Abstract
The purification, characterization, and synthesis of conantokin-R (Con-R), an N-methyl-d-aspartate (NMDA) receptor peptide antagonist from the venom of Conus radiatus, are described. With the use of well defined animal seizure models, Con-R was found to possess an anticonvulsant profile superior to that of ifenprodil and dizocilpine (MK-801). With voltage-clamp recording of Xenopus oocytes expressing heteromeric NMDA receptors from cloned NR1 and NR2 subunit RNAs, Con-R exhibited the following order of preference for NR2 subunits: NR2B ≈ NR2A > NR2C ≫ NR2D. Con-R was without effect on oocytes expressing the α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) receptor subunit GluR1 or the kainate receptor subunit GluR6. In mouse cortical neurons voltage-clamped at −60 mV, Con-R application produced a slowly developing block of inward currents evoked by 10 μM NMDA and 1 μM glycine (IC50 = 350 nM). At 3 μM, Con-R did not affect γ-aminobutyric acid- or kainate-evoked currents. Con-R prevented sound-induced tonic extension seizures in the Frings audiogenic seizure-susceptible mice at i.c.v. doses below toxic levels. It was also effective at nontoxic doses in CF#1 mice against tonic extension seizures induced by threshold (15 mA) and maximal (50 mA) stimulation, and it partially blocked clonic seizures induced by s.c. pentylenetetrazol. In contrast, MK-801 and ifenprodil were effective only at doses approaching (audiogenic seizures) or exceeding (electrical and pentylenetetrazol seizures) those required to produce significant behavioral impairment. These results indicate that the subtype selectivity and other properties of Con-R afford a distinct advantage over the noncompetitive NMDA antagonists MK-801 and ifenprodil. Con-R is a useful new pharmacological agent for differentiation between the anticonvulsant and toxic effects of NMDA antagonists.
Epilepsy is characterized by chronic recurrent seizure activity that affects 1 to 3% of the population in the United States. In recent years, a greater understanding of the genetics and molecular biology underlying certain genetic and acquired seizure disorders has led to the identification of a number of molecular defects that contribute to the initiation and propagation of seizure activity. For example, alterations in voltage- and receptor-gated ion channels, neurotransmitter release, uptake, and receptor function have all been implicated in experimental and human epilepsy (Dichter, 1994; McNamara, 1994; Meldrum, 1995). Of the various defects that have been identified thus far, substantial evidence indicates that enhanced central nervous system (CNS) glutamatergic neurotransmission contributes to seizure activity (Meldrum, 1994; Brusa et al., 1995; Behr et al., 1998;Chapman, 1998; Mathern et al., 1998).
Once released from the presynaptic terminal, glutamate activates both ionotropic [α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA), N-methyl-d-aspartate (NMDA), and kainate receptors] and metabotropic glutamate receptors (mGluR1–8; Hollmann and Heinemann, 1994). The NMDA receptor is thought to be either tetrameric or pentameric in structure and composed of two major subunit families (i.e., NR1 and NR2). There are four subunits within the NR2 family (NR2A–D) and eight possible splice variants of NR1.
The NMDA receptor complex expresses several pharmacologically distinct modulatory sites at which receptor function can be modified (i.e., glutamate and glycine recognition sites, a polyamine site or sites, an open-channel blocking site, a Zn2+ site or sites, an H+ site, and a redox site or sites). Indeed, drugs that affect several of these molecular targets possess potent anticonvulsant activity in experimental animal seizure models. Unfortunately, results from a limited number of add-on clinical trials conducted in patients with therapy-resistant partial seizures have been rather disappointing (Meldrum, 1995). For example, trials withd(−)-(E)-4-(3-phosphonoprop-2-enyl)piperazine-2-carboxylic acid (d-CPPene, a competitive NMDA antagonist), and dextromethorphan and MK-801 (two noncompetitive NMDA antagonists) had to be discontinued due to the emergence of intolerable side effects (Troupin et al., 1986; Fisher et al., 1990; Sveinbjornsdottir et al., 1993). Because the pharmacology of the NMDA receptor is highly dependent on subunit composition and given the differential regional expression of the different NR2 subunits and NR1 splice variants in the brain, the possibility exists of developing subunit-selective NMDA antagonists that are devoid of the adverse effects that have plagued the less selective NMDA antagonists.
