DNA methylation in endometrial cancer

Abstract

Endometrial cancer is the most commonly diagnosed gynecological cancer, and it has been shown to be a complex disease driven by abnormal genetic, and epigenetic alterations, as well as environmental factors. Epigenetic changes resulting in aberrant gene expression are dynamic and modifiable features of many cancer types. A significant epigenetic change is aberrant DNA methylation. In this review, we review evidence on the role of aberrant DNA methylation, examining changes in relation to endometrial carcinogenesis, and report on recent advances in the understanding of the contribution of aberrant DNA methylation to endometrial cancer with the emphasis on the role of dietary/ lifestyle and environmental factors, as well as opportunities and challenges of DNA methylation in endometrial cancer management and prevention.

Introduction

Endometrial cancer is the most commonly diagnosed gynecological cancer. It is the 4^th most common cancer for women, with 42,160 new cases and 7,780 deaths occurring in the United States in 2009.1 Based on differences in clinicopathologic characteristics, there are two subtypes of endometrial carcinoma. Type I, endometrioid endometrial carcinoma, accounts for approximately 80% of cases, occurs frequently in pre- and peri-menopausal women, and is significantly related to a history of unopposed estrogen exposure or other hyperestrogenic risk factors. Type I tumors are usually well differentiated, and most patients present with early-stage of the disease and have a favorable prognosis. In contrast, type II, non-endometrioid endometrial carcinoma, is more common in older postmenopausal women, often poorly differentiated, and not associated with hyperestrogenic factors. Patients with type II tumors are more likely to have metastasis and are at high risk of relapse.2 Type I tumors commonly have near-diploid karyotypes, microsatellite instability and mutations in PTEN, K-RAS and CTNNB1 (β-catenin) genes; whereas type II tumors are characterized more often by p53 mutation, overexpression of Her-2/neu and an aneuploid karyotype.2

Although multiple risk factors, such as age, overweight/obesity and postmenopausal hormone therapy, have been identified, the understanding of the etiologies of the two subtypes of endometrial cancer continues to evolve. Similar to other cancer sites, endometrial cancer has been shown to be a complex disease driven by abnormal genetic, and epigenetic alterations, as well as environmental factors. As a common molecular alteration in human neoplasia, epigenetics is defined as heritable changes in gene expression without alteration of the nucleotide sequence.3 Epigenetic changes are dynamic and modifiable upon treatment with pharmacological agents. In the last ten years, it has become increasingly apparent that epigenetic regulation of gene expression is at least as important to carcinogenesis as more studied genetic disruptions including aneuploidy, point mutations and variation in gene copy number, both gain or loss. While epigenetics refers to a broader class of changes, we focus here on the role of DNA methylation, a significant epigenetic change, examining those changes in relation to endometrial carcinogenesis. We report on recent advances in the understanding of the contribution of aberrant DNA methylation to endometrial cancer with the emphasis on the role of dietary/lifestyle and environmental factors, as well as opportunities and challenges of DNA methylation in endometrial cancer management and prevention.

Aberrant DNA Methylation in Endometrial Cancer

The best-known epigenetic event is aberrant DNA methylation. In humans, mediated by the three known active DNA cytosine methyltransferase (DNMT1, 3A and 3B),4 a methyl group (-CH₃) is covalently bonded to cytosine residues of the CpG dinucleotides, resulting in 5-methylcytosine. DNMT1 maintains attachment of methyl groups to hemimethylated DNA during replication, whereas DNMT3A and DNMT3B can catalyze de novo methylation of DNA.5,6 These enzymes cooperatively regulate the dynamic methylation of DNA. CpG dinucleotides are frequently clustered in CpG islands, regions of DNA rich in CpG sites (60–70%).7,8 These island are usually in or around the promoter regions and often unmethylated in normal tissue,7,8 and most CpG sites in the human genome are methylated.9 DNA methylation is found in normal tissue and contributes to control of transcription,10 affecting processes such as normal development,11,12 silencing of genes on the X-chromosome in females10,13 and gene imprinting.14 Aberrant DNA methylation has been associated with most human tumors,15,16 as well as other non-neoplastic diseases and with aging.17,18

