Journal of Neurophysiology

Dopamine Enhancement of NMDA Currents in Dissociated Medium-Sized Striatal Neurons: Role of D1 Receptors and DARPP-32

Jorge Flores-Hernández, Carlos Cepeda, Elizabeth Hernández-Echeagaray, Christopher R. Calvert, Eve S. Jokel, Allen A. Fienberg, Paul Greengard, Michael S. Levine

Abstract

Dopamine (DA), via activation of D1 receptors, enhancesN-methyl-d-aspartate (NMDA)-evoked responses in striatal neurons. The present investigation examined further the properties of this enhancement and the potential mechanisms by which this enhancement might be effected. Dissociated medium-sized striatal neurons were obtained from intact rats and mice or mutant mice lacking the DA and cyclic adenosine 3′,5′ monophosphate (cAMP)-regulated phosphoprotein of M R 32,000 (DARPP-32). NMDA (10–1,000 μM) induced inward currents in all neurons. In acutely dissociated neurons from intact rats or mice, activation of D1 receptors with the selective agonist, SKF 81297, produced a dose-dependent enhancement of NMDA currents. This enhancement was reduced by the selective D1 receptor antagonist SKF 83566. Quinpirole, a D2 receptor agonist alone, produced small reductions of NMDA currents. However, it consistently and significantly reduced the enhancement of NMDA currents by D1 agonists. In dissociated striatal neurons, in conditions that minimized the contributions of voltage-gated Ca2+ conductances, the D1-induced potentiation was not altered by blockade of L-type voltage-gated Ca2+ conductances in contrast to results in slices. The DARPP-32 signaling pathway has an important role in D1 modulation of NMDA currents. In mice lacking DARPP-32, the enhancement was significantly reduced. Furthermore, okadaic acid, a protein phosphatase 1 (PP-1) inhibitor, increased D1-induced potentiation, suggesting that constitutively active PP-1 attenuates D1-induced potentiation. Finally, activation of D1 receptors produced differential effects on NMDA and gamma aminobutyric acid (GABA)-induced currents in the same cells, enhancing NMDA currents and inhibiting GABA currents. Thus simultaneous activation of D1, NMDA, and GABA receptors could predispose medium-sized spiny neurons toward excitation. Taken together, the present findings indicate that the unique potentiation of NMDA receptor function by activation of the D1 receptor signaling cascade can be controlled by multiple mechanisms and has major influences on neuronal function.

INTRODUCTION

Dopamine (DA) modulates responses evoked by activation of glutamate receptors in medium-sized spiny striatal neurons (Cepeda et al. 1993; Levine et al. 1996a,b; Price et al. 1999; Yan et al. 1999). Our laboratory has shown in striatal slices that the direction of this modulation depends on both the glutamate receptor subtype subject to DA modulation and the DA receptor subtype preferentially activated (Cepeda and Levine 1998;Cepeda et al. 1993; Levine et al. 1996a,b). Activation of D1 family receptors enhances responses, in particular those mediated byN-methyl-d-aspartate (NMDA) receptors, whereas activation of D2 family receptors reduces responses, in particular those mediated by non-NMDA receptors (Cepeda et al. 1993; Levine et al. 1996b).

We have been studying the mechanisms involved in the enhancement of NMDA responses by D1 receptors. One possibility is that DA modulation of voltage-gated Ca2+ conductances contributes to D1 enhancement of NMDA receptor-mediated responses because previous studies have shown that activation of D1 receptors enhances L-type Ca2+ currents (Hernandez-Lopez et al. 1997; Surmeier et al. 1995). We provided support for this alternative by demonstrating that blockade of L-type Ca2+ channels in striatal slices markedly reduced the D1-mediated potentiation of responses due to NMDA receptors (Cepeda et al. 1998a). However, blockade of L-type Ca2+ channels did not completely abolish the enhancement, suggesting that other mechanisms contribute to the modulation (Cepeda et al. 1998a).

Another mechanism by which DA could modulate NMDA receptors is by changing the state of phosphorylation of the NMDA receptor. DA alters cyclic adenosine 3′,5′ monophosphate (cAMP) production. Activation of D1 receptors increases formation of cAMP and stimulates protein kinase A (PKA). Stimulation of the cAMP-PKA cascade with forskolin enhances NMDA-evoked responses (Blank et al. 1997; Colwell and Levine 1995). Furthermore, D1 receptor activation and forskolin increase NMDAR1 subunit phosphorylation through a synergistic mechanism involving increased phosphorylation and decreased dephosphorylation of the receptor subunit (Rajadhyaksha et al. 1998; Snyder et al. 1998). The DA and cAMP-regulated phosphoprotein of M R32,000 (DARPP-32) (Ouimet et al. 1984; Walaas and Greengard 1984) is partly responsible for these effects (Fienberg et al. 1998). DARPP-32 is phosphorylated by cAMP-dependent PKA on a single threonine residue, Thr34, and transformed into a potent inhibitor of protein phosphatase-1 (PP-1) (Hemmings et al. 1984). In turn, PP-1 regulates the phosphorylation state of many neurotransmitter receptors and voltage-gated ion channels (Greengard et al. 1999). Recent evidence supports a role for PP-1 in regulating α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptor channels in medium-sized striatal neurons (Yan et al. 1999).

The present study was designed to further examine the properties of, and the mechanisms underlying D1 modulation of NMDA currents. We used dissociated neurons because adequate voltage clamp can more easily be obtained than from neurons in slices. In our previous study using striatal slices (Cepeda et al. 1998a), interpretation of the findings was complicated by space-clamp limitations in cells with extended processes. In addition, we examined the contribution of DARPP-32 and PP-1 to the modulation.

