INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The striatum is the main input nucleus of the basal ganglia, a group of subcortical structures critically involved in motor and cognitive function. Spiny efferent neurons comprise approximately 95% of all neurons in striatum. These neurons receive excitatory afferents from cerebral cortex and thalamus (Kemp and Powell 1971; Sadikot et al. 1992), which activate postsynaptic N-methyl-D-aspartate receptors (NMDARs) and non-NMDARs (Herrling et al. 1983).

NMDARs consist of an NR1 subunit and any of four NR2 subunits (NR2A-D) (Dingledine et al. 1999). Whereas the NR1 subunit is expressed ubiquitously throughout the brain, expression of NR2 subunits is under spatial and temporal regulation (Monyer et al. 1994). Although all spiny efferent neurons express NR2A and NR2B subunits (Standaert et al. 1999), the relative expression of NR2A subunit mRNA is greater laterally (Ganguly and Keefe 2001; Watanabe et al. 1993). Despite the unknown stoichiometry of native NMDARs, the differential expression of the NR2A subunit mRNA in medial and lateral striatum suggests that the subunit composition of NMDARs will vary between these regions.

Incorporation of different ratios of NR2 subunits into postsynaptic receptors across striatum is likely to be important for information processing, as many functional properties of NMDARs are governed by NR2 subunit composition. For example, NMDARs containing the NR2A subunit demonstrate the highest affinity for competitive antagonists (Buller et al. 1994). Receptors containing the NR2B subunit, meanwhile, have greater affinity for agonists, including glycine (Buller et al. 1994; Kutsuwada et al. 1992; Stern et al. 1992). Consistent with these properties, receptor binding studies conducted on striatal tissue show greater binding of competitive antagonists in lateral striatum, a region rich in NR2A mRNA expression. In comparison, agonist binding is uniform throughout striatum, paralleling the NR2B receptor mRNA distribution (Buller et al. 1994; Monaghan et al. 1988). Thus there appear to be distinct NMDAR populations in striatum.

While the data reviewed above suggest the presence of distinct NMDARs in different regions of striatum, it is not currently known whether the properties of the postsynaptic receptors vary. Although deactivation kinetics of NMDARs were previously characterized in striatum (Götz et al. 1997), this study did not report the striatal subregion from which the recordings were conducted. Furthermore, the study by Götz and colleagues used the nucleated patch and rapid agonist exchange techniques, and therefore surveyed primarily extrasynaptic receptors at a single, early developmental time point. Developmental studies and experiments in expression systems show that the decay-time kinetics of NMDAR-mediated excitatory postsynaptic currents (NMDAR-EPSCs) and deactivation kinetics, respectively, become faster as the ratio of NR2A:NR2B increases (Flint et al. 1997; Kirson and Yaari 1996; Vicini et al. 1998). Furthermore, it has been proposed that NMDAR-EPSCs which have faster decay-time kinetics shift the probability and degree of induction of long-term potentiation (LTP) (Philpot et al. 2001; Tang et al. 1999). However, whether such differences are evident in the same neuronal type of the adult brain expressing different ratios of NR2A:NR2B is unknown.

The purpose of this study was therefore to test whether the biophysical and pharmacological properties of NMDAR-EPSCs would differ based on location of the recorded neuron within the striatum and the developmental time point at which the slice was prepared. We report that the decay-time kinetics of the NMDAR-EPSCs were significantly faster in dorsolateral striatum in slices obtained from adult animals. In addition, application of either glycine or D-serine selectively enhanced the peak amplitude of NMDAR-EPSCs in dorsolateral striatum. These data indicate that there are functionally distinct synaptic NMDARs in ventromedial versus dorsolateral striatum in the adult rodent brain. Furthermore, these results suggest that processing of excitatory input through NMDARs is different in the ventromedial and dorsolateral striatum, two regions that differ in the processing of limbic and sensorimotor information.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

Male Sprague-Dawley rats (approximately 150 g) were used in adult experiments. P4-6 rats were used in all neonatal experiments. Rats were housed in groups in a room controlled for temperature and lighting. Rats had free access to food and water. All animal care and experimental manipulations were approved by the Institutional Animal Care and Use Committee of the University of Utah and were in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.

Striatal slice preparation

Acute brain slices were obtained as previously described, with slight modifications (Wilcox et al. 1996). Briefly, rats were anesthetized with pentobarbital (50 mg/kg) and decapitated. The brains were rapidly removed and placed into an ice cold, oxygenated (95% O₂-5% CO₂) sucrose Ringer solution, pH = 7.4, containing (in mM) 200 sucrose, 3 KCl, 1.4 NaH₂PO₄, 2 MgSO₄, 26 NaHCO₃, 10 glucose, and 2 CaCl₂. The brain was divided along the midline. The brain hemisphere was then glued caudal-side down to a Vibraslicer chuck (Campden Instruments). Coronal sections (300-350 µm) containing the striatum were collected and placed in a holding chamber at room temperature containing oxygenated Ringer solution with 126 mM NaCl in place of the sucrose, pH = 7.37-7.41. The osmolality of the Ringer solution was 295-305 mOsm. Sections remained in the Ringer solution for 1 h before recording.

