| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Department of Medical Microbiology and Parasitology, Shanghai Second Medical University, Shanghai, China 2Institute of Immunology, Second Military Medical University, Shanghai, China 3Shanghai Institutes for Biological Sciences, Chinese Academy of Sciences, Shanghai, China
Correspondence Xiao-Kui Guo xkguo{at}shsmu.edu.cn
Received February 13, 2004
Accepted June 14, 2004
To understand the biological processes within host cells induced by VacA, isogenic strains of Helicobacter pylori (NCTC 11638 or 11638-
vacA) were used to stimulate gastric cancer cells SGC7901, and differentially expressed genes in host cells were identified using cDNA microarray technology. More than 300 genes were found to alter their mRNA expression at different time points, among which 68 were related to the cytoskeleton, 87 were associated with cell cycle, cell death and proliferation, IL8 expression was also found to be up-regulated. Cells co-cultured with broth-culture supernatant (BCS) of NCTC 11638 showed more alteration in microtubule cytoskeleton morphology, as observed by laser scanning confocal microscopy, and a lower apoptosis rate, detected by flow cytometry, compared with those co-cultured with BCS of 11638-vacA. The supernatants of cells co-cultured with NCTC 11638 showed significantly higher IL8 expression than those co-cultured with 11638-vacA. It is concluded that VacA disrupts cytoskeletal architecture by influencing the expression of cytoskeleton-associated genes. VacA breaks the balance between cell proliferation and cell death by inducing the maladjustment of genes related to cell cycle. VacA is also able to induce the inflammatory response.
These authors contributed equally to this work. Abbreviations: BCS, broth-culture supernatant; ROS, reactive oxygen species.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
In order to understand the biological processes within host cells induced by H. pylori VacA, analysis of the alteration of gene expression in host cells is effective, as host responses contribute much to pathogenesis besides the bacterial factors. Microarray expression profiling and the development of data-mining tools offer high throughput screening of the differential expression of many genes in pathogenic cells (DeRisi et al., 1996). In some studies, microarrays were used to identify host molecular pathways that a single virulence determinant affects by enabling comparative analysis of the host transcriptional response to infection with virulent or avirulent mutant bacteria (Bach et al., 2002; Detweiler et al., 2001). These studies are a good paradigm for investigating the pathogenesis of any single virulence determinant.
In this study, gastric cancer cells were infected with a wild-type H. pylori VacA+ strain or the isogenic VacA– mutant strain. Differentially expressed genes were identified using cDNA microarray technology. Different effects on cytoskeletal alteration, programmed cell death and IL8 secretion were evaluated.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
vacA was constructed by substitution of a kanamycin-resistant gene for a short fragment of vacA through homologous recombination, as previously described by Yuan et al. (2003a). H. pylori strains were cultured routinely on brain heart infusion (BHI) agar plates with 5 % sheep blood in mixed air containing 10 % CO2, 5 % O2 and 85 % N2 at 37 °C. For the preparation of cell-free supernatants from H. pylori broth cultures (BCS), H. pylori was cultured in BHI broth with 10 % FBS in mixed air at 37 °C with agitation (200 r.p.m.) for 48 h. The cultures were centrifuged (15 000 g, 30 min, 4 °C) and filtered through a 0.2 µm syringe filter.
H. pylori infection of SGC7901 gastric cancer cells.
SGC7901 gastric cancer cells were maintained in DMEM (Gibco-BRL) with 10 % FBS. The day before H. pylori infection or cell/BCS co-culture, fresh medium with 2 % FBS was substituted. Eighteen hours later, cells grown to 90 % confluency were infected with H. pylori isogenic strains at a m.o.i. of 10 in culture medium for 2, 6, 10 or 24 h, or co-cultured with different BCS.
Total RNA isolation.
SGC7901 gastric cancer cells co-cultured with H. pylori strain NCTC 11638 or 11638-vacA were collected at 2, 6, 10 and 24 h after infection for RNA isolation. Total RNA was isolated using Trizol reagent (Life Technologies) according to the manufacturer's instructions.
Construction of microarrays.
The cDNA microarrays were designed by Shanghai BioStar Genechip (www.chinagenenet.com). In this study, microarrays with 8464 human cDNAs were used, including full-length and partial complementary DNAs representing novel, known and control genes.
Preparation and hybridization of fluorescent-labelled cDNA.
