To test the hypothesis that H. pylori infection disrupts host plasma membrane integrity, we used membrane-impermeant fluorescein isothiocyanate-dextran (FDx) [14] as a marker to detect membrane disruption. Human gastric cancer cell lines, AGS and SC-M1, were infected with H. pylori; uninfected cells were used as control. FDx was observed to label the cytoplasm of cells infected with H. pylori NTUH-GC05 (Figure 1A, AGS Hp-infected/FDx and SC-M1 Hp-infected/FDx), but not the cytoplasm of non-infected control cells (Figure 1A, AGS and SC-M1 controls with Ca²⁺/FDx). High percentages of FDx-labeled cells were identified in infected populations (AGS, 95% ± 4.3%, n = 3 slides; SC-M1, 90% ± 10%, n = 3 slides) vs. control, non-infected populations (AGS, 3.7% ± 3.9%, n = 3 slides; SC-M, 3% ± 5.2%, n = 3 slides). To exclude the possibility that the presence of intracellular dextrans was due to H. pylori cytotoxin-induced fluid phase endocytosis or pinocytosis, we treated AGS cells with cytochalasin B, which blocks these two modes of glucose uptake [15]. In non-infected cells treated with cytochalasin B (Figure 1B, Cyto B/FDx), there was apparently less labeling with FDx than in non-infected cells not treated with cytochalasin B (Figure 1B, Control/FDx). After H. pylori infection, FDx could be detected as puncta (fluorescence hot spots) in the cytoplasm despite cytochalasin B treatment (Figure 1C). Strikingly, if Ca²⁺, which is required for a membrane repair response, was omitted from the medium in the presence or absence of cytochalasin B, the FDx labeling was of a more diffuse, global nature (Figure 1C) as against the well- defined, punctate labeling of the cytoplasm observed in the presence of Ca²⁺. These results suggest that H. pylori induces plasma membrane disruptions, allowing the entry of FDx into the cytoplasm of AGS and SC-M1 epithelial cells by an endocytosis-independent mechanism. However, a repair response, elicited in the presence of Ca²⁺, restricts the diffusion of FDx into the cytoplasm of gastric epithelial cells via this route.

VacA secreted by H. pylori causes cellular vacuolation via pinocytosis [16]. CagA induces many signaling pathways in the host cell [11]. To determine whether the entry of FDx into the cytoplasm of infected cells was mediated by VacA or CagA, the cells were infected with mutant strains (HpΔvacA, HpΔcagA, or HpΔvacA/ΔcagA) (Figure 2). In the presence of Ca²⁺, FDx was present at membrane disruption sites in cells infected with mutant strains, which was consistent with the observations in cells infected with the wild-type strain. In the absence of Ca²⁺, FDx was present diffusely in the cytoplasm of the infected cells (Figure 2A–C). Cytochalasin B treatment had no effect on the patterns of FDx labeling in mutant strain-infected cells (Figure 2D–F). Thus, VacA and CagA activities did not account for the FDx entry into H. pylori-infected cells.

Consistent with previous reports [17], we found that the level of intracellular Ca²⁺ was increased in H. pylori-infected cells (Figure 3A). Annexins are a family of Ca²⁺- and phospholipid-binding proteins that have been found to exhibit a conserved core domain and play a role in membrane trafficking, vesicle aggregation, and, more recently, membrane repair [5,8,18–20]. AnxA1 was one of the first annexins to be implicated in repair [5]. In non-infected cells, AnxA1 was dispersed throughout the cell (Figure 3B, Blank/AnxA1), but it became concentrated at the cell perimeter upon treatment with a calcium ionophore (Figure 3B, +Iono/AnxA1; Supporting Information Figure S1). To determine whether H. pylori infection activates membrane repair, we monitored the localization of AnxA1 within infected cells. Infection with H. pylori induced a dramatic change in AnxA1 localization, such that the staining was observed to accumulate around the plasma membrane (Figure 3C), consistent with the translocation of AnxA1 from the cytoplasm to the plasma membrane for repair [5].

