Effect of Riluzole on the Neurological and NeuropathologicalChanges in an Animal Model of Cardiac Arrest-InducedMovement Disorder1

ANUMANTHA G. KANTHASAMY,2,3 RICHARD J. YUN,2 BANG NGUYEN,2 and DANIEL D. TRUONG2

Parkinson’s and Movement Disorders Institute, Long Beach Memorial Medical Center, Long Beach, California

Accepted for publication October 7, 1998 This paper is available online at http://www.jpet.org

ABSTRACTPosthypoxic myoclonus and seizures precipitate as secondaryneurological consequences in ischemic/hypoxic insults of thecentral nervous system. Neuronal hyperexcitation may be dueto excessive activation of glutamatergic neurotransmission, aneffect that has been shown to follow ischemic/hypoxic events.Therefore, riluzole, an anticonvulsant that inhibits the release ofglutamate by stabilizing the inactivated state of activated volt-age-sensitive sodium channels, was tested for its antimyo-clonic and neuroprotective properties in the cardiac arrest-induced animal model of posthypoxic myoclonus. Riluzole(4–12 mg/kg i.p.) dose-dependently attenuated the audiogenicseizures and action myoclonus seen in this animal model. His-tological examination using Nissl staining and the novel Fluoro-

Jade histochemistry in cardiac-arrested animals showed anextensive neuronal degeneration in the hippocampus and cer-ebellum. Riluzole treatment almost completely prevented theneuronal degeneration in these brain areas. The neuroprotec-tive effect was more pronounced in hippocampal pyramidalneurons and cerebellar Purkinje cells. These effects were seenat therapeutically relevant doses of riluzole, and the animalstolerated the treatment well. These findings indicate that thepathogenesis of posthypoxic myoclonus and seizure may in-volve excessive activation of glutamate neurotransmission, andthat riluzole may serve as an effective pharmacological agentwith neuroprotective potential for the treatment of neurologicalconditions associated with cardiac arrest in humans.

Myoclonus is a neurological disorder which is character-ized by sudden, brief, shock-like involuntary movementscaused by active muscle contractions or inhibitions (Fahn,1986; Hallett, 1987). These symptoms can be precipitated byvarious pathological conditions affecting the central nervoussystem (CNS). In particular, posthypoxic myoclonus can oc-cur after hypoxic and ischemic episodes (Fahn, 1986). Theacute form of posthypoxic myoclonus is most often associatedwith seizures and occurs spontaneously. It is a life-threaten-ing condition that requires immediate medical treatment(Wijdicks et al., 1994). The delayed form of posthypoxic my-oclonus is stimulus-sensitive; the muscle jerks are expressed

as a type of action myoclonus. This condition affects motorfunction to different degrees of severity ranging from mild toserious debilitation. Therapeutic treatment for posthypoxicmyoclonus has been limited, mainly due to the lack of a clearunderstanding of the pathophysiological mechanisms in-volved. Lack of a suitable animal model has hindered boththe understanding of biochemical mechanisms underlyingthe disorder and better clinical management of the resultingsymptoms. Recently, we have developed an animal model ofposthypoxic myoclonus by inducing 8- to 10-min cardiac ar-rest in rats (Truong et al., 1994; Kanthasamy et al., 1996b).This animal model exhibits behavioral and pharmacologicalcharacteristics that resemble those found in human posthy-poxic myoclonus (Jaw et al., 1994; Truong et al., 1994; Ma-tsumoto et al., 1995a,b; Truong et al., 1995; Kanthasamy etal., 1996a,b; for review: Kanthasamy et al., 1996b). Cur-rently, it is the most realistic animal model available to studypathophysiological mechanisms and to develop novel phar-macotherapies for the treatment of this neurological disor-der.

The precise mechanisms underlying posthypoxic myoclo-nus are not clearly understood, although imbalances in oneor more neurotransmitter systems seem to be involved (Mat-sumoto et al., 1995a,b; Kanthasamy et al., 1996a,b). Gluta-

Received for publication July 13, 1998.1 This work was supported by the Myoclonus Research Foundation and

clinical funds from the Parkinson’s and Movement Disorders Institute. Thehistology work was made possible, in part, through access to the OpticalBiology Shared Resource of the Cancer Center Support Grant CA-62203 at theUniversity of California, Irvine. Support from the National Institutes ofHealth-Society for Advancement of Chicanos and Native Americans in ScienceBiomedical Research for undergraduate students through the University ofCalifornia, Irvine is gratefully acknowledged. This work was presented in partat the 27th Annual Meeting of the Society for Neuroscience, New Orleans, LA,October 1997.

2 Current address: Department of Neurology, College of Medicine, Univer-sity of California Irvine, Irvine, CA 92697.

3 Current address: Department of Community & Environmental Medicine,College of Medicine, University of California Irvine, Irvine, CA 92697.

ABBREVIATIONS: CNS, central nervous system; NMDA, N-methyl-D-aspartate

0022-3565/99/2883-1340$03.00/0THE JOURNAL OF PHARMACOLOGY AND EXPERIMENTAL THERAPEUTICS Vol. 288, No. 3Copyright © 1999 by The American Society for Pharmacology and Experimental Therapeutics Printed in U.S.A.JPET 288:1340–1348, 1999

1340

at Univ of C

alifornia-Riverside Library T

ech. Services/S

erials on Septem

ber 4, 2012jpet.aspetjournals.org

Dow

nloaded from

mate acts as a major excitatory neurotransmitter in the CNSwhere it mediates fast, excitatory synaptic neurotransmis-sion, and it may play a role in neuronal communication andCNS pathology (Garthwaite and Meldrum, 1990; Lipton andRosenberg, 1994). Excessive glutamatergic neurotransmis-sion may be due to an increase in extracellular glutamate, aneffect that has been shown to follow ischemic/hypoxic events(Butcher et al., 1990; Globus et al., 1991). Overactivity ofglutamatergic neurotransmission can lead to rapid changesin neuronal biochemistry that culminate in excitotoxic celldeath in susceptible neuronal populations (Choi and Roth-man, 1990). Thus, excitatory amino acid-mediated neuronaloverexcitation is thought to be an underlying pathologicalmechanism both in neuronal damage and in the neurologicalconsequences of the posthypoxic state. In this context, wehave previously shown that N-methyl-D-aspartate (NMDA)receptor antagonists (Matsumoto et al., 1995b) and nitricoxide synthase inhibitors are effective in attenuating posthy-poxic myoclonus (Truong et al., 1995), indicating that theglutamate system may play a crucial role in the pathophys-iology of this disorder.

