European Journal of Neuroscience, Vol. 8, pp. 2019-2028, 1996 @ European Neuroscience Association

The Competitive Transport In hi bitor L-trans-pyrrolidine- 2,4-dicarboxylate Triggers Excitotoxicity in Rat Cortical Neuron-Astrocvte Co-cultu res via G lutamate Release rather than Upfake Inhibition

Andrea Volterral, Paola Bezzil, Barbara Lodi Rizzini', Davide Trotti', Kyrre Ullensvang2, Niels C. Danbolt2 and Giorgio Racagnil 'Centre of Neuropharmacology, Institute of Pharmacological Sciences, University of Milan, Via Balzaretti 9, 201 33 Milan, Italy 2Department of Anatomy, Institute of Basic Medical Sciences, University of Oslo, P.O. Box 1105 Blindern, N-0137 Oslo, Norway

Keywords: glutamate transporters, non-vesicular release, heteroexchange, excitatory amino acid neurotransmission, neuron- astrocyte communication

Abstract

We studied the early and late effects of ~-trans-pyrrolidine-2,4-dicarboxylate (PDC), a competitive inhibitor of glutamate uptake with low affinity for glutamate receptors, in co-cultures of rat cortical neurons and glia expressing spontaneous excitatory amino acid (EM) neurotransmission. At 100 or 200 pM, PDC induced different patterns of electrical changes: 100 pM prolonged tetrodotoxin-sensitive excitation triggered by synaptic glutamate release; 200 pM produced sustained, tetrodotoxin-insensitive and EM-mediated neuronal depolarization, overwhelming synaptic activity. At 200 pM, but not at 100 pM, PDC caused rapid elevation of the glutamate concentration ([Glu],) in the culture medium, resulting in NMDA receptor-mediated excitotoxic death of neurons 24 h later. The increase in [Glu], was largely insensitive to tetrodotoxin, independent of extracellular Ca2+, and present also in astrocyte-pure cultures. By the use of glutamate transporters functionally reconstituted in liposomes, we showed directly that PDC activates carrier-mediated release of glutamate via heteroexchange. Glutamate release and delayed neurotoxicity in our cultures were suppressed if PDC was applied in a Na'-free medium containing Li'. However, replacement of Na' with choline instead of Li+ did not result in an identical effect, suggesting that Li' does not act simply as an external Na' substitute. In conclusion, our data indicate that alteration of glutamate transport by PDC has excitotoxic consequences and that active release of glutamate rather than just uptake inhibition is responsible for the generation of neuronal injury.

Introduction

Glutamate uptake is essential for the maintenance of the extracellular glutamate concentration ([Glu],) at the threshold of excitatory amino acid (EAA) receptor activation (Lipton and Rosenberg, 1994). Thereby it provides a high signal-to-noise ratio for excitatory transmission and prevents harmful receptor overstimulation. The uptake process is mediated by glycoproteins present in the plasma membrane of neuronal and glial cells. At least four different transporters have now been cloned (Kanai and Hediger, 1992; Pines et al., 1992; Storck et al., 1992; Fairman et al., 1995), and they have highly differentiated brain distributions (Rothstein et al., 1994; Chaudhry et al., 1995; Lehre et al., 1995). They appear to be regulated by protein kinase C (Casado et al., 1993), arachidonic acid (Trotti et al., 1995; Zerangue et al., 1995) and endogenous oxidants (Volterra et ab, 1994; Trotti et al., 1996). Whether changes in glutamate uptake influence fast excitatory transmission is somewhat controversial (Isaacson and Nicoll, 1993; Sarantis et al., 1993; Barbour et al., 1994; Maki et al., 1994; Mennerick and Zorumski, 1994; Tong and Jahr, 1994). On

the other hand, transport alterations have been associated with the neuropathologies of ischaemia (Szatkowski and Attwell, 1994) and amyotrophic lateral sclerosis (Rothstein et al., 1992, 1995).

Pharmacological agents able to interfere with glutamate uptake provide a tool for understanding the physiopathological significance of transport changes. One of the most interesting compounds is L-

trans-pyrrolidine-2,4-dicarboxylate (PDC; Bridges et al. 1991), a potent competitive uptake inhibitor with negligible effects on E M receptor currents, at least at concentrations S200 pM (Isaacson and Nicoll, 1993; Sarantis et al., 1993; Maki et al., 1994). When applied in combination with glutamate to cultured hippocampal neurons, PDC was able to potentiate the neurotoxic action of the transmitter, apparently by slowing down its cellular uptake (Robinson et al., 1993). However, PDC also promoted active efflux of ~-[~H]aspartate from preloaded cultured granule cells (Griffiths el al., 1994). There- fore, the type of transport alterations induced by this agent and their specific influences on excitatory transmission and toxicity are not

Correspondence to: Dr Andrea Volterra, as above

Received I 1 January 1996, revised 25 April 1996. accepted 3 May 1996

2020 Transport alteration induces neurotoxic glutamate release

well understood. The purpose of the present work is to investigate these aspects by monitoring the effects of PDC on electrical activity, [Glu], and long-term neuronal survival in co-cultures of cortical nerve cells and glia that express spontaneous EAA neurotransmission.

Materials and methods

Tissue culture Astrocytic cell culture

Primary astroglial cell cultures were obtained from the cerebral cortices of newborn (postnatal day 1) rats by mechanical dissocia- tion as previously described (Volterra et al., 1992). Cells were first grown for 2 weeks in 75 cm2 flasks (37"C, 5% co2/95% humidified air atmosphere) in Eagle's Minimal Essential Medium (MEM) plus 4 mM glutamine, 0.6% wt/vol glucose, 100 U/ml penicillinktrepto- mycin and 20% fetal calf serum (FCS). Then they were enriched in flat type-I astrocytes (>98%) following the procedure of McCarthy and de Vellis (1980). and replated at 2 X lo5 viable celldml of MEM plus 10% FCS in 35 mm Petri dishes, and the monolayers were allowed to reach confluence (=7 days). Typical glial protein content (measured by the method of Lowry eral., 1951) was 200-300 pg/dish.

Neuron-astrocyte co-cultures

Cortical cultures containing both neurons and glia were prepared from rat embryos (embryonic day 17) by mechanical dissociation essentially as described by Hartley et al. (1993). The cell suspension in MEM (plus 2 mM glutamine, 0.3% wt/vol glucose, 1.5 U/ml penicillidstreptomycin, 12 mM NaHC03) with 10% FCS plus 10% horse serum was adjusted to 5 X lo5 viable cells/ml and plated in 35 mm Petri dishes already containing a confluent astrocytic mono- layer. After 7 days, co-cultures were exposed to cytosine arabinoside (Ara-C; 5 pM) for 24 h and had the medium replaced with MEM plus 10% horse serum. They were used for experiments after 5-7 more days in vitro, when an extensive synaptic network had formed (Fig. 3C). In a few cases, co-cultures were selectively depleted of nerve cells 1 day before the experiment by repeatedly aspirating and replacing the culture medium with a Pasteur pipette. This procedure reduces the overall protein content of the co-culture (typically 6 W 800 pg/dish) by -100 pg.

