Dopamine (DA) is released in large amounts during cerebral ischemia and may exacerbate tissue damage. Riluzole (54274 RP) is a recently developed agent that depresses glutamate neurotransmission in the central nervous system (CNS) and that may protect against ischemic injury in some animal models. Because glutamate stimulates the release of DA in the striatum, the authors hypothesized that riluzole could antagonize DA release in this structure.
Assay for DA release consisted of superfusing 3H-DA preloaded synaptosomes with artificial cerebrospinal fluid (1 ml/min, 37 degrees C) and measuring the radioactivity obtained from 1-min fractions over 22 min, first in the absence of any treatment (spontaneous release, 8 min), then in the presence of depolarizing agents combined with riluzole (0.1-100 microM, 5 min), and finally with no pharmacologic stimulation (9 min). The following depolarizing agents were tested: KCl (9, 15 mM), veratridine (0.01-1 microM), N-methyl-D-aspartate (NMDA, 0.1-1 mM), kainate (0.1-1 mM), and nicotine (0.01-0.5 mM). Assay for DA uptake was performed by measuring the radioactivity incorporated in synaptosomes incubated with 3H-DA (44 nM; 5 min; 37 degrees C).
All depolarizing agents produced a significant, concentration-related increase from basal 3H-DA release. Riluzole was found to decrease the release induced by veratridine (1 microM), NMDA (1 mM), and kainate (1 mM) in a significant, concentration-related manner (IC50 = 9.5 microM, 1.6 microM, and 5.8 microM for veratridine, NMDA, and kainate, respectively). In contrast, it did not affect the release elicited by either KCl or nicotine. Riluzole had no significant effect on the specific 3H-DA uptake.
Riluzole produced a potent blockade of the release of DA mediated by activation of presynaptic sodium channels, NMDA, and kainate receptors. Depression of glutamate transmission together with blockade of DA release may contribute to the actions of this agent in vivo.
Cerebral ischemia causes axon terminals to release excitotoxic glutamate and open N-methyl-D-aspartate (NMDA) channels, which allow Ca2+ entry into the neurons and trigger a cascade of intracellular events leading to neuronal damage. [1–3] Riluzole (2-amino-6-trifluoro-methoxy benzothiazole, 54274 RP, PK 26124)4 is a recently developed agent that exhibits anesthetic properties and that may protect against ischemic neuronal death in some animal models. [5–7] It has been shown to depress glutamate release at mammalian central nervous system (CNS) synapses.  Although its mechanism of action has not been clearly established, depressant actions on the conduction of presynaptic glutamate fibers,  sodium channels,  or NMDA-induced Ca2+ currents [6,11] have been reported.
In the rat striatum, a functionally important supraspinal brain area, glutamate terminals originating from the cerebral cortex make synaptic contact with DA nerve endings arising from the nigrostriatal pathway.  Mimicking activation of the corticostriatal glutamate fibers by local application of glutamate agonists in this brain area results in a potent stimulation of dopamine (DA) release mediated by NMDA and AMPA receptors located on the presynaptic DA terminals. [13–16] Alternatively, DA may be released by activation of voltage-operated Ca2+(VOCC) or sodium channels (VOSC) or nicotinic receptors present on the DA terminals (Figure 1).  Because DA is released in large amounts into the striatum during cerebral ischemia and may exacerbate tissue damage, [18–19] we hypothesized that riluzole may influence the release or the high affinity uptake of this transmitter.
Materials and Methods
Handling procedures, as written in the Guide for the Care and Use of Laboratory Animals* were followed throughout. Experiments were performed on male Sprague-Dawley rats (Iffa-Credo, L'Arbresle, France) weighing 200–225 g and housed on a 12:12 light/dark cycle with food and water ad libitum. Approval was obtained from the Institutional Animal Care and Use Committee at the University of Paris 7.
