Similarity of aberrant DNA methylation in Barrett's esophagus and esophageal adenocarcinoma

Eric Smith¹ , Neville J De Young¹ , Sandra J Pavey² , Nicholas K Hayward² , Derek J Nancarrow² , David C Whiteman³ , B Mark Smithers⁴ , Andrew R Ruszkiewicz⁵ , Andrew D Clouston⁴ , David C Gotley⁴ , Peter G Devitt¹ , Glyn G Jamieson¹ and Paul A Drew⁶

¹Discipline of Surgery, School of Medicine, The University of Adelaide, Royal Adelaide Hospital, Adelaide, South Australia, 5005, Australia

²Oncogenomics, Queensland Institute of Medical Research, Brisbane, 4006, Australia

³Division of Population Studies and Human Genetics, Queensland Institute of Medical Research, Brisbane, 4006, Australia

⁴School of Medicine, University of Queensland, Brisbane, Queensland, 4072, Australia

⁵Division of Tissue Pathology, Institute of Medical and Veterinary Science, Adelaide, South Australia, 5000, Australia

⁶School of Nursing and Midwifery, Flinders University, Bedford Park, South Australia, 5042, Australia

author email corresponding author email

Molecular Cancer 2008, 7:75doi:10.1186/1476-4598-7-75


Published:	2 October 2008

Abstract

Background

Barrett's esophagus (BE) is the metaplastic replacement of squamous with columnar epithelium in the esophagus, as a result of reflux. It is the major risk factor for the development of esophageal adenocarcinoma (EAC). Methylation of CpG dinucleotides of normally unmethylated genes is associated with silencing of their expression, and is common in EAC. This study was designed to determine at what stage, in the progression from BE to EAC, methylation of key genes occurs.

Results

We examined nine genes (APC, CDKN2A, ID4, MGMT, RBP1, RUNX3, SFRP1, TIMP3, and TMEFF2), frequently methylated in multiple cancer types, in a panel of squamous (19 biopsies from patients without BE or EAC, 16 from patients with BE, 21 from patients with EAC), BE (40 metaplastic, seven high grade dysplastic) and 37 EAC tissues. The methylation frequency, the percentage of samples that had any extent of methylation, for each of the nine genes in the EAC (95%, 59%, 76%, 57%, 70%, 73%, 95%, 74% and 83% respectively) was significantly higher than in any of the squamous groups. The methylation frequency for each of the nine genes in the metaplastic BE (95%, 28%, 78%, 48%, 58%, 48%, 93%, 88% and 75% respectively) was significantly higher than in the squamous samples except for CDKN2A and RBP1. The methylation frequency did not differ between BE and EAC samples, except for CDKN2A and RUNX3 which were significantly higher in EAC. The methylation extent was an estimate of both the number of methylated alleles and the density of methylation on these alleles. This was significantly greater in EAC than in metaplastic BE for all genes except APC, MGMT and TIMP3. There was no significant difference in methylation extent for any gene between high grade dysplastic BE and EAC.

Conclusion

We found significant methylation in metaplastic BE, which for seven of the nine genes studied did not differ in frequency from that found in EAC. This is also the first report of gene silencing by methylation of ID4 in BE or EAC. This study suggests that metaplastic BE is a highly abnormal tissue, more similar to cancer tissue than to normal epithelium.