Electrophysiological patch-clamp recordings The current-clamp mode of the whole-cell patch-clamp technique was used to monitor membrane potential changes from neuronal cells co-cultured with astrocytes (9-14 days in vitro). Recordings were performed with a List EFT7 patch-clamp amplifier (List Electronic, Darmstadt, Germany). Patch pipettes, made from borosilicate glass capillary tubes with a Flaming Brown P-80PC puller (Sutter Instru- ments, San Rafael, CA), had a resistance of 4-5 Mi2 when filled with intracellular solution (in mM): 140 KCI, 0.5 CaC12, 5 EGTA, 5 Mg-ATP, 10 HEPES, pH adjusted to 7.3 with KOH. Nerve cells typically had a resting potential between -60 and -65 mV. Junction potentials were compensated for before seal formation. During the experiments cultures were continuously superfused with the following extracellular solution (in mM): 140 NaC1, 3.4 KCl, 1.8 CaCI2, 1 MgC12, 20 glucose, 5 HEPES (pH adjusted to 7.4 with NaOH); 20 pM bicuculline was also present to block GAJ3AA receptors and to study excitatory transmission in better isolation. Changes performed in some experiments included omission of Mg", addition of tetrodo- toxin (1 pM) and replacement of Na' with Li' (see below). Pharmacological agents were locally applied to cells through a six-

barrel microperfusion system (mild flow), allowing several conditions to be compared in the same cell.

Exposure of cell cultures to pharmacological agents Exposure of cell cultures to pharmacological agents was carried out according to Hartley et al. (1993) with slight modifications. The culture medium was replaced by quadruple exchange with a HEPES- buffered control salt solution (HCSS) containing (in mM): 120 NaCI, 3.4 KCl, 1.8 CaCI2, 1 MgCI2, 25 glucose, 20 HEPES Na salt, pH 7.4. Agents, at final concentration, were then applied in HCSS (37"C, 30 min, unless otherwise specified). Exposures were terminated by rapid replacement (= 1000-fold dilution) of the incubation medium with MEM containing (in mM): 2 glutamine, 12 NaHC03.25 glucose, 1.5 U/ml penicillin/streptomycin without FCS; and cultures were placed in the incubator for another 24 or 48 h. The incubation medium, divided into aliquots, was used for measuring cell-released [Glu], or lactate dehydrogenase (LDH) activity. In the experiments in which Na' was replaced by Li', HCSS contained 120 mM LiCl and 20 mM HEPES in acid form. The pH was adjusted to 7.4 with LiOH.

Bioluminescent assay of [Glu], The [Glu], value of cortical cultures was measured with the bio- luminescent assay described by Fosse et al. (1986) using an LKB Wallac (model 1251) luminometer. Briefly, 200 p1 of glutamate- specific reagent mixture [containing: 25 mM KPi, 100 pM dithio- threitol, 5 pg/ml luciferase (P. Fischerii), 400 mU/ml flavin mono- nucleotide (FMN) reductase (P. Fischerii), 2.5 pM FMN, 40 pg/ml Triton X-100, 20 pM tetradecanal, 2 mM NAD, 250 pM ADP, 0.5 mg/ml glutamate dehydrogenase and 16% glycerol] were allowed to react with 50 p1 of either HCSS incubation medium or L-glutamate (in 25 mM KPi, for standard curves) in polystyrene cuvettes at a fixed temperature (27.5"C) in the luminometer until a plateau of light emission was reached (-5 min). The background (from the reaction mixture and incubation medium) was subtracted. The values of [Glu], in the incubation medium were extrapolated by linear regression fitting of a standard L-glutamate curve (0.1-5 pM). In the experiments in which we determined the basal [Glu], value in neuron-astrocyte co-cultures and in astrocyte-pure cultures, the absolute values were confirmed by HPLC with electrochemical detection.

Assessment of overall cell damage Overall cell injury in cortical cell cultures was evaluated 24 or 48 h after pharmacological exposures, as described by Choi et al. (1987). Cultures were first inspected by phase-contrast microscopy and photographs were taken for morphological examination. A quantitative estimate of cell damage was obtained by measuring LDH activity released by injured cells in the culture medium during the post- treatment period (Koh and Choi, 1987). In a few experiments extracellular LDH was determined directly in the medium of incuba- tion with PDC. In all cases, 0.15 ml samples of cell supernatants were reacted with 1 ml of LDH diagnostic kit, and enzyme activity was measured spectrophotometrically and expressed as units/l/min according to the calculation protocol of the kit manual. Release of LDH due to pharmacological treatment in each experiment was compared with values for controls for both minimal cell damage (HCSS incubation) and maximal neuron damage (500 pM glutamate, 10 min).

PfHjaspartate uptake into neuron-astrocyte co-cultures Sodium-dependent uptake of EAA in co-cultures was determined using the non-metabolizable glutamate analogue D-aspartate as

Transport alteration induces neurotoxic glutamate release 2021

radiotracer. Treatments were performed as described above (section headed Exposure of cell cultures to pharmacological agents). Co- cultures were incubated for 5 min at 37°C with HCSS containing 0.5 pCi o-[%H]aspartate, adjusted to 1 pM with cold D-aspartate. Uptake was found to be linear with time. Details of the uptake assay and the calculation of specific Na+-dependent uptake are reported in a previous work (Volterra et al., 1992).

L-pyglutamate transport in liposomes Transport function was studied on CHAPS-solubilized proteins from crude rat brain synaptic plasma membranes reconstituted in liposomes as previously described (Danbolt et al., 1990; Trotti et al., 1995). Briefly, membranes were mixed with CHAPS-containing buffer and centrifuged. The supernatant was mixed with reconstitution mixture (cholate, phospholipids and salt) and gel-filtered on Sephadex G-50 to remove detergent. Liposomes form spontaneously during this gel- filtration and the solution with which the column is equilibrated becomes the internal medium of the liposomes. In some experi- ments (Table I , experiment A) liposomes were prepared with internal K+ (100 mM KCl, 50 mM K-HEPES, pH 7.5). These liposomes were then actively loaded (Danbolt et al., 1990) with labelled glutamate by incubation with 0.2 pM ~-[~H]glutamate in NaSCN buffer (100 mM NaSCN, 50 mM Na-HEPES, pH 7.5). To start release, after 3 min the liposomes were gel-filtered on columns containing either the above NaSCN buffer or LiSCN buffer (100 mM LiSCN, 50 mM Li-HEPES, pH 7.5). PDC (0 or 200 pM) was added to the gel-filtered liposomes. The reaction was terminated after 3 min by filtration through Millipore filters (HAW, 0.45 pm pores). The radioactivity in the liposomes retained on the filters was determined by liquid scintillation counting. In other experiments (Table 1, experiment B) liposomes were prepared with Na+ buffer (100 mM NaCl, 50 mM Na-HEPES, pH 7.5) with or without 1 mM PDC as the internal medium. External PDC was removed by gel filtration of the liposomes in Na+ buffer without PDC. The transport reaction was initiated by the addition of 0.1 pM (external) labelled glutamate and terminated after 3 min as above.