Riluzole (2-amino-6-trifluoromethoxy benzothiazole) is ananticonvulsant that primarily intervenes with glutamate-mediated excitation by stabilizing voltage-dependent sodiumchannels in their inactivated configuration (Herbert et al.,1994; Doble, 1996). In addition to its anticonvulsant proper-ties, riluzole exhibits neuroprotective action in various invitro and in vivo models (Malgouris et al., 1989; Pratt et al.,1992; Dessi et al., 1993; Estevez et al., 1995;). This drug hasbeen shown to prevent anoxic injury in cultured cerebellarneurons (Dessi et al., 1993), reduce glutamate neurotoxicityin motoneurons (Estevez et al., 1995), and act as a neuropro-tectant in rodent models of cerebral ischemia (Malgouris etal., 1989; Pratt et al., 1992). It has also been recently ap-proved for treatment of amyotrophic lateral sclerosis, a pro-gressive motor neuron disease thought to involve hyperglu-tamatergic neurotransmission (Lacomblez et al., 1996).Because of its therapeutic potential in glutamate-mediatedoverexcitation, we have tested the antimyoclonic and neuro-protective effects of riluzole in the posthypoxic animal model,using both behavioral and histological measures. Also, wehave examined the utility of Fluoro-Jade histochemistry, anovel histological method, for neuroprotective studies in thecardiac arrest model.

Materials and MethodsAnimals. Adult male Sprague-Dawley rats (200–250 g; Zivic-

Miller Laboratories, Inc., Alison Park, PA) were used in these exper-iments. The rats were housed two per cage in a temperature-con-trolled room (23°C) with a 12:12-h light/dark cycle. The animals wereallowed free access to food and water. All experimental procedureswere approved by the Institutional Animal Care and Use Committeeat the University of California Irvine, and Memorial Health Services.

Animal Model of Cardiac Arrest-Induced Posthypoxic My-oclonus. The cardiac arrest procedure was performed as describedpreviously (Kanthasamy et al., 1996a,b). Briefly, each rat was anes-thetized with ketamine (85 mg/kg i.p.) and xylazine (15 mg/kg i.p.).Atropine (0.04 mg/kg i.p.) was administered to minimize respiratorysecretion. Methoxyflurane was provided as supplemental anesthesiaif necessary. The trachea was intubated with an 18-gauge catheter,which was then attached to a ventilator (settings: 425 ml/min N2O,175 ml/min O2, 60 strokes/min, 5 cm of H2O positive end-expiratory

pressure). The rat was placed on a heating pad, and electrocardio-gram electrodes were attached. Body temperature was maintainedat 37 6 0.5°C with a rectal temperature probe controlled by aservo-feedback circuit. The left femoral artery and vein were cathe-terized to monitor arterial blood pressure and for the administrationof drugs, respectively. Cardiac arrest was initiated and maintainedby mechanically obstructing all the major blood vessels, includingthe aorta, by hooking them with an L-shaped loop and simulta-neously compressing the chest. Cessation of ventilation and drop inblood pressure were confirmed by electrocardiogram and blood pres-sure tracings monitored on the polygraph. Resuscitation began at 8min after the cardiac arrest by resuming ventilation (100 strokes/min, 100% O2), manual thoracic compressions, and i.v. injection of 10mg/kg epinephrine and 4 mEq/kg sodium bicarbonate. Upon success-ful resuscitation, each rat was weaned from the ventilator, the cath-eters were removed, and wounds were sutured. The animal wasplaced in an oxygen tent on a heating pad until it completely recov-ered from the surgical coma. Normally, the animals recovered fromsurgery and began feeding themselves 3 to 5 h postsurgery. Thiscardiac arrest procedure yielded a survival rate of approximately90% with minimal postoperative care.

Riluzole Treatment. The cardiac-arrested rats were divided intofour groups receiving 0, 4, 8, or 12 mg/kg riluzole in 0.01 N hydro-chloric acid, i.p. This dose range was selected because it has beenshown to be effective in other animal studies (Pratt et al., 1992;Doble, 1996). Three doses of riluzole were administered at separatetimes. The first dose was given 2 h after the resuscitation, the timeat which the animals started recovering from the cardiac arrest. Thesecond dose was given 24 h after the cardiac arrest procedure and theseizure score was estimated 15 min after the administration of thesecond dose. The third dose was given 15 min before evaluation ofaudiogenic myoclonus, which was evaluated 48 h after surgery. Inseparate experiments, just the first dose of riluzole (12 mg/kg) wasadministered to determine the efficacy of single dose treatment.Also, the first dose was eliminated, whereas the subsequent doseswere given to elucidate whether neurodegeneration was due to theanti-ischemic or antiseizure properties of riluzole.

Behavioral Testing. After 24 h of postsurgical recovery, theposthypoxic rats were tested for seizure activity. Either drug orvehicle was injected, and a quantitative measure of seizure activitywas determined (Truong et al., 1995) for each rat by a blindedobserver using the following rating scale: 0, no seizure; 1, runningonly/no convulsion; 2, running phase with generalized clonus involv-ing forelimbs, hindlimbs, pinnae, and/or vibrissae; 3, tonic flexion ofneck, trunk, and forelimbs; 4, convulsion with complete tonic exten-sion of hindlimbs; and 5, maximal convulsion followed by loss ofconsciousness.