METHODS

Animals

All procedures were carried out in accordance with the National Institutes of Health's Guide for Care and Use of Laboratory Animals and were approved by the Institutional Animal Care and Use Committee at the University of California at Los Angeles. Experiments used Sprague-Dawley rats (21–40 days old; n = 78), DARPP-32 knockout [148 ± 6 (SE) days of age, 137–166 age range,n = 6] and wild-type (WT) mice of the same strain as the DARPP-32 knockouts (145 ± 5 days of age,134–164 age range,n = 8). The mice were generated at the Rockefeller University and shipped to UCLA. The generation and characterization of these mice have been described (Fienberg and Greengard 2000; Fienberg et al. 1998).

Dissociation procedure

Striatal neurons from rats or DARPP-32 knockout and WT mice were acutely dissociated using procedures slightly modified from those previously described (Bargas et al. 1994;Flores-Hernández et al. 2000; Surmeier et al. 1995). Briefly, rats and mice were anesthetized with Halothane, perfused transcardially, and decapitated. Brains were quickly removed and placed in ice-cold, low-Ca2+sucrose solution that contained (in mM) 250 sucrose, 2.5 KCl, 0.1 CaCl2, 1 Na2HPO4, 4 MgSO4, 11 glucose, and 15N-[2-hydroxyethyl]piperazine-N-[2-ethanesulfonic acid] (HEPES), pH = 7.4, 300–305 mOsm/l. Brain blocks were cut to obtain 400-μm coronal slices using a DSK microslicer (Osaka, Japan) while bathed in the same solution. Slices were then incubated for 1–6 h at room temperature (20–22°C) in a NaHCO3 buffered Earle's balanced salts solution (EBSS) bubbled with 95% O2-5% CO2 and supplemented with (in mM) 1 pyruvic acid, 0.005 glutathione, 0.1 NG-nitro-l-arginine, and 1 kynurenic acid, pH = 7.4 with NaOH, 300–305 mOsm/l. After ≥1 h incubation, a slice was placed in low-Ca2+ isethionate solution containing (in mM) 140 Na isethionate, 2 KCl, 2 MgCl2, 0.1 CaCl2, 23 glucose, and 15 HEPES, pH = 7.4, 300–305 mOsm/l, bubbled with O2 and supplemented as described in the preceding text. With the aid of a microscope, the dorsal striatum was dissected. Dissections were limited to the tissue rostral to the anterior commissure to reduce the possibility of contamination from the globus pallidus.

The tissue was placed in an oxygenated cell-stir chamber (Wheaton, Millville, NJ) containing Papain (1–2 mg/ml, Calbiochem) in a HEPES-buffered Hank's balanced salt solution (HBSS, Sigma Chemical) at 35°C bubbled with O2 and supplemented as before, pH = 7.4 with NaOH, 300–305 mOsm/l. After 20–40 min of enzyme digestion, tissue was rinsed three times with the low-Ca2+-isethionate solution and mechanically dissociated with a graded series of fire-polished Pasteur pipettes. The cell suspension was then plated into a 35-mm Lux petri dish mounted on the stage of an upright fixed-stage microscope (Zeiss Axioskop, Thornwood, NY) containing HEPES-buffered HBSS saline.

Whole cell recordings

Whole cell recordings employed standard techniques (Bargas et al. 1994). Electrodes were pulled from Corning 8250 glass (A-M Systems, Carlsborg, WA) and heat polished prior to use. The internal solution consisted of (in mM) 175N-methyl-d-glucamine (NMDG), 40 HEPES, 2 MgCl2, 10 ethylene glycol-bis(β-aminoethyl ether)-N, N,N′,N′-tetraacetic acid (EGTA), 12 phosphocreatine, 2 Na2ATP, 0.2 Na3GTP, and 0.1 leupeptin, pH = 7.2–7.3 with H2SO4, 265–270 mOsm/l. The external solution consisted of (in mM) 127 NaCl, 20 CsCl, 5 BaCl2, 2 CaCl2 12 glucose, 10 HEPES, 0.001 tetrodotoxin, and 0.02 glycine, pH = 7.3 with NaOH, 300–305 mOsm/l. Barium was used as the main charge carrier through Ca2+ channels and to better block K+ conductances. Addition of 2 mM Ca2+ to the external solution helped preserve the physiological internal environment and also to favor processes dependent on activation of second messengers (Greengard et al. 1999). For example, previous studies have shown that NMDA receptor channel function is regulated by endogenous Ca2+-dependent phosphatases (Lieberman and Mody 1994). One potential concern is the anomalous mole-fraction effect, one of the properties of multi-ion channels. For example, for Ca2+ channels, when Ba2+ and Ca2+ are combined, the current produced is lower than with either cation alone (Hille 2001). This effect would further reduce the contribution of Ca2+ conductances to D1-induced modulation. To determine if this combination of ions affected NMDA responses, we tested solutions with Ba2+ alone versus the combination of Ba2+ and Ca2+ on currents evoked by NMDA and found no difference in the evoked NMDA current.

Recordings were obtained with an Axon Instruments 200A patch-clamp amplifier and controlled with a Pentium clone running pClamp (v. 8.01) with a DigiData 1200 series interface (Axon Instruments, Foster City, CA). Electrode resistance was typically 2–4 MΩ in the bath. After seal rupture, series resistance (<20 MΩ) was compensated (70–90%) and periodically monitored. Recordings were made from medium-sized cells that had short (<75 μm) proximal dendrites (Fig.1). Unless otherwise noted, the cell membrane potential was held at −40 mV. This holding potential was used because it reduced activation of Ca2+ channels and allowed examination in the same cell of both NMDA and GABA currents.

