Abstract
Helicobacter pylori colonizes the stomach of around 50% of humans. This chronic infection can lead to gastric pathologic conditions such as gastric ulcers and gastric adenocarcinomas. The strong inflammatory response elicited by H. pylori is characterized by the induction of the expression of several cytokines. Among those, IL-18 is found highly upregulated in infected individuals, and its expression correlates with the severity of gastric inflammation. IL-18 is produced as inactive proform and has to be cleaved by the multiprotein complex inflammasome to be active. In immune cells, the NLRC4 inflammasome, which is activated by flagellin or bacterial secretion systems, was shown to be dispensable for H. pylori–induced inflammasome activation. However, apart from immune cells, gastric epithelial cells can also produce IL-18. In this study, we analyzed the role of the NLRC4 inflammasome during H. pylori infection. Our results indicate that NLRC4 and a functional type IV secretion system are crucial for the production of IL-18 from human and murine gastric epithelial cells. In vivo, Nlrc4−/− mice failed to produce gastric IL-18 upon H. pylori infection. Compared with wild type mice, Nlrc4−/− mice controlled H. pylori better without showing strong inflammation. Moreover, H. pylori–induced IL-18 inhibits β-defensin 1 expression in a NF-κB–dependent manner, resulting in higher bacterial colonization. At the same time, inflammasome activation enhances neutrophil infiltration, resulting in inflammation. Thus, NLRC4 inflammasome activation and subsequent IL-18 production favors bacterial persistence by inhibiting antimicrobial peptide production and, at the same time, contributes to gastric inflammation.
Introduction
Helicobacter pylori infects the stomach of almost half of the world’s population. Despite inducing strong innate and adaptive immune responses, the bacterium is not cleared and can persist in its niche. Eventually, chronic inflammation leads to the development of more severe pathologies such as gastric ulcers, lymphomas, or carcinomas.
Different inflammatory signaling cascades contribute to H. pylori–associated gastric inflammation and gastric pathologic conditions. Of those, activation of the NLRP3 inflammasome has been shown to be crucial for the gastric production of IL-1β and IL-18 (1), whereas little is known about the function of the NLRC4 inflammasome in the context of H. pylori infection. NLRC4 senses intracellular flagellin or components of bacterial secretion systems. Notably, H. pylori possesses two flagellin subunits, FlaA and FlaB, which constitute the filament of the flagella, as well as a type IV secretion system (T4SS) by which bacterial products can be injected into host cells. Thus, these virulence factors can potentially trigger the NLRC4 inflammasome.
The expression of IL-1β and IL-18 is upregulated in H. pylori–infected individuals (2, 3), reflecting the important role of the inflammasome in H. pylori–induced pathology. IL-1β is mainly released in response to H. pylori by immune cells, and its production depends on the NLRP3 inflammasome (1, 4–8), whereas IL-18 is produced by immune cells (9, 10) as well as epithelial cells upon H. pylori infection (2, 11). In immune cells, IL-18 production was dependent on the NLRP3 inflammasome (5). However, the inflammasome involved in IL-18 secretion by epithelial cells during H. pylori infection has not been characterized. Interestingly, early epithelial IL-18 expression in the gut was found to be important for tissue repair and the induction of antimicrobial peptides, thereby precluding the dissemination of commensal bacteria, dysbiosis, and inflammation (12–14).
Production of antimicrobial peptides, including defensins, represents a first line of defense at epithelial cell surfaces because they prevent mucosal colonization with bacteria, viruses, and fungi. β-Defensins (BDs) are mainly produced by epithelial cells and can be found throughout the gastrointestinal tract. They work via charge-mediated binding to membranes to form pores and channels. This results in membrane rupture and lysis of the pathogen (15, 16). Further, defensins can also function as chemokines (17–20) and modulators of immune responses (21–23). In a second wave, innate and adaptive immune cells induce epithelial cells to produce antimicrobial peptides to fight infections (24, 25). BDs are also important in the context of H. pylori infection. Thus, H. pylori–infected individuals show increased levels of human (h)BD-2, -3, and -4 in the stomach (26–33), whereas the influence of H. pylori infection on the expression of hBD-1, which is constitutively expressed in the stomach, remains unclear (26–29, 31, 32). Notably, little is known on the molecular mechanisms leading to the regulation of defensins during H. pylori infection, apart from the involvement of H. pylori T4SS and NF-κB signaling (27, 28, 34–37).
In the current study, we have addressed the role of NLRC4 in gastric epithelial cells during H. pylori infection. Our results show that the NLRC4 inflammasome modulates the immune response toward H. pylori and reveal a novel mechanism by which H. pylori through the NLRC4 inflammasome blocks the expression of BD-1, which would favor bacterial colonization.
Materials and Methods
Bacterial strains and culture conditions
The H. pylori strains PMSS1 and SS1 (38) were grown on Wilkins-Chalgren blood agar plates supplemented with 10% horse serum and Dent supplement (Oxoid) under microaerobic conditions (10% CO2, 5% O2, and 85% N2; 37°C).
Cultivation, isolation, and stimulation of cells
MKN7 [JCRB1025 (39)] human gastric cancer cells were used in this study. Short tandem repeat analysis was performed for cell authentication (40). Gastric epithelial cells were grown in DMEM/10% FCS/1% penicillin/streptomycin (Life Technologies). Murine bone marrow–derived dendritic cells (BMDCs) were generated from bone marrow as described (1) in RPMI 1640 medium containing 20 ng/ml GM-CSF (PeproTech), and supplemented with 10% FCS (Sigma-Aldrich), 1% penicillin/streptomycin (Life Technologies), and 50 μM 2-ME (Sigma-Aldrich). For isolation of primary gastric epithelial cells, stomachs were cut in small pieces and washed thoroughly with cold chelating buffer (5.6 mM Na2HPO4, 8 mM KH2PO4, 96.2 mM NaCl, 1.6 mM KCl, 43.4 mM sucrose, 54.9 mM d-sorbitol, and 0.5 mM dl-DTT [pH 7]). After incubation in chelating buffer containing 10 mM EDTA for 30 min at 4°C, glands were squeezed out with a glass slide (41). After disruption of the glands, cells were resuspended in Advanced DMEM/F12 supplemented with 10% FCS, 1% HEPES, 1% GlutaMAX, and 1% penicillin/streptomycin (Life Technologies) and plated on 10 μg/mm2 rat tail collagen I–coated cell culture plates. After overnight culture, media were exchanged to antibiotic-free media for infection or cytokine stimulation. Cells were incubated for 3 h or overnight with H. pylori at a multiplicity of infection (MOI) 10 or 50 or treated with recombinant human or murine IL-18 (R&D Systems). All cells were maintained at 37°C in a CO2 atmosphere.
