VEGF reverts the cognitive impairment induced by a focal traumatic brain injury during the development of rats raised under environmental enrichment

VEGF

reverts

the

cognitive

impairment

induced

by

a

focal

traumatic

brain

injury

during

the

development

of

rats

raised

under

environmental

enrichment

N.

Ortuzar

_,

_I.

_Rico-Barrio

_,

_H.

_Bengoetxea

_,

_E.G.

_{Argando ˜na}

_,

_J.V.

_Lafuente

a_{LaboratoryofClinicalandExperimentalNeuroscience(LaNCE),DepartmentofNeuroscience,FacultyofMedicineandOdontology,Universityofthe}

BasqueCountry(UPV/EHU),BarrioSarriena,E-48940Leioa,Spain

b_{DepartmentofMedicine,UnitofAnatomy,UniversityofFribourg.RueAlbertGockel1,CH-1700FribourgSwitzerland}

The roleof VEGF in the nervous systemis extensive; apart from itsangiogenic effect, VEGF has beendescribedasaneuroprotective,neurotrophicandneurogenicmolecule.Similareffectshavebeen describedforenrichedenvironment(EE).Moreover,bothVEGFandEEhavebeenrelatedtoimproved spatialmemory.Ouraimwastoinvestigatetheneurovascularandcognitiveeffectsof intracerebrally-administered VEGF and enriched environment during the critical period of the ratvisual cortex development.ResultsshowedthatVEGFinfusionaswellasenrichedenvironmentinduced neurovas-cularandcognitiveeffectsindevelopingrats.VEGFadministrationproducedanenhancementduring thelearningprocessofenrichedanimalsandactedasanangiogenicfactorbothinprimaryvisualcortex (V1)anddentategyrus(DG)inordertocounteractminipumpimplantation-induceddamage.Thisfact revealedthatDGvascularizationiscriticalfornormallearning.Incontrasttothisenrichedenvironment actedontheneuronaldensityoftheDGandV1cortex,andresultsshowedlearningenhancementonly innon-operatedrats.Inconclusion,VEGFadministrationonlyhaseffectsifdamageisobserveddueto injury.Oncecontrolvalueswerereached,nofurthereffectsappeared,showingaceilingeffect.Ourresults stronglysupportthatinadditiontoneurogenesis,vascularizationplaysapivotalroleforlearningand memory.

1. Introduction

Growthfactorsanditsreceptorsarekeyregulatorsforcentral nervoussystemdevelopmentaswellasforhomeostasis mainte-nance.Thesemoleculesexerteffectsoverallcomponentsofthe neurogliovascularunit (NVU),suchasvessels,neuronsandglia,

∗ Correspondingauthorat:LaboratoryofClinicalandExperimentalNeuroscience

(LaNCE).DepartmentofNeuroscience.FacultyofMedicineandOdontology.

Univer-sityoftheBasqueCountry(UPV/EHU).BarrioSarrienas/n.E48940,Leioa,Spain.

Tel.:+34946015606;fax:+34946015055.

E-mailaddress:naiara.ortuzar@ehu.es(N.Ortuzar).

and some authorsrefer to thesemolecules asangioglioneurins [1,2].VascularEndothelialGrowthFactor(VEGF)isthearchetypal angioglioneurinandactsonbothvascularandneuraldevelopment. TheroleofVEGFinthenervoussystemisextensive.Apartfrom itsangiogeniceffectindevelopmentalandpathological angiogene-sis[3],ithasbeendescribedasaneuroprotective,neurotrophicand neurogenicmolecule[4–7].TheneuroprotectivefunctionofVEGF appearstobeduetoacombinationofdirectneuroprotectiveeffects and thestimulationofangiogenesis.In addition,previous stud-ieshavedescribeditsneurogeniceffectinthesubventricularzone (SVZ)[8]andinthesubgranularzone(SGZ)ofthedentategyrus[9]. IthasbeenpostulatedthatVEGFpromotestheproliferationand dif-ferentiationofneuronalprecursors,releasingneurotrophicfactors



Published in "%HKDYLRXUDO%UDLQ5HVHDUFK±"

which should be cited to refer to this work.

