Helicobacter pylori Infection Activates Src Homology-2 Domain–Containing Phosphatase 2 To Suppress IFN-γ Signaling

Type II immune IFN-γ, which is predominantly produced by NK cells, T cells, and APCs, is an important cytokine for inflammation and immune defense against microbial infections and tumorigenesis (1). The binding of IFN-γ initially leads to IFN-γ receptor (IFNGR) dimerization, followed by activation of JAK1/2 and recruitment of STAT1. JAK2 mediates STAT1 phosphorylation, leading to a STAT1 homodimer complex that translocates into the nucleus and turns on a variety of IFN regulatory factors (IRFs) (2). Under IFN-γ stimulation, cytokines and other inflammatory modulators show cross talk, leading to multifaceted cellular responses (3). For IFN-γ homeostasis, negative regulators include suppressor of cytokine signaling (SOCS) 1 and SOCS3, which can bind to IFNGRs to attenuate JAK/STAT signaling (4, 5), and Src homology-2 (SH2) domain–containing phosphatase (SHP) 2, which can dephosphorylate JAK2 and STAT1 (6, 7). Normally, SHP2 is inactive due to mutual allosteric inhibition, with its N-terminal SH2 (N-SH2) domain interacting with its protein phosphatase domain. To activate SHP2, SHP2-binding proteins bind competitively to the N-SH2 domain and lead to strong SHP2 activation, whereas some intramolecular responses may cause tyrosine phosphorylation of SHP2 at the C terminus, inducing an allosteric interaction with the N-SH2 domain (8). Infection by pathogens may subvert IFN-γ responses for both host immunomodulation and immunopathogenesis (9, 10).

Helicobacter pylori, a Gram-negative spiral bacteria, is the most common pathogen in the world and may cause immunopathogenesis in the form of atrophic gastritis, intestinal metaplasia, and gastric carcinogenesis through a complex mechanism involving the induction of the major virulence factors cytotoxin-associated gene A (CagA) and vacuolating cytotoxin A (VacA) (11–13). During H. pylori infection, a type IV secretion system is formed; proteins encoded by genes within the cag pathogenicity island can export CagA into host cells. Under exogenous CagA overexpression, CagA directly mediates SHP2 activation by binding to SH2 domains, causing SHP2 conformation change and activation without phosphorylating C-terminal tyrosine residues 542 and 580 (14). CagA has been suggested to be oncogenic because CagA-activated SHP2 is required for activation of the Ras pathway, which can stimulate cell proliferation, differentiation, and survival (8, 15, 16). CagA transgenic mice spontaneously develop gastrointestinal and hematopoietic neoplasms (17). Thus, CagA-positive H. pylori strains are associated with a higher risk of developing gastric cancer than CagA-negative strains. However, the effects of CagA on IFN-γ immunosurveillance are unknown.

To escape immune defenses, H. pylori may have developed several mechanisms, including phagocytosis resistance, modulation of dendritic cell activity, and alterations of T cell responses (11, 12, 18). An increase in serum level of IFN-γ occurs in H. pylori–infected patients and experimental animal models and is a predictor of clinical outcome (19, 20), and IFN-γ is also increased in PBMCs infected with H. pylori (21). The pathogenic and antimicrobial roles of IFN-γ were identified because mice deficient in IFN-γ, T-bet (a transcription factor of IFN-γ), or IRF1 (a downstream modulator of IFN-γ) fail to develop gastric inflammation but have increased susceptibility to H. pylori colonization (12, 22–26). Moreover, polymorphisms in IFNGRs affect H. pylori infection in human (27). IFN-γ also confers antimicrobial activity directly against CagA expression in H. pylori infection (28) and prevents gastric carcinogenesis by inducing epithelial cell autophagy and T cell apoptosis (29). However, H. pylori may disrupt IFN-γ signaling through an unknown mechanism (30). In addition to acting as a tumor-promoting factor, SHP2 is a well-known negative regulator of IFN-γ/STAT1 signaling. Therefore, we hypothesize that H. pylori infection causes IFN-γ–mediated gastric inflammation while developing resistance to escape from IFN-γ–mediated bacterial clearance and anticancer. In this study, using a clinically isolated H. pylori strain, HP238, and its cagA gene isogenic mutant strain, HP238^CagAm, we investigated the regulation of CagA-activated SHP2 on IFN-γ resistance during H. pylori infection.

