Role of Serotonin in the Paradoxical Calming Effect of Psychostimulants on Hyperactivity

See allHide authors and affiliations

Science  15 Jan 1999:
Vol. 283, Issue 5400, pp. 397-401
DOI: 10.1126/science.283.5400.397


The mechanism by which psychostimulants act as calming agents in humans with attention-deficit hyperactivity disorder (ADHD) or hyperkinetic disorder is currently unknown. Mice lacking the gene encoding the plasma membrane dopamine transporter (DAT) have elevated dopaminergic tone and are hyperactive. This activity was exacerbated by exposure to a novel environment. Additionally, these mice were impaired in spatial cognitive function, and they showed a decrease in locomotion in response to psychostimulants. This paradoxical calming effect of psychostimulants depended on serotonergic neurotransmission. The parallels between the DAT knockout mice and individuals with ADHD suggest that common mechanisms may underlie some of their behaviors and responses to psychostimulants.

The catecholamine dopamine (DA) is present in both the central and peripheral nervous systems, where it controls a variety of different physiological processes (1). Perhaps one of the most important regulators of dopaminergic function is the DA transporter (DAT) (2). This transporter is located on the plasma membrane of DA neurons, where it controls the concentrations of DA by rapidly removing the transmitter from the extracellular space and localizing it into the cytoplasm (3). A long-term interest in this transporter derives from its role as an endogenous target for psychostimulants, antidepressants, and several neurotoxins (3). Disruption of the DAT gene in mice (4) results in a phenotype that includes behavioral abnormalities, neuroendocrine dysfunction, dwarfism, and altered sensitivities to certain drugs (2, 5). A prominent characteristic of these mice is their marked hyperactivity (4). This hyperkinetic behavior is consistent with the loss of transporter function and the resultant high concentrations of extracellular DA in the striata of these mice (2).

Recently, an association between polymorphisms in the noncoding regions of the human DAT and ADHD has been suggested (6). ADHD is a condition that is manifested by impulsivity, hyperactivity, and inattention (7). These symptoms are also defined as hyperkinetic disorder (HKD) (8). It is estimated that 3 to 6% of school-aged children are affected by this condition (9). Since the 1930s, treatment has involved the use of psychostimulant compounds that paradoxically serve to attenuate the hyperactivity and often improve cognitive performance (9, 10). Although psychostimulants preferentially inhibit the DAT, they also block the action of other monoamine transporters (3, 5,11). As a result, extracellular concentrations of DA, norepinephrine (NE), and serotonin (5-HT) can be elevated. Although dysregulation of each of these monoamine systems has been postulated to be involved in ADHD-HKD, it is commonly believed that the DA system is preferentially implicated in the etiology and pharmacotherapy of these disorders (9, 12).

Drugs that block the DAT result in a pronounced increase in extracellular DA in brain regions that mediate enhanced locomotion and stereotypic behavior (2–4, 11). DAT knockout (KO) mice were placed into a novel environment, and locomotor, rearing, and stereotypic activities were monitored (13). DAT-KO animals exhibited substantially higher levels of activity than wild-type mice (Fig. 1, A to C). Whereas the DAT-KO mice showed minimal habituation to the novel environment over 240 min of observation, habituation occurred within the first 30 to 40 min for the wild-type animals. Hyperactivity in the DAT-KO animals appeared to be novelty driven because locomotor activity was about 12-fold higher in the novel environment (Fig. 1D). Moreover, repetitive exposure to the open field augmented the activity of the DAT-KO mice over seven consecutive days of testing, whereas locomotor responses of the wild-type mice showed habituation (14). These results suggest that the DAT-KO animals might be less able to adapt to novel stimuli than the wild-type controls.

