Human gastric epithelial immortalized GES-1 cells representing the mature (adult) gastric epithelium (a gift from Prof. Jia, Shandong University, China) were cultured in DMEM medium (Corning, Tewksbury, MA, USA) supplemented with 10% FBS with 5% CO₂ at 37°C. The cells were subcultured every second day. Prior to the bacterial infection study, the cells were incubated in FBS‐free DMEM medium overnight at 37°C in 5% CO₂. Human fetal stomach cells (HFSC, Innoprot, USA) were purchased from Smartec Scientific Corp., Taiwan. The HFSC cells were cultured in Epithelial Pro-Conditioned Cellutions Medium (D‐PRO‐015, DV Biologics, CA) with 5% CO₂ at 37°C. The medium was changed every second day, and the cells were subcultured before they overgrew until they were induced to human stomach fetal epithelium (HSFE) cells, a cell line to represent immature (child) gastric epithelium. Before the study conduction, the cells were incubated in D‐PRO medium and DMEM (1:1 mix) with 10% FBS. Prior to the bacterial infection study, the cells were starved in FBS‐free DMEM medium overnight at 37°C in 5% CO₂.

H. pylori strain HP238 isolated from a clinical patient was used, which has been shown to express CagA, VacA, and BabA proteins in previous studies (20). The bacteria were maintained on CDC blood agar (BBL, Microbiology Systems, Cockeysville, MD) or in Brucella broth containing 10% horse serum (Gibco BRL, Life Technologies, Rockville, MD) at 37°C and incubated under micro‐aerophilic conditions (10% CO₂, 5% O₂ and 85% N₂) for 24-48 hours. The bacteria were transferred to phosphate-buffered saline (PBS) before infecting the cells. Growth density was measured spectrophotometrically at 600 nm. The infectious dose of bacteria was 1×10⁸ bacteria/ml at an OD of 1. Cells were infected at a multiplicity of infection (MOI) of 100:1.

GES-1 and HSFE cells (1 x 10⁶/well) were grown in 6-well culture dishes until approximately 80% confluence. After washing with PBS, H. pylori (HP238, MOI 100) was added to the wells without centrifugation and incubated for 30 minutes to 8 hours. The culture supernatant of each well was collected and stored for further use.

Cells were pretreated with pharmacological MAPK inhibitors for p38 (SB203580), JNK (SP600125), and ERK (PD98059) (all from Cell Signaling Technology) with variable concentrations, and DMSO as a negative control for 1 hour before H. pylori infection. siRNA against p38, ERK, JNK and control siRNA (scrambled siRNA, all from Life Technologies, Thermo Fisher Scientific, Carlsbad, CA, USA) ( Table 1 ) were transfected into the GES-1 and HSFE cells, respectively, with Lipofectamine RNAi MAX according to the manufacturer’s instructions.

After H. pylori infection, each well with GES-1 and HSFE cells was washed three times with PBS to remove unbound bacteria. The HSFE and GES‐1 cells were detached by 0.2% EDTA at 37°C for 5 minutes, then fixed with 70% alcohol at 4°C. The cells were incubated with each primary antibody (B0471: anti-H. pylori, Dako; 2-25LE: anti-Lewis b, abcan; CD15s: anti-sialyl‐Lewis x, BD) at 4°C for 30 minutes, and then incubated with the secondary antibody Alexa Fluor 488, 555 (abcan, Invitrogen) at 4°C for 30 minutes. After washing with DPBS, fluorescence was measured using BD FACSCalibur (Cell Quest). The percentage or mean fluorescence intensity was measured using FlowJo v10 (Genetech Biotech Co. Ltd. Tw). In brief, observation of adhesion intensity for H. pylori was based on flow detection of one cell at a time to measure its anti-H. pylori fluorescence value. We used infectious dose MOI 100 in each test and assumed the mean intensity could be represented the bacterial adhesion amount. All tests were done in triplicate.

