Transient and rapid activation of Akt/GSK-3β and mTORC1 signaling by D3 dopamine receptor stimulation in dorsal striatum and nucleus accumbens

Authors

  • Marie-Josèphe Salles,

    1. Institut National de la Santé et de la Recherche Médicale (Inserm), Paris, France
    2. Université Pierre et Marie Curie-Paris 6, Paris, France
    3. Institut du Fer à Moulin, Paris, France
    4. Psychopharmacology Department, Institut de Recherches Servier, Croissy sur Seine, France
    Search for more papers by this author
  • Denis Hervé,

    Corresponding author
    1. Université Pierre et Marie Curie-Paris 6, Paris, France
    2. Institut du Fer à Moulin, Paris, France
    • Institut National de la Santé et de la Recherche Médicale (Inserm), Paris, France
    Search for more papers by this author
  • Jean-Michel Rivet,

    1. Psychopharmacology Department, Institut de Recherches Servier, Croissy sur Seine, France
    Search for more papers by this author
  • Sophie Longueville,

    1. Institut National de la Santé et de la Recherche Médicale (Inserm), Paris, France
    2. Université Pierre et Marie Curie-Paris 6, Paris, France
    3. Institut du Fer à Moulin, Paris, France
    Search for more papers by this author
  • Mark J. Millan,

    1. Psychopharmacology Department, Institut de Recherches Servier, Croissy sur Seine, France
    Search for more papers by this author
  • Jean–Antoine Girault,

    1. Institut National de la Santé et de la Recherche Médicale (Inserm), Paris, France
    2. Université Pierre et Marie Curie-Paris 6, Paris, France
    3. Institut du Fer à Moulin, Paris, France
    Search for more papers by this author
  • Clotilde Mannoury la Cour

    1. Psychopharmacology Department, Institut de Recherches Servier, Croissy sur Seine, France
    Search for more papers by this author

Address correspondence and reprint requests to Denis Hervé, Institut du Fer à Moulin Inserm UMR-S 839, 17 rue du Fer a Moulin, 75005 Paris, France. E-mail: denis.herve@inserm.fr

Abstract

D2/D3 dopamine receptors (D2R/D3R) agonists regulate Akt, but their effects display a complex time-course. In addition, the respective roles of D2R and D3R are not defined and downstream targets remain poorly characterized, especially in vivo. These issues were addressed here for D3R. Systemic administration of quinelorane, a D2R/D3R agonist, transiently increased phosphorylation of Akt and GSK-3β in rat nucleus accumbens and dorsal striatum with maximal effects 10 min after injection. Akt activation was associated with phosphorylation of several effectors of the mammalian target of rapamycin complex 1 (mTORC1): p70S6 kinase, ribosomal protein-S6 (Ser240/244), and eukaryotic initiation factor-4E binding protein-1. The action of quinelorane was antagonized by a D2/D3R antagonist, raclopride, and the selective D3R antagonist S33084, inactive by themselves. Furthermore, no effect of quinerolane was seen in knock-out mice lacking D3R. In drd1a-EGFP transgenic mice, quinelorane activated Akt/GSK-3β in both neurons expressing and lacking D1 receptor. Thus, the stimulation of D3R transiently activates the Akt/GSK-3β pathway in the two populations of medium-size spiny neurons of the nucleus accumbens and dorsal striatum. This effect may contribute to the influence of D3R ligands on reward, cognition, and processes disrupted in schizophrenia, drug abuse, and Parkinson's disease.

Abbreviations used
4E-BP1

eukaryotic initiation factor 4E binding protein 1

DA

dopamine

DARPP-32

dopamine- and cAMP-regulated phosphoprotein of Mr 32 kDa

EGFP

enhanced green fluorescent protein

eIF4E

eukaryotic initiation factor 4E

ERK1/2

extracellular signal-regulated kinase 1 and 2

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

GSK-3β

glycogen synthase kinase-3β

mTORC1

raptor-mammalian target of rapamycin complex 1

NAcc

nucleus accumbens

NS

not significant

p70S6K

p70 S6 kinase

p90RSK

p90 ribosomal S6 kinase

PKA

protein kinase A

rpS6

ribosomal protein S6

R

receptor

S33084

(3aR,9bS)-N[4-(8-cyano-1,3a,4,9b-tetrahydro-3H-benzopyrano[3,4-c]pyrrole-2-yl)-butyl] (4-phenyl)benzamide)

WT

wild type

All dopamine (DA) receptors identified in mammals are G protein-coupled receptors that respond to DA by activating complex intracellular signaling pathways, the dissection of which is crucial to understand how DA regulates brain functions (Herve & Girault 2005). DA receptors (R) are divided into two categories: the D1 class (D1R and D5R) activates the intracellular production of cAMP through its coupling to Gs/olf G protein, while the D2 class (D2R, D3R, and D4R) inhibits cAMP production and regulates ion channels through Gi/o (Beaulieu et al. 2011). D1R and D2R are expressed abundantly in the dorsal striatum and nucleus accumbens (NAcc, or ventral striatum), the main targets of the mesencephalic DA neurons. D3R expression is high in the NAcc shell, intermediate in the NAcc core and low, but detectable, in the dorsal striatum (Bouthenet et al. 1991; Le Moine and Bloch 1996; Surmeier et al. 1996; Sokoloff et al. 2006). D1R and D2R are largely segregated in two populations of the principal striatal neurons, the GABAergic medium-size spiny neurons (MSNs), which display different projection areas (Gerfen et al. 1990; Le Moine and Bloch 1996; Bertran-Gonzalez et al. 2010). In the NAcc, D3R is present in the two populations but with a preference for the D1R-expressing MSNs (Le Moine and Bloch 1996; Surmeier et al. 1996; Schwartz et al. 1998).