Decreased DNA methylation (hypomethylation) is an early event in carcinogenesis, and one of the first epigenetic alterations.19 Hypomethylation can occur at normally methylated DNA sequences as repeated sequences, as well as both encoding regions and introns of genes.15,20 It is associated with early stage genetic instability and upregulation of gene expression.20,21 Another form of aberrant DNA methylation, hypermethylation of CpG islands in the promoter region of particular genes, appears to be significant in carcinogenesis for a number of tumor sites. Promoter hypermethylation is associated with gene silencing, and can affect carcinogenesis particularly when the affected gene is a tumor suppressor genes or other genes involved in the cell cycle, DNA mismatch repair, cell-to-cell interaction, steroid receptor, apoptosis and angiogenesis.22 There is evidence that aberrant DNA methylation is an early and widespread alteration in endometrial tumorigenesis, which is implicated in loss of expression of a variety of critical genes. A recent study of methylation profile in endometrial tumorigenesis showed that, among 24 tumor suppressor genes, the number of promoter methylated loci increased in the progression from normal endometrium to simple hyperplasia to complex hyperplasia (complex hyperplasia without atypia/complex hyperplasia with atypia).23 In addition, aberrant DNA methylation of some tumor suppressor genes was evident before endometrial carcinoma diagnosis in women with the DNA mismatch repair gene mutation.23 This study provides important evidence into the timing and molecular alterations of the critical events in endometrial carcinogenesis, which may be useful to identify DNA methylation profile for early detection of endometrial cancer. In the following review, we summarize the genes that are frequently silenced by DNA methylation in endometrial cancer (Table 1) and discuss how this mechanism may contribute to endometrial carcinogenesis.

Promoter Hypermethylation of Genes in Endometrial Cancer

DNA mismatch repair gene in endometrial cancer.

Germline mutations in DNA mismatch repair genes hMLH1, hMSH2, hMSH3, hMSH6 and hPMS2 have been identified in majority hereditary nonpolyposis colorectal cancer (HNPCC)/Lynch syndrome and in associated familial endometrial cancer.24 A significant characteristic of HNPCC is the presence of microsatellite instability (MSI), changes in the lengths of repetitive genetic loci.24 MSI has also been observed to be present in approximately 20% of sporadic endometrial carcinomas,25–30 while hMLH1 and hMSH2 mutations are rare (less than 10%) in sporadic endometrial cancers with the MSI⁺ phenotype.31–34 However, reduced protein expression of hMLH1 and other mismatch repair genes is a common finding in endometrial cancers lacking detectable mutations.31,35–38 There is evidence that promoter methylation is associated with lack of expression of hMLH1 in human cancers and in mismatch repair-defective human tumor cell lines.39–42 A strong association between hMLH1 promoter methylation and transcriptional silencing and MSI⁺ phenotype was also reported in sporadic endometrial cancer, particularly in the endometroid type.27,31,38,43–45 In addition, the demethylation of the hMLH1gene in several cell lines, including one endometrial carcinoma cell line, using the agent 5-aza-2′-deoxycytidine resulted in reaction of hMLH1expression and restoration of the activity of mismatch repair genes.46 hMLH1 promoter methylation has also been shown to be an early event in the procession from normal endometrium to carcinoma, and as a feature of a subset of precursor lesion.47–51 However, hMLH2 methylation is very rare (1.4% of 138 studied cases) in endometrial cancer.44

Methylation of steroid receptor genes in endometrial cancer.

It is known that the endometrium is highly responsive to hormonal stimuli. The cyclic production of estrogen and progesterone during the menstrual cycle, and declining sex steroid hormone levels after menopause are directly correlated with endometrial proliferation and/or atrophic morphological changes. The majority of known risk factors for endometrial cancer are thought to be directly or indirectly related to exposure to hormones, particularly estrogen. A number of studies have evaluated the association of promoter methylation of the estrogen receptor (ER) and progesterone receptor (PR) genes in endometrial cancer.

The ER gene, located at chromosome 6q25.1, has a CpG island in its promoter and exon 1 regions. Reduced ER RNA and protein expression levels have been found in human endometrial cancer tissues and cell lines,52,53 while have not been consistently reported to be associated with aberrant promoter methylation of ER.53–57 The expression of three isoforms of ERα (ERα-A, ERα-B and ERα-C) and ERβ genes in endometrial cancer cell lines was investigated by Sasaki et al.53 They found no ERα-C expression, as well as restoration of the expression with 5-aza-2′-deoxycytidine treatment. They further found promoter methylation of ERα-C isoform in 94% of endometrial cancer tissues.53 However, Shiozawa et al.56 detected ER promoter methylation in only 24% of 25 endometrial cancer cases; methylation was correlated with ER-negative status of the tumors. Navari et al.55 also did not find any association between loss of ER expression and de novo methylation of the ER gene. Moreover, neither Shiozawa56 nor Navari55 detected alterations in the methylation patterns of different ER isoforms. Although the specific role of the three ERα isoforms in endometrial cancer is still not clear, it was hypothesized that each isoform has a specific character due to the existence of elaborate mechanisms regulating estrogenic effects. Further studies investigating promoter methylation of ERα isoforms in endometrial cancer are needed.

