Effect of carbamazepine and gabapentin on excitability in the trigeminal subnucleus caudalis of neonatal rats using a voltage-sensitive dye imaging te

Transcription

1 Effect of carbamazepine and gabapentin on excitability in the trigeminal subnucleus caudalis of neonatal rats using a voltage-sensitive dye imaging technique Akiko Matsumoto 1, Hirofumi Arisaka 1, Yuki Hosokawa 2, Shigeki Sakuraba 1, Takeo Sugita 1, Nobuo Umezawa 1, Yuki Kaku 3, Kazu-ichi Yoshida 1, Shun-ichi Kuwana 4 * 1 Division of Anesthesiology, Department of Clinical Care Medicine, Kanagawa Dental College, Yokosuka City, Kanagawa , Japan 2 Department of Anesthesiology, Kitasato University School of Medicine, Sagamihara City, Kanagawa , Japan 3 Center for Medical Sciences, Ibaraki Prefectural University of Health Sciences, Ibaraki-ken , Japan 4 Faculty of Health Sciences, Uekusa Gakuen University, Ogura-cho, Wakaba-ku, Chiba City , Japan Equal contributors *Correspondence: s-kuwana@uekusa.ac.jp 1

2 addreses: 2

3 Abstract Background Theantiepilepticdrugscarbamazepineandgabapentinareefectiveintreating neuropathicpainandtrigeminalneuralgia.inthepresentstudy,toanalyzetheefectsof carbamazepineandgabapentinonneuronalexcitationinthespinaltrigeminal subnucleuscaudalis(sp5c)inthemedulaoblongata,werecordedtemporalchangesin nociceptiveaferentactivityinthesp5coftrigeminalnerve-atachedbrainstemslicesof neonatalratsusingavoltage-sensitivedyeimagingtechnique. Results Electricalstimulationofthetrigeminalnerverootletevokedchangesinthefluorescence intensityofdyeinthesp5c.theopticalsignalswerecomposedoftwophases,afast componentwithasharppeakfolowedbyalong-lastingcomponentwithaperiodof morethan500ms.thisevokedexcitationwasnotinfluencedbyadministrationof carbamazepine(10,100and1000µm)orgabapentin(1and10 µm),butwasincreased byadministrationof100µmgabapentin.thisevokedexcitationwasincreasedfurther inlow Mg 2+ (0.8mM)conditions,andthisefectoflow Mg 2+ concentrationwas antagonizedby30µmdl-2-amino-5-phosphonopentanoicacid(ap5),a 3

4 N-methyl-D-aspartate(NMDA)receptorblocker.Theincreasedexcitationinlow Mg 2+ conditionswasalsoantagonizedbycarbamazepine(1000µm)andgabapentin (100µM). Conclusions Carbamazepineandgabapentindidnotdecreaseelectricalyevokedexcitationinthe Sp5cincontrolconditions.Furtherexcitationinlow Mg 2+ conditionswasantagonized bythenmdareceptorblockerap5.carbamazepineandgabapentinhadsimilarefects toap5onevokedexcitationinthesp5cinlow Mg 2+ conditions.thus,weconcluded thatcarbamazepineandgabapentinmayactbyblockingnmdareceptorsinthesp5c, whichcontributestoitsanti-hypersensitivityinneuropathicpain. Keywords Carbamazepine,Gabapentin,Ratbrainstem,Spinaltrigeminalnucleus, Voltage-sensitivedyeimaging 4

5 Background Theantiepilepticagentscarbamazepineandgabapentinareefectiveagainst neuropathicpainandtrigeminalneuralgia[1,2].however,theactionofthese antiepilepticdrugsonneuronalactivityinthebrainmaynotbesimple.alargebodyof evidenceindicatesthatcarbamazepinemayinteractwithdiferenttypesofionchannels andsynaptictransmision[3,4].themoleculartargetsforionchannelshavegeneraly beenvoltage-gatedna + channels[5],ca 2+ channels[6]andk + channels[7].an increasingnumberoffindingsindicatethatcarbamazepineinducestheinhibitionof glutamaterelease[8],inhibitionofanadenosinereceptor[9]andmodulationof neuromodulatorlevels,suchasthoseofserotonin,dopaminandcyclicadenosine monophosphate(amp)[3].inaddition,theefectsofgabapentinonneuralactivityare notexplainedbyasinglemechanism[10,11].althoughgabapentinisastructural analogueofγamino-butyricacid(gaba),ithasnoafinityforgabareceptors.the maintargetofgabapentinissynaptictransmision,whereitinhibitsvoltage-gatedca 2+ channelsinthepresynapticmembrane,whichinhibitsthereleaseofglutamateand substancep[12].recentevidencesuggeststhatotheractionsofgabapentininclude inhibitionofglutamatergicn-methyl-d-aspartate(nmda)receptors[13]andan increaseingabarelease[14]. 5

6 Tostudytheneuralmechanismsoftrigeminalneuralgia,trigeminalnerve-atached brainstempreparationsfromneonatalratsareused[15,16].inelectrophysiological studies,activity-dependentneuronalhyperexcitability,so-caledcentralsensitization, hasbeenreportedinthespinaltrigeminalsubnucleuscaudalis(sp5c),whichreceives nociceptiveinformationfromtheorofacialarea.furthermore,suchstudieshaveshown thatnmdareceptorscontributesubstantialytopolysynaptictransmisioninthesp5c andtolong-termpotentiation. ToanalyzethespatialdynamicsofneuronalexcitationpropagationintheSp5c,it isposibletouseopticalimaginganalysisandvoltage-sensitivedyes[17,18].a previousstudyusingtrigeminalnerve-atachedbrainstemslicesfrompostnatalrats showedthatsynaptictransmisionviaunmyelinatedaferentsinthesp5cwasmediated substantialybynmdareceptors[19].inthepresentstudy,weexaminedtheefectsof carbamazepineandgabapentinonexcitabilityinthesp5cofneonatalratsusingan opticalimagingtechnique.furthermore,weconfirmedthecontributionofnmda receptorstoenhancedexcitabilityinthesp5cinlow Mg 2+ concentrationconditions. Results InfluenceofcarbamazepineonevokedexcitationintheSp5c 6

