High-affinity [3H] kainic acid binding to brain membranes

J Pharmacol Toxicol 42 (1999) 121–125

High-affinity [3H] kainic acid binding to brain membranes: a re-evaluation of ligand potency and selectivity Nicola Crawford, Tracey K. Lang, D. Steven Kerr*, David J. de Vries Department of Pharmacology, University of Otago Medical School, P.O. Box 913, Dunedin, New Zealand Received August 11, 1999; accepted January 19, 2000

Abstract [3H]Kainic acid ([3H]KA) is a widely used tool for studying the KA class of excitatory amino acid receptors. [3H]KA of significantly higher specific activity has become available permitting use of radioligand concentrations below the dissociation constant (KD) of the high-affinity binding site. We employed low radioligand (0.05–0.2 nM) and receptor concentrations (0.01 nM) to gain new insights into the binding characteristics of the high-affinity KA binding site in a standard preparation of lyzed synaptosomal membranes from the cerebral cortex of male Sprague-Dawley rats. Under these conditions, KA binds to a single class of high-affinity sites with a KD of 1.0 ⫾ 0.3 nM. The potencies of competing agents are considerably higher than published reports. Specifically, domoic acid, glutamate, and glutamine exhibit IC50 values for displacing [3H]KA of 0.37 ⫾ 0.02, 94 ⫾ 13, and 1500 ⫾ 500 nM, respectively. Domoate (1 ␮M) was tested against a panel of 32 central nervous system binding sites and found to be inactive at each, indicating this toxin displays considerable selectivity. This study illustrates the remarkable potency of domoic acid and underlines the importance of performing radioligand binding studies at concentrations of constituents that permit characterization of high-affinity interactions. © 2000 Elsevier Science Inc. All rights reserved. Keywords: Kainic acid; Domoic acid; In vitro assay; Radioligand binding

1. Introduction Kainic acid (KA) is the prototypical agonist for the KA class of excitatory amino acid receptors. The initial study with [3H]KA by Simon and coworkers (Simon et al., 1976) featured [3H]KA with a specific activity of 0.4 Ci/mmol, included at a routine concentration of 40 nM. They identified specific [3H]KA binding in several sites within the mammalian central nervous system (CNS) with a KD for KA of 60 nM. Subsequently, London & Coyle (1979) used 4 Ci/mmol [3H]KA and were able to identify an additional, higher-affinity binding site with a KD of 4 to 16 nM. More recently, studies on the molecular biology of ionotropic glutamate receptors have identified several receptor clones displaying a range of pharmacologic properties. These may be grouped into families featuring selective high-affinity and selective low-affinity KA binding and AMPA (2-amino-3-hydroxy5-methyl-isoxazole-4-propionate)/KAnonselectivity. Additional sources of receptor diversity are introduced by the propensity of receptor subtypes to form heteromeric channels and to undergo post-transcriptional editing (reviewed in Hollman & Heinemann, 1994; Jorgensen et al., 1995; Bleakman & Lodge, 1998). Reconciling the receptor subtypes identi* Corresponding author. E-mail address: [email protected] (D.S. Kerr).

fied by molecular biological techniques with functional data is proving a slow process hampered by a paucity of receptor selective compounds. A key question in excitatory amino acid research is the identification of the site at which agonists exert their excitotoxicity. One of the most potent excitotoxins is domoic acid (DOM), which has been responsible for dramatic cases of human (Teitelbaum et al., 1990) and wildlife intoxication. DOM acts primarily at the KA class of glutamate receptors, and there has been a working assumption that this is the site at which both it and KA initiate their toxic effects. Increasingly, this assumption is being questioned due to the disparity in relative potencies between binding and functional studies. DOM has been reported to be 1.5 to 2-fold more potent in competing for high-affinity [3H]KA binding than KA (Slevin et al., 1983; Hampson et al., 1992). In contrast, DOM exerts effects in in vivo toxicity studies at 8-fold and in electrophysiologic studies at 20-fold greater potency than KA (Huettner, 1990; Tasker et al., 1991). Tasker and coworkers (1996) argue that the high-affinity binding site is not relevant to the toxicity of DOM. Rather, it is a lowaffinity glutamate receptor, antagonized with some selectivity by the compound NS-102, which mediates these effects (Tasker et al., 1996; Johansen et al., 1993). Therefore, one goal of this study was to closely examine the relative potency of DOM at the high-affinity [3H]KA binding site and

