Journal of Neurophysiology

Dopamine D-1/D-5 Receptor Activation Is Required for Long-Term Potentiation in the Rat Neostriatum In Vitro

J.N.D. Kerr, J. R. Wickens

Abstract

Dopamine and glutamate are key neurotransmitters involved in learning and memory mechanisms of the brain. These two neurotransmitter systems converge on nerve cells in the neostriatum. Dopamine modulation of activity-dependent plasticity at glutamatergic corticostriatal synapses has been proposed as a cellular mechanism for learning in the neostriatum. The present research investigated the role of specific subtypes of dopamine receptors in long-term potentiation (LTP) in the corticostriatal pathway, using intracellular recording from striatal neurons in a corticostriatal slice preparation. In agreement with previous reports, LTP could be induced reliably under Mg2+-free conditions. This Mg2+-free LTP was blocked by dopamine depletion and by the dopamine D-1/D-5 receptor antagonist SCH 23390 but was not blocked by the dopamine D-2 receptor antagonist remoxipride or the GABAA antagonist picrotoxin. In dopamine-depleted slices, the ability to induce LTP could be restored by bath application of the dopamine D-1/D-5 receptor agonist, SKF 38393. These results show that activation of dopamine D-1/D-5 receptors by either endogenous dopamine or exogenous dopamine agonists is a requirement for the induction of LTP in the corticostriatal pathway. These findings have significance for current understanding of learning and memory mechanisms of the neostriatum and for theoretical understanding of the mechanism of action of drugs used in the treatment of psychotic illnesses and Parkinson's disease.

INTRODUCTION

Activity-dependent synaptic plasticity is a widely used model for learning and memory mechanisms of the brain (Bliss and Collingridge 1993). Although most extensively studied in the hippocampus, activity-dependent synaptic plasticity has also been described in several other brain areas including the neostriatum. The neostriatum is a brain region involved in certain types of learning including reward-related (Beninger 1983) and motor (Graybiel 1995) learning. It receives inputs from all regions of the cerebral cortex (McGeorge and Faull 1989) via an extensive glutamatergic projection (McGeer et al. 1977). The neostriatum also receives a major dopaminergic projection from the substantia nigra, which terminates in close proximity to the corticostriatal inputs (Smith et al. 1994). A number of models have proposed activity-dependent synaptic plasticity in the corticostriatal pathway as a mechanism for learning-related functions of the neostriatum (Beninger 1983; Groves 1983;Miller 1981; Wickens 1990). These models assume that synaptic plasticity in the corticostriatal pathway is regulated by the dopamine inputs from the substantia nigra.

The corticostriatal pathway is of critical importance for the function of the neostriatum. Corticostriatal inputs to the neostriatum synapse directly on the spiny projection neurons (Somogyi et al. 1981), which are the output neurons of the neostriatum (Preston et al. 1980). Thus the corticostriatal synapses are the direct connection between the input and output of the neostriatum. Furthermore the spiny projection neurons are relatively quiescent, and their firing activity occurs in response to excitation by cortical inputs (Wilson and Groves 1981;Wilson et al. 1983). Thus the efficacy of the corticostriatal synapse is a major determinant of the action potential activity of the spiny projection neurons, and plasticity in these synapses is a candidate mechanism for the learning functions of the neostriatum.

Both long-term potentiation (LTP) and long-term depression (LTD) have been described in the corticostriatal pathway. LTD can be induced by high-frequency stimulation (HFS) of the cortical afferents to the neostriatum (Calabresi et al. 1992b; Lovinger et al. 1993; Wickens et al. 1996, 1998), and the requirements for its induction have been extensively characterized (Calabresi et al. 1992a,b, 1994, 1995). In contrast to LTD, neostriatal LTP cannot be induced reliably by HFS in standard solutions. It was first reported after HFS in slices bathed in Mg2+-free fluid (Walsh 1991). This form of LTP was subsequently shown to be blocked byN-methyl-d-aspartate (NMDA) receptor antagonists (Calabresi et al. 1992c), suggesting that the unmasking of LTP in Mg2+-free fluid was due to removal of the voltage-dependent Mg2+ block of the NMDA channels (Nowak et al. 1984).

