Abstract
This study set out to profile the activity of (S)-desmethylzopiclone (SEP-174559) at subtypes of the γ-aminobutyric acid type-A (GABAA) receptor and other neurotransmitter receptor ion channels. Recombinant receptors were expressed in human embryonic kidney 293 cells and examined functionally by patch-clamp recording with fast perfusion of agonist and drug solutions. Micromolar concentrations of SEP-174559 potentiated GABAA receptor currents evoked by subsaturating concentrations of GABA. The potentiation was related to a leftward shift in the GABA dose-response curves, suggesting the drug acts to increase GABA binding affinity. The potentiation strictly required the presence of the γ2 subunit; no enhancement was seen for receptors containing instead the γ1 subunit or lacking a γ subunit altogether. SEP-174559 and its parent compound, racemic zopiclone, were not selective between α1-, α2-, or α3-bearing GABAAreceptors. Within the nicotinic receptor superfamily, SEP-174559 did not affect serotonin type-3 receptor function but was found to inhibit nicotinic acetylcholine (nACh) receptors. The inhibition of nACh receptors was noncompetitive and was mimicked by zopiclone, alprazolam, and diazepam. In the glutamate receptor superfamily, SEP-174559 inhibited N-methyl-d-aspartate (NMDA) receptor currents but did not affect non-NMDA receptors. These data confirm that SEP-174559 has benzodiazepine-like actions at γ2-bearing subtypes of the GABAA receptor and suggest additional actions of benzodiazepine-site ligands at nACh and NMDA receptors.
GABAAreceptors mediate most of the inhibitory synaptic transmission in the central nervous system and are the principal target of neuroactive drugs used in the treatment of anxiety, insomnia, and epilepsy. Based on their genetic and structural relatedness, GABAA receptors belong to the nicotinic superfamily of neurotransmitter receptor ion channels, which also includes nicotinic acetylcholine (nACh), serotonin type-3 (5-HT3), and glycine receptors. GABAA receptors are GABA-gated Cl− ion channels typically composed of two α (1–6), two β (1–3), and one γ (1–3) or δ1 subunit in a pentameric assembly (Sieghart, 1995; Barnard et al., 1998; Mehta and Ticku, 1999). Among the many drugs that act at these receptors are the benzodiazepines, allosteric modulators that enhance GABAA receptor signaling by binding to a site at the α-γ subunit interface (Gunther et al., 1995; Sigel and Buhr, 1997). Transgenic studies indicate that benzodiazepines exert their sedative effects and partly their anticonvulsant effects through α1β2γ2 receptors (Rudolph et al., 1999; McKernan et al., 2000) and their anxiolytic effects through α2β2γ2 receptors (Low et al., 2000). Thus, it may be possible for a drug to elicit only a desired action by targeting a particular receptor subtype.
Zopiclone is a cyclopyrrolone, a class of nonbenzodiazepine drugs that have high affinity for the benzodiazepine binding site but, relative to benzodiazepines, produce comparable anxiolytic effects with less sedation, muscle relaxation, or addictive potential (Piot et al., 1990;Sanger et al., 1994; Karle and Nielsen, 1998). Racemic zopiclone, which consists of (R)- and (S)-enantiomers, acts at the benzodiazepine site to enhance GABAA receptor binding and function (Im et al., 1993; Davies et al., 2000). Although the selectivity of zopiclone is not fully characterized, it is known to act at γ2-bearing GABAA receptors, including α1β2γ2 (Im et al., 1993; Reynolds and Maitra, 1996; Davies et al., 2000). Recent behavioral studies in rodents suggest that the sedative and anxiolytic activities of (R,S)-zopiclone are produced mainly by the (S)-enantiomer (Carlson et al., 2001), which is metabolized in vivo to (S)-desmethylzopiclone (SEP-174559). SEP-174559 is active on its own and was found selectively to produce anxiolytic and anticonvulsant effects; doses 100 times higher were required to impair locomotor activity or motor coordination (Carlson et al., 2001). Yet, the activity of SEP-174559 and its selectivity for various subtypes of the GABAA receptor have not been described.
The present study profiles the benzodiazepine-like actions of SEP-174559 in comparison with zopiclone at various GABAA receptor subtypes and further defines the spectrum of effects of benzodiazepine-site ligands at other neurotransmitter receptor ion channels in both the nicotinic and glutamate receptor superfamilies.
