[1] 20w Homocysteic Acid as a Putative Excitatory Amino Acid Neurotransmitter: I. Postsynaptic Characteristics at N- Methyl-D-Aspartate-Type Receptors on Striatal Cholinergic Interneurons

[1] 235w The actions of the stereoisomers of homocysteic acid (HCA) were characterized at N-methyl-paspartate (NMDA)-type receptors which mediate excitatory amino acid-evoked ['Hlacetylcholine ([ 3H]ACh) release from striatal cholinergic interneurons. Like NMDA, L-HCA and D-HCA evoked the release of ['HIACh formed from ['Hlcholine in striatal slices. The concentration-response curve for L-HCA was virtually superimposable on that for NMDA, yielding an equal ECso value (56.1 p M ) and maximal response. However, D-HCA was weaker, with an ECS0 value of 8 l. l pM, and an apparently smaller maximal response. L-HCA-evoked ['HIACh release was inhibited by the same categories of compounds which inhibit NMDA-evoked ['HI ACh release: the divalent ion Mgz+ (ICsO = 25.8 p M ) ; competitive NMDA antagonists 2-amino-7-phosphonoheptanoate (IC50 = 5 1.2 p M ) and 3-(2-carboxypiperazin-4-yl)propyl-1 -phosphonic acid (ICsO = 20.1 pM); and the dissociative anesthetics tiletamine (IC50 = 0.59 p M ) and MK-801 (ICsO = 0.087 pM).Like NMDA, L-HCA produced a tachyphylaxis in this system. Tachyphylaxis to NMDA resulted in a decreased response to L-HCA, and conversely, tachyphylaxis to L-HCA resulted in a decreased response to NMDA. The results suggest that L-HCA is an agonist at the NMDA-type receptor and may represent an endogenous ligand for this excitatory amino acid receptor. Key Words: Homocysteate-Acetylcholine release-N-Methyl-D-aspartate-type receptors-Striatum-Excitatory amino acid. Lehmann J. et al. Homocysteic acid as a putative excitatory amino acid neurotransmitter: I. Postsynaptic characteristics at N-methyl-paspartate-type receptors on striatal cholinergic interneurons.

[1] 121w L-HCA induced a concentration-dependent increase in the release of [3H]ACh when added to the medium in a 2-min pulse, as indicated by the ratio of evoked release (SI) over the baseline of spontaneous release (SpI; Fig. 1). The EC5o of L-HCA (56.1 p M ) was not significantly different from that of NMDA (42.9 p M ) in these experiments (Fig. 1, Table 1). D-HCA produced much less [3H]ACh release than equivalent concentrations of L-HCA or NMDA (Fig. 1). This difference was due primarily to a decreased efficacy of D-HCA compared to L-HCA or NMDA (Table 1). All subsequent release experiments were performed with the L-enantiomer, which is the naturally occurring enantiomer, as well as that which caused a greater effect.

[2] 94w When two 2-min pulses of L-HCA (50 pM) were applied, two peaks of [3H]ACh release were observed (Fig. 2). The evoked release above the baseline of spontaneous outflow for the second stimulation (SII) divided by that for the first stimulation (Sl) yielded a ratio (SII/ SI) that was precise and reproducible. When a competitive antagonist of NMDA-type receptors, such as AP7, was added to the medium 18 min before SI1, the spontaneous release was unaffected, but the L-HCAevoked [ 3H]ACh release was inhibited selectively (Fig. 2). This inhibition was quantifiable by the ratio SII/SI.

[3] 67w The inhibition of L-HCA-evoked [3H]ACh release was concentration-dependent for the competitive NMDA-type receptor antagonists AP7 and CPP, as well as for the noncompetitive inhibitors of NMDA-type receptor function, tiletamine and MK-80 1 (Fig. 3). The ICso values are shown in Table 2. Magnesium, which acts at the cation channel linked to the NMDA-type Table 2. Each value is the mean f SEM of at least four determinations.

[4] 40w TABLE 2. Inhibition of L-HCA-evoked ['H]ACh release by competitive antagonists and noncompetitive inhibitors of the NMDA-type receptor IC50, N Compound (95% confidence limits) MK-80 1 0.087 (0.063-0.120) Tiletamine 0.59 (0.40-0.86) CPP 20.1 (14.1-28.6) AP7 5 1.2 (38.3-68.4) MgC1z 25.8 (18.0-36.9)

[5] 29w receptor to impede its function, also produced a concentration-dependent inhibition of L-HCA-evoked [3H]ACh release (Fig. 4). In all cases, these compounds failed to affect the spontaneous outflow of [3H]ACh.

