The excitotoxicity of heterocyclic dicar☐yclic acids in rat hippocampal slices: structure-activity relationships

Brain Research, 571 (1992) 145-148 © 1992 Elsevier Science Publishers B.V. All rights reserved. 0006-8993/92/$05.00

145

BRES 25010

The excitotoxicity of heterocyclic dicarboxylic acids in rat hippocampal slices: structure-activity relationships Avital Schurr, C.A. West and B.M. Rigor Department of Anesthesiology, University of Louisville School of Medicine, Louisville, KY 40292 (U.S.A.) (Accepted 15 October 1991) Key words: Heterocyelie diearboxylate; Hippocampal slice; Hypoglycemia; Hypoxia; Neuronal damage; N-Methyl-D-aspartateagonist

The structural resemblance of certain heterocyclic diearboxylates to aspartate and glutamate led investigators to study their potency as agonists and antagonists of the N-methyl-n-aspartate (NMDA) receptor. The sensitivity of hypoxie rat hippocampal slices to NMDA ligands is several fold greater than that of normoxie slices. In the present study, the excitotoxic potency of heterocyclie dicarboxylates was assessed electrophysiologieally by measuring their ability to enhance hypoxie and hypoglycemic neuronal damage in the rat hippocampal slice preparation. Four compounds were tested: qninolinate (QUIN), 4,5-imidazole-diearboxylate (IZDA), 1,2,3-triazole-4,5-dicarboxylate (TZDA), and 2,3-pyrazinediearboxylate (PZDA). QUIN was the most toxic drug in enhancing both hypoxic and hypoglycemie neuronal damage. IZDA and TZDA were slightly less toxic than QUIN, while PZDA was innocuous. The effect of the 3 active drugs was blocked by the NMDA competitive antagonist DL-2-amino-5-phosphonovalerate.The sequence -N-CH(COOH)-CH(COOH)- appears to be a prerequisite for a heterocyelie diearboxylate to exert NMDA-type agonistie properties. A 5-membered ring heterocyelie compound which contains more than one nitrogen atom in its ring retains its NMDA-type toxicity while a 6-membered ring with more than one nitrogen atom (PZDA) does not. Investigators, hoping to unravel the structure-binding relationships of ligand-receptors, are frequently faced with the painstaking task of testing numerous chemically related compounds once a ligand for a given receptor has been described. The logic behind this process is the postulate that the interact{on between the receptor and its agonist or antagonist resembles that of a lock and a key. In the 1970's, the most investigated receptor system was that of the opiates. In the 1980's, the glutamatergic receptor system was the center of attention. The excitotoxic hypothesis s is the main reason for the continuing interest in this excitatory neurotransmitter system today. The premise of this hypothesis holds a major role for excitatory amino acids (EAA) in the mechanism of various brain disorders, including cerebral ischemia-hypoxia and hypoglycemia, epilepsy, Alzheimer's disease, and others, To date several receptor subtypes of glutamate (GLU) have been described 6A2 of which the N-methyl-D-aspartate (NMDA) is probably the most studied. The resemblance of certain cyclic dicarboxylates such as quinolinate (QUIN), c/s-, and trans-2,3-piperidinedicarboxylate, c/s2,4-piperidinedicarboxylate, and c/s-2,3-piperazinedicarboxylate to NMDA, aspartate (ASP), and G L U has led investigators to study their potency as agonists and antagonists of the N M D A receptor L3. The outcome of

these studies highligths the importance of careful selection not only of the compounds one should test but also of the system in which they are to be tested. While the fLrst study found c/s-2,3-piperidinedicarboxylate to be an antagonist that blocks NMDA-evoked excitation in neurons of the rat cerebral cortex in vivo 1, the second study, using assays for binding and uptake sites of acidic amino acids, showed this compound to be an agonist of the same receptor 3. Using the rat hippocampal slice preparation, we have recently demonstrated the ability of ASP, G L U , and N M D A to enhance hypoxic and hypoglycemic neuronal damage 9. Our findings confirmed earlier observations in cell cultures regarding the dependence of N M D A agonists' excitotoxicity on neuronal energy levels2'4'5'7. In a study just completed 1~ we assessed the potencies of 2,3pyridinedicarboxylate (Quinolinate, QUIN) and 6 other pyridinedicarboxylates as excitotoxins employing the energy-compromised (hypoxic) hippocarnpal slice preparation. This hypersensitized system was able to discriminate between seemingly subtle modifications in either the position or the nature of the side groups attached to the pyridine ring. The aim of the present study was to increase our understanding of the agonistic and antagonistic structural requirements of the N M D A receptor. We employed the

Correspondence: A. Schurr, Department of Anesthesiology, University of Louisville School of Medicine, Louisville, KY 40292 U.S.A.

