buckinghamshire, United Kingdom). The protein levels of phospho-IB-␣, phospho-JNK, IKK␣, and p100/52 were determined by densitometry using KODAK 1D Image Analyzer software and normalized with the level of actin, TF-IID, or TRAF2.
Immunofluorescence. AGS cells were seeded in Lab-Tek chamber slides (Nalge Nunc International, Naperville, IL) and infected with H. pylori for the indicated time. The cells were washed twice with PBS, fixed in 2% paraformaldehyde for 30 min, washed with PBS, and permeabilized with 0.2% Triton X-100 for 1 h. After blocking with 10% normal goat serum, the cells were incubated overnight at 4°C with the polyclonal anti-IKK␣ antibody diluted in PBS. The cells were then washed three times with PBS and incubated with Alexa Fluor 488 (Molecular Probes, Eugene, OR) for 1 h. The nuclei were visualized by staining with propidium iodide. Images were obtained using the LSM510 confocal laser scanning microscope (Carl Zeiss, Oberkohen, Germany).
TUNEL assay. To investigate the effect of H. pylori infection on cell apoptosis, we used the TdT-mediated dUTP-biotin nick end-labeling (TUNEL) assay. AGS cells, which were treated with control or IKK siRNAs in Lab-Tek chamber slides, were maintained in serum-free medium for 24 h and infected with H. pylori at an MOI of 100 for a further 8 h. The cells were washed three times with PBS, and apoptotic cells were stained with Apoptag (Serologicals Inc., Norcross, GA) in accordance with the manufacturer's instructions. Apoptotic cells were visualized by fluorescein isothiocyanate, and the nuclei were stained with propidium iodide, followed by microscopic examination with the LSM510 confocal laser scanning microscope. The number of apoptotic cells, in a total of 1,500 to 2,000 cells in each well, was counted in three independent experiments, and the percentage of apoptotic cells was calculated.
Quantification of chemokines by ELISA. The interleukin-8 (IL-8) and GRO␣ concentrations in the culture supernatants were measured by enzyme-linked immunosorbent assay (ELISA) as specified by the manufacturer (Techne, Minneapolis, MN). AGS cells were plated in 24-well plates, transfected with siRNAs for 48 h, and infected with H. pylori for a further 8 h. The culture supernatants were then aspirated and stored at Ϫ70°C until they were subjected to the ELISA. The concentrations of IL-8 and GRO␣ were determined using standard curves obtained with the respective recombinant proteins. The values are represented as the averages Ϯ standard deviations (SD) of three independent experiments.
Microarray procedures. For RNA preparation, AGS cells were transfected with control or IKK-specific siRNAs for 48 h. The cultures were supplemented with fresh medium and subsequently infected with H. pylori for 3 h. The RNA was extracted using Isogen (Wako, Osaka, Japan), the samples were treated with DNase for 1 h, and then the samples were purified using an RNA purification kit (QIAGEN  29,000 open reading frame oligo probes. Briefly, 5 g of total RNA was amplified using the Amino Allyl MessageAmp aRNA kit (Ambion, Austin, TX). Antisense RNA (5 g) from control or experimental samples, e.g., unstimulated versus infected or control siRNA-transfected versus IKK siRNA-transfected, were labeled with Cy5 or Cy3, respectively. The two fluorescently labeled probes were mixed and applied to a microarray, followed by incubation under humidified conditions at 60°C overnight. Fluorescent images of the hybridized microarrays were scanned with a fluorescence laser confocal slide scanner (Affymetrix 