Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

NMDA selectively activates a major subclass of glutamatergic excitatory amino acid receptors in the vertebrate CNS. Numerous studies have described a significant role for NRs in synaptic plasticity and in a number of neurodegenerative diseases and pathological states, including epilepsy and neuronal cell death after ischemic injury (Olney, 1990; Collingridge and Lester, 1989). Prolonged activation of NRs under such conditions is a major source of extracellular calcium entry into the cell. This results in cellular calcium overload, which is thought to play a substantial role in triggering processes that result in cellular neurodegeneration and death (Olney, 1990).

The NR is a ligand-gated cation channel that is activated by the coagonists glutamate (or selectively in vitro by NMDA) and glycine acting at a strychnine-insensitive glycine site (Wong and Kemp, 1991). It is further regulated by many other modulatory sites, which include a voltage-dependent block of the channel by Mg²⁺, a voltage-independent action of Zn²⁺, the redox state of the receptor, arachidonic acid, ethanol, neurosteroids, pH and polyamines (Rock and MacDonald, 1995; Romano and Williams, 1994; Yoneda and Ogita, 1991; Olney, 1990). The NR complex consists of heteromeric assemblies of subunits (Hollmann and Heinemann, 1994). Two classes of subunits, designated NR1 and NR2, have been cloned from rat brain. Eight isoforms of the NR1 subunit and four types of the NR2 subunit, which have been designated NR2A, NR2B, NR2C and NR2D, have been found (Zukin and Bennett, 1995; Hollmann and Heinemann, 1994). The distribution, developmental regulation and pharmacology of the various combinations of these subunits are of great interest (Portera-Cailliau et al., 1996; Buller et al., 1994; Lynch et al., 1995; Laurie and Seeburg, 1994; Williams, 1994).

Ifenprodil and eliprodil are neuroprotective agents whose mechanism of action has been ascribed to their NMDA antagonist properties (Scatton et al., 1994). They have been shown to be noncompetitive inhibitors of NRs in functional studies and binding assays (Reynolds and Miller, 1989), and it has been suggested that ifenprodil is an antagonist of the stimulatory effects of polyamines (Scatton et al., 1994; Carter et al., 1990). Studies using recombinant NRs have demonstrated that ifenprodil is a selective antagonist at the NR1a/NR2B subtype of the NR (Williams, 1993; Nicolas and Carter, 1994; Williams et al., 1993). Recently, these findings with ifenprodil have been confirmed, and haloperidol and trifluperidol have been shown to have a similar selectivity for the NR1a/NR2B receptor subtype expressed in Xenopus oocytes (Ilyin et al., 1995). Although the antagonist action of haloperidol at the NR has been attributed to an interaction with the strychnine-insensitive glycine site (Fletcher and MacDonald, 1993), results from Ilyin et al. (1996) suggest that haloperidol's action on NRs is mediated by a noncompetitive allosteric modulatory site expressed by isoforms of the receptor containing the NR2B subunit.

Radiolabeled TCP and (+)MK-801 have been extensively used as biochemical probes to determine the functional state of the channel when the agonist recognition sites or the various modulating sites are occupied with agonists or antagonists (Yoneda and Ogita, 1991; Reynolds and Miller, 1990). Binding studies using radiolabeled MK-801 and ifenprodil in rat brain have shown that ifenprodil distinguishes between two discrete populations of NRs (Williams, 1993; Scatton et al., 1994; Reynolds and Miller, 1989). The data strongly suggest that the high-affinity inhibition of the binding of these radioligands to native NRs by ifenprodil is mediated by an NR subtype that contains the NR2B subunit. Moreover, it has been demonstrated that both the NR1 and the NR2 subunits regulate interaction of the receptor with polyamines (Williams et al., 1994; Zhang et al., 1994; Williams, 1994).