Infection of mice with H. pylori
Six- to ten-wk-old male and female mice were used for in vivo experiments and four- to eight-wk-old ones for the isolation of primary cells. Nlrc4−/− mice (42) were kindly provided by Florian Greten (II. Medizinische Klinik und Poliklinik, Klinikum Rechts der Isar der Technische Universität München, München, Germany). Wild type littermates or cohoused offspring from littermates were used as controls. Age- and sex-matched mice were used as uninfected controls. Mice were bred under specific pathogen-free conditions and housed at the Institute for Medical Microbiology, Immunology and Hygiene (Technische Universität München). Experiments were conducted in compliance with European guidelines for the care and use of laboratory animals and were approved by the local authorities (Regierung von Oberbayern [AZ 55.2.1-54-2532-155-12]). Prior to infection, mice were fasted for 4 h. Mice were infected with H. pylori bacteria (2 × 108 CFU) thrice by oral gavage in Brucella broth containing 10% FCS every other day. Mice were sacrificed after 1 mo, and gastric tissue was collected for quantitative bacterial culture, flow cytometry, mRNA expression analysis, protein expression, and histologic examination. The same section of every stomach was assigned to the same downstream processing to reduce sampling error.
Assessment of bacterial colonization
For quantitative assessment of H. pylori colonization, gastric tissue was weighed and homogenized in Brucella broth containing 10% FCS. The number of CFU was determined by plating serial dilutions on Wilkins-Chalgren Dent agar plates, supplemented with bacitracin (200 μg/ml), nalidixic acid (10 μg/ml), and polymyxin B (3 μg/ml).
Preparation of gastric single-cell suspensions for flow cytometry
For isolation of gastric immune cells, stomach sections were digested in RPMI containing 10% FCS (Life Technologies), 1 mg/ml collagenase IV (Sigma-Aldrich), and 200 μg/ml DNase I (Roche) for 30 min at 37°C under shaking. For LIVE/DEAD discrimination, Ethidium Monoazide Bromide (Invitrogen) was used. Cells were stained with CD45-APCeF780, CD11b-eF450, CD11c-PECy7, and Ly6G-PE or with CD45-APC, CD3-PECy7, CD4-eF450, CD8-APCeF780, and TCR γδ-FITC Abs (Invitrogen, BioLegend). After the staining of cell surface markers, cells were fixed, permeabilized (Transcription Factor Staining Buffer Set; Invitrogen), and stained with a PE-conjugated Foxp3 Ab (Invitrogen). Data were acquired on a CyAn ADP nine color analyzer (Beckman Coulter). Flow cytometry data were analyzed using FlowJo (Tree Star). Singlet events (based on forward and side scatter) and viable (Ethidium Monoazide Bromide–negative) cells were selected prior to analysis.
Real-time quantitative PCR
Gastric tissue was homogenized, and total RNA was isolated using the RNeasy Mini Kit (QIAGEN) according to the manufacturer’s instructions, including on-column DNase digestion. RNA was retrotranscribed to cDNA using a combination of random hexamer primers and Moloney murine leukemia virus reverse transcriptase (Promega). Subsequently, cDNAs were amplified using a CFX384 quantitative PCR cycler (Bio-Rad Laboratories). The following primers were used: Tnfaip3 (A20) sense primer 5′-TGGTTCCAATTTTGCTCCTT-3′; Tnfaip3 antisense primer 5′-CGTTGATCAGAGTCGTG-3′; Defb1 sense primer 5′-TGAGCATAAAGGACGAGCGA-3′; Defb1 antisense primer 5′-CTCAGGACCAGGCAGATGTT-3′; Defb3 sense primer 5′-GTCTCCACCTGCAGCTTTTAG-3′; Defb3 antisense primer 5′-AGGAAAGGAACTCCACAACTGC-3′; Foxp3 sense primer 5′-AGGAGCCGCAAGCTAAAAGC-3′; Foxp3 antisense primer 5′-TGCCTTCGTGCCCACTGT-3′; Gapdh sense primer 5′-GCACAGTCAAGGCCGAGAAT-3′; Gapdh antisense primer 5′-GCCTTCTCCATGGTGGTGAA-3′; Ifng sense primer 5′-TCAAGTGGCATAGATGTGGAAGAA-3′; Ifng antisense primer 5′-TGGCTCTGCAGGATTTTCATG-3′; Il-10 sense primer 5′-CTAGAGCTGCGGACTGCCTTC-3′; Il-10 antisense primer 5′-CCTGCTCCACTGCCTTGCTCTTAT-3′; Il-17A sense primer 5′-GCTCCAGAAGGCCCTCAGA-3′; Il-17A antisense primer 5′-AGCTTTCCCTCCGCATTGA-3′; Il-18 sense primer 5′-ACTGTACAACCGCAGTAATAC-3′; Il-18 antisense primer 5′-AGTGAACATTACAGATTTATCCC-3′; Il-23 sense primer 5′-ATGCTGGATTGCAGAGCAGTA-3′; Il-23 antisense primer 5′-ACGGGGCACATTATTTTTAGTCT-3′, Nlrc4 sense primer 5′-GAAGAATCCTGTGATCTCCAAGAG-3′; Nlrc4 antisense primer 5′-GATCAAATTGTCAAGATTCTGTGC-3′; Tgfb sense primer 5′-ATCCTGTCCAAACTAAGGCTCG-3′; Tgfb antisense primer 5′-ACCTCTTTAGCATAGTAGTCCGC-3′; Tnfa sense primer 5′-CGATGGGTTGTACCTTGTC-3′; Tnfa antisense primer 5′-CGGACTCCGCAAAGTCTAAG-3′; GAPDH sense primer 5′-GAAGGTGAAGGTCGGAGT-3′; GAPDH antisense primer 5′-GAAGATGGTGATGGGATTTC-3′; TNFAIP3 sense primer 5′-TCCTCAGGCTTTGTATTTGAGC-3′; TNFAIP3 antisense primer 5′-TCTCCCGTATCTTCACAGCTT-3′; DEFB1 sense primer 5′-TTGTCTGAGATGGCCTCAGGTGGTAAC-3′; DEFB1 antisense primer 5′-ATCGGGCAGGCAGAATAGAGACA-3′; NLRC4 sense primer 5′-TGAACTGATCGACAGGATGAAC-3′; and NLRC4 antisense primer 5′-GTCTCCAGTTTTTCAACCCAAG-3′. Fold change was calculated using the 2−ΔΔCT method with GAPDH as an endogenous control for the normalization of target gene expression, and normalized to PMSS1-infected wild type mice. Gene products were separated on 2% agarose gels.