The rats were tested for auditory stimulus-induced myoclonus aspreviously described (Truong et al., 1994; Kanthasamy et al.,1996a,b) 24 h after seizure testing. Rats were placed in clear Plexi-glas cages (44 3 22 cm) at least 10 min before the behavioral testingand were then presented with 45 clicks (95 dB, 0.75 Hz, 40 ms) of ametronome stimulus. The involuntary muscle jerks to each clickwere scored based on the following criteria: 0, no jerks; 1, ear twitch;2, ear and head jerk; 3, ear, head, and shoulder jerk; 4, whole bodyjerk; and 5, whole body jerk of such severity that it caused a jump.The cumulative score of 45 clicks yielded the total myoclonus scorefor each animal. The myoclonus scores for each animal were deter-mined at 15, 30, 60, 120, and 180 min after administration of thedrug or vehicle, and were compared to baseline scores determined foreach animal before the surgical procedure.

Histological Studies. After the behavioral studies, the rats wereinjected with a lethal dose of sodium pentobarbital (100 mg/kg i.p.)and perfused intracardially with saline followed by 10% Formalin in0.1 M phosphate buffer (pH 7.4). The rats were then decapitated, andtheir brains were removed and immersed in Formalin solution tofurther fix the brain tissue. The brains were then embedded inparaffin wax. Serial coronal sections (6-mm thickness) were taken

1999 Riluzole Attenuates Posthypoxic Neurological Deficits 1341

at Univ of C

alifornia-Riverside Library T

ech. Services/S

erials on Septem

ber 4, 2012jpet.aspetjournals.org

Dow

nloaded from

from various sections of the brain, stained for Nissl substance usingcresyl violet, and examined for pathological changes.

Similar sections were alternatively stained with Fluoro-Jade, anewly developed fluorescent marker that selectively stains degener-ating neurons (Schmued et al., 1997; Freyaldenhoven et al., 1997).Briefly, slides containing sections of brain tissue were dewaxed inxylene, immersed in 100% ethanol for 5 min, 70% alcohol for 2 min,and then rinsed with two 1-min changes of dd-H20. The slides werethen incubated with freshly prepared 0.06% potassium permanga-nate for 17 min and rinsed again with two 1-min changes of dd-H20.This was followed by a 30-min incubation in 0.001% Fluoro-Jadesolution at room temperature and two 1-min rinses with dd-H20. Theslides were air dried with a blow dryer for 10 min at low heat, placedin xylene for 2 min, covered with Permount, and stored in a darkarea. The sections were viewed under a fluorescence research micro-scope (model BH2; Olympus) using a fluorescein-5-isothiocyanatefilter.

Quantitative Histological Analysis. The image processing andquantitative analysis of histological data were performed using Im-age-Pro plus software (Media Cybernetics, Inc., Silver Spring, MD)in combination with a SPOT digital camera (Diagnostic Instruments,Inc., Sterling Heights, MI). The images acquired from the slidesrepresenting the specific color signals (violet for cresyl violet andgreen fluorescence for Fluoro-Jade) were compared with the slidesthat were devoid of the staining (reagent-blank slides) to construct aMagro algorithm with specific hues. The cells represented by theselected color were automatically defined, encircled, numbered, andmeasured.

Statistical Analysis. Data were expressed as mean 6S.E.M. andstatistical significance was determined by ANOVA with the Dun-nett’s test in the case of multiple comparisons or with Student’s t testin case of simple comparisons. Differences were accepted as signifi-cant at p , .05 or less.

ResultsCardiac Arrest-Induced Neurological Deficits. The

cardiac arrest-induced posthypoxic myoclonus rat serves as asuitable experimental model for the present study, becausethis model exhibits not only substantial reproducible neuro-pathological changes but also behaviorally expresses the neu-rological syndrome. After cardiac arrest, the rats exhibitedspontaneous seizures within 12 h and then showed audio-genic seizures for the next 24 to 48 h. The types of seizuresgenerated after cardiac arrest were tonic, partial with wildrunning behavior, and generalized clonic-tonic with loss ofconsciousness. After the seizure period, the rats exhibitedstimulus-sensitive action myoclonus that persisted for a rel-atively longer period, up to 3 weeks postsurgery.

Antiepileptic Effect of Riluzole in Cardiac Arrest-Induced Posthypoxic Myoclonus Animal Model. Be-cause postanoxic seizure precipitates as an early neurologicalsymptoms of severe ischemic insults, antiepileptic propertiesof riluzole were evaluated in the audiogenic seizures phase ofthe posthypoxic animal model of myoclonus. Riluzole dose-dependently reduced the audiogenic seizure in the cardiacarrest animals (Fig. 1). At a dose of 8 mg/kg, the seizure scorewas significantly (p , .05) reduced, whereas at a higher doseof 12 mg/kg, the drug almost completely blocked seizureactivity (p , .01). Also, riluzole treatment reduced the inci-dence of the development of spontaneous seizures, whichoccurs 12 to 24 h postsurgery (data not shown).

Antimyoclonic Effect of Riluzole in Cardiac Arrest-Induced Posthypoxic Myoclonus Animal Model. To de-termine the antimyoclonic properties of riluzole, the drug

was tested for its ability to attenuate stimulus-sensitive my-oclonus. Riluzole dose-dependently attenuated myoclonus inthe posthypoxic rats as shown in Fig. 2. Significant reduc-tions (p , .05) in myoclonus scores were seen at dosages of 8and 12 mg/kg within 30 min of injection. The rats showedbehavioral improvements up to 30 min postinjection, afterwhich the animals exhibited relatively stable behavior. Atthe 4-mg/kg dose, the myoclonus score dropped initially, butreturned to levels similar to control levels within 1 h. Impor-tantly, riluzole at these doses (4–12 mg/kg, i.p.) did not causeany noticeable behavioral side effects. Such side effects arecommonly a problem with drugs that interfere in glutama-tergic transmission (Koek and Colpaert, 1990; Carter, 1994).In addition, unlike NMDA antagonists (e.g., MK-801), thedrug treatment did not markedly alter the body temperature(baseline 5 36.3°C 6 0.6°C versus riluzole (12 mg/kg) treat-ment 5 35.9°C 6 0.7°C).

