Characterization of Novel Aryl-Ether, Biaryl, and Fluorene Aspartic Acid and Diaminopropionic Acid Analogs as Potent Inhibitors of the High-Affinity Glutamate

    Discovery Neuroscience (J.D., H.B.M., T.A.C., B.J., Q.L., D.K., S.L.) and Chemical & Screening Sciences (A.G., C.G., K.F., J.B.), Wyeth Research, Princeton, New Jersey
    Neurocrine Biosciences, San Diego, California (R.P., J.W., A.F.)

    Abstract

    In this study, we describe the pharmacological characterization of novel aryl-ether, biaryl, and fluorene aspartic acid and diaminopropionic acid analogs as potent inhibitors of EAAT2, the predominant glutamate transporter in forebrain regions. The rank order of potency determined for the inhibition of human EAAT2 was N4-[4-(2-bromo-4,5-difluorophenoxy)phenyl]-L-asparagine (WAY-213613) (IC50 = 85 ± 5 nM) > N4-(2'-methyl-1,1'-biphenyl-4-yl)-L-asparagine (WAY-213394) (IC50 = 145 ± 22 nM) = N4-[7-(trifluoromethyl)-9H-fluoren-2-yl]-L-asparagine (WAY-212922) (IC50 = 157 ± 11 nM) = 3-{[(4'-chloro-2-methyl-1,1'-biphenyl-4-yl)carbonyl]amino}-L-alanine (WAY-211686) (IC50 = 190 ± 10 nM). WAY-213613 was the most selective of the compounds examined, with IC50 values for inhibition of EAAT1 and EAAT3 of 5 and 3.8 e, respectively, corresponding to a 59- and 45-fold selectivity toward EAAT2. An identical rank order of potency [WAY-213613 (35 ± 7 nM) > WAY-213394 (92 ± 13 nM) = WAY-212922 (95 ± 8 nM) = WAY-211686 (101 ± 20 nM)] was observed for the inhibition of glutamate uptake in rat cortical synaptosomes, consistent with the predominant contribution of EAAT2 to this activity. Kinetic studies with each of the compounds in synaptosomes revealed a competitive mechanism of inhibition. All compounds were determined to be nonsubstrates by evaluating both the stimulation of currents in EAAT2-injected oocytes and the heteroexchange of D-[3H]aspartate from cortical synaptosomes. WAY-213613 represents the most potent and selective inhibitor of EAAT2 identified to date. Taken in combination with its selectivity over ionotropic and metabotropic glutamate receptors, this compound represents a potential tool for the further elucidation of EAAT2 function.

    Glutamate is the predominant excitatory neurotransmitter in the mammalian central nervous system. Glutamate transmission is mediated via interaction with the ligand-gated ion channel receptors, termed the ionotropic receptors, and the seven-transmembrane domain G-proteineCcoupled receptors, termed metabotropic glutamate receptors (Barnard, 1997; Schoepp et al., 1999). Activation of these receptors is responsible for the physiological actions of glutamate, whereas paradoxically, overstimulation of the ionotropic receptors contributes to the excitotoxic actions attributed to glutamate. Therefore, synaptic glutamate levels must be tightly regulated to maintain the integrity of synaptic transmission and to limit or prevent the pathophysiological activity of this excitatory neurotransmitter.

    A family of high-affinity Na+-dependent glutamate transporters expressed in the plasma membranes of both neurons and astroglia is responsible for the clearance of extracellular glutamate by mediating the cellular uptake of glutamate in a process driven largely by the energy of the transmembrane Na+ gradient (Danbolt, 2001). Five members of this transporter family have been identified by molecular cloning and are designated EAAT1/GLAST, EAAT2/GLT-1, EAAT3/EAAC1, EAAT4, and EAAT5 (Kanai and Hediger, 1992; Pines et al., 1992; Storck et al., 1992; Arriza et al., 1994, 1997; Fairman et al., 1995). Among the transporter subtypes, EAAT2 accounts for the bulk of transport activity measured in forebrain preparations. The pharmacological profile of glutamate transport measured in rat forebrain preparations is well correlated with the pharmacology of the cloned rat GLT-1 and human EAAT2 subtypes expressed in mammalian cell lines (Dunlop et al., 1999; Tan et al., 1999). Moreover, transgenic mice with a targeted deletion of the EAAT2 transporter lose more than 90% of the forebrain glutamate transport capacity and exhibit increased vulnerability to an experimental head trauma insult (Tanaka et al., 1997). Experiments using antisense oligonucleotides to knock down levels of EAAT2 resulted in reduced glutamate transport capacity, elevated extracellular glutamate, and increased neurodegeneration (Rothstein et al., 1996), providing another line of evidence in support of a major role of EAAT2 in glutamate clearance.

