Behavioral sensitization to amphetamine results from an uncoupling between noradrenergic and serotonergic neurons Lucas Salomon*, Christophe Lanteri*, Jacques Glowinski, and Jean-Pol Tassin† Institut National de la Sante´et de la Recherche Me´dicale Unite´114, Colle`ge de France, 11, Place Marcelin Berthelot, 75231 Paris Cedex 05, France Edited by Richard D. Palmiter, University of Washington School of Medicine, Seattle, WA, and approved March 23, 2006 (received for review February 1, 2006) In rodents, drugs of abuse induce locomotor hyperactivity, and locomotor stimulation and behavioral sensitization are entirely repeating injections enhances this response. This effect, called dependent on the stimulation of two nondopaminergic mono- behavioral sensitization, persists many months after the last ad- aminergic receptors, ␣1b-adrenergic and 5-HT2A (11). Knockout ministration, thus mimicking long-term sensitivity to drugs ob- mice for the ␣1b-adrenergic receptor (␣1b-AR KO) (12) and the served in human addicts. We show here that, in naı¨ve animals, antagonists of ␣1-adrenergic and 5-HT2A receptors (prazosin noradrenergic and serotonergic systems, besides their behavioral and SR46349B, respectively) were used to define ␣1b-adrenergic activating effects, inhibit each other by means of the stimulation and 5-HT2A components in drug-induced locomotor activity. We ␣ of 1b-adrenergic and 5-HT2A receptors and that this mutual have shown that prazosin blocks most of the morphine-evoked inhibition vanishes with repeated injections of d-amphetamine; locomotor response in WT mice and that, as expected, mor- this uncoupling may be responsible for behavioral sensitization phine-evoked locomotor response in ␣1b-AR KO mice was not and for an increased reactivity of dopaminergic neurons. First, after affected by prazosin (11). Surprisingly, morphine-evoked loco- repeated d-amphetamine injections, a d-amphetamine challenge motor response was 3-fold higher in ␣1b-AR KO mice than in induces a dramatic increase in cortical extracellular norepinephrine WT mice when both species were treated with prazosin (11). (NE) levels. This increased cortical NE release still occurs after 1 Because SR46349B entirely inhibited morphine-induced loco- month of withdrawal but is diminished or blocked if sensitization motor response in ␣1b-AR KO mice, it was suggested that is performed in the presence of prazosin, SR46349B, or both 5-HT receptors could compensate for the genetic deletion of ␣ 2A 1-adrenergic and 5-HT2A receptor antagonists, respectively. A ␣1b-adrenergic receptors (11). However, when both species were strong correlation between increases in cortical extracellular NE repeatedly treated with morphine, morphine-evoked locomotor levels and the expression of behavioral sensitization was found. response in presence of prazosin increased in WT mice and Second, repeated d-amphetamine injections induce an increased became similar to that observed in ␣1b-AR KO mice (11). This reactivity of serotonergic neurons measured by cortical extracel- finding suggests that prazosin limits the 5-HT2A component of lular serotonin (5-HT) levels after the administration of a 5-HT morphine-evoked locomotor activity in naı¨ve WT mice and that ␣ releaser, p-chloroamphetamine. Third, knockout mice for 1b- this limitation disappears when animals are sensitized. A possi- ␣ adrenergic ( 1b-AR KO) or 5-HT2A (5-HT2A-R KO) receptor, respec- bility could be that, in addition to their behavioral activating tively, exhibit a behavioral and biochemical hyperreactivity to the effects, noradrenergic and serotonergic systems are coupled (i.e., ␣ acute injection of p-chloroamphetamine ( 1b-AR KO; 5-HT levels) limit or stimulate each other) in naı¨ve animals and become and d-amphetamine (5-HT2A-R KO; NE levels). Uncoupling between independent after repeated injections of psychostimulants or noradrenergic and serotonergic neurons may occur not only in opiates, explaining accordingly the development of behavioral addiction but also during chronic stressful situations, thus facili- sensitization. tating the onset of mental illness. To test the first hypothesis, i.e., whether 5-HT2A receptors compensate for the genetic deletion of ␣1b-adrenergic receptors d-amphetamine ͉ microdialysis ͉ norepinephrine ͉ serotonin ͉ behavioral and also whether ␣1b-adrenergic receptors compensate in 5-HT sensitization 2A receptor knockout mice (5-HT2A-R KO) (13), densities of ␣1b- adrenergic and 5-HT2A binding sites were measured by autoradiog- sychostimulants and opiates, two major groups of drugs of raphy in 5-HT2A-R KO and ␣1b-AR KO mice and compared with Pabuse, produce locomotor stimulant effects that become those of WT mice. Binding sites were studied in the prefrontal enhanced with repeated intermittent injections. This enhanced cortex (PFC) because of the role of cortical ␣1b-adrenergic recep- behavioral response, named behavioral sensitization, is enduring tors in stimulant-induced locomotor activity (14, 15) and because and can last up to 1 year after drug exposure (1). Studies of the both receptors are colocalized in layers III–V in this structure (14, neurobiological basis of behavioral sensitization have focused, 16). The absence of obvious interactions between ␣1b-adrenergic despite conflicting data (2–6), on the midbrain dopamine (DA) and 5-HT2A receptors led us to study, as mentioned above, the system because of evidence suggesting that this system mediates possibility of a modification of a mutual relationship between locomotor stimulation as well as the ability of drugs to elicit noradrenergic and serotonergic neurons after repeated treatments craving and to lead to abuse (7). Indeed, it was established that most drugs abused by humans increase DA release in the nucleus accumbens, a structure innervated by midbrain DA neurons (8). Conflict of interest statement: No conflicts declared. Moreover, animals readily self-administer agents that increase This paper was submitted directly (Track II) to the PNAS office. DA transmission, such as amphetamine and cocaine (9). Fur- Abbreviations: DA, dopamine; NE, norepinephrine; 5-HT, serotonin; PCA, p-chloroamphet- thermore, it has been proposed that the rewarding properties of amine; VTA, ventral tegmental area; PFC, prefrontal cortex; ␣1b-AR KO, knockout for the ␣ opiates, such as morphine or heroin, are produced by the 1b-adrenergic receptor; 5-HT2A-R KO, knockout for the 5-HT2A receptor. disinhibition of midbrain DA cells firing via the stimulation of *L.S. and C.L. contributed equally to this work. ␮-opiate receptors located on GABAergic midbrain interneu- †To whom correspondence should be addressed. E-mail: jean-pol.tassin@college-de- rons that negatively regulate DA cells firing (10). Recently, france.fr. however, we have shown that psychostimulant͞opiate-induced © 2006 by The National Academy of Sciences of the USA 7476–7481 ͉ PNAS ͉ May 9, 2006 ͉ vol. 103 ͉ no. 19 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0600839103 Downloaded by guest on September 28, 2021 Fig. 1. Autoradiography of ␣1b-adrenergic and 5-HT2A binding sites in ␣1b-AR KO and 5-HT2A-R KO mice and a comparison with WT mice. The higher amplitude of the 5-HT2A locomotor component observed in ␣1b-AR KO mice (11), when compared with WT, is not due to an increased density in 5-HT2A receptor binding sites. Similarly, the density of cortical ␣1b-adrenergic binding sites is not modified Ͻ in 5-HT2A-R KO mice. 3H Kt, tritiated ketanserin; 3H Pz, tritiated prazosin. ***, P 0.001 when compared with corresponding controls. Fig. 2. Repeated treatment with d-amphetamine induces both a behavioral sensitization and a potentiation in the increase of cortical extracellular NE levels that lasts for at least 1 month. Cortical extracellular NE levels are expressed as a with psychostimulants or opiates. We present here data obtained percentage of the respective mean basal value (acute, 0.64 Ϯ 0.1 pg of NE every with repeated d-amphetamine treatments. 20 min; repeated 4-days withdrawal, 0.55 Ϯ 0.09 pg every 20 min; 1-month First, mice locomotor responses and prefrontocortical extracel- withdrawal, 0.56 Ϯ 0.1 pg every 20 min). Effects of acute saline injection, which lular norepinephrine (NE) levels were determined in different did not significantly modify basal cortical extracellular NE levels, are not shown conditions, i.e., after acute and repeated treatments with d- for the sake of clarity. Each group contained at least seven animals to determine amphetamine, after 4 days or 1 month of withdrawal and when locomotor activity and five animals in microdialysis experiments. repeated d-amphetamine treatments were done in presence of prazosin, SR46349B, or a mixture of both antagonists. Then, ␣1b-adrenergic and 5-HT2A receptors, i.e., between noradren- extracellular prefrontocortical serotonin (5-HT) levels were deter- ergic and serotonergic neurons. mined in naı¨ve animals and in those having received repeated d-amphetamine treatments. However, because d-amphetamine did Effects of Acute and Repeated d-Amphetamine on d-Amphetamine- not modify cortical 5-HT extracellular levels in our experimental Induced Locomotor Responses and Cortical Extracellular NE Levels. conditions, reactivity of serotonergic neurons was estimated after Fig. 2 illustrates that, as expected, acute injection of d- the injection of a 5-HT releaser, p-chloroamphetamine (PCA), a amphetamine increases both locomotor activity [F(1,273) ϭ compound analogous to ecstasy and known to induce locomotor 166.6; P Ͻ 0.0001] and extracellular NE levels in the PFC hyperactivity and behavioral sensitization in mice (17). [ϩ167%; F(1,5) ϭ 172.7; P Ͻ 0.0001, when compared with basal NEUROSCIENCE Finally, locomotor and biochemical responses to d-amphetamine values] (18). After a 4-day withdrawal period, animals that ␣ or PCA were measured in 5-HT2A-R KO or 1b-AR KO mice, received repeated injections of d-amphetamine not only exhib- respectively. ited an important behavioral sensitization [ϩ241%; F(1,20) ϭ 315.1; P Ͻ 0.0001, when compared with acute] but also a Results dramatic and similar potentiation in extracellular NE levels ␣ Absence of Direct Interactions Between 1b-Adrenergic and 5-HT2A [ϩ259%; F(1,5) ϭ 116.9; P Ͻ 0.0001, when compared with Receptors.