1056-8719/00/$ – see front matter © 1999 Elsevier Science Inc. All rights reserved. PII: S 1 0 5 6 - 8 7 1 9 ( 0 0 ) 0 0 0 4 0 - X

122

N. Crawford et al. / J Pharmacol Toxicol 42 (1999) 121–125

ascertain if potent effects of DOM have been overlooked through methodological problems. Another explanation for the potent toxicity of DOM is that it exerts additional effects at sites that have yet to be described. There is evidence from both in vivo and in situ studies that DOM acts at a site independent of KA (Tasker & Strain, 1992; Xi & Ramsdell, 1996). To explore this possibility, we undertook broad screening of DOM against a range of radioligand binding sites. Specific activity of a radioligand, and hence radioligand concentration, is a limitation in accurately estimating the potencies of labeled and unlabeled ligands, particularly when working with tritiated compounds. This difficulty has been described in many studies using both theoretical and practical systems (Chang et al., 1975; Goldstein & Barrett, 1987; de Vries et al., 1988). Disregard of this practical limitation typically leads to gross underestimates of the potency of labeled and unlabeled ligands. We have re-evaluated the binding parameters of [3H]KA using the currently available specific activity of 58 to 60 Ci/mmol, together with low concentrations of radioligand and receptor, in order to gain a better understanding of this binding site.

2. Methods Radioligand binding studies were performed using standard techniques as described previously (Ryan et al., 1993). Male Sprague-Dawley Rats were sacrificed by CO2 asphyxiation, cerebral cortices (including hippocampi) dissected, and washed synaptosomal membranes prepared by differential centrifugation and repeated freeze-thawing. In an assay buffer of 50 mM Tris-HCl, pH 7.4 at 22⬚C, [3H]KA (58 to 60 Ci/mmol; NEN) and membranes were incubated at room temperature for 60 min in the presence and absence of a range of inhibitors, in triplicate. One hundred percent inhibition was defined as that observed in the presence of 100 ␮M L-glutamate. Under routine conditions, [3H]KA was employed at concentrations of 0.1 to 0.2 nM and membranes at a concentration of 35 ␮g protein/mL, in a final volume of 5 mL. Incubations were terminated by vacuum filtration on glass fibre filters (Whatman GF/B; Maidstone, England). Filters were washed with 20 mL of assay buffer and placed in vials with scintillant (BCS; Amersham) containing 10% H2O for 10 to 14 h before counting at 33% efficiency. Radioligand binding data were analyzed and modeled using two programs, Prism (GraphPad, San Diego, CA, USA) and EBDA/LIGAND (McPherson, 1985; Munson and Rodbard, 1980). Results obtained with both were similar and values presented here were obtained using the former. All data are presented as mean ⫾ SEM, unless otherwise indicated. 2.1. Materials Kainic acid and glutamine were obtained from Sigma Chemical Company (St. Louis, MO, USA). DOM, NS-102, cyano-7-nitro-quinoxaline-2,3-dione (CNQX), quinolinic acid, (⫾)-4-bromohomoibotenic acid, AMPA, aniracetam,

and cyclothiazide were obtained from Research Biochemicals International (Natick, MA, USA). L-glutamate and cysteine were purchased from BDH Chemicals Ltd. (Poole, UK). 5,7-dichlorokynurenic acid and 1S, 3R, ACPD were obtained from Tocris (Bristol, UK) and MK-801 from Merck Reagents (Gibbstown, NJ, USA). Aspartic acid (Sigma) and L-AP4 (Tocris) were gifts from R. Mills and R. Sayer, respectively (Department of Physiology, University of Otago, Dunedin, New Zealand). All other chemicals were of research grade.

