Riluzole (RP 54274) is an experimental benzothiazole with anesthetic properties, but little is known about its synaptic or cellular actions.
The authors investigated riluzole effects on synaptic response of CA 1 pyramidal neurons in rat hippocampal brain slices. Electrophysiologic recordings of population spikes (PS), excitatory postsynaptic potentials (EPSP), and fiber volleys were studied. Paired pulse stimulation (120 ms interpulse interval) was used to measure effects on gamma-amino butyric acid (GABA)-mediated synaptic inhibition, and stimulus trains (33 Hz) were used to test for use-dependent effects.
Synaptically evoked PS discharge was blocked in a concentration-dependent manner by riluzole (2.0-20 microM), similar to effects produced by other anesthetics. Paired pulse inhibition was not altered by riluzole. In contrast, 20 microM thiopental produced a marked increase in paired pulse inhibition. Riluzole (5.0 microM) produced a 46.6 +/- 19.8% depression of glutamate-mediated EPSPs, which could account for most of the mate-mediated EPSPs, which could account for most of the depression of PS discharge (54.2 +/- 12.6%) produced by this concentration. Riluzole produced a 36 +/- 17% depression of fiver volley amplitudes, which, based on input/output analysis, could completely account for the depression of EPSPs. The depression of fiber volley amplitudes showed a marked use-dependence; the second and subsequent action potentials in a train were progressively depressed by riluzole to a greater extent than the first action potential.
Riluzole produced a potent block of excitatory synaptic transmission via depression of presynaptic conduction in glutamatergic nerve fibers. The use-dependent depression observed resembled that produced by some local anesthetics on nerve conduction and sodium channels. The presynaptic action, together with a lack of effect on gamma-amino butyric acid-mediated inhibition, provides a unique mechanism of action for a general anesthetic.
Key words: Animals: rats. Anesthetics: thiopental. Brain: hippocampus, CA 1; cortex; cortical. Neurophysiology: action potential; conduction; postsynaptic; nerve; sodium channel. Pharmacology: gamma-aminobutyric acid; depressant; glutamate; hypnotic; narcotic; riluzole. Synapse: excitation; inhibition; transmission.
DEPRESSION of glutamate-mediated excitatory synaptic transmission is thought to play an important role in the central nervous system depression produced by several anesthetics. Depressed glutamate responses at N-methyl-D-aspartate receptors appear to be a major mechanism of action for dissociative anesthetics like ketamine. [2,3]Similarly, volatile anesthetics and barbiturates have been shown to depress glutamate-mediated synapse, [4-6]and this could combine with enhanced gamma-amino butyric acid (GABA)-mediated inhibition to produce central nervous system depression. [7,8]Recent interest has focused on anesthetics that reduce excitatory synaptic transmission by depressing glutamate release from nerve terminals. Riluzole (2-amino-6-trifluoro-methoxy benzothiazole, RP 54274, PK 26124) has been shown to depress glutamate release at mammalian central nervous system excitatory synapses [10,11]and produces potent hypnotic, anesthetic, and minimum alveolar concentration-sparing effects in rats. Riluzole's mechanism of action is not known; however, effects on presynaptic sodium channels, N-methyl-D-aspartate-induced CA sup ++ currents, [14-16]GABA uptake, and glutamate release have been suggested. In the current study, we investigated riluzole's effects on: (1) synaptically evoked population spike responses, (2) GABA-mediated paired pulse inhibition, (3) glutamate-mediated synaptic excitation, and (4) presynaptic action potential conduction (Figure 1), to determine which actions contribute to depressed synaptically evoked discharge of CA 1 pyramidal neurons in rat hippocampal slices.
