Unraveling Potential Biomarkers of Alcoholic Dementia through Integrated Bioinformatics Analysis and Experimental Validation

摘要

Objective: Alcohol is one of the most frequently abused substances worldwide. Chronic excessive drinking or alcoholism can lead to alcohol dependence and alcohol poisoning, causing permanent brain damage and cognitive impairment. In China, about 40 million people suffer from alcohol dependence, with 1.1 million annual alcohol-related deaths, posing a substantial societal burden. The pathogenesis of alcoholic dementia remains poorly understood, and effective treatments are lacking. This study employed bioinformatics and molecular biology approaches to investigate the pathogenesis of alcoholic dementia, identify therapeutic targets, and propose novel treatment strategies. Methods: Gene expression datasets from human brain tissues were obtained from the Gene Expression Omnibus(GEO) database using keywords including “Alcohol”, “Alcoholism”, and “Dementia”. Differential expression analysis was performed using the “limma” package in R 4.1.1 to identify differentially expressed genes(DEGs). Protein-protein interaction(PPI) networks and gene co-expression networks were constructed using Cytoscape 3.6.1 to identify hub genes. Functional enrichment analyses, including Gene Ontology(GO) and KEGG pathway analyses, were conducted through the MetaScape database to elucidate the biological roles and signaling pathways of the DEGs. Machine learning algorithms were applied to pinpoint target genes from the hub genes, and single-sample gene set enrichment analysis(ssGSEA) was used to explore correlations between target genes and immune cell infiltration. A mouse model of alcoholic dementia was established using 3xTg-AD mice and wild-type controls exposed to a two-bottle choice drinking paradigm. Behavioral experiments were conducted to assess cognitive function. RT-qPCR and Western blot experiments were performed to detect the transcriptional and protein expression levels of target genes in the model mice. Furthermore, structure-based virtual screening was performed via the DSigDB database for molecular docking and drug discovery. Finally, an in vitro model of alcoholinduced neuroinflammation was established using SH-SY5Y cells stimulated with ethanol, and the neuroprotective effects of candidate compounds compounds were validated using the CCK-8 assay. Results: The GSE5281 and GSE62699 datasets were retrieved from the GEO database. GSE5281 contained 161 human brain samples from cognitive impaired donors, while GSE62699 contained 72 human brain samples from alcohol-dependent individuals, 36 of which were controls. DEG analysis identified 480 up-regulated genes and 762 down-regulated genes in the GSE5281 dataset, and 118 up-regulated genes and 101 down-regulated genes in the GSE62699 dataset. After deduplication, 68 overlapping genes were identified. Functional enrichment analysis indicated that DEGs were enriched in 1,232 GO categories in GO enrichment analysis and 569 signaling pathways in KEGG pathways, most notably neurodegenerative pathways and cAMP signaling pathways. 11 key genes(STAT3, ESR1, IL6, JUN, TP53, TNF, AKT1, EGFR, HSP90 AA1, SRC, TP5). Machine learning algorithms further refined STAT3 and JUN as core target genes, which also showed significant correlations with immune cell infiltration. In an alcoholic dementia mouse model, both mRNA and protein levels of STAT3 and JUN were significantly up-regulated compared to wild-type controls. Molecular docking screening identified two compounds, acitretin and pimaric acid, that target STAT3 and JUN. CCK-8 assays demonstrated that both compounds alleviated alcohol-induced neuroinflammation in SHSY5Y cells, with pimaric acid exhibiting superior neuroprotective activity compared to acitretin. Both in vitro and in vivo experiments confirmed the therapeutic potential of the target genes.Conclusion: This study provides a theoretical foundation for identifying potential target genes in alcoholic dementia through an integrative approach combining bioinformatics methods, machine learning, computer-aided drug design, and experimental validation. STAT3 and JUN may serve as promising therapeutic targets for alcoholic dementia, while pimaric acid represents a candidate compound for further drug development and structural optimization.

关键词

alcoholic dementia / cAMP signaling pathways / neurodegenerative diseases / GEO / machine learning