Advertisement

Helicobacter pylori Depletes Cholesterol in Gastric Glands to Prevent Interferon Gamma Signaling and Escape the Inflammatory Response

Open AccessPublished:December 19, 2017DOI:https://doi.org/10.1053/j.gastro.2017.12.008

      Background & Aims

      Despite inducing an inflammatory response, Helicobacter pylori can persist in the gastric mucosa for decades. H pylori expression of cholesterol-α-glucosyltransferase (encoded by cgt) is required for gastric colonization and T-cell activation. We investigated how cgt affects gastric epithelial cells and the host immune response.

      Methods

      MKN45 gastric epithelial cells, AGS cells, and human primary gastric epithelial cells (obtained from patients undergoing gastrectomy or sleeve resection or gastric antral organoids) were incubated with interferon gamma (IFNG) or interferon beta (IFNB) and exposed to H pylori, including cagPAI and cgt mutant strains. Some cells were incubated with methyl-β-cyclodextrin (to deplete cholesterol from membranes) or myriocin and zaragozic acid to prevent biosynthesis of sphingolipids and cholesterol and analyzed by immunoblot, immunofluorescence, and reverse transcription quantitative polymerase chain reaction analyses. We compared gene expression patterns among primary human gastric cells, uninfected or infected with H pylori P12 wt or P12Δcgt, using microarray analysis. Mice with disruption of the IFNG receptor 1 (Ifngr1–/– mice) and C57BL6 (control) mice were infected with PMSS1 (wild-type) or PMSS1Δcgt H pylori; gastric tissues were collected and analyzed by reverse transcription quantitative polymerase chain reaction or confocal microscopy.

      Results

      In primary gastric cells and cell lines, infection with H pylori, but not cgt mutants, blocked IFNG-induced signaling via JAK and STAT. Cells infected with H pylori were depleted of cholesterol, which reduced IFNG signaling by disrupting lipid rafts, leading to reduced phosphorylation (activation) of JAK and STAT1. H pylori infection of cells also blocked signaling by IFNB, interleukin 6 (IL6), and IL22 and reduced activation of genes regulated by these signaling pathways, including cytokines that regulate T-cell function (MIG and IP10) and anti-microbial peptides such as human β-defensin 3 (hBD3). We found that this mechanism allows H pylori to persist in proximity to infected cells while inducing inflammation only in the neighboring, non-infected epithelium. Stomach tissues from mice infected with PMSS1 had increased levels of IFNG, but did not express higher levels of interferon-response genes. Expression of the IFNG-response gene IRF1 was substantially higher in PMSS1Δcgt-infected mice than PMSS1-infected mice. Ifngr1–/– mice were colonized by PMSS1 to a greater extent than control mice.

      Conclusions

      H pylori expression of cgt reduces cholesterol levels in infected gastric epithelial cells and thereby blocks IFNG signaling, allowing the bacteria to escape the host inflammatory response. These findings provide insight into the mechanisms by which H pylori might promote gastric carcinogenesis (persisting despite constant inflammation) and ineffectiveness of T-cell–based vaccines against H pylori.

      Graphical abstract

      Keywords

      Abbreviations used in this paper:

      αCAG (cholesteryl-6’-O-aycl-α-glucoside), αCPG (cholesteryl-6’-O-phosphatidyl-α-glucoside), ALI (air-liquid interface), CagA (cytotoxin-associated gene A), hBD3 (human β-defensin 3), IFN (interferon), IFNG (interferon gamma), IL (interleukin), IRF (interferon regulatory factors), JAK (Janus Kinase), mβCD (methyl-β-cyclodextrin), MAPK (mitogen-activated protein kinase), MOI (multiplicity of infection), NF-κB (nuclear factor κB), PEG (polyethylene glycol), RT-qPCR (reverse transcription quantitative polymerase chain reaction), STAT (signaling transducer and activator of transcription)
      See Covering the Cover synopsis on page 1216.

       Background and Context

      For reasons not well-understood, Helicobacter pylori can persist life-long in the stomach despite causing a strong inflammatory response, increasing the risk for serious sequelae such as ulcers and cancer.

       New Findings

      By extracting cholesterol from host cells, H pylori blocks the assembly of IFN and other cytokine receptors, rendering the infected mucosa unable to respond properly to inflammatory signals from stroma and immune cells.

       Limitations

      Bacterial mutants for the cgt gene, which is required for cholesterol extraction, are unable to establish an in vivo infection, preventing a detailed analysis of its effects in the animal model.

       Impact

      This study provides important insight into the pitfalls of past vaccination approaches for H pylori. It also explains how cholesterol rich diet could negatively influence the inflammatory condition of the infected stomach mucosa.
      About half the world’s population is chronically infected with the gram-negative bacterium Helicobacter pylori, which is implicated in severe gastric disease, including peptic ulcer and adenocarcinoma.
      • Bauer B.
      • Meyer T.F.
      The human gastric pathogen helicobacter pylori and its association with gastric cancer and ulcer disease.
      Colonization takes place in the gastric mucus and eventually involves bacterial adherence to the glandular epithelium.
      • Sigal M.
      • Rothenberg M.E.
      • Logan C.Y.
      • et al.
      Helicobacter pylori activates and expands Lgr5(+) stem cells through direct colonization of the gastric glands.
      The infection is characterized by a rapid and strong nuclear factor κB (NF-κB)–driven response,
      • Zimmermann S.
      • Pfannkuch L.
      • Al-Zeer M.A.
      • Bartfeld S.
      • et al.
      ALPK1- and TIFA-dependent innate immune response triggered by the helicobacter pylori type IV secretion system.
      which leads to the recruitment and activation of immune cells: CD4+ T cells, dominated by the Th1 lineage
      • Bamford K.B.
      • Fan X.
      • Crowe S.E.
      • et al.
      Lymphocytes in the human gastric mucosa during Helicobacter pylori have a T helper cell 1 phenotype.
      • Salama N.R.
      • Hartung M.L.
      • Muller A.
      Life in the human stomach: persistence strategies of the bacterial pathogen Helicobacter pylori.
      play a decisive role in controlling the Helicobacter load via secretion of interferon gamma (IFNG).
      • Sayi A.
      • Kohler E.
      • Hitzler I.
      • et al.
      The CD4+ T cell-mediated IFN-gamma response to Helicobacter infection is essential for clearance and determines gastric cancer risk.
      This Th1 response is also fueled by type I IFNs (α/β) released from gastric epithelial cells themselves.
      • Watanabe T.
      • Asano N.
      • Fichtner-Feigl S.
      • et al.
      NOD1 contributes to mouse host defense against Helicobacter pylori via induction of type I IFN and activation of the ISGF3 signaling pathway.
      Th17 and Th22 T cells, characterized by interleukin (IL)17 and IL22 production, also promote the inflammatory milieu, helping to control infection.
      • Horvath Jr., D.J.
      • Washington M.K.
      • Cope V.A.
      • et al.
      IL-23 Contributes to control of chronic Helicobacter pylori infection and the development of T helper responses in a mouse model.
      • Zhuang Y.
      • Cheng P.
      • Liu X.F.
      • et al.
      A pro-inflammatory role for Th22 cells in Helicobacter pylori-associated gastritis.
      Yet, H pylori is able to escape full elimination by host immunity through an unknown mechanism, resulting in a severe chronic inflammatory condition that represents a crucial aspect of its pathogenesis.
      Gastric epithelial cells display receptors for type I (α/β) and type II (γ) interferons, the subunits of which (IFNAR1/IFNAR2 and IFNGR1/IFNGR2, respectively) are assembled in specialized cholesterol-rich membrane microdomains, known as lipid rafts.
      • Takaoka A.
      • Mitani Y.
      • Suemori H.
      • et al.
      Cross talk between interferon-gamma and -alpha/beta signaling components in caveolar membrane domains.
      Microdomain-dependent receptor activation
      • Blouin C.M.
      • Hamon Y.
      • Gonnord P.
      • et al.
      Glycosylation-dependent IFN-gammaR partitioning in lipid and actin nanodomains is critical for JAK activation.
      triggers JAK (Janus Kinase) and STAT (signaling transducer and activator of transcription) signaling via STAT1/2 phosphorylation and nuclear translocation to promote downstream expression of genes involved in inflammation and defense, including interferon regulatory factors (IRF), which further amplify the IFN response through positive feedback via STATs.
      • Ivashkiv L.B.
      • Donlin L.T.
      Regulation of type I interferon responses.
      Gastric epithelial cells also express receptors for IL22
      • Zhuang Y.
      • Cheng P.
      • Liu X.F.
      • et al.
      A pro-inflammatory role for Th22 cells in Helicobacter pylori-associated gastritis.
      , which can stimulate the production of antimicrobial factors that defend against mucosal pathogens.
      • Zheng Y.
      • Valdez P.A.
      • Danilenko D.M.
      • et al.
      Interleukin-22 mediates early host defense against attaching and effacing bacterial pathogens.
      The effector mechanisms controlling colonization by H pylori remain sparsely understood. Epithelial cells can produce antimicrobial peptides (AMP), such as hBD3 (human β-defensin 3), which effectively kills H pylori
      • Bauer B.
      • Pang E.
      • Holland C.
      • et al.
      The Helicobacter pylori virulence effector CagA abrogates human beta-defensin 3 expression via inactivation of EGFR signaling.
      • Nuding S.
      • Gersemann M.
      • Hosaka Y.
      • et al.
      Gastric antimicrobial peptides fail to eradicate Helicobacter pylori infection due to selective induction and resistance.
      and is induced through MAPK (mitogen-activated protein kinase) and JAK/STAT signaling and stimulation with IFNG or IL22.
      • Bauer B.
      • Pang E.
      • Holland C.
      • et al.
      The Helicobacter pylori virulence effector CagA abrogates human beta-defensin 3 expression via inactivation of EGFR signaling.
      • Fahlgren A.
      • Hammarstrom S.
      • Danielsson A.
      • et al.
      beta-Defensin-3 and -4 in intestinal epithelial cells display increased mRNA expression in ulcerative colitis.
      • Joly S.
      • Organ C.C.
      • Johnson G.K.
      • et al.
      Correlation between beta-defensin expression and induction profiles in gingival keratinocytes.
      • Dixon B.R.
      • Radin J.N.
      • Piazuelo M.B.
      • et al.
      IL-17a and IL-22 induce expression of antimicrobials in gastrointestinal epithelial cells and may contribute to epithelial cell defense against Helicobacter pylori.
      Strikingly, despite continued inflammation, hBD3 is not detected in infected human gastric biopsies.
      • Nuding S.
      • Gersemann M.
      • Hosaka Y.
      • et al.
      Gastric antimicrobial peptides fail to eradicate Helicobacter pylori infection due to selective induction and resistance.
      • Bauer B.
      • Wex T.
      • Kuester D.
      • et al.
      Differential expression of human beta defensin 2 and 3 in gastric mucosa of Helicobacter pylori-infected individuals.
      We have previously shown that H pylori prevents hBD3 expression through a CagA (cytotoxin-associated gene A)-dependent mechanism.
      • Bauer B.
      • Pang E.
      • Holland C.
      • et al.
      The Helicobacter pylori virulence effector CagA abrogates human beta-defensin 3 expression via inactivation of EGFR signaling.
      However, how hBD3 remains blocked despite IFNG and IL22 stimulation is unclear.
      H pylori, which is auxotroph for cholesterol, extracts the lipid from host membranes to incorporate it into its outer membrane as an α-glucosylated derivative, using the enzyme Cgt encoded by the gene HP0421 (cgt).
      • Lebrun A.H.
      • Wunder C.
      • Hildebrand J.
      • et al.
      Cloning of a cholesterol-alpha-glucosyltransferase from Helicobacter pylori.
      • Wang H.J.
      • Cheng W.C.
      • Cheng H.H.
      • et al.
      Helicobacter pylori cholesteryl glucosides interfere with host membrane phase and affect type IV secretion system function during infection in AGS cells.
      • Wunder C.
      • Churin Y.
      • Winau F.
      • et al.
      Cholesterol glucosylation promotes immune evasion by Helicobacter pylori.
      Cgt is the first in a series of H pylori enzymes that cause additional modifications to generate cholesteryl-6’-O-aycl-α-glucoside (αCAG) or cholesteryl-6’-O-phosphatidyl-α-glucoside (αCPG).
      • Dixon B.R.
      • Radin J.N.
      • Piazuelo M.B.
      • et al.
      IL-17a and IL-22 induce expression of antimicrobials in gastrointestinal epithelial cells and may contribute to epithelial cell defense against Helicobacter pylori.
      Cholesterol glucosylation and extraction from host cells result in lipid raft destruction and/or alteration of the membrane architecture,
      • Wang H.J.
      • Cheng W.C.
      • Cheng H.H.
      • et al.
      Helicobacter pylori cholesteryl glucosides interfere with host membrane phase and affect type IV secretion system function during infection in AGS cells.
      • Wunder C.
      • Churin Y.
      • Winau F.
      • et al.
      Cholesterol glucosylation promotes immune evasion by Helicobacter pylori.
      which has been linked to immune evasion and bacterial persistence.
      • Wunder C.
      • Churin Y.
      • Winau F.
      • et al.
      Cholesterol glucosylation promotes immune evasion by Helicobacter pylori.
      • Beigier-Bompadre M.
      • Moos V.
      • Belogolova E.
      • et al.
      Modulation of the CD4+ T-cell response by Helicobacter pylori depends on known virulence factors and bacterial cholesterol and cholesterol alpha-glucoside content.
      • Du S.Y.
      • Wang H.J.
      • Cheng H.H.
      • et al.
      Cholesterol glucosylation by Helicobacter pylori delays internalization and arrests phagosome maturation in macrophages.
      We previously reported that a cholesterol-rich diet leads to a reduction of the H pylori load concomitant with an increased Th1 response.
      • Wunder C.
      • Churin Y.
      • Winau F.
      • et al.
      Cholesterol glucosylation promotes immune evasion by Helicobacter pylori.
      Here, we provide insight into the underlying mechanism by showing that the IFN response is subverted by H pylori in gastric epithelial cells. This is caused by Cgt-dependent cholesterol depletion, resulting in the destruction of lipid rafts, failure of IFN receptor subunit assembly and, ultimately, lack of downstream signaling. Similarly, Cgt blocks IL6 and IL22 signaling. Our data provide evidence for the highly effective destruction of the responsiveness of H pylori-infected epithelium, even in the presence of strong cytokine signals from the adjacent micro-environment, thus impairing mucosal defense.