Fig. 1.

Examples of medium-sized neurons, with patch electrodes attached, acutely dissociated from a rat (A) and a DARPP-32 knockout mouse (B).

Drug application

Drugs were applied with a gravity-fed “three-pipe” system. The array of application capillaries (ca. 150 μm ID) was positioned a few hundred micrometers from the cell under study. Solution changes were performed by changing the position of the array with a DC drive system controlled by a SF-77B perfusion system (Warner Instruments, Hamden, CT) synchronized by pClamp. The solution changes were complete within <100 ms. Generally, NMDA (1–1,000 μM) was applied for 3 s every 15–20 s.

Drugs were prepared as concentrated stocks in deoxygenated water. DA receptor ligands used were SKF 81297 (D1 agonist), SKF 83566 (D1 antagonist), and quinpirole (D2 agonist; RBI, Natick, MA). DA receptor agonist solutions were protected from ambient light. Other drugs used were nifedipine, an L-type Ca2+ channel antagonist (Sigma), the NMDA receptor antagonist 2-amino-phosphonovalerate (AP5, Tocris, Baldwin, MO), Ro 32–0432 [a cell permeable protein kinase C (PKC) inhibitor], okadaic acid (PP-1/2A inhibitor), and norokadaone (an inactive form of okadaic acid). The latter three drugs were obtained from Calbiochem (La Jolla, CA).

Statistical analyses

Values in the figures and text are presented as means ± SE. Dose-response curves were fit with a sigmoid function (Origin, Microcal Software, Northampton, MA). Differences among group means were assessed with appropriate t-tests or ANOVAs for independent or repeated measures. Welch's approximation to the t-test for unequal variances was used when group variances were not homogeneous (Welch 1947). For post hoc evaluations using ANOVAs, the Bonferroni t-test was used because this test is one of the more conservative approaches using multiple comparisons.

RESULTS

Experiments were performed only on medium-sized neurons (Fig. 1). Acutely dissociated medium-sized neurons in both rats and mice were easy to distinguish from large-sized putative cholinergic cells based on somatic cross-sectional size and membrane capacitance. In most cells in rats and mice, the initial dendritic segments were preserved and in some cases secondary branches showed spines. No attempt was made to further identify and separate the medium-sized cells into different subpopulations of projection cells or projection cells and interneurons. Because the medium-sized interneurons represent only a small percentage of cells in the striatum, we assumed that most recordings were obtained from projection neurons.

NMDA currents in striatal neurons from rats

Application of NMDA for 3 s induced inward currents in all neurons tested (Fig. 2 A). As pointed out in the preceding text, neurons were held at −40 mV to inactivate most Ca2+ conductances and Mg2+ was removed from the bath to maximize the response to NMDA. To establish a concentration-response relationship, NMDA was applied at multiple concentrations in 13 neurons (1, 10, 50, 100, 200, and 1000 μM; not all neurons were tested with each concentration; Fig. 2 A). Small, non- or slowly desensitizing currents began to be observed at 10 μM and increased in amplitude in a concentration-dependent manner. As concentration increased, two types of responses occurred. Neurons displayed a fast, desensitizing component which appeared at 50 μM and at higher concentrations of NMDA. Other neurons displayed an initial peak but they did not desensitize as rapidly. Because both types of responses were modulated approximately equally by application of D1 agonists, data were pooled. Concentration-response functions were generated for peak NMDA currents by normalizing the peak amplitude to the saturating concentration (1,000 μM). There was a statistically significant increase in peak current with concentration (F = 24.2, df = 5/50,P < 0.001). The resulting best fit sigmoidal concentration-response function for the peak current had an EC50 of 28 μM (Fig. 2 B1) and a Hill coefficient of 1.23. Normalized peak current density was computed by dividing by cell capacitance (Alzheimer et al. 1993) and was best fit by a sigmoidal function with comparable EC50s and Hill coefficients (Fig.2 B2). Because current and current density functions for responses induced by NMDA were similar and there was no significant variation in neuron capacitance across different experimental groups within rats or mice, we will present only data for evoked currents. As is typical, NMDA currents were blocked by AP5 (Fig. 2 A2). At 10 μM AP5, the peak current induced by 100 μM NMDA was reduced by 38 ± 8% (n = 5), at 50 μM AP5 the reduction was 72 ± 12% (n = 5), and at 100 μM AP5 it was 81 ± 11% (n = 7; F = 14.8, df = 3/14, P < 0.01).

Fig. 2.

N-methyl-d-aspartate (NMDA) dose-response relationships in rat medium-sized striatal neurons. A1: a representative set of whole-cell currents evoked by 6 different concentrations of NMDA in the same cell. For these and other traces, NMDA was applied at the horizontal line, and the voltage was held at −40 mV in Mg2+-free solution. A2: whole- cell inward currents evoked by NMDA (100 μM) reduced by the receptor antagonist 2-amino-phosphonovalerate (AP5, 50 μM). B, 1 and 2: peak currents and peak current densities with best fit functions plotted. Concentrations on the abscissa are log coordinates.

Activation of D1 receptors enhances NMDA currents

Based on the concentration-response functions we chose 100 μM of NMDA as the concentration to examine modulation by DA receptor agonists because it was above the EC50 for the peak currents and produced consistent responses in each neuron examined. To establish a concentration-response relationship, the selective D1 receptor agonist SKF 81297 was applied at multiple concentrations in 11 neurons (1, 10, and 100 nM and 1 and 10 μM; not all neurons were tested with each concentration; Fig.3 A). Application of SKF 81297 significantly enhanced peak NMDA currents in a concentration-dependent fashion (F = 9.91, df = 4/22, P < 0.001; Fig. 3 B). The enhancement had an EC50 of 188 nM and a Hill coefficient of 0.59. In addition, there was a concentration-dependent increase in the decay time constant (F = 3.63, df = 5/26, P < 0.025), indicating that the decay of the peak current was significantly slower in the presence of the D1 agonist. The largest increases in decay time constants were 36 ± 10 and 34 ± 12% at 1- and 10-μM concentrations of SKF 81297, respectively. Because the greatest modulation occurred for D1 agonist concentrations of 1 and 10 μM, alterations in D1 modulation in subsequent experiments used one or both of these concentrations to examine the properties of the modulation.

