Chromatin accessibility mapping for functional GWAS risk variants of neuropsychiatric and neurodegenerative disorders in hiPSC-derived neural cells

Despite genome-wide association studies (GWAS) of late-onset Alzheimer’s disease (LOAD) have identified 75 genetic risk loci 2-7, the disease causal mechanism underlying most GWAS risk loci remains largely unknown. A major challenge is how to identify a putative causal disease variant among multiple risk variants at individual GWAS risk locus and functionally link the risk allele to LOAD-relevant cellular phenotypes. Leveraging our recent approach for identifying functional noncoding GWAS risk variants that show allele-specific open chromatin (ASoC), we systematically identified putative causal LOAD risk variants in human iPSC-derived neurons and glia cells, and established a causal link between PICALM risk allele to unappreciated microglia (MG)-specific role of PICALM in lipid droplet (LD) accumulation and phagocytosis. Our ASoC mapping identified functional risk variants for 26 AD risk loci, most of which are MG-specific. At the PICALM locus, the LOAD risk allele of the ASoC SNP rs10792832 altered chromatin accessibility to PU.1 binding and reduced PICALM expression in MG, thereby leading to impaired phagocytosis of Amyloid beta (Aβ) and myelin debris. Interestingly, transcriptomic profiling of MG carrying PICALM risk allele not only supported abnormal phagocytosis but also revealed enrichment of gene pathways related to cholesterol synthesis and LD formation. Further genetic and pharmacological perturbations in MG established the causal link of the PICALM risk allele with PICALM reduction, LD accumulation and phagocytosis deficits. Our work elucidates an MG-specific role of PICALM in regulating lipid/cholesterol accumulation that is potentially linked to AD pathophysiology, providing a functional neurobiological basis for developing novel clinical interventions.