Single nucleus transcriptomic analysis of the rat nucleus accumbens reveals novel medium spiny neuron subtypes and distinct patterns of polygenic enrichment for substance use phenotypes

The nucleus accumbens (NAc) is known to play important roles in drug-taking and -seeking behaviors. The primary output pathways of the NAc consist of GABAergic medium spiny neurons (MSNs), usually expressing dopamine type 1 receptors (D1 MSNs) and dopamine type 2 receptors (D2 MSNs). These groups of MSNs in the NAc are known to have functional differences and additional evidence suggests that there are distinct subtypes within the major groups. However, details on MSN subtypes are limited and the potential roles of these populations in substance use disorders are largely unknown. The goal of the current study is to make use of the largest currently available rat NAc single nucleus transcriptomic data set to present high resolution subclustering of MSNs and polygenic enrichment for substance use phenotypes in MSN subtypes. Brains from male control Brown Norway rats (n=11) were snap frozen and NAc samples were punched. Single nucleus RNAseq libraries were generated using the 10x Genomics Chromium Controller and sequenced on a NovaSeq 6000. The R package Seurat was used for quality control and to cluster nuclei based on transcriptomic data. MSNs (n=47,591) were identified based on known marker genes and subclustered. MSNs from an independent rat NAc replication data set were mapped to the discovery data using the Seurat MapQuery function. Correlation of gene expression between clusters in the discovery and replication data sets was calculated by Pearson correlation coefficients. Gene-level summary statistics were obtained for genome-wide association studies of tobacco use disorder, opioid use disorder, alcohol use disorder, and alcohol consumption using MAGMA. Polygenic enrichment for each phenotype was scored for all nuclei with scDRS. Subclustering of MSNs resulted in identification of 37 transcriptomically-distinct populations, including D1 and D2 MSNs and separate non-D1/D2 populations. All subclusters were able to be defined by expression of only 1-3 marker genes and were not statistically biased towards any individual sample. Mapping of an independent rat NAc data set onto the discovery data revealed the presence of all 37 subclusters in the replication samples, supporting their validity. Unique populations of MSNs were significantly enriched for expression of genes associated with each of the substance use phenotypes analyzed. These data indicate that MSNs consist of a large number of subtypes with unique gene expression profiles, which likely represent different functional niches within the NAc. Subtype-specific differences in polygenic enrichment for various substance use phenotypes suggest these phenotypes may engage distinct MSN populations. Future work should focus on targeted manipulation of the relevant MSN subtypes to determine their functional significance in the context of substance use.