The Physiology, Signaling, and Pharmacology of Dopamine Receptors

1. Amplification of Protein Kinase A Signaling by DARPP-32.

Several substrates of PKA, such as CREB, ionotropic glutamate receptors (AMPA and NMDA), and certain ion channels, have been shown to be affected by dopamine receptor stimulation (Greengard, 2001). Among the PKA substrates, the 32-kDa dopamine and cAMP-regulated phosphoprotein (DARPP-32) is one of the most extensively studied molecules involved in dopamine receptor signaling. DARPP-32 is a multifunctional phosphoprotein that is predominantly expressed in MSNs. In these cells, DARPP-32 acts as an integrator involved in the modulation of cell signaling in response to multiple neurotransmitters, including dopamine (Svenningsson et al., 2004). Phosphorylation of this molecule at threonine 34 by PKA or protein kinase G activates the protein phosphatase 1 (PP1) inhibitory function of DARPP-32 (Hemmings et al., 1984a,b). The direct regulation of DARPP-32 phosphorylation at threonine 34 by dopamine receptors has recently been demonstrated in vivo using BAC transgenic mice that overexpressed immunoprecipatable-tagged DARPP-32 proteins specifically in D1 or D2 dopamine receptor positive MSNs (Bateup et al., 2008). The results from studies conducted using these mice have shown that enhanced dopamine receptor stimulation results in an increased phosphorylation of DARPP-32 in response to PKA activation in D1 dopamine receptor-expressing neurons, whereas stimulation of D2 dopamine receptors reduces the phosphorylation of DARPP-32 at threonine 34, presumably as a consequence of a reduction in PKA activation (Bateup et al., 2008) and/or the dephosphorylation of threonine 34 by the calmodulin-dependent protein phosphatase 2B (PP2B; also known as calcineurin) that is activated by increased intracellular Ca²⁺ after activation of D2 dopamine receptors (Nishi et al., 1997).

In contrast to PKA, cyclin-dependent kinase 5 (CDK5) has been shown to phosphorylate DARPP-32 at threonine 75, preventing the inhibition of PP1 by DARPP-32 and converting DARPP-32 to an inhibitor of PKA (Bibb et al., 1999). Furthermore, DARPP-32 is also phosphorylated at serine 137 by casein kinase 1 (CK1) (Desdouits et al., 1995) and at serine 97/102 by casein kinase 2 (CK2) (Girault et al., 1989a). The phosphorylation of DARPP-32 by CK1 reduces threonine 34 phosphorylation, whereas CK2 enhances the phosphorylation of DARPP-32 by PKA. Furthermore, the dephosphorylation of serine 97/102 by protein phosphatase 2A (PP2A) also results in an accumulation of DARPP-32 in the nucleus, where it prevents the dephosphorylation of histone H3 by PP1 and leads to enhanced gene expression in response to D1 dopamine receptor stimulation (Stipanovich et al., 2008).

Studies of DARPP-32 function in vivo and in striatal slice preparations have shown that this regulator acts as an amplification mechanism for PKA signaling. In MSNs, the phosphorylation state of multiple PKA targets, such as ionotropic glutamate and GABA receptors, is the result of an equilibrium between PKA and PP1 activity (Greengard et al., 1999; Greengard, 2001), and similar mechanisms can have important effects on the regulation of synaptic plasticity by dopamine receptors in other brain areas, such as frontal cortex (Xu et al., 2009). By inhibiting PP1, DARPP-32 tips this equilibrium toward the phosphorylated state and enhances the efficacy of PKA-mediated signaling. Because the kinases that regulate DARRP-32 can be activated in response to multiple hormones, neuropeptides, and neurotransmitters, it is not surprising that DARPP-32 has been shown to coordinate different signaling modalities (Greengard, 2001). In agreement with this finding, modulation of the inhibition of PP1 by DARPP-32 in response to nondopaminergic drugs, such as cannabinoids (Andersson et al., 2005) and caffeine (Lindskog et al., 2002), can affect PKA signaling responses after the activation of dopamine receptors and can modulate dopamine-associated behaviors.

Mice lacking DARPP-32 display deficits in their responses to dopaminergic drugs, such as cocaine (Fienberg et al., 1998), which elevates the extracellular concentration of dopamine by blocking the dopamine transporter (DAT). Similar results were obtained using knockin mice that expressed a mutant DARPP-32 lacking threonine 34 (Svenningsson et al., 2003; Zhang et al., 2006). Taken together, these findings confirm the role of phospho-threonine 34 DARPP-32 in the development of dopamine-dependent behaviors. However, a thorough ethological characterization of DARPP-32 knockout mice showed that several dopamine-associated behaviors, including basal locomotion, are not overtly disrupted in these mice (Nally et al., 2003, 2004; Waddington et al., 2005). By comparison, observations made in dopamine-depleted animals have shown that basal locomotion is essentially abolished in the absence of the stimulation of dopamine receptors (Zhou and Palmiter, 1995; Sotnikova et al., 2005). Moreover, behavioral responses to drugs that change dopamine levels, such as apomorphine and cocaine, are only partially affected in DARPP-32 knockout mice (Fienberg et al., 1998; Nally et al., 2004). For example, DARPP-32 knockout mice have been shown to exhibit a deficiency in their acute locomotor response only to intermediate doses (i.e.,10 mg/kg) of cocaine, whereas their responses to higher drug doses (i.e., 20 mg/kg) were essentially intact (Fienberg et al., 1998). Recent characterization of dopamine-mediated behaviors in mice lacking DARRP-32 in D1 dopamine receptor-expressing neurons revealed a slight reduction in spontaneous locomotion and cocaine-induced hyperactivity (Bateup et al., 2010). In contrast, the specific ablation of DARPP-32 in neurons that express D2 dopamine receptors resulted in enhanced locomotor responses to cocaine and an enhanced level of basal locomotor activity (Bateup et al., 2010). This latter observation is interesting because mice lacking D2 dopamine receptors or the predominantly postsynaptic D2L variant have been shown to display an overall reduction in basal locomotor activity and in their responsiveness to cocaine (Baik et al., 1995; Vargas-Pérez et al., 2004; Doi et al., 2006; Welter et al., 2007). These findings suggest that a reduction in phospho-threonine 34 DARPP-32 in response to D2 dopamine receptor stimulation may contribute to some of the pro-locomotor effects that are associated with D2 dopamine receptor activation. In general, the persistence of dopamine behaviors in DARPP-32 knockout mice suggests that although DARPP-32 functions as an important modulator of dopamine receptor signaling, it is not the only modulator/effector of dopamine-related actions, and its activity may be compensated for by other signaling mechanisms.

The behavioral and physiological impacts of the phosphorylation of DARPP-32 by CDK5 have also been studied. Repeated exposure to cocaine in normal rats and the persistent elevation of extracellular dopamine levels in mice that lack the dopamine transporter (DAT-KO mice) enhance the expression of CDK5 and its coactivator p35 in the striatum (Bibb et al., 2001; Cyr et al., 2003). Moreover, transgenic mice that overexpress the transcription factor ΔFosB, which is accumulated after long-term over-stimulation of D1 dopamine receptors in DAT-KO or cocaine-treated animals, also displayed increased CDK5 expression, providing a mechanism for the regulation of the CDK5 gene expression by dopamine (Bibb et al., 2001; Cyr et al., 2003). Enhanced CDK5 expression has been shown to antagonize the development of cocaine sensitization (Bibb et al., 2001); however, whether this effect is due to a postsynaptic inhibition of PKA via phospho-threonine 75 DARPP-32 or to a presynaptic function of CDK5 in the regulation of dopamine release remains controversial (Chergui et al., 2004).