Patch-clamp recordings

Slices were transferred into the recording chamber, which was perfused with fresh, oxygenated, Mg²⁺-free Ringer solution at room temperature (21-23°C) via a gravity-feed system at approximately 4 ml/min. The whole cell patch-clamp technique was used to record from single striatal neurons. Neurons were patch-clamped using the blind technique (Blanton et al. 1989). Microelectrodes with a resistance of 3-6 M were pulled using a P-87 micropipette puller (Sutter Instruments). For the majority of the experiments, the internal solution in the borosilicate glass microelectrodes consisted of (in mM) 130 K gluconate, 10 KCl, 10 HEPES, 1 EGTA, and 0.1 CaCl₂. The pH was 7.28-7.31 and the osmolality was 285-295 mOsm. In a subset of experiments, CsCl (140 mM) was substituted for K gluconate and KCl, and QX-314 (10 mM) was added to either internal solution.

EPSCs were elicited using local, minimal stimulation to mitigate voltage- and space-clamp errors (Stevens and Wang 1994; Wilcox et al. 1996). Briefly, a bipolar stimulating electrode was placed in the vicinity of the recording electrode (<300 µm) either in the dorsolateral or the ventromedial striatum. The stimulating electrode was used to deliver current pulses (100 µs) of sufficient amplitude to produce the smallest EPSC (25-40 pA), which could be reliably evoked at low frequency (0.1 Hz). To isolate NMDAR-EPSCs, the Ringer solution also contained 10 µM 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) to block the non-NMDA component of glutamate neurotransmission, 50 µM picrotoxin to block activation of GABA_A receptors, and 10 µM glycine to maximally activate NMDARs. For experiments under nominal glycine conditions, no glycine was included in the Ringer solution and 10 µM 6-nitro-7-sulfamoylbenzo(f)-quinoxaline-2,3-dione (NBQX) was substituted for CNQX.

Data were acquired with an Axopatch 1D amplifier and the CLAMPEX8 software package interfaced to a Digidata 1200 acquisition board (Axon Instruments). Signals were filtered at 5 kHz and sampled at 10 kHz.

Inclusion criteria

Only recordings that did not exhibit substantial changes in either holding current or resistance at the electrode tip were used for analysis. All cells required <100 pA to be clamped to 70 mV. Cells with resting membrane potentials above 55 mV were omitted from data analysis. For analysis in current-clamp mode, the membrane potential was monitored in response to current injection. In all cases, this measurement was taken near the middle of the current injection. In depolarizing current injections, during which an action potential was generated, the measurement was taken immediately prior to the action potential. A total of 153 cells meeting these inclusion criteria were used in the present study.

Data analysis

The following parameters were determined for averaged EPSCs in all experiments: risetimes, peak amplitudes, decay time constants, and weighted  (_w). All data are presented as mean ± SE. Data were tested for significance using a Student's t-test. Statistical significance was set at P  0.05. To evaluate differences in peak amplitude, statistics were done on peak amplitudes expressed as percent of control using a two-way ANOVA. For ifenprodil studies, control values were defined as the peak amplitude of the NMDAR-EPSC prior to administration of ifenprodil. For experiments where the concentration of glycine or D-serine was manipulated, control values were defined as the peak amplitude of the NMDAR-EPSC in the presence of no added glycine. The decay time constants were fit with a double exponential equation: I(t) = I_f × exp(t/_f) + I_s × exp(t/_s), where I_f is the amplitude of the fast component, I_s is the amplitude of the slow component, and _f and _s are the fast and slow time constants, respectively. Weighted time constants were calculated using the equation: _w = [I_f/(I_f + I_s)] × _f + [(I_s/(I_f + I_s)] × _s (Stocca and Vicini 1998).

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Basic properties of striatal neurons and isolation of synaptic NMDAR-EPSCs

In recordings from cells in both the ventromedial and dorsolateral striatum, basic cell membrane properties were very similar between regions and consistent with previous reports for spiny efferent neurons (Tepper et al. 1998). Prolonged injection of current produced the signature depolarization ramp (Fig. 1A). Additionally, the change in membrane potential as a function of current injection was nearly indistinguishable between the two regions of striatum (Fig. 1B) and showed decreased input resistance in response to hyperpolarizing current injection (Tepper et al. 1998). Isolated EPSCs were confirmed to be mediated by NMDARs, as shown by the reversible block of the EPSCs in all cells tested by bath application of the NMDAR antagonist D,L-2-amino-5-phosphonovaleric acid (D,L-APV; 50 µM; peak amplitude was 3.1 ± 2.8% of control; n = 6; Fig. 1C).