3. Results The effect of dilution on the inhibition profile of KA is illustrated in Fig. 1. Using [3H]KA at a concentration of 1 nM and varying the unlabeled KA concentration yielded a displacement curve with an IC50 of 2.4 ⫾ 0.4 nM, a slope of 0.97 ⫾ 0.09, and a calculated receptor concentration of 4.47 ⫾ 0.05 pM (n ⫽ 3). Increasing the assay volume by a factor of 10, thus decreasing both radioligand and receptor concentrations, resulted in a decrease in binding parameters to an IC50 of 1.1 ⫾ 0.1 nM, a slope of 0.81 ⫾ 0.08, and a receptor concentration of 1.56 ⫾ 0.19 pM (n ⫽ 3). Expressed as a dissociation constant (KD), this value corresponds to 1.0 ⫾ 0.3 nM and the receptor density may be expressed as 44.3 ⫾ 5.5 fmol/mg protein. Under these latter assay conditions, total radioactivity associated with the filter was 303 ⫾ 10 to 685 ⫾ 13 cpm, representing 0.2% to 0.8% of the radioligand added. Of the filter associated radioactivity, 23 ⫾ 12 to 97 ⫾ 30 cpm was not displaceable by 100 ␮⌴ L-glutamate. In all cases, greater than 85% specific binding was observed. These assay conditions were subsequently employed as standard conditions.

Fig. 1. Inhibition of [3H]KA (0.1 nM, 䊏; 1.0 nM, 䉱) binding from rat cortical membrane preparation by KA. Different concentrations of KA were incubated with the membrane preparation and [3H]KA at room temperature for 60 min, and binding terminated by vacuum filtration. Data are presented as mean ⫾ SEM of 3 experiments, with each experiment performed in triplicate.

N. Crawford et al. / J Pharmacol Toxicol 42 (1999) 121–125

Under standard conditions, the inhibition profiles of DOM, glutamate, and glutamine were assessed (Fig. 2; Table 1). Each compound in this series could displace all specific binding as defined with 100 ␮M L-glutamate. Detailed analysis of this data was undertaken using both individual experiments and pooled data, and computerized nonlinear curve fitting was employed to provide the values presented. The corresponding IC50 obtained for DOM, glutamate, and glutamine are 0.37 ⫾ 0.02, 94 ⫾ 13, and 1500 ⫾ 500 nM, respectively. Although a minority of individual DOM competition experiments were best fit by a multisite model, this trend was not consistent and single site analysis provided superior fits for all other compounds tested. Table 2 presents further pharmacologic characterization of [3H]KA binding against a range of compounds at 100 ␮M. Although most had no effect on [3H]KA binding, aspartic acid did produce moderate inhibition. In contrast, both AMPA and 4-bromohomoibotenic acid exhibited strong inhibitory behavior. These AMPA agonists reduced [3H]KA binding to 26.9% and 29.5% of control values, respectively. All drugs were soluble in the recommended solvents. However, NS-102, which dissolved readily at 1.3 mg/ mL DMSO, came out of solution on addition to the assay. At lower concentrations, NS-102 also showed no effect (data not shown). To explore the possibility that DOM interacts with sites that have, as yet, gone undescribed, we contracted a broad screening of DOM against a suite of validated radioreceptor binding assays (Cerep, France). At a concentration of 1 ␮M, domoic acid did not significantly inhibit radioligand binding at any of the 32 sites examined: adenosine (A1, A2); adrenoceptors (␣1, ␣2, ␤1); angiotensin (AT1); benzodiazepine (central); cholecystokinin (CCKA); dopamine (D1, D2); dopamine uptake site; GABA (nonselective); glucocorticoid (nonselective); glycine (strychnine- insensitive); his-

123

Table 1 Inhibition profiles of various compounds at the high-affinity [3H]KA binding site employing subnanomolar concentrations of radiolabel and receptor

Domoic acid Kainate L-glutamate Glutamine

KI (nM)

Slope

0.30 ⫾ 0.05 1.05 ⫾ 0.38 85 ⫾ 15 1365 ⫾ 589

0.77 ⫾ 0.06 0.81 ⫾ 0.08 1.17 ⫾ 0.12 0.97 ⫾ 0.17

Each drug was tested with at least 11 concentrations of inhibitor in triplicate and results are the mean ⫾ SEM of three experiments.