Figure 1

Parameters of locomotor activity in an open-field environment and accompanying extracellular striatal DA concentrations in wild-type (WT) and DAT-KO mice. (A to C) Locomotion (horizontal activity), rearings (vertical activity), and stereotypy in a novel environment for wild-type and DAT-KO mice. Activity was recorded every 5 min over a 4-hour period; n = 10 for wild-type mice andn = 16 for DAT-KO animals. (D) Locomotor activity in either a familiar or a novel environment for wild-type and DAT-KO mice. Activity was monitored for 1 hour; n = 10 for wild-type and n = 10 for DAT-KO animals. ∗,P < 0.05—activity in familiar versus novel environment. No significant differences between the activities of DAT-KO and wild-type mice were discerned in their home cages. (E) Extracellular DA concentrations monitored by microdialysis in striatum of freely moving mice in home cages and the novel, open-field environment. Samples were taken every 20 min;n = 4 for wild-type and n = 5 for DAT-KO mice. (F) Effects of haloperidol on locomotor activity of the DAT-KO mice in a novel environment. After a 30-min exposure to the open field, DAT-KO mice and wild-type controls (14) were injected with either saline (SAL) or haloperidol (HAL); n = 14 for the DAT-KO mice given saline andn = 6 for the DAT-KO mice given haloperidol. Haloperidol effectively inhibited locomotion in both genotypes [total distance traveled by wild-type mice for 210 min was 219 ± 52 cm (n = 6, haloperidol-treated) compared with 640 ± 98 cm (n = 10, saline-treated); ∗, P< 0.05 (14)].

Extracellular DA concentrations were sampled by microdialysis in freely moving mice in both the familiar and novel environments (15). Despite the fact that DAT-KO mice have fivefold higher concentrations of extracellular DA (2), no differences were discerned as a function of time when samples were taken from mice in either environment (Fig. 1E). Dopaminergic neurotransmission, however, was required for the presence of locomotor activity because haloperidol, a DA receptor antagonist, suppressed activity in both the wild-type (14) and DAT-KO mice (Fig. 1F). These data suggest that although the enhanced spontaneous locomotor activity in the DAT-deficient animals was not due to additional augmentation of dopaminergic neurotransmission in striatum, it was dependent on dopaminergic tone (see also Fig. 4C). These findings indicate that additional mechanisms may be responsible for the novelty-induced hyperactivity of the DAT-KO mice.

An aspect of ADHD-HKD that may accompany or occur independently of the hyperactivity involves cognition. Animals were tested in an eight-arm radial maze with a win-shift paradigm over 21 consecutive days (16). In this procedure, only the first entry in each arm was rewarded. Initially, performance was very poor for both genotypes (Fig. 2A). As training progressed, wild-type mice attained high levels of performance within the first several sessions. As a group, the DAT-KO mice were significantly impaired. These differences in performance did not appear to be due to possible motivational differences because animals from both groups consumed practically identical amounts of food in the maze (17).

Figure 2

Assessment of spatial cognitive performance in wild-type and DAT-KO mice. Spatial learning and memory were assessed with an eight-arm radial maze during win-shift testing over 21 sessions. (A) Performance as assessed by the number of entries made into each of the eight arms of the apparatus before making an error (entries to repeat). (B) Number of perseverative errors made per animal across the seven session blocks of testing. (C) Latency to enter a given arm (the time it took to enter all eight arms or 300 s divided by the total number of arms entered during this time). Wild-type (n = 9) (○) and DAT-KO mice (n = 9) (•); ∗, P < 0.05.

Another manner of evaluating performance in the radial maze is to examine the response patterns that are made while acquiring the task. One type of response is that of perseveration (18). DAT-KO mice made significantly more perseverative errors than the wild-type animals, and these mutants showed very little reduction in these types of errors over all sessions of training (Fig. 2B). Activity in the maze was also assessed. During any given session, most of the DAT-KO mice were unable to solve the maze within the allotted 300 s. By the fifth through the seventh test blocks, wild-type animals were reliably solving the maze within 120 to 160 s (17). When the time to complete the task was divided by the total number of arm entries, DAT-KO mice exhibited longer latencies to enter a given arm than their wild-type controls (Fig. 2C). In contrast to the open-field tests where the DAT-KO mice were hyperactive, these maze data imply that the mutants were less active than the wild-type mice in this setting. In fact, mutants were often observed to spend more time engaged in other extraneous activities (for example, sniffing, rearing, and so forth) than the controls (17). These results suggest that the DAT-KO mice are deficient in spatial learning and that they might have more difficulty than wild-type animals in suppressing inappropriate responses.

