INTRODUCTION

Neuronal G protein-coupled receptors (GPCRs) represent a large class of proteins that orchestrate a diverse array of physiological processes critical for maintaining normal brain function. Central to their regulation are a family of G protein-coupled receptor kinases (GRKs) that are actively engaged following agonist-mediated receptor activation and function in part to dampen intracellular signaling via homologous desensitization. Seven GRK family members have been identified to date, four of which (GRKs 2, 3, 5, and 6) are highly expressed and in diverse cellular populations in the brain (Premont and Gainetdinov, 2007). The physiological functions of GRKs 3–6 have been previously examined using isoform-specific KO mice to reveal critical roles for these kinases in the regulation of muscarinic, dopamine (DA), and opioid receptor function in adult animals (Gainetdinov et al, 2003; Gainetdinov et al, 1999; McLaughlin et al, 2004). In contrast, germline deletion of GRK2 in mice results in hypoplasia of the ventricular myocardium and embryonic lethality (Jaber et al, 1996). This has consequently hampered efforts to understand the physiological importance of GRK2 function in the adult brain, leaving relatively little known.

DA is a catecholamine neurotransmitter involved in a diverse array of functions in mammals, including the control of movement, addiction, emotion, and cognition. Consistent with its vital role as a neuromodulator, dopaminergic dysfunction has been implicated in various pathological states, such as Parkinson’s disease, Huntington’s disease, Tourette’s syndrome, and attention deficit and hyperactivity disorder (Beaulieu and Gainetdinov, 2011). Two distinct classes of DA-responsive GPCRs, D1-like (D1R and D5R) and D2-like (D2R, D3R, and D4R), are expressed within the basal ganglia and signal through opposing G-protein-dependent pathways following activation by DA (Bateup et al, 2008). Previous work from our lab also suggests that striatal D2 receptors signal in a G-protein-independent manner via the scaffolding protein βarrestin-2 to modulate striatal protein kinase B (Akt) and GSK3β signaling (Beaulieu et al, 2005; Beaulieu et al, 2007). Genetic and/or pharmacological disruption of downstream effectors engaged by DA receptors (eg, DA- and cAMP-regulated phosphoprotein of molecular weight 32 kDa (DARPP-32) and GSK3β) robustly alters motor behavior and sensitivity to antipsychotic and psychostimulant drugs (Bateup et al, 2010; Beaulieu et al, 2004; Urs et al, 2012). Thus, a more comprehensive understanding of proteins involved in DA receptor signaling may yield valuable mechanistic insights regarding several important clinically relevant drugs and provide novel therapeutic targets.

Several studies in heterologous cells have demonstrated that GRK2 can profoundly regulate the trafficking and agonist-induced signaling of D1 and D2 receptors (Ito et al, 1999; Iwata et al, 1999; Namkung et al, 2009a; Namkung et al, 2009b; Sedaghat and Tiberi, 2011; Tiberi et al, 1996). However, evidence for this regulation in vivo or even in systems with more physiologically relevant levels of GRK2 (eg, cultured neurons) remains scarce. Studies utilizing pharmacological interventions that perturb the DA system, such as the psychostimulant cocaine, or potent neurotoxins, such as MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine) and 6-hydroxydopamine, have found region-specific alterations in GRK2 levels (Ahmed et al, 2008a; Bezard et al, 2005; Schroeder et al, 2009). Additionally, it was found that fragile X mental retardation protein deficiency in mice causes a coincident increase in GRK2 levels in the frontal cortex and D1R hyperphosphorylation (Wang et al, 2008). Although these studies provide a potential link between GRK2 and the DA system, the physiological consequences of altered GRK2 levels to DA-dependent behaviors remain largely unknown.

To examine the possibility that GRK2 regulates DA-dependent behaviors, we generated mice with the conditional deletion of GRK2 in either DA-receptive neurons (D1R-, D2R-, and A2AR-expressing) or in midbrain DA neurons (DA transporter (DAT)-expressing neurons) and utilized behavioral and electrophysiological approaches to determine the impact of GRK2 deletion. Our results reveal cell-type-specific effects of GRK2 deletion on motor behavior, sensitivity to cocaine-induced locomotion, cocaine-induced striatal signaling, and DA uptake inhibition, DA release and uptake, and D2 autoreceptor function.

MATERIALS AND METHODS

Generation of Experimental Animals

All animals studies were conducted with an approved protocol from the Duke University Institutional Animal Care and Use Committee in accordance with the National Institutes of Health guidelines. Floxed Grk2 mice were obtained from Dr Gerald Dorn II (Washington University, St Louis, Missouri) (Matkovich et al, 2006) and Rosa26-floxed-stop-EYFP reporter mice were obtained from Dr Brigid Hogan (Duke University, Durham, NC) (Srinivas et al, 2001). The D1Rcre (founder line EY262), D2Rcre (founder line ER44), and A2ARcre (founder line KG139) bacterial artificial chromosome (BAC) transgenic mouse lines were purchased from GENSAT (The Rockefeller University). The DATcre knock-in mouse line was obtained from Dr Xiaoxi Zhuang (University of Chicago, Chicago, IL) and has been previously published (Zhuang et al, 2005). Experimental animals were obtained by crossing homozygous floxed Grk2 (129/C57BL/6J) mice to homozygous floxed mice that were also hemizygous (from BAC transgenic lines) or heterozygous (for the knock-in mouse line) for cre. These breedings yielded homozygous floxed Grk2 mice that were also cre-positive (KO animals) and floxed Grk2 littermates (control animals). A summary of the mouse lines generated can be found in Supplementary Table S1. Drug and experiment-naive adult mice (2–6 months of age) and of mixed sex were used for all studies. Animals were placed on a 12-h light/dark cycle in a temperature- and humidity-controlled environment with ad libitum access to food and water.

Drugs

Cocaine hydrocholoride, quinpirole hydrochloride, sulpiride, and apomorphine hydrochloride hemihydride were purchased from Sigma-Aldrich, and SKF81297 was purchased from Tocris Biosciences. All drugs were dissolved in isotonic, sterile saline and injected as described at a volume of 10 ml/kg body weight.

