Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Native NMDA receptors are composed of an NR1 subunit (can occur as eight different splice variants), which combines with NR2A-D subunits to form heteromeric receptors (for reviews see McBain and Mayer, 1994; Kemp and Kew, 1998). NMDA receptors are thought to be tetramers that are composed of two NR1 and two NR2 subunits (Laube et al., 1998), which is compatible with earlier electrophysiological evidence demonstrating that NMDA receptor activation requires occupation of two independent glycine binding sites and two independent glutamate sites (Benveniste and Mayer, 1991; Clements and Westbrook, 1991). The glutamate and glycine binding sites are believed to be located on the NR2 and NR1 subunits, respectively (Wafford et al., 1995; Hirai et al., 1996; Laube et al., 1997; Anson et al., 1998). However, the proposed tetrameric stoichiometry of NMDA receptor remains controversial because evidence also exists for a pentameric structure (for review see Kemp and Kew, 1998).

In the adult rodent and human brain the NR1 subunit is widely distributed throughout the brain, whereas the NR2 subunits are expressed in a distinct spatio-temporal manner (Watanabe et al., 1993; Monyer et al., 1994; Rigby et al., 1996; Wenzel et al., 1997). The predominant NR2 subunits in the forebrain are NR2A and NR2B, with NR2C expressed largely in the cerebellum and various select nuclei, and NR2D expression is confined to the diencephalon and midbrain (for reviews see McBain and Mayer, 1994; Gill et al., 1999). Coexpression of NR1 with one or more of the NR2 subunits yields receptors with distinct functional and pharmacological properties that appear to resemble those of native receptors (Priestley et al., 1995; Meddows et al., 2001).

Many compounds have been described that can interact with the NMDA receptor at distinct sites, such as the ion-channel site, the glutamate transmitter recognition site, and the glycine-modulatory site (see Kemp and Kew; 1998). Ligands acting at the glutamate or at the glycine binding site show low selectivity between the different subtypes of receptors (Priestley et al., 1995, Grimwood et al., 1996).

Over the last decade much attention has been given to the role of NMDA receptors in neurodegeneration following acute ischemic and traumatic brain injury (for review see Gill et al., 1999). The marked neurotoxic potential of the NMDA receptor appears to result from its high permeability to calcium, a known mediator of cell damage, its high affinity for glutamate, and its relative lack of desensitization during prolonged activation. The nonselective NMDA antagonists have been studied extensively in the last 10 years both in animal models and in the clinical setting (Muir and Lees, 1995; Lees, 1996; for reviews see Gill et al., 1999). In animal models these compounds have been investigated for possible application in neurodegenerative conditions such as stroke, traumatic brain injury, Parkinson's disease, and epilepsy, and for the treatment of acute pain. Nonselective NMDA antagonists showed robust neuroprotective effects in animal models of stroke. However, the use of nonselective NMDA receptor antagonists showed that blockade of NMDA receptors, besides its potential beneficial effects, produced profound CNS side-effects that have limited their therapeutic utility to date (Muir and Lees, 1995). In humans, these side-effects range from light-headedness, dizziness, paresthesia, and agitation at low doses, through nystagmus, hallucinations, somnolence, and blood pressure increases at moderate doses, to catatonia and "dissociative anesthesia" at high doses (Muir and Lees, 1995). These side-effects have limited the doses that could be administered in clinical trials, and as a consequence, plasma drug concentrations that are fully neuroprotective in animal models have been difficult to achieve in humans.

To try to eliminate these side-effects we have targeted the NR2B subunit of the NMDA receptor to develop selective compounds. Antagonists that are selective for the NMDA receptors containing NR2B subunits, such as ifenprodil, CP 101,606, and Ro 25-6981, are able to inhibit 50 to 60% of the ³H-MK-801 binding to all of the NMDA receptors in adult rat forebrain (Fischer et al., 1997; Mutel et al., 1998). Furthermore, we have found that these antagonists bind with a higher affinity to activated and desensitized states of the receptor relative to the unliganded resting state and, thus, display an activity- or state-dependent mode of action (Kew et al., 1996; Fischer et al., 1997, 1998). It is predicted that such compounds will preferentially block NMDA receptors, which are continuously activated by sustained high glutamate levels in ischemic brain areas, while leaving those physiologically activated in normal brain areas relatively unaffected.

