DNA methylation characteristics associated with chemotherapy resistance in epithelial ovarian cancer.

OBJECTIVE

The high mortality rate of epithelial ovarian cancer (EOC) is often attributed to the frequent development of chemoresistance. DNA methylation is a predictive biomarker for chemoresistance.

METHODS

This study utilized DNA methylation profiles and relevant information from GEO and TCGA to identify different methylated CpG sites (DMCs) between chemoresistant and chemosensitive patients. Subsequently, we constructed chemoresistance risk models with DMCs. The genes corresponding to candidate DMCs in chemoresistance risk models were further analyzed to identify different methylated gene symbols (DMGs) associated with chemoresistance. The DMGs that showed a strong correlation with the corresponding DMCs were analyzed through immunohistochemistry.

RESULTS

Compared to chemosensitive EOC patients, chemoresistant patients showed 423 hypermethylated CpGs and 1445 hypomethylated CpGs. The chemoresistance risk models based on DMCs have shown the improved predictive ability for chemoresistance in EOC (AUC = 65.0-76.2%). The methylations of cg25510164, cg13154880, cg15362155 and cg08665359 were strongly associated with decreased risk of chemoresistance. Conversely, the methylation of cg08872590 and cg14739437 significantly increased the risk. We identified 13 DMGs, from 47 DMCs corresponding genes, between chemosensitive and chemoresistant samples. Among the DMGs, the expression levels of DDR2 and OPCML exhibited strong correlations with the corresponding DMCs. DDR2 and OPCML both showed enhanced expression in chemoresistant ovarian microarray tissue.

CONCLUSIONS

Hypomethylated CpGs may play a significant role in DNA methylation associated with chemoresistance in EOC. The epigenetic modification of DDR2 could have important implications for the development of chemoresistance. Our study provides valuable insights for future research on DNA methylation in the chemoresistance of EOC.