Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

The N-methyl-D-aspartate (NMDA) receptor is a glutamate-gated, multisubunit channel complex that is involved in a variety of neural processes, including synaptic transmission, synaptic plasticity, and cell death (Watkins and Evans, 1981; Choi, 1992; Bliss and Collingridge, 1993). NMDA receptor subunits are distributed heterogeneously throughout the central nervous system and it is likely that different NMDA receptor complexes underlie distinct physiological, as well as pathological, roles. With the development of subunit-specific antagonists it should be possible to determine the specific function of NMDA receptor subpopulations and reduce the adverse side effects of antagonists used therapeutically.

Native NMDA receptors contain subunits from both the NR1 and NR2 families of receptor subunits (Moriyoshi et al., 1991; Kutsuwada et al., 1992; Monyer et al., 1992, 1994; Ishii et al., 1993). In addition, some NMDA receptors may contain an NR3 subunit that appears to negatively modulate channel activity (Das et al., 1998). The NR1 gene undergoes alternative RNA splicing to give rise to eight splice variants (Durand et al., 1992; Sugihara et al., 1992; Hollmann et al., 1993). Four closely related NR2 genes code for the four NR2 subunits, NR2A-D (Ikeda et al., 1992; Ishii et al., 1993; Monyer et al., 1994). Consistent with evidence that the glutamate recognition site is located on the NR2 subunit (Laube et al., 1997; Anson et al., 1998), the differing NR2 subunits confer distinct glutamate site pharmacological properties on the NR1/NR2 heterodimer (Buller et al., 1994; Laurie and Seeburg, 1994; Buller and Monaghan, 1997).

To date, the two major forebrain NR2 subunits (NR2A and NR2B) have primarily been shown to have very similar glutamate recognition site pharmacological profiles with only a difference in their relative affinity for agonists and antagonists. Compared with the NR2B subunit, the NR2A subunit confers higher affinities for antagonists and lower affinities for agonists (Buller et al., 1994; Laurie and Seeburg, 1994; Priestley et al., 1995; Grimwood et al., 1996; Kendrick et al., 1996). However, with more recently developed antagonists (Jane et al., 2000), we find that recombinant NR2A and NR2B-containing receptors expressed in Xenopus oocytes display differing antagonist pharmacological profiles with three antagonists displaying higher affinities for NR2B- than NR2A-containing receptors (Buller and Monaghan, 1997). With the radioligand L-[³H]glutamate, it has been possible to describe native NMDA receptor populations pharmacologically corresponding to NR2B-, NR2C-, and NR2D-containing receptors (Beaton et al., 1992; Monaghan et al., 1998). However, native receptors having the distinctive pharmacological properties described for recombinant NR2A-containing NMDA receptors have not yet been observed. Thus, either L-[³H]glutamate is not a suitable ligand for native NR2A-containing NMDA receptors, or NR2A subunits display altered pharmacological properties when assembled into native NMDA receptor complexes.

In an effort to compare the pharmacology of native NR2A- and NR2B-containing receptors, we have examined the binding properties of the radioligand [³H]CGP39653 in the cerebellum and medial striatum. Because [³H]CGP39653 ([³H](E)-2-amino-4-propyl-5-phosphono-3-pentenoic acid) is an NMDA antagonist with low affinity for NR2C-containing receptors (Laurie and Seeburg, 1994), binding in cerebellum, a brain region exclusively containing NR2A and NR2C subunits (Monyer et al., 1992; Watanabe et al., 1993), should be limited to the NR2A subunit. [³H]CGP39653 binding in the medial striatum, however, should be predominately to NR2B-containing receptors because this region has high relative expression levels of NR2B subunits (Watanabe et al., 1993; Buller et al., 1994). Consistent with these predictions, our results indicate that [³H]CGP39653 labels two distinct populations of NMDA receptors that have pharmacological and anatomical properties that correspond to NR2A and NR2B subunits.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Quantitative Receptor Autoradiography. Following halothane anesthesia, brains were removed from adult male Sprague-Dawley rats (200-250 g) and immediately frozen under powdered dry ice. Horizontal sections (6 µm) were cut with a cryostat, thaw mounted onto gelatin-subbed slides, and stored at 20°C (<24 h). Slides were then air dried and preincubated for 30 min at 0°C in 50 mM Tris-acetate, pH 7.6, followed by an additional preincubation for 20 min at 30°C in fresh buffer. Sections were incubated with 20 nM [³H]CGP39653, or the indicated range of [³H]CGP39653 concentrations, in 50 mM Tris-acetate, pH 7.6, containing 1 mM MgCl₂ for 30 min at 0°C. Nonspecific binding, defined by including 500 µM CGS-19755 (cis-4-phosphonomethyl-2-piperidine carboxylic acid), represented ~5% of total binding. Sections were washed for 20 s in ice-cold buffer and quickly dried by an airstream. The slides were placed against ³H-sensitive film (Amersham) with ³H-standards (Amersham, Arlington Heights, IL) and were allowed to expose for 5 weeks at 4°C.