AnxA4 has been reported to be overexpressed in gastric cancer patients infected with H. pylori [21]. To determine whether AnxA4 plays an important role in responding to H. pylori infection, we monitored the localization of AnxA4 in AGS and SC-M1 cells using time-lapse microscopy. Cells were transfected with enhanced green fluorescent protein (EGFP)-AnxA4 fusion protein and then infected with Hoechst-labeled H. pylori. In the absence of H. pylori infection, AnxA4 was uniformly dispersed throughout the cell, including the cytoplasm and nucleus (Figure 3B, Blank/AnxA4). Following H. pylori contact with the cell surface, AnxA4 gradually (over an interval of 99 to 168 minutes) accumulated at the sites of infection on the plasma membrane (Figure 3D).

To determine whether the observed AnxA4 translocation to the plasma membrane promoted a repair process, we overexpressed AnxA4 by transfecting cells with expression vectors or knocking down AnxA4 expression using small interfering RNAs (siRNAs). As expected, there was a 26-fold increase in FDx labeling (n = 15) in the cytoplasm of infected cells transfected with an empty vector over that observed in infected cells transfected with a vector encoding AnxA4 (Over-AnxA4; Figure 4A). Additionally, there was a 17-fold increase in FDx labeling (n = 15) in AnxA4-knockdown (siAnxA4) cells over control siRNA (siControl)-transfected cells (Figure 4B). This indicates that AnxA4 promotes membrane repair following H. pylori infection. Resealing of the plasma membrane is facilitated by the recruitment of intercellular vesicles derived from the endoplasmic reticulum, Golgi compartment, and lysosomes [22,23]. We found that the surface of infected cells displayed increased levels of the lysosomal marker LAMP-2 (Figure 5A). Furthermore, AnxA4 increased LAMP-2 expression on the cell surface of infected cells (Figure 5B,C). Thus, these results indicate that AnxA4 is involved in a H. pylori-induced repair response that displays the expected exocytotic events.

Long-term chronic H. pylori infection could lead to increased epithelial cell proliferation, ultimately promoting the development of gastric cancer [24]. Thus, we speculate that continuing cell membrane disruptions that induce a repair response might be related to gastric carcinogenesis. To explore the relationship between membrane repair response and cell proliferation, we used glass beads and electroporation to create membrane injury and induce membrane repair responses [25,26]. FDx was detected in the cytoplasm of gastric cells that were loaded on glass beads or were electroporated, indicating membrane injury (Figure 6B,D, and Supporting Information Figure S2B). AGS cells damaged by the glass beads showed an increased growth rate compared with non-treated control cells (p < 0.01; Figure 6A,B). Similarly, electroporated AGS cells displayed an increased growth rate compared with non-electroporated control cells (p < 0.01; Figure 6C,D). The non-transformed human gastric cells Hs 738.St/Int also showed a similar response after electroporation (Supporting Information Figure S2). These results suggest that one of the repair responses to plasma membrane injury is cell proliferation.

Following a bacterial infection, the activation of host defense mechanisms to reduce damage is crucial for host survival. Recent reports have shown that host cells are actively involved in membrane repair responses to bacterial infections. Host cells have been reported to induce plasma membrane repair response to membrane permeabilization caused by two human gastrointestinal pathogens, Salmonella enterica serovar Typhimurium and Yersinia pseudotuberculosis [3]. To avoid host immune responses, Mycobacterium tuberculosis, a virulent lung pathogen, impairs plasma membrane repair [1].

H. pylori is a human gastric pathogen and a risk factor for gastric malignancy. However, the link between H. pylori and plasma membrane repair response is currently unknown. Here, cytochalasin B-resistant staining of the cytoplasm with a membrane-impermeant probe, FDx, develops sites of H. pylori interaction with the plasma membrane of host gastric cells. This staining was observed after infection with both wild-type and virulence factor-deficient H. pylori strains. It is worth noting that the FDx-labeling pattern produced as a result of H. pylori-induced membrane disruptions was punctate and localized to the site of bacterium/cell interaction (Figures 1 and 2), whereas the FDx-labeling pattern obtained through larger plasma membrane disruptions, such as those produced by glass beads (Figure 6B), was diffuse and more global. Similar phenomena were observed in damaged BS-C-1 cells and sea urchin eggs [27,28].

