Anxiety disorders are the most frequent mental illness experienced by adolescents worldwide (Polanczyk et al., 2015). Exposure therapy for anxiety is based on the principle of extinction, in which fear to a stimulus can be reduced by repeated presentations of that stimulus without an aversive outcome. Adolescents attain poorer outcomes following exposure therapy for anxiety compared with children (Southam-Gerow et al., 2001; Bodden et al., 2008). Consistent with this, studies in humans and rodents showed that extinction is impaired in adolescents compared to adults and juveniles/children (Kim et al., 2011; Pattwell et al., 2012; Baker and Richardson, 2015). Specifically, extinction memory is “forgotten” in adolescent rodents that show high freezing when tested in the same context as extinction, compared to adult rodents. It has therefore been suggested that adolescent resistance to exposure-based therapies relates to deficits in extinction learning at this age (Hartley and Casey, 2013; Kim and Ganella, 2015).

The prefrontal cortex (PFC) is a critical neural region underlying extinction deficits in adolescent rodents (Kim et al., 2011; Pattwell et al., 2012; Baker and Richardson, 2015). This is not surprising, as the PFC, especially the infralimbic cortex (IL), has repeatedly been identified to be particularly important for the recall of extinction across different learning paradigms and different ages (Quirk et al., 2000; Laurent and Westbrook, 2008; Peters et al., 2009; Mueller et al., 2010; Orsini et al., 2011; Sierra-Mercado et al., 2011; Gass and Chandler, 2013; Abraham et al., 2014). To further understand adolescent vulnerability to anxiety disorders, we aim to explore the role of prefrontal dopamine 1 receptor (D1R) and dopamine 2 receptor (D2R) in fear extinction during adolescence compared to adulthood. Dopamine signaling displays a unique maturation profile during adolescence, over and above many other neurotransmitter systems in the PFC (O’Donnell, 2010; Wahlstrom et al., 2010; Kim et al., 2016). The density of dopaminergic fiber infiltration of the PFC increases throughout adolescence until early adulthood in rodents (Kalsbeek et al., 1988) and non-human primates (Rosenberg and Lewis, 1995). Dopamine synthesis also peaks in the PFC during adolescence (Andersen et al., 1997), along with dopamine receptor density in the PFC (Tarazi and Baldessarini, 2000). Importantly, age-related discontinuities in the function of the dopamine system have been suggested to underlie the adolescent behavioral phenotype observed across human and non-human mammals (Laviola et al., 2003). This includes differences in impulsivity and cue reactivity, behavioral responses critically implicated in extinction learning (Laviola et al., 2003; Pattwell et al., 2013).

Dopamine exerts its effects via five distinct receptors, which are subdivided into two families: D1-like and D2-like receptors (Andersen et al., 1990). The D1-like subfamily comprises D1R and D5R, and the D2-like includes D2R, D3R and D4R (Missale et al., 1998). The most abundant dopamine receptor subtypes in the central nervous system are D1R and D2R (Jaber et al., 1996), with both showing expression in the medial prefrontal cortex (mPFC; Vincent et al., 1993). As members of different subfamilies, D1R and D2R show distinct profiles in terms of downstream signal transduction and physiological effects (Jackson and Westlind-Danielsson, 1994; Beaulieu and Gainetdinov, 2011).

Findings in adult rats show complex involvement of prefrontal D1R and D2R signaling in fear extinction. For example, infusion of the D1R antagonist SCH-23390 into the IL of PFC impairs long-term fear extinction (Hikind and Maroun, 2008). Consistent with this, transgenic mice lacking D1R show normal fear conditioning but delayed extinction up to 90 days post-conditioning (El-Ghundi et al., 2001). By comparison, the role of D2R signaling in extinction is less clear. For instance, one study showed that pre-extinction systemic treatment with the D2R agonist quinpirole blocked extinction of conditioned fear (Nader and LeDoux, 1999), while another showed largely no effect across a range of doses (0.25, 0.5, 2.0 mg/kg), though one dose (1.0 mg/kg) impaired long-term extinction (Ponnusamy et al., 2005). However, pre-extinction systemic or intracerebroventricular (i.c.v.) injection of the D2R antagonist haloperidol has also been found to increase conditioned stimulus (CS)-elicited freezing during extinction and at test the next day (Holtzman-Assif et al., 2010). However, systemic D2R antagonism with sulpiride facilitates extinction both within-session and at test the next day (Ponnusamy et al., 2005), while one known study investigating intra-IL D2R antagonism using raclopride found impaired long-term fear extinction in adult rats the next day (Mueller et al., 2010).