      Materials and Methods

       Ethical Permissions

      Human gastric tissue specimens were obtained from individuals undergoing gastrectomy or sleeve resection, under the ethics approval by the Charité Ethics Committee (EA1/058/11 and EA1/129/12). Animal experiments were performed in mice maintained under pathogen-free conditions based on approval by the Ethics Committee for Animal Experimentation of the State of Berlin (G0205/12).

       Cell Culture Infection and Treatment

      Bacteria were collected from plates and resuspended in RPMI 1640 (serum-free). All cells were grown to 60% confluency, washed twice with phosphate-buffered saline, and serum starved with RPMI 1640 for 16 hours prior to experiments. Infection was carried out under serum-starved conditions at 37 °C, 5% CO2 for the indicated times at multiplicity of infection (MOI) 20 or 50. Treatment with methyl-β-cyclodextrin (mβCD; Sigma, St. Louis, MO) was carried out for 5 hours at indicated concentrations. Biosynthesis of sphingolipids and cholesterol was inhibited by treating cells with myriocin (50 μmol/L; Sigma) for 72 hours and zaragozic acid (50 μmol/L, Cayman Chemical: Ann Arbor, MI) for the last 18 hours. In selected experiments, bacteria were treated with water-soluble cholesterol (Sigma) at 1 mg/mL for 1 hour prior to infection. Alternatively, polyethylene glycol (PEG)-cholesterol (10 mg/mL; Sigma) was added to cultures during the last hour of infection, while mock-infected cells were treated with an equal volume containing cholesterol. After infection, recombinant human IFNG (R&D Systems: Minneapolis, MN; 10 ng/mL), IFNB (PBL Assay Science, Piscataway, NJ; 2300 U/mL), IL6 (Peprotech, Rocky Hill, NJ; 25 ng/mL) or IL22 (Peprotech; 50 ng/mL) were added to selected wells and maintained for indicated times until the end of the experiment.

       Primary Gastric Epithelial Air-Liquid Interface Cultures

      Cells from organoids or freshly isolated glands were seeded in trans-well inserts (Millipore (Merck): Darmstadt, GE; PIHP01250) in 24-well plates and wells filled with 400 μL primary cell culture medium. Once cells had formed a confluent monolayer, medium on top of the cells was removed to start air-liquid interface (ALI) culture. Cultures were kept at 37°C, 5% CO2 in a humidified incubator. Ten days later cells were infected by placing 50 μL of bacteria in phosphate-buffered saline (MOI 100) on top of the filters for 3 days.
      For further information see Supplementary Materials and Methods. All RT-qPCR primers listed in Supplementary Table 1.

       Data Deposition

      Microarray data have been deposited in the Gene Expression Omnibus (GEO; www.ncbi.nlm.nih.gov/geo/) of the National Center for Biotechnology Information under GEO accession number GSE76589.

      Results

       H pylori Blocks JAK/STAT Signaling upon IFNG Treatment

      To assess the influence of H pylori on the response to IFNG, we infected MKN45 gastric epithelial cells with strains P12 and P1 for 6 or 24 hours, followed by treatment with IFNG for 30 minutes. Immunoblotting revealed STAT1 phosphorylation in response to IFNG treatment in non-infected cells, cells infected with heat-killed bacteria, and cells infected for 6 hours (Figure 1A). Surprisingly, after prolonged infection (24 hours) IFNG stimulation failed to activate STAT1, irrespective of the H pylori strain (Figure 1A). To analyze the dynamics of this phenotype, we performed a time course experiment, showing that STAT1 phosphorylation was diminished after 16 hours and completely blocked from 24 until 96 hours of infection with P12 (Supplementary Figure 1A). To investigate the underlying mechanisms, we thus chose the 24 hour time point of infection as a reference. In AGS cells, too, 24 hours of infection inhibited the response to IFNG (Supplementary Figure 1B). MOI 10 was sufficient to partially block IFNG signaling in MKN45 cells within 24 hours and a complete block was observed with MOI 50 (Supplementary Figure 1C), which was therefore chosen for further experiments.
      Figure thumbnail gr1
      Figure 1Prolonged infection prevents IFNG-induced activation of JAK1, JAK2, and STAT1 through Cgt activity. MKN45 cells were infected with the indicated strains (HK, heat killed) at MOI 50 and treated (+) or mock-treated (-) with IFNG (10 ng/mL, 30 minutes) at the indicated times post-infection (p.i.). (A and B) Immunoblot analysis of total protein level and phosphorylation of JAK1 or STAT1, respectively. (C) Analysis of JAK2 phosphorylation (red) by confocal IF, maximum projection. Cells were non-infected (NI) or infected with indicated H pylori strains (green) and treated with IFNG (10 ng/mL, 30 minutes). Scale bar = 5 μm. (D and E) Immunoblot analysis JAK1 and STAT1 phosphorylation and total protein levels of MKN45 cells infected or mock-infected (NI) for 24 hours at MOI 50 with the indicated H pylori strains and treatment with IFNG (10 ng/mL, 30 minutes).
      Upon IFNG stimulation, an activated IFNGR complex phosphorylates JAK1 and JAK2 kinases, which in turn phosphorylate STAT1. After 24 hours of infection with P12 or P1, neither JAK1 nor JAK2 was phosphorylated any more in MKN45 cells (Figure 1B and 1C, middle panel; Supplementary Figure 1D). Notably, JAK1 (Figure 1B) and JAK2 (data not shown) were activated in non-infected MKN45 cells even without IFNG stimulation; however, this was also diminished after infection. Infection conditions did not compromise cell viability (Supplementary Figure 1E).

       Cgt is Required for Inactivation of IFNG-JAK/STAT1 Pathway

      To identify the bacterial factor involved in the block of IFNG signaling, we infected MKN45 cells with H pylori wild type and mutant strains. As a recent report linked CagA translocation to STAT1 dephosphorylation via SHP-2 activation,
      • Wang Y.C.
      • Chen C.L.
      • Sheu B.S.
      • et al.
      Helicobacter pylori infection activates Src homology-2 domain-containing phosphatase 2 to suppress IFN-gamma signaling.
      we tested deletion mutants for cagPAI, which encodes the entire type IV secretion system, cagA and cgt. While the cagA and cagPAI mutant strains still inhibited JAK/STAT1 signaling upon IFNG stimulation, the cgt mutant did not (Figure 1D and Supplementary Figure 1B). Immunofluorescence analysis showed that P12Δcgt (Figure 1C and Supplementary Figure 1D) did not block JAK2 activation upon IFNG treatment, despite adhering to epithelial cells at levels comparable to wild type. Similarly, STAT1 signaling upon IFNG treatment was still inhibited in a Cgt-dependent manner at 48 h.p.i. (Supplementary Figure 1F). Although genetic complementation of the cgt-mutant strain only partially reconstituted Cgt expression (Supplementary Figure 1G), it was enough to restore the ability to impair JAK/STAT1 activation (Figure 1E) in MKN45 cells. Finally, we evaluated the consequences of disrupted JAK/STAT1 signaling by analyzing expression of downstream genes after IFN treatment, by infecting MKN45 cells for 24 hours with P12 wild type or P12Δcgt prior to treatment with IFNG for 2.5 or 5 hours, followed by reverse transcription–quantitative polymerase chain reaction (RT-qPCR). Expression of IRF1, as well as the T-cell attractant chemokines MIG and IFNG-inducible protein 10 (IP)-10 (encoded by CXCL9 and CXCL10 genes, respectively), was strongly up-regulated in non-infected cells upon IFNG treatment (Supplementary Figure 1H). Cells infected with P12Δcgt also responded to IFNG, but after infection with wt P12 the response was significantly reduced. Accordingly, Cgt activity blocks the JAK/STAT1 pathway, as well as transcription of downstream genes involved in amplification of the IFNG response.