7 Theinfluenceofdrugadministrationonevokedexcitationwasexaminedusing trigeminalnerve-atachedbrainstemsagitalslicepreparations,asshowninfigure1. Figure2showstheinfluenceofcarbamazepineatdiferentconcentrations(10,100or 1000µM)onevokedexcitationintheSp5c.Duringsuperfusionwithmock cerebrospinalfluid(csf)(fig.2a),theopticalsignalswerecomposedoftwophases,a fastcomponentwithasharppeakfolowedbyalong-lastingcomponentwithaperiod ofmorethan500ms.thetimedelayfromstimulationtothepeakofthefirst componentwas36.8±8.3ms(n=6).thedistancefromthenerverootlettothesp5c wasapproximately4.0mm.therefore,theconductionvelocitywasapproximately 0.11m/s. Theintensityandtimecourseoftheopticalsignaldidnotaltersignificantlyafter switchingtosuperfusionwith1000µmcarbamazepine-containingmockcsf(fig.2b). Theleft-handpanelsinFigs.2Aand2Bshowfluorescentsignalimagesat165msafter electricalstimulationduringsuperfusionwithmockcsfandwith1000µm carbamazepine-containingmockcsf,respectively.theright-handpanelsinfigs.2a and2bshowtime-coursesoffluorescentsignalchangesduringsuperfusionwithmock CSF(Fig.2A)andwith1000µMcarbamazepine-containingCSF(Fig.2B), respectively. Whenelectricalstimulationwasappliedinthepresenceof1000µM 7

8 carbamazepine,thetime-courseofsp5cexcitementshowedasharppeakafter stimulationfolowedbyaslowdecline,similartothatinthecontrol.toanalyzesignal amplitudesduringsuperfusionwith10,100and1000µmcarbamazepine,fluorescence signalamplitudewasindicatedasthepercentamplitudeofthecontrolconditions.the peakamplitudeofthefirstcomponentseemedtodecreaseslightlyto96.0±12.5% duringsuperfusionof1000µmcarbamazepine,butthediferencewasnotsignificant (Fig.2C).Thepeakamplitudeofthefirstcomponentwasnotafectedbyany concentrationofcarbamazepine(fig.2c).toevaluatetheefectofcarbamazepineon thelong-lastingcomponent,wemeasuredtheevokedsignalat165msand385msafter stimulation(figs.2dand2e),butthesignalamplitudeofthelong-lastingcomponent wasalsounafectedbycarbamazepine(figs.2dand2e). InfluenceofgabapentinonevokedexcitationintheSp5c Thedose responserelationshipofgabapentinwiththeevokedexcitationwasexamined usingtrigeminalnerve-atachedbrainstemsagitalslicepreparations(fig.3). When electricalstimulation wasappliedinthepresenceof1or10µmgabapentin,the time-courseofsp5cexcitementshowedasharppeakafterstimulationfolowedbya slowdecline,similartothatinthecontrolconditions.thesignalamplitudesofthefast andlong-lastingcomponentswerenotafectedby1or10µmgabapentin(figs.3c,3d 8

9 and3e).ofnote,100µmgabapentinsignificantlyincreasedthepeakamplitudeofthe fastcomponentto178±5.66%(p<0.01).signalamplitudesat165msand385msafter stimulationincreasedto179±23.5%(fig.3d)and240±31.2%,respectively(fig.3e). Thus,100µMgabapentininducedfurtherexcitationintheelectricaltrigeminalnerve rootstimulation-inducedexcitementofthesp5c. Influence oflow Mg 2+ concentration and an NMDA antagonist on evoked excitationinthesp5c Theinfluenceoflow Mg 2+ concentration(0.8mm)onevokedexcitationinthesp5c wasobservedusingtrigeminalnerve-atachedbrainstemsagitalslicepreparations (Fig.4). Whenelectricalstimulationwasappliedduringsuperfusionwithlow Mg 2+ concentrationsolution,thepeakofthefastcomponentafterstimulationincreasedto177 ±30.6%.Furthermore,thesignalamplitudeat165msand385msafterelectrical stimulationincreasedmarkedlyover250%duringsuperfusionwithlow Mg 2+ concentrationsolution(fig.4),suggestingthatthelow Mg 2+ concentrationincreased electricaltrigeminalnerverootstimulation-inducedexcitementinthesp5c.the increaseinthelong-lastingcomponentwithlow Mg 2+ concentrationtreatmentwas greaterthanthatofthefastcomponent.apreviousstudyperformedbytakuma[19] showedthatincreasedexcitationinthesp5cwithlow Mg 2+ concentrationtreatmentwas 9