Interestingly, findings on the expression of D1R and D2R in the PFC across adolescence are also varied. For example, D1R and D2R density in the PFC has been reported to be high at postnatal day (P) 40, then decline into adulthood across P60, P80, P100 and P120, with D1R declining more dramatically compared to D2R (Andersen et al., 2000). This is consistent with positron emission tomography (PET) findings in humans age 10–30 years, which show that D1R binding in the PFC decreases from adolescence into adulthood (Jucaite et al., 2010). Another human study reports a peak in D1R gene expression during adolescence compared to infancy and adulthood, with no changes in D2R gene expression across those ages (Weickert et al., 2007). In contrast, studies have also found no change in D1R gene expression or binding in PFC from P21 to P60 in rodents (Leslie et al., 1991; Tarazi et al., 1999; Tarazi and Baldessarini, 2000; Araki et al., 2007). A human study found no significant difference in PFC D1R gene expression between adolescents and adults (Rothmond et al., 2012). A study that used microarrays showed no changes relating to dopamine signaling, including D1R and D2R expression, across subjects aged 0–49 years (Harris et al., 2009). Taken together, further investigation into the changes in PFC D1R and D2R expression in adolescence and adulthood appear warranted, especially in light of the emerging role of PFC dopamine signaling in fear extinction.

To investigate the role of prefrontal dopamine signaling in adolescent vs. adult extinction, we first examined fear conditioning and extinction in adolescent and adult rats. Using real-time quantitative PCR (qPCR) analyses, we then measured extinction-related changes in mPFC D1R and D2R gene expression in adolescent and adult rats. Building on these results, we then acutely enhanced D1R or D2R signaling in IL during extinction using SKF-81297 or quinpirole, respectively. Dopamine receptor agonists were chosen so that parallels can be made to existing FDA-approved dopamine receptor agonists, which are more readily administered compared to dopamine receptor antagonists in adolescent humans (Kirino, 2012; Kim and Lawrence, 2014). Thus, results from the present chapter have strong translational potential for improving extinction-based treatments for anxiety, as well as adding to literature on the mechanisms of extinction across development.

Here we show that adolescent rats display a deficit in long-term extinction of a conditioned fear response compared to adult rats. We also showed that D1R/D2R ratio is decreased following extinction in adolescents and increased in adults. Further, enhancing IL D2R signaling using quinpirole improved long-term extinction in adolescent rats but delayed extinction acquisition in adult rats, while increasing IL D1R signaling using SKF-81297 had no effects at any age. Present findings further highlight that adolescent extinction impairments relate to developmental changes in mPFC function, and identify for the first time that maturation of PFC dopamine signaling plays a role.

Our behavioral findings are consistent with previous studies that report impaired fear extinction in adolescents compared to adults in both rodents and humans (McCallum et al., 2010; Kim et al., 2011; Pattwell et al., 2012; Baker and Richardson, 2015). These data add to a growing literature suggesting that adolescence is characterized by impairments in cue extinction more broadly, since adolescent rats also display deficits in extinction of a cocaine-associated context (Brenhouse and Andersen, 2008) and a cocaine-associated cue (Zbukvic et al., 2016). Importantly, the present findings recapitulate clinical evidence that extinction-based therapy for anxiety disorders is less effective in adolescents compared to other ages (Southam-Gerow et al., 2001; Bodden et al., 2008).

Prior to extinction, adolescents displayed increased D1R and D1R/D2R ratio mRNA compared to adults. This is consistent with studies that report a peak in D1R gene expression (Rothmond et al., 2012; Garske et al., 2013) and receptor expression (Andersen et al., 2000; Brenhouse et al., 2008) in the PFC during adolescence. In particular, previous findings indicate that early life adversity can exacerbate an adolescent peak in D1R expression on PFC projection neurons in rats (Brenhouse et al., 2013). Therefore, the increased adolescent D1R/D2R ratio observed in the present study may be a result of fear conditioning. By comparison, we found no age difference in D2R gene expression, in line with reports that D2R expression in the mPFC reaches stable adult levels by adolescence (Tarazi et al., 1998).