The conantokins represent a family of peptides isolated from the venom of predatory marine cone snails of the genus Conus. The cone snails' venom contains a diverse mixture of pharmacologically active structurally constrained peptides referred to as conotoxins (Olivera, 1997). Many of the biologically active peptide fractions have been found to be highly selective for a specific isoform of a receptor or ion channel. The cone snails likely use their venom for many purposes that may include prey capture, defense, and escape from predators (Olivera, 1997). Previous studies have demonstrated that conantokin-G (Con-G) isolated from the venom of Conus geographus is a potent and selective antagonist of NMDA-evoked currents (Hammerland et al., 1992) and NMDA-induced Ca2+ transients (Haack et al., 1993). In the present study, we used electrophysiological recording techniques to examine the ability of conantokin-R (Con-R) isolated from the venom of Conus radiatus to block NMDA receptors in native receptors in cultured cortical neurons and recombinant receptors expressed inXenopus oocytes. Con-R differs in structure from Con-G in several important characteristics. First, Con-R is 27 amino acids in length compared with the 17 amino acids found in Con-G. Second, it possesses only four of the five γ-carboxyglutamate (Gla) residues of Con-G, and last, only 4 of the 17 amino acids in conantokin-G are preserved in Con-R. In the present investigation, we demonstrate that Con-R is a potent NMDA antagonist with a unique NR2 subunit selectivity profile. Given this unique selectivity profile, we compared its anticonvulsant and toxicity profile with that of ifenprodil and MK-801. The results of this study demonstrate that Con-R is a potent anticonvulsant peptide in a battery of animal seizure models. Furthermore, the results suggest that the subunit-selectivity of Con-R may provide certain therapeutic advantages over less selective NMDA antagonists.
Materials and Methods
Purification of Con-R
Lyophilized C. radiatus venom (1.28 g) obtained from the Philippines was extracted with 1.1% acetic acid as previously described (McIntosh et al., 1982). The extract was fractionated by HPLC on a preparative column (500 × 21.5 mm) of molecular sieve Bio-Sil TSK 125 eluted with 1.1% acetic acid at a flow rate of 10 ml/min. The fractions that eluted from 18 to 24 min were pooled and dried in a Savant Speed Vac concentrator. The residue was taken up in a small volume of 0.1% trifluoroacetic acid (TFA) and then fractionated on a semipreparative Vydac reverse-phase C18column (4.6 × 25 cm, 5 μm) with a 0.12% min−1 gradient of acetonitrile in 0.1% TFA.
Sequence Analysis
The peptide was reduced and alkylated before sequencing. Con-R was dissolved in 50 μl of 0.25 M Tris · HCl, 6 M guanidine hydrochloride, and 2 mM EDTA, pH 7.5, and reduced with 2 μl of 10% β-mercaptoethanol at room temperature (RT) for 30 min. The peptide was alkylated using 2 μl of 20% 4-vinyl pyridine in ethanol at RT in the dark for 15 min. The solution was acidified with 5 μl of 10% TFA before desalting by reverse phase HPLC with a Brownlee Aquapore RP-300 C-8 column (2.1 × 100 mm) and a gradient of acetonitrile in 0.1% TFA. Peptides were immobilized on a glass fiber filter and treated with TFA in the presence of 3 mg of polybrene and 0.2 mg of NaCl before sequencing using an Applied Biosystems model 477A pulsed liquid phase automated protein sequencer.
Amino Acid Analysis
Hydrolysis was performed in 4 N methanesulfonic acid at 110°C for 24 h and analyzed using a Perkin-Elmer HPLC analyzer with automated postcolumn OPA derivatization and fluorescence detection.
Subsequent to its purification and analysis, synthetic Con-R was prepared by Cognetix Inc. All subsequent studies were conducted with synthetic peptide.