Quantitative Analysis of Autoradiograms. Quantification was accomplished by computer-assisted image analysis of the autoradiographs (Imaging Research, St. Catherines, Ontario, Canada). Binding densities were calibrated from standard curves of [³H]Microscale (Amersham) radioactive standards. With iterative, nonlinear regression analysis (ISI Software; GraphPad, San Diego, CA) dose-response curves were generated for the [³H]CGP39653 saturation analysis and the competitive NMDA antagonists. The derived [³H]CGP39653 K_d values were used to determine antagonist K_i values (Cheng and Prusoff, 1973). All experiments were performed at least three times.

Compounds. PBPD [(±)-cis-4-(4-phenylbenzoyl)piperazine-2,3-dicarboxylic acid] was synthesized and purified by the authors; (R)-2-amino-5-phosphonopentanoate (D-AP5) was obtained from Tocris (Bristol, UK); LY233536 [(±)-6-(1H-Tetrazol-5-yl-methyl) decahydroisoquinoline-3-carboxylic acid] was kindly provided by Dr. Ornstein, (Eli Lilly, Indianapolis, IN); EAB515 {(S)--amino-5-(phosphonomethyl)[1,1'-biphenyl]-3-propanoic acid} and D-CPPene [(R)-(E)-4-(3-phosphonoprop-2-enyl)piperazine-2-carboxylic acid] were kindly provided by Drs. Aebischer and Mueller (Novartis Pharmaceuticals, Basel, Switzerland). [³H]CGP39653 was obtained from NEN (Boston, MA).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Saturation Analysis of [³H]CGP39653 Binding. [³H]CGP39653 displayed single-site, saturable high-affinity binding (12-23 nM) with fairly similar affinities in the differing brain regions examined (Fig. 1; Table 1). These K_d values are comparable to those found for [³H]CGP39653 binding to recombinant NR2A- and NR2B-containing NMDA receptors (4-22 nM) (Laurie and Seeburg, 1994; Kendrick et al., 1996). [³H]CGP39653 displayed a B_max of 4.7 pmol/mg protein in the hippocampal CA1 stratum radiatum, a value comparable to that previously described for other ligands of NMDA receptors (3-4 pmol/mg protein) (Monaghan et al., 1984; Monaghan and Cotman, 1985). The central nervous system pattern of labeling observed for 20 nM [³H]CGP39653 differed from NMDA-sensitive L-[³H]glutamate binding (Monaghan and Cotman, 1985) in that [³H]CGP39653 binding was not observed in the glomerular layer of the olfactory bulb and [³H]CGP39653 labeling in the midline thalamic nuclei was not greater than in the lateral thalamic nuclei. [³H]CGP39653 (and L-[³H]glutamate) displayed relatively greater binding in the medial striatum and lateral septum than does the NR2A-selective radioligand [³H]CPP (Monaghan et al., 1988; Buller et al., 1994).

View larger version (24K): [in this window] [in a new window]   Fig. 1.   Saturation analysis of [3H]CGP39653 binding to NMDA receptors in rat brain. Rat brain tissue sections were incubated with increasing concentrations of [3H]CGP39653 and the levels of specific binding in the CA1 stratum radiatum (CA1), cerebellar granule cell layer (CRB), medial striatum (STR), and the ventral posterior medial thalamic nucleus (VPM) were determined by quantitative autoradiography. Corresponding Kd and Bmax values are in Table 1.

