Vacuolating Cytotoxin in Helicobacter pylori Water-Soluble Proteins Upregulates Chemokine Expression in Human Eosinophils via Ca²⁺ Influx, Mitochondrial Reactive Oxygen Intermediates, and NF-κB Activation

Abstract

Helicobacter pylori-infected gastric mucosa is characterized by infiltration of inflammatory cells such as neutrophils and eosinophils. However, little information is available on the relationship between H. pylori virulence factors and chemokine expression in eosinophils. This study investigates the role of vacuolating cytotoxin (VacA) in chemokine expression from human eosinophils. Eosinophils were isolated from the peripheral blood of healthy volunteers using a magnetic cell separation system. VacA⁺ H. pylori water-soluble proteins (WSP) induced higher expression of interleukin-8, growth-related oncogene alpha, monocyte chemotactic protein 1, and RANTES (regulated on activation, normal, T-cell expressed and secreted) than Vac⁻ WSP in human eosinophils, as assessed by quantitative reverse transcription-PCR and enzyme-linked immunosorbent assay. Purified VacA not only increased chemokine expression but also activated p65/p50 NF-κB heterodimers and phosphorylated IκB kinase (IKK) α/β signals in human eosinophils. Inhibition of NF-κB and IKK significantly decreased the chemokine expression in VacA-stimulated eosinophils. Furthermore, VacA-induced NF-κB activation and chemokine release from eosinophils were dependent on Ca²⁺ influx and mitochondrial generation of reactive oxygen intermediates (ROI). These results suggest that NF-κB and IKK signals via Ca²⁺ influx and mitochondrial ROI play a role in the up-regulation of chemokine expression in eosinophils stimulated with H. pylori VacA.

The Helicobacter pylori-infected gastric mucosa of patients with active chronic inflammation is characterized by the infiltration of various inflammatory cells, including neutrophils, eosinophils, lymphocytes, and cells of macrophage lineage (13, 14, 36). Even when H. pylori has been eradicated, the infiltrated eosinophils decrease slowly and do not return to normal levels within 1 year after treatment (13). Consistent with this, a 4-year histological follow-up study of gastric mucosa after H. pylori eradication showed that neutrophil infiltration disappeared within 2 months after eradication, while mononuclear cell infiltration persisted into the second year (46). In addition, the severity of chronic gastritis was significantly correlated with the eosinophil score (36). These studies suggest that infiltrated eosinophils may contribute to prolonged inflammation in H. pylori-infected gastric mucosa. However, there have been no reports elucidating the mechanism of prolonged inflammatory reaction by eosinophils and, in particular, the role of H. pylori virulence factors in chemokine expression by eosinophils.

Chemokines are small polypeptides with molecular masses from 8 to 12 kDa that have potent chemoattractant activity for leukocytes (1). Chemokines have four highly conserved cysteine residues in their primary structure and are classified into four subfamilies called CXC, CC, C, and CX3C, according to the number of amino acids other than cysteine (X) inserted between the first two cysteine residues (1). Members of the CXC chemokine subfamily, with one amino acid insertion between the first two cysteine residues, are potent chemoattractants for neutrophils, but they have no discernible effects on monocytes. Interleukin-8 (IL-8), for example, belongs to this CXC chemokine subfamily. CC chemokines, in which two cysteine residues lie side by side, attract mainly mononuclear cells (1). Many of these chemokine genes are targets of the nuclear transcription factor-kappa B (NF-κB) (47). NF-κB dimers are held in the cytoplasm in an inactive state by inhibitory proteins called IκBs (35). IκB kinase (IKK) is known to directly phosphorylate IκB, which then undergoes ubiquitin-mediated proteolysis, thereby releasing NF-κB dimers to translocate to the nucleus. Eosinophils can also secrete IL-8 in an NF-κB-dependent manner (51). Considering that chronic gastritis induced by H. pylori infection shows the infiltration of eosinophils and neutrophils, there is a possibility that these infiltrated cells continuously secrete chemokines. However, eosinophil-induced chemokine expression remains unclear, although chemokine expression in neutrophils has been well known in H. pylori infection (27, 28).