Conventional treatment of ADHD-HKD generally involves the use of psychostimulants. Although these drugs serve to enhance activity in normal individuals, they exert a calming effect in ADHD-HKD patients. Psychostimulant effects on the activities of wild-type and DAT-KO animals were evaluated in the open-field environment (13). Methylphenidate, dextroamphetamine, and cocaine enhanced locomotor activity in wild-type mice (Fig. 3, A to C). By contrast, activity in the DAT-KO animals was substantially attenuated (Fig. 3, E to G) (19). Similar effects on rearing and stereotypic activities were observed (14). The effects of methylphenidate on locomotion in wild-type mice were almost immediate, whereas those in the DAT-KO mice were much more delayed and long-lasting. Moreover, additional experiments revealed that the inhibitory effect of methylphenidate on the hyperactivity of the DAT-KO animals was dose-dependent [20 to 60 mg/kg, intraperitoneal (ip)] (14). The same doses in wild-type controls were excitatory (14), with the enhancement in activity following an inverted U-shaped function (20).

Figure 3

Effects of various drugs on locomotor activity in a novel environment in wild-type and DAT-KO mice. (A toD) Wild-type and (E to L) DAT-KO mice were placed in the open-field apparatus for an initial period of 30 min and then were injected with the psychostimulants: methylphenidate (MPH), d-amphetamine (AMPH), and cocaine (COC); the NET inhibitor [nisoxetine (NIS)]; or various serotonergic agents: a SERT inhibitor [fluoxetine (FLU)], a mixed serotonin receptor agonist [quipazine (QUI)], or serotonin precursors [5-hydroxytryptophan (5-HTP) and l -tryptophan (TRP)]; controls were given saline (SAL; 10 ml/kg, ip). Mice were returned to the apparatus and activity was measured. Drug-naı̈ve animals were used for each drug protocol and n represents the number of animals used. ○, control; •, treated mice. All serotonergic drugs (I to L) potently inhibited the activity of DAT-KO mice, whereas in wild-type mice (14) only 5-HTP produced a significant decrease in the activity [total distance traveled by wild-type mice for 120 min was 221 ± 52 cm (n = 6, 5-HTP–treated) compared with 525 ± 79 cm (n = 10, saline-treated), P < 0.05]. In an additional set of experiments, lower doses of d-amphetamine (0.75 mg/kg, ip), quipazine (0.5 mg/kg, ip), 5-hydroxytryptophan (10 mg/kg, ip), andl -tryptophan (10 mg/kg, ip) were all able to significantly decrease the activity of DAT-KO mice. With the exception of amphetamine, no effects of these treatments on the activity of wild-type mice were observed (14). s.c., subcutaneous.

Extracellular DA concentrations in striatum were measured by microdialysis in freely moving animals after administration of methylphenidate (15). DA concentrations were markedly augmented in wild-type mice, whereas those for the DAT-KO animals were unperturbed by the methylphenidate (see Fig. 4A), amphetamine, or cocaine treatments (5). The time courses for the increase and decline in striatal DA concentrations in wild-type mice approximated the changes in locomotor activity (see Fig. 3A). By contrast, the DA concentrations in mutants, which were already increased, were unchanged in response to methylphenidate (see Fig. 4A) despite the attenuation in locomotor activity (see Fig. 3E). This association of striatal DA concentrations with activity in the wild-type mice and the dissociation of these two parameters in the DAT-KO animals suggests that different mechanisms may be involved in these responses to psychostimulants.