[10]orexertingadirectmitogenicactiononneuralprecursors[11]. FurtherstudiesdescribedVEGFasadirectmediatorofimproved cognitioninrodents[12–14]andasinducinglong-termsynaptic enhancementinhippocampalneurons[15].Nonetheless,arecent study hasshown that VEGF promoteshippocampus-dependent memoryindependentlyofitseffectsonneurogenesisand angio-genesis,byincreasingthesynapticstrength[16].

Enrichedenvironment(EE)isanexperimentalparadigmdefined as“acombinationofcomplexandinanimatesocialstimulation” [17]thatinducescellular,molecularandbehaviouraleffects,not onlyinstandardconditionsbutalsoinpathologicalones[18,19]. WhereasinitialstudiesshowedthatEEmodifiedcorticalweight [20,21],subsequentstudiesdescribedanincreaseofnuclearand somasizeinneurons,dendriticbranchingorsynapsissize[22–24] among other effects. Moreover, EE also increases neurogene-sis in the hippocampus [13] as it increases the expression of angioglioneurinssuchasVEGForBDNF[25–27].Therefore,EEhas strongeffectsontheplasticityofneuralconnections,especiallyin thevisualcortex[28].

Thepostnataldevelopmentofthevisualcortexismodulatedby environmentalexperience.Sensorymodificationsduringanearly criticalperiodresultinsubstantialplasticityandareacrucialfactor inestablishingthematurecircuitry[29].Inrats,thecriticalperiod forthevisualsystemislocatedbetweenthethirdandthefifth post-natalweeksandthemaximumpeakofexperience-inducedchanges occursduringthefourthandthefifthweeks[30,31].

The aim of the present study was to investigate the neu-rovascular and cognitiveeffects of VEGF infusionand enriched environmentduringthecriticalperiodofthevisualcortexin devel-opingrats.

2. Materialandmethods

2.1. Animalsandhousing

10Long–Evansrats(Harlan,Barcelona,Spain)wereusedforeachexperimental

group.Animalswereraisedindifferentrearingconditions:

1.Standard conditions (SC): rats raised in a standard laboratory cage

(500mm×280mm×140mm)with12-hlight/darkcycle(lightsonat8a.m.)

2. Enriched environment(EE):rats raisedina largecage(720mm× 550mm×

330mm)furnishedwithcolourfultoysanddifferentlyshapedobjects(shelters,

tunnels)thatwerechangedevery2days(with12-hlight/darkcycle;lightsonat

8a.m.).

2.2. Minipumpimplantation

ExperimentswereperformedonLong–Evansrats72hpriortothebeginning

ofthecriticalperiod(P18,30–40g). Animalswereanaesthetizedwithavertin

(1ml/100g)andasagittalincisionwasmademidwaybetweentheeyes.Firstly,

theskinwasretracted,andthentheperiosteum.Subsequentlyasubcutaneous

pocketwasopenedintheanimal’sbackintowhichtheosmoticminipump(Mod.

1004,Alzet,Cupertino,CA,USA)wasinserted.Abraininfusionkit(Mod.AlzetBrain

InfusionKitIII,Alzet)wasfixedtotheskullwithcyanoacrylate.Minipumpswere

connectedviaPVCtubingtotheinfusionkitandthecannulawasimplanted1mm

lateraltothesagittalsutureand1mmanteriortolambdaintothelefthemisphere.

Totaloperatingtimewasapproximately25min.

Differentexperimentalgroupswereused:

a)Agroupofnon-operatedratsascontrol

b) PBSinfusion

c) VEGFinfusion

VEGF(Ref:sc-4571,SantaCruzBiotechnologyInc,Germany)wasadministered

(100ng/ml)atadeliveryrateof0.11␮l/hwiththecannulaplacedinthemiddle

corticallayers.Withineachgroup,thefollowingvisualstimulationconditionswere

studied:

- SC-SC:Ratsraisedinstandardconditions(SC)beforeandafterminipump

implan-tation.

