Abstract
Glutamic acid activates ionotropic glutamate receptors that mediate excitatory transmission in the central nervous system. The introduction of a methyl group at position 4 of glutamic acid imparts selectivity for kainate receptors, relative to other (N-methyl-d-aspartate and α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid) ionotropic glutamate receptors. Among the stereoisomers of 4-methylglutamic acid, the potency of the (2S,4R)-isomer (SYM 2081) to inhibit [3H]kainic acid binding to both wild-type (rat forebrain) and recombinant (GluR6) kainate receptors (IC50 values of ∼32 and 19 nM, respectively) was comparable to that of kainic acid (IC50 values of ∼13 and 28 nM, respectively). SYM 2081 was ∼800- and 200-fold less potent as an inhibitor of radioligand binding to wild-type (rat forebrain) α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid andN-methyl-d-aspartate receptors, respectively. Preexposure of human embryonic kidney 293 cells stably expressing GluR6 receptors to low concentrations of SYM 2081 (30–300 nM) resulted in a reversible blockade of the rapidly desensitizing currents produced by kainate application. At higher concentrations, SYM 2081 (EC50 of ∼1 μM) elicited kainate-like, rapidly desensitizing, inward currents. Pretreatment of recombinant GluR6 receptors with concanavalin A both abolished the effect of SYM 2081 to block kainate-induced currents and revealed nondesensitizing currents induced by SYM 2081 alone. The latter observations provide strong support for the hypothesis that SYM 2081 blocks kainate-induced currents through a process of agonist-induced desensitization. SYM 2081 also activated α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid receptor currents in primary cultures of cerebral cortex and, consistent with data obtained by radioligand binding, was ∼5-fold less potent than kainate (EC50 values of 325 and 70 μM, respectively) in this measure. SYM 2081 is a high-affinity, selective, kainate agonist that may prove useful both as a probe to examine the physiological functions of kainate receptors and as the prototype of a novel class of therapeutic agents.
Ionotropic glutamate receptors are distributed throughout the mammalian central nervous system, where they subserve neurotransmission at the majority of excitatory synapses. Molecular cloning studies have identified more than a dozen discrete subunits that comprise this heterogeneous family of ligand-gated ion channels (Hollmann and Heinemann, 1994; Schoepfer et al., 1994). Based on pharmacological and electrophysiological criteria, the receptors have been subclassified into NMDA, AMPA and kainate receptors.
Although converging lines of evidence have implicated activation of NMDA and, to a lesser extent, AMPA receptors in the neuropathologies associated with stroke, head injury and seizures (Sheardown et al., 1990; Meldrum, 1992; Collingridge and Watkins, 1994), the role of kainate receptors in both physiological and pathophysiological processes remains unclear. Five cDNAs have been cloned (GluR5-7, KA1 and KA2) (Bettler et al., 1990, 1992; Egebjerg et al., 1991; Werner et al., 1991; Herb et al., 1992; Sommer et al., 1992) that, when expressed in heterologous cells, display most of the characteristics described for native kainate receptors, including rapid desensitization by kainate. There is good evidence that receptors on neurons of sensory dorsal root ganglia are composed of GluR5 and KA2 subunits (Bettler et al., 1990; Herb et al., 1992; Sommer et al., 1992; Partin et al., 1993), whereas kainate receptors in hippocampus are likely to be composed of the GluR6 subunit (Ruanoet al., 1995), possibly in combination with KA1 or KA2 (Wisden and Seeburg, 1993). Despite the widespread distribution of both [3H]kainate binding sites (London and Coyle, 1979; Coyle, 1983; Honoré et al., 1986) and mRNAs encoding kainate receptors (Wisden and Seeburg, 1993) in brain, with few exceptions (Huettner, 1990; Paternain et al., 1995) it has been difficult to demonstrate kainate receptor-mediated currents in neurons. This may be due to the rapid desensitization produced by kainate in dorsal root ganglia and cultured hippocampal neurons (Huettner, 1990;Lerma et al., 1993; Ruano et al., 1995), the “masking” of kainate responses by larger, AMPA receptor-mediated currents (Paternain et al., 1995) and the paucity of selective, high-affinity kainate receptor ligands.