    A pharmacological elucidation of the functional role of EAAT2 has been hampered by the lack of potent and selective transport inhibitors. Widely used pharmacological tools such as the pyrrolidine dicarboxylates (Bridges et al., 1991), aminocyclobutane dicarboxylates (Fletcher et al., 1991), and carboxycyclopropyl glycines (Nakamura et al., 1993) suffer from combinations of poor selectivity across EAAT subtypes, rather modest potencies in the micromolar range, and cross-reactivity with glutamate receptor targets. More recently, the threo--hydroxyaspartate derivative threo--benzyloxyaspartate (TBOA) and the novel heptane dicarboxylate WAY-855 have been reported as novel EAAT inhibitors (Lebrun et al., 1997; Shimamoto et al., 1998; Dunlop et al., 2003), and these agents provide some improvement over the earlier compounds. Subsequent structure activity relationship studies with TBOA have led to the identification of nanomolar potent EAAT2 inhibitors, with the most selective agent, (2S,3S)-3-[3-(4-methoxybenzoylamino)benzyloxy]aspartate), exhibiting 39-fold selectivity over EAAT3 (Shimamoto et al., 1998). Such agents are likely to be better tools for the characterization of transporter function.

    In this study, we present the in vitro pharmacological characterization of novel chemotypes possessing potent and EAAT2-preferring inhibitory activity. Compounds were designed by appending a lipophilic load on a relatively polar aspartic acid or aspartic acid-like (diaminopropionic acid) headpiece. On the basis of our hypothetical binding site interaction model, generated from a set of literature and inhouse EAAT inhibitors, additional aromatic centers observed in these novel compounds are implicated in the enhancement of EAAT2 potency and concomitant improvement in selectivity over EAAT1 and EAAT3. The synthesis and structure-activity relationship of these series of compounds will be presented elsewhere. These novel compounds offer advantages over existing EAAT2 inhibitors in terms of potency, selectivity, and ease of synthesis. WAY-213613 was found to be the most potent and selective of the compounds described herein, inhibiting EAAT2 with an IC50 of 85 nM and exhibiting 59- and 45-fold selectivity over EAAT1 and EAAT3, respectively. Moreover, WAY-213613 was devoid of ionotropic and metabotropic glutamate receptor activity, indicating it to be a highly EAAT2-preferring transporter ligand.

    Materials and Methods

    Uptake in Stable Cell Lines. Stable HEK cell lines expressing each of the human glutamate transporter subtypes EAAT1eC3 were maintained in T175 flasks at 37°C in a humidified atmosphere with 5% CO2. Cells were detached and replated at 50,000 cells/well in 96-well culture plates the day before measurements of glutamate uptake. Uptake assays were performed in Dulbecco's phosphate-buffered saline (D-PBS) in the presence of 1 e glutamate and 0.2 e藽i/ml L-[3H]glutamate (specific activity, 20eC30 Ci/mmol; 1 mCi/ml) in a final volume of 100 e蘬. Cells were washed with D-PBS before incubation with substrate in the absence or presence of compounds for 20 min at room temperature. Assays were stopped by aspiration followed by two ice-cold D-PBS washes. Cells were solubilized with 0.5 N NaOH before the addition of Microscint 20 for the determination of 3H accumulation in the wells using a TopCount scintillation counter (PerkinElmer Life and Analytical Sciences, Boston, MA). Uptake was linear for incubation times up to 30 min (data not shown); thus, data were analyzed as true rates. Nonspecific uptake was corrected for by performing all experiments in the absence and presence of sodium (replacement of sodium chloride with choline chloride). Sodium-independent uptake accounted for 10% of the total uptake and was subtracted before any further data calculation.

    Uptake in Rat Cortical Synaptosomes. P2 synaptosomal fractions isolated from rat cerebral cortex were prepared by homogenization of tissue in ice-cold isolation medium (310 mM sucrose and 10 mM HEPES, pH 7.4) followed by centrifugation at 1000g for 5 min. The resulting supernatant was collected and centrifuged at 20,000g for 20 min to obtain the crude synaptosomal P2 pellet, which was used for uptake studies at a protein concentration of 1 mg/ml in HEPES-buffered saline (HBS). L-[3H]Glutamate (specific activity, 20eC30 Ci/mmol; 1 mCi/ml) uptake was assayed in a final volume of 300 e蘬 of HBS containing 50 e蘥 of synaptosomal protein, 1 e L-glutamate, and 0.25 e藽i/assay L-[3H]glutamate in the absence and presence of drug for the determination of IC50 values. Kinetic experiments were performed in the presence of 0.25 e藽i/assay L-[3H]glutamate over a range of nonradioactive L-glutamate concentrations in both the absence and presence of compounds. Uptake assays were incubated for 5 min at room temperature and were terminated by filtration on a 96-well Unifilter (Whatman GF/B; Whatman, Clifton, NJ) plate followed by two washes with ice-cold buffer. Microscint 20 was added directly to filter plates for the determination of 3H retained on the filters using a TopCount scintillation counter. Uptake was linear over the 5-min incubation period, and data were corrected to represent sodium-dependent uptake.

