Interindividual variation in human cortical cell type abundance and expression

INTRODUCTION Interindividual variation is a known feature of the human brain. In neocortical regions, broad classes of neurons and non-neuronal cells vary in abundance and gene expression in healthy adult humans, yet variation remains unexplored within finer divisions of cell types. Characterizing the effects of biological factors such as age, sex, ancestry, disease, or genetic variants on high-resolution cell types requires molecular profiling of single nuclei from large donor cohorts.RATIONALE Single-nucleus RNA-sequencing (snRNA-seq) combined with whole-genome sequencing (WGS) applied to a large cohort of donors enable comprehensive assessment of interdonor variation in the neocortex of nonaged adults. Integrated analyses of demographic characteristics, single-nucleotide polymorphisms (SNPs), and gene expression can implicate genes associated with biological factors and gene regulatory regions in specific cell types. Such work can also serve as an important baseline to interpret cellular variation in neurological and psychiatric diseases that affect cortical function, and for understanding changes with age and associated neurodegenerative disorders.RESULTS We present comprehensive transcriptomic (snRNA-seq) and genomic (WGS) profiling of cortical tissues from 75 human brain donors comprising nearly 400,000 nuclei covering all major neocortical cell types. We show quantitatively that nearly all cells collected from these adult donors can be confidently assigned to a cell type in a taxonomy defined by using only a few donors. This demonstrates a highly consistent cellular architecture across individuals and confirms the viability of cell type mapping for larger-scale studies. These highly conserved cell types showed substantial variation in gene expression and abundances between individuals driven by multiple biological factors. Underlying medical conditions in our donor cohort affect cellular abundance. For example, PVALB-expressing interneurons show decreased abundance in epilepsy cases, reflecting previously reported cell loss in this disease. We found differences in gene expression across individuals at the finest cell type resolution. Gene networks in excitatory neurons and glia were particularly variable across donors, irrespective of medical condition or brain region. Deep-layer neuronal types that communicate with distant brain regions showed higher variation than that in superficial types. A substantial proportion of variation in gene expression is explained by donor, including contributions from age, sex, ancestry, and disease state. Furthermore, genomic variation was significantly associated with variable gene expression, with most cell types containing cis-expression quantitative trait loci. Yet much variation remains unexplained by measured factors, paving the way for larger studies of similar design.CONCLUSION By profiling the human brain across 75 adult individuals through use of snRNA-seq and WGS, we assessed variation in cortical cellular abundance and gene expression at cell type-level resolution. This study indicates a highly consistent cellular makeup across human individuals but with substantial variation that reflects donor characteristics, disease condition, and genetic regulation that will provide a comprehensive reference for future studies of disease.