Fax: +81 3 5565 1753
Early Detection and Diagnosis
The presence of a methylation fingerprint of Helicobacter pylori infection in human gastric mucosae
Article first published online: 24 SEP 2008
Copyright © 2008 Wiley-Liss, Inc.
International Journal of Cancer
Volume 124, Issue 4, pages 905–910, 15 February 2009
How to Cite
Nakajima, T., Yamashita, S., Maekita, T., Niwa, T., Nakazawa, K. and Ushijima, T. (2009), The presence of a methylation fingerprint of Helicobacter pylori infection in human gastric mucosae. Int. J. Cancer, 124: 905–910. doi: 10.1002/ijc.24018
- Issue published online: 11 DEC 2008
- Article first published online: 24 SEP 2008
- Accepted manuscript online: 24 SEP 2008 12:00AM EST
- Manuscript Accepted: 11 SEP 2008
- Manuscript Received: 15 JUN 2008
- The Ministry of Health, Labour and Welfare Japan (for the Third-term Comprehensive Cancer Control Strategy Pioneering Basic Research and Cancer Research)
- DNA methylation;
- Helicobacter pylori;
- molecular epidemiology
Aberrant DNA methylation is deeply involved in human cancers, but its inducers and targets are still mostly unclear. Helicobacter pylori infection was recently shown to induce aberrant methylation in gastric mucosae, and produce a predisposed field for cancerization. Here, we analyzed the presence of target genes in methylation induction by H. pylori and the mechanism for the gene specificity. Noncancerous gastric mucosae were collected from 4 groups of individuals (with and without a gastric cancer, and with and without current H. pylori infection; N = 11 for each group), and methylation of promoter CpG islands of 48 genes that can be methylated in gastric cancer cell lines was analyzed by methylation-specific PCR. In total, 26 genes were consistently methylated in individuals with current or past infection by H. pylori, whereas 7 genes were not methylated at all. In addition, 14 genes were randomly or intermediately methylated in individuals with gastric cancers and the remaining 1 gene was methylated in all the cases. The methylation-susceptible genes had significantly lower mRNA expression levels than the methylation-resistant genes. H. pylori infection did not induce mRNA and protein expression of DNA methyltransferases; DNMT1, DNMT3A or DNMT3B. Gene specificity was present in the induction of aberrant DNA methylation by H. pylori infection, and low mRNA expression, which could precede methylation, was one of the mechanisms for the gene specificity. These findings open up the possibility that a methylation fingerprint can be used as a novel marker for past exposure to a specific carcinogenic factor. © 2008 Wiley-Liss, Inc.
Aberrant DNA methylation is deeply involved in human cancer development and progression.1 In some cancer types, such as gastric cancers, tumor-suppressor genes are more frequently inactivated by aberrant DNA methylation than by mutations.2 Nevertheless, only limited information is available for inducers of aberrant DNA methylation, which include aging, viral infection and ulcerative colitis.3, 4 Also, almost no information is available for gene specificity in methylation induction by a specific factor. Using cancer tissues, it is very difficult to clarify an association between a specific inducer and methylation of a gene. Aberrant methylation of a gene can be present in cancer tissues because its methylation conferred a growth advantage although it was a rare and random event, or because its methylation was carried over from a precursor tissue to a cancer tissue since it was frequently induced in the precursor tissue. In contrast, using a noncancerous tissue, one can assess an effect of a methylation inducer by the fraction of cells with methylation in the polyclonal tissue.
Gastric mucosa infected by Helicobacter pylori is a useful model to examine the possible presence of gene specificity in methylation induction by a specific factor since H. pylori infection was recently shown to induce aberrant DNA methylation potently in gastric mucosae.5 Moreover, the fraction of DNA molecules with aberrant methylation (methylation level) in gastric mucosae of individuals without current H. pylori infection was correlated with gastric cancer risk,5, 6 indicating that methylation in noncancerous tissues is related to gastric carcinogenesis. So far, 6 CpG islands in gene promoter regions methylated in gastric cancers7 were analyzed, and all were methylated in gastric mucosae with current and past infection with H. pylori. However, it is unknown whether these 6 genes are preferentially methylated by H. pylori infection or H. pylori infection induces methylation of random genes.
