Integrative analysis of DNA methylation and gene expression profiles identified potential breast cancer-specific diagnostic markers.

Xinhua Liu,Yonglin Peng,Ju Wang

doi:10.1042/bsr20201053

Xinhua Liu, Yonglin Peng + Show 1 more

Open Access

https://doi.org/10.1042/bsr20201053

Copy DOI

Journal: Bioscience reports	Publication Date: May 27, 2020
Citations: 13	License type: CC BY 4.0

Affiliation: Tianjin Medical University

Abstract

Breast cancer is a common malignant tumor among women whose prognosis is largely determined by the period and accuracy of diagnosis. We here propose to identify a robust DNA methylation-based breast cancer-specific diagnostic signature. Genome-wide DNA methylation and gene expression profiles of breast cancer patients along with their adjacent normal tissues from the Cancer Genome Atlas (TCGA) were obtained as the training set. CpGs that with significantly elevated methylation level in breast cancer than not only their adjacent normal tissues and the other ten common cancers from TCGA but also the healthy breast tissues from the Gene Expression Omnibus (GEO) were finally remained for logistic regression analysis. Another independent breast cancer DNA methylation dataset from GEO was used as the testing set. Lots of CpGs were hyper-methylated in breast cancer samples compared with adjacent normal tissues, which tend to be negatively correlated with gene expressions. Eight CpGs located at RIIAD1, ENPP2, ESPN, and ETS1, were finally retained. The diagnostic model was reliable in separating BRCA from normal samples. Besides, chromatin accessibility status of RIIAD1, ENPP2, ESPN and ETS1 showed great differences between MCF-7 and MDA-MB-231 cell lines. In conclusion, the present study should be helpful for breast cancer early and accurate diagnosis.

Highlights

Breast cancer (BRCA) is one of the most common cancers and the second cause of cancer-related death among women [1]
Training set, which was used for screening breast cancer (BRCA)-specific diagnostic markers, is composed of tumor and adjacent normal tissues of BRCA and other 10 cancers, including BLCA, COAD, GBM, HNSCC, KIRC, LIHC, LUAD, LUSC, READ, and UCEC in the Cancer Genome Atlas (TCGA)
CpG island methylator phenotype (CIMP), which is used for describing CpG island promoter methylation of specific tumors and closely correlated with tumorigenesis and prognosis, was initially developed in colorectal cancer and has been extremely applied in many other cancers including BRCA

Summary

Introduction

Breast cancer (BRCA) is one of the most common cancers and the second cause of cancer-related death among women [1]. Lots of risk factors, including environmental and genetic aspects, have been increasingly revealed to be carcinogenic in BRCA. Imaging techniques, including mammography, computed tomography, magnetic resonance imaging, and so on, have been used as the primary means for BRCA diagnosis and patients monitoring until now [4,5]. Several biochemical-based methods have been developed for overcoming potential limitations of imaging methods, such as high spending and low sensitivity or specificity in special cases [6]. Lots of BRCA-specific biochemical biomarkers, in addition to the pan-cancer markers, have been identified and proved their potential applications in BRCA diagnosis and stage monitoring. MicroRNA, a type of regulatory small molecule that without protein-coding potential, which plays pivotal roles in post-transcriptional gene expression regulations, is emerging as BRCA diagnostic and prognostic biomarkers [10,11,12,13].

Methods

Results

Conclusion