Values are means ± SD. Statistical analysis of data analyses were performed using Prism version 5 (GraphPad Software, San Diego, CA). Two sets of the data were analyzed by an unpaired Student t test. Three or more sets of data were analyzed by one-way ANOVA with Tukey multiple-comparison posttest. Statistical significance was set at p < 0.05.

The reagents and Abs used were SHP2 inhibitor NSC 87877, reactive oxygen species (ROS) scavenger N-acetylcysteine (NAC), DMSO, and DAPI (Sigma-Aldrich, St. Louis, MO); recombinant human IFN-γ (PeproTech, Rocky Hill, NJ); anti–phospho-CagA at Tyr⁹⁷² and CagA (Santa Cruz Biotechnology, Santa Cruz, CA); Abs against phospho-STAT1α/β at Tyr⁷⁰¹, STAT1α/β, IRF1, phospho-SHP2 at Tyr⁵⁴², SHP2, SOCS1, SOCS3, caveolin-1, E-cadherin, and rabbit IgG (Cell Signaling Technology, Beverly, MA); Abs against IFNGR1 and IFNGR2 (Abcam, Cambridge, MA); mouse mAb specific for β-actin and tubulin (Chemicon International, Temecula, CA); and Alexa Fluor 488– or 594– and HRP-conjugated goat anti-mouse, goat anti-rabbit, and donkey anti-goat IgG (Invitrogen, Carlsbad, CA). All drug treatments were assessed for cytotoxic effects using cytotoxicity assays prior to experiments. Noncytotoxic dosages were used in this study.

The CagA-positive H. pylori clinical isolate strain HP238, the isogenic cagA mutant strain HP238^CagAm, and the standard strain ATCC43504 were obtained from the Dr. Jiunn-Jong Wu (Department of Medical Laboratory Science and Biotechnology, National Cheng Kung University) and exhibited according to previous studies (31). As compared with ATCC43504, which normally presents an EPIYA-ABCCC motif, the genomic sequences of the clinical isolate HP238 contain are currently under investigation. According to the previous studies, HP238 expresses CagA, VacA, and BabA proteins (32). Most strains of H. pylori in east Asia, particularly in Taiwan, encoded CagA containing the EPIYA-ABD motif (33). H. pylori were heat-killed (HK) by boiling at 100°C for 10 min, as described previously (34).

Human gastric adenocarcinoma cells MKN45 (JCRB0254; Japanese Collection of Research Bioresources) and AZ-521 (JCRB0061), human gastric epithelial immortalized GES-1 cells (Institute of Biochemistry and Cell Biology at the Chinese Academy of Sciences; Shanghai, China), and human leukemia cell line U937 (ATCC CRL-1953.2) were grown routinely on plastic in RPMI 1640 medium (RPMI; Invitrogen Life Technologies, Rockville, MD), with l-glutamine and 15 mM HEPES supplemented with 10% heat-inactivated FBS (Invitrogen Life Technologies), 50 U penicillin, and 50 μg ml⁻¹ streptomycin. Cells were maintained in a humidified atmosphere with 5% CO₂ and 95% air. To address the question of whether H. pylori infection causes IFN-γ resistance, in addition to cell lines tested in this study, the isolated primary murine gastric epithelial cells from C57BL/6J mouse gastric tissues were also used. The isolation procedure was carried out according to the previous study (35). Eight- to 12-week-old progeny of wild-type C57BL/6J mice from The Jackson Laboratory (Bar Harbor, ME) were used in our experiments. They were fed standard laboratory food and water ad libitum in the Laboratory Animal Center of the National Cheng Kung University. The animals were raised and cared for according to the guidelines set by the National Science Council of Taiwan. The experimental protocols adhered to the rules of the Animal Protection Act of Taiwan and were approved by the Laboratory Animal Care and Use Committee of National Cheng Kung University. The morphology and E-cadherin expression of isolated cells were also determined by microscopic observation (IX71; Olympus, Tokyo, Japan) and immunostaining (as described below), respectively.

To evaluate cell damage, lactate dehydrogenase activity was assayed using a colorimetric assay (Cytotoxicity Detection Kit; Roche Diagnostics, Lewes, U.K.) according to the manufacturer’s instructions. A microplate reader (SpectraMax 340PC; Molecular Devices, Sunnyvale, CA) was used to measure the absorbance at 620 nm with a reference wavelength of 450 nm, and data were analyzed using the Softmax Pro software (Molecular Devices).