Materials PDC, L-glutamate, NMDA, 2-amino-5-phosphonopentanoate (APV), 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and ( 2 ) a-methyl-4- carboxyphenyl-glycine (MCPG) were from Tocris (Bristol, UK). GYKI 52466 was a kind gift of Drs I. Tarnawa and S. Farkas (Institute of Drug Research, Budapest, Hungary). GYKI 52466 and CNQX were dissolved in 0.3% (voVvol) DMSO (DMSO alone had no effect on [Glu], or LDH). ~.[~H]glutamate and ~-[~H]aspartate (50 Ci/mmol) were from Amersham, Sephadex G-50 was from Pharmacia (Uppsala, Sweden), and CHAPS, tetrodotoxin, EGTA, bicuculline, DNase and Ara-C were from Sigma, while most of the salts were from Merck. Cell culture media, MEM, FCS, horse serum, penicillidstreptomycin and glutamine were from ICN Flow (Costa Mesa, CA). The reaction mixture for the bioluminescent assay contained LKB's NAD(P)H monitoring kit (catalogue no. B-1243-104, Turku, Finland) + glutamate dehydro- genase (GDH, Sigma) andFMN reductase (Boehringer). The diagnostic kit for LDH (catalogue no. 3399) was from Merck.

Results Different effects on EAA transmission by PDC at increasing concentrations

After 9-14 days in vitro neurons co-cultured with astrocytes appeared highly interconnected (Fig. 3C). In a 1 mM Mg2+ buffer containing

CONTROL A

CW1RU B PDC I00 v+4

POC T T X C

PDC CNOX+APV D

I I h h dm*

FIG. 1 . Excitatory activity in neuron-astrocyte co-cultures and effect of 100 pM PDC. (A) Representative pattern of spontaneous excitatory events recorded from a nerve cell (V, = 4i4 mV) after 12 days in co-culture. The cell was bathed in a medium containing 20 pM bicuculline with (left) or without (right) 1 mM Mg2+. Notice that in the absence of Mg2+ most of the postsynaptic depolarizations trigger action potential discharges associated with paroxysmal depolarization shift-like phenomena. Recordings in B, C and D are all from the same nerve cell (V, = 4 0 mV) at subsequent times and are typical of nine different experiments. (B) Changes in excitatory activity during application of 100 pM PDC. The short excitatory events are transformed into prolonged bursts of activity separated by resting periods. On the right (WASH): 1 min after washout of PDC the initial activity is restored. (C) Three minutes later, perfusion of tetrodotoxin ('lTX; 1 pM) reversibly abolishes the excitatory activity and PDC has now no effect on membrane potential. (D) About 5 rnin later postsynaptic excitation is blocked by ionotropic EAA receptor antagonists (10 pM CNQX + 50 pM APV); also in this condition PDC is without effect.

20 pM bicuculline they showed variable patterns of spontaneous synaptic activity including large excitatory potentials, often reaching threshold to trigger one or more action potentials (Figs 1 and 2). Removal of Mg2+ (Fig. 1A) led to a substantial increase in activity (i.e. increased frequency and multiple action potential discharges). Addition of tetrodotoxin (1 pM) or a mixture of CNQX (10 pM) and APV (50 pM) suppressed the activity, revealing its dependence on glutamate release and EAA receptor activation (Figs 1 and 2). Superfusion of PDC onto the cells strongly affected excitatory transmission. The characteristics of the PDC effect were very different at 100 and 200 pM. At 100 pM, PDC shifted the neuronal firing from a quite regular pattern to an epileptiform pattern, causing the appearance of prolonged afterpotential depolarizations (20-40 mV) with bursts of paroxysmal activity superimposed, separated by resting intervals (Fig. IB). In the presence of 200 pM PDC spontaneous spiking was rapidly abolished. The nerve cells were unable to repolarize after firing and progressively adjusted to a steadily depolarized level (- +30 mV), which was maintained throughout the application of PDC (in some cases >3 min; Fig. 2A). The electrical changes induced by PDC (100 or 200 pM) were reversed by washout of the compound. In the presence of tetrodotoxin, which abolished

2022 Transport alteration induces neurotoxic glutamate release

PDC 200 pt.4 WASH CONTROL A

PDC B TTX

FIG. 2. Sustained neuronal depolarization induced by 200 pM PDC and its blockade by EAA antagonists but not by tetrodotoxin (TTX). Recordings in A. B and C were taken at different times during the same experiment, representative of eight experiments. (A) Sustained membrane depolarization induced by 200 pM PDC. In this case PDC was applied immediately after an action potential: the cell could not repolarize and its membrane potential was driven 32 mV above resting potential ( V , = -58 mV), a level maintained throughout the period of PDC perfusion. On the right, washout of PDC restores the suppressed activity. (B) About 5 min later, with the excitatory activity abolished by tetrodotoxin. 200 pM PDC still causes slow membrane depolarization (+26 mV). In different experiments, the level of this depolarization and its time to peak vaned significantly. (C) After suppression of the activity with CNQX + APV. PDC had no effect.

circuital transmission, 100 pM PDC had no effect on membrane potential (Fig. IC), while 200 pM PDC still induced membrane depolarization, although less prominent than without tetrodotoxin (+17.8 ? 5.6 mV at peak, n = 16; Figs 2B and 7A). This depolarization was prevented by treatment with a combination of the EAA receptor antagonists CNQX and APV (Fig. 2C).

PDC at 200 pM increases [Glul0 and causes delayed neuroexcitotoxicity The basal glutamate concentration in the medium of our co- cultures was consistently -1 pM. We evaluated the effects of 100 and 200 p.M PDC on this [Glu], value. After 30 min of application, 100 pM PDC slightly increased [Glu],, whereas 200 pM PDC increased it 5-fold (Fig. 3A). This dramatic difference might reflect very different inhibitory potencies of the two PDC concentrations on EAA uptake. However, 100 pM PDC inhibited ~-[~H]aspartate uptake (see Materials and methods) by 35 2 4% and 200 pM PDC inhibited it by 53 ? 3% (n = 3). We then studied the increase in extracellular LDH as an index of cell damage (Koh and Choi, 1987). During application of 200 pM PDC (15 or 30 min), the LDH activity in the culture medium remained identical to the control, suggesting that the [Glu], elevation was not due to non-specific leakage of the transmitter from damaged cells. However, 24 h later extracellular LDH was significantly increased (Fig. 3A). In addition, a picture of diffuse nerve damage had developed: few cell bodies were spared and the neuritic network appeared highly disorganized, with the processes extensively blebbed. No relevant damage was observed in the glial layer underneath (Fig. 3D). Accordingly, 200 FM PDC did not induce an increase in LDH in astrocyte-pure cultures (not shown). Treatment with 100 pM PDC left the appearance of our co-cultures as well as their LDH level identical to that of controls. Therefore, the early [Glu], changes produced by PDC and the late neurotoxicity were closely correlated. Moreover, the two events were linked in a cause-effect relationship. This was shown by incubating 200 pM PDC with glutamate dehydrogenase (GDH, 100 U/ml + 3 mM NAD+ + 25 mM ADP), the enzyme that metabolizes glutamate to a-ketoglutarate. Thus, with GDH present, [Glu], decreased to <0.2 pM instead of increasing, and PDC was now completely

ineffective in inducing delayed neuronal damage. a-Ketoglutarate (300 pM) alone was neither toxic nor protective against PDC toxicity (Fig. 3B).

