INTRODUCTION

Drug use often begins during adolescence (Johnston et al, 2010; SAMHSA, 2012), and early onset of drug abuse has been hypothesized to increase the risk of later drug addiction (Kandel et al, 1992; Palmer et al, 2009). One of the greatest challenges to the treatment of drug addiction is vulnerability to relapse (O’Brien, 1997; Chung and Maisto, 2006). In an animal model of relapse, drug-seeking after abstinence from intravenous (i.v.) drug self-administration can be measured as reinstatement of lever-pressing triggered by contextual cues associated with the drug-taking environment (context-induced), discrete drug-paired cues (cue-induced), or reexposure to the drug itself (drug-primed; reviewed in Shaham et al, 2003). Despite reports suggesting that adolescence is a period of heightened vulnerability to drugs (eg, Spear (2000) and Crews et al (2007) for reviews), we have found that male rats that self-administered heroin, morphine, or cocaine as adolescents exhibit attenuated cue-induced reinstatement of drug-seeking after forced abstinence, compared with rats that self-administered in adulthood (Doherty et al, 2009; Li and Frantz, 2009; Doherty and Frantz, 2012a). This result was independently replicated with cocaine-seeking (Anker and Carroll, 2010). It suggests that adolescent rats are less sensitive to some enduring drug effects, an interpretation supported by age differences in several cocaine- and nicotine-related behaviors (Balda et al, 2006; Kerstetter and Kantak, 2007; Shram et al, 2008), as well as fewer signs and symptoms of drug withdrawal among adolescent rodents compared with adults (Doherty and Frantz, 2012b); Doremus et al, 2003; O'Dell et al, 2006; Hodgson et al, 2009).

To identify the substrates involved in this attenuated drug-seeking among adolescent rats, the present study focused on the medial prefrontal cortex (mPFC), a region involved in reinstatement among adult subjects (Kalivas and Volkow, 2005). Stimulus-induced neural activity can be visualized with the immediate early gene product, Fos (Harlan and Garcia, 1998). Fos expression in the context of drug self-administration predicts neural plasticity (Nestler, 2001), and context- or cue-induced reinstatement of heroin-seeking in adult rats elevates mPFC c-fos mRNA (Koya et al, 2006; Kuntz et al, 2008). Thus, the aim of the present study was to test the hypothesis that the extent of mPFC activation during a 1-h heroin-seeking test after forced abstinence from heroin self-administration would be lower among rats that self-administered heroin as adolescents compared with adults.

MATERIALS AND METHODS

Subjects

Male Sprague–Dawley rats (Charles River, Raleigh, NC) arrived at post-natal day (P) 22 (n=12) or P70–74 (n=10) and were housed in groups of 2–3 (12 h light/dark cycle, lights off 0700 hours). Behavioral testing occurred at the same time every day (0900–1400 hours). Body weights were recorded daily. All procedures complied with the NIH Guide for Care and Use of Laboratory Animals (seventh edition, 1998) and were approved by the Institutional Animal Care and Use Committee of Georgia State University.

Drugs

Heroin HCl (gift from NIDA), Brevital (1% methohexital sodium; King Pharmaceuticals, Bristol, TN), and Timentin antibiotic (GlaxoSmithKline, Research Triangle Park, NC) were all dissolved in sterile saline and filtered (0.25-μm membrane) before i.v. administration.

Heroin Self-Administration and Heroin-Seeking Procedures

Equipment

Heroin self-administration was conducted in operant conditioning chambers equipped with two retractable levers (Med Associates, St Albans, VT). Active lever presses delivered infusions via polyethylene tubing to the catheter portal on each subject’s back. The i.v. catheters were made as described (Caine et al, 1993), with a shorter length of tubing for adolescents (Shahbazi et al, 2008).

I.v. catheter implantation

Adolescent (P28–29) and adult (P76–81) rats were anesthetized with an isoflurane–oxygen vapor mixture (1.5–5% ) and were catheterized according to Caine et al (1993) and Shahbazi et al (2008). Rats received the antibiotic timentin (100 mg/ml) and heparinized saline (30 USP U/ml) every day to deter infection and maintain catheter patency (twice daily for 2 days post surgery). Catheter patency was tested weekly with Brevital. If the catheter was defective, the subject was not included in analysis. Acquisition of lever-pressing commenced 5–7 days after surgery.