2. Metabotropic Neurotoxicity of cAMP-Mediated Dopamine Receptor Signaling.

Increased CDK5 expression in response to D1 dopamine receptor stimulation may also be associated with the degeneration of MSNs in response to excessive dopamine receptor stimulation. A subpopulation (30%) of hyperdopaminergic DAT-KO mice have been shown to sporadically develop progressive locomotor dysfunctions characterized by a loss of the locomotor hyperactivity that is typical of DAT-KO mice and by the development of dyskinesia, paralysis, and death (Cyr et al., 2003). This motor disorder resembles a phenotype that has been observed in the lines of transgenic mice that express the variants of huntingtin that are associated with Huntington's disease (Levine et al., 2004), suggesting that MSNs were affected in the DAT-KO mice that developed these symptoms (symptomatic DAT-KO). An examination of the nigrostriatal system of the symptomatic mice revealed a specific loss of ∼30% of the MSNs, a phenomenon that was not observed in the DAT-KO mice that did not develop locomotor dysfunctions (Cyr et al., 2003). Furthermore, terminal deoxynucleotidyl transferase dUTP nick-end labeling (TUNEL)-positive MSNs containing detectable levels of activated caspase 3 were also observed, indicating an activation of the apoptotic processes in these cells. The symptomatic mice displayed an enhanced level of CDK5 and p35 expression compared with their nonsymptomatic littermates. This up-regulation of CDK5 was accompanied by the accumulation of a hyperphosphorylated form of the microtubule-associated τ protein (Cyr et al., 2003). Hyperphosphorylation of τ on residues that are phosphorylated by CDK5 has been shown to be associated with Alzheimer's disease (Patrick et al., 1999; Dhavan and Tsai, 2001) and other neurodegenerative disorders, including amyotrophic lateral sclerosis and Huntington's disease (Jellinger, 1998; Nguyen et al., 2001; Beaulieu and Julien, 2003).

An increase in the level of dopamine neurotransmission has been shown to enhance the motor and neuropathological phenotypes in a knockin mouse model expressing a mutant form of huntingtin that contains 92 CAG repeats (Cyr et al., 2006). In addition, activation of the D1 dopamine receptor accelerates the formation of mutant huntingtin nuclear aggregates in SK-N-MC neuroblastoma cells (Robinson et al., 2008). Overstimulation of ionotropic neurotransmitter receptors, in particular glutamate receptors, is well known to exert adverse excitotoxic effects (Choi, 1987). Taken together, these observations concerning the neurotoxicity that is caused by the overstimulation of D1 dopamine receptors suggest that the overstimulation of metabotropic receptors might have “metabotoxic” effects that render neurons more susceptible to other insults, such as those caused by the expression of mutant huntingtin.

3. Coincidence Detection by Mitogen-Activated Protein Kinases.

The complexity of the signaling network that is regulated by cAMP downstream from dopamine receptors also provides mechanisms for a context-dependent regulation of cellular responses to dopamine. For example, it is possible that many of the consequences of dopamine receptor stimulation occur only under specific conditions that require the coactivation of other types of receptors. These types of cellular responses that are dependent on the co-occurrence of different forms of stimulation are designated coincidence detectors and are known to play a central role in the regulation of synaptic plasticity. MAP kinases have been shown to act as important coincidence detectors that integrate the actions of dopamine with those of other neurotransmitter systems.

Many MAP kinases have been shown to be signaling intermediates that are involved in the regulation of dopamine-associated behaviors (Berhow et al., 1996). Results obtained from heterologous cell culture systems suggest that both D1- and D2-class dopamine receptors can regulate the MAP kinases extracellular-signal regulated kinases 1 and 2 (ERK1 and ERK2) (Beom et al., 2004; Chen et al., 2004b; Wang et al., 2005; Kim et al., 2006). In vivo, the administration of amphetamine (Beaulieu et al., 2006; Valjent et al., 2006b), cocaine (Berhow et al., 1996; Valjent et al., 2000), or the D2-class dopamine receptor antagonist haloperidol (Pozzi et al., 2003) has been shown to enhance striatal ERK1 and ERK2 phosphorylation. Moreover, ERK2 phosphorylation has also been shown to increase in response to elevated dopamine levels in the striatum of DAT-KO mice (Beaulieu et al., 2006). Pharmacological and genetic characterizations of the dopamine receptor types that are involved in the activation of striatal ERK have revealed that D1 dopamine receptors are essential for the activation of these kinases in MSNs (Valjent et al., 2000, 2005; Zhang et al., 2004; Liu et al., 2006a). In contrast, D2-class dopamine receptors, in particular D3 dopamine receptors, have been shown to mediate the inhibition of ERK-mediated signaling in the striatum (Zhang et al., 2004).

The activation of ERK by D1 dopamine receptors in response to cocaine administration has been shown to be dependent upon the stimulation of ionotropic glutamate receptors and to be preventable by the NMDA receptor antagonist known as MK-801 (Valjent et al., 2005). The regulation of ERK phosphorylation by D1 dopamine receptors and NMDA receptors may result from the convergence of several signaling modalities that are regulated by these two types of receptors. NMDA receptor stimulation results in the activation of the ERK kinase MEK. In the absence of D1 dopamine receptor stimulation, the action of MEK on ERK activity is counteracted by the activity of striatal-enriched tyrosine phosphatase (STEP), resulting in a zero-sum equilibrium in which the overall activity of ERK remains unchanged. The activity of STEP depends on its dephosphorylation by PP1. Activation of the D1 dopamine receptor/PKA/DARRP-32 signaling cascade leads to the inactivation of PP1 and the consequent inactivation of STEP (Valjent et al., 2000, 2005). This modification of the equilibrium of STEP and MEK activity after D1 dopamine receptor stimulation allows for the activation of ERK by MEK (Fig. 2A). The coregulation of ERK by D1 dopamine receptors and NMDA receptors has indicated that ERK might act as a signal integrator for dopamine and glutamate neurotransmission during the development of the behavioral responses to drugs of abuse.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 2.

Signaling networks regulated by dopamine in D1-class dopamine receptor responding neurons. A, regulation of Gα_s/cAMP/PKA signaling by D1-class dopamine receptors. B, regulation of Gα_q/PLC signaling by D1-class dopamine receptors or D1:D2 dopamine receptor heterodimers. The action of other neurotransmitters, growth factors, and neurotrophins has been included to illustrate the role of these intermediates as signal integrators. Blue arrows indicate activation, red T-arrows indicate inhibition, and green arrows indicate the amplification of an already activated function. Black arrows indicate actions that can either be activatory or inhibitory on the function of specific substrates. D1R, D1 dopamine receptor; D2R, D2 dopamine receptor; D5R, D5 dopamine rceptor. For a detailed description of these different signaling pathways, see sections IV.A and IV.B.

An inhibition of brain ERK activity has been shown to affect short- and long-term responses to psychostimulants. At high doses, the administration of the blood-brain barrier-permeant MEK inhibitor α-[amino[(4-aminophenyl)thio]methylene]-2-(trifluoromethyl)benzeneacetonitrile (SL327) results in a massive reduction in ERK phosphorylation and antagonizes the locomotor actions of cocaine and amphetamine (Valjent et al., 2000, 2006b; Beaulieu et al., 2006). At low doses, this drug does not affect short-term locomotor responses, but it does interfere with the development of long-term changes in behavioral and synaptic plasticity that are associated with neuronal adaptation and addiction to drugs of abuse (Valjent et al., 2000, 2006b). The mechanism underlying these long-lasting behavioral outcomes is not fully understood. Multiple groups have reported that ERK-mediated signaling is essential for the activation of the transcription factors CREB, Zif268, and c-Fos and for the phosphorylation of histone H3 in response to cocaine (Valjent et al., 2000, 2006a; Brami-Cherrier et al., 2005; Miller and Marshall, 2005). Activation of this cascade is believed to be associated with the development of long-lasting changes in gene expression and in the development of behavioral adaptations. The epigenetic regulation of gene expression by ERK may result in part from the activation of the mitogen- and stress-activated protein kinase 1 in the dorsal striatum and the nucleus accumbens (Brami-Cherrier et al., 2005). Although experiments involving the local administration of MEK inhibitors have shown that ERK signaling in the nucleus accumbens is important for the development of long-lasting adaptations to cocaine (Miller and Marshall, 2005), ERK signaling in other brain regions, such as the amygdala (Lu et al., 2005; Radwanska et al., 2005), also seems to play a role in these phenomena.

