Male Sprague-Dawley rats (Charles River Labs, Hollister, CA) arrived at P17 with dams or at P75. Rats were group-housed in an AAALAC-accredited vivarium on a 12 hr light/dark cycle with unlimited access to food and water. Pups were weaned on P21. All experiments were performed during the light cycle and carried out in accordance with the Institutional Animal Care and Use Committee at the University of California, Irvine.

Fluoxetine HCl (Sigma, St. Louis, MO) was dissolved in sterile water. Nicotine hydrogen tartrate (Sigma, St. Louis, MO) was dissolved in sterile saline and adjusted to pH 7.2 – 7.4; doses were calculated as free base. Cocaine HCl (Sigma, St. Louis, MO) (–),-quinpirole HCl (Tocris Bioscience, Bristol, UK), and WAY-100,635 (Sigma, St. Louis, MO) were dissolved in sterile saline. S-15535 (Sigma, St. Louis, MO) was dissolved in sterile saline that was acidified with lactic acid, and pH was adjusted to as close to neutrality as possible (pH > 5.0). L-741,626 (Tocris Bioscience, Bristol, UK) was dissolved in 50% ethanol/saline. Propofol was purchased from Abbot Laboratories (Chicago, IL).

Rats, aged P24 or P82, were anesthetized with equithesin (0.0035 ml/g) and surgically implanted with an indwelling catheter in their jugular vein (37). During 3 days of recovery, catheters were flushed daily with heparinized saline solution (1000 units heparin per 30 ml bacteriostatic saline) to maintain patency. Catheter patency was verified by testing with propofol (5 mg/kg, i.v.) after drug pretreatment, and only animals showing rapid anesthesia were included in analyses.

Adolescent (P28–31) and adult (P86–89) rats received a drug pretreatment for 4 consecutive days (32). Daily intravenous injections of fluoxetine (1 mg/kg), WAY-100,635 (100 µg/kg), fluoxetine + WAY-100,635, nicotine (60 µg/kg), or saline were administered during adolescence or adulthood. Behavioral testing was conducted the following day on either P32 or P90 ( Supplementary Figure 1 ).

One day after pretreatment, at P32 for adolescents and P90 for adults, locomotor activity was recorded in an open field chamber (43.2 × 43.2 × 30.5 cm³) by 16 photobeams along the sides of each wall (MED Associates, Inc., St. Albans, VT). Following 30 min of habituation in the test chamber, rats received an injection of either quinpirole (0.4 mg/kg, i.p.), a selective D₂ and D₃ receptor agonist, or vehicle and returned for 30 min monitoring of horizontal activity. For acute 5-HT_1AR antagonist studies, rats were given a single infusion of WAY-100,635 (100 µg/kg, i.v.), S-15535 (300 µg/kg, i.v.), or vehicle 20 min prior to locomotor testing.

One day after pretreatment, at P32 for adolescents and P90 for adults, acquisition of cocaine self-administration was evaluated in an operant chamber containing two nose poke holes (MED Associates, Inc., St. Albans, VT). Rats self-administered cocaine (0.5 mg/kg, i.v., 20 µl, 1.1s) in a single 1 hr session on a fixed ratio 1 (FR1) schedule. During each infusion, the cue light above the reinforced hole was illuminated; afterward the house light shut off for a 20 sec timeout where responses were counted but had no effect. To control for nonspecific activity, responses for the non-reinforced hole were recorded but had no programmed consequences. For acute 5-HT_1AR antagonist studies, rats were given a single infusion of WAY-100,635 (100 µg/kg, i.v.), L-741,626 (2 mg/kg, i.p.), or vehicle 20 min prior to self-administration testing.

On P32 or P90, 24 hrs after the last pretreatment infusion, brains were collected, rapidly frozen, and stored at -70°C until processing. Coronal sections of 20 µm thickness were cut in a cryostat maintained at -20°C. Sections were mounted onto ice-cold, gelatin-coated glass slides (Fisher Scientific, Waltham, MA). Slides were air dried for 30 min at room temperature and processed for 8-OH-DPAT-stimulated [³⁵S]GTPγS binding.

