ACA

CG

IUd

Aanctm

ofiarfradgcat

(dNirvmfegowtavp

*EA�oaPdracopw

Neuroscience 137 (2006) 241–251

0d

LPHA-LINOLENIC ACID AND RILUZOLE TREATMENT CONFEREREBRAL PROTECTION AND IMPROVE SURVIVAL

FTER FOCAL BRAIN ISCHEMIA

Kttt

Klc

Smcs3piwttn

6(npeseE

ldeBaL

iaio2stpsi

. HEURTEAUX,* C. LAIGLE, N. BLONDEAU,

. JARRETOU AND M. LAZDUNSKI

nstitut de Pharmacologie Moléculaire et Cellulaire, UMR 6097, CNRSniversité de Nice Sophia Antipolis, Institut Paul Hamel, 660 Routees Lucioles, Sophia-Antipolis, 06560 Valbonne, France

bstract—We investigated here the effects of alpha-linoleniccid and riluzole, both activators of the 2P-domain K� chan-el family TREK/TRAAK, in a model of focal ischemia clini-ally relevant to stroke, not only assessing neuronal protec-ion, but also long term survival. Moreover, all the drug treat-ents were initiated post-ischemia.

Mice were subjected to transient middle cerebral arterycclusion (1 h) and reperfusion according to the intraluminallament model. Drugs were injected into the jugular veinccording to three protocols: (i) a single dose of 4 mg/kgiluzole or 500 nmol/kg alpha-linolenic acid at different reper-usion time; (ii) a three-day therapy (a single dose of 2 mg/kgiluzole and 250 nmol/kg alpha-linolenic acid given 1–2, 48nd 72 h after reperfusion); (iii) a three-week therapy (a singleose of 2 mg/kg riluzole and 250 nmol/kg alpha-linolenic acidiven once a week during three weeks after reperfusion. Aombined treatment with 2 mg/kg riluzole�250 nmol/kglpha-linolenic acid injected 2 h after reperfusion was alsoested.

A single dose of riluzole (4 mg/kg) or alpha-linolenic acid500 nmol/kg) injected up to 3 h after reperfusion reducedrastically the stroke volume by 75% and 86%, respectively.eurological deficits 24 h after ischemia were significantly

mproved by alpha-linolenic acid500 or riluzole4 with a neu-ological score of 1.8 as compared with 2.5 observed inehicle-treated mice. Alpha-linolenic acid- and riluzole treat-ent were associated with a reduction in cytopathological

eatures of cell injury, including DNA fragmentation and Baxxpression in the cortex and the caudate putamen. With re-ard to the survival rate at 30 days, the best protections werebtained with the alpha-linolenic acid-injection in the three-eek therapy as well as with a single dose of the combined

reatment (2 mg/kg riluzole�250 nmol/kg alpha-linoleniccid). Palmitic acid, a saturated fatty acid that does not acti-ate the 2P-domain K-channel TREK/TRAAK family, did notrovide any neuroprotection.

Corresponding author. Tel: �33-493957784.-mail address: heurteau@ipmc.cnrs.fr (C. Heurteaux).bbreviations: ALA, alpha-linolenic acid; ALA250, 250 nmol/kg-linolenic acid; ALA500, 500 nmol/kg �-linolenic acid; ANOVA, analysisf variance; MABP, mean arterial blood pressure; MCA, middle cerebralrtery; MCAO, middle cerebral artery occlusion; PALM, palmitic acid;AL500, 500 nmol/kg palmitic acid; pCO2, partial pressure of carbonioxide; pO2, partial pressure of oxygen; RLZ, riluzole; RLZ2, 2 mg/kgiluzole; RLZ4, 4 mg/kg riluzole; TRAAK, TWIK-1 related arachidoniccid-stimulated potassium channel; TREK, TWIK-1-related potassiumhannel; TTC, 2,3,5-triphenyltetrazolium chloride; TUNEL, terminal de-xynucleotidyl transferase (TdT)-mediated 2=-deoxyuridine 5=-triphos-

dhate-biotin nick-end labeling; TWIK-1, tandem of pore domains in aeak inward rectifying potassium channel.

306-4522/06$30.00�0.00 © 2005 Published by Elsevier Ltd on behalf of IBRO.oi:10.1016/j.neuroscience.2005.08.083

241

Taken together, these data suggest that the TREK/TRAAK-channel family may be a promising target for neuroprotec-

ion, and that riluzole and alpha-linolenic acid could be ofherapeutic value against focal ischemia/reperfusion injury tohe brain. © 2005 Published by Elsevier Ltd on behalf of IBRO.

ey words: neuroprotection, polyunsaturated fatty acids, ri-uzole, focal ischemia, long-term survival, 2P-domain K�

hannels.

troke remains the third leading cause of death and theost common cause of adult disability in industrialized

ountries. Until now, the only treatment for acute ischemictroke that has a proven efficacy is i.v. thrombolysis withinh of onset. However, only a small percentage of the

atients are eligible for this therapy. Anticoagulant therapys also performed for the prevention of stroke in patientsith high risk factors and for halting the evolution of a

hrombotic stroke (Hankey and Warlow, 1999). It remainsherefore of foremost importance to develop efficaciouseuroprotective agents.

The neuroprotective effects of riluzole (RLZ, 2-amino--trifluoromethoxy benzothiazole) and alpha-linolenic acidALA) have been well established in different models ofeurodegeneration. RLZ currently in clinical use amongatients with amyotrophic lateral sclerosis (Bensimont al., 1994) induces major protective effects against brain,pinal cord and retina injuries (Malgouris et al., 1989; Prattt al., 1992; Martin et al., 1993; Barneoud et al., 1996;ttaiche et al., 1999; Lang-Lazdunski et al., 1999).

Both in vitro and in vivo studies have shown alpha-inolenic acid-benefit for neurons in hyperexcitability-in-uced neuronal death such as epileptic seizures (Voskuylt al., 1998; Leaf et al., 1999; Lauritzen et al., 2000;londeau et al., 2001). It also prevents neuronal cell deathnd paraplegia after transient spinal cord ischemia (Lang-azdunski et al., 2003).