10 partialyantagonizedbythenmdaantagonistap5. Wealsoexaminedadditional superfusionwith30µmap5inlow Mg 2+ conditions.figure4cshowedthat administrationofap5atenuatedtheevokedexcitementinthesp5candrestoreditto thecontrollevel.thepeakamplitudewas126±19.2%duringsuperfusionwith30µm AP5inlow Mg 2+ conditions(fig.4d).theamplitudeofthelong-lastingcomponent wasalsorestoredbyadditionalsuperfusionwith30µmap5inlow Mg 2+ conditions. Theamplitudesat165msand385msafterstimulationdecreasedto104±21.4%(Fig. 4E)and91.4±31.5%(Fig.4F),respectively. Influence of carbamazepine on evoked excitationinthe Sp5cinlow Mg 2+ conditions Figure5showstheefectofcarbamazepineinlow Mg 2+ conditionsonevoked excitationinthesp5c.superfusionwithalow Mg 2+ concentrationsolutionpotentiated theevokedexcitation.theefectoftheseconditionsonthelong-lastingcomponent (205±67.3%at165msand237±40.8%at385msafterstimulation)wasmoremarked thanthatonthefastcomponent(162±63.8%).additionaladministrationof1000µm carbamazepineinlow Mg 2+ conditionsatenuatedtheevokedexcitationtothelevelof thecontrolconditions(fig.5).thepeakamplitudewas130±40.2%duringsuperfusion (Fig.5D).Theamplitudeofthelong-lastingcomponentwasalsorestoredbyadditional 10

11 superfusionwithcarbamazepineinlow Mg 2+ conditions.theamplitudesat165msand 385msafterstimulationdecreasedto113±58.5%(Fig.5E)and119±60.7%(Fig.5F), respectively.theefectofcarbamazepineinlow Mg 2+ conditionswassimilartothatof AP5. InfluenceofgabapentinonevokedexcitationintheSp5cinlowMg 2+ conditions Figure6showstheefectofgabapentinonevokedexcitationintheSp5cinlow Mg 2+ conditions.superfusionwithlow Mg 2+ concentrationsolutioncausedpotentiationofthe evokedexcitation.theefectoflow Mg 2+ concentrationsolutiononthelong-lasting component(148±14.5%at165msand186±21.5%at385msafterstimulation)was moremarkedthanthatonthefastcomponent(128±23.0%).additionaladministration of100µmgabapentininlow Mg 2+ conditionsatenuatedtheevokedexcitationtothe levelofthecontrolconditions(fig.6).thepeakamplitudewas119±22.2%during superfusion(fig.6d).theamplitudeofthelong-lastingcomponentwasalsorestored byadditionalsuperfusionwithcarbamazepineinlow Mg 2+ conditions.amplitudesat 165 msand385 msafterstimulationdecreasedto97.8±10.7%(fig.6e)and115± 12.4%(Fig.6F),respectively.Theefectsofgabapentininlow Mg 2+ conditionswas similartothoseofap5andcarbamazepine. 11

12 Discussion Applyingthevoltage-sensitivedyeimagingmethodwithisolatedneonatalratbrainstem slicesalowedelectricaltrigeminalnervestimulation-inducedexcitementofthesp5cto betemporalyandspatialyvisualized.soleadministrationofcarbamazepineandoflow concentrationsofgabapentindidnotafectevokedexcitationinthesp5c,although administrationofahighconcentrationofgabapentinemphasizedtheevokedexcitation inthesp5c.superfusionwithalow Mg 2+ solutionpotentiatedtheevokedexcitation. Theefectoflow Mg 2+ conditionsonthelong-lastingcomponentwasmoremarkedthan thatonthefastcomponent.additionaladministrationofap5,carbamazepineor gabapentininlow Mg 2+ conditionsatenuatedtheevokedexcitationtothelevelofthe controlconditions,showingthattheemphasizedexcitementbylow Mg 2+ concentration wasantagonizedbyap5,carbamazepineorgabapentin. Single-pulsestimulationofthetrigeminalnerverootletinducedanoptical responseinthesp5c.theopticalsignalswerecomposedoftwophases,afast componentwithasharppeakfolowedbyalong-lastingcomponentwithaperiodof approximately500ms.thesefindingswereconsistentwiththoseoftakuma[19],who foundthatthespatiotemporalpropertiesofopticalsignalswerecorelatedwiththoseof thefieldpotentialrecordingsbyusingsimilarpreparationstoours.theopticalsignalof 12

13 thefastcomponentwaseliminatedbytreatmentwith 6-cyano-nitro-quinoxaline-2,3-dione(CNQX).Thelong-lastingcomponentincreasedin amplitudeinlow Mg 2+ conditionsbutwassignificantlyreducedbyap5.inan electrophysiologicalstudyusingwhole-celpatch-clamptechniques,onoderaetal.[16] reportedthatstimulationofthemandibularnerveat0.03hzevokedcompound excitatorypostsynapticpotentials(epsps)ofneuronsinthesp5c.compoundpotentials consistentwithmonosynapticepsps,whichhadahighthresholdandwithpolysynaptic EPSPs,wasatenuatedbyhighfrequency(33 50Hz)stimulation.Inlow Mg 2+ conditions,thefastmonosynapticepspcomponentwasabolishedbysimultaneous applicationofcnqxandap5,andtheslowpolysynapticepspwaslargelyatenuated byapplicationofap5.inthepresentstudy,weconfirmedtheresultthatap5atenuated thefastcomponentandlargelyatenuatedthelong-lastingcomponentinlow Mg 2+ conditions.theresultthatenhancedexcitationinlow Mg 2+ conditionswasatenuated byapplicationofap5suggestedthatlow Mg 2+ conditionsinducedtheactivationof NMDAreceptorsinthesecondaryneuronsoftheSp5c.Furthermore,wedemonstrated thatcarbamazepineandgabapentinhadsimilarefectstoap5onevokedexcitationin thesp5cinlow Mg 2+ conditions.thisresultstronglysuggestedthatcarbamazepineand gabapentinactasantagonistsofnmdareceptors.therefore,blockageofnmda 13

