Novel insights into the immune cell landscape and gene signatures in autism spectrum disorder by bioinformatics and clinical analysis

Abstract

The pathogenesis of autism spectrum disorder (ASD) is not well understood, especially in terms of immunity and inflammation, and there are currently no early diagnostic or treatment methods. In this study, we obtained six existing Gene Expression Omnibus transcriptome datasets from the blood of ASD patients. We performed functional enrichment analysis, PPI analysis, CIBERSORT algorithm, and Spearman correlation analysis, with a focus on expression profiling in hub genes and immune cells. We validated that monocytes and nonclassical monocytes were upregulated in the ASD group using peripheral blood (30 children with ASD and 30 age and sex-matched typically developing children) using flow cytometry. The receiver operating characteristic curves (PSMC4 and ALAS2) and analysis stratified by ASD severity (LIlRB1 and CD69) showed that they had predictive value using the “training” and verification groups. Three immune cell types – monocytes, M2 macrophages, and activated dendritic cells – had different degrees of correlation with 15 identified hub genes. In addition, we analyzed the miRNA-mRNA network and agents-gene interactions using miRNA databases (starBase and miRDB) and the DSigDB database. Two miRNAs (miR-342-3p and miR-1321) and 23 agents were linked with ASD. These findings suggest that dysregulation of the immune system may contribute to ASD development, especially dysregulation of monocytes and monocyte-derived cells. ASD-related hub genes may serve as potential predictors for ASD, and the potential ASD-related miRNAs and agents identified here may open up new strategies for the prevention and treatment of ASD.