Distinct Roles of PDE4 and PDE10A in the Regulation of cAMP/PKA Signaling in the Striatum

Phosphodiesterase (PDE) is a critical regulator of cAMP/protein kinase A (PKA) signaling in cells. Multiple PDEs with different substrate specificities and subcellular localization are expressed in neurons. Dopamine plays a central role in the regulation of motor and cognitive functions. The effect of dopamine is largely mediated through the cAMP/PKA signaling cascade, and therefore controlled by PDE activity. We used in vitro and in vivo biochemical techniques to dissect the roles of PDE4 and PDE10A in dopaminergic neurotransmission in mouse striatum by monitoring the ability of selective PDE inhibitors to regulate phosphorylation of presynaptic [e.g., tyrosine hydroxylase (TH)] and postsynaptic [e.g., dopamine- and cAMP-regulated phosphoprotein of M r 32 kDa (DARPP-32)] PKA substrates. The PDE4 inhibitor, rolipram, induced a large increase in TH Ser40 phosphorylation at dopaminergic terminals that was associated with a commensurate increase in dopamine synthesis and turnover in striatum in vivo. Rolipram induced a small increase in DARPP-32 Thr34 phosphorylation preferentially in striatopallidal neurons by activating adenosine A2A receptor signaling in striatum. In contrast, the PDE10A inhibitor, papaverine, had no effect on TH phosphorylation or dopamine turnover, but instead robustly increased DARPP-32 Thr34 and GluR1 Ser845 phosphorylation in striatal neurons. Inhibition of PDE10A by papaverine activated cAMP/PKA signaling in both striatonigral and striatopallidal neurons, resulting in potentiation of dopamine D1 receptor signaling and inhibition of dopamine D2 receptor signaling. These biochemical results are supported by immunohistochemical data demonstrating differential localization of PDE10A and PDE4 in striatum. These data underscore the importance of individual brain-enriched cyclic-nucleotide PDE isoforms as therapeutic targets for neuropsychiatric and neurodegenerative disorders affecting dopamine neurotransmission.


Introduction
Dopamine plays a central role in the regulation of motor and cognitive functions. The cAMP/protein kinase A (PKA) signaling cascade is essential for dopamine neurotransmission. Dopamine, acting on D 1 receptors, stimulates cAMP/PKA signaling via G s/olfmediated activation of adenylyl cyclase (Hervé et al., 2001), whereas dopamine, acting on D 2 receptors, inhibits cAMP/PKA signaling via G i -mediated inactivation of adenylyl cyclase (Stoof and Kebabian, 1981). In mammalian striatum, the synthesis of dopamine by tyrosine hydroxylase (TH) (Harada et al., 1996;Dunkley et al., 2004) and the release of dopamine from nigrostriatal dopaminergic terminals (Zhu et al., 2004;Seino and Shibasaki, 2005) are regulated by the cAMP/PKA signaling cascade. In postsynaptic striatal neurons, DARPP-32, a dopamine-and cAMP-regulated phosphoprotein of M r 32 kDa, is a major target for the cAMP/PKA signaling cascade (Greengard et al., 1999;Svenningsson et al., 2004). DARPP-32 is expressed in both the D 1 receptor-enriched striatonigral and D 2 receptor-enriched striatopallidal neurons (Bateup et al., 2008). Phosphorylation at Thr34 by PKA converts DARPP-32 into a potent inhibitor of the widespectrum protein phosphatase-1 (PP-1). The inhibition of PP-1 thereby controls the phosphorylation state and activity of many downstream physiological effectors, including various neurotransmitter receptors and voltage-gated ion channels. Mice lacking DARPP-32 are deficient in their molecular, electrophysiological, and behavioral responses to dopamine, drugs of abuse, and antipsychotic medication, indicating an essential role for DARPP-32 in dopaminergic signaling (Fienberg et al., 1998). Dopaminergic signaling is controlled by phosphodiesterases (PDEs), which degrade cAMP and downregulate cAMP/PKA signaling. PDEs are encoded by 21 genes and subdivided into 11 families according to structural and functional properties (Bender and Beavo, 2006). The brain expression and subcellular localization of PDE families are tightly regulated. Multiple PDEs are expressed in neurons, each with distinct roles in cAMP and cyclic GMP (cGMP) signaling. Several PDE families are expressed in striatum . For example, PDE1B is abundantly expressed in striatum (Polli and Kincaid, 1994). Mice lacking PDE1B exhibit increased DARPP-32 phosphorylation at Thr34, indicating that PDE1B normally downregulates cAMP/ PKA signaling in striatal neurons (Reed et al., 2002). The role of other PDEs, such as PDE4 and PDE10A, in regulating the DARPP-32 signaling pathway is unknown. PDE10A is highly enriched in striatum (Fujishige et al., 1999;Coskran et al., 2006;Xie et al., 2006). Inhibition of PDE10A by papaverine increases phosphorylation of cAMP-dependent substrates, including the cAMP-response element-binding protein (CREB) and extracellular receptor kinase (ERK), by activating cAMP/PKA signaling (Siuciak et al., 2006b). PDE4B, another striatal-enriched PDE, likely plays a regulatory role in dopaminergic neurotransmission because inhibition of PDE4 by rolipram stimulates dopamine synthesis (Kehr et al., 1985;Schoffelmeer et al., 1985;Yamashita et al., 1997a). However, the precise role of PDE4 in dopaminergic neurotransmission is currently unknown.