[6] 113w The NMDA-type receptor apparently desensitizes upon prolonged exposure to agonists. This characteristic may be used to determine if agonists are acting through the same set of receptors. When NMDA (50 p M ) was added to the medium during the entire second half of a superfusion experiment, there was an immediate increase in [3H]ACh release (Fig. 5). This release declined gradually with time, illustrating tachyphylaxis. When 18 min later the second 2-min pulse of L-HCA (50 p M ) was added, the response to this amino acid was greatly reduced, illustrating crosstachyphylaxis (Fig. 5). The same tachyphylaxis was produced by L-HCA itself (Fig. 6), which likewise produced a cross-tachyphylaxis to NMDA (Fig. 6).

[1] 218w HCA has been proposed as a putative neurotransmitter which acts at excitatory amino acid receptors in the brain, with the major evidence as follows: (1) 2 V 5.0 1 0-0 Control W-m Addition of NMDA at t = 60 rnin A-A Addition of NMDA, no 2nd L-HCA stirnulotion Addition of NMDA 50 pM V T . . . ~ 1 . 0 ; I . I . 1 c * . -I 30 60 90 Time of superfusion, minutes FIG. 5. NMDA-induced cross-tachyphylaxis to L-HCA. Two pulses (represented by 0) of L-HCA (50 M), each of 2 min duration, resulted in two distinct peaks of evoked [3H]ACh release under control conditions (0). Addition of NMDA (50 pM; D), starting at 60 rnin and kept in the medium until the end of the experiment, resulted first in an elevation of fractional release, and subsequently in a greatly reduced response to the second application of L-HCA (50 f l ) , showing that NMDA induced a cross-tachyphylaxis to L-HCA. In a third group (A), addition of NMDA (50 M ) starting at 60 min resulted first in an elevation of fractional release, which then decreased with time even though NMDA (50 pM) was always present, showing that NMDA induced tachyphylaxis. Each point represents the mean f SEM of four observations.

[2] 13w HCA is released from brain slices by depolarizing stimuli (Do et al., 1986b).

[3] 75w (2) Electrophysiological experiments show that HCA produces bistable depolarizing shifts similar to those produced by NMDA, suggesting an interaction with the NMDA-type receptor (Herrling et al., 1983; Do et al., 1986~). (3) An uptake system for [35S]HCA in brain has been demonstrated (Cox et al., 1977). (4) L-HCA potently induces a pattern of cytopathology in the ex vivo chick retina which mimics the pattern of NMDA-induced, but not kainate-or quisqualate-induced, neurotoxicity (Olney et al., 1987).

[4] 107w The results obtained in this study support the conclusions obtained by electrophysiological techniques and in studies of neurotoxicity, i.e., that HCA acts as an agonist at NMDA-type receptors. In the present study, HCA was found to have a small degree of stereospecificity, with the D-enantiomer being approximately 50% less potent than the L-enantiomer and showing a lower maximal response as well. L-HCA has been reported previously to be 3.85-fold more potent than D-HCA in [3H]CPP-binding studies (Murphy et al., 1987). Stereospecificity of agonists regarding both potency and intrinsic activity has been observed previously for other agonists, such as the dopamine receptor agonist apomorphine (Lehmann et al., 1983).

[5] 78w Like NMDA-evoked [3H]ACh release from the striatum, L-HCA-evoked [3H]ACh release was antagonized by competitive NMDA-type receptor antagonists and dissociative anesthetics (Fig. 3) and magnesium (Fig. 4). The potencies of these compounds against L-HCA were not substantially different from their potencies against NMDA (Lehmann and Scatton, 1982;Lehmann et al., 1987;Lehmann and Wood, 1988). Because the pharmacological profile of antagonism of L-HCA matches that of antagonism of NMDA, it is fair to conclude that L-HCA interacts with the NMDA-type receptor.