The culture supernatants were centrifuged at 35,000 rpm for 5 minutes to remove bacteria and cell debris, and then stored at -80°C until use. The concentrations of IL-6 and IL-8 were measured at 0-4 hours by ELISA (R & D System) according to the manufacturer’s instructions. The absorbance of each micro-plate was read on a spectrophotometer using 450 nm as the primary wavelength and 570 nm as the reference.

GES-1 and HSFE cells were cocultured with H. pylori for the appropriate times (0.5-8 hours). Total protein of the cells was lysed using RIPA (Millipore), and the protein concentration was quantified using a BCATM Protein Assay Kit (PIERCE). Samples containing equal amounts of total protein were subjected to SDS-PAGE (Bio-Rad system). Proteins were transferred to a 0.25-μm Transblot nitrocellulose membranes (Bio-Rad), and the membranes were blocked using 5% (w/v) milk in TBST. Immunodetection of phosphorylated or total MAPKs was performed by incubating the membranes with anti-p-p38, anti-p-ERK1/2, anti-p-JNK, anti-p38, anti-ERK, and anti-JNK primary antibodies (all from Cell Signaling Technology) at a dilution of 1/1000, except for anti-actin antibody (Millipore) which was used at a dilution of 1/2000 in 5% (w/v) milk in TBST. Secondary goat anti-rabbit and anti-mouse (Chemicon/Millipore) antibodies were used at a dilution of 1/10000 in 5% milk in TBST. Western blots were developed using ECL detection reagent (Millipore) and exposed to Amersham HyperfilmTM MP (GE Healthcare). The quantification of MAPK activity was measured by densitometry from 0.5 to 4 hours.

The p38 (CRK) (NM_016823) Human Tagged ORF Clone was purchased from OriGene Technologies Inc. (Rockville, MD, USA). The procedure for generating GES-1 cells overexpressing p38 was conducted following the manufacturer’s instructions. In brief, the TransIT-X2 was warmed to room temperature and gently vortexed before use. A total of 1.5 ml serum-free culture medium was added to a sterile test tube, followed by the addition of 15 μl plasma DNA (1 μg/μl) and gentle pipetting to mix thoroughly. Next, 45 μl TransIT-X2 was added to the diluted DNA mixture and mixed thoroughly. The mixture was incubated at room temperature for 15 to 30 minutes, and then drops of the TransIT-X2/DNA complex were added to each area of the dish, with gentle shaking to allow for even distribution. The dishes were then incubated for 24-72 hours. The cells were collected for tests in vitro.

Patients with dyspepsia receiving esophagogastroduodenoscopy examination with biopsies for H. pylori infection were enrolled. Topographical specimens from the gastric antrum, body and cardia of each H. pylori-infected or non-infected child and adult (each group n=4) were used to analyze the expression of MAPK activity, which was determined by immunohistochemical staining. We selected only for patients with chronic gastritis and excluded who had peptic ulcers or gastric cancers. The age and sex were matched between H. pylori-infected and non-infected children and adults. Tissue sections were treated with primary antibodies against p-p38 (1:100), p-JNK (1:400), and p-ERK (1:500) (all from Cell Signaling Technology) at 4°C overnight in a humidified chamber. Setting step followed the manufacturer’s instructions (ECL System, Millipore Corporation, USA). Specifically bound peroxidase was detected by Chemiluminescent HRP Substrate (Max Polymer Detection System, NovolinkTM), and then exposed to X-rays (GE Healthcare, UK) for an optimal duration.

MAPK expression was analyzed in SE cells, DG cells, and LP mononuclear cells and stromal cells. The intensity of each MAPK expression was scored (range 0–4) according to the distribution of positively stained SE cells, DG cells, and LP mononuclear cells and stromal cells. The quantitative score of each MAPK expression on the topographic portion was as follows: score 0, negative or less than 1% of the stained cells; score 1, 1% to 25% of the stained cells; score 2, 25% to 50% of the stained cells; score 3, 50% to 75% of the stained cells; and score 4, 75% to 100% of the stained cells ( Supplemental Figure 1 ).