Total cell lysates were extracted, and proteins were separated using SDS-PAGE and then transferred to a polyvinylidene difluoride membrane (Millipore, Billerica, MA). After blocking, blots were developed with the indicated Abs and developed using an ECL Western blot detection kit (Millipore) according to the manufacturer’s instructions. For membrane/cytosolic protein analysis, protein fractions were isolated with a Compartmental Protein Extraction Kit (Calbiochem, San Diego, CA) according to the manufacturer’s instructions. The relative signal intensity was quantified using ImageJ software (version 1.41o) from W. Rasband (National Institutes of Health, Bethesda, MD). The changes in the ratio of proteins compared with the normalized value of untreated cells (indicated protein/β-actin or phosphorylated protein/total protein/β-actin) were also shown.

For the luciferase reporter assay, the cells were transiently cotransfected using GeneJammer (Stratagene) with IRF1 promoter-driven luciferase reporter (0.2 μg) and 0.01 μg Renilla luciferase-expressing plasmid (pRL-TK; Promega). At 24 h after transfection, the cells were treated with IFN-γ for 1 h, lysed, and then harvested for luciferase and Renilla measurements using a luciferase assay system (Dual-Glo; Promega). For each lysate, the firefly luciferase activity was normalized to the Renilla luciferase activity to assess transfection efficiencies.

To detect the expression of E-cadherin, IFNGR1, IFNGR2, CagA, phosphorylated SHP2 (Tyr⁵⁴²), and SHP2, we fixed, stained, and analyzed the cells as described previously (36). Cells were stained with primary Abs and then incubated with a mixture of Alexa Fluor 488– or 594–conjugated goat anti-mouse or rabbit IgG. Cells were analyzed using flow cytometry (FACSCalibur; BD Biosciences, San Jose, CA) with excitation set at 488 nm; emission was detected with the FL-1 channel (515–545 nm) and the FL-2 channel (525–625 nm). Samples were analyzed using CellQuest Pro 4.0.2 software (BD Biosciences), and quantification was conducted using the WinMDI 2.8 software (The Scripps Institute, La Jolla, CA). Small cell debris was excluded by gating on a forward scatter plot. For confocal microscopy, DAPI (5 μg ml⁻¹) was used for nuclear staining. The cells were then visualized using a confocal laser scanning microscope (Digital Eclipse C1 si-ready; Nikon, Tokyo, Japan). A differential interference contrast image was also collected by acquiring a series of sections along the optical (z) axis of the microscope.

Protein was downregulated using lentiviral expression of short hairpin RNA (shRNA) targeting human SHP2 (TRCN0000005003 containing the following shRNA target sequence: 5′-CGCTAAGAGAACTTAAACTTT-3′) and a negative control construct (luciferase shRNA). shRNA clones were obtained from the National RNAi Core Facility (Institute of Molecular Biology/Genomic Research Center, Academia Sinica, Taipei, Taiwan). Lentiviruses were prepared and cells were infected according to previously described protocols (36). Briefly, MKN45 cells were transduced with a lentivirus at an appropriate multiplicity of infection (MOI) in complete growth medium supplemented with polybrene (Sigma-Aldrich). After transduction for 24 h and puromycin (Calbiochem) selection for 3 d, protein expression was monitored using Western blot analysis.

For coimmunoprecipitation, 100 μg cell lysate from cells with H. pylori infection was incubated together with 5 μg protein G (Amersham Biosciences) and 2 μg anti-CagA IgG overnight at 4°C. The expression of SHP2 and phosphorylated SHP2 (Tyr⁵⁴²) was determined using Western blotting, as described above.

Intracellular oxidative stress was measured using dichlorodihydrofluorescein diacetate oxidation. The cells were exposed to 20 μM 5-(and-6)-chloromethyl-2′,7′- dichlorodihydrofluorescein diacetate, acetyl ester (CM-H₂DCFDA) (Invitrogen) for 1 h. The cells were then analyzed using the FL-1 channel (515-545 nm) with an FACSCalibur cell sorter. After another wash with PBS, the cells were analyzed using flow cytometry (FACSCalibur) with an excitation wavelength of 488 nm. The ROS levels are reported as the mean fluorescence intensity of the total cells using CellQuest Pro 4.0.2 software, and quantification was performed with WinMDI 2.8 software. Small cell debris was excluded by gating on a forward scatter plot.