14 receptorsofsecondaryneuronsinthesp5cmaycontributetotheclinicalefectivenes ofcarbamazepineandgabapentin. Inhibitionofionchannelsandsynaptictransmisionhavebeenreportedasthe pharmacologicalactionsofcarbamazepineonthenervoussystem(seeintroduction).in controlconditions,soleapplicationofcarbamazepinedidnotafectevokedexcitement inthesp5c,suggestingthatcarbamazepinedidnotmodulatethenormaltransductionof actionpotentialandsynaptictransmision. WhenexcesiveactivationoftheNMDA receptorisinducedinthepathophysiologyofseveralneurologicalconditions,suchas trigeminalneuropathicpain,applicationofcarbamazepinewasefectivebecauseofa blockintheactivationofnmdareceptors[20,21]. Regardingthepharmacologicalactionofgabapentinonthenervoussystem,the inhibitionofactivepotentialinductionthroughbindingtotheα2δsubunitof voltage-dependentca 2+ channelshasbeenreported[22].becausegabapentinhasahigh afinityfortheα2δsubunit,itwasconsideredtoinhibitthereleaseofneurotransmiters byinhibitingca 2+ ioninflux[11].gabapentinhasalsobeenshowntoinducethe modulationofothertargetsincludingnmdareceptors.inthepresentstudy,we suggestedthattheefectivesiteofgabapentinwasnmdareceptorsinthesp5c. Oneunanticipatedfindingwasthatsoleapplicationof100µMgabapentin 14

15 emphasizedtheevokedexcitationincontrolconditions.petrofetal.[14]reportedthat gabapentinincreasedgabalevelsinthebrainsofepilepticpatients.denselypacked GABAergicneuronsintheSp5chavebeendescribedpreviously[23,24].Because GABAergictransmisioninvariousbrainregionsofimmatureanimalsisexcitatoryand inhibitorysynapticpotentialsappearrelativelylaterindevelopment[25],weshould considerthatnotonlynon-gabaergicbutalsogabaergicinterneuronsinthisregion wereexcited.itmaybeposiblethatahighconcentrationofgabapentininducedgaba release,andgabaactsasanexcitatorytransmiterinthesp5c.however,application ofbicuculinefurtherincreasedtheevokedexcitationafterapplicationof100µm gabapentin(unpublisheddata),suggestingthatgabaergictransmisionwasinhibitory inourpreparations.therefore,theposibilitythatgabaergictransmisioninthesp5c isexcitatorymayberuledout.toclarifythemechanismsofexcitationbygabapentin, furtherstudiesarerequired. Inthepresentstudy,weestimatedtheconductionvelocityasapproximately 0.11m/sforthepeakofthefastcomponent.Thisvaluewassimilartothevalue (~0.18m/s)calculatedbypreviousopticalmeasurements[19].Thesevalueswereslow comparedwithapreviouselectrophysiologicalmethod,inwhichtheconduction velocitywascalculatedtobe0.37m/s[16].conductionvelocitiesobtainedbyoptical 15

16 andelectrophysiologicalstudiesfalintotherangeofthatofc-fibers.ontheotherhand, trigeminalsensoryneuronschangedelectrophysiologicalpropertiesduringearly postnatalmaturation.cabanesetal.[26]showedthattrigeminalganglionneuronsin neonatalmicehaduniformlyslowconductionvelocitiesandseparatedaccordingto theirconductionvelocityintoaδandcneuronsduringthe3-weekpostnatal developmentperiod.thus,itisdificulttoshowwhichaxontypeofatrigeminalnerve wasmainlystimulatedinthepresentstudy. Inconclusion,antiepilepticdrugscarbamazepineandgabapentindidnotdecrease electricalyevokedexcitationinthesp5cincontrolconditions.furtherexcitationin low Mg 2+ conditionswasatenuatedbythenmdareceptorantagonistap5. CarbamazepineandgabapentinhadsimilarefectstoAP5onevokedexcitationinthe Sp5cinlow Mg 2+ conditions.thus,weconcludedthatcarbamazepineandgabapentin actbyblockingnmdareceptorsinthesp5c,whichcontributestoits anti-hypersensitivityactioninneuropathicpainandtrigeminalneuralgia. Methods Preparations 16

17 Al procedureswereconductedinaccordancewiththeguidelinesoftheuekusagakuen UniversityLaboratoryAnimalCareandUseCommitee.Datawereobtainedfrom54 neonatal Wistarrats(2 3daysold).Theisolationofbrainstem-spinalcordpreparations hasbeendescribedindetailpreviously[27].inbrief,ratsweredeeplyanesthetizedwith diethyletherandthebrainstemwasisolatedinadisectingchamberatroom temperature.thechamberwasfiledwithmockcsfequilibratedwithagasmixture (5%CO 2 ino 2 ;ph7.4).thecompositionofthemockcsfwasasfolows(inmm): NaCl,126;KCl,5;CaCl 2,2; MgSO 4,2;NaH 2 PO 4,1.25;NaHCO 3,26andglucose,30. Thecerebrumwasquicklyremovedbytransectionattheupperborderoftheinferior coliculus.eachtrunkofthebilateraltrigeminalnervesthatrunthroughthecraniobasal bonewasisolatedtoalengthof1mm,enablingittobepuledintoasuctionelectrode. Subsequently,thetrigeminalnerve-atachedbrainstem-spinalcordwascutcaudalyat thelevelofthec3roots(fig.1a).furthermore,theisolatedmedulawassectioned sagitalyat1 1.5mmlateralfromthemidsagitalplanewithahandmadeslicer(Fig. 1B).Thetrigeminalnerve-atachedbrainstemsagitalslicewasplacedinarecording chamber(volume1.0ml)withthemedialsideupandcontinuouslysuperfused(flow 4 6mL/min)at26 CwithoxygenatedmockCSF. 17