In addition, long-lasting regulation of epigenetic mechanisms might contribute to the development of unwanted side effects in response to treatment with l-DOPA (Santini et al., 2007, 2009a). In 6-hydroxydopamine-lesioned/l-DOPA-treated mice, DARPP-32-dependent ERK activation is associated with the development of dyskinesia, whereas inhibition of ERK signaling during long-term l-DOPA treatment counteracts the induction of dyskinesia (Santini et al., 2007). It is noteworthy that ERK activation in dyskinetic mice was accompanied with changes in mitogen- and stress-activated protein kinase 1 activation and histone H3 phosphorylation that are reminiscent of responses to D1 dopamine receptor-mediated ERK activation after repeated cocaine administration (Santini et al., 2007, 2009a). However, in addition to ERK signaling, other types of cell-signaling mechanisms that regulate gene expression and synaptic plasticity, such as D1 dopamine receptor-mediated activation of the mTOR complex 1, also seem to be involved in dyskinesia (Santini et al., 2009b). Understanding these mechanisms and their role in dyskinesia and addiction may be important for controlling the unwanted side effects of drugs that act on dopamine neurotransmission.

Modulation of ERK signaling has also been implicated in the regulation of hyperactivity resulting from psychostimulants in DAT-KO mice (Beaulieu et al., 2006). In normal animals, psychostimulants such as amphetamine and methylphenidate increase extracellular levels of dopamine, leading to the development of locomotor hyperactivity and to an augmented activation of ERK. Increased extracellular dopamine levels also result in an elevation in the level of ERK activity and in the development of a hyperactive phenotype in DAT-KO mice (Giros et al., 1996; Gainetdinov et al., 1999b; Beaulieu et al., 2004). However, the administration of psychostimulants in these mice results in a reduction of locomotor activity that is reminiscent of the action of psychostimulants in ADHD (Gainetdinov et al., 1999b; Beaulieu et al., 2006). This paradoxical effect of psychostimulants in the DAT-KO mice can also be mimicked by drugs that enhance serotonergic neurotransmission, suggesting that psychostimulants act through a serotonergic mechanism in DAT-KO mice (Gainetdinov et al., 1999b). A characterization of the signaling mechanisms that are affected by amphetamine and methylphenidate revealed that, in contrast to their effects in normal control mice, psychostimulants inhibit ERK in the striatum of DAT-KO mice (Beaulieu et al., 2006). It is noteworthy that the administration of the serotonergic drugs fluoxetine and nonselective serotonin agonist 5-carboxamidotryptamine has also been shown to result in an inhibition of striatal ERK in DAT-KO mice. Furthermore, administration of ERK inhibitor SL327 recapitulated the actions of psychostimulants in DAT-KO animals, whereas it had no effect on locomotion in wild-type (WT) mice that were treated with the NMDA receptor antagonist dizocilpine maleate (MK-801) (Beaulieu et al., 2006). These data suggest that ERK can integrate the actions of dopamine and serotonin at the level of striatal MSNs. Moreover, these observations raise the possibility that the inhibitory action of psychostimulants on dopamine-dependent hyperactivity results from an altered regulation of striatal ERK phosphorylation. However, further studies will be needed to clarify whether this paradoxical action of psychostimulants on ERK-mediated signaling can provide a mechanism of action of these drugs for the management of ADHD.

4. Interaction with Epac Proteins.

In addition to their effects on PKA, dopamine receptors may also exert physiological actions by acting on other cAMP-regulated molecules. Exchange proteins that are directly activated by cAMP (Epac1 and Epac2) are cAMP-regulated signaling proteins that are highly enriched in the striatum (Gloerich and Bos, 2010). Recent evidence suggests that Epac2 is involved in D1 dopamine receptor-mediated synapse remodeling and depression in cultured rat cortical neurons (Woolfrey et al., 2009). In addition to this finding, there are also reports of cAMP-dependent effects of dopamine receptor signaling that seems to be independent of PKA and Epac (exchange protein directly activated by cAMP) activity (Podda et al., 2010), suggesting that other cAMP-responsive molecules are effectors of dopamine receptor signaling.

1. Receptor Signaling through Gα_q.

There is evidence that, in addition to their effects on cAMP-regulated signaling, dopamine receptors can also couple to Gα_q to regulate phospholipase C (PLC) (Fig. 2B). Activation of PLC leads to the production of inositol trisphosphate (IP₃) and diacylglycerol (DAG). This activation results in the activation of PKC by DAG and an increased mobilization of intracellular calcium in response to IP₃. Although several lines of evidence indicate that dopamine can regulate PKC activity or intracellular calcium signaling (Felder et al., 1989; Friedman et al., 1997; Lee et al., 2004; Sahu et al., 2009), the exact mechanisms involved in this effect are still controversial. As early as 1989, Felder et al. reported that the D1 dopamine receptor agonist SKF 82526 stimulates PLC activity independently of cAMP in renal tubular membranes. It is noteworthy that this result of D1-class receptor activation has been observed in D1-KO mice (Friedman et al., 1997) but not in mice lacking the D5 dopamine receptor (Sahu et al., 2009). Furthermore, the expression of D1 dopamine receptors in transfected HEK293 cells did not affect intracellular calcium signaling, whereas the expression of D5 dopamine receptors in the same cells induced extensive calcium mobilization after stimulation (So et al., 2009). Although these observations do not make it possible to exclude completely the contribution of D1 dopamine receptors in Gα_q-mediated signaling, they suggest either that the D5 dopamine receptor is the main regulator of this signaling in vivo or that the D1 dopamine receptor needs to interact with other proteins to couple to Gα_q.

The molecular determinants that regulate the alternative coupling of D1-class receptors to Gα_q are still not completely understood. Among the possible mechanisms, it has been postulated that proteins that interact with dopamine receptors may regulate the coupling of these receptors to different G proteins in vivo (Bergson et al., 2003) or that the activation of dopamine receptors may enhance the signaling of other GPCRs through the PLC/IP₃ pathway (Dai et al., 2008). Alternatively, it has also been shown that D1/D2 dopamine receptor heterodimers can regulate DAG and IP₃ signaling in transfected cells (Lee et al., 2004). The physiological relevance of this observation has remained controversial because studies conducted in BAC-transgenic mice that express fluorescent gene-reporter proteins have shown that D1 and D2 dopamine receptors are not coexpressed in most striatal neurons in mice (Heiman et al., 2008; Shuen et al., 2008; Valjent et al., 2009). However, the coexpression of D1 and D2 dopamine receptors has been reported after immunohistological detection in the nucleus accumbens of 8-month-old mice (Rashid et al., 2007) in addition to several other locations in the basal ganglia (Perreault et al., 2010). There is also evidence for the regulation of BDNF production through a calcium-dependent D1 dopamine receptor signaling mechanism in the nucleus accumbens of adult rats (Hasbi et al., 2009) and for a potential contribution of D1/D2 dopamine receptor heterodimers in the regulation of dopamine-responsive behaviors in adult rats (Perreault et al., 2010). Finally, a recent study of dopamine receptor expression in the prefrontal cortex of BAC transgenic mice that express fluorescent gene-reporter proteins in neurons with D1 and D2 dopamine receptors showed that most D1 dopamine receptor-positive pyramidal neurons in the prefrontal cortex also express low levels of D2 dopamine receptors (Zhang et al., 2010). Therefore, it is possible that D1/D2 dopamine receptor heterodimers occur in vivo and regulate calcium-dependent cell signaling in some neuronal populations (Fig. 2B).

