1

Monoaminergic descending pathways contribute to modulation ofneuropathic pain by increasing-intensity treadmill exercise afterperipheralnerveinjury

VictorM.Lopez-Alvarez*,MariaPuigdomenech*,XavierNavarroandStefanoCobianchi

InstituteofNeurosciences,DepartmentofCellBiology,PhysiologyandImmunology,UniversitatAutònomadeBarcelona,andCentrodeInvestigaciónBiomédicaenRedsobreEnfermedadesNeurodegenerativas(CIBERNED),Bellaterra,Spain.

*Bothauthorscontributedequallytothiswork.

Correspondingauthor:Dr.StefanoCobianchi,UnitatdeFisiologiaMèdica,FacultatdeMedicina,UniversitatAutònomadeBarcelona,E-08193Bellaterra,Spain.E-mail:stefano.cobianchi@uab.cat

Thisisanon-finalversionofanarticlepublishedinfinalformin

ExperimentalNeurology,299(2018)42–55, doi: 10.1016/j.expneurol.2017.10.007

©2018. This manuscript version is made available under the CC-BY-NC-ND 4.0 license

http://creativecommons.org/licenses/by-nc-nd/4.0/

2

Abstract

Thisstudycharacterizestheimpactofincreasing-intensitytreadmillexercise(iTR)on noradrenergic (NE) and serotonergic (5HT) modulation of neuropathic pain.Following sciatic nerve transection and repair (SNTR) rats developed significantmechanical and thermal hyperalgesia thatwas partially preventedby iTR performedduringthefirst2weeksafterinjury.

Markeddecreaseintheexpressionof5HT2Aandα1Aandβ-,butnotα2AadrenergicreceptorsinthespinalcorddorsalhornwasassociatedtoSNTRandrecoveredbyiTR,particularlyinlaminaII.iTRsignificantlyincreased5HT2Ainperiaqueductalgrey(PAG),raphe magnus (RM) and dorsal raphe nucleus (DRN), with a pattern suggestingreorganizationofserotonergicexcitatoryinterconnectionsbetweenPAGandDRN.iTRalso increased the expression of α1A in locus coeruleus (LC) and DRN, and β2 in LC,indicatingthatexerciseenhancedactivityofNEneurons,likelybyactivatingautologousprojectionsfromDRNandPAG.

iTR hypoalgesia was antagonized by blockade of β2 and 5HT2A receptors withadministration of butoxamine and ketanserin. The neurotoxin DSP4 was injected toinduce depletion of NE projections from LC before starting iTR. DSP4 treatmentworsened mechanical hyperalgesia, but iTR hypoalgesia was similarly produced.Moreover,5HT2AexpressioninLCfurtherincreasedafterDSP4injection,alltheseresultssuggesting an intrinsic regulation of 5HT andNE activity between PAG, DRN and LCneuronsactivatedbyiTR.

Finally, iTR significantly reduced microglial reactivity in LC and increased non-microglial BDNF expression, an effect that was reverted by butoxamine, implicatingBDNFregulationincentral5HT/NEactionsonneuropathicpain.

Keywords

Treadmill,exercise,training,neuropathicpain,hyperalgesia,serotonin,noradrenalin,locuscoeruleus.

Abbreviations

iTR,increasing-intensitytreadmillexercise;5HT,serotonergic;NE,noradrenergic;SNTR,sciaticnervetransectionandrepair;i.p.,intraperitoneal;RM,raphemagnus;DRN,dorsalraphenucleus;PAG,periacqueductalgrey;LC,locuscoeruleus;Bu,butoxamine;Ke,ketanserin;DSP4,N-(2-chloroethyl)-N-ethyl-2-bromobenzylamine;BDNF,brain-derivedneurotrophicfactor.

3

1. Introduction

Increasing-intensitytreadmillexercise(iTR),performedduringthefirstdaysafterinjury,hasbeenshowntosignificantlyreducehyperalgesiainneuropathicpainmodels,suchassciaticnervechronicconstrictioninjury(CCI)(Cobianchietal.,2010)andsectionandsuturerepair(SNTR)(Cobianchietal.,2013).iTRdecreasedtheexpressionofBDNFandotherneurotrophinsalthoughwithoutimpairingnerveregeneration(Cobianchietal., 2013, Lopez-Alvarez et al., 2015). iTR-induced decrease of BDNF expressionwasassociated to reduction ofmicrogliosis and restoration of the expression of chloridetransporters in the primary sensory neurons and along the central pain pathways(Lopez-Alvarezetal.,2015;Modoletal.,2014).Littleisknown,however,onthecentralmechanismsbywhichiTRinduceshypoalgesia,andwehypothesizethatitmayactbyactivatingdescendingpathwaysforinhibitionofpaintransmissionatthespinallevel.

Brain areas can modulate the ascending pain signals by serotonergic andnoradrenergic projections to spinal cord neurons, which can facilitate or inhibit theafferent sensory neurons. Periaqueductal grey (PAG) areas receive pain andtemperature fibers and activate defensive and stress responses by sending axons toboth locus coeruleus (LC) and raphe magnus nucleus (RM), whose antinociceptiveoutputtriggerdescendinginhibition(Millan,2002).SerotonergicRMandnoradrenergicLC projections normally activate spinal enkephalinergic and GABA/glycinergicinterneurons. Enkephalin released from terminals of enkephalinergic dorsal horninterneurons acts on the opioid receptors located on the central processes ofnociceptiveprimaryafferents,reducingCa2+entryintotheirterminalsanddecreasingthe release of nociceptive neurotransmitters such as glutamate and substance P(Ossipovetal.,2010).Similarly,activationofdorsalhorninterneuronscontainingGABAorglycinealsoinhibitthespinaltransmissionofnoxioussensorysignals,andpreviousstudies indicate that spinalGABAergic inhibition is reducedafterexperimentalnerveinjury (Mooreetal., 2002).The lossof tonic inhibitionby spinal interneurons isalsoassociated with dysregulation of NKCC1/KCC2 chloride cotransporters expression,inducing an inversion of GABAergic depolarizing currents in neuropathic conditions(Modoletal.,2014),thatispreventedbyiTR(Lopez-Alvarezetal.,2015).

Besidesthedemonstratedperipheraleffects(Cobianchietal.,2010,2013;Lopez-Alvarezet al., 2015;Udinaetal., 2011),wehypothesized that iTRmayactivatepaincentral inhibitionnormallygatingthenociceptiveinputtosupraspinal,medullaryandcortical areas,which are decreased after peripheral nerve injury. By stimulating thedescending noradrenergic and serotonergic projections to the dorsal horn, specificexercisetrainingmayresultintheactivationofinhibitorycircuitsinthedorsalhornandintheconsequentinhibitionofsecond-orderspinothalamicneuronsbypresynapticandpostsynapticmechanisms.

Westudiedtheexpressionofnoradrenaline(NE)andserotonin(5HT)receptorsinsensory neurons of the spinal cord dorsal horn, that participate in painmodulation.AmongNEreceptors,α1AreceptorsareexpressedinGABAergicandglycinergicneurons

4

ofdorsalhornlaminaII,wheretheymayparticipateinendogenousinhibitionofafferentpainbyexcitinginhibitoryinterneurons(Babaetal.,2000).α2AreceptorsareexpressedinthespinalcordpredominantlyontheterminalsofprimaryC-fibersafferents,wheretheyinhibitnociception(Stoneetal.,1998).β2receptorisanexcitatoryadrenoreceptorexpressed in dorsal horn neurons (Nicholson et al., 2005), which activation inducesantinociceptiveeffects(Yalcinetal.,2009a,b,2010;Zhangetal.,2016).Wealsoassessedchanges in serotonergic 5HT2A receptor since it is involved in spinal chloridehomeostasis,whichdysregulationisassociatedwithspinaldisinhibitionandneuropathicpain(GackiereandVinay,2014;Jacobsetal.,2002).Underhyperalgesicstatesthe5HT2Areceptorwasfoundtobeexpressedin laminaeI-IIINK1R-positiveprojectionneurons(Mantyhetal.,1997)andinlaminaIIgalanin-containingneuronsexpressingGABAergicboutons(Tiongetal.,2011),receivingemerginginterestasapotentialtargetfortreatingnerveinjury-inducedpainandspasticity(Bosetal.,2013).