The PR gene, located at chromosome 11q13, also has a CpG island in its first exon. The PR gene encodes two distinct subtypes, PR-A (94 kDa) and PR-B (114 kDa), with distinct functions.58 Previous studies showed that the ratio of PR-A to PR-B expression is abnormal in endometrial cancer, leading to an inappropriate response to progesterone.58,59 Consistent with the transcriptional levels of the two PR isoforms, in one study, over 70% of the samples were promoter methylated for PR-B in endometrial cancer, primarily in tissues from type I cancers; however, PR-A was unmethylated in all cancer and normal endometrial samples.60 It has been suggested that progesterone acts principally through PR-B to inhibit endometrial cancer cell invasiveness modulated by adhesion molecules.61 Evidence from studies of endometrial cancer cell lines indicated that these changes may be the result of aberrant DNA methylation of these genes. Such cell culture studies have shown no PR-B expression, although PR-A expression was observed.60 Other studies have shown a positive association between promoter hypermethylation of PR-B and reduced PR mRNA expression in these cancer cell lines;62,63 DNMT inhibitor 5-aza-2′-deoxycytidine, as well as histone deacetylase (HDAC) inhibitor trichostatin A, led to demethylation and restoration of PR expression.

Methylation of tumor suppressor genes in endometrial cancer.

Promoter hypermethylation of tumor suppressor genes is a major event in the origin of many cancers, and has been the focus of attention in the last decade. A number of studies established a list of tumor suppressor genes frequently hyper-methylated in endometrial cancer. The phosphatase and ten-sin homolog (PTEN, also known as MMAC1/TEP1) gene on chromosome 10q23.3, responsible for Cowden syndrome64 and BannyanZonana syndrome,65 encodes a dual-specificity phosphatase able to dephosphorylate both tyrosine phosphate and serine/threonine phosphate residues.66 It is reported to be important for the inhibition of cell migration and spreading and focal adhesion.66 Inactivation of PTEN has been observed in a number of human cancers, including endometrial cancer. PTEN mutations occur in to 26–80% of endometrial cancers, making it the most commonly known genetic alteration associated with this disease.67,68 Promoter hypermethylation, as the alternative mechanism of PTEN allelic inactivation, was first demonstrated as an occurrence in sporadic colorectal tumors with microsatellite instability, reported to be on the order of 19% of tumors.69 A number of studies have also shown that PTEN promoter methylation in about 20% of sporadic type I endometrial carcinoma,70–72 and that this methylation is significantly associated with metastatic disease and with MSI phenotype.72 However, results from Zysman et al.73 suggested that it is PTEN pseudogene, and not PTEN itself, which is predominantly methylated in endometrial cancer cell lines and in endometrial tumors. More studies with the ability to make distinction between promoter methylation of PTEN and its pseudogene are needed to understand better the mechanisms of PTEN inactivation in endometrial carcinogenesis.

Located on chromosome 9p21, the p16 gene encodes a cyclindependent kinase inhibitor, which can block the cell cycle and arrest the growth of deregulated cancer cells.74 Loss of p16 expression resulting from homozygous deletion, mutation or promoter methylation is a common feature of many human cancers and of cancer cell lines. Promoter methylation of p16 gene has been observed in between 11% and 75% of sporadic endometrial cancers;51,75–78 however, other studies have reported much lower frequencies of p16 methylation.50,79–81 This variability may be a reflection of variation in the sensitivity of the assays used to assess p16 methylation, differences in primer design for the same assay, and differences in sample size or differences in the populations under study. Furthermore, there is limited and inconsistent data regarding the correlation between that p16 promoter hypermethylation and clinicopathological features of endometrial cancer. Although Wong et al.75 reported that p16 promoter methylation was associated with advanced stage and poorer survival of endometrial cancer, no correlation was observed between promoter methylation of p16 and clinicpathological features and prognosis of endometrial cancer in another recent study.76 In addition, some studies found that p16 gene inactivation but rare p16 methylation occurs in a subgroup of aggressive endometrial carcinomas with poor prognosis,80,81 which suggests that the molecular mechanisms of p16 inactivation in endometrial cancer remain unclear.

