Identification of Dysregulated Mechanisms and Potential Biomarkers in Ischemic Stroke Onset

Bing Feng,1,* Xinling Meng,2,* Hui Zhou,1,* Liechun Chen,3 Chun Zou,3 Lucong Liang,3 Youshi Meng,3,4 Ning Xu,3,4 Hao Wang,3,4 Donghua Zou3,4 1Department of Neurology, The People’s Hospital of Guiping, Guigang, Guangxi, 537200, People’s Republic of China; 2Department of Endocrinology, The People’s Hospital of Guiping, Guigang, Guangxi, 537200, People’s Republic of China; 3Department of Neurology, The Fifth Affiliated Hospital of Guangxi Medical University, Nanning, Guangxi, 530022, People’s Republic of China; 4Department of Neurology, The First People’s Hospital of Nanning, Nanning, Guangxi, 530022, People’s Republic of China*These authors contributed equally to this workCorrespondence: Hao Wang; Donghua ZouThe Fifth Affiliated Hospital of Guangxi Medical University, 89 Qixing Road, Nanning, Guangxi, 530022, People’s Republic of ChinaTel +86 7712636189Email bswanghao2010@163.com; danvor0922@hotmail.comObjective: Ischemic stroke (IS) is a major cause of severe disability. This study aimed to identify potential biomarkers closely related to IS diagnosis and treatment.Methods: Profiles of gene expression were obtained from datasets GSE16561, GSE22255, GSE112801 and GSE110993. Differentially expressed mRNAs between IS and controls were then subjected to weighted gene co-expression network analysis as well as multiscale embedded gene co-expression network analysis. The intersection of the two sets of module genes was subjected to analyses of functional enrichment and of microRNAs (miRNAs) regulation. Then, the area under receiver operating characteristic curves (AUC) was calculated to assess the ability of genes to discriminate IS patients from controls. IS diagnostic signatures were constructed using least absolute shrinkage and selection operator regression.Results: A total of 234 common co-expression network genes were found to be potentially associated with IS. Enrichment analysis found that these genes were mainly associated with inflammation and immune response. The aberrantly expressed miRNAs (hsa-miR-651-5p, hsa-miR-138-5p, hsa-miR-9-3p and hsa-miR-374a-3p) in IS had regulatory effects on IS-related genes and were involved in brain-related diseases. We used the criterion AUC > 0.7 to screen out 23 hub genes from IS-related genes in the GSE16561 and GSE22255 datasets. We obtained an 8-gene signature (ADCY4, DUSP1, ATP5F1, DCTN5, EIF3G, ELAVL1, EXOSC7 and PPIE) from the training set of GSE16561 dataset, which we confirmed in the validation set of GSE16561 dataset and in the GSE22255 dataset. The genes in this signature were highly accurate for diagnosing IS. In addition, the 8-gene signature significantly correlated with infiltration by immune cells.Conclusion: These findings provide new clues to molecular mechanisms and treatment targets in IS. The genes in the signature may be candidate markers and potential gene targets for treatments.Keywords: ischemic stroke, bioinformatics analysis, miRNAs, immune inflammatory response, candidate markers