Antagonist Pharmacology of Metabotropic Glutamate Receptors Coupled to Phospholipase D Activation in Adult Rat Hippocampus: Focus on (2R,19S,29R,39S)-2-(29-Carboxy-39-phenylcyclopropyl)glycine Versus 3,5-Dihydroxyphenylglycine

Metabotropic glutamate (mGlu) receptors coupled to phospholipase D (PLD) appear to be distinct from any known mGlu receptor subtype linked to phospholipase C or adenylyl cyclase. The availability of antagonists is necessary for understanding the role of these receptors in the central nervous system, but selective ligands have not yet been identified. In a previous report, we observed that 3,5-dihydroxyphenylglycine (3,5-DHPG) inhibits the PLD response induced by (1S,3R)-1aminocyclopentane-1,3-dicarboxylate in adult rat hippocampal slices. We now show that the antagonist action of 3,5-DHPG (IC50 5 70 mM) was noncompetitive in nature and nonselective, because the drug was also able to reduce PLD activation elicited by 100 mM norepinephrine and 1 mM histamine. In the search for a selective and more potent antagonist, we examined the effects of sixteen stereoisomers of 2-(29-carboxy-39phenylcyclopropyl)glycine (PCCG) on the PLD-specific transphosphatidylation reaction resulting in the formation of [H]phosphatidylethanol. The (2R,19S,29R,39S)-PCCG stereoisomer (PCCG-13) antagonized the formation of [H]phosphatidylethanol induced by 100 mM (1S,3R)-1-aminocyclopentane1,3-dicarboxylate in a dose-dependent manner and with a much lower IC50 value (25 nM) compared with 3,5-DHPG. In addition, increasing concentrations of PCCG-13 were able to shift to the right the agonist dose-response curve but had no effect when tested on other receptors coupled to PLD. The potent, selective, and competitive antagonist PCCG-13 may represent an important tool for elucidating the role of PLDcoupled mGlu receptors in adult hippocampus. Glutamate receptors of the ionotropic (iGlu) and metabotropic (mGlu) types are known to mediate the excitatory and potentially neurotoxic effects of glutamate in the central nervous system. iGlu receptors are ligand-gated ion channels, whereas mGlu receptors are coupled to a variety of effector systems through GTP-binding proteins (Conn et al., 1994). Based on sequence homology, agonist pharmacology, and coupling to second-messenger systems, cloned mGlu receptors have been subdivided into three groups (for a review, see Conn and Pin, 1997). Group I receptors (mGlu 1 and mGlu 5, and their splice variants) are coupled to activation of PLC in a number of heterologous expression systems, whereas group II (mGlu 2 and mGlu 3) and group III (mGlu 4, mGlu 6, mGlu 7, and mGlu 8) receptors are both negatively coupled to adenylyl cyclase. It has been demonstrated that the nonselective mGlu receptor agonist (1S,3R)-1-aminocyclopentane-1,3-dicarboxylate [(1S,3R)-ACPD] is also able to stimulate PLD activity in neonate and adult hippocampal slices (Boss and Conn, 1992; Holler et al., 1993). Recent studies suggest that the glutamatergic activation of phospholipase D (PLD) in immature tissue is indirectly promoted by group I mGlu receptors via protein kinase C (PKC) activation (Klein et al., 1997, 1998), whereas in adult hippocampus, the pharmacology of PLD-coupled mGlu receptors does not appear to correspond to the profile of known subtypes coupled to This work was supported by the University of Florence and the European Community (Biomed 2 Project BMH4-CT96-0228 and Biotechnology Project BIO4-CT96-0049). ABBREVIATIONS: (1S,3R)-ACPD, (1S,3R)-1-aminocyclopentane-1,3-dicarboxylate; 3,5-DHPG, 3,5-dihydroxyphenylglycine; iGlu, ionotropic glutamate; IP, inositol phosphate; L-CCG-I, L(2S,19S,29S)-(carboxycyclopropyl)glycine; (1)-MCPG, (1)-a-methyl-4-carboxyphenylglycine; mGlu, metabotropic glutamate; PCCG, 2-(29-carboxy-39-phenylcyclopropyl)glycine; PEt, phosphatidylethanol; PKC, protein kinase C; PLC, phospholipase C; PLD, phospholipase D. 0026-895X/99/040699-09$3.00/0 Copyright © The American Society for Pharmacology and Experimental Therapeutics All rights of reproduction in any form reserved. MOLECULAR PHARMACOLOGY, 55:699–707 (1999). 699 at A PE T Jornals on A uust 7, 2017 m oharm .aspeurnals.org D ow nladed from PLC or adenylyl cyclase (Boss et al., 1994; Pellegrini-Giampietro et al., 1996a). PLD is the key enzyme in a signal transduction pathway that hydrolyzes phosphatidylcholine and leads to the formation of the second messengers phosphatidic acid and diacylglycerol (for reviews, see Klein et al., 1995; Morris et al., 1996; Exton, 1997). A number of neurotransmitter receptors, including muscarinic, a1-adrenergic, histamine H1, and mGlu receptors, are known to be coupled to both PLC and PLD: their agonists may therefore induce the formation of diacylglycerol by directly stimulating either phosphoinositide or phosphatidylcholine hydrolysis. Because phorbol esters activate PLD, it has also been suggested that