The significance of glycolysis index and its correlations with immune infiltrates in Alzheimer's disease

Abstract

Alzheimer's disease (AD) is a common neurodegenerative disorder without an effective treatment, and results in an increasingly serious health problem. However, its pathogenesis is complex and poorly understood. Nonetheless, the exact role of dysfunctional glucose metabolism in AD pathogenesis remains unclear. We screened 28 core glycolysis-related genes and introduced a novel metric, the glycolysis index, to estimate the activation of glycolysis. The glycolysis index was significantly lower in the AD group in four different brain regions (frontal cortex, FC; temporal cortex, TC; hippocampus, HP; and entorhinal cortex, EC) than that in the control group. Combined with differential expression and over-representation analyses, we determined the clinical and pathological relevance of glycolysis in AD. Subsequently, we investigated the role of glycolysis in the AD brain microenvironment. We developed a glycolysis-brain cell marker connection network, which revealed a close relationship between glycolysis and seven brain cell types, most of which presented abundant variants in AD. Using immunohistochemistry, we detected greater infiltrated microglia and higher expression of glycolysis-related microglia markers in the APP/PS1 AD model than that in the control group, consistent with our bioinformatic analysis results. Furthermore, the excellent predictive value of the glycolysis index has been verified in different populations. Overall, our present findings revealed the clinical and biological significance of glycolysis and the brain microenvironment in AD.

Keywords: Alzheimer’s disease; brain cell markers; glycolysis index; microglia; prognostic indicator.

Figure 1

Glycolysis index and its clinical relevance in AD (A-C) GSEA analysis on AD vs. control samples of the glycolysis gluconeogenesis (KEGG) gene set in FC. (A) Glycolysis and gluconeogenesis (Wikipathways) in FC and (B) glycolysis (Hallmark) in TC (C) The green curve denotes the running ES for the gene set as the analysis moves down the ranked list of genes. Positive correlation, NES >0; Negative correlation, NES <0. Adj. P, adjusted p-value. These callouts also apply to the Supplementary Figure 1 . (D) The 28 core glycolysis genes. Cells in red denote the genes identified as leading edge genes in these brain regions. (E) AD samples reveal lower glycolysis indices than that in the control in different brain regions. ***p<0.001. (F) The trends of glycolysis index in different Braak NFT stages using the GSE84422 dataset. Blue dots represent the individual samples. NES, Normalized enrichment score; GSEA, gene set enrichment analysis; AD, Alzheimer’s disease; FC, frontal cortex; TC, temporal cortex; ES, enrichment score; NFT, neurofibrillary tangles; and KEGG, Kyoto Encyclopedia of Genes and Genomes.

Figure 2

Screening glycolysis-related genes, and their relevant signaling pathways and functions in AD (A) Volcano plots of DEGs in AD vs. control samples. (B) Volcano plots of DEGs between AD samples with low and high glycolysis indices. The genes with FDR <0.05 and |log₂ FC|>0.5 are considered significant. (C) Venn diagram of the overlapping genes between the two differential expression analyses. (D) The GO enrichment analysis of intersection DEGs. (E) The KEGG enrichment analysis of intersection DEGs. DEG, differentially expressed genes; AD, Alzheimer’s disease; FDR, false discovery rate; GO, gene ontology; and KEGG, Kyoto Encyclopedia of Genes and Genomes.

Figure 3

The association between glycolysis and the brain microenvironment (A) Heatmap depicts the abundance of six brain cells calculated by CIBERSORTx in AD and control samples. Rho: the Spearman’s rho between the abundance and glycolysis index. The significance of the differences between AD and control are measured by the directed log P values of the Wilcoxon’s rank-sum tests. A positive value denotes higher abundance in the AD group. (B) Relative abundance of six brain cell types in different brain regions between control and AD samples. (C) The quantification of microglia using the ES of microglia marker genes in different brain regions. (D) Spearman’s correlation analysis between the glycolysis index and the ES of microglia markers. Different colors indicate different regions, and it is the same as the color in (C). *p<0.05, ***p<0.001. AD, Alzheimer’s disease; ES, enrichment score.

Figure 4

Mining of glycolysis-related brain cell markers (A) The correlations between glycolysis-related genes and various brain cell marker genes. The thickness of the line denotes the number of metrics supporting the correlation. The red line denotes a positive correlation, whereas the blue line denotes a negative correlation. (B) The GO enrichment analysis of glycolysis-related brain cell markers. (C) The KEGG enrichment analysis of glycolysis-related brain cell markers. (D) Boxplots of SALL1 expression in different brain regions between the control and AD samples. (E) Spearman’s correlation analysis between the glycolysis index and SALL1 expression. **p<0.01, ***p<0.001. AD, Alzheimer’s disease; GO, gene ontology; SALL1, Spalt Like Transcription Factor 1; and KEGG, Kyoto Encyclopedia of Genes and Genomes.

Figure 5

The expression of Iba1 and Sall1 in C57BL/6J mice and APP/PS1 AD model mice (A, C) IHC staining for Iba1 and Sall1 in different regions between the control and AD samples. (B, D) The expression levels of Iba1 and Sall1 in different regions between the control and AD samples, n=4. *p<0.05, **p<0.01. AD, Alzheimer’s disease; SALL1, Spalt Like Transcription Factor 1; and Iba1, ionized calcium-binding adapter molecule 1.

Figure 6

Distributions of glycolysis indices in normal brain tissues (A) The sex difference of glycolysis indices. (B) Glycolysis indices in different age groups. (C) Distributions of glycolysis indices in the different parts of the brain. ***p<0.001.