Locomotor Activity and Cocaine Sensitization

Locomotor activity was measured at 5-min intervals in an automated Omnitech Digiscan open-field (20 × 20 cm2) apparatus. Animals were tested during the light phase of the cycle and habituated to the room before testing. For the acute cocaine experiments, mice were habituated in the open field for 30 min, removed, injected with the indicated dose of cocaine, and immediately placed back into the apparatus. Sensitization experiments were performed as previously described (Gainetdinov et al, 2003). Briefly, on day 1 of the protocol, mice were injected with the indicated dose of cocaine and placed into the open field. On days 2–5, mice were injected with the same dose of cocaine and placed back into their home cage. After a 1-day reprieve from treatment, mice were challenged on day 7 with the same dose of cocaine and monitored in the open field for 90 min postinjection.

Conditioned Place Preference

Mice were tested using a three-compartment apparatus (Med Associates) as previously described (Daigle and Caron, 2012).

Fast-Scan Cyclic Voltammetry (FSCV)

Mice received one injection daily of either vehicle or cocaine (20 mg/kg, i.p.) for 5 consecutive days. Approximately 40 h after the last injection, mice were killed for slice voltammetry experiments. Note that while this is the same day that mice in the behavioral sensitization experiments received cocaine challenge and locomotor activity assessment, no injections were given to mice in the voltammetry study on day 7. Rather, all drugs were bath applied on brain slices as described in detail below. After killing, multiple coronal slices (400 μm) containing the NAc and CPu were prepared from each animal with a vibrating tissue slicer while immersed in oxygenated artificial cerebrospinal fluid (aCSF) containing (in mM): NaCl (126), KCl (2.5), NaH2PO4 (1.2), CaCl2 (2.4), MgCl2 (1.2), NaHCO3 (25), glucose (11), and L-ascorbic acid (0.4), and pH was adjusted to 7.4. Once sliced, tissue was transferred to the testing chambers containing bath aCSF (32 °C), which flowed at 1 ml/min. A cylindrical carbon fiber microelectrode (100–200 mM length, 7 mM radius) and a bipolar stimulating electrode was placed into the dorsolateral CPu, and DA was evoked by a single, rectangular, electrical pulse (750 μA, 4 ms) applied every 5 min. Extracellular DA was monitored at the carbon fiber electrode every 100 ms using FSCV by applying a triangular waveform (−0.4 to +1.2 to −0.4 V vs Ag/AgCl, 400 V/s) (Kennedy et al, 1992). Once the extracellular DA response was stable (ie, did not exceed 10 % variation in peak height for three successive stimulations), a single experiment was performed on each slice containing the CPu as described below.

After obtaining stable DA to single-pulse stimulation, multiple concentrations of cocaine (0.3, 1.0, 3.0, 10.0, and 30.0 μM) were bath applied cumulatively to the slice while continuing single-pulse stimulations every 5 min. Each concentration of cocaine was allowed to stabilize before adding the next concentration (typically 45 min at each concentration, or eight stimulations). For a separate set of multiple-pulse experiments without cocaine, stimulation parameters were adjusted from the single pulse to include five-pulse stimulations across frequencies that mimic both tonic and phasic firing of DA neurons, including 5, 10, 20, 25, and 100 Hz frequencies. Following the five-pulse series of stimulations, the D2R antagonist sulpiride (10 μM) was then added to the bath, and DA evoked by single-pulse stimulation was allowed to stabilize to the drug before repeating the multiple-pulse protocol. Immediately following the completion of each experiment, recording electrodes were calibrated by recording their response (in electrical current; nA) to a known concentration of DA in aCSF (3 mM) using a flow-injection system. This value was then used to convert electrical current to DA concentration. Evoked levels of DA were modeled using Michaelis–Menten kinetics, as a balance between release and uptake (Wightman et al, 1988). Michaelis–Menten modeling provides parameters that describe the amount of DA released following stimulation (ie, the peak height of the signal) and the maximal rate of DA uptake (Vmax). We followed standard voltammetric modeling procedures by setting the apparent Km parameter to 160 nM for each animal based on well-established research on the affinity of DA for the DAT (Wu et al, 2001), while baseline Vmax values were allowed to vary as the baseline measure of rate of DA uptake. All modeled data were required to exceed a goodness-of-fit ⩾R2=0.98, and all Vmax values fell within an acceptable range of physiological rates of uptake described in previous work. In experiments that incorporated the use of cocaine, the time constant tau (measured as 2/3 return to baseline from peak height) and peak height of the signal was calculated for each concentration of cocaine as a percentage of baseline tau or peak height in order to account for changes in cocaine-induce DA uptake inhibition and release. All voltammetry data were collected and modeled using the Demon Voltammetry and Analysis Software (Yorgason et al, 2011).

Immunohistochemistry

Staining of free-floating brain sections was performed as previously described (Daigle et al, 2011). Briefly, mice were anesthetized with chloral hydrate (500 mg/kg i.p.) and perfused with ice-cold PBS followed by 10% formalin solution (Sigma Aldrich). Brains were postfixed overnight in 10% formalin and sectioned (50 μm) on a vibratome. Following permeabilization and blocking, sections were incubated overnight at 4 °C in one or more of these antibodies: GRK2 (1 : 500; Santa Cruz Biotechnology), rabbit anti-GFP (1 : 5000; abcam), chicken anti-GFP (1 : 500; abcam), DAT (DA transporter) (1 : 300; Millipore), and met-Enkephalin (1 : 250; Millipore). Primary antibody was detected with Alexa dye-conjugated secondary antibodies (Life Technologies). Images (1024 × 1024 pixels) were acquired on either a Zeiss LSM-510 confocal microscope or an Olympus FluoView FV1000 confocal microscope.

Western Blotting Analyses

Determination of phosphoprotein levels in dorsal striatal extracts was performed as previously described (Daigle et al, 2011). Blots were probed with antibodies directed against pT308 Akt (1 : 100), Akt (1 : 1500), pS9 GSK3 (1 : 100), or GSK3 (1 : 500) (all from Cell Signaling Technology), pT34 DARPP-32 (1 : 300; Phosphosolutions), or DARPP-32 (1 : 1000; BD Transduction Laboratories). Primary antibodies and direct fluorescent signal from secondary antibodies were detected using an Odyssey Infrared imaging system (LI-COR Biosciences). Phosphoprotein levels were normalized to the corresponding level of total protein and then normalized to the average for the wild-type group (Grk2f/f) for each separate experiment.

Tyramide Signal Amplification (TSA) Immunohistochemistry

Free-floating brain sections were incubated for 5 min in PBS containing 0.5% NaBH2, 10% MeOH, and 3% H202. Sections were then washed in PBS followed by incubation for 15 min in PBS containing 1.2% Triton X-100, blocked in PBS containing 5% normal goat serum, 2% bovine serum albumin, and 0.2% Triton X-100 for 1 h, and then incubated overnight at 4 °C in the indicated antibody. Sections were incubated the following day in either an Alexa-dye-conjugated (Invitrogen) or an HRP-conjugated secondary antibody. The latter was detected by brief incubation with an Alexa488-labelled tyramide reagent (Invitrogen).