       Inhibition of IFNG Response is Linked to Host Cholesterol Depletion, Lipid Raft Disruption, and IFNGR Assembly

      Cholesterol acquisition by H pylori has been linked to the destruction of lipid rafts in epithelial cells.
      • Wunder C.
      • Churin Y.
      • Winau F.
      • et al.
      Cholesterol glucosylation promotes immune evasion by Helicobacter pylori.
      Lipid rafts serve as platforms for cell signaling cascades, including IFN signaling.
      • Arish M.
      • Husein A.
      • Kashif M.
      • et al.
      Orchestration of membrane receptor signaling by membrane lipids.
      As previously described,
      • Wunder C.
      • Churin Y.
      • Winau F.
      • et al.
      Cholesterol glucosylation promotes immune evasion by Helicobacter pylori.
      we observed cholesterol depletion in cells infected with wild-type H pylori (Supplementary Figure 2A). To connect this to the block of the IFNG response, we treated cells with mβCD, which depletes cholesterol from eukaryotic membranes. Treatment did not affect cellular viability (Supplementary Figure 2B) but impaired both JAK1 and STAT1 phosphorylation (Figure 2A). We also inhibited the biosynthesis of sphingolipids and cholesterol by combined treatment with myriocin and zaragozic acid to disrupt lipid raft function.
      • Blouin C.M.
      • Hamon Y.
      • Gonnord P.
      • et al.
      Glycosylation-dependent IFN-gammaR partitioning in lipid and actin nanodomains is critical for JAK activation.
      Myriocin and zaragozic acid treatment indeed significantly blocked the induction of IRF1 expression in response to IFNG (Supplementary Figure 2C). In addition, when we coated H pylori with water-soluble cholesterol before infection to abolish cholesterol transfer from host cells, they did not block the cellular response to IFNG treatment (Figure 2B). Cholesterol coating also restored the capacity of wt H pylori-infected cells to up-regulate IRF1 in response to IFNG, at levels comparable to non-infected cells (Supplementary Figure 2D). Cholesterol coating did not affect bacterial viability (Supplementary Figure 2E). To control for any effects on the initial host-pathogen interplay (eg, by masking bacterial adhesins), we repeated the experiment by adding PEG-cholesterol to the medium during the last hour of infection. This partially rescued the capacity of infected cells to respond to IFNG treatment, further highlighting the importance of cholesterol as a mediator of inflammation in response to H pylori (Figure 2C).
      Figure thumbnail gr2
      Figure 2Cholesterol depletion by H pylori blocks IFNG signaling by disrupting lipid rafts, impairing IFNGR complex assembly. (A) Immunoblot analysis of JAK1/STAT1 phosphorylation after 5 hours mβCD treatment at indicated concentrations or in non-treated cells (-), with or without IFNG treatment (10 ng/mL, 30 minutes). (B) Immunoblot analysis of JAK1 and STAT1 phosphorylation upon IFNG treatment (+). MKN45 cells were infected for 24 hours with the indicated strains or non-infected, then treated with IFNG (+) (10 ng/mL, 30 minutes). Cholesterol-coating of bacteria was started 1 hour before infection and applied to non-infected cells accordingly. (C) Immunoblot analysis of STAT1 phosphorylation upon IFNG and/or PEG-cholesterol treatment. MKN45 cells were infected for 24 hours with H pylori wt or non-infected. In selected wells (+), PEG-cholesterol (5 mg/mL) was added during the final hour of infection. Then, cells were treated with IFNG (10 ng/mL, 30 minutes) (+) or mock-treated (-). Quantification: band intensity ratio STAT1pY:actin. (D) Confocal microscopy analysis (x-y sections with lateral views along dotted lines: x-z, bottom; y-z right) of the lipid raft marker GM1 (green) in membranes of cells non-infected or infected for 24 hours with indicated strains. Nuclei stained with DAPI (blue). Scale bar: 5 μm. (E) Immunoblot analysis of detergent-resistant membranes for IFNGR1, IFNGR2, caveolin, or dot blot analysis with CTxB-HRP. Cells were non-infected or infected for 24 hours with indicated strains, followed by treatment (+) or mock-treatment (-) with IFNG (10 ng/mL, 30 minutes). (F and G) Immunoprecipitation (IP) and immunoblot analysis of IFNGR subunits 1 and 2. IFNGR2 was pulled down using IFNGR2 antibody. MKN45 cells were infected with H pylori P12 wt or coated with water-soluble cholesterol before infection. Results representative of 2 independent experiments.
      The IFNGR subunits 1 and 2 need to merge at cholesterol-rich micro-domains
      • Blouin C.M.
      • Hamon Y.
      • Gonnord P.
      • et al.
      Glycosylation-dependent IFN-gammaR partitioning in lipid and actin nanodomains is critical for JAK activation.
      to respond to IFNG stimulation and initiate downstream signaling via JAK1 and JAK2.
      • Takaoka A.
      • Mitani Y.
      • Suemori H.
      • et al.
      Cross talk between interferon-gamma and -alpha/beta signaling components in caveolar membrane domains.
      • Sen S.
      • Roy K.
      • Mukherjee S.
      • et al.
      Restoration of IFNgammaR subunit assembly, IFNgamma signaling and parasite clearance in Leishmania donovani infected macrophages: role of membrane cholesterol.
      We investigated the impact of H pylori on the integrity of lipid rafts by assessing potential alterations in the assembly of functional IFNGR. According to immunofluorescence analysis, surface accumulation of glycosphingolipid ganglioside GM1, a constituent marker of lipid rafts, is lost after infection with P12 wt and P12ΔcagA (Figure 2D, quantification of the relative membrane GM1 signal in Supplementary Figure 2F). In contrast, infection with P12Δcgt did not notably alter GM1 distribution (Figure 2D and Supplementary Figure 2F). In addition, we performed a membrane fractionation to separate lipid rafts as detergent-resistant membranes. In control cells, fractions containing cholesterol-rich micro-domains, identified by the presence of raft markers GM1 or caveolin, partitioned in top fractions (Supplementary Figure 2G). Combining these fractions for Western blot analysis showed that GM1 as well as IFNGR1 were lost in detergent-resistant membranes upon infection with wild-type P12 but not P12Δcgt, regardless of IFNG treatment (Figure 2E). In total lysates, IFNGR1 and IFNGR2 were detected in all conditions (Supplementary Figure 2H). Finally, we performed immunoprecipitation of IFNGR subunit 2 to test co-precipitation of IFNGR1. Immunoprecipitation specificity was controlled for by absence of the transferrin receptor CD71, a protein not associated with lipid rafts or the IFNGR complex (Supplementary Figure 2I). In line with recent findings,
      • Blouin C.M.
      • Hamon Y.
      • Gonnord P.
      • et al.
      Glycosylation-dependent IFN-gammaR partitioning in lipid and actin nanodomains is critical for JAK activation.
      MKN45 cells showed constitutive oligomerization of IFNGR subunits even in the absence of IFNG (Figure 2F). Despite this, prolonged wt H pylori infection but not P12Δcgt abolished assembly of receptor subunits. Coating H pylori with exogenous cholesterol prior to infection restored assembly of IFNGR subunits (Figure 2G). Together, these data suggest that subversion of JAK/STAT1 signaling by H pylori takes place at the very top of the pathway, by preventing assembly of IFNG receptor subunits through lipid raft destruction.

       H pylori Blocks the IFNG Response In Vivo in a Cgt-Dependent Manner

      Reportedly, H pylori cgt mutants are unable to colonize mice.
      • Wang H.J.
      • Cheng W.C.
      • Cheng H.H.
      • et al.
      Helicobacter pylori cholesteryl glucosides interfere with host membrane phase and affect type IV secretion system function during infection in AGS cells.
      • Wunder C.
      • Churin Y.
      • Winau F.
      • et al.
      Cholesterol glucosylation promotes immune evasion by Helicobacter pylori.
      Because Cgt activity is required for the block of the IFNG response, which in turn protects against H pylori in vivo,
      • Sayi A.
      • Kohler E.
      • Hitzler I.
      • et al.
      The CD4+ T cell-mediated IFN-gamma response to Helicobacter infection is essential for clearance and determines gastric cancer risk.
      we investigated whether the IFNG response is linked to bacterial clearance. After 3 days of infection PMSS1Δcgt was already undetectable in wild-type mice (Figure 3A). Mice infected with PMSS1 or PMSS1Δcgt presented similar IFNG levels in the stomach; however, the induction of the IFNG downstream response gene IRF1 was substantially higher in PMSS1Δcgt-infected mice (Figure 3B). Next, we infected mice for 2 weeks with PMSS1 to allow a stronger Th1 response to develop. We observed a corpus-predominant gland occupation by H pylori in all analyzed mice (Figure 3C). Notably, the IFNG expression (Figure 3D, left) was higher compared with the less colonized antrum. However, IRF1 expression remained at similar levels (Figure 3D, right). Infection with wild-type H pylori thus leads to increased levels of IFNG, but fails to increase expression of interferon response genes, consistent with the notion of a Cgt-dependent block. Finally, although we reasoned that H pylori PMSS1Δcgt might be able to infect Ifngr1-knock out mice (Figure 3A, right), this was not the case. Knockout mice did show increased colonization by wild-type PMSS1, indicating that additional defense-related pathways apart from IFNG might contribute to the clearance of PMSS1Δcgt.
      Figure thumbnail gr3
      Figure 3IFN response triggered by H pylori in vivo is ineffective. (A) Bacterial clearance in C57BL6 wt and Ifngr1 -/- mice orally infected with PMSS1 wt or PMSS1Δcgt. Each point represents CFU counts for individual mice. Mean ± SD is also represented. **P <.01. (B) RT-qPCR analysis of IFNG and IRF1 expression in stomachs of infected wt and Ifngr1 -/- mice, shown as fold-change to respective expression of mGAPDH. Data represented as mean ± SD. ns, non-significant, *P <.05. (C) Confocal microscopy image of longitudinal stomach sections of mice orally infected with PMSS1 for 2 weeks, labelled for actin (phalloidin, red), H pylori (green), and nuclei (DAPI, blue). Scale bars: 1000 μm (upper) and 100 μm (lower). (D) RT-qPCR analysis of IFNG and IRF1 expression in samples taken from corpus or antrum of the mice used in (C), results represented as in (B).