ELISA
Cell supernatants and stomach extracts were analyzed for the secreted cytokines IL-1β (Invitrogen) and IL-18 (R&D Systems) by ELISA according to the supplier’s instructions. To determine IL-1β and IL-18 protein levels in stomach extracts, stomach tissue was homogenized in PBS with proteinase inhibitors (Roche). After determining the whole protein content by Pierce BCA Protein Assay (Thermo Fisher Scientific), 10 μg of protein were used for performing IL-1β and IL-18 quantification.
Western blot
Cell lysates were obtained by lysing cells in 1× SDS buffer (1). Proteins from cell-free supernatants were extracted by TCA precipitation, resuspended in Laemmli buffer, and 5× SDS buffer. Equal amounts of lysates were loaded on SDS-PAGE gels, separated, and transferred to nitrocellulose or PVDF membranes. Membranes were incubated with the specific Ab according to the manufacturer’s instructions overnight. Abs against GAPDH, p-STAT3, STAT3, p-ERK, ERK, p-p38, p38, and NLRC4 (Cell Signaling Technology), IL-1β (R&D Systems), IL-18 (BioVision), and caspase-1 (AdipoGen Life Sciences) were used as primary Abs. HRP-coupled secondary Abs were from Promega. Quantification of band density was performed using LabImage 1D software (INTAS), and data were normalized to unstimulated cells.
Histologic examination of gastric tissue
For histologic evaluation H&E, chloroacetate esterase staining to detect polymorph neutrophils and mast cells (43, 44), and immunohistochemistry were performed. Pathology score was determined according to the updated Sydney score system (45) by a pathologist blinded for samples.
For immunohistochemistry, sections were dewaxed in xylene and gradually rehydrated. Heat-induced Ag retrieval was done in 0.01 M sodium citrate (pH 6). After blocking the slides in 5% goat serum for 1 h at room temperature, slides were incubated with primary Abs against CD3 (Thermo Fisher Scientific) or murine BD-1 (mBD-1) (Bioss Antibodies) overnight at 4°C following the manufacturer’s instructions. HRP-conjugated secondary Ab (Promega) was applied for 1 h, and samples were developed using SignalStain DAB Substrate (Cell Signaling Technology) and counterstained with hematoxylin. Six high power fields (20× magnification) were randomly scored for each sample using the virtual slide scanning system VS120 (Olympus).
Statistical analysis
Data are presented as mean ± SD or median. Data were analyzed by nonparametric tests (Mann–Whitney U test or Kruskal–Wallis test followed by uncorrected Dunn multiple comparisons test) or with parametric tests (Student t test or one-way ANOVA followed by Tukey multiple comparison test). GraphPad Prism 8 software (GraphPad) was used to determine statistical significance. Statistical significance was established when the p value was *p ≤ 0.05, **p < 0.01, or ***p < 0.001.
Results
NLRC4 inflammasome is essential for IL-18 maturation in gastric epithelial cells upon H. pylori infection
Epithelial NLRC4 and IL-18 have been described to be important players in intestinal immunity and inflammation. We hypothesized that NLRC4 expressed in gastric epithelial cells (Supplemental Fig. 1A) might also be an important modulator of the immune responses elicited by H. pylori. Therefore, we isolated primary gastric epithelial cells from wild type and Nlrc4−/− mice and infected them with the CagPAI-proficient H. pylori strain PMSS1. Murine gastric epithelial cells expressed pro–IL-18 in response to the infection regardless of their genotype (Fig. 1A). However, compared with noninfected cells, a higher secretion of mature IL-18 was observed in wild type cells (Fig. 1B), indicating that the NLRC4 inflammasome is required for IL-18 maturation in gastric epithelial cells. In contrast, neither the expression nor the maturation of IL-1β was induced in primary murine gastric epithelial cells (Supplemental Fig. 1B, 1C). Notably, a lack of NLRC4 in BMDCs did not affect the expression and maturation of IL-1β or IL-18 upon H. pylori infection (Supplemental Fig. 1D, 1E).
