Introduction

Behavioral consequences of d-amphetamine injections are generally considered as the consequence of an increased release of dopamine (DA) in the brain. This increased DA release is due to the blockade or the reversal of the DA transporter located on dopaminergic nerve terminals (Besson et al. 1971; Von Voigtlander and Moore 1973). Because the locomotor hyperactivity induced by the systemic injection of d-amphetamine is inhibited either by bilateral 6-hydroxydopamine lesions of mesolimbic dopaminergic neurons or by the application of neuroleptics into the nucleus accumbens, it was suggested that d-amphetamine-induced psychomotor activation mainly results from an increased DA transmission in this structure (Pijnenburg et al. 1975; Kelly et al. 1975). However, d-amphetamine is also a potent norepinephrine (NE) releaser and many studies have indicated that blocking noradrenergic transmission with an α1-adrenergic antagonist, such as prazosin, hampers d-amphetamine-induced locomotor hyperactivity (Snoddy and Tessel 1985; Dickinson et al. 1988; Blanc et al. 1994). More recently, it was shown that mice lacking the α1b-adrenergic receptor (α1b-AR KO) were severely affected in their locomotor response and behavioral sensitization to d-amphetamine (Drouin et al. 2002). Finally, microdialysis experiments indicated that the increased release of DA induced by d-amphetamine in the nucleus accumbens was absent in α1b-AR KO mice (Auclair et al. 2002). Altogether, these data confirmed the hypothesis of the existence of interactions between noradrenergic and dopaminergic neurons (Antelman and Caggiula 1977; Kokkinidis and Anisman 1978, 1979; Ogren et al. 1983; Tassin et al. 1986; Taghzouti et al. 1988; Lategan et al. 1990; Shi et al. 2000; Drouin et al. 2002) and, more precisely, of a coupling between noradrenergic and dopaminergic transmissions (Darracq et al. 1998).

Further experiments nevertheless indicated that, even in α1b-AR KO mice, d-amphetamine was still inducing a slight locomotor response, suggesting that other factors than noradrenergic and dopaminergic transmissions were implied in d-amphetamine-induced behavioral effects. Because some data had shown that the systemic injection of MDL 100907, a specific 5-HT2A receptor antagonist, reduces the locomotor hyperactivity induced by amphetamine (Moser et al. 1996; Sorensen et al. 1993; O’Neill et al. 1999), 5-HT2A receptors appeared good candidates to participate in the behavioral effects of d-amphetamine. Moreover, Porras et al. (2002) and Auclair et al. (2004a), using another 5-HT2A receptor antagonist, SR46349B (Rinaldi-Carmona et al. 1992), have found that the blockade of 5-HT2A receptors decreases the d-amphetamine-induced increase in extracellular DA levels in the nucleus accumbens. Finally, in more recent experiments, we found that 5-HT2A and α1b-adrenergic receptors entirely mediate not only DA release but also locomotor response and behavioral sensitization induced by d-amphetamine (Auclair et al. 2004b). These data therefore suggested that both receptors participate in the same regulating pathway.

To test this hypothesis, i.e., whether 5-HT2A receptors control d-amphetamine-induced locomotor activity in a way similar to that observed for α1b-adrenergic receptors, mice deprived of 5-HT2A receptors (5-HT2A-R KO) were studied. Their locomotor responses after acute and repeated treatment with d-amphetamine were monitored and compared to those obtained in wild-type (WT; 5-HT2A +/+) and α1b-AR KO mice. The effects of antagonists of α1b-adrenergic and 5-HT2A receptors, prazosin and SR46349B, respectively, were also assessed in 5-HT2A-R KO and WT mice. Finally, because effects of prazosin and SR46349B on d-amphetamine-induced locomotor hyperactivity were different between naïve and sensitized mice, dose–response curves for both receptor antagonists were performed. Data indicate that 5-HT2A receptors blockade do not have the same consequences in naïve mice than in those sensitized to d-amphetamine.