To study the effects of H. pylori infection on IFN-γ signaling in gastric epithelial cells, STAT1 phosphorylation at tyrosine 701 and IRF1 transactivation were assessed using Western blotting and a luciferase reporter assay, respectively, according to our previous report (37). Expression of CagA protein was used as the positive signal for H. pylori infection. Our results showed that 6 h postinfection, viable (Fig. 1A, left panel) but not HK (Fig. 1A, right panel) wild-type H. pylori strain HP238 caused a marked decrease in IFN-γ–induced STAT1 phosphorylation. The effects of H. pylori infection on IFN-γ resistance was also confirmed in cells treated with a higher dose of IFN-γ (data not shown). Further, viable HP238 infection caused a significant (p < 0.05) decrease in IFN-γ–induced IRF1 transactivation as early as 1 h (Fig. 1B, top panel), and this induction was extended to 6 h (Fig. 1B, bottom panel). Checking the expression of STAT1, a selected downstream targeting gene of IRF1 (Supplemental Fig. 1), confirmed that H. pylori infection attenuated not only STAT1 phosphorylation 1 h posttreatment but also protein expression at the late stage of treatment (untreated, 1.00; IFN-γ, 1.21; and HP238 plus IFN-γ, 1.04) (Fig. 1C). Using standard H. pylori strain ATCC43504, results also showed that 6 h postinfection of H. pylori caused a decrease in IFN-γ–induced STAT1 phosphorylation (Fig. 1D, left panel) and IRF1 transactivation significantly (p < 0.05; Fig. 1D, right panel). In addition to MKN45 cells, another gastric epithelial cell line AZ-521 (Fig. 1E) and immune monoctyic U937 cells (Fig. 1F) were studied for confirming the similar effect of H. pylori infection on blocking IFN-γ–induced STAT1 phosphorylation and IRF1 transactivation. Furthermore, by using isolated primary murine gastric epithelial cells (Fig. 1G, left panel), H. pylori infection caused inhibition on IFN-γ–induced STAT1 phosphorylation (Fig. 1G, right panel). These findings show that H. pylori infection results in IFN-γ resistance.

To investigate the molecular basis for the IFN-γ resistance caused by H. pylori infection, flow cytometric analysis was first used to detect the surface expression of IFNGR1 and IFNGR2 to determine whether H. pylori infection affected these receptors. No change in IFNGR1 or IFNGR2 expression was found after H. pylori infection (Fig. 2A). We next investigated whether H. pylori infection causes alterations in negative regulators of IFN-γ homeostasis, including SHP2, SOCS1, and SOCS3 (2). Western blotting showed that the protein levels of SHP2, SOCS1, and SOCS3 were unchanged after H. pylori infection; however, viable (Fig. 2B, left panel) but not HK (Fig. 2B, right panel) HP238 induced tyrosine phosphorylation of SHP2 at the C-terminal residue 542. To verify the activation of SHP2, the downstream signals of SHP2 were confirmed by the detection of activation of MEK/ERK (data not shown) (8, 15, 16). These results demonstrated that H. pylori infection causes SHP2 activation.

To evaluate the essential role of active SHP2 in IFN-γ resistance, we inhibited SHP2 pharmacologically using the selective inhibitor NSC 87877, which binds to the catalytic cleft of SHP2 (37). Western blotting and reporter assay showed that 6 h post–H. pylori infection, IFN-γ–induced STAT1 phosphorylation (Fig. 3A) and IRF1 transactivation (Fig. 3B) in cells for 1 h posttreatment and IFN-γ–induced STAT1 expression 6 h posttreatment (untreated, 1.00; IFN-γ, 1.90; HP238 plus IFN-γ, 1.48 versus NSC87877, 1.00; NSC87877 plus IFN-γ, 1.45; and NSC87877 plus HP238 plus IFN-γ, 1.46) (Fig. 3C) was suppressed in a SHP2-regulated manner. To further validate this finding, we next inhibited SHP2 genetically using a lentiviral-based shRNA approach (37). The H. pylori infection–mediated suppression of IFN-γ–induced STAT1 phosphorylation (Fig. 3D), IRF1 transactivation (Fig. 3E), and STAT1 expression (untreated, 1.00; IFN-γ, 2.47; HP238 plus IFN-γ, 1.48 versus shSHP2, 1.00; shSHP2 plus IFN-γ, 3.04; and shSHP2 plus HP238 plus IFN-γ, 3.12) (Fig. 3F) were also reversed in the absence of SHP2. Following SHP2 inhibition pharmacologically, and genetically, the basal level of STAT1 was decreased through an unknown mechanism. Furthermore, SHP2-mediated IFN-γ resistance at 1 h post–H. pylori infection was also demonstrated (data not shown). These findings clarify a novel strategy through which H. pylori infection causes IFN-γ resistance via SHP2 activation.