Production of IL-18 by gastric epithelial cells depends on the NLRC4 inflammasome and a functional T4SS. (A) Primary gastric epithelial cells were isolated from wild type or Nlrc4−/− mice and were infected with H. pylori PMSS1 at an MOI of 10. After overnight infection, IL-18p24 protein levels were detected by Western blot. GAPDH was used as a loading control. One representative blot of three independent experiments is shown. (B) IL-18 was measured by ELISA in the supernatants from primary gastric epithelial cells treated as indicated in (A). Results are presented as mean ± SD from six independent experiments. (C) MKN7 cells were infected with PMSS1 or SS1 at an MOI of 10 overnight, and IL-18 was quantified by ELISA in the culture supernatants. Results are shown as mean ± SD from three independent experiments. (D) Relative Il-18 mRNA levels determined by quantitative PCR and (E) IL-18 protein levels measured in gastric extracts by ELISA in wild type or Nlrc4−/− mice infected with H. pylori PMSS1 for 1 mo. Horizontal lines indicate the median. Data are pooled from up to three independent experiments with 3–10 mice per group. (F) Protein levels of IL-18p24, IL-18p18, caspase-1p45, and caspase-1p20 in gastric extracts detected by Western blot. GAPDH was used as a loading control. (G) Wild type mice were infected with H. pylori PMSS1 or SS1 for 1 mo. IL-18 protein levels were measured in gastric extracts by ELISA. Horizontal lines indicate the median. Data from one experiment with four to six mice per group are shown. (H) IL-18p24, IL-18p18, caspase-1p45, and caspase-1p20 protein levels were detected by Western blot in gastric extracts. GAPDH was used as a loading control. (B and C) ANOVA followed by Tukey multiple comparison test; (D, E, and G) Kruskal–Wallis test followed by Dunn test. *p ≤ 0.05, **p < 0.01, ***p < 0.001.
Production of IL-18 by gastric epithelial cells depends on the NLRC4 inflammasome and a functional T4SS. (A) Primary gastric epithelial cells were isolated from wild type or Nlrc4−/− mice and were infected with H. pylori PMSS1 at an MOI of 10. After overnight infection, IL-18p24 protein levels were detected by Western blot. GAPDH was used as a loading control. One representative blot of three independent experiments is shown. (B) IL-18 was measured by ELISA in the supernatants from primary gastric epithelial cells treated as indicated in (A). Results are presented as mean ± SD from six independent experiments. (C) MKN7 cells were infected with PMSS1 or SS1 at an MOI of 10 overnight, and IL-18 was quantified by ELISA in the culture supernatants. Results are shown as mean ± SD from three independent experiments. (D) Relative Il-18 mRNA levels determined by quantitative PCR and (E) IL-18 protein levels measured in gastric extracts by ELISA in wild type or Nlrc4−/− mice infected with H. pylori PMSS1 for 1 mo. Horizontal lines indicate the median. Data are pooled from up to three independent experiments with 3–10 mice per group. (F) Protein levels of IL-18p24, IL-18p18, caspase-1p45, and caspase-1p20 in gastric extracts detected by Western blot. GAPDH was used as a loading control. (G) Wild type mice were infected with H. pylori PMSS1 or SS1 for 1 mo. IL-18 protein levels were measured in gastric extracts by ELISA. Horizontal lines indicate the median. Data from one experiment with four to six mice per group are shown. (H) IL-18p24, IL-18p18, caspase-1p45, and caspase-1p20 protein levels were detected by Western blot in gastric extracts. GAPDH was used as a loading control. (B and C) ANOVA followed by Tukey multiple comparison test; (D, E, and G) Kruskal–Wallis test followed by Dunn test. *p ≤ 0.05, **p < 0.01, ***p < 0.001.
The induction of IL-18 expression upon H. pylori infection was also evaluated in human gastric epithelial cells, which also express NLRC4 (Supplemental Fig. 1F, 1G). Thus, MKN-7 cells were infected with different H. pylori strains, and levels of mature IL-18 were determined by ELISA. MKN-7 cells secreted high levels of IL-18 when infected with H. pylori PMSS1 (Fig. 1C). Interestingly, the mouse-adapted SS1 strain, which lacks a functional T4SS, induced IL-18 to a much lesser extent (Fig. 1C). As observed in murine gastric epithelial cells, IL-1β was not produced by human epithelial cells in response to the infection (Supplemental Fig. 1H).
We further analyzed the expression of IL-18 in response to H. pylori infection in the stomach of wild type and Nlrc4−/− mice 1 mo postinfection. We observed similar mRNA levels of Il-18 and pro–IL-18 protein levels in uninfected animals and after 1-mo H. pylori infection in wild type and Nlrc4−/− mice (Fig. 1D, 1F), whereas the secretion of mature IL-18 was reduced in Nlrc4−/− mice (Fig. 1E, 1F). No caspase-1 activation was observed in infected Nlrc4−/− mice (Fig. 1F). Interestingly, the SS1 strain, which lacks a functional T4SS, induced IL-18 and caspase-1 activation to a lesser extent than the PMSS1 strain also in vivo (Fig. 1G, 1H). These results indicate that the NLRC4 inflammasome is important for IL-18 maturation induced by H. pylori infection and that a functional T4SS is required.
Reduced neutrophil recruitment is observed in the stomach of H. pylori–infected Nlrc4−/− mice
Given that the NLRC4 inflammasome is required for epithelial IL-18 secretion, we next analyzed how a lack of NLRC4 would affect the immune response to H. pylori infection. After 1-mo infection, we observed lower bacterial load in the stomach of Nlrc4 knockout mice (Fig. 2A). Unexpectedly, Nlrc4−/− mice also showed lower gastric inflammation when compared with wild type H. pylori–infected animals (Fig. 2B), indicating that the NLRC4 inflammasome plays a role in the immune response elicited by H. pylori. This was corroborated by the lower percentage of infiltrating leukocytes (CD45+ cells) observed in mice lacking NLRC4 (Fig. 2C). We next examined the local gastric inflammatory response to H. pylori, focusing first on the infiltration of innate immune cells, macrophages, neutrophils, and dendritic cells to the stomach of the infected mice. No differences were detected in macrophage infiltration between wild type and Nlrc4−/− mice (Supplemental Fig. 2), whereas lower neutrophil infiltration was observed in the gastric tissue of H. pylori–infected Nlrc4−/− animals (Fig. 2D, 2E), which could relate to the lower inflammatory score observed in Nlrc4−/− mice. In addition, less infiltration of myeloid dendritic cells was detected in the stomach of Nlrc4−/− mice (Fig. 2F).