    Oocyte Studies. EAAT mRNAs were injected into defolliculated stage IV or V Xenopus laevis oocytes (10 ng/oocyte), and the membrane currents were recorded 3 to 7 days later. Recording solution contained 96 mM NaCl, 2 mM KCl, 1 mM CaCl2, 1 mM MgCl2, and 5 mM HEPES, pH 7.4. Two-electrode voltage-clamp recordings were performed at room temperature with an NPI TURBO TEC-03X amplifier, EPC-9, and PULSE program for analog-to-digital interface and data acquisition. Microelectrodes were filled with 3 M KCl solution and had resistance of approximately 1 Me? Cells were clamped at eC40 mV and depolarized to eC80 mV for 100 ms, at which the current was measured at the end of the step protocol. EAAT current amplitude was determined by measuring the difference of the step currents before and after the substrate (10 e glutamate) applied to the bath. The current was low-passeCfiltered at 200 Hz and digitized at 5 kHz. Compounds were dissolved in dimethyl sulfoxide as stock and diluted to final concentration with less than 1% dimethyl sulfoxide in the bath. Data were averaged from all oocytes tested and are expressed as mean ± S.E.

    Ionotropic Glutamate Receptor Assays. Evaluation of the compounds for effects on ionotropic glutamate receptors was performed by measuring their effects on NMDA and AMPA-induced depolarization of primary cultures of rat hippocampal neurons using a membrane potential-sensitive fluorescent dye [fluorometric imaging plate reader (FLIPR) membrane potential dye; Molecular Devices, Sunnyvale, CA]. Minced hippocampi dissected from embryonic day 18 embryos were incubated at 37°C in 0.01% papain (Worthington Biochemicals, Freehold, NJ), 0.1% dispase (Roche Products, Hertfordshire, UK), and 0.01% DNase (Sigma Chemical, St. Louis, MO) dissolved in Hanks' balanced salt solution. Individual neurons were plated in a 96-well format at a density of 100,000 cells/well. Neurons were cultured at 37°C in 5% CO2 in Neurobasal Media (Invitrogen, Carlsbad, CA) supplemented with 2 mM glutamine and 1x B27 supplement (Invitrogen). Assays were performed 2 weeks after plating. After removing the culture media, 200 e蘬 of FLIPR membrane potential dye was added to each well, and plates were incubated at room temperature for 0.5 to 1 h. Agonist activity was evaluated by direct application of the compounds to the plate, after dye loading, and after monitoring 15-s baseline fluorescence online. For antagonist assays, compounds were included during the dye loading step with subsequent stimulation by the addition of 50 e蘬 of depolarizing solution containing 100 e NMDA or AMPA added to each well during FLEX station runs after a 15-s reading of baseline. As a positive control, 100 e NMDA or AMPA was added to confirm the agonist activity, and 30 e Ro 25-6981 or NBQX was used to block NMDA- and AMPA-mediated responses, respectively.

    Metabotropic Glutamate Receptor Selectivity Studies. Stable Chinese hamster ovary cell lines expressing either the human mGlu receptors 1, 2, or 4 subtypes were used for functional studies using the measurement of agonist-stimulated mobilization of intracellular calcium with the FLIPR. For the Gi-coupled mGlu receptors 2 and 4, functional calcium coupling was facilitated by stable coexpression of a G-protein chimera comprising Gq with the C-terminal five amino acids replaced with those from Gi (Kowal et al., 2003). Cells were maintained and passaged upon reaching approximately 80% confluence. Cells were plated 24 h before the experiment in poly-D-lysineeCcoated 96-well plates at a density of approximately 50,000 cells/well. In preparation of the assay, the confluent monolayer of cells was washed twice with Hanks' balanced salt solution supplemented with 20 mM HEPES and 2.5 mM probenecid (FLIPR buffer), and then the cells were loaded by adding 4 e Fluo-4 AM (Molecular Probes, Eugene, OR) in FLIPR buffer for 1 h at 37°C. After loading, the cells were rinsed twice with FLIPR buffer, and intracellular calcium increases after agonist application were detected by measuring increases in fluorescence with the FLIPR. For evaluation of antagonist activity, LY341495 was included during the dye-loading step and was stimulated subsequently by the addition of an approximate EC80 concentration of the agonists L-quisqualate, LY354740, and L-AP4 for mGlu receptors 1, 2, and 4.

    Synaptosomal Exchange. Rat cortical P2 synaptosome fractions were resuspended in HBS at a protein concentration of 1 mg/ml and incubated in the presence of 1 e藽i/ml D-[3H]aspartate (10eC15 Ci/mmol; 1 mCi/ml) for 1 h at room temperature with constant shaking. Synaptosome suspension (100 e蘬) equilibrated with 3H-label was diluted into 100 e蘬 of HBS containing compound at final concentration of 100 e immediately after the 1-h loading period. Under this dilution paradigm, the equilibrium established between internal and external 3H label is disrupted such that efflux of the label is now favored as the system tends toward re-establishing equilibrium. Therefore, substrate inhibitors, themselves translocated by the transport system, will favor 3H efflux via heteroexchange, whereas nontransportable inhibitors either will have no effect if efflux is limited or will block efflux. Synaptosomes were incubated with drug for 5 min at room temperature, followed by centrifugation to separate pellet and supernatant. An aliquot of supernatant was subsequently removed for the determination of 3H efflux, a direct index of drug-mediated exchange.