We next investigated the role of CagA, a virulence factor of H. pylori that directly induces SHP2 activation (14), in developing IFN-γ resistance. HP238 with a genetic isogenic mutation of cagA (HP238^CagAm) was used (31). Western blotting showed a lack of CagA expression in HP238^CagAm-infected gastric epithelial cells MKN45 (Fig. 4A) and GES-1 cells (Fig. 4B). The hummingbird effect by CagA-bearing H. pylori was confirmed by monitoring the morphological changes (Supplemental Fig. 2) (8, 15, 16). Compared with wild-type HP238, which caused defects in IFN-γ–activated STAT1 phosphorylation (Fig. 4A) and IRF1 transactivation (Fig. 4C), infection with HP238^CagAm failed to cause such effects, indicating that CagA determines IFN-γ resistance during H. pylori infection. The increased IRF1 transactivation in IFN-γ–treated HP238^CagAm-infected cells may be resulted from the effect of bacterial component LPS, which can facilitate IFN-γ signaling as describe previously (38). Western blot analysis also showed that HP238^CagAm infection caused less SHP2 phosphorylation both in MKN45 (Fig. 4A) and GES-1 cells (Fig. 4B), indicating the partial regulatory role of CagA in SHP2. However, the presence of phosphorylated SHP2 in HP238^CagAm-infected cells indicated an alternative model of CagA-independent SHP2 activation. Our findings illustrate that CagA has an essential role in IFN-γ resistance; however, its controversial role in SHP2 activation needs further investigation.

CagA is a membrane-associated protein that binds to SHP2 and then causes SHP2 activation without tyrosine phosphorylation at the C terminus (8). We unexpectedly found that HP238^CagAm infection failed to induce IFN-γ resistance but still caused tyrosine phosphorylation of SHP2. Therefore, CagA is hypothesized to be dispensable for SHP2 phosphorylation, but SHP2 may be required for IFN-γ resistance if CagA is concurrently present. To investigate these hypotheses, we next studied whether CagA protein was important for the cellular regulation of tyrosine-phosphorylated SHP2. Flow cytometric analysis first confirmed that both HP238 and HP238^CagAm caused both early (1 h postinfection) and long-term (6 h postinfection) SHP2 phosphorylation at tyrosine 542 (Fig. 5A). In HP238-infected cells, CagA coexpressed with tyrosine-phosphorylated SHP2. Immunostaining showed that CagA was colocalized with SHP2 (Fig. 5B) and tyrosine-phosphorylated SHP2 (Fig. 5C) at 1 h postinfection and significantly upregulated at 6 h postinfection with HP238, but not HP238^CagAm. The imaging results also showed that CagA/SHP2 and CagA/phosphorylated SHP2 are localized to the membrane. To further verify the location of CagA and its interaction with tyrosine-phosphorylated SHP2, membrane/cytosolic protein extraction analysis showed increased levels of tyrosine-phosphorylated SHP2 and CagA on the cell membrane of H. pylori–infected cells (Fig. 5D). Coimmunoprecipitation demonstrated both early (1 h postinfection) and extended (6 h postinfection) interactions of CagA with both SHP2 and phosphorylated SHP2 (Fig. 5E). These results indicate that membrane-associated CagA also interacts with phosphorylated SHP2.