Since patterns of basal D1R and D2R mRNA expression in the cortex are found to correlate with receptor binding (Weiner et al., 1991), the current findings imply a markedly different prefrontal dopaminergic environment pre-extinction depending on age, with adolescent mPFC networks likely dominated by D1R activity relative to D2R activity compared to adults. Computational modeling predicts that when the PFC is dominated by D1R relative to D2R signaling, this produces a state of net inhibition (Seamans and Yang, 2004). Notably, the present findings in adolescents are similar to reports of rats with lesions of the mPFC, where fear conditioning and within-session extinction learning are intact but long-term extinction is impaired (Quirk et al., 2000; Garcia et al., 2006). Moreover, it appears that the mPFC is not recruited as efficiently during fear extinction in adolescence compared to adulthood (Kim et al., 2011; Pattwell et al., 2012; Baker and Richardson, 2015). In humans, the intense emotionality of adolescents is thought be at least partly due to an under-recruitment of the PFC (Somerville et al., 2010). The present findings suggest that the mPFC dopaminergic profile may contribute to adolescent emotionality.

Since we observed age differences in prefrontal dopamine receptor gene expression following extinction learning, we then sought to examine functional differences in dopamine signaling across adolescence vs. adulthood. Therefore, we investigated the immediate and long-term effect of enhancing D1R or D2R signaling in the IL of the PFC at the time of extinction learning. This subregion of the PFC is homologous to Brodmann Area 25 in the human brain (Gass and Chandler, 2013). Critically, both the IL of rodents and the corresponding ventromedial PFC in humans have been strongly implicated in extinction learning and retrieval, whereas the prelimbic subregion (PrL) is involved in fear expression (Quirk et al., 2000; Phelps et al., 2004; Sierra-Mercado et al., 2011). Changes in dopamine receptor gene expression following extinction, and the opposite effects observed in adolescent vs. adult rats following the manipulation of prefrontal D2R activity suggest a fundamental dissociation in dopaminergic signaling in the mPFC in relation to extinction across adolescence. Adolescent data are consistent with previous findings that intra-IL quinpirole improves extinction of a discrete cocaine-associated cue in adolescent rats (Zbukvic et al., 2016), suggesting a role for prefrontal D2R signaling across extinction learning broadly. In contrast, intra-IL quinpirole delayed the acquisition of extinction in adult rats, with no effect on long-term extinction. This is consistent with previous reports of systemic quinpirole delaying extinction in adult rats (Nader and LeDoux, 1999; Ponnusamy et al., 2005). However, systemic or ICV treatment with the D2R antagonist haloperidol (Holtzman-Assif et al., 2010), or intra-IL or systemic treatment with the D2R antagonist raclopride (Mueller et al., 2010) at the time of extinction have been shown to impair the retrieval of extinction the next day in adult rats. Disparities between findings may be due to the specificity of agonists and antagonists used, and/or route of administration (Gehlert et al., 1992; Bowery et al., 1994; Tseng and O’Donnell, 2007). Present findings suggest that the level of D2R signaling in the adult IL may be naturally optimal compared to adolescents for extinction learning, and further activation by quinpirole might disrupt this balance. Overall, divergent effects of intra-IL quinpirole add to gene expression data suggesting that that the IL of the PFC is involved in extinction learning during adolescence as well as adulthood, however that extinction may involve differential D2R signaling across development. In contrast, we found the D1R agonist SKF-81297 had no effects on within-session or long-term extinction in either adolescent or adult rats. Previous studies using adult rodents report that attenuating IL D1R signaling impairs extinction (Hikind and Maroun, 2008; Fricks-Gleason et al., 2012). By comparison, systemic treatment with a D1R agonist enhanced extinction of cued and contextual fear in adult rats, though the anatomical targets of that effect were not clear (Abraham et al., 2016). It may be that enhancing IL D1R activity is not sufficient for extinction in adolescence and adulthood.