428 Array Scanner; Santa Clara, CA). The images were analyzed using the ImaGene 4.2 software (Bio-Discovery, Marina Del Rey, CA) according to the manufacturer's instructions. To control for labeling differences and to reduce hybridization errors, all of the reactions were carried out in duplicate, whereby the fluorescent dyes were switched. Quantitative real-time PCR. RNA was prepared as described above from AGS cells that were infected with H. pylori for the indicated time. For quantitative PCR, cDNA was prepared using a combination of oligo(dT), random primers, and the Impron II Reverse Transcription System (Promega). Each PCR was carried out in triplicate in a 25-l volume that contained the SYBR Green Master mix (Applied Biosystems, Foster City, CA) and using the ABI Prism 7000 Sequence Detection System (Applied Biosystems). The following PCR conditions were used: 15 min at 95°C for the initial denaturation, followed by 45 cycles of 95°C for 30 s and 60°C for 30 s. Relative quantification of gene expression was performed using GAPDH mRNA as the internal standard. Two independent experiments were performed with similar results, and a representative was shown. The oligonucleotide primers for IL-8, A20, c-IAP2, MCL1, and survivin have been described previously (12,35,42,48). The primer sequences for other genes, which were designed using the Primer Express software (Applied Biosystems), were as follows: XIAP sense, 5Ј-AGTGGTAGTCCTGTTTCAGCATCA-3Ј, and antisense, 5Ј-CCGCACGGTATCTCCTTCA-3Ј; GADD45␤ sense, 5Ј-CAC GCTCATCCAGTCCTTCTG-3Ј, and antisense, 5Ј-CCGACACCCGCACGAT-3Ј; BCL10 sense, 5Ј-TTTTTTGAGACAGTCTTGCTCTATCG-3Ј, and antisense, 5Ј-AGCATGGGAGGCAGAAGTTG-3Ј.
Statistical methods. Statistical analysis was performed using the Student's t test, two sided, and Dunnett's post hoc tests for multiple comparisons. Differences were considered statistically significant with P Ͻ 0.05.

Role of IKKs on NF-B signaling in H. pylori-infected AGS cells.
To investigate the roles of IKK␣ and IKK␤ in H. pyloriinfected AGS cells, we initially performed a reporter assay for NF-B-dependent transcription using kinase mutant forms of IKK␣ and IKK␤. As reported previously (11,26), cotransfection of dominant-negative IKKs decreased H. pylori-mediated NF-B reporter activity (Fig. 1A), which indicates that the overexpression of dominant-negative IKK␣ or dominant-negative IKK␤ inhibits NF-B activation. We also assessed the effect of IKK gene silencing on this signaling pathway. The siRNAs for IKK␣, IKK␤, or nonsilencing control RNA were transfected in AGS cells, and NF-B reporter activity was analyzed with or without H. pylori infection. Similar to the effect of the dominant-negative molecules, IKK␣ and IKK␤ gene silencing decreased by 50% the H. pylori-induced NF-B reporter activation (Fig. 1B).
We then performed immunoblots for phosphorylated IB␣ to reveal the upstream event that leads to NF-B activation. As shown in Fig. 1C, IKK␤ silencing dramatically reduced H. pylori-mediated IB␣ phosphorylation. In contrast, IKK␣ silencing had a very limited effect on IB␣ phosphorylation, although the siRNA for IKK␣ apparently decreased the IKK␣ protein level. These siRNAs for IKKs had only slight effects on H. pylori-induced JNK phosphorylation. Thus, we believe that IKK␣ is involved in the NF-B signaling activation induced by H. pylori through a mechanism that is distinct from IB phosphorylation, which is transduced via IKK␤ activation.

Nuclear translocation of IKK␣, but not p100 processing, is induced by H. pylori infection of AGS cells.