Aliquots of 30 µg total RNA were fluorescently labelled with Cy5- or Cy3-dCTP (Amersham Pharmacia) by reverse transcription in the presence of 10 µg oligo(dT) and 1 µl SuperScript II (50 U µl–1; Gibco-BRL). The labelled cDNAs were purified using MicroSpin S-200 columns (Amersham Pharmacia) and lyophilized. The probes were resuspended in 20 µl hybridization solution containing 8 µg poly(dA), 2 µg yeast tRNA and 10 µg human CotIDNA (Gibco-BRL). After heating to 95 °C for 2 min and then cooling to room temperature, the mixture was applied to slides and covered by a coverslip. Slides were incubated in a humid cabinet for 16–18 h at 42 °C in an incubator. Slides were washed at 60 °C for 10 min in solutions of 2x SSC with 0.2 % SDS, 0.1x SSC with 0.2 % SDS and 0.1x SSC sequentially, and were then dried at room temperature.
Array scanning and data processing.
Every slide was scanned at 10 µm resolution on a GenePix 4000B scanner (Axon Instruments) at variable PMT voltage to obtain maximal signal intensities with no more than 1 % probe saturation. The images were processed with GenePix Pro 3.0. Ratios were normalized by a linear regression between ln(Cy5) and ln(Cy3) of all the genes on the microarray. Genes exhibiting a 2.5-fold or greater change in expression level at at least one time point and exceeding 200 in signal intensity were considered true outliers.
Preparation of 32P-labelled probes.
The plasmids containing cDNA clones used for preparing probes were provided by Shanghai BioStar Genechip. Two hundred nanograms of plasmids were used as templates for PCR amplification. PCR products were purified using QIAquick gel extraction kit (Qiagene). The probes were labelled with [
-32P]dCTP by random priming using Strip-EZ DNA labelling system (Ambion).
Northern blot analysis.
A 10 µg sample of total RNA per lane was subjected to electrophoresis on 1.2 % agarose gels containing 2.2 M formaldehyde. RNAs were transferred onto Zeta-probe blotting membranes (Bio-Rad) using a vacuum blotter (model 785, Bio-Rad) and baked under vacuum at 80 °C for 2 h. Membranes were hybridized for 16 h at 60 °C with ULTRArray hybridization solution (Ambion) containing cDNA probes labelled with [-32P]dCTP by random priming (Strip-EZ DNA labelling system; Ambion). The hybridized membranes were serially washed at 55 °C using 2x SSC/0.1 % SDS solution, and then exposed to a phosphorimager. Blots were scanned and quantified by a phosphorimager in combination with Optiquant software version 2.50 (Cyclone Storage Phosphor System; Packard Instruments).
Human IL8 immunoassay using ELISA kit.
Cell culture supernatants collected as above were centrifuged at 400 g at room temperature for 10 min. Then, IL8 in each supernatant was assayed in duplicate according to the manufacturer's instruction (R&D Systems).
Flow cytometry for detection of apoptosis.
Cells were digested by 0.05 % trypsin, washed with PBS, and fixed in 70 % ethanol at 4 °C. Following filtration, cells were stained by propidium iodide (PI) containing 100 µg RNase ml–1 and apoptosis was detected by flow cytometry (Aoki et al., 2002; Yuan et al., 2003b)
Fluorescent staining and laser scanning confocal microscopy of microtubules and actin.
Cells from monolayers were fixed in 4 % citromint, processed with 0.3 % Triton X-100, and washed with PBS in the intervening time. Subsequently, cells were incubated with a primary antibody, monoclonal mouse anti-ß-tubulin (Sigma) for 1 h, and then incubated with a second antibody, FITC-conjugated goat anti-mouse (Sigma) for another 1 h. Following immunofluorescent staining, cells were stained by rhodamine phalloidin (Molecular Probes) for 40 min. Afterwards, cells were observed with a laser scanning confocal microscope (LSM-510; Zeiss). The cytoskeletal elements were examined for their overall morphology, orientation, distribution and disruption.
|
RESULTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
vacA. More than 300 genes altered their mRNA expression levels at different time points. According to their functions related to the pathogenesis of H. pylori infection, some of the genes were selected and classified for further study, among which, 68 were related to cytoskeleton (Table 1) and 87 were associated with cell cycle, cell death and proliferation (Table 2). IL8 mRNA expression was also found to be up-regulated by more than fivefold 10 h after the infection.