Experimental Procedures
Cell Cultures and Transfections.
Human embryonic kidney 293 fibroblasts (HEK293, CRL 1573; American Type Culture Collection, Manassas, VA) were cultured in minimal essential medium supplemented with 10% fetal bovine serum and 2 mM glutamine (Invitrogen, Carlsbad, CA) and incubated at 37°C in a 5% CO2 environment. Cells were plated into 25-cm2 flasks (Falcon Plastics, Oxnard, CA) and passaged twice weekly to fresh flasks. Excess cells were removed, plated onto poly-d-lysine-coated 35-mm dishes (Nalge Nunc, Naperville, IL), and cotransfected the following day with cDNA plasmids encoding receptor subunits and enhanced green fluorescent protein at a 9:1 ratio. Transfections were done using the LipofectAMINE PLUS reagents (Invitrogen). Patch-clamp recordings were obtained 12 to 48 h post-transfection. Studies used mammalian cDNA expression vectors encoding rat GABAA receptor α1 (accession no. L08490), α2 (L08491), α3 (L08492), β2 (X15467), γ1 (X57514), and γ2 (L08497) subunits, rat nACh receptor α3 (L31621) and β4 (U42976) subunits, the mouse 5-HT3A receptor subunit (M74425), rat NMDA receptor NR1 (X63255) and NR2B (M91562) subunits, rat GluR1flip AMPA receptor subunit (M38060), and rat GluR6 kainate receptor subunit (Z11715).
Patch-Clamp Recording.
Cells were continuously superfused with standard extracellular solution containing 150 mM NaCl, 3 mM KCl, 5 mM HEPES, 1 mM MgCl2, 1.8 mM CaCl2, 10 mM glucose, and 0.1 mg ml−1 phenol red, pH 7.3. Recording microelectrodes were fabricated from thin-walled borosilicate glass capillary tubes (TW150F; World Precision Instruments, New Haven, CT) having resistances of 2 to 4 MΩ when filled with an internal solution containing 135 mM CsCl, 10 mM CsF, 10 mM HEPES, 5 mM EGTA, 1 mM MgCl2, and 0.5 mM CaCl2, pH 7.2. Whole-cell and outside-out patch recordings were performed in voltage clamp at a holding potential of −70 mV using an Axopatch 200B amplifier (Axon Instruments, Union City, CA). Current signals were filtered at 1 to 3 kHz with an eight-pole Bessel filter (Cygnus Technology, Delaware Water Gap, PA), digitized at 1 to 20 kHz, and stored on a Macintosh PowerPC-G3 computer using an ITC-16 interface (InstruTECH Corporation, Port Washington, NY) under the control of the data acquisition and analysis program Synapse (Synergistic Research Systems, Silver Spring, MD).
Rapid Solution Exchange.
Rapid drug applications were achieved in two ways. In whole-cell recordings of most receptors, a multivalve solution exchange system was used. Solutions were driven by a syringe pump at a rate of 2 ml min−1 through a flowpipe having either four or eight inputs glued together within ∼1 mm of a common output. Solution flow-through in each channel was controlled by a set of upstream three-way solenoid valves (Lee Co., Westbrook, CT). Switching between control and drug solutions was achieved by opening and closing the appropriate valves under computer control. The rate of solution exchange using this system was ≤5 ms as determined from open-tip junction currents but was further limited by cell diameter. In outside-out patch recordings with GluR1 and GluR6 receptors, the ultrafast solution exchange needed to resolve the fast response, and kinetic properties of these receptors was achieved using an LSS-3100 piezotranslator (Burleigh, Fishers, NY). Control and agonist solutions were driven simultaneously at a rate of 0.3 ml min−1 through the two parallel barrels of a theta tube. The membrane patch was positioned in the control stream near the solution interface and a piezotranslator was used to rapidly move the theta tube ∼50 μm such that the solution interface moved across the patch. Inputs were then switched to solutions containing drug and retested. The rate of solution exchange in this system was ≤100 μs as determined from open-tip junction currents measured at the end of each experiment.
Data Analysis.
Agonist dose-response data were normalized to saturation values for each cell and fit by the logistic equationI = Imax/(1 + (EC50/[agonist])nH), where Imax is the maximal current at saturating agonist concentrations, EC50 is the concentration of agonist that elicits a half-maximal response, andnH is the Hill coefficient or steepness factor; these parameters were derived using an iterative least-squares fitting algorithm.
