Methylation in esophageal carcinogenesis

Abstract

Genetic abnormalities of proto-oncogenes and tumor suppressor genes have been demonstrated to be changes that are frequently involved in esophageal cancer pathogenesis. However, hypermethylation of CpG islands, an epigenetic event, is coming more and more into focus in carcinogenesis of the esophagus. Recent studies have proved that promoter hypermethylation of tumor suppressor genes is frequently observed in esophageal carcinomas and seems to play an important role in the pathogenesis of this tumor type. In this review, we will discuss current research on genes that are hypermethylated in human esophageal cancer and precancerous lesions of the esophagus. We will also discuss the potential use of hypermethylated genes as targets for detection, prognosis and treatment of esophageal cancer.

Key Words: Methylation; Esophageal cancer; Tumor suppressor gene; Carcinogenesis

INTRODUCTION

Esophageal cancer is one of the least studied and deadliest cancers, with a remarkable geographical distribution and a low likelihood of cure[1]. Therefore, the current challenges in the management of esophageal cancer are to obtain a better understanding of the underlying molecular biological alterations to provide new treatment options. Cancer of the esophagus exists in two main forms with different etiological and pathological characteristics-esophageal squamous cell carcinoma (ESCC) and esophageal adenocarcinoma (EAC)[1]. It is well known that esophageal carcinogenesis is a multistage and progressive process which includes basal cell hyperplasia (BCH), dysplasia (DYS), carcinoma in site (CIS) and advanced esophageal carcinoma[1,2]. A variety of genetic lesions are involved in esophageal carcinogenesis, including gene amplifications, loss of heterozygosity(LOH) or homozygous deletions, mutations, and chromosomal rearrangements[1,2]. From the above mentioned genetic lesions , mutations are greatly focused on. The vast majority of esophageal cancers have mutations of the TP53 and p16 genes at an early stage followed by mutations in genes such as APC, Rb and cyclin D₁ at later stages of progression[3]. Furthermore, the alteration of the gene mutations may show different types in ESCCs and EACs. For example, ESCCs and EACs show distinct patterns of TP53 mutations, namely a high prevalence of G > A transitions at CpG sites in EACs whereas in ESCCs a higher prevalence of G to T transversions and mutations at A:T base pairs is present[4].