In addition to occurring in Mg2+-free conditions, striatal LTP can be induced in normal bathing solutions if dopamine is applied in pulses timed to coincide with cortical HFS (Wickens et al. 1996). A similar phenomenon occurs in response to substantia nigra stimulation in the intact animal (Reynolds and Wickens 2000) but not in dopamine-depleted animals, suggesting that LTP can be induced under more physiological conditions and that it is a dopamine-dependent phenomenon. There are also indications that Mg2+-free LTP is blocked by chronic dopamine depletion by 6-hydroxydopamine (Centonze et al. 1999). These findings suggest a possible link between the induction of neostriatal LTP under Mg2+-free conditions and LTP associated with endogenous dopamine release or direct application of dopamine.

The present experiments used intracellular recording techniques to investigate the role of specific dopamine receptor subtypes in corticostriatal LTP. We report a stimulation protocol that reliably induces LTP of the corticostriatal pathway in Mg2+-free conditions. We also show that this form of LTP is blocked reliably by a dopamine D-1/D-5 receptor antagonist but not by a dopamine D-2 receptor antagonist or GABAA receptor antagonist. Finally, we show that LTP is prevented by dopamine depletion but can be restored by a dopamine D-1 receptor agonist.

METHODS

Male Wistar rats (190–240 g) were deeply anesthetized with ether and decapitated. The brain was quickly removed and chilled in ice-cold artificial cerebrospinal fluid (ACSF, see following text). After cooling for 3 min, the brain was removed from solution, the hemispheres were separated, and a block of brain tissue was prepared by sectioning one hemisphere in a horizontal plane 45° to the base of the brain. The block containing the neostriatum and overlying cortex was fixed to the stage of a Campden vibroslice, and 400-μm slices were cut in which the cortex, neostriatum, and corticostriatal connecting fibers were preserved (Arbuthnott et al. 1985; Kawaguchi et al. 1989). Slices were maintained at room temperature before being transferred to a recording chamber in which they were superfused with ACSF containing (in mM) 124 NaCl, 2.5 KCl, 2.0 MgSO4, 2.5 CaCl2, 1.25 NaH2PO4, 26 NaHCO3, and 11 glucose that was gassed with 95% O2-5% CO2 mixture and maintained at a temperature of 35 ± 0.1°C (mean ± SD). During slice preparation, an ACSF solution in which sucrose (248 mM) was substituted for NaCl was used to maximize the yield of good impalements (Aghajanian and Rasmussen 1989). Slices were allowed to equilibrate in the recording chamber for at least 1 h before use.

Intracellular records were obtained from neostriatal cells using glass microelectrodes filled with 2 M potassium acetate solution (95–120 MΩ). For inclusion in the study, cells were required to meet the following criteria: resting membrane potential more negative than −80 mV and stable throughout the recording period (at least 50 min) without use of holding current; action potential overshoot greater than 10 mV; and, action potential onset delayed at least 50 ms from the onset of a just-suprathreshold current pulse. Cells were rejected from the study at the start of the experiment if they did not meet these criteria. These criteria were adopted after extensive preliminary work had shown that less stringent criteria resulted in greater variability in both cellular properties and LTP. A total of 29 cells meeting these inclusion criteria were used in the present study.