The vacuolating cytotoxin (VacA) is one of the major virulence factors in the pathogenesis of H. pylori-related diseases (32, 33). H. pylori strains expressing higher activity of VacA are correlated with increased severity of gastritis (18, 30, 40). VacA has been reported to have immunosuppressive activity, including the inhibition of T-cell proliferation (3, 12). However, VacA also has proinflammatory activities in immune cells (37, 44), suggesting that VacA induces chemokine expression. Considering that the severity of chronic gastritis is significantly correlated with the eosinophil score (36), there is a possibility that VacA activity may be correlated with the increased chemokine expression by eosinophils. Therefore, expression of chemokines from VacA-stimulated eosinophils may be involved in the prolonged inflammatory reaction. To date, little information is available on the relationship between VacA stimulation and chemokine expression in eosinophils. In this study, we investigate the role of VacA in induction of chemokine expression by human eosinophils and find that a signaling pathway including Ca²⁺ influx, reactive oxygen intermediates (ROI), IKK, and p65/p50 heterodimeric NF-κB activation plays a key role.

MATERIALS AND METHODS

H. pylori strains.

The high toxin-producing H. pylori strain 60190 (ATCC 49503; CagA⁺ vacA s1a/m1) was used for the purification of VacA. A non-toxin-producing isogenic mutant (VacA⁻ mutant strain 60190) was created as previously described, except that natural transformation was used to transform H. pylori 60190 with the pVacA-negative:km plasmid, which was constructed according to the protocol of Cover et al. (6). Immunoblotting confirmed that no VacA or VacA fragments were expressed in the mutant strain. H. pylori was cultured under microaerophilic conditions (5% O₂, 10% CO₂, 85% N₂), and the virulence factors of the H. pylori strains, including CagA protein and production of VacA, were determined as previously described (24).

Preparation of H. pylori water-soluble protein.

H. pylori water-soluble protein (WSP) was prepared according to a previous protocol (27, 28). Protein concentrations in H. pylori WSP were determined by the Bradford method (Bio-Rad, Hercules, CA), and urease activity was measured by a coupled enzyme assay as previously described (9, 28). The Limulus amoebocyte lysate assay (Associates of Cape Cod, Inc., Falmouth, MA) was used for the detection of endotoxic activity of H. pylori lipopolysaccharide in H. pylori water-soluble proteins (WSP).

Preparation of H. pylori broth culture supernatants and VacA purification.

A VacA-producing H. pylori strain (60190) was grown in sulfite-free brucella broth containing 5% fetal bovine serum (FBS; Gibco, Grand Island, NY) (5). Cultures were incubated on a rotary shaker for 48 h at 37°C under microaerophilic conditions, and the cultures were then centrifuged. Broth culture supernatants were concentrated 30-fold by ultrafiltration (Millipore, Bedford, MA) and passed through a 0.2-μm-pore-size filter. VacA was purified from broth culture supernatants and was then dialyzed in phosphate-buffered saline (PBS) as described previously (5, 11). Immediately before use on cells, purified VacA was activated. Acid activation of VacA was accomplished by dropwise addition of 250 mM HCl to the purified toxin until a pH of 2.0 was reached (8). The purification of VacA protein was kindly supported by Patrice Boquet at Laboratoire de Bacteriologie, Hopital de l'Archet 2, Nice, France (11).

Isolation of peripheral blood eosinophils.

Eosinophils were isolated from the peripheral blood of healthy volunteers using a magnetic cell separation system (Miltenyi Biotec, Bergisch Gladbach, Germany), as described previously (43). In brief, venous blood (30 ml), anticoagulated with heparin (10 U/ml), was diluted with PIPES [piperazine-N,N′-bis(2-ethanesulfonic acid)] buffer (25 mM PIPES, 50 mM NaCl, 5 mM KCl, 25 mM NaOH, 5.4 mM glucose, pH 7.4) at a 1:1 ratio. Diluted blood was layered onto a Histopaque solution (density, 1.083 g/ml) (Sigma Chemical Co., St. Louis, MO) and centrifuged at 100 × g at 4°C for 30 min. The supernatant and mononuclear cells at the interface were removed carefully. Erythrocytes in the sediment were lysed by exposure to two cycles of sterile distilled water. Isolated granulocytes were washed with PIPES buffer containing 1% FBS, and an approximately equal volume of anti-CD16 antibody conjugated with magnetic particles (Miltenyi Biotec) was added to the cell pellet. After 30 min of incubation on ice, cells were loaded onto the separation column positioned in the magnetic cell separation system magnetic field. Cells were eluted with 15 ml of PIPES buffer with 1% FBS. The purity of eosinophils counted by Randolph's stain was >98%. Purified eosinophils were used immediately for experiments.

