Comprehensive analysis of mitochondrial dysfunction and necroptosis in intracranial aneurysms from the perspective of predictive, preventative, and personalized medicine

Abstract

Mitochondrial dysfunction and necroptosis are closely associated, and play vital roles in the medical strategy of multiple cardiovascular diseases. However, their implications in intracranial aneurysms (IAs) remain unclear. In this study, we aimed to explore whether mitochondrial dysfunction and necroptosis could be identified as valuable starting points for predictive, preventive, and personalized medicine for IAs. The transcriptional profiles of 75 IAs and 37 control samples were collected from the Gene Expression Omnibus (GEO) database. Differentially expressed genes (DEGs), weighted gene co-expression network analysis, and least absolute shrinkage and selection operator (LASSO) regression were used to screen key genes. The ssGSEA algorithm was performed to establish phenotype scores. The correlation between mitochondrial dysfunction and necroptosis was evaluated using functional enrichment crossover, phenotype score correlation, immune infiltration, and interaction network construction. The IA diagnostic values of key genes were identified using machine learning. Finally, we performed the single-cell sequencing (scRNA-seq) analysis to explore mitochondrial dysfunction and necroptosis at the cellular level. In total, 42 IA-mitochondrial DEGs and 15 IA-necroptosis DEGs were identified. Screening revealed seven key genes invovled in mitochondrial dysfunction (KMO, HADH, BAX, AADAT, SDSL, PYCR1, and MAOA) and five genes involved in necroptosis (IL1B, CAMK2G, STAT1, NLRP3, and BAX). Machine learning confirmed the high diagnostic value of these key genes for IA. The IA samples showed higher expression of mitochondrial dysfunction and necroptosis. Mitochondrial dysfunction and necroptosis exhibited a close association. Furthermore, scRNA-seq indicated that mitochondrial dysfunction and necroptosis were preferentially up-regulated in monocytes/macrophages and vascular smooth muscle cells (VSMCs) within IA lesions. In conclusion, mitochondria-induced necroptosis was involved in IA formation, and was mainly up-regulated in monocytes/macrophages and VSMCs within IA lesions. Mitochondria-induced necroptosis may be a novel potential target for diagnosis, prevention, and treatment of IA.

Keywords: 3P medicine; Bioinformatics; Intracranial aneurysm; Mitochondrial dysfunction; Necroptosis.

Fig. 1

The flow chart of this study

Fig. 2

Identification of DEGs and IA-related genes in IA. A The PCA results of GSE75436. B Volcano plot showing DEGs in the IA and the normal samples. C Heat map of DEGs. D Sample clustering diagram of WGCNA. E Soft-thresholding filtering. F Clustering dendrogram of genes. G Correlation heatmap of gene modules and clinical features

Fig. 3

The function crossover of gene between IA Mito DEGs and IA-Necroptosis DEGs. A Venn diagram showing the overlap of DEGs, IA-related genes, and mitochondria-related genes. B Venn diagram showing the overlap of DEGs, IA-related genes, and Necroptosis-related genes. C Expression of 44 IA-Mito DEGs and 15 IA-Necroptosis DEGs in GSE75436. D and E GO enrichment results of 42 IA-Mito DEGs (D) and 15 IA-Necroptosis DEGs (E). F and G KEGG enrichment results of 42 IA-Mito DEGs (F) and 15 IA-Necroptosis DEGs (G)

Fig. 4

The function crossover of phenotype scores between mitochondrial dysfunction and necroptosis. A LASSO regression screened out key genes of mitochondrial dysfunction from the 42 IA-Mito DEGs. B LASSO regression screened out key genes of necroptosis from the 15 IA-Mito DEGs. C Boxplot of phenotype scores of mitochondrial dysfunction and necroptosis between IA and normal samples. D The correlation matrix of phenotype scores of mitochondrial dysfunction and necroptosis. E GO enrichment analysis is based on the GSEA algorithm in key genes. F KEGG enrichment analysis is based on the GSEA algorithm in key genes

Fig. 5

The ssGSEA algorithm for analyzing immunocyte infiltration A Heatmap of 28 immune cell types. B The boxplot plot of 28 immune gene-sets content. C The correlation matrix of immune cells. D Correlation diagram for phenotype scores and immunocyte expression

Fig. 6

Construction of interaction network and validation of key genes. A The construction of PPI network among IA-Mito DEGs, IA-Necroptosis DEGs, and key genes. B Prediction of TFs and microRNAs for key genes of mitochondrial dysfunction and necroptosis. C Expression of 11 key genes in the validating datasets GSE15629 and GSE122897. D Evaluation of IA diagnosis value of key genes by the RF and SMO algorithms

Fig. 7

The expression of mitochondrial dysfunction and necroptosis at the cellular level. A Integration of multiple sample data using the R package harmony. B UMAP plot is colored by different cell types. C UMAP plot is colored by different cell types and divided by IA and sham groups. D Bar chart showing the percentage of different types of cells between IA and sham groups. E UMAP plot is colored by the expression of mitochondrial dysfunction. F Bar chart showing the expression disparity of mitochondrial dysfunction between IA and sham groups. G UMAP plot is colored by the expression of necroptosis. H Bar chart showing the expression disparity of necroptosis between IA and sham groups. I GO enrichment results of IA-Mito DEGs in the previously obtained PPI network. J Mechanism diagram of mitochondria-induced necroptosis in IAs