Data and Statistical Analyses

All statistical analyses were performed using the GraphPad Prism 4.0 software (GraphPad Software). The results presented are means with the SEMs. An unpaired Student’s t-test, one-way ANOVA, or two-way repeated measures (RM) ANOVA was used for the different comparisons as indicated in the text and figure legends.

RESULTS

Selective GRK2 Deletion Alters Spontaneous Locomotion and Causes a Bidirectional Change in Sensitivity to Cocaine

To determine whether GRK2 regulates DA receptor function in vivo and, in turn, DA-dependent behaviors, we generated conditional knockout (KO) mice in which GRK2 was selectively deleted from D1R- and D2R-expressing neurons in the murine brain. This was accomplished by crossing BAC transgenic mice in which Cre recombinase was driven under the D1R and D2R promoter elements (D1Rcre and D2Rcre mice, respectively) to mice with a floxed Grk2 locus (Grk2f/f) to yield conditional KO mice (D1RcreGrk2f/f and D2RcreGrk2f/f) and littermate controls (Grk2f/f). As D2Rs are expressed in cholinergic neurons within the striatum, striatopallidal neurons, and midbrain DA projection neurons, we expected GRK2 deletion in all of these subpopulations in D2RcreGrk2f/f mice. To confirm Cre-mediated recombination in the intended cellular populations, D1Rcre and D2Rcre mice were also crossed to Rosa26-EYFP (R26-EYFP) reporter mice, and the brains were fixed and processed for immunostaining against EYFP (anti-GFP) and met-Enkephalin (anti-Enk), a marker of striatopallidal MSNs (Supplementary Figures S1 and S2). We detected Cre-mediated recombination in the expected brain regions (including strong anti-GFP staining in striatonigral MSNs and cortical layer VI neurons for the D1Rcre line and in striatopallidal MSNs for the D2Rcre line), consistent with the known expression patterns for endogenous D1 and D2 DA receptors (Supplementary Figures S1A, 1B, S2A, and 2B). In addition, co-labeling of EYFP with Enk was observed in brain sections from D2RcreR26-EYFP but not D1RcreR26-EYFP mice (Supplementary Figures S1A, S1B S2A, and S2B). Deletion of GRK2 from D1R- and D2R-expressing neurons was also verified using an isoform-specific GRK2 antibody (Ahmed et al, 2008b) and an antibody directed against GFP. Here we observed, in part, deletions of GRK2 within the cortex in the D1R line (Supplementary Figure S1C) and within striatal cholinergic neurons (Supplementary Figure S2C) and DA neurons in the D2R line (data not shown). Collectively, these results confirm that D1RcreGrk2f/f and D2RcreGrk2f/f mice exhibit selective deletion of GRK2 in genetically defined D1R- and D2R-expressing cellular populations, respectively.

To investigate the role of GRK2 in DA-dependent behaviors, we first analyzed spontaneous locomotion and the effects of cocaine in mice with the conditional deletion of GRK2 from D1R-expressing neurons. Although no difference in baseline locomotor activity between genotypes was observed in drug-naive mice, the acute administration of cocaine to D1RcreGrk2f/f mice resulted in significantly enhanced psychomotor activation relative to control animals (interaction, F(23,897)=2.71, p<0.0001; genotype, F(1,897)=9.39, p=0.0039; Figure 1a and b). Analysis of the sum of the distance travelled 90 min postinjection of 10 or 20 mg/kg cocaine (i.p.) revealed a 2.5- and a 1.6-fold increase, respectively, in the activity of D1RcreGrk2f/f mice compared with control animals (Figure 1c). To determine whether GRK2 deficiency in D1R-expressing neurons altered sensitization to cocaine, we next evaluated the locomotor activity of D1RcreGrk2f/f mice following repeated, daily injections of cocaine (20 mg/kg, i.p.). As expected, D1RcreGrk2f/f mice were hypersensitive to the initial drug exposure, displaying increased activity on day 1 of the treatment regimen (interaction, F(23,506)=2.19, p=0.0012; genotype, F(1,506)=11.05, p=0.0031; Figure 1d, left and p=0.0039; Figure 1e). However, repeated cocaine injections produced robust sensitization in both genotypes, with no significant difference in activity observed between groups on day 7 (Figure 1d, right and Figure 1e). These data suggest that one function of GRK2 in D1R-expressing neurons may be to negatively regulate psychomotor activation induced by acute cocaine treatment.

Figure 1
figure 1

Deletion of GRK2 from D1R-expressing neurons promotes hypersensitivity to cocaine. (a) Spontaneous locomotor activity of control and D1RcreGrk2f/f mice (n=9 mice per group) analyzed in 30-min intervals over a 2-h period. (b) Locomotor activity of mice (Grk2f/f mice, n=20; D1RcreGrk2f/f mice, n=21) administered cocaine (20 mg/kg, i.p.). *p<0.05, **p<0.01, ***p<0.001, comparison of drug effect between genotypes by two-way RM ANOVA. (c) Cocaine dose-response for D1RcreGrk2f/f (n=6–21) and Grk2f/f (n=6–20) mice. **p<0.01, comparison relative to control mice by unpaired Student’s t-test. (d) Cocaine sensitization in D1RcreGrk2f/f and Grk2f/f mice (n=8–12 mice for each group). Locomotor activity of cocaine-treated mice on days 1 and 7 are shown. *p<0.05, **p<0.01, ***p<0.001 for comparison of the drug effect between genotypes by two-way RM ANOVA. (e) Cumulative distance traveled by cocaine-treated mice in the first 30 min postinjection of cocaine. ##p<0.01 for comparison of the drug effect between genotypes on day 1 and ***p<0.001 for comparison of the drug effect on day 1 vs day 7 by unpaired Student’s t-test.