Preparation of Striatal Synaptosomes
Animals were killed by stunning and decapitation. After longitudinal hemisection of the brain, striata were quickly removed and processed to a purified synaptosomal preparation as previously reported in detail. [17,20] Synaptosomes were diluted up to 0.16 mg/ml in an ice-cold artificial cerebrospinal fluid (CSF) containing (in mM): NaCl, 126.5; NaHCO3, 27.5; KCl, 1; KH2PO4, 0.5; CaCl2, 1.1; MgCl2, 0.83; Na2SO4, 0.5; glucose, 5.9; pargyline, 10–2; ascorbic acid, 1, adjusted at pH 7.3 with 95%/5%[v/v] O2-CO sub 2 mixture. In Ca2+(Mg2+, respectively)-free experiments, CaCl2(MgCl2, respectively) was omitted from the CSF.
Dopamine Release Assay
Uptake of DA was performed by incubating (10 min; 37 [degree sign] Celsius) the synaptosomes with [2,5,6]3H-DA (Amersham, Little Chalfont, UK; > 44 nmol/l, 24 Ci/mmol, 1 mCi/ml). Aliquots (1 ml) were pipetted into each of the superfusion chambers (0.3 ml volume each) and embedded in Whatman GF/F glass filters (0.70-micro meter retention capacity) by light suction, then superfused at a 1 ml/min flow rate using a superfusion device equipped with an automatic fraction collector (Brandel, Gaithersburg, MD). The 1 ml/min flow rate was selected because of its ability to prevent reuptake of3H-DA by the synaptosome preparation. [18,20] After a 30-min washing step (37 [degree sign] Celsius), serial fractions were collected every minute during three consecutive periods: the mean value of the radioactivity (counts/min measured over the first 8 min by liquid scintillation spectrometry using Aquasol-2 (New England Nuclear, Boston, MA) was considered the basal (spontaneous)3H-DA release. During the next 5 min of superfusion, either no treatment (time-dependent control) or various pharmacologic agents were delivered to the preparation, and radioactivity again was estimated in each fraction. Because of a 3-ml dead volume occupied by the inside of the connectubes, pharmacologic agents delivered at a 1-ml/min flow rate reached synaptosomes only after a delay, which averaged 180 s. For the final 9 min, radioactivity was determined from 1-min fractions in the absence of any treatment, although in the experiments on the nicotine-induced release, fractions were collected every 30 s because of the rapid desensitization of the nicotinic response.  The influence of a drug on the release of3H-DA was assessed by calculating the difference between the release observed in the presence and absence (time-dependent control) of the pharmacologic agent used. This was expressed as a fractional increase from spontaneous release.
To allow comparison with previous studies, the concentrations of the pharmacologic agents tested were: KCl (9, 15 mM), NMDA (50 micro Meter, 100 micro Meter, 500 micro Meter, and 1 mM in a Mg2+-free superfusion medium, Sigma, L'Isle d'Abeau, France), kainate (50 micro Meter, 100 micro Meter, 500 micro Meter, and 1 mM, Sigma), the noncompetitive NMDA receptor antagonist MK801 (100 micro Meter, Merck-Sharp and Dohme, Darmstadt, Germany), the AMPA/kainate receptor antagonist 6-cyano-7-nitro-quinoxaline-2,3-dione (CNQX, 100 micro Meter, Research Biochemical Incorporated, Illkirch, France), nicotine (10, 50, 100, and 500 micro Meter, Sigma), the nicotinic receptor antagonist mecamylamine (10 micro Meter, Sigma), veratridine (0.01, 0.1, 0.5, and 1 micro Meter, Sigma), and the blocker of the VOSC tetrodotoxin (TTX, 0.1 micro Meter Sigma). In these experiments, mecamylamine, MK801, CNQX, and TTX were coapplied together with nicotine (0.5 mM), NMDA (1 mM), kainate (1 mM), and veratridine (1 micro Meter), respectively. In addition, to further assess the role played by the VOSC, the influence of TTX (0.1 micro Meter) on the release elicited by nicotine, NMDA, and kainate was also examined. Riluzole was dissolved in dimethylsulphoxide (Merck-Sharp and Dohme) and tested over a 10 sup -7 - 10 sup -4 M concentration range.