The EAA receptor types involved in the neurotoxic action of PDC were then investigated. Antagonists selective for the different receptor families were added at maximal inhibitory concentrations in conjunc- tion with PDC. As shown in Figure 3B, APV (100 pM), the prototype NMDA antagonist, completely prevented the delayed increase in LDH. However, CNQX (10 pM) and GYM 52466 (50 pM), two AMPAkainate blockers of the competitive and non-competitive type (Donevan and Rogawski, 1993), as well as (-t)MCPG (500 pM), a competitive inhibitor at most metabotropic subtypes (Watkins and Collingridge, 1994), also significantly reduced the effect of PDC.

Neurotoxic [ Glu] elevation is via a tetrodotoxin-insensitive and Cz?' -independent mechanism

As shown in Figures 1 and 2, tetrodotoxin ( I pM) abolished the circuital excitatory activity in our co-cultures, leaving only small fluctuations of the baseline potential, probably reflecting miniature excitatory postsynaptic potentials. A 30 min incubation with the toxin slightly decreased basal [Glu], (by 30%). but did not significantly prevent the rapid rise in [Glu], and the delayed neurotoxicity induced by 200 pM PDC (Fig. 4). In other experiments, we added the Ca2' chelator EGTA (3 mM) to the incubation medium in order to block Ca2+ entry into the cells and exocytotic release. Like tetrodotoxin, EGTA slightly decreased basal [Glu], without affecting the increase in [Glu], induced by PDC (Fig. 4). However, in contrast to tetrodotoxin, it completely abolished the delayed neuronal injury. Therefore, external Ca2+ influx is required for expression of PDC neurotoxicity but not for induction of the increase in [Glu],. Overall, the above data indicate that PDC cannot promote neurotoxic [Glu], elevation simply by slowing the re-uptake of the glutamate pool released by synaptic activity in the culture.

Lithium blocks all the effects of PDC Although insensitive to the above treatments, the increase in [Glu],, and its neurotoxic consequences were completely prevented by

Transport alteration induces neurotoxic glutamate release 2023

A, # m 500 - 0 [ G I U ] ~ 0 a n LDH L 1- al >

al m 0

0 300 -

PDC (gM) 100 200 2 00 30' 15' 30'

FIG. 3. Increase in [Glu], and delayed excitotoxicity in neuron-astrocyte co-cultures exposed to 200 pM PDC. (A) Early changes in [Glu], and late changes in extracellular LDH induced by 100 or 200 pM PDC. Hollow bars show the percentage increase over basal [Glu], measured at the end of PDC application (protocols: 30 min, 100 pM: 15 min, 200 pM; 30 min, 200 pM). Solid bars show the percentage increase over basal LDH measured 24 h after the same treatments. Notice the close correlation between early glutamate elevation and delayed cell toxicity. A two-tailed Student's r-test indicates that 15 or 30 min of incubation with 200 but not with 100 pM PDC significantly increased [Glu], (*) and LDH (#) with respect to controls. Data represent the average 2 SEM of at least five experiments in triplicate. (B) Pharmacological characterization of PDC neurotoxicity. The late LDH accumulation induced by 30 min of exposure to 200 pM PDC is defined as 100% (solid bar). This is -70% of the effect produced by 500 pM glutamate (10 min), which gives complete destruction of all neurons. NMDA at 500 pM is nearly equipotent to glutamate in Mg2+-free medium, but 8-fold less potent than PDC in a medium containing 1 mM Mg2+. One-way analysis of variance and Tukey's test for multiple comparisons indicate that all the treatments tested significantly reduced the neurotoxic action of PDC (*). Among the EAA receptor antagonists (open bars), APV provides the greatest protection (#), similar to that given by GDH. The metabolic products of the reaction between glutamate and GDH, a-ketoglutarate and ammonium ion (a-KGT, 300 pM) are non-toxic. Data (?SEM) are from at least four experiments in triplicate. (C, D) Representative phase-contrast photomicrographs of co-cultures taken 24 h after 30 min of exposure to control buffer (C) or 200 pM PDC (D). Scale bar, 50 pm. In C is seen the high-density network of interconnected nerve cells typical of our cultures. In D the neuritic processes have lost their organization and only a small number of neuronal cell bodies can be recognized. In some areas, the glial cell layer underneath has become visible.

application of 200 ph4 PDC in a medium lacking Na' and containing Li+ as substitute (Fig. 5A). To our suxprise, the use of a different Na+ substitute, choline instead of Li', did not produce the same effect. Thus, with choline buffer we observed about a 2-fold increase in basal [Glu],, not seen with Li', and a further increase with PDC (to >300% of control). Moreover, incubation of the cells in a choline buffer itself caused a delayed increase in LDH which was not significantly potentiated by PDC.

We then compared the depolarizing effect of 200 ph4 PDC on nerve cells that were alternately bathed in normal Na' and Li+ buffer (containing 1 pM tetrodotoxin; Fig. 5B). On average, PDC depolarization peaked at 21.4 2 5.9 mV in Na+ (n = 8) but at only 4.1 2 1.9 mV in Li'. In contrast, the depolarizing action of exogenous glutamate (30 pM) was identical in Li' and in Na' buffer (>30 mV, n = 7).

The above data strongly suggest that the increase in [Glu],, neuronal depolarization and late neurotoxicity induced by PDC are all related consequences of the same, Li'-sensitive mechanism(s) activated by PDC.

PDC promotes LP -sensitive [Glu],, elevation also in glial cultures The cellular compartments releasing glutamate were identified by comparing the effect of PDC in normal co-cultures and in co-cultures depleted of neurons but containing a similar amount of protein (see Materials and methods). PDC at 200 pM increased [Glu], in both types of culture; however, in the normal co-cultures [Glu], increased to 4- to 5-fold above control (Fig. 3A), while in the neuron-depleted co-cultures it was less than doubled (Fig. 6A). This experiment indicates not only that nerve cells make the largest contribution to the elevation of [Glu],, but also that glial cells participate in the overall effect. Therefore, we further investigated the mechanism of [Glu], elevation induced by PDC in pure type-1 astrocytic cultures, i.e. in the absence of classical vesicular glutamate release. By analogy with co-cultures, the basal [Glu], value for astrocytes (-0.3 pM) was not modified by 100 pM PDC. However, higher concentrations of PDC (200-1000 pM) increased it dose-dependently (up to -4-fold; Fig. 6A). This increase in [Glu], was tetrodotoxin- and EGTA- insensitive (not shown). Incubation of PDC in Na+-free media

2024 Transport alteration induces neurotoxic glutamate release

containing either Li+ or choline had a very different outcome depending on the Na' substitute used. Thus, in Li +-buffer the increase in [Glu], induced by PDC was almost abolished (-85%), while in choline-buffer it was only partly reduced (-25%; Fig. 6B).