Antagonist Pharmacology of [³H]CGP39653 Binding. Dose-response curves were generated for NMDA receptor antagonists against [³H]CGP39653 binding in the rat brain autoradiographic preparation. Qualitative analysis of antagonist displacement of [³H]CGP39653 binding revealed two general patterns of displacement (Fig. 2). After partial [³H]CGP39653 displacement by D-CPPene or D-AP5, binding preferentially remained in the most superficial layers (I-III) of the cerebral cortex and in the anterior cingulate cortex. Layer IV parietal cortex was particularly sensitive to displacement by D-CPPene and D-AP5. In contrast, partial displacement by EAB515, LY233536, and PBPD revealed autoradiograms with a more homogeneous cortical lamination pattern and relatively higher binding levels in specific lateral thalamic nuclei (the medial and lateral ventral posterior nuclei and the laterodorsal nuclei). In addition, EAB515, LY233536, and PBPD, but not D-CPPene and D-AP5, were noticeably weaker displacers of binding in the cerebellar granule cell layer compared with forebrain.

View larger version (78K): [in this window] [in a new window]   Fig. 2.   Antagonist displacement of [3H]CGP39653 binding to horizontal sections of rat brain show two patterns of displacement. High binding levels are indicated by red and progressively lower levels by yellow, green, and blue (as indicated by the calibration bar). In the presence of 1 µM D-AP5, [3H]CGP39653 binding remains in cortical layers I-III and Va of parietal cortex (PC), throughout the striatum, and lateral septum (LS), and anterior cingulate (AC). In contrast, [3H]CGP39653 binding not displaced by 30 µM PBPD is found more broadly through the cerebral cortex with the more heavily labeled superficial layer now including layer IV, in the ventral posterior medial (and lateral) thalamic nucleus (VPM), and in the cerebellar granule cell layer (CB). High levels of both populations are found in the hippocampus (HC).

View larger version (11K): [in this window] [in a new window]   Fig. 3.   Dose-response curves for the inhibition by D-AP5 and LY233536 of [3H]CGP39653 binding in the cerebellar granule cell layer (CB), medial striatum (MStr), and ventral posterior medial thalamic nucleus (VPM). Although D-AP5 displayed similar affinities in each region studied, LY233536 displayed significantly lower affinity in the cerebellum. All measurements in this representative experiment were taken from adjacent sections of the same brain; experiments were performed four times. For averaged results, see Table 1.

View larger version (33K): [in this window] [in a new window]   Fig. 4.   Comparison of antagonist potencies at native and recombinant NR2A- and NR2B-containing receptors. For each antagonist, their potencies at inhibiting [3H]CGP39653 binding in the cerebellum and striatum are expressed as a ratio (cerebellar Ki/medial striatum Ki; solid bars). For comparison, the NR1-NR2A Ki/NR1-NR2B Ki potency ratio of each antagonist at recombinant NMDA receptors expressed in Xenopus oocytes also is shown (hatched bars; data from Buller and Monaghan, 1997). For the five drugs, two potency ratios were highly correlated (r = 0.89).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

To directly compare NR2A and NR2B-containing native NMDA receptors, we used [³H]CGP39653 because this ligand can radiolabel recombinant NR1/NR2A and NR1/NR2B receptors (Kendrick et al., 1996) but has 10- to 40-fold lower affinity at NR1/NR2C and NR1/NR2D receptors (Laurie and Seeburg, 1994). Because the medial striatum has predominately the NR2B type of NR2 subunits, this region should reflect NR2B properties, whereas the cerebellar granule cell layer has only NR2A and NR2C subunits and hence should reflect only NR2A properties when labeled with [³H]CGP39653. As discussed below, anatomical and pharmacological findings from this study support the NR2A and NR2B identification of [³H]CGP39653 binding sites.

