ABSTRACT
p16 hypermethylation in Barrett's carcinogenesis has been evaluated in studies which did not take into account sample heterogeneity and yielded qualitative (methylated/unmethylated) instead of accurate quantitative (percentage of CpG methylation) data. We aimed to measure the degree of p16 methylation in pure samples representing all the steps of Barrett's tumorogenesis and to evaluate the influence of sample heterogeneity in methylation analysis. METHODS: 77 paraffin-embedded human esophageal samples were analyzed. Histological grading was established by two pathologists in: negative for dysplasia, indefinite for dysplasia, low-grade dysplasia, high-grade dysplasia and adenocarcinoma. Areas of interest were selected by laser-capture microdissection. p16 methylation was quantified by pyrosequencing. An adjacent section of the whole sample was also analyzed to compare methylation data. RESULTS: After microdissection, we obtained 15 samples of squamous epithelium, 36 non-dysplastic Barrett's esophagus, 3 indefinite for dysplasia, 24 low-grade dysplasia, 4 high-grade dysplasia and 12 adenocarcinoma. Squamous epithelium showed the lowest methylation rates: 6% (IQR 5-11) vs. 11%(7-39.50) in negative/indefinite for dysplasia, p<0.01; 10.60%(6-24) in low-grade dysplasia, p<0.05; and 44.50%(9-66.75) in high-grade dysplasia/adenocarcinoma, p<0.01. This latter group also exhibited higher methylation rates than Barrett's epithelium with and without low-grade dysplasia (p<0.05). p16 methylation rates of microdissected and non-microdissected samples did not correlate unless the considered histological alteration comprised >71% of the sample. CONCLUSIONS: p16 methylation is an early event in Barrett's carcinogenesis which increases with the severity of histological alteration. p16 methylation rates are profoundly influenced by sample heterogeneity, so selection of samples is crucial in order to detect differences.