In this study,we investigated the changes inducedby iTRonnoradrenergic andserotonergic circuitry that may be related to antinociception. For this purpose, weanalyzedtheexpressionofadrenergicα1A,α2Aandβ2receptors,andserotonergic5HT2Areceptorafter injury to the sciaticnerve in rats, and their changesunder increasing-intensity exercisewhen neuropathic pain is prevented.We also relate the exercise-inducedhypoalgesiatotheexpressionofβ2receptorandthereductionofmicrogliosisinnoradrenergicneurons.

2. Materialsandmethods

2.1. AnimalsandsurgeryAdultfemaleSprague-Dawleyrats(240±30g)werehousedinstandardcageswith

access to food and water ad libitum under a light–dark cycle of 12 hours. All theexperimental procedureswere approvedby the Ethics Committeeof theUniversitatAutonomadeBarcelonaandfollowedtheguidelinesoftheEuropeanCommissiononAnimalCare(EUDirective2010/63/EU).Ratswereanesthetizedbyintraperitoneal(i.p.)injection of ketamine (10 mg/kg, Imalgene 500; Rhone-Merieux, Lyon, France) andxylazine(1mg/kg,Rompun;Bayer,Leverkusen,Germany).

Rats were submitted to a sciatic nerve transection and repair (SNTR), a wellcharacterized model that allows the evaluation of neuropathic pain and nerveregeneration (Cobianchietal.,2014).Therightsciaticnervewasexposedat themidthigh, transected at 92mm from the tip of the third toe, and repairedby epineuralsutures (10-0). The wound was closed in two layers and disinfected with povidoneiodine.Ratswerekeptinawarmenvironmentuntiltheirrecoveryfromanesthesia.

2.2. ExperimentaldesignSevendaysbefore surgery, all theanimalswerehabituated to theexperimental

device for treadmill locomotion (Treadmill LE 8706; LETICA, Barcelona, Spain) and

5

pretrainedtothetask,byleavingthemtoexplorethestoppedtreadmillfor5minutesandthentrainedinasingleiTRsession.EachiTRsessionconsistedof1hourrunning,startingatalocomotionspeedof10cm/sthatwasincreased2cm/severy5minutes,untilamaximalspeedof32cm/s(Cobianchietal.,2013;Lopez-Alvarezetal.,2015).Allratswereevaluatedduringfollow-upwithsensorytestsperformedduringthemorning,whereastreadmillrunningsessionswereperformedduringtheafternoon.

Atday3after surgery, animalswere randomly selected to followornot the iTRtraining.Trainingsessionswereperformeddailyover12consecutivedaysfromday3to14afterinjury.

SNTRratsweredividedinseveralgroups:agroupofratsperformediTR(SNTR-iTRgroup,n=10),anda secondgroup remainedsedentary (SNTR-sedgroup,n=8).Othergroups performed or not iTR with pharmacological blockade of β2-receptors withbutoxamine(Sigma-Aldrich,8mg/Kginsaline,i.p.;SNTR-iTR+Bugroup,n=6,andSNTR-sed+Bugroup,n=6),orblockadeof5HT2A-receptorswithketanserin(Sigma-Aldrich,8mg/Kg in saline, i.p.; SNTR-sed+Ke group, n=6, and SNTR-iTR+Ke group, n=6). Thesedrugswereadministeredeachdayoftraining,30minutesbeforestartingtheexercise.Doseswerechosenoverthebasisofpreviousstudiesusingrats.Controlgroupsforbothdrugs were injected with saline vehicle only. A naive group of rats was added forcomparisonwithinjuredrats(n=6).Finally,othertwogroupsofanimalswereinjectedwithN-(2-chloroethyl)-N-ethyl-2-bromobenzylamine(DSP4,Sigma-Aldrich,50mg/kginsaline,i.p.),aneurotoxinthatselectivelyinducesdegenerationofNEneuronsintheLC,thusdepletingNEprojectionsoriginatingfromtheLC(Jonssonetal.,1981;PrietoandGiralt,2001).DSP-4wasinjectedwithin10minofpreparation(Grzannaetal.,1989),anditsadministrationwasperformed4daysbeforetheinjurytoensureeffectfromthebeginningofthetraining.OnegroupofDSP4injectedratsfollowediTR(SNTR-iTR+DSP4group,n=8),andanotherremaineduntrained(SNTR+DSP4group,n=10).

2.3. NociceptivetestsforpainthresholdmeasurementThree days before surgery, all the injured animals were habituated to the

experimentaldevices, and then tested forbaselinenociceptive thresholds recording.Thenociceptivebehaviortestsformechanicalandthermalstimuliwereperformedonbothhindpawsbeforeandatdifferentdaysafterinjury(dpi),theexperimenterbeingblindtoassignmentofratstothedifferentgroups.Lateralandmedialsitesofthepawwere tested to differentiate changes in sensory thresholds produced respectively bysciatic nerve injury from those due to saphenous nerve sprouting (Cobianchi et al.,2014).

SensitivitytomechanicalstimuliwasmeasuredbymeansofanelectronicVonFreyalgesimeter(Bioseb,Chaville,France).Ratswereplacedonawirenetplatforminplasticchambers. Then, a non-noxious pointed probewas gently applied to each test site,slowlyincreasingthepressure.Thethresholdwasexpressedastheforce(ingrams)atwhichratswithdrewthepawinresponsetothestimulus.Acutoffforcewassetat40g,

6

when the stimulus lifted the paw without response. The mechanical nociceptivethresholdwascalculatedasthemeanof3measurementspertestsite,witha3minuteintervalbetweeneachmeasurement.

ThermalsensitivitywasassessedbymeansofaPlantartestalgesimeter(UgoBasile,Comerio,Italy).Ratswereplacedintoaplasticboxwithanelevatedplexiglassfloor.Thebeamofa lampwaspointedatthesametestsitesasabove inthehindpawplantarsurface.Intensitywassettolowpower(40mW/cm2)withaslowheatingrate.Acutofftimeforthestimuliwassetat20secondstopreventtissuedamage.Heatpainthresholdwascalculatedasthemeanof3trialspertestsite,witha5minuteintervalbetweentrials,andexpressedasthelatency(inseconds)ofpawwithdrawalresponse.Forbothmechanical and thermal thresholds, values are reported as the percentage ratiobetweentheipsilateralinjuredandthecontralateralnormalpawateachtestday.Thisallowsarepresentationofneuropathicpaininducedbytheinjurysincenosignificantvariationsofcontralateralthresholdswerefoundinpreviousstudies(Cobianchietal.,2013;Lopez-Alvarezetal.,2015).

2.4. ImmunohistochemistryofspinalcordandbrainAttheendoffollow-upat14dpi,allratswereeuthanizedandspinalcordandbrain

samplescollectedforimmunohistochemicalassays.Lumbar(L4-L5)spinalcordsegmentandbrainwereremovedfromperfusedanimalsandkept infixativefor1hourand4hours respectively, and then cryoprotected with 30% sucrose in PBS. Transversesections, 25µm thick,were cutona cryostat andmountedon slides. Sectionswereblockedwithspecificserumandincubatedovernightwithprimaryantibodiesin0.3%TritonX-100 inPBS,used for staining5HTandNE receptors in spinal cordandbrainsections (anti-β2 receptors, rabbit, 1:500, SantaCruz; anti-α1Areceptors, goat, 1:500,SantaCruz;anti-α2Areceptors,goat,1:500,SantaCruz;anti-5HT2Areceptors,goat,1:500,Santa Cruz), along with antibodies to identify 5HT and NE neurons (anti-tyrosinhydroxylase(TH),mouse,1:500,SantaCruz;anti-tryptophanhydroxylase(TPH),mouse,1:500,SantaCruz),microglialcells(anti-Iba1,goat,1:200,Abcam)andBDNF(anti-BDNF,sheep,1:200,Millipore).