Fig. 3.

NMDA currents are potentiated by application of the full D1 dopamine (DA) receptor agonist SKF 81297 in a concentration-dependent manner in neurons from rats. A: examples of potentiation of NMDA (100 μM) current by four concentrations (10 and 100 nM and 1 and 10 μM) of SKF 81297 in a single neuron. Each pair of responses represent the control response and then the response in the presence of the D1 agonist at each concentration (←). Data are not shown inA for the 1 nM concentration shown graphically inB. B: concentration-response functions of the mean percent modulation for peak currents plotted with a best-fit function. Concentrations on the abscissa are log coordinates.

To provide evidence for specificity, in a separate group of neurons, the effects of SKF 81297 (1 μM) were examined in the presence of the selective D1 receptor antagonist SKF 83566 (1 μM; n = 15 cells; Fig. 4 A). Alone at this concentration, SKF 83566 had little net effect on the NMDA current (−4.4 ± 2.3% average decrease with both increases and decreases in peak NMDA current occurring). Higher concentrations of the D1 antagonist alone produced more consistent decreases in peak NMDA current and were thus not used. The magnitude of the D1-induced enhancement was reduced from 22 ± 2.3 (range of +42.5 to +14% increase in the presence of SKF 81297 alone; all cells displayed increases) to 15 ± 2.8% by the antagonist (all but 1 cell showed decreases in agonist-induced potentiation in the presence of the antagonist; t = 2.81, df = 14, P< 0.02; Fig. 4 B).

Fig. 4.

A: example of antagonism of SKF 81297-induced enhancement of NMDA current by the specific D1 receptor antagonist SKF 83566 (1 μM) in a neuron from a rat. Left: the response to NMDA (100 μM). Middle: the enhancement by application of SKF 81297. Right: attenuation of the enhancement in the presence of SKF 83566. - - -, helps visualize the changes. B: bar graphs showing mean percent modulation of the peak NMDA current by SKF81297 alone (left bar) and reduction of mean percent SKF 81297 modulation by SKF 83566 (right bar; ± SE). Asterisk indicates that the decrease is statistically significant (see text for statistics).C: example of decreases of SKF 81297-induced enhancement of NMDA current by the specific D2 receptor agonist quinpirole (10 μM) in a neuron from a rat. Left: the response to NMDA (100 μM). Middle: the increase produced by SKF 81297 (1 μM). Right: quinpirole reduces the enhancement produced by SKF 81297 in the same neuron. - - -, helps visualize the changes. D: bar graphs showing mean percent modulation of the peak NMDA current (±SE) by SKF 81297 alone (left bar) and the decrease in the mean percent modulation (±SE) produced by quinpirole (right bar). Asterisk indicates the decrease is statistically significant (see text for statistics). E: blockade of L type Ca2+currents does not reduce the modulation of NMDA currents by SKF 81297 in neurons from rats. Left: response to NMDA and its potentiation in the presence of SKF 81297 (10 μM).Right: nifedipine alone (10 μM) produces little change in the NMDA current and does not alter the increase produced by SKF 81297. F: bar graphs showing that percent modulations of the NMDA current in the presence or absence of nifedipine were similar.

Activation of D2 receptors reduced the enhancement of NMDA currents produced by D1 receptor activation

We have shown previously that activation of D2 DA receptors can decrease responses mediated by activation of NMDA receptors using current-clamp recordings in striatal slices (Cepeda et al. 1993; Levine and Cepeda 1998). However, these effects could have been mediated by pre- and/or postsynaptic actions of D2 receptors as we have previously shown a presynaptic action of this receptor subtype using D2 receptor-deficient mice (Cepeda et al. 2001). To determine if activation of D2 receptors alters NMDA currents, we applied the D2 agonist quinpirole (10 μM) in a separate group of neurons (n = 11). Alone, quinpirole produced slight and variable changes in mean peak amplitude of NMDA current (average −4 ± 2% reduction with both increases and decreases in peak NMDA current occurring). We then applied quinpirole in the presence of SKF 81297 to determine if it affected the enhancement of NMDA currents (Fig. 4 C). In this group of neurons, SKF 81297 (1 μM) produced a 25 ± 4% enhancement of the NMDA response (t = 3.38, df = 10, P < 0.01). When quinpirole was then applied in the presence of the D1 agonist, the D1-induced enhancement of the mean peak NMDA current was reduced by ∼50% to 12 ± 2% (all cells showed a reduction in the agonist-induced potentiation; t = 5.35, df = 10, P < 0.001; Fig. 4 D).