    Molecular Modeling. A data set was assembled using literaturereported EAAT inhibitors and the novel compounds disclosed in the present study (Fig. 1). It was assumed that the compounds interact with the transporter(s) in a similar manner, a reasonable assumption given that the compounds have similar mechanism of action as nontransportable, competitive inhibitors of high-affinity glutamate uptake. A "common feature hypothesis" in Catalyst 4.6 (Accelrys, San Diego, CA) was used to develop the pharmacophore model. The Catalyst pharmacophore model contains the location constraints (three-dimensional coordinates and their tolerance) of commonly used chemical features such as hydrogen bond acceptor, hydrogen bond donor, negative ionizable groups such as carboxylic, acid and positive ionizable groups such as amino groups and ring aromatics. Before importing structures into Catalyst, each molecule was fully minimized using MMFF94 force field (Halgren, 1996), and the conformational expansion was performed using the optimal algorithm implemented in Catalyst for each molecule, with a maximum of 250 conformers per molecule and an energy range of 15 kcal/mol.

    Data Analysis. IC50 values for the dose-response curves were generated using a four-parameter logistic curve-fitting model using Origin 7.5 (OriginLab Corp, Northampton, MA) using the following equation: y = [(A1 eC A2)/1 + (x/x0)nH] + A2, where A1 is the maximum, A2 is the minimum, x0 is the IC50 value, and nH is the Hill slope coefficient.

    Results

    Inhibition of Human EAAT Subtypes. Figure 1 illustrates the structures of WAY-213613, WAY-213394, WAY-212922, and WAY-211686. Potent EAAT2 inhibitory activity was observed for all four compounds as determined by the inhibition of L-[3H]glutamate uptake in a stable HEK cell line expressing the human EAAT2 subtype (Fig. 2 and Table 1). WAY-213613 was the most potent of the compounds, with an IC50 value of 85 ± 5 nM. Primary selectivity determinations for the compounds using HEK cells stably expressing either the human EAAT1 or EAAT3 subtype are also illustrated in Fig. 2 with estimated IC50 values presented in Table 1. In addition to being the most potent compound, WAY-213613 was the most selective of the four, exhibiting 59- and 45-fold selectivity over EAAT1 and EAAT3, respectively. The uptake of L-glutamate measured in the EAAT-expressing HEK cell lines in the absence of drug was 10.7, 158.6, and 8.0 pmol/min/mg for EAAT1, -2, and -3, respectively.

    Data are presented as mean ± S.E.M. from three independent experiments.

    Synaptosomal Uptake Studies. Additional pharmacological characterization in the form of inhibition of L-[3H]glutamate uptake into rat brain cortical synaptosome fractions, a preparation in which EAAT2 contributes the bulk of the measured substrate accumulation, was performed for the compounds in addition to the nonselective and widely used pharmacological agent L-trans-pyrollidine-2,4-dicarboxylate (trans-2,4-PDC). Concentration-response curves are illustrated in Fig. 3. Glutamate uptake in the synaptosomal preparation in the absence of added compound (control) was 2 to 2.5 nmol/min/mg. Estimated IC50 values for the inhibition of synaptosomal L-[3H]glutamate uptake (Table 1) and rank order of potency were essentially identical with those determined for inhibition of EAAT2, with WAY-213613 being the most potent (IC50 = 35 ± 7 nM) of the compounds evaluated, whereas the potency of trans-PDC was identical with that reported by others (IC50 = 1.6 ± 0.2 e). Kinetic studies were also performed with the synaptosomal preparation to establish mode of inhibition. A double-reciprocal plot (Line-weaver-Burk transformation) of the effect of multiple concentrations of WAY-213613 on the saturable uptake of L-[3H]glutamate is presented in Fig. 4, indicative of a competitive mechanism of inhibition. Calculated KI values in the presence of 3, 30, and 300 nM WAY-213613 were 15, 41, and 55 nM, respectively, consistent with the inhibitory potency determined for WAY-213613 in the determination of IC50 values described above. Similar experiments performed with WAY-213394, WAY-212922, and WAY-211686 indicated a competitive mode of inhibition for these compounds (data not shown).

    Oocyte Studies. On the basis of its superior potency and EAAT2 selectivity, the effect of WAY-213613 on glutamate-induced currents in oocytes was examined. Application of 30 e WAY-213613 (or indeed the other three compounds described in this study) to EAAT2 injected oocytes failed to induce a transporter-like current, indicating the compound to be a nonsubstrate (data not shown). Inward currents evoked by the application of glutamate to EAAT2-injected oocytes were blocked in a concentration-dependent manner by WAY-213613 (Fig. 5) with an estimated IC50 value of 130 nM. WAY-213613 produced a similar concentration-dependent block of glutamate-induced currents in either EAAT1- or EAAT3-injected oocytes (Fig. 5), with IC50 values of 48 and 4 e, respectively, reflecting 370- and 31-fold selectivity toward EAAT2.