18 Voltage-sensitivedyeimaging Thevoltage-sensitivedyeimagingtechniquehasbeendescribedindetailpreviously [28].Inbrief,forstaining,preparationswerekeptfor30mininmockCSFcontaining thevoltage-sensitivedyedi-4-anepps(7.5mg/mlin0.1%dmso, MolecularProbes, Eugene,Oregon,USA),beforebeingkeptforatleast30mininnormalmockCSF. Afterstaining,excesdyewasremovedbysuperfusionofthepreparationwithdye-free solution.after30minofwashing,opticalimaginganddataanalysiswereperformed usinga MiCAM02hardwareandsoftwarepackage(BrainVision,Tokyo,Japan).For opticalimaging,weusedafixed-stageuprightfluorescencemicroscope(measurescope UM-2,Nikon,Tokyo,Japan)withalowmagnificationobjectivelens(XLFluor4 /340, Olympus,Tokyo,Japan)andahigh-resolution MiCAM02camera. Torecordthevoltage-sensitivedyesignals,weusedlightfroma150 Whalogen lampcontroledbyanelectromagneticshuter(orielinstruments,stratford,usa). Changesinfluorescenceofthedyeweredetectedbythecamerathrougha nm excitationfilter,adichroicmiror,anda590nmabsorptionfilter(mbe1405,nikon). Thecameracapturedimagesof88 60pixels,andthesizeoftheareawas mm.Opticalsignalsfrom3 3pixels(approximately0.03mm 2 )wereaveragedand areshowedintheimages. 18

19 Totalframeacquisitionwassetto511.Samplingtimewas2.2ms/frame;therefore,the totalrecordingtimewas1124.2ms.neuronalactivitywasevokedbysquarepulse electricalstimuli(1.0ms, mA)deliveredtothetrigeminalnerverootletviaaglas suctionelectrode.acquisitionwastriggeredbytheelectricalstimulus.thetriggersignal wasactivatedafterone-quarterofthetotalrecordingtime,corespondingto284ms afterstartingacquisitionwhenwesetthetotalacquisitiontimeto1124.2ms.signal amplitudewasnormalizedusingthedf/fmethod,wherefisthetotalfluorescentsignal anddfcorespondstothechangeinfluorescenceobservedfolowingevoked modificationofthemembranepotential.toimprovethesignal-to-noiseratio,we averagedsignalsdetectedin10consecutivetrialsat0.3hz.toanalyzetheintensityof thesignals,wemeasuredthepeakamplitude(at30 40msafterstimulation)and amplitudesat40%(at165msafterstimulation)andat60%(at385msafter stimulation)ofthetotalrecordingtime(1124.2ms). Measuringpointsat165msand 385msafterstimulationwereselectedarbitrarily. Drugadministration Carbamazepine(SigmaAldrich,SaintLouis, MO,USA)wasaddedtoperfusedmock CSFat10,100and1000µM.Gabapentin(SigmaAldrich)wasaddedtoperfusedmock 19

20 CSF at1.0,10and100µm.nmdareceptorantagonist DL-2-amino-5-phosphonopentanoicacid(AP5,SigmaAldrich)wasaddedtoperfused mockcsfat30µm.theseconcentrationsweredeterminedfrompreliminary experimentsandpreviousstudies[16,29].opticalrecordsusingelectricalstimulation weretaken20minafterthestartofsuperfusionwithcontrolmockcsfandweretaken 20minafterswitchingtodrug-containingmockCSF.Toinduceactivationofthe NMDAreceptor,weusedlow Mg 2+ concentrationsolution(inmm):nacl,126;kcl,5; CaCl 2,2.6; MgSO 4,0.8;NaH 2 PO 4,1.25;NaHCO 3,26andglucose,30.Intheseseriesof experiments,opticalrecordsusingelectricalstimulationweretaken20minafterthe startofsuperfusionwithcontrolmockcsfandweretaken20minafterswitchingto low Mg 2+ concentrationsolution,andthen,taken20minafterswitchingtolow Mg 2+ solutioncontaining30µmap5,1000 µmcarbamazepineor100 µmgabapentin. Dataanalysis OpticalsignalamplitudesobtainedbeforesuperfusionwithmockCSFcontainingdrugs wasdefinedasthecontrolvalue.thelevelofstatisticalsignificanceforthediference betweenthemeanvalueofeachvariableobtainedduringapplicationofdiferent concentrationsofdrugwasconductedbyanovafolowedbypair-wisecomparisons 20

21 usingthetukey-kramermethodformultiplecomparisonsasindicated.alstatistical analyseswereconductedusingstatcel(omspublisher,japan).alvalueswerereported asmean±sd,andalp values<0.05wereconsideredsignificant. Competinginterests Theauthorsdeclaredthattheyhavenocompetinginterests. Authors'contributions AMandHAcariedoutexperiments,colectedandanalyzeddata,anddraftedthe manuscript.amandhacontributedequalytothisstudy.yhcariedoutexperiments andconductedthestudy.ssandkyhelpeddesignandconductthestudy,anddrafted themanuscript.tsandnucariedoutexperimentsandcolectedandanalyzeddata.sk designedandsupervisedthisstudy,explainedthedata,anddraftedthemanuscript.al authorsreadandapprovedthefinalmanuscript. Acknowledgements ThisworkwassupportedbyJSPSKAKENHIgrantnumbers , ,by thescienceresearchpromotionfundofthepromotionand MutualAidCorporationfor 21