Effect of Con-R on NMDA-Evoked Currents in Cultured Neurons
Whole-cell voltage-clamp recordings from cultured cortical neurons were used to examine the effect of synthetic Con-R on NMDA-evoked currents. Cortical cells were cultured from 15-gestational-day-old Swiss-Webster mouse fetuses and used 1 to 3 weeks after plating (Hertz et al., 1990; Skeen et al., 1994). Recordings were carried out at RT (23°C) according to previously described techniques (Hamill et al., 1981) in a control bathing solution containing 142 mM NaCl, 1.5 mM KCl, 0.1 mM CaCl2, 10 mM HEPES, 10 mM glucose, and 20 mM sucrose (320 mOsm, pH 7.4). The bathing solution also contained 1 μM strychnine to block the glycine receptor and 200 to 500 nM tetrodotoxin to block voltage-gated ion channels. Whole-cell voltage-clamp recordings were obtained with an Axopatch 200 amplifier (Axon Instruments, Burlingame, CA) using patch electrodes (2–3 MΩ) filled with an intracellular solution containing 153 mM CsCl, 10 mM EGTA, 10 mM HEPES, and 4 mM MgCl2 (290 mOsm, pH 7.4). Currents were filtered at 1 to 2 KHz, digitally sampled at 1 KHz, and acquired on computer using Axotape software (Axon Instruments). Currents were also acquired to a chart recorder.
Cells were voltage-clamped at −60 mV unless otherwise noted. Agonist and Con-R-containing solutions were applied using a rapid perfusion system that consisted of a gravity-fed multibarreled microperfusion pipette that was positioned 200 to 400 μm from the cell. Solution exchange was controlled by solenoid valves with a Valvebank 8 controller (Automate Scientific). GABAA receptor currents were evoked using 2 μM γ-aminobutyric acid (GABA), whereas NMDA- and non-NMDA-evoked currents were evoked by 10 μM NMDA (in the presence of 1 μM glycine) and 100 μM kainate, respectively. Agonists were applied for 2 to 5 s and separated by a 20- to 30-s wash period. With this protocol, GABA and GluR currents were relatively stable for the duration of the recording period. Concentration-effect data were fit to the logistic equation
Effect of Con-R on Glutamate-Evoked Currents at Recombinant GluRs Expressed in Xenopus Oocytes
cRNA Synthesis.
Template was prepared from circular plasmid cDNA by linearizing each clone with a suitable restriction enzyme. cRNA was prepared from 1 μg of linearized template using an in vitro transcription kit (Stratagene, La Jolla, CA) with a modified standard protocol that uses each of the nucleotides at 800 μM (except for GTP, 200 μM), 400 μM m7GpppG (Pharmacia, Piscataway, NJ) for capping, and an extended reaction time of 4 h with T3 or T7 RNA polymerase. All cRNAs were trace-labeled with [α-32P]UTP (Amersham, Arlington Heights, IL) to allow for quality checks by gel electrophoresis and calculation of the yield.
Electrophysiological Recordings from XenopusOocytes.
Frog oocytes of stages V/VI were obtained by surgically removing parts of the ovaries of X. laevis anesthetized with tricaine (2 g/l). The removed ovaries were chopped and incubated with 815 U/ml (2.8 mg/ml) collagenase type I (Worthington Biochemicals, Freehold, NJ) and 2200 U/ml (0.15 mg/ml) trypsin at 20°C for 2 h in calcium-free Barth's solution (see below) with slow agitation to remove the follicular cell layer and then washed extensively with Barth's solution [88 mM NaCl, 1.1 mM KCl, 2.4 mM NaHCO3, 0.3 mM Ca(NO3)2, 0.3 mM CaCl2, 0.8 mM MgCl2, 15 mM HEPES, pH 7.6, with NaOH]. Oocytes were maintained in Barth's solution supplemented with 100 μg/ml gentamycin, 40 μg/ml streptomycin, and 63 μg/ml penicillin. Oocytes were injected with 10 ng of cRNA for homomeric receptors and 5 ng of cRNA for each subunit of heterodimeric receptors 4 to 24 h after collagenase treatment using a 10-μl Drummond microdispenser. At 5 to 7 days after RNA injection, oocytes were recorded in amphibian Ringer's solution (115 mM NaCl, 2.5 mM KCl, 1.8 mM CaCl2, 10 mM HEPES, pH 7.2, with NaOH, 0.1 mg/ml BSA) under voltage clamp at a holding potential of −70 mV with a Turbo Tec-10CD amplifier (NPI). Voltage electrodes had resistances of 1 to 3 MΩ and were filled with 3 M KCl; current electrodes had resistances of approximately 1 MΩ and were filled with 3 M CaCl2. Agonist (5 μM glutamate and 10 μM glycine for NMDA receptors, 5 μM glutamate for the AMPA receptor, and 1 μM kainate for the kainate receptor) was applied by superfusion at a flow rate of 0.25 ml/min in a 50-μl recording chamber. Oocytes were perfused 3 min with agonist alone, followed by a 3-min perfusion of agonist and Con-R together and a 3-min perfusion of agonist alone. To estimate IC50 values, four different Con-R concentrations (A) were applied. Steady-state values of the evoked currents (I) were measured and fitted with the SigmaPlot program (Jandel Scientific, San Rafael, CA) to the equation
Anticonvulsant and Behavioral Toxicity Profile
The anticonvulsant profile of Con-R was characterized in four (one sensory, two electrical, and one chemical) well defined animal seizure models (White et al., 1995). The results obtained with Con-R were compared with those obtained in side-by-side studies with the noncompetitive NMDA antagonist dizocilpine (MK-801) and the NR2B-selective polyamine antagonist ifenprodil.