Here we report the in vitro and in vivo profile of a novel, highly selective and potent blocker of NMDA receptors containing the NR2B subunit, 1-[2-(4-hydroxy-phenoxy)-ethyl]-4-(4-methyl-benzyl)-piperidin-4-ol (Ro 63-1908; Alanine et al., 1999; Fig. 1). This compound has also been described by Zhou et al. (1999).

View larger version (9K): [in this window] [in a new window]   Fig. 1.   Structure of Ro 63-1908 [1-[2-(4-hydroxy-phenoxy)ethyl]-4-(4-methylbenzyl)piperidin-4-ol].

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Materials

Drugs used in binding, electrophysiological, and toxicity experiments were purchased from the following sources: D-2-amino-5-phosphonopentanoic acid (Tocris Cookson, Inc., Ballwin, MO), flunarizine (Sigma/RBI, Natick, MA), ifenprodil (Sigma/RBI), MK-801(dizocilpine; Sigma/RBI), 2,3-dihydroxy-6-nitro-7-sulfamoylbenzo(f)quinoxaline (Tocris Cookson), [1-(2-thienyl)cyclohexyl] piperidine (TCP) (Sigma/RBI), tetrodotoxin (Latoxan, Valence, France), and veratridine (Sigma/RBI). Ro 25-6981 and Ro 63-1908 were synthesized in the Chemistry Department of Hoffmann-La Roche. All other chemicals were obtained from Sigma-Aldrich (St. Louis, MO) or Fluka (Buchs, Switzerland).

Binding Experiments

Binding experiments with rat forebrain membranes and the radioligand ³H-MK-801 (dizocilpine) were performed as described previously by Ransom and Stec (1988) and Fischer et al. (1997). Briefly, membranes were prepared from whole brain of male rats (150-200 g) by homogenization and centrifugation (48,000g for 10 min). For binding experiments, the membranes were washed three times, and the final pellet was re-homogenized in 25 volumes of buffer (original wet weight) and used as such in the assay. The final protein concentration in the assay was 200 µg/ml. The incubation was performed in the presence of 1 nM added glutamate, glycine, and spermidine. The ligand, ³H-MK-801, (+)-[3-3H(N)] (PerkinElmer Life Sciences, Boston, MA), 20 Ci/mmol, was used at 5 nM final concentration. Nonspecific binding was determined in the presence of 100 µM TCP. After 2 h of incubation at room temperature, the suspension was filtered (Whatman GF/B, soaked in 0.1% polyethylenimine for 2 h) and washed five times with 3 ml of cold 5 mM Tris-HCl (pH 7.4). The filters were counted with 10 ml of Ultima-gold (Packard BioScience, Meriden, CT) in a Tri-Carb 2500 TR scintillation counter after agitation. The dpm values were transformed to percentage of specific binding. Each experiment was repeated at least three times.

Binding with ³H-Ro 25-6981 to rat tissue was performed as previously described by Mutel et al. (1998). Briefly, rat brain membranes, prepared as above, were washed three times in a Tris-HCl (5 mM, pH 7.4) cold binding buffer. The final pellet was resuspended in the same buffer and used at a final concentration of 200 µg of protein/ml. ³H-Ro 25-6981 (5 nM), specific activity, 20.7 Ci/mmol, was used, and the nonspecific binding was determined using 10 µM 1-(4-chlorophenyl)-2-methyl-6-methoxy-7-hydroxy-1,2,3,4-tetrahydroisoquinolin (Ro 04-5595). The incubation time was 2 h at 4°C. The assay was stopped by filtration on Whatman GF/B glass fiber filters (Unifilter 96, Canberra Packard S.A., Zurich, Switzerland), and the radioactivity on the filter was measured as above. Each experiment was repeated at least three times.