Genetic mutation of genes that inhibit the formation of tumors has long been known to be one of the main driving forces in the development of cancer[1]. However, recent data have focused our attention to the contribution of epigenetics to tumorigenesis. In tumorigenesis of the esophagus, the epigenetic inactivation of genes is as an important a driving force as the inactivation of genes by mutation[5]. ‘Epigenetic’ events, i.e. heritable changes in gene function which cannot be explained by changes in DNA sequence, are composed of histone acetylation, the chromatin structure and DNA methylation[5]. DNA methylation seems to be the most important mechanism for “epigenetic change” at present[5,6]. Through a process of post-replicative covalent modification catalysed by DNA methyltransferases(DNMTs), the DNA of mammalian cells contains a ‘fifth base’, namely 5-methylcytosine. The most frequent target for this modification is cytosine in the context of the dinucleotide CpG[5]. Throughout the genome CpG dinucleotides are found at one-fifth of their predicted frequency[6]. In marked contrast to the genome-wide underrepresentation of CpGs, there are regions of the genome termed CpG islands which have maintained their expected frequency of the dinucleotides. And the CpG islands are often found within the promoter of genes[6,7]. It has been known for many years that, in general terms, there is an inverse relationship between the density of promoter methylation and the transcriptional activity of a gene[8,9]. The mechanism of gene silencing by promoter hypermethylation has recently been shown to be related to the recruitment of repressor protein complex, resulting in de-acetylation of the chromatin and histone, thus barring access to the active transcription complex. However, the actual mechanisms by which DNA methylation modulates gene expression have remained elusive[7,10]. The assays for detection of cytosine methylation can be divided into two groups: restriction enzyme-based and bisulfite treatment-based[11,12]. The former employs the inhibition of certain restriction enzymes by methylation of their recognition sites as an indicator for the presence of methylation. The latter translates the epigenetic information of cytosine methylation in primary sequence differences by converting unmethylated cytosine to uracil whereas methylated cytosine remains unaltered. The bisulfite-converted genomic DNA can be analyzed by a wide variety of PCR-based methods[11,12]. Methylation is needed for the normal development of cells. And genome stability and normal gene expression are largely maintained by a fixed and predetermined pattern of DNA methylation[13]. Aberrant methylation confers a selective growth advantage that results in cancerous growth[13]. From various lines of evidence, it is known that the methylation pattern of the cancerous cell is associated with a broad genomic hypomethylated state that is often accompanied by a more regional and locus-specific hypermethylated pattern[7]. The presence of alterations in the profile of DNA methylation in cancer was initially thought to be exclusively a global hypomethylation of the genome that would possibly lead to massive overexpression of oncogenes whose CpG islands were normally hypermethylated[14]. Nowadays, however, this is considered to be an unlikely or, at least, incomplete scenario. The popularity of the concept of demethylation of oncogenes leading to their activation is in clear decadency[14,15]. Hypermethylation of CpG islands located in the promoter regions of tumor suppressor genes(TSGs) is now firmly established as an important mechanism for gene inactivation[16,17,18]. The particular genes that are hypermethylated in tumor cells are strongly specific to the tissue of origin of the tumor. A profile of CpG island hypermethylation exists according to the tumor type[15]. The mechanism responsible for this type of pattern remains largely unclear. Moreover, accumulating evidence indicates that CpG island hypermethylation is an early event in cancer development and, in some cases, may precede the neoplastic process[19]. Therefore, such profiles would provide invaluable insight into mechanisms underlying the evolution of each tumor type and will provide new molecular markers. This review will focus on the current understanding of DNA methylation abnormalities in esophageal cancer and discuss how this knowledge contributes to our understanding of the pathogenesis of esophageal cancer.

STUDIES OF GENE PROMOTER HYPER-METHYLATION IN ESOPHAGEAL CANCER

Putative tumor suppressor genes, involving apoptosis, cell adherence, DNA repair, and the cell cycle, have been investigated for hypermethylation by various techniques in esophageal cancer. Below, we have chosen to describe those genes that have been most extensively studied in the past and that have been shown to undergo epigenetic changes in esophageal carcinoma.

p14^ARF, p15 and p16

The 9p21 chromosomal band is one of the most frequently altered genomic regions in human cancers[20]. Within a short distance of 50 kb, a gene cluster consisting of three genes, p14^ARF, p15 and p16, is harbored. All of which have putative tumor suppressor roles[20,21,22]. Inactivation of p14^ARF, p15 and p16 genes has been observed in many types of human cancers including ESCC[21,22,23]. For example, the results from immunohistochemical analysis indicated that p16 expression was present in only 3 out of 22 ESCC cases[24]. Some studies showed that germline mutations in the p16 gene might be related to familial melanoma[25], but another study found the mutation of the p16 gene in esophageal cancer was rare[26]. Hemizygous and homozygous deletion at 9p21 are widely considered to be one of the primary mechanisms of p16/p15 inactivation[26]. Recently, however, aberrant methylation of the CpG islands at the promoter regions of p16 and p15 genes was reported in many cancers and was associated with loss of transcription[8,27]. Abbaszadegan et al[28] assessed a large family with clustering of ESCC in northeastern Iran and found aberrant p16 promoter methylation in 64% of ESCC family members and none in normal volunteers. By analyzing the p14^ARF, p15, and p16 genes individually in 40 ESCCs, Xing et al[29] detected aberrant promoter methylation of the p16 gene in 40%, of p14^ARF in 15%, and of p15 in 13% tumor samples. They further detected homozygous deletion of p16 in 18%, of p14^ARF in 33%, and of p15 in 40% tumor samples, and detected no mutation in the p14^ARF and p16 genes[29]. Hardie et al[30] reported that hypermethylation of the p16 promoter is detected in 85% (18/21) of EACs while p16 mutations are uncommon (1.9%; 1/54). Another report found that in 50 Barrett’s esophagus-associated EACs, hypermethylation of p16 was present frequently (54%), but hypermethylation of p14^ARF was absent[27]. The above results suggest that p14^ARF, together with p15, is a primary target of homozygous deletion, whereas p16 is the hypermethylation hotspot in human esophageal cancer.