    Ionotropic Glutamate Receptor Assays. Using a membrane potential-sensitive fluorescent indicator dye, we established that the addition of either NMDA, in the absence of Mg2+, or AMPA, in the presence of 250 e蘥/ml concanavalin A, to primary cultures of hippocampal neurons produced an increase in dye fluorescence indicative of depolarization. Under these experimental conditions, AMPA stimulation can result in the activation of both AMPA (directly) and kainate (by indirect release of glutamate) receptors. The depolarizing responses to NMDA or AMPA could be selectivity antagonized by either Ro 25-6981 or NBQX, respectively (Fig. 6, insets). In contrast, addition of 50 e WAY-213613, WAY-213394, WAY-212922, or WAY-211686 failed to produce a depolarization in these neurons indicating that the compounds failed to activate either NMDA or non-NMDA receptors (data not shown). In addition, the compounds were evaluated for their ability to antagonize depolarizing responses to either NMDA (Fig. 6A) or AMPA (Fig. 6B) in cultured hippocampal neurons. In the case of WAY-213613, the compound exhibiting the best EAAT2 selectivity, <40% inhibition of depolarizing response to either NMDA or AMPA was observed at the highest concentration tested (50 e), whereas no antagonist effect was observed at 5 e. From these data, it can be concluded that WAY-213613 exhibits good selectivity over ionotropic receptors; approximately 60- and 140-fold compared with its effect on L-[3H]glutamate uptake by the cloned EAAT2 subtype or L-[3H]glutamate uptake in rat cortical synaptosome fractions (EAAT2 predominant), respectively. WAY-211686 exhibited the most activity toward blocking NMDA-stimulated responses with 31 and 73% inhibition observed at 5 and 50 e, respectively. In the case of AMPA-stimulated responses, WAY-213394 exhibited 20 and 62% inhibition at 5 and 50 e, respectively.

    Metabotropic Glutamate Receptor Assays. WAY-213613 was evaluated for agonist and antagonist activity at human mGlu receptors 1, 2, and 4, examples of group I, II, and III mGlu receptors, respectively. As illustrated in Fig. 7, agonist responses were elicited at each receptor subtype by appropriate ligands; 1 e L-quisqualate for mGlu receptor 1, 0.1 e LY354740 for mGlu receptor 2, and 1 e L-AP4 for mGlu receptor 4. In contrast, 100 e WAY-213613 failed to activate any of the mGlu receptors examined (Fig. 7). In addition, agonist-stimulated responses were blocked by the mGlu receptor antagonist LY341495, used at concentrations appropriate to each mGlu receptor studied, whereas 100 e WAY-213613 failed to exhibit antagonist activity at any of the subtypes.

    Synaptosomal Exchange. Electrophysiological studies in oocytes indicated that WAY-213613 behaved in a manner consistent with a nonsubstrate inhibitor, failing to activate an inward transporter-like current when applied to EAAT2-injected oocytes yet blocking the glutamate-induced currents. A biochemical index that can be used for the determination of substrate versus nonsubstrate activity is the capacity for drug-mediated heteroexchange of accumulated D-[3H]aspartate substrate (Griffiths et al., 1994; Koch et al., 1999; Dunlop, 2001). Cortical synaptosomes were equilibrated with D-[3H]aspartate before the addition of exogenous unlabeled D-aspartate or WAY-213613 to assess drug-mediated 3H efflux. As illustrated, the addition of exogenous D-aspartate to D-[3H]aspartate-loaded synaptosomes stimulated the efflux of the 3H-label (Fig. 8); however, WAY-213613 failed to increase 3H efflux over control levels, consistent with the property of a nonsubstrate inhibitor.

    Molecular Modeling. A pharmacophore model containing a number of key features—hydrogen bond acceptor sites, positive ionizable (protonatable nitrogen), negative ionizable (acidic function), and two hydrophobic centers—was developed using both literature-reported EAAT inhibitors and the novel compounds presented in this study (Fig. 9). Figure 9 illustrates the fit of L-TBOA and WAY-213613 to the identified pharmacophore, indicating both overlapping and distinct conformity to the model. A major distinction between L-TBOA and WAY-213613 is the conformity of the former to all features with the exception of the hydrophobic center 4.11, whereas WAY-213613 fits all features with the exception of the hydrogen bond acceptor. We find that the newly described inhibitors presented in the current study lack a distal acidic group identified by others (Campiani et al., 2001, 2003) as an important pharmacophoric site corresponding to the hydrogen bond acceptor feature in our model. Moreover, the compounds reported in the current study share in common the distal hydrophobic center 4.11. These two distinctive features are candidates to be considered important in conferring EAAT2 selectivity of our newly identified agents. It is important to note that in the current model, the folded form of the aspartate framework in TBOA is well-superimposed with the folded form of glutamate (Fig. 9), whereas others have proposed the extended form for the aspartate framework in their model (Koch et al., 1999; Campiani et al., 2001, 2003). Each of these approaches has used different modeling scenarios, and although we have predicted the folded form of the aspartate framework, the other features of the molecule present a very similar configuration in all modeling studies. Moreover, close inspection of the models predicted by Koch et al. (1999) and Campiani et al. (2001, 2003) suggests that these do not predict exactly the same configuration of the aspartate framework. Our model provides a plausible alternative hypothesis for EAAT binding and should be considered in the rational design of new compounds.