To determine the efficacy of a single dose of riluzole, 12mg/kg (i.p.) was tested. As shown in Fig. 3, a single dose of 12mg/kg riluzole treatment 2 h after cardiac arrest surgery didnot significantly attenuate anoxic seizures (A) and posthy-poxic myoclonus (B). This single dose treatment showed only

Fig. 1. Antiseizure effect of riluzole in posthypoxic myoclonic rats. Ratswere administered either vehicle or riluzole (4, 8, or 12 mg/kg i.p.) 2 hafter resuscitation and again 15 min before testing. Seizure score wasdetermined using the rating scale described in Materials and Methods.Data represent mean 6 S.E.M. of 6 to 10 animals. Significant reductionsin seizure activity were observed at the 8- and 12-mg/kg dosages ascompared with vehicle treated group (*p , .05; **p , .01).

Fig. 2. Antimyoclonic effect of riluzole in posthypoxic myoclonic rats.Rats were administered either vehicle or riluzole (4, 8, or 12 mg/kg i.p.)and audiogenic myoclonus was evaluated at different time points over a3-h period. Data represent mean 6 S.E.M. of 6 to 10 animals. At the 8-and 12-mg/kg dosages, the drug significantly reduced myoclonus at alltime points tested (*p , .05; **p , .01).

1342 Kanthasamy et al. Vol. 288

at Univ of C

alifornia-Riverside Library T

ech. Services/S

erials on Septem

ber 4, 2012jpet.aspetjournals.org

Dow

nloaded from

a minimal neuroprotective effect in the cerebellar Purkinjecells and hippocampal neurons (data not shown).

Neuroprotective Effect of Riluzole on HippocampalPyramidal and Cerebellar Purkinje Neurons in Post-hypoxic Myoclonus Model. To determine the neuroprotec-tive effect of riluzole against posthypoxic neuronal damage,the brains were histologically evaluated after the behavioralmeasurements. Representative photographs of histologicalchanges in brain sections containing dorsal hippocampusstained with cresyl violet are presented in Fig. 4. Controlanimals exhibited normal cellular architecture of pyramidalcells in hippocampus (Fig. 4, A and B). Hippocampal sectionsof posthypoxic rats treated with vehicle showed a severe lossof the pyramidal cells in region CA1 and CA2, as well aslosses in the dentate granule neurons (Fig. 4, C and D).Damage to CA1 neurons was more dramatic than that ofother regions. Systemic injection of riluzole (12 mg/kg i.p.)almost completely blocked the neuronal damage to pyramidalcells of the hippocampus (Fig. 4, E and F). In Fig. 5, controlanimals show normal morphological characteristic of Pur-kinje cells as stained with cresyl violet (Fig. 5A). These arethe round bipolar cells that lie between the molecular andgranule layers. Cardiac arrest caused almost complete loss ofPurkinje cells (Fig. 5B). Riluzole treatment dramatically res-cued the Purkinje cells from cardiac arrest-induced ischemic

injury (Fig. 5C). There was no clear cell loss noted in theother areas of the brains stained with cresyl violet.

Figure 6 summarizes the quantitative data analysis ofcresyl violet-stained cells in the hippocampus (A) and cere-bellum (B) of riluzole-treated rats. Cardiac arrest reducedthe hippocampal pyramidal cells and cerebellar Purkinjecells to 6 and 21% of noncardiac arrested rats, respectively.Riluzole treatment dose-dependently attenuated the neuro-nal damage. At the 12-mg/kg dose, riluzole protected 80%(p , .01) of neuronal loss in both hippocampus and cerebel-lum.

We have also evaluated the neuroprotective efficacy ofriluzole by eliminating the initial dose (2 h postcardiac ar-rest). This experiment was conducted to determine whetherthe neuroprotective effect of riluzole is related to its anti-ischemic or anticonvulsant properties. The cardiac-arrestedrats that received riluzole treatment (12 mg/kg i.p.) only at24 and 48 h postarrest were not protected from ischemicneuronal injury either in the hippocampus or cerebellum.The quantification of neuronal damage by cresyl violet-stain-ing showed a hippocampal pyramidal neuronal loss of 93 66% and the cerebellar Purkinje neuronal loss of 78 6 5% inthe vehicle-treated group. In comparison, the riluzole-treatedgroup showed a 90 6 7% loss of pyramidal neurons in thehippocampus and 81 6 8% Purkinje cell loss in the cerebel-lum.

Usefulness of Fluoro-Jade Histochemistry for Neu-roprotective Studies in Cardiac Arrest-Induced Post-hypoxic Myoclonus Model. The Fluoro-Jade histofluores-cence technique stained only degenerating neurons indicatedby a bright fluorescence and was remarkably helpful in iden-tifying defective cells in a given population (Schmued et al.,1997). This technique more clearly localizes damaged neu-rons, especially in the cerebellum, as compared to the Nisslstaining. A representative Fluoro-jade labeling in the hip-pocampus and cerebellum is presented in Figs. 7 and 8,respectively. No Fluoro-jade positive fluorescence stainingwas noted in either in hippocampal (Fig. 7A) or cerebellarregions (Fig. 8A) of control animals. Hippocampal and cere-bellar sections from the posthypoxic rats showed manyFluoro-Jade positive cells in the pyramidal (Fig. 7B) andPurkinje cell layers (Fig. 8B), whereas a very few or no suchpositive cells were seen in similar sections of the riluzole-treated posthypoxic rats (Figs. 7C and 8C). Although the lossof cells in hippocampus and cerebellum was quite obviouswith cresyl violet staining, this technique showed subtle orno morphological changes in many areas that clearly showedneuronal damage with Fluoro-Jade. These areas include theparietal cortex, hind limb cortex, thalamus, indusium gri-seum, and zona incerta (data not shown). In addition, Fluoro-Jade staining displayed degenerating axons and dendritesthat were not seen with cresyl violet stain.