Self-administration

Adolescent rats were P35 at the start of behavioral testing (body weight 143 g±1.7), and adults were P82–87 (body weight 388 g±5.2). In four treatment groups (adolescent-onset saline, n=6; adolescent-onset heroin, n=6; adult-onset saline, n=5; and adult-onset heroin, n=5), rats spontaneously acquired lever-pressing maintained by heroin (or saline) on a fixed ratio one schedule of reinforcement in 13 sessions over 14 days (0.05 and 0.025 mg/kg per infusion; 3 h daily sessions). Infusions were signaled by a compound stimulus (cue light above active lever, 5-s burst of white noise, house light off). The infusion volume and duration were adjusted daily based on individual body weight (see Doherty and Frantz (2012a) for details).

Heroin-seeking

After 12 days of forced abstinence in the home cages, we tested heroin-seeking for 1 h under extinction conditions. Heroin-seeking was defined as responding on the previously active lever (Shalev et al, 2001). The compound stimulus of discrete drug-paired cues was not presented in this test. Thus, measured behavior could be called ‘extinction responding in the absence of discrete drug-paired cues’. Although reexposure to the drug self-administration environment (context) is likely to be the stimulus driving any lever-pressing we might observe in these conditions, our methods differ from those of Shaham et al (2003) that provide precise tests of context-induced reinstatement that involve daily exposure to an alternative environment during forced abstinence (eg, Bossert et al, 2011).

Immunohistochemistry

Ninety minutes after the start of the heroin-seeking test, rats were anesthetized and intracardially perfused with phosphate buffer (0.1 M), followed by 4% paraformaldehyde. Brains were removed, postfixed in 4% paraformaldehyde, and submerged in 30% phosphate-buffered sucrose before cryosectioning (40 μm, four series). Sections were collected in cryoprotectant and stored at −20 °C.

Fos-ir detection

Brain sections from every experimental group were processed simultaneously in the same multi-well tissue rack to ensure equal treatment. Immunolabeling for Fos occurred on every eighth section through the rostro-caudal extent of the prelimbic and infralimbic mPFC, as well as a portion of the primary motor cortex (M1). We chose M1 as an anatomical control because motor cortex has not been implicated as a region with selective control of drug-seeking. Sections were incubated with anti-Fos antibody for 48 h at 4 °C (1 : 500 rabbit anti-Fos; Santa Cruz Biotechnology, Santa Cruz, CA), followed by biotinylated secondary antibody incubation for 1.5 h (1 : 600; Vector Laboratories, Burlingame, CA). Fos was visualized with an ABC kit followed by cobalt-enhanced 3,3′-diaminobenzidine tetrahydrochloride. Nonspecific staining was uniformly low. Sections were mounted on coded, gelatin-coated slides, dehydrated, counter-stained with cresyl violet, cleared with xylene, and coverslipped with permount.

Stereological Analysis of Fos-ir Neurons

Following Cooke (2011), the total number of Fos-ir+ and Fos-ir− neurons was estimated with the optical fractionator probe in StereoInvestigator (MBF Biosciences, Williston, VT) by a researcher blind to experimental groups. Contours of the prelimbic and infralimbic areas, and a central portion of M1, were traced in one hemisphere at × 2.5 magnification. These contours were used to estimate the regional volume of each area according to the formula:

Where T=average section thickness, x=sectioning ratio, and A1−m=areas of the traced contours. Neurons were counted at × 100 under brightfield illumination using the standard rules with regard to inclusion (Mouton, 2002; see Table 1) on six sections per hemisphere for the prelimbic area (estimated at 4.7, 4.2, 3.7, 3.2, 2.7, and 2.2 mm anterior to bregma; plates 5–10 in Paxinos and Watson (1998)), and three sections for infralimbic and M1 areas (3.2, 2.7, and 2.2 mm anterior to bregma; plates 8–10; Figure 1a). This scheme resulted in an equivalent number of neurons counted in each age group. Fos-ir+ neurons possessed dark-brown cytoplasm and black nuclei, whereas Fos-ir− neurons possessed light-violet clear cytoplasm and dark-violet nuclei (Figure 1b). The estimated number of neurons, E(N) was calculated according to the formula:

Table 1 Parameters for Stereological Analysis of Cortical Neurons
Full size table
Figure 1
figure 1

(a). Neurons were counted in the prelimbic (black), infralimbic (dark gray), and M1 (light gray) cortices. (b). Photomicrographs at × 100 of prelimbic (left) and infralimbic (right) neurons, indicating Fos-ir+ neurons (black arrows), Fos-ir− neurons (gray arrows), and glial cells (gray triangles). Scale bar=10 μm.

PowerPoint slide

Full size image

Where Q−=number of counted cells, ssf=section sampling fraction, asf=area sampling fraction, and tsf=thickness sampling fraction.

Data Analysis

Heroin self-administration

Intake and lever presses were analyzed separately using mixed-measures analysis of variance (ANOVA) with age (adolescent vs adult) and treatment (heroin vs saline self-administration) as between-subjects factors, and session, dose (0.05 vs 0.025 mg/kg per infusion), and lever (active vs inactive) as within-subjects repeated measures. Total heroin intake was compared across age groups using t-tests.

Heroin-seeking

Lever presses during the 1-h heroin-seeking test were analyzed using a three-way ANOVA, with age and treatment as between-subjects factors, and lever as the within-subjects factor. In addition, responding on the active lever in heroin-treatment groups was expressed relative to active lever responding in age-matched saline controls (fold change from saline), then compared with t-tests.

Fos-ir

Estimates of total neuron number in prelimbic, infralimbic, and M1 were analyzed separately using two-way ANOVAs, with age and treatment as between-subjects factors. The percentage of Fos-ir+ to total neurons in each brain region was calculated and tested for normality. Distributions were non-normal in IL and M1, so these data were log transformed before further analysis, which consisted of two-way ANOVAs, with age and treatment as between-subjects factors. Fos-ir expression across the rostro-caudal axis was compared by calculating the fold change in Fos-ir+ expression in heroin- vs saline-treated rats. A two-way ANOVA tested for differences across age (between-subjects factor) and bregma level (within-subjects repeated measure).

To compare neuronal activation across brain regions, data from all brain regions were log transformed and fold change of Fos-ir+ from saline was analyzed using a two-way mixed-measures ANOVA, with age as a between-subjects factor and brain region as a within-subjects factor, at the three rostro-caudal levels at which all three brain regions were analyzed. Data were analyzed using SPSS (SPSS, Chicago, IL) and expressed as means±SEM. In all cases, follow-up ANOVAs and post-hoc tests were conducted as appropriate, with P<0.05 considered significant.

RESULTS

Heroin self-administration

Adolescent and adult rats self-administered similar amounts of heroin, with similar patterns of acquisition (Figure 2). As expected, the number of infusions per session differed according to drug treatment (F1,18=24.7, P<0.001), session (F12,216=6.4, P<0.001), and a treatment-by-session interaction (F12,216=14.8, P<0.001), but there was no main effect of age nor interactions with age. Heroin intake increased gradually across sessions, and rats self-administering heroin earned more infusions than those receiving saline, especially in the second week. Regardless of age, when the dose per infusion was decreased, heroin infusions per session increased (main effect of dose, F1,9=38.2, P<0.001). Furthermore, heroin intake (mg/kg) did not differ by age (Figure 2, inset).

Figure 2
figure 2

Infusions of heroin or saline on a fixed ratio one schedule of reinforcement. Inset: total heroin intake (mg/kg) by adolescent (Adol) or adult rats. All points and bars represent mean±SEM.