Here, we identify distinct roles for PDE4 and PDE10A in cAMP/PKA signaling in striatonigral and striatopallidal neurons and at dopaminergic terminals. PDE10A predominantly regulates DARPP-32 phosphorylation in the same direction as a dopamine D 2 antagonist in striatopallidal neurons, whereas PDE4 predominantly regulates TH phosphorylation at dopaminergic terminals. Thus, PDE4 and PDE10A have distinct roles in striatal dopaminergic neurotransmission conferred by their discrete cellular localization.
Frozen tissue samples were sonicated in boiling 1% SDS and boiled for an additional 10 min. Small aliquots of the homogenate were retained for protein determination by the BCA protein assay method (Pierce). Equal amounts of protein (40 g) were loaded onto 4 -12% polyacrylamide Bis-Tris gels (#345-0124; Bio-Rad), separated by electrophoresis, and transferred to nitrocellulose membranes (0.2 m) (Schleicher and Schuell).
Immunoprecipitations of Flag-and Myc-tagged DARPP-32 in neostriatal slices from D 1 -DARPP-32-Flag/D 2 -DARPP-32-Myc mice. D 1 -DARPP-32-Flag/D 2 -DARPP-32-Myc transgenic mice express Flag-and Myc-tagged DARPP-32 under the control of dopamine D 1 and D 2 receptor promoters, respectively (Bateup et al., 2008). In the striatum, Flagtagged DARPP-32 was shown to be expressed selectively in D 1 receptorenriched striatonigral neurons, and Myc-tagged DARPP-32 selectively in D 2 receptor-enriched striatopallidal neurons. Using antibodies against Flag and Myc tags, we can selectively immunoprecipitate DARPP-32 from D 1 receptor-and D 2 receptor-expressing neurons and analyze the phosphorylation state of DARPP-32 in a neuronal type-specific manner. In each experiment, six striatal slices were prepared from one mouse, and were divided into three treatment conditions. In each treatment condition, six slices, collected from three mice (two slices from each mouse), were used for the analysis of DARPP-32 phosphorylation. Six striatal slices were sonicated in 720 l of immunoprecipitation (IP) lysis buffer [50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 1% SDS, 100 nM okadaic acid, phosphatase inhibitor mixture (#P5726; Sigma-Aldrich), and protease inhibitor mixture (#11873580001; Roche)]. After determination of protein concentration, 15 g of protein was saved for the analysis of DARPP-32 phosphorylation in total striatal homogenate, and the residual homogenates were used for IPs. In each IP from striatal homogenate, 50 l of washed EZView Red anti-Flag M2 affinity gel (Sigma-Aldrich) and 45 l of anti-Myc antibody (Novus Biologicals) coupled to magnetic beads (3 g of Myc antibody for every 5 l of magnetic beads) (Dynabeads M-280 Tosylactivated; Invitrogen) were added. The homogenate/antibody mixture was gently rotated overnight at 4°C. After the overnight incubation, the Myc magnetic beads were separated from the homogenate/antibody mixture using a magnetic particle concentrator (Invitrogen), and then the Flag affinity gels were separated by centrifugation. The Myc magnetic beads and Flag affinity gels were washed with 1ϫ PBS three times. After the final wash, 30 l of sample buffer was added, and samples were boiled for 2 min.
Preparation of striatal tissues for analysis of protein phosphorylation in vivo. Mice were injected intraperitoneally with rolipram (10 mg/kg), papaverine (30 mg/kg), or haloperidol (0.1, 0.3, or 1.0 mg/kg) or with vehicle control (5 ml/kg body weight). The vehicle for rolipram contained the following (final concentration): 5% DMSO, 5% Tween 20, 15% polyethylene glycol 400 (PEG 400), and 75% water. The vehicle for papaverine was saline, and the vehicle for haloperidol was acidified saline titrated to pH 5.5. At the indicated time points after injection, the mice were killed by focused microwave cranial irradiation (4.5-5.0 kW for 1.3 s) using a small animal microwave (Muromachi Kikai), which inactivates protein kinases and phosphatases to preserve in vivo levels of protein phosphorylation (O'Callaghan and Sriram, 2004). Striata were dissected from each mouse brain, frozen in liquid nitrogen, and stored at Ϫ80°C until assayed.
Frozen samples of brain tissue were processed as described for slices. Equal amounts of protein (15-30 g) were loaded on 10% polyacrylamide BIS-Tris gels (Bio-Rad), separated by electrophoresis, and transferred to nitrocellulose membranes (0.2 M) (Schleicher and Schuell).
The membrane was incubated with a goat anti-mouse or rabbit Alexa 680-linked IgG (1:5000 dilution; Invitrogen) or a goat anti-mouse or rabbit IRDye800-linked IgG (1:5000 dilution; Rockland). Fluorescence at infrared wavelengths was detected by the Odyssey infrared imaging system (LI-COR) and quantified using Odyssey software. In an individual experiment, samples from control and drug-treated slices were analyzed on the same immunoblot. For each experiment, values obtained for slices were normalized to values for either the control or the drug-treated slices, as described in the figure legends. Normalized data from multiple experiments were averaged and statistical analysis was performed as described in the figure legends.
Analysis of the levels of dopamine and its metabolites in striatal tissue in vivo. Mice were injected intraperitoneally with rolipram (10 mg/kg, in 5% DMSO plus 5% Tween 20 plus 15% PEG 400 plus 75% water), papaverine (30 mg/kg, in saline), or haloperidol (0.3 mg/kg, in acidified saline titrated to pH 5.5) or with vehicle solution (5 ml/kg body weight). Thirty minutes after injection, the mice were killed by focused microwave cranial irradiation as described above. Striata were dissected from each mouse brain, frozen in liquid nitrogen, and stored at Ϫ80°C until assayed.