The analysis of the binding assays was done by fitting the function below to concentration-response relations from individual experiments:
with y, labeled receptor; P_low, proportion of low-affinity binding sites; K_dlow, affinity of low-affinity binding sites; and K_dhigh, affinity of high-affinity binding sites.

Electrophysiology

cDNA clones coding for the subunits NR1C (also termed NMDAR1-2a or NR1₀₀₁), NR2A, and NR2D of the NMDA receptor were isolated from a rat gt11 cDNA library (see Fischer et al., 1997). The clone for the subunit NR2B of the rat NMDA receptor was kindly provided by Dr. S. Nakanishi (Kyoto, Japan). The cDNA for NR1F (NMDAR1-2b or NR1₁₀₁) was derived from NR1C by molecular engineering (Fischer et al., 1997). The methods for subcloning, nuclear injection into Xenopus oocytes, and current recording from the oocytes under voltage clamp were described previously (Fischer et al., 1997; Kew et al., 1998). Current responses were evoked once every 2.5 min by applying approximately half-maximal concentrations of the NMDA receptor coagonists L-glutamate (0.5-2.7 µM) and glycine (0.07-0.4 µM). After an initial series of control responses, Ro 63-1908 (0.01-10 µM) was added to the basal saline and the agonist containing solutions superfusing the oocyte. At the lower concentrations the compound was applied up to 30 min for approaching equilibrium inhibition.

Measurements of intracellular free Na⁺ and Ca²⁺ concentrations were performed using imaging methods as described by Fischer et al. (1997). To calculate the mean values from these experiments, at least three independent experiments with triplicates were performed.

Cortical Neurons. Cells were used after 10 to 14 days in vitro (see Fischer et al., 1997). Whole-cell voltage-clamp recordings were performed as described (Kew et al., 1996). Briefly, cells were continuously superfused at room temperature with saline containing 149 mM NaCl, 3.25 mM KCl, 2 mM CaCl₂, 2 mM MgCl₂, 10 mM HEPES, 11 mM D-glucose, pH 7.35, with an osmolarity of 350 mOsm adjusted with sucrose. Patch pipettes had a resistance of approximately 2 to 4 M when filled with a solution containing 120 mM CsF, 10 mM CsCl, 11 mM EGTA, 0.5 mM CaCl₂, 10 mM HEPES, pH 7.25, with osmolarity adjusted to 330 mOsm with sucrose. Whole cell recordings were made at a holding potential of 60 mV using an Axopatch 200A amplifier (Axon Instruments, Foster City, CA). Pipette seal resistances were typically >10 gigaohms, and pipette capacitance transients were minimized both before and after membrane breakthrough. No series resistance compensation was applied. Drugs were diluted from concentrated stock into a modified version of the superfusion solution lacking MgCl₂ and including 30 µM glycine. Drugs were applied to cells by fast superfusion from double- or triple-barreled capillary assemblies (internal capillary diameter 320 µm). Currents were filtered (cut-off frequency 5 kHz), digitized using a Digidata 1200 digitizer (Axon Instruments), and captured on-line using pCLAMP6 software (Axon Instruments).

In Vitro Toxicity Experiments

Cortical neuronal cultures and subsequent toxicity experiments [glutamate, kainate, and oxygen-glucose deprivation (OGD)] were performed as described by Fischer et al. (1997).

In Vivo Characterization

Pharmacokinetic Analysis of Ro 63-1908. Pharmacokinetic studies with Ro 63-1908 were performed in rats, dogs, mice, and cynomolgus monkeys following intravenous administration of bolus doses in the range of 2 to 10 mg/kg. Systemic clearance values were estimated under steady-state conditions, following continuous intravenous infusions in rats (7.5-15 mg/kg/h) and in cynomolgus monkeys (1.5-5 mg/kg/h). The plasma levels of Ro 63-1908 were measured using the high-performance liquid chromatography column-switching method with fluorescence detection (Wyss et al., 2000). Brain penetration of Ro 63-1908 was also examined, with cerebrospinal fluid and brain levels being analyzed using the same high-performance liquid chromatography method.

