Elucidating cellular contributions to tau-mediated neurodegeneration using drosophila and single-cell transcriptomics.

Abstract

Alzheimer's Dementia (AD) is neuropathologically defined by aggregations of amyloid-beta plaques and tau neurofibrillary tangles in the brain, with tau pathology being most strongly correlated with cognitive decline. Systems biology approaches profiling post-mortem human AD brain tissue have highlighted involvement of key processes such as innate immunity and translational regulation. However, interpretation of cross-sectional post-mortem gene expression data is confounded by co-morbidities and mixed brain pathologies at autopsy. Moreover, characterizing longitudinal, age-related changes in Alzheimer's dementia (AD) is of particular importance because clinical manifestation of AD is a decades-long process, and age is the strongest known risk factor. We previously completed RNA-sequencing and proteomics of several highly tractable, longitudinal, Drosophila models of tauopathy to characterize tau- and age-specific perturbations. However, cellular heterogeneity confounds interpretation of bulk tissue data. We thus leverage single cell RNA sequencing (scRNA-seq) to explore cell-specific contributions to tau- and age-mediated gene expression. Whole brains of adult Drosophila pan-neuronally expressing a mutant variant of human Tau (R406W) and controls were dissociated for scRNA-seq (10x Genomics Chromium) at 1-, 10-, and 20- days of age. Perturbations in cellular composition and gene expression were quantified. To address cross-species conservation of tau-mediated scRNAseq signatures, we leverage existing human AD brain scRNA-seq data. As expected, we highlight evidence for tau-mediated degeneration of Kenyon cells, a neuronal population in the Drosophila "mushroom body" that is required for learning and memory, along with a concomitant increase in glial cell numbers (e.g. gliosis). Gene set enrichment analysis reveal overrepresentation of acetylcholine receptor regulation pathways in vulnerable mushroom body neurons. Interestingly, regularized regression analyses highlight innate immune activity as predictive of tau-related neuronal reduction. Cluster-specific transcriptional signatures between flies and humans demonstrate strong correlations among inferred cell identities. Our results comprise a powerful single cell transcriptomic resource for studying tau-mediated disruption of gene expression and dynamic age-dependent changes at cellular resolution in a tractable model system that is amenable for high throughput functional validation.