In this study, we compared the inhibition of haloperidol and trifluperidol to that of ifenprodil and eliprodil by using [³H]TCP binding to native NRs in rat brain under conditions that made it possible to detect high-affinity binding to receptors expressing the NR2B subunit. We also characterized the inhibition of these agents in [³H]ifenprodil binding to determine whether their pharmacology at a site directly labeled with ifenprodil reflected the pharmacology of the electrophysiological assays (Ilyin et al., 1996; Nicolas and Carter, 1994; Schoemaker et al., 1995). In addition, we determined the effects of GBR 12909, a high-affinity sigma site ligand that has been used to mask the binding of [³H]ifenprodil to sigma sites (Schoemaker et al., 1990), and of the polyamine spermidine, which has been shown to activate selectively NRs that express the NR2B subunit (Williams, 1994), on the inhibition of haloperidol in this assay.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

All buffers and reagents used in assay incubations or to dissolve drugs were prepared with water purified through a Milli-Q reverse osmosis system (Millipore Corp., Bedford, MA) and treated with UV emissions. Before their use in the assays, buffers were further filtered through a sterile Corning filtration unit (Corning Glass Works, Corning, NY) containing a 0.2-µ filter. The buffer used to rinse the membranes on the assay filters was prepared with purified water, but it was not refiltered and was stored no longer than 5 days. Stock solutions of the drugs (usually 10 mM) were dissolved in 20 mM HEPES-KOH buffer, pH 7.4 (assay buffer), with the addition of 1 to 5 µl of glacial HAc, if needed to keep them in solution. For eliprodil, the stock solution was buffer with the addition of 10% dimethyl sulfoxide. All subsequent dilutions from stock were made in buffer.

Membrane Preparation

The preparation of an extensively washed, lysed, frozen/thawed, buffy coat membrane was based on previously described methods (Jones et al., 1989). Whole adult rat brains minus the brainstem and cerebellum were purchased frozen (Zivic-Miller Laboratories, Inc., Zelienople, PA) and stored at 80°C. Brains were thawed at room temperature, and two forebrains were homogenized in 35 to 40 ml of ice-cold 0.32 M sucrose in a Potter-Elvehjem glass homogenizer, by means of a motor-driven Teflon pestle until the homogenate was uniform (about 8 strokes up and down). Homogenates were centrifuged at 285 × g for 10 min at 4°C. The pellets were discarded, and the supernatants were recentrifuged at 18,000 × g for 20 min at 4°C. Supernatants were slowly decanted to retain the soft buffy coat on the pellet. Each pellet was briefly disrupted with a Polytron homogenizer, setting 6, in 35 ml each of ice-cold filtered water. Homogenates were centrifuged at 7000 × g for 20 min at 4°C. The soft buffy coat was removed by shaking over parafilm and was decanted into clean centrifuge tubes. The buffy coat preparation was centrifuged at 34,500 × g for 20 min at 4°C, and the supernatant was slowly decanted as before. The pellets were disrupted with the Polytron in 35 ml each of 50 mM HEPES-KOH buffer, pH 7.4, and centrifuged at 34,500 × g for 10 min at 4°C; the resulting supernatants were discarded. This step was repeated three more times. The final pellets were stored at 80°C until used.

On the day of the assay, pellets were thawed and disrupted with the Polytron in 35 ml each of the assay buffer at pH 7.4. Homogenates were incubated at 37°C for 30 min in a shaking water bath, followed by centrifugation at 40,000 × g for 10 min at 4°C. This wash step without the incubation was repeated three more times. For use in the assays, each pellet was resuspended by using the Polytron in 20 ml of the assay buffer and pooled.

To prepare rat pup brain membranes, we decapitated 3-day-old rat pups (Charles River, Portage, MI) removed the brains and dissected away the cerebellum and brainstem. Ten brains per centrifuge tube were homogenized and prepared as described for adult brains.

Binding Studies

[³H]TCP binding. Triplicate incubations were carried out in a volume of 0.5 ml in 1.3-ml polyproylene tubes (Marsh Biomedical Products Inc., Rochester, NY) for 10 min at room temperature. Incubations contained test agents, membranes (100-200 µg protein) and 2 nM [³H]TCP in 20 mM HEPES-KOH buffer, pH 7.4 (assay buffer). Assays were started by addition of the membranes. Bound radioligand was separated by filtration under reduced pressure with a Tomtec Mach II, 96-well cell harvester (Tomtec Inc., Orange, CO). Filtration was through Whatman GF/B glass-fiber filters (Whatman Ltd., Maidstone, England), which had been soaked for at least 15 min in 0.1% polyethylenimine and allowed to air dry. The filters were rinsed with 3 ml of ice-cold assay buffer within 6 sec. Air was allowed to pass through the filters for an additional 10 sec to remove residual moisture. The filter mat was supported on a cold (20°C) Teflon support, and filters from individual wells were separated and placed in Mini Poly-Q vials (Beckman Instruments Inc., Fullerton, CA) and filled with 4 ml of scintillation cocktail (Beckman Ready Protein⁺). Radioactivity retained on the filter was determined by liquid scintillation spectrophotometry. Nonspecific binding was defined as the binding in the presence of 100 µM (+)MK-801. In the presence of 10 µM glutamate, glycine and spermidine specific binding was 90%.