Materials.
Reagents were from Sigma-Aldrich (St. Louis, MO) or Sigma/RBI (Natick, MA). SEP-174559, (R,S)-zopiclone, and the zopiclone enantiomers were kindly provided by Sepracor (Marlborough, MA). Agonist and drug stocks were prepared in 0.05 M HEPES buffer, pH 7.4, except (R,S)-zopiclone, (S)-zopiclone, (R)-zopiclone, alprazolam, and diazepam were prepared at 20 mM in dimethyl sulfoxide; the final concentration of dimethyl sulfoxide was ≤0.1%, which was tested and found by itself to have no activity on the receptors examined in this study. Agonist solutions for dose-response analyses were prepared freshly by serial dilution.
Results
Racemic zopiclone [(R,S)-zopiclone] has been shown to enhance GABAA receptor currents via the benzodiazepine site (Im et al., 1993; Davies et al., 2000). The molecular actions of its principal metabolite, SEP-174559, have not been described. To profile the activity and GABAAreceptor selectivity of SEP-174559, GABA-evoked currents were tested by whole-cell patch-clamp recording in transfected HEK293 cells expressing various recombinant GABAA receptor subtypes. Effects on related and unrelated neurotransmitter receptor ion channels were also tested to assess the specificity of drug action. In many cases, the actions of (R,S)-zopiclone or other benzodiazepine-site ligands were also tested for comparison.
GABAA Receptor Actions.
To test the influence of the GABAA receptor γ subunit, HEK293 cells were cotransfected with α1 and β2 subunits alone or in combination with γ1 or γ2 subunits. Responses were evoked by application of 10 μM GABA. Representative traces are presented in Fig.1A. Data are summarized in Table1. In cells expressing α1β2γ2 receptors, coapplication of 2 μM SEP-174559 enhanced GABA-evoked currents by 11 ± 3% (n = 4). Coapplication of 20 μM SEP-174559 produced a greater increase of 32 ± 4% (n = 11). Likewise, 2 and 20 μM (R,S)-zopiclone enhanced GABA responses by 34 ± 8% (n = 5) and 53 ± 11% (n = 11), respectively. In cells expressing α1β2 alone, neither SEP-174559 nor (R,S)-zopiclone had any effect up to 20 μM (n = 6). Also, no enhancement was seen for γ1-bearing receptors. Rather, α1β2γ1 receptors were modestly inhibited by 20 μM SEP-174559 by 9 ± 1% (n = 10). This inhibition of the α1β1γ1 subtype was unexpected but not unique to cyclopyrrolones because 20 μM alprazolam and 20 μM diazepam also inhibited α1β1γ1 currents by 83 ± 5 (n = 4) and 84 ± 7% (n = 4), respectively.
To test the influence of the α subunit composition, HEK293 cells were cotransfected with α1, α2, or α3 in combination with β2 and γ2 subunits. Responses were evoked by 10 μM GABA and tested by coapplication of 20 μM SEP-174559 or 20 μM (R,S)-zopiclone. GABA-evoked responses were enhanced in all cases regardless of α subunit composition (Fig. 1B); however, the magnitude of potentiation followed the rank order of α3β2γ2 > α2β2γ2 > α1β2γ2 (Table 1). This rank order did not result from differences in drug affinities at the GABAA subtypes. Drug dissociation constants (Kd) were estimated from the kinetics of onset (kon) and recovery (koff) from potentiation using the equation Kd =koff/kon, and these values are given in Table 2. Rather, the rank order reflected differences in GABA affinities of the various subtypes. To determine this, GABA dose-response curves were generated in the absence and presence of 20 μM SEP-174559 in cells expressing α1, α2, or α3 in combination with β2 and γ2. Cells were also examined that expressed α1β2γ1 receptors that were modestly inhibited by SEP-174559. The EC50 for receptor activation by GABA followed the rank order of α3β2γ2 > α2β2γ2 > α1β2γ2 (Table3). GABA dose-response curves were left-shifted by SEP-174559 in all cells expressing γ2-bearing receptors but not altered, or slightly right-shifted, in cells expressing α1β2γ1 receptors (Fig.2). The Hill coefficient of the dose-response curve was not altered in any case, suggesting the actions of SEP-174559 did not involve a change in cooperativity or in the number of functional GABA binding sites. The magnitude of the leftward-shift of the dose response was entirely consistent with the magnitude of SEP-174559 potentiation at any given concentration of GABA. For example, when α1β2γ2 receptors were activated by 2 μM GABA (EC10), the potentiation by 20 μM SEP-174559 was 103 ± 17% (n = 5) and by 20 μM (R,S)-zopiclone was 227 ± 44% (n = 5). Moreover, SEP-174559 and (R,S)-zopiclone did not affect α1β2γ2 (n = 4), α2β2γ2 (n = 3), or α3β2γ2 (n = 3) receptor currents evoked by saturating GABA (1 mM). These data indicate that SEP-174559 enhances GABA binding affinity at all of the γ2-bearing GABAA subtypes tested under our conditions and provide no indication of any α-subtype specificity.
