Epigenetic patterns in the progression of esophageal adenocarcinoma.

Esophageal adenocarcinoma (EAC) arises after normal squamous mucosa undergoes metaplasia to specialized columnar epithelium (intestinal metaplasia or Barrett's esophagus), which can then ultimately progress to dysplasia and subsequent malignancy. Epigenetic studies of this model have thus far been limited to the DNA methylation analysis of a few genes. In this study, we analyzed a panel of 20 genes using a quantitative, high-throughput methylation assay, METHYLIGHT: We used this broader approach to gain insight into concordant methylation behavior between genes and to generate epigenomic fingerprints for the different histological stages of EAC. Our study included a total of 104 tissue specimens from 51 patients with different stages of Barrett's esophagus and/or associated adenocarcinoma. We screened 84 of these samples with the full panel of 20 genes and found distinct classes of methylation patterns in the different types of tissue. The most informative genes were those with an intermediate frequency of significant hypermethylation [ranging from 15% (CDKN2A) to 60% (MGMT) of the samples]. This group could be further subdivided into three classes, according to the absence (CDKN2A, ESR1, and MYOD1) or presence (CALCA, MGMT, and TIMP3) of methylation in normal esophageal mucosa and stomach, or the infrequent methylation of normal esophageal mucosa accompanied by methylation in all normal stomach samples (APC). The other genes were less informative, because the frequency of hypermethylation was below 5% (ARF, CDH1, CDKN2B, GSTP1, MLH1, PTGS2, and THBS1), completely absent (CTNNB1, RB1, TGFBR2, and TYMS1), or ubiquitous (HIC1 and MTHFR), regardless of tissue type. Each class undergoes unique epigenetic changes at different steps of disease progression of EAC, suggesting a step-wise loss of multiple protective barriers against CpG island hypermethylation. The aberrant hypermethylation occurs at many different loci in the same tissues, suggestive of an overall deregulation of methylation control in EAC tumorigenesis. However, we did not find evidence for a distinct group of tumors with a CpG island methylator phenotype. Finally, we found that normal and metaplastic tissues from patients with evidence of associated dysplasia or cancer had a significantly higher incidence of hypermethylation than similar tissues from patients with no further progression of their disease. The fact that the samples from these two groups of patients were histologically indistinguishable, yet molecularly distinct, suggests that the occurrence of such hypermethylation may provide a clinical tool to identify patients with premalignant Barrett's who are at risk for further progression.