Up-regulated Smad5 Mediates Apoptosis of Gastric Epithelial Cells Induced by Helicobacter pylori Infection*

The gastric pathogen Helicobacter pylori activates epithelial cell signaling pathways, and its infection induces changes in the expression of several genes in infected human gastric tissues. Recent studies have indicated that the ability of H. pylori to regulate epithelial cell responses depends on the presence of an intact cag pathogenicity island (cagPAI). We investigated altered mRNA expression of gastric epithelial cells after infection with H. pylori, both cagPAI-positive and cagPAI-negative strains, by cDNA microarray, reverse transcription PCR, and Northern blot analysis. Our results indicated that cagPAI-positive H. pylori strains (ATCC 43504 and clinical isolated strains) significantly activated Smad5 mRNA expression of human gastric epithelial cells (AGS, KATOIII, MKN28, and MKN45). We further examined whether the up-regulated Smad5 was related to apoptosis of gastric epithelial cells induced byH. pylori. Smad5 RNA interference completely inhibitedH. pylori-induced apoptosis. These results suggest that Smad5 is up-regulated in gastric epithelial cells through the presence of cagPAI of H. pylori and that Smad5 mediates apoptosis of gastric epithelial cells induced by H. pyloriinfection.


The gastric pathogen Helicobacter pylori activates epithelial cell signaling pathways, and its infection induces changes in the expression of several genes in infected human gastric tissues. Recent studies have indicated that the ability of H. pylori to regulate epithelial cell responses depends on the presence of an intact cag pathogenicity island (cagPAI). We investigated altered mRNA expression of gastric epithelial cells after infection with H. pylori, both cagPAI-positive and cag-PAI-negative strains, by cDNA microarray, reverse transcription PCR, and Northern blot analysis. Our results indicated that cagPAI-positive H. pylori strains (ATCC 43504 and clinical isolated strains) significantly activated Smad5 mRNA expression of human gastric epithelial cells (AGS, KATOIII, MKN28, and MKN45). We further examined whether the up-regulated Smad5 was related to apoptosis of gastric epithelial cells induced by H. pylori. Smad5 RNA interference completely inhibited H. pylori-induced apoptosis. These results suggest that Smad5 is up-regulated in gastric epithelial cells through the presence of cagPAI of H. pylori and that Smad5 mediates apoptosis of gastric epithelial cells induced by H. pylori infection.
Helicobacter pylori is a human pathogen that infects the gastric mucosa and causes an inflammatory process leading to gastritis, gastric ulceration, duodenal ulceration, mucosa-associated lymphoid tissue lymphoma, and gastric cancer (1). The pathogenesis of gastroduodenal diseases caused by this bacterium is not well understood. Since the whole genome of H. pylori was sequenced in 1997, several putative virulence factors, including VacA (2), IceA, OipA (3), HrgA (4), lipopolysaccharide, and the neutrophil-activating protein (5), have been elucidated. The cag pathogenicity island (cagPAI), 1 a complex of genes coding ϳ30 proteins, has been reported to be a major virulence factor of H. pylori. The cagPAI is acquired by horizontal transfer and is found in about 50 -70% of H. pylori isolates in Western countries and in more than 90% of H. pylori isolates in Asian countries, including Japan (6,7). This lesion codes for the type IV secretion machinery system forming a cylinder-like structure connected to epithelial cells (8). Many virulence gene products or other interactive proteins might be transferred into the host cells via this system. Peptic ulceration and gastric cancer occur in some people with H. pylori infection, but the majority remain asymptomatic. Although differences among the degrees of gastric mucosal damage caused by different strains should be an important factor for development of various clinical outcomes, these strain differences do not provide a complete explanation for individual differences in H. pylori infection-induced gastric mucosal injury. Therefore, it is presumed that host responses also play an important role in the outcome of H. pylori infection, interacting with virulence factors and environmental factors. Recent studies have shown that H. pylori induced various cellular responses, proliferation, apoptosis (9), cytoskeletal rearrangement (10), modification of intracellular signaling molecules (11), vacuolation (12), and cytokine secretion (13). In this study, we investigated the altered gene expression of host cells infected with cagPAI-positive or cagPAI-negative H. pylori strain and the association between the altered gene expression and the cellular responses.