The FHIT gene

The FHIT gene is located at chromosome 3p14.2 and encodes a polypeptide of 147 amino acids[31]. FHIT allelic deletions and reduced or absent FHIT protein expression have been observed in a variety of tumors suggesting a putative tumor suppressor function[31,32]. In ESCCs, the CpG island in the FHIT promoter region was hypermethylated in 25 of the 36 (69%) analyzed cases, significantly correlated with the deletion of FHIT protein expression[33]. Methylated ESCC cell lines exhibit re-expression of the FHIT gene and demethylation in the CpG islands after treatment with demethylating agent 5-aza-2’-deoxycytidine[34]. Another report showed that aberrant methylation of FHIT was found in 85 of 257 (33%) ESCCs[35]. These findings suggest that methylation of the 5’ CpG islands of the FHIT gene is closely associated with transcriptional inactivation and might be involved in tumor development of the esophagus.

The RARβ₂ gene

The retinoic acid receptor-beta2 (RARβ₂) gene located at 3p24 has been intensively studied in many cancers and found to have defective function, thus making it a candidate TSG[36]. RARβ₂ expression was detected in 88% (14/16) of normal esophageal tissues and only 54% (84/162) of esophageal carcinomas[36]. And 14 of 20 (70%) ESCC samples had hypermethylation of the RARβ₂ promoter[37]. Another group reported that 34 of 47 (73%) primary resected ESCC samples showed RARβ₂ methylation[38]. After 5-aza-2’-deoxycytidine treatment the expression of RARβ₂ was reversed in two RARβ₂-downregulated ESCC cell lines[39]. These results identified methylation as the underlying mechanism for this frequent loss of RARβ₂ in esophageal cancer.

The APC gene

The adenomatous polyposis coli (APC) gene, located on chromosome 5q21, is a TSG in the wnt signaling pathway[40]. APC shows frequent LOH in esophageal carcinomas, and the prevalence of mutations in the APC gene in esophageal carcinomas is low[41]. Hypermethylation of the promoter region of the APC gene occurred in abnormal esophageal tissue in 48 of 52 (92%) patients with EAC, in 16 of 32 (50%) patients with ESCC, but not in matching normal esophageal tissues[41]. So methylation of the promoter region of this gene constitutes an alternative mechanism of gene inactivation in esophageal carcinoma.

The ER gene

The estrogen receptor (ER) gene at chromosome 6q, which has growth and metastasis suppressor activity in many different cell types, is widely expressed in tissues other than breast, and is methylated in 51% of EAC patients[42].

The MGMT gene

The human enzyme O⁶-methylguanine-DNA methyltransferase (MGMT) is located on chromosome band 10q26, and protects the cell from guanine methylation by irreversibly transferring the alkyl group of the O⁶-methylguanine to a specific cysteine residue within the molecule[43]. Approximately 20% of tumor cell lines lack MGMT activity and are highly sensitive to alkylating agents[44]. In established cancer cell lines, MGMT expression appears to be correlated with methylation in the promoter of the gene[44]. The gene has been shown to be methylated in 46/119 (39%) ESCC, but all 21 normal esophageal tissues had unmethylated MGMT[45]. Another report studied the role of DNA hypermethylation in the loss of expression of MGMT during the development of ESCC, and found that 5 of 17 (29%) normal esophagus, 10 of 20 (50%) BCH, 8 of 12 (67%) DYS, and 13 of 18 (72%) ESCC samples had DNA hypermethylation in the MGMT promoter region, showing a gradual increase with the progression of carcinogenesis, while the frequency of the loss of MGMT mRNA and protein expression progressively decreased from normal to BCH, DYS, and ESCC, and it was highly correlated with MGMT promoter hypermethylation[46].