The middle cerebral artery occlusion (MCAO) using thentraluminal suture technique has been shown to reliablynd rapidly induce neurological deficits and hemispheric

nfarcts restricted to the territory of the MCA 24 h after thenset of ischemia (Ginsberg and Busto, 1989; Sharp et al.,000). Using this model described as the closest to humantroke and monitoring the long-term recovery, the aim ofhe present work is to demonstrate the relevance of therotective effect of alpha-linolenic acid and RLZ againsttroke for future clinical study. Both drug treatments were

nitiated post-ischemia, to determine their therapeutic win-

ows. Three-day and three-week therapies effects were

ttdataTwdapttd

A

AtLwsicatd

P

Gtmaepip(oop(sFtiup

I

FbZ(lgBAibicw

ccbfltpwlt

D

Taiwc(5tl(aptifis

icafowjUdi

N

Natdwma

D

TtftCtrfpoueeci

C. Heurteaux et al. / Neuroscience 137 (2006) 241–251242

hen analyzed to characterize a potential improvement inhe survival of animals and their cerebral protection. Therug combination was also tested to evaluate whether thessociation of these drug actions could enhance the pro-ective effects. To test the link between alpha-linoleniccid, riluzole and the 2P-domain K� channel TREK/RAAK family, palmitic acid (PALM), a saturated fatty acid,hich does not activate these channels (Lesage and Laz-unski, 1999, 2000; Patel and Honore, 2001) was injecteds negative control and his effects analyzed. Moreover, weaid a particular attention to the comparison of the short-erm and the long-term effects of these agents adminis-ered in single or combined doses and in repeated single-rug treatment over 3 days and 3 weeks.

EXPERIMENTAL PROCEDURES

nimals

ll experiments were conducted according to the NIH guide andhe Society of Neurosciences guidelines for the Care and Use ofaboratory Animals (NIH publications No. 80-23, 1985). All effortsere made to minimize the number of animals used and theiruffering. Adult male C57/Bl6 mice, weighing 22–26 g were usedn this study. The animals housed under controlled laboratoryonditions with a 12-h light/dark cycle, a temperature of 21�2 °C,nd a humidity of 60–70% for at least one week prior to drugreatment or surgery. The mice had free access to standard rodentiet and tap water.

hysiological parameters

eneral anesthesia was induced with 3% isoflurane and main-ained with 1% isoflurane by means of an open facemask for eachouse. Mice were allowed to breathe spontaneously. A subset ofnimals (n�5 per group) was monitored for physiological param-ters including mean arterial blood pressure (MABP), rectal tem-erature, arterial blood gases and pH before, during and after

schemia. The right femoral artery was catheterized with PE-10olyethylene tubing and connected to a blood pressure transducerHarvard Apparatus, Les Ulis, France) for continuous monitoringf MABP (mm Hg). A heparinized blood sample (75 �l) was thenbtained from the catheterized femoral artery and blood partialressure of oxygen (pO2), partial pressure of carbon dioxidepCO2) and pH were measured using an Acid-Base Laboratoryystem (ABL 555, Radiometer-Copenhagen, Neuilly/Plaisance,rance). Core temperature was monitored continuously with a

hermometer (3-mm probe diameter, Harvard Apparatus), insertednto the rectum and maintained at physiological temperaturessing a thermostatically controlled heating blanket (Harvard Ap-aratus).

nduction of transient focal cerebral ischemia

ocal ischemia was induced by occlusion of the left middle cere-ral artery (MCA) using an intraluminal filament technique (Ding-hou et al., 2002) based on modifications of the original rat model

Longa et al., 1989). After a midline neck incision was made, theeft common and external carotid arteries were isolated and li-ated with a 4-0 silk suture (Ethicon, Johnson and Johnson Intl,russels, Belgium). A temporary Yasargil aneurysm clip (BMH31,esculap, Tuttlingen, Germany) was temporarily placed on the

nternal carotid artery. A 6-0 nylon monofilament (Ethicon),lunted at tip with an open flame, was introduced through a small

ncision into the common carotid artery and 13 mm distal to thearotid bifurcation for occlusion of the origin of the MCA. Animals

ere kept at 37 °C for one hour, after which time the thread was o

arefully withdrawn to allow reperfusion of the MCA territory. Toontrol the MCAO severity regional CBF (rCBF) was determinedy laser-Doppler flowmetry (Perimed, Craponne, France) using aexible 0.5-mm fiber optic extension to the master probe fixed onhe intact skull over the ischemic cortex. Sham-operation waserformed inserting the thread into the common carotid arteryithout advancing it to occlude the MCA. The animals were al-

owed to regain full consciousness on a heating pad before re-urning to the cage.

rug treatments

he experimental protocol is shown in Fig. 1. It includes one studyt short term, where the animals were analyzed at 24 h post-

schemia (Fig. 1A) and three studies at long term (Fig. 1B, C, D)here the mice were evaluated one month following MCA oc-lusion. In study A, mice received a single dose of 4 mg/kg riluzoleRLZ4) or 500 nmol/kg alpha-linolenic acid (ALA500), or00 nmol/kg palmitic acid (PAL500) at different times after MCAOo determine the therapeutic window. The doses used were se-ected based on our previous studies in different animal modelsLang-Lazdunski et al., 1999; Lauritzen et al., 2000; Blondeau etl., 2001) and from pilot studies on mice searching for the bestrotection (data not shown). Single doses of RLZ and ALA wereested at different time points to establish therapeutic windows. Its based on these results that we chose as reference treatment axed dose for a determined time point of injection after reperfu-ion.

Next, the efficiency of a single dose of RLZ4 or ALA500njected one (RLZ4) or two (ALA500) hours after reperfusion wasompared with a dose of 2 mg/kg riluzole (RLZ2) or 250 nmol/kglpha-linolenic acid (ALA250) given 1–2, 48 and 72 h after reper-usion (Fig. 1B) or 1–2 h after reperfusion following MCAO andnce a week during the next 2 weeks (Fig. 1C). In study C, miceere treated with a combined treatment of RLZ2�ALA250 in-

ected 2 h after reperfusion following MCAO (Fig. 1D). RLZ (RBI,SA), ALA and PALM (Biomol, USA) were injected as a bolusirectly in the jugular vein. Sham-operated animals and mice

njected with vehicle were used as controls.

eurological deficits

eurological deficits of mice were assessed 24 h post-ischemia inblinded fashion according to a scoring scale in a postural reflex

est developed by Bederson et al. (1986): Grade 0: no visibleeficits; Grade 1: forelimb flexion; Grade 2: unidirectional circlinghen the animal is pulled by the tail; Grade 3: circling and rollingovement; Grade 4: decreased level of consciousness. Mice withneurological deficit above grade 3 were excluded from the study.