Dopamine and its metabolites were quantified by high-performance liquid chromatography with electrochemical detection (HPLC-EC; Waters). Tissues were homogenized in 300 l of ice-cold 0.2 M perchloric acid, containing 1 M dihydroxybenzylamine as internal standard, and centrifuged at 10,000 g for 10 min at 4°C. The supernatant was filtered through a 0.2 m membrane, and an aliquot (10 l) was injected from a temperature-controlled (4°C) automatic sample injector (Waters 717plus Autosampler) connected to a Waters 515 HPLC pump. Catecholamines were separated on a C18 reverse-phase column (LC-18 RP; Waters SYMMETRY, 25 cm ϫ 4.6 mm; 5 m), electrochemically detected (Waters 464 Pulsed Electrochemical Detector; range 10 nA, potential ϩ0.7 V), and analyzed using Millennium software (Waters). The mobile phase, pH 3.0, for isocratic separation of dopamine consisted of dibasic sodium phosphate (75 mM), octane sulfonic acid (1.7 mM), acetonitrile (10% v/v), and EDTA (25 M). Flow rate was maintained at 1 ml/min. Dopamine, 3,4-dihydroxyphenylacetic acid (DOPAC), and homovanillic acid (HVA) standards (0.5-25 pmol) were prepared in 0.2 M perchloric acid containing dihydroxybenzylamine. Recovery of each analyte was adjusted with respect to the internal standard and quantified from a standard curve. The levels of dopamine and its metabolites were expressed as micrograms per gram of wet tissue.
Treatment of neostriatal slices with rolipram for 60 min increased the levels of phospho-Thr34 DARPP-32 and phospho-Ser845 GluR1 at a high concentration of 100 M by approximately twofold, but not significantly at lower concentrations ( Fig. 1 A, B, left and center panels). In the analysis of the time course, rolipram at 10 M slightly increased DARPP-32 Thr34 phosphorylation at 2 min of incubation and GluR1 Ser845 phosphorylation at 30 and 60 min. Treatment with rolipram increased ERK2 Thr202/Tyr204 phosphorylation in a dose-and timedependent manner similar to that seen with papaverine (Fig. 1C).
These results clearly demonstrate that the effects of the PDE10A inhibitor, papaverine, on the phosphorylation of postsynaptic PKA substrates in striatal neurons, DARPP-32 ( p Ͻ 0.01 for both dose-response and time course, two-way ANOVA) and GluR1 ( p Ͻ 0.01 for both dose-response and time course, two-way ANOVA), were much larger than those of the PDE4 inhibitor, rolipram.

Effect of papaverine and rolipram on TH Ser40 and synapsin I Ser9 phosphorylation in neostriatal slices
We next examined the effect of papaverine and rolipram on the phosphorylation of presynaptic PKA substrates, TH at Ser40 and synapsin I at Ser9, which are known to play an important role in the synthesis of dopamine (Harada et al., 1996;Dunkley et al., 2004) and the release of neurotransmitters (Fiumara et al., 2004), respectively. Treatment with rolipram for 60 min increased the levels of phospho-Ser40 TH and phospho-Ser9 synapsin I maximally at a concentration of 10 M by threefold and fourfold, respectively, with a half-maximal effect of ϳ100 nM (Fig. 2 A, B, left and center panels). Rolipram at 10 M induced a maximal effect on TH Ser40 and synapsin I Ser9 phosphorylation within 10 min of incubation ( Fig. 2 A, B, right panels).
Treatment with papaverine at a high concentration of 10 M for 60 min increased the levels of phospho-Ser40 TH and phospho-Ser9 synapsin I by approximately twofold, but not at lower concentrations (Fig. 2 A, B, left and center panels). Papaverine at 1 M did not affect TH Ser40 and synapsin I Ser9 phosphorylation at any incubation time (Fig. 2 A, B, right panels). The effects of the PDE4 inhibitor, rolipram, on the phosphorylation of TH ( p Ͻ 0.01 for both dose-response and time course, twoway ANOVA) and synapsin I ( p Ͻ 0.01 for both dose-response and time course, two-way ANOVA) were much larger than those of the PDE10A inhibitor, papaverine.
These results suggest that the inhibition of PDE10A by papaverine induces activation of cAMP/PKA signaling in medium spiny neurons, leading to the phosphorylation of DARPP-32 and GluR1, and that the inhibition of PDE4 by rolipram induces activation of cAMP/PKA signaling mainly at presynaptic, dopaminergic terminals, leading to the phosphorylation of TH and synapsin I. ERK2 is likely expressed both in medium spiny neurons and at dopaminergic terminals, and therefore, ERK2 phosphorylation is similarly affected by papaverine and rolipram.  for the indicated times (right panels). Changes in the phosphorylation of DARPP-32 at Thr34 (the PKA site), GluR1 at Ser845 (the PKA site), and ERK at Thr202/Tyr204 (the MEK site) were determined by Western blotting using phosphorylation-state-specific antibodies. Typical immunoblots are shown in left panels. Data represent means Ϯ SEM for 5-13 experiments. *p Ͻ 0.05, **p Ͻ 0.01 compared with untreated slices for papaverine; § p Ͻ 0.05, § § p Ͻ 0.01 compared with untreated slices for rolipram; one-way ANOVA followed by Newman-Keuls test.
dopamine D 1 and adenosine A 2A receptor signaling to the papaverine-induced increase in DARPP-32 Thr34 phosphorylation, the effect of papaverine was examined in the presence of a dopamine D 1 receptor antagonist, SCH23390, or an adenosine A 2A receptor antagonist, ZM241385 (Fig. 3A). The papaverineinduced increase in DARPP-32 Thr34 phosphorylation was attenuated both by SCH23390 and by ZM241385. However, the papaverine-induced increase was not affected by an inhibitor of soluble guanylyl cyclase, ODQ, indicating that papaverine did not activate cGMP/protein kinase G (PKG) signaling (Siuciak et al., 2006b) coupled to DARPP-32 Thr34 phosphorylation in this slice preparation (Nishi et al., 2005).
