Identification of Disease-Related Genes That Are Common between Alzheimer’s and Cardiovascular Disease Using Blood Genome-Wide Transcriptome Analysis

Abstract

Accumulating evidence has suggested a shared pathophysiology between Alzheimer’s disease (AD) and cardiovascular disease (CVD). Based on genome-wide transcriptomes, specifically those of blood samples, we identify the shared disease-related signatures between AD and CVD. In addition to gene expressions in blood, the following prior knowledge were utilized to identify several candidate disease-related gene (DRG) sets: protein–protein interactions, transcription factors, disease–gene relationship databases, and single nucleotide polymorphisms. We selected the respective DRG sets for AD and CVD that show a high accuracy for disease prediction in bulk and single-cell gene expression datasets. Then, gene regulatory networks (GRNs) were constructed from each of the AD and CVD DRG sets to identify the upstream regulating genes. Using the GRNs, we identified two common upstream genes (GPBP1 and SETDB2) between the AD and CVD GRNs. In summary, this study has identified the potential AD- and CVD-related genes and common hub genes between these sets, which may help to elucidate the shared mechanisms between these two diseases.

Pan-Cancer Analysis Reveals the Prognostic Potential of the THAP9/THAP9-AS1 Sense–Antisense Gene Pair in Human Cancers

“…Some genes that are overrepresented in these clusters are GPBP1, API5, PIK2C3, GOSR1 and PRPF40A. GPBP1, also known as Vasculin, is a promoter binding protein that is reported to have roles in atherosclerosis [ 55 ], hypertension, hypercholesterolemia [ 56 ], and Alzheimer’s disease [ 57 , 58 ]. AP15 prevents apoptosis in the absence of growth factor [ 59 , 60 , 61 , 62 ], while the BECN1-PIK3C3 complex plays a crucial role in autophagy [ 63 ].…”

Identification of Cardiovascular Disease-Related Genes Based on the Co-Expression Network Analysis of Genome-Wide Blood Transcriptome

“…A previous study retrieved 11 blood CVD gene expression datasets [ 16 ] from the Gene Expression Omnibus (GEO) database [ 19 ]. The integration of numerous transcriptomic datasets can cause several problems, including the loss of many transcripts (also termed as probe or probe-set) and disease-related signatures (such as fold change (FC) between two statuses) [ 18 ].…”

“…The integration of numerous transcriptomic datasets can cause several problems, including the loss of many transcripts (also termed as probe or probe-set) and disease-related signatures (such as fold change (FC) between two statuses) [ 18 ]. Therefore, Lee and Lee [ 16 ] selected the blood CVD gene expression datasets for the final analysis because of the good quality of these datasets, as measured by MetaQC tool [ 20 ]. In this study, four blood datasets were selected as representative CVD datasets.…”

“…Peripheral blood is readily available from both humans and mice, and recent evidence has shown that it can reflect the transcriptomic signature of other tissues [ 14 , 15 ]. Lee and Lee [ 16 ] utilized blood transcriptome data and identified several CVD-related gene sets using statistical methods, disease-gene databases, and eQTL and GWAS summary statistics. However, only differential expression (DE) analysis was conducted in this study [ 16 ], and, therefore, it could not explain the co-dysregulated patterns among disease status genes.…”

“…Lee and Lee [ 16 ] utilized blood transcriptome data and identified several CVD-related gene sets using statistical methods, disease-gene databases, and eQTL and GWAS summary statistics. However, only differential expression (DE) analysis was conducted in this study [ 16 ], and, therefore, it could not explain the co-dysregulated patterns among disease status genes. Several studies have performed differential co-expression (DcoE) analyses to identify altered gene–gene interactions under disease conditions [ 17 , 18 ].…”