Inactivation of COX-2, HMLH1 and CDKN2A Gene by Promoter Methylation in Gastric Cancer: Relationship with Histological Subtype, Tumor Location and Helicobacter pylori Genotype

Objective: We aimed to evaluate the inactivation of COX-2 , HMLH1 and CDKN2A by promoter methylation and its relationship with the infection by different Helicobacter pylori strains in gastric cancer. Methods: DNA extracted from 76 H. pylori -positive gastric tumor samples was available for promoter methylation identification by methylation-specific PCR and H. pylori subtyping by PCR. Immunohistochemistry was used to determine COX-2, p16 INK4A and HMLH1 expression. Results: A strong negative correlation was found between the expression of these markers and the presence of promoter methylation in their genes. Among cardia tumors, negativity of p16 INK4A was a significant finding. On the other hand, in noncardia tumors, the histological subtypes had different gene expression patterns. In the intestinal subtype, a significant finding was HMLH1 inactivation by methylation, while in the diffuse subtype, CDKN2A inactivation by methylation was the significant finding. Tumors with methylated COX-2 and HMLH1 genes were associated with H. pylori vac A Received: February 14, 2011 Accepted after revision: May 10, 2011 Published online: August 17, 2011 Markênia Kélia Santos Alves, MS Departamento de Patologia e Medicina Legal Universidade Federal do Ceará, Faculdade de Medicina Rua Alexandre Baraúna, 949, Porangabussu, Fortaleza, CE 60183-630 (Brazil) Tel. +55 85 3366 8306, E-Mail markenialves @ yahoo.com.br © 2011 S. Karger AG, Basel 1015–2008/11/0785–0266$38.00/0 Accessible online at: www.karger.com/pat D ow nl oa de d by : 54 .7 0. 40 .1 1 10 /1 7/ 20 17 7 :1 4: 11 P M COX-2 , HMLH1 and CDKN2A Inactivation by Promoter Methylation Pathobiology 2011;78:266–276 267 genes in GC. Its product, p16 INK4A , acts during the G1 phase of the cell cycle by inhibiting progression to the next cell cycle phase through the selective inhibition of the formation of complexes cyclin D/cyclin-dependent kinase 4 or 6 [5] . The HMLH1 gene is a human homologue of the gene mut L of Escherichia coli , located on chromosome 3p21.3. This gene codes for a protein with the same name that belongs to the DNA mismatch repair system [6] . Inactivation of the CDKN2A and HMLH1 genes has been documented in several tumors such as colorectal and kidney, besides GC [4, 5, 7–9] . The restoration of expression of these genes by demethylating agents [10, 11] indicates that these alterations may potentially be therapeutic targets. Additionally, COX-2 overexpression has been related to the intense inflammatory response and GC development [12] , but some studies have shown a low expression of the COX-2 gene in GC associated with gene promoter hypermethylation [11, 13] . This gene, officially called prostaglandin-endoperoxide synthase 2 (PTGS-2) , is located on chromosome 1q25.2–q25.3 and encodes an enzyme by the same name; PTGS-2 is involved in the synthesis of prostaglandins, which mediate cell signaling and are involved in inflammation process [12] . Although there are studies indicating hypermethylation in promoter regions of genes COX-2 , CDKN2A and HMLH1 as an important mechanism of gene silencing and suppression of their function in gastric carcinogenesis, there are still many gaps to be considered, such as its importance in each histological subtype, in tumor location and the factors that promote this process. Studies in vitro and in vivo and of bacterial eradication have associated the presence of Helicobacter pylori to the induction of hypermethylation in promoter regions of genes important for GC progression, such as COX-2 , CDH1 , HMLH1 , CDKN2A and RUNX3 [14–16] , albeit with controversial results [17–19] . H. pylori has a high genetic variability, which has been related to virulence and clinical outcome of gastric diseases [20–22] . Two well-established virulence factors are the presence of the cytotoxin-associated antigen A ( cag A) gene, located within the right portion of the cag pathogenicity island ( cag PAI), and vacuolating cytotoxin A ( vac A), mostly the vac As1m1 allelic combination [21, 22] . cag -PAI also harbors the cag E and vir B11 genes, located in the right and left regions of the island, respectively. These genes are known to play a role in constructing the type IV