Quantitative RT-PCR analysis and chemokine enzyme-linked immunosorbent assay (ELISA).

Freshly isolated human eosinophils or a human eosinophil cell line, EoL-1 cells (Riken BioResource Center RCB0641, Tsukuba, Japan) (41), were stimulated with H. pylori WSP or VacA in 24-well plates, after which total cellular RNA was extracted using Trizol reagent (Gibco BRL, Gaithersburg, MD). Chemokine mRNA levels were measured by quantitative reverse transcription-PCR (RT-PCR) using internal standards. The oligonucleotide primers used for PCR amplification and the sizes of the PCR products obtained from target cellular RNA and synthetic standard RNA were described in a previous report (17). PCR amplification consisted of 28 to 35 cycles of 1 min of denaturation at 95°C, 2.5 min of annealing, and extension at 60°C (IL-8 and growth related oncogene alpha [GRO-α]), 65°C (monocyte chemotactic protein 1 [MCP-1] and RANTES), or 72°C (β-actin). Real-time PCR to detect IL-8 and β-actin mRNAs was carried out using an ABI PRISM 7700 Sequence Detection System (Perkin-Elmer Applied Systems, Foster City, CA) and SYBR green fluorescent dye according to the manufacturer's instructions. Probes, reagents, and TaqMan cytokine gene expression plates were used as recommended by the manufacturer (Applied Biosystems, Foster City, CA).

Chemokines in culture supernatants were determined by Quantikine immunoassay kits (R&D Systems). Chemokine proteins were tested in triplicate.

In some experiments, human eosinophils were preincubated with an NF-κB essential modifier (NEMO)-binding domain (NBD) peptide (200 μM; Peptron, Daejeon, Korea) or an inhibitor of NF-κB, calpain-1 inhibitor(25 μM; Calbiochem, La Jolla, CA), before addition of VacA. An NBD peptide can block association of NEMO with the IKK complex and inhibit NF-κB activation (26, 34). To suppress NF-κB activity, recombinant retrovirus containing dominant-negative IκBα (retrovirus-IκBα-AA) was transfected into EoL-1 cells, as described previously (22, 25). At 48 h after transfection, the cells were stimulated with VacA.

EMSAs.

Cells were harvested and nuclear extracts were prepared as described previously (21). The concentrations of proteins in the extracts were determined by a Bradford assay (Bio-Rad, Minneapolis, MN). Electrophoretic mobility shift assays (EMSAs) were performed according to the manufacturer's protocol (Promega, Madison, WI) (25). To identify specific members of the NF-κB family, a supershift assay was performed as described above, except that rabbit antibodies (1 μg/reaction) against NF-κB proteins, including p50 (H-119X), p52 (K-27X), p65 (sc-8008X), c-Rel (B-6X), and Rel B (C-19X) (all from Santa Cruz Biotechnology), were added during the binding reaction (25).

Immunoblots.

Cells were washed with ice-cold PBS and lysed in 0.5 ml/well lysis buffer (150 mM NaCl, 20 mM Tris, pH 7.5, 0.1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 10 μg/ml aprotonin). Fifteen to 50 micrograms of protein/lane was size-fractionated on 6% polyacrylamide minigels (Mini-Protein II; Bio-Rad) and electrophoretically transferred to nitrocellulose membranes (0.1-μm-pore-size filter). Specific proteins were detected using mouse anti-human IκBα (Santa Cruz Biotechnology), IKKα, IKKβ, phospho-IKKα/IKKβ, and actin (Cell Signaling Technology, Beverly, MA) as primary antibodies, and peroxidase-conjugated anti-mouse immunoglobulin G (Transduction Laboratories, Lexington, KY) as a secondary antibody. Specifically bound peroxidase was detected by enhanced chemiluminescence (ECL system; Amersham Life Science, Buckinghamshire, England) and exposure to X-ray film.

Adenoviral infection and NF-κB-luciferase reporter assay.

EoL-1 cells were infected for 16 h with an adenoviral vector encoding an NF-κB-luciferase reporter gene (Ad5κB-LUC) and incubated with VacA, after which whole-cell lysates were prepared as previously described (29). Luciferase activity was determined according to the manufacturer's instructions (Tropix Inc., Bedford, MA) (21).

Fluorocytometric analysis.