The E-cadherin gene

E-cadherin gene on chromosome 16q22.1 encodes a Mr 120 000 transmembrane glycoprotein expressed on the surface of epithelial cells. In epithelial tissues, E-cadherin mediates homophilic, Ca²⁺-dependent intercellular adhesion that is essential for the maintenance of normal tissue architecture[47]. Loss of E-cadherin expression occurs in a variety of human tumors and is correlated with invasion and metastasis, and activation of E-cadherin results in the growth inhibition of tumor cell lines[48]. E-cadherin can be targeted by both genetic and epigenetic means. Moreover, the hypermethylation of E-cadherin was seen frequently in most tumor types, but mutations only frequently in a small number of specific subtypes[48]. In esophageal carcinoma, downregulation of E-cadherin is common and is associated with an increase in invasive and metastatic potential, but mutations of the gene are rare[49]. E-cadherin was methylated in 26 of 31 (84%) EAC specimens, 16 of 20 (80%) ESCC samples and 4 of 6 ESCC cell lines[49,50,51]. And treatment of E-cadherin-negative carcinoma cells with the demethylating agent, 5-aza-2’-deoxycytidine, induced re-expression of the gene[51]. These data suggest that epigenetic silencing via aberrant methylation of the E-cadherin promoter is the critical mechanism for inactivation of this gene in esophageal cancer.

The TSLC1 gene

The TSLC1 (tumor suppressor in lung cancer) gene located on 11q23.2 was first characterized as a TSG in human non-small cell lung cancer(NSCLC) and termed TSLC1[52]. The tumor suppressor role of this gene has been demonstrated in the cell lines of NSCLC, hepatocellular carcinoma, pancreatic cancer and ESCC[52,53]. Loss of TSLC1 expression was observed in 75% of the ESCC cell lines and 50% of the primary tumors from ESCC patients[53]. Ito et al[53] examined the methylation status of six cytosine residues of CpG sites in a putative promoter sequence upstream from the TSLC1 translation initiation site by bisulfite sequencing in four cell lines, including KYSE270, which expressed TSLC1, and KYSE410, KYSE520, and KYSE960, which did not express it. They also found all of the cytosine residues in KYSE270 DNA were unmethylated, whereas all of the six cytosine residues in KYSE520 DNA and five residues in KYSE410 and KYSE960 DNA were methylated. Especially, the cytosine residues in KYSE520 DNA were all hypermethylated. However, the report about the status of the promoter methylation of TSLC1 gene in esophageal cancer tissue has not been available.

The RASSF1A gene

Many known RAS effectors are oncoproteins on their own. Less is known about Ras effectors possessing tumor suppressor properties[54]. Recently, a new family of genes encoding a putative Ras effector, the Ras-association domain family 1 (RASSF1) gene, has been identified within the critical lung and breast cancer deletion region at 3p21.3. The RASSF1 locus encodes several major transcripts by alternative promoter selection and alternative mRNA splicing: RASSF1A, RASSF1B and RASSF1C. Many studies have suggested that RASSF1A was a new putative TSG[54,55]. RASSF1A acts as a negative effector of Ras in a pro-apoptotic signaling pathway. Interestingly, mutational inactivation of this gene is very rare (< 2%), and the main mechanism of its inactivation is through promoter methylation and LOH[56]. The RASSF1A isoform is highly epigenetically inactivated in lung, breast, ovarian, kidney, prostate, thyroid, esophagus and several other carcinomas[54]. Hypermethylation of RASSF1A was detected in 73% of ESCC cell lines and 51% of primary ESCCs, whereas only 4.3% of RASSF1A hypermethylation were detected in corresponding noncancerous tissues[38]. There was a statistically significant correlation between the presence of hypermethylation and tumor stage[57]. Wong et al[58] also found that RASSF1A was partially methylated in 3/7 (43%) esophageal cancer cell lines; 22/64 (34%) ESCCs and 3/64 (4.7%) corresponding non-tumor samples; and was not methylated in 2 immortalized normal oesophageal epithelial cell lines and 6 normal esophageal epithelium samples. These findings suggest that epigenetic silencing of RASSF1A gene expression by promoter hypermethylation could play an important role in ESCC carcinogenesis.

