DNA mismatch repair gene methylation in gastric cancer in individuals from northern Brazil

Gastric cancer is one of the most common malignancies. DNA methylation is implicated in DNA mismatch repair genes deficiency. In the present study, we evaluated the methylation status of MLH1, MSH2, MSH6 and PMS2 in 20 diffuseand 26 intestinal-type gastric cancer samples and 20 normal gastric mucosal of gastric cancer patients from Northern Brazil. We found that none of the nonneoplastic samples showed methylation of any gene promoter and 50% of gastric cancer samples showed at least one methylated gene promoter. Methylation frequencies of MLH1, MSH2, MSH6 and PMS2 promoter were 21.74%, 17.39%, 0% and 28.26% respectively in gastric cancer samples. MLH1 and PMS2 methylation were associated with neoplastic samples compared to nonneoplastic ones. PMS2 methylation was associated with diffuseand intestinal-type cancer compared with normal controls. Intestinal-type cancer showed significant association with MLH1 methylation. Diffuse-type cancer was significantly associated with MSH2 methylation. Our findings show differential gene methylation in tumoral tissue, which allows us to conclude that methylation is associated with gastric carcinogenesis. Methylation of mismatch repair genes was associated with gastric carcinogenesis and may be a helpful tool for diagnosis, prognosis and therapies. However, MSH6 does not seem to be regulated by methylation in our samples.


Introduction
Gastric cancer is the fourth most common malignancy and the second most common cause of cancer death worldwide (Parkin et al., 2002).In Northern Bra-zil, the state of Pará has a high incidence rate of this neoplasia and its capital, Belém, was ranked eleventh in number of gastric cancers per inhabitant among all cities worldwide with cancer records (INCA, 2005).
DNA methylation is the most studied epigenetic alteration, occurring through the addition of a methyl radical to the cytosine base adjacent to guanine (Bird, 2002).In cancer, DNA methylation of the promoter region of a normal tumor-suppressor gene leads to the aberrant silencing of its functions.
Genetic pathways involved in gastric cancer devel-opment are not clearly understood, although several genetic and epigenetic alterations have been implicated in DNA mismatch repair (MMR) genes deficiency, resulting in microsatellite instability (MSI) phenotype.The human MMR system repairs DNA replication errors or physicochemical damage.Microsatellite regions are susceptible to mutation due to slippage of DNA polymerase during DNA replication.Failure to excise these errors may lead to frameshift mutations in many target genes (Bacani et al., 2005).MLH1, MSH2, MSH6 and PMS2 are major MMR genes implicated in genetic stability (Gologan and Sepulveda, 2005).
The MSI phenotype that occurs in tumors has been demonstrated to be due to deficiency of one of the main DNA MMR proteins (MLH1 and MSH2) (Boland, 2000).Hypermethylation of MLH1 promoter CpG island was found responsible for the development of most gastric cancers exhibiting MSI (Fleisher et al., 1999).MLH1 may form heterodimers with PMS2 acting as molecular matchmakers, recruiting other proteins of MMR complex (Kunkel and Erie, 2005).However, PMS2 methylation status has never been studied in gastric carcinogenesis.MSH2 has also been implicated in gastric carcinogenesis (Kim et al., 2001;Kitajima et al., 2003).MSH2 may form heterodimers with MSH6 repairing single base-base and 1-2 base insertion-deletion mismatches (Kunkel and Erie, 2005).Few studies assessed MSH2 methylation frequency (Wu et al., 2000;Fang et al., 2004), but MSH6 methylation status has never been studied in gastric carcinogenesis.
The knowledge of the methylation status of DNA MMR can help understanding a major gastric carcinogenesis pathway.Tumor-specific epigenetic alterations can be used as molecular markers of malignancy, and thus provide more specific therapy.In the present study, we evaluated the methylation status of MLH1, MSH2, MSH6 and PMS2 promoters, as well as their potential association with clinical and pathological characteristics, such as gender, age, histopathology, tumor extension and presence of lymph nodes or distant metastasis.

Samples
The study included 66 gastric tissue samples.Of them, 20 were nonneoplastic gastric mucosa of gastric cancer patients (at a distance form the primary tumor) and 46 sporadic gastric cancer samples.Gastric samples were surgically obtained at João de Barros Barreto University Hospital (HUJBB) in the state of Pará.Patients had never been submitted to either chemotherapy or radiotherapy prior to surgery, nor had any other diagnosed cancer.All patients signed an informed consent approved by Research Ethics Committees of HUJBB and Ribeirão Preto Medical School Clinics Hospital.All 46 gastric cancer samples were classified according to Laurén (1965): 20 were diffuse-type and 26 were intestinal-type.Tumors were Primer sequences (5'-3') for MSP M: methylated sequence; U: unmethylated sequences.