etermination of infarct volume

o assess the infarct volume in the short-term study (single injec-ion of ALA, RLZ or PALM), mice were killed at 24 h after reper-usion. Their brains were removed and sectioned into six 1 mm-hick coronal slices using a tissue chopper (Phymep, France).oronal brain slices were immediately immersed into 2% 2,3,5-

riphenyltetrazolium chloride (TTC, Sigma, France) for 20 min atoom temperature in the dark followed by fixation in a 4% para-ormaldehyde solution overnight prior to analysis as describedreviously (Ding-Zhou et al., 2002). The striatal and cortical areasf infarction, outlined in light were measured on each sectionsing a computer image analysis system and corrected for braindema according to Golanov and Reis (1995). Infarct volume,xpressed in mm3 was calculated by a linear integration of theorrected lesions areas. The same TTC staining at 24 h post-schemia was also applied to assess the first effect on infarct size

f the combined administration used in the long-term study.

teiwfwdftTtmwsbUup

T

T5pBi2(pttpaofQ

Fmt m (RLZ2

C. Heurteaux et al. / Neuroscience 137 (2006) 241–251 243

In addition Cresyl Violet, a dye that stains the Nissl bodies inhe stellate somas of viable neurons was used to confirm thevolution and the extent of infarct volume and analyze cell survival

n ischemic mice. Coronal frozen sections of brain (10 �m-thick)ere added to a solution of 1% Cresyl Violet in 0.25% acetic acid

or 3 min, rinsed, dehydrated and mounted with Entellan. Sectionsere analyzed under light microscopy. Cell counting using stan-ard methods (Blondeau et al., 2002) were performed in a blindashion in striatal and cortical penumbra at 30 days to determinehe level of cerebral damage. Staining coronal brain sections withTC and Cresyl Violet after 24 h and 30 days of recovery, respec-

ively has been shown to provide a reliable estimate of the infarctargins and the boundary between penumbra and core, whichas estimated by the difference between the size of both infarctstudied at 24 and 3 weeks post-ischemia. Histologically, theoundaries between core and penumbra can be well evaluated.nlike the core region where many neurons are dead, the pen-mbra region appears with intact neurons and cells presenting

ig. 1. Flowchart illustrating the different paradigms. Timing of MCAortality rate, TTC-staining, Bax expression, TUNEL staining and cell c

herapy paradigm (RLZ2 or ALA250). (C) Three-week therapy paradig

yknosis. w

UNEL staining

erminal deoxynucleotidyl transferase (TdT)-mediated 2=-deoxyuridine=-triphosphate-biotin nick-end labeling (TUNEL) staining waserformed by using the In Situ Cell Death Detection Kit (Roche,asel, Switzerland), which allows to demonstrate apoptotic cells in

schemic brain tissue as described previously (Blondeau et al.,002). Coronal 10 �m frozen sections were rehydrated in ethanol95, 70 and 50%) followed by PBS and bathed in 0.3% hydrogeneroxide/methanol to inactivate endogenous peroxidase. Sec-ions were then permeabilized in 0.3% Tween/PBS and washedwice in PBS. Labeling of 3=-OH terminal DNA fragments was thenerformed at 37 °C for 2 h by using the TUNEL reaction mixtureccording to the manufacturer’s protocol. Positive control wasbtained by pre-incubation of a section with DNAse I (20 �g/ml)or 15 min at 37 °C before incubation with biotinylated dUTP.uantitative (number of TUNEL-positive cells/0.5 mm2) analyses

sion, ALA and RLZ treatment at different doses, neurological score,A) Single dose paradigm (RLZ4 or ALA500 or PAL500). (B) Three-dayor ALA250). (D) Combined therapy paradigm (ALA250�RLZ2).

O/reperfuounting. (

ere performed in the ipsilateral cortical penumbra. TUNEL-pos-

ia

W

Wctgr(TT4Pfgfmg(b(Wc1h(rsu

I

Fmafwr1s(p(wE

S

DrtvdlfP

Ap

Tpt

sriM1tgacrtcmaRRcaigwusptrwtu

mriA

wfsBmtcns

T(d

MpppRN

*

C. Heurteaux et al. / Neuroscience 137 (2006) 241–251244

tive cells displayed a brown staining within the nucleus in thepoptotic cells.

estern blot analysis

hole-cell protein extraction was performed using standard pro-edures. Samples were obtained from the MCA territory cortex onhe ischemic sides and from nonischemic controls. (n�3 perroup). Fresh brain tissue was cut into pieces after 24 h ofeperfusion and homogenized in four volumes of cold lysis buffer20 nmol/l Tris pH: 7.5, 137 mmol/l NaCl, 2 mmol/l EDTA, 1%riton X-100, 10% glycerol, and protease inhibitor cocktail) on ice.he homogenates were centrifuged at 12,000�g for 30 min at°C. The supernatant was stored at �70 °C until further use.

rotein concentrations were measured using conventional Brad-ord’s method. Fifty microgram proteins from each experimentalroup were applied to 10% SDS PAGE gels and electrophoresed

or 1 h at 100 mA. Proteins were transferred onto nitrocelluloseembrane (Hybond) in blotting buffer (156 mmol/l Tris, 1 mol/llycine, PBS) for 90 min at 80 mA and blocked with 5% skim milkRegilait) in PBS for 2 h at room temperature. The blotted mem-rane was incubated with the rabbit polyclonal Bax antibodySanta Cruz, SC493, diluted 1/1000) for 2 h at room temperature.

estern blots were incubated with horseradish peroxidase-onjugated anti-rabbit IgG (Jackson ImmunoResearch, diluted/15,000) for 1 h at room temperature and revealed using en-anced chemiluminescence Western blotting detection reagentsPierce). To control for sample loading, stripped membranes wereehybridized with a �-tubulin antibody as internal control (data nothown). Films with specific bands were scanned and quantifiedsing an imaging densitometer.

mmunohistochemistry

rozen brain sections 25 �m thick were immersed in 0.3% H2O2/ethanol for 10 min, permeabilized in 0.1% Triton/PBS for 10 minnd blocked with 3% goat serum/PBS for 2 h at room temperature

ollowed by a single rinse in PBS. Sections were incubated thenith the primary antibody overnight. The antiserum used was the

abbit polyclonal Bax antibody (Santa Cruz, SC493, diluted/300). After the primary incubation and three rinses in PBS,ections were then incubated in biotinylated horse anti-rabbit IgGJackson ImmunoResearch, diluted 1/15,000) for 2 h at room tem-erature. Bax expression was visualized by 3,3=-diaminobenzidineDAB) staining using the VectaStain ABC kit (Biovalley). All sectionsere washed in PBS and then in distilled water, and mounted withntellan.

tatistical analysis

ata were expressed as mean�S.D. of n observations, where nepresents the number of animals or samples. Comparison be-ween multiple groups was evaluated by one-way analysis ofariance (ANOVA) followed by a PLSD (protected least significantifference) Fisher’s test. When measures were repeated (physio-

ogical variables) data were analyzed by a two-factorial ANOVAor repeated measures. Significance was assumed at a value of�0.05.