Besides the above mentioned genes, there are hypermethylations of some other genes involving esophageal cancer, including hMLH1[59,60], VHL[38], TIMP3[42,61], DAP-kinase[42], pRb[62], ECRG4[63], Chfr[64], HLA class I[65], EYA4[66], CDH13[67], SFRP1[68] and PGP9.5[69]. Table 1 gives a summary of the profile of gene hypermethylation in human esophageal cancer.

Table 1 Compilation of genes hypermethylated in esophageal cancer n (%).

Gene	Entity of pathology	Incidence of methylation		Reference
		Cancer tissue	(Adjacent)nonmalignant tissue
p14ARF	ESCC	6/40 (15)	ND	29
	EAC	0/50 (0.0)	ND	27
p15	ESCC	6/34 (18)	ND	26
	ESCC	5/40 (13)	ND	29
p16	ESCC	17/34 (50)	ND	26
	ESCC	18/28 (64)	0/30 (0.0)	28
	ESCC	16/40 (40)	ND	29
	EAC	27/50 (54)	ND	27
	EAC	8/21 (38)	ND	27
	EAC	16/41 (39)	10/41 (24)	42
FHIT	ESCC	25/36 (69)	ND	33
	ESCC	85/257 (33)	ND	35
RARβ2	ESCC	14/20 (70)	2/17 (12)	37
	ESCC	34/47 (73)	18/47 (38)	38
APC	EAC	48/52 (92)	0/52 (0.0)	41
	ESCC	16/32 (50)	0/32 (0.0)	41
	EAC	28/41 (68)	3/41 (7.3)	42
	EAC	39/50 (78)	ND	27
ER	EAC	21/41 (51)	5/41 (12)	42
MGMT	ESCC	46/119 (39)	0/21 (0.0)	45
	EAC	24/41 (56)	10/41 (24)	42
E-Cadherin	EAC	26/31 (84)	ND	49
	EAC	27/41 (66)	5/41 (12)	42
	ESCC	16/20 (80)	ND	50
TSLC1	ESCC	28/56 (50)	ND	53
RASSF1A	ESCC	24/47 (51)	2/47 (4.3)	38
	ESCC	22/64 (34)	3/64 (4.7)	58
hMLH1	ESCC	0/30 (0.0)	ND	59
	ESCC + EAC	79/124 (64)	ND	60
VHL	ESCC	6/47 (13)	0/47 (0)	39
TIMP3	EAC	8/41 (19)	0/41 (0.0)	42
	EAC	71/79 (90)	0/79 (0.0)	61
DAP-kinase	EAC	8/41 (19)	2/41 (4.9)	42
pRb	EAC	10/30 (33)	ND	62
ECRG4	ESCC	12/15 (80 )	3/20 (15)	63
Chfr	ESCC + EAC	7/43 (16)	ND	64
HLA class I	ESCC	13/29 (45)	ND	65
EYA4	EAC	33/40 (83)	2/58 (3.4)	66
CDH13	ESCC+EAC	5/37 (14)	ND	67
SFRP1	EAC	37/40 (93)	3/30 (10)	68
PGP9.5	ESCC	21/50 (42)	ND	69

ND: Not done.

From the above mentioned reports, we find that there is obvious different methylation frequency in a gene from different authors in some cases, which may be due to different assay methods and different specimen resources. And the geographical difference may be due to variable carcinogens in the different areas. Furthermore, there seems to be obvious difference of methylation frequency of some genes between ESCCs and EACs. For example, hypermethylation of the promoter region of the APC gene occurred in 48 of 52 (92%) patients with EAC, but in 16 of 32 (50%) patients with ESCC[41], which suggests that hypermethylation of the APC gene has distinct roles in ESCC and EAC.