In addition to the mechanism of CagA-mediated SHP2 activation directly, tyrosine phosphorylation at the C terminus of SHP2 may cause it to bind to the N-SH2 domain and promote activation (8); modulators other than CagA may regulate this phosphorylation step. Recent findings indicate that ROS can induce tyrosine phosphorylation of SHP2 at the C terminus through the formation of lipid rafts (39, 40). Notably, H. pylori infection also causes ROS generation (41). Both CagA- and ROS-activated SHP2 may concurrently affect IFN-γ–mediated defenses during H. pylori infection. We next used CM-H₂DCFDA to detect ROS generation in H. pylori–infected cell. The results showed that viable, but not HK, HP238, or HP238^CagAm induced ROS generation within 1 h, and this effect extended to 6 h (Fig. 6A). NAC, an ROS scavenger, was used to inhibit ROS. Notably, pretreatment with NAC reversed HP238- or HP238^CagAm-induced SHP2 phosphorylation (HP238, 1.85 versus HP238 plus NAC, 0.80; and HP238^CagAm, 1.45 versus HP238^CagAm plus NAC, 1.19), confirming that either CagA or ROS regulates SHP2 activation in H. pylori infection (Fig. 6B). The effect of NAC appears to be less inhibitory for CagA-deficient strains in SHP2 phosphorylation may be resulted from the increased ROS generation in HP238^CagAm-infected cells as showed in Fig. 6A. Western blotting showed that inhibiting ROS also decreased the H. pylori–induced suppression of IFN-γ–induced STAT1 phosphorylation (Fig. 6C) and STAT1 expression (Fig. 6D). These results demonstrate that, in addition to CagA, ROS is required for H. pylori–induced SHP2 phosphorylation and that ROS/SHP2-mediated IFN-γ resistance is only induced while CagA is concurrently expressed.

In this study, we sought to determine the mechanism of H. pylori–induced IFN-γ resistance. Several experiments were conducted in the context of H. pylori infection to: 1) study IFN-γ–activated STAT1/IRF1 signaling; 2) explore the role of SHP2 in IFN-γ resistance; and 3) verify the role of bacterial CagA and cellular oxidative stress for SHP2 phosphorylation and IFN-γ resistance. Our findings, as summarized in Supplemental Fig. 3, demonstrate a molecular mechanism through which H. pylori suppresses IFN-γ signaling. Under IFN-γ stimulation, JAK2/STAT1 signaling may mediate IRF1 expression to support antimicrobial activity. During H. pylori infection (CagA⁺), membrane-associated CagA binds to and directly activates SHP2. Additionally, ROS are generated in response to H. pylori infection independent of CagA. ROS may trigger an unknown signaling entity that facilitates tyrosine phosphorylation of SHP2. In the potential model of SHP2 activation as identified in this study, the concurrent CagA-mediated interaction between SHP2 at the membrane and ROS-facilitated SHP2 phosphorylation effectively causes IFN-γ resistance. These results not only explore the interregulation of SHP2-binding protein and ROS on SHP2 activation, as not previously discussed (8), but also make it possible to understand the pathological role in IFN-γ resistance during H. pylori infection.

The pathogenesis of H. pylori infection is complex, and additional pathogenic factors, host factors, and their intramolecular cross talk continue to be discovered (18, 42–47). Although H. pylori infection causes chronic gastric inflammation, the immune escape of H. pylori infection, particularly under high levels of IFN-γ, has been less commonly reported. Gastric epithelial cells are major cells in inflamed sites of H. pylori infection. Although the results showed in this study that H. pylori infection causes IFN-γ resistance in tested gastric cancer cells MKN45 and AZ521 as well as in monocytic U937 cells, such effects of immune evasion caused by H. pylori need to validate in human primary gastric epithelial cells and/or immune cells infected with H. pylori. In addition to inflammatory responses, CagA-mediated cell transformation may prompt gastric tumorigenesis. Regarding the antimicrobial and anticancer properties of IFN-γ, this study was aimed to investigate the effect of H. pylori infection on IFN-γ signal transduction, especially in STAT1/IRF1 signaling, which is key for IFN-γ responses. The dual roles of IFN-γ are important for H. pylori-caused gastric inflammation and bacterial clearance (12, 22–26). Consistent with these findings, Sayi et al. (48) further demonstrated the tumorigenic risk caused by IFN-γ, although IFN-γ also triggers clearance during H. pylori infection. For antimicrobial activity, in addition to activating macrophages for phagocytosis, CagA protein is inhibited by IFN-γ through an as-yet-unknown mechanism, as reported previously (28). To demonstrate the antimicrobial effect of IFN-γ, we confirmed that pretreatment with IFN-γ decreased CagA expression, as detected by Western blot analysis (Supplemental Fig. 4). Additionally, H. pylori infection may escape from phagocytosis by an active method that causes macrophages to undergo apoptosis (11, 12, 18, 49). These findings suggest that H. pylori develops strategies to escape from the immune defenses of IFN-γ and macrophages, and the presence of high levels of IFN-γ may pathologically affect gastric inflammation.