The quantitative analysis of Fluoro-jade positive cells ispresented in Fig. 9. Riluzole afforded a significant (p , .01)neuroprotection in both the hippocampus (Fig. 9A) and cer-ebellum (Fig. 9B) as compared with vehicle-treated groups.The effect was dose-related; however, the neuroprotectiveeffect of riluzole as seen by this stain was more dramatic inthe cerebellum than in the hippocampus.

Fig. 3. Effect of a single dose of riluzole on the posthypoxic myoclonus.Rats were administered either vehicle or riluzole (12 mg/kg i.p.) 2 h afterthe cardiac arrest surgery. Audiogenic seizure (A) and myoclonus score(B) measured 24 and 72 h after cardiac arrest, respectively. Data repre-sent mean 6 S.E.M of six animals in each group.

1999 Riluzole Attenuates Posthypoxic Neurological Deficits 1343

at Univ of C

alifornia-Riverside Library T

ech. Services/S

erials on Septem

ber 4, 2012jpet.aspetjournals.org

Dow

nloaded from

DiscussionThe experimental results indicate that riluzole has signif-

icant antimyoclonic and antiepileptic effects in cardiac ar-rest-induced movement disorders, and that the drug offersneuroprotection at the cellular level from posthypoxic degen-erative mechanisms as seen by staining with cresyl violetand Fluoro-Jade reagent. The antimyoclonic action of riluzoleis readily achieved at therapeutic doses and the effect isdose-dependent, indicating the pharmacological validity ofthe study. Riluzole is an antiexcitotoxic agent that has beenshown to be effective in both in vitro and in vivo models of

glutamate overactivation (Malgouris et al., 1989; Benoit andEscande, 1991; Estevez et al., 1995). Together, the presentstudy suggests that excessive activation of glutamatergicneurotransmission after an ischemic insult may play a role inthe posthypoxic myoclonus and seizure, and attenuation ofthis neuronal hyperexcitability by pharmacological meansmay be beneficial for treating the neurological and neuro-pathological changes associated with cardiac arrest.

The mechanism(s) by which riluzole confers its antimyo-clonic effect and neuroprotection is not precisely known.However, previous studies (Martin et al., 1993; Herbert et al.,

Fig. 4. Neuroprotective effect of riluzole on hippocampal sections as seen by staining with cresyl violet. A, dorsal hippocampus in control rats appearshistologically normal. B, the cell bodies of hippocampal pyramidal neurons in CA1 region are densely distributed. C, dorsal hippocampus of cardiacarrested rats shows severe loss of cells. D, there is an almost complete degeneration of pyramidal neurons in CA1 region. E, dorsal hippocampus ofcardiac arrested rat treated with riluzole (12 mg/kg i.p.) shows a remarkable neuroprotective effect. F, the riluzole treatment has almost completelyblocked the cardiac arrest-induced neuronal degeneration in pyramidal cells. Original magnification, A, C, and E 5 403; B, D, and F 5 1003.

1344 Kanthasamy et al. Vol. 288

at Univ of C

alifornia-Riverside Library T

ech. Services/S

erials on Septem

ber 4, 2012jpet.aspetjournals.org

Dow

nloaded from

1994; Doble, 1996) have revealed that riluzole may mediateantiexcitatory amino acid properties via one or more of thefollowing mechanisms: 1) inhibition of glutamate releasefrom presynaptic glutamatergic nerve terminals, 2) stabili-

zation of voltage-dependent sodium channels in their inac-tive conformation, 3) noncompetitive blockade of postsynap-tic NMDA ionotropic channels, and 4) activation of a Gprotein-dependent pathway. The first two mechanisms ap-pear to be interrelated in terms of synaptic transmissionbecause a prolonged inactivation of sodium channels willreduce the ability of the cells to depolarize (firing rate) andthereby block the release of glutamate. Previous studies haveshown that riluzole inhibits the release of glutamate both invitro and in vivo (Cheramy et al., 1992; Martin et al., 1993).In addition, both glutamate release and sodium channel in-activation are blocked by pertussis toxin (Hubert et al.,1994), suggesting that activation of a G protein-related mech-anism may also play a role in the biological effect of riluzole.Radioligand-binding studies do not support the notion thatriluzole inhibits glutamate receptors because the drug doesnot bind to NMDA, a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA), kainate, or metabotropic receptors(Benoit and Escande, 1991).

In global cerebral ischemia, glutamate release occurs dur-ing the occlusion period; however, the levels return to base-line in the reperfusion period. In our cardiac arrest model,which resembles the global ischemic model, glutamate re-lease probably occurs during the cardiac arrest period. Be-cause the first dose of riluzole was administered 2 h post-resuscitation, riluzole probably does not act by the inhibitionof the acute release of glutamate, but rather by attenuationof delayed accumulation of synaptic glutamate through inhi-bition of activated sodium channels in the posthypoxic state.The loss of cellular bioenergetics and ionic homeostasis areearly events of ischemia/hypoxia. Na1 influx initiates thecellular depolarization that contributes to deregulation of anumber of energy-dependent physiological events, includingionic homeostasis and neurotransmitter reuptake. It hasbeen suggested that the beneficial effect of Na1 channelblockers could be the consequence of a reduction in energydemand resulting in improved resistance to ischemia andpreservation of Ca11 homeostasis (Urenjak and Obreno-vitch, 1996). There is also substantial evidence indicatingthat increased levels of synaptic glutamate in the posthy-

Fig. 5. Neuroprotective effect of riluzole on the cerebellar Purkinje cells as seen by staining with cresyl violet. A, the spherical cell bodies of Purkinjecells in control rats are aligned nicely between the granular and molecular layers. B, cardiac arrested animals show dramatic loss of all the Purkinjeneurons. C, the neurons are completely rescued by riluzole treatment (12 mg/kg i.p.). Original magnification, 403.