In this study to analyze the presence of gene specificity for methylation induction, firstly we examined the methylation status of 48 promoter CpG islands in the noncancerous gastric mucosae of 4 groups of individuals (with and without a gastric cancer, and with and without current H. pylori infection). The 48 genes were selected as genes that can be methylation-silenced in gastric cancer cell lines8 because the vast majority of CpG islands in gene promoter regions are not methylated at all in noncancerous tissues, and we had to newly select genes that have better chances to be methylated in noncancerous tissues. Secondly, we analyzed an association between susceptibility to methylation induction and mRNA expression levels in normal tissue without and with H. pylori infection.
Material and methods
Tissue samples and DNA/RNA extraction
For methylation analysis, (noncancerous) gastric mucosa samples were collected from 4 groups of individuals (with and without a gastric cancer, and with and without current H. pylori infection; N = 11 for each group, average age = 60.8 ± 13.8 years). For analysis of mRNA expression that determines gene specificity of methylation induction, we need to analyze the mRNA expression level in gastric mucosae free of methylation, which, once induced, will cause decreased gene transcription to avoid confusion between cause and consequence. Therefore, samples were collected from 11 healthy volunteers, who were considered to have less chance for methylation induction by H. pylori than elderly individuals (7 males and 4 females; 6 with H. pylori infection and 5 without; average age = 34.8 ± 3.1 years). Biopsy specimens were taken from one standard site of the stomach (antral regions in the lesser curvature) using sterilized biopsy forceps (Olympus, Tokyo, Japan). H. pylori infection status was analyzed by culture test (Eiken, Tokyo, Japan) and rapid urease test (Otsuka, Tokushima, Japan). All the materials were obtained with written informed consents, and the procedures were approved by the institutional review board. High molecular weight DNA was extracted by the standard phenol/chloroform method and total RNA was isolated using ISOGEN (Nippon Gene, Tokyo, Japan) and an RNeasy Mini kit (Qiagen, Valencia, CA).
Cell lines and 5-aza-dC treatment
Gastric cancer cell lines, AGS and KATO-III, were obtained from the Japanese Collection of Research Bioresources (Tokyo, Japan) and the American Type Culture Collection (Manassas, VA). For treatment with a demethylating agent, 5-aza-2′-deoxycitidine (5-aza-dC, Sigma, St. Louis, MO), cells were seeded on day 0, media containing 0.3 μM 5-aza-dC was freshly added on days 1 and 3, and cells were harvested on day 5. Genomic DNA and total RNA were isolated in the same way as the primary samples.
Bisulfite treatment and methylation-specific PCR
Bisulfite treatment was performed as previously described.9 Briefly, DNA samples (1 μg each) digested by BamHI were denatured in 0.3 N NaOH at 37°C for 15 min. The samples underwent 15 cycles of 30-sec denaturation at 95°C and 15-min incubation at 50°C in 3.1 N sodium bisulfite (pH 5.0) and 0.5 mM hydroquinone. The samples were desalted with the Wizard DNA Clean-Up system (Promega, Madison, WI), and desulfonated in 0.3 N NaOH. DNA was ethanol precipitated and dissolved in 40 μL of TE buffer.
Methylation-specific PCR (MSP) was performed with a primer set specific to the methylated or unmethylated sequence (M or U set), respectively,8 using 2 μL of the sodium bisulfite-treated DNA. A region upstream of a putative transcriptional start site (200 bp or less) was analyzed, and CpG maps of all the genes are shown in the Supporting Information Figure 1. DNA methylated with SssI methylase was used to determine a specific condition of PCR for the M set, and DNA amplified by a GenomiPhi DNA amplification kit (GE Healthcare Bio-Sciences) was used for the U set. A number of PCR cycles that would yield a minimal visible band was determined using these fully methylated DNA (for M primers) and fully unmethylated DNA (for U primers), and a further 4 cycles were added for actual analysis of test samples. Methylation levels were classified as none (−), low (+), high (++) according to the intensity of the band for methylated DNA molecules compared with that for unmethylated DNA, respectively.