In this study, we found that native NMDA receptors of the medial striatum displayed pharmacological properties that were distinct from NMDA receptors in the cerebellum. Specifically, the antagonists LY233536, EAB515, and PBPD displayed higher affinity for the NR2B-enriched medial striatum, whereas D-CPPene and D-AP5 displayed a slightly higher affinity (but not statistically significant) for apparent NR2A-containing NMDA receptors of the cerebellum. These findings are consistent with the pattern of antagonist selectivity that we previously reported for recombinant NR2A and NR2B subunits expressed in Xenopus oocytes (Buller and Monaghan, 1997). As shown in Fig. 4, there is a strong correlation between the antagonist profiles of the native NMDA receptors that appear to contain NR2A and NR2B subunits and the antagonist profiles of recombinant NR2A and NR2B-containing NMDA receptors. Thus, these findings strongly suggest that native NR2A and NR2B subunits have the same antagonist specificities as do recombinant NR2A- and NR2B-containing NMDA receptors, respectively.

Relevant to the above-mentioned conclusions, anatomical results from the present study support our initial hypothesis that [³H]CGP39653 can be used to selectively label NR2A and NR2B subunits. [³H]CGP39653 labeled only those brain regions expressing NR2A or NR2B subunits. For example, both the cerebellar granule cell layer, which has NR2A and NR2C subunits, and the medial striatum and lateral septum, which predominately have NR2B subunits, were labeled in the present study. Furthermore, the two distinct patterns of anatomical selectivity displayed by the NMDA receptor antagonists correspond well to NR2A and NR2B subunit distributions. D-AP5 and D-CPPene preferentially displaced [³H]CGP39653 binding in the NR2A-enriched deep cerebral cortical layers (particularly layer IV) and the ventral posterior medial and lateral thalamus, whereas leaving an NR2B-like distribution (Fig. 2a; cf. Figs. 1 and 2, Buller et al., 1994 and Fig. 1, Monaghan et al., 1998). In contrast, EAB515, PBPD, and LY233536 preferentially displaced [³H]CGP39653 binding in NR2B-enriched regions (e.g., the anterior cingulate cortex, the medial striatum, and lateral septum) leaving a, NR2A-like distribution (Fig. 2b).

Consistent with studies with recombinant receptors (Laurie and Seeburg, 1994), [³H]CGP39653 did not appear to label native NR2C and NR2D subunits. The absence of NR2C and NR2D labeling is supported by the finding that D-CPPene, which has low affinity for NR2C- and NR2D-containing receptors (Buller et al., 1994; Laurie and Seeburg, 1994), did not display a second low-affinity component in inhibiting [³H]CGP39653 binding in brain regions expressing NR2C and NR2D subunits (Ishii et al., 1993; Monyer et al., 1994). Furthermore, [³H]CGP39653 binding was absent from the glomerular layer of the olfactory bulb; this region contains both NR2C and NR2D subunits, but no detectable NR2A or NR2B (Buller et al., 1994). The absence of NR2C labeling explains why D-AP5 and D-CPPene display a relatively high affinity for cerebellar [³H]CGP39653 binding sites, whereas these antagonists display a markedly low affinity for cerebellar NMDA receptors labeled by L-[³H]glutamate (Beaton et al., 1992; Buller et al., 1994). Thus, multiple lines of evidence indicate that [³H]CGP39653 selectively labels NR2A- and NR2B-containing receptors under the conditions used in the present study. In the absence of Mg²⁺, [³H]CGP39653 will only bind to NR2A-containing receptors (Laurie and Seeburg, 1994). The addition of 1 mM Mg²⁺ is necessary to increase the affinity of [³H]CGP39653 for NR2B-containing receptors to enable radioligand binding (Kendrick et al., 1996), but Mg²⁺ effects on recombinant NR2C (and NR2D) have not been reported. From the present experiments, it appears that Mg²⁺ enhancement of affinity for NR2C subunits, if it occurs, cannot compensate for the 17- and 40-fold lower affinity (Laurie and Seeburg, 1994) that CGP39653 has for NR2C compared with NR2B- and NR2A-containing receptors, respectively.

