Host response to Helicobacter pylori infection before initiation of the adaptive immune response

Abstract

Helicobacter pylori persistently colonizes the human stomach. In this study, immune responses to H. pylori that occur in the early stages of infection were investigated. Within the first 2 days after orogastric infection of mice with H. pylori, there was a transient infiltration of macrophages and neutrophils into the glandular stomach. By day 10 postinfection, the numbers of macrophages and neutrophils decreased to baseline levels. By 3 weeks postinfection, an adaptive immune response was detected, marked by gastric infiltration of T lymphocytes, macrophages, and neutrophils, as well as increased numbers of H. pylori-specific T cells, macrophages, and dendritic cells in paragastric lymph nodes. Neutrophil-attracting and macrophage-attracting chemokines were expressed at higher levels in the stomachs of H. pylori-infected mice than in the stomachs of uninfected mice. Increased expression of TNFα and IFNγ (Th1-type inflammatory cytokines) and IL-17 (a Th17-type cytokine) was detected in the stomachs of H. pylori-infected mice, but increased expression of IL-4 (a Th2-type cytokine) was not detected. These data indicate that a transient gastric inflammatory response to H. pylori occurs within the first few days after infection, before the priming of T cells and initiation of an adaptive immune response. It is speculated that inappropriate waning of the innate immune response during early stages of infection may be a factor that contributes to H. pylori persistence.

Helicobacter, innate, adaptive, T lymphocytes, cytokines, chemokines

Introduction

Helicobacter pylori is a spiral-shaped Gram-negative bacterium that persistently colonizes the human stomach. Helicobacter pylori infection is a risk factor for development of peptic ulcer disease, gastric adenocarcinoma, and gastric mucosa-associated lymphoid tissue (MALT) lymphoma (Peterson, 1991; Peek & Blaser, 2002; Suerbaum & Michetti, 2002). Helicobacter pylori is typically acquired in childhood and can persist for many decades despite the development of an adaptive immune response (Algood & Cover, 2006). Helicobacter pylori-specific antibodies are detectable in human sera and gastric juice, and H. pylori-specific T cells are detectable in the stomachs of H. pylori-infected persons (Bamford et al., 1998; Mattsson et al., 1998; Perez-Perez et al., 1999). Infiltration of the gastric mucosa with neutrophils, macrophages, and lymphocytes (termed superficial gastritis) is a hallmark of H. pylori infection (Agnihotri et al., 1998; Bamford et al., 1998; Sommer et al., 1998; Quiding-Jarbrink et al., 2001; Goll et al., 2005).

Numerous studies have characterized the immune response to H. pylori in humans who are persistently infected with this organism. However, relatively little is known about host responses to H. pylori that occur during the very early stages of infection. In particular, very little is known about innate immune responses to H. pylori that occur before development of an adaptive immune response. The goal of the current study was to analyze host responses to H. pylori that occur immediately following infection in a mouse model. Specifically, the objectives were to detect gastric mucosal inflammatory responses mediated by innate immune defenses, and to identify potential sites of T cell priming that might be important for initiation of an adaptive immune response.

Materials and methods

Culture of H. pylori

A mouse-passaged derivative of H. pylori strain SS1 was used in all experiments. Bacteria were grown on trypticase soy agar (TSA) plates containing 5% sheep blood. Alternatively, bacteria were grown in Brucella broth containing 5% heat-inactivated fetal bovine serum (FBS) and 10 µg mL⁻¹ vancomycin. Cultures were grown at 37^oC in either room air supplemented with 5% CO₂, or under microaerobic conditions generated by a CampyPak Plus^* Hydrogen+CO₂ with an integral palladium catalyst (BD Biosciences, Santa Cruz, CA).

Infection of mice with H. pylori

Helicobacter-free male FVB/N mice (Taconic, Germantown, NY), 8 weeks old, were used in all experiments. The Vanderbilt University Institutional Animal Care and Use Committee approved all the animal protocols used in this study. Sentinel mice housed in the same room tested negative for Helicobacter infection, pinworms, mouse parvovirus, and several other murine pathogens. One day before infection of mice, H. pylori were inoculated into liquid medium and were cultured for 18 h under microaerobic conditions, as described above. Fasting mice were orogastrically inoculated with a suspension of 10⁸ CFU H. pylori (in 0.5 mL of Brucella broth) or with Brucella broth alone as control.