Recent studies on IKK␣ or IKK␤ knockout cells have revealed the specific functions of IKKs in cytokine signaling. To elucidate the role of IKK␣ in H. pylori infection, we examined the IKK␣specific signaling pathway, namely the processing of p100, and the nuclear translocation of IKK␣. As reported for other cell lines (9), LT induced both IB␣ and p100 phosphorylation in AGS cells ( Fig. 2A). In LT-treated cells, the p100 level gradually decreased and that of p52 increased, which indicates that LT activates the NF-B alternative pathway in this cell line. In contrast, H. pylori infection induced only IB␣ phosphorylation. Neither p100 phosphorylation nor processing to p52 was observed in H. pylori-treated cells ( Fig. 2A). This suggests that the alternative pathway of NF-B activation, which includes p100 processing to p52, is not induced by H. pylori in AGS cells. To confirm these findings, we also investigated the effect of H. pylori infection or LT treatment on IKK-silenced cells. As shown in Fig. 2B, p52 protein induced by LT was severely reduced by IKK␣ siRNA but not by IKK␤ siRNA, which demonstrates the essential role of IKK␣ in the alternative pathway. In contrast, H. pylori infection did not increase the p52 protein level in any cell type, although the basal p52 protein level was slightly reduced in IKK-silenced cells. We also found that the p100 protein level was increased significantly in H. pylori-infected IKK␣-silenced cells but not in IKK␤-silenced cells. This result also indicates that p100 is a target gene of NF-B in H. pylori-infected cells, especially via the IKK␤-dependent classical pathway. Collectively, these results clearly demonstrate that H. pylori does not induce IKK␣-dependent p100 phosphorylation or its processing to p52 in AGS cells, despite p100 induction via the IKK␤dependent classical pathway. We also investigated whether H. pylori induces IKK␣ nuclear translocation. Nuclear and cytosolic fractions of H. pyloriinfected cells were analyzed by immunoblotting for IKK␣. As shown in Fig. 3A, H. pylori induced IKK␣ nuclear accumulation in a time-dependent manner. Nuclear accumulation of IKK␣ was observed 30 min after infection and increased for 1.5 h, after which the level remained the same. The time course of IKK␣ nuclear translocation was similar to that of p50 nuclear translocation.
We also performed immunofluorescence staining to confirm IKK␣ nuclear translocation. In uninfected AGS cells, IKK␣ was localized, mainly in the cytosol. However, upon infection, nuclear staining of IKK␣ was observed in about 15 to 20% of the cells (Fig. 3B and C). These results indicate that H. pylori activates IKK␣ and induces its nuclear translocation but does not induce p100 processing in AGS cells.
Factors associated with H. pylori-induced IKK␣ nuclear translocation. H. pylori activates the intracellular signaling pathways of epithelial cell lines in MOI-dependent and cag PAI-dependent manners (21,22,26). Thus, we investigated whether these bacterial factors also affect IKK␣ nuclear translocation. AGS cells were infected with H. pylori at the indicated MOI for 2 h, and nuclear extracts were subjected to immunoblotting for IKK␣. As shown in Fig. 4A, IKK␣ nuclear accumulation was observed in cells that were infected with H. pylori at an MOI of 10. The levels of IKK␣ and p50 in the nucleus increased in relation to increases in the infection ratio, up to an MOI of 100.
We also investigated the roles of bacterial virulence factors in IKK␣ nuclear translocation using cagA, cagE, cag PAI, and vacA mutant strains. Immunoblot analysis revealed that gene disruption of cagE or cag PAI reduced IKK␣ nuclear accumulation (Fig. 4B). In contrast, in cagA and vacA mutant-infected cells we observed almost the same level of nuclear IKK␣ and p50 as in wild-type-infected cells. These results indicate that the cag PAI molecular transportation system is required for IKK␣ activation. Although the CagA and VacA proteins are bacterial cytotoxins that enter epithelial cells (32,36), these bacterial toxins themselves do not induce either NF-B activation or IKK␣ nuclear translocation. We then assessed the upstream signaling event for IKK complex activation. Several studies have revealed that TAK1 transduces cytokine signaling to the IKK complex. Therefore, we used siRNAs for TAK1 and IKKs to examine the importance of these molecules for H. pylori-induced IKK␣ nuclear translocation. As shown in Fig. 4C, TAK1 silencing reduced the nuclear translocation of IKK␣ as well as that of p50. In contrast, IKK␤ silencing had no effect on IKK␣ localization while p50 nuclear translocation was severely inhibited. These results indicate that TAK1 is an important signaling intermediate for H. pylori-induced NF-B activation, which bifurcates upstream of the stimulus to both IKK␣ and IKK␤.