Materials and methods

Animals

Experiments were performed on adult male mice (25–35 g). 5HT2A-R KO mice (Fiorica-Howells et al. 2002; Weisstaub et al. 2006) were back-crossed on a C57BL6 genetic background for at least seven generations. The homozygous 5-HT2A +/+ mice were used as controls all over the study and referred to WT mice in the text. α1b-AR KO mice (Cavalli et al. 1997) were also back-crossed on a C57Bl6 genetic background for at least seven generations. When α1b-AR KO were used (see Fig. 2), it was verified that locomotor response to amphetamine was similar in homozygous α1b-AR +/+ and 5-HT2A +/+ mice. All mice, housed by groups of eight in plastic cages, were maintained on a 12-h light/dark cycle (lights on at 07:00) with food and water freely available.

Drugs

d-Amphetamine sulfate (Sigma Aldrich, L’Isle d’Abeau-Chesne, France) was dissolved in saline. Prazosin hydrochloride (Sigma Aldrich) was sonicated in water. SR46349B hemifumarate {(1Z,2E)-1-(2-fluoro-phenyl)-3-(4-hydroxyphenyl)-prop-2-en-one-O-(2-dimethylamino-ethyl)-oxime hemifumarate} was a generous gift from Laboratories Sanofi-Synthelabo (Montpellier, France). It was dissolved with a drop of lactic acid, neutralized with 1 M of NaOH, and sonicated in saline. All drugs were injected intraperitoneally (0.3 ml per 100 g). Doses are expressed as salts. d-Amphetamine was given at 2 mg/kg i.p. When not specifically specified, doses of prazosin (1 mg/kg, i.p.) and SR46349B (1 mg/kg, i.p.) were kept identical to previous experiments (Auclair et al. 2004a, b; Drouin et al. 2002; Salomon et al. 2006).

Locomotor activity

Acute treatment

Mice were introduced in a circular corridor (4.5 cm width, 17 cm external diameter) crossed by four infrared beams (1.5 cm above the base) placed at every 90° (Imetronic, Pessac, France). The locomotor activity was counted when animals interrupted two successive beams and thus had traveled a quarter of the circular corridor. In each session, the spontaneous activity was recorded for 60 min (habituation to the experimental procedure). Then, mice received intraperitoneal saline or pretreatment (prazosin or/and SR46349B), and their activity was recorded for 30 min. Finally, mice were injected intraperitoneally with d-amphetamine (2 mg/kg), and their locomotor responses were recorded for an additional 120-min period. Doses of d-amphetamine were chosen as the highest ones, which never induced stereotypy after repeated treatments.

Repeated treatment

Mice were injected on four consecutive days, and their locomotor activity was recorded with the same protocol as that for an acute treatment. To test the effect of the pretreatment on the expression of behavioral sensitization, mice received every single day only saline and d-amphetamine (2 mg/kg). Four days after the last injection, mice received prazosin (1 mg/kg) and/or SR46349B (1 mg/kg) followed 30 min later by d-amphetamine. To test the effect of the pretreatment on the development of behavioral sensitization, mice received every single day a pretreatment (prazosin [1 mg/kg] and/or SR46349B [1 mg/kg]) plus d-amphetamine. Four days after the last injection, mice received only saline and a d-amphetamine injection.

Statistics

Statistical analysis was performed using Graph Pad Prism 3.0 software, (San Diego, CA). Locomotor activity was described either in function of time or in function of pretreatment. Two-way ANOVA (genotype×treatment) were performed to study the effects of the different parameters and their possible interactions on the whole acute and repeated responses on 1 h of the three mice strains. Bonferroni posttests were done to compare more precisely the different strains. To study the global effect of treatments and pretreatments in naïve and sensitized animals and during the development of behavioral sensitization, means of the locomotor activity were as well calculated per 1 h and compared with one-way ANOVA followed by Tukey’s test. Pharmacological treatments correspond to independent groups of animals. Significant differences were set at P < 0.05.