Processing of mouse stomachs

The stomach was removed from each mouse by excising between the esophagus and the duodenum. The paragastric lymph nodes (located on the lesser curvature of the glandular stomach) were dissected away from the gastric wall and placed in RPMI medium. The forestomach (nonglandular portion) was removed from the glandular stomach and discarded. The glandular stomach was opened, rinsed gently in phosphate-buffered saline (PBS), and cut longitudinally. For experiments performed ≥3 days postinfection, one-half of the stomach was placed into RPMI medium, one-fourth of the stomach was placed into 0.5 mL RNALater (Ambion, Austin, TX) and stored at −20 °C, and the remaining fourth was placed into 1 mL Brucella broth containing 10% FBS for H. pylori culture. For experiments performed at days 1 and 2 postinfection, the whole stomach was used for flow cytometric analysis.

Culture of H. pylori from mouse stomach

The portion of the stomach in Brucella broth-10% FBS was homogenized using a disposable pestle (Kimble–Kontes, Vineland, NJ). Serial dilutions of the homogenate were plated on TSA plates containing 5% sheep blood, 10 µg mL⁻¹ nalidixic acid, 100 µg mL⁻¹ vancomycin, 2 µg mL⁻¹ amphotericin, 200 µg mL⁻¹ bacitracin, and 2500 U mL⁻¹ polymyxin. After 5–7 days of culture under microaerobic conditions, H. pylori colonies were counted.

RNA extraction and real-time reverse transcriptase (RT)-PCR

RNA was isolated from the stomach using the Trizol isolation protocol (Invitrogen, Carlsbad, CA) with slight modifications. The stomach tissue was homogenized in 1 mL of Trizol reagent and then two chloroform extractions were performed. Following an isopropanol precipitation, the RNA was washed with 70% ethanol and treated with RNAse inhibitor (Applied Biosystems, Foster City, CA) for 45 min. Following resuspension of the RNA at 65 °C for 15 min, the RNA preparations were further purified using the Qiagen RNA isolation kit and RNAse-free DNAse treatment as directed by the manufacturer (Qiagen Inc., Valencia, CA).

The RNA was reverse transcribed using the iScript cDNA kit (Biorad, Hercules, CA). For real-time RT-PCR, the relative gene expression method was used (Giulietti et al., 2001). Hypoxanthine guanine phosphoribosyl transferase (HPRT) served as the normalizer, and uninfected stomach as the calibrator. All samples were analyzed in triplicate, along with ‘no reverse transcriptase’ controls, using a MyiQ Single-Color Real-Time PCR Detection System (BioRad). Levels of cytokine expression are indicated as ‘relative units,’ based on comparison of tissue from H. pylori-infected mice with tissue from uninfected mice (calibrator tissue) (Giulietti et al., 2001). Each primer was used at a final concentration of 400 nM. The primer sequences were as follows: IFNγ (forward) 5′-TCAAGTGGCATAGATGTGGAAGAA-3′, IFNγ (reverse) 5′-TGGCTCTGCAGGATTTTCATG-3′, TNFα (forward) 5′-CATCTTCTCAAAATTCGAGTGACAA-3′, TNFα (reverse) 5′-TGGGAGTAGACAAGGTACAACCC-3′, IL-4 (forward) 5′-ACAGGAGAAGGGACGCCAT-3′, IL-4 (reverse) 5′-GAAGCCCTACAGACGA-GCTCA-3′, IL-17 (forward) 5′-CAGGACGCGCAAACATGA-3′, and IL-17 (reverse) 5′-GCAACAGCATCAGAGACACAGAT-3′, macrophage inflammatory protein 2 (MIP-2) (forward) 5′-GCCAAGGGTTGACTTCA-3′, MIP-2 (reverse) 5′-TGTCTGGGCGCAGTG-3′, monocyte chemoattractant protein 1 (MCP-1) (forward) 5′-CTTCCTCCACCACCATGCA-3′, MCP-1 (reverse) 5′-CCAGCCGGCAACTGTGA-3′, monocyte chemoattractant protein 5 (MCP-5) (forward) 5′-GGAGGATCACAAGCAGCCAGT-3′, MCP-5 (reverse) 5′-TCAGCACAGATCTCCTTATCCAGTAT-3′.

