Integration of cell-specific gene expression and chromatin accessibility facilitates localization of neurodegenerative risk in microglia

Abstract Genome-wide association studies (GWAS) have been highly successful in nominating genomic loci underlying risk of common phenotypes, including neurodegenerative diseases (NDDs). However, due to linkage disequilibrium across the genome, a persistent challenge in interpretation of GWAS is to break down loci to genes and variants and thus nominate disease mechanisms. Here, we use iPSC-derived cells containing natural population-level variation to address GWAS loci across NDDs including Alzheimer’s disease (AD), Parkinson’s disease (PD) and Dementia with Lewy Bodies (DLB). We used two differentiation protocols yielding different cell types in the brain, neurons and microglia, previously shown to contribute to NDD pathogenesis. Using meta-analyses to combine QTL mapping in differentiated iPSCswith published human brain datasets, we observed multiple loci associated with NDDs that are often restricted to either neurons or microglia. We find that identification of colocalized GWAS and QTL signals, rather than simple expression, supports microglia as having a strong contribution to disease risk. The BIN1 locus contains an eQTL that overlaps genetic risk shared between AD and DLB and has five risk-overlapping peaks that are specific to microglia. We tested these peaks for enhancer activity using a perturb-seq based method in microglia. Our results show one of the nominated peaks controls BIN1 expression in microglia but also modifies expression of other genes at the locus. These results expand on the previous hypothesis that microglia play a causal, rather than reactive, role in NDDs and show that iPSC-derived cells are a useful model to experimentally dissect GWAS loci that colocalize with QTL. cis-QTL for differentiated fore-brain neurons (iFBn) and microglia (iMGL) cis-eQTL results - each archive contains the TensorQTL cis map csv and per chromosome parquet files iFBn_eQTL.tar.gz iMGL_eQTL.tar.gz cis-caQTL results - each archive contains the TensorQTL cis map csv and per chromosome parquet files iFBn_caQTL.tar.gz iMGL_caQTL.tar.gz Correlations of gene's and their cis-proximal ATAC peaks - each archive contains the TensorQTL cis map csv and per chromosome parquet files iMGL_gene-ATAC.tar.gz iFBn_gene-ATAC.tar.gz Meta-analyses of cis-eQTL for broad neuron types and microglia - meta-analyses of public post-mortem cell-type specific cis-eQTL and cis-eQTL from differentiated specific cell-types excitatory_meta_eqtl.parquet inhibitory_meta_eqtl.parquet microglia_meta_eqtl.parquet Colocalization scores of cis-eQTL, meta-cis-eQTL for specific cell-types, and public bulk cis-eQTL per brain region from MetaBrain for GWAS common neurodegenerative risk loci for Alzheimer's, Lewy Body Dementia, and Parkinson's disease. eqtl_coloc_scores.csv caqtl_coloc_scores.csv