Genome-wide DNA methylation profiling reveals diagnostic markers in parathyroid tumours.

PURPOSE: Parathyroid carcinoma (PC) and adenoma (PA) differ in clinical features and molecular mechanisms. We aimed to investigate differences of DNA methylation pattern and epigenetic impacts on gene expression in parathyroid tumours. METHODS: Genome-wide DNA methylation patterns of PC and PA from 24 patients were evaluated using an Infinium Methylation EPIC BeadChip array. Integrated analysis of differentially methylated positions (DMPs) and transcript expression in 5 PC and 6 PA samples was performed. RT-qPCR and pyrosequencing were used for further validation in an expansion cohort. RESULTS: Compared to PA, 85,317 DMPs were identified in PC, including 73,221 hyper-methylated and 12,096 hypo-methylated. Negative correlation between DNA methylation and transcript expression, which comprised 80% of these DMPs, suggested epigenetic regulation of 41 methylation-transcript pairs (p < 0.05). Based on DMR analysis, methylation-transcript pairs of PDZK1, FAM69A, SRGAP3, TAL1 were further selected for validation. PDZK1 and cg13019092 were the most predominantly correlated (r = -0.72, p < 0.001). Combined cg15495124 and cg13019092 potentially discriminated PC from PA in our study. In addition, cg15495124 and cg13019092 correlated with CDC73 abnormalities. CONCLUSION: PC exhibited frequent hyper-methylated DMPs compared to PA. DNA methylation may help identify the key epigenetic factors contributing to parathyroid tumorigenesis and support differential diagnosis between PC and PA.