Role of L-type Ca2+ conductances

In medium-sized spiny neurons and in cortical pyramidal cells L-type Ca2+ conductances play a role in the enhancement of NMDA currents by activation of D1 receptors (Cepeda et al. 1998a; Wang and O'Donnell 2001). Previously, we demonstrated that in the striatal slice blockade of L-type Ca2+ conductances markedly reduced D1-mediated potentiation of NMDA currents, suggesting that Ca2+ conductances in dendrites contributed to the effects of the D1 agonist (Cepeda et al. 1998a). Here we demonstrate in dissociated striatal neurons, in conditions that minimize the contribution of voltage-gated Ca2+conductances [most of the dendrites have been removed, holding potential was −40 mV, which inactivates L-type Ca2+ conductances (Hille 2001) and better space clamp can be obtained], blockade of L-type Ca2+ conductances had almost no effect on D1-induced modulation. Application of SKF 81297 (10 μM,n = 7) (Fig. 4 E, left) produced a 42.7 ± 13.6% increase in the peak NMDA current (Fig.4 F, left bar graph). Application of the L-type Ca2+ channel antagonist nifedipine (10 μM) alone produced inconsistent effects on NMDA currents (−6.7 ± 6.9% reduction in mean peak current with both increases and decreases in peak NMDA current occurring; Fig. 4 E, right). When SKF 81297 (10 μM) was applied in the presence of nifedipine the modulation induced by the D1 agonist also was not affected (47.6 ± 15.5%; Fig. 4 F, right bar graph). Thus the increase in NMDA current induced by activation of NMDA receptors is not dependent on activation of L-type Ca2+ currents in acutely isolated striatal neurons under the present conditions indicating that other mechanisms contribute to the modulation.

Modulation of NMDA currents was reduced in DARPP-32 knockout mice

To provide information about the mechanism underlying D1 potentiation of NMDA currents, we examined D1 modulation in DARPP-32 knockout mice. Although these mice were older than the rats, our goal was to examine mechanisms underlying the modulation that are more easily done in mice in which genetic alterations can be taken advantage of. Thus we did not directly compare data obtained from mice and rats in this study because of the age difference. The role of DARPP-32 in the modulation of striatal neuronal excitability has been well documented (Fienberg and Greengard 2000; Fienberg et al. 1998). In DARPP-32 knockout mice, the inhibition of PP-1 by DARPP-32 should be absent, and we would predict that D1 potentiation of NMDA-induced responses would be reduced. NMDA currents were examined in 25 neurons from WT and 23 neurons from DARPP-32 knockout mice (Fig.5, A and B). There was a slight reduction in the mean peak amplitude of the NMDA current in neurons obtained from DARPP-32 knock out compared with WT mice (442 ± 51 vs. 370 ± 47 pA, respectively) that was not statistically significant, however. We examined the enhancement of NMDA currents by the D1 agonist at two concentrations, 1 and 10 μM (n = 12 cells in WTs; n = 14 cells in knockouts). In WT mice, the increases in NMDA current produced by application of SKF81297 were 20.3 ± 2.7 and 32.4 ± 4.6% at 1 and 10 μM, respectively (Fig. 5, A, top, andB). In DARPP-32 knockout mice, the corresponding increases were 15.2 ± 1.7 and 16.9 ± 3.2%, respectively (Fig.5 A, bottom, and B). At both concentrations, the enhancement of the NMDA current was greater in WT than in DARPP-32 knockout mice (t = 5.69, df = 11,P < 0.01; t = 3.66, df = 13,P < 0.01 for 1 and 10 μM concentrations of the D1 agonist, respectively; Fig. 5 B). Thus D1-induced potentiation is significantly reduced in the DARPP-32 knockout mice but is not absent, suggesting that the DARPP-32 protein is part of a series of mechanisms that contributes to D1 potentiation of NMDA currents.

Fig. 5.

A, top: the current response to NMDA alone and its enhancement by SKF 81297 (1 μM) in a wild-type (WT) mouse.Bottom: the current response to NMDA and the smaller enhancement produced by SKF 81297 in a DARPP-32 knockout mouse.B: bar graphs showing mean percent modulation (±SE) produced by 1 and 10 μM SKF 81297 in WT and DARPP-32 knockout mice. Both decreases in modulation were statistically significant in the DARPP-32 knockout mice (asterisks; see text for statistics).C: okadaic acid (OA), a protein phosphatase 1 (PP-1) inhibitor, enhances D1 modulation of NMDA currents in neurons from WT and DARPP-32 knockout mice. Left: trace from a WT mouse shows the response to NMDA (100 μM). Middle: trace shows the increase produced by SKF 81297 (1 μM).Right: trace shows that OA (100 nM) increased the enhancement produced by SKF 81297 in the same neuron. Dashed horizontal line helps visualize the changes. D: bar graphs showing mean percent modulation of the peak NMDA current (±SE) by SKF 81297 alone and in the presence of OA in WT mice (left 2 bars) and in the DARPP-32 knock out mice (right 2 bars). Asterisks indicate the increases produced by OA were statistically significant (see text for statistics).

To further examine the role of PP-1 in the enhancement of NMDA current, we applied the PP-1 inhibitor okadaic acid (Greengard et al. 1999) at 100 nM, a concentration that blocks PP-1 (Caporaso et al. 2000; Ishihara et al. 1989). Alone, okadaic acid produced a small but inconsistent increase in the NMDA current in WT mice (3.4 ± 2.8%,n = 4). In the presence of the D1 agonist (1 μM), okadaic acid significantly increased the D1-induced potentiation from 20.3 ± 2.7 to 29.9 ± 3.2% in the WT mice (t = 6.31, df = 4, P < 0.01; Fig. 5,C and D). In DARPP-32 knockout mice, alone okadaic acid also produced a small inconsistent change in NMDA (7.0 ± 1.9%, n = 8). In the presence of the D1 agonist (1 μM), okadaic acid significantly increased the potentiation from 15.2 ± 1.7 to 28.3 ± 2.9% (t = 6.25, df = 7, P < 0.001; Fig. 5 D).