PowerPoint slide

Full size image

In direct contrast to the D1R conditional KO line, D2RcreGrk2f/f mice exhibited increased spontaneous locomotor activity (genotype, F(1,102)=9.14, p=0.0047; Figure 2a) and reduced responsiveness to acute cocaine (20 mg/kg, i.p.) treatment (Figure 2b). Analysis of the fold increase in locomotor activity induced by cocaine (fold activation) revealed a robust reduction in responsiveness of D2RcreGrk2f/f mice to multiple independent doses of the stimulant relative to control animals (p<0.0001 and p=0.0033, 20 and 30 mg/kg doses, respectively; Figure 2b). A cocaine sensitization paradigm was used to determine whether D2RcreGrk2f/f mice could mount a response to a sensitizing regimen of cocaine (20 mg/kg, i.p.). This analysis revealed that D2RcreGrk2f/f mice were significantly less active following cocaine treatment on both days 1 and 7 relative to control animals (day 1: interaction, F(23,299)=4.02, p<0.0001 and day 7: interaction, F(23,299)=3.24, p<0.0001; genotype, F(1,299)=8.78, p=0.011; Figure 2c and d). Interestingly, peak activity (35 min time point) on day 1 vs day 7 for conditional KO mice was increased (393±77 vs 960±145 cm), indicating some degree of sensitization to the locomotor response. However, analysis of the total distance travelled 30 min postinjection revealed a significant reduction in activity of D2RcreGrk2f/f mice relative to control animals on both days (day 1, p=0.0167; and day 7, p=0.0036; Figure 2d). These results strongly suggest that GRK2 in D2R-expressing neurons has a role during spontaneous locomotion and is required for the locomotor-stimulating effects of cocaine.

Figure 2
figure 2

Deletion of GRK2 from D2R-expressing neurons enhances spontaneous locomotion and decreases sensitivity to cocaine. (a) Spontaneous locomotion of Grk2f/f (n=19) and D2RcreGrk2f/f (n=17) mice analyzed in 30-min intervals over a 2-h period. *p<0.05, comparison between genotypes by two-way RM ANOVA. (b) Cocaine-induced psychomotor activation of Grk2f/f (n=8–16) and D2RcreGrk2f/f (n=8–15) mice expressed as a percentage of activation (or fold increase) relative to vehicle-treated animals of the same genotype. **p<0.01 and ***p<0.001, comparison of the drug effect between genotypes by unpaired Student’s t-test. (c) Cocaine sensitization in D2RcreGrk2f/f and Grk2f/f mice (n=7–8 mice for each group). Locomotor activity in cocaine-treated mice was analyzed on days 1 and 7. *p<0.05, **p<0.01, ***p<0.001, comparison of the drug effect between genotypes by two-way RM ANOVA. (d) Cumulative distance traveled by cocaine-treated mice in the first 30 min postinjection of cocaine. *p<0.05 and **p<0.01, comparison of the drug effect between genotypes on days 1 and 7 by unpaired Student’s t-test. (e) Apomorphine-induced vertical activity in control and D2RcreGrk2f/f mice (n=10–12 mice for each group). *p<0.05, comparison of the drug effect between genotypes by unpaired Student’s t-test.

PowerPoint slide

Full size image

To determine whether the rewarding properties of cocaine were altered by the selective deletion of GRK2, we used a conditioned place preference paradigm. Independent cohorts of D1RcreGrk2f/f and D2RcreGrk2f/f mice were conditioned to two different doses of cocaine (10 and 20 mg/kg, i.p.), and their preference for the drug-paired chamber was analyzed. Robust conditioned responses were observed, and no significant differences were identified between genotypes, suggesting that GRK2 in D1R- and D2R-expressing neurons does not have a role in associative learning and cocaine reward (Supplementary Figure S3A and B).

The phenotypes observed in D1RcreGrk2f/f and D2RcreGrk2f/f mice are not due to the BAC transgenes, because both D1Rcre and D2Rcre hemizygous mice showed no spontaneous locomotor phenotype (Supplementary Figures S1D, 1F S2D, and F) and exhibited normal locomotor activity responses to cocaine (Supplementary Figures S1E, S1F S2E, and S2F). These results argue against any contribution of BAC transgene expression or integration site to DA-related phenotypes in these animals and are consistent with other published observations using these BAC transgenic animals (Bateup et al, 2010; Guzman et al, 2011).

Enhanced DARPP-32 Signaling in the Striatum of Cocaine-Treated D1RcreGrk2f/f Mice

To determine whether the hypersensitivity phenotype of D1RcreGrk2f/f mice was accompanied by alterations in striatal signaling, we next analyzed phospho-DARPP-32 levels because (1) cocaine has been shown to increase the phosphorylation of DARPP-32 at threonine 34 (T34) selectively in striatonigral neurons (Bateup et al, 2008) and (2) because DARPP-32 regulates basal and cocaine-induced locomotion (Bateup et al, 2010). Here we observed that acute cocaine administration promoted a robust increase in striatal pT34 DARPP-32 levels in both genotypes (interaction, F(1,17)=34.47, p<0.0001); however, the effect was significantly enhanced in D1RcreGrk2f/f mice relative to control animals (210±11% vs 149±17% of control, respectively; interaction, F(1,17)=5.06, p=0.03; Figure 3a). In contrast, a similar magnitude response to cocaine was observed in D2RcreGrk2f/f mice (Figure 3b). Evaluation of basal pT34 DARPP-32 levels in drug-naive D1RcreGrk2f/f and D2RcreGrk2f/f mice revealed no significant differences relative to matched controls (data not shown). These results suggest that GRK2 deletion in D1R-expressing neurons selectively enhances striatal D1R signaling.

Figure 3
figure 3

Cocaine-induced DARPP-32 phosphorylation is enhanced in the striatum of D1RcreGrk2f/f mice. (a and b) Mice were injected with either vehicle or cocaine (20 mg/kg, i.p.) and killed 15 min postinjection. pT34 DARPP-32 and total DARPP-32 levels were measured in extracts prepared from the dorsal striatum. Representative western blots (upper panels; V=vehicle and C=cocaine) and the corresponding densitometric analyses are shown. n=4–8 mice for each treatment group. pT34 DARPP-32 levels were normalized to the levels of DARPP-32 in the respective striatal extracts. ***p<0.001, vehicle vs cocaine treatment within genotypes and #p<0.05 for comparison between genotypes of the drug effect by two-way ANOVA.

PowerPoint slide

Full size image

To determine whether D2R-mediated Akt and GSK3 signaling in striatopallidal neurons was disrupted by selective GRK2 deletion, we analyzed the levels of phosphorylated threonine 308 (T308) Akt and serine 9 (S9) GSK3 in the striatum of both conditional KO lines. This analysis revealed no significant differences in the basal levels of striatal pT308 Akt and pS9 GSK3 in D1RcreGrk2f/f and D2RcreGrk2f/f mice relative to matched control animals (Supplementary Figure S4A–D). Furthermore, we observed no differences between genotypes in striatal Akt and GSK3 signaling following acute cocaine exposure (data not shown). These results suggest that GRK2 deficiency in D1R- or D2R-expressing neurons does not alter βarrestin-2-dependent D2R signaling in striatopallidal neurons, and appear to be consistent with the observation that GRK2 phosphorylation is not required for the association of D2R with βarrestin-2 (Namkung et al, 2009a).