We next examined the effect of a dopamine D 1 receptor agonist, SKF81297 (1 M), and an adenosine A 2A receptor agonist, CGS21680 (5 M), on DARPP-32 Thr34 phosphorylation in the absence or presence of papaverine (Fig. 3B). The effects of SKF81297 and CGS21680 on DARPP-32 Thr34 phosphorylation were enhanced by papaverine. These results suggest that PDE10A regulates both dopamine D 1 receptor-stimulated cAMP/PKA signaling in striatonigral neurons and adenosine A 2A receptorstimulated cAMP/PKA signaling in striatopallidal neurons.
Dopamine D 2 receptors are expressed in striatopallidal neurons, couple to G i , and thereby inhibit cAMP/PKA signaling in striatopallidal neurons (Nishi et al., 1997). Because dopamine D 2 receptors play a central role in the regulation of psychomotor functions by dopamine, the role of PDE10A in dopamine D 2 receptor signaling was examined (Fig. 3C). Treatment with a dopamine D 2 receptor agonist, quinpirole, decreased the level of phospho-Thr34 DARPP-32 to 50% of control, whereas treatment with a dopamine D 2 receptor antagonist, raclopride, did not affect DARPP-32 Thr34 phosphorylation. Treatment with papaverine (10 M) increased DARPP-32 Thr34 phosphorylation by 14-fold. In the presence of papaverine, quinpirole failed to decrease the level of phospho-Thr34 DARPP-32. These results suggest that the PDE10A inhibitor, papaverine, attenuates the effect of a dopamine D 2 receptor agonist in striatopallidal neurons. The inhibitory effect of quinpirole on TH Ser40 phosphorylation was not attenuated by papaverine at dopaminergic terminals (data not shown), where papaverine has little effect.

PDE4 regulates adenosine A 2A receptor signaling in striatopallidal neurons
A possible role of PDE4 in dopamine D 1 receptor and adenosine A 2A receptor signaling in striatal neurons was examined. Treatment of slices with rolipram (100 M) slightly increased the level of phospho-Thr34 DARPP-32, but the effect was not statistically significant in this series of experiments (Fig. 4). Both CGS21680 (5 M) and SKF81297 (1 M) increased DARPP-32 Thr34 phos- Changes in the phosphorylation of TH at Ser40 (the PKA site) and synapsin I at Ser9 (the PKA/CaMKI site), which are selectively expressed at presynaptic terminals, were determined by Western blotting using phosphorylation-state-specific antibodies. Typical immunoblots are shown in left panels. Data represent means Ϯ SEM for 5-13 experiments. **p Ͻ 0.01 compared with untreated slices for papaverine; § § p Ͻ 0.01 compared with untreated slices for rolipram; one-way ANOVA followed by Newman-Keuls test.
phorylation by approximately fourfold. Rolipram enhanced the stimulatory effect of CGS21680 on DARPP-32 Thr34 phosphorylation, but not that of SKF81297. These results suggest that PDE4 preferentially regulates adenosine A 2A receptor-stimulated cAMP/PKA signaling in striatopallidal neurons, in addition to cAMP/PKA signaling at dopaminergic terminals (Fig. 2).

Expression patterns of PDE10A and PDE4B in striatal neurons
The expression patterns of PDE10A and PDE4B in striatal neurons were analyzed by immunohistochemistry. DARPP-32 was used as a marker of medium spiny neurons (Ouimet et al., 1992). The expression of PDE10A was detected in all DARPP-32positive striatal neurons (Fig. 5A), and a punctate pattern of PDE10A staining was observed. Strong immunoreactivity of PDE4B was detected in a subset of DARPP-32-positive neurons (Fig. 5B).
To determine the subset of medium spiny neurons that expresses PDE4B, PDE4B expression was analyzed using neostriatal tissues from D1-DARPP-32-Flag/D2-DARPP-32-Myc mice. In these mice, Flag-tagged DARPP-32 is expressed in striatonigral neurons, and Myc-tagged DARPP-32 is expressed in striatopallidal neurons (Bateup et al., 2008). In confirmation of a previous report (Bateup et al., 2008), there was no overlap of the expression of Flag-tagged and Myc-tagged DARPP-32 (Fig. 6 A). In this study, we found that the expression of PDE4B was higher in Myc-positive striatopallidal neurons than in Flag-positive striatonigral neurons (Fig. 6 B, C).