The statistics and figure were performed by software Prism6. The Student’s t-test was applied as appropriate for the parametric differences. The differences were considered to be significant at P<0.05.

HSFE cells were induced using Epithelial Pro‐Conditioned Cellutions Medium from human fetal stomach cells (HFSC) according to the manufacturer’s instructions (DV Biologics, Costa Mesa, CA). Figure 1 shows the primary cultures of HFSC cells after seeding under a phase contrast microscope at 10X at 24 and 96 hours. The cell morphology was characterized by an elongated cytoplasm and stratification with D Pro-Conditioned Medium ( Figure 1A ) and modified 2/3 E-Pro and 1/3 DMEM Medium ( Figure 1B ). Most of the cells were positive for anti-MUC5AC (pink) monoclonal antibodies, a marker of gastric epithelial cells, with Ber-EP4 (blue) nuclear staining ( Figure 1C ) and negative for vimentin (data not shown).

Comparisons of the cell colonization rate and bacterial density between the GES-1 and HSFE cells after H. pylori challenge at various time periods are shown in Figure 2 . The results showed that both bacterial colonization rate and density were positively correlated with incubation period. Moreover, both H. pylori colonization rate ( Figure 2A ) and density ( Figure 2B ) were significantly higher in the GES-1 cells than in the HSFE cells (P < 0.001).

Figure 3 shows Lewis b (Le^b) and sialyl-Lewis x (sLe^x) antigen expressions after H. pylori challenge in the GES-1 and HSFE cells at 0-4 hours over time. H. pylori infection induced Le^b and sLe^x antigen expressions in both GES-1 and HSFE cells in a time-dependent manner ( Figures 3A–D ). The percentage of Le^b antigen expression in the GES-1 cells rapidly increased to a full expression (100%) after H. pylori challenge, however this was slower in the HSFE cells ( Figure 3A ). Moreover, the intensity of Le^b antigen expression after challenge was significantly higher in the GES-1 cells than in the HSFE cells at each time period ( Figure 3B ). In addition, H. pylori induced a time-dependent increases in the percentage and intensity of sLe^x antigen expression in the GES-1 cells. However, the sLe^x antigen expression after challenge was significantly lower in the HSFE cells than in the GES-1 cells ( Figures 3C, D ).

H. pylori infection induced the expressions of IL-8 and IL-6 cytokines in both the GES-1 and HSFE cells ( Figures 4A, B ). Interestingly, the cytokine levels induced by H. pylori were much higher in the HSFE cells than in the GES-1 cells. In addition, the Western blot results showed that H. pylori-induced MAPK phosphorylation activity reached the highest level at 1 hour, and that the induction levels were different between the GES-1 and HSFE cells ( Figures 4C, D ). In the GES-1 cells, H. pylori induced a persistently high p-JNK/JNK ratio, however the p-p38/p38 ratio was only transiently activated at 1 hour. The phosphorylation of ERK was not obviously induced ( Figure 4E ). However, in the HSFE cells, H. pylori activated MAPK activity at 1 hour, and then the levels decreased ( Figure 4F ). Comparing MAPK activity between the GES-1 and HSFE cells, H. pylori induced a significantly higher p-JNK/JNK ratio in the GES-1 cells than in the HSFE cells ( Figure 4G ). However, H. pylori induced higher p-p38/p38 ( Figure 4H ) and p-ERK/ERK ( Figure 4I ) ratios in the HSFE cells than in the GES-1 cells.