Results

5HT2A-R KO mice exhibit higher locomotor response to d-amphetamine than WT mice; effects of prazosin and SR46349B

d-Amphetamine (2 mg/kg) induced increased locomotion in both WT and 5HT2A-R KO mice [two-way ANOVA (genotype×treatment) F(2,39) = 15.66; P < 0.0001]. Interestingly, we found also a significant effect of genotype (F(1,39) = 14.64; P < 0.0005) and the interaction between genotype and treatment (F(2,39) = 4.46; P = 0.0118). More precisely, 5HT2A-R KO mice were surprisingly found hyperreactive to d-amphetamine when compared with WT littermates (Bonferroni posttest: t = 2.83; P < 0.05). This difference was not related to the hyperreactivity of 5HT2A-R KO mice to novelty when compared with WT mice (Student’s t test t = 0.71; P = 0.5) nor to an effect of stress induced by injection (Fig. 1). As expected, in WT mice, pretreatment changed the locomotor response to d-amphetamine (one-way ANOVA F(2,21) = 9.723; P = 0.001). Prazosin (1 mg/kg), a α1-adrenergic receptor antagonist, decreased WT mice locomotor response to d-amphetamine by 93% (Tukey’s posttest; P < 0.01). SR46349B (1 mg/kg), a selective 5HT2A receptor antagonist, induced an 80% loss of the locomotor response (Tukey’s posttest; P < 0.01). In 5HT2A-R KO mice, pretreatment also modified the locomotor response (one-way ANOVA F(2,18) = 9.696; P = 0.0014). Prazosin completely blocked the important locomotor effects of d-amphetamine (−99%, Tukey’s posttest; P < 0.01); however, SR46349B did not modify, as expected, the locomotor response induced by d-amphetamine (Tukey’s posttest; P > 0.05), strongly suggesting that, in our conditions, SR46349B is specifically acting on 5-HT2A receptors.

Fig. 1
figure 1

Effects of prazosin and SR46349B on the acute d-amphetamine-induced locomotor response in WT and 5-HT2A-R KO mice. Prazosin (1 mg/kg, i.p.), SR46349B (1 mg/kg, i.p.), or saline were injected 30 min before d-amphetamine treatment (2 mg/kg, i.p.). Locomotor activity is expressed as quarter turns per 5 min and is presented as histograms of locomotor activity during 60 min after d-amphetamine injections. Saline injection that induces a very small locomotor response in both strains is not shown for sake of clarity. Each group contained at least seven and no more than nine animals. Double asterisks indicate P < 0.01, and triple asterisks indicate P < 0.001 significantly different from corresponding controls

Full size image

Similar behavioral sensitization to d-amphetamine in WT and 5HT2A-R KO mice but not in α1b-AR KO mice

Figure 2 presents the time courses of behavioral sensitization to d-amphetamine in three strains of mice. As shown on Fig. 1, d-amphetamine (2 mg/kg) induced an increased locomotion, which increased with repeated injections [two-way ANOVA (genotype × treatment) F(4,90) = 8.60; P < 0.0001 for the treatment effect]. A significant effect of genotype was found (F(2,90) = 61.55; P < 0.0001) but no interaction between the genotype and treatment. Comparing the acute response, the three strains responded differently (one-way ANOVA, F(2,20) = 8.23; P = 0.023). As shown on Fig. 1, 5-HT2A-R KO mice exhibited a higher locomotor response than WT mice (+110%, Tukey’s posttest, P < 0.05). On the contrary, and in agreement with previous data, d-amphetamine induced a lower locomotor activity in α1b-AR KO mice than in WT mice (−63%, Tukey’s posttest; P < 0.05). After four daily injections and a 4-day withdrawal period, the d-amphetamine test injection induced a potentiated locomotor activity in WT mice [+292%, one-way ANOVA, F(4,35) = 7.2; P = 0.0002] and in 5HT2A-R KO mice [+93%, one-way ANOVA F(4,28) = 2.72; P = 0.05] but not in α1b-AR KO mice [one-way ANOVA, F(4,27) = 2.72; P = 0.56], revealing a behavioral sensitization in the two first strains but not in the third one. Interestingly, after behavioral sensitization to d-amphetamine, whereas locomotor response to d-amphetamine in the three strains was still different [one-way ANOVA, F(2,20) = 8.23; P = 0.023], locomotor response to d-amphetamine was similar in 5HT2A-R KO and WT mice (Tukey’s posttest; P > 0.05). Locomotor response to d-amphetamine in sensitized α1b-AR KO mice was found to be significantly lower than in sensitized WT mice (−80%, Tukey’s posttest; P < 0.001).