Con-R and MK-801 were dissolved in 0.9% NaCl, and ifenprodil was dissolved in dimethyl sulfoxide (final concentration, 12.5%). All test substances were administered i.c.v. to awake mice in a volume of 5 μl via a 10-μl Hamilton syringe.
Sound-Induced Seizures.
Male and female Frings audiogenic seizure-susceptible mice (18–25 g), obtained from an in-house colony at the University of Utah (Salt Lake City, UT) were housed in a temperature- and humidity-controlled environment and allowed free access to both food and water.
Treated mice were placed into a round Plexiglas jar (diameter, 14.5 cm; height, 30 cm) and, at the predetermined time of peak effect (TPE) of Con-R, exposed to a sound stimulus of 110 decibels (11 KHz) for 25 s, during which the animal was observed for the presence or absence of hindlimb tonic extension. Animals not displaying hindlimb tonic extension were considered protected. The severity of the seizure was quantified by assigning a numerical score to the observed response: no response, 0; wild running for <10 s, 1; wild running >10 s, 2; clonic seizure, 3; forelimb extension/hindlimb flexion, 4; and forelimb and hindlimb extension, 5 (White et al., 1992).
Maximal Electroshock (MES) and Threshold Tonic Extension (TTE) Seizures.
For this study, male CF#1 mice (18–25 g; Charles River, Wilmington, MA) were used (White et al., 1995). At the previously determined time to peak effect (TPE) of Con-R, MK-801, and ifenprodil, a drop of electrolyte solution (0.5% butacaine sulfate in 0.9% saline) was placed on the eyes of each mouse, the corneal electrodes were applied, and the electrical stimulus (50 and 15 mA, 60 Hz for MES and TTE, respectively) was delivered for 0.2 s via an apparatus similar to that originally designed by Woodbury and Davenport (1952). Abolition of the hindlimb tonic extensor component after drug treatment was taken as the end point for this test.
Subcutaneous Pentylenetetrazol (PTZ)-Induced Seizures.
At the TPE after i.c.v. administration of various doses of Con-R, 85 mg/kg PTZ was administered s.c. to Con-R-treated CF#1 mice. Individual mice were observed for the presence or absence of a minimal clonic seizure of the forelimbs and/or vibrissae for 30 min (White et al., 1995).
Determination of Median Effective (ED50) and Median Toxic (TD50) Doses.
All quantitative studies were conducted at the TPE of each test substance. In the determination of the ED50 by the respective anticonvulsant procedure, groups of seven or eight mice were injected with varying doses of each substance until at least two points were established between the dose level that protected 0% of the animals and the dose level that protected 100% of the animals. For the determination of the TD50, groups of mice were injected with increasing doses of Con-R, MK-801, or ifenprodil and tested for their ability to maintain balance on a rotating rod (2.5 cm in diameter rotating at 6 rpm). These data were then subjected to probit analysis (Finney, 1971), and the ED50, TD50, and 95% confidence intervals were calculated.
Results
Purification and Biochemical Characterization of Con-R
The purification of Con-R from the venom of C. radiatusis shown in Fig. 1. Automated Edman degradation and mass spectrometry of the pure peptide revealed the following sequence assignment (see Materials and Methods)
(Scheme 1), where γ is γ-carboxyglutamate and the other letters are standard symbols for amino acids.