[³H]Ifenprodil binding. Binding assays were carried out in the presence of 2 nM [³H]ifenprodil as described for [³H]TCP binding except that the protein concentration was reduced by half, and the incubation time was increased to 2 hr to allow the binding to reach steady state. Ifenprodil (1 mM) was used to define the nonspecific binding.

Data Analysis

Binding curves were statistically analyzed for a best one- or two-site competition fit using GraphPad Prism software (GraphPad Software Inc., San Diego, CA). The normalized data were fit by nonweighted nonlinear regression to either or Control data were entered as 100%, and no parameters were constrained. Inhibition curves were compared by ANOVA with post-test comparisons of the log^IC₅₀ by using Dunnett's multiple-comparisons post-test or Student's nonpaired, two-tailed t test (GraphPad InStat software).

Materials. TCP, [piperidyl-3,4-³H(N)] (specific activity, 45-50 Ci/mmol) and ifenprodil, [phenyl-³H] (specific activity, 66.2 Ci/mmol) were purchased from Dupont NEN Research Products (Boston, MA). Ifenprodil tartrate, trifluperidol hydrochloride and GBR-12909 dihydrochloride were purchased from Research Biochemicals International (Natick, MA). Spermidine trihydrochloride was purchased from United States Biochemical Corp. (Cleveland, OH). HEPES, glutamate and glycine were purchased from Sigma Chemical Co. (St. Louis, MO). Haloperidol was obtained from McNeil Laboratories (Raritan, NJ) or Research Biochemicals International. Eliprodil was synthesized by Thomas Malone (Parke-Davis Pharmaceutical Research, Ann Arbor, MI), and (+)MK-801 was synthesized by Leonard Lescosky (Parke-Davis).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

[³H]TCP Binding

Our preliminary experiments suggested that the sites inhibited with high affinity by ifenprodil were most clearly distinguished by using [³H]TCP under nonequilibrium conditions. We therefore characterized the assay, using a 10-min incubation at room temperature, by examining the interactions of the coagonists glutamate and glycine with the polyamines spermine and spermidine and the channel blocker (+)MK-801. The addition of 10 µM glutamate, that of 10 µM glycine and that of 10 µM spermidine increased the binding of [³H]TCP to 249%, 257% and 209% of the basal binding, respectively (fig. 1A). The addition of glycine in combination with glutamate and that of spermidine in combination with glutamate enhanced the binding by 602% and 514%, respectively. Spermidine added in combination with glycine enhanced the binding by 401%, and spermidine in combination with glutamate and glycine increased the binding to 882% of basal. In the presence of 10 µM glutamate and glycine, spermine and spermidine exhibited biphasic concentration-response curves (fig. 1B). Both polyamines stimulated the binding between 1 and 30 µM and inhibited the binding at higher concentrations. Putrescine gave no enhancement (data not shown). Spermidine stimulated the binding to 187% of control and was more efficacious than spermine. Spermine was somewhat more potent than spermidine but stimulated the binding only to 159% of control. In the presence of 10 µM glutamate, glycine and spermidine, the potent channel blocking drug (+)MK-801 inhibited in a monophasic manner with an IC₅₀ value of 7.3 nM (fig. 1C).