RESULTS

single dose of ALA500 or RLZ4 induced cerebralrotection at 24 h and 30 days after MCAO

o determine whether acute administration of ALA or RLZrotects against ischemic stroke, mice were subjected to

ransient (1 h) MCAO and injected intraveniously with a t

ingle dose of ALA500 or RLZ4 at different times aftereperfusion. This paradigm corresponds to Fig. 1A. Admin-stration of RLZ or fatty acids (ALA or PAL) did not affect

ABP, arterial pO2 or pCO2 and arterial blood pH (Table). The intraluminal suture method resulted in an infarct inhe right MCA perfused region. Representative photo-raphs of brain slices stained with TTC at 24 h after MCAOre shown in Fig. 2A. As indicated by the white area,erebral infarcts in ALA500- and RLZ4-treated mice wereeduced. The infarct spread into the dorsomedial cortex inhe region of the anterior cerebral artery/MCA and theaudate-putamen was particularly inhibited by both treat-ents. Quantitative assessment of infarct volume 24 hfter MCAO revealed the therapeutic window of ALA500 orLZ4 treatment (Fig. 2B). Administration of ALA500 orLZ4 up to 6 h after reperfusion drastically decreasederebral infarction. Maximal ALA-induced protection waschieved with a dose of 500 nmol/kg injected 2 h post-

schemia with a 86% reduction as compared with vehicle-roup (P�0.001), an effect that was lost when treatmentas initiated 12 h after reperfusion. Administration of RLZp to 3 h after reperfusion at a dose of 4 mg/kg reduced thetroke volume by approximately 75% (P�0.001) as com-ared with control mice, with no significant difference be-ween reperfusion times. In contrast, PALM500, a satu-ated fatty acid, had no beneficial effect on the infarct sizehen injected 2 h after reperfusion, a time corresponding

o the best cerebral protection obtained with ALA, a poly-nsaturated fatty acid (Fig. 2B).

Based on these results, we chose as reference treat-ent the dose of RLZ4 or ALA500, injected 1 or 2 h after

eperfusion, respectively. In this context, neurological def-cits 24 h after ischemia were also significantly improved byLA500 or RLZ4 (Table 1).

The protection induced by ALA500- or RLZ4-injectionas also associated with a reduction in cytopathological

eatures of cell injury, including cell shrinkage (cresyl violettaining, not shown), DNA-fragmentation (Fig. 3A), andax expression (Fig. 3B–D). TUNEL method revealedassive DNA fragmentation by 24 h in the ipsilateral cor-

ex of vehicle-treated mice, which underwent 1 h-MCAO asompared with sham-operated animals. Fig. 3A shows theumber of TUNEL-positive cells, characterized byhrunken cytoplasm and dark nuclei in the ipsilateral cor-

able 1. Physiological parameters 30 minutes after injection of drugsn�5 per group) and neurological deficit score after 1-hour MCAO inrug- and vehicle-treated mice (24 hours following reperfusion)

VEH LIN RLZ PALM

ABP (mm Hg) 90�2 92�3 94�6 92�4aCO2 (mm Hg) 39�6 44�3 46�2 41�3aO2 (mm Hg) 444�41 420�9 462�29 452�27H (arterial) 7.30�0.02 7.35�0.01 7.31�0.02 7.33�0.03T (°C) 37.4�0.3 37.2�0.4 37.2�0.4 37.4�0.3eurologicalscore

2.48�0.11 1.79�0.14* 1.78�0.13* 2.59�0.15

P�0.05 vs. vehicle-treated ischemic mice.

ex of different groups. ALA500- or RLZ4-injection signifi-

cbAsan2puiWaprct

oB3AiWmtcWcmc3R

FoTT r MCAO.D vehicle-t

C. Heurteaux et al. / Neuroscience 137 (2006) 241–251 245

antly reduced the percentage of positive cells with aetter protection with RLZ (13.9% in RLZ4- and 36.5% inLA-groups vs 49% in vehicle-group, P�0.001). Expres-ion of the protein Bax, known to have a role in promotingpoptosis pathways, is regarded as a diagnostic marker ofeurons undergoing cell death (Tsujimoto and Shimizu,000). The induction of the Bax protein in the corticalenumbra of vehicle- and treated-mice has been assessedsing Western blotting (Fig. 3B and 3C) and immunochem-

stry (Fig. 3D) at 3 h after reperfusion following MCAO. Theestern blotting analysis, using the polyclonal anti-Bax

ntibody revealed a single major band at 21 kDa, the sizeredicted for the Bax-� protein (Fig. 3B). Fig. 3C summa-izes the quantification of Bax expression in ipsilateralortical extracts from the complete set of mice. Bax induc-

ig. 2. Effects of a single injection of ALA, RLZ or PALM on infarct sizf TTC-stained from mice killed 24 h after ischemia. One-hour MCAOhe white areas represent the infarct regions in these sections. In cherapeutic window of ALA500 and RLZ4 on infarct volume 24 h afteata are expressed as means�S.E.M. * P�0.05; ** P�0.001 versus

ion remained undetectable in protein extracts from cortex c

f sham-operated groups. One hour MCAO induced strongax expression in cortical extracts of vehicle-treated miceh after reperfusion. In contrast, injection of a dose of

LA500 or RLZ4 after MCAO led to a significant decreasen Bax expression in the cortical extracts of treated mice.

hile Bax expression increased 5.9-fold in vehicle-treatedice, injection of RLZ4 or ALA500 significantly prevented

his increase. Immunostaining on brain sections (Fig. 3D)onfirmed the findings obtained with Western blot analysis.eak Bax immunoreactivity decorated the cytoplasm and

ell processes of neurons in the cortex of sham-operatedice, but increased Bax immunoreactivity occurred in the

ytoplasm of penumbra neurons of the ipsilateral cortex ath following reperfusion. Post-treatment with ALA500 orLZ4 strongly reduced Bax immunoreactivity in penumbra

sections after 1-hour reversible MCAO. (A) Representative samplesion animals treated with vehicle or PALM showed severe infarctions.LA500 RLZ4 treatment had a significant reduction in infarction. (B)ALA, RLZ or PALM was i.v. injected (n�15 per experimental group).reated ischemic mice.

e in brain/reperfusontrast, A

ortical neurons of ischemic brains (Fig. 3D).