We note that in the final experiment of the present study, we observed a small spontaneous recovery of extinguished freezing 24 h following extinction in adult rats that received saline infusion into the IL before extinction. This was not observed in our first experiment, as well as in our and other groups’ previous studies (Quirk, 2002; McCallum et al., 2010; Kim et al., 2011; Orsini et al., 2011). A careful examination of the literature revealed that pre-extinction infusion of saline or vehicle into the IL or the PrL, but not other brain regions, may cause this effect in adult rats (Sierra-Mercado et al., 2011). This small but significant spontaneous recovery due to pre-extinction vehicle infusion appears to also be present even when the freezing was well extinguished to baseline (i.e., ~0%), and when vehicle has also been infused during test to provide identical physiological contexts for extinction and test (Laurent and Westbrook, 2008). In studies that do not involves such saline or vehicle infusions into the IL, adult rodents maintain low levels of freezing following extinction (Quirk, 2002; McCallum et al., 2010; Kim et al., 2011; Orsini et al., 2011). Therefore, we believe that the result of our final experiment does not affect the interpretation of our data. Additionally, there was a significant effect of age during conditioning for qPCR experiment, whereas there was only a trend in the behavioral experiment. Based on the results of microinfusion experiments, we believe that adolescent rats showed reduced freezing during conditioning compared to adults in the present study. While these results are inconsistent with previous findings (McCallum et al., 2010; Kim et al., 2011), there are more recent studies that do not report conditioning data in extinction across adolescence (Pattwell et al., 2012, 2016). Importantly, freezing levels during extinction were comparable in adolescent and adults in all experiments of the present study.

Further studies are required to confirm age differences in D1R vs. D2R protein expression following fear conditioning and extinction, although antibodies that are specific to D1R or D2R, and not any other dopamine receptor subtypes are notoriously lacking. Specifically, D1R and D2R proteins show striking structural similarities, which have implications for visualization and quantification in brain tissue. For instance, sequence similarity searching using the Basic Local Alignment Search Tool (BLAST) database reveals that D1R and D2R share 77% of their amino acid sequence (Agostino, 2012). It follows that D1R and D2R display similar ligand binding profiles (Levey et al., 1993). This means that commercially available antibodies for D1R and D2R are liable to display cross-reactivity (Michel et al., 2009; Hutchings et al., 2010). Immunostaining of D2R in particular has historically shown conflicting results across previous literature, with some studies reporting extensive labeling throughout all layers of cortex (Ariano et al., 1993), while others have shown little to no staining (Levey et al., 1993; Sesack et al., 1994). Therefore, present qPCR findings offer an exciting first step to elucidating how the maturing prefrontal dopamine system may contribute to extinction across adolescence.

Epidemiological data indicate that persistence of mental health problems among adolescents relates more to recurrence rather than chronicity of youth-onset disorders (Kessler et al., 2012). Our data suggest that this may be due, at least in part, to extinction impairments at this age. Not only do we demonstrate that adolescent rats are impaired in long-term extinction, we also show for the first time that extinction produces unique changes dopamine receptor gene expression across development. Importantly, present findings highlight D2R as a promising pharmacological target to improve exposure therapy in adolescents, which is consistent with our recent observation that the partial agonist of the D2R, aripiprazole, can significantly facilitate long-term extinction in adolescent rats (Ganella et al. under review). While behavioral therapies that involve exposure therapy are the most effective way to treat anxiety disorders, less than one in five adolescents have received therapy for their anxiety (Merikangas et al., 2011). We propose that an effective pharmacological adjunct that acutely accompanies behavioral therapy could significantly reduce the amount of treatment necessary during this vulnerable period, and reduce chronic use of medication. Given that neural correlates of adolescent behavior are highly conserved across species (Spear, 2000), present findings represent an important step to developing more effective treatments for adolescents living with anxiety disorders.

JHK and ICZ conceptualized and designed the study. ICZ, CHJP and DEG acquired the data. ICZ and JHK analyzed the data. ICZ, AJL and JHK interpreted the data. All authors were involved in critically revising the work for important intellectual content and in final approval of the version to be published. All authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

This work was supported by an Australian Postgraduate Award awarded to ICZ, Baker Foundation Fellowship awarded to DEG, a Principal Research Fellowship (APP1020737) from the National Health and Medical Research Council (NHMRC) of Australia awarded to AJL, NHMRC Career Development Fellowship (grant number APP1083309) awarded to JHK and NHMRC Project grant (APP1063140) awarded to JHK and AJL. Australian Research Council Discovery Grant (DP150102496) awarded to JHK.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.