       Cholesterol Depletion by H pylori Inhibits Type I IFN, IL6, and IL22 Signaling and Downstream hBD3 Response

      Lipid raft integrity is also known to be required for type I IFNs and IL6 signaling.
      • Takaoka A.
      • Mitani Y.
      • Suemori H.
      • et al.
      Cross talk between interferon-gamma and -alpha/beta signaling components in caveolar membrane domains.
      • Sehgal P.B.
      • Guo G.G.
      • Shah M.
      • et al.
      Cytokine signaling: STATS in plasma membrane rafts.
      In congruence, we observed that H pylori P12 also inhibited IFNB-induced STAT1 signaling in MKN45 cells, and this effect was dependent on the presence of cgt but not cagA or cagPAI (Figure 4A). Also, mβCD treatment inhibited IFNB signaling in a dose-dependent manner (Supplementary Figure 3A). Similarly, wild-type infection and mβCD treatment, but not infection with the cgt mutant, prevented STAT3 phosphorylation upon IL6 treatment (Figure 4B). IL22 is a crucial cytokine for mediating the epithelial defense against mucosal pathogens. Binding to receptors in epithelial cells triggers STAT3 activation, inducing expression of antimicrobial factors.
      • Zheng Y.
      • Valdez P.A.
      • Danilenko D.M.
      • et al.
      Interleukin-22 mediates early host defense against attaching and effacing bacterial pathogens.
      Again, 24-hour infection with wt P12 but not a P12Δcgt mutant strain inhibited IL22 signaling transduction in epithelial cells (Figure 4C); the block was also observed upon mβCD treatment (Figure 4C). Genetic rescue of the cgt mutant partially restored the bacterial capacity to block IL22 signaling (Supplementary Figure 3B). Because IL22 signaling has not previously been reported to depend on cholesterol, we demonstrated restoration of IL22 signaling by adding PEG-cholesterol to MKN45 cells before IL22 stimulation (Supplementary Figure 3C).
      Figure thumbnail gr4
      Figure 4Cgt prevents IFNB-, IL6-, and IL22-induced signaling in epithelial cells upon prolonged infection and impairs hBD3 expression. (A) Immunoblot analysis of STAT1 pathway activation in MKN45 cells. Cells were infected for 24 hours with indicated strains or non-infected, then treated (+) or mock-treated (-) with IFNB (2300 U/mL) for 30 minutes. (B and C) Immunoblot analysis of STAT3 activation in MKN45 cells infected for 24 hours with indicated strains or treated with mβCD (5 mg/mL) for 5 hours, followed by 30 minutes with (B) IL6 (100 ng/mL) or (C) IL22 (50 ng/mL). (D and E) RT-qPCR analysis of hBD3. MKN45 cells were non-infected or infected with indicated strains for 24 hours. IFNG (10 ng/mL) (D) or IL22 (50 ng/mL) (E) were then added to selected wells for the specified times. To represent the net response to cytokines, results are normalized against relative gene expression of corresponding non-infected, P12 wt and P12Δcgt infected cells. Data represented as mean ± SD. ns, non-significant, **P <.01.
      Because IFNG and IL22 are reported to induce epithelial expression of hBD3,
      • Ostaff M.J.
      • Stange E.F.
      • Wehkamp J.
      Antimicrobial peptides and gut microbiota in homeostasis and pathology.
      a potent defensin against H pylori,
      • Nuding S.
      • Gersemann M.
      • Hosaka Y.
      • et al.
      Gastric antimicrobial peptides fail to eradicate Helicobacter pylori infection due to selective induction and resistance.
      • Bauer B.
      • Wex T.
      • Kuester D.
      • et al.
      Differential expression of human beta defensin 2 and 3 in gastric mucosa of Helicobacter pylori-infected individuals.
      we tested the impact of infection on the cytokine-induced hBD3 response. Infection of MKN45 cells for 24 hours with P12Δcgt but not P12 wt, followed by treatment with IFNG for 5 hours, induced a significant increase of hBD3 transcription (Figure 4D, left), which was even more dramatic after 24-hour IFNG treatment (Figure 4D, right). In contrast, the increase over time in non-infected and P12 wt-infected cells was minimal. Similarly, IL22 treatment also induced hBD3 expression in P12Δcgt-infected cells, albeit to a lower extent, following a similar time course as IFNG (Figure 4E). In summary, H pylori blocks the production of hBD3 in epithelial cells stimulated with IFNG or IL22. Interestingly, co-stimulation with a cgt-deficient strain induced substantially higher hBD3 expression compared with cytokine treatment alone. Overall, these data confirm that cholesterol depletion by H pylori not only inhibits the response to IFNG, but also type I IFNs, IL6, and IL22.

       Inhibition of JAK/STAT Signaling in Human Primary Gastric Epithelial Cells

      We next used human gastric primary cultures
      • Schlaermann P.
      • Toelle B.
      • Berger H.
      • et al.
      A novel human gastric primary cell culture system for modelling Helicobacter pylori infection in vitro.
      • Bartfeld S.
      • Bayram T.
      • van de Wetering M.
      • et al.
      In vitro expansion of human gastric epithelial stem cells and their responses to bacterial infection.
      to validate our observations during authentic host-pathogen interaction. Cells derived from primary gastric antral organoids were seeded on plastic and infected with H pylori under serum starvation before IFNG or IFNB treatment for 30 minutes. Similar to results obtained with MKN45, H pylori P12, but not P12Δcgt, prevented JAK1/STAT1 phosphorylation (Figure 5A). These differences were also observed at higher MOI and in cells isolated from the corpus region (Supplementary Figure 4A), independent of donors (Supplementary Figure 4B). In contrast to cancer cell lines, primary gastric epithelial cells did not exhibit constitutive JAK1 activation (Figure 5A). Block of IFNG signaling was observed at 24 hours post-infection, but not at earlier time points (Supplementary Figure 4C) or with a low MOI (Supplementary Figure 4D). Cholesterol-coated bacteria, however, failed to block JAK/STAT1 signaling (Figure 5B). Further, block of IFN signaling, as determined by the up-regulation of the type II IFN-activated gene CXCL9 in response to IFNG stimulation, was observed in H pylori P12 wt, P12ΔcagA, P12ΔcagPAI, or P12ΔvacA, but not in P12Δcgt-infected cells (Supplementary Figure 4E). Moreover, H pylori P12 wt, but not P12Δcgt, inhibited STAT3 phosphorylation upon IL22 treatment at 24 hours post-infection (Figure 5C). Genetic rescue of P12Δcgt restored its capacity to block the response to IFNG and IL22 in primary epithelial cells (Supplementary Figure 4F and G).
      Figure thumbnail gr5
      Figure 5Cgt blocks IFN signaling in primary gastric epithelial cells. Human primary gastric cells were infected with indicated strains at MOI 20 for 24 hours or non-infected. (A) Immunoblot analysis of JAK1/STAT1 pathway activation. After infection cells from antrum were mock-treated or treated for 30 minutes with IFNG (5 ng/mL) or IFNB (2300 U/mL). Quantification: band intensity ratio STAT1pY: actin. (B) Cholesterol-coating of bacteria was started 1 hour before infection and applied to non-infected cells accordingly. Afterwards cells were mock-treated or treated for 30 minutes with IFNG (5 ng/mL). Result representative of 2 independent experiments. (C) Immunoblot analysis of STAT3 activation. Cells from antrum were infected for 24 hours with indicated strains and after infection treated for 30 minutes with IL22 (50 ng/mL). (D and E) Primary human cells were infected with H pylori P12 wt or P12Δcgt for 26.5 hours and microarray results compared with those obtained in non-infected (n.i.) cells. Results shown are from 2 independent experiments with cells obtained from 2 independent donors. (D) Gene Ontology processes enrichment (Panther) from up-regulated genes in n.i. vs. P12 wt data set. In all cases P < .05. (E) NF-κB-response genes differentially modulated in n.i. vs. P12 wt and n.i. vs P12Δcgt datasets. In all cases, fold change > 1.5 and P value <.05.
      To assess the global cellular response to infection with H pylori P12 wt or P12Δcgt, we performed microarray analysis with human primary gastric cells infected with P12 wt or P12Δcgt. Gene Ontology term enrichment analysis revealed responsiveness for genes involved in ‘response to external stimuli,’ ‘signal transduction,’ and ‘immune response,’ which was similar for both wt and P12Δcgt-infected cells (Figure 5D, and data not shown). Many of the pro-inflammatory genes up-regulated upon infection with either strain are related to NF-κB signaling (Figure 5E). Therefore, while H pylori effectively blocks the response to IFNG, IFNB, and IL22 in normal human gastric epithelial cells, initial sensing of the pathogen by NF-κB
      • Zimmermann S.
      • Pfannkuch L.
      • Al-Zeer M.A.
      • Bartfeld S.
      • et al.
      ALPK1- and TIFA-dependent innate immune response triggered by the helicobacter pylori type IV secretion system.
      remains largely unaffected.

       H pylori Generates Micro-Niches of Diminished Inflammatory Response

      To spatially resolve the effects H pylori exerts, we examined the resulting cellular phenotypes over longer periods of time using a novel infection model of human gastric primary cells in air-liquid interphase culture.
      • Boccellato F.
      • Woelffling S.
      • Imai-Matsushima A.
      • et al.
      Polarised epithelial monolayers of the gastric mucosa reveal insights into mucosal homeostasis and defence against infection.
      This model enables longer infection times and resembles the in vivo situation more closely because it features greater cell type diversity, a protective layer of mucin, and cell polarization (Supplementary Figure 5A). After 3 days of infection with P12 wt at MOI 100, cells were treated with IFNG for 30 minutes and analyzed by immunofluorescence. Interestingly, IFNG treatment led to nuclear translocation of phospho-STAT1 (Figure 6A, middle panel). Micro-colonies of spiral-shaped bacteria (red) formed infection foci (Figure 6B and Supplementary Figures 5A and B) correlating with areas of reduced phospho-STAT1 exhibition (Figure 6A, bottom panel). At lower magnification, areas of reduced STAT1 activation appeared to clearly correlate with infected areas, while non-infected areas of the same filter responded normally to IFNG (Figure 6B). Next, we quantified the number of phospho-STAT1–positive cells in infected compared with non- or less well-infected microscopic fields of the same filter. The vast majority of infected areas displayed phospho-STAT1 levels below the activation threshold, while non-infected fields exhibited a robust IFNG response (Supplementary Figure 5C). Infection with the P12Δcgt mutant at MOI 200 led to similar colonization densities, yet the infected monolayer exhibited full STAT1 phosphorylation upon IFNG similar to non-infected cells (Supplementary Figure 5D). Similarly, H pylori infection also hampered STAT3 phosphorylation upon IL22 treatment in infected areas (Figure 6C). Thus, using an advanced epithelial cell culture model, we demonstrate that H pylori prevents infected cells from responding to IFNG or IL22. This favors the formation of micro-colonies at micro-niches devoid of STAT1 or STAT3 activation. Although non-infected sites of the same culture retain full responsiveness, this does not halt micro-colony formation at protected sites.
      Figure thumbnail gr6
      Figure 6Block of cytokine responses in infected ALI cultures. Primary human cells grown on filters were infected with P12 wt at MOI 100 for 3 days or mock-infected (NI), then mock-treated or treated for 30 minutes with IFNG (10 ng/mL) (A and B) or IL22 (50 ng/mL) (C) and fixed with PFA. Filters were labelled with DAPI (DNA, blue) and antibodies against H pylori (red) and analyzed by whole-mount IF. (A) Analysis of STAT1 phosphorylation. Filters were labelled with phalloidin (green) and STAT1-pY (white). Scale bar: 20 μm. (B) Analysis of STAT1 phosphorylation, lower magnification. Filters were labelled with antibodies against STAT1-pY (white). Scale bar: 50 μm. (C) Analysis of STAT3 phosphorylation. Filters were labelled with antibodies against STAT3-pY (white). Scale bar: 20 μm.