View larger version (17K): [in this window] [in a new window]   Fig. 1. Basic membrane properties of striatal neurons are similar in ventromedial and dorsolateral striatum. A: representative traces from current-clamp recordings from ventromedial (left) and dorsolateral (right) striatal neurons following injection of current of varying amplitudes. Membrane potential was adjusted to 70 mV for both cells. Resting membrane potential was 64 and 62 mV, respectively. Bottom: waveforms for current injection, which started at 0.25 nA and increased by 0.05 nA until +0.10 nA. B: summary of changes in membrane potential as a function of current injection for all cells tested in ventromedial and dorsolateral striatum. Data for ventromedial cells are mean values ± SE (n = 14-21 cells) and are represented by . Data for dorsolateral cells are mean values ± SE (n = 9-15) and are represented by . Error bars that are not visible are smaller than the symbol. C: reversible blockade of NMDAR-EPSCs by the NMDAR antagonist D,L-2-amino-5-phosphonovaleric acid (D,L-APV; 50 µM; averages of 35 traces). Local minimal stimulation of striatum in close proximity (<300 µm) to recorded cells elicits a long-lasting EPSC. On bath application of D,L-APV, there was a block of the EPSC, confirming that it was mediated by the NMDAR. Recordings were conducted at a membrane potential of 70 mV in the presence of 50 µM picrotoxin, 10 µM CNQX, and 10 µM glycine in Mg2+-free solution.

Kinetics of NMDAR-EPSCs in ventromedial and dorsolateral adult striatum

Previous studies examining different timepoints in postnatal development have shown that NMDARs comprised of different subunits give rise to NMDAR-EPSCs with different decay-time kinetics (Flint et al. 1997; Kirson and Yaari 1996). Additionally, heterologous expression systems expressing NMDARs comprised of different subunits exhibit different deactivation kinetics following rapid application of agonist (Chen et al. 1999; Vicini et al. 1998). However, differences in NMDAR decay-time kinetics have not previously been demonstrated in a single neuronal type in a given structure in the adult CNS in which different levels of NR2 subunits are expressed. Therefore to test the hypothesis that the differential expression of the NR2A subunit in medial versus lateral striatum would lead to differences in the decay-time kinetics of NMDAR-EPSCs in these two regions, we used local, minimal stimulation to assess the decay-time kinetics of NMDAR-EPSCs in neurons in ventromedial and dorsolateral striatum (Figs. 2 and 3). As shown in Fig. 2A, neurons recorded in ventromedial striatum had evoked NMDAR-EPSCs that had significantly longer decay-time constants when compared with neurons recorded in the dorsolateral striatum. A double exponential was required to fit the decay of the NMDAR-EPSCs in both regions of the striatum. The average values of the fast and slow components of the decay-time constant (_f and _s, respectively) of the NMDAR-EPSCs were significantly faster in recordings from dorsolateral striatum (Table 1). Furthermore, the relative fractional contribution of the fast component of the decay (%Fast) was significantly greater in EPSCs from the dorsolateral striatum than from the ventromedial striatum (Table 1). Taken together, these values produced an average _w of the decay-time constant that was significantly faster for NMDAR-EPSCs in dorsolateral cells (Fig. 3; Table 1). Additionally, the risetime kinetics of NMDAR-EPSCs of neurons in the dorsolateral striatum were significantly faster than those in the ventromedial striatum (Fig. 2C; Table 1). When QX-314 (10 mM) was included in the internal pipette solution to block voltage-activated Na⁺-channels and postsynaptic GABA_B receptor-mediated effects, the differences in _w of the decay-time constant of NMDAR-EPSCs in ventromedial versus dorsolateral striatum were still observed (_w =373 ± 41 ms vs. 173 ± 26 ms; P = 0.002; n = 4 and 6, respectively) (Nathan et al. 1990). Furthermore, using a CsCl internal solution with QX-314, we voltage clamped the membrane potential of the neurons in both regions to +40 mV and observed that the profound difference in decay time kinetics of NMDAR-EPSCs in each region persisted (_w =315 ± 3 ms vs. 237 ± 26 ms in ventromedial vs. dorsolateral striatum, respectively; P = 0.04; n = 3, Fig. 2B).