The role of IKKs in H. pylori-induced epithelial cell responses. It has been reported that NF-B activation induced by H. pylori mediates cytokine production and antiapoptosis in gastric epithelial cells. Therefore, we assessed whether IKK␣ activation in gastric cells affects these cellular responses. IL-8 production by H. pylori-infected AGS cells was measured by ELISA. In the control cells, approximately 2,400 pg/ml IL-8 was produced after 8 h of H. pylori infection. However, cells treated with the IKK␣ or IKK␤ siRNA showed severely decreased levels of IL-8 production. In the IKK␣-silenced cells, H. pylori induced about 1,200 pg/ml IL-8, which was approximately half the level induced in the control cells (Fig. 5A). Another chemokine observed in H. pylori-infected gastric mucosa, GRO␣, has also been reported to have chemotactic activities for neutrophils (47). The production of GRO␣ by H. pylori-infected cells was also reduced by IKK␣ or IKK␤ silencing (Fig. 5B). These results indicate that both IKK␣ and IKK␤ are necessary for chemokine production.
We also assessed the role of each IKK on cellular apoptosis. Using TUNEL staining, we evaluated the effect of H. pylori infection on serum starvation-induced apoptosis. In control siRNAtransfected cells, about 0.6% of the cells were apoptotic (Fig. 5C). The percentage of apoptotic cells was similar after IKK␣ silencing. However, in the case of IKK␤ silencing, H. pylori infection enhanced the apoptosis of AGS cells (2.2% Ϯ 0.7%; P Ͻ 0.05 compared to control transfected cells). These results indicate that the antiapoptotic effect of H. pylori is transduced mainly through IKK␤ activation.
The role of IKKs in H. pylori-induced gene transcription. We next investigated the IKK␣ target genes in H. pylori-infected AGS cells. Cells that were treated with control or IKKspecific siRNAs were infected with H. pylori for 3 h, and the transcriptional profiles were determined by microarray analysis. In the control oligonucleotide-transfected AGS cells, H. pylori infection up-regulated 181 out of 21,000 genes. The 181 genes included those for immune responses, antiapoptosis, and signal transduction; representative genes are shown in Table 1. Using siRNA, we found 15 of the genes with enhanced expression were down-regulated more than 20% by IKK␣ silencing, and 25 of the genes were down-regulated by IKK␤ silencing (Table 2). Interestingly, 12 out of 15 of the IKK␣-regulated genes were identical to IKK␤-regulated genes. These results, based on microarray experiments, indicate that most of the IKK␣ target genes in H. pylori-infected AGS cells are similar to IKK␤ target genes, which are activated via the NF-B classical pathway. In addition, it appears that the induction of several genes, such as NF-B1 and CXCL2, requires signaling via To confirm the microarray data and to evaluate sequential changes in gene induction, we performed real-time PCR for several genes. As shown in Fig. 6, the expression of IL-8 (A), cIAP2 (B), and MCL1 (C) was up-regulated by H. pylori infection but was effectively inhibited by the siRNAs for either IKK. These results are in accordance with the microarray data. A20 (D), which negatively regulates NF-B activity and is involved in antiapoptosis, was also up-regulated by H. pylori infection in a time-dependent manner and down-regulated by IKK silencing. A20 induction was not observed in the current microarray experiments, although it was observed in a previous study (28). In contrast to these H. pylori-inducible genes, other antiapoptotic genes were not affected (E to H). Thus, IKK␣ and IKK␤ appear to be equally involved in H. pylori-induced antiapoptotic gene expression, although the antiapoptotic phenotypes observed in cells silenced for individual IKK subunits were different (Fig. 5C).