Electrophysiological traces were recorded using an Axoprobe 1A intracellular amplifier (Axon Instruments). Traces were digitized at a sampling rate of 10 kHz per channel using pClamp software (Axon Instruments) and saved to hard disk for analysis off-line. Postsynaptic potentials (PSPs) were evoked by stimulation of the deeper layers of the cortex and adjacent white matter (bipolar electrodes, monophasic constant current pulses, 0.1 ms, 150 μA max). Stimulus intensity was adjusted to give initial postsynaptic potential (PSP) amplitudes of 10–15 mV. After a 10-min baseline period in normal solution, perfusion was switched to a Mg2+-free solution. During exchange of Mg2+-free for normal solution, PSPs were recorded for 20 min by which time the PSP amplitude and waveform had stabilized. Thus recordings were made for at least 30 min before HFS.

The cortical stimulus current used during HFS was the same as that used for test pulses and thus was not a suprathreshold stimulus. To ensure a conjunction of both presynaptic activity and action potential firing of the postsynaptic cell, HFS of the cortex (trains of 50 pulses at 100 Hz repeated 6 times at 10-s intervals) was paired with depolarization of the postsynaptic neuron using an intracellular current pulse. Prior to HFS, the depolarizing current pulse for each cell was adjusted to an intensity that ensured action potential firing in response to cortical stimulation (520-ms pulse, 0.2–1.2 nA). Test responses were recorded for at least 20 min following HFS. No intracellular holding current was used to maintain the membrane potential, and cells were discarded if the membrane potential changed during the recording period. Only one cell per slice was used to avoid effects of prior HFS on synaptic plasticity.

Measurements of PSP peak amplitude and slope were made using in-house programs based on Axograph 2.0 software (Axon Instruments). Peak PSP values were measured from the resting membrane potential to the maximum depolarization during the PSP. The PSP slope was the maximum rate of rise, obtained from the maximum value of slope of the moving regression line fitted to eight consecutive sample points (corresponding to 0.8 ms) between onset and peak of the PSP.

Measurements of cellular properties (input resistance and action potential characteristics) were made shortly after impalement and repeated 10 min before and 20 min after HFS. Input resistance was determined from the slope of a regression line fitted to the membrane potentials produced by a series of subthreshold depolarizing current pulses. Threshold for action potential firing was defined as the point on the voltage trajectory at which the rate of depolarization exceeded 8 mV/ms. Action potential amplitude was defined as the difference between threshold and the peak of the action potential waveform. Action potential duration was measured at the voltage midway between threshold and peak potentials. The amplitude of the afterhyperpolarization potential (AHP) was defined as the difference between threshold and the minimum of the hyperpolarization that followed each action potential. Threshold was used as the baseline for AHP measurements rather than resting membrane potential because the equilibrium potential for AHPs is more depolarized than the hyperpolarized resting membrane potential of spiny projection neurons.

Remoxipride (10 μM, Sigma), SCH 23390 (10 μM, RBI), SKF 38393 (5 μM, RBI), and picrotoxin (50 μM, Sigma) were dissolved to their desired final concentration in the Mg2+-free superfusing fluid. Dopamine-depleted slices were prepared from animals injected with alpha-methyl para tyrosine (AMPT, 300 mg/kg ip, RBI) 2.5 h before slice preparation. The administration of AMPT depletes up to 86% of the releasable stores of dopamine (White et al. 1993) by inhibiting tyrosine hydroxylase, an enzyme catalyzing the rate limiting step in the production of dopamine (Cumming et al. 1994). Previous work has shown that the protocol used in the present experiments abolishes dopamine release within 45 min (Williams and Millar 1990). All animals in the AMPT group showed reduced motor activity (consistent with dopamine depletion) prior to slice preparation, and recording was completed within 5 h of the AMPT injection.

Statistical analysis of synaptic plasticity was based on the percentage change in response from baseline values (average of 5 min of test responses prior to HFS). Between-group differences were tested for statistical significance using a one-way ANOVA followed by Student-Newman multiple comparison procedure. Statistical analysis of cellular properties used a two-tailed t-test for independent samples (between group comparison) and a paired t-test (within group comparison of cellular properties 10 min before and 20 min after HFS). The probability level for statistical significance was set at P = 0.05.