View larger version (26K): [in this window] [in a new window]   Fig. 2. Decay kinetics and risetimes of synaptic NMDAR-EPSCs from neurons in dorsolateral striatum are substantially faster than those from neurons in ventromedial striatum of adult rats. A: local minimal stimulation of striatum in close proximity (<300 µm) to the recorded cell elicits a long-lasting NMDAR-EPSC in ventromedial and dorsolateral striatum. The average of 35 EPSCs evoked at 0.1 Hz (Vh= 70 mV) can be adequately described by a double exponential. The EPSCs in ventromedial and dorsolateral striatum are superimposed and normalized to the peak amplitude for comparison. Note the prolonged decay time of EPSCs evoked in neurons recorded in the ventromedial region of the striatum. B: in a separate set of similar experiments, CsCl and QX-314 were included in the internal solution. Under these conditions, EPSCs were evoked via local, minimal stimulation at 0.1 Hz at a holding potential of +40 mV. As was the case in A, the decay time of EPSCs evoked in both the dorsolateral and ventromedial striatum can be fit with double exponentials. When averaged EPSCs are superimposed and normalized to the peak amplitude, it is readily apparent that the decay time kinetics for EPSCs recorded in ventromedial neurons are significantly prolonged. C: expanding the timescale of the NMDAR-EPSCs in A reveals different 10-90% risetimes in ventromedial vs. dorsolateral striatum. The risetimes from ventromedial and dorsolateral are normalized for comparison. In the normalized traces in A-C, NMDAR-EPSCs from ventromedial and dorsolateral neurons are labeled for clarity. Currents were recorded at a membrane potential of 70 mV in A and C and +40 mV in B in the presence of 50 µM picrotoxin, 10 µM CNQX, and 10 µM glycine in Mg2+-free solution.

View larger version (9K): [in this window] [in a new window]   Fig. 3. Average decay-time kinetics of NMDAR-EPSCs from cells in ventromedial striatum are significantly longer than those of NMDAR-EPSCs from cells in dorsolateral striatum. Summary scatterplot of w (weighted decay time constant) of synaptic NMDAR-EPSCs of neurons in ventromedial and dorsolateral striatum. Bars indicate average values. , values from individual ventromedial cells. , values from individual dorsolateral cells. *Significantly different from w for ventromedial cells, P < 0.0001, unpaired t-test.

Kinetics of NMDAR-EPSCs in ventromedial and dorsolateral neurons in neonatal striatum

Expression of NR2 subunits is developmentally regulated (Monyer et al. 1994). Whereas the NR2B subunit is expressed in the striatum at birth, levels of NR2A subunit protein are not detectable until P7 (Sheng et al. 1994; Wang et al. 1995). To confirm the role of different NR2 subunit combinations in the biophysical properties of NMDAR-EPSCs in ventromedial versus dorsolateral striatum, we conducted whole cell patch-clamp experiments under the same conditions (low frequency, local, minimal stimulation) in slices from neonatal rats. Unlike adult striatum (Fig. 2), recordings from neurons in ventromedial and dorsolateral striatum from P4-6 rat pups revealed no significant differences in _f, _s, %Fast, and _w decay-time kinetics, as well as rise time kinetics (Fig. 4; Table 1). However, the rise times (Table 1) and decay-time kinetics are significantly slower in both regions of neonatal striatum than they are in the corresponding regions of adult striatum, suggesting incorporation of NR2A subunit into NMDARs throughout adult striatum (_w = 688 ± 53 vs. 488 ± 22 ms in ventromedial; P < 0.001; 583 ± 41 vs. 231 ± 15 ms in dorsolateral; P < 0.0001). The slower decay-time and risetime kinetics of EPSCs evoked in cells of both regions of striatum in slices obtained from pups support the hypothesis that the differences between NMDAR-EPSCs in ventromedial versus dorsolateral striatum in adults arise as a consequence of differences in NR2 subunit incorporation into the NMDARs.

View larger version (14K): [in this window] [in a new window]   Fig. 4. Decay-time kinetics of synaptic NMDAR-EPSCs from neurons in dorsolateral striatum are the same as those from neurons in ventromedial striatum of P4-6 neonatal rats. Sample traces showing the decay kinetics of synaptic NMDAR-EPSCs of neurons in ventromedial and dorsolateral striatum of neonatal rats. Average of 35 EPSCs evoked at 0.1 Hz can be adequately described by a double exponential (gray lines). These EPSCs are normalized to the peak amplitude and superimposed for comparison (right). In the normalized traces, NMDAR-EPSCs from ventromedial and dorsolateral neurons are labeled. For clarification, the normalized dorsolateral EPSC is shown in gray. Currents were recorded at a membrane potential of 70 mV in the presence of 50 µM picrotoxin, 10 µM CNQX, and 10 µM glycine in Mg2+-free solution.

Effects of ifenprodil on NMDAR-EPSCs in ventromedial and dorsolateral neurons in adult striatum

