Aberrant Expression Profiles of lncRNAs and Their Associated Nearby Coding Genes in the Hippocampus of the SAMP8 Mouse Model with AD

Abstract

The senescence-accelerated mouse prone 8 (SAMP8) mouse model is a useful model for investigating the fundamental mechanisms involved in the age-related learning and memory deficits of Alzheimer's disease (AD), while the SAM/resistant 1 (SAMR1) mouse model shows normal features. Recent evidence has shown that long non-coding RNAs (lncRNAs) may play an important role in AD pathogenesis. However, a comprehensive and systematic understanding of the function of AD-related lncRNAs and their associated nearby coding genes in AD is still lacking. In this study, we collected the hippocampus, the main area of AD pathological processes, of SAMP8 and SAMR1 animals and performed microarray analysis to identify aberrantly expressed lncRNAs and their associated nearby coding genes, which may contribute to AD pathogenesis. We identified 3,112 differentially expressed lncRNAs and 3,191 differentially expressed mRNAs in SAMP8 mice compared to SAMR1 mice. More than 70% of the deregulated lncRNAs were intergenic and exon sense-overlapping lncRNAs. Gene Ontology (GO) and pathway analyses of the AD-related transcripts were also performed and are described in detail, which imply that metabolic process reprograming was likely related to AD. Furthermore, six lncRNAs and six mRNAs were selected for further validation of the microarray results using quantitative PCR, and the results were consistent with the findings from the microarray. Moreover, we analyzed 780 lincRNAs (also called long "intergenic" non-coding RNAs) and their associated nearby coding genes. Among these lincRNAs, AK158400 had the most genes nearby (n = 13), all of which belonged to the histone cluster 1 family, suggesting regulation of the nucleosome structure of the chromosomal fiber by affecting nearby genes during AD progression. In addition, we also identified 97 aberrant antisense lncRNAs and their associated coding genes. It is likely that these dysregulated lncRNAs and their associated nearby coding genes play a role in the development and/or progression of AD.

Keywords: Alzheimer’s disease; lncRNA-associated nearby genes; lncRNAs.

Figure 1

Learning and Memory Abilities of SAMP8 Mice (A) Escape latency of SAMP8 (n = 10) mice and SAMR1 (n = 10) littermates in the Morris water maze test. ∗∗p < 0.01, ∗∗∗p < 0.001. (B) Drawings are representations of single-mouse distances to the platform location. (C) Number of crossings in the probe trial test. ∗∗p < 0.01. (D) Time spent in the target quadrant in the probe trial test. ∗∗p < 0.01.

Figure 2

Overview of lncRNA Expression Profiles in Hippocampal Tissues of 8-Month-Old SAMP8 Mice Compared with Age-Matched SAMR1 Mice (A) Hierarchical clustering of differentially expressed lncRNAs. Green indicates low intensity, black indicates medium intensity, and red indicates strong intensity. (B) Boxplot of differentially expressed lncRNAs in each group. (C) Scatterplot of lncRNA signal values visualizing the variation (or reproducibility) between the two groups. The green lines represent the fold change lines. The lncRNAs above the top green line and below the bottom green line demonstrated more than a 2-fold change of lncRNA expression between the two compared samples. (D) Volcano plot of the differential expression of lncRNAs. The vertical lines correspond to 2-fold upregulation and downregulation, respectively. The horizontal line represents a p value of 0.05, and the red points on the plot represent the differentially expressed lncRNAs with statistical significance. (E) Pie chart shows the number of upregulated and downregulated lncRNAs with the length and fold change. Green indicates upregulated lncRNA, orange indicates downregulated lncRNA, blue bar indicates length of lncRNA, and brown bar indicates fold change of lncRNA. (F) Number of lncRNAs in the different subgroups classified by fold change (FC). Blue and orange bars indicate the number of upregulated and downregulated lncRNAs, respectively. (G and H) Pie chart shows the classification of the lncRNAs. According the genomic positions, upregulated (G) and downregulated (H) lncRNAs were classified as bidirectional, exon sense-overlapping, intergenic, intron sense-overlapping, intronic antisense, and natural antisense. (I) Top 10 significantly upregulated and downregulated lncRNAs in the microarray data.

Figure 3

Differential Expression Profiles of mRNAs (A–C) Boxplot (A), scatterplot (B), and volcano plot (C) of the differentially expressed mRNAs in the two groups that were compared. (D) The pie chart shows the number of upregulated and downregulated mRNAs with the length and fold change. Green indicates upregulated mRNA, orange indicates downregulated mRNA, blue bar indicates length of mRNA, and brown bar indicates fold change of mRNA. (E) Number of mRNAs in the different subgroups classified by fold change (FC). Blue and orange bars indicate the number of upregulated and downregulated mRNAs, respectively. (F) Top 10 significantly upregulated and downregulated mRNAs in the microarray data.