2. Regulation of Gβγ Signaling and Ion Channels by D2 Dopamine Receptors.

D2-class receptors also modulate intracellular calcium levels (Fig. 3, A and B) by acting on ion channels or by triggering the release of intracellular calcium stores (Nishi et al., 1997; Missale et al., 1998). Some of these actions of D2-class receptors on calcium signaling seem to be related to signaling responses that are mediated by the Gβγ subunits of heterotrimeric G proteins, which are separated from Gα subunits after receptor activation. Gβγ subunits that are regulated by D2 dopamine receptors have been shown to activate PLC and to increase the cytoplasmic calcium concentration in MSNs (Hernandez-Lopez et al., 2000). It is noteworthy that this mechanism also reduces the level of activity of the L-type calcium channels in these cells, indicating that stimulation of D2 dopamine receptors has complex effects on calcium-mediated biological processes in these neurons (Fig. 3B). It is noteworthy that D2 dopamine receptor-regulated Gβγ subunits are also involved in the regulation of N-type calcium channels in striatal interneurons (Yan et al., 1997). Finally, the role of the Gβγ subunits is not limited to calcium channels. These proteins are also involved in the association/regulation of the D2 dopamine receptor and G protein-coupled inwardly rectifying potassium channels (GIRKs) (Kuzhikandathil et al., 1998; Lavine et al., 2002). The activation of GIRKs after activation of the D2 dopamine receptor (Fig. 3B) and other GPCRs has an inhibitory effect in neurons and could potentially mediate several functions of dopamine in vivo. For a recent review of the potential implications of the activation of GIRKs in human pathologic conditions, see Lüscher and Slesinger (2010).

• Download figure
• Open in new tab
• Download powerpoint

Fig. 3.

Signaling networks regulated by dopamine in D2-class dopamine receptor responding neurons. A, regulation of Gα_s/cAMP/PKA signaling by D2-class receptors. B, regulation of Gβγ signaling by D2-class receptors. C, regulation of β-arrestin 2/Akt/GSK-3 signaling by D2 and D3 dopamine receptors. The action of other neurotransmitters, growth factors, and neurotrophin has been included to illustrate the role of many of these intermediates as signal integrators. Blue arrows indicate activation, red T-arrows indicate inhibition, and green arrows indicate the amplification of an already activated function. Black arrows indicate actions that can either be activatory or inhibitory on the function of specific substrates. PDK1, phosphatidylinositol-dependent kinase 1; Pi3K, phosphatidylinositol 3-kinase; RTK, receptor tyrosine kinase. For a detailed description of these different signaling pathways, see sections IV.A, IV.B, IV.D, and IV.F.

1. Evidence for the Involvement of Regulators of G Protein Signaling.

The regulation of G protein-mediated dopamine receptor signaling by RGSs has been studied extensively both in vitro and in vivo. Compelling evidence exists that indicates that the RGS9-2 subtype, which is particularly enriched in the striatum, plays an important role in the regulation of D2 dopamine receptor signaling. This regulation was clearly demonstrated in direct biochemical and electrophysiological experiments in heterologous cellular systems and striatal neurons (Granneman et al., 1998; Cabrera-Vera et al., 2004; Kovoor et al., 2005; Seeman et al., 2007; Martemyanov and Arshavsky, 2009). Furthermore, the functional role of this regulation was observed in studies that showed an altered regulation of RGS9-2 protein levels after short-term or repeated exposure to psychostimulants and D2 dopamine receptor ligands in normal animals (Seeman et al., 2007). In addition, this finding was supported by the demonstration of the enhanced psychomotor and rewarding effects of psychostimulants in mice lacking RGS9-2 (Rahman et al., 2003; Traynor et al., 2009). In contrast, animals with an increased expression of RGS9-2 have been shown to exhibit a diminished response to psychostimulants, D2 (but not D1) dopamine receptor agonists and to undesirable dyskinetic effects of long-term l-DOPA treatment in experimental models of PD (Gold et al., 2007; Traynor et al., 2009). The results of a recent report have indicated that the function of striatal RGS9-2 is controlled by its association with an additional subunit, R7BP, and have demonstrated motor coordination deficits and locomotor supersensitivity to morphine, but not cocaine, in mice lacking R7BP. In addition, the sensitivity of locomotor stimulation to cocaine seemed to depend on RGS7 because its interaction with R7BP is dictated by RGS9-2 expression. Thus, a cooperative role in the regulation of dopamine signaling in the striatum has been proposed for two RGS proteins, RGS7 and RGS9-2, which are balanced by a common subunit R7BP (Anderson et al., 2010). Other members of the RGS family that are enriched in the primary dopaminergic areas of the brain, RGS2 and RGS4, may also contribute to the regulation of dopamine receptors; however, the evidence supporting this role has not yet been confirmed through direct testing in vivo (Burchett et al., 1999; Taymans et al., 2003, 2004; Schwendt et al., 2006; Anderson et al., 2010).

2. Evidence for Additional Regulatory Mechanisms.

Other regulatory proteins have been identified as potential modulators of D2 dopamine receptor cAMP-mediated signaling, such as prostate apoptosis response-4 (Par-4), a leucine zipper-containing protein that plays a role in apoptosis (Park et al., 2005). The results of this study have shown that recombinant Par-4 directly interacts with D2 dopamine receptors via a calmodulin-binding motif that is situated in the third cytoplasmic loop of the receptor. A reduction in Par-4 expression in a heterologous cell system or in cultured neurons from knockin mice expressing a mutant Par-4 that lacked an interaction domain for the D2 dopamine receptor resulted in a reduction of D2 dopamine receptor signaling via cAMP. Last, knockin mice expressing a Par-4 mutant that does not interact with D2 dopamine receptors have been shown to display behavioral abnormalities in tests that model depression in rodents (Park et al., 2005). However, these behavioral paradigms have been traditionally associated with serotonergic or adrenergic, rather than dopaminergic, neurotransmission (Xu et al., 2000; Cervo et al., 2005; Crowley et al., 2005). In the absence of a characterization of the actions of dopaminergic drugs in Par-4 mutant mice, the role of this molecule in the regulation of dopamine signaling in vivo is not yet completely understood.

1. Interactions with Calcium Channels.

D1 dopamine receptors and N-type calcium channels are highly expressed and colocalized on apical dendrites in the prefrontal cortex. A recent study of the regulation of N-type calcium channels by D1 dopamine receptors showed that the second intracellular loop of D1 dopamine receptors can directly interact with the COOH terminal region of N-type calcium channel Cav2.2 subunits in vitro. Furthermore, these full-length molecules also interact in native tissue, and the activation of D1 dopamine receptors results in a protein interaction-dependent inhibition/internalization of N-type calcium channels in the prefrontal cortex (Kisilevsky et al., 2008). A similar interaction of N-type calcium channels with potentially similar effects on the regulation of N-type calcium channels has also been reported for transfected cells (Kisilevsky and Zamponi, 2008).

2. Direct Interactions with Ionotropic Receptors.

Dopamine receptors can also interact directly with ionotropic glutamate and GABA receptors. The COOH terminal domain of the D1 dopamine receptor has been shown to interact with the NR1 and the NR2A subunit of the glutamatergic NMDA receptor in transfected cells and hippocampal cultured neurons (Lee et al., 2002; Fiorentini et al., 2003). The D2 dopamine receptor also interacts with the NR2B subunit of the NMDA receptor in response to cocaine in the postsynaptic density microdomain of excitatory synapses in striatal neurons (Liu et al., 2006b). Finally, the COOH terminal domain of the D5 dopamine receptor has been shown to interact with the second intracellular loop of the GABA-A γ2-receptor subunit in the rat hippocampus (Liu et al., 2000). It is noteworthy that the interaction of dopamine receptors with GABA-A receptors seems to be specific to the D5 dopamine receptor, because this interaction does not occur for the D1 dopamine receptor.