Quantitative reverse transcription PCR
cDNA was synthesized from 1 μg of total RNA using a Superscript II kit (Life Technologies, Rockville, MD) with a random primer. Real-time PCR was performed using an iCycler Thermal Cycler (Bio-Rad Laboratories, Hercules, CA) with SYBR Green I (BioWhittaker Molecular Applications, Rockland, ME). The number of molecules of a specific gene in a sample was measured by comparing its amplification with that of standard samples, which contained 101–107 copies of the gene. The standard samples were produced by PCR amplification and purification using Zymo-Spin I™ Columns (Zymo Research, Orange, CA). The amount of the standard samples was measured by OD 260 nm and also by quantification of band intensities after electrophoresis. The mRNA quantity of each gene was normalized to that of β-actin. The primers and PCR conditions are shown in the Supporting Information Table 1. The difference of mRNA expression levels between 2 groups of genes was analyzed by the Welch t-test method (both sided).
Western blot analysis
Each 100 μg whole-cell lysate sample was subjected to SDS-PAGE (10% acrylamide gel) and blotted to PVDF membrane. DNMT1 and DNMT3A were detected using rabbit polyclonal antibody against human DNMT1 (NEB, Beverly, MA), human DNMT3A (Cell Signaling Technology, Danvers, MA), respectively at 1/1,000 dilution. DNMT3B was detected using goat polyclonal antibody against human DNMT3B (Santa Cruz Biotechnology, Santa Cruz, CA) at 1/500 dilution. Horse radish peroxidase-conjugated secondary antibody (antirabbit; Cell Signaling Technology, antigoat; Santa Cruz Biotechnology) was used at 1/5,000 dilution.
Confirmation of gene silencing due to promoter CpG islands
The 48 genes consisted of 32 randomly and 16 arbitrarily selected genes from 421 genes that had been identified as methylation-silenced genes in a gastric cancer cell line using microarray analysis of cells treated with 5-aza-dC, and MSP analysis.8 First, we analyzed mRNA expression of 7 of the 48 genes before and after 5-aza-dC treatment using real-time RT-PCR (Fig. 1). It was confirmed that no or little mRNA expression was present in cell lines without unmethylated DNA molecules and that mRNA expression was upregulated by the 5-aza-dC treatment.
Gene specificity in methylation induction by H. pylori infection in gastric mucosae
We then analyzed the methylation status of the promoter CpG islands of the 48 genes in the (noncancerous) gastric mucosae of 4 groups of individuals; those with and without H. pylori infection and with and without a gastric cancer. Since MSP can produce inconsistent results if inappropriately performed, we carefully selected a PCR cycle for each primer set so that false positive and negative results were not produced. We scored the methylation status as negative, weakly positive or positive by comparing the band density with that of a fully methylated control (representative results in Fig. 2a).
When all the genes were aligned in the order of number of samples with methylation (Fig. 2b), the 48 genes were divided into 3 groups: (i) 7 genes that were completely unmethylated in any of the 4 groups (genes 1–7 in Fig. 2b; methylation-resistant genes), (ii) 14 genes that were methylated randomly or more frequently in individuals with cancers (genes 9–21; intermediate genes); and (iii) 26 genes that were consistently methylated in the individuals with H. pylori infection or with a gastric cancer (genes 22–47; methylation-susceptible genes). The remaining 1 gene, PLAGL1, was methylated in all the individuals. This demonstrated that some genes are resistant to methylation induction by H. pylori infection while others are susceptible, namely the presence of gene specificity in methylation induction.