Audiogenic Seizures. Male DBA/2J mice (21 days old) were exposed to a 14-kHz sinusoidal tone at 110 dB for 60 s, which induced wild running progressing to tonic extensions of the limbs. Groups of eight mice were treated with either Ro 63-1908 (1.87, 3.75, 7.5, 15, and 30 mg/kg i.p.) or MK-801 (0.03 to 1 mg/kg i.p.) 30 min before exposure to the sound. Animals not exhibiting a tonic seizure within the 60 s of sound exposure were considered protected.

Seizures Induced by NMDA. NMDA (0.15 µg/2 µl) was injected i.c.v. to MORO mice (18-20 g), and the animals were observed for 5 min. With this treatment, mice exhibit wild running, clonic-tonic seizures (75% of the mice), and death (40% of the mice). Ro 63-1908 (0.3, 1, 3, 10, and 30 mg/kg i.v.) or MK-801 (0.01-1 mg/kg,i.v.) were injected to groups of 8 mice 15 min before administration of NMDA. Animals not exhibiting clonic and tonic seizures were considered protected.

Motor Coordination. Mice (DBA/2J and MORO mice) were first trained to remain for 2 min on a revolving Rotarod apparatus (accelerating units increase from 3.5 to 35 rpm in 5 min). Ro 63-1908 and MK-801 were given i.v. or i.p. 15 or 30 min before the test. The latency time to fall off the Rotarod was determined (cut-off time used was 2 min). The same groups of mice were then used in the seizure tests. Results in the dose-response graph are given as the mean latency ± S.E.M. in percentage of the control group. All experiments were done on a "blind" basis.

Locomotor Activity

The computerized Digiscan 16 Animal Activity Monitoring System (Omnitech, Columbus, OH) was used to quantify motor activity in male Sprague-Dawley rats. Immediately after the administration of Ro 63-1908 (1, 3, 10, and 30 mg/kg i.v.), rats were placed into the activity monitor for 4 h, and different parameters were measured, such as horizontal activity (total number of interruptions of the horizontal sensors during a given period). The mean from eight different animals was determined.

Middle Cerebral Artery Occlusion

Permanent focal cerebral ischemia was induced in adult (10-11weeks) male Fischer 344/Ico old SPF rats (Iffa Credo, L'Arbresele, France), weighing between 220 and 230 g. The animals were food-deprived for approximately 12 h before surgery. The left femoral artery and vein were cannulated to enable monitoring of blood pressure, blood gases, and continuous infusion of the compound. The left middle cerebral artery (MCA) was permanently occluded using bipolar coagulation in anesthetized (3% isoflurane in 70% air and 30% oxygen) animals as described by Shiraishi and Simon (1989). Following surgery the animals were allowed to recover from the anesthesia and placed in a box with a swivel device to enable continuous infusion of the vehicle or test compound. Ro 63-1908 or vehicle was administered within 5 to 7 min after MCA occlusion as a bolus dose over 2 min, followed by an infusion over 5 h. A dose-response relationship for Ro 63-1908 was determined; bolus doses of 0.14, 0.7, 2.8, and 5.6 mg/kg followed by infusions of 0.28, 1.4, 5.6, and 11.2 mg/kg/h for 5 h were administered as described above (n = 9-14 per group). The bolus + infusion dosing regimen is akin to what may be used for neuroprotective agents in patients.

Measurement of Infarct Volume. The animals were deeply anesthetized with isoflurane 48 h after MCA occlusion and decapitated. Brains were immersion-fixed in 3.8% buffered formaldehyde for 5 days and automatically processed for embedding in paraffin (Bavimed 2050Z2; Haska AG, Bern, Switzerland) through a series of ethanol baths (70-100%) into xylol and paraffin over 21 h. Serial coronal sections (10 µm thick, three brains per block) were cut at 1-mm intervals through the forebrain and stained with toluidine blue. The boundaries of the infarcted cerebral cortex and neostriatum were marked by an unbiased observer and traced from a monitor (at 21× magnification) onto a digitizing tablet interfaced to an image analysis system (DiaSys Datalab; H. Meyer, Thöringen, Switzerland). The infarct volumes were calculated by numeric integration of the infarct areas on sequential slices without correction for edema. Statistical significance was calculated with ANOVA (BMDP; Statistical Solutions, Cork, Ireland), followed by a Bonferroni correction.