View larger version (15K): [in this window] [in a new window]   Fig. 1.   The effects of glutamate, glycine, polyamines and (+)MK-801 on the binding of [3H]TCP. A) The enhancing effects of 10 µM glutamate, glycine and spermidine alone and in combination. The bars represent (mean dpm ± S.E.M.) from 10 separate experiments: A, basal (1143 ± 143); B, glutamate (2847 ± 265); C, glycine (2938 ± 239); D, glutamate and glycine (6885 ± 754); E, spermidine (2391 ± 292); F, spermidine and glutamate (5873 ± 477); G, spermidine and glycine (4579 ± 318); H, glutamate and glycine and spermidine (10,088 ± 1021). [3H]TCP binding was carried out as described in "Materials and Methods" for 10 min at room temperature. B) Concentration-effect curves for the polyamines spermine () and spermidine () on [3H]TCP binding. [3H]TCP binding was carried out as described in "Materials and Methods" for 10 min at room temperature in the presence of 10 µM glutamate and glycine. Values are the mean ± S.E.M. from three separate experiments. Specific binding in the absence of drug (dpm ± S.E.M.): spermine, 6824 ± 302; spermidine, 7208 ± 113. C) Inhibition of [3H]TCP binding by (+)MK-801. IC50 = 7.8 nM; Hill slope = 1.0. Binding was carried out in the presence of 10 µM glutamate, glycine and spermidine as described in "Materials and Methods" for 10 min at room temperature. Values are the mean ± S.E.M. from three separate experiments. Specific binding in the absence of drug (dpm ± S.E.M.) was 10,878 ± 777.

To examine the pharmacology of the NR2B-associated sites, which are inhibited by ifenprodil with high affinity, we compared the inhibition of [³H]TCP binding in the presence of 10 µM glutamate, glycine and spermidine by ifenprodil and eliprodil to that of haloperidol and trifluperidol (fig. 2). All of these agents inhibited the binding with two affinity states. The selectivity for the high-affinity sites over the low affinity-sites covaried with the potency (table 1). Ifenprodil was the most potent at the high-affinity sites (IC_50H = 0.093 µM) and the most selective for high-affinity binding (IC_50L/IC_50H > 1100). Haloperidol was the least potent (IC_50H = 1.9 µM) and the least selective (IC_50L/IC_50H = 68). Eliprodil (IC_50H = 0.79 µM) was about 8 times less potent than ifenprodil and a little more than 100-fold selective for the high-affinity sites.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Inhibition of [3H]TCP binding by ifenprodil (), trifluperidol (), eliprodil (), and haloperidol () in adult rat forebrain membranes. Binding was carried out as described in "Materials and Methods" in the presence of 10 µM glutamate, glycine and spermidine. The values are the mean ± S.E.M. from at least three independent experiments. Specific binding in the absence of drug (dpm ± S.E.M.): ifenprodil, 9324 ± 201; trifluperidol, 9264 ± 956; eliprodil, 11368 ± 79; haloperidol, 9352 ± 418. Data for the individual curves are shown in table 1.

Almost all of the sites labeled with [¹²⁵I]MK-801 in the 3-day-old rat pup brain are inhibited with high affinity by ifenprodil; the low-affinity sites appear as the rat matures (Williams et al., 1993). To examine whether the butyrophenones exhibit a similar pharmacology, we studied the binding in the 3-day-old rat pup brain. In agreement with previous findings, ifenprodil inhibited 88% of the [³H]TCP binding sites in the rat pup brain with high affinity (fig. 3). Eliprodil, trifluperidol and haloperidol inhibition curves were best fit with one-site competition curves. The potencies of the agents in the rat pup brain were similar to those of the high-affinity sites in adult rat brain (table 1).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Inhibition of [3H]TCP binding by ifenprodil (panel A), eliprodil (panel B), trifluperidol (panel C) and haloperidol (panel D) in 3-day-old rat pup forebrain membranes. Binding was carried out as described in "Materials and Methods" in the presence of 10 µM glutamate, glycine and spermidine. The values are the mean ± S.E.M. from three independent experiments. Specific binding in the absence of drug (dpm ± S.E.M.): ifenprodil, 1871 ± 277; eliprodil, 1428 ± 298; trifluperidol, 1400 ± 270; haloperidol, 1234 ± 141. Data for the individual curves are shown in table 1.

[³H]Ifenprodil Binding

Ifenprodil is known to bind with high affinity to sigma sites as well as to NRs, and the NMDA portion of ifenprodil binding is displaced by polyamines (Schoemaker et al., 1995). Figure 4 shows the inhibition of [³H]ifenprodil binding by spermidine and the dopamine uptake blocker and sigma ligand GBR-12909. GBR-12909 inhibited [³H]ifenprodil binding in a biphasic manner (fig. 4A). The inhibition appeared to be maximal at 40% of control; however, GBR-12909 did not remain in solution at concentrations greater than 10 µM. Spermidine inhibited [³H]ifenprodil binding in a biphasic manner, only 17% of the binding remaining at 10 mM (fig. 4B). We then compared the inhibition of ifenprodil, eliprodil and haloperidol in the absence and in the presence of either 3 µM GBR-12909 or 1 mM spermidine (fig. 5).