DA receptors regulate the activity of Akt (also known as protein kinase B), a serine/threonine kinase involved in various cellular functions, including cell growth and proliferation (Manning and Cantley 2007; Beaulieu et al. 2011). Akt activation leads to that of the mammalian target of rapamycin complex 1 (mTORC1) via tuberous sclerosis complex 2 (TSC2) and Rheb (Bhaskar and Hay 2007). Activated mTORC1 phosphorylates two substrates, p70S6 kinase (p70S6K), leading in turn to phosphorylation of the ribosomal protein S6 (rpS6) on Ser235/236 and Ser240/244, and eIF4E binding protein 1 (4E–BP1), an inhibitor of the initiation factor eIF4E (Mamane et al. 2006; Foster and Fingar 2010). This signaling cascade controls the translation of terminal oligopyrimidine (TOP) mRNAs (Thoreen et al. 2012). Recent studies support a role for mTORC1, p70S6K, and 4E-BP1 in synaptic plasticity and cognitive processing, for example by contributing to long-term potentiation (LTP) and long-term depression (LTD) (Hoeffer and Klann 2010). However, mTORC1 has an inverse-U dose–response relationship with over-stimulation reducing cognitive function (Tang et al. 2002; Myskiw et al. 2008; Hoeffer and Klann 2010; Stoica et al. 2011; Millan et al. 2012). Glycogen synthase kinase-3β (GSK-3β) is another effector of Akt, which is widely expressed in the brain (Leroy and Brion 1999) and implicated in processes underlying neuronal survival, development, and synaptic plasticity (Peineau et al. 2008).

Depending on its duration, the activation of D2-type DA receptors differentially regulates Akt-dependent pathways. In various models of cells in culture, stimulation of D2R and/or D3R rapidly activates Akt signaling (Brami-Cherrier et al. 2002; Nair and Sealfon 2003; Mannoury la Cour et al. 2011). Similarly, in vivo selective D2R/D3R agonist or psychostimulants, which elevate DA extracellular levels, induce the phosphorylation of Akt and both isoforms of GSK-3(α/β) in the NAcc or dorsal striatum shortly after their administration (Brami-Cherrier et al. 2002; Svenningsson et al. 2003; Mannoury la Cour et al. 2011). In contrast, prolonged stimulation of D2-type receptors (over 30–60 min) decreases Akt phosphorylation specifically at Thr308 residue in the dorsal striatum (Beaulieu et al. 2004, 2005). This reduction of Akt activity is essentially mediated by D2R, as shown by its absence in D2R-deficient mice (Beaulieu et al. 2007b), through the formation of a D2R/β-arrestin2/protein phosphatase 2A/Akt complex(Beaulieu et al. 2005, 2011).

Interestingly, as suggested by results obtained in D3R knockout mice (Beaulieu et al. 2007b), D3R could also participate in Akt signaling, an hypothesis reinforced by the quick increase in Akt phosphorylation after stimulation of D3R in various cell lines (Chen et al. 2009; Mannoury la Cour et al. 2011; Collo et al. 2012). However, very little is known concerning the influence of D3R in vivo on Akt-dependent pathways. This is an important issue since D3R in subcortical structures, like the NAcc, contributes to the control of reward, cognitive processing, and motor function (Sokoloff et al. 2006; Heidbreder and Newman 2010; Millan et al. 2010). The present study focused on the potential influence of short-term stimulation of D3R on Akt, GSK-3β, and mTORC1-dependent signaling in the NAcc and dorsal striatum of rats and mice. We employed as an experimental tool the highly selective D2R/D3R agonist, quinerolane, which has been shown to activate Akt via D3R in cell lines (Mannoury la Cour et al. 2011). A specific role for D3Rs was evaluated using the selective D3R antagonist, S33084 (Millan et al. 2000a; Collo et al. 2012) and knock-out mice genetically deprived of D3Rs (Accili et al. 1996). Finally, we clarified the cellular localization of these effects employing drd1a-enhanced green fluorescent protein (EGFP) transgenic mice that permit the identification of neuronal populations expressing or not D1R in MSNs (Bertran-Gonzalez et al. 2008).

Materials and methods

Animals

We used male Wistar Han rats (Charles River Laboratories, L'Arbresle, France, 226–250 g), homozygous mutant mice lacking DA D3R generated on C57BL/6J background (The Jackson Laboratory, Bar Harbor, ME, USA, B6.129S4-Drd3Tm1Dac/J) (Accili et al. 1996), and C57BL/6J wild-type (WT) mice. drd1a-EGFP transgenic mice (mixed Swiss-Webster and C57BL/6N background) were generated by the GENSAT [Gene Expression Nervous System Atlas, (Gong et al. 2003)]. In these mice, a bacterial artificial chromosome (BAC) expressed EGFP under the control of the D1R promoter. Animals were housed in a 12 h light–dark cycle, in stable conditions of temperature (22°C), with free access to food and water. All the experiments were in accordance with the guidelines of the French Agriculture and Forestry Ministry for handling animals (decree 87–848).

Drugs and in vivo treatment

Quinelorane dichlorhydrate and raclopride tartrate were purchased from Sigma-Aldrich (Saint Quentin Fallavier, France). S 33084 [(3aR,9bS)-N[4-(8-cyano-1,3a,4,9b-tetrahydro-3H-benzopyrano[3,4-c]pyrrole-2-yl)-butyl] (4-phenyl)benzamide was synthetized by G. Lavielle (Servier, Paris, France). Groups of animals received a single injection of quinelorane (0.16 mg/kg for rats or 0.63 mg/kg for mice, s.c., dissolved in H2O) 5, 10, 20, or 90 min before being killed. For antagonist experiments, raclopride (0.16 mg/kg, s.c.), S33084 (0.63 mg/kg, s.c.), or vehicle (1 mL/kg, s.c.) were administered 30 min before quinelorane or vehicle.

