Cocaine'omics: Genome-wide and transcriptome-wide analyses provide biological insight into cocaine use and dependence

Abstract

We investigated the genetic and molecular architecture of cocaine dependence (CD) and cocaine use by integrating genome-/transcriptome-wide analyses. To prioritize candidates for follow-up investigation, we also sought to translate gene expression findings across species. Using data from the largest genome-wide association study (GWAS) of CD to date (n = 3176, 74% with CD), we assessed genomic heritability, gene-based associations, and tissue enrichment. We detected a significant single-nucleotide polymorphism heritability of 28% for CD and identified three genes (two loci) underlying this predisposition: the C1qL2 (complement component C1 q like 2), KCTD20 (potassium channel tetramerization domain containing 20), and STK38 (serine/threonine kinase 38) genes. Tissue enrichment analyses indicated robust enrichment in numerous brain regions, including the hippocampus. We used postmortem human hippocampal RNA-sequencing data from previous study (n = 15, seven chronic cocaine users) to follow up genome-wide results and to identify differentially expressed genes/transcripts and gene networks underlying cocaine use. Cross-species analyses utilized hippocampal gene expression from a mouse model of cocaine use. Differentially expressed genes/transcripts in humans were enriched for the genes nominally associated with CD via GWAS (P < 0.05) and for differentially expressed genes in the hippocampus of cocaine-exposed mice. We identified KCTD20 as a central component of a hippocampal gene network strongly associated with human cocaine use, and this gene network was conserved in the mouse hippocampus. We outline a framework to map and translate genome-wide findings onto tissue-specific gene expression, which provided biological insight into cocaine use/dependence.