View larger version (11K): [in this window] [in a new window]   Fig. 4.   Inhibition of [3H]ifenprodil binding by GBR-12909 and spermidine. A) Inhibition by GBR-12909. IC50 values: high-affinity fraction (41%), 0.0004 µM; low-affinity fraction, 0.223 µM. Values are the mean ± S.E.M. from 3 to 5 independent experiments. Specific binding in the absence of drug [dpm ± S.E.M. (N)] was 16,567 ± 1748 dpm. B) Inhibition by spermidine. IC50 values: high-affinity fraction (49%) = 13 µM; low-affinity fraction = 3710 µM. Values are the mean ± S.E.M. from 3 or 4 independent experiments. Specific binding in the absence of drug [dpm ± S.E.M. (N)] was 16,709 ± 2340 dpm.

View larger version (14K): [in this window] [in a new window]   Fig. 5.   A) Inhibition of [3H]ifenprodil binding by ifenprodil (), eliprodil (), trifluperidol () and haloperidol (). B) Inhibition of [3H]ifenprodil binding by ifenprodil (), eliprodil () and haloperidol () in the presence of 3 µM GBR-12909. C) Inhibition of [3H]ifenprodil binding by ifenprodil (), eliprodil () and haloperidol () in the presence of 1 mM spermidine. Binding was carried out as described in "Materials and Methods." Specific binding in the absence of drug (dpm ± S.E.M.) from N independent experiments was as follows. A: ifenprodil, 20,094 ± 2120 (5); eliprodil, 17,275 ± 3109 (5); trifluperidol, 17,385 ± 2842 (5); haloperidol, 15,083 ± 2296 (5). B: ifenprodil, 11,135 ± 376 (3); eliprodil, 10,715 ± 658 (3); haloperidol, 9709 ± 1243 (3). C: ifenprodil, 6514 ± 545 (5); eliprodil, 9703 ± 2542 (3); haloperidol, 6959 ± 1308 (4). Values for the individual curves are shown in table 2.

In the absence of blocking agents, the potencies of the compounds to inhibit [³H]ifenprodil binding with high affinity correlated with their potencies to inhibit [³H]TCP binding and to inhibit glutamate-induced current at the NR1a/NR2B receptor expressed in Xenopus oocytes (fig. 5A and table 1). The high-affinity fraction of the binding was 64% to 74% of control. In the presence of GBR-12909, there was no significant change in the IC₅₀ values for either the high-affinity or the low-affinity fraction of the binding for any agent (fig. 5B). The fraction of the binding inhibited with high affinity was modestly reduced by 4%, 8% and 13% for ifenprodil, eliprodil and haloperidol, respectively (table 2). In the presence of 1 mM spermidine, the potency of haloperidol to inhibit the high-affinity fraction of the binding was increased 2.8-fold (fig. 5C). There was no effect on the low-affinity fraction of the binding of haloperidol or on the potencies of ifenprodil and eliprodil for either fraction. A marked reduction of 32%, 28% and 22% in the fraction of high-affinity sites was observed for ifenprodil, eliprodil and haloperidol, respectively, in the presence of spermidine.

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

In the present study, we have confirmed previous investigations (Nicolas and Carter, 1994; Ilyin et al., 1995; Williams et al., 1993) that suggest that an ifenprodil-sensitive, high-affinity, inhibitory site of the binding of radiolabeled open-channel blockers reflects inhibition at isoforms of the NR that contain the NR2B subunit, and we have extended the pharmacology of this site to include the butyrophenones haloperidol and trifluperidol. The data supported the conclusion from electrophysiological studies (Ilyin et al., 1996) that haloperidol selectively inhibits the NR1a/NR2B subtype of the NR and that haloperidol and ifenprodil share overlapping binding sites.