- SC-EE:Ratsraisedinstandardconditions(SC)untilminipumpimplantation(P18)

andinanenrichedenvironment(EE)afterimplantationuntilP46.

Aftersurgeryhadbeencompleted,foodandwaterwereprovidedadlibitum.

MinipumpswereleftinpositionforfourweeksuntilP46(150–200g).

2.3. Ethicsstatements

AllanimalexperimentswereperformedinaccordancewiththeEuropean

Com-munityCouncilDirective(2010/63/EU)andapprovedbytheEthicsCommitteefor

AnimalWelfare(CEBA)oftheUniversityoftheBasqueCountry.

2.4. Morriswatermaze

SpatiallearningandmemorywereassessedinaMorriswatermazetaskduring

thelast11daysoftheinfusionperiod.Inthistask,animalslearnedtolocatethe

positionofthesubmergedplatformusingextra-mazespatialcues.

Thewatermazeconsistedofacircularswimmingpool,170cmdiameterand

0.6mheight,filledwithwater(22±1◦_{C)madeopaquewithnon-toxicwhitepaint.}

Visualcueswerefixedonthewallsconstantlyvisiblefromthepool,whichwas

con-ceptuallydividedintofourquadrantsandhadfourpointsdesignedasstartingpoints

(north,south,eastandwest).Duringfivedays(days1–5),animalsweretestedfor

place-learningacquisitionwiththeescapeplatform(11cmdiameter,47cmheight)

locatedinthemiddleofthesoutheastquadrant,2cmbelowwatersurface.Fourtrials

perdaywereperformed(30min.inter-trialinterval),introducingtheratsrandomly

fromeachofthefourstartingpositionswhilefacingthewall,andallowingthemto

swimuntiltheylocatedtheplatform.Animalsthatfailedtofindtheplatformwithin

120swereguidedtotheplatformandlefttherefor60s,aswerethesuccessfulones.

Onthe7thday,theplatformwasremovedandasingleprobetrialwasperformed

(120s)toassessspatialmemoryretentionfortheplatformlocation.Timespentin

thetargetareaaswellasintheotherquadrantswasmeasured.Nextday,asession

withacuedvisibleplatformwascarriedoutinwhichtheplatformwasmarkedby

ablackflagasacontrolproceduretodiscardmotorormotivationalsensorydeficits

betweenexperimentalgroups.Finally,duringdays9–11,animalsweresubjected

toareversaltest.Thelocationofthehiddenplatformwaschangedtonortheast

andanimalsweretestedforplace-learningacquisitionofthenewlocation.Allthe

trialswererecordedandtracedwithanimagetrackingsystem(SMART,PanlabSL,

Barcelona,Spain)connectedtoavideocameraplacedabovethepool.Thelatency

andtravelledpathlengthtofindtheplatformweremeasuredineachtrialandthe

meanvalueforeachdaywascalculated.

2.5. Fixationandtissueprocessing

Ratswereanaesthetizedwith6%chloralhydrate.Afteranaesthesia,the

can-nulawasremovedandtheanimalsweretranscardiallyperfusedwithasaline

solutionfollowedbyafixativecontaining4%paraformaldehydein0.1M

phosphate-bufferedsalinesolution(PBS).Afterperfusion,brainswerestoredovernightat4◦_C

infreshfixative.Thenextday,thebrainwasstoredin30%sucrosein0.1MPBS

untilthetissuessank.Then,sampleswerecoronallycutat50␮mwithacryotome

andserially-collectedsectionswerekeptin0.1MPBStobehistochemicallyand

immunohistochemicallyprocessedusingthefree-floatingmethod.