NLRC4 influences H. pylori gastric colonization and neutrophil infiltration. (A) Wild type and Nlrc4−/− mice were infected with H. pylori PMSS1 and sacrificed after 1 mo. Colonization level per milligram stomach was determined by plating serial dilutions of stomach extracts. CFU, colony forming units. (B) Inflammatory score assessed in the stomach of uninfected and H. pylori–infected mice. (C) Flow cytometry analysis of leukocyte population (CD45+ of live cells) in the gastric mucosa. A representative plot of each group is depicted. (D) Flow cytometry analysis of neutrophils (CD45+/Ly6G+ of live cells) in the stomach of infected and naive mice. A representative plot of each group is depicted. (E) Numbers of neutrophils (and mast cells) (chloroacetate esterase [CAE] positive cells per high power field [HPF]) infiltrating the stomach of uninfected and H. pylori–infected mice. Representative images are shown. (F) Flow cytometry analysis of dendritic cells (CD45+/CD11b+/CD11c+ of live cells) in the gastric mucosa. A representative plot of each group is depicted. Data are pooled from up to four (A–E) or three (F) independent experiments with 3–10 mice per group. Horizontal lines indicate the median. (A) Mann–Whitney U test; (B–F) Kruskal–Wallis test followed by Dunn test. *p ≤ 0.05, **p < 0.01, ***p < 0.001.
NLRC4 influences H. pylori gastric colonization and neutrophil infiltration. (A) Wild type and Nlrc4−/− mice were infected with H. pylori PMSS1 and sacrificed after 1 mo. Colonization level per milligram stomach was determined by plating serial dilutions of stomach extracts. CFU, colony forming units. (B) Inflammatory score assessed in the stomach of uninfected and H. pylori–infected mice. (C) Flow cytometry analysis of leukocyte population (CD45+ of live cells) in the gastric mucosa. A representative plot of each group is depicted. (D) Flow cytometry analysis of neutrophils (CD45+/Ly6G+ of live cells) in the stomach of infected and naive mice. A representative plot of each group is depicted. (E) Numbers of neutrophils (and mast cells) (chloroacetate esterase [CAE] positive cells per high power field [HPF]) infiltrating the stomach of uninfected and H. pylori–infected mice. Representative images are shown. (F) Flow cytometry analysis of dendritic cells (CD45+/CD11b+/CD11c+ of live cells) in the gastric mucosa. A representative plot of each group is depicted. Data are pooled from up to four (A–E) or three (F) independent experiments with 3–10 mice per group. Horizontal lines indicate the median. (A) Mann–Whitney U test; (B–F) Kruskal–Wallis test followed by Dunn test. *p ≤ 0.05, **p < 0.01, ***p < 0.001.
Th17 immune response is enhanced in H. pylori–infected Nlrc4−/− mice
Our results suggested that the lower inflammation detected in Nlrc4−/− infected mice could be attributed to reduced infiltration of neutrophils to the stomach; however, this observation could not explain the concomitant reduced bacterial burden observed in mice lacking the NLRC4 inflammasome. The cellular adaptive immune response is considered to be crucial for clearing of the infection. Therefore, we characterized by flow cytometry T cell subsets infiltrating the stomach of the mice to identify the cells responsible for the reduced colonization seen in Nlrc4−/− infected mice. Although the percentage of CD3+ infiltrating cells was similar (Fig. 3A, 3B), an increased percentage of infiltrating CD4+ cells was detected in the gastric tissue of Nlrc4−/− H. pylori–infected mice, whereas CD8+ cell infiltrates were significantly lower when compared with wild type infected mice (Fig. 3C, Supplemental Fig. 3A). Because it had been suggested that H. pylori–induced IL-18 promotes the establishment of a regulatory T cell (Treg) response (10), we analyzed the expression of Il-10 as a marker for Tregs. However, we could not detect differences in its expression between H. pylori–infected Nlrc4−/− and wild type mice (Fig. 3D). In addition, Tgfb expression, which is also a marker for Tregs, was higher in mice lacking NLRC4 (Fig. 3E), whereas Foxp3 expression and the percentage of infiltrating CD4+Foxp3+ cells did not differ between the genotypes (Supplemental Fig. 3B, 3C). This indicates that wild type and Nlrc4−/− mice show similar Treg responses upon H. pylori infection. We next analyzed Ifng and Il-17 mRNA levels as surrogates for Th1 and Th17 responses, respectively. Although we could not detect differences in Ifng or Tnfa expression, typical markers for Th1 cells (Fig. 3D, 3E), we did observe increased levels of Il-17 in the absence of NLRC4 (Fig. 3D). This was accompanied by increased levels of Tgfb and a trend of higher Il-23 levels, both linked to the induction and maintenance of Th17 cells (46, 47) (Fig. 3E). Percentage of the IL-17–producing cell subset, γδ T cells, did not differ between H. pylori–infected wild type and Nlrc4−/− mice (Supplemental Fig. 3D).
The NLRC4 inflammasome dampens Th17 inflammatory responses. Wild type and Nlrc4−/− mice were infected with H. pylori PMSS1 and sacrificed after 1 mo. (A) Flow cytometry analysis of T cell population (CD45+/CD3+ of live cells) in the gastric mucosa. A representative plot of each group is depicted. (B) Number of CD3+ cells per high power field. Representative images of CD3 expression in the stomach mucosa from each group are shown. (C) Flow cytometry analysis of CD4+ and CD8+ T cell infiltration. A representative plot of each group is depicted. Frequencies of CD45+/CD3+/CD4+ and CD45+/CD3+/CD8+ live cells are shown. (D and E) mRNA levels of Il-10, Ifng, Il-17, Tgfb, Tnfa, and Il-23, cytokines associated with T helper responses, were analyzed by quantitative PCR. Horizontal lines indicate the median. Data are pooled from up to four independent experiments with 3–10 mice per group. (A–E) Kruskal–Wallis test followed by Dunn test. *p ≤ 0.05, **p < 0.01, ***p < 0.001.