Immunoblotting

After killing by decapitation, the animal brains were rapidly removed, frozen in liquid nitrogen and stored at −80°C until use. They were cut using a Leica refrigerated microtome until reaching the desired levels, and tissue samples were punched out from dorsal striatum and NAcc (both core and shell subdivisions) using cooled 1.8 mm and 0.8 mm diameter tubes for rats and mice, respectively. Frozen samples were sonicated in 95°C Laemmli buffer (~ 3 μg/μL of protein) and then incubated for 5 min at 95°C. Homogenates were loaded on 15-well 4–12% polyacrylamide gels (NuPAGE Novex Bis-Tris gels) (Invitrogen, Cergy-Pontoise, France) and transferred onto nitrocellulose membranes. To saturate non-specific sites, the membranes were incubated 1 h in blocking buffer (5% weight/volume [w/v] skimmed milk in Tris-buffered saline (TBS)-Tween). Rabbit polyclonal antibodies specific of the phosphorylated forms of Akt (pSer473 and pThr308 residues), GSK-3β (pSer9), extracellular signal-regulated kinase 1 and 2 (ERK1/2) (pThr202 and/or pTyr204 for ERK1, pThr185 and/or pTyr187 for ERK2), rpS6 (pSer240/244 and pSer235/236), and 4E-BP1 (pThr37/46) were used. The phosphorylation of p70S6K (pThr389) was detected with a phospho-specific mouse monoclonal antibody. All the antibodies were purchased from Cell Signaling Technology (Danvers, MA, USA) and used at 1/1000 dilution, except phospho-Akt (pSer473 and pThr308) antibodies that were used at the 1/600 and 1/500, respectively. Secondary antibodies were anti-rabbit or anti-mouse antibodies coupled to horseradish peroxidase conjugate for detection with ECL chemiluminescence (GE-Healthcare, Les Ulis, France), or IRDye 700DX-conjugated anti-rabbit and IRDye 800CW-conjugated anti-mouse IgGs (1/5000, Rockland Immunochemical, Gilbertsville, PA, USA) for detection with Odyssey-LI-COR infrared fluorescent detection system (LI-COR, Lincoln, NE, USA). After elution of phospho-specific antibodies using Re-Blott (1X) solution (Millipore, Saint Quentin en Yvelines, France), the membranes were re-probed for the various total proteins (phosphorylated and non-phosphorylated) and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) (monoclonal anti-GAPDH rabbit antibody, 1/10 000, Cell Signaling Technology). The signals (optical density or fluorescence intensity) for each phospho and total protein were normalized to the GAPDH signal in each sample. Phosphorylation levels of each protein were evaluated by the ratio between measures for phospho-specific and total proteins, and results are expressed in percentage of the mean of vehicle-treated rats or mice.

Immunohistofluorescence

Ten min after an acute treatment with quinelorane (0.63 mg/kg, s.c.), drd1a-EGFP mice were perfused transcardially with 4% (w/v) paraformaldehyde in 0.1 M sodium phosphate buffer, pH 7.5 under deep anesthesia (pentobarbital, 500 mg/kg, i.p.; Sanofi, Paris, France). Brains were post-fixed overnight in the same solution at 4°C. Thirty-micrometer thick sections were cut with a vibratome (Leica Microsystèmes SAS, Nanterre, France) and stored at −20°C in a solution containing 30% (v/v) ethylene glycol, 30% (v/v) glycerol and 0.1 M sodium phosphate buffer. Sections were processed as previously described (Bertran-Gonzalez et al. 2008). After permeabilization by a 20-min incubation in 0.2% (v/v) Triton X-100 in TBS-NaF (0.25 M Tris, 0.5 M NaCl, 50 mM NaF, pH 7.5), sections were blocked during 1 h in TBS-NaF with 3% (w/v) Bovine Serum Albumin (BSA). They were then incubated overnight at 4°C with the anti-pSer9-GSK-3β (1/200, Cell Signaling Technology). After three washes, sections were incubated for 45 min with goat Cy3-coupled (1 : 400; The Jackson Laboratory) secondary antibodies. Sections were rinsed twice in TBS and twice in Tris buffer (0.25 M Tris) before mounting in Vectashield with 4′-6-diamidino-2-phenylindole (DAPI) (Vector Laboratories, Burlingame, CA, USA).

Double-labeled images (212 × 212 μm, 4–10 per animal, 10 animals per group) were acquired in the center of dorsal striatum and the NAcc core just ventral to the anterior commissure using the EGFP and Cy3 fluorescence channels of sequential laser-scanning confocal microscope (60X objective, Fluoview FV10; Olympus, Institut du Fer à Moulin Imaging Facility, Tokyo, Japan) set at fixed parameters for each experiment. The neurons were considered as positive for pSer9-GSK-3β when their immunofluorescence was higher than threshold intensity. The same threshold was applied automatically to all the pictures by a home-written software using Image J software (NIH, Bethesda, MD, USA) and its level was fixed by an observer unaware of the treatment received by the mice. The number of phospho-positive neurons was counted in the EGFP-positive and negative populations in each image and the mean density (expressed as number of neurons per mm2) for each animal was used for statistical analysis.

Statistical analysis

Data from western blots were analyzed with one-way or two-way anova followed by Bonferroni's post hoc comparison test using GraphPad Prism 5 software (San Diego, CA, USA). Data from cell counting were analyzed using non-parametric Friedman's test on value ranks (Zar 1999) followed by post hoc analysis using corrected Mann–Whitney test on values (with α set at 0.0125 as significance threshold). Values in graphs were expressed as mean ± SEM.