DISCUSSION
In this report, we have examined the roles of IKK␣ and IKK␤ in H. pylori-infected gastric cancer cells. Both of these kinases are involved in NF-B activation and inflammatory cytokine production. IKK␤ is considered to act as a physiological IB kinase during H. pylori infection, while IKK␣ does not have this activity. Our results reveal that H. pylori induces the nuclear translocation of IKK␣, which may be one of the important roles of IKK␣ in gastric cancer cells. Chemokine expression induced by H. pylori infection was repressed by both IKK siRNAs, although the antiapoptotic effects were abrogated only in IKK␤-silenced cells. Thus, IKK␣ and IKK␤ seem to regulate independent cell responses through different mechanisms of NF-B activation in H. pylori-infected gastric cancer cells.
Although both IKK␣ and IKK␤ were discovered as stimulusdependent kinases of IB that are structurally related to each other, their roles in cell biology may be different. IKK␤ is a Microarray analysis was performed on RNA samples from H. pylori-infected AGS cells, which were treated with the control or IKK-specific siRNA. Among the genes induced by H. pylori infection (Table 1), those genes that were suppressed Ͼ20% by IKK siRNA treatment in duplicate were considered to be IKK-regulated genes. The GenBank accession number, the common name, the average fold change in control siRNA transfected cells and the mean inhibition rate by IKKs are shown. The 30% inhibition level represents 0.7-fold expression in IKK-silenced H. pylori-infected cells. ND represents data that were not available, owing to the absence of results from at least one microarray.
considered to be an essential signal transducer in cytokinemediated NF-B activation, thereby promoting cell survival and preventing apoptosis (24,25,49). However, in our analysis of H. pylori-infected cells, not only the siRNA for IKK␤ but also the siRNA for IKK␣ reduced NF-B reporter activity. Therefore, we investigated the role of IKK␣ in the NF-B pathway in AGS cells, especially with respect to IKK␣-specific signaling. We found that H. pylori induces the nuclear translocation of IKK␣, which was first reported in TNF-␣-treated cells (1,46). Similar to the classical NF-B activation mode (21,26), IKK␣ nuclear translocation is induced by H. pylori in cag PAIand MOI-dependent manners. Many bacterial components, such as peptidoglycan, lipopolysaccharide, and flagellin, are known to target cellular receptors, called Toll-like receptors, and to induce IB␣ phosphorylation and NF-B activation (reviewed in references 17 and 29). However, it has not been established whether these bacterial components induce IKK␣ nuclear translocation and inflammatory gene expression. Since the nontoxic H. pylori cag PAI mutant did not induce this type of signaling, the IKK␣ activation observed for cag-positive strains in our experiments is possibly associated with severe gastric disease. Interestingly, H. pylori did not induce the activation of the alternative NF-B pathway in AGS cells. Recent studies have shown that certain stimuli, such as LT, BAFF, and CD40, induce p100 processing to p52, which then translocates into the nucleus together with RelB (3,9,39). Furthermore, lipopolysaccharide activates the alternative pathway in pre-B cells or primary dendritic cells (31). In our experiment, LT stimulation induced p100 phosphorylation and increased p52 in the AGS cells. Thus, this cell line has a normal response with respect to alternative pathway signal transduction but is defective for activation of the H. pylori-mediated alternative pathway. H. pylori also failed to phosphorylate p100 in AGS cells. Collectively, these results indicate that H. pylori does not activate IKK␣ kinase activity in this cell line, in spite of the ability of IKK␣ to undergo nuclear translocation (Fig. 7). These results also suggest that epithelial cell lines, such as AGS and MKN45, are not stimulated by H. pylori lipopolysaccharide to activate either the classical or alternative NF-B pathway (27). In contrast to the unresponsiveness of epithelial cells, we have found that H. pylori induces activation of the alternative pathway in lymphocytes and fibroblasts (34). Thus, it appears that the ability of H. pylori to activate the alternative NF-B pathway is cell type dependent.