RESULTS

Switching of the ACSF to a Mg2+-free solution was associated with an increase in the slope and peak of the test responses. These changes reached a steady state and stabilized within 20 min of changeover from normal to Mg2+-free ACSF. In these slices, cortical HFS paired with depolarization of the postsynaptic cell reliably induced a long-term increase in both the slope and peak of the cortically evoked test responses, which we refer to as Mg2+-free LTP. Averaging across all cells tested (n = 6), there was an increase in the PSP peak amplitude (22.3 ± 3.4%) and slope (15.3 ± 2.9%) measured 20 min after HFS. These findings are illustrated in Fig. 1.

Fig. 1.

Striatal long-term potentiation (LTP) in Mg2+-free solution. ↑, time of high-frequency stimulation (HFS). ▭, perfusion with Mg2+-free solution. - - -, baseline taken as the average of last 5 min responses before HFS. A: measurements of the slope (○) and peak amplitude (□) of postsynaptic responses, showing at each time point the average of 6 consecutive responses evoked at 0.1 Hz. B: response to cortical test stimulus just prior to HFS and 20 min after HFS. i: traces at time points b and c shown overlaid; ii: initial part of traces on expanded timebase so that differences in slope can be seen. Each trace is an average of 6 consecutive responses. C: superimposed responses to a series of current pulses 10 min before HFS (i, time a) and 20 min after HFS (ii,time c). All data from same cell.

To test whether the Mg2+-free LTP was due to potentiation of a reversed inhibitory postsynaptic potential (IPSP), LTP was measured in the presence of picrotoxin, a GABAA receptor antagonist. The group average (n = 3) showed an increase in the PSP peak amplitude (40.8 ± 5.1%) and slope (28.4 ± 5.4%) measured 20 min after HFS that was not significantly different from the LTP seen in the Mg2+-free control group (data not shown).

Subtype-specific dopamine receptor antagonists were used to test whether dopamine receptors play a role in Mg2+-free LTP. The D-1/D-5 antagonist SCH 23390 (10 μM) abolished Mg2+-free LTP (Fig.2). In slices treated with SCH 23390, there was no change in the group average (n = 5) measures of PSP peak amplitude (−1.6 ± 2.2%) or slope (−3.3 ± 1.6%) measured 20 min after HFS.

Fig. 2.

Effects of SCH 23390 on striatal LTP in Mg2+-free solution. Same layout as in Fig. 1.

To determine whether dopamine D-2 receptors play a role in Mg2+-free LTP, the D-2 receptor antagonist remoxipride was also tested (Fig. 3). Averaging across all slices tested (n = 4), remoxipride (10 μM) did not prevent LTP of the PSP peak amplitude (26.9 ± 6.7%) or slope (24.7 ± 3.4%) measured 20 min after HFS.

Fig. 3.

Effects of remoxipride on striatal LTP in Mg2+-free solution. Same layout as in Fig. 1.

Group average data for the control (Mg2+-free), dopamine D-1/D-5 receptor antagonist (SCH 23390) and dopamine D-2 receptor antagonist (remoxipride) are shown in Fig.4. The difference between Mg2+-free control and SCH 23390 groups was significant (P < 0.05). There was no significant difference between Mg2+-free control and remoxipride groups. Although the magnitude of Mg2+-free LTP appeared to be greater in the remoxipride-treated slices, this difference was not significant. Thus dopamine D-1/D-5 receptor activation, but not D-2 receptor activation, was found to be a requirement for neostriatal LTP in Mg2+-free conditions.

Fig. 4.

Dopamine D-1/D-5 receptor antagonists, but not D-2 antagonists, block Mg2+-free LTP in the neostriatum. Percentage change from baseline, taken as the average of last 5 min responses before HFS, group averages across all cells in each treatment group.A: change in postsynaptic potential (PSP) slope.B: change in PSP peak amplitude. □, Mg2+-free controls (n = 6); ▴, Mg2+-free plus remoxipride (n = 4); ●, Mg2+-free plus SCH 23390 (n = 5). *, significant difference from controls (P < 0.05).