bTwoRmcamM3MsrrhCRA7j

ti

Tmi

AcooodaaAptc(pm

Frseo(Ere eated isca ree sepa

C. Heurteaux et al. / Neuroscience 137 (2006) 241–251246

Importantly, the reduction of the infarct volume inducedy ALA500- or RLZ4-injection, previously observed byTC staining, was still occurring at 30 days after MCAOith Nissl staining. The lesion volume, revealed by the lackf Nissl staining, was drastically reduced in ALA500- orLZ4-treated mice, as compared with vehicle-treated ani-als (data not shown). The infarct size reduction was asso-

iated to better cell preservation after ALA500 or RLZ4dministration that lasted for at least one month as deter-ined by cell counting in well-known structures injured byCAO (i.e. dorsomedial cortex and caudate-putamen) at0 days postischemia (Fig. 4A). As expected, one hourCAO produced severe cell loss. Compared with the

ham-operated group, only 57.9 and 43% penumbra neu-ons of cerebral cortex and striatum were preserved,espectively. This cellular loss was almost completely in-ibited by ALA500- or RLZ4-injection in both structures.ortical cell survival was around 95% in ALA500- andLZ4-treated groups (P�0.001). Administration ofLA500 or RLZ4 limited striatal cell destruction to 5.9 and.9% of neurons, respectively (P�0.001). In contrast, in-

ig. 3. Decrease of TUNEL-positive cells and Bax expression folloepresenting percentage of TUNEL-positive cells/total in the ipsilateralides per mouse). * P�0.05, ** P�0.001 versus vehicle (VEH)-treatedxtracts isolated from sham-operated (Sham), VEH-injected brains or br PALM500 was i.v. injected 1 (for RLZ4) or 2 h (for ALA500 or PALMC) Quantification of Bax expression obtained in Western blotting. Txperimental Procedures. Values are expressed as a percentage

epresentative of three separate experiments (n�4 per experimentaxpression of Bax protein in cortical penumbra of RLZ4- or ALA500-trt 3 h after reperfusion following MCAO. Data are representative of th

ection of the saturated fatty acid PALM500 failed to protect f

he brain and no rescue of the cell number was observedn both structures (Fig. 4A).

hree-day and three-week therapies reducedortality and increased cell survival following

schemia at 30 days after MCAO

s indicated in the paradigms of Fig. 1B and 1C, the effi-iency of a single dose of RLZ4 or ALA500 injected 1 (RLZ4)r 2 h (ALA500) after reperfusion was compared with a dosef RLZ2 or ALA250 given 1–2, 48 and 72 h after reperfusionr 1–2 h after reperfusion following MCAO and once a weekuring the next two weeks of reperfusion. The protection wasssessed by neuronal counting and mortality rate one monthfter MCAO. In terms of cell survival, the weekly injection ofLA250 or RLZ2 during three weeks also induced a potentrotection (Fig. 4B and 4C). It limited the striatal cell destruc-ion to 4.6 and 5.7% of neurons (Fig. 4B), and induced aortical cell survival of 96.5 and 95.1%, respectivelyP�0.001) (Fig. 4C). In terms of mortality (Fig. 5), the bestrotection against ischemia was obtained with a weekly treat-ent of ALA250 during three weeks (70% survival rate)

acute injection of ALA500 or RLZ4 after MCAO. (A) Histogramsral cortex (per 0.5 mm2) 24 h after reperfusion (n�10 per group, 10mice. (B) Western blotting analysis of Bax protein induction in cortical

ated with a dose of RLZ4, ALA500 or PAL500 (PALM). RLZ4, ALA500er reperfusion. Brains were removed from mice 3 h after reperfusion.fic bands were quantified using an imaging system as described in-operated mice. Data are expressed as mean�S.E.M. Data are* P�0.05, ** P�0.001 versus sham-operated mice. (D) Decreasedhemic mice as compared with VEH- or PALM-injected ischemic micerate experiments (n�5 per experimental group). Scale bar�100 �m.

wing thel penumbischemicrains tre500) aft

he speciof sham

l group).

ollowed by RLZ-treatment at a dose of 2 mg/kg with the

se

t

FRcicg2(

C. Heurteaux et al. / Neuroscience 137 (2006) 241–251 247

ame periodicity (60% survival rate). Compared with the ref-

ig. 4. Comparative long-term protective effects between single, threLZ on cell death one month after 1-hour reversible MCAO. Histogaudate-putamen in (i) ipsi- and (c) contralateral sides) assessed inschemia. (A) Single dose paradigm: RLZ4 or ALA500 was i.v. injectombined therapies paradigms. Days 1–3: a dose of RLZ2 or ALA250 given 1–2 h after reperfusion and once each week during the next two wh after reperfusion. Data are expressed as means�S.E.M. (n�15 pe

VEH) ischemic mice.

rence treatment (a single dose of ALA500 or RLZ4), the r

hree-day therapy did not significantly improve the survival

three-week therapy and administration of a combined dose of ALA orresent the neuronal density per 0.5 mm2 brain structure (cortex orlet-stained sections (cortex and caudate-putamen) one month afterh after reperfusion, respectively. (B, C) Three-day, three-week and

, 48 and 72 h after reperfusion. Weeks 1–3: a dose of RLZ2 or ALA250reperfusion. Combined: a single dose of RLZ2�ALA250 injected onceental group, 10 slides per animal). ** P�0.001 versus vehicle-treated

e-day orrams repcresyl-vioed 1 or 2iven 1–2eeks of

r experim

ate of ischemic mice (Fig. 5).

Cb

Atgtg

wvttcsi

Frow ose of A

FTca

C. Heurteaux et al. / Neuroscience 137 (2006) 241–251248

ombined doses of RLZ2 and ALA250 induced theest brain resistance to ischemia

s indicated in paradigm D, we next tested the hypothesishat combined administration of a dose of RLZ2�ALA250iven 2 h after MCAO could improve the protection against

he deleterious effects of ischemia. Representative photo-raphs of brain slices stained with TTC at 24 h after MCAO

ig. 5. Comparative effects between three-day or three-week therapieate of mice one month after 1-hour reversible MCAO. Single dose: af RLZ2 or ALA250 given 1–2, 48 and 72 h after reperfusion. Weekseek during the next two weeks of reperfusion. Combined: a single d

ig. 6. Effects of combined administration of ALA250 and RLZ2 on inTC-stained from mice killed 24 h after ischemia. The white areas re

orresponding infarction volumes 24 h after ischemia. A single dose of ALA250re expressed as means�S.E.M. (n�15 per experimental group). ** P�0.001

ith the corresponding quantitative assessment of infarctolume are shown in Fig. 6. As indicated by the white area,he cerebral infarct in mice treated with a combined injec-ion of RLZ2�ALA250 was significantly reduced. Cellounting in the peri-infarct zone 30 days after MCAOhows that a combined injection of RLZ2 and ALA250

nduced a penumbra cortical and striatal cell survival of

ministration of a single or combined dose of ALA and RLZ on survivalLZ4 or ALA500 injected 1 or 2 h after reperfusion. Days 1–3: a dosese of RLZ2 or ALA250 given 1–2 h after reperfusion and once each

LA250�RLZ4 injected once 2 h after reperfusion (n�15 per group).

in brain sections after 1-hour MCAO. (A) Representative samples ofhe infarct regions in these sections. (B) Histograms representing the

s and addose of R1–3: a do

farct sizepresent t

�RLZ2 was i.v. injected 2 h after reperfusion (n�15 per group). Dataversus vehicle-treated (VEH) ischemic mice.