8-OH-DPAT-stimulated [³⁵S]GTPγS (Perkin Elmer, Boston, MA) autoradiography was as described by Hensler et al. (2007) (38) with slight modification. Slide-mounted sections were equilibrated in assay buffer (50 mM HEPES; 3 mM magnesium perchlorate; 0.2 mM EGTA; 100 mM NaCl; and 0.2 mM dithiothreitol; pH 7.4) for 10 min at room temperature. Sections were pre-incubated in 5’-guanylate diphosphate (2 mM GDP) for 15 min at room temperature, and then incubated for 60 min in HEPES buffer containing GDP (2 mM) and [³⁵S]GTPγS (40 pM), with or without 8-OH-DPAT (1 µM). Basal [³⁵S]GTPγS binding was determined in the absence of 8-OH-DPAT. Nonspecific [³⁵S]GTPγS binding was determined in the presence of unlabeled GTPγS (10 µM). The incubation was stopped by two 5 min washes in ice-cold HEPES buffer (50 mM, pH 7.4), followed by a brief rinse in ice-cold deionized water. Sections were air dried at room temperature for 1 hr and exposed to Kodak Biomax MR film for 48 hrs.

Digitized autoradiograms were quantified with a MicroComputer Imaging Device (MCID Imaging Research, St. Catherine, Ontario, Canada; Supplementary Figure 2 ). Autoradiograms of 8-OH-DPAT-stimulated [³⁵S]GTPγS binding were simultaneously exposed with [¹⁴C] standards (39). Standard calibration curves were generated from [¹⁴C] standards of optical density against radioligand concentration (dpm/mg tissue) and used to transform regional density values into relative radioactivity measures. Nonspecific binding of [³⁵S]GTPγS was subtracted from basal binding and from specific, agonist-stimulated, binding. Specific binding was expressed as percent change from basal. Areas that were analyzed included subregions of the cortex (prelimbic, infralimbic, ventrolateral orbital, cingulate, agranular insular, primary motor, and secondary motor); amygdala (basolateral, central, and medial); striatum (caudate putamen, nucleus accumbens shell and core); hypothalamus (paraventricular nucleus and lateral); dorsal and ventral hippocampus; dorsal and median raphe; bed nucleus of the stria terminalis; periventricular thalamus; and ventral tegmental area.

No statistical methods were used to predetermine sample sizes. Sample sizes used were similar to those reported in previous publications (32–36). Analysis of digitized autoradiograms were performed blind.

Locomotor activity data are expressed as mean ± SEM and were analyzed by 3-way ANOVA comparing Age × Pretreatment × Drug or 2-way ANOVA comparing Pretreatment × Drug. All significant main effects were further analyzed by one-way ANOVA with Bonferroni post hoc comparisons. For acquisition of cocaine or sucrose self-administration, reinforced (R) and non-reinforced (NR) response data are expressed as mean ± SEM and were analyzed by 2-way ANOVA for Pretreatment × R/NR Responses, with repeated measures on Responses. Significant effects of reinforcement were analyzed by Bonferroni-corrected paired t-tests for each pretreatment group. Anatomical data for each brain region are expressed as mean ± SEM and were analyzed separately using 2-way ANOVA comparing Pretreatment × Age. Any significant main effects were further compared using Dunnett’s post hoc test. All analyses were completed using SPSS Statistics software. Outliers (≥2 SD from group mean) were removed prior to analysis.

As shown in Figure 1 fluoxetine pretreatment of adolescents, but not adults, increased quinpirole-induced locomotor activity via a 5-HT_1AR mechanism. Overall effects of Age (F_1,126 = 96.189, p < 0.001), Pretreatment (F_3,126 = 3.255, p < 0.05), and Drug (F_1,126 = 111.913, p < 0.001) were seen, as well as significant interactions of Age × Pretreatment (F_3,126 = 4.696, p < 0.01), Age × Drug (F_1,126 = 88.538, p < 0.001), Pretreatment × Drug (F_3,126 = 4.301, p < 0.01], and Age × Pretreatment × Drug (F_3,126 = 4.937, p < 0.01). Quinpirole induced significant locomotor activity in adolescents (p < 0.001; Figure 1A ), which was enhanced by fluoxetine pretreatment (p < 0.01). Co-administration of the 5-HT_1A receptor antagonist, WAY-100,635, during pretreatment blocked fluoxetine enhancement of quinpirole-induced locomotion in adolescents (p < 0.001). As has been shown previously (33), quinpirole did not induce significant locomotor activity in adults; nor did fluoxetine or WAY-100,635 pretreatment influence locomotor activity in this age group ( Figure 1B ).