The presence of Gla moieties was suspected because on Edman degradation, these characteristically yielded a very small amount of glutamate (far lower than would be expected if Glu were actually present at that position). The mole ratio of amino acids in the acid hydrolysate (2 Asx, 5 Glx, 2 Gly, 1 Arg, 5 Ala, 1 Pro, 1 Tyr, 2 Val, 1 Met, 2 Cys, 1 Ile, 1 Leu, 3 Lys) was congruent with the above sequence having five residues of Gla. In addition, the mass value (ave MH+ = 3098) was consistent with positions 3, 4, 11, and 15 being Gla residues, with the C-terminal proline as the free acid and the Cys residues in a disulfide linkage.
The sequence assignment above was confirmed by chemical synthesis. Both the free carboxyl and amidated C-terminal forms of the peptide were chemically synthesized. The HPLC elution patterns of mixed natural and synthetic peptides were examined. The native peptide coeluted with Con-R with a free C terminus and not with the amidated analog. Synthetic Con-R induced a sleep-like state in young mice identical with that elicited by the native material. Furthermore, a cDNA clone has been isolated from C. radiatus venom ducts consistent with the sequence given above. Thus, all the data are consistent with a 27-amino acid peptide containing four Gla residues and a disulfide linkage.
Electrophysiological Characterization of Con-R
Mouse Cortical Neurons.
Given the sequence homology and similar in vivo biological activity of Con-R to other members of the conantokin family, the activity of the peptide as an NMDA receptor antagonist was assessed. In the initial studies, the activity of Con-R on native NMDA receptors was characterized in whole-cell recordings from cultured mouse cortical neurons. At a holding potential of −60 mV, 10 μM NMDA (in the presence of 1 μM glycine) evoked inward current responses that showed little rundown with repeated application. As shown in Fig. 2A, Con-R (1 μM) application produced a slowly developing but potent block of inward currents evoked by 10 μM NMDA, requiring several minutes to achieve steady-state block. Con-R block was not use dependent because steady-state block could be achieved in the absence of agonist. A series of experiments similar to that in Fig. 2A were carried out with concentrations of Con-R from 30 nM to 3 μM. The fractional block values were determined at or near steady state. The concentration-block data are shown in Fig. 2B. A sigmoidal curve was fit to the data according to eq. 1 and gave an IC50 value of ∼350 nM (nH = 1.1). As shown in Fig.3, Con-R was selective for the NMDA receptor. For example, Con-R had minimal effect on kainate or GABA-evoked currents at a concentration (3 μM) that produced almost complete block of NMDA-evoked currents.
Effect of Con-R on Expressed GluRs in X. laevisOocytes.
To date, no systematic examination of the specificity of any conantokin for various subtypes of NMDA receptors has been reported. The efficacy of Con-R on two different NMDA receptor subtypes expressed in X. laevis oocytes is shown in Fig.4. Con-R, at a concentration of 3 μM, inhibited ∼95% of the current produced by the NR1-1b/NR2B NMDA subtype in response to glutamate and glycine (Fig. 4A). Similar experiments with lower concentrations of peptide gave an IC50 value of 1.03 ± 0.07 μM for the NR1-1b/NR2B subtype. Similar results were obtained when NR2B was coexpressed with the NR1 splice variant NR1-1a (Table1). At a concentration of 3 μM, Con-R significantly blocked (72%) glutamate-evoked currents in oocytes expressing NR1-1a plus NR2A and NR1-1b plus NR2A (Table 1). At higher concentrations, inhibition of the glutamate-evoked currents was observed in oocytes expressing NR1-1b plus NR2C subunits (IC50 = 7.3 ± 2.8 μM). In contrast, no inhibition of glutamate-evoked currents was observed in oocytes expressing NR1-1a/NR2C subunits at 100 μM Con-R. The affinity of Con-R appears to be >50-fold higher for the NR1-1b/NR2B subunit than for the NR1-1b/NR2D combination, which was unaffected even at 10 μM Con-R (Fig. 4B). These results indicate that Con-R is an NMDA receptor antagonist, which shows a unique NR2 subunit selectivity profile. At the various NMDA NR1-1b/NR2 receptor subunits examined, the rank order of inhibition was NR2B ≈ NR2A > NR2C ≫ NR2D. At the concentrations tested, Con-R did not inhibit the AMPA receptor GluR1 (Fig. 4C) or the kainate receptor GluR6 (Fig. 4D).