Thefollowingdaysectionswereincubatedfor2hourswithAlexaFluor488and/orAlexa Fluor 594 conjugated secondary antibodies (1:200, Life Technologies). AfterwashinginPBS,sectionswerecoverslippedwithasolutionofDAPI(1mL/mL)inMowiolmounting medium. Positive and negative controls of antibodies were previouslycheckedfornon-specificlabeling.Foreachmarker,threesectionsfrom4samplespergroupwereusedforquantificationofimmunolabeling.Theexperimenterwasblindedtothesamplegroup.AnatomicalstructuresofinterestwerepreviouslyidentifiedundermicroscopeinspinalcordandbrainatlascoordinatesusingspecificsectionsandNisslorHematoxylinstaining.10Xand20XimageswerecapturedwithaZeissLSM700confocalmicroscope,andanalyzedusingImageJsoftware(NIH,USA).Thresholdingoffluorescentsignalwasadjustedoverthebackground levelofanegativecontrol (samplesstained

7

withoutprimaryantibody).Theintegrateddensityofimmunoreactivitywascalculatedwithin regions of interest drawn with ImageJ software tool. For spinal cordimmunoreactivity,dataareshownforbothipsilateralandcontralateralsides.Forbrainimmunoreactivity,dataareshownonlyforprojectionareasoftheinjuredside.

2.5. DataanalysisDataarepresentedasmean±SEM.Statisticalanalysisofnociceptivethresholdswas

made by two-way analysis of variance (ANOVA)with group and time after injury asfactors, followed by Bonferroni’s post hoc comparisons. Statistical comparisons forimmunofluorescencedataweremadebyone-wayandtwo-wayANOVAsfollowedbyTukey’s post hoc test when necessary. The level of statistical significance was 5%(p<0.05)inalltheanalyses.3. Results

3.1. iTRreducedhyperalgesiaafterSNTRChangesinsensorythresholdswererecordedatbothmedialandlateralsitesofthe

hindpaw to monitor the contribution to hyperalgesia of collateral sprouting ofsaphenousnerve,andofdenervationbytheinjuredsciaticnerve(Cobianchietal.,2013;Lopez-Alvarezetal.,2015).AsshowedinFig.1A-B,mechanicalandthermalthresholdsin uninjured contralateral paws were not significantly affected by injury with anipsi/contra%thresholdratioofabout100%at3dpi.HoweverSNTRproducedamarkeddecreaseofthemeanmechanicalthresholdatthemedialsidefrom3to14dpi(Fig.1A;48 to32%ofcontralateral inSNTR-sedgroup).Thedecreaseof thermal thresholdatmedialsidewascomparativelylower(Fig.1B;92to69%inSNTR-sedgroup).Statisticalanalysesforbothgroupandtimefactors,andalsotheirinteraction,showedsignificanteffects for thechangesobserved inmeanmedialmechanical (Group factor:F=26.39,p<0.0001;Timefactor:F=71.45,p<0.0001;Interaction:F=9.35,p<0.0001)andthermalthresholds (Group factor: F=17, p=0.0001; Time factor: F=4, p<0.0114; Interaction:F=8.83, p<0.0001). iTR significantly prevented the reduction of bothmechanical andthermalthresholdsat7and14dpiintheSNTR-iTRgroup(Fig.1;p<0.05andp<0.001,SNRT-iTRvs.SNRT-sedBonferroniposttest).

Ontheotherside,thelateralsideoftheinjuredhindpawwasunresponsiveduringthe14daysfollowingSNTR,sincethisterritoryremainsdenervateduntilatleast4weeksafterinjury(Cobianchietal.,2014).Sincethedevelopmentofthermalhyperalgesiawasslowerthanthemechanical,wedidnotanalyzechangesofthermalthresholdsinfurtherexperiments,andfocusedonthechangesobservedatthemedialsideofthepaw.

3.2. iTRcounteractedthedecreaseofα1Aandb2adrenergicand5HT2Aserotonergic

receptorsexpressioninthedorsalhornafterSNTR

8

WestudiedtheexpressionofadrenergicandserotonergicreceptorsinlaminaeI,IIandIIIofthespinalcorddorsalhorn.Wefoundα1AreceptormainlyexpressedinlaminaIIandatlowerintensityinlaminaeIandIII(Fig.2A).AfterSNTR,α1Aexpressionstronglydecreasedinalllaminaeofbothdorsalhornipsilateralandcontralateraltoinjury(Fig.2A,B).iTRproducedanincreaseofα1AimmunoreactivityinlaminaeII-IIIofbothdorsalhorns,althoughonlysignificantlyintheipsilateralside(Fig.2B;p<0.001,SNTR-iTRvs.SNTR-sed).Ontheotherhand,theexpressionofα2Areceptorwasalmostunchangedinthedorsalhornofinjuredsedentaryandexercisedrats(datanotshown).

Theβ2receptorwasexpressedinalllaminaeI-IIIofthedorsalhorninthenaïverats(Fig. 3A). SNTR reduced the expression of β2 receptor in both the ipsilateral andcontralateral sidesof lamina I, II and III (p<0.001naive vs. SNTR-sed; Fig. 3A,B). Theexpressionofβ2wassignificantlyrecoveredbyiTRinthelaminaIIbilaterally(p<0.001SNTR-sedvs.SNTR-iTR)andinthecontralaterallaminaI(p<0.01).

Innaïveanimals,the5HT2AreceptorwasmoredenselyexpressedinlaminaII,evenifspotclustersofimmunofluorescentlabelingwerepresentinalldorsalhornlaminae(Fig. 4A). After SNTR, 5HT2A immunoreactivity decreased significantly in lamina IIbilaterally(p<0.001naivevs.SNTR-sed;Fig.4B)andintheipsilaterallaminaI(p<0.05naivevs.SNTR-sed).Interestingly,iTRsignificantlyrecovered5HT2AreceptorexpressioninlaminaII(p<0.01SNTR-sedvs.SNTR-iTR),andintheipsilaterallaminaI(p<0.05,SNTR-sedvs.SNTR-iTR;Fig.4B).

TheseresultsindicatethatnerveinjurydeterminesadecreaseofNEand5HTtonein the dorsal horn neurons that may reflect the drive from normal to neuropathicconditions,andhighlighttheiTRpotentialtoreversethesechanges.

3.3. Changesin5HT2A,α1Aandβ2receptorsexpressioninmidbrainareasinducedby

SNTRandafteriTRSinceiTRshowedtosignificantlyimpacttheexpressionof5HT2Aserotonergicand

α1Aandβ2adrenergicreceptorsinthelumbardorsalhornofspinalcord,weinvestigatedthe changes that peripheral nerve injury and iTR treatment determined on theexpressionofthesereceptorsathigherintegratingbraincentersthatparticipateinpaindescendingmodulation,suchtheperiaqueductalgreymatter(PAG),thelocuscoeruleus(LC),thedorsalraphe(DRN)andtheraphemagnusnucleus(RM)(Fig.5A).

Thenormalexpressionofα1Areceptorwasfoundinmediumtolarge-sizeneuronsofLCandDRNandonterminalsofNEprojectionstoRM(Fig.5B,naïvegroup).Slightlystronger stainingwas observed in theDRN after injury (Fig. 5B, SNTR-sed group), incontrastwiththereducedexpressioninthedorsalhorn.iTRsignificantlyincreasedα1Aimmunoreactivity in LC and DRN, and in adjacent areas, such as the subcoeruleusnucleus, and inmore neurons of dorsomedial DRN areas (Fig. 5B, SNTR-iTR group).ComparedwithLCandDRN,wheretheincreaseofα1Aimmunoreactivitywassignificant(Fig.5C;SNTR-iTRvs.naïve,p<0.01,andSNTR-iTRvs.SNTR-sed,p<0.05),onlyslightlyincreasedamountof immunopositiveclusterswasobservedsurroundingthesomaof

9

RMneurons.Thechangesinthepatternofα1AreceptorexpressionacrossLCandraphenucleisuggestthatautologousα1AreceptorsinLCmayhavebeenactivatedinapositivelooptoincreaseLCactivityunderexercisetraining.

Autologousβ2receptorswerealsofoundinLCandsparseimmunoreactivityinDRNand RM of naïve rats (Fig. 6A, naïve group). Nerve injury did not induce significantchangesinβ2receptors(Fig.6A,SNTR-sedgroup).iTRratsshowedimmunoreactivityinalargerareaofLC(Fig.6A-B,SNTR-iTRvs.SNTR-sedp<0.01;SNTR-iTRvs.naïvep<0.05),and also a slight not significant higher density in DRN and RM, as terminals ofnoradrenergicprojectionsfromLC(Fig.6B).TheseresultsconfirmedthatiTRenhancedtheactivityofNELCneurons.