HCSS l T X EGTA X n I :

FIG. 4. Influence of tetrodotoxin (TTX) and EGTA on the increases in [Glu], and LDH triggered by 200 pM PDC. Upper panel shows [Glu], (expressed as percentage of basal) in different media after a 30 min incubation without (open bars) or with (solid bars) 200 pM PDC. The media were control buffer alone (HCSS), buffer plus 1 pM tetrodotoxin, and buffer plus 3 mM EGTA. Data are averages ? SEM of ten (HCSS), seven (tetrodotoxin) and four (EGTA) experiments in triplicate. One-way analysis of variance plus Tukey's test for multiple comparisons indicated that (*) PDC induced a significant increase in [Glu], in all media tested, and that (#) in the medium with tetrodotoxin the increase was lower. However, this was true only if the increase was referred to the basal [Glu], without tetrodotoxin and not to that with tetrodotoxin (because tetrodotoxin decreases basal [Glu],). Lower panel shows extracellular LDH accumulation measured 24 h after the above treatments in the same experiments. PDC caused a significant increase in LDH over control in HCSS and in tetrodotoxin-containing medium (*) but not in medium with EGTA. In the presence of tetrodotoxin the elevation was lower (#).

No+ Li+ Choline PI

0 m I U I T

B

PDC activates exchange-release of glutamate from reconstituted glutamate transporters

The molecular basis underlying the PDC-induced increase in glutamate was investigated using glutamate transporters reconstituted in liposomes (see Materials and methods). As shown in Table 1 (experiment A), liposomes loaded with Kf and ~-['H]glutamate released most of their content of labelled glutamate after a 3 min incubation with external PDC (200 pM) in Na' buffer. When PDC was in Li' buffer, the release of internal glutamate was reduced (by -50%) but not abolished. Identical results were obtained using unlabelled glutamate instead of PDC (not shown). In another type of experiment (Table 1, experiment B), liposomes were prepared with or without 1 mM internal PDC and exposed to external ~-[~H]glutamate in the absence of transmembrane ion gradients (i.e. 150 mM Na+ on both sides of the liposomal membrane). Liposomes with internal PDC accumulated external ~-[~H]glutamate, whereas liposomes without PDC did not take up external ~-[~H]glutamate. These data provide a direct molecular demonstration that PDC activates a transporter- mediated exchange process with glutamate. Such a mechanism could account for initiation of release of internal glutamate from cells.

Discussion

Two different steps of glutamate transport alteration with PDC Exposure of cortical neuron-astrocyte co-cultures to the competitive glutamate uptake inhibitor PDC (100 or 200 pM) leads to changes in spontaneous EAA neurotransmission, cell-released glutamate and long-term neuronal survival. Our data strongly indicate that the observed dysfunctions result from the alterations induced by PDC in the transport of endogenous glutamate rather than from a direct action of the compound on EAA receptors. Thus, at least two of our findings are incompatible with the latter possibility: (i) suppression of PDC neurotoxicity by GDH, an enzyme that catalyses the metabolic transformation of glutamate but not of PDC (Fig. 3B); and (ii)

Na+ PDC GLUT

Li+ POC GLUT

n

FIG. 5. Effects of 200 pM PDC on [Glu],, LDH and neuronal membrane potential in Na+-free media containing Li' or choline. (A) Upper panel shows percentage increase over basal [Glu], observed after incubations (30 min) in normal Naf buffer or in Na+-free buffer containing Li' or choline. Open bars (CTRL) show effects of the Na' substitutions themselves: solid bars (PDC) show effects of 200 pM PDC. Lower panel shows LDH changes measured 24 h after the above treatments in the same experiments. One-way analysis of variance plus Tukey's test for multiple comparisons indicated that (i) Na' substitution with choline, but not with Li', resulted in a significant increase in both [Glu], and LDH (+), and (ii) the effects of PDC on [Glu], and LDH were significantly lower in Li+- as well as in choline-containing medium (*), but (iii) the extent of reduction was much bigger in Li' (#). Data represent the mean 2 SEM of six experiments in triplicate. (B) Effects of PDC (200 pM) and glutamate (30 pM) on neuronal membrane potential in Na+ buffer (upper trace) and in Li+ buffer (lower trace). Recordings are from the same cell (representative of eight) at different times. Tetrodotoxin ( 1 pM) was present throughout the experiment. In Na+ buffer the cell was depolarized by both PDC (+20 mV) and glutamate (+32 mV). Notice that, despite an identical perfusion condition, PDC had an onset of action much slower than glutamate. In Li+, the effect of PDC was abolished, while the depolarization induced by glutamate was identical to that seen in Naf. In all experiments, when the bathing solution was switched from Na+ to Li+ buffer the resting potential adjusted to a slightly depolarized value (on average, +8.7 ? 2.3 mV, n = 8). This depolarization was insensitive to the ionotropic EAA receptor antagonists (+7 .3 ? 2 mV, n = 4, with CNQX + APV) and was also seen in a few experiments with choline instead of Li+.

Transport alteration induces neurotoxic glutamate release 2025

inhibition by Lif of the depolarizing effect of PDC but not of that of exogenous glutamate (Fig. 5B).

The types of effects induced by PDC at 100 and 200 pM are quite different. At 100 pM, PDC shifts the normal pattern of spontaneous excitatory activity seen in our cultures towards an epileptiform behaviour (Fig. IB). If applied during resting intervals or in the presence of tetrodotoxin, PDC has no evident action. This is revealed when the cell depolarizes or fires an action potential in response to synaptic release of glutamate, because the excitatory events now appear greatly prolonged. This type of change is consistent with a mechanism of reuptake inhibition by PDC, which would extend the time spent by glutamate in areas close to the release sites with

POC(FM) 100 200 1000

LL

0 50

30 L1 10

Li + Choline

FIG. 6. PDC-induced [Glu], elevation in glial cultures: effect of external Na' substitution with Li' or choline. (A) The bars represent the percentage increase over basal [Glu], observed in the medium of astrocyte-pure cultures exposed for 30 rnin to increasing concentrations of PDC. The increase produced by 200 and lo00 pM PDC was significant (*. two-tailed Student's r-test). Data are the average * SEM of four experiments in triplicate. (B) The [Glu], elevation induced by lo00 pM PDC (defined as 100%) was differently attenuated in Na+-free media containing Lif or choline. In both cases the reduction reached statistical significance (*). However, inhibition with Li' was much higher than with choline (#, one-way analysis of variance + Tukey's test for multiple comparisons). Data are mean 2 SEM of six experiments in triplicate.

increased activation of the EAA receptors in those areas. However, during the superfusion of PDC, these bursting episodes eventually come to an end, giving place to silent periods where the cell potential remains at its resting value. Therefore, in the presence of 100 pM PDC, the residual capacity to remove glutamate (through diffusion or uptake) must be enough to bring the local [Glu], value back to the threshold for EAA receptor activation before the next synaptic discharge. We find that 100 pM PDC causes a 35% reduction in D- [3H]aspartate uptake by our cultures. The Ki values of PDC on neuronal (granule cells, 40 pM) and astrocytic (100 pM) uptake reported by other authors are also in line with incomplete uptake blockade by 100 pM PDC (Griffiths et al., 1994). In agreement, continued PDC application (30 min) does not significantly modify the [Glu], in neuron-astrocyte co-cultures or in astrocyte-pure cultures and does not promote extracellular LDH accumulation or major signs of cell suffering within the next 24-48 h.

