Thomas Jhou

Clinical and preclinical studies indicate that adaptations in corticostriatal neurotransmission significantly contribute to heroin relapse vulnerability. In animal models, heroin self-administration and extinction produce cellular adaptations in both neurons and astrocytes within the nucleus accumbens (NA) core that are required for cue-induced heroin seeking. Specifically, decreased glutamate clearance and reduced association of perisynaptic astrocytic processes with NAcore synapses allow glutamate release from prelimbic (PrL) cortical terminals to engage synaptic and structural plasticity in NAcore medium spiny neurons. Normalizing astrocyte glutamate homeostasis with drugs like the antioxidant N-acetylcysteine (NAC) prevents cue-induced heroin seeking. Surprisingly, little is known about heroin-induced alterations in astrocytes or pyramidal neurons projecting to the NAcore in the PrL cortex (PrL-NAcore). Here, we observe functional adaptations in the PrL cortical astrocyte following heroin self-administration (SA) and extinction as measured by the electrophysiologically evoked plasmalemmal glutamate transporter 1 (GLT-1)-dependent current. We likewise observed the increased complexity of the glial fibrillary acidic protein (GFAP) cytoskeletal arbor and increased association of the astrocytic plasma membrane with synaptic markers following heroin SA and extinction training in the PrL cortex. Repeated treatment with NAC during extinction reversed both the enhanced astrocytic complexity and synaptic association. In PrL-NAcore neurons, heroin SA and extinction decreased the apical tuft dendritic spine density and enlarged dendritic spine head diameter in male Sprague–Dawley rats. Repeated NAC treatment during extinction prevented decreases in spine density but not dendritic spine head expansion. Moreover, heroin SA and extinction increased the co-registry of the GluA1 subunit of AMPA receptors in both the dendrite shaft and spine heads of PrL-NAcore neurons. Interestingly, the accumulation of GluA1 immunoreactivity in spine heads was further potentiated by NAC treatment during extinction. Finally, we show that the NAC treatment and elimination of thrombospondin 2 (TSP-2) block cue-induced heroin relapse. Taken together, our data reveal circuit-level adaptations in cortical dendritic spine morphology potentially linked to heroin-induced alterations in astrocyte complexity and association at the synapses. Additionally, these data demonstrate that NAC reverses PrL cortical heroin SA-and-extinction-induced adaptations in both astrocytes and corticostriatal neurons.