tamine (H1); muscarinic (nonselective); neurokinin (NK1); neuropeptide Y (nonselective); nicotinic (nonselective); NMDA (CGS19755); opiate (nonselective; ␴); phencyclidine; 5-hydroxytryptamine (nonselective; 5-HT1B; 5-HT2A; 5-HT uptake); vasopressin (V1); Ca⫹2 channel (N-type); K⫹ channel (Ca⫹2-dependent); K⫹ channel (voltage-dependent). 4. Discussion In order to obtain accurate estimates of the potency of ligands in any radioligand binding study, it is essential to employ concentrations of constituents (radioligand and receptor) that are less than the KD of the ligand under examination. Our assay conditions are typical of numerous studies with [3H]KA with respect to membrane preparation and buffer. However, by employing subnanomolar concentrations of radioligand and receptor, we have obtained inhibitory potencies for KA and other compounds at significantly lower concentrations than previously reported. Recent published results for the IC50 of KA receptors in brain membranes are in the range 5 to 20 nM (Hampson et al., 1992; Smith & McIlhinney, 1992; Zhou et al., 1997). We attribute

Table 2 Mean % specific binding of 0.1 nM [3H]KA in the presence of various compounds at 100 ␮M

Fig. 2. Inhibition of [3H]KA (0.1 nM) from rat cortical membranes by domoic acid (䊉), L-glutamate (䊏), and glutamine (䉱). Concentrations of drug on the abscissa represent the final concentration in the assay. Data are represented as the mean ⫾ SEM of three experiments, with each experiment performed in triplicate.

Compound (100 ␮M)

% Specific binding

AMPA 4-bromohomoibotenic acid Aspartic acid 1S, 3R-ACPD Cysteine Quinolinic acid L-AP4 MK-801 Aniracetam Dichlorokynurenic acid Cyclothiazide NS-102

26.9 ⫾ 2.2 29.5 ⫾ 4.3 78.7 ⫾ 3.9 86.1 ⫾ 12 90.6 ⫾ 14 93.5 ⫾ 8.8 100 ⫾ 9.4 102 ⫾ 3.0 106 ⫾ 5.3 110 ⫾ 4.3 114 ⫾ 13 99.0 ⫾ 5.0

Nonspecific binding was determined in the presence of 100 ␮ M L-glutamate. Data are expressed as the mean ⫾ SEM of three experiments, each performed in triplicate.

124

N. Crawford et al. / J Pharmacol Toxicol 42 (1999) 121–125

our result for KA (IC50 1.1 nM, KD 1.0 nM) to avoidance of radioligand competition and ligand depletion artefacts associated with radioligand and receptor concentrations greater than 10% of the relevant KD (Chang et al., 1975; Goldstein & Barrett, 1987). All of the competing ligands displayed higher inhibitory potencies in this system than previously reported. The role of assay conditions in determining the measured inhibitory potency of ligands is especially marked for very potent drugs, particularly when their true dissociation constant is below that of the radioligand and receptor concentrations (Goldstein & Barrett, 1987). Historically, the initial report on the potency of domoic acid at [3H]KA binding sites featured an IC50 of 37 nM using 50 nM [3H]KA (Slevin et al., 1983). More recently, values of 4.9 nM (5 nM [3H]KA; Hampson et al., 1992) and 7.7 nM (2 nM [3H]KA; Johansen et al., 1993) have been typical. Our novel result, an IC50 of 0.37 nM for DOM in the presence of 0.1 nM [3H]KA, is consistent with the increase in apparent potency of a competing agent when the radiolabel is included at a lower concentration than its KD; conditions that are assumed in most experimental protocols and analyses. Application of the Cheng & Prusoff (1973) equation to provide a simple correction permitting calculations of realistic inhibitory constants was hampered in previous studies by overestimation of the KD for [3H]KA. It is possible that domoic acid exerts an effect at the high-affinity KA binding site at concentrations even lower than those reported here. Additional experiments employing assay conditions featuring half of the ligand and receptor concentrations used above failed to identify any further decrease in the KA dissociation constant (data not shown). However, such experiments are technically difficult due to the remaining limitation of the specific activity of [3H]KA and hence radioactivity bound. Access to a radioligand with superior potency and/or specific activity would clarify this question. Given the considerable effects that varying assay conditions exert on results obtained in [3H]KA binding experiments, this study calls for a re-examination of the absolute potency of glutamatergic ligands at the high-affinity KA binding site in many published reports. However, it does serve to confirm the relative potencies published for this site (Toms et al., 1997). In contrast, matching this pharmacologic profile with that obtained with transfected cells expressing individual homomeric receptor subtypes remains difficult. Homomeric high-affinity binding sites do not display more sensitivity to DOM than KA, whereas low-affinity sites are 2 to 30-fold more sensitive (Hollman & Heinemann, 1994). One issue affecting comparisons may be the data gathered on the nominally low-affinity sites (GluR5-7) where the KD was measured using high [3H]KA concentrations. For example, a study on homomeric GluR5 transfected cells employed 50 nM [3H]KA and obtained a KD of 70 nM (Sommer et al., 1992). For reasons outlined in this study, the estimate of affinity for KA in previous studies