Results

Quinelorane induces a transient increase of Akt and GSK-3β phosphorylation in the rat dorsal striatum and nucleus accumbens

To selectively stimulate D2R and D3R in vivo, rats were treated with 0.16 mg/kg of quinelorane. This dose was high enough to stimulate both receptor types in vivo (Gobert et al. 2004; Collins et al. 2007), although the agonist displays about a 100-fold higher affinity for D3R than D2R (Sokoloff et al. 1992). Quinelorane increased Akt phosphorylation levels at both Ser473 and Thr308 residues in the dorsal striatum and NAcc. The effect was maximal 10 min after the administration of quinelorane and of a similar amplitude in the dorsal striatum and NAcc when the results were expressed as percent of their respective control values (Fig. 1a and b). Akt phosphorylation returned to baseline 20 min after agonist injection and no significant effect was observed at 5 min (Fig. 1a and b).

Figure 1.

Time-course of quinelorane-induced protein phosphorylation in rat dorsal striatum and nucleus accumbens. Groups of rats were injected with vehicle or quinelorane (0.16 mg/kg, s.c.) and killed at the indicated time-points. Phosphorylation of Akt at Thr308 (p-T308-Akt, a) and Ser473 (p-S473-Akt, b), GSK-3β (p-S9-GSK-3β, c), p70S6K (p-T389-p70S6K, d), 4E-BP1 (p-T37/46-4E-BP1, e), rpS6 at Ser240/244 (p-S240/244-rpS6, f), ERK1/2 (p-ERK1/2, g), and rpS6 at Ser235/236 (p-S235/236-rpS6, h) was evaluated by immunoblotting and quantified. Phosphorylation levels are normalized to the total amount of the corresponding protein in the same sample and expressed as a percentage of the mean basal levels in control (vehicle-treated) rats (Cont). Data are the mean ± SEM for each time-point (n = 4 to 6 rats per group). One-way anova for the striatum: p-T308-Akt/Akt, F(3,16) = 4.8, < 0.05; p-S473-Akt/Akt, F(3,16) = 5.4, < 0.05; p-S9-GSK-3β/GSK-3β, F(3,16) = 3.7, p< 0.05, p-T389-p70S6K/p70S6K, F(3,18) = 10.6, < 0.05; p-T37/46-4E-BP1/4E-BP1, F(3,18) = 1.0, NS; p-S240/244-rpS6/rpS6, F(3,15) = 7.2, < 0.05; p-ERK1/2 / ERK1/2, F(3,22) = 0.8, NS; p-S235/236-rpS6/rpS6, F(3,16) = 0.4, NS. For the nucleus accumbens: p-T308-Akt/Akt, F(3,17) = 6.1, < 0.05; p-S473-Akt/Akt, F(3,16) = 17.7, < 0.05; p-S9-GSK-3β/GSK-3β, F(3,19) = 7.3, < 0.05; p-T389-p70S6K/p70S6K, F(3,17) = 9.7, < 0.05; p-T37/46-4E-BP1/4E-BP1, F(3,16) = 5.9, < 0.05; p-S240/244-rpS6/rpS6, F(3,17) = 3.4, < 0.05; p-ERK1/2 / ERK1/2, F(3,21) = 0.5, NS; p-S235/236-rpS6/rpS6, F(3,20) = 0.3, NS. Post hoc comparison (Bonferroni's test): *< 0.05, **< 0.01, quinelorane- versus vehicle-treated in the dorsal striatum; #< 0.05; ##< 0.01; ###< 0.001, quinelorane- versus vehicle-treated in the nucleus accumbens. Not significant, NS.

A similar time-course of phosphorylation was observed for GSK-3β in both structures, with a maximal increase of phospho-Ser9-GSK-3β 10 min after administration of the agonist (Fig. 1c). The total levels of Akt and GSK-3β were not modified by quinelorane treatment as compared to basal conditions in both the dorsal striatum and NAcc (data not shown).

Quinelorane induces a transient phosphorylation of mTORC1 substrates

Activation of Akt is able to stimulate mTORC1, leading to the phosphorylation of several substrates, including p70S6K (at Thr389) and 4E–BP1 (at Thr37/46), and indirectly that of rpS6 (at Ser240/244) (Dunlop et al. 2009). In consistency with the Akt activation, phosphorylation of p70S6K at Thr389, 4E-BP1 at Thr37/46, and rpS6 at Ser240/244 was increased in the dorsal striatum and NAcc 10 min after quinelorane administration (Fig. 1d, e, f). These effects were transient and phosphorylation returned to baseline after 20 min with a time-course very similar to that observed for Akt. Thus, the increased phosphorylation of Akt seen after quinelorane was parallel to those of several effectors of the mTORC1 pathway.

Interestingly, the mitogen-activated protein kinases ERK1/2 can stimulate mTORC1 through the activation of p90RSK and Raptor (Carriere et al. 2008). However, the increased phosphorylation of mTORC1 effectors in our experimental conditions did not result from the ERK1/2 activation since no significant change of ERK1/2 phosphorylation levels was detected in either dorsal striatum or NAcc after quinelorane treatment (Fig. 1g). In addition, phosphorylation of rpS6 at Ser235/236 (Fig. 1h), a site that can be phosphorylated by the ERK1/2-activated kinase p90RSK (Roux et al. 2007), was not altered by the administration of quinelorane. Thus, our data are compatible with an Akt-induced activation of mTORC1.

