Multimodal interrogation of cell type-specific transcriptional states and brain-wide activity profiles in maladaptive opioid use

Abstract

Opioid use disorder is thought to arise from persistently dysregulated neural activity in brain regions associated with reward and motivation, but the precise molecular and circuit mechanisms underlying individual differences in susceptibility to escalated opioid use are not well understood. To address this gap, we have utilized a rat model of IV heroin self-administration in which rats display robust individual differences in self-administration. Multiple addiction metrics are collected throughout the paradigm and used to generate an addiction risk profile for each rat, which can then be classified as high or low risk. We compared brain-wide cFOS between rats with varying addiction risk histories, with and without re-exposure to a drug-paired cue, demonstrating brain-wide neural activity differences between high and low addiction risk rats after cued reinstatement. Using a generalized linear model, we tested for significant relationships between addiction metrics and cFOS z-scores in each brain region, revealing specific signatures of neural activity associated with unique behavioral patterns. High addiction risk rats were additionally found to be uniquely characterized by a loss of correlated cFOS z-scores in cortical and parahippocampal brain regions during protracted withdrawal from heroin. A central node in opioid addiction circuitry is the nucleus accumbens (NAc), which integrates glutamatergic input from diverse brain regions with dopaminergic input from the ventral tegmental area. To characterize the transcriptional state of neurons throughout this circuit, we conducted snRNAseq in prelimbic cortex (PL), NAc, and VP after self-administration and protracted withdrawal. We hypothesized that subsets of NAc medium spiny neurons (MSNs) would be significantly perturbed in rats with high addiction risk behaviors, while PL neurons were expected to display more uniform changes. Nuclei isolation and library prep were conducted using 10X Genomics Next GEM pipeline, and data analysis was completed using the R package Seurat. Data was integrated with high-quality mouse scRNAseq datasets to facilitate identification of shared and distinct neuronal subtypes for each brain region, producing an atlas of known and novel cell types. Analysis of each neuronal subtype revealed significant transcriptional alterations in multiple NAc D1 and D2 expressing MSN sub-clusters, as well as L2/3 and L4/5 IT neurons in PL, including a conserved response in genes associated with postsynaptic densities across these distinct cell types. In contrast, VP neurons were largely unchanged by heroin exposure, despite high expression of the μ-opioid receptor in VP. WGCNA analysis revealed neuronal subtype-specific modules of gene expression with differential expression by heroin addiction risk group and highlighted the importance of the metabotropic glutamate receptor mGluR5 in delineating high and low addiction risk gene expression patterns in all three brain regions. Using an intersectional CRISPR/Cas9 viral strategy, we knocked out mGluR5 from D1 MSNs in the NAc, which are known to alter their activity in acute opioid reward and during opioid withdrawal. Grm5 KO rats were less likely to engage in escalated opioid intake if they were classified as low risk before this intervention, and displayed signs of stabilization in their addiction risk phenotype if they were previously categorized as high risk. In summary, we observed a brain-wide signature of altered neuronal activity in high-risk rats, as well as cell-type specific loss of correlated gene expression in PL and NAc. We hypothesize that the bidirectional changes detected in Grm5 (upregulated in PL, downregulated in NAc) may represent a protective homeostatic change induced after escalated opioid intake, accompanied by an opioid-use induced decline in NAc glutamatergic tone from higher-order cortical regions. Eliminating mGluR5 from D1 MSNs was sufficient to attenuate heroin escalation in both low-risk and high-risk rats. Thus, using a multi-modal approach, we have expanded our knowledge of the brain regions, circuits, and cell types that become dysregulated after escalated opioid use, and specifically identified a subtype of metabotropic glutamate receptor in the NAc as an important player in the regulation of neural plasticity across cycles of opioid use, abstinence, and relapse.