      Discussion

      A hallmark of gastric infections with H pylori is the strong NF-κB–driven response of the epithelium. This initiates a chronic inflammatory condition fueled by the recruitment of immune cells, which produce IFNs
      • Bamford K.B.
      • Fan X.
      • Crowe S.E.
      • et al.
      Lymphocytes in the human gastric mucosa during Helicobacter pylori have a T helper cell 1 phenotype.
      • Sayi A.
      • Kohler E.
      • Hitzler I.
      • et al.
      The CD4+ T cell-mediated IFN-gamma response to Helicobacter infection is essential for clearance and determines gastric cancer risk.
      and other cytokines, including IL22.
      • Zheng Y.
      • Valdez P.A.
      • Danilenko D.M.
      • et al.
      Interleukin-22 mediates early host defense against attaching and effacing bacterial pathogens.
      These cytokines trigger a second wave of epithelial responses, characterized by the release of defensins and other bactericidal factors, to restrict pathogen growth.
      • Ostaff M.J.
      • Stange E.F.
      • Wehkamp J.
      Antimicrobial peptides and gut microbiota in homeostasis and pathology.
      As we report here, H pylori evolved a powerful means to prevent infected epithelial cells from responding to this cytokine burst – shutting down bactericidal activity right at the site of infection. This intriguing phenomenon, exerted by the bacterial enzyme Cgt, is in line with previous observations that point to an unexplained inhibition of IFNG-induced nuclear translocation of STAT1 during H pylori infection.
      • Mitchell P.
      • Germain C.
      • Fiori P.L.
      • et al.
      Chronic exposure to Helicobacter pylori impairs dendritic cell function and inhibits Th1 development.
      Our findings now provide the mechanistic clue to our earlier characterization of cgt as a critical determinant of H pylori immune evasion.
      • Wunder C.
      • Churin Y.
      • Winau F.
      • et al.
      Cholesterol glucosylation promotes immune evasion by Helicobacter pylori.
      The gastric mucosa of patients exhibits a marked prevalence of pro-inflammatory cytokines and mediators that provoke innate and Th1-, Th17-, and Th22-driven immune responses.
      • Horvath Jr., D.J.
      • Washington M.K.
      • Cope V.A.
      • et al.
      IL-23 Contributes to control of chronic Helicobacter pylori infection and the development of T helper responses in a mouse model.
      • Zhuang Y.
      • Cheng P.
      • Liu X.F.
      • et al.
      A pro-inflammatory role for Th22 cells in Helicobacter pylori-associated gastritis.
      • Rad R.
      • Dossumbekova A.
      • Neu B.
      • et al.
      Cytokine gene polymorphisms influence mucosal cytokine expression, gastric inflammation, and host specific colonisation during Helicobacter pylori infection.
      • Wen S.
      • Felley C.P.
      • Bouzourene H.
      • et al.
      Inflammatory gene profiles in gastric mucosa during Helicobacter pylori infection in humans.
      However, IFNG and IL17 merely control, rather than clear, established infections. Similarly, IL22 induces the production of relevant antimicrobial factors in gastric epithelial cells, but surprisingly IL22 knockout and wt mice show similar H pylori colonization rates.
      • Dixon B.R.
      • Radin J.N.
      • Piazuelo M.B.
      • et al.
      IL-17a and IL-22 induce expression of antimicrobials in gastrointestinal epithelial cells and may contribute to epithelial cell defense against Helicobacter pylori.
      Previous work identified the H pylori factors VacA and GGT as immune modulators that impair the activation and proliferation of T cells.
      • Salama N.R.
      • Hartung M.L.
      • Muller A.
      Life in the human stomach: persistence strategies of the bacterial pathogen Helicobacter pylori.
      • Gebert B.
      • Fischer W.
      • Weiss E.
      • et al.
      Helicobacter pylori vacuolating cytotoxin inhibits T lymphocyte activation.
      They skew normal T-cell maturation from Th1 and Th17 toward a Treg phenotype and thereby elicit a considerable degree of local and systemic tolerance.
      • Oertli M.
      • Noben M.
      • Engler D.B.
      • et al.
      Helicobacter pylori gamma-glutamyl transpeptidase and vacuolating cytotoxin promote gastric persistence and immune tolerance.
      Although this deficit of T-cell maturation can be rescued by vaccination, vaccine-driven T-cell activation normally achieves only a reduction of the bacterial load concomitant with an increased inflammation.
      • Koch M.
      • Meyer T.F.
      • Moss S.F.
      Inflammation, immunity, vaccines for Helicobacter pylori infection.
      This paradox points to a block of the T-cell–mediated immunity at the effector side of host epithelial defense, which can be explained by the action of Cgt. Because Cgt acts only on infected cells, this gives rise to an intriguing scenario: while inflammation prevails in the infected tissue, Cgt generates protected islands of diminished defense, promoting pathogen survival. Indeed, this notion is corroborated by observations with chronically infected mice challenged with a secondary H pylori infection
      • Keilberg D.
      • Zavros Y.
      • Shepherd B.
      • et al.
      Spatial and temporal shifts in bacterial biogeography and gland occupation during the development of a chronic infection.
      : newly incoming bacteria were unable to colonize infected mice except for a few glands already occupied by the primary infection, pointing toward an overall anti-microbial environment in the infected stomach. Together these observations suggest that H pylori forms protected micro-niches surrounded by an otherwise inflamed and microbicidal milieu (Figure 7).
      Figure thumbnail gr7
      Figure 7Summary of cholesterol-dependent subversion of immune response. Extraction of host cell cholesterol by H pylori’s Cgt disrupts lipid rafts and thereby prevents assembly of IFNG receptors. In infected cells this leads to a shutdown of JAK/STAT1 signaling, inhibiting the production of T-cell chemo-attractants and antimicrobial peptides like hBD3. With this mechanism, H pylori creates micro-niches of locally diminished inflammation, despite a marked global immune response.
      Amongst the T-cell chemotactic factors affected by Cgt are MIG and IP-10, which contribute to protection against H pylori.
      • Watanabe T.
      • Asano N.
      • Fichtner-Feigl S.
      • et al.
      NOD1 contributes to mouse host defense against Helicobacter pylori via induction of type I IFN and activation of the ISGF3 signaling pathway.
      • Flach C.F.
      • Mozer M.
      • Sundquist M.
      • et al.
      Mucosal vaccination increases local chemokine production attracting immune cells to the stomach mucosa of Helicobacter pylori infected mice.
      Homeostasis of the mucosal colonization by H pylori, avoiding an excessive bacterial load, is thought to be achieved through secretion of a variety of innate factors, including defensins, mucins, and oxidative metabolites.
      • Nuding S.
      • Gersemann M.
      • Hosaka Y.
      • et al.
      Gastric antimicrobial peptides fail to eradicate Helicobacter pylori infection due to selective induction and resistance.
      • Ermak T.H.
      • Giannasca P.J.
      • Nichols R.
      • et al.
      Immunization of mice with urease vaccine affords protection against Helicobacter pylori infection in the absence of antibodies and is mediated by MHC class II-restricted responses.
      • Kawakubo M.
      • Ito Y.
      • Okimura Y.
      • et al.
      Natural antibiotic function of a human gastric mucin against Helicobacter pylori infection.
      Many of these are produced by the epithelium in response to IFNG and IL22.
      • Zheng Y.
      • Valdez P.A.
      • Danilenko D.M.
      • et al.
      Interleukin-22 mediates early host defense against attaching and effacing bacterial pathogens.
      • Fahlgren A.
      • Hammarstrom S.
      • Danielsson A.
      • et al.
      beta-Defensin-3 and -4 in intestinal epithelial cells display increased mRNA expression in ulcerative colitis.
      • Joly S.
      • Organ C.C.
      • Johnson G.K.
      • et al.
      Correlation between beta-defensin expression and induction profiles in gingival keratinocytes.
      hBD3, one of the most potent factors against H pylori, is induced in an IFNG-dependent manner in the context of chronic gastrointestinal inflammation.
      • Nuding S.
      • Gersemann M.
      • Hosaka Y.
      • et al.
      Gastric antimicrobial peptides fail to eradicate Helicobacter pylori infection due to selective induction and resistance.
      • Fahlgren A.
      • Hammarstrom S.
      • Danielsson A.
      • et al.
      beta-Defensin-3 and -4 in intestinal epithelial cells display increased mRNA expression in ulcerative colitis.