Dihydrorhodamine-123 (DHR-123; Molecular Probes, Eugene, OR) was used to detect mitochondrial generation of ROI in eosinophils using fluorocytometry, as previously described (15). Briefly, eosinophils were washed and resuspended in 1 ml of PBS at a concentration of approximately 10⁶ cells/ml. Ten minutes prior to analysis, DHR-123 was added to a final concentration of 5 μM. In some experiments, eosinophils were preincubated for 30 min with an antioxidant such as butylated hydroxyanisole (BHA; 200 μM) or butylated hydroxytoluene (BHT; 200 μM) or Ca²⁺ chelator [1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid tetra(acetoxymethyl ester)] ([BAPTA-AM] 50 μM; Calbiochem) before the addition of VacA.

RESULTS

One of the H. pylori WSP components that up-regulates the expression of chemokines is a vacuolating cytotoxin.

We asked whether treatment with H. pylori WSP obtained from VacA⁺ or VacA⁻ H. pylori strains could make a difference in the expression of CXC and CC chemokines in human eosinophils. As shown in Fig. 1, all H. pylori WSP increased the release of such chemokines as IL-8, GRO-α, MCP-1, and RANTES from eosinophils. However, the chemokine secretion by VacA⁻ mutant H. pylori WSP was lower than that by VacA⁺ WSP. For example, VacA⁻ mutant WSP reduced IL-8 production by ∼ 68% compared to VacA⁺ H. pylori WSP. The urease activity and specific urease activity in VacA⁺ H. pylori WSP were 91 U/ml and 27.6 U/mg of protein, respectively. The urease activity and specific urease activity in VacA⁻ mutant WSP was 92 U/ml and 28.2 U/mg of protein, respectively. The endotoxin activity of all H. pylori WSP preparations was less than 1 U/ml.

FIG. 1.

Chemokine secretion by eosinophils treated with H. pylori WSP (HPWP). Human eosinophils (10⁶ cells) were incubated for 18 h with H. pylori WSP (30 μg/ml) obtained from VacA⁺ H. pylori 60190 or a VacA⁻ mutant. Chemokine levels in culture supernatants were determined by ELISA. Data are means ± SEM (n = 7). Asterisks indicate that values for H. pylori WSP obtained from the VacA⁺ H. pylori strain are significantly different from those obtained from the VacA⁻ mutant (P < 0.05).

We next measured chemokine secretion in human eosinophils stimulated with VacA protein purified from H. pylori 60190. The magnitude of the chemokine response was dependent on the concentration of purified VacA. Thus, stimulation of eosinophils with increasing concentrations of VacA was paralleled by increased IL-8 release. After 18 h of stimulation with concentrations of 0.1, 1, 10, and 50 μg/ml, IL-8 release increased 45.5 ± 3.5 ng/ml, 119.5 ± 12.4 ng/ml, 660.3 ± 20.4 ng/ml, and 621.7 ± 31.3 ng/ml, respectively, relative to levels of unstimulated controls (52.9 ± 8.7 ng/ml) (mean ± standard error of the mean [SEM]; n = 3).

H. pylori VacA induces NF-κB activation and IκB degradation in human eosinophils.

To determine whether VacA could activate NF-κB signals in human eosinophils, DNA binding studies were performed using nuclear extracts after stimulation with VacA (10 μg/ml). Stimulation of eosinophils with VacA increased NF-κB DNA bindings, as shown by EMSA, and IκBα degradation was observed by immunoblot analysis (Fig. 2A). The magnitude of the NF-κB response was dependent on the concentration per cell of stimulating VacA in an eosinophil cell line, EoL-1. Stimulation with increasing concentrations of VacA was paralleled by increased luciferase activity in EoL-1 cells transfected with Ad5κB-LUC. Three hours after stimulation at concentrations of 0.1, 1, 10, and 50 μg/ml, reporter gene activity of NF-κB increased 0.9-, 1.6-, 3.6-, and 3.4-fold, respectively, relative to unstimulated controls (mean value; n = 3).

FIG. 2.