Lack of association between CpG island configuration and methylation susceptibility
The 48 genes analyzed here all had CpG islands in their promoter regions. However, based on recent reports,10 there was a possibility that, even among CpG islands, their configurations (number of CpG sites, G+C content, and CpG score) might influence the susceptibility of individual genes to methylation induction by H. pylori. Therefore, we examined their configurations in 300 bp upstream regions from transcription start sites (Fig. 2b), which corresponded to the nucleosome-free region and whose methylation is critical for gene silencing.11, 12
The number of CpG sites in the region was 29.2 ± 10.4 (mean ± standard deviation) and 25.4 ± 9.3 for the susceptible and resistant genes, respectively (p = 0.38). The G + C content was 68.4 ± 7.4 and 66.4 ± 7.9% for the susceptible and resistant genes, respectively (p = 0.52). The CpG score was 0.82 ± 0.18 and 0.75 ± 0.21 for the susceptible and resistant genes, respectively (p = 0.40). In short, no significant difference was present between the 2 groups.
Involvement of low mRNA expression levels in gene specificity in methylation induction
To investigate an association between the gene specificity in methylation induction and mRNA expression levels in gastric mucosae, we analyzed mRNA expression levels of all of the 7 methylation-resistant and 15 methylation-susceptible genes, which were randomly selected from the 26 methylation-susceptible genes. To compare mRNA expression levels among different genes, the numbers of cDNA molecules were measured by quantitative RT-PCR after accurate measurement of the weights (converted into the numbers of DNA molecules) of standard DNA samples of all the genes. mRNA expression levels were analyzed in the gastric mucosae of young healthy individuals with and without H. pylori infection, who were considered to have no or little methylation of the genes analyzed.
The average mRNA expression level of methylation-resistant genes was much higher than that of methylation-susceptible genes among individuals without H. pylori infection (4.3 × 10−2vs. 7.3 × 10−3; p = 0.0008) and also among individuals with H. pylori infection (2.9 × 10−2vs. 5.1 × 10−3; p = 0.0012) (Fig. 3). Three of the 7 resistant genes and 5 of the 15 susceptible genes showed a significant decrease of mRNA expression levels by H. pylori infection, but no genes showed significantly increased mRNA expression.
Expression levels of DNA methyltransferase
To gain an insight into how H. pylori infection induces aberrant methylation, we analyzed mRNA expression levels of maintenance DNA methyltransferase, DNMT1, and de novo methyltransferases, DNMT3A and DNMT3B, in the gastric mucosae with and without H. pylori infection. However, no significant increase in their mRNA expression levels was observed (Fig. 4a). Further, at the protein level, expression levels of DNMT1, DNMT3A and DNMT3B were below the detection limit even in the gastric mucosae with H. pylori (Fig. 4b), indicating no increase was induced by H. pylori infection.
The presence of gene specificity for aberrant DNA methylation induction by a specific carcinogenic factor was demonstrated for the first time in this study. Also, genes susceptible to methylation had significantly lower mRNA expression levels than resistant genes. For clarification of the relationship between a methylation-inducing factor and gene specificity, use of noncancerous gastric tissue, which is polyclonal, was important because gene silencing due to promoter methylation can result in over- or under-presence of methylation in cancer tissues. Methylation in noncancerous tissues is also reported in the colonic mucosae of patients with ulcerative colitis13, 14 and liver tissues of patients with hepatocellular carcinomas,15 but limited numbers of genes have been analyzed so far.
Methylation of specific genes can persist for a lifetime, and there is a possibility that the methylation profile can be used as a methylation fingerprint of H. pylori infection in the past, as specific p53 and APC mutations are used to assess past exposure to specific carcinogens.16, 17 Use of DNA methylation has an advantage over mutations because methylation can be present in a significant fraction of cells in noncancerous tissues, and can be detected sensitively and reproducibly. The noncancerous gastric mucosae of cases with a gastric cancer without current H. pylori infection, most of which are considered to have had past exposure to H. pylori,18 showed the same methylation profile as individuals with current H. pylori infection. This finding indicated that the methylation profile induced by H. pylori infection can persist even after discontinuation of H. pylori infection. Although eradication of H. pylori was reported to decrease incidences of individuals with methylation,19, 20 the decrease is only partial, not to zero, and highly variable among individuals (manuscript in preparation).