Figure 4

GO Analysis of Differentially Upregulated Expression of mRNAs (A, D, and G) Pie chart shows the top 10 significant enrichment terms. (B, E, and H) Bar plot shows the top 10 enrichment scores (−log₁₀ (p value)). (C, F, and I) Bar plot shows the top 10 fold enrichment values ((count/pop. hits)/(list. total/pop. total)). “Count” stands for the number of DE genes associated with the listed ID of gene ontology term; “Pop.Hits” stands for the number of background population genes associated with the listed ID of gene ontology term; “List.Total” stands for the total number of DE genes; “Pop.Total” stands for the total number of background population genes. (A–C) biological process (BP). (D–F) Cellular component (CC). (G–I) Molecular function (MF).

Figure 5

GO Analysis of Differentially Downregulated mRNAs (A, D, and G) Pie chart shows the top 10 significant enrichment terms. (B, E, and H) Bar plot shows the top 10 enrichment scores (−log₁₀ (p value)). (C, F, and I) Bar plot shows the top 10 fold enrichment values ((count/pop. hits)/(list. total/pop. total)). “Count” stands for the number of DE genes associated with the listed ID of gene ontology term; “Pop.Hits” stands for the number of background population genes associated with the listed ID of gene ontology term; “List.Total” stands for the total number of DE genes; “Pop.Total” stands for the total number of background population genes. (A–C) Biological process (BP). (D–F) Cellular component (CC). (G–I) Molecular function (MF).

Figure 6

Pathway Analysis of mRNAs with Dysregulated Expression (A) Number of pathways of mRNAs with dysregulated expression. Pathway analysis mapped the genes to KEGG pathways. A p value ≥0.05 denotes the significance correlation of the pathway to the conditions. (B and C) Pathway analysis using KEGG for the differentially expressed transcripts and schematic diagrams of the two gene categories. Pathways corresponding to the upregulated transcripts (B) and pathways corresponding to the downregulated transcripts (C) are shown. The x and y axes represent the top 10 significantly enriched pathways and their scores (−log₁₀ (p value)), respectively.

Figure 7

Validation of lncRNAs and mRNAs Using qRT-PCR (A–F) Comparison of the expression levels of lncRNAs between the microarray and qRT-PCR results. Three differentially upregulated (A–C) and downregulated (D–F) lncRNAs were validated by qRT-PCR. ∗∗∗p < 0.001. (G–L) Comparison of the expression levels of mRNAs between the microarray and qRT-PCR results. Three differentially upregulated (G–I) and downregulated (J–L) mRNAs were detected by qRT-PCR. The y axis represents the relative fold changes in expression across eight samples (SAMR1 = 4; SAMP8 = 4). ∗∗∗p < 0.001.

Figure 8

Analysis of lncRNAs and Their Nearby Coding Genes (A) Pie chart shows the number of lincRNAs (the length and fold change ranged from 60 to 43,829 and 2 to 61.5, respectively) that had nearby coding genes (<300 kb). The lincRNAs with nearby coding genes (distance between the lncRNA and coding gene <300 kb) were identified. Blue indicates upregulated lincRNAs, and orange indicates downregulated lincRNAs. (B) AK158400 had the most nearby coding genes. AK158400 had 13 nearby coding genes. (C) Top 10 lincRNAs (according to fold change) with their nearby coding genes. (D) Pie chart shows the number of antisense lncRNAs (the length and fold change ranged from 217 to 4,595 and 2 to 23, respectively) with associated coding genes. Blue indicates upregulated antisense lncRNAs, and orange indicates downregulated antisense lncRNAs. (E) Top 10 antisense lncRNAs (according to fold change) with their associated coding genes.

Figure 9

E230001N04Rik Regulates Its Nearby Coding Genes Srpk1 and Fkbp5 Levels and Tau Level in Alzheimer’s Disease (A and B) Expression patterns of lncRNAs in okadaic acid-induced (A) and Aβ42-induced (B) HT22 cell models. (C and D) Expression patterns of lncRNAs and their associated nearby coding genes in microarray data and two AD cell models by qRT-PCR validation. (C) Okadaic acid-induced HT22 cell models. (D) Aβ42-induced HT22 cell models. (E) The expression of Sirt2 in AD cell models by silence of Gm19897. (F and G) Expression of Srpk1 (F) and Fkbp5 (G) in AD cell models by silence of E230001N04Rik. (H) Expression of tau in okadaic acid-induced HT22 cell models by silence of E230001N04Rik. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001. NS, not significant.

Figure 10

Flowchart of the Experiment