Most agonists and antagonists of non-NMDA receptors show only limited selectivity between the AMPA and kainate receptor subtypes (Wonget al., 1994; Wilding and Huettner, 1996). For example, although kainic acid is roughly 100-fold selective for kainate receptors over AMPA receptors (Huettner, 1990; Patneau et al., 1994; Paternain et al., 1995), the agonist elicits large sustained (nondesensitizing) currents due to its activation of AMPA receptors (Patneau and Mayer, 1991) in hippocampal and neocortical neurons. This activity complicates the use of kainate as a selective compound (Paternain et al., 1995). The most selective antagonists, such as the quinoxalinedione derivative ACEA-1011 and the isatin oxime NS-102, display only 10- to 20-fold preference for kainate receptors over AMPA receptors (Verdoorn et al., 1994;Wilding and Huettner, 1996). We have recently reported the stereoselective synthesis of the four diastereomers of 4-methylglutamate (Gu et al., 1995). Among these isomers, the (2S,4R)-form (SYM 2081) displayed an affinity for kainate binding sites in brain comparable to that of kainic acid. Here, we demonstrate that SYM 2081 is a high-affinity, competitive inhibitor of [3H]kainate binding to recombinant GluR6 receptors and that SYM 2081 potently blocks kainate responses through desensitization of the ion channel. Furthermore, these effects are observed at concentrations 2 to 3 orders of magnitude lower than those required to affect ligand binding to other (NMDA and AMPA) ionotropic glutamate receptors.
Materials and Methods
Membrane preparation for [3H]kainate and [3H]AMPA binding.
Cerebral cortical membranes from adult, male, Sprague-Dawley rats (175–300 g; Taconic Farms, Germantown, NY) were prepared essentially as described (London and Coyle, 1979; Honoré et al., 1986). All procedures were carried out at 0–4°C unless otherwise indicated. Tissues were disrupted (Polytron homogenizer, setting 6, 30 sec) in 10 volumes of ice-cold 50 mM Tris-HCl buffer (pH 7.4). The homogenate was centrifuged at 25,000 × g for 20 min, and the resulting pellet was “washed” three times by successive resuspensions and recentrifugations. The pellet was resuspended in 10 volumes of Tris-HCl buffer, incubated in a water bath (37°C) for 30 min and recentrifuged. The resulting pellet was resuspended in 10 volumes of buffer and frozen at −70°C for at least 24 hr before use. On the day of assay, membranes were thawed, resuspended in 50 mM Tris-HCl and centrifuged at 25,000 × g for 20 min. The pellet was resuspended in 50 mM Tris-HCl for [3H]kainate binding and in 30 mM Tris-HCl buffer containing 2.5 mM CaCl2 and 100 mM KSCN for [3H]AMPA binding.
Membranes were also prepared from HEK293 cells that stably express the fully edited version (VCR) of GluR6 (Filipkowski et al., 1994) under the control of an inducible promoter (Howe et al., 1995). Cells were obtained from J. R. Howe, Yale University (New Haven, CT), and maintained in continuous culture in Eagle minimal essential medium containing 10% fetal bovine serum, 2 μg/ml tetracycline and 0.5 mg/ml G-418. Tetracycline was omitted from the growth medium to induce the expression of the GluR6 subunit. Cells grown in flasks (150-cm2) to 75 to 100% confluency were washed twice with 10 ml of phosphate-buffered saline and removed with a glass rod. The cell suspension was centrifuged (200 × g for 5 min), and the resulting pellet was stored at −70°C. The frozen pellet (typically from five to eight flasks) was thawed, homogenized in 40 ml of ice-cold 50 mM Tris-HCl buffer (pH 7.4) with a Polytron homogenizer (setting 6, 30 sec) and centrifuged at 25,000 × g for 20 min. The resulting pellet was homogenized in an equal volume of buffer and recentrifuged. Cells were washed twice by resuspension and centrifugation, and the final pellet was suspended in 20 ml of 50 mM Tris-HCl for [3H]kainate binding. Protein content was measured with the bicinchoninic acid protein assay reagent (Pierce, Rockford, IL), using bovine serum albumin as a standard.