      We have previously observed that EGFR-dependent induction of hBD3 is down-regulated upon translocation of H pylori CagA via the activation of SHP-214 and CagA-dependent activation of SHP-2 also interferes with IFNG signaling.
      • Wang Y.C.
      • Chen C.L.
      • Sheu B.S.
      • et al.
      Helicobacter pylori infection activates Src homology-2 domain-containing phosphatase 2 to suppress IFN-gamma signaling.
      Thus, CagA acts synergistically with Cgt in preventing IFN signaling and hBD3 synthesis, with Cgt expressed also in strains that lack the cagPAI T4SS. Our results, however, show a minimal contribution of CagA in blocking JAK/STAT signaling, possibly because of variations in the active motifs of CagA proteins in our European strain as compared with Asian strains,
      • Naito M.
      • Yamazaki T.
      • Tsutsumi R.
      • et al.
      Influence of EPIYA-repeat polymorphism on the phosphorylation-dependent biological activity of Helicobacter pylori CagA.
      such as the one used by Wang and colleagues.
      • Wang Y.C.
      • Chen C.L.
      • Sheu B.S.
      • et al.
      Helicobacter pylori infection activates Src homology-2 domain-containing phosphatase 2 to suppress IFN-gamma signaling.
      Consistently, biopsies from European patients exhibited poor bactericidal activity against H pylori and lack of hBD3, regardless of whether they were cagPAI/cagA positive or negative.
      • Nuding S.
      • Gersemann M.
      • Hosaka Y.
      • et al.
      Gastric antimicrobial peptides fail to eradicate Helicobacter pylori infection due to selective induction and resistance.
      • Bauer B.
      • Wex T.
      • Kuester D.
      • et al.
      Differential expression of human beta defensin 2 and 3 in gastric mucosa of Helicobacter pylori-infected individuals.
      Lipid rafts are also targeted by other pathogens known to block IFN signaling (such as West Nile virus), which modifies the distribution of lipid rafts in the membrane
      • Mackenzie J.M.
      • Khromykh A.A.
      • Parton R.G.
      Cholesterol manipulation by West Nile virus perturbs the cellular immune response.
      and Leishmania donovani, which disrupts lipid rafts in macrophages to inhibit assembly of IFNGR subunits.
      • Sen S.
      • Roy K.
      • Mukherjee S.
      • et al.
      Restoration of IFNgammaR subunit assembly, IFNgamma signaling and parasite clearance in Leishmania donovani infected macrophages: role of membrane cholesterol.
      Here, using H pylori, we achieve striking mechanistic insights into microbial cholesterol depletion from host cells, an apparently common, yet little appreciated virulence strategy. Accordingly, cholesterol depletion prevents the partitioning of IFNGR1 to lipid rafts and association with IFNGR2 subunits.
      • Sen S.
      • Roy K.
      • Mukherjee S.
      • et al.
      Restoration of IFNgammaR subunit assembly, IFNgamma signaling and parasite clearance in Leishmania donovani infected macrophages: role of membrane cholesterol.
      In MKN45 cells, we observed an association of IFNGR2 and IFNGR1 subunits even in the absence of infection, consistent with the abnormal, constitutive JAK1/2 activation. However, this association collapsed upon H pylori infection, rendering these cells unresponsive, independently of the presence of interferons.
      Our in vivo experiments confirmed a Cgt-dependent block of the IFNG response, with the results obtained in Ifngr1 knockout mice indicating that additional pathways contribute to preventing colonization with the PMSS1Δcgt mutant. This finding is consistent with the inhibitory action of Cgt on multiple lipid raft-dependent pathways, including IL6 and IFNB,
      • Takaoka A.
      • Mitani Y.
      • Suemori H.
      • et al.
      Cross talk between interferon-gamma and -alpha/beta signaling components in caveolar membrane domains.
      • Sehgal P.B.
      • Guo G.G.
      • Shah M.
      • et al.
      Cytokine signaling: STATS in plasma membrane rafts.
      • Mackenzie J.M.
      • Khromykh A.A.
      • Parton R.G.
      Cholesterol manipulation by West Nile virus perturbs the cellular immune response.
      and IL22, in addition to IFNG. Vice versa, Cgt may also cause receptor activation such as for EGFR and TGFBR, where cholesterol depletion increases the number of molecules available for ligand binding.
      • Arish M.
      • Husein A.
      • Kashif M.
      • et al.
      Orchestration of membrane receptor signaling by membrane lipids.
      • Westover E.J.
      • Covey D.F.
      • Brockman H.L.
      • et al.
      Cholesterol depletion results in site-specific increases in epidermal growth factor receptor phosphorylation due to membrane level effects. Studies with cholesterol enantiomers.
      • Chen C.L.
      • Liu I.H.
      • Fliesler S.J.
      • et al.
      Cholesterol suppresses cellular TGF-beta responsiveness: implications in atherogenesis.
      Apart from this, many signaling routes, particularly those involving the pro-inflammatory pathway NF-κB, appear to function normally in H pylori-infected cells, even at a time when cholesterol depletion has progressed.
      We have substantiated our findings in primary epithelial cells from different donors. Importantly, the use of cells from human gastric organoids allowed us to address biological questions in a mutation-free background.
      • Schlaermann P.
      • Toelle B.
      • Berger H.
      • et al.
      A novel human gastric primary cell culture system for modelling Helicobacter pylori infection in vitro.
      Although primary cell models are increasingly used to address questions of infection biology,
      • Bartfeld S.
      Modeling infectious diseases and host-microbe interactions in gastrointestinal organoids.
      little data related to H pylori are available yet. Here, we also utilized an advanced ALI model, which offers several improvements: (i) cultures grow under the influence of Wnt to maintain stemness; (ii) cells differentiate toward a diversity of cells, including pit, neck, chief, parietal, and neuroendocrine cells; (iii) cell polarization and secretion of a protective mucus layer supports an authentic equilibrium between cells and bacteria, enabling long-term infection.
      • Boccellato F.
      • Woelffling S.
      • Imai-Matsushima A.
      • et al.
      Polarised epithelial monolayers of the gastric mucosa reveal insights into mucosal homeostasis and defence against infection.
      As we have previously shown, dietary cholesterol supplementation significantly reduces the H pylori burden in mice.
      • Wunder C.
      • Churin Y.
      • Winau F.
      • et al.
      Cholesterol glucosylation promotes immune evasion by Helicobacter pylori.
      The decreased colonization came at the cost of substantially increased IFNG-driven inflammation. Our current results, together with the established link between lipid rafts and IFN signaling,
      • Takaoka A.
      • Mitani Y.
      • Suemori H.
      • et al.
      Cross talk between interferon-gamma and -alpha/beta signaling components in caveolar membrane domains.
      • Sen S.
      • Roy K.
      • Mukherjee S.
      • et al.
      Restoration of IFNgammaR subunit assembly, IFNgamma signaling and parasite clearance in Leishmania donovani infected macrophages: role of membrane cholesterol.
      are in complete agreement with previous in vivo findings.
      • Wunder C.
      • Churin Y.
      • Winau F.
      • et al.
      Cholesterol glucosylation promotes immune evasion by Helicobacter pylori.
      Accordingly, cholesterol supply in the context of H pylori infection enhances JAK/STAT signaling together with the release of antibacterial effectors. Such increased Th1 responses, however, are also thought to promote preneoplastic lesions.
      • Sayi A.
      • Kohler E.
      • Hitzler I.
      • et al.
      The CD4+ T cell-mediated IFN-gamma response to Helicobacter infection is essential for clearance and determines gastric cancer risk.
      • Syu L.J.
      • El-Zaatari M.
      • Eaton K.A.
      • et al.
      Transgenic expression of interferon-gamma in mouse stomach leads to inflammation, metaplasia, and dysplasia.
      This notion, in turn, is consistent with the situation in patients suffering from increased blood cholesterol levels, especially in the form of low-density lipoprotein-cholesterol, who often exhibit severe H pylori-induced gastritis.
      • Kucukazman M.
      • Yavuz B.
      • Sacikara M.
      • et al.
      The relationship between updated Sydney System score and LDL cholesterol levels in patients infected with Helicobacter pylori.
      Accordingly, a low-cholesterol diet may reduce the pathology of H pylori gastric infections. This places Cgt function, cholesterol metabolism, and inflammation at the crossroads of gastric pathogenesis and cancer.