Flow cytometric analysis of the stomach and lymph node cells

To analyze the cellular infiltrates in mouse stomachs, one-half of the glandular stomach was digested for 30 min with 1 mg mL⁻¹ dispase (Roche, Indianapolis, IN), 0.25 mg mL⁻¹ collagenase A, and 25 Units mL⁻¹ DNAse (both from Boehringer Mannheim, Mannheim, Germany) at 37 °C. The suspension was passed through a 70 µm cell strainer (BD). Cells were harvested by centrifugation, washed, and then live cells were counted using a hemocytometer and trypan blue exclusion staining. The paragastric lymph node preparations were passed through a 70 µm cell strainer without enzymatic digestion, and were then processed and analyzed in the same manner as the stomach cells. The samples were stained with 2 µg mL⁻¹ anti-CD4, anti-CD8, and anti-CD3, or 1.5 µg mL⁻¹ anti-Gr1, 2 µg mL⁻¹ anti-CD11b and anti-CD11c (BD) in a volume of 100 µL fluorescence-activated cell sorting (FACS) buffer (PBS, pH 7.4, containing 0.1% sodium azide, 0.1% bovine serum albumin, and 20% mouse serum). Cells were washed, resuspended in FACS buffer, and collected on a BD LSR II flow cytometer (BD Biosciences). By gating on the live cells (identified based on forward and side scatter properties) and subsequently gating on positively stained cells (based on the fluorescence using specific antibodies compared with isotype control antibodies), the proportion of live cells corresponding to each of the analyzed cell types was determined. Analysis was performed using denovo software, Thornhill, ON. The numbers of each cell type in a sample were calculated by multiplying the total number of live cells (determined based on counting with a hemocytometer and trypan blue staining) by the proportion of positively stained cells detected by flow cytometry.

Dendritic cell (DC) preparation

DCs were derived from the bone marrow of uninfected FVB/N mice. The bone marrow was flushed from the femurs and tibias of euthanized mice with Dulbecco's modified Eagle's medium (DMEM), as described previously (Bodnar et al., 2001). Red blood cells were lysed by treatment of the bone marrow suspension with red blood cell lysis buffer (0.17 M Tris–0.16 M ammonium chloride). The cells were washed twice with PBS containing 2% FBS, counted, resuspended in DC medium (10% FBS, 1% sodium pyruvate, and 1%l-glutamate in DMEM), and cultured overnight in a 25-mL flask. The following day, nonadherent and semiadherent cells were removed, centrifuged, and then resuspended (10⁶ cells mL⁻¹) and cultured for 5–7 days in DC medium containing 20 ng mL⁻¹ granulocyte macrophage colonystimulating factor (GM-CSF) and 20 ng mL⁻¹ IL-4 (PeproTech, Rocky Hill, NJ) in p75 filter flasks (Sarstedt, Newton, NC). DCs were cocultured with H. pylori strain SS1 harvested from broth cultures (at a multiplicity of infection of 1–2 bacteria cell⁻¹) for 18 h, before use in ELISPOT assays.