To maintain IFN-γ homeostasis, SOCS1 and SOCS3 can be induced through STAT-mediated transcriptional induction (4, 5); in contrast, SHP2 is later activated by IFN-γ stimulation through an unknown mechanism. Thus, hijacking these negative regulators may be a strategy for pathogen-derived IFN-γ resistance (9). For example, infection by human CMV and Leishmania impairs IFN-γ signaling by activating the tyrosine phosphatases SHP2 and SHP1, respectively (50, 51). The possible mechanisms for pathogen-induced SHP2 activation remain unclear. In H. pylori–infected gastric epithelial cells, we found that SHP2 is activated, and this is accompanied by the presence of tyrosine phosphorylation, whereas there was no change in the expression of SOCS1 or SOCS3. Genetic and pharmacological inhibition of SHP2 showed the essential role of SHP2 in H. pylori–induced IFN-γ resistance. The regulation of SHP2 activation is therefore of interest to study its effects on IFN-γ resistance. The membrane-associated protein CagA binds to SHP2 and then causes SHP2 activation, and these steps occur without tyrosine phosphorylation at the C terminus (8). Unexpectedly, we found that HP238^CagAm infection failed to induce IFN-γ resistance, indicating that IFN-γ resistance depends on CagA, but still partly caused tyrosine phosphorylation of SHP2. The finding that SHP2 is still partly activated by CagA mutant H. pylori is inconsistent with some previous studies (14); however, they used CagA overexpression in gastric epithelial cells to study the effects of CagA on SHP2 activation, and the results clearly showed CagA directly binds to SH2 domains and then causes SHP2 conformation change and activation. Of course, mutated CagA did not activate SHP2 in their model. However, no bacterial strains with or without CagA mutation were investigated in their work. We agreed to the findings that CagA is important for SHP2 activation; however, the activation and modification of SHP2 in vivo are also multifaceted. Under H. pylori infection, a variety of bacterial effects may affect on cellular activation of SHP2. According to our findings in this study, we found a natural infection of H. pylori–induced ROS generation that synergizes with CagA to determine tyrosine phosphorylation of SHP2. In other words, the natural infection used in this study differs from the approach of CagA overexpression. H. pylori–induced CagA and SHP2 may not only act as oncogenic factors, as reported previously (8, 15, 16), but also promote IFN-γ resistance, as first demonstrated in this study.

Genetic studies have demonstrated the essential role of IFN-γ–mediated bacterial clearance in vivo. However, H. pylori may develop IFN-γ resistance as a survival strategy. An earlier study showed that, under post–H. pylori infection, there are defects in IFN-γ/TNF-α–induced IFN-induced protein 10 and monokines induced by IFN-γ expression (52). Mitchell et al. (30) provide the first evidence showing IFN-γ resistance caused by H. pylori infection. However, the molecular mechanism underlying this resistance has not been documented. Thus far, this resistance has been found to be independent of the effects of several genes, including CagA, CagE, and vacA. Our findings confirm that IFN-γ resistance is caused by H. pylori. However, inconsistent with the study by Mitchell et al. (30), which used clinical isolates with CagA mutations, we further showed CagA-dependent IFN-γ resistance specially using an isogenic mutant strain. The different strains of H. pylori may cause the different responses to IFN-γ. We suggest that CagA-mediated IFN-γ resistance occurs only within CagA-positive cells because CagA mediates SHP2 activation (8). Inconsistent with the previous study that H. pylori infection primes epithelial cells to facilitate IFN-γ–induced inflammation (53), the different conclusions from these studies may result from the differing MOI used in these studies. We showed that, at a higher MOI (MOI of 100), cells are resistant to IFN-γ signaling. However, at a lower MOI (MOI of 10), the pathogen–cell contact and/or the soluble mediators secreted from infected cells may prime the activation of IFN-γ in CagA-negative cells (53). Our preliminary data showed there was no inhibition on IFN-γ–activated STAT1 in H. pylori–infected MKN45 cells under the lower MOI (data not shown). In addition to causing severe gastric inflammation, IFN-γ may promote bacterial internalization, as demonstrated in Escherichia coli–infected gut epithelia (54). The differences in these models need further validation.