Membrane preparation for [3H]MK-801 and [3H]CGP 39653 binding.
Rat forebrain (whole brain minus cerebellum and brainstem) membranes were prepared essentially as described by Nowak et al. (1993). Tissues were disrupted with a Polytron homogenizer in 10 volumes of 5 mM HEPES/4.5 mM Tris buffer (pH 7.6) containing 0.32 M sucrose. The homogenate was diluted to 50 volumes with assay buffer (5 mM HEPES/4.5 mM Tris buffer, pH 7.6) and centrifuged at 1,000 × gfor 10 min. The supernatant was decanted and centrifuged at 25,000 × g for 20 min. The resulting pellet was homogenized in 50 volumes of buffer and centrifuged at 8,000 × g for 20 min. The supernatant and “buffy” pellet coat were collected and centrifuged at 25,000 × g for 20 min. The resulting pellet was suspended in assay buffer containing 1 mM Na4EDTA, and the suspension was recentrifuged. This washing procedure was repeated four times, with EDTA being absent from the last cycle. The resulting pellet was resuspended in 5 volumes of assay buffer, frozen over solid CO2 and stored at −70°C. On the day of assay, tissues were thawed, diluted 10-fold with assay buffer and centrifuged at 25,000 × g for 20 min. The resulting pellet was resuspended in 50 volumes of assay buffer for NMDA receptor binding assays.
[3H]Kainate binding.
Assays were routinely performed in a total volume of 500 μl containing membrane suspension (∼200 μg of protein), [3H]kainate (final concentration, 5–10 nM), test compounds and 50 mM Tris-HCl buffer (pH 7.4) to volume. Nonspecific binding was defined using 0.6 mM glutamate. Assays were initiated by the addition of [3H]kainate, incubated at 0–4°C for 60 min and terminated by rapid filtration (Brandel M-24R cell harvester) through Whatman GF/C glass fiber filters, followed by two 3-ml washes with ice-cold assay buffer. Saturation isotherms in HEK293-GluR6 membranes were performed using [3H]kainate concentrations from ∼2 to 150 nM.
[3H]AMPA binding.
[3H]AMPA binding was assayed essentially as described by Honoré et al. (1988), in 30 mM Tris-HCl buffer (pH 7.4) containing 2.5 mM CaCl2 and 100 mM KSCN. Assays contained 250 μl of membrane suspension (∼200 μg of protein), 50 μl of test compounds, 50 μl of [3H]AMPA (final concentration, 10 nM) and buffer to a final volume of 500 μl. Nonspecific binding was assessed with 0.6 mM glutamate. Assays were terminated after 30 min (0–4°C) as described above.
[3H]CGP 39653 binding.
Membrane suspensions (∼100 μg of protein) were incubated with 5 nM [3H]CGP 39653, test compounds and assay buffer (5 mM HEPES/4.5 mM Tris, pH 7.6), in a final volume of 500 μl. Nonspecific binding was determined with 1 mM l-glutamate. After a 90-min incubation (0–4°C), assays were terminated by rapid filtration over Whatman GF/B filters as described above.
[3H]MK-801 binding.
Assays consisted of 250 μl of membrane suspension (50–100 μg of protein), 50 μl of test compound, 50 μl of [3H]MK-801 (final concentration, 5 μM) and assay buffer (5 mM HEPES/4.5 mM Tris, pH 7.6), in a volume of 500 μl. Nonspecific binding was assessed with phencyclidine hydrochloride (100 μM). Assays were incubated at room temperature for 2 hr and terminated by rapid filtration over Whatman GF/B glass fiber filters that had been presoaked in 0.03% polyethyleneimine, as described above.
Electrophysiology.