We next tested the same concentration of okadaic acid in striatal neurons (n = 6) obtained from rats. Alone okadaic acid produced inconsistent effects on the peak NMDA current (−4.3 ± 1.1%). Application of 1 μM SKF 81297 induced a 14.8 ± 1.7% increase in the NMDA (100 μM) current. In the presence of the D1 agonist, okadaic acid significantly increased the potentiation to 22.2 ± 2.7% (t = 3.43, df = 5,P < 0.05). Application of the inactive form of okadaic acid, norokadaone (100 nM) produced no change in D1-induced potentiation [14.5 ± 2.8 vs. 11.7 ± 3.2% (n = 3) before and during application of norokadaone, respectively].

A recent report demonstrated a functional role for PKC in D1 modulation of NMDA-induced responses in nucleus accumbens neurons (Chergui and Lacey 1999). Ro 32–0432, a selective PKC inhibitor, blocked D1-induced potentiation of NMDA-induced responses. To determine if PKC contributed to D1-induced potentiation of NMDA responses in striatal neurons, we examined the effects of Ro 32–0432 in the dissociated cells. In the presence of Ro 32–0432 (5 μM) alone, NMDA-induced peak currents were essentially unchanged (−3.6 ± 3.0%, n = 8). SKF 81297 (1 μM) induced a 16.6 ± 2.2% (n = 11) increase in the peak response. When Ro 32–0432 was combined with SKF 81297, the potentiation increased (23.9 ± 4.3%, n = 11). Therefore in dissociated striatal neurons antagonism of PKC does not appear to block D1-induced potentiation.

D1 agonists produced differential effects on NMDA and GABA currents in the same neurons

A previous study (Flores-Hernández et al. 2000) demonstrated that activation of D1 receptors reduces GABA currents in medium-sized striatal neurons. To determine if GABA and NMDA currents are modulated in the same neurons by activation of D1 receptors, we examined the effects of the D1 agonist on both types of currents. GABA (100 μM, 3 s) was applied similarly to NMDA at a holding potential of −40 mV. The concentration of GABA chosen was the same as used in a previous study (Flores-Hernández et al. 2000). As demonstrated in that study, GABA induced a rapid peak outward current followed by a steady-state outward current (Fig.6 A, top). The mean amplitude of the peak current was 484 ± 50 pA (n= 13). When SKF 81297 (10 μM) was applied, the peak current was significantly reduced by 29.7 ± 3.8% (t = 4.29, df = 12, P < 0.001; Fig. 6 B,left bar). In the same neurons, when SKF 81297 (10 μM) was applied, NMDA (100 μM) currents were significantly enhanced by 24.3 ± 5.8% (t = 6.96, df = 12,P < 0.001; Fig. 6 A, bottom, andB, right bar). To determine if there was a relationship between D1 modulation of GABA and NMDA currents, the percent of modulations for each amino acid were correlated. There was a low negative Pearson product-moment correlation coefficient (R = −0.285), suggesting a moderate relationship between the ability of D1 receptor activation to modulate both NMDA and GABA currents. This effect was not statistically significant, however. Thus a net effect of activation of D1 receptors at this holding potential is to enhance excitation not only by increasing NMDA current but also by reducing GABA-mediated currents.

Fig. 6.

A, top: currents activated by GABA (100 μM) alone (left), in the presence of SKF 81297 (10 μM;middle), and GABA (100 μM) alone again (wash;right). Bottom: currents activated by NMDA (100 μM) alone (left), in the presence of SKF 81297 (10 μM; middle), and NMDA (100 μM) alone again (Wash; right) in the same neuron from a rat. Note that SKF 81297 attenuates GABA-induced currents while it enhances NMDA-induced currents. B: bar graphs showing that application of SKF 81297 inhibits GABA currents while concomitantly enhancing NMDA currents in the same group of neurons. Asterisks indicate the decreases and increases in modulation were statistically significant (see text for statistics).

DISCUSSION

A number of interesting findings have emerged from these experiments. D1 receptor activation enhances NMDA receptor-mediated currents in dissociated medium-sized spiny neurons in a concentration-dependent manner. This effect is reduced by application of a D1 receptor antagonist. We have shown previously that the D1-induced potentiation of NMDA responses is also blocked in D1 receptor-deficient mice (Levine et al. 1996a). While D2 receptor activation had little direct effect on postsynaptic NMDA currents in the present experiments, it reduced the D1-induced potentiation, indicating that D1 receptors needed to be activated before D2 receptor effects would be observed. In this preparation, the D1-induced potentiation is independent of L-type voltage-gated Ca2+ conductances, as in our experimental conditions, these channels were inactivated at membrane potentials of −40 mV and additional blockade of L-type channels did not significantly alter the enhancement. The present findings demonstrate that the DARPP-32 signaling pathway also has an important role in the modulation of NMDA currents by D1 receptors. In DARPP-32 knockout mice the enhancement was significantly reduced. Furthermore, okadaic acid, a PP-1/2A inhibitor, increases D1-induced potentiation suggesting that constitutively active PP-1/2A attenuated D1-induced potentiation. Thus phosphorylation of NMDA receptor subunits by activation of the D1 signaling cascade is a potential mechanism contributing to the modulation demonstrated in these experiments (Greengard et al. 1999). Finally, activation of D1 receptors produces differential effects on NMDA and GABA currents in the same cells, enhancing NMDA currents and inhibiting GABA currents.