PowerPoint slide

Full size image

Heroin-seeking after forced abstinence

Rats that self-administered heroin as adolescents (adolescent-onset group) exhibited less heroin-seeking after abstinence than adults, although all rats that self-administered heroin exhibited more lever-pressing after abstinence than their saline-treated counterparts (Figure 3). A three-way ANOVA confirmed that lever-pressing differed according to treatment (F1,18=39.02, P<0.001) and lever (F1,18=21.83, P<0.001), and that adolescent-onset groups pressed the active lever less than adults (age-by-lever interaction, F1,18=7.47, P<0.05). A follow-up two-way ANOVA on active lever-pressing confirmed a significant age-by-treatment interaction (F1,18=5.87, P<0.05); and among those rats that self-administered heroin, the adolescent-onset group pressed fewer times on the active lever than adults (P<0.05), and heroin-experienced rats pressed more than saline controls (adolescent onset: P<0.01; adult onset: P<0.001). When active lever-pressing was analyzed as a fold change among heroin groups vs saline controls, relative responding was attenuated in adolescent-onset compared with adult-onset groups (P<0.001; Figure 3, inset).

Figure 3
figure 3

Active and inactive lever presses during a 1-h heroin-seeking test. Heroin-treated rats pressed on the active lever more than the inactive lever (##P<0.01; ###P<0.001), and more on the active lever than their saline-treated counterparts. The adolescent-onset heroin group pressed less on the active lever than the adult-onset heroin group (*P<0.05). Inset: relative heroin-seeking was attenuated in the adolescent-onset groups, compared with adults (***P<0.001). For adolescent saline, adolescent heroin, adult saline, and adult heroin-treatment groups, n=6, 6, 5, and 5, respectively. All bars represent mean±SEM.

PowerPoint slide

Full size image

Prelimbic Fos-ir

Heroin-seeking was associated with a greater percentage of Fos-ir+ neurons in the prelimbic cortex of adult-onset groups compared with their saline-treated counterparts, but no difference in FOS expression was observed in adolescence-onset heroin groups vs controls. These results were supported by a two-way ANOVA on the percentage of prelimbic Fos-ir+ to total neurons, showing a significant interaction of age-by-treatment (F1,18=7.1, P<0.05); post-hoc Tukey’s pairwise comparisons confirmed a significant treatment effect in the adult-onset groups (P<0.01). When the percentage of Fos-ir+ neurons to total was expressed as a fold change from saline, the adult-onset heroin group exhibited more Fos-ir+ neurons than the adolescent-onset heroin group, particularly toward the caudal pole of prelimbic area (Figure 4g; age-by-rostro-caudal level interaction, F5,45=3.13, P<0.05), and follow-up independent samples t-tests (with Bonferroni’s correction) at each level confirmed that the adult-onset group had more Fos-ir+ neurons at levels estimated to be 3.7–2.2 mm anterior to bregma (P<0.05).

Figure 4
figure 4

(a–c). Neuron number in prelimbic, infralimbic, and M1 cortices. Adolescent-onset groups (P61) had more neurons than adult-onset groups (P107–112) in infralimbic, but not prelimbic or M1. (d–f). The adolescent-onset heroin group had a lower percentage of Fos-ir+ neurons than adults, in prelimbic (d) and infralimbic (e) areas (*). In contrast, the adult-onset heroin group had more Fos-ir+ neurons than saline controls in prelimbic (d) and infralimbic (e) areas (#). (g–i): Fold change in relative number of Fos-ir+ neurons in heroin-seeking rats compared with their saline controls in prelimbic, infralimbic, and M1 areas, expressed as a function of rostro-caudal level, with less Fos labeling in the adolescent-onset group than adults (*). Also in the infralimbic area of adult rats (h), the number of Fos-ir+ neurons was higher at +2.7 bregma than the other rostro-caudal levels (##). For adolescent saline, adolescent heroin, adult saline, and adult heroin-treatment groups, n=6, 6, 5, and 5, respectively. All bars and points in a–i represent mean±SEM (*, #P<0.05; **, ##P<0.01; ***P<0.001).