2.6. Butyrylcholinesterasehistochemistry

SectionswerehistochemicallyprocessedforButyrylCholinesterasetovisualize

thevascularpattern.Sectionswerewashedtwicein0.1MTris-maleatebuffer(TMB)

(pH6),acetylcholinesterasewasinhibitedinaBW284CS1

(1,5-bis(4-allyldimethyl-ammoniumphenyl)-pentan-3-onedibromide)(Ref:A-9013,Sigma–Aldrich,Spain)

0.05M 20min bath and sections were incubated overnight in the following

incubationsolution:butyrylthiocholineiodide(Ref:108150250,AcrosOrganics,

Barcelona,Spain)1mg/ml,5%sodiumcitrate0.1M,10%coppersulphate30mM,

10%BW284CS10.05mM,10%potassiumferricyanide5mMand65%TMB0.1M.

Thenextday,sectionswerewashedinTMB,mountedongelatine-coatedslides,

dehydratedandcovered.

2.7. Immunohistochemistry

Sections were immunohistochemically processedforNeuN (Ref:MAB377,

ChemiconInternationalInc,Spain,1:400).Firstly,sectionswerewashedin0.1MPBS

andthenincubatedwith4%H2O2/methanolfor20mininordertoquench

endoge-nousperoxidaseactivity.Afterwards,non-specificsiteswereblockedbyaddingBSA

5%intheincubationmediumfor1h.Sectionswerethenincubatedovernightat4◦_C

withprimaryantibody.Thefollowingday,sectionswerewashedin0.1MPBSand

incubatedwithbiotinylatedsecondaryantibodies(EliteABCkit,Vector

Laborato-ries,Burlingame,CA,USA).Theimmunohistochemicalreactionwasrevealedbythe

avidin-biotincomplex(EliteABCkit,VectorLaboratories)usingdiaminobenzidine

asachromogen.Sectionsweremountedongelatine-coatedslides,dehydratedand

covered.

2.8. LEAlectinstaininganddouble-labelimmunofluorescence

Free-floatingsectionsofprimaryvisualcortexandthedentategyruswere

used for both LEA lectin stainingand double-label immunofluorescence. For

microvesselstaining,Lycopersiconesculentum(tomatolectinLEA)histochemistry



Fig.1. EffectsofVEGFadministrationonspatiallearningandmemoryintheMorrisWaterMazetaskofanimalsrearedinstandardconditions(SC-SC)andinenriched

environment(SC-EE).5acquisitions,1cued,1removaland3reversalsessionswereperformedindifferentexperimentalgroups(Control,PBSandVEGF).Escapelatencies

(a,b)andtravelledpathlength(c,d)toreachtheplatformduringthelearningprocess.Percentageoftimespentineachquadrantwhenplatformwasremoved(e,f).Escape

latency(g,h)toreachtheplatformduringthereversaltest.Mean±SEM,n=10.+_{SignificancebetweenVEGFandControl(}+_p<0.05;++_{p<0.01).*SignificancebetweenControl}

andPBS(*p<0.05;**p<0,01;***p<0.001).#_{SignificancebetweenVEGFandPBS(}#_p<0.05;###_p<0.001).

wasused.Sectionswerewashedin0.1MPBS(pH7.4).Then,non-specificsiteswere

blockedbyaddingBSA4%intheincubationmediumfor1hfollowedbyovernight

incubationwithLEA-TRICTantibody(Ref:L9511,Sigma–Aldrich, Spain,1:200)

in1%BSAin0.1MPBScontaining0.5%TritonX-100.Afterwards,sectionswere

washedinPBS.DoubleimmunolabellingusingGFAP(glialfibrillaryacidicprotein),

NeuNandLEAasmarkersforastrocytes,neuronsandendothelialcells,respectively,

andforVEGFwerecarriedouttodetermineco-localizationwitheachcelltype.