METHYLATION IN SERUM DNA FROM ESOPHAGEAL CANCER PATIENTS

Despite advances in diagnosis and treatment of various cancers, early detection and treatment of cancer remain a challenge. One potential early detection biomarker is DNA methylation of the promoter region of certain cancer-associated genes[70]. Genetic analysis has shown that cell-free circulating DNA in plasma or serum of cancer patients shares similar genetic alterations to those described in the corresponding tumor[70,71]. Numerous studies have demonstrated the presence of promoter hypermethylation of tumor suppressor genes in the serum DNA of patients with various cancers[72]. Hypermethylated APC DNA was observed in the plasma of 13 of 52 (25%) EACs and 2 of 32 (6.3%) ESCCs[41]. Hibi et al[73] found that aberrant promoter methylation of the p16 gene was detected in 31 of 38 (82%) ESCCs, and 7 of these 31 (23%) patients with a p16 alteration in the primary tumor had the same methylation changes in the corresponding serum DNA. This study yielded a promising result: a tumor associated DNA alteration could be detected in the serum of 18% of ESCC patients (7 of 38 patients) using p16 methylation as a target. Moreover, the clinical sensitivity of this assay can be potentially improved by incorporating other possibly methylated target genes, which have been estimated in other tumor types. For example, Esteller et al[74] analyzed primary NSCLCs and serum from 22 patients for the methylation pattern of four TSGs (DAPK, GSTP1, p16, and MGMT). Methylation of at least one of these genes was detected in 68% of NSCLCs. Comparing primary tumors with methylation and matched serum samples, 73% of the matched serum samples were found to be methylated[74]. In addition, none of the sera from patients with tumors not demonstrating methylation were positive[74]. Therefore, combined detection of aberrant promoter hypermethylation of cancer-related genes in serum may be useful for esophageal cancer diagnosis or the detection of recurrence.

PROGNOSTIC SIGNIFICANCE OF GENE HYPERMETHYLATION

In the past few years, numerous attempts have been made to establish a genetic technique for reliably predicting tumor prognosis, but these attempts have been hindered by two main problems. First, only a few genes are somatically mutated in solid tumors and, second, because cell populations of primary neoplasms are heterogeneous, no single marker can accurately predict the behavior of the tumor[13]. Fortunately, emerging evidence suggests a possible prognostic value of gene promoter hypermethylation[75]. Lee et al[35] reported that aberrant methylation of the FHIT promoter in ESCC was found to be significantly associated with a poor prognosis for stage 1-2 cases. Mandelker et al[69] reported that PGP9.5 methylation was an independent prognostic factor for ESCC survival (P = 0.03) . Kawakami et al[41] reported that high plasma levels of methylated APC DNA were statistically significantly associated with reduced EAC patient survival. Brock et al[42] analyzed the methylation status of seven genes (including APC, E-cadherin, MGMT, ER, p16, DAP-kinase and TIMP3) of 41 EAC samples and found that DNA methylation of some genes individually showed only trends toward diminished survival, whereas patients whose tumors had > 50% of their gene profile methylated had both significantly poorer survival and earlier tumor recurrence than those without positive methylation. The data suggest that combined detection of methylation status for multiple genes is an effective strategy for prediction of esophageal tumor behavior. Although some genes that are frequently inactivated by methylation and are of prognostic impact for esophageal cancer patients have already been found, additional genes need to be identified. Thus, patients with a worse prognosis could be selected. These patients might benefit from a more aggressive treatment strategy.