    To gain further insight into the molecular basis contributing to the pharmacological specificity toward EAAT2 inhibition, the superimposition of the compounds presented in this study with other known EAAT inhibitors, including kainate, dihydrokainate, and L-TBOA, was used to generate composite molecular volumes of the novel agents. The superimpositions of the known and novel ligands are presented (Fig. 10, A and B, respectively), indicating significant lipophilic volume occupied by the novel compounds. Access to this distal lipophilic pocket is achieved consistently with all four novel compounds represented in this study.

    Discussion

    High-affinity glutamate transporters termed EAATs, play a critical role in maintaining the fine balance between the physiological actions of glutamate as an excitatory neurotransmitter and preventing the excessive accumulation of extracellular glutamate responsible for excitotoxic cell death. Although this fundamental function of the EAATs has been well established, a complete elucidation of the relative contribution of individual subtypes and a better understanding of their physiological roles has been hampered by the unavailability of potent and selective pharmacological tools.

    In the present study, we provide a pharmacological characterization of novel aryl-ether, biaryl, and fluorene aspartic acid and diaminopropionic acid analogs as nanomolar potent inhibitors of the high-affinity glutamate transporter EAAT2. Four compounds are described (WAY-213613, WAY-213394, WAY-212922, and WAY-211686). The aryl-ether aspartic acid analog WAY-213613 was both the most potent and EAAT2-selective of the compounds studied, exhibiting an IC50 value for inhibition of glutamate uptake by the cloned EAAT2 subtype of 85 nM and a 59- and 45-fold separation in potency for EAAT1 and EAAT3, respectively. In comparison, the biaryl substituted aspartic acid analog WAY-213394, although exhibiting similar EAAT2 potency (145 nM) and retaining selectivity toward EAAT3 (45-fold), exhibited reduced selectivity toward EAAT1 (14-fold). Using a fluorene ring scaffold in the aspartic acid analog WAY-212922 resulted in a compound with potent EAAT2 inhibitory activity accompanied by reduced selectivity over EAAT1 and EAAT3 of 22- and 12-fold, respectively. Finally, the biaryl-substituted diaminopropionic acid analog WAY-211686 also exhibited potent EAAT2 inhibitory activity but suffered from the poorest selectivity of the compounds examined (9- and 8-fold for EAAT1 and EAAT3, respectively).

    The inhibitory potency of the compounds was also confirmed in a native preparation of rat cortical synaptosomes, a model proposed on the basis of pharmacological studies to reflect predominantly EAAT2 activity (Dunlop et al., 1999; Tan et al., 1999). Each of the four compounds exhibited similar inhibitory potencies in the synaptosomal preparation and an identical rank order of potency, confirming their property as potent glutamate transport inhibitors and providing further evidence for the identity between EAAT2-mediated uptake and uptake by the transport system expressed in rat cortical synaptosomes. Additional studies in the synaptosomal preparation were undertaken to determine that the compounds behaved in a manner consistent with competitive inhibitors, suggesting that they overlap with the glutamate recognition site on the transporter. Given that the compounds retain the carboxylic acid moiety, it is perhaps not surprising that they would exhibit significant overlap with the glutamate-binding site.

    A third index used to estimate the inhibitory activity of the most potent and selective compound WAY-213613 was to evaluate its effect on glutamate-induced currents in oocytes expressing each of the transporters. It was observed that the inhibitory potencies for blocking glutamate-induced currents in both EAAT2- and EAAT3-injected oocytes were identical with those determined in the cell line uptake studies, whereas a 10-fold weaker IC50 value was observed in EAAT1-injected oocytes compared with EAAT1-expressing HEK cells. It is possible that such a discrepancy arises from the indirect index of glutamate translocation provided by current measurements in oocytes reflecting both ion conductances coupled to glutamate translocation and uncoupled ion flux. Studies in oocytes also allowed for the determination that WAY-213613 and the other compounds failed to activate a transporter-like current when applied to oocytes, indicating that they were not substrates for the transporter. Such an observation was confirmed with the independent biochemical index of 3H-substrate exchange with WAY-213613 failing to promote the exchange of accumulated D-[3H]aspartate. Given the bulky nature of the substituent group in WAY-213613, it is likely that this impedes substrate translocation despite significant overlap with the glutamate recognition site. Taken together, the above observations indicate WAY-213613 to be a potent, EAAT2-preferring, competitive nonsubstrate inhibitor of EAAT2.

    To assess the potential usefulness of WAY-213613 as a pharmacological tool to study EAAT2, it is of importance to provide an assessment of the compound at other principal sites of glutamate action, specifically the ionotropic and metabotropic glutamate receptors. Our studies demonstrate that WAY-213613 was devoid of either agonist or antagonist activity at the ionotropic NMDA and AMPA receptors expressed natively in cultured hippocampal neurons, or at the human mGlu receptors 1, 2, and 4 heterologously expressed in stable cell lines. These results confirm the specificity of WAY-213613 for the glutamate transport system.