The blockade of Mg2+-free LTP by a dopamine D-1/D-5 receptor antagonist suggests that endogenous dopamine may be involved in this form of LTP. The requirement for endogenous dopamine was tested in slices depleted of releasable dopamine by pretreatment with AMPT. Pretreatment of slices with AMPT blocked Mg2+-free LTP (Fig.5, A and C). Averaging across all slices tested (n = 6), there was no change in the PSP peak amplitude (−9.4 ± 10.2%) or slope (−13.2 ± 10.2%) measured 20 min after HFS. The difference between the Mg2+-free control group and the AMPT-treated group was significant (P < 0.05).

Fig. 5.

Mg2+-free LTP is blocked by dopamine depletion and restored by dopamine D-1 receptor agonist. A: alpha-methyl para tyrosine (AMPT) pretreatment blocks Mg2+-free LTP of PSP peak amplitude. B: AMPT pretreatment blocks Mg2+-free LTP of PSP slope. C: after AMPT pretreatment, addition of the D-1 agonist SKF 38393 restores Mg2+-free LTP of PSP peak amplitude. D: SKF 38393 restores Mg2+-free LTP of PSP slope.E: group average of peak amplitude data from AMPT-treated controls and slices treated with SKF 38393.F: slope data for same slices as in E. *, significant difference between groups (P < 0.05).

Application of the dopamine D-1 receptor agonist SKF 38393 restored the ability of AMPT-treated slices to show Mg2+-free LTP (Fig. 5, B and C). There was LTP of PSP peak amplitude (20.5 ± 7.8%) and slope (18.4 ± 6.6%) measured 20 min after HFS (n = 5). The difference between the control (dopamine-depleted) group and the SKF 38393 (dopamine-depleted) group was significant (P < 0.05).

An apparent short-term facilitation of both PSP slope and peak was observed in slices exposed to SCH23390 (Fig. 2). This short-term facilitation was not significantly different in magnitude from the initial potentiation seen in control slices, slices exposed to Remoxipride, or AMPT-treated slices exposed to SKF 38393 (Fig. 4). Short-term facilitation was, however, completely abolished in slices from AMPT-treated animals (Fig. 5).

Changes (or the absence of changes) in postsynaptic response measures (slope and peak of the PSP) could be secondary to changes in cellular properties other than synaptic efficacy. This possibility was tested by measuring selected cellular properties 10 min before and 20 min after HFS. No significant changes were observed in resting membrane potential, input resistance, action potential threshold, action potential amplitude and duration, or AHPs (Figs. 1 C,2 C, and 3 C and Table1).

View this table:
Table 1.

Cellular properties before and after HFS

Between-group differences in the induction of Mg2+-free LTP could also in theory be secondary to differences in cellular properties affecting neuronal excitability. However, there were no significant between group differences in cellular properties as a result of treatment with SCH 23390, remoxipride, AMPT, or AMPT plus SKF 38393 (Table 1). To exclude possible differences in the level of depolarization during HFS, the intensity of the applied current and the responses of the postsynaptic cell to HFS plus depolarization were also compared between groups. There were no significant differences between groups in the intensity of the current pulse used during HFS, the number of action potentials fired, or the level of depolarization produced in the postsynaptic neuron during HFS (data not shown).

DISCUSSION

The main finding of the present study was that dopamine D-1/D-5 receptor activation is a necessary requirement for Mg2+-free LTP in the neostriatum. The present result is in agreement with previous studies showing that LTP could reliably be induced by HFS of the corticostriatal pathway in slices bathed in Mg2+-free solution (Calabresi et al. 1992c; Walsh 1991). In addition, we found that Mg2+-free LTP was blocked by the dopamine D-1/D-5 receptor antagonist SCH 23390 and not blocked by the dopamine D-2-selective antagonist remoxipride. To our knowledge, this is the first study showing that dopamine D-1/D-5 receptor activation is necessary for Mg2+-free LTP in the neostriatum.