IFNγ ELISPOT

ELISPOT plates (96-well filtration plates; Millipore MultiScreen MAIPS4510; Millipore, Bedford, MA) were pretreated with 35% ethanol and then washed three times with sterile PBS. A rat anti-IFNγ capture antibody (clone R4-6A2; BD Pharmingen) was bound to the wells by incubating overnight at a concentration of 10 µg mL⁻¹. The ELISPOT plate was washed with PBS/0.1% Tween 20, and the plate was blocked with DMEM containing 20% FBS for 2 h. Single-cell suspensions from the lymph nodes were prepared as described above. Lymph node cells (100 000 cells well⁻¹) were incubated with medium alone, H. pylori-pulsed DCs, or control DCs (not exposed to H. pylori), for 36 h at 37 °C in 5% CO₂ in the presence of 20 U mL⁻¹ IL-2. The cells were then removed by washing, and a biotinylated anti-IFNγ detection antibody (clone XMG1.2; BD Pharmingen) was added for 2 h at 37 °C at 5 µg mL⁻¹, followed by five washes and the addition of streptavidin-peroxidase for 1 h at room temperature in a humidified chamber (Vectastain ABC kit; Vector Laboratories, Burlingame, CA). After additional washing, the plate was developed using the 3-amino-9-ethylcarbazole substrate kit as directed by the manufacturer (Vectastain ABC kit; Vector Laboratories). ELISPOT plates were read on an immunospot CTL plate reader (Cellular Technology, Cleveland, OH).

Statistical analysis

Four to six mice per group per time point were used for all of the studies. Statistical analysis was performed using one-way anova, followed by a Student–Neuman–Keuls post hoc test to compare results from different groups of mice. For analyses of bacterial numbers and cell numbers, the data were normalized by log transformation before statistical analysis.

Results

Helicobacter pylori colonization of the mouse stomach

To analyze immune responses to H. pylori, FVB/N mice were infected with a single dose (10⁸ CFU) of H. pylori strain SS1. The density of bacterial colonization of the stomach progressively increased until about 20 days postinfection (Fig. 1). Helicobacter pylori was recovered from the stomachs of mice as late as 6 months postinfection (7.8 × 10⁵±2.1 × 10⁵ CFU per stomach, data not shown). Helicobacter pylori was recovered from the stomachs of 100% of the mice that were infected during the course of this study.

Figure 1

Helicobacter pylori colonization of the mouse stomach. FVB/N mice were infected orogastrically with H. pylori on day 0 and sacrificed at the indicated timepoints postinfection. Serial dilutions of stomach homogenates were cultured as described in ‘Materials and methods.’ The day 0 data point indicates the number of organisms inoculated into the stomach. Other points represent the calculated number of CFU per whole stomach (mean±SEM), based on analysis of five to six mice at each time point. The SEM was <0.3 × 10⁵ at timepoints 3, 10, and 21 days postinfection, and therefore the error bars are not visible at these timepoints. The results are representative of three independent experiments.

Macrophage and neutrophil responses to H. pylori infection

Multiple enteric pathogens, including Citrobacter rodentium, Yersinia enterocolitica, Salmonella enterica, and Listeria monocytogenes, are known to elicit gastrointestinal inflammatory responses within several days after infection, and the most abundant infiltrating cell types are macrophages and neutrophils (Fukai & Maruyama, 1979; Heesemann et al., 1993; Cheminay et al., 2004; Coburn et al., 2005; Wei et al., 2005). To analyze macrophage and neutrophil responses to H. pylori, flow cytometric methods were used to assess infiltration of the mouse stomach with these cells. A transient increase in the numbers of neutrophils and macrophages in the stomach occurred within the first 2 days after H. pylori infection (Fig. 2). Fewer neutrophils and macrophages were detected in the stomach at days 3 and 10 postinfection than at days 1 and 2 postinfection. By 2–3 weeks postinfection, a second phase of neutrophil and macrophage infiltration into the stomach was detected. In this second phase of infiltration, fewer neutrophils and macrophages were detected than at days 1 and 2 postinfection.

Figure 2

Neutrophil and macrophage infiltration of the stomach following Helicobacter pylori infection. Mice were infected orogastrically with H. pylori and were sacrificed at the indicated timepoints postinfection. Mock-infected control mice received Brucella broth alone and were sacrificed at multiple subsequent timepoints. Gastric homogenates were stained with fluorescently labeled antibodies specific for neutrophils (CD11b⁺Gr1⁺) or macrophages (CD11b⁺Gr1⁻), and the cells were then analyzed by flow cytometry. The total numbers of each cell type per whole stomach were calculated as described in ‘Materials and methods.’ Day 0 data points indicate the mean numbers of cells detected in the stomachs from mock-infected mice. (a) Total number of neutrophils per whole stomach. (b) Total number of macrophages per whole stomach. Each bar represents the mean±SEM based on experiments using five to six mice at each time point. The results are representative of two independent experiments. ^*P≤0.05, ^**P≤0.01 (compared with mock-infected mice).