Theimmunolabelingof5HT2AreceptorinnaïveratsshowedpositiveclustersinPAGandDRN regions and rarely in the RMnucleus extending laterally into the adjacentreticular formation,whereare large-sizeserotonergicneurons (Fig.7A,naïvegroup).Peripheralnerveinjurydecreasedtheexpressionof5HT2AinPAGandDRN(Fig.7A-B,SNTR-sedgroup).Thepatternofexpressionappeareddifferentininjuredanimals,withhigherpresenceofscattered5HT2AclustersthroughdorsomedialareasofRMthaninthePAGcomparedwithnaïvesamples.iTRstronglyincreased5HT2Areceptorexpressionin PAG and DRN (Fig. 7A-B, SNTR-iTR vs. SNTR-sed p<0.01, and SNTR-iTR vs. naïvep<0.05).Therewasstrongimmunolabelinginthesomaofmanylarge-sizeneuronsintheseareasandevenindendritesofPAGandofDRNneuronsprojectingtoothernuclei.Spotclustersindicatingterminal5HT2AcontactsweresignificantlyincreasedintheRM(Fig.7A-B,SNTR-iTRvs.SNTR-sedp<0.05,andSNTR-iTRvs.naïvep<0.01).This resultsuggestsareorganizationofserotonergicconnectionsbetweenPAGandraphenuclei,withthepossiblefunctiontoincreasetheiractivitytodescendingoutput.

3.4. Blockadeofb2or5HT2AreceptorsantagonizedtheiTRinducedhypoalgesiaafter

SNTRWewantedtotestiftheactivationofβ2and5HT2Areceptorswasdirectlyinvolved

inthepreventionofmechanicalhyperalgesiaobservediniTRratsafterSNTR.Inhibitionof β2 receptorswith butoxamine, a selective β2 blocker, before starting iTR sessionsantagonized its hypoalgesic effect at both 7 and 14 dpi (Fig. 8A, SNTR-iTR vs. SNTR-iTR+Bup<0.05).However,butoxaminetreatmentalonehadnoeffectsonmechanicalhyperalgesia,indicatingaspecificβ2activitycontributionontheiTReffects.

Similarly,treatmentwithketanserin,aninhibitorof5HT2receptors(alsowithweakɑ1adrenergicblockingproperties),significantlyantagonizedtheiTRhypoalgesiceffect,butonlyat14dpi(Fig.8B).iTRhypoalgesiawasonlypartiallybluntedat7dpi(Fig.8B,SNTR-iTR vs. SNTR-iTR+Ke, p<0.001 at 14 dpi). Ketanserin treatment alone partiallyreduced SNTR inducedmechanical hyperalgesia, indicating that 5HT2 receptors (andmaybeɑ1receptorsblockedbyketanserin)maymediatemechanismsofsensitizationafternerveinjury.

10

3.5. DepletionofNEoutputfromlocuscoeruleuswithDSP4didnotmodifythehypoalgesiceffectofiTR

TheNEspecificneurotoxinDSP4wasinjectedtoinducedepletionofNEprojectionsfrom the LC (Grzanna et al., 1989), thus preventing the activation ofNE descendingprojectionsduringiTR(Fig.9).InratsreceivingDSP4injectionthehyperalgesiawasevenslightly increased compared with untreated rats (Fig. 9A). However, iTR treatmentsimilarly produced hypoalgesia in injured rats (p<0.01 SNTR-iTR+DSP4 vs. SNTR-sed+DSP4),despitethedepletionofNEoutputsignificantlyreducedtheexpressionofβ2andα1AreceptorsintheLC,whichwasnotcompensatedbyiTR(Fig.9B-C).Thisresultsuggests that the activation of NE descending projections may be not the onlymechanisminducingpainreliefbyiTR,andpointsoutthepossibilitythatmesencephalic5HTsystemplayasupportiveormodulatoryroleforNE.

Toconfirmthishypothesis,welookedatthe5HT2AexpressioninLC.Sparselabelingof5HT2AreceptorswasfoundintheLCofnaïverats(Fig.10A),probablycorrespondingto terminals of PAG excitatory and DRN inhibitory projections that regulate the LCactivity(Haddjerietal.,1997;Kimetal.,2004).SNTRdecreasedtheexpressionof5HT2Areceptors,whichwaspreventedinthegroupsperformingiTR(Fig.10A-B).Theinhibitionofβ2receptorsbybutoxaminedidnothaveanyeffectontheexpressionof5HT2AintheLCofinjuredrats.Incontrast,injectionofDSP4inducedasignificantincreaseof5HT2Aimmunoreactivity (p<0.01and0.05SNTR-sedvs.SNTR-sed+DSP4andSNTR-iTR+DSP4respectively),evenhigherthanbyiTRalone.ThestrongeffectinducedbyDSP4ontheexpressionof5HT2AreceptorsmaysuggesttheactivationofDRN5HTtoincreasetheinhibitionofdepletingLCneurons,whileiTRexerciseseemstorestorethenormaltonicPAGandDRN5HTcontrolonNEactivity.AlltogethertheseresultsdemonstratethatiTRhypoalgesiaisconveyedbybothβ2and5HT2Areceptorslikelythroughparallelactivationof5HT2AinPAGandDRNalongtoadrenergicLCneuronstobalancetheimpairmentoftheiractivityduetoinjury.

3.6. iTRreducedmicrogliosisinlocuscoeruleusdependentonactivationofβ2receptor

afterSNTRSinceactivationofmicrogliahasbeendescribedtoberegulatedbyβ2adrenergic

receptoractivation(Fujitaetal.,1998;O’Donnelletal.,2012),andplaysanactiverolein the pathway that leads to BDNF regulation and KCC2 dephosphorylation inneuropathicpainstates(Modoletal.,2014),weevaluatedthemicroglialactivationafterSNTR (Fig. 11).We focused in LC in order to analyze the reaction ofmicroglial cellsadjacenttonoradrenergicneuronsprojectingtothespinalcord.Iba1immunoreactivityintheLCshowedsignificantmicrogliaactivationafternerveinjurycomparedtonaïverats(Fig.11A-B,p<0.01,naïvevs.SNTR-sed).Interestingly,iTRsignificantlyreducedIba1immunoreactivity (p<0.05, SNTR-sed vs. SNTR-iTR). However, the effect of iTR wasrevertedbybutoxamineinjection(p<0.001,naïvevs.SNTR-sed+BuandSNTR-iTR+Bu).Butoxamine treated groups showed highly ramified microglial cells suggesting that

11

inhibitionofβ2receptorsmayfavorproliferationofmicroglia(Fig.11A).Theseresultsshow that β2 receptors are involved in the iTR reductionofmicroglial reaction afternerve injury in the LC, and this mechanism could represent a potential shift fromhyperalgesictohypoalgesicstate.

Sincetheexpressionofinjury-inducedmicroglialBDNFreactivityinthespinalcorddorsalhornneuronsisassociatedtoneuropathicpainandisreducedbyiTR(Cobianchiet al., 2013; Lopez-Alvarez et al., 2015; Udina et al., 2011), we also analyzed BDNFexpression in the LC. BDNF immunoreactivitywas scant in LC of naïve rats, and notcolocalizedwith Iba1 (Fig. 11A, naïve group). The level of BDNF expressionwas notinfluencedbynerveinjury,butitwaspresentinreactivemicroglia(Fig.11A,SNTR-sedgroup),andincreasedabovenormallevelsbyiTR(Fig.11A,SNTR-iTRgroup).Incontrast,BDNFimmunoreactivitywasloweredbybutoxaminetreatment(Fig.11A,SNTR-sed+BuandSNTR-iTR+Bu),althoughchangeswerenotsignificantbetweenthedifferentgroups(Fig.11B).ThisresultshowsthatiTRincreasesBDNFexpressioninmidbrainareasbutnotdependentonβ2receptorsdifferentlyfromtheactivationofmicroglia.4. Discussion

In this study we aimed to analyze the contribution of 5HT and NE descendingpathwaystothehypoalgesiceffectofiTRonperipheralneuropathicpain.WeshowedthatiTRreducedhyperalgesiawasassociatedwithparallelrecoveryintheexpressionofserotonergic5HT2Aandnoradrenergicα1Aandβ2receptorsinsensoryneuronsoflumbarspinalcordandinbrainstemareasknowntointegratesensoryandmotorinputsinordertomodulatethedescendingresponses.