Virtually all neurons examined displayed potentiation in response to application of D1 agonists. This is at variance with the hypothesis that D1 and D2 receptors are segregated onto different populations of medium-sized spiny striatal neurons (Gerfen 1992). This hypothesis remains controversial (Surmeier et al. 1993). For example, a previous study using single-cell reverse transcriptase coupled with PCR demonstrated that almost all striatal neurons displayed subtypes of both D1 and D2 family receptors, either the D1 or D5 receptor and one or more of the D2, D3, or D4 subtypes (Surmeier et al. 1996). More recently, additional evidence for co-localization of D1 and D2 receptors on medium-sized striatal spiny neurons has been provided (Aizman et al. 2000). However, it should be pointed out that our studies, at least in rats, used younger animals, and there may be maturational differences in DA receptor segregation that occur after early developmental periods. Based on the evidence cited in the preceding text, it is not particularly surprising that virtually all neurons displayed D1-induced potentiation and that it could be reduced by a D2 family agonist. What did vary from neuron-to-neuron was the magnitude of the potentiation or its attenuation, a variable that could reflect the density of D1 or D2 family receptors expressed by specific neurons or the degree to which dendritic processes remained after the dissociation. Although there was a clear dose-response relationship within individual neurons (see Fig. 3), the degree of modulation at the higher concentrations of the D1 agonist varied.

Multiple mechanisms for D1 receptor modulation of NMDA receptors

Multiple mechanisms are capable of mediating the D1 agonist-induced potentiation of NMDA-evoked responses. Elimination of each mechanism alone decreases the potentiation but does not eliminate it, suggesting these mechanisms operate independently but synergistically.

One of the mechanisms involves the D1/cAMP/PKA cascade. Forskolin, an activator of this cascade enhances NMDA responses (Blank et al. 1997; Colwell and Levine 1995). Furthermore, D1 receptor activation increases NR1 subunit phosphorylation via the DARPP-32 signaling cascade (Fienberg et al. 1998;Rajadhyaksha et al. 1998; Snyder et al. 1998). DARPP-32 when phosphorylated by PKA acts as a potent inhibitor of PP-1. Mutant mice that lack DARPP-32 have a significant reduction in D1-induced NMDA receptor modulation. Although potentiation still exists, increasing the concentration of the D1 agonist does not increase the potentiation as much in the knock out as it does in the WT. Okadaic acid, a PP-1/2A inhibitor further increased potentiation providing additional evidence for the importance of this cascade. The finding that okadaic acid restored D1 modulation in the DARPP-32 mutant suggests the system remains intact with little compensatory change in the mutant mice. We also observed that NMDA currents were slightly decreased in the DARPP-32 knockout mice. While this effect was not statistically significant, possibly DARPP-32 plays a small constitutively active role in maintaining the levels of excitability of NMDA receptors and/or in their state of phosphorylation. In summary, our evidence indicates that increased NMDA receptor phosphorylation (probably of the NR1 subunit) mediated by activation of PKA and inhibition of PP-1/2A can contribute to the increase in membrane current (Greengard et al. 1999).

In the current study, we have presented data from both rats and mice, albeit of different ages. We purposely did not make direct comparisons because of the age differences. Although we used younger rats, our developmental studies indicate that by ∼3 wk of age, responses to glutamate receptor agonists and dopaminergic modulation are relatively mature (Cepeda et al. 1998a,b; Colwell et al. 1998; Hurst et al. 2001). However, in both species, D1-induced potentiation was observed and at least the effects of okadaic acid were similar, suggesting that comparable mechanisms may underlie the potentiation.

Another mechanism contributing to D1 potentiation of NMDA responses involves L-type Ca2+ conductances. Blockade of L-type Ca2+ conductances reduces the enhancement of NMDA currents in striatal slices (Cepeda et al. 1998a). There is good evidence that D1 receptor activation enhances L-type Ca2+currents in medium-sized striatal neurons possibly via direct phosphorylation of the channel by PKA and inhibition of PP-1 (Surmeier et al. 1995), causing increased intracellular Ca2+. Although we could not rule out space-clamp limitations in our previous study in the slice (Cepeda et al. 1998a), in the present experiments, we attempted to minimize the contribution of voltage-gated conductances to determine if the enhancement occurred independent of their modulation. In dissociated cells using a holding potential of −40 mV, which would further decrease the influences of L-type conductances, the addition of the L-type Ca2+ channel antagonist nifedipine did not alter D1-induced potentiation. These findings indicate that the role of L-type Ca2+ channels was minimized in dissociated cells. In addition, most of the dendrites were removed by the dissociation procedure. L-type Ca2+ conductances have a more important role when dendrites are present (Cepeda et al. 1998a). In fact, L-type Ca2+ channels are present on dendritic spines and interact with NMDA receptors (Carlin et al. 2000; Segal 1995).

A recent study demonstrated in cells from nucleus accumbens slices that PKC inhibition blocked the enhancement of NMDA currents by D1 agonists (Chergui and Lacey 1999). However, in the present study, we demonstrate that this effect does not occur in dissociated striatal neurons, indicating that signaling mechanisms may differ depending on neuronal location or the type of preparation.

Other mechanisms that can contribute to the enhancement of NMDA responses by D1 agonists include modulation of ionic pumps and regulation of intracellular Ca2+ stores (Bertorello et al. 1990; Mahan et al. 1990) as we have reviewed previously (Cepeda and Levine 1998). A recent report provided evidence for a unique D1-NMDA receptor interaction in the control of membrane trafficking (Dunah and Standaert 2001). D1 receptor activation caused rapid movement of NMDA receptor subunits between intracellular and postsynaptic sites. Conversely, activation of NMDA receptors recruits functional D1 receptors from the interior of the cell to the plasma membrane, in particular dendritic spines, without affecting the distribution of D2 receptors, and this effect is abolished by incubation of cells in Ca2+-free medium (Scott et al. 2002). Finally, there is recent evidence that there is direct protein-protein coupling between DA receptors and subunits of GABAA and NMDA receptors (Liu et al. 2000a,b). Overall, these multiple mechanisms permit a high degree of synergy and redundancy in signaling through activation of D1 and NMDA receptors.