D1RcreGrk2f/f Mice Exhibit Enhanced DA Release, Uptake, and Sensitivity to Cocaine

To investigate whether GRK2 deficiency in D1R- or D2R-expressing neurons caused perturbations to the DA system, we measured evoked DA release and uptake using FSCV in dorsal striatal slices prepared from the conditional KO mice. We observed that, in slices from drug-naive D1RcreGrk2f/f mice, electrically evoked DA release (either to single or multiple stimuli) was robustly increased (∼2-fold at each frequency) relative to that measured in slices from control animals (genotype, F(1,30)=24.08, p=0.0027; Figure 4a and b). Additionally, as shown in Figure 4c, we found that the maximal rate of DA uptake (Vmax) was dramatically increased in slices from D1RcreGrk2f/f mice (p=0.038). This robust increase in Vmax levels was similar to what we frequently observe in cocaine-treated wild-type animals (data not shown) and may be due to an increase in cell-surface DAT levels as has been previously described for cocaine (Brown et al, 2001; Daws et al, 2002).

Figure 4
figure 4

Enhanced baseline and cocaine-induced DA release in D1RcreGrk2f/f mice. (a) Representative DA traces (no drug applied) from electrical stimulations that mimic tonic (one pulse) and phasic (five pulses at 25 Hz) firing of DA neurons in slices prepared from drug-naive Grk2f/f (black) and D1RcreGrk2f/f mice (red). (b) Frequency–DA response curves for slices from control and D1RcreGrk2f/f mice (n=3–5 mice for each group). *p<0.05, **p<0.01, and ***p<0.001, Grk2f/f vs D1RcreGrk2f/f by two-way RM ANOVA. (c) Maximal rates of DAT uptake (Vmax) in slices prepared from drug-naive Grk2f/f (n=11) and D1RcreGrk2f/f (n=8) mice. *p<0.05, Grk2f/f vs D1RcreGrk2f/f by unpaired Student’s t-test. (d) Cocaine response (expressed as percentage of predrug baseline) demonstrates the relationship between bath-applied cocaine (10 μM) and electrically stimulated DA release. Slices were prepared from drug-naive Grk2f/f (n=5) and D1RcreGrk2f/f (n=4) mice. *p<0.05, Grk2f/f vs D1RcreGrk2f/f by unpaired Student’s t-test. (e) Frequency–response curves (expressed as percentage of one-pulse stimulation) before (solid lines) and after (dotted lines) bath application of sulpiride (10 μM). n=4–6 mice for each group.

PowerPoint slide

Full size image

To determine whether D1RcreGrk2f/f mice were also hypersensitive to cocaine at the neurochemical level, we measured evoked DA release in the presence of cocaine (10 μM) in slices from both genotypes. This analysis revealed a significant increase in evoked DA levels in striatal slices prepared from D1RcreGrk2f/f mice (p=0.03; Figure 4d). Interestingly, we did not observe a difference in the rate of DA uptake inhibition (τ) between genotypes (data not shown), indicating that GRK2 deficiency in D1R-expressing neurons does not alter the effectiveness of cocaine at blocking DA reuptake.

As D2 autoreceptors present on DA neurons have been shown to tightly regulate striatal DA release (Bello et al, 2011), we next investigated whether the neurochemical phenotype observed in D1RcreGrk2f/f mice was due to altered D2 autoreceptor activity. Evoked DA release in striatal slices was measured in the presence of the D2R antagonist sulpiride (10 μM) and expressed as a percentage of the baseline value (ie, release at one pulse; Figure 4e). Here we observed that sulpiride did not affect DA release in sections prepared from either genotype. To further evaluate D2 autoreceptor function in this mouse line, we measured evoked DA release in the presence of quinpirole, a direct D2R agonist. This analysis revealed no significant differences in quinpirole-induced inhibition of DA release between genotypes, regardless of treatment history (Supplementary Figure S5). Collectively, these results strongly suggest that GRK2 deficiency in D1R-expressing neurons does not alter D2 autoreceptor function.

Reduced DA Release and Increased D2 Autoreceptor Activity in D2RcreGrk2f/f Mice

Conversely, measurement of DA release evoked by a single or multiple stimuli in slices prepared from drug-naive D2RcreGrk2f/f mice revealed a robust decrease in baseline DA release relative to that measured in control slices (genotype, F(1,35)=39.35, p=0.0004; Figure 5a and b). The depression of baseline DA levels was not due to alterations in the rate of DA uptake, because measured Vmax levels were similar in slices between D2RcreGrk2f/f and control animals (Figure 5c). To determine whether altered D2 autoreceptor receptor activity in D2RcreGrk2f/f mice may underlie the hypodopaminergia phenotype, we bath applied sulpiride to slices prepared from drug-naive mice of both genotypes following electrical stimulation. As expected, stimulated DA release in D2RcreGrk2f/f slices was reduced relative to controls (solid lines, black vs green; interaction, F(4,40)=4.95, p=0.0025; genotype, F(1,40)=22.08, p=0.0008; Figure 5d). But interestingly, while sulpiride treatment had no effect in control slices, D2 autoreceptor antagonism relieved the blunted DA release observed in KO slices (dashed vs solid green lines; interaction, F(4,32)=2.74, p=0.04; genotype, F(1,32)=28.98, p=0.0007; Figure 5d). These findings suggest that under basal conditions, GRK2 regulates D2 autoreceptor function in this cell population, and the absence of this kinase leads directly to increased receptor activity and dampened evoked DA release.

Figure 5
figure 5

Depressed baseline DA release and increased D2 autoreceptor activity in D2RcreGrk2f/f mice. (a) Representative DA traces (no drug applied) from electrical stimulations that mimic tonic and phasic firing of DA neurons in slices prepared from drug-naive Grk2f/f (black) and D2RcreGrk2f/f mice (green). (b) Frequency–DA response curves for control and conditional KO mice (n=4–5 mice for each group). *p<0.05, **p<0.01, and ***p<0.001, Grk2f/f vs D2RcreGrk2f/f mice by two-way RM ANOVA. (c) Maximal rates of DAT uptake in drug-naive Grk2f/f (n=6) and D2RcreGrk2f/f (n=10) mice. (d) Frequency–response curves before (solid lines) and after (dotted lines) bath application of sulpiride (10 μM). n=4–7 mice for each group. **p<0.01 and ***p<0.001, vehicle vs sulpiride-treated D2RcreGrk2f/f mice by two-way RM ANOVA.