The actions of the (R)- and (S)-zopiclone enantiomers were also examined in cells expressing α1β2γ2, α2β2γ2, or α3β2γ2 receptor subtypes. Responses were evoked by 10 μM GABA and tested by coapplication of 20 μM (R)-zopiclone or (S)-zopiclone. Both (R)- and (S)-enantiomers were found to enhance GABAA receptor currents (Fig.3). The (S)-enantiomer was nearly twice as effective at all γ2-bearing subtypes tested (Table1). The actions of (R)- and (S)-zopiclone enantiomers also differed in the rate of recovery from potentiation after drug removal. Potentiation by (R)-zopiclone was quickly relieved, reminiscent of SEP-174559, whereas the slow recovery from potentiation by (S)-zopiclone seems to underlie the prolonged higher affinity action of the racemic.
Drug Actions on Other Receptors in Nicotinic Superfamily.
Given their genetic and structural relatedness to GABAA receptors, it was prudent to investigate whether SEP-174559 is active at nACh receptors or 5-HT3 receptors. The activity of SEP-174559 was tested by whole-cell patch-clamp recording in transfected HEK293 cells expressing either 5-HT3-A or α3β4 nACh receptors. SEP-174559 did not affect 5-HT3receptor currents. Responses were evoked by application of 100 μM 5-HT. Coapplication of 20 μM SEP-174559 did not alter the magnitude or decay time course of 5-HT-evoked currents (n = 4). SEP-174559 also did not affect responses evoked by 10 μM 5-HT (n = 3), which we determined is near the EC50 for these receptors. Alprazolam (20 μM) also did not affect 5-HT3 receptor currents evoked by 10 μM 5-HT (n = 2) or 100 μM 5-HT (n = 2). In contrast, SEP-174559 did inhibit nACh receptors currents. Responses were evoked by application of 1 mM ACh. Coapplication of 20 μM SEP-174559 reduced the magnitude of ACh-evoked currents by 35 ± 2% (n = 12) (Fig.4). Similar inhibition by 34 ± 3% (n = 8) was produced by 20 μM (R,S)-zopiclone. Higher concentrations of both drugs (100 μM) further reduced ACh-evoked responses by 74 ± 2 (n = 7) and 71 ± 2% (n = 3), respectively. The mechanism of block by SEP-174559 and (R,S)-zopiclone was further examined and found to be noncompetitive in that the magnitude of inhibition was not dependent on agonist concentration (Fig. 5C). In three cells tested at various membrane potentials, the inhibition also was not voltage-dependent (Fig. 5, A and B). Together, these data suggest an allosteric mechanism of action by which drug binding affects channel opening or conductance rather than an open-channel blocking mechanism. Because such inhibition of nicotinic acetylcholine receptors (nAChRs) by benzodiazepine-site ligands has not been reported previously, it was important to determine whether this effect was unique to cyclopyrrolones. It was not. Coapplication of 20 μM alprazolam inhibited 1 mM ACh-evoked currents by 78 ± 3% (n = 4). Likewise, 2 and 20 μM diazepam inhibited ACh-evoked currents by 21 ± 2 and 94 ± 1% (n = 5), respectively, suggesting an inhibitory potency (IC50) on the order of 5 to 10 μM under these conditions.
Drug Actions on Receptors in Glutamate Receptor Superfamily.