Ifenprodil is an NMDAR antagonist that has a 400-fold greater affinity for NR1/NR2B receptors than for NR1/NR2A (Williams 1993). We therefore used this agent to investigate the pharmacological properties of synaptic NMDAR-EPSCs in neurons in ventromedial and dorsolateral striatum. The low levels of NR2A mRNA expression in ventromedial striatum led to the hypothesis that ifenprodil would have a greater effect on NMDAR-EPSCs in the NR2B-predominate ventromedial striatum. Figure 5A shows the averages of 35 NMDAR-EPSCs from cells of the ventromedial and dorsolateral adult striatum in the presence or absence of ifenprodil. In all cases, drug administration was followed by a wash period, where NMDAR-EPSC peak amplitudes returned to >95% of control amplitudes. At a number of different concentrations (1, 3, and 10 µM), ifenprodil did not differentially inhibit NMDAR-EPSCs in ventromedial versus dorsolateral striatum of adult rats (Fig. 5A). Particularly interesting was the fact that 3 µM ifenprodil had little effect on NMDAR-EPSC peak amplitudes (Fig. 5A), because this concentration of ifenprodil produces a profound decrease in NMDAR-EPSC amplitude at early timepoints in development (no NR2A), and loses efficacy across development as NR2A expression increases (Kew et al. 1998; Kirson and Yaari 1996; Williams 1993). Even at a higher concentration of ifenprodil (10 µM), there was still no significant difference between the amount of inhibition of the NMDAR-EPSCs in ventromedial and dorsolateral striatal neurons (Fig. 5A; 38% inhibition in ventromedial striatum vs. 46% in dorsolateral striatum, P = 0.3). At each of these concentrations of ifenprodil, the kinetics of NMDAR-EPSCs were not altered. This minimal decrease was not due to an ifenprodil-mediated decrease in glutamate release, because 3 µM ifenprodil had no significant effect on non-NMDAR-EPSCs (peak amplitude of non-NMDA component of EPSC was 99 ± 2.3% of control, n = 3; data not shown).

View larger version (25K): [in this window] [in a new window]   Fig. 5. Ifenprodil does not discriminate between synaptic NMDAR-EPSCs of striatal neurons in ventromedial vs. dorsolateral striatum, and much more potently reduces the amplitude of NMDAR-EPSCs recorded in brain slices obtained from P4-P6 neonates. A: (left) example traces of the effects of 3 µM ifenprodil on the peak amplitude of NMDAR-EPSCs from neurons in the ventromedial (top) and dorsolateral (bottom) striatum of adult rats. Right: summary bar graphs showing effects of ifenprodil on NMDAR-EPSC peak amplitude in adult rats. Black bars represent ventromedial cells and open bars represent dorsolateral cells. The y axis represents NMDAR-EPSC peak amplitude expressed as a percent of control. The x axis is the concentration of ifenprodil. Individual bars represent means ± SE of 6 cells in each region at both 1 and 3 µM, and the average of 4 cells in each region at 10 µM. B: (left) example traces of the effects of 3 µM ifenprodil on the peak amplitude of NMDAR-EPSCs from neurons in the ventromedial (top) and dorsolateral (bottom) striatum of P4-P6 rat neonates. Right: summary bar graphs showing the effects of 3 µM ifenprodil on NMDAR-EPSC peak amplitude in slices from rat pups. The y axis represents NMDAR-EPSC peak amplitude expressed as a percent of control. The x axis is striatal subregion from which recordings were made. Individual bars represent means ± SE of 6 cells in each region. All currents were recorded at a membrane potential of 70 mV, in the presence of 50 µM picrotoxin, 10 µM CNQX, and 10 µM glycine in Mg2+-free solution.

Previous studies conducted on hippocampal cells in culture have shown that decreasing the concentration of the required coagonist glycine results in increased sensitivity of NMDA-evoked currents to blockade by ifenprodil (Kendrick et al. 1998). However, elimination of exogenously added glycine failed to alter the inhibition of NMDAR-EPSCs by 10 µM ifenprodil in either location of striatum (41.0 ± 3.8% inhibition in ventromedial striatum vs. 38.7 ± 6.3% in dorsolateral striatum, P = 0.5, n = 3 for both ventromedial and dorsolateral striatum; data not shown).

Effects of ifenprodil on NMDAR-EPSCs in ventromedial and dorsolateral neurons in neonatal striatum

Ifenprodil has been shown to effectively block NMDAR-EPSCs in the brain slices prepared from neonatal animals (Kirson and Yaari 1996). Again, this is a time point at which no NR2A subunit is expressed in brain. In the present studies, administration of 3 µM ifenprodil to P4-6 neonatal rats resulted in dramatic attenuation of NMDAR-EPSC peak amplitudes in both ventromedial and dorsolateral striatum (Fig. 5B). The marked sensitivity of NMDAR-EPSCs to ifenprodil in recordings from neonatal striata, regardless of cell location, provides further evidence that the presence of the NR2A subunit in adult striatal NMDARs alters the function of these channels.