Sectionswerewashedin0.1MPBSandwereincubatedwithblockingsolution

(5%BSAin0.1MPBS)for1hfollowedbyovernightincubationwithacocktail

ofprimaryantibodiesin1%BSAin0.1MPBScontaining0.5%TritonX-100.The

antibodiesusedweremonoclonalmouseanti-GFAP(Ref:G-3893,Sigma–Aldrich,

Spain,1:400),monoclonalmouseanti-NeuN(Ref:MAB377,ChemiconInternational

Inc,1:400),andpolyclonalrabbitanti-VEGF(Ref:sc-152,SantaCruzBiotechnology,

Inc.,USA,1:400).Afterrinsing,sectionswereincubatedfor2hwiththefollowing

fluorochromeconjugatedsecondaryantibodies(goatantimouseAlexaFluor®_488,

goatantirabbitAlexaFluor®_{568;Invitrogen,Spain,1:400)in1%BSAin0.1MPBS}

containing0.3%TritonX-100.Hoechst-33258wasaddedtocounterstainthenuclei.

Sectionswererinsed,mountedongelatine-coatedslides,andcover-slippedin

aque-ousmedium.Forallmethods,negativecontrolsinwhichtheprimaryantibodies

wereomittedwereincludedineachstainingrun.Imageswereacquiredforconfocal

fluorescencemicroscopywithanOlympusFluoviewFV500confocalmicroscope

usingsequentialacquisitiontoavoidoverlappingoffluorescenceemissionspectra.

2.9. Morphometricprocedures

Vascularandneuronaldensitiesweremeasuredbytheopticaldissectormethod

inButyrylCholinesterasehistochemistryandNeuNimmunohistochemistry,

respec-tively,withtheaidoftheMercatorImageAnalysissystem (ExploraNova,La

Rochelle,France).Forthispurpose,probesof50␮m×50␮mseparatedby120␮m

werelaunchedintoapreviously-delimitedareacorrespondingtolayerIVofthe

primaryvisualcortexandthegranularcelllayerofthedentategyrus.Positivecells

andvesselswerecountediffoundtobepresentinsidetheprobeandiftheydidnot

touchtheforbiddenXandYaxes.10animalsand10histologicalsectionsperanimal

wereusedforeachexperimentalgroup,and2measurementsweretakenforeach

section(ipsilateralandcontralateralhemisphere),makingatotalof200

measure-mentsforeachgroup.Measurementsofeachsliceofthecortexweretakenandthe

meanvalueperanimalwascalculated.

2.10. Statisticalanalysis

StatisticalanalysiswasperformedusingSPSSstatisticalsoftware(version19.0,

IBM,Spain).Priortoanalysis,datawereexaminedfornormaldistributionusing

theKolmogorov–SmirnovtestandforhomogeneityofvariancesusingLevene’s

test.Theeffectsofexperimentalconditionswereevaluatedusingtheone-way

ANOVAanalysiswiththeBonferronicorrectionforequalvariancesorTamhane’sT2

correctionforunequalvariances.Dataaredescribedasmean±SEM.Significance

wasdeclaredatp<0.05.

3. Results

3.1. EffectsofVEGFadministrationinlearningandmemory 3.1.1. Standardconditions(SC-SC)

Spatial memoryacquisition testing showed that the escape latencyaswellastotalpathlengthoftheVEGF-infusedgroupwere



Table1

Effectofenrichedenvironmentonspatiallearningandmemoryineachexperimentalgroup(Control,PBS,VEGF).Meanvalueofescapelatencyandtravelledpathlength±SEM

andpvaluebetweenenvironmentally-enriched(SC-EE)andstandardcondition(SC-SC)rearedgroupswithineachexperimentalgroup.