[³H]CGP39653 displayed a similar, or slightly higher, affinity for medial striatal NMDA receptors than for cerebellar NMDA receptors. In contrast, CGP39653 displays a 2- to 6-fold higher affinity for recombinant NR2A- than NR2B-containing NMDA receptors (Laurie and Seeburg, 1994; Kendrick et al., 1996). This finding could potentially be accounted for by the presence of NR2C subunits in the cerebellum. [³H]CGP39653 affinity was determined in a saturation assay that included concentrations up to 160 nM. Although we could not resolve a low-affinity component, it is possible that at the higher concentrations, some low-affinity [³H]CGP39653 binding to NR2C subunits occurs that lowers the apparent affinity of [³H]CGP39653. This hypothesis is consistent with the finding that [³H]CGP39653 displayed a higher affinity in the lateral thalamus, which contains both NR2A and NR2B but negligible NR2C, than in the medial striatum, which contains primarily NR2B subunits. Because the potency of the five competitive inhibitors in the cerebellum displayed a high correlation to the recombinant NR2A pharmacological profile, it appears that at 20 nM, [³H]CGP39653 did not significantly label NR2C subunits. It is also possible that cerebellar NR2A subunits may coassemble with NR2C subunits in the cerebellum (Wafford et al., 1993) and that in this complex, NR2A subunits may have a reduced [³H]CGP39653 affinity.

The pharmacological pattern found for apparent NR2A-containing receptors described herein has not been described for native NMDA receptors labeled with L-[³H]glutamate (Beaton et al., 1992; Buller et al., 1994), or with any other radioligand. In this study, we find that the apparent NR2A-containing NMDA receptors of the cerebellum displayed the distinctive pharmacological properties displayed by recombinant NR2A-containing NMDA receptors. Consequently, the glutamate recognition site of NR2A in the cerebellum does not appear significantly altered in a native NMDA receptor complex.

In brain regions where NR2A and NR2B subunits are coexpressed, it is possible that these subunits are found in the same receptor complex (Sheng et al., 1994). Qualitatively (Fig. 2), we can observe the two distinct antagonist displacement patterns corresponding to NR2A and NR2B subunits. Thus, it appears that their possible coassembly does not significantly alter their respective glutamate-binding site pharmacological properties on NR2 subunits or that these heteromeric complexes represent a relatively minor population. Although the NR2A expression pattern could be observed in the thalamus after partial displacement with NR2B-preferring antagonists, the antagonist K_i values for lateral thalamus, which contains both NR2A and NR2B subunits, were most similar to that found for NR2B. It is likely that the higher affinity of L-glutamate for the abundant NR2B-containing receptors (Kutsuwada et al., 1992; Buller et al., 1994; Laurie and Seeburg, 1994; Priestley et al., 1995) masks the relatively lower levels of NR2A labeling by L-[³H]glutamate.

NR2A and NR2B subunits have been difficult to differentiate with glutamate site antagonists. Previous studies have shown that glutamate site antagonists, in general, display higher affinity for NR2A- than NR2B-containing receptors (Buller et al., 1994; Laurie and Seeburg, 1994; Priestley et al., 1995; Grimwood et al., 1996; Kendrick et al., 1996). However, these studies used competitive antagonists composed of straight chain, piperadine, or piperazine structures (e.g., D-AP5, CGS19755, and D-CPPene) and these "classic" antagonists all display the potency series NR2A > NR2B > NR2C > NR2D. In contrast, in the current study, we included the more "atypical" antagonists LY233536, PBPD, and EAB515 (Andaloro et al., 1996); each of these contain two extra cyclic groups that may increase rigidity (LY233536) or may increase interactions with a hydrophobic pocket in the ligand binding site (PBPD and EAB515). Also potentially relevant is the observation that LY233536 and EAB515 are distinct in having greater NMDA receptor activity in their S-isomers (Müller et al., 1992; Ornstein and Klimkowski, 1992) (the stereoisomers of PBPD have not been resolved).

Because the glutamate recognition sites of NR2A and NR2B subunits are pharmacologically distinct, it should be possible to identify, or develop, antagonists of yet greater subtype selectivity. Such compounds may have useful clinical/research applications. NMDA receptor antagonists have several potential therapeutic uses; however, the clinical application of these antagonists has been hindered by the adverse side effects associated with these compounds. Given the differing distributions of NR2A and NR2B subunits in the brain, it is likely that antagonists selective for distinct NMDA receptor subunits would display different therapeutic/adverse effect profiles. Consistent with this hypothesis, LY233536 and EAB515 have been noted to have atypical behavioral effects compared with classic NMDA receptor antagonists (Urwyler et al., 1996; D. Leander, personal communication). Thus, the development of competitive antagonists with greater selectivity for NR2A or NR2B subunits should yield compounds with distinct effects on brain function.