Anticonvulsant Profile of Con-R
Effect of Con-R, Ifenprodil, and MK-801 on Audiogenic Seizures.
Initial studies to characterize the anticonvulsant action of Con-R were conducted in the Frings audiogenic mouse. In this model, sound-induced seizures are characterized by wild running followed by a tonic extensor phase in response to a high-intensity sound stimulus. Con-R, ifenprodil, and MK-801 all produced a time-dependent inhibition of audiogenic seizures after i.c.v. administration of 0.0125, 25, and 3 nmol, respectively. At these doses, maximal protection for Con-R, ifenprodil, and MK-801 was observed 60, 5, and 5 min after i.c.v. administration, respectively. Maximum toxicity after i.c.v. administration was observed after 15, 5, and 5 min for Con-R, ifenprodil, and MK-801, respectively.
Anticonvulsant efficacy in the Frings mouse was quantified at the TPE for all three compounds. The data summarized in Table2 demonstrate that Con-R was the most potent of the compounds tested. Thus, compared with MK-801 and ifenprodil, Con-R was 50- and ∼2000-fold more potent against tonic extension seizures than MK-801 and ifenprodil, respectively (Table 2). As is evident from the results summarized in Table 2, Con-R displayed marked anticonvulsant activity at doses that were devoid of behavioral toxicity as estimated by the Rotarod test. Thus, when the median TD50 was evaluated in relation to the ED50, Con-R was clearly superior to the other drugs examined. For example, the protective index (PI = TD50/ED50) for ifenprodil, MK-801, and Con-R was calculated to be <1, 2.2, and 17.5, respectively.
In addition to blocking the tonic extensor component of the audiogenic seizure, Con-R also eliminated the running and clonic phases at doses devoid of behavioral toxicity. In contrast, toxic doses of MK-801 and ifenprodil were able to block only the tonic extensor component of the seizure (results not shown).
Effect of Con-R, Ifenprodil, and MK-801 on Electrically Induced Tonic Extension Seizures.
Con-R, ifenprodil, and MK-801 were also evaluated for their ability to block tonic extension induced by maximal (50 mA) and threshold (15 mA) electroshock stimulation after i.c.v. administration to CF#1 mice. Results summarized in Table3 demonstrate that Con-R possesses the ability to block tonic extension induced by both threshold and MES stimulation. Con-R was equipotent in both models, and efficacy was observed at doses that did not produce Rotarod impairment. In contrast, ifenprodil was only partially efficacious against TTE seizures and was inactive against MES seizures at a behaviorally toxic dose of 100 nmol (Table 3). Likewise, toxic doses of MK-801 were required to block TTE seizures (TD50 = 1.8 nmol, ED50 = 4 nmol). In the MES test, a maximum protection of 25% was observed at an MK-801 dose that was 6-fold greater than the TD50 (Table 3). Thus, compared with MK-801 and ifenprodil, Con-R displayed a more favorable PI in the two electrical models used.
Effect of Con-R, Ifenprodil, and MK-801 on s.c. PTZ-Induced Clonic Seizures.
Con-R, ifenprodil, and MK 801 were all tested for their ability to block s.c. PTZ-induced clonic seizures. The results summarized in Table 3 demonstrate that Con-R was only partially effective against clonic seizures induced by s.c. administered PTZ. No greater protection was afforded by higher doses of Con-R. In fact, at a behaviorally toxic dose of 400 pmol, only two of eight mice were protected.
Ifenprodil and MK-801 were both more efficacious than Con-R against clonic seizures induced by s.c. PTZ. However, behaviorally toxic doses of both compounds were required to block s.c. PTZ seizure activity (Table 3).
Discussion
A number of investigations have implicated a role for the excitatory neurotransmitter glutamate in the initiation and propagation of seizure activity (McNamara, 1994; Chapman, 1998). However, the results from clinical studies in epilepsy patients with competitive NMDA antagonists have been rather disappointing (Loscher and Schmidt, 1994; Meldrum, 1995). In the trials that have been conducted to date, efficacy in epilepsy patients was not evident at doses that produced marked psychotomimetic side effects.