ABERRANT DNA METHYLATION IS AN EARLY EVENT IN ESOPHAGEAL CARCINOGENESIS

In many tumors, it has been proved that aberrant DNA methylation frequently occurs in precancerous tissue as well as cancer tissue, and both factors, genetic and epigenetic, lie at the origin of carcinogenesis[76]. The relative contribution of each varies significantly in different human tumors[76]. Nie et al[77] compared hypermethylation of p16, p15, p14, HLA-A, -B, -C, hMLH1, E-cadherin, FHIT and VHL genes in precancerous esophageal tissues and found that in 48 biopsy samples with BCH or DYS, the most frequent hypermethylated genes were p16 (19%) and p14^ARF (15%), and seventeen out of these 48 samples (35%) contained hypermethylation of at least one gene. In the resected tissues, 52% of the BCH and 81% of the tumors showed hypermethylation of at least one gene. Another study reported that 2 of 17 (12%) normal esophagus, 9 of 21 (43%)BCH, 7 of 12 (58%) DYS, and 14 of 20 (70%) ESCC samples had hypermethylation of the RARβ₂ promoter region[37]. As to progression of EAC, it has been reported that, methylation of the p16 promoter was detected in 18 of 22 (82%) EAC and 10 of 33 (30%) premalignant lesions, whereas no methylation of the p16 promoter was found in normal esophageal epithelia[78]. Hardie et al[30] reported hypermethylation of the p16 promoter was detected in 77% (14/18) of Barrett’s epithelia, and in 85% (18/21) of EACs. These data suggest that aberrant DNA methylation participates early in the development of esophageal cancer. Recently, the lab of professor Yang CS reported that EGCG, the major polyphenol from green tea, inhibited DNMT activity and reactivated several methylation-silenced genes, including p16, RARβ₂, MGMT and hMLH1, in human esophageal cancer KYSE 510 cells, accompanied by the expression of mRNA of these genes[79]. The result suggests that methylation might be a new target of chemopreventive activity. In the last two decades, it has been proven that many drugs, such as tamoxifen, aspirin, COX-2 inhibitors, possess positive chemopreventive activity against esophageal cancer[2]. However, the exact mechanisms have not been elucidated so far. Therefore, it will be very attractive to examine the effect of these drugs on promoter methylation status of key genes in esophageal cancer cells, esophageal cancer tissue, and especially precancerous tissue of the esophagus.

HYPERMETHYLATION AS A TARGET OF THERAPEUTIC INTERVENTION

It has been reported that demethylating agents 5-azacytidine and 5-aza-2’-deoxycytidine can restore the normal demethylated state of several types of TSGs and increase their expression in various cancers, including esophageal cancer, in vitro and in vivo[13,14,15,39,54]. Since methylation and transcriptional status are inversely correlated, the use of demethylating agents appears to be a promising option for the treatment of tumors. Methylation of genes in tumor cells could provide a tumor-specific target for new therapies[80,81,82]. In fact, these demethylating agents have exhibited significant activity in the treatment of patients with myelodysplastic syndrome, chronic myeloid leukaemia and acute myeloid leukaemia[83,84]. However, preliminary experience with these agents in solid tumors has been relatively low[85]. Esophageal tumor shows a high prevalence of TSG hypermethylation, and the above studies demonstrated that gene expression could be restored after treatment of esophageal tumor cells with demethylating agents in vitro. However, up to date the clinical trial about demethylating agents in esophageal cancer is unavailable. Although it is too early to make any expectation about the effect of these drugs on esophageal cancer, this is a very promising concept and needs to be tested in clinical trials.

CONCLUSION AND PERSPECTIVES

Esophageal carcinogenesis is a stepwise process of the accumulation of genetic and epigenetic abnormalities. It has become clear that promoter hypermethylation of TSGs is as important for this multistep process as genetic changes in the progression of esophageal carcinogenesis. The steadily growing list of genes inactivated by promoter hypermethylation in esophageal carcinoma provides not only new insights into the molecular basis of the diseases but also a long list of interesting candidate genes for the development of molecular markers which might contribute to the improvement of diagnosis and also prognosis. In addition, the fact that methylation can be reversed in vitro and the effect of the demethylating agent 5-aza-2’-deoxycytidine in vitro raise hope for new treatment strategies for esophageal cancer patients. Furthermore, understanding of the significance of aberrant DNA methylation in the precancerous stage may show that a new strategy, correction of aberrant DNA methylation, can prevent esophageal cancer in people with premalignant lesions, such as Barrett’s esophagus, BCH and DYS.

ACKNOWLEDGMENTS

We acknowledge the help given by Li-Dong Wang, Professor of Pathology and Oncology, College of Medicine, Zhengzhou University, China.