TheimmunohistochemicalcharacterizationofNEreceptorsexpressioninspinalcorddorsalhornrevealedthattheyaremainlyexpressedinlaminaeII-III,wheretheywerereducedafterSNTR.Therecoveryofα1Aandβ2butnotofα2AreceptorsobservedafteriTRmayberelatedtotheinhibitionoftheafferentpain.Babaetal.(2000)demonstratedthe contribution of α1 receptors to antinociception by activation of GABAergic andglycinergicinhibitoryinterneurons.Ourresultssuggestthatthedownregulationofα1Areceptor in dorsal horn interneurons contributes to hyperexcitability and spinaldisinhibitionafterperipheralnerveinjury.iTRtrainingupregulatedtheα1AreceptorinthelaminaII,aregionrichofGABAergicandglycinergicinterneurons,thuspromotingtheirinhibitoryaction.

iTRalsorevertedthereductionofβ2receptor,particularlyinlaminaIIofthedorsalhorn.Theβ2receptorispresentinthecentralterminalsofnociceptiveafferentsandondendrites of superficial dorsal horn neurons (Nicholson et al., 2005; Patterson andHanley,1987;Mizukami,2004).Interestinglycapsaicintreatmentinratsdecreasedtheβadrenergicbindingsitesinthespinalcord(PattersonandHanley,1987),aswehavefound after sciatic nerve injury. The facts that iTR increasedβ2 expression, and thatinhibitionofβ2withbutoxamineantagonizedthehypoalgesiceffectofexercise,supporttheviewthatβ2receptoriscontributingtomediatethehypoalgesiainducedbyiTR.This

12

suggestionisinlinewiththeantiallodyniceffectdemonstratedbyYalcinetal.(2009a,b,2010)bypharmacologicalstimulationofβ2aftersciaticlesion.

Activationof5HT2Areceptorisknowntoenhanceglycineand/orGABAresponsesinspinalneurons(Xuetal.,1996;Liuetal.,2000),andtoshiftthechlorideequilibriumpotential in the hyperpolarizing direction (Bos et al., 2013), restoring endogenousinhibition at the dorsal horn. We have found that iTR increased 5HT2A expression,suggestingthatactivationof5HT2AmediatedtonicinhibitionmaycontributetotheiTReffectonreducinghyperalgesiaafternerveinjury.However,anincreaseindorsalhorn5HT2Adensitywasreportedinamodelofpersistentpain(VanSteenwinckeletal.,2009),and upregulation and activation of 5HT2A was observed after spinal nerve ligation,associatedtospinalhyperexcitability(Airaetal.,2010).Furthermore,administrationofaselective5HT2AagonistinthespinalnerveligationmodelsignificantlyincreasedC-fiberevokedpotentials(Airaetal.,2010).

On the other hand, we showed that immunoreactivity to 5HT2A after SNTRsignificantlydecreased in laminaeI-III,similarlytoβ2receptor,andthepreventionof5HT2AreceptoractivationduringexercisebypretreatmentwithketanserinreducedtheiTRhypoalgesic effect. Then if restorationofboth spinal 5HT2A andβ2 receptors is adirectconsequenceofiTR,5HT2Amaybenotdirectlyinvolvedasβ2inthemechanismofearlypainsuppression.Thismaybeexplainedbythemultiplerolesthat5HT2Areceptormay play on neuropathic pain conditions, by differently acting on neuronalsubpopulationsofdorsalhorntoinhibitorfacilitatepainsignals.Afterinjury,ashiftof5HT2A receptors to inhibitory interneurons was shown (Aira et al., 2010). Since iTRcounteracted the spinal loss of 5HT2A, we suggest that our treadmill protocol maypotentiatetheactivityofinhibitoryinterneuronsthatinterveneinnociception.

Besides changes in the spinal cord, antinociceptionmay be conveyed by iTR byactivationof 5HT2A receptor aswell as adrenoreceptors athigherbrain centers. It isknownthatprojectionsfromPAG-RMandLCtothespinalcordplayapivotalroleinpaincontrol.PainandtemperaturefibersprojecttoPAGthroughthespinomesencephalictract, andPAG can control afferentpainbymeansof parallel actionsonRMand LC(Basbaum and Fields, 1978). In these nuclei, antinociceptive mechanisms can beactivatedthroughμ1opioid,5HT2Aand5HT2Cserotonergic,andα1Aadrenergicreceptorsof the LC (de Freitas et al., 2016). β2 receptor is also expressed in areas directlyparticipatinginpain(Nicholsonetal.,2005),andhumangeneticstudiesconfirmedthecontribution of β2 to chronic pain disorders (Diatchenko et al., 2006).Moreover, β2receptorisessentialfortheantiallodynicactionofantidepressantdrugs(Yalcinetal.,2009a,b), since the absence or blockade of β2 suppressed the hypoalgesic effect ofantidepressantdrugsonmechanicalallodynia(Yalcinetal.,2009b).

TheactionsplayedbyiTRonmesencephalicnucleitoactivatecentralpaincontrolmay be multiple, and conveyed through synaptic regulation of NE and 5HTinterconnections. Indeed,PAG isknownalso tobeexcitedor inhibitedby5HT2Aand5HT1AreceptorsexpressingfibersfromDRN,whichformaregulatorycircuitrynegatively

13

modulatedbyGABAandopioids(JolasandAghajanian,1997;Liuetal.,2000).TheDRNisthe largestserotonergicnucleusprovidingmostsupplyof5HTtoforebrain,and itsdorsalsubnucleiarelocatedadjacenttothePAG,wheredenseclustersof5HTneuronsprojecttotheRM(KwiatandBasbaum,1990;ChoandBasbaum,1991)andtotheLC(Kimetal.,2004).Activationofthebrainstem5HTsystemmodulatesbothnociceptiveandmotorspinalcordcircuits(JordanandSlawinska,2011;Pearlsteinetal.,2005),andexercise increases the release of 5HT in supraspinal areas associated to reducingmechanicalhyperalgesia(Gerinetal.,2008;Korbetal.,2010;Jacobsetal.,2002).Inarecent work, Bobinski et al. (2015) found increased brainstem levels of 5HT, itsmetabolitesand5HT1B/2A/2C receptorsaftera2-weeks low-intensity treadmill trainingfollowing sciatic nerve injury. Interestingly, the inhibition of 5HT but not ofcatecholamines reversed the hypoalgesic effect of such low-intensity exercise. Atdifference, our study suggests that NE antinociception can be further activated byincreasingtheintensityofexercise.Indeed,theiTRtraininginducedanincreaseofα1A,β2and5HT2AreceptorsexpressioninthePAG-DRNandPAG-LCpathways.5HTandNEsystemsmay interact in theseareasduringexercise.Earlier studies revealed thatNEcauses an increase in 5HT neuronal firing in DRN, mediated via activation of α1adrenoceptorlocatedon5HTneurons(Dayetal.,1997;VandermaelenandAghajanian,1983).Wehypothesizethattheactivationofα1AreceptorintheDRNmayformpartofaregulatoryloopthatisactivatedbyiTRbetweenDRNandLCtoenhanceLCandRMdescending pain-suppressing neurons. NE axons could also activate α1 receptors onGABAergic and glycinergic inhibitory interneurons leading to inhibition of pain-relayneurons(Pertovaara,2006).

On the other hand, pharmacological stimulation of β2 receptor suppressedneuropathicpainaftersciaticnerveinsult(Yalcinetal.,2010).WefoundthatiTRstronglyincreasedtheexpressionofβ2receptorinthedorsalhornlaminaIIinterneuronsandintheLC,andthatactivationofβ2receptorwasnecessarytoinduceiTRhypoalgesia.Tounderstandwhichpartofthiscircuitryisinvolvedinthehypoalgesiceffect,wedepletedtheNEoutputfromLCbyusingtheneurotoxinDSP4(Jonssonetal.,1981;PrietoandGiralt, 2001). Administration of DSP4 has been shown to revert antinociception indifferentpainmodels(Zhongetal.,1985;Kudoetal.,2010)andtoregulate5HTagonistsinducedanalgesia(Garciaetal.,2003).Inourmodel,pretreatmentwithDSP4increasedthehyperalgesiabutdidnotblocktheeffectofiTR.SincetheactivityofLCneuronsissubordinatedtotheactivationof5HTprojectionsfromPAGandDRN,wesuggestthatiTRtrainingcantriggerNE-induceddescendinginhibitionbyreestablishingalso5HT2Areceptor activity in the brainstem. Our results highlight the 5HT and NE reciprocalactionsunderneuropathicpainandblockadeofNEneurons.