The functional consequences of these interactions of dopamine and ionotropic receptor functions are diverse. The interaction of the D5 dopamine receptor with GABA-A reduces GABA-A receptor-mediated whole-cell currents (Liu et al., 2000). Likewise, the interaction of the D2 dopamine receptor with the NR2B NMDA receptor subunit disrupts the association of the NMDA receptor with CaMKII and inhibits the phosphorylation of this receptor and the resulting NMDA receptor-mediated currents (Liu et al., 2006b). Finally, the diverse interactions of the D1 dopamine receptor with the NMDA receptor have several functional consequences for NMDA receptor functions and appear to be important in regulating working memory in rats (Nai et al., 2010).

Initial observations indicate that the disruption of the D1/NR1 interaction after D1 dopamine receptor activation allows for the recruitment of calmodulin and phosphatidylinositol 3-kinases to NR1, resulting in the activation of cell survival mechanisms. In addition, the interaction of the D1 dopamine receptor with NR2A inhibits NMDA currents by reducing cell-surface expression (Lee et al., 2002). However, the functional interactions between the D1 receptor and the NMDA receptor are complex (Yao et al., 2008). In addition to these consequences of physical receptor interactions, activation of the D1 dopamine receptor has been shown to increase NMDA receptor activation or otherwise affect synaptic plasticity through mechanisms that involve protein kinases (Dunah and Standaert, 2001; Flores-Hernández et al., 2002; Chen et al., 2004a) or dopamine receptor interactions at a neural circuit level (Xu and Yao, 2010).

There is also evidence that the interaction with NMDA receptors affects D1 dopamine receptor cell surface expression and trafficking, resulting in increased plasma membrane insertion of D1 dopamine receptors (Fiorentini et al., 2003; Pei et al., 2004). However, this action of the NMDA receptor on the D1 dopamine receptor is probably restricted to specific neuronal populations in vivo because mice that exhibit a 95% reduction in the expression of the NR1 NMDA receptor subunit (Mohn et al., 1999) have normal striatal D1 dopamine receptor functioning (Ramsey et al., 2008).

The interaction of the D1 dopamine receptor and the NMDA receptor is also regulated by proteins that associate, either directly or indirectly, with both of these receptor types. For example, the synaptic scaffolding protein PSD-95 not only plays a role in stabilizing glutamate receptors in the postsynaptic density but is also down-regulated in several mouse models of psychostimulant or dopamine supersensitivity (Yao et al., 2004). It is noteworthy that PSD-95 directly interacts with the COOH terminal of D1 dopamine receptors to reduce its expression at the cell surface through enhanced dynamin-dependent endocytosis (Zhang et al., 2007). This increased internalization of D1 dopamine receptors in the presence of PSD-95 correlates with a reduced D1-NMDA receptor interaction (Zhang et al., 2009). This interplay between NMDA receptors, D1 dopamine receptors, and PSD-95 supports a model for the regulation of D1 dopamine receptor expression at the cell surface. Its interaction with NMDA receptors would anchor the D1 dopamine receptors at the cell surface and would interfere with their internalization after activation. In turn, competition between PSD-95 and NMDA receptors for interaction with the COOH terminal of D1 dopamine receptors might modulate internalization by dissociating D1 dopamine receptors from NMDA receptors, thereby liberating the D1 dopamine receptors and allowing for their internalization (Zhang et al., 2009). This type of interaction between receptors, ion channels, and their respective or shared associated proteins provides ample possibilities for the regulation of receptor functions in response to local conditions at the subcellular level.

1. G Protein-Coupled Receptor Kinases.

GRK2 is widely expressed in the brain, including primary dopaminergic areas (Arriza et al., 1992) and can phosphorylate and regulate D1, D2, and D3 dopamine receptors in in vitro cellular systems (Tiberi et al., 1996; Iwata et al., 1999; Kim et al., 2001; Kabbani et al., 2002; Sedaghat et al., 2006; Banday et al., 2007). Indirect evidence for the role of this kinase in the regulation of dopamine receptors was provided by the observation of alterations in the GRK2 expression levels in the striatum of animals with experimental parkinsonism (Bezard et al., 2005), rats treated with antipsychotics (Ahmed et al., 2008b), and rats receiving long-term cocaine treatment (Schroeder et al., 2009). Whereas a deletion of the GRK2 gene in mice results in embryonic lethality, presumably as a result of cardiac hypoplasia (Jaber et al., 1996), mice heterozygous for this mutation are viable and were used to investigate the role of GRK2 in dopamine receptor functioning (Gainetdinov et al., 2004). Locomotor effects of the psychostimulant amphetamine, which interacts with the DAT to elevate extracellular dopamine and thus indirectly induces dopamine receptor stimulation (Jones et al., 1998), and the direct D1/D2 dopamine receptor agonist apomorphine were tested in GRK2 heterozygous (HET) mice. These treatments caused similar behavioral responses in the mutant and WT mice, suggesting that the impact of a partial loss of GRK2 on dopamine receptor-mediated responses is limited (Gainetdinov et al., 2004). However, cocaine treatment caused a minor enhancement in locomotor activity in the mutants at only one dosage, suggesting that some populations of dopamine receptors may be supersensitive. A more pronounced level of GRK2 deficiency might be necessary to demonstrate a potential involvement of this kinase in dopamine receptor regulation in physiological settings, and future studies with mouse lines bearing region- or cell-specific deletions of GRK2 in the brain would be particularly informative. Although it is unclear at present which dopamine receptor subtype(s) could be regulated by GRK2, recent evidence suggests that the fragile X mental retardation protein interacts with GRK2 to cause alterations in the regulation and function of the D1 dopamine receptor (Wang et al., 2008). Furthermore, the neuronal calcium sensor-1 (NCS-1) has been shown to directly interact with GRK2 to modulate the desensitization of D2 dopamine receptors, suggesting that the abnormalities in NCS-1 expression described in schizophrenia and bipolar disorder could cause changes in the NCS-1-dependent GRK2 regulation of the D2 dopamine receptor (Kabbani et al., 2002). In addition, the possibility of a constitutive regulation of D2 dopamine receptor expression and signaling by GRK2 that may be independent of receptor phosphorylation has been demonstrated (Namkung et al., 2009). Further detailed investigations are necessary to test these potential mechanisms in in vivo physiological settings.

Similar to GRK2, GRK3 has a ubiquitous pattern of expression in the brain, and in vitro studies using cell culture have indicated a potential role of GRK3 in the regulation of D1, D2, and D3 dopamine receptors (Tiberi et al., 1996; Kim et al., 2001; Kabbani et al., 2002; Sedaghat et al., 2006). Alterations in the GRK3 expression in several brain areas were observed after the long-term treatment of rats with the antipsychotics haloperidol and clozapine or in experimental animal models of PD (Ahmed et al., 2008a,b). In mice lacking GRK3, significantly reduced locomotor responses to cocaine and apomorphine were observed, indicating that this kinase is unlikely to be involved in the desensitization of locomotion-controlling dopamine receptors but rather is “positively” involved in dopamine receptor signaling. Alternatively, GRK3 may be involved in the desensitization of D3 dopamine autoreceptors, thereby providing negative presynaptic control of dopamine release (Kim et al., 2005).