D2R and/or D3R mediate the quinelorane-induced activation of Akt in the rat dorsal striatum and nucleus accumbens

The involvement of D2R and/or D3R in the effects of quinelorane on Akt and GSK-3β phosphorylation at 10 min was investigated by pre-treating the rats with the selective D2R/D3R antagonist, raclopride, 30 min before the agonist administration. Pre-treatment with raclopride abolished the quinelorane-elicited increase in Akt phosphorylation at both Ser473 and Thr308 and GSK-3β at Ser9 in the dorsal striatum and NAcc (Fig. 2). No significant effect was observed when raclopride was administered alone. In addition, none of these treatments had a significant influence on the total levels of Akt and GSK-3β (data not shown).

Figure 2.

Inhibition by raclopride of quinelorane-induced increase in Akt and GSK-3β phosphorylation in the rat dorsal striatum and nucleus accumbens. Groups of rats received two injections, first with raclopride (0.16 mg/kg s.c.) or vehicle, and 30 min later with quinelorane (0.16 mg/kg s.c.) or vehicle. Rats were killed 10 min after the second injection and phosphorylation of Akt (at Thr308 and Ser473) and GSK-3β (at Ser9) was measured. Representative immunoblots of phospho- and total proteins in the dorsal striatum (a) and nucleus accumbens (c). Quantification of the phospho/total protein ratio in the dorsal striatum (b) and nucleus accumbens (d). Histograms correspond to the mean + SEM of data obtained in two independent experiments (= 5 to 11 rats per group). One-way anova for the striatum (c): p-T308-Akt/Akt, F(3,34) = 4.8, < 0.05; p-S473-Akt/Akt, F(3,26) = 5.1, < 0.05; p-S9-GSK-3β/GSK-3β, F(3,27) = 4.6, < 0.05. For the nucleus accumbens (d): p-T308-Akt/Akt F(3,35) = 5.9, < 0.05; p-S473-Akt/Akt F(3,35) = 4.8, < 0.05; p-S9-GSK-3β/GSK-3β F(3,29) = 3.9, < 0.05. Post hoc comparison (Bonferroni's test): *< 0.05, **< 0.01, quinelorane- versus vehicle-treated; #< 0.05, ##< 0.01, quinelorane- versus raclopride + quinelorane-treated.

The inhibitory action of raclopride was also investigated on the signaling pathways downstream from mTORC1. In both dorsal striatum and NAcc, no significant effect of quinelorane was observed on the phosphorylation levels of Thr389-p70S6K, Ser240/244-rpS6, and Thr37/46-4E–BP1 in raclopride-pre-treated rats (Fig. 3), whereas a marked increase of phosphorylation was seen in animals receiving quinelorane alone.

Figure 3.

Inhibition by raclopride of quinelorane-induced increase in mTORC1 signaling in the rat striatum and nucleus accumbens. Groups of rats were treated as described in the legend to Fig. 2. The phosphorylation of 4E-BP1 at Thr37/46, p70S6K at Thr389, and rpS6 at Ser240/244, three targets of mTORC1, was measured. Representative immunoblots of phospho- and total proteins in the dorsal striatum (a) and nucleus accumbens (c). Quantification of the phospho/total protein ratio in the dorsal striatum (b) and nucleus accumbens (d). Histograms correspond to the mean + SEM of data obtained in two independent experiments (= 7 to 11 rats per group). One-way anova for the dorsal striatum (b): p-T37/46-4E-BP1/4E-BP1, F(3,40) = 5.1, < 0.05; p-T389-p70S6K/p70S6K, F(3,32) = 5.7, < 0.05; p-S240/244-rpS6/rpS6, F(3,30) = 4.0, < 0.05. For the nucleus accumbens (d): p-T37/46-4E-BP1/4E-BP1. F(3,37) = 5.1, < 0.05, p-T389-p70S6K/p70S6K F(3,32) = 7.6, < 0.05; p-S240/244-rpS6/rpS6, F(3,34) = 3.7, < 0.05. Post hoc comparison (Bonferroni's test): *< 0.05, **< 0.01, quinelorane-treated versus vehicle-treated; #< 0.05, quinelorane- versus raclopride + quinelorane-treated.

The quinelorane-induced phosphorylation of GSK-3β occurs in both D1R- and D2R-expressing neurons

Since activation of signaling pathways in response to drugs often occurs in specific neuronal subpopulations in the striatum (Bertran-Gonzalez et al. 2008), we sought to determine in which types of neurons the Akt/GSK-3β pathway was activated following quinelorane administration using drd1a-EGFP mice in which EGFP is expressed under the control of the D1R promoter in the striatonigral MSNs that bear D1R. In contrast, the EGFP-negative MSNs correspond to the striatopallidal neurons that bear D2R, because very few neurons co-express both receptors (~ 5% in the dorsal striatum and NAcc core) (Bertran-Gonzalez et al. 2008; Valjent et al. 2009). The phosphorylation of GSK-3β at Ser9 was used to evaluate the activation of Akt/GSK-3β pathway, since we found that the corresponding phospho Ser9-GSK-3β-specific antibody gave rise to unambiguous immunofluorescence in tissue sections. In drd1a-EGFP mice, the injection of 0.63 mg/kg of quinelorane induced a significant increase in the number of positive neurons for phospho-Ser9-GSK-3β immunoreactivity in the mouse dorsal striatum and NAcc core, with a phospho-signal remaining exclusively in the cytoplasm (Fig. 4a). The number of phospho-Ser9-GSK-3β positive cells was similarly increased in the populations of neurons expressing or not D1R in the dorsal striatum and NAcc (Fig. 4a and b). We noticed that the density of phospho-positive neurons was higher in the NAcc than in the dorsal striatum.

Figure 4.

