Identifying the diagnostic value of inflammatory genes in intracranial aneurysms based on machine learning

Abstract

AbstractBackground Intracranial aneurysm (IA) is a cerebrovascular disease that can be caused by a variety of factors. Clinical trials have indicated that inflammation and inflammatory cells play critical roles in the pathophysiology of IA. Nonetheless, the roles of inflammation-related genes (IRGs) in IA remain unclear. Methods The GSE75436 and GSE54083 datasets were acquired from the Gene Expression Omnibus (GEO) database, and the IRGs were extracted from the MSigDB database. First, the two GEO datasets were combined, and the batch effects were removed. The differentially expressed inflammation-related genes (DEIRGs) were identified by overlapping the IRGs with the set of differentially expressed genes (DEGs) between IA and control samples. The functions of the DEIRGs were evaluated by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses. Then, ROC curve analysis was used to verify the diagnostic ability of the DEIRGs, and diagnostic models were constructed with 7 machine learning methods. Furthermore, gene set enrichment analysis (GSEA) was performed to explore the potential biological functions of the biomarkers. Immune cell relevance was assessed by single-sample gene set enrichment analysis (ssGSEA). In addition, a TF–mRNA‒miRNA network was established, and potential biomarkers were predicted. Ultimately, the mRNA expression levels of the biomarkers were validated by quantitative real-time PCR (qRT‒PCR). Results In total, 35 DEIRGs were retrieved by overlapping the 964 DEGs and 200 IRGs. Functional enrichment analysis revealed that the DEIRGs were significantly enriched in the regulation of the inflammatory response, immune receptor activity, and lipid and atherosclerosis pathways. Moreover, 13 genes with an AUC greater than 0.85 were selected for diagnostic model construction by the RF algorithm, and 7 biomarkers were obtained in the final model. GSEA indicated that these 7 biomarkers were mainly associated with inflammation. The significantly differentially abundant immune cells exhibited positive correlations with the biomarkers. Subsequently, we proposed that SERPINE might be modulated by TBX3, MLX, and NR1I3 and that SLC11A2 might be modulated by hsa-miR-6838-5p, hsa-miR-4784, and hsa-miR-3663-5p. In addition, 22 drugs were predicted to interact with the biomarkers, including fluoxetine, aleplasinin, and orlistat. Finally, qRT‒PCR results showed that the expression levels of the 7 biomarkers were significantly higher in IA tissue than in superficial temporal artery tissue. Conclusion This research provides a new perspective for understanding the molecular mechanism of IA pathogenesis and valuable evidence for the pathological diagnosis of IRGs.