Interactions of modulatory agents of the NR are complex, and the results of binding assays can vary depending on the conditions of the binding assay, the radioligand chosen and the particular method of receptor preparation. We verified that the [³H]TCP binding assay, carried out under our nonequilibrium conditions, exhibited the well-established characteristics for this type of assay (Reynolds and Miller, 1990; Wong and Kemp, 1991; Romano and Williams, 1994; Yoneda and Ogita, 1991). Using a 10-min incubation time at room temperature and a well-washed, frozen-thawed membrane preparation, we found that the [³H]TCP assay detected the previously reported high- and low-affinity states of ifenprodil, potent inhibition by the noncompetitive antagonist (+)MK-801 and the combined agonist effects of glutamate and glycine. It also detected the glycine-independent enhancement and biphasic concentration-response curves of polyamines. The assay could thus be used as a biochemical correlate of the functional state of the channel and to discriminate binding preferential to the NR2B subunit (Scatton et al., 1994; Reynolds and Miller, 1989; Williams et al., 1993).

Ifenprodil, eliprodil, trifluperidol and haloperidol all inhibited the binding in a biphasic manner. The fraction of the high-affinity sites for all the compounds was between 39% and 48%, and the IC₅₀ values to inhibit the high-affinity portion of the binding were correlated with their potency at the NR1a/NR2B subtype of the NR. These findings with ifenprodil are in agreement with earlier studies by Reynolds and Miller (1989) and those of Williams (1993), which describe a correlation between the high-affinity inhibition of [¹²⁵I]MK-801 binding to the NMDA channel by ifenprodil and functional blockade of the NR1a/NR2B subtype of the NMDA receptor expressed in oocytes. The potency and selectivity of ifenprodil for the high-affinity site were greater in this study. This is probably because of the differences in assay conditions and the radioligand used. The order-of-magnitude difference in potency between ifenprodil and haloperidol, which can be observed in both the binding data and the Xenopus oocyte data (table 1), would be consistent with the conclusion of the earlier studies that the high-affinity fraction of the binding is to receptors expressing the NR2B subunit.

Williams found that ifenprodil distinguishes primarily the high-affinity site in the rat pup brain and that the low-affinity site appears as the rat matures (Williams et al., 1993). The rat pup forebrain is thought to express predominantly NRs containing the NR2B subunit. This corresponds with the low-level expression of the NR2A receptor protein that has been reported in the young rat forebrain (Portera-Cailliau et al., 1996). To investigate further whether the butyrophenones can distinguish the NR2B subunit in native NRs in rat brain membranes, we examined these agents in the 3-day-old rat pup brain. Ifenprodil inhibited almost all of the sites (88%) with high affinity. Eliprodil, trifluperidol and haloperidol distinguished only one inhibitory site. The IC₅₀ values at this high-affinity fraction were similar to those in the adult brain. These results supported the conclusion that these butyrophenones are selective for receptor subtypes that contain the NR2B subunit of the NR.

To determine whether the selectivity of haloperidol and trifluperidol for the NR2B isoform of the receptor was mediated by the same site as ifenprodil, we compared all these agents in [³H]ifenprodil binding. Radiolabeled ifenprodil has been shown to label a number of different binding sites, and results differ somewhat, depending on the particular binding conditions employed (Scatton et al., 1994; Schoemaker et al., 1995). We therefore carried out [³H]ifenprodil binding by using the same well-washed membrane preparation, buffer and temperature conditions that were used in the [³H]TCP binding, except that the binding was allowed to reach steady state. The characteristics of the [³H]ifenprodil assay under these conditions were in substantial agreement with those of previous investigators (Dana et al., 1991; Hashimoto et al., 1994; Schoemaker et al., 1995). All the agents inhibited the binding in a biphasic manner, reaching a maximal inhibition of 84% of the control binding at a concentration of 1 mM. One mM ifenprodil was thus chosen to define the nonspecific binding. The IC₅₀ value of ifenprodil at the high-affinity sites was 10 nM, which is consistent with the suggestion that this fraction represents the fraction bound to the NR2B subunit (Nicolas and Carter, 1994; Scatton et al., 1994).