Regulation of DARPP-32 Thr34 phosphorylation in striatonigral and striatopallidal neurons by papaverine and rolipram
Neuronal type-specific regulation of DARPP-32 Thr34 phosphorylation by papaverine and rolipram was investigated using neostriatal slices from D 1 -DARPP-32-Flag/D 2 -DARPP-32-Myc mice. Flag-and Myc-tagged DARPP-32 (D 1 -Flag and D 2 -Myc) was immunoprecipitated from D 1 receptor-enriched striatonigral and D 2 receptor-enriched striatopallidal neurons, respectively, and the phosphorylation states of DARPP-32 at Thr34 in the two types of neurons were analyzed (Fig. 7). Treatment of neostriatal slices from D 1 -DARPP-32-Flag/D 2 -DARPP-32-Myc mice with papaverine (10 M) increased the level of phospho-Thr34 DARPP-32 by approximately sevenfold in total striatal homogenate. Papaverine increased the levels of phospho-Thr34 Flag-and Myc-tagged DARPP-32 by twofold and sixfold, respectively. Treatment of slices with rolipram (100 M) increased the level of phospho-Thr34 DARPP-32 by ϳ1.9-fold in total striatal homogenate. Rolipram increased the levels of phospho-Thr34 Flag-and Myc-tagged DARPP-32 by 2.2-fold and 2.6-fold, respectively. We analyzed the relative stoichiometry of the phosphorylation of Flag-and Myc-tagged DARPP-32 at Thr34 in slice preparations under basal conditions, and found that the Thr34 phosphorylation of Myc-tagged DARPP-32 was ϳ4.5-fold higher than that of Flag-tagged DARPP-32. Taking this stoichiometry into consideration, the phosphorylated level of DARPP-32 at Thr34 is 13.5-fold higher after papaverine treatment, and 5.3fold higher after rolipram treatment, in the striatopallidal compared with the striatonigral neurons. and adenosine deaminase (30 g/ml for 60 min), the effects of a dopamine D 1 agonist, SKF81297 (1 M for 5 min) and an adenosine A 2A receptor agonist, CGS21680 (5 M for 2 min), on DARPP-32 Thr34 phosphorylation were examined. Adenosine deaminase, additionally included in the incubation medium to decease tissue content of adenosine, reduced the basal and papaverine-induced levels of phospho-Thr34 DARPP-32. Data represent means Ϯ SEM for 6 -19 experiments. C, The effect of a dopamine D 2 receptor agonist, quinpirole (1 M for 10 min), and a dopamine D 2 receptor antagonist, raclopride (1 M for 10 min), on DARPP-32 Thr34 phosphorylation was examined in the absence (left) or presence (right) of papaverine (10 M for 70 min). Data represent means Ϯ SEM for 7-13 experiments. **p Ͻ 0.01 compared with control; § § p Ͻ 0.01 compared with papaverine alone; † † p Ͻ 0.01 compared with CGS21680 alone; ¶ ¶ p Ͻ 0.01 compared with SKF81297 alone; one-way ANOVA followed by Newman-Keuls test.

Effect of papaverine and rolipram on TH phosphorylation at presynaptic dopaminergic terminals in vivo
The role of PDE10A and PDE4 in the regulation of the phosphorylation state of the presynaptic PKA substrate, TH at Ser40, was examined in intact animals. A single injection of papaverine (30 mg/kg, i.p.) or rolipram (10 mg/kg, i.p.) did not affect the basal level of phospho-Ser40 phosphorylation (Fig. 8A) up to 60 min after intraperitoneal injection. We examined whether either of these phosphodiesterase inhibitors might potentiate the ability of a neuroleptic compound to elevate TH phosphorylation state at Ser40. A dose-response study shown in Figure 8A (inset) revealed that the neuroleptic drug, haloperidol, induced a submaximal increase in TH Ser40 phosphorylation at a concentration of 0.3 mg/kg, i.p. Rolipram (Fig. 8C), but not papaverine (Fig. 8B), enhanced the haloperidol-induced increase in TH Ser40 phosphorylation.

Effect of papaverine and rolipram on GluR1 phosphorylation in striatal neurons in vivo
The role of PDE10A and PDE4 in the regulation of the phosphorylation state of the postsynaptic PKA substrate, GluR1 at Ser845, was examined in intact animals. The basal level of phospho-Ser845 GluR1 was increased maximally by 2.3-fold at 15 min after a single injection of papaverine (30 mg/kg, i.p.). Rolipram (10 mg/kg, i.p.), given as a single injection, induced a much smaller effect on the level of phospho-Ser845 GluR1 at 30 min after injection (Fig. 8 D). The effects of papaverine and rolipram were examined in combination with haloperidol. Haloperidol (0.3 mg/kg, i.p.) alone increased GluR1 Ser845 phosphorylation by ϳ1.5-fold. Papaverine clearly potentiated the haloperidolinduced increase in GluR1 Ser845 phosphorylation, increasing levels to ϳ3.4-fold (Fig. 8 E). In contrast to the effect of papaverine, rolipram induced only a small additional effect on the GluR1 Ser845 phosphorylation, enhancing the haloperidol-induced increase from ϳ1.4fold to ϳ1.7-fold (Fig. 8 F). Phosphorylation of DARPP-32 was also analyzed, but changes in DARPP-32 Thr34 phosphorylation were not detected in response to papaverine, rolipram and haloperidol alone, or in combination (data not shown).
These results in vivo demonstrate that the inhibition of PDE4 by rolipram predominantly enhances haloperidol-induced activation of cAMP/PKA signaling at dopaminergic terminals, leading to TH Ser40 phosphorylation, and that the inhibition of  PDE10A by papaverine selectively enhances haloperidol-induced activation of cAMP/PKA signaling in striatal neurons, leading to GluR1 Ser845 phosphorylation.