Fig. 6. Quantitative analysis of the neuroprotective effect of riluzole inthe hippocampus (A) and the cerebellum (B) measured by cresyl violetmethod. Values are the mean 6 S.E.M. from six to eight animals. Aster-isks indicate a significant difference relative to the vehicle-treated groupat **p , .01.

1999 Riluzole Attenuates Posthypoxic Neurological Deficits 1345

at Univ of C

alifornia-Riverside Library T

ech. Services/S

erials on Septem

ber 4, 2012jpet.aspetjournals.org

Dow

nloaded from

poxic period is primarily mediated by reversal of the Na1-dependent glutamate transporter rather than by the vesicu-lar release (Taylor et al., 1995). Therefore, blockade of theNa1 channel subsequently restores the levels of glutamateand bioenergy, and thereby mitigates excitotoxic signalingcascades (activation of proteases, phospholipases, kinases,nitric oxide synthase, etc.) that otherwise lead to neuronaldamage.

A recent patch-clamp study demonstrates that riluzole, atthe therapeutic concentration, selectively inhibits the inacti-vated state of activated sodium channels through a use-dependent fashion (Herbert et al., 1994). Because of this, thenormal sodium channels are insensitive to the inhibitoryeffect of riluzole. Thus, it appears that riluzole’s preferentialaffinity for the activated Na1 channel during cardiac arrestprevents excessive neuronal firing only in susceptible cellswithout disturbing the synaptic transmission in normal cells.This offers a favorable therapeutic index as compared toconventional sodium channel inhibitors that block both openand closed channels. To further confirm the effect of riluzoleon Na1 channels, we tested lamotrigine, another anticonvul-sant that mediates its pharmacological action through Na1

channel inhibition (Cheung et al., 1992; Xie et al., 1995).Under a similar treatment paradigm, lamotrigine attenuatedneurological and neuropathological changes in the cardiacarrest model at the high doses (A.K., T. Tith, B. Nguyen, A.Tron, D. Truong, submitted). However, at the higher dose (25mg/kg i.p.) the drug produced behavioral side effects includ-ing motor uncoordination and hypotonia. These side effectsmay be due to lack of selectivity for activated Na1 channelsover closed channels at the higher doses (Leach et al., 1986).Recent studies have shown that many of a new class ofanticonvulsants that block Na1 channels possess neuropro-tective properties in stroke and traumatic head injury models(Taylor and Meldrum, 1995). Taken together, the selectiveinactivation of activated sodium channels may account forthe observed neuroprotective effects of riluzole in cardiac

arrest-induced posthypoxic myoclonus model without delete-rious side effects.

The ability of riluzole to prevent neuronal degeneration inboth the hippocampal pyramidal cells and cerebellar Pur-kinje cells in our animal model reflects the potent neuropro-tective potential of the drug on diverse neuronal populations.Riluzole’s neuroprotective properties have also been seen inthe gerbil model of global ischemia, in which this drug wasfound to block necrosis of pyramidal cells in the CA1 region(Malgouris et al., 1989; Pratt et al., 1992). Because the limbicsystem plays a crucial role in epileptogenesis, the neuropro-tective effect of riluzole in the hippocampus can be attributedto the observed anticonvulsant properties of the drug. In thecerebellum of posthypoxic rats, degeneration of Purkinje cellsmade evident by the histological analyses indicate that path-ways between sensory inputs to the cerebellar cortex anddeep cerebellar and lateral vestibular nuclei may be dis-rupted; these events could cause dysfunctional neurotrans-mission to motor nuclei and subsequent impairment of motorcontrol (Guyton, 1991). Thus, the ability of riluzole to blockexcessive glutamatergic neurotransmission in the hippocam-pus and cerebellum may contribute to the overall prophylac-tic action of the drug against pathological motor functionafter ischemic or hypoxic insult.

The clinical potential for riluzole for use in treating post-hypoxic myoclonus is enhanced by its lack of effect on cardio-vascular activity and its few side effects compared with otherantiglutamatergic agents. The lack of cardiovascular effectmay be attributed to the insensitive nature of low frequencyNa1 channels in the cardiac cells to riluzole (Doble, 1996). Animplication of the present study is that administration ofriluzole after 2 h postcardiac arrest may provide an excellenttherapeutic window for treatment of neuronal damage andbehavioral deficits after severe cardiac arrest.

The present study demonstrates that the initial 2-h dose iscrucial to obtain a neuroprotective effect in the cardiac arrestmodel. This emphasizes the importance of the initial avail-

Fig. 7. Neuroprotective effect of riluzole on hippocampal sections as stained histochemically with Fluoro-Jade. This novel reagent labels degeneratingcells in green, whereas normal cells appear unstained. A, CA1 region of hippocampus in the control animals show no Fluoro-Jade positive cells. B, CA1region of hippocampus in cardiac arrested rats shows extensive labeling in nearly all pyramidal neurons. C, CA1 region of hippocampus in cardiacarrested rats treated with riluzole (12 mg/kg) shows few Fluoro-Jade positive cells. Original magnification, 1003.