Fig. 2
figure 2

Development of behavioral sensitization to d-amphetamine in WT, 5-HT2A-R KO, and α1b-AR KO mice. Mice were given every day one injection of d-amphetamine (2 mg/kg) and received, after a 4-day withdrawal, a last injection of d-amphetamine (2 mg/kg) on the test day. Locomotor activity is expressed as the sum of quarter turns per 60 min after d-amphetamine injection. Each group contained at least seven and no more than nine animals. An asterisk indicates P < 0.05, and triple asterisks indicate P < 0.001 significantly different from corresponding controls

Full size image

Effects of prazosin and SR46349B in sensitized 5HT2A-R KO and WT mice

Repeating amphetamine injections induces an increase in locomotor response to d-amphetamine in both WT and 5-HT2A-R KO mice (Fig. 3). In sensitized mice, significant effects of the treatment and the genotype [two-way ANOVA (genotype × treatment), treatment: F(2,37) = 22.50; P < 0.0001 and genotype: F(1,37) = 5.01; P = 0.03] were found, but no interaction was revealed between the two parameters. As described above, after repeated treatment with d-amphetamine, 5HT2A-R KO and WT mice exhibited the same locomotor response to d-amphetamine (Bonferroni posttest: t = 0.24; P > 0.05). In WT mice, pretreatments modified significantly the locomotor responses to d-amphetamine [one-way ANOVA, F(2,21) = 10.21; P = 0.0008]. Prazosin (1 mg/kg) and SR46349B (1 mg/kg) partially blocked the locomotor response to d-amphetamine (−67%, Tukey’s posttest; P < 0.001 and −39%; P < 0.05, for prazosin and SR46349B, respectively). We also found a significant effect of pretreatment in 5-HT2A-R KO mice [one-way ANOVA, F(2,17) = 17.37; P < 0.0001]. Interestingly, SR46349B had no effect (Tukey’s posttest P > 0.05); however, prazosin, in contrast to that obtained in naïve animals (Fig. 1), only reduced the locomotor response by 63% (Tukey’s posttest; P < 0.001).

Fig. 3
figure 3

Effects of prazosin and SR46349B on the repeated d-amphetamine-induced locomotor activity in WT and 5-HT2A-R KO mice. Prazosin (1 mg/kg, i.p.), SR46349B (1 mg/kg, i.p.), or saline were injected 30 min before d-amphetamine treatment (2 mg/kg, i.p.) on the test day. Locomotor activity is expressed in quarter turns per 5 min and is presented as histograms of locomotor activity during 60 min after d-amphetamine injections. Each group contained at least seven and no more than nine animals. An asterisk indicates P < 0.05, and triple asterisks indicate P < 0.001 significantly different from corresponding controls

Full size image

Dose–response curves of the effects of prazosin and SR46349B on locomotor responses to d-amphetamine in naïve and sensitized 5-HT2A-R KO and WT mice

To analyze why prazosin does not completely block locomotor response to d-amphetamine in sensitized 5HT2A-R KO mice, dose–response curves were performed in naïve and sensitized 5HT2A-R KO and WT mice. Figure 4 summarizes data obtained after injection of prazosin or SR46349B in WT and 5-HT2A-R KO mice. It confirms that each antagonist, prazosin or SR46349B, blocks by more than 80% d-amphetamine-induced locomotor response in naïve WT mice (80 + 80% > 100%, which makes the sum of both effects more than additive), whereas these effects are smaller and become additive (respectively, 63 and 39%, which makes almost 100%) in sensitized WT mice. Increasing the dose of prazosin from 1 to 3 mg/kg does not lead to a higher inhibition of locomotor response to d-amphetamine in sensitized and naïve WT mice (Tukey’s posttest; P > 0.05 in both cases). Finally, it is confirmed on this figure that, as shown on Figs. 1 and 3 and up to 1 mg/kg, SR46349B has no effect on d-amphetamine-induced locomotor response in 5-HT2A-R KO mice. In 5-HT2A-R KO naïve mice, 1 mg/kg of prazosin blocks completely d-amphetamine-induced locomotor response, an effect which is similar to that obtained with 3 mg/kg of prazosin [one-way ANOVA, F(3,24) = 21.16; P < 0.0001; Tukey’s posttest between 1 and 3 mg/kg of prazosin; P > 0.05]. In 5-HT2A-R KO sensitized mice, a complete inhibition of d-amphetamine-induced locomotor response is only obtained with 3 mg/kg of prazosin, an effect significantly different from the partial inhibition now obtained with 1 mg/kg of prazosin [−63%, one-way ANOVA, F(3,24) = 28.22; P < 0.0001; Tukey’s posttest between 1 and 3 mg/kg of prazosin; P < 0.05).