Effect of papaverine and rolipram on dopamine metabolism in the striatum in vivo
We examined whether the enhancement of haloperidol-induced TH phosphorylation by rolipram reflects a functional effect of the inhibitor on dopamine synthesis and turnover in the striatum in vivo. The effect of the drug on the rate of dopamine metabolism was quantitated by measuring the relative tissue concentrations of dopamine and its major metabolites, DOPAC and HVA in the striatum (Table 1). An increase in the ratio of DOPAC/dopamine and/or HVA/dopamine indicates an increase in striatal dopamine biosynthesis. A single injection of rolipram (10 mg/kg, i.p.) or papaverine (30 mg/kg, i.p.) per se did not affect the tissue content of dopamine, DOPAC, or HVA or the ratio of DOPAC/ dopamine or HVA/dopamine. Haloperidol (0.3 mg/kg, i.p.) increased the tissue content of DOPAC and HVA and the ratios of DOPAC/dopamine and HVA/dopamine, as previously reported (Boyar and Altar, 1987). Rolipram further enhanced the haloperidol-induced increase in DOPAC content and DOPAC/ dopamine ratio, but did not affect the haloperidol-induced increase in HVA content or HVA/dopamine ratio. Papaverine did not affect the haloperidol-induced increase in DOPAC content or DOPAC/dopamine ratio, but slightly reduced the haloperidolinduced increase in HVA content and HVA/dopamine ratio. The results suggest that PDE4 plays a functional role in modulating do-pamine synthesis and the turnover of dopamine to DOPAC at dopaminergic terminals in response to haloperidol administration.

Discussion
The present study demonstrates that PDE10A and PDE4 are differentially expressed in neuronal subtypes in the striatum and play distinct roles in dopaminergic neurotransmission (see Fig.  9). PDE4 predominantly regulates cAMP/PKA signaling at dopaminergic terminals in the striatum. The inhibition of PDE4 by rolipram increases TH phosphorylation and dopamine synthesis, leading to an increase in dopaminergic tone. PDE4 also regulates cAMP/PKA signaling in medium spiny neurons, preferentially in striatopallidal neurons. In contrast, PDE10A exclusively regulates cAMP/PKA signaling in medium spiny neurons. The inhibition of PDE10A by papaverine activates cAMP/PKA signaling in medium spiny neurons comprising striatopallidal and striatonigral projections, leading to the inhibition of dopamine D 2 receptor signaling in striatopallidal neurons and the potentiation of dopamine D 1 receptor signaling in striatonigral neurons. Both PDE10A and PDE4 have been proposed as therapeutic targets for psychotic disorders Hebb and Robertson, 2007). Elucidation of the distinct roles of PDE10A and PDE4 in dopaminergic neurotransmission reveals the mechanisms by which selective PDE10A and PDE4 inhibitors may exert antipsychotic activity.

Role of PDE10A in striatal neurons
Using both pharmacological and genetic tools, we have shown that PDE10A regulates cAMP/PKA signaling in both striatopallidal and striatonigral neurons. In striatopallidal neurons, PDE10A inhibition by papaverine activated cAMP/PKA signaling by simultaneously potentiating adenosine A 2A receptor signaling and inhibiting dopamine D2 receptor signaling. In striatonigral neurons, PDE10A inhibition by papaverine also activated cAMP/ PKA signaling, leading to the potentiation of dopamine D 1 receptor signaling. PDE10A inhibition had greater effects on signaling in striatopallidal neurons. This observation was unanticipated because immunocytochemical data clearly show PDE10A to be abundant in most medium spiny neurons (Figs. 5,6). The basis for the predominant effect of papaverine in striatopallidal cells is unclear. One possible explanation is that tonic activity of adenosine A 2A receptor signaling may have contributed to the large effects observed in striatopallidal cells, although adenosine deaminase was included during preincubation of slices. Alternatively, there may be functional differences in PDE10A inhibition in the two medium spiny cell populations that reflect subtle differences in PDE10A abundance or PDE10A subcellular targeting of PDE10A within striatopallidal and striatonigral neurons. More detailed neuroanatomical studies would be needed to resolve this issue.
The present study suggests that PDE10A exerts preferential effects on cAMP-mediated, relative to cGMP-mediated, signaling in striatal neurons. For example, although PDE10A is reported to hydrolyze cAMP and cGMP with near equal affinity (Bender and Beavo, 2006;Siuciak et al., 2006b), papaverine had no effect on the modulation of cGMP/PKG/phospho-Thr34 DARPP-32 signaling (Fig. 3A). Thus, the PDE10A inhibitor appears to effectively counteract dopamine D 2 receptor signaling in striatopallidal neurons and potentiate dopamine D 1 receptor signaling in striatonigral neurons mainly via cAMP-mediated effects.
Under in vivo conditions, papaverine stimulates cAMP/PKA signaling, leading to the phosphorylation of GluR1 at Ser845 in striatal neurons (Fig. 8 D), as observed in slice preparations. It is Typical immunoblots for detection of phospho-Thr34 DARPP-32 (P-T34 D32) and total DARPP-32 (total D32) in the same membrane are shown at the top. The levels of phospho-Thr34 DARPP-32 and total DARPP-32 were quantified by the Odyssey infrared imaging system, and the data (phospho-Thr34 DARPP-32/total DARPP-32) were normalized to values obtained with untreated slices. Data represent means Ϯ SEM for four experiments. *p Ͻ 0.05, **p Ͻ 0.01 compared with control; one-way ANOVA followed by Newman-Keuls test.
likely that the ability of papaverine to activate cAMP/PKA signaling is suppressed by high dopamine D 2 tone, because the effect of papaverine was enhanced in the presence of a typical antipsychotic, haloperidol (Fig. 8 E). Haloperidol treatment has recently been shown to activate cAMP/PKA signaling selectively in striatopallidal neurons by inhibiting dopamine D 2 receptors (Bateup et al., 2008). It is noteworthy that the PDE10A inhibitor and the antipsychotic drug inhibit dopamine D 2 receptor signaling by different mechanisms and act synergistically to affect phosphoproteins.
