Table_1_Elucidating the Molecular Mechanism of Ischemic Stroke Using Integrated Analysis of miRNA, mRNA, and lncRNA Expression Profiles.DOCX

Objective: This study aimed to investigate the possible molecular mechanisms associated with ischemic stroke through the construction of a lncRNA-miRNA-mRNA network. miRNA expression profile in GSE55937, mRNA and lncRNA expression profiles in GSE122709, and mRNA expression profile in GSE146882 were downloaded from the NCBI GEO database. After the identification of the differentially expressed miRNA, lncRNA, and mRNA using GSE55937 and GSE122709 in ischemic stroke vs. control groups, a protein-protein interaction (PPI) network was constructed. The lncRNA-miRNA, lncRNA-mRNA, and miRNA-mRNA pairs were predicted, and a lncRNA-miRNA-mRNA network was constructed. Additionally, the gene-drug interactions were predicted. Characteristic genes were used to construct a support vector machine (SVM) model and verified using quantitative reverse transcription polymerase chain reaction. In total 38 miRNAs, 115 lncRNAs, and 990 mRNAs were identified between ischemic stroke and control groups. A PPI network with 371 nodes and 2306 interaction relationships was constructed. The constructed lncRNA-miRNA-mRNA network contained 7 mRNAs, 14 lncRNAs, such as SND1-IT1, NAPA-AS1, LINC01001, LUCAT1, and ASAP1-IT2, and 8 miRNAs, such as miR-93-3p and miR-24-3p. The drug action analysis of the seven differential mRNAs included in the lncRNA-miRNA-mRNA network showed that four genes (GPR17, ADORA1, OPRM1 and LPAR3) were predicted as molecular targets of drugs. The area under the curve of the constructed SVM model was 0.886. The verification results of the relative expression of RNA by qRT-PCR were consistent with the results of bioinformatics analysis. LPAR3, ADORA1, GPR17, and OPRM1 may serve as therapeutic targets of ischemic stroke. lncRNA-miRNA-mRNA regulatory axis such as SND1-IT1/NAPA-AS1/LINC01001-miR-24-3p-LPAR3/ADORA1 and LUCAT1/ASAP1-IT2-miR-93-3p-GPR17 may play important roles in the progression of ischemic stroke.

研究目的：本研究旨在通过构建长链非编码RNA（long non-coding RNA, lncRNA）-微小RNA（microRNA, miRNA）-信使RNA（messenger RNA, mRNA）调控网络，探究与缺血性脑卒中（ischemic stroke）相关的潜在分子机制。从美国国家生物技术信息中心（National Center for Biotechnology Information, NCBI）基因表达综合数据库（Gene Expression Omnibus, GEO）下载了GSE55937数据集的miRNA表达谱、GSE122709数据集的mRNA与lncRNA表达谱，以及GSE146882数据集的mRNA表达谱。利用GSE55937与GSE122709数据集，对缺血性脑卒中组与对照组间的差异表达miRNA、lncRNA及mRNA进行筛选后，构建了蛋白质-蛋白质相互作用（protein-protein interaction, PPI）网络。预测得到lncRNA-miRNA、lncRNA-mRNA及miRNA-mRNA调控对，并构建lncRNA-miRNA-mRNA调控网络。此外，还对基因-药物相互作用进行了预测。以特征基因为基础构建支持向量机（support vector machine, SVM）模型，并通过定量逆转录聚合酶链反应（quantitative reverse transcription polymerase chain reaction, qRT-PCR）对模型进行验证。最终在缺血性脑卒中组与对照组间筛选得到38个差异表达miRNA、115个差异表达lncRNA及990个差异表达mRNA。构建的PPI网络共包含371个节点与2306条相互作用关系。所构建的lncRNA-miRNA-mRNA调控网络共包含7个mRNA、14个lncRNA（如SND1-IT1、NAPA-AS1、LINC01001、LUCAT1及ASAP1-IT2）以及8个miRNA（如miR-93-3p与miR-24-3p）。对该调控网络中包含的7个差异mRNA进行药物作用分析后发现，其中4个基因（GPR17、ADORA1、OPRM1及LPAR3）被预测为药物分子靶点。所构建的SVM模型的曲线下面积为0.886。qRT-PCR检测得到的RNA相对表达量验证结果与生物信息学分析结果一致。LPAR3、ADORA1、GPR17及OPRM1有望成为缺血性脑卒中的治疗靶点。诸如SND1-IT1/NAPA-AS1/LINC01001-miR-24-3p-LPAR3/ADORA1及LUCAT1/ASAP1-IT2-miR-93-3p-GPR17这类lncRNA-miRNA-mRNA调控轴，可能在缺血性脑卒中的疾病进展中发挥重要作用。