1346 Kanthasamy et al. Vol. 288

at Univ of C

alifornia-Riverside Library T

ech. Services/S

erials on Septem

ber 4, 2012jpet.aspetjournals.org

Dow

nloaded from

ability of the neuroprotective agent in the critical onset phaseof the ischemic neurodegenerative process. This result alsosuggests that the neuroprotective effect of riluzole may pri-marily be due to the anti-ischemic properties of the drug.Another advantage of using riluzole for treatment of posthy-poxic neurological consequences is that the drug is orallyactive and readily crosses the blood-brain barrier. Further-more, its marked ability to attenuate myoclonus in rats atdosages well tolerated in humans (Lacomblez et al., 1996)augments its potential for therapeutic use in the treatment ofhuman forms of this disorder. Because of its favorable phar-macokinetic and pharmacodynamic properties, the drug hasrecently been approved for treatment of progressive neuro-degeneration in amyotrophic lateral sclerosis, a neurologicaldisease in which overactivity of the glutamate system has

been well documented (Leigh and Meldrum, 1996). Althoughseveral antiglutamatergic agents have been shown to be ef-fective in attenuating neuronal in damage animal models ofischemic/hypoxic neuronal injury (Boast et al., 1988; Fadenet al., 1989), these drugs produce severe contraindications,including psychotomimetic effects, learning deficits, and mo-tor uncoordination (Koek and Colpaert, 1990; Carter, 1994).In addition, several of these compounds do not cross theblood-brain barrier very effectively, posing a problem in ad-ministering these drugs in clinical settings.

In conclusion, this study demonstrates that riluzole dra-matically reduces neurobehavioral and histological changesin the cardiac arrest-induced animal model of posthypoxicmyoclonus. It also suggests that excessive activation of neu-ronal excitation may play an important role in the pathogen-esis of posthypoxic myoclonus, and that riluzole may providea therapeutic benefit in the medical management of neuro-logical complications associated with cardiac arrest.

Acknowledgments

We thank Dr. Issac Opole, Jarrad Wagner, Amy Tran, Tevy Tith,Thuy Thoung, and Eddie Ramirez for their assistance with thebehavioral and histological experiments. R.J.Y. is a recipient of asummer fellowship from the Myoclonus Foundation.

ReferencesBenoit E and Escande D (1991) Riluzole specifically blocks inactivated Na1 channels

in myelinated nerve fibre. Pfluegers Arch 419:603–609.

Fig. 8. Neuroprotective effect of riluzole on cerebellar sections as stainedwith Fluoro-Jade. In controls, cerebellar neurons show no Fluoro-Jadelabeling (A). In cardiac-arrested rats, cerebellar Purkinje neurons showintense labeling (B). Riluzole treatment has completely blocked the de-generation of Purkinje neurons (C). Original magnification, 1003.

Fig. 9. Quantitative analysis of the neuroprotective effect of riluzole inthe hippocampus (A) and the cerebellum (B) as measured by Fluoro-Jademethod. Values are the mean 6 S.E.M. from six to eight animals. Aster-isks indicate a significant difference relative to the vehicle-treated groupat *p , .05; **p , .01.

1999 Riluzole Attenuates Posthypoxic Neurological Deficits 1347

at Univ of C

alifornia-Riverside Library T

ech. Services/S

erials on Septem

ber 4, 2012jpet.aspetjournals.org

Dow

nloaded from

Boast CA, Gerhardt SC, Pastor G, Lehmann J, Etienne PE and Liebeman JM (1988)The N-Methyl-D-Aspartate antagonists CGS 19755 and CPP reduce ischemic braindamage in gerbils. Brain Res 442:345–348.

Butcher SP, Bullock R, Graham DI and McCulloch J (1990) Correlation betweenamino acid release and neuropathologic outcome in rat brain following middlecerebral artery occlusion. Stroke 21:1727–1733.

Carter AJ (1994) Many agents that antagonize the NMDA receptor-channel complexin vivo also causes disturbances of motor coordination. J Pharmacol Exp Ther269:573–580.

Cheramy A, Barbeito L, Godeheu G and Glowinski, J (1992) Riluzole inhibits therelease of glutamate in the caudate nucleus of the cat in vivo. Neurosci Lett147:209–212.

Cheung H, Kamp D and Harris E (1992) An in vitro investigation of the action oflamotrigine, a novel potential antiepileptic drug: II. Neurochemical studies on themechanism of action. Epilepsy Res 13:107–112.

Choi DW and Rothman SM (1990) The role of glutamate neurotoxicity in hypoxic-ischemic neuronal death. Annu Rev Neurosci 13:171–182.

Dessi F, Ben-Ari Y and Charriaut-Marlangue C (1993), Riluzole prevents anoxicinjury in cultured cerebellar granule neurons. Eur J Pharmacol 250:325–328.

Doble A (1996) The pharmacology and mechanism of action of riluzole. Neurology47:S233–S241.

Estevez AG, Stutzmann JM and Barbeito L (1995) Protective effect of riluzole onexcitatory amino acid-mediated neurotoxicity in motor neuron-enriched cultures.Eur J Pharmacol 280:47–53.

Faden AI, Demediuk P, Panter SS and Vink R (1989) The role of excitatory aminoacids and NMDA receptors in traumatic brain injury. Science (Wash DC) 244:789–800.

Fahn S (1986) Posthypoxic action myoclonus: Literature review and uptake. AdvNeurol 26:49–84.

Freyaldenhoven TE, Ali SF and Schmued LC (1997) Systemic administration ofMPTP induces thalamic neuronal degeneration in mice. Brain Res 759:9–17.

Garthwaite J and Meldrum BS (1990) Excitatory amino acid neurotoxicity andneurodegenerative disease. Trends Pharmacol Sci 11:379–387.

Globus M-T, Busto R, Martinez E, Valdes I, Dietrich WD and Ginsberg, MD (1991)Comparative effect of transient global ischemia on extracellular levels of gluta-mate, glycine, and gamma-aminobutyric acid in vulnerable brain regions in therat. J Neurochem 57:470–478.

