Outputs from the nucleus accumbens (NAc) include projections to the ventral pallidum and the ventral tegmental area and subtantia nigra in the ventral mesencephalon. The medium spiny neurons (MSN) that give rise to these pathways are GABAergic and consist of two populations of equal number that are segregated by differentially expressed proteins, including D1- and D2-dopamine receptors. Afferents to the ventral pallidum arise from both D1- and D2-MSNs, whereas the ventral mesencephalon is selectively innervated by D1-MSN. To determine the extent of collateralization of D1-MSN to these axon terminal fields we used retrograde labeling in transgenic mice expressing tdTomato selectively in D1-MSN, and found that a large majority of D1-MSN in either the shell or core subcompartments of the accumbens collateralized to both output structures. Approximately 70% of D1-MSNs projecting to the ventral pallidum collateralized to the ventral mesencephalon, whereas >90% of mesencephalic D1-MSN afferents collateralized to the ventral pallidum. In contrast, <10% of dorsal striatal D1-MSNs collateralized to both the globus pallidus and ventral mesencephalon. D1-MSN activation is required for conditioned cues to induce cocaine seeking. To determine which D1-MSN projection mediates cued cocaine seeking, we selectively transfected D1-MSNs in transgenic rats with an inhibitory Gi-coupled DREADD. Activation of the transfected Gi-DREADD with clozapine-N-oxide administered into the ventral pallidum, but not into the ventral mesencephalon, blocked cue-induced cocaine seeking. These data show that, although accumbens D1-MSNs largely collateralize to both the ventral pallidum and ventral mesencephalon, only D1-MSN innervation of the ventral pallidum is necessary for cue-induced cocaine seeking.SIGNIFICANCE STATEMENT Activity in D1 dopamine receptor-expressing neurons in the NAc is required for rodents to respond to cocaine-conditioned cues and relapse to drug seeking behaviors. The D1-expressing neurons project to both the ventral pallidum and ventral mesencephalon, and we found that a majority of the neurons that innervate the ventral pallidum also collateralize to the ventral mesencephalon. However, despite innervating both structures, only D1 innervation of the ventral pallidum mediates cue-induced cocaine seeking.

Relapse to drug use can be initiated by drug-associated cues. The intensity of cue-induced drug seeking in rodent models correlates with the induction of transient synaptic potentiation (t-SP) at glutamatergic synapses in the nucleus accumbens core (NAcore). Matrix metalloproteinases (MMPs) are inducible endopeptidases that degrade extracellular matrix (ECM) proteins, and reveal tripeptide Arginine-Glycine-Aspartate (RGD) domains that bind and signal through integrins. Integrins are heterodimeric receptors composed of αβ subunits, and a primary signaling kinase is focal adhesion kinase (FAK). We previously showed that MMP activation is necessary for and potentiates cued reinstatement of cocaine seeking, and MMP-induced catalysis stimulates β3-integrins to induce t-SP. Here, we determined whether β3-integrin signaling through FAK and cofilin (actin depolymerization factor) is necessary to promote synaptic growth during t-SP. Using a small molecule inhibitor to prevent FAK activation, we blocked cued-induced cocaine reinstatement and increased spine head diameter (dh). Immunohistochemistry on NAcore labeled spines with ChR2-EYFP virus, showed increased immunoreactivity of phosphorylation of FAK (p-FAK) and p-cofilin in dendrites of reinstated animals compared with extinguished and yoked saline, and the p-FAK and cofilin depended on β3-integrin signaling. Next, male and female transgenic rats were used to selectively label D1 or D2 neurons with ChR2-mCherry. We found that p-FAK was increased during drug seeking in both D1 and D2-medium spiny neurons (MSNs), but increased p-cofilin was observed only in D1-MSNs. These data indicate that β3-integrin, FAK and cofilin constitute a signaling pathway downstream of MMP activation that is involved in promoting the transient synaptic enlargement in D1-MSNs induced during reinstated cocaine by drug-paired cues.SIGNIFICANCE STATEMENT Drug-associated cues precipitate relapse, which is correlated with transient synaptic enlargement in the accumbens core. We showed that cocaine cue-induced synaptic enlargement depends on matrix metalloprotease signaling in the extracellular matrix (ECM) through β3-integrin to activate focal adhesion kinase (FAK) and phosphorylate the actin binding protein cofilin. The nucleus accumbens core (NAcore) contains two predominate neuronal subtypes selectively expressing either D1-dopamine or D2-dopamine receptors. We used transgenic rats to study each cell type and found that cue-induced signaling through cofilin phosphorylation occurred only in D1-expressing neurons. Thus, cocaine-paired cues initiate cocaine reinstatement and synaptic enlargement through a signaling cascade selectively in D1-expressing neurons requiring ECM stimulation of β3-integrin-mediated phosphorylation of FAK (p-FAK) and cofilin.

The brain µ-opioid receptor (MOR) is critical for the analgesic, rewarding, and addictive effects of opioid drugs. However, in rat models of opioid-related behaviors, the circuit mechanisms of MOR-expressing cells are less known because of a lack of genetic tools to selectively manipulate them. We introduce a CRISPR-based Oprm1-Cre knock-in transgenic rat that provides cell type-specific genetic access to MOR-expressing cells. After performing anatomic and behavioral validation experiments, we used the Oprm1-Cre knock-in rats to study the involvement of NAc MOR-expressing cells in heroin self-administration in male and female rats. Using RNAscope, autoradiography, and FISH chain reaction (HCR-FISH), we found no differences in Oprm1 expression in NAc, dorsal striatum, and dorsal hippocampus, or MOR receptor density (except dorsal striatum) or function between Oprm1-Cre knock-in rats and wildtype littermates. HCR-FISH assay showed that iCre is highly coexpressed with Oprm1 (95%-98%). There were no genotype differences in pain responses, morphine analgesia and tolerance, heroin self-administration, and relapse-related behaviors. We used the Cre-dependent vector AAV1-EF1a-Flex-taCasp3-TEVP to lesion NAc MOR-expressing cells. We found that the lesions decreased acquisition of heroin self-administration in male Oprm1-Cre rats and had a stronger inhibitory effect on the effort to self-administer heroin in female Oprm1-Cre rats. The validation of an Oprm1-Cre knock-in rat enables new strategies for understanding the role of MOR-expressing cells in rat models of opioid addiction, pain-related behaviors, and other opioid-mediated functions. Our initial mechanistic study indicates that lesioning NAc MOR-expressing cells had different effects on heroin self-administration in male and female rats.SIGNIFICANCE STATEMENT The brain µ-opioid receptor (MOR) is critical for the analgesic, rewarding, and addictive effects of opioid drugs. However, in rat models of opioid-related behaviors, the circuit mechanisms of MOR-expressing cells are less known because of a lack of genetic tools to selectively manipulate them. We introduce a CRISPR-based Oprm1-Cre knock-in transgenic rat that provides cell type-specific genetic access to brain MOR-expressing cells. After performing anatomical and behavioral validation experiments, we used the Oprm1-Cre knock-in rats to show that lesioning NAc MOR-expressing cells had different effects on heroin self-administration in males and females. The new Oprm1-Cre rats can be used to study the role of brain MOR-expressing cells in animal models of opioid addiction, pain-related behaviors, and other opioid-mediated functions.