NF-κB activation and IκB degradation in human eosinophils stimulated with H. pylori VacA. Human eosinophils were stimulated with purified VacA (10 μg/ml). (A) NF-κB DNA binding activity was assessed by EMSA at the indicated times. Immunoblots for concurrent IκBα levels under the same conditions are shown beneath each EMSA time point. The results are representative of five repeated experiments. +, the positive control in which eosinophils were treated with tumor necrosis factor α (20 ng/ml) for 1 h; −, the negative control. (B) Activation of specific NF-κB subunits in human eosinophils stimulated with VacA. Supershift assays were performed using antibodies to p50, p52, p65, c-Rel, and Rel B. (C) Phospho-p65 activities in cytoplasmic and nuclear fractions in VacA-stimulated eosinophils. Human eosinophils were incubated with VacA (10 μg/ml) for the indicated times, and cytoplasmic and nuclear fractions were obtained. Protein expression of phospho-p65 and actin was assessed by immunoblotting. The results are representative of at least three repeated experiments.

To identify the specific subunits that comprise the NF-κB signal detected by EMSAs in VacA-stimulated eosinophils, supershift assays were performed. As shown in Fig. 2B, antibodies to p65 and p50 shifted the NF-κB signals significantly. However, anti-p52, anti-c-Rel, or anti-Rel B antibodies did not shift the NF-κB signal. In addition, phospho-p65 activities were increased in VacA-stimulated nuclear and cytoplasmic fractions (Fig. 2C). These results suggest that NF-κB activation by VacA stimulation may be mediated predominantly by heterodimers of p65/p50.

Inhibition of NF-κB activity suppresses chemokine expression in human eosinophils stimulated with H. pylori VacA.

Since VacA stimulation activated NF-κB signals in human eosinophils (Fig. 2), we further evaluated whether blocking the activation of NF-κB could influence the expression of chemokines. Addition of calpain-1 inhibitor to VacA-stimulated eosinophils significantly decreased chemokine secretion (Fig. 3). In another experiment, an EoL-1 human eosinophil cell line was transfected for 48 h with retrovirus-IκBα-AA. These transfected EoL-1 cells were then stimulated with VacA (10 μg/ml). Retrovirus-IκBα-AA completely blocked NF-κB activity in VacA-stimulated EoL-1 cells, while the control retrovirus containing a green fluorescent protein (GFP) plasmid did not change NF-κB activity (Fig. 4A). In addition, retrovirus-IκBα-AA significantly inhibited VacA-induced mRNA expression of IL-8, GRO-α, MCP-1, and RANTES (Fig. 4B). Since real-time PCR is more accurate than quantitative RT-PCR using standard RNA, we performed real-time PCR for IL-8 mRNA. As a result, the relative expression of IL-8 mRNA was determined to be 0.03 ± 0.01 in unstimulated controls, 0.58 ± 0.08 in cells stimulated with VacA alone, 0.17 ± 0.03 in cells stimulated with retrovirus-IκBα-AA plus VacA, and 0.61 ± 0.06 in cells stimulated with retrovirus-GFP plus VacA (mean ± SEM; n = 3).

FIG. 3.

The effects of calpain-1 inhibitor on chemokine production in human eosinophils stimulated with H. pylori VacA. Human eosinophils (10⁶ cells) were pretreated with an inhibitor of NF-κB, calpain-1 inhibitor (25 μM), for 1 h, and then purified VacA (10 μg/ml) was added into the cells for an additional 18 h. Chemokine levels in culture supernatants were determined by ELISA. Data are means ± SEM (n = 5). Asterisks indicate values of VacA plus calpain-1 inhibitor that are significantly different from VacA alone (P < 0.05).

FIG. 4.

Expression of chemokine genes in VacA-stimulated EoL-1 eosinophil cell lines transfected with retrovirus containing IκBα superrepressor. (A) EoL-1 cells were transfected with either retrovirus containing IκBα-superrepressor (IκBα-AA) or control virus (GFP). At 48 h after transfection, the cells were stimulated with VacA (10 μg/ml) for 1 h. NF-κB binding activity was assayed by EMSA. (B) Transfected EoL-1 cells were then stimulated with VacA (10 μg/ml) for 6 h. For quantification of chemokine mRNA transcripts, total RNA was reverse transcribed using an oligo(dT) primer and synthetic internal RNA standards and amplified by PCR. Data are presented as numbers of mRNA transcripts/μg of total RNA (mean ± standard deviation; n = 5). The β-actin mRNA levels in stimulated and unstimulated cells remained relatively constant throughout the same period (∼5 × 10⁵ transcripts/μg of total RNA). Asterisks indicate statistical significance in comparison with control virus-transfected cells stimulated with VacA (P < 0.05). □, control; ▪, VacA; ▤, VacA-stimulated cells transfected with retrovirus-IκBα-AA; ▒, VacA-stimulated cells transfected with control virus (retrovirus-GFP).