Control PBS VEGF

SC-EEvs.SC-SC SC-EEvs.SC-SC SC-EEvs.SC-SC

Latency(s) p Latency(s) p Latency(s) p

Acq1 86±7vs.101±7 0.863 84±7vs.75±7 0.988 68±5vs.73±5 1.000 Acq2 22±6vs.62±6 0.000 56±5vs.53±6 1.000 38±7vs.26±5 0.886 Acq3 18±4vs.27±4 0.227 46±3vs.36±3 0.940 17±3vs.20±3 0.951 Acq4 13±2vs.26±2 0.003 20±2vs.21±2 1.000 13±1vs.16±1 0.795 Acq5 11±2vs.22±2 0.001 20±2vs.20±2 1.000 9±1vs.12±2 0.795 Cue 7±1vs.15±1 0.002 10±1vs.12±1 0.999 8±1vs.10±1 0.510 Rev1 41±5vs.24±5 0.427 33±5vs.39±5 0.994 35±4vs.27±4 0.096 Rev2 7±1vs.16±1 0.000 20±1vs.12±1 0.035 7±1vs.13±1 0.033 Rev3 7±1vs.15±1 0.000 13±1vs.12±1 0.784 5±0.8vs.13±1 0.000 Control PBS VEGF

SC-EEvs.SC-SC SC-EEvs.SC-SC SC-EEvs.SC-SC

Latency(s) p Latency(s) p Latency(s) p

Acq1 508±48vs.525±46 1.000 501±46vs.490±48 1.000 367±46vs.471±46 0.169 Acq2 110±37vs.350±30 0.000 319±33vs.122±38 0.008 153±36vs.131±38 1.000 Acq3 98±23vs.139±22 0.494 232±21vs.173±22 0.931 112±21vs.118±25 1.000 Acq4 86±25vs.153±23 0.013 244±24vs.110±24 0.246 89±22vs.97±24 1.000 Acq5 61±21vs.133±19 0.001 192±18vs.90±19 0.225 57±19vs.64±20 0.998 Cue 49±9vs.82±9 1.000 82±8vs.58±10 1.000 57±8vs.57±9 0.812

decreasedcomparedtonon-operatedcontrolgroupsandthegroup infusedwithPBSduringthelearningprocess(Fig.1a,c).The VEGF-infusedgroupshowedstatistically-significantdifferencesinescape latencycomparedtonon-operatedcontrolgroup(Acq1,p=0.005; Acq2,p=0.001;Acq5,p=0.003)andthePBS-infusedgroup(Acq 2,p=0.009)duringtheacquisitionphase.TheVEGF-infusedgroup alsoshowedadiminishedtotalpathlengthcomparedtothe non-operatedcontrolgroup(Acq2,p=0.001;Acq5,p=0.002),asdid thenon-operatedcontrolgroupcomparedtothePBS-infusedgroup (Acq2,p=0.001).Theprobetrialwasperformed24hafterthefinal dayoftrainingandnodifferencesbetweenexperimentalgroups wereobserved inthetarget quadrant(40%in thecontrol, PBS-infusedandVEGF-infusedgroups;p=1.000)(Fig.1e).Inaddition, whentheplatformlocationwaschanged,nostatisticaldifferences werefoundinescapelatenciesbetweenthestudiedgroups(Fig.1g). 3.1.2. Enrichedenvironment(SC-EE)

TheVEGF-infusedgroupshowedstatistically-decreasedescape latencyandtotalpathlengthcomparedtothePBS-infusedgroup during thelearning process(Fig.1b, d). Betweenthethird and fifthdaysofacquisition,theVEGF-infusedgrouppresentedlower latencycomparedtothePBS-infusedgroup(p=0.000andp=0.010 respectively).Differencesinpathlengthwerealsostatistically sig-nificant (Acq3, p=0.037and Acq 5,p=0.022). In addition, the non-operated controlgroupalsoshowedstatistically-significant lower latency than the PBS-infused group in the third day of acquisition (p=0.001), whereas the differences in path length weresignificantduringthesecond(p=0.002),third(p=0.009)and fifthday(p=0.029) ofthelearningprocess. Onthecuedvisible platformday,thePBS-infusedgroupalsoshoweda statistically-significant differencecompared tothe control group in escape latency(p=0.026)butnodifferenceswereobservedintotalpath length. Furthermore, in the probe trial performance, the PBS-infused groupspent less time in the target quadrant than the VEGF-infused group and the non-operated control group (33% control,22%PBS-infusedand 35%VEGF-infusedgroup;p=0.043 and p=0.295respectively)(Fig.1f).Whentheplatformlocation waschanged,theVEGF-infusedandcontrolgroupsalsopresented significantly lower escape latency than the PBS-infused group (p=0.000inbothcases)(Fig.1h).