Chemokine expression in the stomachs of H. pylori-infected mice

It was hypothesized that infiltration of neutrophils and macrophages into the mouse stomach might be mediated by specific chemotactic factors. To investigate whether such factors were present, real-time RT-PCR was used to examine expression of MIP-2 (a neutrophil chemotactic chemokine), MCP-1, and MCP-5 (macrophage chemotactic chemokines) in the gastric tissue. The expression of all three of these chemokines was detected (Fig. 3). The expression of these chemokines increased during the 3 weeks postinfection, in association with increasing numbers of H. pylori in the stomach (Fig. 1).

Figure 3

Chemokine expression in the gastric tissue of Helicobacter pylori-infected mice. Mice were infected orogastrically with H. pylori and sacrificed at the indicated timepoints postinfection. Mock-infected control mice received Brucella broth alone and were sacrificed at multiple subsequent timepoints. Real-time RT-PCR was performed on RNA extracted from mouse stomach tissue to detect MIP-2, MCP-1, and MCP-5 expression. The levels of chemokine expression are presented relative to the expression of HPRT, a housekeeping control gene, and are normalized to the expression level of the chemokine in a mock-infected stomach (relative units=1). Each point or bar represents the mean±SEM based on experiments using five to six mice. Day 0 data points indicate the mean results for mock-infected mice. ^*P≤0.05, ^**P≤0.01 (compared with mock-infected mice).

Lymphocyte infiltration in the stomachs of H. pylori-infected mice

Infiltration of the stomach by T lymphocytes was also analyzed. The number of T lymphocytes infiltrating the stomach increased markedly by about day 20 postinfection (Fig. 4). T helper cells (CD4⁺CD3⁺ lymphocytes) were more abundant in the stomach than were cytotoxic (CD8⁺) lymphocytes (Fig. 4a). There was also a significant increase in dendritic cells in the stomach by day 20 postinfection, compared with the number of dendritic cells in the stomachs of infected mice at earlier timepoints or the number of dendritic cells in the stomachs of mock-infected mice (data not shown).

Figure 4

T cell infiltration and cytokine production in the mouse stomach following Helicobacter pylori infection. FVB/N mice were infected with H. pylori and sacrificed at the indicated timepoints postinfection. Mock-infected control mice received Brucella broth alone and were sacrificed at multiple subsequent timepoints. Stomach homogenates were stained with fluorescently labeled antibodies specific for T lymphocytes (CD3, CD4, or CD8), and analyzed by flow cytometry. Day 0 data points indicate the mean results for mock-infected mice. (a) Total numbers of CD4⁺CD3⁺ cells and CD8⁺CD3⁺ cells per whole stomach. (b) IFNγ, TNFα, IL-4, and IL-17 expression in the stomachs of H. pylori-infected mice. Real-time RT-PCR was performed on RNA extracted from mouse stomach tissue. The levels of cytokine expression are presented relative to expression of HPRT, a housekeeping control gene, and are normalized to the expression level of the cytokine in a mock-infected stomach (relative units=1). Each point or bar represents the mean±SEM based on experiments using five to six mice. The results are representative of three independent experiments. ^*P≤0.05, ^**P≤0.01 (compared with mock-infected mice).

Cytokine expression in the stomachs of H. pylori-infected mice

Infiltration of the stomach with T cells at day 20 postinfection is consistent with the development of an adaptive immune response to H. pylori. To investigate the expression of cytokines in the stomach during early stages in the development of an adaptive immune response, RNA was extracted from the glandular stomach tissue of H. pylori-infected mice and mock-infected control mice, and the expression of IFNγ, IL-4, and TNFα was quantified using real-time RT-PCR. By days 22–34 postinfection, the levels of IFNγ and TNFα expression in the stomachs of H. pylori-infected mice were significantly higher than in the stomachs of mock-infected mice (Fig. 4b). No significant increase in IL-4 expression was detected in the stomachs of H. pylori-infected mice compared with mock-infected mice. This pattern of cytokine expression is consistent with a Th1-predominant response.