      Acknowledgments

      The authors thank Jörg Angermann, Kirstin Hoffmann, Stefanie Mülllerke, and Ina Wagner for technical assistance, Robert Hurwitz for generating the Cgt antibody, Toni Aebischer, Bianca Bauer, Michael Fehlings, June Ghosh-Guha, and Eric Perret for fruitful discussions, and Rike Zietlow for expert editing of the manuscript.

      Supplementary Material

      Figure thumbnail fx2
      Supplementary Figure 1(A) Time course analysis of IFNG induced STAT1 phosphorylation by immunoblot. MKN45 cells were infected with H pylori P12 wild type at MOI 50 or non-infected (NI). After indicated infection times, cells were treated (+) with IFNG (10 ng/mL, 30 minutes). (B) Immunoblot analysis of IFNG induced STAT1 phosphorylation in AGS gastric epithelial cell line upon infection with indicated H pylori strain at MOI 50 and in non-infected cells. After 24 hours of infection, cells were treated (+) with IFNG (10 ng/mL, 30 minutes). (C) MOI titration analysis of IFNG induced STAT1 phosphorylation by immunoblot. MKN45 cells were infected with H pylori P12 wild type at indicated MOIs or non-infected (NI). After 24 hours of infection, cells were treated (+) with IFNG (10 ng/mL, 30 minutes). (D) Analysis of JAK2 phosphorylation (red) by confocal IF at low magnification, confocal section. Cells were non-infected (NI), or infected with indicated H pylori strains (bacteria labeled in green). In all cases, cells were treated with IFNG (10 ng/mL, 30 minutes). Scale bar: 10 μm. The result is representative of 3 independent experiments. (E) Cellular viability of MKN45 cells infected with indicated H pylori strains, or heat-killed P12 (P12 HK) at indicated MOIs for 24 hours. Viability was measured via a fluorescence-based ATP assay (Promega, Madison, WI). All readings were normalized against non-infected (NI) cells. Cells treated for 5 minutes with Triton X-100 serve as control. Data shows mean ± SD of 2 independent biological experiments. SD is derived from 3 technical replicates. (F) Immunoblot analysis of IFNG induced JAK1 and STAT1 phosphorylation in MKN45 gastric epithelial cells after 48 hours of infection with indicated H pylori strains at MOI 50 and in non-infected cells. At the end of infection, cells were treated (+) or mock-treated (-) with IFNG (10 ng/mL) for 30 minutes. Quantification shows band intensity ratio of STAT1pY:actin. (G) Immunoblot analysis of CGT expression after genetic complementation of CGT mutant strain. 4x106 bacteria of H pylori P12 wt, P12Δcgt, ΔcagA, ΔcagPAI, and 2 clones of the recomplemented strains Δcgt/cgt were lysed in Laemmli buffer and checked for CGT expression by using an in-house created monoclonal anti-CGT antibody. (H) Response to IFNG for genes IRF1, CXCL9, and CXCL10 quantified by RT-qPCR analysis. MKN45 cells were non-infected or infected with indicated H pylori strains for 24 hours. IFNG (10 ng/mL) was then added to selected wells for the specified times. To represent the net response to IFN, results were normalized against relative gene expression of corresponding non-infected, P12 and P12Δcgt infected cells. Data are represented as mean ± SD; ns, non-significant, *P < .05, **P < .01.
      Figure thumbnail fx3
      Supplementary Figure 2(A) Cholesterol quantification of MKN45 upon H pylori infection. Cells were non-infected (NI) or infected with indicated strains of H pylori for 24 hours at MOI 50. After this period, cholesterol was quantified using Amplex kit (Invitrogen Carlsbad, CA). All readings were normalized against protein content and further normalized against cholesterol levels in non-infected cells. Data show mean ± SD of 2 independent biological experiments with 2 technical replicates each. (B) Cellular viability of MKN45 cells treated with mβCD (5 mg/mL) for 6 hours. Viability was measured via a fluorescence-based ATP assay (Promega). All readings were normalized against the NI and mock-treated cells. Cells treated for 5 minutes with TritonX-100 served as control. Data show mean ± SD of 2 independent biological experiments with 2 technical replicates each. (C) Response to 2.5-hour treatment with IFNG (10 ng/mL) after mock treatment or treatment with myriocin (72 hours, 50 μmol/L) plus zaragozic acid (18 hours, 50 μmol/L) (MZ). IRF1 mRNA levels obtained by qRT-PCR were normalized against respective GAPDH mRNA expression, and further normalized against non-infected (NI) cells. Data show mean ± SD of 2 independent biological experiments with 2 technical replicates each. For statistical analysis, MZ-treated samples were compared with mock, IFNG-treated controls: ns non-significant, **P <.01. (D) Response to 2.5-hour treatment with IFNG for IRF1 gene in cells infected with bacteria with or without cholesterol coating. MKN45 cells were non-infected or infected with indicated H pylori strains for 24 hours. IFNG (10 ng/mL) was then added to selected wells for the 2.5 hours. To represent the net response to IFN, results were normalized against relative gene expression of respective corresponding non-infected, P12 and P12Δcgt infected cells. Data are represented as mean ± SD; ns, non-significant, **P < .01. (E) Viability of cholesterol-coated bacteria. Bacteria (108/mL) in serum-free medium were coated or mock-coated with exogenous cholesterol (1 mg/mL) for 1 hour. Colony forming units (CFU)/mL were quantified 3 days later. Data are expressed as percentage bacterial survival relative to the untreated P12 (without cholesterol coating). Data show mean ± SD of 2 independent biological experiments with 2 technical replicates each. (F) Confocal microscopic analysis of the lipid raft marker GM1 (stained with CTxB-Alexa488, green) in cellular membranes of cells non-infected (NI) or infected for 24 hours with indicated H pylori strains. Left panel corresponds to the quantification of relative superficial GM1 signal. Results are shown as a box-plot (2nd and 3rd quartiles, median represented as horizontal line inside the box) of the ratio between cellular membrane and correspondent cytoplasm average intensities and further normalized to the average intensity ratio of non-infected cells of each corresponding experiment. Data correspond to 3 independent experiments; n = number of analyzed cells; **P < .01. Right panel, representative images used for quantification. Scale bar: 10 μm. (G) Immunoblot analysis of detergent resistant membrane fractions (1–10, from top to bottom) of uninfected MKN45 cells after density gradient centrifugation for distribution of caveolin or GM1 distribution detected by dot blot with cholera toxin B conjugated to HRP. Fractions inside the box were pooled for further analysis. (H) Immunoblot analysis of whole cell lysate of MKN45 cells for IFNGR1, IFNGR2, or caveolin. Cells were non-infected (NI) or infected for 24 hours with indicated H pylori strains. After infection time, cells were treated (+) or mock-treated (-) with IFNG (10 ng/mL, 30 minutes). (I) Immunoprecipitation (IP) and immunoblot analysis of IFNGR subunits 1, 2, and CD71. IFNGR2 was pulled down using anti-IFNGR2 antibody.
      Figure thumbnail fx4
      Supplementary Figure 3(A) Immunoblot analysis of STAT1 phosphorylation after 5 hours mβCD treatment at the concentrations indicated or in non-treated cells (-). IFNB treatment (2300 U/mL, 30 minutes) was applied in selected wells. The result is representative of 3 independent experiments. Quantification shows band intensity ratio of STAT1pY:actin (B) Immunoblot analysis of STAT3 phosphorylation of MKN45 cells infected with indicated H pylori strains or mock-infected (NI) for 24 hours at MOI 50 with the indicated H pylori strains, followed by treatment with IL22 (50 ng/mL) for 30 minutes. (C) Immunoblot analysis of STAT3 activation in MKN45 cells. Cells were infected for 24 hours with H pylori P12 wild type or mock infected. In selected wells PEG-cholesterol was added during the final hour of infection. Cells were then treated with IL22 (50 ng/mL, 30 minutes) (+) or mock-treated (-). Quantification shows band intensity ratio of STAT1pY:actin.
      Figure thumbnail fx5
      Supplementary Figure 4Experiment performed in 2D primary gastric human cells from different donors infected with indicated H pylori strains at MOI 20 or 50 for 24 hours or non-infected (NI). (A) Immunoblot analysis of STAT1 pathway activation. Cells from the corpus region were infected with indicated H pylori strains at different MOIs. After infection time cells were mock-treated or treated with IFNG (5 ng/mL). (B) Immunoblot analysis of the activation of JAK1/STAT1 pathway. After infection, cells were mock-treated or treated 30 minutes with IFNG (γ) (5 ng/mL) or IFNB (β) (2300 U/mL). (C) Time course analysis of IFNG-induced STAT1 phosphorylation by immunoblot. Cells were infected with H pylori P12 wt at MOI 20 or non-infected (NI). After indicated infection times, cells were treated (+) with IFNG (5 ng/mL, 30 minutes). (D) Immunoblot analysis of STAT1 phosphorylation. Cells were infected with H pylori wild type at different MOIs or left uninfected. After infection time cells were mock-treated or treated with IFNG (5 ng/mL). (E) Response to 2.5-hour treatment with IFNG (5 ng/mL) after infection with indicated H pylori strains or non-infected (NI). CXCL9 mRNA levels obtained by qRT-PCR were normalized against respective GAPDH mRNA levels, and further normalized against non-infected (NI) cells. Data show mean ± SD of 4 independent biological experiments (performed with cells from 2 different donors) with 2 technical replicates each. For statistical analysis, infected samples were compared with uninfected, IFNG-treated controls: ns non-significant, **P < .01. Immunoblot analysis of STAT1 (F) or STAT3 (G) phosphorylation of primary cells mock-infected (NI) or infected with H pylori P12 wt, P12Δcgt or 2 clones of the recomplemented strains Δcgt/cgt for 24 hours at MOI 50, followed by treatment with (F) IFNG (5 ng/mL, 30 minutes) or (G) IL22 (50 ng/mL, 30 minutes). Quantification of (F) shows band intensity ratio of STAT1pY:actin.
      Figure thumbnail fx6
      Supplementary Figure 5Blockage of cytokine response in infected ALI cultures. Human gastric primary cells were grown on polycarbonate filters, infected with H pylori and analyzed by confocal IF. (A) Schematic representation of an infected ALI monolayer. Microscopic image corresponds to a paraffin section of cells infected with P12 at MOI 100 for 3 days. Monolayers were labelled with DAPI and antibodies against cagA (red) and E-cadherin (green) followed by IF analysis. (B) Detail at high magnification showing infecting H pylori bacteria (labeled in red) from B. Right image corresponds to the square-delimited area on the left image. Scale bars: 50 μm (left), 5 μm (right). (C) Representative microscopic fields from infected filters used for quantification of STAT1 pY-positive cells and quantification of nuclear STAT1 pY-positive cells in infected fields. After 3 days of infection with H pylori wild type (MOI 100), cells were treated for 30 minutes with IFNG (10 ng/mL). Filters were labelled with DAPI (DNA, blue), and antibodies against STAT1 pY (white) and H pylori (red), followed by whole mount IF. Scale bar: 100 μm. For quantification, microscopic fields homogeneously infected (number of cells n = 9548) were compared with non- or minimally infected ones (number of cells n = 15,682). Results were obtained from 4 independent infected filters and are expressed as percentage of nuclear STAT1 pY-positive cells determined by ImageJ software. **P < .01. (D) Analysis of STAT1 phosphorylation. After 3 days of infection with H pylori P12Δcgt (MOI 200), cells were treated for 30 minutes with IFNG (10 ng/mL), and labelled with DAPI (DNA, blue), and antibodies against STAT1 pY (white) and H pylori (red), followed by whole mount IF. Scale bar: 50 μm.
      Supplementary Table 1Primers used in this Study
      Primer nameSequence (5′-3′)
      GAPDH FwdGGTATCGTGGAAGGACTCATGAC
      GAPDH RevATGCCAGTGAGCTTCCCGTTCAG
      IRF1 FwdATGCCCATCACTCGGATGC
      IRF1 RevCCCTGCTTTGTATCGGCCTG
      CXCL9 FwdATTGGAGTGCAAGGAACCCC
      CXCL9 RevGGGCTTGGGGCAAATTGTTT
      CXCL10 FwdGTGGCATTCAAGGAGTACCT
      CXCL10 RevGCATCGATTTTGCTCCCCTC
      IRF7 FwdGAGCTGTGCTGGCGAGAAG
      IRF7 RevTTGGTTGGGACTGGATCTGC
      HP0421 FwdGATCGTCGACATGGTTATTGTTTTAGTCGT
      HP0421 RevGATCAGATCTTTATGATAAGGTTTTAAAGA
      Mouse IRF1 FwdATCTCGGGCATCTTTCGCTT
      Mouse IRF1 RevTCTGCATCTCTAGCCAGGGT
      Mouse GAPDH FwdTCACCATCTTCCAGGAGCG
      Mouse GAPDH RevAAGCAGTTGGTGGTGCAGG

