Transcriptional signatures of hippocampal tau pathology in primary age-related tauopathy and Alzheimer's disease

Abstract

In primary age-related tauopathy (PART) and Alzheimer's disease (AD), tau aggregates share a similar structure and anatomic distribution, which is distinct from tau pathology in other diseases. However, transcriptional similarities between PART and AD and gene expression changes within tau-pathology-bearing neurons are largely unknown. Using GeoMx spatial transcriptomics, mRNA was quantified in hippocampal neurons with and without tau pathology in PART and AD. Synaptic genes were down-regulated in disease overall but up-regulated in tau-pathology-positive neurons. Two transcriptional signatures were associated with intraneuronal tau, both validated in a cortical AD dataset. Genes in the up-regulated signature were enriched in calcium regulation and synaptic function. Notably, transcriptional changes associated with intraneuronal tau in PART and AD were similar, suggesting a possible mechanistic relationship. These findings highlight the power of molecular analysis stratified by pathology and provide insight into common pathways associated with tau pathology in PART and AD.

Keywords: Alzheimer’s disease; CP: Neuroscience; neurodegeneration; neurofibrillary tangles; primary age-related tauopathy; tau; transcription.

Figure 1.. Transcriptional changes associated with tau pathology in PART and AD

(A) GeoMx workflow: tissue is labeled with antibodies and mRNA probes. A region of interest (ROI) is selected and then segmented by antibody labeling. A 10 μm ultraviolet light cleaves the probe barcodes, which are iteratively collected, sequenced, and analyzed from each segment. (B) Hippocampal sections stained for tau pathology (yellow, AT8) and nucleic acid (blue, SYTO13). ROIs were selected within CA1, segmented, and masked for neurons with (red, tau positive) or without (blue, tau negative) tau pathology. Transcript-specific barcodes were sequenced from each segment. (C) Higher power of positive (red) and negative (blue) neurons. (D) Principal-component analysis with PC1 accounting for 99.741% and PC2 accounting for 0.217% of the variance. (E) Differentially expressed genes in tau-positive neurons. (F) Expression of genes altered in tau-positive neurons. Significance by mixed-effects linear modeling after FDR correction. Error bars are 95% CI. AD, Alzheimer’s disease; Ctl, control; CPM, counts per million; FC, fold change; FDR, false discovery rate; Neg, tau pathology negative; Oligo, oligonucleotide; PART, primary age-related tauopathy; PC, principal component; Pos, tau pathology positive; T−, tau pathology negative; T+, tau pathology positive. Scale bars represent 50 μm. *FDR < 0.05 and **FDR < 0.01.

Figure 2.. Synaptic gene changes associated with tau pathology in PART and AD

Synaptic genes were identified with SynGO annotation. (A) Comparison in (B) and (C); control was compared to all neurons in either PART or AD (with and without tau pathology). (B and C) Decreased synaptic gene expression in PART (B) and AD (C) vs. control shown by fold change of SynGO genes (top) with histogram (below). (D) Comparison in (E) and (F); tau-positive neurons compared to tau-negative neurons in PART or AD. (E and F) Increased synaptic gene expression in tau-positive neurons in PART (E) and AD (F) shown by fold change of SynGO genes (top) with histogram (below). (G and H) Representative synaptic and calcium gene changes. Significance by mixed-effects linear model after FDR correction. *FDR < 0.05. (I) Correlation of fold change in tau-positive neurons in PART and AD (simple linear regression, p < 0.0001, R² = 0.45). AD, Alzheimer’s disease; Ctl, control; CPM, counts per million; FC, fold change; Neg, tau pathology negative; PART, primary age-related tauopathy; Pos, tau pathology positive; T−, tau pathology negative; T+, tau pathology positive. (B, C, E, and F) The curves were generated by Gaussian least squares fit, and the mean is significantly different from 0 by an extra sum of squares F test; the dotted line is at 0. Error bars are SEM. (G and H) Error bars are 95% confidence interval (CI).

Figure 3.. Differential gene expression by tau pathology and neuritic amyloid plaque burden

Gene expression was quantified using hybridization chain reaction (HCR) in cases with varying amyloid plaque burden. (A and D) Representative HCR images with 2 DEGs measured per section. (B, C, E, and F) Tau pathology was associated with increased DEG expression (tau effect: B, SYT1 p = 0.000131; C, PRNP p = 1.88–8; E, APLP2 p = 0.000665; and F, CALM1 p = 3.01–5). Amyloid effect was significant for APLP2 (E, quadratic amyloid effect: p = 0.0222). Amyloid and tau interaction was significant for PRNP (C, cubic interaction term: p = 0.0258). Statistics by mixed-effects linear model. Neuritic plaque score: 0, none; 1, sparse; 2, moderate; and 3, frequent. Scale bars represent 10 μm. AD, Alzheimer’s disease; FC, fold change compared to tau negative; Neg, tau pathology negative; NP, neuritic plaque score; PART, primary age-related tauopathy; Pos, tau pathology positive. (B–F) Error bars are 95% CI. Points and outliers not shown.

Figure 4.. CoGAPS analysis by tau pathology in hippocampal neurons and published cortical neuron dataset

(A and B) Coordinated gene association in pattern set (CoGAPS) score for up-regulated (tau up) pattern by disease group and tau pathology (A) and heatmap of the top 20 genes (B). (C) Dot plot of biological processes enriched in up-regulated pattern. Dot size reflects number of genes, while color reflects p value after FDR correction. (D and E) CoGAPS score for down-regulated (tau down) pattern by disease group and tau pathology (D) and heatmap of the top 20 genes (E). (F–J) Data from cortical neurons in AD were analyzed for synaptic gene changes and expression of CoGAPS patterns. (F) Excitatory neuron group 2 (Ex2: CUX2-COL5A2) showing fold change of SynGO genes (top) with histogram (below). The curve was generated by Gaussian least squares fit. Significance is the mean of synaptic genes vs. all genes by an extra sum of squares F test; dotted line is at the mean of all genes. (G and H) Expression of up-regulated pattern (G) and down-regulated pattern (H) in tau-positive excitatory cortical neurons. (I and J) Patterns by excitatory neuronal group. AD, Alzheimer’s disease; All, all quantified genes; Ctl, control; FC, log2 fold change; FDR, false discovery rate corrected p value; Neg, tau pathology negative; PART, primary age-related tauopathy; Pos, tau pathology positive; Reg., regulation; Syn, synaptic genes; T+, tau positive. (F) Error bars are SEM. (A, D, and G–J) Error bars are 95% CI.