22 PrivateSchoolsofJapan,andbytheScienceResearchFundoftheUekusaGakuen University. Authordetails 1 DivisionofAnesthesiology,DepartmentofClinicalCare Medicine,KanagawaDental Colege,YokosukaCity,Kanagawa ,Japan. 2 DepartmentofAnesthesiology, KitasatoUniversitySchoolof Medicine,SagamiharaCity,Kanagawa ,Japan. 3 Centerfor MedicalSciences,IbarakiPrefecturalUniversityofHealthSciences, Ibaraki-ken ,Japan 4 FacultyofHealthSciences,UekusaGakuenUniversity, Ogura-cho, Wakaba-ku,ChibaCity ,Japan References 1.LemosL,AlegriaC,OliveiraJ, MachadoA,OliveiraP,AlmeidaA: Pharmacologicalversusdecompresionapproachesfortreatmentoftrigeminal neuralgia:clinicaloutcomesanddirectcosts.jpainres2011,4: LemosL,FontesR,FroresS,OliveiraP,AlmeidaA:Efectivenesofthe asociationbetweencarbamazepineandperipheralanalgesicblockwith ropivacainetorthetreatmentoftrigeminalneuralgia.jpainres2010, 22

23 3: WorleyPF,BarabanJM.:Siteofanticonvulsantactiononsodiumchannels: autoradiographicandelectrophysiologicalstudiesinratbrain.procnatlacad SciUSA.1987,84: AmbrósioAF,Soares-Da-SilvaP,CarvalhoCM,CarvalhoAP:Mechanismsof actionofcarbamazepineanditsderivatives,oxcarbazepine,bia2-093,and BIA2-024.NeurochemRes 2002,27: UemuraY,FujitaT,OhtsuboS,HirakawaN,SakaguchiY,KumamotoE: Efectsof variousantiepilepticsusedtoaleviateneuropathicpainoncompoundaction potentialinfrogsciaticnerves:comparisonwiththoseoflocalanesthetics. BiomedResInt2014,ArticleID540238,9pages 6.TodorovicSM,LingleCJ:PharmacologicalpropertiesofT-typeCa2+currentin adultratsensoryneurons:efectsofanticonvulsantandanestheticagents. Neurophysiol 1998,79: ZonaC,TancrediV,PalmaE,PironeGC,Avoli M:Potasiumcurrentsinrat corticalneuronsincultureareenhancedbytheantiepilepticdrug carbamazepine.canjphysiolpharmacol1990,68: Okada M,KawataY, MizunoK, WadaK,KondoT,KanekoS:Interactionbetween 23

24 Ca 2+,K +,carbamazepineandzonisamideonhippocampalextracelular glutamatemonitoredwithamicrodialysiselectrode.brjpharmacol 1998, 124: BiberK,FiebichBL,Gebicke-HärterP,vanCalkerD:Carbamazepine-induced upregulationofadenosinea1-receptorsinastrocyteculturesafectscouplingto thephosphoinositolsignalingpathway.neuropsychopharmacology1999, 20: ChengJK,ChiouLC: Mechanismsoftheantinociceptiveactionofgabapentin.J PharmacolSci2006,100: KukkarA1,BaliA,SinghN,JaggiAS:Implicationsandmechanismofactionof gabapentininneuropathicpain.archpharmres2013,36: SutonKG, MartinDJ,PinnockRD,LeeK,ScotRH:Gabapentininhibits high-thresholdcalciumchannelcurrentsinculturedratdorsalrootganglion neurons.brjpharmacol 2002,135: KimYS,ChangHK,LeeJW,SungYH,KimSE,Shin MS,YiJW,ParkJH,KimH, KimCJ:ProtectiveefectofgabapentinonN-methyl-D-aspartate-induced excitotoxicityinrathippocampalca1neurons.jpharmacolsci2009,109:

25 14.PetrofOA,HyderF,RothmanDL, MatsonRH:Efectsofgabapentinonbrain GABA,homocarnosine,andpyrrolidinoneinepilepsypatients.Epilepsia2000, 41: Hamba M,OnoderaK,TakahashiT:Long-termpotentiationofprimaryaferent neurotransmisionattrigeminalsynapsesofjuvenilerats.eurjneurosci2000, 12: OnoderaK,Hamba M,TakahashiT:Primaryaferentsynapticresponses recordedfromtrigeminalcaudalneuronsinamandibularnerve-brainstem preparationofneonatalrats.jphysiol2000,524: SeoK,FujiwaraN,TakeuchiK, MaedaT,SomeyaG:Opticalimagingof excitationpropagationevokedbystimulationtothetrigeminalcaudalis. Neuroreport 2001,12: SeoK,FujiwaraN,TakeuchiK, MaedaT,SomeyaG:Repetitiveaferent stimulationpropagatesexcitationinthetrigeminalcaudalis.neuroreport 2003, 14: TakumaS:Efectofneonatalcapsaicintreatmentonneuralactivityinthe medularydorsalhornofneonatalratsevokedbyelectricalstimulationtothe trigeminalaferents:anoptical,electrophysiological,andquantitativestudy. 25