We then evaluated whether H. pylori colonization intensity was associated with the differential phosphorylation activity of MAPKs between the GES-1 and HSFE cells. Figure 5 shows the impact of MAPK activation after H. pylori challenge on the colonization intensity in the GES-1 and HSFE cells. The GES-1 and HSFE cells were transfected with 10 ρg siRNA of p38, JNK, and ERK, with 8 μl RNAi MAX Lipofectamine. The siRNA transfection efficiency was confirmed by Western blot ( Figure 5A ). In the GES-1 cells, knockdown of JNK and ERK activity significantly impaired the H. pylori colonization density. However, p38 knockdown in the GES-1 did not affect H. pylori colonization density. In contrast to the GES-1 cells, knockdown of ERK, JNK, and p38 activity significantly increased H. pylori colonization intensity in the HSFE cells ( Figure 5B ).

As p-JNK activity was shown to have the greatest impact on the difference in H. pylori colonization density between the GES-1 and HSFE cells, we further investigated whether p-JNK activity affected differences in Lewis antigen expressions between the GES-1 and HSFE cells using siRNA and a p-JNK inhibitor ( Figures 6A, B ). Knockdown of JNK activity significantly decreased Le^b antigen intensity in GES-1 cells after H. pylori challenge but significantly increased the Le^b antigen intensity in HSFE cells ( Figure 6A ). However, only sLe^x antigen intensity was significantly upregulated in the HSFE cells after H. pylori challenge by JNK knockdown ( Figure 6B ). Moreover, pretreatment with a p-JNK inhibitor also decreased Le^b and H. pylori colonization intensities in GES-1 cells ( Figure 6C ) and increased the Le^b, sLe^x, and H. pylori colonization intensities in HSFE cells ( Figure 6D ). Both effects displayed a dose-dependent manner.

To confirm that H. pylori activates different gastric MAPK expressions in children and adults, we investigated the intensity of MAPK expressions in gastric biopsies using immunohistochemistry staining. Figure 7 shows the mean intensity scores of p-JNK, p-p38, and p-ERK on gastric antrum between children and adults with and without H. pylori infection (each n = 4). The intensity scores were compared topographically in superficial epithelium (SE), deep glandular (DG), and lamina propria (LP) cells. The results showed that the children without H. pylori infection had higher antral p-p38 scores in SE, DG, and LP cells than adults without infection ( Figure 7B ). The p-JNK scores in children were also higher than those in adults in DG and LP cells. Importantly, the mean p-JNK score in the SE cells was equivalent between the children and adults ( Figure 7A ). In addition, the mean p-ERK score was not different between the adults and children topographically ( Figure 7C ). In adults with H. pylori infection, the mean p-JNK score was significantly higher in the DG cells compared to those without infection ( Figure 7A ). Furthermore, H. pylori-infected adults also had a higher p-p38 score in the SE cells than non-infected subjects ( Figure 7B ). In children, H. pylori infection did not significantly upregulate p-JNK, p-p38, or p-ERK scores, except for a significantly higher p-JNK score in the LP cells than in those without infection ( Figure 7A ). There were no significant differences in p-ERK scores between the adults and children with and without H. pylori infection ( Figure 7C ).

As the pediatric gastric biopsies generally expressed higher p38 activity than those of the adults, and as H. pylori induced higher p38 phosphorylation in the HSFE cells than in the GES-1 cells, we further investigated whether the overexpression of p38 in GES-1 cells (GES-1^p38+) could abolish H. pylori-induced JNK phosphorylation ( Figure 8 ). The results showed that GES-1 cells infected with H. pylori had a significantly higher p-JNK/JNK ratio than the controls at 0.5 hours (P<0.001) and 1 hour (P<0.001). Moreover, we confirmed again that H. pylori-infected GES-1^p38+ cells significantly reduced p-JNK activity at 0.5 hours (P<0.001) and 1 hour (P<0.001) compared to GES-1 cells after H. pylori infection ( Figure 8 ). These results indicated that the higher gastric p38 activity in children counteracted the JNK phosphorylation after H. pylori infection, and that this may contribute to the lower gastric Lewis antigen expression and bacterial density in children than in adults with H. pylori infection.