EXPERIMENTAL PROCEDURES
Bacterial Strains and Cell Lines-Biopsy specimens were obtained from Japanese patients in Hokkaido University Hospital and were cultured on H. pylori-selective agar plates (Eiken Chemical Co., Ltd., Tokyo, Japan) under microaerophilic conditions (5% O 2 , 10% CO 2 , 85% N2, at 37°C; Aaero Pack Systems, Mitsubishi Gas Chemical, Osaka, Japan) for up to 5 days. Biopsies were obtained with informed consent from all patients under protocols approved by our ethics committee. The organisms were identified as H. pylori by spiral morphology and positive oxidase, urease, and catalase reactions. One colony on the agar was collected and cultured again under the same microaerophilic conditions in brain heart infusion broth (Nissui, Osaka, Japan) containing 5% (v/v) horse serum for up to 3 days. Aliquots were stored at Ϫ80°C in 10% phosphate-buffered saline containing 20% (v/v) glycerol. After thawing of aliquots of the frozen culture, bacterial suspensions were cultured at 37°C in brain heart infusion broth containing 10% fetal calf serum (Invitrogen) under microaerophilic conditions as described above on a gyratory shaker at 170 rpm for 24 -36 h to the plateau phase. The human gastric cell lines AGS, KATOIII, MKN 28, MKN 45 were obtained from the Japanese Cancer Research Resources Bank (Tokyo, Japan) and were maintained in a complete medium consisting of RPMI 1640 medium supplemented with 10% fetal calf serum, 2 mM L-glutamine, 0.2 mg of ampicillin/ml, and 100 g of kanamycin/ml.
Co-culture of Epithelial Cells with H. pylori-Human gastric epithelial cell lines were cultured in RPMI 1640 containing 10% fetal calf serum without antibiotics and used at a final concentration of 5 ϫ 10 5 /ml. Bacterial suspensions were cultured at 37°C in brain heart infusion broth containing 10% fetal calf serum under microaerophilic conditions as described above on a gyratory shaker at 170 rpm for 24 -36 h to the plateau phase. The bacteria were then suspended in sterile phosphate-buffered saline. After centrifugation, the bacteria * This work was supported in part by a Grant-in-Aid for Scientific Research from the Japanese Ministry of Education, Science, Sports and Culture (to T. S. and M. A.) and a Grant-in-Aid for Cancer Research from the Japanese Ministry of Health and Welfare (to T. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
were resuspended at a final concentration of 1ϫ10 7 colony-forming units/ml in RPMI 1640 supplemented with 10% fetal calf serum and used immediately. Gastric epithelial cells alone or cells with bacteria were cultured in tissue culture dishes (Falcon; Becton Dickinson) at 37°C in a humidified incubator in an atmosphere of 95% air and 5% CO 2 . The cells were washed with phosphate-buffered saline three times after 4, 8, 12, and 24 h. Total cellular RNA was extracted from the cells by using Isogen reagent (Nippon Gene, Tokyo, Japan) according to the manufacturer's instructions, and the amount was measured by absorbance at 260 nm.
cDNA Microarray Procedure-Poly(A) RNA was isolated from total cellular RNA (100 g) using an MagExtractor (Toyobo, Tsuruga, Japan) according to the manufacturer's instructions. Total cellular RNA was incubated with oligo(dT) magnetic beads (in the kit), and then nonspecific substance was removed by washing. 2 g of mRNA was reverse transcribed into cDNA by reverse transcriptase, ReverTraAce (Toyobo), in the presence of a cDNA synthesis primer. Biotin-labeled probes were generated by binding of biotin-16-deoxyuridine triphosphate during synthesis of cDNA. The human cDNA expression filters, human cancer filters (Toyobo) were prehybridized at 62°C for 30 min in 20 ml of PerfectHyb solution (Toyobo). After denaturalization, cDNA probes were hybridized to the filters overnight at 62°C. The membranes were washed three times with solution 1 (2ϫ SSC and 0.1% SDS) and three times with solution 2 (0.1ϫ SSC and 0.1% SDS) for 5 min at 62°C. Specific signals on the filters were detected by using a chemiluminescence detection kit, Imaging High (Toyobo), according to the manufacturer's instructions. CDP-Star was used as the chemiluminescent substrate. Images and quantitative data of gene expression levels were obtained using a Fluor-S Multiimager system (Nippon Bio-Rad Laboratories, Tokyo, Japan) and quantified into intensity of signals by using ImaGene (BioDiscovery, Inc., Los Angeles, CA).