Nuclear translocation of IKK␣ is reported to regulate gene expression by modifying histone function in TNF-stimulated cells. The kinase activity of IKK␣ is considered to be essential for this process (1,46). In contrast, for keratinocyte differentiation and normal morphological development, which are also reported to be dependent on IKK␣, the kinase activity is not required, although its nuclear translocation is indispensable (41). In this process, IKK␣ is associated with the suppression of the fibroblast growth factor family of genes, possibly via an indirect mechanism (41). Since it is difficult to determine the essential role of IKK␣ in vivo, which may depend on the type and strength of the stimulus, the cell type, and cell environment, we have investigated IKK␣ function in H. pylori-infected gastric cells. In our experiments, IKK␣ appeared to act as a positive regulator of gene expression, thereby resembling IKK␤, since in microarray experiments H. pylori-induced expression of chemokines and antiapoptotic genes was repressed to a similar extent by IKK␣ or IKK␤ silencing. In this study, IKK␣ nuclear translocation was observed within 30 min of H. pylori infection, which is similar to the time required for IB phosphorylation by IKK␤ and which is clearly different from the kinetics of alternative pathway activation by other ligands, which usually takes several hours (9,31). Furthermore, we have clarified that TAK1 is an important upstream molecule for both IKK␣ nuclear translocation and IKK␤-dependent p50 nuclear translocation. As TAK1 is reported to be the critical activator of IKK in cytokine stimulation (33,42), it is possible that TAK1 is the common upstream molecule for the IKK␤-dependent classical pathway and IKK␣ nuclear translocation in H. pylori-infected cells. Furthermore, we found that both IKKs were involved in NF-B activation and chemokine production in H. pylori-infected cells. Thus, we speculate that both IKK␣ nuclear translocation and IKK␤dependent IB phosphorylation are required for sufficient gene expression by H. pylori (Fig. 7).
The antiapoptotic responses induced by H. pylori seemed to be transduced via IKK␤. To elucidate these IKK phenotypic differences, we carried out a comprehensive and sequential analysis of the antiapoptotic genes in IKK silencing cells. Similar to previous reports on cDNA array experiments of H. pylori-induced gene expression, we found that genes associated with immune responses and signal transductions, such as IL-8, CXCL1, CXCL2, IkBa, p105, and ICAM-1, were up-regulated in AGS cells by H. pylori infection (7,13,28). Most of these genes were shown to be induced by cag-positive H. pylori infection. This is consistent with our results demonstrating that these genes were suppressed by IKK silencing, as IKK activation by H. pylori was dependent on cag PAI. Furthermore, it has been reported that H. pylori upregulates antiapoptotic genes like MCL1, cIAP2, A20, and GG2.1 (13,28,40,48). We also found a critical role of IKKs in antiapoptotic gene regulation (Table 2). However, in spite of the differences in antiapoptotic effects (Fig. 5C), we could not find the differences in antiapoptotic gene regulation between IKK␣-silencing cells and IKK␤-silencing cells. Thus, it is possible that IKK␤ affects antiapoptosis not through gene regulation but through other biological processes, such as posttranscriptional modification via its kinase activity. Previous reports on IKK␤ knockout cells have shown that the inactivation of NF-B signaling enhances JNK activity and affects proapoptosis (44). In our experiments using IKK-silenced cells, the enhancement of JNK activity was not apparent. Furthermore, H. pylori infection did not enhance the expression of the XIAP gene (Fig. 6E), which is reported to activate JNK. Since our experiments failed to clarify the IKK␤dependent antiapoptotic mechanism, further investigations of IKK␤ will facilitate the understanding of gastric diseases associated with dysregulated apoptosis.
In conclusion, we have investigated the function of IKK␣ in the H. pylori infection model using AGS cells. IKK␣ is translocated into the nucleus upon infection, and chemokine expression is induced via IKK␣. These results suggest that IKK␣ activation in the gastric mucosa is associated with severe inflammation and inflammation-induced carcinogenesis in vivo, as is IKK␤.