SHP2 is a well-known negative regulator of IFN-γ/STAT1 signaling (6, 7). This study demonstrates that SHP2 is alternatively regulated by CagA and ROS and that SHP2 has a role in IFN-γ resistance caused by H. pylori infection. Our findings, as analyzed by imaging and biochemical approaches, further showed that membrane-associated CagA interacts with tyrosine-phosphorylated SHP2 in H. pylori infection. In the alternative model of SHP2 activation, CagA and ROS concurrently facilitate SHP2 activation (Supplemental Fig. 3). The possible regulatory effects of ROS on SHP2 modification need further investigation. However, ROS regulation of SHP2 is controversial. Supporting a positive role for ROS on SHP2 activation, evidence shows that oxidative stress induces tyrosine phosphorylation of SHP2 at the C terminus through a mechanism involving the formation of lipid rafts and caveolin1 expression (39, 40). Similar results have been demonstrated in response to exogenous hydrogen peroxide treatment (55) and the inhibition of lysyl oxidase (56).

Notably, infection of H. pylori causes ROS generation through heat shock protein 90/Rac1–regulated NADPH oxidase activation (41), whereas LPS of H. pylori enhances the expression of NADPH oxidase components, including gp91, p22, and p67 (57). However, HK H. pylori did not induce ROS generation as well as SHP2 phosphorylation, indicating an activation mechanism for NADPH oxidase is occurred only in response to viable H. pylori infection. A current study showed that VacA causes GSH decrease followed by ROS generation to induce autophagy-mediated CagA degradation (58). It is still unclear why does HP238^CagAm infection cause higher level of ROS. Without CagA, it may be hypothesized that the ROS-promoting effect of VacA can be enhanced through a physiological feedback response in HP238^CagAm-infected cells. Based on these findings, HK H. pylori failed to express VacA for ROS generation. Essentially, ROS are involved in H. pylori–induced gastric inflammation and gastric carcinogenesis (59, 60). The molecular mechanism through which ROS regulates tyrosine phosphorylation of SHP2 remains unclear. The Src family protein tyrosine kinase Fyn may act upstream of SHP2 phosphorylation at the C-terminal tyrosine residues (61), and ROS fusion tyrosine kinase may activate SHP2, causing subsequent tyrosine phosphorylation of SHP2 (62). These studies support the possibility that ROS may indirectly facilitate SHP2 phosphorylation (8). However, in contrast to the positive role of ROS on SHP2, SHP2 might also be transiently and directly inactivated by ROS under signaling of growth factors and the TCR (63, 64). ROS-mediated SHP2 activation needs further investigation.

The possible regulation on SHP2 is also suggested in multifaceted manners. Studies in our laboratory have shown that glycogen synthase kinase (GSK)-3β has a suppressive role on SHP2 phosphorylation, while pharmacologically inhibiting GSK-3β facilitated SHP2 activation followed by SHP2-mediated IFN-γ resistance (36, 65). Notably, in H. pylori infection, PI3K/Akt signaling may inactivate GSK-3β followed by β-catenin stabilization to promote gastric tumorigenesis (66, 67). We suggest that inactivating GSK-3β by H. pylori infection may also cause SHP2 activation, followed by IFN-γ resistance.

In conclusion, based on the findings of this study, both CagA- and ROS-regulated SHP2 activation may simultaneously facilitate IFN-γ resistance during H. pylori infection. The results confirm a negative regulatory role for SHP2 in IFN-γ signal transduction and suggest an immune escape strategy for H. pylori infection that involves interference with IFN-γ signaling in a CagA/SHP2-mediated manner. IFN-γ resistance may cause persistent infection by H. pylori in gastric epithelial cells, followed by gastric tumorigenesis, which involves bacterial virulence factors and IFN-γ–prompted gastric inflammation. The primary cells associated with H. pylori infection in gastric tissues, a suitable animal model, the clinical relevance, and the various pathogenic strains of H. pylori are needed to test this possibility, as we have demonstrated in vitro in this study. Our study not only explores an alternative model of SHP2 activation but also increases our understanding of the molecular mechanism of IFN-γ resistance caused by H. pylori infection.

We thank the Immunobiology Core, Research Center of Clinical Medicine, National Cheng Kung University Hospital, for providing training, technical support, and assistance with experimental design and data analysis using the Flow Cytometry Core facilities.

The authors have no financial conflicts of interest.