PowerPoint slide

Full size image

Infralimbic Fos-ir

Regardless of drug treatment, the total estimated number of neurons (Fos-ir+ plus Fos-ir−) in the infralimbic area was higher for adolescent-onset compared with adult-onset groups (main effect of age: F1,18=16.56, P<0.001; Figure 4b). As in the prelimbic cortex, more neurons were Fos-ir+ in the infralimbic area of the adult-onset heroin group compared with their saline counterparts, an effect not observed in the adolescent-onset groups (log-transformed data; Figure 4e). This was confirmed by a significant interaction of age-by-treatment (F1,18=7.60, P<0.05), leading to a significant age effect within groups that self-administered heroin (P<0.01), and a treatment effect within adult-onset groups (P<0.05). When the percentage of Fos-ir+ neurons to total was expressed as a fold change from saline, a significant interaction of age-by-rostro-caudal level (F2,18=4.37, P<0.05) led to follow-up independent-sample t-tests (with Bonferroni’s correction) at each rostro-caudal level to confirm that the adult-onset group had more Fos-ir+ neurons at each level (P<0.05; Figure 4h). Within the adult group, a one-way ANOVA on rostro-caudal level was significant (F2,8=29.71, P<0.001), and Bonferroi-corrected paired-samples t-tests confirmed that Fos-ir+ was greater at +2.7 than at other levels (P<0.01).

M1 Fos-ir

The log-transformed percentage of Fos-ir+ to total neurons in M1 did not differ by age group or drug treatment (Figure 4c,f). It follows that the percentage of Fos-ir+ neurons to total, expressed relative to saline counterparts, did not differ by age, although one rat in the adolescent heroin group had more Fos-ir+ neurons than his counterparts, thereby raising the group means at rostro-caudal levels estimated to be at 3.2 and 2.2 mm anterior to bregma (Figure 4i).

Comparison of Fos-ir across brain regions

With the percentage of Fos-ir+ neurons to total among heroin groups expressed as a fold change from saline (and log transformed), neuronal activation was compared across brain regions at levels estimated to be 3.2, 2.7, or 2.2 mm anterior to bregma (Figure 4g, h). At approximately +3.2, differences across age and brain region were observed, such that a significant age-by-brain region interaction occurred (F2,18=4.55, P=0.025). Although neuronal activation among adolescents did not differ by brain region, differences occurred in adults, such that a follow-up one-way ANOVA across brain regions was significant (F2,8=10.30, P=0.006). Specifically, IL activation was greater than M1 activation, according to follow-up t-tests with Bonferroni’s correction (P<0.05). At +2.7, differences across age and brain region were also observed, such that a two-way age-by-brain region ANOVA included significant main effects of age (F1,9=13.29, P<0.01) and brain region (F2,18=12.29, P<0.001), as well as a significant interaction (F2,18=5.90, P<0.05). As at +3.2, neuronal activation did not differ across brain regions in adolescents, but a follow-up one-way ANOVA across brain regions was significant for adults (F2,8=10.54, P<0.01). Specifically, activation was greater in the infralimbic region than prelimbic or M1 areas, according to t-tests with Bonferroni’s correction (P<0.05). At approximately +2.2, no age or brain region differences were observed.

DISCUSSION

Despite similar levels of heroin self-administration by adolescent and adult male rats, the younger age group showed less heroin-seeking after 12 days of forced abstinence. This adolescent-onset group also failed to show significant neural activation in the prelimbic or infralimbic mPFC during the heroin-seeking test, whereas the adult-onset heroin self-administration group showed two to six times more Fos-ir+ neurons than their saline counterparts in both mPFC subregions. These results suggest that activation of the mPFC may mediate drug-seeking in rats that initiated self-administration during adulthood, but not adolescence. In addition, the estimated total number of neurons in the infralimbic cortex was greater in younger vs older subjects, regardless of drug treatment.

In the present experiment, both adolescent and adult rats acquired and maintained heroin self-administration. Relatively stable numbers of infusions earned by both adolescents and adults in heroin self-administration groups contrast with the gradual decline in infusions earned by saline control groups. Moreover, when the heroin dose was reduced by half, the number of infusions earned almost doubled, confirming that heroin reinforced lever-pressing. Although no age differences in the number of heroin infusions per session were observed in the present experiment, we reported slightly higher rates of heroin self-administration among adolescents compared with adults in a previous experiment (Doherty and Frantz, 2012a). Inconsistent patterns of self-administration have prompted us to question the reliability of age differences in the acutely reinforcing effects of opiates in the self-administration procedure (Doherty and Frantz, 2012a).