Another tumor suppressor gene, Ras-association domain gene family 1A (RASSF1A) is located at chromosome 3p21.3 and is known to induce cell cycle arrest through the Rb-mediated checkpoint by inhibiting the accumulation of cyclin D1.82,83 Alteration of this tumor suppressor gene is frequently found in tumors from a variety of sites.82,83 RASSF1A promoter methylation has been reported to occur in 33–85% of endometrial cancers and is associated with reduced expression of RASSF1A.84–90 Arafa et al.84 recently observed frequent hypermethylation of RASSF1A gene promoter (36%) in normal endometrium adjacent to endometrioid carcinoma, suggesting that RASSF1A promoter methylation is an early event in type I endometrial carcinogenesis. Hypermethylation of this gene has been found to be correlated with loss of heterozygosity (LOH); in a study of cervical cancer, 8 of 12 (67%) with hypermethylated RASSF1A gene showed concomitant LOH at 3p21.91 LOH at 3p is a frequent occurrence in endometrial cancer.92 Aberrant promoter methylation of RASSF1A gene combined with LOH at 3p may play a more important role in endometrial carcinogenesis. Moreover, promoter methylation of RASSF1A has also been inconsistently found to be associated with advanced stage, recurrence and survival for endometrial cancer.86–89,93 These inconsistent associations may be due to small sample size, different subtypes of endometrial cancer under study, and short follow-up period for recurrent and survival analyses.

Adenomatous polyposis coli (APC), a tumor suppressor gene, regulates β-catenin in the Wnt signaling pathway, and the aberrations in the Wnt pathway appear to impact the initiation and progression of several human cancers.94,95 Alterations in the APC gene or in the β-catenin gene may affect the Wnt pathway and are thought to be associated with endometrial carcinogenesis.96,97 However, mutations of the APC gene are not common events in endometrial cancers.96,98,99 APC gene promoter methylation has been demonstrated in around 20–45% of endometrial cancers,50,78,90,93,98–101 and more frequent APC promoter methylation in tumors with MSI than in those without MSI was reported in several studies.98–100 No significant associations of APC promoter methylation with the cliniopathological factors or recurrence and distant metastases have been observed in endometrial cancers.50,78

β-catenin can complex with molecules including both APC and E-cadherin and is also involved in cell adhesion, together with E-cadherin, α-catenin and γ-catenin.102 A few studies have evaluated promoter methylation of E-cadherin, a possible tumor suppressor gene, in endometrial cancer.50,99,103,104 Results of these studies have not been consistent. Sito et al.104 reported that the aberrant methylation in promoter region of E-cadherin gene is associated with poor differentiation and myometrial invasion in endometrial carcinomas, suggesting a possible role of E-cadherin in endometrial cancer progression. However, no association between E-cadherin hypermethylation and clinicopathological or immunohistochemical features of endometrial cancer was found in other studies.50,99,103 Pijnenborg et al.99 did not find E-cadherin gene promoter methylation in the tested endometrial tumors, although the absence of E-cadherin expression was detected and found to be associated with the development of distant metastases.

Aberrant MGMT methylation in endometrial cancer.

O-6-methyguanine-DNA-methyltransferase (MGMT), a DNA repair gene, reverses DNA damage via alkylation by removing the methyl group from the O6-guanine and hence protects against DNA mutations. MGMT promoter methylation has been found for a number of cancer sites and precancerous lesions, including colorectal, gastric, lung and glioblastoma; however, it has only been examined in a few studies of endometrial cancer.77,84,101,105 Methylation of MGMT was detected in 48% of synchronous carcinomas of the uterine corpus and ovary;77 however, low frequency or absence of MGMT promoter methylation in singly occurring endometrial cancer was reported in several other studies.84,101,105

Methylation and inactivation of other genes in endometrial cancer.

Retinoic acid receptor (RAR) α, -β and -γ and retinoid X receptor are members of the intercellular receptor superfamily. The RARβ₂ gene has been found to be silenced by promoter methylation in both type I endometrial cancers and endometrial hyperplasia,84,106 suggesting aberrant methylation of RARβ₂ gene as an early epigenetic alteration of endometrial carcinogenesis. Other genes, such as cell cycle inhibitor gene 14-3-3σ,107 paternally expressed gene 3 (PEG3) involved in apoptosis,108 the detoxifying enzyme glutathione S-transferase P1 (GSTP1),109 homeobox genes HOPX, HOXA10 and HOXA11,110–112 RUNX3 tumor suppressor gene,113 and metallothionein 1E gene (MT-1E),114 have been found to be inactivated by aberrant methylation in promoter region in endometrial cancer; however, the impacts of promoter methylation of these genes on endometrial carcinogenesis have not been well known and need to be further investigated.