Currents generated by activation of kainate receptors were recorded from HEK293 cells expressing the GluR6 subunit, prepared as described under “Membrane Preparation in Materials and Methods.” Additional experiments were performed on 1- to 2-week-old primary cultures of rat cerebral cortical neurons prepared as previously described (Baughman et al., 1991). Patch pipettes (3–5 MΩ) contained 140 mM CsCl, 2 mM MgCl2, 1.1 mM tetrasodium ethylene glycol bis(β-aminoethyl ether)-N,N,N′,N′-tetraacetic acid, 1.0 mM CaCl2 and 10 mM HEPES (pH 7.2). Cells were bathed in a solution containing 150 mM NaCl, 4 mM KCl, 2 mM CaCl2, 1 mM MgCl2 and 10 mM HEPES (pH 7.4). Currents were filtered at 0.5 to 1.0 kHz and digitized (EPC-9; HEKA Elektronik) using a sampling frequency of 1 to 2 kHz. Agonists were applied to cells using a gravity-fed perfusion system consisting of six microcapillary tubes (0.32 mm o.d.; J&W Scientific) placed concentrically into the lumen of a 1.7-mm o.d. glass tube. The outer tube was pulled around the inner tubes to form a nozzle having a final o.d. of 0.3 mm. Dead space between the ends of the perfusion tubes and the tip of the nozzle was 2 to 3 μl. Solution exchange times measured 4 to 6 msec (10–90% of steady-state current) at the tip of a patch pipet placed in the position of a cell. For some experiments, a single pair of microcapillary tubes was used to deliver agonists. Results using either method were comparable and were pooled. The bath was constantly perfused at a low rate with control solution. Concentration-response curves were fitted with the logistic equationI = 1/[1 + (EC50/A)n], whereA is the concentration of the drug and n is the Hill coefficient. Fits were made with a Marquardt-Levenberg, nonlinear, least-squares, curve-fitting algorithm (Kaleidagraph).
Materials.
[3H]Kainate (specific activity, 58 Ci/mmol), [3H]AMPA (specific activity, 63 Ci/mmol), [3H]MK-801 (specific activity, 20.3 Ci/mmol) and [3H]CGP 39653 (specific activity, 42 Ci/mmol) were purchased from Du Pont-New England Nuclear (Boston, MA). The isomers of 4-methylglutamate were prepared as described (Gu et al., 1995). Eagle basal medium was supplied by (Grand Island Biological Co., NY. Fetal bovine serum was purchased from Quality Biological, Inc. (Gaithersburg, MD). Other cell culture reagents, glutamate, kainate and AMPA were purchased from Sigma Chemical Co. (St. Louis, MO). Phencyclidine HCl was supplied by the National Institute on Drug Abuse (Rockville, MD). All other materials were supplied by standard commercial sources.
Results
The effects of the four isomers of 4-methylglutamate were initially examined with wild-type glutamate receptors prepared from rat forebrain (table 1). Although all isomers exhibited selectivity for kainate receptors, compared with glutamate, the potency of SYM 2081 to inhibit [3H]kainate binding (IC50, 32 ± 3 nM) was comparable to that of kainic acid (IC50, 13 ± 2 nM). [3H]Kainate labels at least two populations of binding sites in rat brain (Honoré et al., 1986; Johansen et al., 1993). Because the initial assay conditions used primarily label a “high-affinity” [3H]kainate binding site, the potency of SYM 2081 was also determined in the presence of 20 mM Ca++, which optimizes radioligand binding to a population of “low-affinity” kainate binding sites (Honoré et al., 1986; Johansen et al., 1993). Under these conditions, the IC50 for SYM 2081 was 212 ± 16 nM (n = 4), compared with 62 ± 2.3 nM (n = 4) for kainic acid. Additional radioligand binding studies were performed in HEK293 cells that stably express the GluR6 isoform of the kainate receptor. SYM 2081 inhibited [3H]kainate binding to GluR6 receptors with aKi of 9.8 ± 3.5 nM (n = 3), compared with 14.3 ± 3.8 nM for kainic acid (n = 4) (fig. 1). Consistent with a competitive mode of action at GluR6, SYM 2081 (15 nM) increased theKd of [3H]kainic acid (from 6.6 ± 0.5 nM to 25.7 ± 1.8 nM) without significantly increasing the B max (187.6 ± 31.7 vs. 192.6 ± 9.3 fmol/mg of protein).