Brainandspinalexpressionofneurotrophins,particularlyBDNF,canbemodulatedby activation of NE and 5HT pathways, with relevant effects onmotor and sensoryrecovery after nerve injuries. NE activation via β adrenergic receptors seems to beessential forexercise-inducedBDNFregulation(Garciaetal.,2003),andβadrenergic

14

blockade significantly attenuates the increase of BDNFmRNAdue to exercise in thecortex(Ivyetal.,2003).Ontheotherhand,BDNFreleasedbyactivatedmicrogliatriggersneuropathicmechanisms,suchasdownregulationofchloridecotransporterKCC2(Coullet al., 2005; Ferrini and De Koninck, 2013), associated to disinhibition at spinalinterneurons(Modoletal.,2014).iTRincreasedtheBDNFexpressioninLCparalleltodecreasing microgliosis and BDNF-expressing microglia, as we previously showed inassociationtotherecoveryofKCC2levels(Lopez-Alvarezetal.,2015).Theblockadeofβ2receptoralsoincreasedmicrogliosis,suggestingthatactivationofβ2receptorduringiTRplaysaroleinmodulationoftheneuroinflammatoryresponsetonerveinjury.5HT2AreceptoractivityhasbeendescribedtodownregulatetheBDNFexpression intheratbrain(Vaidyaetal.,1997).ThereductionofneurotrophinsNGFandBDNFobservedinDRGsensoryneuronsafteriTR(Cobianchietal.,2013;Lopez-Alvarezetal.,2015)couldbeassociatedtothesustainedincreaseofβ2and5HT2Areceptorsexpressioninbrainandspinalcord,whicharereducedbysciaticnerveinjury.Theseresultsindicatethattheincreasedneurotrophicfactorproductioninsensoryneuronsafterperipheralnerveinjurycanbereverselymodulatedby increasedintensityexercise inordertopreventmaladaptiveplasticchangesassociatedtoneuropathicpain.

5. Conclusions

Theresultsofthisstudybringnewknowledgeonthecontributionofdescendinginhibition, by increasing activity of serotonergic and noradrenergic projections frombrainstem centers, to the beneficial effects of specific exercise training programs inreducing neuropathic pain. Future studies aimed at understanding which neuronalpopulations andwhichmolecularmechanismsunderlying pain inhibition and centralbrainstem circuits are still needed to foster the therapeutic possibilities for treatingsensorimotordisorderswithspecificexerciseprograms.

ConflictofintereststatementTheauthorshavenoconflictsof interesttodeclare.Thisworkwassupportedby

Grant EPIONE (FP7-602547) from the European Commission (EC), and TERCEL andCIBERNEDfundsfromtheFondodeInvestigacionSanitariaofSpain.

AcknowledgementsThe authors are grateful to the technical help of Nuria Barba for the confocal

microscopy,andMònicaEspejoforlabmanagement.

15

References

Aira,Z.,Buesa,I.,Salgueiro,M.,Bilbao,J.,Aguilera,L.,Zimmermann,M.,Azkue,J.J.,2010.Subtype-

specificchangesin5-HTreceptor-mediatedmodulationofCfibre-evokedspinalfieldpotentials

aretriggeredbyperipheralnerveinjury.Neuroscience168,831-841.

Baba H., Goldstein P.A., Okamoto M., Kohno T., Ataka T., Yoshimura M., Shimoji K., 2000.

Norepinephrine facilitates inhibitory transmission in substantia gelatinosa of adult rat spinal

cord(part2):effectsonsomatodendriticsitesofGABAergicneurons.Anesthesiology92,485-

492.

BasbaumA.I.,FieldsH.L.,1978.Endogenouspaincontrolmechanisms:reviewandhypothesis.Ann

Neurol4,451-462.

BobinskiF.,FerreiraT.A.,CordovaM.M.,DombrowskiP.A.,daCunhaC.,SantoC.C.,PoliA.,Pires

R.G.,Martins-SilvaC., SlukaK.A., SantosA.R.,2015.Roleofbrainstemserotonin inanalgesia

producedbylow-intensityexerciseonneuropathicpainaftersciaticnerveinjuryinmice.Pain

156,2595-2606.

BosR.,SadlaoudK.,BoulenguezP.,ButtigiegD.,LiabeufS.,BrocardC.,HaaseG.,BrasH.,VinayL.,

2013. Activation of 5-HT2A receptors upregulates the function of the neuronal K-Cl

cotransporterKCC2.ProcNatlAcadSciUSA110,348-353.

ChoH.J.,BasbaumA.I.,1991.GABAergiccircuitryintherostralventralmedullaoftheratandits

relationshiptodescendingantinociceptivecontrols.JCompNeurol303,316-328.

CobianchiS.,Casals-DiazL.,JaramilloJ.,NavarroX.,2013.Differentialeffectsofactivitydependent

treatments on axonal regeneration and neuropathic pain after peripheral nerve injury. Exp

Neurol240157-167.

CobianchiS.,deCruzJ.,NavarroX.,2014.Assessmentofsensorythresholdsandnociceptivefiber

growthaftersciaticnerveinjuryrevealsthedifferentialcontributionofcollateralreinnervation

andnerveregenerationtoneuropathicpain.ExpNeurol255,1-11.

CobianchiS.,MarinelliS.,FlorenzanoF.,PavoneF.,LuvisettoS.,2010.Short-butnotlong-lasting

treadmill running reduces allodynia and improves functional recovery after peripheral nerve

injury.Neuroscience168,273-287.

CoullJ.A.,BeggsS.,BoudreauD.,BoivinD.,TsudaM.,InoueK.,GravelC.,SalterM.W.,DeKoninck

Y., 2005. BDNF from microglia causes the shift in neuronal anion gradient underlying

neuropathicpain.Nature438,1017-1021.

DayH.E.,CampeauS.,WatsonS.J.,Jr.,AkilH.,1997.Distributionofalpha1a-,alpha1b-andalpha

1d-adrenergicreceptormRNAintheratbrainandspinalcord.JChemNeuroanat13,115-139.

deFreitasR.L.,MedeirosP.,daSilvaJ.A.,deOliveiraR.C.,deOliveiraR.,UllahF.,KhanA.U.,Coimbra

N.C.,2016.Themu1-opioidreceptorand5-HT2A-and5HT2C-serotonergicreceptorsofthelocus

16

coeruleus are critical in elaborating hypoalgesia induced by tonic and tonic-clonic seizures.

Neuroscience336,133-145.

DiatchenkoL.,AndersonA.D.,SladeG.D.,FillingimR.B.,ShabalinaS.A.,HigginsT.J.,SamaS.,Belfer

I.,GoldmanD.,MaxM.B.,WeirB.S.,MaixnerW.,2006.Threemajorhaplotypesofthebeta2

adrenergicreceptordefinepsychologicalprofile,bloodpressure,andtheriskfordevelopment

ofacommonmusculoskeletalpaindisorder.AmJMedGenetBNeuropsychiatrGenet141b,

449-462.

FerriniF.,DeKoninckY.,2013.MicrogliacontrolneuronalnetworkexcitabilityviaBDNFsignalling.

NeuralPlast2013,429815.

FujitaH.,TanakaJ.,MaedaN.,SakanakaM.,1998.Adrenergicagonistssuppresstheproliferation

ofmicrogliathroughbeta2-adrenergicreceptor.NeurosciLett242,37-40.

Gackiere F., Vinay L., 2014. Serotonergicmodulation of post-synaptic inhibition and locomotor

alternatingpatterninthespinalcord.FrontNeuralCircuits8,102.

GarciaC.,ChenM.J.,GarzaA.A.,CotmanC.W.,Russo-NeustadtA.,2003.Theinfluenceofspecific

noradrenergic and serotonergic lesions on the expression of hippocampal brain-derived

neurotrophicfactortranscriptsfollowingvoluntaryphysicalactivity.Neuroscience119,721-732.

GerinC.,TeilhacJ.R.,SmithK.,PrivatA.,2008.Motoractivityinducesreleaseofserotonininthe

dorsalhornoftheratlumbarspinalcord.NeurosciLett436,91-95.

Grzanna R., Berger U., Fritschy J.M., Geffard M., 1989. Acute action of DSP-4 on central

norepinephrineaxons:biochemicalandimmunohistochemicalevidencefordifferentialeffects.

JHistochemCytochem37,1435-1442.