What has remained consistent across our studies of heroin, morphine, and cocaine self-administration, however, is that male rats that initiate drug self-administration during adolescence exhibit attenuated drug-seeking after forced abstinence compared with rats that initiate self-administration during adulthood (Doherty et al, 2009; Li and Frantz, 2009; Doherty and Frantz, 2012a). Our studies reveal attenuated responsivity among younger rats on measures of extinction responding in the absence of discrete drug-paired cues (as in the present study), and cue-induced reinstatement after extinction (see references above). Given that the compound stimulus of discrete drug-paired cues was not included in the present test of drug-seeking, reexposure to the drug self-administration environment (context) is likely the stimulus driving the lever-pressing we observed. These results are consistent with less robust cue-induced reinstatement of cocaine-seeking observed over a shorter experimental time course (Anker and Carroll, 2010); and with reports of faster extinction and less reinstatement of a cocaine-conditioned place preference among adolescents vs adults (Balda et al, 2006, but see Brenhouse et al, 2008), as well as faster extinction of nicotine-seeking (Shram et al, 2008) and less nicotine-paired cue-conditioning among adolescent vs adult rats (Schochet et al, 2004). Cognitive impairment on an amydgala-dependent task after cocaine intake (but not an orbitofrontal cortex-dependent task) is also less severe in younger rats (Harvey et al, 2009; Kerstetter and Kantak, 2007). In further support of reduced enduring effects of heroin in adolescent rats compared with adults, somatic and locomotor signs of spontaneous withdrawal from repeated systemic heroin injections are attenuated, as are the effects of heroin on body weight and food intake (Doherty and Frantz, 2012b). Similarly, fewer affective and/or somatic signs of withdrawal from heroin, morphine, nicotine, or alcohol adolescent rodents as compared with adults (eg, Doherty and Frantz, 2012b; Doremus et al, 2003; O'Dell et al, 2006; Hodgson et al, 2009). Although several studies suggest greater sensitivity among adolescent subjects (or adolescent-onset groups), a growing body of literature suggests that adolescence may be associated with protection from some acute and enduring effects of drug exposure under certain conditions. The parameters that define those conditions, however, remain to be determined.

Given the present replication of the finding that adolescent-onset groups exhibit less robust heroin-seeking after forced abstinence than adults, we went on to examine neural activation in the mPFC. Using Fos immunohistochemistry to mark activated neurons, and stereological techniques to estimate total numbers of Fos-ir+ and Fos-ir− neurons, we confirmed earlier reports that overall neuron number declines with age in the ventral mPFC (Andersen et al, 2000; Markham et al, 2007). Reductions in both cell number and cortical volume occur during the transition from adolescence into adulthood in rodents, non-human primates, and humans (Sowell et al, 1999; Andersen et al, 2000; Markham et al, 2007). Our differentiation of the mPFC into prelimbic and infralimbic regions revealed that the reduction in ventral mPFC cell number between P35 and P90 reported by Markham et al, (2007) may be specific to the infralimbic cortex, and that it does not occur until after P61. Anatomical changes during adolescence correlate with emergence of adult-like executive and decision-making behaviors (Ernst and Fudge, 2009), but their role in drug-seeking, if any, remains to be determined.

Our observation that Fos activation was greater in the infralimbic than prelimbic cortex of adult rats during heroin-seeking is consistent with a greater role for the infralimbic than prelimbic region in heroin-seeking (Van den Oever et al, 2008; Bossert et al, 2011), but contradicts theories based mainly on cocaine-related behavior that the infralimbic cortex inhibits drug-seeking, whereas the prelimbic cortex promotes it (Peters et al, 2009). For both subregions, we also observed activation in only a small percentage of neurons during heroin-seeking, similar to the 5–7% activation reported previously (Bossert et al, 2011). Together with the aforementioned confirmation of age-related declines in neuron number, these similarities across data sets help to validate the present methodology.