may be skewed, such that measured affinity reflects radioligand concentration, rather than the true KD. Some [3H]KA binding studies have obtained shallow inhibition profiles for competitors, suggesting the presence of distinct high- and low-affinity binding sites (Ryan et al., 1993; Hampson et al., 1992; Johansen et al., 1993). The current study featured slopes with values close to unity and data that did not consistently reflect two distinct sites using standard multisite analysis. Inclusion of radioligand concentrations in excess of the true KD, as employed in earlier studies, renders data analysis by standard methods invalid and leads directly to low slopes (Munson & Rodbard, 1980; Goldstein & Barrett, 1987). We suggest it is not possible to employ [3H]KA under a single set of conditions that would provide reliable information about the high-affinity site as well as other sites. In our hands, the high-affinity binding site behaves as an essentially homogenous site; that is, all unlabeled ligands are competing with a single affinity state of [3H]KA. The screening program at 32 CNS binding sites indicates domoic acid displays considerable selectivity. Reference should be made to AMPA receptors as potential targets for DOM and KA. Although both compounds are approximately 1000-fold less potent at the [3H]AMPA binding site, DOM does display a 20-fold higher potency than KA and a nanomolar potency of DOM at a subset of [3H]AMPA sites has been reported (Hampson et al., 1992; Smith & McIlhinney, 1992). Definitive evidence on the specificity and selectivity of DOM would be facilitated by availability of this compound with a radiolabel of high specific activity. The subnanomolar potency of DOM revealed in this study confirms the likelihood that some of this compound’s in vivo effects are mediated by the high-affinity KA binding site. The relevance of this site to the toxic process remains an open question. Our work also supports the proposition that this assay has potential as a highly sensitive screen for this toxin. Acknowledgments We gratefully acknowledge the advice and useful editorial comments by Dr. Ilona Kokay, Dept. of Anatomy and Structural Biology, University of Otago School of Medical Sciences.

References Bleakman, D., & Lodge, D. (1998). Neuropharmacology of AMPA and kainate receptors. Neuropharmacology 37, 1187–1204. Chang, K. J., Jacobs, S., & Cuatrecasas, P. (1975). Quantitative aspects of hormone-receptor interactions of high affinity: Effect of receptor concentration and measurement of dissociation constants of labelled and unlabelled hormones. Biochim Biophys Acta 406, 294–303. Cheng, Y. C., & Prusoff, W. H. (1973). Relationship between the inhibition constant (KI) and the concentration of inhibitor which causes 50 per cent inhibition (I50) of an enzymatic reaction. Biochem Pharmacol 22, 3099–3108.