|
|
Confirmation of the differential expression of genes by Northern blot
To confirm the differential expression profiling of the cDNA microarrays, two genes, TUBB4Q and HSPA1L were chosen for Northern blot analysis. As shown in Fig. 1, the expression levels of these two genes were lower in cells treated with H. pylori NCTC 11638 than those treated with 11638-vacA. The level of ß-actin transcript was approximately the same in both samples, providing an assessment of RNA content in each sample used for Northern blot analysis.
|
IL8 protein expression was induced upon VacA stimulation
IL8 in the culture supernatants of SGC7901 cells co-incubated with NCTC 11638 or 11638-vacA at different time points were detected by ELISA in duplicate. The supernatants of cells co-incubated with NCTC 11638 showed significantly higher IL8 expression levels than those co-incubated with 11638-vacA from 6 h post-infection (Fig. 2).
|
Cells co-cultured with BCS of H. pylori NCTC 11638 show relatively lower apoptosis rates than those co-cultured with BCS of 11638-vacA
Cells were collected at 12, 18, 24, 30, 36, 42 and 48 h after co-culture with BCS of H. pylori NCTC 11638 or 11638-vacA, followed by apoptosis detection by flow cytometry. This experiment was carried out in duplicate. As shown in Fig. 3, cells co-cultured with BCS of H. pylori NCTC 11638 showed a relatively lower apoptosis rate than those co-cultured with BCS of 11638-vacA.
|
Cells co-cultured with BCS of H. pylori NCTC 11638 show more alteration in cytoskeleton morphology, especially in the assembly of tubulin and architecture of the microtubule cytoskeleton
The intracellular distribution of the microtubule (tubulin-based) cytoskeleton visualized with a monoclonal anti-tubulin antibody was captured by laser scanning confocal microscopy in SGC7901 cell monolayers co-cultured with BCS of H. pylori isogenic strains, at different time points. Cells co-cultured with BCS of 11638-vacA showed intact microtubules, which were dispersed in a radial fashion throughout the cytosol. On the other hand, cells co-cultured with BCS of vacA+ H. pylori NCTC 11638 showed collapse and disruption of the microtubule cytoarchitecture (Fig. 4). The staining pattern for filamentous actin did not show any significant differences.
|
|
DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
vacA strain. Therefore, differential expression may be considered as the result of H. pylori VacA stimulation.
VacA and cytoskeleton
It was demonstrated that toxin-producing H. pylori delays healing of acute gastric lesions (Konturek et al., 2000), and in vitro interferes with EGF-induced signal transduction, which is essential for gastric mucosal healing (Fujiwara et al., 1997). The underlying mechanism was determined by Pai et al. (1999), who demonstrated that VacA disrupted cytoskeletal architecture that is essential for the maintenance of cell structural integrity and epithelial barrier function. In the present microarray analysis, it was discovered that cytoskeleton-related gene expression changed a great deal.
Although various actin-related genes altered their expression, cells co-cultured with BCS of a vacA+ H. pylori strain showed a staining pattern for filamentous actin similar to those co-cultured with BCS of the vacA– isogenic strain. This may be because the morphological change of actin filaments is a tardigrade course, and 36 h is not long enough to observe their disruption. However, the microtubule cytoskeleton showed obvious disruption upon VacA stimulation early, by 12 h post-co-culture, consistent with the results of differential expression. This suggests that the morphological change of microtubules occurs much faster than that of actin filaments.
The mRNA expression of some HSP70s (HSPA1L, HSPA8) was down-regulated. Overexpression of constitutive HSP70 was found to significantly maintain microtubular integrity after simulated ischemia (Bluhm et al., 1998). Additionally, HSP70 may stabilize actin filaments in cultured cells (Macejak & Luftig, 1991), modulate actin formation and dynamics associated with the TCP-1 complex (Kubota et al., 1994), and interact with intermediate filaments (Liao et al., 1995). A relationship has been established between HSP70 and H. pylori infection. Konturek et al. (2001) found that gastric mucosal HSP70 mRNA and protein expression were decreased in humans and mice during infection with H. pylori. Together with the present results, it is reasonable to assume that it is VacA that made a pivotal contribution to the loss of HSP70.
The mRNA expressions of some other microtubule-associated genes were decreased, resulting in the destruction of the microtubule cytoskeleton observed in this study. For example, members of the - and ß-tubulin classes form heterodimers and are the principle components of microtubules. TUBB4Q is one of the members of the family with high sequence homology to functional ß-tubulin gene (van Geel et al., 2000).