We further propose that H. pylori induces Ca²⁺ entry and consequent membrane fusion events. These results mimic the findings from an earlier report: when Ca²⁺ was injected into the cell cytoplasm via a microneedle, the Ca²⁺-containing solution was sequestered behind a newly formed membrane barrier as it left the microneedle [29]. This barrier was presumed to be formed via homotypic fusion events involving one or more cytoplasmic organelles.

Some virulence factors used by H. pylori have been reported that they can cause the disruption of cell junction structures, e.g., CagA [12,30] and urease [31]. Lai et al. and Amieva et al. have shown that the disruption of apical-junctions in primary gastric epithelial cells can be caused by CagA [12,30]. Moreover, the disruption of the tight junctions in cells can be caused by ammonium which is produced by H. pylori urease [31]. A recent study has shown that the bacterial toxin streptolysin O (SLO) can induce endocytosis to remove the lesions of the plasma membrane [32]; however, H. pylori does not possess SLO. Although in the present study we demonstrated that neither VacA nor CagA is involved in membrane disruption, the mechanism by which sequestration is mediated in H. pylori-infected cells remains to be elucidated.

AnxA1 participates in the plasma membrane repair response and exhibits a Ca²⁺ sensitivity similar to that of AnxA4 for translocation to the plasma membrane [5,33]. In the present study, we showed for the first time that AnxA4 is involved in plasma membrane repair; it localizes to the membrane disruption sites, and repair is apparently enhanced by AnxA4 overexpression and inhibited by gene silencing. We also found that AnxA4 overexpression significantly increases cell growth rate, whereas AnxA4 knockdown significantly decreases the growth rate (Supporting Information Figure S3). In a previous study, we showed that AnxA4 expression is increased in H. pylori-associated tumors [21]. Moreover, we have shown that disruption of the plasma membrane induced by glass beads or electroporation can activate epithelial cell proliferation.

These results suggest that long-term H. pylori-induced plasma membrane repair could have oncogenic potential. In previous studies, AnxA4 has been reported that is associated with anti-apoptosis and a biomarker of colorectal cancer for tumor diagnosis [34,35]. It also has been observed to be overexpressed in other cancers, including prostate cancer [36], renal carcinoma [37], pancreatic adenocarcinoma [38] and clear cell carcinoma of ovary [39,40]. In this study, the altered location and expression of AnxA4 could be considered as a marker for H. pylori-induced plasma membrane repair response. We suggest that targeting of AnxA4 might be a new therapeutic strategy for H. pylori-induced carcinogenesis. This study could provide a new direction in cancer-drug development.

The H. pylori (NTUH-GC05) strain (vacA+/cagA+) from the stomach of a male gastric cancer patient at the National Taiwan University Hospital was obtained in 1991. The H. pylori strain NTUH-GC05 and GC05ΔcagA were collected from the National Taiwan University Hospital (NTUH), as described elsewhere [12,41]. H. pylori was grown on Columbia blood agar base (BD Difco) containing 5% sheep blood and incubated for 2–3 days in microaerophilic conditions (5% O₂, 10% CO₂, 85% N₂) at 37 °C.

Human stomach adenocarcinoma AGS cells (CRL-1739, ATCC) and SC-M1 cells (cultured from a poorly differentiated adenocarcinoma that showed no metastasis to lymph nodes or adjacent organs) [42] were grown in 90% RPMI 1640 medium (Biological Industries) supplemented with 1% penicillin/streptomycin and 10% fetal bovine serum (Biological Industries). Hs 738.St/Int fibroblast cells (CRL-7869, ATCC) are non-transformed, human fetal gastric/intestinal cells and were grown in 90% Dulbecco’s modified Eagle’s medium (Gibco-Invitrogen) with 10% fetal bovine serum. Cells were cultured at 37°C in a controlled humidified atmosphere in an incubator containing 5% CO₂.

To generate a vacA mutant and a vacA/cagA double mutant, the vacA gene was amplified using PCR with the primer pair vacA-F and vacA-R (Supporting Information Table S1), and the resulting fragments were cloned into a pGEM-T easy vector (Promega, Madison, WI, USA). The cloned vacA gene was then digested with EcoRV, dephosphorylated, and finally ligated to the kanamycin resistance gene. The resulting plasmid was introduced by natural transformation into H. pylori GC05 and GC05ΔcagA [41]. The vacA insertion mutants were selected using 10 μg/mL kanamycin and verified by PCR using the primer pair vacA-F2 and vacA-R2.