9O2g

Tpmticpo2ro(eaiiebtibiTdtanfo1cwjnflwdcfecetpi(hBvor

ptndNd

wvdcrbmswipattswsvdessrwocridtcuimotada

litshgn1rtb

C. Heurteaux et al. / Neuroscience 137 (2006) 241–251 249

8.7 and 98%, respectively (P�0.001) (Fig. 4B and 4C).ne month after MCAO, the mortality rate was reduced to0% (Fig. 5) compared with 75% in the sham-operatedroup.

DISCUSSION

he purpose of the current study was to investigate therotective effects of ALA and RLZ in a clinically relevantodel associating transient focal ischemia and post-

reatment. After 24 h of MCA occlusion, histological changesndicating ischemic brain damage have been observed inaudate putamen, pallidum, basolateral cortex, hippocam-us and parts of thalamus, which is consistent with previ-us studies of intraluminal MCA occlusion (Hata et al.,000; Ding-Zhou et al., 2002). We have tested a dose-ange of both drugs (ALA500 or RLZ4) based on previ-usly published data in other models of neurodegenerationPratt et al., 1992; Lang-Lazdunski et al., 1999; Lauritzent al., 2000). In the first step of the study, we showed thatn i.v. injection with a single dose of ALA500 or RLZ4

nitiated after the onset of MCAO significantly reduced thenfarct size and improved the neurological score, bothxamined at 24h post-ischemia. ALA500 and RLZ4 wereoth efficient up to 6 h following reperfusion, i.e. within a

herapeutic window, which is relevant clinically. Ischemicnjury is associated with progressive DNA fragmentation inoth striatal and cortical neurons within the infarct area

ncluding the core and penumbra. At 24 h post-reperfusion,UNEL labeling appeared intense and abundant in theseamaged structures. Ischemia-induced DNA fragmenta-ion contains not only internucleosomal fragmentation butlso random breakdown products often associated withecrosis. The TUNEL procedure labels all types of DNAragments. It cannot discriminate accurately between ap-ptotic and necrotic cells (Charriaut-Marlangue et al.,996; Unal-Cevik et al., 2004), and therefore is mostlyonsidered as a marker of damaged cells. In that context,e observed that a single dose of ALA500 or RLZ4 in-

ected 1–2 h following MCAO reduced dramatically theumber of TUNEL-positive cells in cortex after 24 h reper-usion. Programmed cell death and apoptosis are regu-ated by members of the Bcl-2 family, including Bcl-2,hich promotes cell survival and Bax, which promotes celleath. Several studies in vivo have demonstrated modifi-ations in the expression levels of members of the Bcl-2amily. Bax upregulation has been found following isch-mia (Krajewski et al., 1995; Ferrer et al., 2003) in asso-iation with Bcl-2 reduction in the infarct core (Matsushitat al., 1998). Bax is a potent proapoptotic molecule, whichriggers activation of terminal caspases and subsequentlyromotes neuronal cell death and has been described to

ncreased in the penumbra 3 h following the reperfusionFerrer et al., 2003). Using Western blotting and immuno-istochemistry, this work confirms the upregulation of theax protein in the penumbra cortical neurons of ischemicehicle-treated mice at 3 h following reperfusion and dem-nstrated that a single dose of ALA500 or RLZ4 strongly

educed Bax expression. Taken together with the fact that s

enumbra is the privileged site for apoptosis, both reduc-ion of TUNEL labeling and Bax reactivity suggest that theeuroprotection induced by ALA500 and RLZ4 involves aecrease of apoptotic pathways occurring in this area.evertheless, we cannot exclude a direct effect of theserugs on apoptosis.

In the second step, to design the best type of treatmente compared neuronal protection, but also long term sur-ival induced by a single dose of ALA500 or RLZ4 to threeay or three week therapies with ALA250 or RLZ2 or aombination therapy of ALA250�RLZ2 injected 2 h aftereperfusion. The results show that the greatest differencesetween paradigms appeared in the survival rate of theice. The model of focal ischemia used in this study is very

evere, since more than 75% of the ischemic mice treatedith vehicle died within the month of reperfusion. Interest-

ngly, three-week therapies with ALA250 or RLZ2 im-roved the survival rate (at one month) as compared withsingle dose of ALA500 or RLZ4. The best protection in

erms of survival rate (83.6%) was obtained with the cock-ail therapy of ALA250�RLZ2 injected 2 h after reperfu-ion, suggesting a combined effect of both drugs. Whene compare the different paradigms and analyze the re-ults obtained in terms of brain protection and mice sur-ival, it seems that salvaging cortical and caudate neuronsid not necessarily improve survival. This point can bexplained by the development of the ischemic edema (va-ogenic and cytotoxic), that we have not measured in thistudy. As mentioned in the Experimental Procedures, neu-ological deficit was assessed at 24 h post-ischemia. Miceith a neurological deficit above grade 4 (decreased levelf consciousness), were excluded from the study to ex-lude ischemic edema, one of the causes of clinical dete-ioration within the first 24 h. The clinical manifestations ofschemic brain swelling generally appear between 2 and 5ays after stroke and this propensity of ischemic brainissue to develop edema and swelling remains the majorause of death within a week. Then, cresyl violet staining,sed to count viable neurons and analyze cell survival in

schemic mice, was performed at three weeks on the ani-als that had survived, probably overcoming the deleteri-us effects of edema. In this context, we cannot excludehat the single combined therapy (ALA250�RLZ2) as wells the three daily injections (ALA250 or RLZ2) reduced theevelopment of edema and consequently reduced brainnd animal death.