VacA and the regulation of cell death and proliferation
The gastric mucosal integrity is maintained through a balance between the proliferation and apoptosis of mucosal cells. In H. pylori-induced gastritis, a mild increase in epithelial proliferation and a mild to moderate increase in apoptosis are common features (von Herbay & Rudi, 2000). It is well accepted that H. pylori-induced apoptosis may play a key role in gastric carcinogenesis by stimulating cell proliferation and/or resulting in gastric atrophy. Thus, alteration of gastric epithelial cell turnover resulting from H. pylori infection is probably an important process involved in the road to carcinogenesis (Xia & Talley, 2001).
The activation of NF-
B has been linked with suppression of apoptosis via the induction of expression of anti-apoptotic genes such as those encoding TRAF and c-IAP. TRAF2 recruits cellular inhibitor of apoptosis protein-1 (cIAP-1) and cIAP-2 (Chen & Goeddel, 2002). In the present study, increased expression of NF-B by 6 h post-infection lead to increased expression of TRAF over the following time-course, suggesting the potential anti-apoptotic effect of VacA through activation of NF-B. TRAFs are able to activate NF-B-inducing kinase (NIK), a member of the MEKK family, which has been shown to activate both IKK and IKKß, leading to the phosphorylation and degradation of IB and activation of NF-B (Maeda et al., 2000); thus establishing a positive feedback in the NF-B signalling pathway. cIAP-2, also named hIAP-1 or MIHC, continuously shows increased expression after infection, suggesting a direct response to VacA stimulation. Through directly binding to and inhibiting caspase activity, this protein may effectively inhibit apoptosis (Goyal, 2001). In addition to being activated by NF-B discussed above, hIAP-1 is known as an NF-B activator, thus establishing a signal amplification loop that promotes cell survival (Chu et al., 1997). It is well known that overexpression of anti-apoptotic determinants can promote tumorigenesis. Therefore, a link may be established between VacA and tumorigenesis via the induction of expression of some anti-apoptotic genes.
DNA damage derived from oxidative stress is another tumorigenic factor attributed to H. pylori infection. The differential expression of aldehyde oxidase 1 (AOX1), which produces hydrogen peroxide and, under certain conditions, can catalyse the formation of superoxide, shows a 16-fold increase upon VacA stimulation. H. pylori is able to directly induce reactive oxygen species (ROS) synthesis in gastric cells (Bagchi et al., 1996), and ROS enhances the expression of oncogenes, stimulates cell proliferation and plays an important role in all stages of carcinogenesis (Trush & Kensler, 1991). In a recent study, ROS production and DNA fragmentation were compared between H. pylori strain 60190, a VacA-producing strain, and its isogenic mutant strain in which the vacA gene had been disrupted. The former induced much greater production of ROS and DNA fragmentation in human gastric mucosal cells than the latter did (Bagchi et al., 2002). NF-B was also involved in oxidative-stress-mediated cell injury. A variety of antioxidants have been demonstrated to inhibit the activation of NF-B (Schreck et al., 1992), and micromolar concentrations of H2O2 could activate NF-B, suggesting that reactive oxygen may act as a second messenger in the activation of transcription factor NF-B (Winrow et al., 1993). Thus, a plausible hypothesis for the link between VacA and gastric carcinogenesis involves ROS-induced DNA damage.
As for the possible role of virulence factors of H. pylori, conflicting results have been reported. In one study, patients infected with cagA+ vacA-s1a strain of H. pylori showed significantly higher gastric epithelial proliferation and lower apoptotic index than those infected with virulence-negative strains. This dissociation of proliferation from apoptosis was suggested as a possible mechanism of carcinogenesis (Peek et al., 1997). Some other studies found much more apoptosis in cells infected with a cagA+ vacA-s1a/m1 strain of H. pylori than those infected with cagA– vacA-s2/m2 strain, or found VacA specifically inhibited cell proliferation (Peek et al., 1995; Rudi et al., 1998). However, Wagner et al. (1997) indicated that inhibition of epithelial cell proliferation was a general phenomenon of H. pylori that was induced by all H. pylori strains, irrespective of the presence of the cytotoxin. Furthermore, it has been proposed that H. pylori may induce hyperproliferation through increasing apoptosis, i.e. increased apoptosis may be the stimulus for a compensatory hyperproliferative and potentially preneoplastic response in chronic H. pylori infection (Moss et al., 1996). In the present differential expression results, we found that many proapoptotic and anti-apoptotic genes increased their mRNA expression at one time, the number of the latter being much more than that of the former; and by 24 h post-infection, mRNA expression of a proliferation-potential-related protein began to increase. In fact, apoptosis detection by flow cytometry determined that VacA does have the potential to inhibit apoptosis, because the cells co-cultured with BCS of vacA+ wild-type H. pylori showed a relatively lower apoptosis rate as compared with those co-cultured with BCS of the vacA– isogenic mutant strain.