Given the limited expression of GRK4 in the brain, it is not surprising that no evidence for the role of this kinase in the regulation of neuronal dopamine receptors has been found (Gainetdinov et al., 2004). However, the role of GRK4 in D1 dopamine receptor regulation, which has important physiological consequences, has been convincingly demonstrated in the kidney. It has been observed that constitutive activity of polymorphic variants of the GRK4 protein that are observed in patients with hypertension can lead to perpetually phosphorylated, inactive D1 dopamine receptors (Felder et al., 2002). Likewise, in vitro cellular systems transfected with GRK4 variants display marked constitutive activity toward the D1 dopamine receptor (Rankin et al., 2006). These and other (Staessen et al., 2008) observations have demonstrated that the hypertension-associated polymorphic variants of GRK4 that are constitutively active lead to persistent D1 dopamine receptor desensitization in the kidney and a loss of dopamine-dependent salt and fluid excretion. Recent evidence also indicates that D3 dopamine receptors in the kidney are phosphorylated and regulated by GRK4 and that this regulation is critical for their signaling in human proximal tubule cells (Villar et al., 2009).

One widely expressed member of the GRK4 subfamily of GRKs, GRK5 (Premont et al., 1995, 1999), can regulate D1 and D2 dopamine receptors in cellular model systems (Tiberi et al., 1996; Iwata et al., 1999; Sedaghat et al., 2006). It has also been reported that long-term treatment with cocaine causes an up-regulation of GRK5 mRNA in the septum (Erdtmann-Vourliotis et al., 2001), and significant alterations in expression have been found in experimental PD animals (Ahmed et al., 2008a), suggesting a possible involvement of this kinase in the regulation of receptor responsiveness to dopaminergic drugs. However, a direct assessment of the potential role of GRK5 in the regulation of dopamine receptors in mice that lack GRK5 revealed that these animals exhibited normal responses to several dopaminergic drugs, indicating that the regulation of dopamine receptors is not affected by a deletion of GRK5. In contrast, significantly altered muscarinic receptor-mediated responses were observed in these mice, indicating a prominent role of this kinase in M2 muscarinic receptor regulation (Gainetdinov et al., 1999a).

GRK6 is highly expressed in many brain areas and in the periphery (Fehr et al., 1997). It is noteworthy that GRK6 seems to be the most prominent GRK in the striatum and in other dopaminergic brain areas (Erdtmann-Vourliotis et al., 2001). A particularly high level of expression of the GRK6 protein was found in the dopamine-receptive GABA-ergic MSNs and in the cholinergic interneurons in the striatum of mice (Gainetdinov et al., 2003). A direct analysis of dopaminergic functioning in mice that lack GRK6 revealed a significantly enhanced responsiveness to primarily dopaminergic psychostimulant drugs and direct dopamine agonists (Gainetdinov et al., 2003; Raehal et al., 2009). Furthermore, an increased level of coupling of striatal D2 dopamine receptors to G proteins and an increased affinity of D2 dopamine receptors for these G proteins was observed in these mice compared with WT mice (Gainetdinov et al., 2003; Seeman et al., 2005). These observations have revealed that D2 dopamine receptors are physiological targets for GRK6-mediated regulation. In addition, these observations suggest that a strategy aimed at a modulation of GRK6 expression or activity could be beneficial in pathological conditions in which dopamine signaling is affected, such as PD. In fact, it has been reported that GRK6 expression was significantly altered in the striatum in animal models of PD and that this expression was sensitive to l-DOPA treatment (Bezard et al., 2005). Intriguingly, lentivirus-mediated overexpression of GRK6 in the striatum in rodent and primate models of Parkinson's disease can alleviate the dyskinesia in these animals that is caused by long-term l-DOPA treatment. At the same time, a reduction in the GRK6 concentration caused by the administration of microRNA with lentiviral vectors increased the level of dyskinesia in parkinsonian rats compared with untreated rats by enhancing the responsiveness of D1 and D2 dopamine receptors (Ahmed et al., 2010). These observations suggest promising new approaches for controlling dyskinesia and motor fluctuations in Parkinson's disease, thereby providing evidence for the potential effectiveness of a therapeutic strategy that aims to target receptor regulatory mechanisms.

2. β-Arrestins.

The nonvisual arrestins, β-arrestins 1 and 2, are ubiquitously expressed in essentially all tissues (Arriza et al., 1992), although β-arrestin 1 is 10- to 20-fold more concentrated in brain compared with β-arrestin 2 (Gurevich et al., 2002). The majority of the evidence that supports the role of β-arrestin 1 in the regulation of D1 and D2 dopamine receptors has been reported in in vitro cellular studies (Kim et al., 2001; Oakley et al., 2001). Striatal β-arrestin 1 levels have been shown to be modulated by dopamine depletion and l-DOPA treatment (Bezard et al., 2005). Mice lacking this regulatory protein appear normal and do not show obvious abnormalities (Conner et al., 1997). However, the locomotor effects of cocaine and apomorphine have been found to be reduced in these mice compared with WT mice (Gainetdinov et al., 2004), suggesting that β-arrestin 1 may be involved in G protein-independent dopamine receptor signaling events rather than being directly involved in dopamine receptor desensitization. Further studies are necessary to investigate this possibility in detail.

Investigations performed in β-arrestin 2 knockout mice have directly demonstrated that the β-arrestin 2 isoform has signaling functions in vivo (Beaulieu et al., 2009). In heterologous cellular systems, β-arrestin 2 has been shown to regulate many GPCRs, including D1, D2, and D3 dopamine receptors (Kim et al., 2001; Oakley et al., 2001; Gainetdinov et al., 2004; Lan et al., 2009a,b). In an initial study, β-arrestin 2-deficient mice were found to exhibit enhanced morphine-induced analgesia compared with WT mice, which indicates that β-arrestin 2 plays an important role in μ-opioid receptor desensitization (Bohn et al., 1999, 2000, 2003, 2004). However, the effects of amphetamine and apomorphine (Gainetdinov et al., 2004; Beaulieu et al., 2005), but not cocaine (Bohn et al., 2003), were reduced in β-arrestin 2 knockout mice compared with WT mice. Furthermore, mice lacking both β-arrestin 2 and DAT have been shown to display a reduction in the typical novelty-induced locomotor hyperactivity phenotype that is characteristic of DAT-KO mice (Beaulieu et al., 2005). These findings indicated that β-arrestin 2 plays a positive role in dopamine receptor regulation. In fact, subsequent in vivo biochemical studies have directly demonstrated that β-arrestin 2 plays a critical role in D2 dopamine receptor signaling (Fig. 1) by regulating the Akt/glycogen synthase kinase 3 (GSK-3) pathway (Beaulieu et al., 2005; Beaulieu et al., 2007a) (see section IV.F for details).

Taken together, these studies have revealed a significant degree of specificity in the regulation of dopamine receptors by GRKs and β-arrestins in vivo (Table 2). However, although a significant amount of data have been collected on the regulatory mechanisms affecting D2 dopamine receptor function, relatively little information exists about the regulation of other receptor subtypes in vivo. It has become evident that the regulation of any receptor subtype is extremely complex, with multiple mechanisms of regulation that may occur simultaneously or consecutively. For example, a significant body of evidence, from in vitro and in vivo studies, indicates that the D2 dopamine receptor can be regulated via heterologous desensitization (Rogue et al., 1990; Namkung and Sibley, 2004), desensitized homologously in a GRK6-dependent manner (Gainetdinov et al., 2003), or involved in a G protein-independent, β-arrestin 2-mediated Akt/GSK-3 signaling cascade (Beaulieu et al., 2009). Furthermore, RGS9-2 mediates the regulation of D2 dopamine receptor-activated G protein signaling (Traynor et al., 2009). Understanding the logic and dynamics of this complexity will require detailed studies in native tissues with naturally expressed cellular complements of the desensitization/regulation machinery that involve in vivo analysis of relevant physiological output. Another important question relates to the fact that no evidence of an arresting role of β-arrestin 1 or β-arrestin 2 on dopaminergic signaling has been observed so far in in vivo functional studies. In addition, whereas an absolute deficiency of both β-arrestin 1 and β-arrestin 2 was found to be embryonic lethal in mice (Kohout et al., 2001), three-allele knockout [β-arrestin 1 KO/β-arrestin 2 HET, and β-arrestin 1 HET/β-arrestin 2 KO] mice exhibited reduced responses to apomorphine compared with WT mice (Gainetdinov et al., 2004), suggesting predominant roles of both β-arrestins as signaling hubs in dopamine receptor signaling (Luttrell and Lefkowitz, 2002; Beaulieu et al., 2007a). Intriguingly, in addition to the established contribution of β-arrestin 2 to the regulation of Akt-mediated signaling responses by D2 dopamine receptor (Beaulieu et al., 2005), a recent investigation has suggested that β-arrestin 2 may also be involved in the regulation of ERK signaling by D1 dopamine receptors (Urs et al., 2011). Further delineation of the negative and positive roles of the β-arrestins in dopamine-mediated functions represents a major challenge for future research on dopamine receptor regulation and signaling.