Dopamine Uptake Assay
Dopamine uptake was performed as previously described. [22,23] Briefly,3H-DA was added to the synaptosomal suspension at 37 [degree sign] Celsius in either the absence of any pharmacologic agent (control) or in the presence of various concentrations of nomifensine (10 sup -8 - 5 x 10 sup -5 M, Research Biochemical Incorporated) or riluzole (10 sup -7 - 10 sup -4 M). After 5-min incubation, the mixture was vacuum-filtered through Whatman GF/F filters and washed with 30 ml of ice-cold CSF. This duration of incubation was determined according to preliminary experiments, indicating that3H-DA increased linearly and reproducibly with time between 3 and 7 min. The radioactivity retained on the filters was extracted with NaOH 0.1 M and measured by liquid scintillation. Specific3H-DA uptake was considered the difference between the radioactivity measured in the presence and absence of nomifensine 10 sup -5 M.
Results (mean +/- SEM) were considered reliable only if they had been reproduced in four independent experiments (each of them performed in triplicate). Concentration-response curves, IC50 values, and Hill coefficients were generated directly from computer using the GraphPAD software (Intuitive Software for Science, San Diego, CA). The function used to fit the curves to the data was the following four-parameter logistic equation:Equation 1where X represents the logarithm of riluzole concentration, A and B, respectively, the minimum (bottom) and maximum (top) of X values. C is the logarithm of the IC50value, and D is the Hill coefficient or slope factor that is positive for curves in which the Y value increases with increasing X and is negative for curves in which the Y value decreases with increasing X. Statistics were performed by analysis of variance (ANOVA) followed by the Student's t test corrected for the number of comparisons. A P value less than 0.05 was considered the threshold for significance.
Pharmacologic Characterization of the Depolarization-evoked sup 3 H-dopamine Release from Striatal Synaptosomes
The3H-DA concentration estimated in the incubation medium during the uptake phase was 250 +/- 15 pmol/mg protein, and the initial basal release rate measured over 5 min was approximately 1.04 + 06 pmol [center dot] mg protein sup -1 [center dot] min sup -1. Basal release rate was not affected by removing Ca2+ or Mg2+ from the superfusion medium. A typical example of the time-course of the3H-DA release observed with the different paradigms used is displayed in Figure 2.
As displayed in Figure 3, all depolarizing agents elicited a significant, concentration-related increase from spontaneous3H-DA release. The3H-DA peak elicited by KCl (9 and 15 mM) was (in fractional units): 1.32 +/- 0.06 (P < 0.01) and 1.61 +/- 0.08 (P <0.01), respectively. It was followed by a rapid return to baseline after cessation of application of the pharmacologic agent. In contrast, the recovery of basal release rate was slower after veratridine, NMDA, kainate, and nicotine application. The effect of KCl was markedly reduced when Ca2+ was removed from the external medium (-87 +/- 8%). The effects of NMDA, kainate, nicotine, and veratridine, respectively, on3H-DA release were markedly reduced by MK 801 (-96 +/- 6%), CNQX (-89 +/- 9%), mecamylamine (-85 +/- 11%), and TTX (-97 +/- 5%), respectively. Application of TTX did not significantly affect the NMDA-, or nicotine-stimulated release, although it slightly, but significantly, attenuated the kainate-induced DA release (Figure 4).
