Association of common genetic variation with Parkinson's disease is driven by microglial modulation of Leucine-rich repeat kinase 2 expression

Previous studies have shown that non-coding genetic variation in the 5' region of the leucine-rich repeat kinase 2 (LRRK2) gene contributes to Parkinson's disease risk and influences LRRK2 expression, with the risk allele associated with increased expression of LRRK2. To interrogate the biological basis of this expression quantitative trait locus (eQTL), we used single nuclei RNA sequencing of human frontal cortex from donors with known LRRK2 risk genotypes to show that the eQTL propagates via microglia, despite robust expression of LRRK2 in other cell types. Second, we found that chromatin in the 5' region of LRRK2 is differentially accessible in microglia compared to other brain cell types using single nuclei ATAC sequencing. We confirmed these effects in a microglia cell model derived from human induced pluripotent stem cells (iPSCs) by generating single cell RNA sequencing and bulk ATAC sequencing datasets. Our results demonstrate that gene expression as a quantitative trait may influence specific cell populations, and that non-coding genetic variants that are associated with other diseases may propagate through restricted cell types. Overall design: Single nuclei RNA sequencing datasets were generated from frontal cortex samples from 15 donors (snRNAseq). Single nuclei ATAC sequencing datasets were generated from frontal cortex samples from 12 of the same group of 15 donors (snATACseq). Single cell RNA sequencing datasets were generated from iPSC-derived microglia from two cell lines (scRNAseq). Bulk ATAC sequencing datasets were generated from iPSC-derived microglia differentiated from four cell lines and also from iPSC-derived forebrain neurons differentiated from one cell line with three technical replicates (bulkATACseq). Single nuclei Multiome sequencing (snMulti-seq) was used to generate both ATAC and gene expression (GEX) datasets from the same nuclei extracted from 3 human substantia nigra samples. A large-scale scRNAseq CRISPRi screen was completed using iPSC-derived microglia differentiated from iPSC constitutively expressing dCas9-KRAB transduced with a pooled lentiviral sgRNA library.