1. Role of the β-Arrestin 2/Akt/Glycogen Synthase Kinase 3 Pathway in D2 Dopamine Receptor-Mediated Functions.

The reduced behavioral responsiveness to dopamine or to drugs that enhance dopamine neurotransmission in β-arrestin 2 knockout mice compared with their WT littermates points toward a role of β-arrestin-containing protein complexes in dopamine receptor signaling. Recent investigations have identified one of these mechanisms and have shown that β-arrestin 2 is a signaling intermediate that is implicated in the cAMP-independent regulation of Akt and GSK-3 by dopamine (Fig. 3C) (Beaulieu et al., 2004, 2005, 2006).

Akt, also termed protein kinase B, is a serine/threonine kinase that is regulated through phosphatidylinositol-mediated signaling. The activation of Akt involves the recruitment of this kinase to the plasma membrane by phosphorylated phosphatidylinositol, phosphorylation at a regulatory threonine residue (threonine 308) by the phosphatidylinositol-dependent kinase 1 and further phosphorylation at another regulatory residue (serine 473) by the rictor-mammalian target of rapamycin (mTOR) protein complexes (Jacinto et al., 2006). Regulation of Akt by phosphorylated phosphatidylinositol has been associated with the action of insulin, insulin-related peptides (e.g., insulin-like growth factor), and neurotrophins (e.g., nerve growth factor, BDNF, and neurotrophin-3) that exert their biological function by stimulating receptor tyrosine kinase (Scheid and Woodgett, 2001). GSK-3α and -3β are two closely related serine/threonine kinases that were originally associated with the regulation of glycogen synthesis in response to insulin (Embi et al., 1980; Frame and Cohen, 2001). These kinases can be inactivated through the phosphorylation of single serine residues, serine 21 (GSK-3α) and serine 9 (GSK-3β), in their regulatory NH₂-terminal domain (Frame and Cohen, 2001). Akt has been shown to inhibit GSK-3α and -3β in response to multiple hormones and growth factors including insulin, insulin-like growth factor, and BDNF (Cross et al., 1995; Frame and Cohen, 2001; Chen and Russo-Neustadt, 2005).

Investigations of altered cell signaling in response to persistently elevated extracellular dopamine levels have identified a reduction in Akt phosphorylation/activity in the striatum of DAT-KO mice compared with WT mice (Beaulieu et al., 2004; Beaulieu et al., 2006). Inactivation of Akt in these mice resulted in concomitant activation of GSK-3α and -3β (Beaulieu et al., 2004). Further characterization of these signaling responses using dopamine depletion or dopamine receptor antagonists [7-chloro-3-methyl-1-phenyl-1,2,4,5-tetrahydro-3-benzazepin-8-ol (SCH23390) and raclopride] in DAT-KO mice has revealed that Akt, GSK-3α, and GSK-3β are regulated by D2-class receptors in these mice (Beaulieu et al., 2004; Sotnikova et al., 2005). Furthermore, the D2 dopamine receptor antagonist and antipsychotic haloperidol has been independently shown to stimulate Akt phosphorylation and to inhibit GSK-3 in normal animals (Emamian et al., 2004). Pharmacological activation of dopamine receptors also results in the modulation of Akt and GSK-3 activity. In addition, administration of amphetamine, an indirect dopamine receptor agonist that induces potent increases in extracellular dopamine (Jones et al., 1998), led to an inhibition of Akt activity and an activation of GSK-3 in healthy C57BL6 mice (Beaulieu et al., 2004). A similar reduction in Akt activity was also reported after treatment of WT mice with the direct D1/D2 dopamine receptor agonist apomorphine, confirming that dopamine regulates the Akt/GSK-3 pathway in vivo (Beaulieu et al., 2004, 2005). However, neither Akt nor GSK-3 activity was affected by direct modulation of cAMP levels in the striatum of mice, indicating that this signaling pathway is not controlled by cAMP (Beaulieu et al., 2004). In contrast, amphetamine and apomorphine administered to β-arrestin 2 KO mice failed to reduce Akt phosphorylation (Beaulieu et al., 2005). Furthermore, mice lacking both β-arrestin 2 and DAT showed no inhibitory action of excessive dopamine on Akt phosphorylation, demonstrating that dopamine receptors regulate Akt through β-arrestin 2 (Beaulieu et al., 2005). Further investigations of the mechanism through which β-arrestin 2 regulates Akt in response to dopamine have shown that stimulation of the D2-class receptors causes the formation of a protein complex (Figs. 1, 3C, and 4) that is composed of Akt, β-arrestin 2, and PP2A and facilitates the dephosphorylation/deactivation (Andjelković et al., 1996) of Akt by PP2A in response to dopamine (Beaulieu et al., 2004, 2005).

Apart from the deficits in the dopamine-associated behaviors observed in β-arrestin 2 KO mice, data from multiple behavioral assays also support the involvement of the β-arrestin 2/Akt/GSK-3 pathway in the regulation of dopamine-associated behaviors. Inhibition of GSK-3 activity has been shown to reduce locomotor hyperactivity in DAT-KO mice and in amphetamine-treated WT animals (Beaulieu et al., 2004; Gould et al., 2004). Confirmation of these pharmacological observations was obtained using genetically engineered animals. GSK-3β KO mice die during embryogenesis, whereas GSK-3β heterozygote mice develop normally without overt phenotypes (Hoeflich et al., 2000). An evaluation of the behavioral actions of amphetamines revealed that GSK-3β heterozygote mice are less responsive to amphetamine over a range of doses than WT mice, supporting the involvement of GSK-3β in the development of dopamine-associated behaviors (Beaulieu et al., 2004). In addition, transgenic mice expressing a GSK-3β mutant that lacks an inhibitory phosphorylation site develop a locomotor hyperactivity phenotype that is reminiscent of DAT-KO mice (Prickaerts et al., 2006), whereas knockin mice lacking GSK-3α and a β inhibitory phosphorylation site are hyper-responsive to amphetamine (Polter et al., 2010). Finally, mice lacking the Akt isoform Akt1 exhibit an enhanced disruption of sensory motor gating (prepulse inhibition) after amphetamine but not after the glutamate NMDA receptor blocker MK-801 compared with WT mice (Emamian et al., 2004). The disruption of sensory motor gating by amphetamine has been used as a behavioral paradigm to model psychosis in rodents and is efficiently blocked by antipsychotics such as haloperidol. Because Akt1 is inhibited after the stimulation of D2-class receptors (Beaulieu et al., 2007b), the increased behavioral effect of amphetamine in Akt1 KO mice compared with WT mice further supports the involvement of Akt inhibition in the development of dopamine-related behavioral responses. However, dopamine regulates more than locomotor activity and sensory motor gating (Carlsson, 1987). Further behavioral characterization of different dopamine-associated behavioral paradigms that use traditional or tissue-specific knockout mice should be conducted to further elucidate the functions of β-arrestin 2, Akt, and GSK-3 in the expression of dopamine-associated behaviors.