PowerPoint slide

Full size image

D2RcreGrk2f/f Mice Exhibit Reduced Neurochemical Sensitivity to Cocaine

We next examined whether the deficiency of GRK2 in D2R-expressing neurons altered sensitivity to cocaine at the neurochemical level. For these studies, mice were injected with either vehicle or cocaine (20 mg/kg, i.p.) once daily for 5 consecutive days, and on day 7 (approximately 40 h after the last injection) striatal slices were prepared from cocaine-sensitized mice (green and blue groups; Figure 6) and vehicle-treated control animals (black and red groups; Figure 6). Cocaine was then bath applied to slices from both treatment groups, and evoked DA release was measured. As expected, repeated drug administration to control mice resulted in a robust dose-dependent enhancement in cocaine-induced DA release and overflow (black vs green lines; genotype, F(1,16)=15.42, p=0.017; Figure 6a and b). In contrast, cocaine at all concentrations tested did not augment evoked DA release in sections from sensitized-D2RcreGrk2f/f mice (red vs blue lines; Figure 6a). Additionally, we analyzed the relationship between increasing concentrations of bath-applied cocaine and rates of DA uptake. These measurements revealed that cocaine was less effective at inhibiting DA uptake in D2RcreGrk2f/f mice regardless of vehicle or cocaine treatment (vehicle groups: interaction, F(4,28)=6.01, p=0.0013; genotype, F(1,28)=8.31, p=0.02 and cocaine groups: interaction, F(4,20)=5.28, p=0.0046; Figure 6b). These findings strongly suggest that GRK2 in D2R-expressing neurons is required for cocaine-induced sensitization of DA release and cocaine-mediated inhibition of DA uptake in the striatum.

Figure 6
figure 6

Reduced cocaine-induced sensitization of DA release and inhibition of DA uptake in D2RcreGrk2f/f mice. (a and b) Cocaine concentration–response curves in slices from Grk2f/f (n=3) and D2RcreGrk2f/f mice treated acutely (vehicle group) or exposed repeatedly to cocaine (cocaine group) demonstrate the relationship between bath-applied cocaine (10 μM) and electrically stimulated DA release (a) or DA uptake inhibition (b). **p<0.01, vehicle vs cocaine-treated Grk2f/f mice in panel (a). ***p<0.001, comparison of acute cocaine treatment groups (black vs red in panel (b) by two-way RM ANOVA. ##p<0.01, cocaine-sensitized Grk2f/f mice vs cocaine-sensitized D2RcreGrk2f/f mice (green vs blue in panel (b) by two-way RM ANOVA.

PowerPoint slide

Full size image

DATcreGrk2f/f Mice are Hyperactive and Less Sensitive to Acute Cocaine Treatment

Because D2RcreGrk2f/f mice have GRK2 deleted in both postsynaptic DA-receptive neurons (ie, striatopallidal MSNs) and presynaptic DA neurons, we next sought to discern the role of GRK2 in these two distinct cell types using a genetic approach. We generated two independent lines of conditional KO mice in which GRK2 was selectively deleted from either presynaptic DA neurons or postsynaptic A2AR-expressing striatopallidal MSNs. This was accomplished by crossing Grk2f/f mice to either DATcre or A2ARcre mouse lines to yield conditional KO mice (DATcreGrk2f/f and A2ARcreGrk2f/f) and control littermates. Immunostaining with cell-type-specific markers (DAT and DARPP-32) confirmed Cre-mediated recombination in the intended cell populations (Supplementary Figures S6A and S7). Interestingly, we did not observe GRK2 expression in A2AR-expressing MSNs within the dorsal striatum or a notable loss of GRK2 staining in the striatum of A2ARcreGrk2f/f KO mice (Supplementary Figure S7A and data not shown). These results strongly suggest that GRK2 is not expressed at measurable levels in MSNs and are consistent with the recent observation that GRK2 is most highly expressed in cholinergic interneurons within the striatum (Bychkov et al, 2012; Daigle and Caron, 2012).

Similar to the D2RcreGrk2f/f line, DATcreGrk2f/f mice were hyperactive over a 2-h period relative to control animals (genotype, F(1,114)=9.20, p=0.0043; Figure 7a). In addition to the marked hyperactivity, DATcreGrk2f/f mice also exhibited reduced sensitivity to acute cocaine administration. Analysis of the fold activation induced by cocaine over a 90-min period revealed a striking reduction in responsiveness of DATcreGrk2f/f mice to two doses of the stimulant relative to control animals (p=0.0045 and p=0.0097, 20 and 30 mg/kg doses, respectively; Figure 7b). In contrast, analysis of locomotor activity following repeated cocaine injections revealed no significant differences between genotypes on days 1 or 7 (Figure 7c and d). Additionally, no difference in cocaine reward was observed in the conditioned place preference paradigm (Supplementary Figure S3C). To confirm that mice heterozygous for the DATcre allele alone do not have altered locomotor behavior, we performed open field studies on DATcre heterozygous animals and wild-type littermates (Supplementary Figure S6). We observed no differences in the total distance travelled by drug-naive or cocaine-treated animals, indicating that the insertion of the IRES-Cre at the endogenous DAT locus does not significantly alter locomotor behavior (Supplementary Figure S6B–D). Taken together, these results indicate that GRK2 in DA neurons has a critical role in the regulation of spontaneous locomotion and acute cocaine-induced psychomotor activation.

Figure 7
figure 7

DATcreGrk2f/f mice exhibit enhanced spontaneous locomotion and reduced sensitivity to cocaine. (a) Spontaneous locomotion of Grk2f/f and DATcreGrk2f/f (n=20 mice per group) mice was analyzed in 30-min intervals over a 2-h period. *p<0.05, comparison between genotypes by two-way RM ANOVA. (b) The change in locomotor activity (expressed as a percentage) of Grk2f/f (n=6–20) and DATcreGrk2f/f (n=8–20) mice following acute cocaine treatment. *p<0.05 and **p<0.01, comparison of the drug effect between genotypes by unpaired Student’s t-test. (c) Cocaine sensitization in DATcreGrk2f/f and Grk2f/f mice (n=19–21 mice for each group). Locomotor activity in cocaine-treated mice on days 1 and 7 are shown. (d) Cumulative distance traveled by cocaine-treated mice in the first 10 min postinjection of cocaine. *p<0.05 and ##p<0.01, comparison of the drug effect between days 1 and 7 in each respective genotype by unpaired Student’s t-test.