Figure 4

Analyses of changes in extracellular DA concentrations in striatum and locomotor activity of DAT-KO and wild-type mice after administration of dopaminergic and serotonergic drugs. (A and B) Effects of methylphenidate or fluoxetine on extracellular DA concentrations in the striatum of freely moving mice measured in home cage environment. At least three samples were collected before each drug treatment (and the mean value was set to 100%). Samples were taken every 20 min for a period of 2 to 3 hours after drug administration; n = 4 for wild-type andn = 5 for DAT-KO mice. In wild-type mice, all points, except for the effect of methylphenidate 160 and 180 min after administration, were significantly different (P < 0.05) from wild-type saline-treated controls (14). There was no significant effect of methylphenidate in DAT-KO mice. Fluoxetine had no significant effect in either genotype. (C) DAT-KO mice were given αMT to block DA synthesis 10 min before exposure to the open-field apparatus. Forty minutes later, mice were administered a mixed D1/D2 dopamine receptor agonist [apomorphine (APO)] to restore locomotor activity (n = 6). A second group of DAT-KO animals were given the αMT, as described above, followed 30 min later by a SERT inhibitor [fluoxetine (FLU)], to enhance extracellular 5-HT concentrations, followed by apomorhine (n = 6).

Psychostimulants have also been reported to interact with the NE (NET) and 5-HT transporters (SERT) (3, 5,11). To test whether NE or 5-HT neurons could be involved in the control of hyperactivity of the DAT-KO mice, we selectively activated each of these systems. Nisoxetine, a selective inhibitor of the NET (21), administered at 4 mg/kg (14) and 10 mg/kg, exerted no effect on locomotion in wild-type (Fig. 3D) or DAT-KO mice (Fig. 3H). Similarly, no significant effects on rearing or stereotypic activity were noted (14). These findings indicate that the NET is unlikely to play an important role in the activating or calming effects of psychostimulants in these mice.

Administration of the SERT inhibitor, fluoxetine, markedly attenuated the activity of the DAT-KO mice (Fig. 3I), whereas it had no effect on wild-type animals (14). These actions were presumably mediated by increased extracellular concentrations of 5-HT due to blockade of the transporter because administration of quipazine, a nonselective 5-HT receptor agonist, also reduced hyperlocomotion in the DAT-KO mice (Fig. 3J). Another method for potentiating central 5-HT neurotransmission is through the increased availability of precursor substrates (22). When mice were treated with 5-hydroxytryptophan or with the dietary 5-HT precursor (l-tryptophan), hyperlocomotion in the DAT-KO mice was again profoundly reduced (Fig. 3, K and L). Similar reductions in other activity parameters (rearing and stereotypic responses) were also observed (14). By comparison, the wild-type animals either did not respond to these serotonergic pharmacological treatments or responded with only marginal and transient reductions in locomotion (14). These pharmacological data emphasize the importance of 5-HT in the regulation of activity levels in the DAT-KO mice.

To determine whether serotonergic agents affect activity through modulation of striatal DA release, we measured extracellular concentrations of DA by microdialysis after fluoxetine treatment. Repeated 20-min samplings after drug administration revealed no significant alterations in striatal DA concentrations in either genotype (Fig. 4B). These data demonstrate that serotonergic neurotransmission can modulate hyperactivity without producing concurrent changes in extracellular striatal DA concentrations.

To examine the mechanism of 5-HT action, we performed additional experiments under conditions in which dopaminergic tone was provided by a direct DA receptor agonist. DAT-KO animals were treated with the tyrosine hydroxylase inhibitor α-methyl-p-tyrosine (αMT) for 40 min, which effectively depleted extracellular DA concentrations by greater than 80% (14) and completely abolished locomotion within 20 min (Fig. 4C). Locomotion was restored when these mice were treated with a direct D1/D2 receptor agonist (apomorphine) to replace the endogenous DA. When fluoxetine was given before apomorphine administration to enhance serotonergic neurotransmission, the locomotor response to apomorphine was completely absent. These data suggest that the 5-HT effects on hyperactivity are localized downstream of dopaminergic neurotransmission.