Investigation of the Effect of Ro 63-1908 on Vacuole Formation

A study was performed to assess the effects of Ro 63-1908 on vacuole formation in the cingulate cortex, since nonselective antagonists have been shown to induce vacuoles within neurons in certain vulnerable cortical brain regions. Animals were administered either saline i.p.(negative control), MK-801, 5 mg/kg i.p (positive control), or Ro 63-1908, 30 mg/kg/h for 5 h (n = 4 per group).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Binding Studies. In adult rat brain membranes, Ro 63-1908 (Fig. 1) inhibited ³H-MK-801 binding in a markedly biphasic manner (Fig. 2) with IC₅₀ values for the high- and low-affinity components of 0.002 and 97 µM, respectively. Using ³H-Ro 25-6981 to label NR2B subunit-containing receptors selectively in the adult forebrain of rat, Ro 63-1908 inhibited with a single high-affinity component and with an IC₅₀ of 10 nM. Ro 63-1908 had only a weak affinity for ₁-adrenergic receptors (Table 1). In an additional 38 radioligand binding assays, Ro 63-1908 was without activity (extrapolated IC₅₀ values >1 µM) with the exception of affinities for dopamine receptors (K_i of 0.2, 0.7, and 3.9 µM for huD₃, huD₄, and huD₂, respectively) and  binding sites (about 80% block at 0.1 µM). In a functional test of dopamine-stimulated guanosine 5'-O-(3-thio)triphosphate binding, Ro 63-1908 acted as a weak antagonist at huD₂ and huD₄ receptors with IC₅₀ values approximately 10 times lower than the K_i values (data not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Inhibition of 3H-MK-801 binding to rat forebrain membranes by Ro 63-1908. A biphasic function (see Materials and Methods) was fitted to the data to describe the concentration dependence of inhibition. Values represent means (±S.D.) from triplicates.

Electrophysiological Characterization of Ro 63-1908. In Xenopus oocytes expressing recombinant rat NMDA receptor subtypes (NR1C + NR2B, NR1F + NR2B, NR1C + NR2A, or NR1C + NR2D), Ro 63-1908 blocked only currents of NR2B-containing receptors with high affinity (Table 1). The presence of the NR1-subunit N-terminal insert (exon 5, present in NR1F, but missing in NR1C) increased the IC₅₀ about 2-fold. A similar small influence of exon 5 has been seen previously with Ro 25-6981 and ifenprodil (Fischer et al., 1997). Thus, Ro 63-1908 shows a high specificity for NR2B-containing NMDA receptor subtypes. In comparison, CP 101,606 was 10 times less potent at blocking NR2B-subunit receptors (NR1C + NR2B) than Ro 63-1908 (IC₅₀ values of 30 and 3 nM, respectively).

View larger version (18K): [in this window] [in a new window]   Fig. 3.   Activity-dependent action of Ro 63-1908 on NMDA receptors. NMDA receptors were stimulated in 6- to 14-day-old cultured cortical neurons in the presence of a saturating (30 µM) glycine concentration. Currents were measured by whole-cell voltage-clamp recordings. A, the effect of Ro 63-1908 depended on the level of receptor activation such that small inward currents elicited by 1 µM NMDA were potentiated in the presence of 1 µM Ro 63-1908 (steady-state current = 286 ± 21% of pre-Ro 63-1908 control, n = 5), whereas much larger currents elicited by 100 µM NMDA were markedly inhibited (steady-state current = 26 ± 2%

Neuroprotection Studies of Ro 63-1908 in Vitro. Ro 63-1908 showed concentration-dependent protection against neurotoxicity induced by either exposure of cultured rat cortical neurons to glutamate (300 µM, 16 h) or by combined deprivation of oxygen and glucose (OGD, 60 min; Fig. 4). Both neurotoxicity models mimic aspects of excitotoxicity in ischemic brain. Neurotoxicity in these tests is mainly mediated by activation of NMDA receptors. The relatively shorter period of the toxic insult in the OGD model [IC₅₀ of 7 ng/ml (0.02 µM)] might explain the higher potency of Ro 63-1908 in comparison with the glutamate toxicity model [IC₅₀ of 68 ng/ml (0.2 µM]. Ro 63-1908 was not protective against exposure to kainate (Table 1) where toxicity is mediated by activation of AMPA receptors.