Haddjeri N., de Montigny C., Blier P., 1997. Modulation of the firing activity of noradrenergic

neuronesintheratlocuscoeruleusbythe5-hydroxtryptaminesystem.BrJPharmacol120,865-

875.

Ivy A.S., Rodriguez F.G., Garcia C., Chen M.J., Russo-Neustadt A.A., 2003. Noradrenergic and

serotonergic blockade inhibitsBDNFmRNAactivation followingexercise andantidepressant.

PharmacolBiochemBehav75,81-88.

JacobsB.L.,Martin-CoraF.J.,FornalC.A.,2002.Activityofmedullaryserotonergicneuronsinfreely

movinganimals.BrainResBrainResRev40,45-52.

Jolas T., Aghajanian G.K., 1997. Opioids suppress spontaneous and NMDA-induced inhibitory

postsynapticcurrentsinthedorsalraphenucleusoftheratinvitro.BrainRes755,229-245.

Jonsson G., Hallman H., Ponzio F., Ross S., 1981. DSP4 (N-(2-chloroethyl)-N-ethyl-2-

bromobenzylamine)--a useful denervation tool for central and peripheral noradrenaline

neurons.EurJPharmacol72,173-188.

17

Jordan L.M., Slawinska U., 2011. Chapter 12--modulation of rhythmic movement: control of

coordination.ProgBrainRes188,181-195.

KimM.A.,LeeH.S.,LeeB.Y.,WaterhouseB.D.,2004.Reciprocalconnectionsbetweensubdivisions

ofthedorsalrapheandthenuclearcoreofthelocuscoeruleusintherat.BrainRes1026,56-67.

KorbA.,BonettiL.V.,daSilvaS.A.,MarcuzzoS.,IlhaJ.,BertagnolliM.,PartataW.A.,Faccioni-Heuser

M.C.,2010.Effectoftreadmillexerciseonserotoninimmunoreactivityinmedullaryraphenuclei

andspinalcordfollowingsciaticnervetransectioninrats.NeurochemRes35,380-389.

KudoT.,KushikataT.,KudoM.,KudoT.,HirotaK.,2010.AcentralneuropathicpainmodelbyDSP-

4inducedlesionofnoradrenergicneurons:preliminaryreport.NeurosciLett481,102-104.

Kwiat G.C., Basbaum A.I., 1990. Organization of tyrosine hydroxylase- and serotonin-

immunoreactivebrainstemneuronswithaxoncollaterals to theperiaqueductalgrayand the

spinalcordintherat.BrainRes528,83-94.

Liu R., Jolas T., AghajanianG., 2000. Serotonin 5-HT(2) receptors activate localGABA inhibitory

inputstoserotonergicneuronsofthedorsalraphenucleus.BrainRes873,34-45.

Lopez-AlvarezV.M.,ModolL.,NavarroX.,CobianchiS.,2015.Early increasing-intensitytreadmill

exercisereducesneuropathicpainbypreventingnociceptorcollateralsproutinganddisruption

ofchloridecotransportershomeostasisafterperipheralnerveinjury.Pain156,1812-1825.

MantyhP.W.,RogersS.D.,HonoreP.,AllenB.J.,GhilardiJ.R.,LiJ.,DaughtersR.S.,LappiD.A.,Wiley

R.G., Simone D.A., 1997. Inhibition of hyperalgesia by ablation of lamina I spinal neurons

expressingthesubstancePreceptor.Science278,275-279.

MillanM.J.,2002.Descendingcontrolofpain.ProgNeurobiol66,355-474.

MizukamiT.,2004.Immunocytochemicallocalizationofbeta2-adrenergicreceptorsintheratspinal

cordandtheirspatialrelationshipstotyrosinehydroxylase-immunoreactiveterminals.Kurume

MedJ51,175-183.

Modol L., Cobianchi S., Navarro X., 2014. Prevention of NKCC1 phosphorylation avoids

downregulation of KCC2 in central sensory pathways and reduces neuropathic pain after

peripheralnerveinjury.Pain155,1577-1590.

MooreK.A.,KohnoT.,KarchewskiL.A.,ScholzJ.,BabaH.,WoolfC.J.,2002.Partialperipheralnerve

injury promotes a selective loss ofGABAergic inhibition in the superficial dorsal hornof the

spinalcord.JNeurosci22,6724-6731.

NicholsonR.,DixonA.K.,SpanswickD.,LeeK.,2005.NoradrenergicreceptormRNAexpressionin

adultratsuperficialdorsalhornanddorsalrootganglionneurons.NeurosciLett380,316-321.

O'Donnell J., Zeppenfeld D., McConnell E., Pena S., Nedergaard M., 2012. Norepinephrine: a

neuromodulatorthatbooststhefunctionofmultiplecelltypestooptimizeCNSperformance.

NeurochemRes37,2496-2512.

18

OssipovM.H.,DussorG.O.,PorrecaF.,2010.Centralmodulationofpain.JClinInvest120,3779-

3787.

Patterson S.I., HanleyM.R., 1987. Autoradiographic evidence for beta-adrenergic receptors on

capsaicin-sensitiveprimaryafferentterminalsinratspinalcord.NeurosciLett78,17-21.

PearlsteinE.,BenMabroukF.,PfliegerJ.F.,VinayL.,2005.Serotoninrefinesthelocomotor-related

alternationsintheinvitroneonatalratspinalcord.EurJNeurosci21,1338-1346.

PertovaaraA.,2006.Noradrenergicpainmodulation.ProgNeurobiol80,53-83.

PrietoM.,GiraltM.T.,2001.EffectsofN-(2-chloroethyl)-N-ethyl-2-bromobenzylamine (DSP4)on

alpha2-adrenoceptorswhichregulatethesynthesisandreleaseofnoradrenalineintheratbrain.

PharmacolToxicol88,152-158.

StoneL.S.,BrobergerC.,VulchanovaL.,WilcoxG.L.,HokfeltT.,RiedlM.S.,EldeR.,1998.Differential

distributionofalpha2Aandalpha2Cadrenergicreceptorimmunoreactivityintheratspinalcord.

JNeurosci18,5928-5937.

TiongS.Y.,PolgarE.,vanKralingenJ.C.,WatanabeM.,ToddA.J.,2011.Galanin-immunoreactivity

identifiesadistinctpopulationofinhibitoryinterneuronsinlaminaeI-IIIoftheratspinalcord.

MolPain7,36.

Udina E., Cobianchi S., Allodi I., Navarro X., 2011. Effects of activity-dependent strategies on

regenerationandplasticityafterperipheralnerveinjuries.AnnAnat193,347-353.

VaidyaV.A.,MarekG.J.,AghajanianG.K.,DumanR.S.,1997.5-HT2Areceptor-mediatedregulation

ofbrain-derivedneurotrophicfactormRNAinthehippocampusandtheneocortex.JNeurosci

17,2785-2795.

VanSteenwinckelJ.,NogheroA.,ThibaultK.,BrisorgueilM.J.,FischerJ.,ConrathM.,2009.The5-

HT2A receptor ismainly expressed in nociceptive sensory neurons in rat lumbar dorsal root

ganglia.Neuroscience161,838-846.

Vandermaelen C.P., Aghajanian G.K., 1983. Electrophysiological and pharmacological

characterization of serotonergic dorsal raphe neurons recorded extracellularly and

intracellularlyinratbrainslices.BrainRes289,109-119.

XuT.L.,NabekuraJ.,AkaikeN.,1996.ProteinkinaseC-mediatedenhancementofglycineresponse

inratsacraldorsalcommissuralneuronesbyserotonin.JPhysiol496(Pt2),491-501.

Yalcin I.,Choucair-JaafarN.,BenbouzidM.,TessierL.H.,MullerA.,HeinL.,Freund-MercierM.J.,

Barrot M., 2009a. beta(2)-adrenoceptors are critical for antidepressant treatment of

neuropathicpain.AnnNeurol65,218-225.

Yalcin I., Tessier L.H., Petit-Demouliere N., Doridot S., Hein L., Freund-MercierM.J., BarrotM.,

2009b.Beta2-adrenoceptorsareessentialfordesipramine,venlafaxineorreboxetineactionin

neuropathicpain.NeurobiolDis33,386-394.

19

YalcinI.,TessierL.H.,Petit-DemouliereN.,WaltispergerE.,HeinL.,Freund-MercierM.J.,BarrotM.,

2010.Chronictreatmentwithagonistsofbeta(2)-adrenergicreceptorsinneuropathicpain.Exp

Neurol221,115-121.