Taken together, this study demonstrates that alpha-inolenic acid and riluzole induce a beneficial effect on thenfarct size, neurological score, neuronal survival and mor-ality rate of ischemic mice in a clinically relevant model oftroke. The neuroprotective properties of RLZ and PUFAsave already been described in other models of neurode-eneration (Malgouris et al., 1989; Pratt et al., 1992; Bar-eoud et al., 1996; Voskuyl et al., 1998; Ettaiche et al.,999; Lang-Lazdunski et al., 1999; Leaf et al., 1999; Lau-itzen et al., 2000). The main interest of this work has beeno compare the neuroprotective effects of RLZ or/and ALA,oth at short (24 h) and long terms (one month). Most

tudies using the MCAO model describe the beneficial

eirwatsaadoadctorcti

ednWdrtepgAaMTa1tT2otTMarcpk

lsmssa

At

FMtc

B

B

B

B

B

B

C

C

D

D

D

D

D

E

F

G

G

H

H

C. Heurteaux et al. / Neuroscience 137 (2006) 241–251250

ffect of a drug by the reduction of the infarction and themprovement of the neurological score at 24 h. As for theapid ischemic preconditioning protection, which disappearshen the reperfusion is extended to 7 days (Perez-Pinzon etl., 1997), it is known that neuronal protection can only beemporary with some therapies (Dietrich et al., 1993). In ourtudy, we analyzed not only the effectiveness of RLZ or/nd ALA on the infarct size and the neurological outcomet 24 h post-ischemia, but also the cell survival in theamaged cerebral structures and the death rate of micene month after ischemia. Looking at cell survival shortlyfter stroke, it appears that the various paradigms (singleose of ALA or RLZ, 3 days or three weeks therapy, orombined treatment) equally induce a remarkable protec-ion of the neurons. On the other hand, for animal survival,nly the 3 week therapy or combined treatment provides aeal improvement. This observation is interesting for laterlinical applications, since it shows that one protocol effec-ive in rescuing cellular populations does not necessarilymprove survival.

It is clear that multiple and complex pathways involvingxcitotoxicity, oxidative and nitrosative stress, periinfarctepolarization, inflammation and apoptosis-like mecha-isms cause cell death in stroke (Dirnagl et al., 1999;hite et al., 2000). In this context, it is reasonable to

evelop combination therapies targeting several of delete-ious mechanisms, simultaneously. In this work, we showhat a combined injection of RLZ and ALA not only inducedffective neuronal survival but more importantly an im-rovement of the animals’ survival. Several researchroups have analyzed the different targets of RLZ andLA. Both molecules are known to reduce hyperexcitabilitynd excitatory neurotransmission (Cheramy et al., 1992;artin et al., 1993; Doble, 1996; Lauritzen et al., 2000).hey both partially inhibit voltage-sensitive Na� channelsnd voltage-sensitive Ca�� channels (Benoit and Escande,991; Hebert et al., 1994; Huang et al., 1997). More impor-antly both activate the background TREK-1, TREK-2 andRAAK potassium channels (Lesage and Lazdunski, 1999,000; Duprat et al., 2000; Patel and Honore, 2001). On thether hand recent work has shown enhanced expression ofhe lipid-sensitive 2P-domain K� channels (the TREK/RAAK channel family) in cortex and hippocampus afterCAO and suggested a correlation between TREK channelsnd acute cerebral ischemia (Li et al., 2005). This idea waseinforced by recent data demonstrating that TREK-1hannels were required for the protective properties ofolyunsaturated fatty acids against global ischemia andainate-induced seizures (Heurteaux et al., 2004).

Our data on the neuroprotective effects of alpha-lino-enic acid and riluzole in a model of focal ischemia areupportive of the idea that a combined therapy with theseolecules may have important clinical implications for a

uccessful therapy of stroke in the future. They alsotrengthen the general opinion that polyunsaturated fattycids have beneficial effects in human health.

cknowledgments—This work was supported by the Centre Na-

ional de la Recherche Scientifique (CNRS) and the Paul Hamel

oundation. We are grateful to the Fondation de la Rechercheédicale for a fellowship to C. Laigle, to Professor M. Plotkine for

raining with the focal ischemia model, and to Dr. Lochlan Rash forareful reading of the manuscript.

REFERENCES

arneoud P, Mazadier M, Miquet JM, Parmentier S, Dubedat P, DobleA, Boireau A (1996) Neuroprotective effects of riluzole on a modelof Parkinson’s disease in the rat. Neuroscience 74:971–983.

ederson JB, Pitts LH, Tsuji M, Nishimura MC, Davis RL, BartkowskiH (1986) Rat middle cerebral artery occlusion: evaluation of themodel and development of a neurologic examination. Stroke17:472–476.

enoit E, Escande D (1991) Riluzole specifically blocks inactivated Nachannels in myelinated nerve fibre. Pflugers Arch 419:603–609.

ensimon G, Lacomblez L, Meininger V (1994) A controlled trial ofriluzole in amyotrophic lateral sclerosis. ALS/Riluzole Study Group.N Engl J Med 330:585–591.

londeau N, Lauritzen I, Widmann C, Lazdunski M, Heurteaux C(2002) A potent protective role of lysophospholipids against globalcerebral ischemia and glutamate excitotoxicity in neuronal cul-tures. J Cereb Blood Flow Metab 22:821–834.

londeau N, Widmann C, Lazdunski M, Heurteaux C (2001) Polyun-saturated fatty acids induce ischemic and epileptic tolerance. Neu-roscience 109:231–241.

harriaut-Marlangue C, Margaill I, Represa A, Popovici T, Plotkine M,Ben-Ari Y (1996) Apoptosis and necrosis after reversible focalischemia: an in situ DNA fragmentation analysis. J Cereb BloodFlow Metab 16:186–194.

heramy A, Barbeiro L, Godeheu G, Glowinski J (1992) Riluzoleinhibits the release of glutamate in the caudate nucleus of the catin vivo. Neurosci Lett 147:209–212.

ietrich WD, Busto R, Alonso O, Globus MY, Ginsberg MD (1993)Intraischemic but not postischemic brain hypothermia protectschronically following global forebrain ischemia in rats. J CerebBlood Flow Metab 13:541–549.

ing-Zhou L, Marchand-Verrecchia C, Croci N, Plotkine M, Margaill I(2002) L-NAME reduces infarction, neurological deficit and blood-brain barrier disruption following cerebral ischemia in mice. EurJ Pharmacol 457:137–146.

irnagl U, Iadecola C, Moskowitz MA (1999) Pathobiology of isch-aemic stroke: an integrated view. Trends Neurosci 22:391–397.

oble A (1996) The pharmacology and mechanism of action of ri-luzole. Neurology 47:S233–S241.

uprat F, Lesage F, Patel AJ, Fink M, Romey G, Lazdunski M (2000)The neuroprotective agent riluzole activates the two P-domain K�channels TREK-1 and TRAAK. Mol Pharmacol 57:906–912.

ttaiche M, Fillacier K, Widmann C, Heurteaux C, Lazdunski M (1999)Riluzole improves functional recovery following ischemia in the ratretina. Invest Ophthalmol Vis Sci 40:729–736.