Northern Blot Analysis-20 g of total RNA was electrophoresed on a 1% agarose gel containing 6.5% formaldehyde and then transferred onto a nylon membrane. A Smad5 probe was made from human Smad5 cDNA that corresponded to its whole coding region. Each probe was labeled with biotin using Biotin-16-dUTP (Roche Diagnostics, Tokyo, Japan). A human ␤-actin probe labeled with biotin was used as a positive control. The membrane was hybridized with the labeled probe for 20 h at 62°C in PerfectHyb (Toyobo). After hybridization, it was washed three times with 2ϫ SSC with 0.1% SDS for 10 min and washed three times with 0.1ϫ SSC with 0.1% SDS for 10 min at 62°C. Positive bands were detected by using chemiluminescence detection kit (Imaging High; Toyobo), and CDP-Star was used as the chemiluminescent substrate according to the manufacturer's instructions.
mRNA Expression by RT-PCR-First strand cDNA templates were synthesized from 2 g of total RNA using ReverTraAce and a random primer (Toyobo) according to the manufacturer's instructions. An aliquot (0.1 l) of Taq DNA polymerase and deoxynucleoside triphosphates (Takara Shuzou Co., Ltd., Shiga, Japan) was mixed with 0.5 l of a first strand cDNA sample and each primer. The primers used were Smad5F (5Ј-CAACACAGCCTTCTGGTTCA-3Ј) and Smad5R (5Ј-TTGA-CAACAAACCCAAGCAG-3Ј) for Smad5 amplification. PCR was performed using a thermal cycler (Takara Shuzou) under the following conditions: an initial denaturation for 5 min at 94°C; 30 s at 94°C, 30 s at 55°C, and 30 s at 72°C; and a final extension at 72°C for 5 min with the number of cycles at which the band intensity increased linearly with the amount of mRNA used. The PCR product was then run on 1.5% agarose gel.
Analysis of Apoptosis-We used two methods to detect apoptosis of epithelial cells induced by H. pylori infection. After co-culture of AGS cells with H. pylori for 72 h, DNA was extracted from the control and treated cells using an apoptosis ladder detection kit (Wako Pure Chemical Industries, Ltd., Osaka, Japan). Each DNA (5 g) was electrophoresed in 2% agarose gels. The gels were photographed under ultraviolet light, and DNA ladder formation was observed. Next, we carried out quantitative analysis of apoptosis. AGS cells were cultured with H. pylori for 72 h in 96-well plates (2ϫ10 4 cells/well). After centrifugation at 1500 rpm for 5 min, the supernatant was removed, and the pellets were frozen at Ϫ80°C for 15 min. Then the terminal deoxynucleotidyl triphosphate-mediated deoxyuridine triphosphate nick end labeling assay was performed using a apoptosis screening kit (Wako Pure Chemical Industries, Ltd.) according to the manufacturer's instructions. The degree of apoptosis was evaluated numerically by measuring the absorbance (490 nm).
Interference of Smad5 mRNA-Two 29-mer DNA oligonucleotides (siRNA oligonucleotide templates) with 21 nucleotides encoding the siRNA and 8 nucleotides complementary to the T7 promoter primer were chemically synthesized, desalted, and purified by reverse phase high pressure liquid chromatography. These sequences were subjected to a BLAST search (NCBI data base) to ensure that only one gene was targeted. Two 21-mer oligonucleotides (sense, 5Ј-AATTACATCCTGC-CGGTGATA-3Ј and antisense, 5Ј-AATATCACCGGCAGGATGTAA-3Ј) encoding Smad5 had no homology to those of Smad1, 2, 3, 4, and 8 in a BLAST search. The two siRNA oligonucleotide templates were hybridized to a T7 promotor primer and were extended by the Klenow DNA polymerase. The sense and antisense siRNA templates were transcribed by T7 RNA polymerase and were hybridized to create doublestranded siRNA using a Silencer siRNA construction kit (Ambion). The control and H. pylori-treated cells were grown in 96-well plates, and cationic lipid-mediated transient transfections were carried out with 50 ng of siRNA/well using GeneSilencer siRNA transfection reagent (Gene Therapy Systems, San Diego, CA). After incubation at 37°C for 24 h, Northern blot analysis was performed to assess the effectiveness of RNA interference, and quantitative analysis of apoptosis was carried out as described above.
Statistics-The data are presented as the means Ϯ S.D. The differences were examined by analysis of variance, and p values Ͻ 0.01 were considered significant.