Effects of Riluzole on sup 3 H-DA Release
Riluzole caused a significant, concentration-related, decrease in the release of3H-DA evoked by NMDA, kainate, and veratridine (Figure 5). The best fit of the curves to the data was always obtained when Hill coefficient was forced to 1. The IC50, values for the inhibitory effect of riluzole were 9.5 micro Meter, 1.6 micro Meter, and 5.8 micro Meter, respectively, for the veratridine-, NMDA-, and kainate-induced, respectively,3H-DA release. In contrast, riluzole failed to significantly alter the release elicited by either KCl (9 and 15 mM) or nicotine (0.5 mM). The attenuation of the NMDA (kainate, respectively)-evoked peaks of radioactivity by riluzole yielded a ceiling effect as 36%(40%, respectively) of the release was unaffected by riluzole concentrations greater than 10–5 M. The vehicle (DMSO) per se did not produce any change in the veratridine-, NMDA-, and kainate-induced peaks.
Effects of Riluzole on sup 3 H-DA Uptake
Under the present experimental conditions,3H-DA uptake averaged 1.98 +/- 0.05 pmol/mg protein after 5 min of incubation. Nomifensine, a classical inhibitor of the DA carrier, produced a concentration-related inhibition of3H-DA uptake, the IC50value being 0.24 micro Meter. In contrast, riluzole failed to affect3H-DA uptake at any of the concentrations tested (Figure 6).
The present study indicates that riluzole is able to block the depolarization-evoked release of DA from striatal nerve terminals, although the ability of riluzole to reduce DA release depends on the nature of the depolarizing stimulus; it markedly decreases the veratridine-elicited release mediated by activation of presynaptic VOSC. It also exhibits a significant inhibitory effect on presynaptic events mediated by ionotropic glutamate receptors because it potently reduces the NMDA- and kainate-evoked DA release. In contrast, riluzole fails to affect the release elicited by activation of the presynaptic nicotinic receptors or involving the VOCC (KCl-evoked release).
Synaptosomes provide a reliable and reproducible model for analyzing, in vitro, the direct presynaptic effects of pharmacologic agents on neurotransmitter release, uptake, or metabolism in various brain areas. [13–17,21–24] However, interpretation of the data should consider that physiologic depolarization, which occurs over milliseconds, may have exhibited different sensitivity to riluzole application compared with pharmacologic depolarization applied during several minutes to the synaptosomal preparation. Also, it cannot be excluded that DA nerve terminals, in situ, subjected to impulse activity and a transmitter-enriched environment may behave differently.
The present results show that riluzole blocks the release of DA elicited by activation of the VOSC. DA release was stimulated by veratridine, an activator of sodium channels, and blocked by TTX application. These data are consistent with previous results indicating that riluzole (10 micro Meter) reduced the increase in intracellular Ca sup 2+ concentration evoked by veratridine (10 micro Meter) on cultured granule cells.  Further, riluzole (1–100 micro Meter) has been shown to reduce the magnitude of the peak sodium currents carried by the cloned rat brain IIA sodium channel alpha-subunits expressed in Xenopus oocytes.  Looking at the mechanisms of riluzole action on sodium channels, several lines of evidence suggest that riluzole may increase inactivation at all potentials by high affinity block of inactivated channels.  Whether this may occur at DA presynaptic nerve endings cannot be inferred from our data. However, this action reported on DA fibers may also apply to other subpopulations of nerve endings, such as the glutamatergic terminals, for which riluzole causes potent depression of the nerve conduction that resembles that produced by some local anesthetics. 
Riluzole did not affect the release of DA stimulated by either KCl or nicotine application to the synaptosomes. The KCl-evoked peak was not observed in the absence of external Ca2+, whereas the nicotine-induced effect was abolished in the presence of mecamylamine, a nicotinic antagonist exhibiting particular efficiency in blocking nicotinic receptors of the striatal DA terminals.  The absence of effect of riluzole on the synaptosomal DA release coupled to activation of presynaptic VOCC or nicotinic receptor-coupled channels suggests that these specific channels are not sensitive to riluzole. although we cannot rule out that riluzole could act at any of a number of sites to depress DA release in the striatum because the process of release is known to involve several proteins associated with vesicle and nerve terminal membranes. Alternatively, it could be expected that the nicotine-induced DA release may have been affected by riluzole because the nicotinic channel is permeant to Na sup + ions. It cannot be excluded that the rather high nicotine concentrations used here may have overwhelmed a potential blocking effect of riluzole on the nicotinic receptor-coupled channel.