errer I, Friguls B, Dalfo E, Justicia C, Planas AM (2003) Caspase-dependent and caspase-independent signalling of apoptosis in thepenumbra following middle cerebral artery occlusion in the adultrat. Neuropathol Appl Neurobiol 29:472–481.

insberg MD, Busto R (1989) Rodent models of cerebral ischemia.Stroke 20:1627–1642.

olanov EV, Reis DJ (1995) Contribution of cerebral edema to theneuronal salvage elicited by stimulation of cerebellar fastigial nu-cleus after occlusion of the middle cerebral artery in rat. J CerebBlood Flow Metab 15:172–174.

ankey GJ, Warlow CP (1999) Treatment and secondary preventionof stroke: evidence, costs, and effects on individuals and popula-tions. Lancet 354:1457–1463.

ata R, Maeda K, Hermann D, Mies G, Hossmann KA (2000) Evolu-tion of brain infarction after transient focal cerebral ischemia in

mice. J Cereb Blood Flow Metab 20:937–946.

H

H

H

K

L

L

L

L

L

L

L

L

M

M

M

P

P

P

S

T

U

V

W

C. Heurteaux et al. / Neuroscience 137 (2006) 241–251 251

ebert T, Drapeau P, Pradier L, Dunn RJ (1994) Block of the rat brainIIA sodium channel alpha subunit by the neuroprotective drugriluzole. Mol Pharmacol 45:1055–1060.

eurteaux C, Guy N, Laigle C, Blondeau N, Duprat F, Mazzuca M,Lang-Lazdunski L, Widmann C, Zanzouri M, Romey G, LazdunskiM (2004) TREK-1, a K(�) channel involved in neuroprotection andgeneral anesthesia. EMBO J 23:2684–2695.

uang C-H, Song J-H, Nagata K, Yeh JZ, Narahashi T (1997) Effectsof the neuroprotective agent riluzole on the high voltage-activatedcalcium channels of rat dorsal root ganglion neurons. J PharmacolExp Ther 282:1280–1290.

rajewski S, Mai JK, Krajewska M, Sikorska M, Mossakowski MJ,Reed JC (1995) Upregulation of bax protein levels in neuronsfollowing cerebral ischemia. J Neurosci 15:6364–6376.

ang-Lazdunski L, Blondeau N, Jarretou G, Lazdunski M, Heurteaux C(2003) Linolenic acid prevents neuronal cell death and paraplegiaafter transient spinal cord ischemia in rats. J Vasc Surg 38:564–575.

ang-Lazdunski L, Vaillant N, Widmann C, Heurteaux C, Lazdunski M(1999) Riluzole improves neurological function following severespinal cord ischemia. J Thorac Cardiovasc Surg 117:881–889.

auritzen I, Blondeau N, Heurteaux C, Widmann C, Romey G, Laz-dunski M (2000) Polyunsaturated fatty acids are potent neuropro-tectors. EMBO J 19:1784–1793.

eaf A, Kang JX, Xiao Y-F, Billman GE, Voskuyl RA (1999) Theantiarrhythmic and anticonvulsant effects of dietary N-3 fatty acids.J Membr Biol 172:1–11.

esage F, Lazdunski M (1999) Potassium channels with two P-domains. In: Potassium ion channels. Molecular structure, function,and diseases: Current topics in membranes, Vol. 46 (Kurachi Yet al., eds), pp 199–222. San Diego, CA: Academic Press.

esage F, Lazdunski M (2000) Molecular and functional properties oftwo pore domain potassium channels. Am J Physiol 279:793–801.

i ZB, Zhang HX, Li LL, Wang XL (2005) Enhanced expressions ofarachidonic acid-sensitive tandem-pore domain potassium chan-nels in rat experimental acute cerebral ischemia. Biochem BiophysRes Commun 327:1163–1169.

onga EZ, Weinstein PR, Carlson S, Cummins R (1989) Reversiblemiddle cerebral artery occlusion without craniectomy in rats. Stroke

20:84–91.

algouris C, Bardot F, Daniel M, Pellis F, Rataud J, Uzan A, BlanchardJ-C, Laduron PM (1989) Riluzole, a novel antiglutamate preventsmemory loss and hippocampal neuronal damage in ischemic ger-bils. J Neurosci 9:3720–3727.

artin D, Thompson MA, Nadler JV (1993) The neuroprotective agentriluzole inhibits release of glutamate and aspartate from slices ofhippocampal area. Eur J Pharmacol 250:473–476.

atsushita K, Matsuyama T, Kitagawa K, Matsumoto M, YanagiharaT, Sugita M (1998) Alterations of Bcl-2 family proteins precedecytoskeletal proteolysis in the penumbra, but not in infarct centresfollowing focal cerebral ischemia in mice. Neuroscience 83:439–448.

atel AJ, Honore E (2001) Properties and modulation of mammalian2P domain K� channels. Trends Neurosci 24:339–346.

erez-Pinzon MA, Xu GP, Mumford PL, Dietrich WD, Rosenthal M,Sick TJ (1997) Rapid ischemic preconditioning protects rats fromcerebral anoxia/ischemia. Adv Exp Med Biol 428:155–161.

ratt J, Rataud J, Bardot F, Roux M, Blanchard JC, Laduron PM,Stutzmann JM (1992) Neuroprotective actions of Riluzole in rodentmodels of global and focal cerebral ischemia. Neurosci Lett140:225–230.

harp FR, Lu A, Tang Y, Millhorn DE (2000) Multiple molecular pen-umbras after focal cerebral ischemia. J Cereb Blood Flow Metab20:1011–1032.

sujimoto Y, Shimizu S (2000) Bcl-2 family: life-or-death switch. FEBSLett 466:6–10.

nal-Cevik I, Kilinc M, Can A, Gursoy-Ozdemir Y, Dalkara T (2004)Apoptotic and necrotic death mechanisms are concomitantly acti-vated in the same cell after cerebral ischemia. Stroke 35:2189–2194. Epub 2004 Jul 2115.

oskuyl RA, Vreugdenhil M, Kang JX, Leaf A (1998) Anticonvulsanteffect of polyunsaturated fatty acids in rats, using the corticalstimulation model. Eur J Pharm 341:145–152.

hite BC, Sullivan JM, DeGracia DJ, O’Neil BJ, Neumar RW, Gross-man LI, Rafols JA, Krause GS (2000) Brain ischemia andreperfusion: molecular mechanisms of neuronal injury. J Neurol Sci

179:1–33.

(Accepted 31 August 2005)(Available online 14 November 2005)