Deciphering the transcriptomic impacts of high-risk copy number variants in the human brain at single cell resolution

Copy number variants (CNVs) are duplications or deletions of specific portions of DNA. Some large, recurrent CNVs confer risk for neuropsychiatric disorders with high penetrance. While CNVs have been studied using neuroimaging, cell cultures, and animal models, little is known about their impact on the human brain transcriptome. We sought to better elucidate the neurobiological architecture of high-risk CNVs through RNA sequencing carried out at single-cell resolution. We collected single-nucleus RNA-sequencing data in dorsolateral prefrontal and anterior cingulate cortex from 13 carriers of neuropsychiatric CNVs (chromosomes 22q11.2, 16p11.2, 1q21.1, 7q11.23 del, and 15q11.2), along with 24 non-carriers matched as much as possible for age, race, sex, and psychiatric diagnosis. Data were processed via CellRanger followed by ambient RNA and doublet removal, quality control and filtering in Seurat, and cell type annotation using Azimuth. We performed differential gene expression analyses across nine major cell types using Dreamlet and functional enrichments using Zenith. As expected, expression of genes in CNV regions was consistent with copy number. Across the transcriptome, results showed more differentially expressed genes (DEGs; FDR < 0.1) in deletion carriers than the respective duplication carriers. Exploration of functional enrichment of DEGs across deletion carriers revealed enrichment of cellular stress and energy metabolism terms. Carriers of the 22q11.2 deletion and 7q11.23 deletion displayed differences in the number of DEGs across cell types. In the 22q11.2 deletion, astrocytes had more than 5x the number of DEGs than the other cell types. These DEGs were enriched for terms related to glycolysis, cellular stress response specifically associated with hypoxia, and circadian regulation of gene expression across cell types, with most prominent results in astrocytes. The 7q11.23 deletion showed almost 10x the number of DEGs in excitatory neurons compared to the other cell types. DEGs were enriched for numerous terms related to oxidative phosphorylation and mitochondrial function significantly across cell types, with many terms most strongly upregulated in excitatory neurons. This is the first transcriptomic interrogations of high-risk CNVs using post-mortem human brain at single-cell resolution. We found that deletions had greater impact on gene expression than duplications, which is consistent with the greater phenotypic impact of deletions. Some CNVs also showed disproportionate effects on specific cell types, which would have been missed in study designs lacking cellular resolution. Our functional enrichment results demonstrated that across major cell types, deletion carriers showed an upregulation of genes associated with cellular stress responses, suggesting neurobiological changes within and across cell types due to CNVs. Future work will further explore convergent and divergent effects across distinct CNVs.