View larger version (15K): [in this window] [in a new window]   Fig. 4.   Protection against glutamate toxicity in cultured cortical neurons. The concentration-dependent protection by Ro 63-1908 against glutamate or OGD-induced toxicity is shown. After 12 days, in vitro cultured cortical neurons were exposed to glutamate (300 µM) and glycine (1 µM) or to OGD. Neuronal cell death was assessed qualitatively by phase-contrast microscopy and quantified by determination of lactate dehydrogenase activity in the cell culture supernatant. The difference between control cultures and glutamate-exposed cultures was taken as 100% lactate dehydrogenase release. At least three independent experiments were performed in quadruplicate. Arithmetic means of the quadruplicates are shown for each experiment.

Pharmacokinetic Properties of Ro 63-1908. Ro 63-1908 penetrated rapidly into the brain; the cerebrospinal fluid/plasma concentration ratio was about 0.3 at 5 min after dosing, showing a rapid equilibration with the free fraction in plasma (see Tables 2 and 3).

Anticonvulsant Properties of Ro 63-1908. Ro 63-1908 was active against sound-induced seizures in DBA/2 mice with an ED₅₀ of 4.50 (2.86-6.61) mg/kg i.p., when administered 30 min before the test. The dose required to produce a full anticonvulsant effect was not associated with a Rotarod deficit (Fig. 5A). At 100 mg/kg i.p., only 10% motor deficit was obtained (data not shown). In MORO mice, Ro 63-1908 dose dependently antagonized seizures induced by i.c.v. administration of NMDA with an ED₅₀ of 2.31 (1.12-4.60) mg/kg i.v. when given 15 min before NMDA. There was no Rotarod deficit detected up to 10 mg/kg i.v. in these animals. At 30 mg/kg i.v., there was a 45% motor deficit on the Rotarod. MK-801 also gave a potent anticonvulsant effect against sound-induced seizures [ED₅₀ of 0.06 (0.005-0.11) mg/kg i.p.] in DBA/2 mice and against NMDA-induced seizures (ED₅₀ = 0.17 mg/kg i.v.) when given 15 min before NMDA. However, there was no separation for MK-801 between the anticonvulsant effect and motor impairment on the Rotarod test, with an ED₅₀ = 0.07 mg/kg i.v.

View larger version (22K): [in this window] [in a new window]   Fig. 5.   Anticonvulsant and locomotor activity of Ro 63-1908. A, effects of Ro 63-1908 against sound-induced seizures and on Rotarod performance in DBA/2J mice. Ro 63-1908 (1-30 mg/kg) was administered i.p., and 25 min after dosing, the animals were tested on the Rotarod, the latency to fall off the Rotarod was measured, and the mice were then exposed to sound. The results are given as the mean latency (percentage of the control group) to fall off the Rotarod and as the percentage of animals demonstrating anticonvulsant activity. Groups of eight mice were used. B, effects of Ro 63-1908 on locomotor activity following acute i.v. administration were measured. Each point represents the mean horizontal activity measured for n = 8 rats.

Locomotor Activity Profile of Ro 63-1908 in Rats. Ro 63-1908 produced an increase of horizontal activity, as shown in Fig. 5B, at doses of 1, 3, and 10 mg/kg i.v. This effect started immediately after the administration of the compound and lasted for 120 min. At 30 mg/kg, an initial decrease of the horizontal activity was observed, which was followed by an increase of the activity (Fig. 5B). The rapid onset of these effects after i.v. injection is indicative of a fast penetration of Ro 63-1908 into brain. In contrast to results with nonselective NMDA receptor open-channel blockers, such as MK-801, rats administered Ro 63-1908 did not exhibit stereotypies such as head weaving and turning behavior.

