Uncovering cell-type-specific driver genes of Alzheimer's Disease by pathology-indexing scRNA-seq, spatial transcriptomics, and CRISPR screens

PROJECT SUMMARY Although amyloid-β plaques and neurofibrillary tangles are the current criteria for pathologic diagnosis of Alzheimer’s Diseases (AD), only 9% of clinically diagnosed AD patients have "pure" AD pathology and most AD cases have mixed pathologies, which significantly increase the odds of dementia . Because diverse intra- and extracellular pathologies and stressors contribute to AD progression, it is essential to track how they affect the various cell types of the brain by cataloging cell-type-specific transcriptomic responses to both intra- and extracellular pathologies in AD pathogenesis. Therefore, this proposal aims to measure the effects of multiple pathologies on each cell type in their native environment, then make this information actionable by computationally identifying the drivers of these effects and testing them in human cell models. To this end, we propose two approaches to simultaneously measure the cell transcriptomes and multiple pathologies in millions of individual cells in their native context. The first approach, “pathology-indexing scRNA-seq,” is designed for intracellular pathologies. It combines single-cell RNA-seq (scRNA-seq) with a set of oligo- barcoded antibodies against intracellular pathologies. This approach enables us to simultaneously measure gene expression and multiple intracellular pathologies all in the same cell. The second approach, “pathology spatial transcriptomics,” is designed for extracellular pathologies. It obtains gene expression of 1~10 cells (55- μM resolution) in spatial registration with extracellular pathology. This enables us to quantify the effects of extracellular pathologies and microenvironment on cell disease states. We will apply these two innovative sequencing technologies to two brain regions, the dorsal lateral prefrontal cortex and hippocampus of postmortem brains of deeply-phenotyped ROSMAP participants. Using univariate, systems biology, and deep learning computational methods, we will identify candidate genes that drive cell-type-specific disease states. To test predicted early driver genes and provide therapeutic targets, we will conduct CRISPR screens in human cortical cell models derived from control and AD isogenic iPSC lines. Collectively, our study will shed important light on the cell-type-specific driver genes in AD pathogenesis, define molecular pathways leading to cell disease-states, and provide experimentally validated targets for preventing the disease-state transition during early AD development.