Full-length anxa4 was amplified by PCR using the primer pair anxa4-F and anxa4-R, and the amplification product was inserted into the HindIII/EcoRI sites of pcDNA 3.1(+) (Invitrogen, Burlington, ON, USA). For immunofluorescence analysis, anxa4 was amplified by PCR using the primer pair anxa4-F2 and anxa4-R2, and the amplification product was inserted into the HindIII/pst I sites of pEGFP-C1 (BD Clontech, Palo Alto, CA, USA). AnxA4-specific siRNA and negative control Stealth siRNA (Stealth™ RNAi) were purchased from Invitrogen. Cells were cultured in six-well plates or on coated cover slips for 24 h. Cells were then transiently transfected with pcDNA 3.1(+)/pEGFP-C1/AnxA4 (8 μg for a six-well plate; 0.4 μg/mL for a 96-well E-plate) or AnxA4 siRNA (100 pmoles for a six-well plate; 10 pmoles for a 96-well E-plate) using Lipofectamine 2000 (Invitrogen, Burlington, IA, USA) according to the manufacturer’s instructions. The efficiency of expression vector and siRNA transfection was determined by immunoblotting. After transfection for 48 h, the differential expression of proteins and genes was detected.

The mouse monoclonal antibodies used in this study were as follows: AnxA1 (sc-12740) from Santa Cruz Biotechnology; LAMP-2 (ab25631) from Abcam; IgG₁ isotype (555746) from BD Biosciences; and α-tubulin (T5168) from Sigma. The rabbit polyclonal antibodies were as follows: H. pylori (AHP602H) from AbD Serotec. The goat polyclonal antibody AnxA4 (sc-1930) was from Santa Cruz Biotechnology.

To determine AnxA4 localization after H. pylori infection, AGS or SC-M1 cells were plated on Lab-Tek™ chamber slide™ system (Nunc, Roskilde, Denmark). After 24-h incubation, cells were transfected with pEGFP-C1/AnxA4 for 48 h prior to infection. H. pylori was resuspended in serum-free culture medium and stained with Hoechst 33258 (Sigma) for 1 h. The stained cells were centrifuged, washed twice, and then resuspended in fresh serum-free culture medium. Cultured cells were replaced with fresh serum-free culture medium (1 mL per well) and infected with H. pylori at a multiplicity of infection (MOI) of 150. Fluorescence images of living cells were captured by fluorescence microscopy (SC-M1: Zeiss Axiovert 200M (Zeiss); AGS: Nikon A1 confocal microscope (Nikon)). Images were processed using MetaMorph software (Version 7.7; Molecular Devices: Sunnyvale, CA, USA, 2010).

Cells (8 × 10⁵ cells/well) were cultured on glass cover slips coated with poly-l-lysine in six-well plates. To observe plasma membrane-disruption sites, cells were treated with cytochalasin B (10 μg/mL) for 1 h prior to infection and then supplemented with fluorescein isothiocyanate-dextran (FDx) (5 mg/mL; 10 kDa) for 30 minutes after infecting the cells with H. pylori for 2.5 h. To examine the effect of the absence of Ca²⁺ on H. pylori-induced plasma membrane disruptions, cells and H. pylori were co-cultured in Ca²⁺-free medium, which was prepared by adding the Ca²⁺ chelator EGTA (5 mM) to remove extracellular calcium from the medium. To determine the cellular location of AnxA1 and AnxA4 in the presence of high intracellular [Ca²⁺]i, the Ca²⁺ ionophore, ionomycin (5 μM), was added to the cells. After infection (MOI = 150) for 3 h, cells were fixed with 4% paraformaldehyde for 15 minutes, permeabilized with 1% Triton-X-100 in PBS (to observe the location of AnxA1 and AnxA4) for 30 minutes, and blocked with 0.1% BSA/PBS overnight at 4 °C. Cells were incubated with mouse anti-AnxA1 antibody (1:100), goat anti-AnxA4 antibody (1:100), or rabbit anti-H. pylori antibody (100 μL) for 1 h at RT. After washing, FITC-labeled anti-mouse (Sigma), TRITC-labeled anti-goat (Sigma), or Cy5-labeled anti-rabbit (Millipore) were used as secondary antibodies at a 1:100 dilution and applied for 1 h at RT. The cover slips were then washed and mounted onto slides. Immunostaining of the cells was observed using confocal microscopy with a Plan-Apochromat 63x/1.40 oil M27 objective (Zeiss LSM 510). Images were analyzed using MetaMorph (Molecular Devices) software.