Guyton AC (1991) The cerebellum, the basal ganglia, and overall motor control, inBasic Neuroscience-Anatomy & Physiology (Guyton AC ed) 2nd ed., pp 224–253,W.B. Saunders Company, Philadelphia.

Hallett M (1987) The pathophysiology of myoclonus. Trends Neurosci 10:69–73.Herbert T, Drapeau P, Pradier L and Dunn RJ (1994) Block of the rat brain IIA

sodium channel alpha subunit by the neuroprotective drug riluzole. Mol Pharma-col 45:1055–1060.

Hubert JP, Delumeau JC, Glowinski, J, Permont J and Doble A (1994) Antagonismby riluzole of entry of calcium evoked by NMDA and veratridine in rat culturedgranule cells: Evidence for a dual mechanism of action. Br J Pharmacol 113:261–267.

Jaw SP, Hussong MJ, Matsumoto RR and Truong DD (1994) Involvement of 5-HT2receptors in posthypoxic stimulus-sensitive myoclonus in rats. Pharmacol BiochemBehav 49:129–131.

Kanthasamy AG, Matsumoto RR and Truong DD (1996a) Animal models of myoc-lonus. Clin Neurosci 3:236–245.

Kanthasamy AG, Vu TQ, Yun RJ and Truong DD (1996b) Antimyoclonic effect ofgabapentin in a posthypoxic animal model of myoclonus. Eur J Pharmacol 297:219–224.

Koek W and Colpaert FC (1990) Selective blockade of N-methyl-D-aspartate

(NMDA)-induced convulsions by NMDA antagonists and putative glycine antag-onists: Relationship with phencyclidine-like behavioral effects. J Pharmacol ExpTher 252:349–357.

Lacomblez L, Bensimon G, Leigh PN, Guillet P, Meininger V and for the Amyotro-phic Lateral Sclerosis/Riluzole Study Group-II (1996) Dose-ranging study of ri-luzole in amyotrophic lateral sclerosis. Lancet 347:1425–1431.

Leach MJ, Marden CM and Miller AA (1986) Pharmacological studies on lam-otrigine, a novel potential antiepileptic drug: II. Neurochemical studies on themechanism of action. Epilepsia 27:490–497.

Leigh PN and Meldrum BS (1996) Excitotoxicity in ALS. Neurology 47:S221–S227.Lipton SA and Rosenberg PA (1994) Excitatory amino acids as a final common

pathway for neurological disorders. N Engl J Med 330:613–622.Malgouris C, Bordot F, Daniel M, Pellis F, Rataud J, Uzan A, Blanchard JC and

Laduron PM (1989) Riluzole, a novel antiglutamate, prevents memory loss andhippocampal neuronal damage in ischaemic gerbils. J Neurosci 9:3720–3727.

Martin D, Thompson MA and Nadler JV (1993) The neuroprotective agent riluzoleinhibits release of glutamate and aspartate from slices of hippocampal area CA1.Eur J Pharmacol 250:473–476.

Matsumoto RR, Aziz N and Truong DD (1995a) Association between brain indola-mine levels and severity of posthypoxic myoclonus in rats. Pharmacol BiochemBehav 50:533–538.

Matsumoto RR, Nguyen DC and Truong DD (1995b) Strychnine-insensitive glycineantagonists attenuate a cardiac arrest-induced movement disorder. Eur J Phar-macol 275:117–123.

Pratt J, Rataud J, Bardot F, Roux M, Blanchard JC, Laduron PM and Stutzmann,JM (1992) Neuroprotective actions of riluzole in rodent models of global and focalcerebral ischemia. Neurosci Lett 140:225–230.

Schmued LC, Albertson C and Slikker Jr W (1997) Fluoro-jade: A novel fluorochromefor the sensitive and reliable histochemical localization of neuronal degeneration.Brain Res 751:37–46.

Taylor CP, Burk SP and Weber ML (1995) Hippocampal slices: Glutamate overflowand cellular damage from ischemia are reduced by sodium channel blockade.J Neurosci Methods 59:121–128.

Taylor CP and Meldrum BS (1995) Na1 channels as targets for neuroprotectivedrugs. Trends Pharmacol Sci 16:309–316.

Truong DD, Matsumoto RR, Schwartz PH, Hussong MJ and Wasterlain, CG (1994)A novel rat cardiac arrest model of post-hypoxic myoclonus. Mov Disorder 9:201–206.

Truong DD, Vu TQ, Matsumoto RR and Kanthasamy AG (1995) Nitric oxide medi-ated neurological disorder in a posthypoxic rats (Abstract). Neuroscience 21:1379.

Urenjak J and Obrenovitch TP (1996) Pharmacological modulation of voltage-gatedNa1 channels: A rational and effective strategy against ischemic brain damage.Pharmacol Rev 48:21–67.

Wijdicks EFM, Parisi JE and Sharbrough FW (1994) Prognostic value of myoclonusstatus in comatose survivors of cardiac arrest. Ann Neurol 35:239–243.

Xie X, Lancaster B, Peakman T and Garthwaite J (1995) Interaction of the antiepi-leptic drug lamotrigine with recombinant rat brain type IIA Na channels and withnative Na channels in rat hippocampal neurons. Pfluegers Arch Eur J Physiol430:437–446.

Send reprint requests to: Anumantha Kanthasamy, Ph.D., Parkinson’s andMovement Disorders Institute, Long Beach Memorial Medical Center Re-search Building, 2801 Atlantic Avenue, P.O. Box 1428, Long Beach, CA 90801-1428. E-mail: akanthas@uci.edu

1348 Kanthasamy et al. Vol. 288

at Univ of C

alifornia-Riverside Library T

ech. Services/S

erials on Septem

ber 4, 2012jpet.aspetjournals.org

Dow

nloaded from