The major finding of the present report is an attenuated mPFC activation during heroin-seeking among rats that self-administered heroin as adolescents, compared with adults. The percentage of Fos-ir+ neurons in the adolescent-onset heroin group did not differ from age-matched saline controls, whereas the adult-onset heroin group exhibited a two- to six-fold increase in Fos-ir+ neurons relative to their controls. The most robust effects occurred at approximately 2.7 mm anterior to bregma. Lower levels of drug-related Fos-ir in adolescents vs adults differs from reports that experimenter-administered nicotine or cocaine acutely elevates c-fos mRNA expression in the mPFC more among younger than older male rats (Schochet et al, 2005; Shram et al, 2007; Cao et al, 2007; Caster and Kuhn, 2009), but self-administration often produces results different from experimenter administration (eg, Robinson et al, 2002). Our results suggest that the mPFC may have a key role in age-dependent enduring effects of heroin self-administration in male rats.

In fact, several lines of evidence support a critical role for the mPFC in age-dependent drug-related behavior. For example, targeted inhibition of Fos-activated neurons in either the prelimbic or infralimbic region reduces heroin-seeking after forced abstinence in adult male rats (Bossert et al, 2011), showing that activation of these neurons is a cause and not a consequence of drug-seeking. Furthermore, the mPFC undergoes significant plasticity during adolescence, making it a good candidate to mediate behaviors that change over the same developmental phase (eg, Spear, 2000; Adriani and Laviola, 2004; Lenroot and Giedd, 2006). Yet, much remains to be addressed in this line of investigation. Although both glutamatergic and GABAergic cells in prelimbic and infralimbic cortex are implicated in heroin-seeking by adult rats (Bossert et al, 2011), double-labeling experiments are necessary to delineate which cell types are differentially activated across age groups, ie, which cells are involved in the enduring motivational salience of drug-associated cues after self-administration during adolescence. Significant changes occur in both glutamatergic and GABAergic cells during adolescence (eg, Andersen et al, 2000; Tseng and O'Donnell, 2007; Brenhouse et al, 2008), potentially contributing to phenomena such as reduced cortical inhibition and coordination in adolescents compared with adults (Sturman and Moghaddam, 2011). Some cells altered by drug intake during adolescence could undergo apoptosis during the transition from adolescence to adulthood, thereby minimizing the enduring effects of early drug exposure. Along these lines, D1 dopamine receptors expressed on glutamatergic projection neurons in the mPFC appear to process the salience of drug-associated contextual cues during reinstatement tests (Brenhouse et al, 2010); thus, a natural decline in D1 receptor expression after adolescence (Andersen et al, 2000; Brenhouse et al, 2008) could also contribute to a loss of salience of drug-associated cues, be they contextual cues or discrete cues paired with drug self-infusions. Resolution between studies like ours demonstrating adolescent resistance to drug effects vs those demonstrating heightened drug sensitivity among adolescents may relate to the duration of drug exposure (eg, whether it ceases during adolescence or continues into adulthood), the intensity of drug exposure (eg, drug doses), and/or the parameters for testing drug effects. For example, we observed similar rates of reinstatement among adolescent- and adult-onset groups when both context and cues were renewed simultaneously (Doherty and Frantz, 2012a). Also, extinction of a cocaine-conditioned place preference occurred at the same rate in adolescent and adult rats when explicit extinction procedures were used (Brenhouse et al, 2010), and the age-dependent effects of cocaine self-administration on cognition vary by which brain region mediates the cognitive task (Kerstetter and Kantak, 2007; Harvey et al, 2009). These latter studies underscore the idea that different mechanisms may mediate extinction responding, context-induced reinstatement, and cue-induced reinstatement of drug-seeking. They may also provide experimental conditions in which to test for these different mechanisms.

Overall, attenuated neuronal activity in the prelimbic and infralimbic areas of the mPFC during heroin-seeking in the adolescent-onset group supports and extends our previous reports on attenuated enduring effects of heroin, morphine, and cocaine on drug-seeking after forced abstinence among young rats (Doherty et al, 2009; Li and Frantz, 2009; Doherty and Frantz, 2012a). Although translation of results from rodents to humans requires extensive research, adolescent male rats may provide a natural model from which neuroprotective factors against long-term drug effects can be identified, thereby providing new targets for the treatment of drug abuse in humans.