[6] 84w Subsequently, we examined the question of whether L-HCA may activate a subpopulation of NMDA-type receptors. This was performed by experiments taking advantage of the phenomenon of tachyphylaxis, or reduced response of the NMDA-type receptor to continuously applied agonist. This method has been used previously to show that quinolinate acts at the same receptor population as NMDA (Lehmann et al., 1985), and that glycine interacts with a group of receptors different and distinct from those activated by NMDA in the striatum (Taylor et al., 1988).

[7] 174w When NMDA was applied continuously, the response declined progressively with time (Fig. 5). When L-HCA was added subsequently, no response was elio ited (Fig. 5). Thus, the tachyphylaxis to NMDA seems to affect also the response to L-HCA. This result suggests that all receptors activated by L-HCA are also activated by NMDA. Conversely, when tachyphylaxis was produced by prolonged exposure to L-HCA, the response to NMDA was also attenuated (Fig. 6). Thus, all receptors activated by NMDA are also activated by L-HCA. Taken together, these results suggest that NMDA and L-HCA act at identical populations of NMDA receptors, and that both produce a tachyphylaxis. The tachyphylaxis which is produced upon continuous exposure to NMDA agonists would seem not to reflect a general unresponsiveness of the striatal cholinergic interneuron. Tachyphylaxis produced by NMDA does not prevent the neuron from responding to glycine, and conversely tachyphylaxis to glycine does not prevent the neuron from responding to NMDA (Taylor et al., 1988). Thus, it is more likely that the tachyphylaxis produced by NMDA is a receptor desensitization.

[8] 81w L-HCA selectively activates NMDA-type receptors assessed electrophysiologically in the striatum and in the cortex, where quisqualate-type receptor activation by other compounds produces a different response (Do et al., 1986~;Zeise et al., 1988). L-HCA has been shown to be an excitotoxin acting virtually exclusively through the NMDA-type receptor in chick retina, where quisqualate-type receptors and kainate-type receptors mediating excitotoxicity are also found (Olney et al., 1987). Both of these pieces of information suggest that L-HCA selectively activates NMDA-type rather than quisqualate-type receptors.

[9] 121w This view of selective interaction of HCA with NMDA-type receptors is challenged, however, by a number of findings. HCA interacts nonstereospecifically with quisqualate-type (but not kainate-type) receptors in binding studies, with the same potency as its effects at the NMDA-type receptor (Murphy and Williams, 1987;Murphy et al., 1987). In medium which contains 1 mM magnesium (and hence suppresses NMDA-type receptor-mediated responses), DL- HCA and quisqualate, but not NMDA or kainate, apparently inhibit the formation of the intracellular sulfate donor [35S]3r-phosphoadenosine-5r-phosphosulfate (Gulat-Marnay et al., 1987). Finally, excitatory responses to L-HCA and quisqualate, but not NMDA, kainate, L-aspartate, or L-glutamate, are enhanced by 8-pchlorophenylglutamate (Davies et al., 1985). Thus, in a number of systems, HCA produces actions like quisqualate, but different from NMDA.

[10] 91w The convulsant action of L-HCA is selectively inhibited by the NMDA-type receptor antagonist AP7, whereas D-HCA-induced convulsions are insensitive to AP7 (Turski et al., 1988). It is worthwhile recalling that, although HCA stereoselectively interacts with the NMDA-type receptor, there is no such stereospecificity at the quisqualate-type receptor (Murphy and Williams, 1987). Therefore, the predominant response to L-HCA in most physiological systems may be due to activation of NMDA-type receptors, but other excitatory amino acid receptors may also be activated at higher concentrations of L-HCA or at equivalent concentrations of the racemate.

[11] 114w The key comparison for the postsynaptic actions of HCA is not with quisqualate or NMDA, but rather with L-glutamate and L-aspartate. For a number of reasons, a substance other than L-glutamate and L-aspartate is suspected to be the endogenous neurotransmitter acting at synapses mediated by excitatory amino acid receptors (Do et al., 1987). In particular, L-HCA activates NMDA-type receptors selectively when applied iontophoretically in the neocortex, whereas L- glutamate shows no such selectivity (Knopfel et al., 1987). Thus, based on postsynaptic actions, L-HCA may be a neurotransmitter which activates NMDAtype receptors. In a related article (Tsai et al., submitted), data addressing the presynaptic evidence for the hypothesis that L-HCA is a neurotransmitter are presented.