Smad5
Up-regulation in Gastric Epithelial Cell Lines-We first examined changes in gastric cellular mRNA expression in response to co-culture with H. pylori (cagPAI-positive, ATCC 43504 strain) at 8 and 24 h by cDNA microarrays in AGS cells. Eleven housekeeping genes were used as internal controls to correct the mRNA abundance. Although the majority of genes indicated only small differences, the expression level of Smad5 mRNA increased dramatically (Fig. 1), with the relative fold changes in density to housekeeping genes being 0. H. pylori infection up-regulated Smad5 mRNA expression of AGS cells after 4 h of co-culture (Fig. 2). We examined several other gastric epithelial cell lines (KATOIII, MKN28, and MKN45) to confirm Smad5 mRNA expression after co-culture with H. pylori by gene-specific RT-PCR. Smad5 mRNA was highly expressed after co-culture with H. pylori in all tested gastric epithelial cell lines (Fig. 3a).
Smad5 Up-regulation in Native Gastric Mucosa-We tested whether Smad5 mRNA was expressed in the 10 native gastric biopsy specimens by RT-PCR. Smad5 mRNA was highly expressed in the native gastric specimens from five patients infected with cagPAI-positive strains, however faintly expressed in the five uninfected gastric specimens (Fig. 3b).
Induction of Apoptosis-DNA fragmentation was induced in AGS cells 72 h after having been co-cultured with cagPAIpositive H. pylori strains (strains 192 and ATCC 43504), whereas no DNA fragmentation was observed with cagPAInegative strains (strains 42 and 273) (Fig. 5a). Quantitative analysis of cellular apoptosis showed that cagPAI-positive H. pylori strains (strains 192, ATCC 43504, 912, 904, and 878) induced significantly greater levels of apoptosis in AGS cells than did cagPAI-negative strains and AGS cells alone (Fig. 5b).
The difference between the results obtained using cagPAIpositive strains and cagPAI-negative strains or AGS cells alone was statistically significant.
Interference of Smad5 Up-regulation and Apoptosis-The upregulated Smad5 mRNA expression was inhibited by Smad5specific RNA interference in all cagPAI-positive strain-infected AGS cells (Fig. 4b). With the suppression of Smad5 mRNA expression, the induction of apoptosis was completely inhibited in the quantitative apoptosis assay (Fig. 6). DISCUSSION The gastric pathogen H. pylori activates epithelial cell signaling pathways after infection. However, the exact signaling pathways are still unknown. The host immune response to H. pylori infection might be of importance with regard to the various clinical outcomes of infection by this organism. We now report that H. pylori can up-regulate the Smad5 expression of gastric epithelial cells and that the Smad5 up-regulation is involved in H. pylori-induced apoptosis of gastric epithelial cells. In addition, it was found that the presence of intact cagPAI is essential for Smad5mediated apoptosis of epithelial cells.
We speculated that a paracrine or autocrine system of TGF-␤ and bone morphogenetic proteins (BMP) from infected H. pylori or AGS cells are involved in the up-regulation of Smad5 expression. In human gastric epithelial cells, AGS had TGF-␤ receptors (TGF-␤RII) and BMP receptors, and up-regulation of Smad5 mRNA expression was observed after exogenous stimulation of TGF-␤1 and BMP by Northern blot analysis (data FIG. 3. H. pylori up-regulated Smad5 expression in other human gastric epithelial cell lines and in native gastric mucosa. a, total RNA was extracted from human gastric epithelial cell lines (KATOIII cells, MKN28 cells, and MKN45 cells) co-cultured with live H. pylori for the indicated time intervals, and the expression of Smad5 mRNA was analyzed by RT-PCR using the specific primers. Smad5 mRNA expressions were up-regulated in all of the tested cells. The fold change of density was indicated. ␤-Actin was amplified as a control in parallel. b, total RNA was extracted from the five native gastric biopsy specimens from five patients infected with cagPAI-positive strains or the five native gastric biopsy specimens from five uninfected patients, and the expression of Smad5 mRNA was analyzed by RT-PCR using the specific primers. Smad5 mRNA were highly expressed in cagPAI-infected gastric mucosa. The fold change of density is indicated. not shown). Although TGF-␤1 and BMP in co-cultured supernatants from H. pylori-infected and uninfected AGS cells were measured by enzyme-linked immunosorbent assay, significant differences were not found (data not shown). Additionally, TGF-␤1 mRNA and BMP mRNA were not upregulated after co-culture with H. pylori in cDNA array experiments. Furthermore, because H. pylori itself did not possess TGF-␤1 or BMP-like genes, which was examined by BLAST search (NCBI data base), it is unlikely that a paracrine or autocrine pathway of TGF-␤ or BMP from AGS infected with H. pylori or direct production of TGF-␤ or BMP from H. pylori is involved in up-regulation of Smad5 expression.