Neuroprotective Effects in Vivo of Ro 63-1908. Ro 63-1908 administered after cerebral ischemia produced a dose-related decrease in the volume of cortical damage with maximum protection of 42% being seen after a dosing regimen of 5.6 mg/kg bolus followed by an infusion of 11 mg/kg/h. In a second experiment, a dosing regimen of 2.8 + 5.6 mg/kg/h gave similar protection of 39%, suggesting that the extent of protection achieved reached a plateau level (Fig. 6A). There was no protection seen in the caudate nucleus because the lenticulostriate artery, which is a branch of the MCA serving the caudate nucleus, is permanently occluded in this model of focal ischemia.

View larger version (40K): [in this window] [in a new window]   Fig. 6.   Neuroprotective effects of Ro 63-1908 in a rat model of focal ischemia. The left MCA was permanently occluded using bipolar coagulation, and Ro 63-1908 was administered 5 min after the occlusion using various dosing regimens. Ro 63-1908 was administered as an i.v. bolus over 2 min followed by an infusion over 5 h. A, there was a dose-related significant neuroprotective effect of Ro 63-1908 against cortical damage (ANOVA, F = 24.46; p = 0.0001) but not against caudate damage (ANOVA, F = 1.22; P > 0.3). The doses of 0.7 mg/kg + 1.4 mg/kg/h and 5.6 mg/kg + 11.2 mg/kg/h gave a significant neuroprotective effect (*, P < 0.01, Bonferroni correction). B, there was a significant neuroprotective effect of Ro 63-1908 against cortical damage at both doses tested (ANOVA, F = 37.23; p = 0.0001); once again, no significant protection was seen against caudate damage. Ro 25-6981 at a dose of 1.25 mg/kg + 2.92 mg/kg/h for 5 h also gave a significant neuroprotective effect. Each bar represents the mean ± S.E.M. for 9-14 animals per group; *, P < 0.01, Bonferroni correction.

Effect of Ro 63-1908 on Vacuoles in the Cingulate Cortex. No vacuoles were seen in any of the sections of cingulate cortex from saline-treated animals, but vacuoles were found in cingulate cortex neurons in every MK-801-treated animal. Vacuoles were not found within cingulate cortex neurons of any animal treated with Ro 63-1908, even though plasma levels far in excess of maximum neuroprotective levels were reached, i.e., between 1650 and 1900 ng/ml.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

Ro 63-1908 is a high-affinity, subtype-selective NMDA receptor antagonist that only blocks NMDA receptors containing the NR2B subunit. The selectivity for NR2B versus NR2A subunits is >20,000-fold. Furthermore, Ro 63-1908 shows a high specificity for NMDA receptor subtypes over other types of receptors and ion channels apart from dopamine receptors and  binding sites. The physiological significance of activities at dopamine receptors and  binding sites is presently unclear but seems unlikely to represent serious concerns for side-effects. Ro 63-1908 was a potent neuroprotectant in excitotoxicity models in vitro that mimic pathophysiological conditions occurring in stroke and traumatic brain injury. Furthermore, Ro 63-1908 produced a dose-dependent neuroprotection of the ischemic infarct in an in vivo rat model of stroke, namely, permanent occlusion of the middle cerebral artery. Maximal protection was reached with plasma levels around 450 ng/ml, and the protection achieved was comparable with that reached with nonselective NMDA receptor blockers such as MK-801 or 3-(2-carboxypiperazine-4-yl)-1-propenyl-1-phosphonic acid (for review see Gill et al., 1999). The activity-dependent mode of action of Ro 63-1908 favors the block of persistently activated NMDA receptors that occurs in ischemic brain areas following stroke and traumatic brain injury (Kew et al., 1996; Gill et al., 1999). This activity dependence, together with subtype selectivity, results in an improved safety profile of Ro 63-1908 compared with nonselective NMDA antagonists such as MK-801 and aptiganel (for reviews see Kemp and Kew, 1999; McBurney, 1997). Thus, there were no significant effects on blood pressure or heart rate in conscious animals and no stereotypic behaviors or circling were seen. In separate behavioral studies in rats, the only observable behavior was increased locomotor activity after acute i.v. bolus injection of Ro 63-1908. An increase in locomotor activity occurred at doses of 1 and 3 mg/kg, whereas an initial decrease was followed by an increase at higher doses (10 and 30 mg/kg). The significance of the locomotor stimulant effect for human behavior is unclear, but the lack of stereotyped behaviors suggests that, like ifenprodil and CP 101,606, this compound may have a reduced psychotomimetic potential compared with nonselective antagonists. Ro 63-1908 prevented sound-induced seizures at doses that did not effect motor coordination as measured on the Rotarod. This is in contrast to the nonselective NMDA receptor antagonists such as MK-801 in which there is no separation between anticonvulsant effects and motor incoordination (Tricklebank et al., 1989). No proconvulsant activity was found with Ro 63-1908.