Compared with glutamate, the 4-methylglutamate analogs were ≥50-fold less potent as inhibitors of [3H]AMPA binding to rat brain membranes (table 1). SYM 2081 was ∼5-fold less potent than kainate as an AMPA receptor ligand (table 1). Consistent with previous results obtained using the unresolved mixture of 4-methylglutamate isomers (Olverman et al., 1988), high concentrations of the individual 4-methylglutamate isomers, including SYM 2081, were low-potency NMDA receptor agonists, compared with glutamate. These isomers enhanced [3H]MK-801 binding with EC50values ranging from 14.3 ± 1.7 μM to >100 μM and inhibited the binding of [3H]CGP 39653 (a competitive glutamate antagonist) with IC50 values ranging from 5.9 to 26.7 μM (table 1).
Characteristic of both wild-type (Huettner, 1990; Lerma et al., 1993; Patneau et al., 1994) and recombinant (Herbet al., 1992; Sommer et al., 1992; Raymondet al., 1993; Verdoorn et al., 1994) kainate receptors, fast application of kainic acid (100 μM) to HEK293 cells expressing GluR6 elicited rapidly desensitizing inward currents (fig.2A, left). Sixty-second preapplications of SYM 2081 using either fast perfusion or slow exchange in the bath reversibly blocked kainate currents (fig. 2A, middle and right). Inhibition measured 72 ± 8% (n = 5) at a concentration of 30 nM and 94 ± 4% (n = 5) at a concentration of 300 nM SYM 2081. Fast applications of higher concentrations of SYM 2081 (EC50, 1.0 ± 0.1 μM; n = 3) produced rapidly desensitizing inward currents (fig. 2B) that resembled those elicited by kainate (fig. 2A). At these higher concentrations, complete desensitization was observed within 1 sec over the entire concentration range of SYM 2081 examined (0.3–30 μM).
In another series of experiments, cells were pretreated with Con A (0.3 mg/ml, for 5 min) to block agonist-induced desensitization (Huettner, 1990; Partin et al., 1993). In Con A-pretreated cells, kainate (100 μM) elicited large nondesensitizing currents that were unaffected by SYM 2081 (fig. 2C). Kainate-evoked currents in the presence of SYM 2081 (30 nM) were 98 ± 4% (n = 4) of control kainate responses in the absence of SYM 2081. In contrast, the same concentration of SYM 2081 reduced kainate-evoked currents to 28 ± 8% (n = 5) of control values in cells that were not treated with Con A.
Because radioligand binding studies demonstrated a lower affinity of SYM 2081 for AMPA receptors, compared with kainate (table 1), parallel concentration-response experiments were performed in rat neocortical neurons. Others have shown that, when stimulated by kainate, these neurons exhibit large sustained currents due to the activation of AMPA receptors (Wilding and Huettner, 1995); currents induced by stimulation of kainate receptors have not been reported for these cells (but seePaternain et al., 1995, for hippocampal neurons). High concentrations of SYM 2081 (>10 μM) produced sustained (nondesensitizing) currents similar in waveform to those evoked by kainate (fig. 3). However, SYM 2081 was ∼4.6-fold less potent than kainate (EC50 of 325 ± 23 vs.70 ± 6 μM, n = 4).
Discussion
The identification of cDNAs encoding a family of kainate receptors (GluR5-7, KA1 and KA2) has resulted in significant insight into their structure and organization (Hollmann and Heinemann, 1994; Schoepferet al., 1994; Bettler and Mulle, 1995). Nonetheless, the physiological functions of kainate receptors remain obscure, due, in part, to a lack of selective, high-affinity ligands. We demonstrate here that introduction of a methyl group at the 4-position of glutamic acid results in a marked increase in potency and selectivity for kainate receptors. Among the four stereoisomers, SYM 2081 exhibited the greatest selectivity for kainate receptors, inhibiting [3H]kainate binding to both high- and low-affinity sites in rat brain with a potency comparable to that of kainic acid.
