Identification of circulating immune landscape in ischemic stroke based on bioinformatics methods

Danyang Li; Lifang Li; Fei Quan; Tianfeng Wang; Si Xu; Shuang Li; Kuo Tian; Meng Feng; Ni He; Liting Tian; Biying Chen; Huixue Zhang; Lihua Wang; Jianjian Wang

doi:10.3389/fgene.2022.921582

Identification of circulating immune landscape in ischemic stroke based on bioinformatics methods

Front Genet. 2022 Jul 25:13:921582. doi: 10.3389/fgene.2022.921582. eCollection 2022.

Authors

Danyang Li¹, Lifang Li¹, Fei Quan², Tianfeng Wang¹, Si Xu¹, Shuang Li¹, Kuo Tian¹, Meng Feng¹, Ni He¹, Liting Tian¹, Biying Chen¹, Huixue Zhang¹, Lihua Wang¹, Jianjian Wang¹

Affiliations

¹ Department of Neurology, The Second Affiliated Hospital of Harbin Medical University, Harbin, China.
² College of Bioinformatics Science and Technology, Harbin Medical University, Harbin, China.

Abstract

Ischemic stroke (IS) is a high-incidence disease that seriously threatens human life and health. Neuroinflammation and immune responses are key players in the pathophysiological processes of IS. However, the underlying immune mechanisms are not fully understood. In this study, we attempted to identify several immune biomarkers associated with IS. We first retrospectively collected validated human IS immune-related genes (IS-IRGs) as seed genes. Afterward, potential IS-IRGs were discovered by applying random walk with restart on the PPI network and the permutation test as a screening strategy. Doing so, the validated and potential sets of IS-IRGs were merged together as an IS-IRG catalog. Two microarray profiles were subsequently used to explore the expression patterns of the IS-IRG catalog, and only IS-IRGs that were differentially expressed between IS patients and controls in both profiles were retained for biomarker selection by the Random Forest rankings. CLEC4D and CD163 were finally identified as immune biomarkers of IS, and a classification model was constructed and verified based on the weights of two biomarkers obtained from the Neural Network algorithm. Furthermore, the CIBERSORT algorithm helped us determine the proportions of circulating immune cells. Correlation analyses between IS immune biomarkers and immune cell proportions demonstrated that CLEC4D was strongly correlated with the proportion of neutrophils (r = 0.72). These results may provide potential targets for further studies on immuno-neuroprotection therapies against reperfusion injury.

Keywords: biomarker; immune; ischemic stroke; machine learning; neuroinflammation.