Difference in DNA methylation in type I and type II endometrial cancers.

There is evidence that type I and II endometrial cancers also contain distinct profiles of aberrant DNA methylation. Promoter hypermethylation of genes including MLH1, APC, MGMT, PTEN and RASSF1A is more frequently detected in type I than type II tumors.72,86,98,115 In addition, promoter hypermethylation of PR-B has been found as a mechanism of loss of PR-B expression in endometrial cancer; and tumors and cell lines applied in those studies included predominantly type I cancers,60,62 although PR negative is more common in type II cancer.116 These findings suggest that promoter hypermethylation may play less of a role in the tumorigenesis of type II cancers.

Some recent studies investigated the expression levels of DNMT1 and DNMT3B in normal endometrium and type I and type II endometrial carcinomas. Compared to normal endometrium, expression levels of both DNMT1 and DNMT3B were significantly increased in type I cancers but downregulated in type II cancers.86,117 Although DNMT1 is known to function as a maintenance methyltraferase, there is evidence that DNMT1 and DNMT3B cooperatively catalyze de novo methylation.6 The lower expression of DNMT1 and DNMT3B may result in global hypomethylation in type II endometrial cancers and contribute to histological differences. More studies are needed to characterize the aberrant DNA methylation profiles in type I and II cancers, to identify different mechanisms for different phenotypes, and to elucidate appropriate prevention approaches.

Environmental Factors, DNA Methylation and Endometrial Cancer

Epidemiology and experimental studies have found a number of agents in diet and environment involved the epigenetic alterations in the process of a variety of human cancers, including endometrial cancer. Those agents are considered “epigenetic carcinogens” (epimutagens).118

Diet.

Diet is an important modifier of DNA methylation profile. The most studied and among the best understood is the relationship between micronutrients involved in one-carbon metabolism and DNA methylation.119–121 Folate plays a central role in one carbon metabolism that provides a methyl group for a variety of biological process including methylation of DNA, RNA and protein, as well as the synthesis of purines and pyrimidines for DNA synthesis.119–121 Vitamins B₂, B₆ and B₁₂, the essential amino acid, methionine, are also involved in one carbon metabolism, potentially affecting genomic DNA methylation and synthesis and thereby causing dysregulation of gene expression.119–121 Low levels of folate, vitamins B₂, B₆, B₁₂ and methionine have been shown to be associated with cancer and cardiovascular disease risk.122,123 Deficiencies of methyl donor intake and excess alcohol consumption have been found to induce promoter hypermethylation and global hypomethylation in animal models.124–126 However, there is limited evidence from epidemiological studies of different cancer sites that these nutrients intake affect DNA methylation;127–131 and very little known for the potential effect of low level of methyl donor intake on endometrial DNA methylation. One concern with interpretation of these inconsistent findings is that each study examined methylation status of a limited number of genes, it cannot be ruled out that methyl donor intake may have impact on promoter methylation of other excluded genes. There is evidence that the folate status may be tissue specific,121 which may explain at least in part the inconsistency of study results. Moreover, diet can also alter histone modification.118 It is possible that diet may impact development of cancer through the mechanisms of both DNA methylation and other epigenetic alterations.

Phytoestrogens are naturally occurring compounds in many foods such as soy and soy products, and associated with reduced endometrial cancer risk in Asian populations and western vegetarian populations. Genistein, one of many phytoestrogens presented in soy, can inhibit cell growth, angiognesis and induces apoptosis in cancer cell lines and animal models;132–134 and exhibit mixed estrogenic and anti-estrogenic properties.135 The underlying molecular mechanisms for these effects of genistein are not well established, but recent studies suggest that genistein may regulate gene activity by modulating epigenetic events. While there are no data on this possible association for endometrial cancer, there is some evidence from other cancer sites that phytoestrogen intake may affect methylation. Day et al.136 reported gene-specific increases in prostate DNA methylation patterns among mice treated with a diet high in genistein compared to mice on a control diet. Several studies also have shown that genistein, lignans, and other related soy isoflavones can reverse hypermethylation and reactivate the silenced genes, including RARβ₂, p16, MGMT, GSTP1 and BTG3 genes, in esophageal, breast and renal cancer cell lines.137–140 These in vitro data showed that genistein can have a dose-dependent inhibition of DNMTA and HDAC activities,137 and increase active chromatin modifications near the transcription start site.140 However, in a recent study conducted in healthy premenopausal women, promoter methylation of RARβ₂ and CCND2 genes in mammary tissue was lower for women with low blood genistein and higher among those with higher circulating genistein.141 These in vitro studies used higher concentrations of genistein than those found in women consuming soy products.137,141 Further studies are needed to evaluate the impact and potential mechanism of dietary phytoestrogens on DNA methylation in endometrial tissue, and identify the proper administration level of phytoestrogens recommended for chemoprevention of endometrial cancer.