The slices treated with SCH 23390 showed an apparent short-term facilitation that was not evident in other groups. This may be an early phase of Mg2+-free LTP that is unmasked by the prevention of the later phase of LTP in the presence of SCH 23390. Previously a role of dopamine D-1/D-5 receptors in the persistence of LTP has been described in the hippocampal CA1 area, where Frey et al. (1991) showed that the presence of SCH 23390 resulted in prevention of late LTP stages (more than 1–2 h). This effect in the hippocampus occurs over a longer time course than that described in the present experiments (hours rather than minutes) but may involve similar cellular mechanisms.

At the doses of SCH 23390 used in the present study (10 μM), it is not possible to rule out an effect on other monoamine receptors such as those for serotonin. However, Mg2+-free LTP was also blocked in slices which were dopamine depleted by pretreatment with AMPT, which does not lead to depletion of serotonin. The only difference between these treatments was the complete abolition of short-term facilitation in the AMPT group, which indirectly suggests a possible link between short-term facilitation and serotonergic effects of SCH 23390; this might warrant further experimental investigation.

It has been reported recently that Mg2+-free LTP was blocked after chronic dopamine depletion with 6-hydroxydopamine (Centonze et al. 1999). Treatment with 6-hydroxydopamine has also been shown to cause loss of synapses in the striatum (Ingham et al. 1998), and this may affect the synapses that are normally potentiated in Mg2+-free LTP. The present results show, however, that in slices acutely depleted of endogenous dopamine by pretreatment with AMPT, HFS in Mg2+-free solutions similarly failed to induce LTP. Together with the previous result, these findings suggest that endogenous dopamine release at the time of HFS is a necessary requirement for Mg2+-free LTP.

Another important finding of the present study is that after acute dopamine depletion with AMPT, Mg2+-free LTP is restored to control levels by bath application of the dopamine D-1 receptor agonist SKF 38393. We are not aware of any previous reports showing that dopamine D-1 receptor activation restores LTP in dopamine-depleted slices. Together with the effects of dopamine depletion and dopamine D-1/D-5 receptor antagonists, the effects of the D-1 receptor agonist strongly implicate dopamine D-1/D-5 receptors in Mg2+-free LTP. The neostriatal dopamine D-1/D-5 receptors are G-protein-coupled receptors that activate adenylyl cyclase leading to intracellular accumulation of cyclic AMP (cAMP). Thus Mg2+-free LTP probably involves elevation of cAMP and triggering of related biochemical cascades present in neostriatal neurons. Elevation of cAMP by direct activation of adenylyl cyclase (Colwell and Levine 1995) or dopamine D-1/D-5 receptors (Price et al. 1999) causes enhancement of responses to glutamatergic agonists. In contrast, the dopamine D-2 receptor is negatively coupled to adenylyl cyclase. We found that the dopamine D-2 receptor antagonist, remoxipride, did not block Mg2+-free LTP. These results are compatible with dopamine D-1/D-5 receptor mediated activation of adenylyl cyclase being necessary for the enhancement of synaptic transmission in Mg2+-free LTP.

The magnitude of Mg2+-free LTP was not reduced in the presence of the GABAA antagonist, picrotoxin. This finding is important because previous studies have not ruled out the possibility that Mg2+-free LTP included a component of LTP of feed-forward inhibitory IPSPs (Calabresi et al. 1992c; Walsh 1991). At the hyperpolarized resting membrane potentials typical of spiny neostriatal neurons, IPSPs may be depolarizing because the equilibrium potential of GABAA-activated conductances is more depolarized (Misgeld et al. 1982). This made it necessary to test if the changes in cortically evoked test responses were due to changes in feedforward IPSPs. The finding that Mg2+-free LTP was not blocked or reduced in magnitude by the GABAA antagonist picrotoxin shows that Mg2+-free LTP is not due to changes in GABAA mediated IPSPs. It implies that the LTP is due to changes in excitatory postsynaptic potentials.