Because native kainate receptors are heterogeneous (Honoréet al., 1986; Johansen et al., 1993), the properties of SYM 2081 were also examined with recombinant GluR6 receptors. These receptors exhibit saturable [3H]kainate binding (Lomeli et al., 1992; Tygesen et al., 1994; Verdoorn et al., 1994) and produce rapidly desensitizing currents in response to kainate and glutamate (Herbet al., 1992; Raymond et al., 1993; Verdoornet al., 1994), characteristic of native kainate receptors (Huettner, 1990; Lerma et al., 1993; Patneau et al., 1994; Ruano et al., 1995). TheKd of [3H]kainate at GluR6 has been previously reported to range between 12.9 and 95 nM (Bettler et al., 1992; Tygesen et al., 1994;Verdoorn et al., 1994), and theKd obtained in this study (6.6 nM) is consistent with the former value. The potency of SYM 2081 to inhibit [3H]kainate binding to GluR6 was somewhat higher than that of kainate itself, and the increase inKd of [3H]kainic acid (without a change in B max) observed in the presence of SYM 2081 is consistent with a competitive mode of action.
Patch-clamp measurements confirmed that SYM 2081 has a potent effect on GluR6 receptors. Preapplication of this compound at a concentration of 30 nM induced a steady-state desensitization that abolished responses to kainate in a reversible manner. At much higher concentrations, SYM 2081 (1 μM) elicited currents that rapidly desensitized, resembling those evoked by kainate. These findings, taken together with the failure of SYM 2081 to block kainate currents in Con A-treated cells, support the hypotheses that SYM 2081 1) acts at the same binding site as kainic acid and 2) inhibits kainate currents via an agonist-induced desensitization. In toto, these findings lead us to conclude that SYM 2081 potently and effectively desensitizes kainate receptors.
SYM 2081 was ∼5-fold less potent than kainate both as an inhibitor of radioligand binding to AMPA receptors in rat brain membranes and as an activator of AMPA receptors in primary cultures of cerebral cortex. The lower affinity of SYM 2081 at AMPA receptors, relative to kainate receptors, may in part explain the absence of seizures in mice after parenteral administration of up to 512 mg/kg (data not shown), whereas far smaller doses of kainate are convulsant (Olney et al., 1974; Coyle, 1983; Sperk et al., 1985). Although it could be argued that SYM 2081 does not cross the blood-brain barrier, steady-state brain concentrations are ∼6 μM after parenteral administration of 200 mg/kg (data not shown). The low affinity of SYM 2081 for AMPA receptors may increase its usefulness in future functional studies of the behavioral consequences of in vivoblockade of kainate receptors.
To date, the most compelling evidence linking activation of glutamate receptors to the neuropathologies associated with stroke, head injury and seizures stems from studies using selective, high-affinity compounds that act at the multiple allosteric regulatory sites on NMDA receptors (Collingridge and Watkins, 1994). In contrast, the involvement of non-NMDA subtypes of glutamate receptors in these and other pathophysiological processes is far less compelling, which may be attributed, at least in part, to the paucity of high-affinity, selective AMPA and kainate receptor ligands. SYM 2081 has significantly higher affinity and selectivity for kainate receptors than do previously described competitive antagonists such as NS-102 and 2,3-dihydroxy-6-nitro-7-sulfamoylbenzo(f)quinoxaline (Honoré et al., 1988; Sheardown et al., 1990; Verdoorn et al., 1994). SYM 2081 represents a prototypical, high-potency, kainate-selective ligand that may be useful both for the elucidation of the physiological roles of kainate receptors and for the development of therapeutic agents.
Footnotes
-
Send reprint requests to: Dr. Kenneth Jones, Synaptic Pharmaceutical Corp., 215 College Rd., Paramus, NJ 07652.
-
↵1 Present address: Department of Molecular Pharmacology and Biological Chemistry, Northwestern University Medical Center, Chicago, IL 60611.
-
↵2 Present address: Synaptic Pharmaceutical Corp., 215 College Rd., Paramus, NJ 07652.
- Abbreviations:
- AMPA
- α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid
- Con A
- concanavalin A
- HEK
- human embryonic kidney
- HEPES
- 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
- NMDA
- N-methyl-d-aspartate
- SYM 2081
- (2S,4R)-4-methylglutamic acid
- Received May 7, 1996.
- Accepted September 17, 1996.
- The American Society for Pharmacology and Experimental Therapeutics