Integrated analysis of gene expression and methylation profiles of 48 candidate genes in breast cancer patients
Zibo Li,Jianfu Heng,Jinhua Yan,Xinwu Guo,Lili Tang,Ming Chen,Limin Peng,Yepeng Wu,Shouman Wang,Zhi Xiao,Zhongping Deng,Lizhong Dai,Jun Wang
DOI: https://doi.org/10.1007/s10549-016-4004-8
2016-01-01
Breast Cancer Research and Treatment
Abstract:Purpose Gene-specific methylation and expression have shown biological and clinical importance for breast cancer diagnosis and prognosis. Integrated analysis of gene methylation and gene expression may identify genes associated with biology mechanism and clinical outcome of breast cancer and aid in clinical management. Methods Using high-throughput microfluidic quantitative PCR, we analyzed the expression profiles of 48 candidate genes in 96 Chinese breast cancer patients and investigated their correlation with gene methylation and associations with breast cancer clinical parameters. Results Breast cancer-specific gene expression alternation was found in 25 genes with significant expression difference between paired tumor and normal tissues. A total of 9 genes ( CCND2, EGFR, GSTP1, PGR, PTGS2, RECK, SOX17, TNFRSF10D , and WIF1 ) showed significant negative correlation between methylation and gene expression, which were validated in the TCGA database. Total 23 genes ( ACADL , APC , BRCA2 , CADM1 , CAV1 , CCND2 , CST6 , EGFR , ESR2 , GSTP1 , ICAM5 , NPY , PGR , PTGS2 , RECK , RUNX3 , SFRP1 , SOX17 , SYK , TGFBR2 , TNFRSF10D , WIF1, and WRN ) annotated with potential TFBSs in the promoter regions showed negative correlation between methylation and expression. In logistics regression analysis, 31 of the 48 genes showed improved performance in disease prediction with combination of methylation and expression coefficient. Conclusions Our results demonstrated the complex correlation and the possible regulatory mechanisms between DNA methylation and gene expression. Integration analysis of methylation and expression of candidate genes could improve performance in breast cancer prediction. These findings would contribute to molecular characterization and identification of biomarkers for potential clinical applications.