secretion system and inducing pro-inflammatory, pro-proliferative epithelial cell signaling [20, 23] and show a high frequency in GC [24] . Another important H. pylori virulence factor is the presence of flagella, coded by several genes, such as fla A and fla B, which have been associated with successful colonization and with the process of infection and persistence in the gastric mucosa [25] . In addition to the well-characterized motility role, the flagella have been associated to other functions such as acting as an export apparatus for virulence factors [26] , in the sensitivity to medium viscosity [27] and as an important immunogenic protein [28] . fla A is the main encoding gene to flagellin of the flagellum filament [29] . So far, few studies have assessed the involvement of H. pylori genotypes regarding their methylation inducing potential of gastric carcinogenesis-related genes. Therefore, studies of the mechanisms by which H. pylori may promote gastric carcinogenesis are important. In view of the genetic diversity shown by H. pylori , it is important to assess the relationship between specific genotypes and their possible carcinogenic mechanisms. Additionally, tumors of the intestinal and diffuse subtypes need to be considered individually, since they are distinct tumors differing not only in clinical and histopathological features, but in their tumorigenic pathways, with distinct genetic and epigenetic alterations [30, 31] . Materials and Methods Clinical Specimens The present study was approved by the Hospital Ethics Committee at the Universidade Federal do Ceará, Brazil, and all subjects signed an informed consent form before inclusion. Samples from 76 patients with gastric adenocarcinoma who had undergone gastrectomy were collected from Walter Cantídio University Hospital and Santa Casa de Misericórdia Hospital, both located in Fortaleza, the capital of Ceará State. The histological classification was done according to Lauren’s classification. DNA Extraction Genomic DNA was extracted from frozen tumor tissue using the cetyltrimethyl ammonium bromide (CTAB) technique, adapted from Foster and Twell [32] . DNA was extracted from only fragments that showed more than 80% tumor cells. DNA quality was analyzed by 1% agarose gel electrophoresis, and the amount was determined using the Nanodrop 3300 fluorospectrometer (Wilmington, Del., USA). Sodium Bisulfite Treatment and Methylation-Specific PCR DNA extracted from tumor tissue was modified by sodium bisulfite to determine the methylation status of the CDKN2A , HMLH1 and COX-2 genes by methylation-specific PCR, as previously described by Ferrasi et al. [17] . The primers targeting the promoter gene regions studied and the annealing temperatures are described in table 1 . PCR was performed in 25  l reaction volume, containing 1 ! Platinum Taq buffer, 3.0 m M MgCl 2 (CDKN2A) or 1.5 m M MgCl 2 (HMLH1 and COX-2), 0.4 m M of each D ow nl oa de d by : 54 .7 0. 40 .1 1 10 /1 7/ 20 17 7 :1 4: 11 P M Alves /Ferrasi /Lima /Ferreira / de Moura Campos Pardini /Rabenhorst Pathobiology 2011;78:266–276 268 dNTP, 0.64  M (CDKN2A), 0.24 m M (HMLH1) or 0.4 m M (COX2) of each primer set, 1 U Platinum Taq DNA Polymerase (Invitrogen, Carlsbad, Calif., USA), and 50 ng of treated DNA. DNA methylated in vitro by Sss-I methylase (New England Biolabs, Beverly, Mass., USA) was used as a positive control. Water and DNA from peripheral lymphocytes of healthy donors were used as negative controls. The PCR products were resolved in a 6% nondenaturing polyacrylamide gel and subsequently submitted to silver staining. For confirmation of the reaction specificity, MS-PCR products from COX-2 , CDKN2A and HMLH1 genes analyzed were cloned with a TOPO TA Cloning Kit (Invitrogen) and both the methylated and unmethylated PCR products were sequenced using an ABI PRISM  BigDye Terminator v.3.0 Cycle Sequencing Kit (Applied Biosystems, Foster City, Calif., USA) and ABI Prism 3100 DNA Sequencer (Applied Biosystems). Detection of H. pylori Infection and vacA Alleles and the Presence of cagA, cagE, virB11 and flaA Genes H. pylori infection was detected by amplification of the ure C gene using primers for PCR, as described by Lage et al. [35] . For the H. pylori -positive samples, the presence of the vac A alleles, cag A, cag E, vir B11 and fla A genes were identified using the primer sets shown in table 2 . PCR