Ro 63-1908 did not induce morphological changes (vacuoles) in the retrosplenial and cingulate cortical neurons following infusion of 30 mg/kg/h of Ro 63-1908 for 5 h. This resulted in plasma concentrations 3-fold higher than the therapeutic level in the stroke model. All MK-801-treated animals, which served as a positive control, had vacuoles in the retrosplenial and cingulate cortex, as shown previously (Olney et al., 1989; Hargreaves et al., 1993). The neuroprotective doses of MK-801 are close to those producing vacuoles in the cingulate cortex (Gill et al., 1991; Hargreaves et al., 1993). Eliprodil, which is also an NR2B subtype-selective NMDA antagonist, did not produce vacuoles in the retrosplenial or cingulate cortex (Carter et al., 1997). Thus, it appears that, unlike nonselective NMDA antagonists, NR2B subtype-selective NMDA antagonists do not induce neuronal vacuolization in cortical neurons.

The nonselective NMDA antagonists show adverse CNS and cardiovascular side effects at maximum protective levels in this model of focal ischemia, and these effects have proven to be dose-limiting in the clinic (McBurney, 1997; Albers et al., 2001). In humans the dose-limiting side effects seen with the nonselective NMDA antagonists were hypertension, agitation, ataxia, sedation, psychosis, and hallucinations (for review see Muir and Lees, 1995; Lees, 1997).

The atypical NMDA antagonist ifenprodil was demonstrated to be neuroprotective in animal models without the adverse side effects of compounds like MK-801, aptiganel, and selfotel (Gotti et al., 1988). Ifenprodil was the first in this class of NMDA antagonists selective for the NR2B subunit of the NMDA receptor (Gotti et al., 1988; Williams, 1993) and was shown to have a novel activity-dependent mechanism of action (Kew et al., 1996, 1998). CP 101,606 is a structural analog of ifenprodil that is also a selective antagonist of NMDA NR2B subunits but, unlike ifenprodil, lacks ₁-adrenoceptor activity (Chenard et al., 1995; Menniti et al., 1997). CP 101,606 was demonstrated to protect against glutamate-induced toxicity in neuronal cultures with a potency similar to that of MK-801 (Menniti et al., 1997). It was also found to be neuroprotective in an acute model of focal ischemia in the cat (Di et al., 1997). In Phase I trials in patients, it did not show any of the cardiovascular or CNS side effects seen with the nonselective NMDA antagonists (Bullock et al., 1999; Merchant et al., 1999). This compound has been in Phase II clinical trials for brain trauma, although recent press reports indicate that clinical studies have now been discontinued due to lack of efficacy.

In conclusion, it appears that activity-dependent NR2B subtype-selective NMDA antagonists, such as Ro 63-1908, show potential for the treatment of stroke and brain trauma. They have advantages over nonselective competitive and ion channel-blocking NMDA antagonists in that they have a much improved tolerability in animals and also, based on initial reports, in humans (Di et al., 1997; Bullock et al., 1999). Therefore, it is conceivable that maximally neuroprotective doses of these compounds could be achieved in humans without cardiovascular or CNS side effects.