PLD activation may be secondary to phosphoinositide hydrolysis via diacylglycerol formation and PKC activation. Compared with PLCproduced diacylglycerol, the formation of diacylglycerol through direct stimulation of the PLD pathway is expected to be 1) slower in time course, 2) more abundant because the substrate phosphatidylcholine is more abundant than phosphatidylinositol in membranes, and 3) long lasting because PKC activated by diacylglycerol desensitizes mGlu receptors coupled to PLC (Catania et al., 1991) but further stimulates PLD. The mutual interaction between PKC and PLD has been described as a “positive feedback loop” by Löffelholz (1989), and it has been proposed as an important mechanism for the generation of increasing amounts of second messengers and prolonged activation of PKC in response to receptor stimulation (Nishizuka, 1998). To understand the role of mGlu receptors linked to PLD activation in the central nervous system, the development of selective antagonists is crucial. In a recent report (PellegriniGiampietro et al., 1996a), we showed that (1)-MCPG (a competitive antagonist of group I and II mGlu receptors) displays mixed agonist/antagonist effects on PLD-coupled mGlu receptors in adult rat hippocampal slices, whereas 3,5-dihydroxyphenylglycine (3,5-DHPG), which is known to be a selective agonist of group I mGlu receptors coupled to PLC activation, was able to antagonize the stimulatory effects of (1S,3R)-ACPD when tested on PLD activity. We now examine in further detail the pharmacological profile of 3,5-DHPG on PLD-coupled mGlu receptors. In addition, we focused our attention on (carboxycyclopropyl)glycines, a valuable source of potent and selective ligands for numerous members of the glutamate receptor family, including mGlu receptors. Considering that the introduction of a hydrophobic moiety, such as a phenyl ring, in position 39 of 2-(29-carboxycyclopropyl)glycine could be useful for mapping the presence of unexplored areas in the recognition site of members of the glutamate receptor family, a complete stereolibrary of 16 2-(29-carboxy-39-phenylcyclopropyl)glycines (PCCGs) was recently synthesized and tested for activity on known mGlu receptors (Pellicciari et al., 1996). In the present study, we investigated whether any of the PCCG isomers lacking activity on known mGlu receptors could selectively antagonize the activation of PLD induced by mGlu receptor agonists. We report on the identification of the (2R,19S,29R,39S)-PCCG stereoisomer (PCCG-13) as a selective, potent and competitive antagonist of PLD-coupled mGlu receptors in the adult rat hippocampus. Experimental Procedures Materials. PCCG-13 was synthesized as described previously (Pellicciari et al., 1996). In a previous preliminary report (PellegriniGiampietro et al., 1996b), the active compound was erroneously indicated as PCCG-16, which differs from PCCG-13 in the stereochemistry at carbon 2. (1S,3R)-ACPD, 3,5-DHPG, and quisqualate were purchased from Tocris Cookson (Bristol, UK). Norepinephrine, histamine, and A23187 were from Sigma Chimica (Milan, Italy). The phosphatidylethanol (PEt) standard was from Avanti Polar Lipids (Pelham, AL). [1,2,3-H]Glycerol (30–60 Ci/mmol) and myo-[2H(N)]inositol (10–25 Ci/mmol) were purchased from Du Pont/NEN (Milan, Italy). Dowex AG-1-X 8 anion exchange resin (100–200 mesh) was from Sigma Chimica, and precoated silica gel 60A (LK6D) plates were from Whatman. Tissue Preparation. Adult (180–200 g) Wistar rats (Nossan strain; Milan) were used. After decapitation, brains were rapidly removed, and the hippocampi dissected and placed into chilled Krebs-bicarbonate buffer (122 mM NaCl, 3.1 mM KCl, 1.2 mM MgSO4, 0.4 mM KH2PO4, 25 mM NaHCO3, 1.3 mM CaCl2, and 10 mM glucose) gassed with 95% O2/5% CO2. Hippocampal slices (350 mm thick) were prepared using a McIlwain tissue chopper and then placed in gassed Krebs-bicarbonate solution for 1 h at 37°C before use. For some experiments, thoracic aortas were removed after decapitation, and aorta rings were prepared as described by Jones et al. (1993). Determination of Agonist-Induced PLD Activity. PLD activity was determined as described previously (Pellegrini-Giampietro et al., 1996a) by making use of the transphosphatidylation reaction between phosphatidylcholine and primary alcohols specifically catalyzed by PLD. Thus, in the presence of exogenously added ethanol, PLD preferentially transfers the alcohol rather than water to the phosphatidyl moiety of phosphatidylcholine, producing PEt in place of phosphatidic acid. Briefly, membrane phospholipids were labeled by incubating hippocampal slices or aorta rings with [H]glycerol (final concentration, 60 mCi/ml) for 2 h at 37°C. Slices or rings were then rinsed and transferred to test tubes (two slices or four rings per test tube) containing 500 ml of drug-containing buffer gently stirred at 37°C by bubbling 