Common and divergent pathways in early stages of glutamate and tau-mediated toxicities in neurodegeneration

Abstract

It has been shown that excitotoxicity and tau-mediated toxicities are major contributing factors to neuronal death in Alzheimer's disease (AD). The excitatory amino acid transporter 2 (EAAT2 or GLT-1), the major glutamate transporter in the brain that regulates glutamate levels synaptically and extrasynaptically, has been shown to be deficient in AD brains, leading to excitotoxicity and subsequent cell death. Similarly, buildup of neurofibrillary tangles, which consist of hyperphosphorylated tau protein, correlates with cognitive decline and neuronal atrophy in AD. However, common genes and pathways that are critical in the aforementioned toxicities have not been well elucidated. To investigate the impact of glutamate dyshomeostasis and tau accumulation on translational profiles of affected hippocampal neurons, we used mouse models of excitotoxicity and tau-mediated toxicities (GLT-1^-/- and P301S, respectively) in conjunction with BAC-TRAP technology. Our data show that GLT-1 deficiency in CA3 pyramidal neurons leads to translational signatures characterized by dysregulation of pathways associated with synaptic plasticity and neuronal survival, while the P301S mutation induces changes in endocytic pathways and mitochondrial dysfunction. Finally, the commonly dysregulated pathways include impaired ion homeostasis and metabolic pathways. These common pathways may shed light on potential therapeutic targets for ameliorating glutamate and tau-mediated toxicities in AD.

Keywords: Alzheimer's disease; Bacterial artificial chromosome; Differentially expressed genes; Excitatory amino acid transporter-1; Glutamate transporter-1; Neurofibrillary tangle; Translating ribosome affinity purification.

Fig. 1.

Generation and validation of transgenic GLT-1^−/− and P301S mice. (A) Schematic diagram showing the breeding scheme and TRAP experiment protocol for GFAP-Cre⁺ GLT-1^loxP/loxP;Gprin3 and P301S;Gprin3-eGFP10a (P301S) mice. Five day old GLT-1^−/− (GFAP-Cre⁺ GLT-1^loxP/loxP;Gprin3) and control (GFAP-Cre⁻ GLT-1^loxP/loxP;Gprin3) pups were subject to 4-OHT injection (33 mg/kg body weight). (B) Representative immunofluorescence images of a Gprin3-eGFPL10a mouse hippocampus stained with NeuN (red) and anti-eGFP (green) antibodies. Images were captured using a 20× objective. (C) A heat map of the read counts (RPKM) of cell type marker genes from TRAP samples shows that neuronal genes had the highest expression, validating the neuronal targeting of BAC-TRAP mice. The scale bar represents log₂RPKM values.

Fig. 2.

GLT-1 deletion is associated with dysregulation of synapse and endocytosis-related pathways. (A) Heatmap of Pearson’s correlation of gene expression data between GLT-1^−/− and control. Data represent Pearson’s r correlation derived from RPKM values. (B) Volcano plot depicting DEGs in GLT-1^−/− versus control hippocampi. Horizontal and vertical dotted lines represent filtering criteria (p ≤ 0.05, log₂FC ≥ 0.5). Red and blue dots represent significantly up and down regulated genes. Grey dots represent genes that were determined to be not significant. (C) Heat map showing RPKM z-scores for the individual replicates of the top 50 protein coding genes. (D) Canonical pathways derived from IPA analyses for significantly dysregulated genes in GLT-1^−/− CA3 neurons. Colour gradient represents z-score. The colors red and blue represent up and down regulated canonical pathways respectively. N = 4 mice for GLT-1^−/− and control groups for panel (A) and (C). N = 4 and 6 for GLT-1^−/− and control groups respectively for panel (B) and (D).

Fig. 3.

The P301S mutation is related to dysregulation of metabolic function and endocytosis pathways. (A) Heatmap of Pearson’s correlation of gene expression data between P301S and control mice. Data represent Pearson’s r correlation derived from RPKM values. (B) Volcano plot depicting DEGs in CA3 neurons of P301S mice at 3 months. Horizontal and vertical dotted lines represent filtering criteria (p ≤ 0.05, log₂FC ≥ 0.5). The red and blue dots represent significantly up and down regulated genes. Grey dots represent genes that were determined to be not significant. (C) Heat map showing RPKM z-scores of the top 50 protein coding genes for individual replicates. (D) Canonical pathways derived from IPA analyses for significantly dysregulated genes in P301S CA3 neurons. Colour gradient represents z-score. The colors red and blue represent up and down regulated canonical pathways respectively. N = 6 and 4 for P301S and control mice respectively for the panel (A) and (C). N = 7 and 6 for P301S and control mice respectively for the panel (B) and (D).

Fig. 4.

Overlaps in gene expression associated with electrophysiological function and metabolism found between transgenic P301S and GLT-1^−/− mice. (A) Stratified heat map showing significant overlaps in gene expression between GLT-1^−/− mice and P301S mice at 3 months of age. The upper right quadrant shows overlap in genes downregulated in both studies. The lower left quadrant shows overlap in genes up regulated in both studies. The lower right and upper left quadrants represent overlaps in genes downregulated in one and upregulated in the other and upregulated in one and downregulated in the other respectively. (B) Canonical pathways derived from IPA for genes significantly dysregulated in both transgenic groups. (C) Venn analysis showing the shared and differentially expressed protein coding genes (p ≤ 0.05, log₂FC ≥ 0.5). 96 (63 down and 30 up) shared protein coding genes were found in GLT-1^−/− and P301S mice (C). (D) Heatmap represents the expression pattern of 20 genes derived from Venn analysis for GLT-1^−/−, P301S, and control mice showing the respective RPKM count z-scores. The colors red and blue represent up and downregulation. N = 4, 7, and 6 for GLT-1^−/−, P301S, and control mice respectively.