To examine the role of endogenous 5-HT in drug reinforcement, we evaluated the effect of fluoxetine pretreatment on subsequent cocaine self-administration in adolescents ( Figure 2 ). A significant effect of Response (F_1,47 = 23.095, p < 0.001) and a Response x Pretreatment interaction (F_3,47 = 14.912, p < 0.001) were observed. Whereas saline-pretreated adolescent rats did not discriminate between reinforced and non-reinforced nosepokes in a novel self-administration chamber, those pretreated with fluoxetine exhibited significantly higher preference for the reinforced hole (p < 0.001) and higher reinforced responses than saline-pretreated controls (p < 0.001). Enhancement of cocaine reinforcement by fluoxetine treatment was blocked by co-administration of WAY-100,635, indicating that the effect was mediated by 5-HT_1A receptors. In contrast, fluoxetine pretreatment had no effect on initial cocaine self-administration in adults ( Supplementary Figure 3 ).

Adolescent treatment with fluoxetine or nicotine induces persistent alteration in 5-HT_1A receptor signaling, even after drug pretreatment has ended. Following adolescent pretreatment with either drug, enhanced quinpirole-induced locomotor activity is observed which is blocked by acute administration of the 5-HT_1A receptor antagonist, WAY-100,635, or partial agonist, S-15535 ( Figure 3 ). A significant overall effect of Antagonist (F_2,131 = 10.195, p < 0.001) and a Pretreatment x Antagonist interaction (F_4,131 = 4.488, p < 0.01) were observed. Pretreatment with nicotine or fluoxetine significantly enhanced quinpirole-induced locomotion in controls (p < 0.001), but not in animals treated acutely with WAY-100,635 or S-15535 ( Figure 3 ).

Following nicotine or fluoxetine pretreatment, enhanced acquisition of cocaine self-administration is seen in adolescent rats, with a significant effect of Pretreatment (F2,62 = 5.748, p < 0.01), Response (F1,62 = 47.807, p < 0.001), and a Response x Pretreatment interaction (F2,62 = 10.882, p < 0.001). Both nicotine and fluoxetine pretreatment significantly increased reinforced responses and preference for the reinforced hole compared to saline-pretreated controls (p < 0.001; Figure 4A ). These pretreatment effects on acquisition of cocaine self-administration were no longer present following acute blockade of 5-HT_1A receptors with WAY-100,635 ( Figure 4B ) or acute blockade of D₂ receptors with L-741,626 ( Figure 4C ). In contrast, pretreatment had no effect on initial sucrose self-administration in adolescents ( Supplementary Figure 4 ).

Given behavioral evidence of changes in 5-HT_1A receptor activity following pretreatment with fluoxetine or nicotine in adolescents, we used a neuroanatomical approach to assess the regional specificity of these functional effects ( Supplementary Table 1 ). With both pretreatments, drug- and region-specific changes were observed in subsequent acute stimulation of [³⁵S]GTPγS binding to G proteins by the 5-HT_1A receptor agonist, 8-OH-DPAT. In the infralimbic cortex (IL), adolescent fluoxetine pretreatment enhanced 5-HT_1A receptor activity, with a significant overall effect of Pretreatment [F_2,52 = 3.754, p<0.05] and an Age × Pretreatment interaction [F_2,52 = 4.630, p<0.05]. Post hoc analysis showed that adolescent rats pretreated with fluoxetine had significantly greater 8-OH-DPAT stimulated [³⁵S]GTPγS binding in the IL compared to saline-pretreated rats (p<0.01; Table 1 ). These effects were not seen in adolescents pretreated with nicotine or in adults.

Adolescent nicotine pretreatment enhanced 5-HT_1A receptor activity in the medial amygdala, with significant overall effects of Pretreatment [F_2,55 = 6.389, p<0.01] and Age [F_1,55 = 5.547, p<0.05]. Post hoc analysis showed that adolescent rats pretreated with nicotine had significantly greater 8-OH-DPAT stimulated [³⁵S]GTPγS binding in this region compared to saline-pretreated rats (p<0.05; Table 1 ). These effects were not seen in adolescents pretreated with fluoxetine or in adults.

The primary motor cortex was the only region to show significant enhancement of 5-HT_1A receptor activity following adolescent pretreatment with either fluoxetine or nicotine, with a significant overall Age × Pretreatment interaction [F_2,54 = 4.919, p<0.05]. Post hoc analysis showed that both nicotine- and fluoxetine-pretreated adolescent rats had significantly higher 8-OH-DPAT stimulated [³⁵S]GTPγS binding than saline-pretreated controls (p<0.05; Table 1 ). Drug pretreatment had no significant effects on 5-HT_1A receptor activity in adult.