N. Crawford et al. / J Pharmacol Toxicol 42 (1999) 121–125 de Vries, D. J., Herald, C. L., Pettit, G. R., & Blumberg, P. M. (1988). Demonstration of sub-nanomolar affinity of bryostatin 1 for the phorbol ester receptor in rat brain. Biochem Pharmacol 37, 4069–4073. Goldstein, A., & Barrett, R. W. (1987). Ligand dissociation constants from competition binding assays: Errors associated with ligand depletion. Mol Pharmacol 31, 603–609. Hampson, D. R., Huang, X. P., Wells, J. W., Walter, J. A., & Wright, J. L. (1992). Interaction of domoic acid and several derivatives with kainic acid and AMPA binding sites in rat brain. Eur J Pharmacol 218, 1–8. Hollman, M., & Heinemann, S. (1994). Cloned glutamate receptors. Ann Rev Neurosci 17, 31–108. Huettner, J. E. (1990). Glutamate receptor channels in rat DRG neurons: activation by kainate and quisqualate and blockade of desensitization by Con A. Neuron 5, 255–266. Johansen, T. H., Drejer, J., Watjen, F., & Nielsen, E. O. (1993). A novel non NMDA receptor antagonist shows selective displacement of low affinity [3H]kainate binding. Eur J Pharmacol 246, 195–204. Jorgensen, M., Tygesen, C. K., & Andersen, P. H. (1995). Inotropic glutamate receptors—focus on non-NMDA receptors. Pharmacol Toxicol 76, 312–319. London, E. D., & Coyle, J. T. (1979). Specific binding of [3H]Kainic acid to receptor sites in rat brain. Mol Pharmacol 15, 492–505. McPherson, G. A. (1985). Analysis of radioligand binding experiments: A collection of computer programs for the IBM PC. J Pharmacol Methods 14, 213–228. Munson, P. J., & Rodbard, D. (1980). LIGAND: A versatile computerized approach for characterization of ligand-binding systems. Anal Biochem 107, 220–239. Ryan, M., Beart, P. M., & de Vries, D. J. (1993). Extracts of marine biota with bioactivity at glutamate receptor subtypes. Pharmacol Comm 3, 225–232. Simon, J. R., Contrera, J. F., & Kuhar, M. J. (1976). Binding of [3H]Kainic

125

acid, an analogue of L-glutamate, to brain membranes. J Neurochem 26, 141–147. Slevin, J. T., Collins, J. F., & Coyle, J. T. (1983). Analogue interactions with the brain receptor labeled by [3H]Kainic acid. Brain Res 265, 169–172. Smith, A. L., & McIlhinney, R. A. (1992). Effects of acromelic acid A on the binding of [3H]-kainic acid and [3H]-AMPA to rat brain synaptic plasma membranes. Br J Pharmacol 105, 83–86. Sommer, B., Burnashev, N., Verdoorn, T. A., Keinanen, K., Sakamann, B., & Seeburg, P. H. (1992). A glutamate receptor channel with high affinity for domoate and kainate. EMBO J 11, 1651–1656. Tasker, A., & Strain, S. M. (1992). Morphine differentially affects domoic acid and kainic acid toxicity in vivo. Neuroreport 3, 789–792. Tasker, R. A., Connell, B. J., & Strain, S. M. (1991). Pharmacology of systemically administered domoic acid in mice. Can J Physiol Pharmacol 69, 378–382. Tasker, R. A., Strain, S. M., & Drejer, J. (1996). Selective reduction in domoic acid toxicity in vivo by a novel non-N-methyl-D-aspartate receptor antagonist. Can J Physiol Pharmacol 74, 1047–1054. Teitelbaum, J. S., Zatorre, R. J., Carpenter, S., Gendron, D., Evans, A. C., Gjedde, A., & Cashman, N. R. (1990). Neurologic sequelae of domoic acid intoxication due to the ingestion of contaminated mussels. N Engl J Med 322, 1781–1787. Toms, N. J., Reid, M. E., Phillips, W., Kemp, M. C., & Roberts, P. J. (1997). A novel kainate receptor ligand [3H]-(2S,4R)-4-methylglutamate: pharmacological characterization in rabbit brain membranes. Neuropharmacol 36, 1483–1488. Xi, D., & Ramsdell, J. S. (1996). Glutamate receptors and calcium entry mechanisms for domoic acid in hippocampal neurons. Neuroreport 7, 1115–1120. Zhou, L. M., Gu, Z. Q., Costa, A. M., Yamada, K. A., Mansson, P. E., Giordano, T., Skolnick, P., & Jones, K. A. (1997). (2S,4R)-4-methylglutamic acid (SYM 2081): a selective, high-affinity ligand for kainate receptors. J Pharmacol Exp Ther 280, 422–427.