Recently, it has been demonstrated that the expression of both NMDA and non-NMDA subunits is altered by disease (Pellegrini-Giampietro et al., 1992; Mathern et al., 1998). For example, expression of the GluR2 subunit of the AMPA receptor is reduced by both ischemia and epilepsy (Pellegrini-Giampietro et al., 1992); whereas the NR2B subunit is increased in the temporal lobe of patients with temporal lobe epilepsy and the hippocampus of rat models of human partial seizures (Mathern et al., 1998). In light of the importance of these two specific subunits in modulating fast (AMPA) and slow (NMDA) excitatory neurotransmission, these changes would be expected to increase CNS excitability. Given that the pharmacology of the NMDA receptor is determined by the subunit composition (Buller et al., 1994; Laurie and Seeburg, 1994; Priestly et al., 1995; Monaghan and Larsen, 1997), it further reasons that a subunit-selective antagonist might be preferred over a less selective antagonist for the treatment of those disorders and diseases where a selective alteration in subunit expression contributes to its pathology. Such an approach would hopefully target the underlying pathology without altering normal CNS function.
In the present study, we characterized the subunit selectivity and the anticonvulsant profile of the novel NMDA receptor-selective peptide Con-R. The results summarized in Fig. 3 demonstrate that Con-R is a somewhat selective antagonist of NMDA-evoked currents in murine cortical neurons and does not modulate currents evoked by the direct application of kainate or GABA. Furthermore, it does not block currents evoked by glutamate activation of homomeric GluR1 AMPA receptors (Fig.4C) or GluR6-containing kainate receptors (Fig. 4D). In addition to its ability to modulate currents through the NR2B heteromeric receptor, Con-R also blocks currents in oocytes expressing NMDA receptors containing the NR2A subunit and, at somewhat higher concentrations, blocks NMDA receptors containing NR2C (Table 1). In contrast, Con-R does not affect glutamate-evoked currents in oocytes coexpressing the NR1-1b plus NR2D subunit (Fig. 4B). Thus, in terms of its subunit-selective action, Con-R clearly differs from the polyamine site antagonist ifenprodil, which is selective for receptors containing the NR2B subunit (Williams, 1993), or MK-801, which is slightly more selective at the NR2A subunit (Laurie and Seeburg, 1994). Furthermore, the inhibition profile of Con-R is different from the competitive NMDA antagonists CGS 19755, CGP 39653, and CPPene, which exert a block at NR1-NR2A, NR1-NR2B, NR1-NR2C, and NR1-NR2D (Laurie and Seeburg, 1994).
The results summarized in Table 2 show that Con-R possesses a unique anticonvulsant profile relative to both the NR2B subunit-selective polyamine antagonist ifenprodil and the channel blocker MK-801. First, Con-R was observed to be more potent, more efficacious, and less toxic than either ifenprodil or MK-801. Furthermore, Con-R was found to be effective in a battery of well defined seizure models at doses that are devoid of behavioral toxicity. For example, Con-R was found to block sound-induced tonic extension and TTE- and MES-induced tonic extension. In addition, it was partially efficacious against s.c. PTZ-induced clonic seizures. These results indicate that Con-R possesses a broad anticonvulsant profile and is likely to be useful against a variety of seizure types that may include both partial and generalized tonic-clonic seizures (White et al., 1998).
The doses of Con-R required to block the various seizure types were well below those that resulted in motor impairment in the Rotarod test. As summarized in Tables 2 and 3, the calculated protective indices for the s.c. PTZ, TTE, MES, and audiogenic seizure tests were ∼2, 2.3, 2.4, and 17.5, respectively. More importantly, none of the doses tested resulted in behavioral deficits characteristic of competitive and noncompetitive NMDA antagonists (e.g., head weaving and darting). The degree of separation between activity and toxicity becomes very impressive when considered in light of the results obtained with the noncompetitive NMDA antagonist MK-801 and the polyamine antagonist ifenprodil. For example, both MK-801 and ifenprodil were virtually inactive even at toxic doses against tonic extension induced by both MES and TTE. Although not conclusive, results obtained with MK-801 suggest that some of the toxicities associated with nonselective NMDA antagonists that were observed in the unsuccessful clinical investigations may result in part from nonselective antagonism of multiple NMDA receptor subunits. More importantly, the excellent protective index of Con-R obtained in Frings mice indicates that its potent anticonvulsant action was not due to a general CNS depression.
