Key points are not available for this paper at this time.
Background Sleep deprivation (SD) is harmful to individuals, but its pathogenesis is not clarified. Objective This study seeks to identify SD-linked autophagy genes via integrated transcriptomic and experimental validation approaches. Methods Primary SD transcriptomic datasets (GSE33302 and GSE9442), derived from murine brain tissue, were retrieved from the Gene Expression Omnibus (GEO) database to identify differentially expressed genes (DEGs). Murine gene symbols were subsequently mapped to their human orthologs to enable downstream bioinformatic analyses and integration with the GeneCards database. The converted DEGs were intersected with autophagy-related genes (ARGs) obtained from GeneCards to identify autophagy-associated DEGs, which were then subjected to functional enrichment analyses. Candidate predictor genes were selected using machine learning (ML) algorithms. Their expression was rigorously validated in both internal and external datasets (GSE9441 and GSE3767), encompassing murine brain tissue and human peripheral blood samples, respectively. In parallel, an SD rat model was established by exposing male Sprague-Dawley rats to continuous sleep deprivation for seven consecutive days. Brain tissues from the prefrontal cortex and hippocampus were harvested, and the expression levels of rat orthologs of the candidate genes were quantified using reverse transcription-quantitative polymerase chain reaction (RT-qPCR). The diagnostic performance of the identified genes was evaluated through nomogram construction and receiver operating characteristic (ROC) curve analysis. In addition, the immune landscape associated with SD was inferred using single-sample gene set enrichment analysis (ssGSEA). The cellular distribution and functional roles of the candidate genes were further explored via single-cell RNA sequencing (scRNA-seq; GSE37665) and gene set enrichment analysis (GSEA). Finally, potential therapeutic targets associated with these genes were predicted. Results Three significantly dysregulated predictor genes: CDKN1A , HSPA5 , and NR4A1 , were identified, and a diagnostic model incorporating these genes demonstrated strong predictive performance. Bioinformatic analysis of immune cell infiltration indicated a potential association between the three key predictor genes and modifications in the immune microenvironment. Moreover, single-cell transcriptomic profiling revealed that these genes were preferentially and highly expressed in endothelial cells, glial cells, and neurons, respectively, implying distinct functional roles across different cellular subpopulations. Conclusion CDKN1A, HSPA5, and NR4A1 emerge as crucial pathogenic biomarkers and potential therapeutic targets for SD. This study provides novel molecular targets for elucidating the mechanisms underlying SD-induced autophagy modulation, immune response, and neurovascular injury.
Gan et al. (Wed,) studied this question.