Quinelorane-induced increase in Ser9-GSK-3β phosphorylation in both the neurons expressing and lacking D1R. Transgenic drd1a-enhanced green fluorescent protein (EGFP) mice expressing EGFP in D1R-containing neurons were treated with vehicle or quinelorane (0.63 mg/kg) and perfused 10 min later. Confocal sections in the nucleus accumbens, showing immunofluorescence (red) for phospho-Ser9-GSK-3β (p-S9-GSK-3β, (a) alone (left panels), or in combination with EGFP fluorescence (green; right panels). Asterisks indicate p-S9-GSK-3β immunoreactivity in D1R-expressing neurons whereas arrows show phospho-GSK-3β immunoreactivity in D1R-lacking neurons. (b) Quantification of p-S9-GSK-3β-positive neurons among the EGFP-positive neurons (D1R+) or negative neurons (D1R−) in the central part of dorsal striatum (Str) and core of nucleus accumbens (NAcc) of vehicle (Veh)- or quinelorane (Quin)-treated drd1a-EGFP mice (= 10 mice per group). The effects of quinelorane and D1R expression were analyzed in each brain area using non-parametric Friedman's test. p-S9-GSK-3β: dorsal striatum, drug effect χ2r(1) = 13.2, p < 0.001, no D1R expression effect χ2r(1) = 1.88, NS; NAcc, drug effect χ2r(1) = 17.8, p < 0.001, no D1R expression effect χ2r(1) = 1.14, NS. Since the drug effects were systematically significant, we analyzed the quinelorane effect in each neuron population using corrected Mann–Whitney test, *p < 0.0125. Scale bars: 20 μm.

D3R activation is involved in the activation of both Akt/GSK-3β and mTORC1 pathways in the rat dorsal striatum and nucleus accumbens

D1R-expressing MSNs clearly displayed an activation of the Akt/GSK-3β pathway in response to quinelorane despite their low or inexistent levels of D2R expression (Bertran-Gonzalez et al. 2008). Since D3R has been reported to be preferentially co-expressed with D1R in MSNs (Le Moine and Bloch 1996; Surmeier et al. 1996; Schwartz et al. 1998), the activation of Akt/GSK-3β and mTORC1 may result, at least in the D1R-expressing MSNs, from D3R stimulation. The implication of D3R in the effects of quinelorane was therefore evaluated in rats and mice after pharmacological blockade and genetic deletion of D3R, respectively.

Pre-treatment of rats with S33084 (0.63 mg/kg), a selective D3R antagonist exhibiting a 100-fold higher affinity for D3R than for D2R (Millan et al. 2000b), injected alone had no significant effect on the phosphorylation levels of any of the proteins studied in either dorsal striatum or NAcc (Fig. 5). However, D3R antagonist pre-treatment significantly reduced the effects of quinelorane on the phosphorylation of Akt at Ser473 and Thr308, GSK-3b at Ser9, p70S6K at Thr389, rpS6 at Ser240/244, and 4E-BP1 at Thr37/46 in these two brain regions (Fig. 5). These results indicated that the quinelorane effects were mediated by D3R.

Figure 5.

Inhibition by a specific D3R antagonist of quinelorane-induced increase in Akt/GSK-3β and mTORC1 signaling in the rat striatum and nucleus accumbens. Groups of rats received two injections, first with the selective D3R antagonist S33084 (0.63 mg/kg) or vehicle, and 30 min later with quinelorane (0.16 mg/kg) or vehicle. Ten min after the second injection, phosphorylation was measured for Akt at Ser473 and Thr308, GSK-3β at Ser9, p-4E-BP1 at Thr37/46, p70S6K at Thr389, and rpS6 at Ser240/244. Representative immunoblots of phospho- and total proteins in the dorsal striatum (a and d). Quantification of the phospho/total protein ratio in the dorsal striatum (b, e) and nucleus accumbens (c, f). Histograms correspond to the mean + SEM of data obtained in two independent experiments (= 5 to 12 rats per group). One-way anova for the striatum: p-T308-Akt/Akt, F(3,37) = 4.6, < 0.05; p-S473-Akt/Akt, F(3,31) = 4.5, < 0.05; p-S9-GSK-3β/GSK-3β, F(3,38) = 6.5, < 0.05; p-T37/46-4E-BP1/4E-BP1, F(3,29) = 4.5, < 0.05; p-T389-p70S6K/p70S6K, F(3,33) = 6.1, < 0.05; p-S240/244-rpS6/rpS6, F(3,31) = 5.9, < 0.05. For the nucleus accumbens: p-T308-Akt/Akt, F(3,33) = 8.3, < 0.05; p-S473-Akt/Akt, F(3,36) = 7.2, < 0.05; p-S9-GSK-3β/GSK-3β, F(3,38) = 7.6, < 0.05; p-T37/46-4E-BP1/4E-BP1, F(3,35)  = 6.3, < 0.05; p-T389-p70S6K/p70S6K, F(3,37) = 6.3, < 0.05; p-S240/244-rpS6/rpS6, F(3,29) = 6.9, < 0.05. Post hoc comparison (Bonferroni's test): *< 0.05, **< 0.01, ***< 0.005, quinelorane- versus vehicle-treated; #< 0.05, ##< 0.01, ###< 0.001, quinelorane- versus quinelorane + S33084-treated.

To further explore the role of D3R, we used mutant mice lacking D3R (D3R-KO) (Accili et al. 1996). The Akt and GSK-3β phosphorylation levels, at Thr308 and Ser9, respectively, appeared increased in vehicle-treated D3R-KO mice when compared to WT mice, this difference reaching significance only in the NAcc (Fig. 6a and b). In WT mice, quinelorane (0.63 mg/kg) increased the phosphorylation of Akt and downstream proteins in both structures whereas, in D3R-KO mice, the quinelorane effects on phosphorylation of all tested proteins were blunted (Fig. 6). Thus, the results in D3R-KO mice strongly indicated that D3R plays a major role in the regulation of the Akt/GSK-3β/mTORC1 pathway in the striatum and NAcc.

