Cell type specific gene expression profiles of the candidate genes for Alzheimer’s disease

Abstract

AbstractBackgroundDecades of genetic studies discovered candidate genes for Alzheimer’s disease (AD). Molecular mechanistic insights, however, are available for few candidate genes. This lack of understanding on gene function has been a challenge to translate the role of candidate genes to discovering treatment and diagnostic biomarkers. Recent advances in single cell profiling provide an opportunity to understand functional significance of candidate genes and their products for the cell types implicated in AD.MethodWe performed an integrative analysis of single cell gene expression profiling datasets (i.e., scRNA‐seq and snRNA‐seq) comprising 11 datasets generated from animal models (0.3 million cells) and 15 datasets prepared with human brain tissues (3.1 million cells) that were retrieved from AD Knowledge Portal and the other public repositories. All human and mouse datasets were merged using Harmony to discover cell subpopulations that were unique to conditions compared to controls and/or wild type. For AD candidate genes, cell types were ranked according to their relative expression levels and correlation structure.ResultClustering of brain cell types such as excitatory and inhibitory neurons, glial and vascular endothelial cells was similar across datasets and species; however, we found that transcriptional signature of microglia at different stage of human AD was distinct from disease‐associated microglia (DAM) observed in the mouse AD models. Of note, several homeostatic microglial genes known to be downregulated in mouse DAM were upregulated in human AD samples. Moreover, microglia were the only cell type that expressed all reported AD candidate genes. Nonetheless, the cell type specific expression pattern of AD candidate genes was different between species and across datasets suggesting biological and technical biases.ConclusionUsing an unbiased approach, we highlight the (dis)similarity between human and mouse AD datasets and potential molecular pathways for AD risk genes within microglia and the other cell types implicated in AD.