In addition to Th1 and Th2 cells, another subset of effector T cells (Th17 cells) has recently been described. The production of IL-17 by Th17 cells may contribute to the pathogenesis of various chronic diseases in which there is an inflammatory component (Witowski et al., 2004; Lubberts et al., 2005; Chen et al., 2006; Yen et al., 2006). To investigate whether IL-17 is produced in gastric tissue in response to H. pylori infection, real-time RT-PCR was performed on mRNA extracted from glandular stomach tissue of H. pylori-infected mice and mock-infected controls. IL-17 expression was detected in the stomachs of H. pylori-infected mice by 3 weeks postinfection (Fig. 4b).

Analysis of the paragastric lymph nodes in H. pylori-infected mice

Based on gross inspection, the size of the paragastric lymph nodes increased substantially during the 3 weeks following H. pylori infection. The total number of cells in the paragastric lymph nodes increased about fourfold (from 0.8 × 10⁶±0.3 × 10⁶ cells in uninfected mice to 3.4 × 10⁶±0.5 × 10⁶ cells in H. pylori-infected mice). Correspondingly, flow cytometric analysis indicated that the number of CD4⁺ and CD8⁺ T cells in the paragastric lymph nodes increased and reached a maximum by about day 20 postinfection (Fig. 5a). Increased numbers of macrophages and dendritic cells were also detected in the paragastric lymph nodes of H. pylori-infected mice (Fig. 5b).

Figure 5

Analysis of paragastric lymph node cellular content following Helicobacter pylori infection. Mice were infected with H. pylori and sacrificed at the indicated timepoints postinfection. Mock-infected control mice received Brucella broth alone and were sacrificed at multiple subsequent timepoints. Paragastric lymph node preparations from each mouse were stained with markers for T lymphocytes, macrophages, and dendritic cells. Day 0 data points indicate the results for mock-infected mice. (a) Total number of CD4⁺CD3⁺ cells or CD8⁺CD3⁺ in the paragastric lymph node of mice (represented in millions). (b) Total number of CD11b⁺Gr1⁻ macrophages, and CD11c⁺ dendritic cells in the paragastric lymph nodes of the mice (represented in thousands). Each scale bar represents the mean±SEM based on experiments using five to six mice. The results are representative of three independent experiments. ^*P≤0.05 and ^**P≤0.01 (compared with mock-infected mice). (c) ELISPOT analysis of paragastric lymph node cells from H. pylori-infected mice. Paragastric lymph node cellular extracts were analyzed by ELISPOT. The graph illustrates analysis of IFNγ production by lymph node cells incubated with H. pylori-pulsed dendritic cells, lymph node cells incubated with unpulsed dendritic cells, and lymph node cells incubated in medium alone. Each bar represents the mean±SEM based on experiments using four to five mice. The results are representative of two independent experiments. ^*P≤0.05 and ^**P≤0.01 (compared with mock-infected mice).

These data suggested that paragastric lymph nodes might be a site of T cell priming during H. pylori infection. To further investigate this possibility, the ability of paragastric lymph node T cells isolated from H. pylori-infected mice to respond to H. pylori antigens was examined. Specifically, an ELISPOT assay was used to test whether the paragastric lymph node T cells produced IFNγ in response to H. pylori antigens, presented by dendritic cells. As a first step, bone marrow-derived dendritic cells from naïve mice were cocultured with intact H. pylori. Paragastric lymph node cells from H. pylori-infected mice were then incubated with DCs that had been cocultured with H. pylori, with control DCs (not pulsed with H. pylori), or with medium alone. IFNγ production was detected by ELISPOT analysis, as described in ‘Materials and methods’. The number of H. pylori-specific T cells expressing IFNγ was increased at 3 weeks postinfection compared with earlier timepoints, and the number of H. pylori-specific T cells expressing IFNγ increased further by 5 weeks postinfection (Fig. 5c). These data suggest that the paragastric lymph node is a site of T cell priming in response to H. pylori infection.