Brief nicotine pretreatment during early adolescence has been shown to uniquely enhance DA-mediated behaviors, including cocaine self-administration, conditioned place preference, and locomotor sensitization as well as ambulatory activity induced by the D₂ receptor agonist, quinpirole^,32–36. We have previously shown that adolescent nicotine enhancement of quinpirole locomotion and cocaine self-administration is both mediated by 5-HT release and blocked by co-administration of selective 5-HT_1AR antagonists (33). These findings are consistent with prior reports of unique activation of raphe by adolescent nicotine treatment (44). We now demonstrate that increased 5-HT activation of 5-HT_1ARs, resulting from fluoxetine blockade of transporter function, also increases quinpirole-induced locomotor activity and cocaine self-administration in adolescents but not adults. In the absence of a drug pretreatment, there are no age differences in the acquisition of cocaine self-administration (44), which indicates that endogenous 5-HT signaling regulates adolescent behavioral sensitivity to cocaine.

Treatment with 5-HT_1AR antagonists during behavioral testing eliminates fluoxetine and nicotine pretreatment effects on quinpirole locomotion and cocaine reinforcement. This finding indicates that persistent enhancement of 5-HT_1AR function mediates the behavioral effects of prior adolescent nicotine and fluoxetine exposure even after discontinuation of drug treatment. Since 5-HT_1ARs are located both pre- and post-synaptically (45), we compared the behavioral effects of two antagonists with differing pharmacological profiles to determine which 5-HT_1ARs mediated adolescent fluoxetine and nicotine effects. As an antagonist, WAY-100,635 is equally effective at pre- and post-synaptic 5-HT_1ARs. In contrast, S-15535 is a partial 5-HT_1AR agonist, which acts functionally as an antagonist at postsynaptic receptors, but as an agonist at presynaptic autoreceptors (46). As with WAY-100635, acute administration of S-15535 immediately before behavioral testing blocked nicotine and fluoxetine enhancement of quinpirole-induced locomotion, indicating an involvement of postsynaptic 5-HT_1ARs.

Prior studies have shown that chronic fluoxetine leads to increased 5-HT system responsiveness after drug cessation, possibly because of 5-HT_1AR autoreceptor desensitization (47). However, this effect is seen in both adolescents and adults, and does not result in changes in baseline monoamine levels. It is therefore unlikely to explain the increased postsynaptic 5-HT_1AR sensitivity that we have observed following subchronic fluoxetine treatment in adolescents. Furthermore, using an in vitro approach to examine receptor-G protein coupling, we have shown unique effects of adolescent fluoxetine and nicotine exposure on 5-HT_1AR function.

Fluoxetine enhanced 5-HT_1AR activity in the IL, a prefrontal cortical region important for emotional learning and habit behavior (48, 49). 5-HT_1ARs in this region have been implicated in modulating emotional response to stressful situations (50). In contrast, nicotine enhanced 5-HT_1AR activity in the medial amygdala, a region of the amygdala implicated in control of sexual and social behaviors (51, 52). 5-HT_1ARs in this region have been implicated in innate fear modulation (53). However, both fluoxetine and nicotine both enhanced 5-HT_1AR activity only in the M1, a cortical motor region that experiences continuous growth in serotoninergic innervation until late adolescence (54). The M1 is actively engaged during initial motor consolidation and plays a critical role in acquisition of skill learning (55). 5-HT_1ARs within M1 decrease corticostriatal activity (56) and modulate DA-mediated motor function (57). Since both fluoxetine and nicotine enhanced 5-HT_1ARs activity within this region, it is possible that M1 may be the site of action of these drugs in enhancing adolescent DA-mediated behaviors.

The induction of locomotor activity by quinpirole in adolescents is mediated by D₂ receptors and is age-specific (36). As confirmed in the present study, adults do not show a locomotor response to quinpirole (33, 58), a finding that is consistent with the maturation of DA systems and D₂ receptor function that occurs during adolescence (1, 3, 5). We have previously shown that adolescent nicotine pretreatment enhances quinpirole-induced locomotor activity via a 5-HT_1AR mechanism and increases D₂ receptor expression and functional activity in the striatum (32, 33). Involvement of the D₂ receptor is further indicated by the finding that D2 receptor blockade or knockdown during nicotine pretreatment of adolescents blocks subsequent enhancement of cocaine self-administration via a microglia-specific mechanism (33). We now show that fluoxetine pretreatment enhances adolescent locomotor response to quinpirole, indicating that increased serotonergic signaling during pretreatment potentiates subsequent D₂ receptor function in this age group. Adolescent nicotine and fluoxetine enhancement of cocaine self-administration is also mediated by increased functional activity of D₂ receptors, since the effect is blocked by treatment with a D₂-specific antagonist, L-741,626, during self-administration testing. The finding that cocaine self-administration in adolescents is mediated by D₂ receptors is in marked contrast to adults, where D₁ receptors mediate cocaine reinforcement (59). Indeed, D₂ receptors do not mediate cocaine self-administration in adults but are, instead, involved in mechanisms that limit intake of high-dose cocaine (60).