The cagPAI region encodes a novel H. pylori secretion system, type IV machinery (16), and this apparatus is essential for the induction of interleukin-8 via an NF-B-dependent transcriptional process in human gastric cells (17,18). It has recently been shown that CagA is injected from the attached H. pylori into host cells via the type IV machinery and that it forms a physical complex with SHP-2, the Src homology 2 domain-containing tyrosine phosphatase, in a phosphorylationdependent manner and stimulates the phosphatase activity (11,19). These findings suggest that protein or gene injection through the type IV machinery is a key mechanism for hostbacterial interaction induced by H. pylori infection. Consequently, it is not surprising that the transcriptional response of gastric epithelial cells is dependent on the presence of cagPAI. We therefore examined the Smad5 expression of AGS cells using cagPAI-positive and cagPAI-negative strains and that of the native gastric mucosa infected with H. pylori. Our results indicated that cagPAI-positive H. pylori strains were able to activate Smad5 mRNA expression and to induce apoptosis of the infected epithelial cells but that cagPAInegative strains were not able to activate Smad5 mRNA expression or induce apoptosis. Although CagA is the only H. pylori protein known to translocate from the bacterium into the cell via the type IV secretion system, it can be assumed that transfer of unknown genes or gene products through the type IV machinery might be necessary for upregulation of the Smad5 gene in host cells.
It has been reported that mutations in Smad4 played a significant role in the progression of colorectal tumors (20) and that a subset of families with juvenile polyposis had germ line mutations in the Smad4 gene and were at increased risk of developing gastrointestinal cancers (21). However, because there has been no report on Smad5 expression in gastrointestinal tract, the role of Smad5 in physiological or pathological status is not known.
Smad family proteins have molecular masses of about 42-65 kDa. Eight different Smads have been identified in mammals and can be classified into three subclasses, receptor-regulated Smads (R-Smads), common mediator Smads (Co-Smads), and inhibitory Smads (I-Smads) (22). Each member of the Smad family plays a different role in signaling pathways. R-Smads can be further subdivided into two subtypes, those phosphorylated after stimulation by TGF-␤ and BMP. Smad5 belongs to the latter group (23). Smad5 was isolated as dwarfin-C and was genetically implicated in TGF-␤-like signaling pathways in Drosophila and Caenorhabditis elegans (24). Suzuki et al. (25) proposed that Smad5 directs the formation of the ventral mesoderm and epidermis in Xenopus embryos. In an antisense oligonucleotide study, Smad5 was shown to mediate the growth inhibitory effect in hematopoietic cells (26), and Yamamoto et al. (27) suggested that Smad5 inhibited myogenic differentiation. Furthermore, BMP actively mediated apoptosis in the embryonic limb (28), and BMP-2 also induced apoptosis in human myeloma cell lines, probably via up-regulation of R-Smads (Smads1, 5, and 8) (29). Many studies have demonstrated that H. pylori induced apoptosis of gastric epithelial cells (30), suggesting that the up-regulated Smad5 mRNA expression might be involved in the apoptosis of gastric epithelial cells induced by H. pylori infection.
We also confirmed that only cagPAI-positive H. pylori strains were capable of inducing up-regulation of Smad5 mRNA as well as having apoptotic effects in human gastric cells. Although virulence factors, VacA, and lipopolysaccharide have been investigated as possible apoptosis-inducing factors (31), the precise intracellular signaling mechanism of apoptosis induced by H. pylori is still unknown. Our results indicated that Smad5 up-regulation might be related to the apotosis induced by cagPAI-positive H. pylori infection as one of the intracellular signaling molecules. We therefore compared the levels of H. pylori-induced apoptosis before and after suppression of Smad5 mRNA expression by RNA interference, and it was found that the induction of apoptosis was reduced to the background level after the interference. These observations suggest that Smad5 up-regulation is a key factor for H. pylori-induced apoptosis. In conclusion, H. pylori up-regulates Smad5 expression through the presence of cagPAI encoding type IV secretion machinery, and up-regulated Smad5 induces apoptotic responses in infected gastric epithelial cells.