To measure intracellular [Ca²⁺]i, cells were infected (MOI = 150) for 3 h, and both infected and non-infected cells were washed in Hanks’ balanced salt solution (HBSS) buffer (137 mM NaCl, 5.33 mM KCl, 4.2 mM NaHCO₃, 0.44 mM KH₂PO₄, 0.34 mM Na₂HPO₄, 5.56 mM glucose; pH 7.3) and then loaded with 4 μM of the Ca ²⁺ indicator, Fluo-3-AM/pluronic acid F-127, for 1 h at 37 °C. Subsequently, cells were harvested using trypsin and then resuspended in HEPES buffer (137 mM NaCl, 5 mM KCl, 1 mM Na₂HPO₄, 5 mM glucose, 1 mM CaCl₂, 0.5 mM MgCl₂, 1 mg/mL BSA, and 10 mM HEPES, pH 7.4). Cells cultured in the presence of ionomycin (5 μM) were used as a positive control for increased Ca²⁺ concentration in the cytoplasm. To determine the cell surface expression of LAMP-2, cells were resuspended in PBS and then fixed with 2% paraformaldehyde, centrifuged, washed in PBS, and blocked with 2% BSA/PBS for 15 minutes. The mouse anti-LAMP-2 antibody and mouse IgG₁ isotype control were used at a dilution of 1:200 for 30 minutes at RT, following which the cells were washed and stained with anti-FITC secondary antibody for 30 minutes at RT. Cells were finally washed and fixed with 4% paraformaldehyde. Fluorescence intensity was determined using the FACSCalibur System (BD Biosciences, San Diego, CA, USA), and the data were acquired by analyzing at least 10,000 cells from each sample.

AGS cells (1 × 10⁴ cells/well for glass bead stimulation; 1.5 × 10⁴ cells/well for electroporation) and Hs 738.St/Int cells (1 × 10³ cells/well) were loaded in each well of a 16-well microtiter E-plate. Each well contained microelectronic sensor arrays at the base to detect the cell index (CI). For transfection experiments, after incubation for 24 h, AGS cells were transfected with expression vectors or siRNAs for 6 h and monitored for a total of 84 h. AGS cells were resuspended in RPMI medium, and Hs 738.St/Int cells were resuspended in DMEM medium containing 10% FBS. The cells (4 × 10⁵ cells in a 400-μL cell suspension) were placed in 0.2-cm cuvettes (Bio-Rad, Hercules, CA, USA) and electroporated twice at 200 V and 100 μF at 5-minute intervals. After electroporation, cells were incubated with 5 mg/mL of FDx for 5 minutes to allow for analysis of cell membrane damage by fluorescence microscopy. Glass beads (Sigma, St. Louis, MO, USA) were used to create membrane injury as previously described [26], with slight modifications. Briefly, glass beads (bead size, 450–600 μm; bead weight, 0.03 g/well in E-plate) were carefully placed onto a plate and gently rocked for 1 minute. The E-plate was placed in the Real-Time Cell Analyzer (RTCA) system and incubated in an incubator containing 5% CO₂ at 37 °C. The level of cell proliferation was represented as CI, which was based on the electrical impedance measured using the xCELLigence system (Roche, Penzberg, Germany).

Data were expressed as mean ± standard deviation (SD). Difference between independent groups was analyzed using a two-tailed Student’s t test. Data obtained from the cell proliferation assay were analyzed using the two-sample Kolmogorov-Smirnov test. A p value of 0.05 indicated statistical significance.