Integrative Analysis of DNA Methylation and microRNA Reveals GNPDA1 and SLC25A16 Related to Biopsychosocial Factors Among Taiwanese Women with a Family History of Breast Cancer

Abstract

Biopsychosocial factors, including family history, influence the development of breast cancer. Malignancies in women with a family history of breast cancer may be detectable based on DNA methylation and microRNA. Objectives: The present study extended an integrative analysis of DNA methylation and microRNA to identify genes associated with biopsychosocial factors. Methods: We identified 3060 healthy women from the Taiwan Biobank and included 32 blood plasma samples for analysis of biopsychosocial factors and epigenetic changes. GEO databases and bioinformatics approaches were used for the identification and validation of potential genes. Results: Our integrative analysis revealed GNPDA1 and SLC25A16 as potential genes. Age, a family history of cancer, and alcohol consumption were associated with GNPDA1 and SLC25A16 based on the current data set and the GEO data set. GNPDA1 and SLC25A16 exhibited significant expression in breast cancer tissues based on UALCAN analysis, where they were overexpressed and underexpressed, respectively. Through a MethSurv analysis, GNPDA1 hypomethylation and SLC25A16 hypermethylation were associated with poor prognoses in terms of overall survival in breast cancer. Moreover, through a MetaCore functional enrichment analysis, GNPDA1 and SLC25A16 were associated with the BRCA1, BRCA2, and pro-oncogenic actions of the androgen receptor in breast cancer. Further, GNPDA1 and SLC25A16 were enriched in known targets of approved cancer drugs as potential genes associated with breast cancer. Conclusions: These two genes might serve as biomarkers for the early detection of breast cancer, especially for women with a family history of breast cancer.

Keywords: DNA methylation; biopsychosocial factors; family history of breast cancer; integrative analysis; microRNA.

Figure 1

Phases of analytical approach and research design. FHBC, family history of breast cancer; QPCR, quantitative polymerase chain reaction; cDNA, complementary DNA; DEM, differentially methylated; DEmiRNA, differetially expressed microRNA LIMMA, linear models of microarray; DESeq2, differential gene expression analysis of RNA-seq2; GEO, Gene Expression Omnibus; UALCAN, University of ALabama at Birmingham CANcer; MethSurv, metylation survival.

Figure 2

Differential analysis of methylated DNA and miRNA. (A) A volcano plot of DEM genes in the FHBC and non-FHBC groups. Significant CpGs in a volcano graphic with logFC ≥ 1.0 (p < 0.05). (B) A volcano plot of DEmiRNA with logFC ≥ 2.0 (p < 0.05).

Figure 3

The identification and validation of potential genes. (A) Screening strategy for CpGs of interest from DEM genes of three data sets. GSE88883 (Data set 1), GSE67919 (Data sets 2a and 2b), current study (Data set 3). (B) Different colored areas represent different data sets. The crossed areas correspond to the common differentially methylated. Data sets from the GEO database were analyzed with GEO2R, and the current study data set was analyzed using LIMMA (p < 0.05). (C) Functional overlap between DEM genes of interest and DEmiRNA of current study. (FH, family history of cancer.)

Figure 4

The expression and methylation level analyses of the potential genes in breast cancer (TCGA database) as well as survival analysis (MethSurv databases). (A,B) A box plot of two potential genes transcripts and beta values in normal (n = 114) and primary tumor (BRCA) tissue (n = 1097). Statistical significance was indicated by t-test *** and p < 0.0001 (left). Individual cancer stages are displayed in box plots (middle). The Kaplan–Meier curves represent the survival analysis of methylated genes in breast cancer (right).

Figure 5

The expression of the GNPDA1 signaling pathway in breast cancer (MetaCore). This platform analyzed 10% of co-expressed genes that are correlated with GNPDA1 with a Spearman partial rho of ≥0.3. (A) Pathway distribution. We found that BRCA1 and BRCA2 were correlated with breast cancer development (with p < 0.05 as a cutoff value). (B) The signaling pathway of BRCA1 and BRCA2 in patients with breast cancer.

Figure 6

Connection between potential biomarkers of breast cancer genes, protein–protein interactions (PPIs) genes, and drugs available for breast cancer.