mixtures, for amplification of the cag E, vir B11and fla A genes, were prepared in a volume of 20  l using Green MasterMix  (Taq DNA Polymerase, dNTPs and MgCl 2 ) according to the manufacturer’s instructions (Promega, Madison, Wisc., USA), with addition of 0.8% Tween 20, 0.3 M ( vir B11), 0.3 M ( cag E) and 0.3 M ( fla A) of each primer and 50 ng of DNA sample. The ure C, cag A, vac A s1/s2, vac A m1 genes were amplified in a 25 l volume containing 2.5  l of 10 ! PCR buffer (Invitrogen, Cergy Pontoise, France), 1% Tween 20, 1.5 m M MgCl 2 (Invitrogen), 200 M of each dNTP (Invitrogen), 1 U Platinum Taq polymerase (Invitrogen), 0.4 M ( ure C, cag A, vac A s1/s2, vac A m1), 0.3 M ( vac A m2) of each primer and 50 ng of DNA sample. The PCR products were resolved by 1% agarose gel electrophoresis with ethidium bromide staining. The sample was considered H. pylori positive when a ure C fragment of 294 bp was amplified. For confirmation of the reaction specificity, PCR products from the ure C gene were cloned in TOPO TA Cloning Kit (Invitrogen) and sequenced using an ABI PRISM BigDye Terminator v.3.0 Cycle Sequencing Kit (Applied Biosystems) and an ABI Prism 3100 DNA Sequencer  (Applied Biosystems). The vac A, cag A, cag E, vir B11 and fla A genes were considered positive when a specific fragment was detected ( table 2 ). DNAse-free water was used as a negative control. DNA preservation has also been confirmed by amplification of different genes in other approaches under study in our laboratory. Random samples were reanalyzed to confirm the results. Immunohistochemistry The detection of p16 INK4A protein was performed using the commercial kit CINTEC Histology (K5340; Dako, Glostrup, Denmark), according to the manufacturer’s recommendations. The proteins HMLH1 and COX-2 were determined by the streptavidin peroxidase method, adapted from the protocol of Hsu et al. [39] . Briefly, for antigen retrieval, deparaffinized sections were pretreated by heating in a microwave oven in 10 m M citrate buffer, pH 6.0, for 20 min. Endogenous peroxidase was blocked by 3% H 2 O 2 , for 10 min. Sections were incubated in a humid chamber overnight at 4 ° C with the following primary antibodies: COX-2 (SC-1746; dilution 1: 50; Santa Cruz Biotechnology, Santa Cruz, Calif., USA) and HMLH1 (SC-581, dilution 1: 100; Santa Cruz Biotechnology). The reaction was detected with the LSAB+ system (DakoCytomation, Carpinteria, Calif., USA), according to the manufacturer’s recommendations. Confirmed cases of p16-positive human breast carcinoma, COX-2-positive colon from patients with Crohn’s disease and HMLH1-positive tonsil were used as positive controls. Immunostaining Analyses The immunohistochemical evaluation was performed independently by two experienced technicians, using direct light microscopy. Any conflicting results were jointly considered for a consensual determination. Protein expression was quantified by manual counting of at least 1,000 tumor cells in 10 different fields at a magnification of ! 400. The labeling index expresses the percentage of nuclear or cytoplasmic positive cells in each tumor sample [40] . Only cases with labeling index equal to or greater than 5% were considered positive. Table 1. P CR primer sets used for MS-PCR Gene Primer Ref. Annealing temperature ° C Size of PCR product bp forward reverse COX-2 M: TTAGATACGGCGGCGGCGGC TCTTTACCCGAACGCTTCCG 19 59 161 U: ATAGATTAGATATGGTGGTGGTGGT CACAATCTTTACCCAAACACTTCCA 65 171 CDKN2A M: TTATTAGAGGGTGGGGCGGATCGC GACCCCGAACCGCGACCGTAA 33 70 150 U: TTATTAGAGGGTGGGGTGGATTGT CAACCCCAAACCACAACCATAA 70 151 H MLH1 M: TATATCGTTCGTAGTATTCGTGT TCCGACCCGAATAAACCCAA 34 66 153 U: TTTTGATGTAGATGTTTTATTAGGGTTGT ACCACCTCATCATAACTACCCACA 64 124 M = Methylated; U = unmethylated. D ow nl oa de d by : 54 .7 0. 40 .1 1 10 /1 7/ 20 17 7 :1 4: 11 P M COX-2 , HMLH1 and CDKN2A Inactivation by Promoter Methylation Pathobiology 2011;78:266–276 269 Statistical Analyses The statistical analyses were carried out using the SPSS version 15.0 program (SPSS Inc., Chicago, Ill., USA). Statistically significant differences were evaluated by the  2 test or Fisher’s exact test. Correlations between immunostaining and COX-2 , HMLH1 and CDKN2A promoter region methylation status were analyzed by Spearman’s rank correlation coefficient. The results were considered statistically significant when the p values were less than 0.05.