Effects of glycine and D-serine on NMDAR-EPSCs in ventromedial and dorsolateral striatum

Glycine is a required coagonist at NMDARs. NMDARs containing the NR2B subunit have an approximately 10-fold greater affinity for this ligand (Kutsuwada et al. 1992; Stern et al. 1992). Additionally, it has been shown previously in expression systems and cell culture, as well as acute hippocampal brain slices, that exogenous application of glycine increases the amplitude of agonist-evoked currents and NMDAR-EPSCs, respectively (Bergeron et al. 1998; Kutsuwada et al. 1992; Stern et al. 1992; Wilcox et al. 1996). Therefore we were interested in testing the effects of different glycine concentrations on NMDAR-EPSCs from neurons in ventromedial versus dorsolateral striatum. These experiments were conducted in nominal (no added) glycine, and 10 µM NBQX was used in place of CNQX to block all non-NMDAR-EPSCs. As seen previously in high glycine (10 µM) conditions, low frequency stimulation (0.1 Hz) of the ventromedial striatum in nominal glycine evoked EPSCs with a longer decay-time constant relative to EPSCs elicited by the same stimulation paradigm in dorsolateral striatum (Table 2). As seen in recordings from the in vitro hippocampal brain slice preparation (Wilcox et al. 1996), the decay-time kinetics of NMDAR-EPSCs did not change following the switch from nominal to 10 µM glycine (Fig. 6; Table 2). However, addition of 10 µM glycine did result in differential effects on the amplitude of NMDAR-EPSCs in ventromedial and dorsolateral striatum (Fig. 6). In ventromedial striatum, there was no change in the peak amplitude of the NMDAR-EPSCs (Fig. 6, A and C; 103.9 ± 3.3% of control; n = 9 in 9 different slices). In dorsolateral striatum, however, bath application of 10 µM glycine resulted in an increase in the amplitude of the NMDAR-EPSCs (Fig. 6, B and C; 145.0 ± 10.4% of control; n = 6 in 6 different slices). A two-way ANOVA revealed a significant interaction between the concentration of glycine and the location of the cell (P = 0.003). On post hoc analysis, the enhancement of the NMDAR-EPSC amplitude following bath application of 10 µM glycine was significant in dorsolateral striatum (P = 0.017), but not in ventromedial striatum (P = 0.79; Fig. 6C).

View larger version (17K): [in this window] [in a new window]   Fig. 6. Administration of 10 µM glycine increases the amplitude of synaptic NMDAR-EPSCs of striatal neurons in dorsolateral, but not ventromedial, striatum. Local minimal stimulation of striatum in close proximity (<300 µm) to the recorded cell elicits a long-lasting NMDAR-EPSC in (A) ventromedial and (B) dorsolateral striatum. The average of 35 EPSCs evoked at 0.1 Hz is shown. In A and B, currents were recorded at a membrane potential of 70 mV in the presence of 50 µM picrotoxin and 10 µM NBQX in Mg2+-free solution. A: representative effect of 10 µM glycine on synaptic NMDAR-EPSC peak amplitude from a neuron in ventromedial striatum. Application of glycine did not change the NMDAR-EPSC peak amplitude (105% of control). B: representative change in the synaptic NMDAR-EPSC peak amplitude from a neuron in dorsolateral striatum following bath application of 10 µM glycine. Bath application of glycine significantly enhanced the NMDAR-EPSC amplitude (138% of control). C: summary graph of enhancement of the peak amplitude of NMDAR-EPSCs of striatal neurons in ventromedial and dorsolateral striatum by administration of 10 µM glycine. Bars indicate average values ± SE from 6 cells in each region, *Significantly different from control NMDAR-EPSCs (P < 0.05).

Although 10 µM glycine produced a selective enhancement of NMDAR-EPSCs in dorsolateral striatum in our experiments, there is evidence that glycine uptake regulates synaptic levels of glycine and therefore the ability to modulate NMDAR activation (Berger et al. 1998). Although a glycine transporter is expressed in striatum (Borowsky et al. 1993; Zafra et al. 1995), it is currently unknown if there are regional variations in either expression or function of the transporter. To control for any potential effect of regional function of gycine transporters, we examined the effect of D-serine on NMDA-R EPSCs in dorsolateral versus ventromedial striatum. D-serine is an agonist at the glycine site of the NMDA receptor that, like glycine, also displays a difference in affinity for NMDA receptors that are comprised of different NR2 subunits. In addition, D-serine does not serve as a substrate for the glycine transporter (Schell et al. 1995; Supplisson and Bergman 1997). As seen with glycine, application of 3 µM D-serine showed no enhancement of EPSCs in ventromedial striatum (102.5 ± 3.4% of control; P = 0.48; n = 3; Fig. 7, A and C), but a significant enhancement in dorsolateral striatum (124.9 ± 7.0% of control; P = 0.016; n = 4; Fig. 7, B and C). These data provide additional support for distinct NMDARs in differential regions in adult striatum.