Fig. 4
figure 4

Prazosin and SR46349B pretreatment dose–responses curves in d-amphetamine naïve and sensitized WT and 5-HT2A-R KO mice. Prazosin (0–3 mg/kg, i.p.), SR46349B (0–1 mg/kg, i.p.), or saline were injected 30 min before d-amphetamine treatment (2 mg/kg, i.p.). This figure shows, among other points, that increasing doses of prazosin from 1 to 3 mg/kg does not modify its inhibitory effect in sensitized WT mice whereas it does in 5-HT2A-R KO mice. Locomotor activity is expressed as the sum of quarter turns per 60 min after d-amphetamine injection. Each group contained at least seven and no more than nine animals

Full size image

Complete blockade of the development of behavioral sensitization to d-amphetamine is obtained with prazosin (1 mg/kg) in 5HT2A-R KO mice but necessitates the addition of SR46349B (1 mg/kg) to prazosin in WT mice

In WT mice, pretreatment with prazosin (1 mg/kg) 30 min before the d-amphetamine injection modifies significantly the locomotor sensitization to d-amphetamine [one-way ANOVA, F(5,36) = 16.68; P < 0.0001] and leads, on the test day, to a partial sensitization of the mice (+126%, Tukey’s posttest P < 0.05 when compared to acute d-amphetamine and −43%, P < 0.05 when compared to repeated saline/d-amphetamine; Fig. 5a). The same protocol with the 5HT2A-R KO mice was sufficient to completely block the development of the behavioral sensitization [one-way ANOVA, F(5,36) = 21.74 P < 0.0001; Tukey’s posttest P > 0.05 when compared to acute d-amphetamine). Interestingly, the WT mice pretreated with prazosin showed on the test day a response to d-amphetamine totally similar to that obtained in naïve 5HT2A-R KO mice (t (1, 10) = 0.25, P = 0.80, Student’s t test). The addition of SR46349B (1 mg/kg) to prazosin in the pretreatment of WT mice (Fig. 5b) completely blocked the development of behavioral sensitization [one-way ANOVA, F(5,39) = 23.36 P < 0.0001; Tukey’s posttest P > 0.05 when compared to acute d-amphetamine), in agreement with previous data (Auclair et al. 2004b).

Fig. 5
figure 5

Effects of prazosin alone or in combination with SR46349B on the development of behavioral sensitization to d-amphetamine in WT and in 5-HT2A-R KO mice. In WT and 5-HT2A-R KO mice, pretreatment [prazosin (1 mg/kg) or saline] was administered up to the fourth day, 30 min before d-amphetamine injection (a). In WT mice, pretreatment [prazosin + SR46349B (1 mg/kg) or saline] was administered up to the fourth day, 30 min before d-amphetamine injection (b). On the test day, only saline and d-amphetamine were injected. Histograms represent the sum of the locomotor activity in quarter turn per 60 min. Each group contained at least seven and no more than nine animals. An asterisk indicates P < 0.05, double asterisks indicate P < 0.01, and triple asterisks indicate P < 0.001 significantly different from corresponding controls

Full size image

Discussion

When animals receive an injection of a α1-adrenergic receptor antagonist before a systemic administration of d-amphetamine, the locomotor hyperactivity usually observed after a d-amphetamine injection is severely affected. In agreement with this finding, α1b-AR KO mice exhibit a diminished locomotor response to d-amphetamine (Drouin et al. 2002). Similarly, when animals receive an injection of a 5-HT2A receptor antagonist before a systemic administration of d-amphetamine, the d-amphetamine-induced locomotor response is decreased by more than 80%. We show now that, paradoxically, 5-HT2A-R KO mice exhibit a twofold higher locomotor response to d-amphetamine than WT littermates. These data suggest that 5-HT2A-R KO mice are constitutively sensitized to d-amphetamine.