In agreement with our results, inhibition of PDE10A by papaverine was previously shown to increase the phosphorylation of CREB and ERK by activating cAMP/PKA signaling in the striatum (Siuciak et al., 2006b). In a behavioral analysis, papaverine treatment (Siuciak et al., 2006b;Becker and Grecksch, 2008) or disruption of the PDE10A gene (Siuciak et al., 2006a) reduced spontaneous locomotor activity, phencyclidine-stimulated locomotor activity, and conditioned avoidance responding. Because inhibition of conditioned avoidance responding serves as a sensitive measure of antipsychotic activity of drugs (Wadenberg and Hicks, 1999), PDE10A inhibitors have been proposed as therapeutic reagents for schizophrenia (Hebb and Robertson, 2007;Menniti et al., 2007). The pharmacological profile of the PDE10A inhibitor, papaverine, to counteract dopamine D 2 receptor signaling and potentiate dopamine D 1 receptor signaling, resembles that of atypical antipsychotics, and therefore supports the idea that PDE10A inhibition would be expected to address the symptoms and cognitive deficits of psychosis.

Role of PDE4 at dopaminergic terminals
This study demonstrates an important role for PDE4 in the regulation of cAMP/PKA signaling at dopaminergic terminals in the striatum. Inhibition of PDE4 activity by rolipram increased the state of phosphorylation of TH at Ser40 in neostriatal slices. In contrast, under in vivo conditions, inhibition of PDE4 by rolipram alone was insufficient to induce the phosphorylation of TH at Ser40. Rolipram did elicit a significant induction of TH phosphorylation when coadministered with the antipsychotic agent, haloperidol. The lack of effect of rolipram alone on TH phosphorylation in vivo may be attributable to high dopamine D 2 tone, which would be likely to exist in the intact animal but not in the isolated striatal slice preparation.
The PKA-dependent phosphorylation of TH at Ser40 increases the catalytic activity of TH (Harada et al., 1996;Dunkley et al., 2004), the rate-limiting step in dopamine biosynthesis (Nagatsu et al., 1964). The inhibition of PDE4 and consequent enhancement of TH phosphorylation at Ser40 in the presence of haloperidol was accompanied by an increase in DOPAC/dopamine ratio in the striatum in vivo. This increase in relative DOPAC concentration reflects an increased metabolism of dopa- ; red open circles), or papaverine (30 mg/kg, i.p.; black closed circles) and killed 15, 30, or 60 min later. Inset, Dose-response curve of haloperidol for phospho-Ser40 TH levels. Mice were treated with either vehicle (saline, 0 time point) or one of three doses of haloperidol (0.1, 0.3, or 1.0 mg/kg, i.p.) and killed 30 min later. **p Ͻ 0.01 compared with vehicle; † † p Ͻ 0.01 compared with haloperidol at 0.3 mg/kg; one-way ANOVA followed by Newman-Keuls test. B, Cotreatment of mice with papaverine does not potentiate haloperidol-induced increases in phospho-Ser40 TH levels. Mice were treated with papaverine alone (30 mg/kg), haloperidol alone (0.3 mg/kg), or both, and killed 30 min later. *p Ͻ 0.05 compared with vehicle; § p Ͻ 0.05 compared with papaverine alone; one-way ANOVA followed by Newman-Keuls test. C, Cotreatment of mice with rolipram potentiates haloperidol-induced increases in phospho-Ser40 TH levels. Mice were treated with rolipram (10 mg/kg) alone, haloperidol alone (0.3 mg/kg), or both, and killed 30 min later. **p Ͻ 0.01 compared with vehicle; † † p Ͻ 0.01 compared with rolipram alone; ¶ ¶ p Ͻ 0.01 compared with haloperidol alone; one-way ANOVA followed by Newman-Keuls test. D, Time course of the effect of papaverine (black closed circles) and rolipram (red open circles) on GluR1 Ser845 phosphorylation. Striatal samples were prepared as described in A. *p Ͻ 0.05 compared with time 0 for papaverine; † p Ͻ 0.05 compared with time 0 for rolipram; one-way ANOVA followed by Newman-Keuls test. E, Cotreatment of mice with papaverine, as described in B, potentiates haloperidol-induced increase in phospho-Ser845 GluR1 level. **p Ͻ 0.01 compared with vehicle; § § p Ͻ 0.01 compared with papaverine alone; ¶ ¶ p Ͻ 0.01 compared with haloperidol alone; one-way ANOVA followed by Newman-Keuls test. F, Cotreatment of mice with rolipram, as described in C, slightly but significantly potentiates haloperidol-induced increases in phospho-Ser845 GluR1 levels. *p Ͻ 0.05, **p Ͻ 0.01 compared with vehicle; † † p Ͻ 0.01 compared with rolipram alone; ¶ p Ͻ 0.05 compared with haloperidol alone; one-way ANOVA followed by Newman-Keuls test. Error bars indicate SEM. mine by the action of monoamine oxidase (MAO) at dopaminergic terminals and reflects an increased dopamine biosynthesis rate in dopaminergic neurons. Released dopamine can also be converted to HVA at extraneuronal sites, through the sequential metabolism by COMT (catechol-O-methyltransferase) and MAO. Rolipram treatment did not affect the HVA/dopamine ratio, indicating that the primary effect of the inhibitor is an increase in the metabolism of newly synthesized dopamine by MAO at dopaminergic terminals, without an accompanying increase in dopamine release. Together, these data are consistent with results from previous studies in which rolipram was found to increase dopamine synthesis without altering dopamine release (Schoffelmeer et al., 1985;Yamashita et al., 1997a,b). The data extend those observations by demonstrating that phosphorylation of TH at Ser40 is a likely mechanism for mediating the action of rolipram on dopamine synthesis.
