The mechanism of mitochondrial metabolic gene PMAIP1 involved in Alzheimer's disease process based on bioinformatics analysis and experimental validation

Abstract

Objectives: This study explored novel biomarkers that can affect the diagnosis and treatment in Alzheimer's Disease (AD) related to mitochondrial metabolism.

Methods: The authors obtained the brain tissue datasets for AD from the Gene Expression Omnibus (GEO) and downloaded the mitochondrial metabolism-related genes set from MitoCarta 3.0 for analysis. Differentially Expressed Genes (DEGs) were screened using the "limma" R package, and the biological functions and pathways were investigated by Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses. The LASSO algorithm was used to identify the candidate center genes and validated in the GSE97760 dataset. PMAIP1 with the highest diagnostic value was selected and its effect on the occurrence of AD by biological experiments.

Results: A sum of 364 DEGs and 50 hub genes were ascertained. GO and KEGG enrichment analysis demonstrated that DEGs were preponderantly associated with cell metabolism and apoptosis. Five genes most associated with AD as candidate central genes by LASSO algorithm analysis. Then, the expression level and specificity of candidate central genes were verified by GSE97760 dataset, which confirmed that PMAIP1 had a high diagnostic value. Finally, the regulatory effects of PMAIP1 on apoptosis and mitochondrial function were detected by siRNA, flow cytometry and Western blot. siRNA-PMAIP1 can alleviate mitochondrial dysfunction and inhibit cell apoptosis.

Conclusion: This study identified biomarkers related to mitochondrial metabolism in AD and provided a theoretical basis for the diagnosis of AD. PMAIP1 was a potential candidate gene that may affect mitochondrial function in Hippocampal neuronal cells, and its mechanism deserves further study.

Keywords: Alzheimer's disease; Apoptosis; Biomarkers; Mitochondrial Metabolism.

Fig. 1

Data preprocessing and identification of DEGs. (A) Boxplot of transcriptome data of GSE5281. (B) Volcano plot of DEGs in GSE5281. The cut-off criteria were |log2Fc|>1 and p < 0.05. The red dots represent the up-regulated genes, and the blue dots denote the down-regulated genes. The grey dots indicate the genes with |log2Fc|<1 and/or p > 0.05. (C) Boxplot of transcriptome data of GSE28146. (D) Volcano plot of DEGs in GSE5281. (E) Venn diagram showing the numbers of overlapped DEGs between GSE5281 and GSE28146.

Fig. 2

Functional enrichment analysis and PPI network. (A) GO functional analysis showing enrichment of DEGs. (B) KEGG pathway enrichment analysis of DEGs. (C) PPI network of DEGs were analyzed using Cytoscape software. (D) PPI network for the top 50 genes.

Fig. 3

Identification of candidate central genes. (A) Venn diagram showing the numbers of overlapped genes between DEG s and MitoCarta. (B) Expression trends of overlapping genes in GSE5281 and GSE28146 datasets. (C) LASSO coefficient profiles of candidate genes. (D) Cross-validation to select the optimal tuning parameter log (λ) in LASSO regression analysis.

Fig. 4

The relationship between 5 genes expression and AD using the method of binomial logistic regression for generalized linear models. (A) ADCK2 (B) COX6B2 (C) PMAIP1 (D) PPA2 (E) YME1L1.

Fig. 5

The AUCs of 5 candidate central genes. (A) ADCK2 (B) COX6B2 (C) PMAIP1 (D) PPA2 (E) YME1L1.

Fig. 6

External dataset validation. (A) The expression patterns of 5 genes from the GSE97760 dataset. * p < 0.05, ** p < 0.01, *** p < 0.001. (B) The AUC of PMAIP1 in the GSE97760 dataset. (C) The AUC of PPA2 in the GSE97760 dataset.

Fig. 7

PMAIP1 expression in Aβ-induced HT-22 cells. (A) The viability of HT-22 cells treated with A-β1-42 (5, 10 and 20 µM) was measured by CCK-8 assay. (B) Protein levels of PMAIP1 measured by western blot assay. ** p < 0.01, *** p < 0.001 versus control.

Fig. 8

Effect of siRNA-PMAIP1 on Aβ-induced apoptosis in HT-22 cells. (A) Protein levels of PMAIP1 after small RNA interfering measured by western blot assay. (B) Cell viability was detected by CCK8 assay. (C) Apoptotic status of HT-22 cells was assayed by flow cytometry. (D) The expression levels of BCL2, Bax and cleaved caspase3 were measured by western blot. *** p < 0.001 versus control. ^# p < 0.5, ^## p < 0.01, ^### p < 0.001 versus siRNA-NC.

Fig. 9

Effect of siRNA-PMAIP1 on Aβ-induced mitochondrial function in HT-22 cells. (A) ROS level was detected by DCFH-DA staining. Original magnification: × 200. (B) MMP was identified by JC-1 staining. JC-1 exists in both aggregates and monomers states. Green fluorescence indicates that JC-1 exists as a monomer at low concentrations, and red fluorescence indicates that JC-1 exists as an aggregate at high concentrations. Original magnification: × 200. (C) The activity of MDA and SOD. *** p < 0.001 versus control. ### p < 0.001 versus siRNA-NC.