However, despite a higher (+110%, Fig. 1) acute response to d-amphetamine than WT mice, d-amphetamine-induced locomotor activity is lower in naïve 5-HT2A-R KO mice than in sensitized WT mice (−47%, Fig. 2), i.e., those having received four injections of d-amphetamine. Actually, repeated injections of d-amphetamine increase d-amphetamine-induced locomotor response in 5-HT2A-R KO mice in such a way that these latter respond to d-amphetamine with the same amplitude than sensitized WT mice (Fig. 2). This means that, although 5-HT2A-R KO mice appear to be constitutively sensitized, repeating d-amphetamine treatments can still increase their locomotor response to d-amphetamine, suggesting that their sensitization to d-amphetamine is only partial.

Analysis of dose–response curves of d-amphetamine-induced locomotor activity after α1-adrenergic or 5-HT2A receptor antagonist treatment shows that, as expected, SR46349B, a 5-HT2A receptor antagonist, has no effect on 5-HT2A-R KO mice d-amphetamine-induced locomotor response, regardless of mice being naïve or sensitized to d-amphetamine. This strongly suggests that SR46349B is, at least at the doses used in this study, specific for 5-HT2A receptors. Moreover, previous studies performed in rats have indicated a tenfold difference of SR46349B affinity for 5-HT2A and 5-HT2C receptors (Rinaldi-Carmona et al. 1992). In sensitized WT mice, it seems that d-amphetamine-induced locomotor response can be dissociated into two components, an α1-adrenergic one that is blocked by prazosin (−63%) and the other one being serotonergic (5-HT2A) and blocked by SR46349B (−39%). In naïve WT mice, however, each antagonist blocks almost completely the d-amphetamine-induced locomotor response, suggesting that both locomotor components are linked or coupled. In other words, in naïve mice, blocking one locomotor component with one antagonist may block the other locomotor component, thus inhibiting completely the locomotor response. Repeated d-amphetamine would dissociate the link between the two locomotor components, as previously proposed (Salomon et al. 2006). This hypothesis may also explain why animals repeatedly treated with d-amphetamine have a higher locomotor response to d-amphetamine than naïve mice and thus exhibit a behavioral sensitization.

Our experiments therefore indicate that repeating d-amphetamine injections diminish the blocking effects of prazosin and SR46349B on d-amphetamine-induced locomotor response. This difference between naïve and sensitized mice response is, however, more important with regards to SR46349B than prazosin. Accordingly, when the same experiments were performed in rats, this difference between naïve and sensitized animals was only found significant with SR46349B (data not shown). This nevertheless indicates that consequences of repeated amphetamine injections on the mutual regulation between noradrenergic and serotonergic neurons are not specific to one species.

We have shown previously that naïve 5-HT2A-R KO mice exhibit acutely a higher cortical NE response to d-amphetamine than WT mice and that this increased response is correlated with behavioral sensitization to d-amphetamine (Salomon et al. 2006). Recent experiments indicate that amphetamine sensitization does not increase further cortical NE response in 5-HT2A-R KO mice (data not shown). However, 3 mg/kg of prazosin is necessary to block completely locomotor response to d-amphetamine in sensitized 5-HT2A-R KO mice, whereas 1 mg/kg of prazosin is enough in naïve 5-HT2A-R KO mice. One possibility could be that repeating amphetamine injections in 5-HT2A-R KO mice has modified the affinity of α1b-adrenergic receptors for prazosin. This hypothesis is presently tested in the laboratory.

Altogether, we show that genetic deletion of 5-HT2A receptors facilitates the locomotor response to d-amphetamine; however, this increased response can still be amplified by repeated d-amphetamine treatments. Behavioral sensitization to d-amphetamine in 5-HT2A-R KO mice may therefore be partly due to a disinhibition of noradrenergic neurons related to the absence of 5-HT2A receptors (constitutive sensitization) and partly to a modification of α1b-adrenergic receptors because of repeated amphetamine treatments (induced sensitization).