Materials and Methods
Experiments were performed on brain slices isolated from male Sprague-Dawley rats (80-120 g). Experimental protocols were approved by the Institutional Animal Care Committee at Stanford University and adhered to published guidelines of the National Institutes of Health, Society for Neuroscience, and the American Physiological Society. Rats were anesthetized with diethyl ether, and their hearts were stopped with a blow to the back of the thorax. Brains were removed and placed in cold (1-2 degrees Celsius), oxygenated artificial cerebrospinal fluid (ACSF, see Materials section). Brains were sectioned in the coronal plane into 450-micro meter thick slices using a vibratome (Vibraslice, Boston, MA). Slices were then hemisected and placed on filter papers at the interface of a humidified carbogen (95% Oxygen2-5% CO2) gas phase and ACSF liquid phase. Slices were allowed at least 1 h for recovery before submersion in ACSF in a recording chamber. The ACSF was saturated with carbogen gas and perfused at a rate of 2.5 ml/min at room temperature (21-24 degrees Celsius). Rapid and accurate solution changes were made using a ValveBank8 computerized perfusion system (AutoMate Scientific, Oakland, CA). Thiopental concentrations were measured using high-pressure liquid chromatography. 
Fiber volleys, population spikes (PS), and field excitatory postsynaptic potentials (EPSP) were evoked (0.5-15 V, 250 micro second, 0.1 Hz) with a bipolar tungsten stimulating electrode placed in stratum radiatum of area CA 1 (Figure 1). Stimulus pulses were delivered from constant-current isolation units (PSIU6) driven by a digitally timed controller (S 8800, Grass Medical Instruments, Quincy, MA). Fiber volleys were recorded with high resistance (1-2 M Omega) glass electrodes filled with ACSF and placed directly in line with stimulating electrodes in stratum radiatum, but separated from the stimulating electrode by 1.0-1.5 mm. Fiber volleys were isolated by blocking synaptic transmission using low calcium (< 0.2 mM) ACSF. Population spikes and EPSP responses were recorded with glass electrodes (0.2-1.0 M Omega) filled with ACSF and placed in stratum oriens or radiatum, respectively. Signals were amplified (5,000-20,000 times), filtered (DC to 10 KHz; Brownlee Precision, Santa Clara, CA) and digitized at 50 KHz using DataWave Technologies software (Longmont, CO) to control an analog to digital converter.
All responses were measured using DataWave software, analyzed and graphed using IGOR Pro (WaveMetrics, Lake Oswego, OR) software. Fiber volley responses were measured as the peak negative to peak positive amplitude. Excitatory postsynaptic potentials amplitudes were measured as the peak negativity from baseline or as the slope of the early response between 20% and 80% of maximum. Population spike responses were measured as the peak negative to peak positive amplitude. All responses were normalized as a percent of control based on the average amplitude during a 10-min recording immediately before riluzole application. Statistical significance was determined by comparing riluzole effects with time-matched control responses using analysis of variance (AXUM, TriMetrix, Seattle, WA).
Rats were obtained from Simonsen Laboratories (Gilroy, CA). Riluzole was provided by Rhone-Poulenc Rorer (Collegeville, PA). Thiopental was obtained from Sigma (St. Louis, MO). All solutions were made up in high-pressure liquid chromatography and spectrophotometric grade water (OmniSolv) supplied by EM Science (Gibbstown, NJ). The ACSF had the following ionic composition (in mM): Sodium sup +, 151.25; Potassium sup +, 3.5; Calcium sup ++, 2.0; Magnesium sup ++, 2.0; Chlorine sup -, 130.5; HCO3sup -, 26; SO4sup -, 2.0; H2PO4sup -, 1.25; and glucose, 10. Chemicals for the ACSF were reagent grade or better and obtained from J.T. Baker (Philadelphia, PA).
Riluzole Produced a Potent Depression of Population Spike Amplitudes
Evoked population spike amplitudes were depressed by low concentrations of riluzole (0.5-10 micro Meter), and a half maximal depression (54.2 plus/minus 12.6%; mean plus/minus SD, n = 9) occurred at 5.0 micro Meter (P < 0.001, analysis of variance compared with control). Initial effects were observed 5-8 min after application, and the half-time for onset was 13 plus/minus 1.7 min, determined from a first-order exponential fit. Steady-state effects were observed by approximately 25 min at a concentration of 5.0 micro Meter. Faster onset times were observed at 10 micro Meter (9 plus/minus 2.9 min), and complete PS depression occurred within 10 min at 20 micro Meter. Recovery from the riluzole-induced depression required more than 2 h. Riluzole-induced PS depression was accompanied by an increase in spike latency (approximately 2.0 ms) and decreases in initial and late positive components (Figure 2), indicative of effects on synaptic input and postsynaptic discharge.