95% O2/5% CO2. Antagonists were applied for 10 min before adding the agonists together with 170 mM ethanol, and the reaction was then carried out for 1 additional hour. Previous experiments in hippocampal slices had shown that a steady-state level of PEt formation was reached within 30 min and was stable for at least 2 h. All experiments were run in triplicate; two control sets of triplicate samples were always included in which 1) only Krebsbicarbonate buffer (background) and 2) only buffer plus 170 mM ethanol (basal PLD activity) were present. The reaction was stopped by adding 2 ml of ice-cold chloroform/ methanol/HCl (100:200:2). The phases were then separated by adding 0.65 ml of chloroform and 0.65 ml of water and, after sonication (30 min), by low-speed centrifugation. Aliquots (1 ml) of the lipid phase were dried under a stream of N2, resuspended in 70 ml of chloroform, and spotted onto precoated silica gel 60A plates together with aliquots of a PEt standard solution. [H]PEt was separated from major phospholipids by thin-layer chromatography using the upper phase of the solvent system ethyl acetate/2,2,4-trimethyl pentane/acetic acid/water (12:5:1:10). Spots were visualized with iodine vapor, and [H]PEt was identified by comparison with the PEt standard. The region corresponding to [H]PEt was scraped off, and radioactivity was counted by liquid scintillation spectrometry. The formation of [H]PEt for each individual sample was expressed as the percentage of radioactivity incorporated into the total lipids present in the organic phase. Because PEt was formed only in the presence of ethanol, the amount of label comigrating with [H]PEt in ethanol-free controls was considered as background and subtracted from the mean of each ethanol-containing triplet. Radioactivity present in ethanol-free (background) samples never exceeded 10% of the radioactivity present in ethanol-containing samples. After sub700 Albani-Torregrossa et al. at A PE T Jornals on A uust 7, 2017 m oharm .aspeurnals.org D ow nladed from tracting the background, basal [H]PEt formation, expressed as [H]PEt/[H]total lipids 3 10, was consistently about 12.0 6 0.8. Data are expressed as percentages of incorporation of label into [H]PEt occurring under agonist-free (basal) conditions. Determination of Agonist-Induced PLC Activity. Agonistinduced phosphoinositide hydrolysis was assayed essentially as described previously (Pellegrini-Giampietro et al., 1996a). Briefly, hippocampal slices incubated for 2 h at 37°C with [H]inositol (final concentration, 20 mCi/ml) were rinsed and then transferred to test tubes (two slices each) with 500 ml of drug-containing buffer gently stirred at 37°C in the presence of 10 mM LiCl by bubbling 95% O2/5% CO2. Antagonists were applied for 10 min before adding the agonists, which were then allowed to react for an additional 15 min. All experiments were run in triplicate. The reaction was stopped by adding 1.88 ml of ice-cold chloroform/methanol (1:2). The phases were separated by adding 0.65 ml of chloroform and 0.65 ml of water and, after brief sonication, by low-speed centrifugation. The upper phase, containing the water-soluble [H]inositol phosphates (IPs) (inositol monophosphate, inositol-1,4-bisphosphate, and inositol1,4,5-trisphosphate), was transferred to Dowex AG 1-X 8 (formate form, 100–200 mesh) anion exchange resin columns. After washing with water and eluting [H]glycerophosphorylinositol with 16 ml of 5 mM sodium tetraborate, 60 mM ammonium formate, the three [H]IPs were eluted together with 16 ml of 1 M ammonium formate, 0.1 M formic acid. The dpm of the corresponding fractions was determined by liquid scintillation spectrometry, and the radioactivity present in [H]IPs was normalized to radioactivity incorporated into [H]glycerophosphorylinositol. Data are expressed as percentages of incorporation of label into [H]IPs occurring under agonistfree (basal) conditions. Data and Statistical Analysis. Agonist dose-response and antagonist inhibition curves were analyzed by nonlinear regression, and EC50 and IC50 values were calculated with the Prism software package (GraphPAD Software, San Diego, CA). The PCCG-13 dissociation constant (KB) was estimated from the results of the (1S,3R)ACPD dose-response curve and the PCCG-13 inhibition curve in the presence of a fixed concentration (Af) of (1S,3R)-ACPD, applying the “null” method and the equation KB 5 IC509/([Af]/EC509 2 1), where EC509 and IC509 are equiactive concentrations derived from agonist and antagonist dose-response curves, respectively, constrained to have the same maximum (Lazareno and Birdsall, 1993). The slope and the pA2 value of the Schild plot of PCCG-13 were calculated by linear regression analysis with the Prism software package. Statistical significance of differences between results was evaluated by performing ANOVA followed by Tukey’s w test for multiple compar-