146 energy-compromised slice preparation to assess the potency of 3 heterocyclic dicarboxylates by comparing their excitotoxieity to that of QUIN. All 4 of these compounds differ from each other in their ring structure rather than in the position or the nature of their side groups, Adult (200-350 g) male Sprague-Dawley rats were used. For each experiment one rat was decapitated and its brain rapidly removed, rinsed with cold (4°C) artificial cerebrospinal fluid (ACSF, see composition below) and dissected. Isolated hippocampi were sliced transversely at 400 gm with a tissue chopper and the resulting s l i c e s were placed in a dual, linear-flow incubation/recording chamber 1°. Each of the two compartments of the dual chamber was supplied with a humidified gas mixture (95% 02/5% CO2) through separate flow meters ( 1 . 5 1/min) and ACSF via a dual peristaltic pump (40 ml/h), Hypoxia was produced by replacing 0 2 in the gas mixture with N 2. The ACSF contained the following (in raM): NaC1 124, KCI 5, NaH2PO 4 3, CaCI2 3.0, MgSO 4 2.0, NaHCO3 23, o-glucose 10, pH 7.3-7.4. Hypoglycemia was produced by perfusing slices with ACSF containing 0.7 mM glucose, instead of the regular 10 mM and 128 mM NaC1, to assure iso-osmolarity. Where indicated, the ACSF was supplemented with QUIN ( 2 , 3 pyridinedicarboxylate; OL-2-amino-5-phosphonovalerate (APV) (all from Sigma Chem. Co., St. Louis, MO); 2,3-pyrazinedicarboxylate (PZDA); 4,5-imidazoledicarboxylate (IZDA); or 1,2,3-triazole-4,5-dicarboxylate (TZDA) (all from Aldrich Chemical Co. Milwaukee, WI). The temperature of the incubation chamber was held at 34 + 0.3°C. Extracellular recordings of population responses in the

stratum pyramidale of the CA1 region were made as described elsewhere TM. CAl-evoked population spikes (neuronal function) were recorded automatically once/ min from a single slice in each compartment of the dual chamber through the whole duration of the experiment. At the end of each experiment, the rest of the slices in each compartment (10-15) were tested for the presence of neuronal function by stimulating the Schaffer collaterals and by recording from the CA1 cell body layer. A waveform analysis program was used to determine the amplitude of the evoked population spike. Any slice exhibiting a population spike of >3 mV was considered to be functional. Slices in which a population spike could not be evoked or in which the spike amplitude was smaller than 3 mV at the end of the experiment were considered neuronally damaged9. Each experiment con-

TABLEI Proportions o f rat hippocampal slices that exhibited neuronal function after treatment with several heterocyclic dicarboxylic acids in the presence or absence of oL-APV, hypoxia or hypoglycemia and 30-rain recovery period Statistical analysis was done using the Z2-test for significant differences. Treatment

Number o f slices

(recovered/

total)

Hypoxia None (control)

Percent

recovered

132/155

85

44/93*

47

33/37 n.s.

89

IZDA (0,5 mM) + APV (0.5 mM)

11/13 n.s.

85

TZDA

12/28"

43

TZDA (0.5 mM) + APV(0.5 mM)

11/13 n.s.

85

PZDA

19/26 n.s.

73

N

QUIN

C

COOfl

~1

(0.1 raM) C008

QUIN (0.1 mM) + APV (0.2 raM) B N

lOO

o

•

Z

!