The NLRC4 inflammasome dampens Th17 inflammatory responses. Wild type and Nlrc4−/− mice were infected with H. pylori PMSS1 and sacrificed after 1 mo. (A) Flow cytometry analysis of T cell population (CD45+/CD3+ of live cells) in the gastric mucosa. A representative plot of each group is depicted. (B) Number of CD3+ cells per high power field. Representative images of CD3 expression in the stomach mucosa from each group are shown. (C) Flow cytometry analysis of CD4+ and CD8+ T cell infiltration. A representative plot of each group is depicted. Frequencies of CD45+/CD3+/CD4+ and CD45+/CD3+/CD8+ live cells are shown. (D and E) mRNA levels of Il-10, Ifng, Il-17, Tgfb, Tnfa, and Il-23, cytokines associated with T helper responses, were analyzed by quantitative PCR. Horizontal lines indicate the median. Data are pooled from up to four independent experiments with 3–10 mice per group. (A–E) Kruskal–Wallis test followed by Dunn test. *p ≤ 0.05, **p < 0.01, ***p < 0.001.
Together, these results indicate that a lack of NLRC4 inflammasome favors Th17 responses, which could favor bacterial clearance.
Nlrc4−/− mice express BD-1 in response to H. pylori infection
Because the NLRC4 inflammasome is involved in the regulation of antimicrobial peptides such as defensins (48, 49), we hypothesized that a lack of NLRC4 in H. pylori–infected mice would alter the secretion of BDs, thus reducing inflammation and favoring bacterial clearance. We first of all studied whether the expression of mBD-1 and murine BD-3 was altered by H. pylori infection in mice because their human orthologs (50, 51) were previously reported to be differentially expressed in response to H. pylori infection (27, 32). In wild type mice expressing Nlrc4, H. pylori infection did not induce the expression of Defb1 (Fig. 4A), whereas upregulation of Defb3 expression was detected in infected mice (Fig. 4B). Interestingly, these effects were dependent on the presence of a functional T4SS in the bacterium because infection with H. pylori SS1 led to the upregulation of Defb1expression (Fig. 4A), whereas no changes in the expression of Defb3 were detected compared with uninfected control mice (Fig. 4B). In the absence of NLRC4, we observed that Defb1 expression was increased upon H. pylori infection (Fig. 4C, 4E), whereas no differences in Defb3 expression were detected between infected wild type and Nlrc4−/− mice (Fig. 4D). We excluded a direct effect of the genetic ablation of Nlrc4 on defensin expression because uninfected wild type and Nlrc4−/− mice showed similar levels of Defb1 and Defb3 in the stomach (Fig. 4C, 4D). Interestingly, similar levels of Defb1 and Defb3 were detected in the stomach of SS1 infected wild type and Nlrc4−/− mice (Fig. 4C, 4D). Increased expression of mBD-1 upon H. pylori infection was confirmed in gastric tissue samples of infected Nlrc4−/− mice (Fig. 4E).
BD-1 expression is regulated by NLRC4 and a functional T4SS. Wild type mice were infected with H. pylori PMSS1 or SS1 and sacrificed after 1 mo. (A and B) mRNA levels of Defb1 and Defb3 were analyzed by quantitative PCR. Data are pooled from one (SS1), two (naive), or four (PMSS1) independent experiments with 3–10 mice per group. Horizontal lines indicate the median. (C and D) mRNA levels of Defb1 and Defb3 were analyzed by quantitative PCR in stomach tissue samples from uninfected and 1-mo H. pylori PMSS1- or SS1-infected wild type and Nlrc4−/− mice. Data are pooled from up to four independent experiments with 3–10 mice per group. Horizontal lines indicate the median. Wild type mice as shown in (A). (E) Representative images of mBD-1 expression detected in the stomach mucosa by immunohistochemistry. (F) Murine primary gastric epithelial cells were infected overnight with PMSS1 or SS1 at an MOI of 10. mRNA levels of Defb1 were analyzed by quantitative PCR. Results are shown as mean ± SD from three independent experiments. (G) MKN7 cells were infected overnight with PMSS1 or SS1 at an MOI of 10. mRNA levels of DEFB1 were analyzed by quantitative PCR. Results are shown as mean ± SD from four independent experiments. (A–D) Kruskal–Wallis test followed by Dunn multiple comparisons test; (F and G) ANOVA followed by Tukey multiple comparison test. *p ≤ 0.05, **p < 0.01, ***p < 0.001.
BD-1 expression is regulated by NLRC4 and a functional T4SS. Wild type mice were infected with H. pylori PMSS1 or SS1 and sacrificed after 1 mo. (A and B) mRNA levels of Defb1 and Defb3 were analyzed by quantitative PCR. Data are pooled from one (SS1), two (naive), or four (PMSS1) independent experiments with 3–10 mice per group. Horizontal lines indicate the median. (C and D) mRNA levels of Defb1 and Defb3 were analyzed by quantitative PCR in stomach tissue samples from uninfected and 1-mo H. pylori PMSS1- or SS1-infected wild type and Nlrc4−/− mice. Data are pooled from up to four independent experiments with 3–10 mice per group. Horizontal lines indicate the median. Wild type mice as shown in (A). (E) Representative images of mBD-1 expression detected in the stomach mucosa by immunohistochemistry. (F) Murine primary gastric epithelial cells were infected overnight with PMSS1 or SS1 at an MOI of 10. mRNA levels of Defb1 were analyzed by quantitative PCR. Results are shown as mean ± SD from three independent experiments. (G) MKN7 cells were infected overnight with PMSS1 or SS1 at an MOI of 10. mRNA levels of DEFB1 were analyzed by quantitative PCR. Results are shown as mean ± SD from four independent experiments. (A–D) Kruskal–Wallis test followed by Dunn multiple comparisons test; (F and G) ANOVA followed by Tukey multiple comparison test. *p ≤ 0.05, **p < 0.01, ***p < 0.001.