Figure 6.

Blockade of quinelorane-induced protein phosphorylation in D3R-lacking mice. D3R-deficient mice as well as their wild-type littermates were treated with quinelorane (0.63 mg/kg) and killed 10 min later. Phosphorylation of various signaling proteins implicated in Akt, GSK-3β, and mTORC1-dependent cascade was measured by immunoblotting. The phospho/total protein ratio was quantified in the dorsal striatum (a) and NAcc (b). Histograms correspond to means + SEM of data (= 7 to 13 mice per group). Two-way anova was performed to test the effects of genotype and treatment (values are in Table S1). Post hoc comparison (Bonferroni's test): *< 0.05, **< 0.01, quinelorane- versus vehicle-treated; #< 0.05, vehicle-treated D3-KO mice versus vehicle-treated wild-type WT mice.

Discussion

This study demonstrates that short-term exposure to a potent D2R/D3R agonist, quinelorane, transiently increases the phosphorylation of Akt, GSK-3β, and effectors of mTORC1 in the striatum and NAcc. The phosphorylation of GSK-3β was observed in D1R- and D2R-expressing MSNs, suggesting a role of D3R and D2R in this effect. The participation of D3R was directly supported by the blockade of the actions of quinerolane by a selective D3R antagonist in rats and genetic invalidation of D3R in mice.

Activation of D2R/D3R elicits transient activation of Akt and downstream pathways

The transient increased in vivo phosphorylation of Akt, GSK-3β, and mTORC1 effectors induced by quinelorane clearly resulted from stimulation of D2R and/or D3R since these effects were completely prevented by pre-treatment with raclopride, a selective D2R/D3R antagonist. Similar D2R/D3R stimulation may participate in the activation of Akt and mTORC1 detected in the striatum shortly after amphetamine or cocaine injection (Brami-Cherrier et al. 2002; Svenningsson et al. 2003; Wu et al. 2011). The direct involvement of D2R and/or D3R is probable since stimulation of transfected D2R or D3R is able to rapidly activate the Akt/GSK-3β pathways in Chinese Hamster Ovary (CHO) cells (Mannoury la Cour et al. 2011).

In our experiments, both Akt and GSK-3β phosphorylation levels peaked 10 min after treatment and returned to baseline within 20 min after quinelorane injection. Beaulieu and colleagues (Beaulieu et al. 2004, 2005) reported that longer stimulation times (30 to 90 min) of D2-type receptors by amphetamine or apomorphine reduced the phosphorylation levels of Akt (specifically at Thr308) and GSK-3β Ser9 below the baseline in the mouse striatum. These delayed effects reflect processes related to D2R desensitization (Beaulieu et al. 2005) that lead to the formation of a complex associating D2R, β-arrestin2, protein phosphatase 2A, and Akt (Beaulieu et al. 2004, 2007a, b). In our experiments, quinelorane tended to decrease Thr308-Akt phosphorylation 90 min after injection but, contrasting with d-amphetamine, this effect was not significant (data not shown), probably because of the short action of quinelorane as compared to the long-lasting DA release induced by d-amphetamine. Altogether, this study and those of Beaulieu et al. strongly suggest that D2R/D3R stimulation induces a biphasic effect on the Akt/GSK-3β pathway characterized by a transient activation followed by a sustained inhibition of signaling (Beaulieu et al. 2011).

Activation of D3R by quinelorane induces phosphorylation of rpS6 at Ser240/244, but not at Ser235/236

In the dorsal striatum and NAcc, quinelorane increased the phosphorylation of rpS6 at Ser240/244 without affecting Ser235/236 sites (see Fig. 1f and h). This lack of effect was surprising since p70S6K was activated by quinelorane and could be expected to phosphorylate both Ser235/236 and Ser240/244 residues of rpS6. However, previously published studies showed that Ser235/236 can be phosphorylated by other kinases than p70S6K, such as p90RSK or PKA (Pende et al. 2004; Roux et al. 2007; Moore et al. 2009; Valjent et al. 2011). Under our conditions, p90RSK was likely unaltered since ERK1/2, its major regulator, was not activated in response to quinelorane. In addition, PKA activity is decreased by quinelorane because of the inhibition of cAMP production caused by D2R/D3R stimulation and the increase of dopamine- and cAMP-regulated phosphoprotein of Mr 32 kDa (DARPP-32) phosphorylation at Thr75, which is an inhibitor of PKA (Bibb et al. 1999). In our experimental conditions and in agreement with previous studies (Bateup et al. 2008), quinelorane led to an increase of DARPP-32 phosphorylation at Thr75 (data not shown). The decrease in PKA activity caused by quinelorane, either via a D2R/D3R-dependent inhibition of cAMP production and/or an increase in phospho-Thr75-DARPP-32 may neutralize any increase of phosphorylation at Ser235/236 induced by p70S6K activation. The difference between the phosphorylation patterns of rpS6 Ser235/236 and Ser240/244 may then result from their different dependence on cAMP/PKA/DARPP-32 signaling.