Methylation-specific PCR (MSP)
Genomic DNA was extracted from gastric samples using phenol-chloroform extraction.Genomic DNA (2 μg) was modified by bisulfite treatment, converting unmethylated cytosines to uracils and leaving methylated cytosines unchanged.MSP was performed on treated DNA as previously described (Herman et al., 1996).Specific primers for MSP (Table 1), located within CpG islands in gene promoter, were designed using the Methprimer software (Li and Dahiya, 2002).
PCR was carried out in a volume of 50 μl with 200 μmol/L of dNTPs, 200 μmol/L of MgCl2, 50 ng of DNA, 200 pmol/L of primers and 1 unit of AmpliTaq GOLD (Applied Biosystems, Foster City, CA, USA).After initial denaturing for 2 min at 94ºC, 35 cycles at 94ºC for 40 s, 50-60ºC for 1 min and 72ºC for 40 s were carried out, followed by a final extension at 72ºC for 5 min.PCR products were separated and visualized using 8% polyacrylamide gel stained with 10% silver nitrate (Fig. 1).Water was used as negative control.MSP results were scored when there was a clearly visible band on electrophoresis gel with methylated and unmethylated primers (Herman et al., 1996).
Sequencing of a methylated and then an unmethylated PCR product of MLH1 was used as a control for DNA bisulfite treatment.For sequencing analysis, ABI PRISM Big Dye Terminator Cycle Sequencing Kit (Perking Elmer, Alameda, CA, USA) was used (Fig. 1).

Statistical analysis
Statistical analyses were performed using Fisher's exact test to assess associations between methylation status and frequency, and clinical and pathological characteristics, such as gender, age, histopathology, tumor extension and presence of lymph nodes or distant metastasis.A 5% significance level was set.

Results
Of 66 patients studied, 47 were males and 19 were females, and mean age was 59 ±9.8775 years (range 27-76).According to Laurén's classification, 20 were diffuse-type and 26 were intestinal-type.All gastric cancer samples were in advanced stage.
Sequencing of a methylated and an unmethylated PCR product of MLH1, used as controls for DNA bisulfite treatment, showed methylated and unmethylated (converted) cytosines, respectively, at CpG-sites when compared to non-treated sequence of DNA.At non-CpG sites, all cytosines were converted to thymine (Fig. 1).This excluded the possibility that successful amplification with methylated primers could be attributable to incomplete bisulfite conversion.
Table 2 shows clinical and pathological characteristics and methylation status of MLH1, MSH2, MSH6 and PMS2 promoters of the samples studied.None of the nonneoplastic samples showed methylation of any gene promoter.Of all gastric cancer samples, 50% showed at least one methylated gene promoter.Methy-lation frequencies of MLH1, MSH2, MSH6 and PMS2 promoters were 21.74%, 17.39%, 0% and 28.26% respectively in gastric cancer samples (Table 3).
MLH1 and PMS2 methylation were associated with gastric cancer samples compared to nonneoplastic samples (p=0.0256 and p=0.0065, respectively).A tendency for MSH2 promoter methylation in gastric cancer compared to normal controls was seen (p=0.0531)(Table 3).
Higher PMS2 methylation frequency was associated with diffuse-(p=0.0471)and intestinal-type gastric cancer (p=0.0065)compared with normal controls.Intestinal-type gastric cancer showed a significant association with higher MLH1 methylation frequency (p=0.0287).In diffuse-type gastric cancer samples, it was detected an association between this cancer type and higher MSH2 methylation frequency (p=0.0471)(Table 3).

TABLE 2.
Clinical and pathological characteristics and methylation status of MLH1, MSH2, MSH6 and PMS2 promoters of the studied samples.
White boxes: samples that showed unmethylated sequences; black boxes: samples that showed methylated sequences.T: size or direct extent of the primary tumor; N: degree of spread to regional lymph nodes; M: presence of metastasis.
In our study, we found that none of the samples with MLH1, MSH2 or PMS2 methylation showed unmethylation sequences using MS-PCR assay.MSP is a sensitive method able to detect at least 0.1% of methylated or unmethylated genes.Our results suggest that methylation of these genes is a common event in a subgroup of gastric cancers and may be related to early carcinogenesis.
We evaluated a CpG island of MLH1 promoter previously described as associated with its gene expression and MSI status (Herman et al., 1998).In the present study, MLH1 methylation was found in 21.74% of gastric cancer samples, which was not significantly different from previously described frequencies (13-41%) (Leung et al., 1999;Toyota et al., 1999;Fang et al., 2003;An et al., 2005;Truninger et al., 2005).Hypermethylation of the MLH1 promoter CpG island was found responsible for the development of the majority of gastric neoplasias exhibiting MSI (Fleisher et al., 1999) and was previously associated with high-degree MSI (MSI-H) in gastric cancer (Leung et al., 1999;Toyota et al., 1999;Wu et al., 2000;Fang et al., 2003;An et al., 2005).
We also found an association between MLH1 methylation and gastric carcinogenesis, especially in intestinal-type samples.Wu et al. (2000) described that MSI-H is significantly more common in antral location, intestinal subtype and H. pylori seropositivity samples.Thus, our data corroborate Wu et al.'s (2000) findings concerning the association between MLH1 methylation and intestinal-type gastric cancer.Kitajima TABLE 3.