We further analyzed Defb1expression in primary gastric epithelial cells. Similar to our observations in vivo, murine gastric epithelial cells infected with SS1 showed higher levels of Defb1 than PMSS1-infected cell (Fig. 4F). When human gastric epithelial cells were infected with H. pylori PMSS1, a downregulation of human DEFB1 levels could be observed (Fig. 4G). In contrast, infection of the cells with H. pylori SS1 led to increased expression of DEFB1 (Fig. 4G).
Together, these results indicate that H. pylori T4SS blocks the expression of BD-1, and this is mediated by the NLRC4 inflammasome.
IL-18 represses BD-1 expression via NF-κB signaling
Because the loss of NLRC4 expression in murine gastric epithelial cells had a major impact on the production of IL-18 in response to H. pylori, we hypothesized that IL-18 might be involved in the regulation of BD-1 expression. To explore this, we treated gastric epithelial cells with rIL-18. After IL-18 treatment, we observed reduced mRNA levels of Defb1 in murine primary gastric epithelial cells (Fig. 5A). To further confirm that epithelial IL-18 is responsible for the observed downregulation of BD-1 expression, we stimulated PMSS1-infected primary gastric epithelial cells from Nlrc4−/− mice with rIL-18. As expected, infection of the cells with H. pylori PMSS1, which does not result in IL-18 production (Fig. 1B), led to increased expression of Defb1 (Fig. 5B). Notably, concomitant IL-18 treatment blocked the upregulation of Defb1 expression induced by H. pylori. These results confirmed that IL-18 is the NLRC4 inflammasome–regulated cytokine inhibiting BD-1 expression during H. pylori infection.
IL-18 blocks BD-1 by NF-κB activation in epithelial cells. (A) Wild type murine gastric primary epithelial cells were stimulated with rIL-18 (2.5 ng/ml), and mRNA levels of Defb1 were analyzed by quantitative PCR after overnight stimulation. Results are shown as mean ± SD from three independent experiments. (B) Primary gastric epithelial cells from Nlrc4−/− mice were infected overnight with PMSS1 at an MOI of 10 and stimulated with rIL-18 (2.5 ng/ml). mRNA levels of Defb1 were analyzed by quantitative PCR. Results are shown as mean ± SD from three independent experiments. (C) Wild type murine gastric primary epithelial cells were stimulated as in (A), and mRNA levels of Tnfaip3 were analyzed by quantitative PCR. Results are shown as mean ± SD from three independent experiments. MKN7 cells were stimulated with rIL-18 (1 ng/ml) overnight. mRNA levels of DEFB1 (D) and TNFAIP3 (E) were analyzed by quantitative PCR. Results are shown as mean ± SD from three independent experiments. (A, C–E) Student t test; (B) ANOVA followed by Tukey multiple comparison test. *p ≤ 0.05, **p < 0.01, ***p < 0.001.
IL-18 blocks BD-1 by NF-κB activation in epithelial cells. (A) Wild type murine gastric primary epithelial cells were stimulated with rIL-18 (2.5 ng/ml), and mRNA levels of Defb1 were analyzed by quantitative PCR after overnight stimulation. Results are shown as mean ± SD from three independent experiments. (B) Primary gastric epithelial cells from Nlrc4−/− mice were infected overnight with PMSS1 at an MOI of 10 and stimulated with rIL-18 (2.5 ng/ml). mRNA levels of Defb1 were analyzed by quantitative PCR. Results are shown as mean ± SD from three independent experiments. (C) Wild type murine gastric primary epithelial cells were stimulated as in (A), and mRNA levels of Tnfaip3 were analyzed by quantitative PCR. Results are shown as mean ± SD from three independent experiments. MKN7 cells were stimulated with rIL-18 (1 ng/ml) overnight. mRNA levels of DEFB1 (D) and TNFAIP3 (E) were analyzed by quantitative PCR. Results are shown as mean ± SD from three independent experiments. (A, C–E) Student t test; (B) ANOVA followed by Tukey multiple comparison test. *p ≤ 0.05, **p < 0.01, ***p < 0.001.
IL-18 has been described to mainly signal via activation of NF-κB signaling. Interestingly, NF-κB was shown to downregulate hBD-1 expression upon H. pylori infection (27). We then analyzed the levels of Tnfaip3 as a surrogate marker for the activation of canonical NF-κB upon treatment with rIL-18, and we observed that indeed treatment with IL-18 led to the activation of NF-κB (Fig. 5C).
Next, we sought to determine whether this effect was also observed in human gastric epithelial cells. Indeed, IL-18 treatment reduced the expression of DEFB1 in MKN7 cells (Fig. 5D). This effect was also mediated by NF-κB activation (Fig. 5E). In contrast, other downstream signaling pathways reported to be induced by IL-18 such as STAT3 or MAPKs p38 and ERK were not altered by IL-18 treatment in gastric epithelial cells (Supplemental Fig. 4).
In summary, our results indicate that IL-18 secreted in response to H. pylori infection represses BD-1 expression via the activation of NF-κB signaling in gastric epithelial cells. This lower BD-1 production leads to attenuated clearing of H. pylori. Further, T4SS-dependent NLRC4 inflammasome activation and IL-18 production contribute to inflammation by recruitment of neutrophils into the stomach (Fig. 6).
Model of NLRC4 inflammasome–dependent immune regulation. (A) H. pylori T4SS–mediated activation of the NLRC4 inflammasome induces the cleavage of pro–IL-18 and the secretion of mature IL-18 from gastric epithelial cells. Secreted IL-18 contributes to the recruitment of neutrophils, resulting in gastric inflammation. Further, IL-18 inhibits Th17/IL-17 response, thereby inducing BD-1 expression. Additionally, IL-18 activates NF-κB, resulting in the downregulation of BD-1. Finally, reduced BD-1 production leads to attenuated killing of H. pylori. (B) In absence of NLRC4, no mature IL-18 is produced. This results in lower infiltration of inflammatory neutrophils, attenuation of IL-18–mediated inhibitory pathways, stronger Th17/IL-17 response, and higher BD-1 expression. High BD-1 production results in enhanced clearing of H. pylori.