In the present study, LTP could reliably be induced by HFS in Mg2+-free solution in all cells tested. The reliability of induction of this form of LTP is consistent with a previous study in which longer HFS trains were applied.Calabresi et al. (1992c) described LTP in Mg2+-free solution after 900 suprathreshold stimulus pulses (applied as 3 trains of 300 pulses at 100 Hz). In contrast, Walsh (1993) found that a somewhat milder induction protocol involving 400 subthreshold stimulus pulses (applied as 4 trains of 100 pulses at 100 Hz) did not reliably induce LTP but induced short-term potentiation lasting from 5 to 45 min. In the present study, LTP lasting for as long as recordings were continued (at least 20 min and up to 40 min) was elicited by HFS consisting of a total of 300 subthreshold stimulus pulses (applied as 6 trains of 50 pulses at 100 Hz) in conjunction with suprathreshold depolarization of the postsynaptic neuron. These results show that Mg2+-free LTP is a robust phenomenon that is readily reproduced in different laboratories despite marked differences in HFS protocols.

Previous work has shown that the induction of Mg2+-free LTP in the neostriatum can be blocked by NMDA receptor antagonists (Calabresi et al. 1992c). In normal bathing solutions, NMDA receptor-operated channels play a minor role in corticostriatal synaptic transmission (Kita 1996). This is because at hyperpolarized membrane potentials such channels are closed by a magnesium block (Nowak et al. 1984) and, as in the present and previous studies (Jiang and North 1991), the resting membrane potential of the neostriatal neurons is strongly hyperpolarized. Thus it is necessary to consider whether the facilitation of LTP in Mg2+-free is due to the greater influx of calcium postsynaptically in Mg2+-free solution. In Mg2+-free conditions, HFS of cortical afferents paired with suprathreshold current injection as used in the present study can be expected to produce influx of calcium via NMDA channels. This failed to produce LTP in slices that had been dopamine depleted by pretreatment with AMPT or slices in which dopamine D-1/D-5 receptors were blocked by a selective antagonist. Thus the present results show that activation of NMDA channels per se is not sufficient for induction of Mg2+-free LTP.

Calabresi et al. (1997) have shown abnormal induction of LTP in slices made from mice lacking dopamine D-2 receptors. In slices made from dopamine D-2 receptor knockout mice, HFS of the corticostriatal pathway under conditions that would normally induce LTD produced LTP. In contrast to the present results showing blockade of Mg2+-free LTP in wild-type rats, the abnormal form of LTP seen in the D-2 receptor-deficient mice was not blocked by SCH 23390. A potentially important factor in these differences is that D-2 receptor-deficient mice have abnormal dopamine function throughout life. Dopaminergic receptor mechanisms can be disrupted by abnormal dopamine receptor stimulation during development. For example, rats given 6-OHDA lesions as neonates show a substantial sub-sensitivity to both D-2 and D-1 antagonists as adults (Duncan et al. 1987). Thus it is plausible that in the D-2 receptor-deficient mice, the dose of SCH 23390 used by Calabresi et al. (1997) may not produce effective D-1 blockage in dopamine receptor knockout animals.

Another possible cause of the difference in response to SCH 23390 in the present experiments and in the D-2 receptor-deficient mice (Calabresi et al. 1997) is that the neurons in the present study were relatively polarized (−95 mV in comparison to −85 mV in the D-2 receptor-deficient mice). No holding current was used in the present experiments, and the resting membrane potential reflects the selection criteria used, the quality of the impalement, the ionic composition of the extracellular fluid, and the effects of the ionic composition of the intracellular electrode solution (Nisenbaum and Wilson 1995). Furthermore during the plasticity-inducing HFS, depolarizing current was applied to ensure suprathreshold depolarization of the recorded cells. This minimizes any potential interaction between resting membrane potential and the production of LTP. In addition, because there was no significant difference in resting membrane potential between the groups in the present experiments (Table 1), the polarized resting membrane potential is unlikely to have influenced the results.