QUIN

<> IZDA 80 ~

[]

60

TZDA

©

~ H

(0.5 mM) ~i

>" 40 ~

COOn

,

0.,

~

C~B

\

20 10 I i0 z NMDARECEPTORAGONIST (/aM)

I~ N

~"

I0 0

f'~

coos

(3.0 m M ) • COOS

I0 a

Fig. 1. The dose-dependent effect of quinolinate (QUIN), 4,5-imidazole-diearboxylate (IZDA), and 1,2,3-triazole-4,5-dicarboxylate (TZDA) on the rate of recovery of rat hippocampal slices from 10rain hypoxia. The dose-dependent effect of N-methyl-D-aspartate (NMDA) is shown for comparison. Slices were exposed to each dose of the drugs for 40 min (30 min pre-hypoxia and 10 min during hypoxia) and then allowed to recover for a 30-min washout/ reoxygenation period,

Hypoglycemia None (control) QUIN (0.2 mM) IZDA (0.5 mM) TZDA (0.5 mM) PZDA (3.0 mM)

76/90 12/25" 11/25" 14/25"* 22/26 n.s.

84 48 44 56 85

* Significantly different from control (P < 0.0005); n.s., not significantly different from control; ** significantly different from control (0.005 < P < 0.01).

147 sisted of a 15-min baseline activity period, either a 40rain drug treatment period (its last 10 rain combined with hypoxia) or a 75-rain hypoglycemic period (its last 40 rain combined with drug perfusion), and a 30-rain washout (recovery) period. The experimental manipulations were performed on slices in one compartment of the dual chamber while slices in the other compartment served as 'controls '9'1°. The incubation of rat hippocampal slices with 100/tM QUIN for 30 rain had no apparent effect on the CA1 population spike. When the concentration of QUIN was doubled, it produced a secondary spike, which might be an indication of the excitatory nature of this compound. The introduction of 500/zM QUIN completely depressed the amplitude of the population spike within 4 min but not before the appearance of short-lived secondary spikes. Both IZDA and T Z D A produced similar effects to those of QUIN, although at somewhat higher concentrations. In contrast, P Z D A was innocuous at all concentrations tested (0.5-3.0 mM). When control slices were exposed to 10 min of hypoxia followed by 30 min of reoxygenation (recovery) period, 80-85% of them recovered their neuronal function (population spike). However, when slices were exposed to hypoxia in the presence of QUIN, IZDA or TZDA, the rate of recovery of neuronal function decreased as the concentration of the drugs was increased from 50 to 1000/~M (Fig. 1). The potency of IZDA and T Z D A in enhancing hypoxic neuronal damage appeared to be lower than that of QUIN (Fig. 1). P Z D A was found to be inactive. All of the 3 active compounds were antagonized by DL-2-amino-5phosphonovalerate (APV), a known competitive antagonist of the NMDA receptor: an effect which indicates that these heterocyclic dicarboxylates are NMDA-type agonists. The use of DL-APV in this study rather than D-APV should explain the relatively high doses that were required for the antagonism of QUIN, IZDA, and TZDA. When slices were exposed to hypoglycemia instead of hypoxia, the outcome was similar; QUIN, IZDA, and T Z D A enhanced hypoglycemic neuronal damage while PZDA was harmless, Table I summarizes the effects of the 4 heterocyclic dicarboxylates on the energy-compromised rat hippocampal slice preparation.