Model of NLRC4 inflammasome–dependent immune regulation. (A) H. pylori T4SS–mediated activation of the NLRC4 inflammasome induces the cleavage of pro–IL-18 and the secretion of mature IL-18 from gastric epithelial cells. Secreted IL-18 contributes to the recruitment of neutrophils, resulting in gastric inflammation. Further, IL-18 inhibits Th17/IL-17 response, thereby inducing BD-1 expression. Additionally, IL-18 activates NF-κB, resulting in the downregulation of BD-1. Finally, reduced BD-1 production leads to attenuated killing of H. pylori. (B) In absence of NLRC4, no mature IL-18 is produced. This results in lower infiltration of inflammatory neutrophils, attenuation of IL-18–mediated inhibitory pathways, stronger Th17/IL-17 response, and higher BD-1 expression. High BD-1 production results in enhanced clearing of H. pylori.
Discussion
Different studies have demonstrated an important role for the NLRP3 inflammasome in the modulation of the immune response against H. pylori infection (1, 4–8). In contrast, a possible role for the NLRC4 inflammasome was unclear. Thus, studies in macrophages showed that although H. pylori flagellin FlaA induced the phosphorylation of NLRC4, it did not lead to its activation (52). We also observed that the lack of NLRC4 in BMDCs did not have any impact on IL-1β maturation in response to H. pylori infection (1). However, and considering the roles of IL-18 and NLRC4 in intestinal homeostasis (12, 13, 53), we sought to determine whether the NLRC4 inflammasome was involved in epithelial responses to H. pylori. We have observed that gastric epithelial cells constitutively express NLRC4, and it is essential for the generation of mature IL-18 in response to H. pylori. Consequently, the lack of NLRC4 during H. pylori infection has a major impact on the immune response. In addition, Nlrc4−/− mice have a reduced bacterial load in the stomach compared with wild type mice. This observation is in line with previous reports showing lower H. pylori gastric colonization in IL-18–deficient mice (9, 10), and suggests that the induction of IL-18 expression, which is increased in the gastric mucosa of infected individuals (2, 3) and which correlates with infiltration of immune cells (2, 54), supports bacterial persistence. When analyzing the possible mechanisms by which IL-18 could favor bacterial colonization, we observed that mice lacking NLRC4 presented higher levels of mBD-1, which could contribute to enhanced clearance of the bacterium observed in these mice. BDs are an important group of antimicrobial peptides involved in host epithelial defense against H. pylori. hBD-2, hBD-3, and hBD-4 have antimicrobial activity against H. pylori, and particularly the expression of hBD-2 and hBD-4 is regulated by NF-κB in a CagPAI-dependent manner upon infection (26–28, 31, 32, 36, 37, 55, 56). Interestingly, the expression of hBD-3 was found to be downregulated during prolonged infection (35). Likewise, the expression of hBD-1, which is constitutively expressed in the stomach (57, 58) and is able to kill H. pylori (29, 55), was reported to be downregulated by H. pylori through the activation of NF-κB. This effect was only detected when H. pylori strains presented an intact CagPAI (27), highlighting the importance of this virulence factor for bacterial persistence. Our results further link the downregulation of BD-1 by H. pylori to the immune response elicited by the bacterium. Thus, IL-18 downregulates the expression of hBD-1 and mBD-1 via the activation of NF-κB. This also requires the T4SS because bacteria lacking a functional T4SS are unable to downregulate the expression of BD-1 in vitro as well as in vivo. These results support an important role for the NLRC4 inflammasome in bacterial immune evasion. By manipulating NLRC4, H. pylori induces the production of IL-18 and thereby NF-κB–mediated downregulation of BD-1, hampering bacterial clearance and favoring persistence. At the same time, the NLRC4 inflammasome has important immunomodulatory effects in the course of H. pylori infection (Fig. 6). In this regard, the lack of NLRC4 leads to an increased Th17 immune response, which is in agreement with previous observations in IL-18–deficient mice (9, 10). However, the significance of Th17 responses during H. pylori infection is somewhat controversial. Whereas some reports suggested that the Th17 response elicited by H. pylori is important for bacterial clearance (59, 60), other studies correlated a lack of IL-17 with reduced H. pylori gastric colonization (61–63). Our results as well as previous studies addressing the role of IL-18 in H. pylori infection (9, 10) support the involvement of H. pylori–driven IL-17 response in bacterial clearance. Interestingly, IL-17 was found to be an important regulator of antimicrobial peptides in the respiratory epithelium (48, 64). Therefore, it cannot be ruled out that in the context of H. pylori, IL-17 might have similar effects on the expression of antimicrobial peptides, thereby controlling bacterial burden. Indeed, a previous study showed that IL-17 in combination with IL-22 contributes to epithelial cell defense against H. pylori by inducing the expression of antimicrobials such as BD-1 (65).
In addition to regulating Th17 responses, IL-18 was reported to be involved in neutrophil recruitment and activation in different tissues (66–68). Supporting this role for IL-18, we observed lower infiltration of neutrophils in Nlrc4−/− mice compared with wild type mice. This was concomitant to reduced inflammation in the stomach. Recruitment and activation of neutrophils are key processes in H. pylori–induced inflammation. In fact, the presence of neutrophils characterizes the activity and severity of gastritis (45). We postulate that IL-18–mediated neutrophil recruitment is an additional mechanism by which the NLRC4 contributes to gastric inflammation during H. pylori infection.
Together, our results provide evidence of an important role of the epithelial NLRC4 inflammasome in H. pylori infection. Thus, by controlling the maturation of IL-18, NLRC4 reduces the host’s immune responses, contributing to bacterial colonization and subsequent gastric inflammation.
Acknowledgements
We thank Stefanie Wüstner and Martin Skerhut for assistance with mouse experiments and Christine Josenhans for providing primers.
Footnotes
This work was supported by a research scholarship from Elite Network of Bavaria to R.P.S., who initiated this work as a Ph.D. student.
The online version of this article contains supplemental material.
References
Disclosures
The authors have no financial conflicts of interest.