At 200 pM, PDC induces sustained neuronal depolarization (>30 mV) that lasts as long as the compound is applied. As a consequence, the synaptic events are overwhelmed and abolished (Fig. 2A). This depolarizing action of PDC (i) is largely independent of the glutamate released by circuital excitatory activity, because it is seen also in the presence of tetrodotoxin; (ii) is mediated by activation of the ionotropic EAA receptors, being sensitive to the blocking action of APV + CNQX; and (iii) is indirect, because it is abolished in Li' buffer, a condition that does not affect the depolariz- ing action of exogenous glutamate (Fig. 5B). The above findings correlate well with the observations that (i) 200 pM PDC rapidly promotes an increase in the ambient [Glu], (Fig. 3A), an indication that the capacity for glutamate clearance has been completely over- whelmed, (ii) this increase in [Glu], is largely tetrodotoxin-insensitive (Fig. 4), and (iii) it is abolished in Li' buffer (Fig. 5A). The above changes cannot be explained by the modest increment in inhibitory potency on uptake shown by 200 pM PDC in comparison with 100 pM (+20%), but they most likely depend on the additional ability of 200 pM PDC to activate carrier-mediated release of glutamate.

PDC induces glutamate release via transporter-mediated heteroexchange

Our study demonstrates that the [Glu], elevation induced by 200 pM PDC largely depends on non-vesicular rather than on exocytotic

TABLE I . ~-[~H]glutamate is exchanged for PDC via glutamate transporters reconstituted in liposomes Experiment A: release of ~-[~H]glutamate ([3H]glu) from preloaded liposomes in response to external PDC

External medium Internal medium [3H]glutamate retained (% of control)

150 mM Na+ (control) 150 mM K+ + [3H]glu 100% 150 mM Li' 150 mM Na+ + 200 pM PDC

150 mM Kf + [3H]glu 150 mM K + + [3H]glu

93 * 9% 44 2 5% 67 * 6% 150 mM Li+ + 200 pM PDC

Experiment B: accumulation of external [3H]glutamate by PDC-containing liposomes in the absence of ion gradients 150 mM K + + [3H]glu

External medium Internal medium [3H]-GLU incorporated (c.p.m.)

150 mM Na+ 150 mM Na' 1011 2 68 150 mM Na' 150 mM Na+ + 1 mM PDC 31 324 ? 1069

In experiment A, liposomes with internal K+ were first allowed to take up [3H]glutamate (see Materials and methods). These actively loaded liposomes were then incubated for 3 min in Na' or Li' buffer 2 200 pM PDC and the remaining radioactivity was determined. [3H]Glutamate retained in liposomes exposed to Na' buffer (control) was defined as 100%. In experiment B liposomes were prepared with or without 1 mM PDC inside. [3H]Glutamate (0.1 pM) was added to the external medium and the amount of radioactivity taken up by the liposomes was measured after 3 min of incubation at room temperature. Note that the ionic compositions of the external and internal media of the liposomes are identical. The data for experiments A and B represent the average 2 SEM of three experiments performed in duplicate.

2026 Transport alteration induces neurotoxic glutamate release

release. Thus (i) the increase in [Glu], is almost insensitive to tetrodotoxin and external Ca2+ chelation (Fig. 4); and (ii) it is also observed in cultures of pure type-1 astrocytes apparently lacking the exocytotic machinery (Fig. 6A). Non-vesicular release in glia has been described either via reversed operation of glutamate uptake (Szatkowski et al., 1990). or via swelling-activated anion transport (Kimelberg et al., 1990). The experiments performed on glutamate transporters reconstituted in liposomes suggest that a carrier-mediated release mechanism is involved in the action of PDC. In particular, they show that PDC promotes carrier-mediated heteroexchange with glutamate (Table 1). Thus, liposomes without any transmembrane ion gradients do not take up any glutamate from the outside since the driving forces of the uptake process are missing. However, if the liposomes are loaded with a transportable amino acid, e.g. PDC, they become able to incorporate external ~-[~H]glutamate (see also Pines and Kanner, 1990; Danbolt, 1994). Since the external L-['Hlglutamate is taken up and concentrated in the liposomes in the absence of ion gradients to fuel the uptake, this must imply that PDC is exchanged with external glutamate. On the other hand, external PDC induces the release of internal ~-['H]glutamate from preloaded liposomes. A similar effect was described by Griffiths et al. (1994), who found that cultured cerebellar granule cells preloaded with ~[~Hlaspar ta te release the label in response to PDC. These authors suggested that the process occurs via an exchange mechanism. Based on the above evidence, it is very likely that PDC initiates glutamate release via exchange through transporters also in our cell cultures. However, it cannot be ruled out that the efflux process in living cells is fuelled by more than one mechanism. For example, activation of the EAA receptors by released glutamate might start a series of ionic, voltage and metabolic changes that could play a role in determining the global efflux (see below).

Blockade of glutamate release by LP

Glutamate uptake depends on external Na' cotransport: removal of extracellular Na' (Na',) abolishes the inward currents generated by glutamate or PDC in glial cells (Sarantis et al., 1993). However, the efflux process triggered by PDC does not appear to be fully dependent on Na',. Thus, in 0 mM "a+],, the release of L-

[3H]glutamate from liposomes is reduced but not abolished. Similarly, Griffiths et al. (1994) report that the Na+,-free condition attenuates without abolishing PDC-induced ~-[~H]aspartate efflux from cultured granule cells. In addition, they find that removal of Na', per se increases basal efflux (see also Taylor et al., 1992). Interestingly, these authors used choline as a substitute for Na',. Indeed, we here confirm that choline increases basal [Glu], and does not fully prevent a further rise with PDC. However, if Li+ instead of choline is used to replace Nafo, the basal [Glu], is unaffected and the PDC-induced efflux is abolished (Fig. 5A). These data imply that Li+ acts not simply as an external Na+ substitute, but via some additional mechanism not shared with choline. Such a mechanism is not clarified at present and will require further, detailed examination. One possibility is that Li' enters the cells to act on transport internally: indeed, the small Li+ ion permeates several channels inaccessible to the much bigger choline. This hypothesis is supported by the finding of Kanner and Bendahan (1982) that exchange-release of ~-[~H]glutamate from synaptic plasma membrane vesicles requires Na'i but not Na',. In addition, Li' could induce some biochemical changes by interference with phosphoinositide metabolism or other transduction systems (Jope and Williams, 1994) and thereby affect transporter function [e.g. glutamate transporters are regulated by protein kinase C-dependent phosphorylation (Casado et al., 1993)l.

Excitotoxic consequences of PDC-induced glutamate release

The observation that both Li+ and GDH provide full protection against the neuronal damage induced by PDC identifies the efflux of endogenous glutamate as the key event responsible for delayed neurotoxicity. The experiments where we compare the [Glu], increase in normal and in neuron-depleted co-cultures indicate that most of the released glutamate originates from neurons and only a minor part from glia. Although astrocytes possess higher transport capacity (Rosenberg et al., 1992). the greater efflux from neurons could be explained by the fact that neuronal and glial transporters operate in different conditions with respect to equilibrium; thus, the glutamate concentration inside the neurons is higher than in glia (Ottersen et al., 1992).