It is perhaps not surprising that the profiles of Con-R and ifenprodil were so markedly different given their somewhat different NR2 subunit selectivities (and possible differences in selectivity for NR1 splice forms). A number of factors could contribute to the observed differences in the anticonvulsant profile of these two NMDA antagonists. First, the anatomical substrate underlying the various seizure types has not been fully elucidated, and although both Con-R and ifenprodil were injected directly into the ventricular space of mice, it is possible that the pharmacokinetic distribution of Con-R differs from that of ifenprodil. Furthermore, the mechanisms through which these two compounds block NMDA currents may be different. Studies with the structurally related conantokin, Con-G, suggest that conantokins are competitive antagonists at the NMDA recognition site (Hammerland et al., 1992; S. Donevan, unpublished observations). In contrast, ifenprodil and other related compounds appear to act through a novel allosteric site by modifying the sensitivity of the receptor to protons. Although ifenprodil is certainly selective for the NMDA subtype of GluRs, it is also active at other ligand and voltage-gated ion channels (Karbon et al., 1990; Chenard et al., 1991; Church and Fletcher, 1995; McCool and Lovinger, 1995). Whether these effects, coupled with an effect at the NR2B receptor, contribute to the relatively low PI of ifenprodil compared with Con-R is not known.
The favorable therapeutic index of Con-R may be due, at least in part, to its subtype selectivity. However, we note that although Con-R has a >10-fold lower affinity than MK801 in vitro, it is more than one order of magnitude more potent than MK-801 in vivo (see Table 3). A definitive explanation of this discrepancy is not available. The presence of multiple Gla residues in conantokins (an essential motif for activity) has been suggested to anchor these peptides to an acidic cell membrane (Myers et al., 1990). The in vitro systems may differ in their membrane compositions from the membrane surrounding relevant NMDA receptor targets in vivo. Another reasonable explanation for the increased in vivo potency is that the peptide anchors to determinants in a subunit that is either missing or inappropriately post-translationally modified in vitro.
Clearly, additional studies are required to fully understand the observed differences between Con-R and other NMDA receptor antagonists such as ifenprodil and MK-801. Nevertheless, the results from the present investigation suggest that it may be possible to develop NMDA antagonists that exert an anticonvulsant action at doses devoid of behavioral toxicity. The present findings clearly support the continued development of compounds that display selectivity for the various NMDA receptor subunits and continued structure-activity studies on the conantokins. Information from these studies will hopefully identify the molecular motif responsible for the activity of Con-R and thereby aid in the design of peptide mimetics with a similar therapeutic profile.
Acknowledgments
We thank Steve Heinemann for kindly providing the NR2B and NR2D subunit cDNA clones and Jen-Francois Hernandez for synthesizing the amidated Con-R.
Footnotes
-
Send reprint requests to: Dr. Baldomero M. Olivera, University of Utah, Department of Biology, 257 South 1400 East, Room 201, Salt Lake City, UT 84112-0840.
-
1 This work was supported by Phase I SBIR Grant IR33 NS36507-01 from the NINDS (H.S.W., R.T.M.), Grant GM48677 from the National Institute of General Medical Sciences (B.M.O.), DFG Sonderforschungsbereich 406 (M.H.), and the Graduiertenkolleg “Organization and Dynamics of Neural Networks” at Göttingen University (I.P.).
- Abbreviations:
- AMPA
- α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid
- Con-R
- conantokin-R
- Con-G
- conantokin-G
- MK-801
- dizocilpine
- NMDA
- N-methyl-d-aspartate
- CNS
- central nervous system
- RT
- room temperature
- GluR
- glutamate receptor
- Gla
- γ-carboxyglutamate
- i.c.v.
- intracerebroventricular
- TPE
- time to peak effect
- TFA
- trifluoroacetic acid
- PI
- protective index
- MES
- maximal electroshock
- TTE
- threshold tonic extension
- PTZ
- pentylenetetrazol
- Received July 2, 1999.
- Accepted September 16, 1999.
- The American Society for Pharmacology and Experimental Therapeutics