Riluzole Did Not Alter gamma-Amino Butyric Acid-mediated Paired Pulse inhibition
Riluzole produced comparable depression of both the first and second PS responses to an identical pair of stimuli (Figure 3). The ratio of second/first PS response was 1.5 in control conditions and 1.9 in the presence of 5.0 micro Meter riluzole. Thiopental produced a greater depression of the second of a pair of PS responses (ratio = 0.65; Figure 3) at all effective concentrations (10-100 micro Meter). The PS depression produced by thiopental was not accompanied by an alteration in the early positivity. Riluzole, however, slowed the rate of rise and depressed peak amplitudes (Figure 3), suggesting that PS depression resulted from depressed excitatory synaptic input.
Riluzole Depressed Excitatory Postsynaptic Potentials Responses
To determine whether riluzole depressed excitatory synaptic inputs to CA 1 neurons, EPSP responses from dendritic fields in stratum radiatum were recorded. Riluzole depressed EPSP amplitudes (Figure 4), and the second EPSP of a paired pulse response was depressed more (52 plus/minus 9.4% of control; n = 10) compared with the first (46.6 plus/minus 19.8%) in the presence of 5 micro Meter riluzole (P < 0.001, for both effects compared with control, P < 0.01 for the first EPSP effect compared with the second. To determine whether this degree of EPSP depression could account for the PS reduction produced by riluzole, the EPSP versus PS relation was studied (Figure 5). In control conditions, the slope for the EPSP versus PS curve was 1.35, so the depression of EPSP amplitudes observed can fully account for reduced PS responses (47 * 1.35 [nearly equal] 63%, compared with the observed 55% depression).
Excitatory postsynaptic potential depression was accompanied by a depression of Schaffer-collateral fiber volleys (Figure 6). Fiber volleys provide a measure of action potential conduction in presynaptic axons. Fiber volley amplitude was depressed by 36 plus/minus 17% (n = 13, P < 0.001) in the presence of 5 micro Meter riluzole, and the second response to a paired pulse was reduced by 47 plus/minus 9%. The fiber volley versus EPSP relation had a slope of 1.2 (Figure 7), so depressed action potential conduction could fully account for riluzole's reduction of EPSP amplitudes (36 * 1.2 [nearly equal] 43%, close to the observed 47% depression). Similarly, the increased depression observed for the second fiber volley correlated well with the increased riluzole effect on the second EPSP (56% calculated vs. 53% depression observed).
Riluzole Depressed Fiber Volleys in a Use-dependent Manner
The increased sensitivity observed for the second fiber volley suggested that a use-dependent depression of axonal conduction could contribute to riluzole's effect. To determine whether riluzole depressed axonal conduction, effects on isolated fiber volleys were studied (Figure 8) after block of synaptic responses (see Methods section). Riluzole depressed isolated fiber volleys in a concentration-dependent manner (0.5-20 micro Meter), with an EC50of 7.5 micro Meter. In the presence of 2 micro Meter riluzole, fiber volleys were depressed by 21 plus/minus 8% for first pulse responses (P < 0.001). The second and subsequent fiber volleys in a 30-Hz train showed a marked increased in sensitivity to riluzole (Figure 8). The second response in a train was depressed an additional 10% (to approximately 30%), and the 12th response was depressed by a further 20% compared with the first, for a total depression of approximately 40%. Onset of the riluzole effect was faster for responses near the end of the train, and recovery was slower.