    Compared with other known glutamate transporter inhibitors, WAY-213613 would seem to be among the most potent and selective of agents reported in the literature. Although kainate and dihydrokainate are known to be selective for EAAT2 compared with the other EAAT subtypes (Arriza et al., 1994), their usefulness as pharmacological tools are hampered by affinity for ionotropic and metabotropic glutamate receptors, respectively, in addition to their weak (20eC50 e) potency. Compounds such as the carboxycycloproyl glycines (Nakamura et al., 1993), aminocyclobutane dicarboxylates (Fletcher et al., 1991), pyrrolidine dicarboxylates (Bridges et al., 1991), and the threo-hydroxyaspartate derivative threo--benzyloxyaspartate (TBOA) (Shimamoto et al., 1998; Shigera et al., 2001) have low micromolar potency and do not discriminate between the different EAAT subtypes. More recently, we described WAY-855 as a novel EAAT2 inhibitor with an IC50 value of 1 e, although this compound only exhibited approximately 10-fold selectivity over EAAT3 (Dunlop et al., 2003). Recent structure-activity relationship studies with TBOA have resulted in the discovery of (2S,3S)-3-[3-(4-cyanobenzoylamino)benzyloxy]aspartate and (2S,3S)-3-[3-(4-methoxybenzoylamino)-benzyloxy]aspartate (Shimamoto et al., 2004) as potent EAAT2 inhibitors (IC50 values of 12 and 36 nM, respectively). However, although these compounds exhibited 22- and 39-fold selectivity over EAAT3, their inhibitory potencies for EAAT1 were essentially the same as those determined for EAAT2. Although WAY-213613 was slightly less potent than either of the TBOA analogs, the compound was more selective over both EAAT1 and EAAT3. By generating a pharmacophore model of known EAAT inhibitors and the novel agents reported here, we have identified a number of features associated with the newly identified compounds distinct from known EAAT inhibitors. These include the absence of a distal acidic group acting as a hydrogen bond acceptor identified in other known EAAT inhibitors and the presence of a hydrophobic center conferring significant lipophilic volume to the novel agents. Although by no means definitive, these observations may provide insight into structural features for selective EAAT2 binding and provide a working model for further validation toward rational design of selective EAAT inhibitors.

    In conclusion, we have identified a number of structurally novel EAAT inhibitors exemplified by WAY-213613, a potent, EAAT2-preferring and nonsubstrate inhibitor of EAAT2. These newly discovered EAAT2 inhibitors will be important tools to further study EAAT2 function in addition to continuing to refine the structural requirements for binding to different transporter subtypes.

    Acknowledgements

    We thank Drs. James Barrett and John Moyer for support.

    doi:10.1124/mol.105.012005.

    References

    Arriza JL, Eliasof S, Kavanaugh MP, and Amara SG (1997) Excitatory amino acid transporter 5, a retinal glutamate transporter coupled to a chloride conductance. Proc Natl Acad Sci USA 94: 4155eC4160.

    Arriza JL, Fairman WA, Wadiche JL, Murdoch GH, Kavanaugh MP, and Amara SG (1994) Functional comparison of three glutamate transporter subtypes cloned from human motor cortex. J Neurosci 14: 5559eC5569.

    Barnard EA (1997) Ionotropic glutamate receptors: new types and new concepts. Trends Pharmacol Sci 18: 141eC148.

    Bridges RJ, Stanley MS, Anderson MW, Cotman CW, and Chamberlin AR (1991) Conformationally defined neurotransmitter analogues. Selective inhibition of glutamate uptake by one pyrrolidine-2,4-dicarboxylate diastereomer. J Med Chem 34: 717eC723.

    Campiani G, De Angelis M, Armaroli S, Fattorusso C, Catalanotti B, Ramunno A, Nacci V, Novellino E, Grewer C, Ionescu D, et al. (2001) A rational approach to the design of selective substrates and potent nontransportable inhibitors of the excitatory amino acid transporter EAAC1 (EAAT3): new glutamate and aspartate analogues as potential neuroprotective agents. J Med Chem 44: 2507eC2510.

    Campiani G, Fattorusso C, De Angelis M, Catalanotti B, Butini S, Fattorusso R, Fiorini I, Nacci V, and Novellino E (2003) Neuronal high-affinity sodium-dependent glutamate transporters (EAATs): targets for the development of novel therapeutics against neurodegenerative diseases. Curr Pharm Des 9: 599eC625.

    Danbolt NC (2001) Glutamate uptake. Prog Neurobiol 65: 1eC105.

    Dunlop J (2001) Substrate exchange properties of the high-affinity glutamate transporter EAAT2. J Neurosci Res 66: 482eC486.

    Dunlop J, Eliasof S, Stack G, McIlvain HB, Greenfield A, Kowal D, Petroski R, and Carrick T (2003) WAY-855 (3-amino-tricyclo[2.2.1.02.6]heptane-1,3-dicarboxylic acid): a novel, EAAT2-preferring, non-substrate inhibitor of high-affinity glutamate uptake. Br J Pharmacol 140: 839eC846.

    Dunlop J, Lou Z, Zhang Y, and McIlvain B (1999) Inducible expression and pharmacology of the human excitatory amino acid transporter 2 subtype of L-glutamate transporter. Br J Pharmacol 128: 1485eC1490.