Discussion

In this study, a mouse model was used to investigate host responses that occur during the initial stages of H. pylori infection. The H. pylori strain used in these studies, SS1, is well adapted for colonization of the mouse stomach, and the levels of gastric colonization detected in the current study at 3 weeks postinfection are consistent with levels of H. pylori strain SS1 colonization that have been described in previous studies (Lee et al., 1990; Akada et al., 2003). The mouse strain used in the current study, FVB/N, has been used in several previous studies, including studies of Helicobacter-induced gastric malignancy (Wang et al., 2000; Fox et al., 2003). Male mice were used in the current study, as well as several previous studies, because male mice are reported to have a higher risk of H. pylori-induced gastric adenocarcinoma than are female mice (Fox et al., 2003). Gastric inflammatory responses to persistent Helicobacter infection in inbred mice are known to vary among different strains of mice (Mohammadi et al., 1996; van Doorn et al., 1999; Mahler et al., 2002), and therefore, it is possible that variations in host responses to H. pylori might be detected at early timepoints if multiple different strains of mice were compared.

Within 2 days after entry of H. pylori into the mouse stomach, a transient gastric infiltration of macrophages and neutrophils was detected. Subsequently, about 20 days postinfection, gastric infiltration of CD4⁺ T cells and dendritic cells was detected, along with macrophages and neutrophils. In a previous study, gastric inflammation was also detected at early timepoints in Helicobacter felis-infected mice (Lee et al., 1990). It is considered likely that the initial gastric inflammatory response is mediated by innate immune defenses, and that the subsequent wave of gastric inflammation reflects the development of an adaptive immune response. The detection of H. pylori-specific T cells in the paragastric lymph node (Fig. 5) provides support for this interpretation. Humoral immune responses have been detected in H. pylori-infected mice by about 1 week postinfection (Ferrero et al., 1998). An analysis of the density of H. pylori colonization indicated that the numbers of organisms in the stomach progressively increased until about 3 weeks postinfection, and then reached a plateau (Fig. 1). These data suggest that the development of an adaptive immune response is required for control of H. pylori proliferation.

There are few well-documented studies of acute H. pylori infection in humans, and therefore, relatively little is known about human immune responses to H. pylori that occur in the early stages of infection. Two studies described the development of neutrophilic gastritis within 5–10 days after ingestion of H. pylori by humans (Marshall et al., 1985; Sobala et al., 1991), and T cell responses have been detected within 4 weeks after ingestion of H. pylori by human volunteers (Nurgalieva et al., 2005). These observations are consistent with the biphasic pattern of gastric inflammatory responses detected in mice in the current study.

An interesting finding in the current study was that the initial inflammatory response to H. pylori (detected during the first few days postinfection) was transient. Neutrophils and macrophages infiltrated the stomach during the first few days postinfection, but the numbers of these cells in the stomach returned to baseline levels by day 10 postinfection. This diminution in intensity of the inflammatory response is somewhat surprising, because the density of H. pylori colonization in the stomach progressively increased between days 3 and 10 postinfection. The mechanistic basis for attenuation of the gastric inflammatory response between days 3 and 10 postinfection is not yet understood. One possibility is that the initial inflammatory response observed in this study was attributable to the introduction of a large bolus of H. pylori into the mouse stomach. The initial inflammatory response may be stimulated directly by bacterial factors such as H. pylori-neutrophil-activating protein (HP-NAP) or lipopolysaccharide, which are known to be activators of neutrophils and macrophages, i.e. the immune cells that are detected in the stomach at early timepoints. As the initial inoculum of H. pylori is cleared from the stomach, the concentration of such bacterial factors would decrease, and this might lead to an attenuation of the inflammatory response. Notably, natural infection of humans probably results from ingestion of a relatively small number of H. pylori compared with the number of bacteria administered to mice in the current study or administered to human volunteers in other studies (Marshall et al., 1985; Graham et al., 1988; Sobala et al., 1991). An alternate explanation for waning of the initial gastric inflammatory response is that H. pylori may produce factors that interfere with recruitment of inflammatory cells or factors that cause lysis or apoptosis of inflammatory cells. It is speculated that inappropriate waning of the innate immune response during the early stages of infection may be a factor that contributes to H. pylori persistence.