H. pylori VacA increased phosphorylated IKK signals in human eosinophils.

VacA stimulation increased the signals of phosphorylated IKKα/β in human eosinophils (Fig. 5A). To determine whether IKK signal might be associated with VacA-induced chemokine expression, human eosinophils were treated with an inhibitory NBD peptide. Secretion of IL-8 and RANTES was significantly inhibited by treatment with the NBD peptide but not by treatment with a mutant-type of NBD peptide (Fig. 5B). In this system, the addition of an NBD peptide suppressed a phosphorylated IKK signal induced by VacA stimulation (Fig. 5C).

FIG. 5.

Inhibition of IKK decreases the expression of chemokines in VacA-stimulated eosinophils. (A) Phosphorylation of IKK in VacA-stimulated eosinophils. Human eosinophils were incubated with VacA (10 μg/ml) for the indicated times. Protein expression of IKKα, IKKβ, phospho-IKKα/β, and actin was assessed by immunoblotting. The results are representative of at least five repeated experiments. (B) Densitometric analysis for the expressed proteins. The values represent the relative densities of each protein compared with actin using the TINA program. •, IKKα; ○, IKKβ; ▴, phospho-IKKα/β. (C) Human eosinophils (10⁶ cells) were treated with an NBD peptide (200 μM) for 1 h, followed by VacA treatment (10 μg/ml). Chemokine production was measured by ELISA 18 h after VacA treatment (mean ± SEM, n = 5; *, P < 0.05 compared with VacA alone). (D) Human eosinophils were incubated with an NBD peptide (200 μM). After 1 h, VacA (10 μg/ml) was added for 10 min. Cell lysates were prepared, size-fractionated, and blotted onto a nitrocellulose membrane. Phospho-IKKα/β and actin were detected with specific antibodies and enhanced chemiluminescence. The results are representative of three independent experiments.

H. pylori VacA-induced NF-κB activation and chemokine release are dependent on mitochondrial ROI generation and Ca²⁺ influx in eosinophils.

Since mitochondrial generation of ROI induced by bacterial toxin is involved in the nuclear translocation of NF-κB (16) and since VacA affects the mitochondria of gastric epithelial cells (10, 49), we asked whether mitochondrial generation of ROI might be involved in IL-8 release from eosinophils exposed to VacA. Human eosinophils pretreated with the antioxidant BHA (200 μM) and then exposed to VacA released significantly diminished amounts of IL-8 (Fig. 6A). Consistent with this, another antioxidant, BHT (200 μM), and the mitochondrial inhibitor rotenone (25 μM) also significantly reduced IL-8 release from eosinophils after 18 h of VacA exposure (control, 56.3 ± 7.7 pg/ml; VacA, 806.7 ± 22.8 pg/ml; VacA plus BHT, 140.7 ± 5.5 pg/ml; VacA plus rotenone, 184.0 ± 10.8 pg/ml; values are means ± SEM for n = 3).

FIG. 6.

Effect of BHA on VacA-induced IL-8 production and ROI generation from eosinophils. (A) Human eosinophils were incubated with either medium alone or medium containing VacA (10 μg/ml) for the indicated time points. In another experiment, cells were first exposed to the antioxidant BHA (200 μM) for 30 min before the addition of VacA. Culture supernatants were collected at different time points, and IL-8 levels were measured by ELISA. Values are the means ± SEM (n = 5). *, P < 0.05 versus controls; **, P < 0.01 versus controls; †, P < 0.01 versus VacA; ‡, P < 0.001 versus VacA. (B) Human eosinophils were loaded with DHR-123, a fluorescent probe for ROI, and then exposed to VacA (10 μg/ml) for 15 min or 30 min. Some cells were exposed to BHA (200 μM), an antioxidant, before the addition of VacA. At the indicated time points, cells were examined by fluorocytometry. The results are representative of at least five repeated experiments.

We next determined whether VacA stimulation might be associated with mitochondrial generation of ROI. VacA increased ROI generation in eosinophils (Fig. 6B). In addition, pretreatment with BHA followed by VacA exposure significantly inhibited the generation of ROI. These results suggest that mitochondrial generation of ROI is involved in the release of chemokines from eosinophils stimulated with VacA. Since VacA could induce NF-κB activation in eosinophils (Fig. 2), we asked whether NF-κB activation might be associated with mitochondrial ROI generation. Preincubation of eosinophils with BHA before exposure to VacA resulted in suppressed activity of NF-κB and phospho-IKKα/β, compared with VacA alone at all time points tested (Fig. 7A). In addition, release of such chemokines as IL-8 and RANTES was also significantly inhibited by BHA pretreatment (Fig. 7B).