26 BrainRes2001,906: LandmarkCJ:Targetsforantiepilepticdrugsinthesynapse.MedSciMonit2007, 13:RA HahmTS1,AhnHJ,RyuS,Gwak MS,ChoiSJ,KimJK,YuJM:Combined carbamazepineandpregabalintherapyinaratmodelofneuropathicpain.brj Anaesth2012,109: D MeloR,DickensonAH:Spinalcordmechanismsofpain.BrJAnaesth 2008, 101: KuwanaS,TsunekawaN,YanagawaY,OkadaY,KuribayashiJ,ObataK: ElectrophysiologicalandmorphologicalcharacteristicsofGABAergic respiratoryneuronsinthemousepre-bötzingercomplex.eurjneurosci 2006, 23: ViggianoA, Monda M,ViggianoA,Chiefari M,AurilioC,DeLucaB:Evidence thatgabaergicneuronsinthespinaltrigeminalnucleusareinvolvedinthe transmisionofinflammatorypainintherat:amicrodialysisand pharmacologicalstudy.eurjpharmacol 2004,496: Ben-AriY:TheGABAexcitatory/inhibitorydevelopmentalsequence. Neuroscience2014,279:

27 26.CabanesC,LópezdeArmentia M,VianaF,BelmonteC:Postnatalchangesin membranepropertiesofmicetrigeminalganglionneurons.jneurophysiol2002, 87: KuwanaS,OkadaY,NatsuiT:Efectsofextracelularcalciumandmagnesium oncentralrespiratorycontrolinthebrainstem-spinalcordofneonatalrat. BrainRes1998,786: OkadaY,ChenZ,YoshidaH,KuwanaS,Jiang W, MaruiwaH: Opticalrecording oftheneuronalactivityinthebrainstem-spinalcord:applicationofa voltage-sensitivedye.advexpmedbiol 2001,499: D MeloR,DickensonAH:Spinalcordmechanismsofpain.BrJAnaesth 2008, 101:

28 Figurelegends Figure1Preparationsusedfor membranepotentialimaging A:Isolatedtrigeminalnerve-brainstempreparation.Thedorsalsideisshownatthefront, andthesagitalsectionalsurfaceisshownbyhatching.b:trigeminalnerve-atached brainstemsagitalslicepreparation.thepreparationwasplacedwiththesagital sectionalsurfaceupwardsformeasurement.electricstimulationwasappliedbysucking thetrigeminalnerverootwithasuctionelectrodeforalpreparations. Sp5c:Spinaltrigeminalsubnucleuscaudalis,Vrootlet:Trigeminalnerverootlet. Figure2EfectofcarbamazepineonevokedexcitationintheSp5c A:RecordingmadeduringsuperfusionwithcontrolmockCSF.Theleftpanels representneuralactivity,whichisindicatedaschangeinfluorescenceintensityusing pseudocolor.therightpanelshowsthetimecourseofthesignalchangeandtimeof electricalstimulation.opticalimageintheleftpanelshows165msafterstimulation,as indicatedbytheverticaldotedlineintherightpanel.twohorizontallinesintheimage aretheanchorsforslicepreparation.b:recordingmadeduringsuperfusionwithmock CSFcontaining1000µMcarbamazepine.C:Peakamplitudesofevokedexcitationin thesp5c(indicatedbyarowsintherightpanelsofaandb)weremeasuredduring 28

29 superfusionwith10,100and1000µmcarbamazepine.d:fluorescencesignal amplitudesat165msafterstimulation(indicatedbyarowsintherightpanelsofaand B)weremeasuredduringsuperfusionwith10,100and1000µMcarbamazepine.E: Fluorescencesignalamplitudesat385msafterstimulation(indicatedbyarowsinthe rightpanelsofaandb)weremeasuredduringsuperfusionwith10,100and1000µm carbamazepine.ineachgraph,fluorescencesignalamplitudeisindicatedasthepercent amplitudeofthecontrolvalueduringsuperfusionwithcontrolmockcsf.dataofeach concentrationwereobtainedfromsixpreparationsandarepresentedasmean±sd. CarbamazepinedidnotinducesignificantchangesinevokedexcitationintheSp5c.NS: notsignificant. Figure3EfectofgabapentinonevokedexcitationintheSp5c A:RecordingmadeduringsuperfusionwithcontrolmockCSF.Theleftpanels representneuralactivity,whichisindicatedaschangesinfluorescenceintensityusing pseudocolor.therightpanelshowsthetimecourseofthesignalchangeandtimeof electricalstimulation.opticalimageintheleftpanelshows165msafterstimulation,as indicatedbytheverticaldotedlineintherightpanel.twohorizontallinesintheimage aretheanchorsforslicepreparation.b:recordingmadeduringsuperfusionwithmock 29