Riluzole was more potent than other anesthetics at depressing CA 1 neuron population spike amplitudes, consistent with the higher potency for riluzole observed in vivo. [12,19]Riluzole depressed responses at lower concentrations (5 micro Meter) than halothane (EC50= 230 micro Meter), thiopental (EC50= 50 micro Meter, Figure 2)* or propofol (EC50= 20 micro Meter). Riluzole exhibited slower onset kinetics compared with these other anesthetics. The long duration of riluzole effects (160 min, Figure 8(C), approximately 4 times longer than thiopental) were also comparable with those seen in vivo, although the markedly different pharmacokinetic properties of the two systems make it difficult to relate drug effects seen in brain slices to those observed in vivo. Sleep times in vivo were more than 3.5 times longer after intraperitoneal riluzole (25 mg/kg; 178 plus/minus 23 min) compared with thiopental (25 mg/kg; 46 plus/minus 4 min) or ketamine (80 mg/kg; 47 plus/minus 2 min). Halothane, in contrast, produced steady-state effects in brain slices within 2-3 min, and recovery occurred within 15 min, comparable with in vivo response times. [20-22]A slow onset and long duration of action are consistent with a use-dependent mechanism of action, where drug binding is dependent on a transient conformation of a receptor/channel. Results from the current study indicate that riluzole depressed CA 1 neuron discharge by blocking presynaptic conduction in a use-dependent manner.
Fiber Volley Depression Was Necessary and Sufficient for Riluzole-induced Population Spike Block
Riluzole-induced depression of synaptically evoked CA 1 neuron discharge was paralleled by an equivalent reduction in EPSP and fiber volley amplitudes (Figure 9). Input/output analyses of fiber volley versus EPSP versus PS relations demonstrated that depressed fiber volley amplitudes fully accounted for the depression of EPSP responses (Figure 7). Excitatory postsynaptic potential depression, in turn, could fully account for the depressed postsynaptic discharge observed (Figure 2, Figure 5, and Figure 6). Therefore, riluzole-induced depression of CA 1 neurons appeared to result from an action at a presynaptic site to block excitatory synaptic input. Enhanced recurrent inhibition was not produced by riluzole, because the ratio of second/first PS responses was not decreased (Figure 5). The second PS response is known to be preferentially depressed by some anesthetics because recurrent GABA-mediated inhibition, resulting from the first population discharge, is enhanced. [23-25]Although volatile anesthetics and barbiturates increase paired pulse inhibition, additional actions at several sites are required to account for depression of CA 1 neuron discharge produced by these anesthetics. These sites include depressed fiber conduction, reduced excitatory input, increased spike threshold, and enhanced GABA-mediated inhibition (Figure 1). [24-28]Riluzole appears to be unique among anesthetics because depressed CA 1 discharge can be attributed to a single site of action: blocked action potential conduction in presynaptic glutamate fibers.
Use-dependent Block of Sodium Channels and Riluzole-induced Depression
Riluzole, like some anesthetics, prevented the memory loss and hippocampal neuron degeneration after ischemia produced by transient carotid artery occlusion in rodents. This neuroprotective effect was also seen in hippocampal slices and was suggested to result from depressed glutamate release, secondary to a block of voltage-activated sodium channels or altered function at N-methyl-D-aspartate receptors. [16,30,31]In Xenopus oocytes, 30 micro Meter riluzole depressed rat-brain-derived sodium channels by stabilizing inactivated channels. Similarly, riluzole blocked sodium channel conductance in frog nerve with an EC50of approximately 50 micro Meter, but with an apparent dissociation constant of 0.29 micro Meter for block of inactivated channels. This selective interaction of riluzole with inactivated sodium channels could explain the use-dependent depression of sodium channel-mediated conduction observed in the current study, although our results cannot address this issue directly. By analogy to the modulated receptor theory, these results suggest that riluzole binds selectively to open or inactivated sodium channels, or both, in a manner analogous to the open channel block produced by some local anesthetics. [15,33,34]The high degree of selectivity for riluzole block of sodium channels (more than 300 times greater than for Potassium sup + channel depression) [13,35]could explain why riluzole may be more useful for general anesthesia compared with traditional local anesthetics. Local anesthetic-induced Potassium sup + channel block leads to hyperexcitable neuronal responses, which appear to counteract the depressant effects produced by Sodium sup + channel block in the central nervous system. The consequences of such channel selectivity for general anesthesia remains to be determined for riluzole and other channel-blocking agents.
The authors thank Heath Lukatch, Catherine Hagan, and Sky Pittson for comments on the manuscript.
*Lukatch HS, Monroe FA, MacIver MB: Thiopental blocks paired-pulse facilitation in hippocampal CA1 neurons (abstract). Society for Neuroscience 1993; 19:275.