Identifying lncRNA-miRNA-mRNA networks to investigate Alzheimer's disease pathogenesis and therapy strategy

Abstract

Alzheimer's disease (AD), the most common cause of dementia, leads to neuronal damage and deterioration of cognitive functions in aging brains. There is evidence suggesting the participation of noncoding RNAs in AD-associated pathophysiology. A potential linkage between AD and lncRNA-associated competing endogenous RNA (ceRNA) networks has been revealed. Nevertheless, there are still no genome-wide studies which have identified the lncRNA-associated ceRNA pairs involved in AD. For this reason, deep RNA-sequencing was performed to systematically investigate lncRNA-associated ceRNA mechanisms in AD model mice (APP/PS1) brains. Our results identified 487, 89, and 3,025 significantly dysregulated lncRNAs, miRNAs, and mRNAs, respectively, and the most comprehensive lncRNA-associated ceRNA networks to date are constructed in the APP/PS1 brain. GO analysis revealed the involvement of the identified networks in regulating AD development from distinct origins, such as synapses and dendrites. Following rigorous selection, the lncRNA-associated ceRNA networks in this AD mouse model were found to be mainly involved in synaptic plasticity as well as memory (Akap5) and regulation of amyloid-β (Aβ)-induced neuroinflammation (Klf4). This study presents the first systematic dissection of lncRNA-associated ceRNA profiles in the APP/PS1 mouse brain. The identified lncRNA-associated ceRNA networks could provide insights that facilitate AD diagnosis and future treatment strategies.

Keywords: APP/PS1 mouse; Alzheimer’s disease (AD); RNA-sequencing; ceRNA network; lncRNA.

Figure 1

The workflow of RNA-seq. Details of the methods used for mRNA-seq, miRNA-seq, and lncRNA-seq are described in Supplementary Materials.

Figure 2

Expression profiles of distinct RNAs. (A–C) Expression profiles of lncRNAs. (A, B) In the volcano plots, green, red, and black points represent lncRNAs that were downregulated, upregulated, and not significantly different in APP/PS1 mice relative to wild-type (WT) control mice at 6 and 9 months, respectively. x-axis: log2 ratio of lncRNA expression levels between AD and WT. y-axis: false-discovery rate values (-log10 transformed) of lncRNAs, P<0.05 (C) Cluster analysis of expression of lncRNAs. Red and blue: increased and decreased expression at 6 and 9 months, respectively. Expression profiles are similarly shown for (D–F) miRNAs, p<0.04 and (G–I) mRNAs, q<0.05.

Figure 3

Validation of expression of lncRNAs by using qPCR. The identified differentially expressed transcripts (lncRNAs, miRNAs, and mRNAs) were divided into three groups. (A) 6yes9no group represents transcripts differential expressed at 6 months but not at 9 months; (B) 6no9yes group represents transcripts not differential expressed at 6 months but differential expressed at 9 months; (C) 6yes9yes group represents transcripts differential expressed at both 6 and 9 months. The expression of lncRNAs was quantified relative to Gapdh expression level by using the comparative cycle threshold (ΔCT) method. Data are presented as means ± SD (n = 3, *p < 0.05, **p < 0.01).

Figure 4

Validation of miRNA expression by using qPCR. (A) 6yes9no group, (B) 6no9yes group, and (C) 6yes9yes group. The expression levels of miRNAs were quantified relative to U6 expression level by using the comparative cycle threshold (ΔCT) method. Data are presented as means ± SD (n = 3, *p < 0.05, **p < 0.01).

Figure 5

Validation of mRNA expression by using qPCR. (A) 6yes9no group, (B) 6no9yes group, and (C) 6yes9yes group. The mRNA expression was quantified relative to Gapdh expression level by using the comparative cycle threshold (ΔCT) method. Data are presented as means ± SD (n = 3, *p < 0.05, **p < 0.01).

Figure 6

The lncRNA-associated ceRNA networks in APP/PS1 mice. CeRNA networks were constructed based on identified lncRNA–miRNA and miRNA–mRNA interactions. The networks include increased lncRNAs, decreased miRNAs, and increased mRNAs in APP/PS1 mice. (A) Grouping (B) 6yes9no group, (C) 6no9yes group, and (D) 6yes9yes group.

Figure 7

Identified lncRNA-associated ceRNA networks in APP/PS1 mice. The ceRNA networks were constructed based on identified lncRNA–miRNA and miRNA–mRNA interactions. The networks include decreased lncRNAs, increased miRNAs, and decreased mRNAs in APP/PS1 mice. (A) 6yes9no group, and (B) 6no9yes group.

Figure 8

Gene Ontology (GO) enrichment annotations of the pathological progression of AD: biological process, cellular component, molecular function. The top terms were synapse (GO:0045202), cytoskeleton (GO:0005856), postsynaptic density (GO:0014069), cell-cell adherens junction (GO:0005913), dendrite (GO:0030425), axon (GO:0030424), and neuron projection (GO:0043005). (A) 6yes9no group, (B) 6no9yes group, and (C) 6yes9yes group. Significantly enriched GO pathways were defined as p values of <0.01. GO analysis was conducted with DAVID (https://david.ncifcrf.gov/summary.jsp) database.

Figure 9

Significantly enriched Kyoto Encyclopedia of Genes and Genomes (KEGG) and Reactome pathways. The identified lncRNA-associated ceRNA-network genes participate in distinct aspects of AD pathology. (A) 6yes9no group, (B) 6no9yes group, and (C) 6yes9yes group. Significantly enriched KEGG pathways with p values of <0.05. Each line represents a gene, and the number of lines indicates the genes enriched. KEGG analysis was conducted with DAVID (https://david.ncifcrf.gov/summary.jsp) database.