While showing that activation of NMDA channels is not sufficient for LTP induction, the present results do not exclude a complex modulatory effect of dopamine D-1/D-5 receptor activation on NMDA receptor-mediated channels, which may also be necessary for LTP induction. Previous work has shown that dopamine application leads to an enhancement of NMDA receptor-mediated responses that is apparently mediated by dopamine D-1/D-5 receptors (Cepeda et al. 1992,1993; Levine et al. 1996). Such enhancement of NMDA receptor-mediated channels may be necessary for LTP to occur in Mg-free solutions, suggesting that both dopamine D-1/D-5 receptor activation and NMDA receptor-mediated channel activation may be necessary. Futher experiments are needed to investigate this possibility.

An alternative explanation for the facilitation of striatal LTP in Mg2+-free conditions is that these conditions favor dopamine release; an effect that itself is mediated by activation of presynaptic NMDA receptors (Desce et al. 1992;Krebs et al. 1991a,b; Roberts and Sharif 1978) presumably located on dopaminergic nerve terminals. The dopaminergic terminals on spiny projection neurons synapse in close proximity to the glutamatergic corticostriatal terminals (Freund et al. 1984; Hersch et al. 1995; Smith et al. 1994; Yung et al. 1995). Thus glutamate released from corticostriatal terminals during cortical HFS might act directly on adjacent dopaminergic terminals to cause dopamine release. Spillover of glutamate under Mg2+-free conditions favors NMDA-mediated release of endogenous dopamine. A model in which Mg2+-free conditions favor LTP by facilitating endogenous dopamine release is compatible with the present findings as well as previous results showing that in the absence of exogenous dopamine NMDA receptor activation is necessary for Mg2+-free LTP (Calabresi et al. 1992c) and that pulsatile application of exogenous dopamine facilitates LTP under physiological conditions (Wickens et al. 1996). To confirm this possibility, further experiments are needed to measure endogenous dopamine release in response to HFS in Mg2+-free and control conditions.

In summary, the present study has confirmed previous reports that Mg2+-free LTP is a reliable and robust phenomenon in the neostriatum. It has extended understanding of this phenomenon by showing that Mg2+-free LTP is dopamine dependent and requires activation of dopamine D-1/D-5 receptors but not dopamine D-2 receptors. Furthermore it has shown that LTP can be restored in dopamine-depleted slices by the application of a dopamine D-1/D-5 agonist. These results are significant for current understanding of learning and memory mechanisms of the neostriatum and are compatible with behavioral evidence that D-1/D-5 receptors are important in reward-related learning (Sutton and Beninger 1999;Wickens and Kotter 1995). The results are also significant for theoretical understanding of the mechanism of the therapeutic effect of dopamine receptor antagonist drugs used in the treatment of psychotic illnesses (Miller et al. 1990) and dopamine receptor agonists in the treatment of Parkinson's disease (Rascol et al. 1999).

Acknowledgments

We thank Prof. W. C. Abraham and Dr. D. Plenz for helpful comments on the manuscript.

This work was supported by the New Zealand Neurological Foundation, the New Zealand Health Research Council, the New Zealand Schizophrenia Fellowship, the HS and JC Anderson Trust, and the New Zealand Lottery Grants Board.

Present address of J.N.D. Kerr: Unit of Neural Networks Physiology, LSN/NIMH, National Institutes of Health, Bethesda, MD 20892-4075.

Footnotes

  • Address for reprint requests: J. R. Wickens, Dept. of Anatomy and Structural Biology, School of Medical Sciences, University of Otago, PO Box 913, Dunedin, New Zealand (E-mail:jeff.wickens{at}stonebow.otago.ac.nz).

REFERENCES

View Abstract