The present study confirms our earlier findings n with regard to the basic structural requirements for an NMDA agonist to interact with its receptor, namely, the sequence -N-CH(COOH)-CH(COOH)-. When this sequence is part of a heterocyclic structure, the carboxylic groups should preferably be positioned one and two carbons, respectively, away from the ring's nitrogen. In an attempt to determine the role of the composition of the heterocyclic ring itself in excitoxicity of NMDA agonists, the structures explored in this study differ from each other by either the number of nitrogen atoms in the ring, their location, or the total number of ring atoms. Thus, QUIN and PZDA differ from each other by the number of nitrogen atoms in their 6-membered ring structure; the first has one such atom and the other has two. The same situation exists between IZDA and TZDA: two 5-membered ring structures, one (IZDA) has two nitrogens, the other (TZDA) has 3. P Z D A and IZDA vary in the total number of atoms (6 and 5, respectively) within their rings. The inactivity of PZDA as compared with the high potency of QUIN indicates the intolerance of the NMDA receptor to a 6-membered ring 2,3-dicarboxylate that contains more than one nitrogen atom. In contrast, the receptor's accomodation of 5-membered ring heterocyclic dicarboxylates, whether they contain two or even 3 nitrogens, is much better as was evident from the equal agonistic potency of IZDA and TZDA. Based on these results, we predict that other 5-membered ring dicarboxylic acids containing either one nitrogen atom in their structure, such as, 2,3-pyrrol-, pyrroline-, and pyrrolidine-dicarboxylate, or two adjacent nitrogens, such as 3,4-pyrazole-, and pyrazoline-dicarboxylate, will be as active as the two tested in this study. As has been demonstrated before 9, NMDA agonists are capable of enhancing both hypoxic and hypoglycemic neuronal damage in the hippocampal slice preparation. Of the heterocyclic dicarboxylates that were tested here, QUIN, IZDA, and T Z D A augmented hypoxic and hypoglycemic neuronal damage whereas P Z D A did neither. These findings indicate that NMDA agonists are involved in the outcome of both hypoxia and hypoglycemia, and that the two neuronal insults share similar cellular mechanisms.

1 Birley, S., Collins, J.E, Perkins, M.N. and Stone, T.W., The effects of cyclic dicarboxylic acids on spontaneous and amino acid-evoked activity of rat cortical neurones, Br. J. Pharmacol., 77 (1982) 7-12. 2 Cox, J.A., Lysko, P.O. and Henneberry, R.C., Excitatory amino acid neurotoxicityat the N-methyl-v-aspartatereceptor in cultured neurons: role of the voltage-dependent magnesium block, Brain Research, 499 (1989) 267-272. 3 Foster, A.C., Collins, J.F. and Schwarcz, R., On the excitotoxic

properties of quinolinic acid, 2,3-piperidine-dicarboxylicacids and structurally related compounds, Neuropharmacology, 22 (1983) 1331-1342. 4 Henneberry, R.C., The role of neuronal energy in the neurotoxicity of excitatory amino acids, Neurobiol. Aging, 10 (1989) 611-613. 5 Lysko, P.G., Cox, J.A., Vigano, M.A. and Henneberry, R.C., Excitatory amino acid neurotoxicityat the N-methyl-D-aspartate receptor in cultured neurons: pharmacologicalcharacterization,

148

Brain Research, 499 (1989) 258-266. 6 Monaghan, D.T., Bridges, R.J. and Cotman, C.W., The e x c i t atory amino acid receptors: their classes, pharmacology, and distinct properties in the function of the central nervous system, Annu. Rev. Pharmacol. Toxicol., 29 (1989) 365-402. 7 Novelli, A., Reilly, J.A., Lysko, EG. and Henncbcrry, R.C., Glutamate becomes neurotoxic via the N-methyl-o-aspartate receptor when intracellular energy levels are reduced, Brain Research, 451 (1988) 205-212. 80lney, J.W., Ho, O.L. and Rhee, V., Cytotoxic effects of acidic and sulfur containing amino acids on the infant mouse central nervous system, Exp. Brain Res., 14 (1971) 61-70. 9 Schurr, A., Lipton, P., West, C.A. and Rigor, B.M., The role of energy metabolism and divalent cations in the neurotoxicity

of excitatory amino acids in vitro. In J. Krieglstein and H. Oberpichler (Eds.), Pharmacology of Cerebral lschemia 1990, Wissenschaftliehe Verlagsgesellschaft, Stuttgart, 1990, pp. 217226. 10 Schurr, A., Reid, K.H., Tseng, M.T., Edmonds, Jr. H.L. and Rigor, B.M., A dual chamber for comparative studies using the brain slice preparations, Comp. Biochem. PhysioL, 82A (1985) 701-704. 11 Schurr, A., West, C.A. and Rigor, B.M., Neurotoxicity of quinolinic acid and its derivatives in hypoxic rat hippocampal slices, Brain Res., in press. 12 Watkins, J.C. and Olverman, H.J., Agonists and antagonists for excitatory amino acid receptors, Trends Neurosci., 10 (1987) 265-272.