30 CSFcontaining100µMgabapentin.C:Peakamplitudesofevokedexcitationinthe Sp5c(indicatedbyarowsintherightpanelsofAandB)weremeasuredduring superfusionwith1,10and100µmgabapentin.d:fluorescencesignalamplitudesat 165msafterstimulation(indicatedbyarowsintherightpanelsofAandB)were measuredduringsuperfusionwith1,10and100µmgabapentin.e:fluorescencesignal amplitudesat385msafterstimulation(indicatedbyarowsintherightpanelsofaand B)weremeasuredduringsuperfusionwith1,10and100µMgabapentin.Ineachgraph, fluorescencesignalamplitudeisindicatedasthepercentamplitudeofthecontrolvalue duringsuperfusionwithcontrolmockcsf.dataofeachconcentrationwereobtained fromsixpreparationsandarepresentedasmean±sd.administrationof1and10µm gabapentindidnotinducesignificantchangesinevokedexcitationinthesp5c,but100 µmgabapentinemphasizedtheevokedexcitation.**p<0.01;nsnotsignificant. Figure4EfectoflowMg 2+ concentrationsolutionandap5onevoked excitationinthesp5c A:RecordingmadeduringsuperfusionwithcontrolmockCSF.B:Recordingmade duringsuperfusionwithlow Mg 2+ concentration(0.8mm)solution.c:recordingmade duringsuperfusionwithlow Mg 2+ solutioncontaining30µmap5.d:peakamplitudes ofevokedexcitationinthesp5c(indicatedbyarowsintherightpanelsofa,bandc) 30

31 weremeasuredduringsuperfusionwithcontrolsolution,withlow Mg 2+ solutionand withlow Mg 2+ solutioncontaining30µmap5.e:fluorescencesignalamplitudesat 165msafterstimulation(indicatedbyarowsintherightpanelsofA,BandC)were measuredduringsuperfusionwithcontrolsolution,withlow Mg 2+ solution,andwith low Mg 2+ solutioncontaining30µmap5.f:fluorescencesignalamplitudesat385ms afterstimulation(indicatedbyarowsintherightpanelsofa,bandc)weremeasured duringsuperfusionwithcontrolsolution,withlow Mg 2+ solution,andwithlow Mg 2+ solutioncontaining30µmap5.ineachgraph,fluorescencesignalamplitudeis indicatedasthepercentamplitudeofthecontrolvalueduringsuperfusionwithcontrol mockcsf.datawereobtainedfromsixpreparationsandarepresentedasmean±sd. Low Mg 2+ solutioninducedsignificantincreasesinevokedexcitationinthesp5c,but thisincreasewasantagonizedbyadditionaladministrationof30µmap5.*p<0.05;** P<0.01;NSnotsignificant. Figure5EfectofcarbamazepineinlowMg 2+ conditionsonevoked excitationinthesp5c A:RecordingmadeduringsuperfusionwithcontrolmockCSF.B:Recordingmade duringsuperfusionwithlow Mg 2+ concentration(0.8mm)solution.c:recordingmade 31

32 duringsuperfusionwithlow Mg 2+ solutioncontaining1000µmcarbamazepine.d: PeakamplitudesofevokedexcitationintheSp5c(indicatedbyarowsintheright panelsofa,bandc)weremeasuredduringsuperfusionwithcontrolsolution,withlow Mg 2+ solutionandwithlow Mg 2+ solutioncontaining1000µmcarbamazepine.e: Fluorescencesignalamplitudesat165msafterstimulation(indicatedbyarowsinthe rightpanelsofa,bandc)weremeasuredduringsuperfusionwithcontrolsolution, withlow Mg 2+ solution,andwithlow Mg 2+ solutioncontaining1000µm carbamazepine.f:fluorescencesignalamplitudesat385msafterstimulation (indicatedbyarowsintherightpanelsofa,bandc)weremeasuredduring superfusionwithcontrolsolution,withlow Mg 2+ solutionandwithlow Mg 2+ solution containing1000µmcarbamazepine.ineachgraph,fluorescencesignalamplitudeis indicatedasthepercentamplitudeofthecontrolvalueduringsuperfusionwithcontrol mockcsf.datawereobtainedfromsixpreparationsandarepresentedasmean±sd. Low Mg 2+ solutioninducedsignificantincreasesinevokedexcitationinthesp5c,but thisincreasewasantagonizedbyadditionaladministrationof1000µmcarbamazepine. *P<0.05;**P<0.01;NSnotsignificant. 32

33 Figure6EfectofgabapentininlowMg 2+ conditionsonevokedexcitation inthesp5c A:RecordingmadeduringsuperfusionwithcontrolmockCSF.B:Recordingmade duringsuperfusionwithlow Mg 2+ concentration(0.8mm)solution.c:recordingmade duringsuperfusionwithlow Mg 2+ solutioncontaining100µmgabapentin.d:peak amplitudesofevokedexcitationinthesp5c(indicatedbyarowsintherightpanelsof A,BandC)weremeasuredduringsuperfusionwithcontrolsolution,withlow Mg 2+ solutionandwithlow Mg 2+ solutioncontaining100µmgabapentin.e:fluorescence signalamplitudesat165msafterstimulation(indicatedbyarowsintherightpanelsof A,BandC)weremeasuredduringsuperfusionwithcontrolsolution,withlow Mg 2+ solution,andwithlow Mg 2+ solutioncontaining100µmgabapentin.f:fluorescence signalamplitudesat385msafterstimulation(indicatedbyarowsintherightpanelsof A,BandC)weremeasuredduringsuperfusionwithcontrolsolution,withlow Mg 2+ solutionandwithlow Mg 2+ solutioncontaining100µmgabapentin.ineachgraph, fluorescencesignalamplitudeisindicatedasthepercentamplitudeofthecontrolvalue duringsuperfusionwithcontrolmockcsf.datawereobtainedfromsixpreparations andarepresentedasmean±sd.low Mg 2+ solutioninducedsignificantincreasesin 33

34 evokedexcitationinthesp5c,butthisincreasewasantagonizedbyadditional administrationof100µmgabapentin.*p<0.05;**p<0.01;nsnotsignificant. 34

35 Figure 1

36 Figure 2

37 Figure 3

38 Figure 4

39 Figure 5

40 Figure 6