Objectives: Using pharmacologic unmasking and genome-wide differential methylation analysis, we identified a novel methylated gene in ovarian cancers.
Methods: Two ovarian cancer cells (OVCAR-3, ES-2) that showed synergistic growth inhibition by 5-aza-dC and cisplatin were selected. After treatment with 5-aza-dC, differential expression profiles were compared using microarray that contained 38,500 genes. Reactivation of candidate genes and their promoter methylation were validated by real-time RT-PCR, MS-PCR and bisulfite sequencing. Methylation status was tested by MS-PCR in 56 patients with epithelial ovarian cancer and compared to the 38 normal ovarian tissues.
Results: We identified 103 candidate genes that were reactivated by 5-aza-dC treatment. Among those, SFN and TGFBI were commonly reactivated in both cells. Since SFN is a well known methylated marker, we selected TGFBI for further validation. Bisulfite sequencing revealed complete promoter methylation in ES-2 and partial methylation in OVCAR-3. In addition, silencing of TGFBI at the transcription level was reversed by 5-aza-dC treatment. TGFBI methylation was observed in 23 out of 38 (60.5%) cases of ovarian cancer, in no normal ovarian tissues (0 of 38, P=0.001), and in 5 out of 18 (27.8%) borderline tumors (P=0.044). In our cohort, we did not observe any association between methylation of TGFBI and clinicopathologic variables or clinical outcomes.
Conclusion: Our results confirm that TGFBI is frequently methylated in ovarian cancer. Its methylation can be used as a novel epigenetic biomarker in discriminating ovarian cancer from non-cancer or borderline tumors.
Copyright 2010 Elsevier Inc. All rights reserved.