FIG. 7.

Effect of BHA on VacA-induced activation of NF-κB and phospho-IKK and the expression of IL-8. (A) Human eosinophils were either exposed to VacA (10 μg/ml) alone or exposed to BHA (200 μM) for 30 min before the addition of VacA. At the indicated time points, nuclear extracts were prepared and analyzed by EMSA using a probe containing a consensus NF-κB binding site. Under the same conditions, protein expression of phospho-IKKα/β and actin were assessed by immunoblotting. Preincubation of eosinophils with the antioxidant BHA before exposure to VacA resulted in diminished NF-κB and phospho-IKK activity compared with VacA alone. The results are representative of at least three repeated experiments. (B) Human eosinophils were preexposed to BHA (200 μM) for 30 min and then treated with VacA (10 μg/ml) for an additional 18 h. Chemokine levels in culture supernatants were determined by ELISA. Data are means ± SEM (n = 5). *, P < 0.05 versus VacA alone.

To examine whether intracellular Ca²⁺ increase might be involved in signaling transduction, VacA-stimulated human eosinophils were pretreated with the Ca²⁺ chelator BAPTA-AM (50 μM). As shown in Fig. 8A, pretreatment with BAPTA-AM 30 min before the addition of VacA significantly decreased ROI production. Similar effects by BAPTA-AM were observed in NF-κB activation (Fig. 8B) and chemokine secretion (Fig. 8C). These results suggest that VacA induces intracellular Ca²⁺ influx, mitochondrial ROI generation, NF-κB activation, and, finally, chemokine expression in human eosinophils.

FIG. 8.

Effect of Ca2⁺ chelation on ROI generation, NF-κB activation, and chemokine expression in VacA-stimulated eosinophils. (A) Human eosinophils were loaded with DHR-123, a fluorescent probe for ROI, and then exposed to VacA (10 μg/ml) for 30 min. Some cells were exposed to BAPTA-AM (50 μM), a calcium chelator, before the addition of VacA. Cells were examined by fluorocytometry. The results are representative of at least five repeated experiments. (B) Under the same conditions, nuclear extracts were prepared and analyzed by EMSA using a probe containing a consensus NF-κB binding site. The results are representative of at least five repeated experiments. (C) Human eosinophils were preexposed to BAPTA-AM (50 μM) for 30 min and then incubated with VacA (10 μg/ml) for an additional 18 h. Chemokine levels in culture supernatants were determined by ELISA. Data are means ± SEM (n = 5). *, P < 0.05 versus controls.

DISCUSSION

In the present study, VacA stimulation up-regulated the expression of CXC chemokines (IL-8 and GRO-α) and CC chemokines (MCP-1 and RANTES) in human eosinophils. CXC chemokines are potent chemoattractants for neutrophils, and CC chemokines can attract eosinophils and lymphocytes (1). Therefore, the chemokines up-regulated by VacA stimulation may contribute to the infiltration of inflammatory cells to mucosa infected with VacA-producing H. pylori. In addition, VacA-stimulated eosinophils may be an important factor in triggering tissue damage, since eosinophils can induce H. pylori-related tissue damage through degranulation processes (19, 39).

VacA⁺ H. pylori WSP-induced chemokine release by eosinophils was higher than that induced by VacA⁻ WSP. These results indicate that VacA is one of the H. pylori WSP components up-regulating the expression of chemokines. However, treatment of eosinophils with VacA⁻ WSP also increased chemokine expression compared with nontreated controls. This increased expression of chemokines may be in part due to urease in WSP, because urease can increase chemokine secretion in immune cells infected with H. pylori (2, 45). However, the high chemokine secretion by VacA⁺ H. pylori WSP compared with VacA⁻ WSP was not due to urease activity, because there was no significant difference in urease activity and specific urease activity between our VacA⁺ and VacA⁻ H. pylori WSP preparations.