The activity of SEP-174559 at NMDA-type glutamate receptors was tested by whole-cell patch-clamp recording in cells expressing the NR1/2B receptor subtype. NMDA receptor currents were examined in Mg2+-free extracellular solution and stimulated by application of 100 μM glutamate plus 10 μM glycine. Coapplication of 20 μM SEP-174559 reduced the magnitude of agonist-evoked currents by 28 ± 4% (n = 5) (Fig.6). Similar inhibition by 23 ± 5% (n = 5) was produced by 20 μM (R,S)-zopiclone. In contrast, SEP-174559 was inactive at AMPA- and kainate-type glutamate receptors. These were tested by examining glutamate-evoked responses in outside-out patches taken from cells expressing the GluR1 AMPA receptor or GluR6 kainate receptor subtypes. Responses were evoked by ultrafast application of 1 mM glutamate. Coapplication of 20 μM SEP-174559 after preexposure to the drug did not alter the peak amplitude or decay time course of either GluR1 (n = 3) or GluR6 (n = 3) receptor currents (Fig. 6).
Discussion
The actions of SEP-174559 were examined at a variety of neurotransmitter receptor ion-channel subtypes. As expected, a benzodiazepine-like enhancement of GABAA receptor currents was confirmed. The facilitation of GABAAresponses by SEP-174559 required the presence of the γ2 subunit, which is prevalent in vivo (Laurie et al., 1992; Wisden et al., 1992;Fritschy and Mohler, 1995) and accounts for most high-affinity benzodiazepine binding (Gunther et al., 1995). In contrast, SEP-174559 was not specific for any particular α subunit. Rather, SEP-174559 was broadly active at α1-, α2-, and α3-bearing receptor subtypes. Estimated Kd values for these receptors were all in the range of 1 to 6 μM, and the mechanism of action of SEP-174559 at GABAA receptors was found to involve an allosteric enhancement of the affinity for GABA binding as has been proposed for other benzodiazepine-site ligands.
It is perhaps not surprising that modulation by (R,S)-zopiclone was also not α-subtype specific. Indeed, (R,S)-zopiclone modulation has been demonstrated for the α1β2γ2 subtype but also for β2γ2 absent an α subunit (Im et al., 1993), suggesting the α subunit is not important for (R,S)-zopiclone modulation. Yet, paradoxically, (R,S)-zopiclone binding has been shown to involve the α-subunit His101 residue, which is a necessary component of the benzodiazepine binding site (Davies et al., 2000). Likewise, our studies also indicate at least some involvement of the α subunit in (R,S)-zopiclone binding because Kd estimates from kinetic analyses of α1-, α2-, and α3-bearing subtypes differed by as much as 7-fold between α1-bearing (highest affinity) and α3-bearing (lowest affinity) subtypes. The present study also confirms that (R)- and (S)-zopiclone enantiomers are both active at GABAA receptors, whereas (S)-zopiclone has relatively higher affinity and likely accounts for much of the action of racemic zopiclone.
In general, SEP-174559 and (R,S)-zopiclone had similar effects on the various receptor subtypes tested. However, in all measures of α1-, α2-, and α3-bearing GABAA receptors, the enhancement produced by (R,S)-zopiclone was consistently greater than that produced by SEP-174559 at equimolar concentrations, and kinetic analyses indicated a 5- to 10-fold higher apparent affinity (lowerKd) of these receptors for (R,S)-zopiclone compared with SEP-174559. The relatively greater enhancement by (R,S)-zopiclone might reflect its higher affinity for GABAAreceptors or a “superagonist” action at the benzodiazepine site as has been suggested previously (Davies et al., 2000). However, because no enhancement is seen at saturating GABA concentrations, it is likely that these drugs differ only in their affinity for binding to GABAA receptors. Thus, the fact that these drugs have similar actions raises two important questions regarding their in vivo effects: 1) might some of the actions of (R,S)-zopiclone in fact result from its metabolites, including SEP-174559? and 2) why is SEP-174559 the more potent and selective anxiolytic agent? Unfortunately, the present study does not answer these questions except to demonstrate the lack of α subtype specificity of both drugs. Satisfactory answers will require additional pharmacokinetic and pharmacological studies.