      References

        • Bauer B.
        • Meyer T.F.
        The human gastric pathogen helicobacter pylori and its association with gastric cancer and ulcer disease.
        Ulcers. 2011; 2011: 1-23
        • Sigal M.
        • Rothenberg M.E.
        • Logan C.Y.
        • et al.
        Helicobacter pylori activates and expands Lgr5(+) stem cells through direct colonization of the gastric glands.
        Gastroenterology. 2015; 148: 1392-1404.e21
        • Zimmermann S.
        • Pfannkuch L.
        • Al-Zeer M.A.
        • Bartfeld S.
        • et al.
        ALPK1- and TIFA-dependent innate immune response triggered by the helicobacter pylori type IV secretion system.
        Cell Rep. 2017; 20: 2384-2395
        • Bamford K.B.
        • Fan X.
        • Crowe S.E.
        • et al.
        Lymphocytes in the human gastric mucosa during Helicobacter pylori have a T helper cell 1 phenotype.
        Gastroenterology. 1998; 114: 482-492
        • Salama N.R.
        • Hartung M.L.
        • Muller A.
        Life in the human stomach: persistence strategies of the bacterial pathogen Helicobacter pylori.
        Nat Rev Microbiol. 2013; 11: 385-399
        • Sayi A.
        • Kohler E.
        • Hitzler I.
        • et al.
        The CD4+ T cell-mediated IFN-gamma response to Helicobacter infection is essential for clearance and determines gastric cancer risk.
        J Immunol. 2009; 182: 7085-7101
        • Watanabe T.
        • Asano N.
        • Fichtner-Feigl S.
        • et al.
        NOD1 contributes to mouse host defense against Helicobacter pylori via induction of type I IFN and activation of the ISGF3 signaling pathway.
        J Clin Invest. 2010; 120: 1645-1662
        • Horvath Jr., D.J.
        • Washington M.K.
        • Cope V.A.
        • et al.
        IL-23 Contributes to control of chronic Helicobacter pylori infection and the development of T helper responses in a mouse model.
        Front Immunol. 2012; 3: 56
        • Zhuang Y.
        • Cheng P.
        • Liu X.F.
        • et al.
        A pro-inflammatory role for Th22 cells in Helicobacter pylori-associated gastritis.
        Gut. 2015; 64: 1368-1378
        • Takaoka A.
        • Mitani Y.
        • Suemori H.
        • et al.
        Cross talk between interferon-gamma and -alpha/beta signaling components in caveolar membrane domains.
        Science. 2000; 288: 2357-2360
        • Blouin C.M.
        • Hamon Y.
        • Gonnord P.
        • et al.
        Glycosylation-dependent IFN-gammaR partitioning in lipid and actin nanodomains is critical for JAK activation.
        Cell. 2016; 166: 920-934
        • Ivashkiv L.B.
        • Donlin L.T.
        Regulation of type I interferon responses.
        Nat Rev Immunol. 2014; 14: 36-49
        • Zheng Y.
        • Valdez P.A.
        • Danilenko D.M.
        • et al.
        Interleukin-22 mediates early host defense against attaching and effacing bacterial pathogens.
        Nat Med. 2008; 14: 282-289
        • Bauer B.
        • Pang E.
        • Holland C.
        • et al.
        The Helicobacter pylori virulence effector CagA abrogates human beta-defensin 3 expression via inactivation of EGFR signaling.
        Cell Host Microbe. 2012; 11: 576-586
        • Nuding S.
        • Gersemann M.
        • Hosaka Y.
        • et al.
        Gastric antimicrobial peptides fail to eradicate Helicobacter pylori infection due to selective induction and resistance.
        PLoS One. 2013; 8: e73867
        • Fahlgren A.
        • Hammarstrom S.
        • Danielsson A.
        • et al.
        beta-Defensin-3 and -4 in intestinal epithelial cells display increased mRNA expression in ulcerative colitis.
        Clin Exp Immunol. 2004; 137: 379-385
        • Joly S.
        • Organ C.C.
        • Johnson G.K.
        • et al.
        Correlation between beta-defensin expression and induction profiles in gingival keratinocytes.
        Mol Immunol. 2005; 42: 1073-1084
        • Dixon B.R.
        • Radin J.N.
        • Piazuelo M.B.
        • et al.
        IL-17a and IL-22 induce expression of antimicrobials in gastrointestinal epithelial cells and may contribute to epithelial cell defense against Helicobacter pylori.
        PLoS One. 2016; 11: e0148514
        • Bauer B.
        • Wex T.
        • Kuester D.
        • et al.
        Differential expression of human beta defensin 2 and 3 in gastric mucosa of Helicobacter pylori-infected individuals.
        Helicobacter. 2013; 18: 6-12
        • Lebrun A.H.
        • Wunder C.
        • Hildebrand J.
        • et al.
        Cloning of a cholesterol-alpha-glucosyltransferase from Helicobacter pylori.
        J Biol Chem. 2006; 281: 27765-27772
        • Wang H.J.
        • Cheng W.C.
        • Cheng H.H.
        • et al.
        Helicobacter pylori cholesteryl glucosides interfere with host membrane phase and affect type IV secretion system function during infection in AGS cells.
        Mol Microbiol. 2012; 83: 67-84
        • Wunder C.
        • Churin Y.
        • Winau F.
        • et al.
        Cholesterol glucosylation promotes immune evasion by Helicobacter pylori.
        Nat Med. 2006; 12: 1030-1038
        • Beigier-Bompadre M.
        • Moos V.
        • Belogolova E.
        • et al.
        Modulation of the CD4+ T-cell response by Helicobacter pylori depends on known virulence factors and bacterial cholesterol and cholesterol alpha-glucoside content.
        J Infect Dis. 2011; 204: 1339-1348
        • Du S.Y.
        • Wang H.J.
        • Cheng H.H.
        • et al.
        Cholesterol glucosylation by Helicobacter pylori delays internalization and arrests phagosome maturation in macrophages.
        J Microbiol Immunol Infect. 2016; 49: 636-645
        • Wang Y.C.
        • Chen C.L.
        • Sheu B.S.
        • et al.
        Helicobacter pylori infection activates Src homology-2 domain-containing phosphatase 2 to suppress IFN-gamma signaling.
        J Immunol. 2014; 193: 4149-4158
        • Arish M.
        • Husein A.
        • Kashif M.
        • et al.
        Orchestration of membrane receptor signaling by membrane lipids.
        Biochimie. 2015; 113: 111-124
        • Sen S.
        • Roy K.
        • Mukherjee S.
        • et al.
        Restoration of IFNgammaR subunit assembly, IFNgamma signaling and parasite clearance in Leishmania donovani infected macrophages: role of membrane cholesterol.
        PLoS Pathog. 2011; 7: e1002229
        • Sehgal P.B.
        • Guo G.G.
        • Shah M.
        • et al.
        Cytokine signaling: STATS in plasma membrane rafts.
        J Biol Chem. 2002; 277: 12067-12074
        • Ostaff M.J.
        • Stange E.F.
        • Wehkamp J.
        Antimicrobial peptides and gut microbiota in homeostasis and pathology.
        EMBO Mol Med. 2013; 5: 1465-1483
        • Schlaermann P.
        • Toelle B.
        • Berger H.
        • et al.
        A novel human gastric primary cell culture system for modelling Helicobacter pylori infection in vitro.
        Gut. 2016; 65: 202-213
        • Bartfeld S.
        • Bayram T.
        • van de Wetering M.
        • et al.
        In vitro expansion of human gastric epithelial stem cells and their responses to bacterial infection.
        Gastroenterology. 2015; 148: 126-136.e6
        • Boccellato F.
        • Woelffling S.
        • Imai-Matsushima A.
        • et al.
        Polarised epithelial monolayers of the gastric mucosa reveal insights into mucosal homeostasis and defence against infection.
        Gut. 2018; (https://doi.org/10.1136/gutjnl-2017-314540. [Epub ahead of print])
        • Mitchell P.
        • Germain C.
        • Fiori P.L.
        • et al.
        Chronic exposure to Helicobacter pylori impairs dendritic cell function and inhibits Th1 development.
        Infect Immun. 2007; 75: 810-819
        • Rad R.
        • Dossumbekova A.
        • Neu B.
        • et al.
        Cytokine gene polymorphisms influence mucosal cytokine expression, gastric inflammation, and host specific colonisation during Helicobacter pylori infection.
        Gut. 2004; 53: 1082-1089
        • Wen S.
        • Felley C.P.
        • Bouzourene H.
        • et al.
        Inflammatory gene profiles in gastric mucosa during Helicobacter pylori infection in humans.
        J Immunol. 2004; 172: 2595-2606
        • Gebert B.
        • Fischer W.
        • Weiss E.
        • et al.
        Helicobacter pylori vacuolating cytotoxin inhibits T lymphocyte activation.
        Science. 2003; 301: 1099-1102
        • Oertli M.
        • Noben M.
        • Engler D.B.
        • et al.
        Helicobacter pylori gamma-glutamyl transpeptidase and vacuolating cytotoxin promote gastric persistence and immune tolerance.
        Proc Natl Acad Sci U S A. 2013; 110: 3047-3052
        • Koch M.
        • Meyer T.F.
        • Moss S.F.
        Inflammation, immunity, vaccines for Helicobacter pylori infection.
        Helicobacter. 2013; 18: 18-23
        • Keilberg D.
        • Zavros Y.
        • Shepherd B.
        • et al.
        Spatial and temporal shifts in bacterial biogeography and gland occupation during the development of a chronic infection.
        MBio. 2016; 7
        • Flach C.F.
        • Mozer M.
        • Sundquist M.
        • et al.
        Mucosal vaccination increases local chemokine production attracting immune cells to the stomach mucosa of Helicobacter pylori infected mice.
        Vaccine. 2012; 30: 1636-1643
        • Ermak T.H.
        • Giannasca P.J.
        • Nichols R.
        • et al.
        Immunization of mice with urease vaccine affords protection against Helicobacter pylori infection in the absence of antibodies and is mediated by MHC class II-restricted responses.
        J Exp Med. 1998; 188: 2277-2288
        • Kawakubo M.
        • Ito Y.
        • Okimura Y.
        • et al.
        Natural antibiotic function of a human gastric mucin against Helicobacter pylori infection.
        Science. 2004; 305: 1003-1006
        • Naito M.
        • Yamazaki T.
        • Tsutsumi R.
        • et al.
        Influence of EPIYA-repeat polymorphism on the phosphorylation-dependent biological activity of Helicobacter pylori CagA.
        Gastroenterology. 2006; 130: 1181-1190
        • Mackenzie J.M.
        • Khromykh A.A.
        • Parton R.G.
        Cholesterol manipulation by West Nile virus perturbs the cellular immune response.
        Cell Host Microbe. 2007; 2: 229-239
        • Westover E.J.
        • Covey D.F.
        • Brockman H.L.
        • et al.
        Cholesterol depletion results in site-specific increases in epidermal growth factor receptor phosphorylation due to membrane level effects. Studies with cholesterol enantiomers.
        J Biol Chem. 2003; 278: 51125-51133
        • Chen C.L.
        • Liu I.H.
        • Fliesler S.J.
        • et al.
        Cholesterol suppresses cellular TGF-beta responsiveness: implications in atherogenesis.
        J Cell Sci. 2007; 120: 3509-3521
        • Bartfeld S.
        Modeling infectious diseases and host-microbe interactions in gastrointestinal organoids.
        Dev Biol. 2016; 420: 262-270
        • Syu L.J.
        • El-Zaatari M.
        • Eaton K.A.
        • et al.
        Transgenic expression of interferon-gamma in mouse stomach leads to inflammation, metaplasia, and dysplasia.
        Am J Pathol. 2012; 181: 2114-2125
        • Kucukazman M.
        • Yavuz B.
        • Sacikara M.
        • et al.
        The relationship between updated Sydney System score and LDL cholesterol levels in patients infected with Helicobacter pylori.
        Dig Dis Sci. 2009; 54: 604-607