The PDE4 isoform responsible for effects at dopaminergic terminals has yet to be characterized. PDE4A and PDE4D mRNA have been detected by in situ hybridization in the substantia nigra (Pérez-Torres et al., 2000). However, a separate immunohistochemical study reported moderate expression of PDE4B and low expression of PDE4D in the same brain region (Cherry and Davis, 1999). Biochemical analysis of TH phosphorylation and dopamine turnover in the striatum from PDE4A, PDE4B, and PDE4D knock-out mice may be required to identity the PDE4 subtype responsible for the regulation of dopamine biosynthesis at dopaminergic terminals observed here.

Role of PDE4 in striatal neurons
PDE4 plays a major role in regulating dopamine synthesis at dopaminergic terminals, and also regulates cAMP/PKA signaling in striatal neurons. The expression of PDE4B at mRNA and protein levels has previously been reported in caudate-putamen (Cherry and Davis, 1999;Pérez-Torres et al., 2000). We used transgenic mice expressing Flag and Myc under the control of dopamine D 1 and D 2 receptor promoters, respectively, to compare the levels of PDE4B expression in the two major, functionally distinct subpopulations of medium spiny neurons by immunohistochemistry. The expression level of PDE4B was higher in striatopallidal neurons than that in striatonigral neurons. Consistent with this observation, we found that PDE4 inhibition selectively potentiated cAMP/PKA signaling in striatopallidal neurons. Rolipram treatment augmented phosphorylation of Thr34 DARPP-32 in response to an adenosine A 2A receptor agonist, but had no effect on phosphorylation mediated by a dopamine D 1 receptor agonist. These results suggest that PDE4 preferentially regulates cAMP/PKA signaling in striatopallidal neurons. Thus, cAMP/PKA signaling in striatopallidal neurons is regulated by at least two PDEs, PDE10A and PDE4, although the impact of PDE4 inhibition is less robust than that of PDE10A inhibition. The increase in cAMP/PKA signaling in striatopallidal neurons elicited by the PDE4 inhibitor, rolipram, would be expected to oppose dopamine D 2 receptor signaling in these cells, similar to the effects of the PDE10A inhibitor, papaverine.
The PDE4 inhibitor, rolipram, like the PDE10A inhibitor, papaverine, inhibits dopamine D 2 receptor signaling consistent  . Differential role of PDE10A and PDE4 in striatal neurons and at dopaminergic terminals. This study provides evidence for differential expression and action of PDE10A and PDE4 in the striatum. PDE10A is expressed in two types of striatal neurons: D 1 receptor-enriched striatonigral and D 2 receptor-enriched striatopallidal neurons. The inhibition of PDE10A by papaverine potentiates the adenosine A 2A receptor-induced increase in DARPP-32 phosphorylation, counteracts the dopamine D 2 receptor-induced decrease in DARPP-32 phosphorylation in striatopallidal neurons, and potentiates the dopamine D 1 receptor-induced increase in DARPP-32 phosphorylation in striatonigral neurons. PDE4 predominantly functions at dopaminergic terminals, and an inhibition of PDE4 by rolipram results in an increase in TH phosphorylation and dopamine synthesis. The inhibition of PDE4 also increases DARPP-32 Thr34 phosphorylation, preferentially in striatopallidal neurons, and potentiates the adenosine A 2A receptor-induced increase in DARPP-32 phosphorylation in these neurons.
with the pharmacological profile of an antipsychotic medication. Interestingly, at the same time, rolipram stimulates dopamine synthesis, indicating that PDE4 inhibition raises dopaminergic tone with mild inhibition of dopamine D 2 receptor signaling in the striatum. This pharmacological profile is similar to that of some atypical antipsychotic medications, and may explain the antipsychotic activity of rolipram observed in animal models of schizophrenia, in which the compound rescued amphetamineinduced reductions in auditory-evoked potentials (Maxwell et al., 2004), MK801 [(ϩ)-5-methyl-10,11-dihydro-5H-dibenzo-[a,d]cyclohepten-5,10-imine maleate]-induced deficits in latent inhibition (Davis and Gould, 2005), and amphetamine-induced deficits in prepulse inhibition (Kanes et al., 2007). Among PDE4 subtypes, PDE4B is a possible target for rolipram, because the inhibition of conditioned avoidance responding by rolipram was attenuated in PDE4B knock-out mice (Siuciak et al., 2007(Siuciak et al., , 2008. Because the nonselective PDE4 inhibitors including rolipram are known to induce the side effect of nausea and vomiting possibly via inhibition of PDE4D (Robichaud et al., 2002), the development of a PDE4B-selective inhibitor may be necessary for clinical trials in the treatment of psychiatric disorders.
In conclusion, the present study employs pharmacological, genetic, and neuroanatomical approaches to dissect the contribution of PDE4 and PDE10A to cAMP/PKA signaling in the striatum. Our studies reveal that PDE10A predominantly regulates cAMP/PKA signaling in medium spiny striatal neurons, as exemplified by the state of phosphorylation of DARPP-32, acting like a dopamine D 2 antagonist in striatopallidal neurons. In contrast, PDE4 predominantly regulates TH phosphorylation at dopaminergic terminals. Thus, PDE4 and PDE10A have distinct roles in striatal dopaminergic neurotransmission conferred by their discrete cellular localization.