Riluzole also reduced markedly DA release evoked by excitatory amino acids (NMDA and kainate). The peak elicited by NMDA application was blocked by application of the NMDA receptor antagonist MK 801, but not TTX, whereas that produced by kainate was blocked by the AMPA receptor antagonist CNQX and slightly, but significantly, attenuated by TTX. This indicates that this amino acid-stimulated release was a result of activation of the NMDA and AMPA receptors present on the presynaptic DA terminals and that VOSCs are not significantly involved in NMDA-evoked DA release, which is consistent with previous findings.  Thus, riluzole also depresses amino acid-mediated transmission by blocking responses involving glutamate ionotropic receptors located on presynaptic DA terminals. Riluzole has been shown to antagonize NMDA receptor-mediated currents in Xenopus oocytes injected with mRNA from rat whole brain cortex, although the effect is less potent (IC50= 18.2 micro Meter).  It had only a weak blocking effect (IC50= 167 micro Meter) on the kainate-mediated responses in this model. The low IC50values reported in our study suggest that the responses mediated by presynaptic NMDA and kainate receptors located on DA nerve endings may be particularly sensitive to riluzole.
It is not clear how riluzole blocks electrophysiologic responses mediated by excitatory amino acids because no interaction between this agent and any of the recognition sites on the NMDA or AMPA receptors has been demonstrated.  Also, riluzole did not interact with either the glycine site or the phencyclidine site of the NMDA ionotropic receptor and did not block metabotropic receptors coupled to phospholipase C. [4,27,28] Recently, Hubert et al. have shown that riluzole antagonized the NMDA-induced increase in intracellularCa2+ via a pertussis toxin sensitive mechanism.  This supports that riluzole can act at G proteins located in the immediate environment of the NMDA receptor itself. The NMDA- and kainate-stimulated DA release could be antagonized only in part by riluzole because a ceiling effect was observed at concentrations greater than 10 micro Meter. Altogether, these data suggest that riluzole blocks presynaptic ionotropic receptors via either direct, noncompetitive, or, more likely, by indirect mechanisms.
The presence of excessive concentrations of glutamate in the synaptic cleft has been involved in the so-called excitotoxic process in the CNS. This is favored by certain conditions including a failure of glutamate uptake, prolonged and excessive depolarization of the nerve terminal, or introduction of glutamic acid from a nonregulated source. [1–3] It has been shown that riluzole decreases glutamate transmission by acting at presynaptic target sites, including conduction velocity and neurotransmitter release. [4,8,9] Our results indicate that, in addition to these previously reported actions, riluzole blocks the responses mediated by the presynaptic VOSC, NMDA, and kainate receptors located on the striatal DA terminals. The fact that riluzole blocks excitatory amino acid-mediated neurotransmission by acting at pre- and postsynaptic target sites and, on the other hand, antagonizes the activity of the VOSC is consistent with one possible mechanism of antiischemic activity.
In conclusion, we provide evidence that riluzole exerts a selective depression of the release of DA elicited by excitatory amino acids and veratridine in the rat striatum in vitro. Interestingly, several lines of evidence indicate that the striatal DA is released in large quantities into the extracellular space during cerebral ischemia and may exacerbate tissue damage. [18,19,29,30] Depression of glutamate transmission together with blockade of DA release may provide a pharmacologic basis to better understand some neuroprotective effects of riluzole in vivo.
*Guide for the Care and Use of Laboratory Animals, Health and Human Services, NIH Publication No 86–23, Revised 1996, pp 58–9.