The detection of IFNγ but not IL-4 in the stomachs of H. pylori-infected mice is consistent with the development of a Th1-predominant response. The development of a Th1-predominant response to H. pylori in mice has been reported in multiple previous studies (Mohammadi et al., 1997; Smythies et al., 2000; Eaton et al., 2001; Sommer et al., 2001). In the current study, the expression of IL-17 was detected in the stomachs of H. pylori-infected mice, which suggests that H. pylori infection may also stimulate a Th17 response. IL-17, a proinflammatory cytokine produced mainly by Th17 cells, stimulates fibroblasts, endothelial cells, macrophages, and epithelial cells to produce proinflammatory cytokines such as IL-1, IL-6, IL-8, and TNFα (Harrington et al., 2006), and may play a role in the pathogenesis of numerous chronic inflammatory diseases, including inflammatory bowel disease, rheumatoid arthritis, and multiple sclerosis (Witowski et al., 2004; Lubberts et al., 2005; Chen et al., 2006; Yen et al., 2006). The role of IL-17 in inflammatory responses to enteric pathogens has not yet been investigated in detail. However, a recent study provided evidence that both Th1 and Th17 responses may contribute to severe intestinal inflammation in Helicobacter hepaticus-infected mice (Kullberg et al., 2006), and gastric expression of IL-17 has been detected in H. pylori-infected humans (Mizuno et al., 2005). It is speculated that IL-17 may play a role in the induction of gastric inflammation during H. pylori infection.

Very little is known about sites where priming of the immune response to H. pylori occurs. In contrast to the intestine, the stomach does not contain Peyer's patches or M cells (Neutra et al., 1996). Gastric epithelial cells upregulate the expression of MHC Class II and co-stimulatory molecules during H. pylori infection (Valnes et al., 1990; Ye et al., 1997; Archimandritis et al., 2000), and potentially these cells have a role in antigen presentation. Monocytes, macrophages, and dendritic cells in the lamina propria of the gastric mucosa may also play important roles in antigen presentation (Suzuki et al., 2002; Voland et al., 2003; Kranzer et al., 2004). Alternatively, priming of the immune response to H. pylori may occur within lymph nodes draining the stomach (Kaparakis et al., 2006), or may occur at intestinal sites or mesenteric lymph nodes in response to H. pylori antigens that are shed from the stomach. In the current study, it is demonstrated that T cells from the paragastric lymph nodes of H. pylori-infected mice secrete IFNγ when stimulated by H. pylori-pulsed dendritic cells. This provides evidence that paragastric lymph nodes are an important site for priming of Th1-type CD4⁺ T cells during H. pylori infection. Further study of this site may provide important insights into the development of adaptive immune responses to H. pylori.

In summary, these results provide important new insights into immune responses to H. pylori that occur during the early stages of infection. The immune responses to H. pylori that occur in experimentally infected animals and in naturally infected humans are typically ineffective in eradicating H. pylori. However, studies in animal models indicate that H. pylori infection can be prevented by prophylactic immunization (Ermak et al., 1998; Pappo et al., 1999; Akhiani et al., 2002, 2004; Garhart et al., 2003), and established infections can potentially be eradicated via therapeutic immunization (Vyas & Sihorkar, 1999; Sutton et al., 2000; Li et al., 2004). An important goal for future studies will be to understand more clearly the immunologic determinants of protective immunity against H. pylori, and to determine why immune responses to H. pylori fail to eradicate the infection in most infected humans.

Acknowledgements

This work was supported by NIH T32 AI-07474, R01 AI39657, R01 DK53623, R01 DK58587, R01 CA77955, R01 DK73902, R01 DK53620 and the Department of Veterans Affairs. The authors are grateful to Jim Higginbotham and Wilfred Ajayi of the Vanderbilt University Immunology Core Lab, which is funded by GCRC Grant MO1 RR-0095 from NCRR/NIH. The authors thank Dr Geraldine Miller and members of the Cover, Wilson, and Peek Laboratories for helpful discussions.

Author notes

Editor: Hans Kusters