PowerPoint slide

Full size image

A2ARcreGrk2f/f Mice Exhibit Normal Locomotor Behavior and Sensitivity to Cocaine

Although we were unable to detect appreciable GRK2 expression in A2AR-expressing striatal MSNs (Supplementary Figures S7 and S8), we could not exclude the possibility that an extremely low, yet functionally relevant level of GRK2 was present in this cell population. We therefore performed the same battery of behavioral experiments in A2ARcreGrk2f/f mice as with the other conditional KO lines (Figure 8). However, analysis of spontaneous locomotor activity revealed no significant differences between genotypes (Figure 8a and c). Further studies with cocaine also revealed no significant differences between genotypes in sensitivity to either an acute cocaine exposure or to a sensitizing regimen of cocaine (Figure 8b–d). Additionally, cocaine-induced conditioned place preference was unaltered in A2ARcreGrk2f/f mice (Supplementary Figure S3D). These results strongly suggest that the hyperactivity and hyposensitivity to cocaine phenotypes observed in D2RcreGrk2f/f mice are not due to a deficiency of GRK2 within postsynaptic striatopallidal MSNs.

Figure 8
figure 8

Deletion of GRK2 from A2AR-expressing neurons does not alter basal locomotor activity or sensitivity to cocaine. (a and b) Spontaneous activity (analyzed in 30-min intervals over a 2-h period) or cocaine-induced (20 mg/kg, i.p.) locomotion of Grk2f/f and A2ARcreGrk2f/f mice (vehicle, n=13–16 mice per group; cocaine, n=15 mice per group). (c) The sum of distance travelled by Grk2f/f and A2ARcreGrk2f/f mice postinjection of vehicle or cocaine (n=8–16 mice per group). (d) Cocaine sensitization in Grk2f/f and A2ARcreGrk2f/f mice (n=13–16 mice for each group). Locomotor activity in cocaine-treated (20 mg/kg, i.p.) mice on days 1 and 7 are shown.

PowerPoint slide

Full size image

DISCUSSION

Although previous studies have implicated GRK2 as an essential mediator of DA receptors in vitro, no investigations to date have examined whether the deficiency of this kinase in the mature brain alters DA receptor function or animal behaviors classically associated with the actions of DA. Here we used a genetic strategy to dissect GRK2 function within the mammalian brain and present evidence in support of a cell-type-specific role for GRK2 in the control of spontaneous locomotion, DA neurotransmission, D1R signaling, D2 autoreceptor function, and the behavioral response to psychostimulants. Collectively, our behavioral, biochemical, and electrophysiological results (largely summarized in Supplementary Table S2) provide novel mechanistic insight into DA receptor function in vivo and the actions of psychostimulant drugs in the murine brain. Although our current efforts were arguably DA-centric, it is plausible that GRK2 deficiency alters the function of additional GPCRs within the studied neuronal populations. This possibility should be examined in greater detail in future studies.

Earlier work with GRK5 and GRK6 germline KO mice revealed a very modest phenotype in unchallenged animals, suggesting that these kinases have little impact on diverse physiological processes under basal conditions (Gainetdinov et al, 2003; Gainetdinov et al, 1999). Similar to these findings, we observed no overt behavioral phenotype in drug-naive mice lacking GRK2 in either cholinergic (Daigle and Caron, 2012) or D1R-expressing neurons (present study). In contrast, the selective deletion of GRK2 in D2R-expressing neurons resulted in increased spontaneous locomotor activity, decreased apomorphine-induced vertical activity, and impaired habituation to a novel environment (data not shown). Basal hyperactivity was also observed in mice lacking GRK2 in DA neurons, but interestingly, not in mice with GRK2 deficiency in postsynaptic A2AR-expressing striatopallidal MSNs. The similar impairment in baseline locomotion between conditional KO mice (ie, D2RcreGrk2f/f mice vs DATcreGrk2f/f mice) strongly suggests that the hyperactivity phenotype of D2RcreGrk2f/f mice is caused by the loss of GRK2 in DA neurons.

Psychostimulant drugs such as cocaine promote robust increases in extracellular DA levels, resulting in enhanced DA receptor signaling and elevated locomotor activity. Because D1 and D2 receptors have a vital role in mediating the locomotor-stimulating effects of cocaine (Chausmer and Katz, 2001; Xu et al, 1994), and because GRK2 can regulate these receptors in vitro (Ito et al, 1999; Iwata et al, 1999; Namkung et al, 2009a; 2009b; Sedaghat and Tiberi, 2011; Tiberi et al, 1996), we hypothesized that GRK2 may regulate behavioral sensitivity to cocaine. Here we report that the deletion of GRK2 in D1R- and D2R-expressing neurons causes opposite effects on acute cocaine-induced locomotion, resulting in hypersensitive or hyposensitive mice, respectively. Interestingly, while all cocaine-mediated effects remained intact following deletion of GRK2 from A2AR-expressing striatopallidal MSNs (present study) or cholinergic neurons (Daigle and Caron, 2012), mice with GRK2 deficiency in DA neurons were also markedly less sensitive to the acute effects of cocaine. Collectively, these results strongly suggest that the reduced sensitivity of D2RcreGrk2f/f mice to cocaine is primarily due to the loss of GRK2 function in DA neurons.

Sensitization to the locomotor-stimulating effects of cocaine was also dramatically reduced by the deficiency of GRK2 in D2R-expressing neurons, but unlike for spontaneous and acute cocaine-induced locomotion, this behavioral impairment was not recapitulated in DATcreGrk2f/f or in A2ARcreGrk2f/f mice. These results suggest that GRK2 function in D2R-expressing neurons outside of the striatum may be required for the full expression of behavioral sensitization to cocaine. Although further studies are still needed to pinpoint a contributing neuronal population, one promising candidate brain region is the medial prefrontal cortex (mPFC). Beyer and Steketee (2000, 2002) reported that quinpirole infusion into the mPFC blocks acute cocaine-induced locomotor activity and sensitization, suggesting that cortical D2R activation negatively regulates the actions of cocaine. Their behavioral results essentially phenocopy what we observe in D2RcreGrk2f/f mice, raising the possibility that increased activity of cortical D2Rs may also contribute to the reduction in the cocaine-induced effects in these conditional KO animals.