    Fairman W, Vandenberg RJ, Arizza JL, Kavanaugh MP, and Amara SG (1995) An excitatory amino acid transporter with properties of a ligand gated chloride channel. Nature (Lond) 375: 599eC603.

    Fletcher EJ, Mewett KN, Drew CA, Allan RD, and Johnston GA (1991) Inhibition of L-glutamate uptake into rat cortical synaptosomes by the conformationally restricted analogue of glutamic acid, cis-1-aminocyclobutane-1,3-dicarboxylic acid. Neurosci Lett 121: 133eC135.

    Griffiths R, Dunlop J, Gorman A, Senior J, and Grieve A (1994). L-trans-Pyrrolidine-2,4-dicarboxylate and cis-1-aminocyclobutane-1,3-dicarboxylate behave as transportable competitive inhibitors of the high-affinity glutamate transporters. Biochem Pharmacol 47: 267eC274.

    Halgren TA (1996) Merck molecular force field. I. Basis, form, scope, parameterization and performance of MMFF94. J Comp Chem 17: 490eC519.

    Kanai Y and Hediger MA (1992) Primary structure and functional characterization of a high-affinity glutamate transporter. Nature (Lond) 360: 467eC471.

    Koch HP, Kavanaugh MP, Esslinger CS, Zerangue N, Humphrey JM, Amara SG, Chamberlin AR, and Bridges RJ (1999). Differentiation of substrate and non substrate inhibitors of the high-affinity, sodium-dependent glutamate transporters. Mol Pharmacol 56: 1095eC1104.

    Kowal D, Nawoschik S, Ochalski R, and Dunlop J (2003) Functional calcium coupling with the human metabotropic glutamate receptor subtypes 2 and 4 by stable co-expression with a calcium pathway facilitating G-protein chimera in Chinese hamster ovary cells. Biochem Pharmacol 66: 785eC790.

    Lebrun B, Sakaitani M, Shimamoto K, Yasuda-Kamatani Y, and Nakajima T (1997) New -hydroxyaspartate derivatives are competitive blockers for the bovine glutamate/aspartate transporter. J Biol Chem 272: 20336eC20339.

    Nakamura Y, Kataoka K, Ishida M, and Shinozaki H (1993) (2S,3S,4R)-2(carboxycyclopropyl) glycine, a potent and competitive inhibitor of both glial and neuronal uptake of glutamate. Neuropharmacology 32: 833eC837.

    Pines G, Danbolt NC, Bjoras M, Zhang Y, Bendahan A, Eide L, Koepsell H, Storm-Mathisen J, Seeberg E, and Kanner BL (1992) Cloning and expression of a rat brain L-glutamate transporter. Nature (Lond) 360: 464eC467.

    Rothstein JD, Dykes-Hoberg M, Pardo CA, Bristol LA, Jin L, Kuncl RW, Kanaly Y, Hediger MA, Wang Y, Schielke JP, et al. (1996) Knockout of glutamate transporters reveals a major role for astroglial transport in excitotoxicity and clearance of glutamate. Neuron 16: 675eC686.

    Schoepp DD, Jane DE, and Monn JA (1999) Pharmacological agents acting at subtypes of metabotropic glutamate receptors. Neuropharmacology 38: 1431eC1476.

    Shigera Y, Shimamoto K, Yasuda-Kamatani Y, Seal RP, Yumoto, N Nakajima T, and Amara SG (2001) Effects of threo-beta-hydroxyaspartate derivatives on excitatory amino acid transporters (EAAT4 and EAAT5) J Neurochem 79: 297eC302.

    Shimamoto K, Lebrun B, Yasuda-Kamatani Y, Sakaitani M, Shigeri Y, Yumoto N, and Nakajima T (1998) DL-threo--Benzyloxyaspartate, a potent blocker of excitatory amino acid transporters. Mol Pharmacol 53: 195eC201.

    Shimamoto K, Sakai R, Takaoka K, Yumoto N, Nakajima T, Amara SG, and Shigeri Y (2004) Characterization of novel L-threo--benzyloxyaspartate derivatives, potent blockers of the glutamate transporters. Mol Pharmacol 65: 1008eC1015.

    Storck T, Schulte S, Hofmann K, and Stoffel W (1992) Structure, expression and functional analysis of a Na+-dependent glutamate/aspartate transporter from rat brain. Proc Natl Acad Sci USA 89: 10955eC10959.

    Tan J, Zelenaia O, Correale D, Rothstein JD, and Robinson MD (1999) Expression of the GLT-1 subtype of Na+-dependent glutamate transporter: pharmacological characterization and lack of regulation by protein kinase C. J Pharmacol Exp Ther 289: 1600eC1610.

    Tanaka K, Watase K, Manabe T, Yamada K, Watanabe M, Takahashi K, Iwama H, Nishikawa T, Ichihara N, Kikuchi T, et al. (1997) Epilepsy and exacerbation of brain injury in mice lacking the glutamate transporter GLT-1. Science (Wash DC) 276: 1699eC1702.