AUTHOR=Wang Qianfei , Liu Dong , Gao Tianci , Tao Yulei , Li Xin , Liu Yuan , Liu Zhiliang , Mei Jianqiang , Chen Fenqiao TITLE=Experimental verification and identifying biomarkers related to insomnia JOURNAL=Frontiers in Neurology VOLUME=14 YEAR=2023 URL=https://www.frontiersin.org/journals/neurology/articles/10.3389/fneur.2023.1189076 DOI=10.3389/fneur.2023.1189076 ISSN=1664-2295 ABSTRACT=Introduction

Insomnia is the most common form of sleep deprivation (SD) observed in clinics. Although there are differences between insomnia and SD, they have similar symptoms and the same animal model. Currently, there is a lack of microarray data on insomnia. Therefore, for now, we are going to apply the SD data to insomnia. Although many studies have explained the possible mechanisms associated with insomnia, no previous studies have considered the key genes associated with insomnia or the relationship between insomnia and immune cells. In this study, we analyzed the relationship between key genes and immune cells by identifying biomarkers for the diagnosis of insomnia. Next, we verified the efficacy of these biomarkers experimentally.

Methods

First, we downloaded four microarrays (GSE11755, GSE12624, GSE28750, and GSE48080) from the Gene Expression Omnibus (GEO) database, which included data from 239 normal human blood samples and 365 blood specimens from patients with SD. Then, we analyzed two groups of differentially expressed genes (DEGs) and used Support Vector Machine Recursive Feature Elimination (SVM-RFE) analysis and the Least Absolute Shrinkage and Selection Operator (LASSO) regression model to investigate these key genes. Next, we used CIBERSORT to investigate the composition of 22 immune cell components of key genes in SD patients. Finally, the expression levels of key biomarkers in sleep-deprived patients were examined by quantitative real-time polymerase chain reaction (qRT-PCR).

Results

A total of 50 DEGs were identified: six genes were significantly upregulated, and 44 genes were significantly downregulated. Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analysis showed that Salmonella infection, NOD-like receptor (NLR) signaling pathway, Kaposi sarcoma-associated herpesvirus infection, and Th17 cell differentiation were significant. Based on machine learning, we identified C2CD2L, SPINT2, APOL3, PKNOX1, and A2M as key genes for SD; these were confirmed by receiver operating characteristic (ROC) analysis. Immune cell infiltration analysis showed that C2CD2L, SPINT2, APOL3, PKNOX1, and A2M were related in different degrees to regulatory T cells (Tregs), follicular T helper cells, CD8 cells, and other immune cells. The qRT-PCR experiments confirmed that the expression levels of C2CD2L concurred with the results derived from machine learning, but PKNOX1 and APOL3 did not.

Discussion

In summary, we identified a key gene (C2CD2L) that may facilitate the development of biomarkers for insomnia.