[1] 34w [methyl-'HICholine chloride (80 Ci/mmol) was obtained from New England Nuclear (Boston, MA, U.S.A.). D-HCA The present address of Dr. P. L. Wood is CNS Diseases, Monsanto AA-A4, 700 Chesterfield Pkwy., Chesterfield, MO 63198, U.S.A.

[2] 68w Abbreviations used: ACh, acetylcholine; AP7, 2-amino-7-phosphonoheptanoate; CPP, 3-(2-carboxypiperazin-4-yl)propyl-l-phosphonic acid; HCA, homocysteic acid; NMDA, N-methyl-Gaspartate. and L-HCA were obtained from Sigma (St. Louis, MO, U.S.A.). MK-801 was obtained from Merck, Sharpe and Dohme (Essex, U.K.). Tiletamine was from Warner-Lambert (Ann Arbor, MI, U.S.A.). 2-Amino-7-phosphonoheptanoate (AP7) was synthesized by Ciba-Geigy Corp. in Basel, Switzerland, and 3-(2-carboxypiperazin-4-yl)propyl-I-phosphonic acid (CPP) was synthesized by Ciba-Geigy Corp. in Summit, NJ, U S A .

[3] 245w HCA-and NMDA-evoked release of ['HIACh from striatal slices was measured as previously described (Lehmann and Scatton, 1982). Following decapitation of male Sprague-Dawley rats (Mbf:(SD) 200-300 g), the striatum was excised and chopped coronally (0.3 mm). [N-methyl-'H]Choline (80 Ci/mmol, 50 nM; New England Nuclear) was added together with slices to 20 ml of Krebs medium [composition (mM): NaCl, 1 18; KCI, 4.2; NaHCO3, 25; NaH2P04, 1 .O; CaC12, 1.3; glucose, 111. After a 30-min incubation at 37°C in this medium, two slices were placed in each of 24 disposable, conical, superfusion chambers and superfused with Krebs medium containing 10 pM hemicholinium-3 at a speed of 0.2 ml/min. In order to generate concentration-response curves for D-HCA, L-HCA, and NMDA, only a single 2-min pulse of agonist (10 pM to 1 mM) was applied at 42 min after the beginning of superfusion. To examine antagonism and cross-tachyphylaxis, a two-pulse paradigm allowing normalization of data for each slice was used, as follows: NMDA (50 p M ) was added beginning at 42 rnin (S,) and 76 rnin (SII) after the start of superfusion and was kept in the medium for a duration of 2 min. Antagonists were added 18 rnin preceding SII and kept in the medium until the end of the experiment. The superfusate was collected in 6-min fractions, to which was added 0.05 ml of 30% ascorbic acid to scavenge oxygen and reduce chemiluminescence. The striatal slices were recovered at the end of the experiment and solubilized

[4] 149w T N = n L-HCA 0 -1 10 100 1000 0.75 1; , , , , , , , , ( , , , , , , , ' I , , , , , , , ~ [Agonist], ,uM FIG. 1. Stereospecific release of [3H]ACh by HCA. The potency and efficacy of L-HCA were comparable to the potency and efficacy of NMDA, whereas D-HCA showed primarily a decreased apparent maximal effect. ECS0 values and maximal effects of these agonists are shown in Table 1. Abscissa: The micromolar concentration of agonist is plotted logarithmically. Ordinate: The elevation of [3H]ACh release is expressed as fractional release in the 641-1 fraction immediately following application of agonist (SI) divided by the baseline of spontaneous release in the 6-min fraction immediately preceding addition of agonist (Spl), to give the ratio SI/Spl. Each point represents the mean i SEM of eight to 24 values.

[1] 48w HCA (Olney et al., 1987) suggest, moreover, that L-HCA activates NMDA-type receptors selectively. We report here the pharmacological characteristics of HCA acting on NMDA-type receptors which mediate ['Hlacetylcholine ( [3H]ACh) release in the rat striaturn, a system permitting quantification of potency, intrinsic activity, and desensitization properties of agonists.

[2] 81w To determine potencies of agonists, the elevation of dpm in the superfusion medium in the fraction immediately following addition of the agonist (SI) divided by the spontaneous release (immediately preceding addition of the agonist; Sp,), i.e., SI/Spl, was calculated. However, in most experiments, fractional release was calculated as previously described (Lehmann et al., 1985), to assess the full spectrum of possible effects of compounds on NMDA-type agonist-evoked ['HIACh release. In order to determine ECS0 and maximal response values, the equation f(