Methylation number and frequency in gastric tissue samples, n (%), and p-values for comparison of gastric cancer with nonneoplastic samples.
M: methylated; U: unmethylated; a p<0.05.
The statistical analysis showed an association between methylation of MLH1, MSH2 or PMS2 promoter in gastric cancer (p<0.0001), and both diffuse-(p=0.0004)and intestinal-type (p=0.0001)compared to nonneoplastic samples.
We analyzed whether DNA methylation was associated with clinical and pathological characteristics, and detected a tendency for PMS2 methylation in intestinal-type gastric cancer with larger tumor extension (T3 and T4) (p=0.0622).

Discussion
There are at least two distinct genetic instabilities in gastric tumorigenesis.One is microsatellite instability (or MSI pathway) and the other is chromosomal instability (or suppressor pathway), including tumors with low-frequency MSI (MSI-L) as well as microsatellite stable (MSS) with accumulation of loss of tumor suppressor genes, such as TP53, RB, APC, MCC and DCC.The MSI pathway consisting of a gastric cancer subset with high-frequency MSI (MSI-H) and the defective repair of mismatched bases induces increased mutation rate at the nucleotide level, and the consequent widespread MSI (Martins et al., 1999;Wu et al., 2000;Fang et al., 2001).In the present study, 50% of gastric cancer samples showed at least one methylated gene promoter and methylation of MLH1, MSH2 or PMS2 promoter was significantly associated with this cancer, suggesting that this tumor develops by a MSI pathway. et al. (2003) also described that loss of MHL1 expression is associated with differentiated gastric cancer samples, which is roughly consistent with intestinal-type cancer according to Laurén's classification.
To the best of our knowledge, only one study evaluated PMS2 methylation status, but no correlation was found between the lack of PMS2 expression and this gene methylation (0% of methylation) in colorectal cancer (Truninger et al., 2005).The present study it is the first to evaluate PMS2 methylation status in gastric cancer.We found PMS2 methylation in 28.26% of gastric cancer samples studied and it was associated with both intestinal-and diffuse-type gastric cancer compared to normal controls.It was also detected a tendency for PMS2 methylation in intestinal-type gastric cancer with larger tumor extension.However, further studies on PMS2 methylation are needed to better understand this finding in gastric carcinogenesis.
MSH2 methylation status was only evaluated in two other studies on gastric carcinogenesis (Wu et al., 2000;Fang et al., 2004).All studies, including the present study, evaluated the same CpG island that was previously associated with MSH2 expression (Wang et al., 2003).Wu et al. (2000) have not found MSH2 methylation in 42 samples studied.Fang et al. (2004) performed bisulfite modification and sequencing of MSH2 in cancerous, paracancerous and non-cancerous tissues from ten patients and found that methylation occurred only at the -166 CpG site.In our samples, we found MSH2 methylation in 17.39% of samples studied and a tendency for MSH2 promoter in gastric cancer compared to normal controls.Thus, MSH2 methylation frequency in our study population was higher than that found in Asian populations.
We also found an association between MSH2 methylation and diffuse-type gastric cancer samples.Kitajima et al. (2003) previously described an association between loss of MSH2 expression and undifferentiated gastric cancer type, largely corresponding to diffuse-type in Laurén's classification, in a Japanese population.
MSH6 was previously described as a potential transcriptionally silenced gene by cytosine methylation (Bearzatto et al., 2000;Gazzoli and Kolodner, 2003).The 5'-flanking region of MSH6 gene was found to comprise seven functional Sp1 transcription factor binding sites, each binding to Sp1 and Sp3 and contributing to promoter activity.The Sp1-4, Sp1-5, Sp1-6, and Sp1-7 sites have a paramount role in this activity and allow DNA methylation to inhibit Sp1 binding (Gazzoli and Kolodner, 2003).
The methylation status of MSH6 has not been reported in any primary tumor.In the present study, we evaluated the methylation status of MSH6 region which comprises the most important Sp1 sites.However, we did not detect MSH6 methylation in any gastric specimen, suggesting that MSH6 methylation is not a major event for gastric carcinogenesis in the samples studied.
The integrity of genetic information depends on the fidelity of DNA replication and the efficiency of several different DNA repair processes.The identification of epigenetic modifications in MMR genes and the determination of the frequency of alterations may be a useful tool for developing more specific cancer therapies.This is the first study evaluating the methylation status of MLH1, MSH2, MSH6 and PMS2 promoters and their hypermethylation frequencies in gastric tissue samples in a population from Northern Brazil.The methylation status of PMS2 and MSH6 promoter had never been investigated before in gastric samples.Our findings show that methylation of MLH1, MSH2 and PMS2 MMR genes is associated with gastric carcinogenesis.

FIGURE 1 .
FIGURE 1. a) Examples of gel electrophoresis using MLH1, MSH2, MSH6 and PMS2 MSP primers.L: size marker; M: methylated; U: unmethylated.b) Methylated and unmethylated samples of MLH1 used as positive control for DNA bisulfite treatment, when compared to untreated sequence.