It is commonly believed that changes in dopaminergic tone are highly related to alterations in locomotor activity (1,3, 4, 11). DAT-KO mice display marked hyperactivity with respect to locomotor, rearing, and stereotypic responses over that of wild-type animals. These responses are most evident when the mice are placed into a novel environment. Hyperlocomotion of DAT-KO mice is similar to the behavior of ADHD-HKD children where many of these individuals have greater difficulty in suppressing their hyperactivity in novel surroundings (9,23).

In the present study, global processes associated with spatial learning and memory were impaired in the DAT-KO mice. Analysis of perseverative errors in the radial maze suggested that these mice might be unable to inhibit inappropriate responses or that they might be inattentive to relevant cues in their environment (18). Interestingly, Barkley (24) has recently proposed that ADHD-HKD patients suffer primarily from an inability to inhibit their behavioral responses such that they are hyperresponsive to various stimuli. Therefore, impairments of cognitive function would be secondary to the hyperkinesis. Similar mechanisms may exist in the DAT-KO mice.

Psychostimulants exerted a calming effect in DAT-KO mice. The studies with nisoxetine suggested that noradrenergic transmission was unlikely to play an important role in this response. By contrast, agents that increased serotonergic neurotransmission were all observed to substantially reduce hyperactivity in the DAT-KO mice. Only minor effects were seen in wild-type mice. The relative magnitude of these effects suggests that levels of existing dopaminergic tone in wild-type and DAT-KO mice may determine the potency of the serotonergic inhibitory effect. Altogether, these data indicate that the primary calming effect of psychostimulants in DAT-KO mice is mediated by the 5-HT system (25).

Psychostimulant therapy in ADHD-HKD patients is somewhat controversial because these drugs have long-term sensitizing effects and a potential for abuse and they may be neurotoxic (26). The findings with the DAT-KO mice provide the tantalizing possibility that hyperkinetic behaviors may be controlled through the precise targeting of 5-HT receptors or even through enhanced availability of 5-HT precursors. 5-HT has been considered as inhibitory to behavioral activation (27) and impulsive behavior (28). Several investigators (29) have suggested that 5-HT may reduce psychostimulant-induced hyperactivity; however, the neuronal circuitry and mechanisms involved in this DA–5-HT interaction have remained elusive (27, 30). This situation has become much more complicated with the identification of multiple 5-HT receptor subtypes that may activate, inhibit, or exert no effects on locomotion (27). Our results reaffirm in a genetic model that 5-HT can constrain hyperactivity and that this effect is localized downstream of DA neurons.

DAT-KO mice are hyperactive, show an impairment in spatial cognitive function, and exhibit paradoxical responses to psychostimulants. Despite these similarities between the mutant mice and humans with ADHD-HKD, it is unlikely that their phenotypes are completely identical. Nonetheless, the results with the DAT-KO mice emphasize the importance of a relative balance of the 5-HT and DA systems for normal motor activity. Therefore, alterations in any of the parameters that control this delicate homeostatic situation might underlie hyperactive states. Despite these reservations, the preponderance of common symptomatologies between the DAT-KO mice and individuals with ADHD-HKD suggests that these mice may not only serve as a useful animal model and as a resource to test new therapies but that they may also provide insights into the basic mechanisms that underlie the etiology of this and other hyperkinetic disorders.

  • * Present address: CNRS UMR5541, University of Bordeaux II Victor Segalen, 146 Rue Leo Saignat, 33076 Bordeaux Cedex, France.

  • To whom correspondence should be addressed. E-mail: caron002{at}


View Abstract

Navigate This Article