D3R mediates quinelorane-induced activation of Akt/GSK-3β and mTORC1 signaling

In immunofluorescence experiments, quinelorane increased the phosphorylation of GSK-3β in striatal neurons either showing or lacking D1R expression. In D1R-positive neurons, the activation of Akt/GSK-3β pathway is unlikely to result from direct activation of D2Rs which are not expressed in most of this MSN subpopulation (Matamales et al. 2009). In contrast to D2R, D3R is preferentially expressed in D1R-positive neurons (Le Moine and Bloch 1996; Surmeier et al. 1996; Schwartz et al. 1998). As quinelorane has no affinity for D1R sites and higher affinity for D3R than for D2R (Sokoloff et al. 1992), its action in these neurons more likely involves D3R. Accordingly, pretreatment with the selective D3R antagonist S33084 abolished the quinelorane-induced activation of Akt/GSK-3β and mTORC1 pathways in the dorsal striatum and NAcc of rats. Similar results were found in D3R-KO mice in which quinerolane effects on Akt/GSK-3β and mTORC1 pathways were blunted, demonstrating that D3R mediates, at least in part, these in vivo effects in striatal neurons. However, these data do not exclude a role of D2R in MSNs in which they are expressed. Additional work with D2R KO mice (highly selective D2R antagonists are not yet available) would be of interest to further define their role.

D3R is unevenly distributed within the striatum, with high levels of expression in the ventral striatum, particularly in the NAcc shell and to a lesser extent in the core, and low levels in the dorsal striatum (Le Moine and Bloch 1996; Surmeier et al. 1996; Schwartz et al. 1998). Despite these different levels of expression, D3R-dependent activations of Akt and mTORC1 pathways appeared similar in all striatal areas when the activations were expressed as percent of controls. However, since the basal levels of phosphorylation of the various targets of Akt or mTORC1 were not compared in the striatum and NAcc, the absolute effects of quinelorane could be different in the two structures.

How are D3Rs coupled to Akt activation? Previous work in cell culture has shown a possible coupling between D3R and Akt through classical mechanisms involving activation of phosphatidyl-inositol-3-kinase (PI3K) and the recruitment and phosphorylation of Akt by PDK1 (Mannoury la Cour et al. 2011). Similar processes may account for Akt activation by quinelorane in vivo. An additional question is the link between mTORC1 and Akt activation. Our results are compatible with Akt being the sole activator of the whole pathway, although we cannot exclude additional convergent signaling mechanisms. D3R could also be associated with inhibition of Akt signaling when it is long-lastingly altered. The basal levels of phosphorylation of Akt and GSK3β were increased in D3R-KO mice, an effect which is not observed after acute treatment of the D3 antagonist. Similar effects have been already reported in D3R-KO mice by Beaulieu et al. (Beaulieu et al. 2007b) who attributed it to a synergism of D3R with the processes of Akt inhibition mediated by prolonged D2R activation.

Possible functional role of the link between D3R and mTORC1 signaling

The role of Akt and mTORC1 in neuronal morphological changes is well established (Jaworski and Sheng 2006; Peineau et al. 2008). In mesencephalic DA neurons, recruitment of the Akt and/or mTORC1 pathways by D2R and D3R was shown to regulate axonal and dendritic arborization (Fasano et al. 2008; Collo et al. 2012). The D3R-dependent activation of these pathways observed in the dorsal striatum and NAcc suggests their possible participation in morphological changes of striatal neurons important for neuronal plasticity.

Many in vivo studies have demonstrated a role of D3R in cognitive processing (Watson et al. 2012a, b). There is evidence for an influence of D3R antagonists on working memory, attention, executive functions coordinated by fronto-cortico-striatal circuits, procedural learning integrated in the dorsal striatum, and incentive learning which involves the NAcc (Glickstein et al. 2005; Laszy et al. 2005; Millan and Brocco 2008; Millan et al. 2010, 2012). In view of the strong implication of Akt/GSK-3β and mTOR signaling in molecular mechanisms underpinning cognition (Hoeffer and Klann 2010; Peng et al. 2010), the present observations suggest that their activation may be involved in the influence of D3R upon cognitive processes depending on the striatum. In addition, although the precise contribution of D3R and D2R in the pathogenesis and management of schizophrenia remains unclear (Sokoloff et al. 2006; Gyertyan and Saghy 2007; Millan and Brocco 2008; Banasikowski et al. 2010; Heidbreder and Newman 2010), the present data provide novel insights into cellular mechanisms potentially involved in the modulation of psychotic states by D3R.

Concluding comments

In conclusion, this study demonstrates that stimulation of D3R activates both Akt/GSK–3β and mTORC1-dependent cellular signals in the NAcc and dorsal striatum of rats and mice, and that the involved D3R are located in D1R-containing cells (MSNs) and possibly in cells lacking D1R (presumptive D2R-containing neurons). These observations are consistent with – and greatly extend – previous in vitro studies of D3R coupling to Akt. Furthermore, they offer a molecular substrate for an improved understanding of the role of D3R in the control of synaptic plasticity as well as reward, motor function, and cognition, which are integrated in the NAcc and dorsal striatum. It will be of interest to evaluate to what extent Akt/GSK-3β and mTORC1 are involved in the putative therapeutic effects of D3R ligands in disorders like schizophrenia, drug abuse, and Parkinson's disease, and to extend this study to other structures expressing D3R including the cortex and cerebellum.

Acknowledgements

Research at Institut du Fer à Moulin was supported by grants from ANR, ERC, FRM, and FRC. MJ Salles was supported by a CIFRE fellowship during her PhD. The group of Denis Hervé and Jean-Antoine Girault is a member of the Paris School of Neuroscience (ENP) and of the Bio-Psy Laboratory of Excellence. The authors are grateful to Renata Coura for her help in statistical analysis.

Conflict of interest

Jean-Michel Rivet, Mark J. Millan, and Clotilde Mannoury la Cour are full time employees of Servier Pharmaceuticals Inc. The other authors declare no conflict of interest.

Ancillary