In dissociated cells in the present study, we found that D2 receptor activation significantly reduced the enhancement of NMDA currents produced by the D1 agonist. This effect cannot be easily explained by the actions of the D2 agonist alone. The D2 agonist produced a 4% average reduction alone but a 13% reduction in the presence of the D1 agonist. Although it is possible that the dissociation procedure disrupts aspects of D2 receptor function, an enabling effect of D1 activation on D2 function has been reported previously for nucleus accumbens neurons (Wachtel and White 1995; White 1987). Thus it appears that D1 receptors may need to be activated before postsynaptic D2-mediated effects become apparent. In the cells that co-localize D1 and D2 family receptors (Surmeier et al. 1996), the antagonism of D1 effects by D2 receptors may provide a protective mechanism against excessive excitation induced when NMDA and D1 receptors are simultaneously activated (Cepeda et al. 1998b).

The potentiation in the NMDA current caused by the D1 agonist was reduced by co-application of the D1 antagonist SKF 83566, suggesting that the potentiation was specific to activation of D1 receptors. The decrease in average potentiation could not be accounted for by the effects of the D1 antagonist alone at the 1 μM concentration. Previous studies have shown that D1 receptor-mediated effects, at least on GABA currents, can be blocked by D1 antagonists in both dissociated and cultured striatal neurons (Flores-Hernández et al. 2000), although in acutely isolated neurons, there is evidence that D1 receptor activation reduces specific K+conductances via allosteric regulation of the channels rather than via coupling to adenylyl cyclase (Nisenbaum et al. 1998).

DA modulation of responses mediated by activation of glutamate and GABA receptors

There is considerable evidence for DA modulation of glutamate- and GABA-evoked responses in striatum, nucleus accumbens, and cerebral cortex (Blank et al. 1997; Cepeda et al. 1992,1993, 1998a; Chergui and Lacey 1999;Colwell and Levine 1995; Flores-Hernández et al. 2000; Galarraga et al. 1997; Gonon and Sundstrom 1996; Harvey and Lacey 1997;Hernandez-Lopez et al. 1997; Hsu et al. 1995; Levine et al. 1996a,b; Price et al. 1999; Seamans et al. 2001; Wang and O'Donnell 2001; Yan et al. 1999; Zheng et al. 1999; but see Calabresi et al. 1995;Nicola and Malenka 1998). The differential effects of D1 receptor activation on NMDA and GABA responses in the same neuron represent an important mechanism for amplification of synaptic input. Previous studies have demonstrated that activation of GABA receptors substantially reduces responses mediated by glutamate receptor activation by a combination of pre- and postsynaptic mechanisms (Calabresi et al. 1991). If DA is released simultaneously and D1 receptors are activated, there will be a simultaneous attenuation of the GABAergic response and a potentiation of the glutamatergic response, effectively amplifying corticostriatal inputs. Thus simultaneous activation of D1 and NMDA receptors would predispose medium-sized spiny neurons toward a more depolarized membrane potential, especially when their membrane potential is in the “up-state” (Wilson and Kawaguchi 1996). If GABA receptors are activated in this condition, the inhibitory response would be attenuated by D1 effects, either prolonging the enhancement or further potentiating excitatory inputs. Interestingly, under these conditions, if DA also activates D2 receptors, the amplification will be reduced.

Functional implications

The findings indicate that D1 receptor activation enhances NMDA currents, in part through a signaling cascade involving DARPP-32 and PP-1. Taken together with our previous evidence that Ca2+ currents can contribute to the modulation of NMDA responses, these results show that multiple pathways appear to be involved in the modulation of NMDA responses by D1 receptors. There are important implications of the D1-NMDA receptor interactions. At a morphological level, the enhancement of NMDA responses by activation of D1 receptors could effectively stabilize and reinforce corticostriatal synaptic connections via interactions in spines. For example, PP-1 regulated by DARPP-32 is particularly abundant in the postsynaptic density fraction (Shields et al. 1985; Sim et al. 1994) and is enriched in dendritic spines (Ouimet et al. 1995), where most excitatory inputs converge. PP-1 interacts with different targeting proteins such as spinophillin and neurabin and may control spine morphology, dynamics, and synaptic efficacy (Harris and Kater 1994). As pointed out in the preceding text, there is now evidence that D1 receptor activation alters trafficking of NMDA subunits (Dunah and Standaert 2001) and NMDA receptor activation recruits D1 receptors into the membrane (Scott et al. 2002). At the functional level, D1-NMDA receptor interactions appear to be involved in striatal synaptic plasticity. They support the induction of long-term potentiation (Charpier and Deniau 1997; Kerr and Wickens 2001), and such potentiation is absent in DARPP-32-deficient mice (Calabresi et al. 2000). In conclusion, it is clear that the unique potentiation of NMDA receptor function by activation of the D1 receptor signaling cascade has major influences on neuronal function.

Acknowledgments

The authors acknowledge D. Crandall and C. Gray for preparation of the illustrations.

This work was supported by National Institutes of Health Grants NS-33538 and NS-35649 to M. S. Levine and MH-40899 and DA-10044 to P. Greengard, by National Association Research in Schizophrenia and Depression to M. S. Levine, and by Office of Naval Research Grant N00149810436 to M. S. Levine.

Footnotes

  • Address for reprint requests: M. S. Levine, Mental Retardation Research Ctr., 760 Westwood Plaza NPI58-258, University of California, Los Angeles, CA 90095 (E-mail:mlevine{at}mednet.ucla.edu).

REFERENCES

View Abstract