Identification of HIBCH and MGME1 as Mitochondrial Dynamics-Related Biomarkers in Alzheimer's Disease Via Integrated Bioinformatics Analysis

Abstract

Mitochondrial dynamics (MD) play a crucial role in the genesis of Alzheimer's disease (AD); however, the molecular mechanisms underlying MD dysregulation in AD remain unclear. This study aimed to identify critical molecules of MD that contribute to AD progression using GEO data and bioinformatics approaches. The GSE63061 dataset comparing AD patients with healthy controls was analysed, WGCNA was employed to identify co-expression modules and differentially expressed genes (DEGs) and LASSO model was developed and verified using the DEGs to screen for potential biomarkers. A PPI network was built to predict upstream miRNAs, which were experimentally validated using luciferase reporter assays. A total of 3518 DEGs were identified (2209 upregulated, 1309 downregulated; |log₂FC| > 1.5, adjusted p < 0.05). WGCNA revealed 160 MD-related genes. LASSO regression selected HIBCH and MGME1 as novel biomarkers with significant downregulation in AD (fold change > 2, p < 0.001). KEGG enrichment analysis highlighted pathways associated with neurodegeneration. Luciferase assays confirmed direct binding of miR-922 to the 3'UTR of MGME1. HIBCH and MGME1 are promising diagnostic biomarkers for AD with AUC values of 0.73 and 0.74. Mechanistically, miR-922 was experimentally validated to directly bind MGME1 3'UTR.

Keywords: bioinformatics; biomechanics; brain; data analysis; network analysis.

FIGURE 1

The volcano plot of significantly different mRNA genes and significant difference mRNA heatmap. (A) The volcano plot of significantly different mRNA genes, including 2209 upregulated genes (B) 1309 downregulated genes, as shown in Figure 1B (|logFC| > 0.585, p adj < 0.05). (C) All differently expressed genes were also illustrated as heatmap.

FIGURE 2

WGCNA analysis and VENN analysis of WGCNA module genes and mitochondrial dynamics genes. (A) A total of 3518 mRNA transcripts across 158 samples were included in subsequent analyses. Weighted gene co‐expression network analysis (WGCNA) was performed with a soft‐thresholding power of 5. (B) This identified four co‐expression modules containing 240, 2,175, 721, and 382 genes, respectively. (C) Modules MEblue and Menturquoise showed significant association with AD pathology, exhibiting 2896 differentially co‐expressed genes. (D) Venn diagram overlap analysis identifies 160 mitochondria dynamics‐associated hub genes in AD, including key players such as HIBCH and MGME1.

FIGURE 3

GO/KEGG analysis and univariate logistic regression analysis of differentially expressed genes. (A) The KEGG and GO analysis results indicated 385 significantly enriched pathways (p < 0.05, Count > 2). (B) The first 20 genes of 160 genes conducted single factor logistic regression were illustrated as the forest map.

FIGURE 4

Diagnostic performance of HIBCH/MGME1 in AD. (A, B) Lasso regression identified 27 candidate genes. (C, D) ROC curves validated HIBCH/MGME1 as robust biomarkers (AUC = 0.73–0.74). (E, F) HIBCH and MGME1 expression is significantly downregulated in AD tissues.

FIGURE 5

Integration of protein‐protein interactions (PPI) and miRNA regulatory networks. (A) PPI network analysis (GeneMANIA) identified key interactors of diagnostic genes, highlighting functional associations (co‐localization, shared domains, and pathways). (B) miRNA‐target prediction identified hsa‐miR‐922 and hsa‐miR‐98 as upstream regulators of MGME1 through cross‐database validation (mirWalk).

FIGURE 6

Small‐molecule drug prediction for candidate genes. (A) Signature Search identified 100 drug chemical small molecules related to expression regulation of HIBCH (Z Score > 1, p < 0.1). (B) Edaravone, aphidicolin, and ranitidine as top regulators of HIBCH (Z Score > 1, p < 0.1).

FIGURE 7

Luciferase reporter assay validation of miRNA‐MGME1 binding. (A, C) miR‐922 directly binds to the MGME1 3′UTR: Wild‐type (WT) versus mutant (MUT) plasmids: Luciferase activity decreased by 42% (p = 0.003). Negative control (NC) group validated system integrity. (B, D) miR‐98‐5p shows no specific binding: Activity reduction < 20% versus NC (p > 0.05), confirming non‐specific effects. miR‐922, but not miR‐98‐5p, post‐transcriptionally suppresses MGME1 expression via direct 3′UTR interaction.