View larger version (19K): [in this window] [in a new window]   Fig. 7. Administration of 3 µM D-serine increases the amplitude of synaptic NMDAR-EPSCs of striatal neurons in dorsolateral, but not ventromedial, striatum. Local minimal stimulation of striatum in close proximity (<300 µm) to the recorded cell elicits a long-lasting NMDAR-EPSC in (A) ventromedial and (B) dorsolateral striatum. The average of 35 EPSCs evoked at 0.1 Hz is shown. In A and B, currents were recorded at a membrane potential of 70 mV in the presence of 50 µM picrotoxin and 10 µM NBQX in Mg2+-free solution. A: representative effect of 3 µM D-serine on synaptic NMDAR-EPSC peak amplitude from a neuron in ventromedial striatum. Application of D-serine did not change the NMDAR-EPSC peak amplitude (103% of control). B: representative change in the synaptic NMDAR-EPSC peak amplitude from a neuron in dorsolateral striatum following bath application of 3 µM D-serine. Bath application of D-serine significantly enhanced the NMDAR-EPSC amplitude (129% of control). C: summary graph of enhancement of the peak amplitude of NMDAR-EPSCs of striatal neurons in ventromedial and dorsolateral striatum by administration of 3 µM D-serine. Bars indicate average values ± SE from 3 cells in each region. *Significantly different from control NMDAR-EPSCs (P < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The present experiments provide important new evidence for both temporal and spatial regulation of functionally distinct synaptic NMDARs in brain slices of the rodent striatum. Not only do the risetimes and decay times of NMDAR-EPSCs become faster as a consequence of development, but there is a regional disparity in the EPSC kinetics as well. NMDAR-EPSCs in the dorsolateral striatum have significantly faster risetimes as well as substantially faster decay times than those recorded in the ventromedial region of the striatum. Additionally, application of the NMDA receptor coagonists glycine and D-serine significantly increases the NMDAR-EPSC peak amplitude only in the dorsolateral region of the striatum. These findings suggest strongly that properties of NMDARs, and thus processing of excitatory input across ventromedial to dorsolateral adult striatum, are not uniform. Such variation likely will give rise to differences in signal processing in these specific regions of striatum, and might be relevant to the different functions of ventromedial (limbic) versus dorsolateral (sensorimotor) striatum.

The differences in the NMDAR-EPSC rise- and decay-time kinetics observed regionally in adult striatum likely arise as a consequence of differences in NR2A subunit incorporation into the synaptic NMDARs. First, irrespective of location in striatum, basic membrane properties of the cells were very similar. Second, the level of NR2A subunit mRNA expression is substantially greater in lateral than medial striatum (Ganguly and Keefe 2001; Watanabe et al. 1993), and this regional distribution of subunit expression matches the regional distribution of the kinetic differences. Third, the activation and deactivation kinetics of NR1/NR2A receptors studied in expression systems are faster than those of NR1/NR2B receptors (Chen et al. 1999; Monyer et al. 1994; Vicini et al. 1998). Likewise, as NR2A subunit expression increases across development, the rise times (Kirson and Yaari 1996) and decay kinetics of the NMDAR-EPSC, both here and in a number of other brain regions, decrease (Carmignoto and Vicini 1992; Flint et al. 1997; Hurst et al. 2001; Stocca and Vicini 1998). Finally, the present results show that there are no differences in the rise times or decay kinetics of NMDAR-EPSCs in ventromedial versus dorsolateral striatum at a developmental timepoint (P4-6) at which there is no NR2A subunit expression in striatum (Sheng et al. 1994; Wang et al. 1995). However, the immaturity of excitatory synapses on striatal neurons at this timepoint (Butler et al. 1998) may contribute to the slow NMDAR-EPSC decay-time kinetics in both regions. Although other cellular mechanisms such as phosphorylation and [Ca²⁺]_i (Shi et al. 2000; Umemiya et al. 2001) can influence NMDAR-EPSC decay-time kinetics, regional differences in the striatum of either calcineurin or Ca²⁺ buffering have not been reported (for example, see Sola et al. 1999). In addition, we have demonstrated here that the kinetics for the EPSCs are still dramatically different at holding potentials of +40 mV, a voltage at which Ca²⁺ influx through NMDA receptor-gated ion channels and voltage-sensitive Ca²⁺ channels is much reduced. This suggests that the kinetics of the EPSCs are due to intrinsic differences of the NMDA receptor rather than regional differences in calcium signaling. Taken together, these findings strongly suggest that the differential regional distribution of the NR2A subunit in adult striatum results in the presence of functionally distinct NMDARs in the same neuron type in different regions of adult striatum.

