Use of marijuana (Cannabis sativa) often begins in adolescence, and heavy adolescent marijuana use is often associated with impaired cognitive function in adulthood. However, clinical reports of long-lasting cognitive deficits, particularly in subjects who discontinue use in adulthood, are mixed. Moreover, dissociating innate differences in cognitive function from cannabis-induced deficits is challenging. Therefore, the current study sought to develop a rodent model of adolescent cannabinoid self-administration (SA), using the synthetic cannabinoid receptor agonist WIN55,212-2 (WIN), in order to assess measures of relapse/reinstatement of drug seeking and long-term effects on cognitive function assessed in a delay-match-to-sample working memory task and a spatial recognition task. Adolescent male rats readily self-administered WIN in 2-h or 6-h sessions/day, but did not demonstrate an escalation of intake with 6-h access. Rats exhibited significant cue-induced reinstatement of WIN seeking that increased with 21 days of abstinence (ie, 'incubation of craving'). Cognitive testing occurred in adulthood under drug-free conditions. Both 2-h and 6-h adolescent WIN SA groups exhibited significantly better working memory performance in adulthood relative to sucrose SA controls, and performance was associated with altered expression of proteins regulating GABAergic and glutamatergic signaling in the prefrontal cortex. Self-administered WIN did not produce either acute or chronic effects on short-term memory, but experimenter administration of WIN in adolescence, at doses previously reported in the literature, produced acute deficits in short-term memory that recovered with abstinence. Thus, SA of a rewarding cannabinoid in adolescence does not produce long-term cognitive dysfunction.

Accurate discrimination among cues signifying reward, danger or safety initiates the proper emotional response in order to guide behavior. Appropriate conditioned inhibition of fear in the presence of a safety cue would allow an organism to engage in reward seeking behaviors. There is currently little known about the mechanisms of reward, fear and safety cue discrimination and how a safety cue can inhibit fear and release reward seeking from inhibition. Here we assess reward, fear and safety cue learning together using a behavioral paradigm that has identified neurons that discriminate among these cues in the basolateral amygdala (BLA) (Sangha, Chadick, & Janak, 2013). Dopamine signaling in the BLA has been implicated in discriminatory reward learning, learned fear responses and fear extinction. We tested the hypothesis that D1 receptor activity will influence reward-fear-safety cue discrimination by using the D1 receptor agonist, SKF-3839, and antagonist, SCH-23390, either systemically or within the BLA during discrimination learning in male Long Evans rats. We show that both the agonist and antagonist interfered with fear suppression in the presence of the safety cue, when administered systemically or when infused directly into the BLA. This indicates that altering D1 receptor activity in the basolateral amygdala impairs fear suppression during a safety cue. Neither the agonist or antagonist had a consistent negative impact on discriminatory reward seeking when infused into the BLA. However, systemic administration of the D1 receptor agonist did reduce reward seeking behavior during a task that included fear and safety cues. We did not observe a negative impact on reward seeking during systemic administration of a D1 receptor agonist in a task that only included reward cue + sucrose and nonreward cue + no sucrose pairings. This indicates the impairments we saw with the systemically applied agonist in the safety-fear-reward cue discrimination task were more likely due to effects on fear and/or motivation rather than on cue discrimination. Together, our data indicate that altered dopamine D1 receptor activity in the BLA may be a potential mechanism that leads to the impairment in fear suppression to the safety signal seen with PTSD patients.

Working memory develops over the course of adolescence, and neuroimaging studies find development-associated changes in the activity of prefrontal cortical brain regions. Establishment of a rodent model of working memory development would permit more comprehensive studies of the molecular and circuit basis for working memory development in health and disease. Thus, in this study, working memory performance was compared between adolescent and adult male Sprague-Dawley rats using an operant-based, delay-match-to-sample working memory task. Adolescent and adult rats showed similar rates of learning the task and similar performance at a low cognitive load (delays ≤ 6 s). However, when the cognitive load increased, adolescents exhibited impaired working memory performance relative to adults, until postnatal day 50 when performance was not significantly different. Despite evidence that cannabinoids disrupt working memory, we found no effect of acute treatment with the cannabinoid receptor agonist, WIN55212,2, at either age. Moreover, expression of glutamate and GABA receptor subunits was examined in the prelimbic and infralimbic prefrontal cortex across development. NMDA receptor subunit GluN2B expression significantly decreased with age in parallel with improvements in working memory. Thus, we show evidence that rats can be used as a model to study the molecular underpinnings of working memory development.