2. Two Signaling Modalities of Slow Synaptic Transmission.

β-arrestin- and G protein-mediated responses to GPCR stimulation are characterized by different kinetics, suggesting that dopamine receptor signaling may be composed of two separate modalities that develop differently over time. In cultured fibroblasts, β-arrestin-mediated ERK signaling has a slower onset and a more prolonged duration than G protein-mediated GPCR signaling (Ahn et al., 2004). In addition, the β-arrestin 2-dependent inhibition of Akt by dopamine in the mouse striatum displays a very different kinetic timeline (Fig. 1) compared with the signaling events that are regulated by the cAMP/PKA pathway (Beaulieu et al., 2004, 2005). After administration of dopaminergic drugs, such as amphetamine and cocaine, the cAMP-dependent phosphorylation of ERK and DARPP-32 peaks and returns to baseline within the first 30 min (Valjent et al., 2000 2005; Svenningsson et al., 2003). In contrast, β-arrestin 2-dependent inhibition of Akt by amphetamine develops progressively over the first 30 to 60 min of the drug's action (Beaulieu et al., 2004, 2005). It is noteworthy that no difference in the behavioral response to amphetamine was observed when mutant mice lacking one copy of the GSK-3β gene and WT mice were compared over the first 30 min after drug administration. However, after this time point, the intensity of the locomotor activation was significantly reduced in the mutant animals. This difference persisted until the behavioral drug effects were no longer observed at 2 h after administration of the drug (Beaulieu et al., 2004). This finding suggests that the regulation and maintenance of a subset of dopamine-associated behaviors and the action of drugs might depend upon two complementary phases of GPCR signaling responses (Fig. 1). The early phase of this response would involve cAMP-mediated and other G protein-mediated responses that have a rapid onset and desensitization. In contrast, the second or late phase of the response would be characterized by β-arrestin-mediated cell signaling that has a slower onset, a longer duration, and no known desensitization mechanism (Fig. 1). β-Arrestin-mediated mechanisms seem to be a key determinant to separate these two phases of cellular responses in slow synaptic transmission because these mechanisms are responsible for both the termination of the early phase and the initiation of the late phase of the receptor signaling responses. G protein and β-arrestin 2 signaling modalities may also provide mechanisms for neurons to distinguish between bursts of repeated neurotransmitter stimulation and prolonged, continuous stimulation of their receptors by dopamine. As a result, the early and late phases of slow synaptic transmission may be compatible with differential responsiveness to rapid/phasic and long-lasting/tonic changes in dopamine concentrations, respectively, and may thus provide molecular mechanisms that support signaling responses to different modalities of dopamine neurotransmission.

3. Glycogen Synthase Kinase-3 Targets Involved in the Effects of Dopamine.

GSK-3 has multiple substrates, including proteins involved in diverse cellular processes (Cohen and Frame, 2001; Frame and Cohen, 2001; Scheid and Woodgett, 2001; Woodgett, 2001). A thorough review of the role of GSK-3 substrates in the regulation of behavior can be found in the literature (Beaulieu et al., 2009; Karam et al., 2010). Direct in vivo evidence supports the role of GSK-3 in the regulation of at least two important brain functions by dopamine (Fig. 4). Ionotropic glutamate receptors are heteromultimeric ion channels for which trafficking and expression are closely related to the regulation of synaptic plasticity (Sheng and Hoogenraad, 2007). Dopamine is known to regulate ionotropic glutamate receptors and their associated proteins in neurons (Esteban et al., 2003; Svenningsson et al., 2004; Yao et al., 2004). Inhibition of GSK-3 prevents the development of long-term depression (LTD) in rat hippocampal slices (Peineau et al., 2007, 2008), whereas GSK-3 activation reduces the expression of the NMDA receptor subunits NR2A/B on the cell surface in hippocampal slices and cultured cortical neurons (Chen et al., 2007; Zhu et al., 2007). In agreement with this finding, GSK-3 activity has recently been shown to be required for the inhibition of synaptic NMDA receptor functioning by D2 dopamine receptors in the rat prefrontal cortex (Li et al., 2009), suggesting a role for GSK-3-mediated D2 dopamine receptor signaling in the regulation of NMDA receptor synaptic plasticity by dopamine. In addition, recent evidence indicates that GSK-3 also contributes to the regulation of LTD by disrupting the normal transport of AMPA receptors at the synapse through the kinesin cargo system (Du et al., 2010). Although the contribution of D2 dopamine receptors to this new type of AMPA receptor regulation has not yet been studied, these results suggest an exciting new mechanism through which these receptors can contribute to the regulation of mood and synaptic plasticity. Overall, the results obtained concerning the regulation of LTD and the inhibition of glutamate receptor functions by GSK-3 isoforms also suggest that activation of these kinases in response to D2 receptor stimulation in a subpopulation of MSNs may prevent the depolarization of these neurons in response to glutamate. The results of a recent study using optogenetic stimulation of MSNs that express D2 receptors (i.e., MSNs involved in the indirect striatopalidal pathway) showed that the depolarization of these neurons reduces locomotion in mice compared with controls (Kravitz et al., 2010). This suggests that the inhibition of excitatory glutamate receptors by GSK-3 in response to dopamine in MSNs expressing D2 dopamine receptors could stimulate locomotion.

• Download figure
• Open in new tab
• Download powerpoint

Fig. 4.

Regulation of Akt/GSK-3 signaling by psychoactive drugs and related network of gene products associated with mental disorders. Proteins labeled in orange are the product of genes associated with an increased risk of developing schizophrenia and/or bipolar disorders. Blue arrows indicate activation, and the red T-arrows indicate inhibition. Black arrows indicate actions that can either be activatory or inhibitory on the function of specific substrates. Behavioral changes in dopaminergic responses have been reported in Akt1 and β-arrestin 2 KO mice and in GSK-3β-HET mice. COMT, catechol-O-methyltransferase; D2R, D2 dopamine receptor; Disc1, Disrupted-in-Schizophrenia 1; NRG1, neuregulin 1; PDK1, phosphatidylinositol-dependent kinase 1; Pi3K, phosphatidylinositol 3-kinase.

D2 dopamine receptor signaling through GSK-3 may also contribute to the regulation of circadian responses by dopamine. The results from studies conducted using mice lacking D2 dopamine receptors have revealed a role for this receptor in the regulation of circadian behavior by light (Doi et al., 2006). Furthermore, D2 dopamine receptor signaling enhances the transcriptional capacity of the circadian transcription factor CLOCK-BMAL1 complex, and the activity of this complex in the regulation of mPer1 gene transcription in response to light is reduced in the retinas of D2-KO mice compared with WT mice (Yujnovsky et al., 2006). The phosphorylation of BMAL1 by GSK-3 results in the ubiquitination and degradation of this transcription factor, and reduced GSK-3 activity in D2-KO mice (Beaulieu et al., 2007b) correlates with an increase in BMAL1 protein levels in the striatum of these mice (Sahar et al., 2010). These findings strongly suggest that molecules involved in circadian regulation might also represent downstream targets of GSK-3-mediated D2 dopamine receptor signaling that are involved in the regulation of behavior. It is noteworthy that such regulation of circadian mechanisms might also contribute to long-lasting changes in gene expression that are involved in the regulation of synaptic plasticity (Borrelli et al., 2008). However, the regulation of other signaling cascades, including MAP kinases, also seems to play an important role in the complex mechanisms through which the D2 dopamine receptor contributes to the regulation of circadian rhythm functions (Yujnovsky et al., 2006).