Whereas the kinetic properties of NMDAR-EPSCs in ventromedial and dorsolateral striatum differ, the present findings indicate that not all of the pharmacological properties of the receptor vary in a similar manner. In slices from adults, bath application of the NR2B-selective antagonist ifenprodil did not differentially affect NMDAR-EPSCs in ventromedial versus dorsolateral striatum. These findings are consistent with our previous in vivo studies that failed to find differential effects of pharmacologically distinct NMDAR antagonists on immediate early gene expression in medial versus lateral striatum (Adams et al. 2000; Keefe and Adams 1998; Keefe and Ganguly 1998). Not only did ifenprodil not differentially affect NMDAREPSCs in dorsolateral versus ventromedial striatum, it produced only limited blockade of the NMDAR-EPSCs in the adult striatum. Consistent with these observations, a number of reports using adult animals have found ifenprodil to be diminished in its capacity to block synaptic NMDAR-EPSCs (Cathala et al. 2000; Ramoa and Prusky 1997; Tovar and Westbrook 1999). These studies suggest that incorporation of the NR2A subunit into the receptor complex in adult animals is responsible for this decreased sensitivity to ifenprodil (Tovar and Westbrook 1999; Williams 2001). If this is the case, then it is likely that all synaptic NMDA receptors in the adult striatum contain an NR2A subunit. Further supporting this hypothesis are the findings herein that the risetimes and decay-time kinetics of the NMDAR-EPSCs are faster in adults than in neonates, irrespective of regional location within the striatum, and that NMDAR-EPSCs are potently blocked by ifenprodil in striatal slices taken at a developmental time point (P4-6) at which the NR2A subunit is not present (Sheng et al. 1994; Wang et al. 1995). Therefore based on the observed developmental changes and regional differences in the kinetic and pharmacological properties of the NMDAR-EPSCs, we hypothesize that the majority of synaptic striatal NMDARs in adult likely contain the NR2A subunit, but that the number of NR2A subunits per receptor is greater in dorsolateral striatum than in ventromedial striatum, giving rise to functionally distinct NMDARs in different regions of the adult striatum.

Although ifenprodil did not differentiate functionally distinct NMDARs in ventromedial and dorsolateral striatum, manipulation of the glycine or D-serine concentration did. NMDARs containing the NR2A subunit have a lower affinity for glycine than do those containing the NR2B subunit (Kutsuwada et al. 1992; Stern et al. 1992). Thus NMDARs in dorsolateral striatum, where NR2A expression is greatest, should have the lowest affinity for glycine in striatum and be most affected by exogenous glycine application. In fact, addition of 10 µM glycine or 3 µM D-serine significantly enhanced NMDAR-EPSCs in dorsolateral striatum. While differences in endogenous glycine levels could exist, regional differences in glycine transporter expression in striatum to date have not been reported (Borowsky et al. 1993; Zafra et al. 1995). The fact that D-serine had no effect on NMDAR-EPSC peak amplitude in ventromedial striatum suggests that differences in endogenous glycine levels due to regional differences in transporter expression or function do not underlie the selective enhancement of EPSCs in the dorsolateral striatum. These data further support the idea that levels of ambient glycine in the slice are not saturating for postsynaptic NMDA receptors expressed in cells of the dorsolateral striatum, an area rich in NR2A subunit expression. Indeed, receptors comprised of NR2A subunits have a lower glycine affinity than those receptors comprised of NR2B subunits. Thus the kinetic differences discussed above and the selective effect of glycine and D-serine manipulations on NMDAR-EPSC peak amplitude in neurons of dorsolateral striatum may reflect a greater overall incorporation of NR2A-subunits into the NMDARs mediating these EPSCs. In addition, these findings suggest that glycine or D-serine availability is likely to significantly and differentially impact NMDAR-mediated functions in different regions of striatum.

The functional relevance of subpopulations of NMDARs with distinct biophysical and pharmacological properties in striatum is currently unknown. However, Partridge et al. (2000) showed that one can induce LTP at corticostriatal synapses in both medial and lateral striatum early in development. As animals age, however, high-frequency stimulation of corticostriatal afferents results in long-term depression in dorsolateral striatum, the area in which NR2A mRNA expression increases notably during development. At this same developmental timepoint, high-frequency stimulation of cortical afferents terminating in the dorsomedial aspect of striatum, where NR2A mRNA expression is not as abundant, continues to produce LTP. At a variety of synapses, such as those in visual cortex and hippocampus, changing the ratio of NR2A:NR2B is hypothesized to produce a shift in the probability and degree of induction of LTP (Kirkwood et al. 1996; Philpot et al. 2001; Tang et al. 1999; but see also Barth and Malenka 2001; Lu et al. 2001). Thus the relative level of NR2A expression in different regions of adult striatum may underlie different types of corticostriatal synaptic plasticity.

In conclusion, the present study provides electrophysiological and pharmacological evidence for different types of synaptic NMDARs in neurons in ventromedial versus dorsolateral striatum. These differences are most likely due to differential regional expression of NR2 subunits. NMDAR-EPSCs in neurons in dorsolateral striatum have substantially faster decay kinetics than those in ventromedial striatum, and the kinetics of NMDAR-EPSCs in both regions are significantly faster than the kinetics of NMDAR-EPSCs in brain slices of neonates. In addition, we have demonstrated that ambient concentrations of glycine are subsaturating in dorsolateral striatum, where a higher amount of NR2A subunit is expressed. These findings are consistent with both the temporal and spatial regulation of levels of expression of mRNA encoding the NR2A subunit in the striatum and may relate to known differences in the processing of cognitive and motor information in these two different regions of the striatum.