Deletion of GRK2 from D1R-expressing neurons caused a robust increase in striatal DA release and uptake in drug-naive mice. Particularly intriguing was the dramatic increase in Vmax levels in these mice, because the rate of uptake here was similar to that observed in control animals exposed to cocaine (data not shown). One explanation for these effects is that the loss of GRK2 in this neuronal population causes an increase in the membrane-associated levels of DAT in the striatum. This perturbation could presumably also enhance DA release, because DA recycling into synaptic vesicles should also be augmented in response to a protracted increase in surface DAT levels. The initial increase in cell-surface DAT levels may be caused by elevated glutamate levels in the VTA that result from increased activity of D1Rs located on afferent glutamatergic terminals originating from forebrain regions (Kalivas and Duffy, 1995; Lu et al, 1997; Sesack and Pickel, 1992). Enhanced glutamate release directly onto DA neurons may increase neuronal firing frequency, which would thereby result in enhanced striatal DA release. If this were the case, cell-surface DAT levels may increase to compensate for the chronic elevation of extracellular DA levels. We also cannot exclude the possibility that GRK2 deletion altered the activity of additional GPCRs in this neuronal population, and these potential perturbations contributed to the profound neurochemical phenotype. In addition to these baseline alterations, enhanced DA release in response to acute cocaine exposure was also observed in striatal slices from D1RcreGrk2f/f mice. This observation strongly suggests that locomotor supersensitivity to cocaine may be due to elevated drug-induced DA release in the striatum of D1RcreGrk2f/f mice.

In contrast, the elimination of GRK2 from D2R-expressing neurons caused a significant reduction in striatal DA release under basal conditions. Remarkably, we found that this depression could be completely reversed by application of sulpiride, a classical D2R antagonist. These results demonstrate that GRK2 deficiency in D2R-expressing neurons leads to overactivity of presynaptic D2 autoreceptors, which thereby results in the persistent suppression of DA release from striatal DA terminals. Enhanced D2 autoreceptor activity may be due to impaired GRK2-mediated desensitization of these presynaptic receptors. Importantly, the absence of GRK6 has also been reported to enhance striatal D2R activity (Gainetdinov et al, 2003), but in contrast to the present findings, this work implicated postsynaptic D2Rs as the principal physiological target of GRK6. Taken together, these data provide the first evidence that the regulation of presynaptic autoreceptors by GRKs occurs in a similar manner to the regulation of postsynaptic receptors, albeit by different GRKs. Furthermore, while it is now evident that multiple GRKs can regulate DA receptors in vivo, GRK2- and GRK6-mediated regulation exhibit neuronal specificity and regulate distinct populations of striatal D2Rs (presynaptic vs postsynaptic).

The reduced behavioral sensitivity of D2RcreGrk2f/f mice to cocaine was paralleled by changes at the neurochemical level. Here we observed that repeated cocaine treatment of D2RcreGrk2f/f mice was less effective at promoting the expected sensitized increase in striatal DA release. Additionally, we uncovered a dramatic impairment in cocaine-mediated uptake inhibition in the striatum of both drug-naive and drug-sensitized animals. Therefore, in addition to the regulation of D2 autoreceptors, GRK2 deficiency may also impact the sensitivity of DAT for cocaine and/or the availability of releasable DA. Detailed investigation will be required in future work to establish the entire collection of presynaptic targets affected by GRK2 deletion.

Mice with the genetic deletion of D2 autoreceptors exhibit hyperactivity when placed in a novel environment and are hypersensitive to the acute locomotor-stimulating effects of cocaine (Anzalone et al, 2012; Bello et al, 2011). Our results are partially consistent with these earlier findings in that the reverse perturbation, increased D2 autoreceptor activity, causes reduced locomotor sensitivity to cocaine. However, both D2RcreGrk2f/f mice and D2 autoreceptor-deficient mice exhibited hyperactivity when exposed to a novel environment. This seemingly paradoxical phenotype in baseline locomotor activity between D2RcreGrk2f/f mice and D2 autoreceptor-deficient mice is likely due to the constitutive deletion of GRK2 from D2R-expressing cells, which could lead to multiple aberrant physiological processes in diverse brain regions (eg, potential cortical D2R hyperactivity in the mPFC). It is also important to note that both chronic hyperdopaminergia and hypodopaminergia cause a remarkable array of adaptive changes at the molecular level, including alterations in DA synthesis, storage, and DA receptor levels and functioning, as well as distinct behavioral impairments (Giros et al, 1996; Jones et al, 1999; Salahpour et al, 2008; Sotnikova et al, 2005). Therefore, the lack of an opposite behavioral outcome in this context (ie, spontaneous locomotion) may also be due to divergent compensatory mechanisms that exist between genetic models and result from alterations in extracellular DA levels.

The mystery of GRK2 function in the adult brain is beginning to become unraveled. This pursuit has been greatly aided by genetic methods that improve on the traditional gene KO approach and allow for the cell-type-specific deletion of proteins in the mature brain. Here we provide evidence that GRK2 deficiency in select neuronal populations in adult animals profoundly alters cocaine-dependent behaviors and striatal DA neurotransmission. Intriguingly, we found that the deletion of GRK2 in D1R- and D2R-expressing neurons differentially affects locomotor sensitivity to cocaine and striatal DA dynamics. D2 autoreceptor function was also significantly impaired in the absence of GRK2, suggesting that this receptor is robustly regulated by GRK2 under physiological conditions. Taken together, it appears that GRK2 has a vital role in the central DA system and thus may represent a novel therapeutic target for DA-related brain disorders.

FUNDING AND DISCLOSURE

This work was supported in part by NIH grants DA030026 (to TLD), MH073853, U19-MH082441, and P30-DA029925 (to MGC), DA031791 (to MJF), DA024095 (to SRJ), and DA030161 (to SRJ). MGC has received funds for sponsored research agreements unrelated to this work from Forest Laboratories and F. Hoffmann La Roche as well as consulting fees from Omeros Corporation. MGC has also received compensation in the form of honoraria for lecturing at various scientific meetings and academic institutions. RRG is supported in part by research grants from F. Hoffmann La-Roche (Basel, Switzerland) and St Petersburg State University (St Petersburg, Russia) grant no. 1.38.201.2014. None of the above presents any conflicts of interest with the results described in the present paper. The other authors declare no conflict of interest.