Other lifestyle factors.

Physical activity has been shown to be associated with modestly reduced risk of endometrial cancer. A recent cross-sectional study in women without breast cancer found that both lifetime and recent physical activity were inversely related with promoter hypermethylation of APC gene in nonmalignant breast tissue.142 No similar studies on endometrial cancer have been reported, but the finding for breast tissue suggests that the protective effect of physical activity on endometrium may be associated with epigenetic changes including DNA methylation.

External steroid hormones.

In the past several decades, external sources of steroid hormones, which influence cell proliferation and therefore the risk of endometrial cancer, have been widely used and contribute to endometrial cancer development, particularly type I tumors. Epigenetic events may play an important role in the physiological response to external steroid hormones and undergo continuous modification and alteration. Progestrone has been given in combination with estrogen in hormone therapy to prevent the increased risk of endometrial cancer associated with unopposed estrogen therapy. The use of combined oral contraceptives (COCs) has generally been seen to reduce the risk of endometrial cancer,143–145 which may be due to the progestrone component of COCs. Recent studies showed that DNMT3A and DNMT3B in human endometrium are under the regulation of both progesterone and estrogen, suggesting DNA methylation may be influenced by sex steroid hormones.146 It is possible the external progesterone mediates a protective effect on endometrium through impacting DNMT activities or downregulation of an epimutagen effect of estrogen by epigenetic modifications. However, more studies are needed to clarify the molecular mechanisms of external progestin on endometrium.

Conclusion

Recent developments in the field of epigenetics, especially studies of DNA methylation, have provided valuable insights for understanding the role of epigenetic alterations in normal cellular processes and abnormal changes leading to endometrial carcinogenesis. These new insights hold tremendous potential in the diagnosis, treatment and prevention of endometrial cancer.

There is accumulating evidence that DNA methylation changes may contribute to carcinogenesis in the endometrium, although evidence regarding these changes induced by dietary/lifestyle and environmental factors in endometrial cancer is quite limited. There is still much to explore regarding the target genes of aberrant methylation, how they contribute to the carcinogenic process and what factors affect that methylation. The emerging powerful technologies147 can be used to quickly identify DNA methylation profiles for different subtypes of endometrial cancer. In addition, epigenetic-epidemiological studies provide opportunities not only to study the contribution of epigenetics to endometrial cancer but to understand the joint impact of genetic, epigenetic and environmental exposures on the risk of endometrial cancer.

Figures and Tables

Table 1 Selected genes frequently silenced by DNA promoter methylation in endometrial carcinoma

Gene	Alternate gene name	Function
APC	Adenomatous polyposis coli	Tumor suppressor gene, cell adhesion, signal transduction, stabilization of the cytoskeleton, regulation of cell cycle and apoptosis
CDH1 (E-cad)	E-cadherin	Epithelial cell-cell adhesion, suppresses invasion and metastasis
ERα	Estrogen receptor α	Steroid receptor, regulation of cell proliferation
MGMT	O-6-methylguanine-DNA methyltransferase	DNA repair gene
hMLH1	Human Mut-L homolog 1	DNA mismatch repair gene
p16	Cyclin-dependent kinase inhibitor 2A	Tumor suppressor gene, cell cycle regulation, involved in senescence
PR-B	Progesterone receptor	Steroid receptor, growth regulation
PTEN	Phosphatase and tensin homolog	Tumor suppressor gene, controls proliferation and apoptosis
RASSF1A	Ras association domain family protein 1	Tumor suppressor gene, cell cycle regulation, microtubule stabilization, cellular adhesion and inhibits tumor formation, apoptosis
RARβ2	Retinoic acid receptor	Apoptosis, involved in senescence, inhibition of proliferation

Acknowledgement

This work was supported by the National Cancer Institute (R01CA 092040).