The progressive increase in the ambient [Glu], induced by PDC generates sustained neuronal depolarization (Fig. 2A), probably via tonic activation of the NMDA receptors (Isaacson and Nicoll, 1993; Sarantis et al., 1993). Application of exogenous glutamate to cortical cultures was found to produce similar electrical changes associated with two NMDA-dependent phenomena: early Ca" entry and delayed neurotoxicity (Hablitz and Burgard, 1993; Hartley et al., 1993). It is very likely that PDC activates the same phenomena, because its neurotoxic action is completely antagonized both by APV, the prototype NMDA blocker, and by EGTA, an external Ca2' chelator (Figs 3B and 4). However, in the case of PDC, toxicity is also blocked by Li+, which prevents glutamate efflux and the initiation of the neurotoxic cascade. AMPAkainate and metabotropic blockers also provide partial protection (Fig. 3B). These receptor types are probably activated by the elevation of extracellular glutamate and participate in the neurotoxic cascade via amplification of local phenomena, e.g. by stimulating the efflux process (Schwartz, 1987; Thomsen et al., 1994) and/or by enhancing NMDA receptor function (Schoepp and Conn, 1993). In addition, they could potentiate the excitatory transmission in the culture.

Non-vesicular glutamate release through transporters is thought to play a major role in ischaemic neuronal degeneration (reviewed in Szatkowski and Attwell, 1994). Loss of the ionic membrane gradients due to metabolic impairment would lead the transport process to run in reversed mode, releasing neurotoxic glutamate concentrations in the extracellular medium (Attwell et al., 1993; Mad1 and Burgesser, 1993). The present study suggests that neurotoxic release of glutamate could also be activated via a different mechanism of transport alteration, in the absence of ionic or metabolic derange- ment. Most likely, in this case, the glutamate transporters take up from the extracellular space a molecule with low or no activity on EAA receptors (PDC), and in exchange release a potent activator (glutamate). It seems reasonable to predict that other exogenous or endogenous compounds that are transportable substrates of the glutamate carriers would trigger the same events.

Acknowledgements

This work was supported by Telethon-Italy (grants 586 and 754 to A. V.), and EC Biomed 2 Contract BMH4-CT95-571 (to A. V. and N. C. D.). We also acknowledge the contribution of the European Science Foundation for Research Fellowship 1995 in Toxicology (to D. T.) and of Norwegian Research Council for Student Research fellowships (to K. U.). We thank Stefano Floridi for his contribution in the early phases of the project, Sabino Vesce for preparing cell cultures and performing HPLC measurements, Dr Daniela Curti for providing access to the luminometer, and Drs Steven A. Siegelbaum, Matthias A. Hediger and Stefano Vicini for helpful comments and discussions.

Transport alteration induces neurotoxic glutamate release 2027

sodium and potassium ion coupled L-glutamic acid transporters from rat brain. Biochemistry, 21, 6327-6330.

Kimelberg, H. K., Goderie, S. K., Higman, S., Pang, S. and Waniewski, R. A. (1990) Swelling-induced release of glutamate, aspartate and taurine from astrocyte cultures. J. Neurosci., 10, 1583-1591.

Koh, J. Y. and Choi, D. W. (1987) Quantitative determination of glutamate mediated cortical neuronal injury in cell culture by lactate dehydrogenase efflux assay. J. Neurosci. Merhods, 20, 83-90.

Lehre, K. P., Levy, L. M., Ottersen, 0. P., Storm-Mathisen, J. and Danbolt, N. C. (1995) Differential expression of two glial glutamate transporters in the rat brain: quantitative and immunocytochemical observations. J . Neurosci., 15, 1835-1853.

Lipton, S. A. and Rosenberg, P. A. (1994) Excitatory amino acids as a final common pathway for neurologic disorders. N. Engl. J. Med., 330, 613-6.22.

Lowry, 0. H., Rosebrough, N. J., Farr, A. L. and Randall, R. J. (195 1) Protein measurements with the Folin phenol reagent. J . B i d . Chem., 193, 265-275.

Madl, J. E. and Burgesser, K. (1993) Adenosine triphosphate depletion reverses sodium-dependent, neuronal uptake of glutamate in rat hippocampal slices. J . Neurosci., 13, 44294444.

Maki, R., Robinson, M. B. and Dichter, M. A. (1994) The glutamate uptake inhibitor ~-trans-pyrrolidine-2,4-dicarboxylate depresses excitatory synaptic transmission via a presynaptic mechanism in cultured hippocampal neurons. J. Neurosci., 14, 6754-6762.

McCarthy, K. D. and de Vellis, J. (1980) Preparation of separate astroglial and oligodendroglial cell cultures from rat cerebral tissue. J. Cell Biol., 85, 890-892.

Mennerick, S. and Zorumski, C. F. (1994) Glial contributions to excitatory neurotransmission in cultured hippocampal cells. Nature, 368, 59-62.

Ottersen, 0. P., Zhang, N. and Walberg, F. (1992) Metabolic compartmentation of glutamate and glutamine: morphological evidence obtained by quantitative immunocytochemistry in rat cerebellum. Neuroscience. 46, 5 19-534.

Pines, G. and Kanner, B. I. (1990) Counterflow of L-glutamate in plasma membrane vesicles and reconstituted preparations from rat brain. Biochernisrry, 29, 11209-11214.

Pines, G., Danbolt, N. C., Bi0ris. M., Zhang, Y., Bendahan, A,, Eide, L., Koepsell, H., Seeberg, E., Storm-Mathisen, J. and Kanner, B. I. (1992) Cloning and expression of a rat brain L-glutamate transporter. Nature, 360, 464467.

Robinson, M. B., Djali, S. and Buchhalter, J. R. (1993) Inhibition of glutamate uptake with ~-trans-pyrrolidine-2,4-dicarboxylate potentiates glutamate toxicity in primary hippocampal cultures. J. Neurochern., 61, 2099-2103.

Rosenberg, P. A., Amin, S. and Leitner, M. (1992) Glutamate uptake disguises neurotoxic potency of glutamate agonists in cerebral cortex in dissociated cell culture. J. Neurosci., 12, 56-61.

Rothstein, J. D., Martin, L. J. and Kunkl, R. W. (1992) Decreased glutamate transport by the brain and spinal cord in amyotrophic lateral sclerosis. N. Engl. J . Med., 326, 1464-1468.

Rothstein, J. D., Martin, L., Levey, A. I., Dykes-Hoberg, M., Jin, L., Wu, D., Nash, N. and Kunkl, R. W. (1994) Localization of neuronal and glial glutamate transporters. Neuron, 13, 7 13-725.

Rothstein, J. D., Jin, L., Van Kammen, M., Levey, A. I., Martin, L. and Kunkl, R. W. (1995) Selective loss of glial glutamate transporter GLT-I in amyotrophic lateral sclerosis. Ann. Neurol., 38, 73-84.

Sarantis, M., Ballerini, L., Miller, B., Angus Silver, R., Edwards, M. and Attwell, D. (1993) Glutamate uptake from the synaptic cleft does not shape the decay of the non-NMDA component of the synaptic current. Neuron, 11, 541-549.

Schoepp, D. D. and Conn, J. (1993) Metabotropic glutamate receptors in brain function and pathology. Trends Pharmacol. Sci., 14, 13-20.

Schwartz, E. (1987) Depolarization without calcium can release gamma- aminobutyric acid from a retinal neuron. Science, 238, 350-355.

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. Narl Acad. Sci. LISA, 89, 10955-10959.

Szatkowski, M. and Attwell, D. (1994) Triggering and execution of neuronal death in brain ischemia: two phases of glutamate release by different mechanisms. Trends Neurosci., 17, 359-365.