Ifenprodil, eliprodil and haloperidol are known to have high affinity for sigma sites (Contreras et al., 1990a; Karbon et al., 1990; Schoemaker et al., 1995). GBR-12909, a high-affinity ligand for both dopamine uptake and sigma sites (Andersen, 1987; Contreras et al., 1990b), has been used to mask the sigma site portion of [³H]ifenprodil binding and to occlude a low-affinity piperazine acceptor site (Beart et al., 1992; Dana et al., 1991; Hashimoto et al., 1994; Schoemaker et al., 1995). We studied the inhibition of ifenprodil, eliprodil and haloperidol in [³H]ifenprodil binding in the presence of 3 µM GBR-12909 to determine the effect of masking these sites. In agreement with others (Hashimoto et al., 1994), [³H]ifenprodil binding under our conditions was inhibited by GBR-12909 with two components. The high-affinity fraction accounted for 41% of the binding and was inhibited with an IC₅₀ value of 0.4 nM. In the presence of 3 µM GBR-12909, there was no significant change in the IC₅₀ values of any agent at either the low- or the high-affinity site, although a small decrease in the fraction of sites inhibited with high affinity occurred. Haloperidol is the most potent of these agents at sigma sites when studied in binding assays using a radiolabeled sigma site-selective ligand (Schoemaker et al., 1995)¹, but it was much less potent than ifenprodil in [³H]ifenprodil binding in the presence and absence of GBR-12909. Others have found that the temperature of the incubation profoundly influences the binding of [³H]ifenprodil to sigma sites (Hashimoto et al., 1994). Under our binding conditions, GBR 12909 inhibited a substantial portion of the binding sites but did not change the affinity of ifenprodil, eliprodil or haloperidol for either fraction of the binding. GBR-12909 did not block glutamate-induced current at the NR1a/NR2B receptor expressed in oocytes at concentrations up to 10 µM.² This suggests that the high-affinity inhibition of these agents is not due to inhibition of [³H]ifenprodil binding to sigma sites. This result is in contrast to some findings (Hashimoto et al., 1994) but in agreement with others (Mercer et al., 1993).

Ifenprodil has been characterized as a polyamine antagonist, and polyamines have been shown to displace the NMDA portion of radiolabeled ifenprodil binding (Schoemaker et al., 1995). In agreement with earlier reports, we found that spermidine could displace [³H]ifenprodil binding in a biphasic manner, only 17% of the binding remaining in the presence of 10 mM spermidine. In the presence of 1 mM spermidine, which reduced the binding more than 50%, we found that the most marked effect was a noncompetitive inhibition of the fraction of the sites inhibited with high affinity by ifenprodil, eliprodil and haloperidol. The reduction in the high-affinity fraction ranged between 22% and 32%. The only effect on the affinity of the agents for either the high- or the low-affinity site was to decrease the IC₅₀ value of haloperidol at the high-affinity fraction 2.8-fold.

The reduction of the high-affinity fraction of [³H]ifenprodil binding by spermidine could be the result of reducing the maximal number of the high-affinity binding sites by an allosteric interaction with ifenprodil's high-affinity binding site on receptors expressing the NR2B subunit. Alternatively, it is possible that in addition to its voltage-independent NR2B-associated binding site, ifenprodil is binding at a separate, lower-affinity, voltage-dependent site near or in the channel pore at NRs expressing either NR2A or NR2B subunits (Williams, 1994). Spermidine has been shown to act as an open-channel blocker (Rock and MacDonald, 1995). Spermidine could be blocking ifenprodil's binding at this lower-affinity channel-associated site in a competitive or a noncompetitive manner. The data do not support a competitive interaction of spermidine and ifenprodil at the high-affinity binding site.

In summary, our experiments showed that the butyrophenones trifluperidol and haloperidol interact with [³H]TCP binding and [³H]ifenprodil binding to NRs in native rat brain membranes in a manner similar to ifenprodil and eliprodil. The polyamine spermidine decreased the fraction of high-affinity sites detected by ifenprodil, eliprodil and haloperidol in [³H]ifenprodil binding. The interaction of spermidine appeared noncompetitive in nature, and spermidine caused a decrease in the IC₅₀ value of haloperidol at the high-affinity fraction of the binding. Other studies, which used electrophysiological techniques and radiolabeled NMDA channel blockers (Ilyin et al., 1996; Williams, 1993; Williams et al., 1993) have shown that ifenprodil and haloperidol selectively interact with the NR1a/NR2B subtype of the NR. Nicolas and Carter (1994) also have reported a polyamine-sensitive, high-affinity [³H]ifenprodil binding site in rat forebrain whose distribution matched that of NR2B mRNA. The high-affinity inhibition defined by these agents in the present study is probably inhibition of binding to rat forebrain NRs that express the NR2B subunit.