VacA stimulation induced NF-κB activation and IκBα degradation in human eosinophils. In addition, inhibition of NF-κB significantly decreased chemokine expression by VacA-stimulated eosinophils, suggesting that NF-κB activation by VacA is directly connected with chemokine expression. NF-κB is a dimeric transcription factor composed of homodimers or heterodimers of Rel protein, of which there are five family members in mammalian cells (RelA [p65], c-Rel, RelB, NF-κB1 [p50], and NF-κB2 [p52]) (47). Heterodimers of p65 and p50 are the predominant NF-κB complexes that translocate to the nucleus after H. pylori infection of intestinal epithelial cells (20, 23). Supershift studies demonstrated that NF-κB activation by VacA stimulation of eosinophils was mediated predominantly by heterodimers of p65/p50. In addition, phospho-p65 activities were increased in VacA-stimulated nuclear and cytoplasmic fractions in a time-dependent manner.

A major pathway of NF-κB activation involves the activation of IKK, which is followed by IκB phosphorylation. Phosphorylated IκB then undergoes ubiquitin-mediated proteolysis, thereby releasing NF-κB dimers to translocate to the nucleus (35). In this study, VacA stimulation increased the levels of phosphorylated IKKα/β in human eosinophils. In addition, NBD peptide, which can block association of NEMO with the IKK complex (34), significantly reduced VacA-induced chemokine up-regulation. The present study showed that the levels of IKKα were relatively constant but that IKKβ was increased in eosinophils stimulated with VacA. Concurrently, levels of phospho-IKKα/β were also increased, suggesting that IKKβ was phosphorylated in eosinophils stimulated with VacA. These results demonstrate that the activation of IKK is a crucial step for NF-κB activation and chemokine expression in human eosinophils following VacA stimulation.

Eosinophils isolated from human volunteers may be stimulated with nonspecific factors such as culture medium, FBS, purification buffer, etc. Therefore, this nonspecific stimulation may induce signals of mitogen-activated protein kinase (MAPK) in eosinophils, because resting eosinophils isolated from human volunteers showed constitutive expression of p38 signal (data not shown). MAPK activation is known to be associated with IKK activation (38). In addition, AP-1 signal activated by H. pylori or bacterial enterotoxin is associated with MAPK and plays a role in IL-8 release (21, 42). Based on these reports, it is possible that minimal activation of IKK may be in part due to MAPK activation.

VacA is known to affect the mitochondria of gastric epithelial cells (10, 49). The present study demonstrates that chemokine release from eosinophils after exposure to VacA is dependent on an oxidative burst originating from the mitochondria and that this process is associated with NF-κB activation. In addition, NF-κB activation in human eosinophils could be induced via a Ca²⁺-dependent mechanism. These results are consistent with a report that H. pylori VacA induced a Ca²⁺-dependent production of proinflammatory cytokines in a mast cell line (7). Moreover, pretreatment with the Ca²⁺ chelator BAPTA-AM significantly decreased ROI production, NF-κB activation, and chemokine release induced by VacA stimulation, suggesting that the intracellular Ca²⁺ influx might precede mitochondrial ROI production and NF-κB activation in human eosinophils. Water-soluble fractions containing VacA protein could induce more or less chemokine release, which often depends on cell type and bacterial strains. In addition, the observation that the inhibition of Ca²⁺ release (using BAPTA-AM) blocks both NF-κB activation and IL-8 production in VacA-treated eosinophils seems to be phenomenological. Therefore, further study is needed to clarify these phenomena.

It is not known whether cytochrome c released by VacA stimulation may be directly associated with NF-κB activation, since cytochrome c release did not necessarily require interaction of VacA with mitochondria in gastric epithelial cells (50). However, elevated cytosolic Ca²⁺ levels can trigger mitochondrial damage in several cell types via the generation of ROI (31, 48). Moreover, exposure of isolated cardiac myocytes to Ca²⁺ resulted in leakage of cytochrome c and generation of ROI (4). Based on these reports, increased intracellular Ca²⁺ levels in eosinophils stimulated with VacA may be critical for the mitochondrial cytochrome c and ROI generation, although their exact roles await further elucidation.

In summary, the VacA of H. pylori WSP up-regulates the expression of chemokines in human eosinophils, and a signaling pathway (Ca²⁺ influx → mitochondrial ROI generation → IKK and p65/p50 heterodimeric NF-κB activation) plays an important role in this process. These results suggest that increased chemokine release from eosinophils can contribute to the infiltration of inflammatory cells and gastric damage in VacA-producing H. pylori infection.