As shown in the present study, fluoxetine and nicotine enhancement of DA-mediated behaviors in adolescents are mediated by functional sensitization of both 5-HT_1AR and D₂ receptors. Enhanced behavioral response following pretreatment with either drug is eliminated by administration of either 5-HT_1AR or D₂ receptor antagonists immediately prior to testing. This suggests that there is a close interaction of 5-HT_1AR and D₂ receptors in adolescent brain that modulates reward circuitry. The mechanisms underlying such coupling of receptor activity are presently unknown but may either involve direct receptor interaction or a circuit based mechanism. Further studies will be required to evaluate the exact nature of this linked receptor response.

The findings in this report are subject to at least three limitations. First, this study was conducted in males since we have previously shown that adolescent nicotine enhancement of cocaine self-administration is seen in both males and females (32, 33). However, further studies will be required to determine whether this is also the case for fluoxetine in females. Second, pretreatment with nicotine does not fully capture the myriad of other psychoactive constituents that are present in tobacco leaves, combustible tobacco products, or e-cigarette vaping liquids, and further investigations are needed in order to characterize how these other psychoactive constituents interact with nicotine and fluoxetine. Finally, since adolescent and adult rats received pretreatment for 4 consecutive days and behavioral testing was conducted the following day, our study is limited to the short-term effects of drug exposure and acute pharmacological interventions. Additional studies exploring the long-term consequences of drug exposure during adolescence could better inform the effectiveness of pharmacological interventions on behavior later in adulthood.

With increasing evidence that aberrant activation of 5-HT systems during adolescence triggers lasting changes in neuronal signaling, use of drugs such as tobacco, e-cigarettes, and antidepressants may permanently alter the course of adolescent brain maturation and neurotransmitter signaling. Although teen smoking is declining, teen vaping has increased greatly in recent years (6, 7) and has been suggested to be a gateway to future tobacco use (61, 62). We have now shown that nicotine disrupts normative limbic development and sensitizes the adolescent brain to DA drugs such as cocaine. Thus, as with tobacco, e-cigarettes may prime behavioral susceptibility to drugs of abuse in teenagers with lifelong consequences for mental health.

Treatments for mental health disorders in teens also pose serious safety concerns. Even though fluoxetine is FDA approved to treat teen depression, the majority of testing regarding its effectiveness and safety has been conducted in adults (63, 64). In addition, fluoxetine’s antidepressant effects require chronic administration in order to be therapeutic, suggesting postsynaptic neuroadaptations in 5-HT receptor signaling (65). However, we have demonstrated that changes in adolescent behavior after nicotine or fluoxetine exposure can be blocked with an acute treatment of a 5-HT_1AR antagonist or partial agonist.

Prior studies have shown that 5-HT_1AR antagonists are safe for clinical use. The delayed antidepressant effects of SSRIs can be reduced by co-treatment of pindolol, a nonselective beta blocker with 5-HT_1AR antagonist activity (66, 67). The 5-HT_1AR partial agonist busipirone – which has long been used to treat anxiety – also demonstrates clinical efficacy when given in combination with SSRIs (68). Recently, vilazodone, which has combined SSRI and 5-HT_1AR partial agonist properties, has been approved for treatment of major depressive disorder in adults (69). Since we have shown that 5-HT_1AR antagonists and partial agonists attenuate adverse behavioral responses to antidepressants in adolescents, treatment for teen depression should take into consideration potential drug therapies with 5-HT_1AR antagonist activity to improve antidepressant efficacy, particularly during the initial period when adolescents are most vulnerable to the confounding effects of SSRI treatment. Furthermore, this body of work indicates that 5-HT_1ARs are also viable drug targets for the treatment of adverse health consequences after adolescent exposure to tobacco, e-cigarettes, and antidepressants.