ZhangF.F.,MoriokaN.,AbeH.,FujiiS.,MiyauchiK.,NakamuraY.,Hisaoka-NakashimaK.,NakataY.,

2016. Stimulation of spinal dorsal horn beta2-adrenergic receptor ameliorates neuropathic

mechanical hypersensitivity through a reduction of phosphorylation of microglial p38 MAP

kinaseandastrocyticc-junN-terminalkinase.NeurochemInt101,144-155.

ZhongF.X., JiX.Q.,TsouK.,1985. IntrathecalDSP4selectivelydepletesspinalnoradrenalineand

attenuatesmorphineanalgesia.EurJPharmacol116,327-330.

Figures

20

Figure 1.Hyperalgesia is reduced by iTR after SNTR. Changes inmechanical (A) andthermal (B)sensory thresholds recordedat themedial (filledcolourbars)and lateral(dottedcolourbars)testsitesinratsafterSNTRthatwereuntrained(-sed)orfolloweddailyiTRfrom3days(-iTR).Valuesarerepresentedasthepercentratiobetweenthemeanipsilateralandthecontralateralpawthresholdat3,7and14dayspost-injury(dpi).*p<0.05,***p<0.001.

21

Figure2.iTRcounteractedthedecreaseofα1AreceptorexpressionindorsalhornafterSNTR.(A)Representativeconfocalimagesofα1Aadrenergicreceptorimmunoreactivityat 14 dpi in the spinal dorsal horn of naïve rats, and in the ipsilateral (ipsi) andcontralateral (contra) dorsal horns of SNTR-sed and SNTR-iTR rats, with graphicrepresentationoflaminaeI-IVasregionsofinterestincludedincapturedimages.Scalebar100μm.(B)Quantificationofα1Aimmunoreactivityintheipsilateralandcontralateraldorsalhornlaminae(I,II,III)ofSNTR-sedandSNTR-iTRratscomparedwithtonaïverats.**p<=0.01,***p<0.001.

22

Figure3. iTRcounteractedthedecreaseofβ2receptorexpressionindorsalhornafterSNTR. (A) Representative confocal images of the β2 adrenergic receptorimmunoreactivityat14dpiinthespinaldorsalhornofnaïverats,andintheipsilateral(ipsi)andcontralateral(contra)dorsalhornsofSNTR-sedandSNTR-iTRrats,withgraphicrepresentationoflaminaeI-IVasregionsofinterestincludedincapturedimages.Scalebar100μm.(B)Quantificationofβ2immunoreactivityintheipsilateralandcontralateraldorsalhornlaminae(I,II,III)ofSNTR-sedandSNTR-iTRratscomparedtonaïverats.**p<=0.01,***p<0.001.

23

Figure4.iTRcounteractedthedecreaseof5HT2AreceptorexpressionindorsalhornafterSNTR. (A) Representative confocal images of the 5HT2A serotonergic receptorimmunoreactivityat14dpiinthespinaldorsalhornofnaïverats,andintheipsilateral(ipsi)andcontralateral(contra)dorsalhornsofSNTR-sedandSNTR-iTRrats,withgraphicrepresentationoflaminaeI-IVasregionsofinterestincludedincapturedimages.Scalebar 100μm. (B) Quantification of 5HT2A immunoreactivity in the ipsilateral andcontralateraldorsalhornlaminae(I,II,III)ofSNTR-sedandSNTR-iTRratscomparedtonaïverats.*p<0.001,**p<=0.01,***p<0.001.

24

Figure5.iTRincreasedtheexpressionof5HT2AreceptorinPAG,DRNandRMafterSNTR.(A)Graphic representation of PAG, RM,DRN and LC nuclei considered as regions ofinterestforquantificationincoronalsections.(B)Representativeconfocalimagesofthe5HT2Aserotonergicreceptorimmunoreactivityat14dpiinPAG,DRNandRMofnaïve,SNTR-sed and SNTR-iTR rats. Scale bar 100μm. (C) Quantification of 5HT2Aimmunoreactivity in PAG,DRN andRMof SNTR-sed and SNTR-iTR rats compared tonaïverats.*p<0.05;**p<0.01.

25

Figure6. iTR increased the expressionofα1A receptor in LCandDRNafter SNTR. (A)Representativeconfocalimagesoftheα1Areceptorimmunoreactivityat14dpiinLC,DRNandRMofnaïve,SNTR-sedandSNTR-iTRrats.Scalebar100μm.(B)Quantificationofα1AimmunoreactivityinLC,DRNandRMofSNTR-sedandSNTR-iTRratscomparedtonaïverats.*p<0.05;**p<0.01.

26

Figure7.iTRincreasedtheexpressionofβ2receptorinLCafterSNTR.(A)Representativeconfocalimagesoftheβ2adrenergicreceptorimmunoreactivityat14dpiinLC,DRNandRMof naïve, SNTR-sed and SNTR-iTR rats. Scale bar 100μm. (B)Quantificationof β2immunoreactivityinLC,DRNandRMofSNTR-sedandSNTR-iTRratscomparedtonaïverats.*p<0.05;**p<0.01.

27

Figure8.Blockadeofβ2or5HT2A receptorsantagonized the iTR inducedhypoalgesiaafterSNTR. (A-B)Changes inmechanical thresholdofSNTR-sedandSNTR-iTRgroupscomparedwithratstreatedwithButoxamine(A,SNTR-sed+BuandSNTR-iTR+Bugroups)orKetanserin(B,SNTR-sed+KeandSNTR-iTR+Kegroups),recordedatmedialtestsitesat3,7and14dayspost-injury(dpi)andrepresentedasthepercentratiobetweentheipsilateralandthecontralateralpaw.*p<0.05;**p<0.01;***p<0.001.

28

Figure 9.Depletion of NE output from locus coeruleus with DSP4 did not affect thehypoalgesiceffectofiTR.(A)ChangesinmechanicalthresholdofSNTR-sedandSNTR-iTRgroupscomparedwithratstreatedalsowithDSP4,recordedatmedialtestsitesat3,7and14dayspost-injury (dpi)and representedas thepercent ratiobetween theipsilateralandthecontralateralpaw.(B)Representativeconfocalimagesofα1Aandβ2adrenergicreceptorsimmunoreactivityat14dpiinLCofSNTR-sed,SNTR-sed+DSP4andSNTR-iTR+DSP4rats.Scalebar100μm.(C)Quantificationofα1Aandβ2immunoreactivityinLCofSNTR-sedandSNTR-iTRgroupscomparedwithratstreatedwithDSP4(SNTR-sed+DSP4andSNTR-iTR+DSP4groups).*p<0.05;**p<0.01;***p<0.001.

29

Figure10.Theexpressionof5HT2AreceptorinlocuscoeruleuswasnormalizedbyiTRandfurtherincreasedbyDSP4injectionafterSNTR.(A)Representativeconfocal imagesofthe5HT2Aserotonergicreceptorimmunoreactivityat14dpiinLCofnaïve,SNTR-sedandSNTR-iTRrats,andthosetreatedwithButoxamineorwithDSP4.Scalebar100μm.(B)Quantificationof5HT2AimmunoreactivityinLCofnaïve,SNTR-sedandSNTR-iTRgroupscomparedwiththesamegroupstreatedalsowithButoxamine(SNTR-sed+BuandSNTR-iTR+Bu groups) or DSP4 (SNTR-sed+DSP4 and SNTR-iTR+DSP4 groups). * p<0.05; **p<0.01.

30

Figure 11. iTR reducedmicrogliosis in locus coeruleus dependent on activation of β2receptor after SNTR. (A) Representative confocal images of Iba1 and BDNFimmunoreactivityat14dpiinLCofnaïve,SNTR-sedandSNTR-iTRgroupscomparedwithgroupstreatedalsowithButoxamine(SNTR-sed+BuandSNTR-iTR+Bugroups).YellowarrowspointtomicroglialcellsexpressingBDNF;redarrowspointtoBDNFlabelingnotcolocalizedwith Iba1. Scalebar100μm. (B)Relativequantificationof Iba1andBDNFimmunoreactivity (IntegratedDensity, InDen) in double-labeled LC samples of naïve,SNTR-sedandSNTR-iTRratscomparedwithratstreatedalsowithButoxamine(SNTR-sed+BuandSNTR-iTR+Bugroups).*p<0.05;**p<0.01;***p<0.001.