Abstract
Alzheimer's disease (AD) is a heterogeneous neurodegenerative disease. Recent studies employing microRNA-seq and genome-wide sequencing have identified some non-coding RNAs that are influentially involved in AD pathogenesis. Non-coding RNAs can compete with other endogenous RNAs by microRNA response elements (MREs) and manipulate biological processes, such as tumorigenesis. However, only a few non-coding RNAs have been reported in the pathogenesis of AD. In this study, we constructed the first competing endogenous RNA (ceRNA) network leveraging whole transcriptome sequencing and a previously studied microRNA-seq of APPswe/PS1ΔE9 transgenic mice. The underlying mechanisms for the involvement of ceRNA in AD were validated using the Dual Luciferase Reporter Assay, detection of transcription levels by quantitative RT-PCR and translation levels by Western blotting, and morphological examination in primary cultured neurons. In the ceRNA network, four lncRNAs (C030034L19Rik, Rpph1, A830012C17Rik, and Gm15477) and five miRNAs (miR-182-5p, miR-330-5p, miR-326-3p, miR-132-3p, and miR-484) are enriched in nine pathways and an AD-related gene pool. Among them, Ribonuclease P RNA component H1 (Rpph1) is upregulated in the cortex of APPswe/PS1ΔE9 mice compared to wild type controls. Rpph1 binds to miR326-3p/miR-330-5p and causes the release of their downstream target Cdc42, which leads to CDC42 upregulation. This effect was disrupted upon mutation of the MRE on Rpph1. Moreover, overexpression of Rpph1 increased dendritic spine density in primary cultured hippocampal pyramidal neurons, whereas knocking down of Rpph1 had the reverse effect. In conclusion, Rpph1 modulates CDC42 expression level in a ceRNA-dependent manner, which may represent a compensatory mechanism in the early stage of the AD pathogenesis.