With regards to the first question, the available data suggest that (R,S)-zopiclone and SEP-174559 have similar molecular mechanisms of action. Given the relatively greater potency of (R,S)-zopiclone compared with SEP-174559, it seems unlikely that the lower affinity metabolite could account for the anxiolytic or anticonvulsant effects of (R,S)-zopiclone in vivo without producing similar impairments in locomotor activity or motor coordination (Carlson et al., 2001). However, in relating drug affinities in vitro to effective doses in vivo, one should like to know something about the bioavailability, plasma concentrations, and brain penetration of the drugs in question. In this regard, zopiclone is rapidly absorbed, distributed in various tissues, including brain, and achieves plasma concentrations of 30 to 90 ng ml−1 (100–300 nM) after therapeutic doses (Gaillot et al., 1983), which is comparable with its effective concentration at GABAAreceptors in vitro (Table 2; Reynolds and Maitra, 1996; Davies et al., 2000). There are as yet no comparable pharmacokinetic data on SEP-174559, but it may be important to consider that SEP-174559 is far more soluble in aqueous solution than zopiclone or benzodiazepines and so might differ dramatically in its ability to act at central GABAA receptors. If so, this could help to explain why two drugs with apparently 10-fold different affinities are generally effective at similar doses, but cannot explain the apparent selectivity of SEP-174559 in behavioral studies to produce anxiolytic and anticonvulsant effects without also producing sedation as does (R,S)-zopiclone (Carlson et al., 2001).
This bears on the second question. Benzodiazepines are thought to exert their sedative and anticonvulsant effects through α1-bearing GABAA receptors (Rudolph et al., 1999; McKernan et al., 2000) and their anxiolytic effects through α2-bearing receptors (Low et al., 2000). Clearly, SEP-174559 is not selective for α2-bearing GABAA receptors. Neither is (R,S)-zopiclone. Both drugs are active at α2-bearing receptors, but also at α1- and α3-bearing receptors. Based on our current understanding then, both drugs would be expected to produce anxiolytic effects (via α2), anticonvulsant effects (via α1), and sedation (via α1). (R,S)-Zopiclone fulfills all of these expectations, whereas SEP-174559 fails to produce sedation except at very high doses (50- to 100-fold higher than required for anxiolytic or anticonvulsant effects) (Carlson et al., 2001). Together, these observations suggest that other (i.e., not α1-, α2-, nor α3-bearing) receptors must be involved in the sedative effects of benzodiazepine ligands. These effects might involve other GABAA subtypes or receptors outside the GABAA receptor family. In addition to the commonly appreciated enhancement of GABAAreceptors, our studies indicate that SEP-174559, (R,S)-zopiclone, and other classical benzodiazepines inhibit γ1-bearing GABAAreceptors, nACh receptors (α3β4 subtype), and NMDA receptors (NR1/2B subtype). Such inhibitory effects have not been reported previously, and their significance to the main effects or side effect profiles of benzodiazepine-site ligands is unclear. These actions are relatively low potency. However, it may be prudent to consider whether these or other similar actions might be related to the sedative, amnesic properties, or addictive potential of benzodiazepine ligands. As such, it may be informative in future drug development efforts at least to assess the actions of benzodiazepine-site ligands at nACh and NMDA receptors.
Acknowledgments
I thank Elizabeth Cornell, Keri Cannon, and Dr. Stephanie Mah for technical contributions to this work, Dr. Lindsay Hough for comments on the manuscript, Drs. Peter Seeburg, Mark Mayer, Stephen Heinemann, David Julius, Elias Aizenman, and Shigetada Nakanishi for the various receptor subunit cDNAs, and Sepracor for SEP-174559 and the zopiclone enantiomers.
Footnotes
-
This work was supported by the Albany Medical College strategic research initiative, the Henry Schaffer Foundation, and a grant from Sepracor.
-
DOI: 10.1124/jpet.102.033886
- Abbreviations:
- GABA
- γ-aminobutyric acid
- nACh
- nicotinic acetylcholine
- 5-HT3
- serotonin type-3
- SEP-174559
- (S)-desmethylzopiclone
- HEK
- human embryonic kidney
- Glu
- glutamate
- AMPA
- α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid
- nAChR
- nicotinic acetylcholine receptor
- NMDA
- N-methyl-d-aspartate
- ACh
- acetylcholine
- Received February 25, 2002.
- Accepted March 27, 2002.
- The American Society for Pharmacology and Experimental Therapeutics