A role for 5-HT2A receptors in the regulation of locus coeruleus noradrenergic neurons has already been described. For example, serotonin exerts a tonic inhibitory influence on locus coeruleus neurons through postsynaptic 5-HT2A receptors that are not located on noradrenergic neurons (Szabo and Blier 2001). Same authors (Szabo and Blier 2002) have also shown that, whereas sustained administration of a serotonin reuptake blocker, such as citalopram or paroxetine, reduces firing activity of noradrenergic neurons, a serotonin reuptake blocker that is also a 5-HT2A receptor antagonist, YM992, increases the synaptic availability of NE. Similarly, Chiang and Aston-Jones (1993) have suggested that 5-HT2A receptor stimulation influences locus coeruleus indirectly and causes tonic activation of a GABAergic input to noradrenergic neurons. In agreement with this hypothesis, Szabo and Blier (2001) have proposed a localization of 5-HT2A receptors on GABAergic nerve terminals in the locus coeruleus.

One question remains, however, unanswered: Despite the inhibiting effects of prazosin and SR46349B on d-amphetamine-induced locomotor response, why are α1b-AR KO and 5-HT2A-R KO mice, respectively, hypo- and hyper-responsive to d-amphetamine?

The first part of the answer is that 2 mg/kg of d-amphetamine releases NE but has no effect on extracellular serotonin, at least in our hands (Salomon et al. 2006). Lack of α1b-adrenergic receptors annihilates the behavioral effects of d-amphetamine related to an increased NE release (Drouin et al. 2002); however, in 5-HT2A-R KO mice, the disinhibition of noradrenergic neurons may increase the stimulation by NE of cortical α1b-adrenergic receptors (Darracq et al. 1998). The possibility that d-amphetamine effects are specific to noradrenergic neurons was confirmed when we tested locomotor responses to a specific serotonin releaser, para-chloroamphetamine (PCA). Indeed, locomotor response to PCA is more than threefold higher in α1b-AR KO mice than in WT mice (Salomon et al. 2006), thus suggesting that, in α1b-AR KO mice, serotonergic neurons are disinhibited as well as are noradrenergic neurons in 5-HT2A-R KO mice.

The second part of the answer concerns the effects of SR46349B and prazosin; both compounds decrease locomotor response to d-amphetamine in WT mice. This means that the stimulation of α1b-adrenergic or/and 5-HT2A receptors should facilitate locomotor responses to d-amphetamine, a finding that disagrees with data obtained in 5-HT2A-R KO mice. This may indicate that the stimulation of one type of receptor has two opposite consequences depending on its localization. For example, stimulation of 5-HT2A receptors increases amphetamine-induced DA release (Ichikawa and Meltzer 1995) or locomotor response through an activation of dopaminergic neurons in the ventral tegmental area (VTA; Doherty and Pickel 2000; Auclair et al. 2004a) or of pyramidal cells in the prefrontal cortex (Pazos et al. 1985; Bortolozzi et al. 2005), whereas stimulation of 5-HT2A receptors located in the locus coeruleus (Szabo and Blier 2001) inhibits noradrenergic neurons. If we consider our previous findings (Darracq et al. 1998; Drouin et al. 2002), a decreased firing of noradrenergic neurons should induce a decreased locomotor response to d-amphetamine. Thus, depending on the localization of 5-HT2A receptors, VTA or locus coeruleus, stimulation of 5-HT2A receptors may, respectively, increase or decrease d-amphetamine locomotor response. In any case, our data show clearly that the chronic absence of a receptor induced by a genetic deletion may have a behavioral consequence opposite to its pharmacological acute blockade. We are not aware of any other similar example in the literature. However, Heisler and Tecott (2000) have shown that the deletion of 5-HT2C receptors unmasks a pharmacological effect on 5-HT1B receptors.

In conclusion, we show that 5-HT2A-R KO mice are hyperresponsive to d-amphetamine and propose that this is due to a disinhibition of noradrenergic neurons activated by d-amphetamine. This increased noradrenergic response may compensate for the absence of the 5-HT2A receptors located in the VTA and/or in the prefrontal cortex. Our data also indicate that repeated amphetamine treatments disrupt the inhibition of noradrenergic neurons by 5-HT2A receptors thus suggesting a physiological basis of behavioral sensitization to d-amphetamine.