VacA and IL8 production
Since Crabtree et al. (1993) reported that IL8 was increased in H. pylori-infected gastric mucosa, considerable evidence has accumulated to confirm that IL8 plays a major role in H. pylori-associated acute inflammatory response (Crabtree et al., 1994; Crowe et al., 1995). Previous studies have provided variable findings regarding the bacterial factors that induce IL8 production in host cells. Backert et al. (2004) suggested that at least two pathways, a cagPAI-dependent and a cagPAI-independent pathway, may exist. Many virulence factors other than cagPAI may be involved in IL8 induction, including VacA, IceA, urease or even ammonia (Straubinger et al., 2003). However, there is controversy regarding the role of VacA as an IL8 inducer. In several studies, H. pylori SS1 strain, which is cagPAI+ but does not produce vacuolating cytotoxin, induced only a modest IL8 response, while two other strains, SPM326 and LC11, with the cagA+ genotype and the ability to produce VacA, could induce much greater IL8 production in vitro (van Doorn et al., 1999; Eaton et al., 2001; Day et al., 2001). In an in vivo study using H. pylori strains that were vacA+ and cagPAI–, IL8 was found to be significantly upregulated in H. pylori-infected cats (Straubinger et al., 2003). de Jonge et al. (2001) also found a strain of H. pylori that is vacA-s1am1 and cag-negative which induced a large amount of IL8 production in KATO-III cells. On the contrary, Kim et al. (2000) suggested that the presence of Vac production may not influence the expression level of IL8 genes in human gastric epithelial cells. In another in-depth study, H. pylori strain CCUG 17874 and its isogenic cytotoxin-negative mutant 17874vacA stimulated the same amount of IL8 in KATO-III cells, thus ruling out the cytotoxin as the stimulatory agent (Huang et al., 1995). However, in our study, we show that VacA+ H. pylori induced up to a fivefold increase in IL8 mRNA expression and significantly higher IL8 protein production compared with the isogenic VacA– strain. Moreover, the up-regulated expressions of NF-B and AOX1 accord with the current dogma that NF-B and oxidative stress are involved in the transcriptional activation of IL8 gene (Keates et al., 1997; Shimada et al., 1999). This result differs from what was observed by Naumann et al. (1999), i.e. the isogenic mutant of a wild-type vacuolating-cytotoxin-producing H. pylori strain carrying a knockout of vacA gene does not affect NF-B activation. What results in such a discrepancy remains unclear. One possibility is that we have used a gastric cancer cell line different from those used in previous studies. Currently, both bacterial factors and host responses to infection are considered to play a role in determining disease outcome (Day et al., 2001). We also hypothesize here that hydrogen peroxide produced by AOX1 modulates IL8 gene expression through an NF-B independent pathway, as was recently described by Ito et al. (2004).
In summary, we report here the comparison of a large screening of differentially expressed genes in gastric cells induced by vacA+ H. pylori or the vacA– isogenic mutant strain. Many genes previously not known to be related to H. pylori cytotoxin VacA have been identified, providing a database resource for future studies in vitro and in vivo.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
B control. Proc Natl Acad Sci U S A 94, 10057–10062.B in gastric epithelial cells. Gastroenterology 113, 1099–1109.[CrossRef][Medline]B through a signaling pathway involving IB kinases, NF-B-inducing kinase, TRAF2, and TRAF6 in gastric cancer cells. Gastroenterology 119, 97–108.[CrossRef][Medline]This article has been cited by other articles:
|
|
 |
|
  J. Hisatsune, M. Nakayama, H. Isomoto, H. Kurazono, N. Mukaida, A. K. Mukhopadhyay, T. Azuma, Y. Yamaoka, J. Sap, E. Yamasaki, et al. Molecular Characterization of Helicobacter pylori VacA Induction of IL-8 in U937 Cells Reveals a Prominent Role for p38MAPK in Activating Transcription Factor-2, cAMP Response Element Binding Protein, and NF-{kappa}B Activation J. Immunol., April 1, 2008; 180(7): 5017 - 5027. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