To establish a methylation profile as a fingerprint of H. pylori infection, the profile must be specific. Unfortunately, few gastric cancers can be considered as those induced solely by another carcinogenic factor, such as Epstein-Barr virus infection21 or high salt intake,22 and the specificity cannot be examined easily. However, since low mRNA expression levels are involved in gene specificity, there is a possibility that different carcinogenic factors induce different methylation profiles through induction of reduced mRNA expression of different genes. Once the specificity of a methylation profile is established, a methylation fingerprint will be very useful for clinicopathological analysis and epidemiology. Among the clinically used tests for H. pylori infection, the culture and rapid urease tests can detect only current H. pylori infection.23, 24 The serum antibody test remains positive in only half the patients as early as 1 year after successful eradication of H. pylori.25, 26
The role of low mRNA expression in methylation induction has been reported.4 De Smet et al. showed that weak transcriptional capacity leads to promoter remethylation by analysis of demethylation and mRNA expression of MAGE-A1 in various cell lines.27 Song et al. showed that decreased promoter activity leads to hypermethylation of a promoter CpG island of an exogenously introduced gene by disrupting its promoter activity.28 We and others previously observed that most genes methylated in cancer tissues had no or little expression in cancer precursor cells.29–32 This study showed that, in normal cells and in vivo, low mRNA expression is important for methylation induction. Also, it was suggested that downregulation by H. pylori infection precedes methylation since 8 of the 22 genes with expression analyses were downregulated by H. pylori infection, but none were upregulated. The 22 genes were selected from those that can be methylated in gastric cancer cell lines and even the resistant genes are considered to be relatively susceptible among the entire genes.
Even among the genes with similarly low mRNA expression levels, some genes were resistant and others were susceptible to methylation induction by H. pylori. As additional factors that determine the gene specificity of methylation induction, histone modification deregulation could be important. For example, a repressive histone modification, methylation at Lys27 of histone H3 (H3K27) induced by Polycomb group proteins, is associated with genes methylated in cancers.33, 34 Active chromatin marks, associated with active mRNA expression, could be important to protect DNA from methylation. At the same time, without H. pylori infection, even the susceptible genes were not methylated, indicating that abnormality in epigenetic regulation was induced by H. pylori infection. The final step of aberrant methylation must be mediated by DNA methyltransferases, and actually overexpression of de novo methyltransferases enhance methylation of specific genes in a mouse model.35 Also, some inflammatory cytokines, such as IL-6, are reported to induce DNA methyltransferases.36 However, contrary to initial expectations, H. pylori infection did not induce either mRNA or protein expression of DNMT1, DNMT3A and DNMT3B in gastric mucosae. Abnormalities in epigenetic regulation induced by H. pylori infection also need to be investigated.
In summary, methylation of specific genes was induced by H. pylori infection in noncancerous gastric membranes, and preceding low mRNA expression was suggested to be involved in the specificity. Use of the specific profile as a methylation fingerprint of past exposure to a specific carcinogenic factor was suggested.
- 15Genetic instability and aberrant DNA methylation in chronic hepatitis and cirrhosis—a comprehensive study of loss of heterozygosity and microsatellite instability at 39 loci and DNA hypermethylation on 8 CpG islands in microdissected specimens from patients with hepatocellular carcinoma. Hepatology 2000; 32: 970–9., , , , , .
- 26Usefulness of serological IgG antibody determinations for confirming eradication of Helicobacter pylori infection. Am J Gastroenterol 1999; 94: 2105–08., , , , , , , , .Direct Link:
Additional Supporting Information may be found in the online version of this article.
|IJC_24018_sm_SuppFig1.tif||7357K||Supporting Information Figure 1. The maps of the genes analyzed. Vertical lines, individual GpC or CpG sites; Open boxes, non-coding and coding exons; Arrows; transcription start sites; and Arrowheads, positions of MSP primers. It can be seen that a region upstream of a putative transcriptional start site in CpG island was analyzed by MSP.|
|IJC_24018_sm_SuppTable.doc||69K||Supporting Information Table.|
Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.