Exploring autophagy-related prognostic genes of Alzheimer's disease based on pathway crosstalk analysis

Abstract

Recent studies have shown that different signaling pathways are involved in the pathogenesis of Alzheimer's disease (AD), with complex molecular connections existing between these pathways. Autophagy is crucial for the degradation and production of pathogenic proteins in AD, and it shows link with other AD-related pathways. However, current methods for identifying potential therapeutic targets for AD are primarily based on single-gene analysis or a single signal pathway, both of which are somewhat limited. Finding other methods is necessary for providing novel underlying AD therapeutic targets. Therefore, given the central role of autophagy in AD and its interplay with its pathways, we aimed to identify prognostic genes related to autophagy within and between these pathways based on pathway crosstalk analysis. The method of pathway analysis based on global influence (PAGI) was applied to find the feature mRNAs involved in the crosstalk between autophagy and other AD-related pathways. Subsequently, the weighted gene co-expression network analysis (WGCNA) was used to construct a co-expression module of feature mRNAs and differential lncRNAs. Finally, based on 2 autophagy-related crosstalk genes (CD40 and SMAD7), we constructed a prognosis model by multivariate Cox regression, which could predict the overall survival of AD patients with medium-to-high accuracy. In conclusion, we provided an effective method for extracting autophagy-related significant genes based on pathway crosstalk in AD. We found the biomarkers valuable to the AD prognosis, which may also play an essential role in the development and treatment of AD.

FIGURE 1

Framework of the experiment.

FIGURE 2

Network connection diagram of 10 pathways and their crosstalk genes. The blue diamond nodes represent different pathways, and the size of the nodes represents the score of the pathway in PAGI. The circular node connected to the blue diamond node represents the genes included in the pathways. The shade of the color represents the level of the gene’s GDS score. The darker the node’s color, the higher the GDS score represents the gene. hsa04141: Protein processing in endoplasmic reticulum. hsa04210: Apoptosis. hsa05010: Alzheimer’s disease. hsa04150: mTOR signaling pathway. hsa04910: Insulin signaling pathway. hsa04722: Neurotrophin signaling pathway. hsa04144: Endocytosis. hsa04930: Type II diabetes mellitus. hsa04020: Calcium signaling.

FIGURE 3

Network construction of coexpressed genes. (A) Analysis of the scale independence and mean connectivity for various soft-threshold powers; (B) the cluster dendrogram of module Eigengenes; (C) dendrogram clustered based on a dissimilarity measure (1-TOM).

FIGURE 4

Venn diagram showing the overlap of crosstalk genes, autophagy-related genes (autophagy), and Alzheimer’s disease pathway genes (AD pathway, hsa05010).

FIGURE 5

The results of pathway enrichment. The horizontal axis is the number of genes in the pathway, and the vertical axis is the pathway list. Red to blue indicates the q-value (adjusted p-value).

FIGURE 6

Prognostic significance analysis of the Alzheimer’s disease training dataset. (A) Kaplan–Meier curve to compare OS of high risk with low-risk samples (p = 0.002); (B) forest plot of multivariate independent prognostic analysis. The square on the horizontal line shows the hazard ratio (HR), and the horizontal line represents the 95% confidence interval; (C) time-dependent receiver operating characteristic (ROC) curve analysis of the risk score model for predicting 3- and 5-year OS; and (D) multi-index ROC curve. The curve area is used to assess the accuracy of the risk model (model AUC = 0.758).

FIGURE 7

Prognostic significance analysis of the Alzheimer’s disease testing dataset. (A) Kaplan–Meier curve to compare OS of high risk with low-risk samples (p < 0.001); (B) forest plot of multivariate independent prognostic analysis. The square on the horizontal line shows the hazard ratio (HR), and the horizontal line represents the 95% confidence interval; (C) time-dependent receiver operating characteristic (ROC) curve analysis of the risk score model for predicting 3- and 5-year OS; and (D) multi-index ROC curve. The curve area is used to assess the accuracy of the risk model (model AUC = 0.737).

FIGURE 8

Relative expression levels of CD40 in Alzheimer’s disease and control sample (“*” represents p < 0.05, “**” represents p < 0.01, “***” represents p < 0.001). (A) frontal cortex of GSE118553; (B) hippocampus of GSE5281; and (C) posterior cingulate of GSE5281.

FIGURE 9

Relative expression levels of SMAD7 in Alzheimer’s disease and control sample (“*” represents p < 0.05, “**” represents p < 0.01, “***” represents p < 0.001). (A) Entorhinal cortex of GSE118553; (B) temporal cortex of GSE118553; and (C) Syn4009614.

SUPPLEMENTAL FIGURE 1

Genes associated with other diseases act to Alzheimer’s disease pathway. The red rectangles represent biological processes involved in genes associated with other diseases.