Szatkowski, M., Barbour, B. and Attwell, D. (1990) Non-vesicular release of glutamate from glial cells by reversed electrogenic uptake. Nature, 348, 443446.

Taylor, C. P., Geer, J. J. and Burke, S. P. (1992) Endogenous extracellular glutamate accumulation in rat neocortical cultures by reversal of the transmembrane sodium gradient. Neurosci. Lett., 145, 197-200.

Thomsen, A., Hansen, L. and Suzdak, P. D. (1994) t glut am ate uptake

Abbreviations

AMPA APV CHAPS

CNQX EAA FCS GDH HCSS HPLC LDH MEM NMDA PDC

a-amino-3-hydroxy-5-methyl-4-isoxazole 2-amino-5-phosphonopentanoate 3-[( 3-cholamidopropyl)dimethylammonio]- I-propane- sulphonate 6-cyan0-7-nitroquinoxaline-2.3-dione excitatory amino acid(s) fetal calf serum glutamate dehydrogenase HEPES-buffered control salt solution high-performance liquid chromatography lactate dehydrogenase Eagle's Minimal Essential Medium N-methyl-o-aspartate ~-trans-pyrrolidine-2,4-dicarboxylate

References

Attwell, D., Barbour, B. and Szatkowski, M. (1993) Nonvesicular release of neurotransmitter. Neuron, 11, 401407.

Barbour, B., Keller, B. U., Llano, I . and Marty, A. (1994) Prolonged presence of glutamate during excitatory synaptic transmission to cerebellar Purkinje cells. Neuron, 12, 1331-1343.

Bridges, R. J., Stanley, M. S., Anderson, M. W., Cotman, C. W. and Chamberlin, A. R. ( 1991) Conformationally defined neurotransmitter analogues. Selective inhibition of glutamate uptake by one pyrrolidine-2,4-dicarboxylate diastereomer. J. Med. Chem., 34, 717-725.

Casado, M., Bendahan, A,, Zafra, F., Danbolt, N. C., Aragon, C., Gimenez, C. and Kanner, B. I. (1993) Phosphorylation and modulation of brain glutamate transporters by protein kinase C. J. Biol. Chem., 268, 27313- 27317.

Chaudhry, F. A,, Lehre. K. P., van Lookeren Campagne, M., Ottersen, 0. P., Danbolt, N. C. and Storm-Mathisen, J. (1995) Glutamate transporters in glial plasma membranes: highly differentiated localizations revealed by quantitative ultrastructural immunocytochemistry. Neuron, 15, 71 1-720.

Choi, D. W., Maulucci-Gedde, M. and Krigstein, A. R. (1987) Glutamate neurotoxicity in cortical cell cultures. J . Neurosci., 7, 357-368.

Danbolt, N. C. (1994) The high affinity uptake system for excitatory amino acids in the brain. Prog. Neurobiol., 44, 377-396.

Danbolt, N. C., Pines, G. and Kanner, B. I. (1990) Purification and reconstitution of the sodium- and potassium-coupled glutamate transport glycoprotein from rat brain. Biochemistry, 29, 6734-6740.

Donevan, S. D. and Rogawski, M. A. (1993) GYKl 52466, a 2.3- benzodiazepine, is a highly selective, noncompetitive antagonist of AMPA/ kainate receptor responses. Neuron, 10, 5 1-59,

Fairman, W. A,, Vanderberg, R. J., Amza, J. L., Kavanaugh, M. P. and Amara, S. G. (1995) An excitatory amino-acid transporter with properties of a ligand-gated chloride channel. Nature, 375, 599-603.

Fosse, V. M., Kolstad, J. and Fonnum, F. (1986) A bioluminescence method for the measurement of L-glutamate: applications to the study of changes in the release of L-glutamate from lateral geniculate nucleus and superior colliculus after visual cortex ablation in rats. J . Neurochem., 47, 340-349.

Griffiths, R., Dunlop, J., Gorman, A., Senior, J. and Grieve, A. (1994) L-trans- pyrrolidine-2,4-dicarboxylate and cis- 1 -aminocyclobutane- 1.3-dicarboxy- late behave as transportable, competitive inhibitors of the high-affinity glutamate transporters. Biochem. Pharrnacol., 47, 267-274.

Hablitz, J. J. and Burgard, E. C. (1993) Glutamate-induced changes in membrane properties and calcium accumulation in neocortical neurons. SOC. Neurosci. Absa , 19, 26.

Hartley, D. M., Kurth, M. C., Bjerkness, L., Weiss. J. H. and Choi, D. W. ( 1993) Glutamate receptor-induced 45Ca2f accumulation in cortical cell culture correlates with subsequent neuronal degeneration. J . Neurosci., 13,

Isaacson, J. S. and Nicoll, R. A. (1993) The uptake inhibitor L-trans-PDC enhances responses to glutamate but fails to alter the kinetics of excitatory synaptic currents in the hippocampus. J. Neumphysiol., 70, 21 87-2191.

Jope, R. S. and Williams, M. B. (1994) Lithium and brain signal transduction systems. Biochem. Pharmacol., 47, 42941.

Kanai, Y. and Hediger, M. A. (1992) Primary structure and functional characterization of a high-affinity glutamate transporter. Nature, 360, 46747 1.

Kanner, B. I. and Bendahan, A. (1982) Binding order of substrates to the

1993-2000.

2028 Transport alteration induces neurotoxic glutamate release

inhibitors may stimulate phosphoinositide turnover hydrolysis in baby hamster kidney cells expressing mGluRla via heteroexchange with L- glutamate without direct activation of mGluRla. J . Neurochern., 63,

Tong, G. and Jahr, C. E. (1994) Block of glutamate transporters potentiates postsynaptic excitation. Neuron, 13, 1 195-1 203.

Trotti, D., Volterra, A., Lehre, K. P., Rossi, D., Gjesdal, 0.. Racagni, G. and Danbolt, N. C. (1995) Arachidonic acid inhibits a purified and reconstituted glutamate transporter directly from the water phase and not via the phospholipid membrane. J. Biol. Chem., 270, 9890-9895.

Trotti, D., Rossi, D., Gjesdal, 0.. Levy, L. M., Racagni, G., Danbolt, N. C. and Volterra, A. (1996) Peroxynitrite inhibits glutamate transporter subtypes. J. Biol. Chem., 271, 59765979.

2038-2047.

Volterra, A., Trotti, D., Cassutti, P., Tromba, C., Salvaggio, A,, Melcangi, R. C. and Racagni, G. (1992) High sensitivity of glutamate uptake to extracellular free arachidonic acid levels in rat cortical synaptosomes and astrocytes. J . Neurochem., 59,600-606.

Volterra, A., Trotti, D., Trornba, C., Floridi, S. and Racagni, G. (1994) Glutamate uptake inhibition by oxygen free radicals in rat cortical astrocytes. J. Neurosci., 14, 2924-2932.

Watkins, J . and Collingridge, G. (1994) Phenylglycine derivatives as antagonists of metabotropic glutamate receptors. Trends Pharmacol. Sci., 15, 333-342.

Zerangue, N., Aniza, J. L., Amara, S. G. and Kavanaugh, M. P. (1995) Differential modulation of human glutamate transporter subtypes by arachidonic acid. J . Biol. Chem., 270, 6433-6435.