3.1.3. Enrichedenvironment(SC-EE)vs.standardconditions (SC-SC)

Duringthelearningprocess,animalsrearedinenriched envi-ronment showed statistically-decreased latency compared to standard-rearedanimalsinthenon-operatedcontrolgroup(Acq 2,p=0.000;Acq4,p=0.003;Acq5,p=0.001andCue,p=0.002). Path length was also statistically diminished in enriched ani-mals compared to the control group that was standard-reared during thelearning process (Acq 2, p=0.000; Acq 4, p=0.013; Acq 5, p=0.001). In the PBS and VEGF-infused groups, no dif-ferencesin theescape latencybetweenSC-EEand SC-SCreared ratswerefound.In contrast,inthePBS-infusedgroup,thepath length for enriched rats was significantly longer than in stan-dardsduringtheseconddayoftheacquisitionphase(p=0.008). Inaddition,whereasinthePBS-infusedgrouptheprobetrial per-formanceshowedstatistically-significantlowertimespentinthe target quadrant of SC-EE reared rats (40% in SC-SC vs. 22% in SC-EE; p=0.002), nosignificantdifferences were foundin non-operatedcontrols(40%inSC-SCand33%inSC-EE,p=1.000)orthe VEGF-infusedgroups(40%inSC-SCand35%inSC-EE;p=1.000). Furthermore,inthenon-operatedcontrolgroup,SC-EErearedrats alsoshowedlowerlatencyduringthereversaltestcomparedto SC-SCrearedrats(p=0.000).Thesameresultswereobtainedin theVEGF-infused group(p=0.033and p=0.000). Incontrast, in thePBS-infusedgroup,lowerescapelatencywasobservedinthe SC-SCrearedgroupduringthereversallearningprocess(p=0.035) (Table1).

3.2. EffectsofVEGFadministrationinthedentategyrus 3.2.1. Standardconditions(SC-SC)

Thevasculardensityofthedentategyrusshowednodifferences amongthestudiedexperimentalgroups.Intheipsilateralcortex, thePBS-infusedgroupshowed3%lowerdensitycomparedtothe non-operated group,whileintheVEGF-infusedgroup,observed density was 3% higher (p=1.000 in both cases).The difference between the PBS-and VEGF-infused groups was 7% (p=1.000), beinghigherinthelatterone.Similarresultswerefoundinthe con-tralateralhemisphere.WhereasthePBS-infusedgrouppresented 5% (p=1.000) lower vascular density, the VEGF-infused group



Fig.2.EffectsofVEGFadministrationonvascularandneuronaldensitiesinthedentategyrus.Quantitativestudybetweendifferentexperimentalgroups(Control,PBSand

VEGF)inanimalsrearedinstandardconditions(SC-SC)andinenrichedenvironment(SC-EE).Horizontalaxesshowtheipsilateralandcontralateralhemispheres.Vertical

axesshownumberofvesselsofthedentategyrus(a,b)andNeuNpositivecells(c,d)permm3_{ofthegranularcelllayerofthedentategyrus.Mean±SEM.***p<0.001.}

showed6%(p=1.000)higherdensitythanthenon-operated con-trolgroup.Finally,theVEGF-infusedgroupshowed11%(p=0.162) highervasculardensitythanthePBS-infusedgroup(Fig.2a).

Neuronaldensitywasquantifiedinthegranularcelllayerof thedentategyrus.Althoughnostatistically-significantdifferences werefound,intheipsilateralhemisphere,thePBS-infusedgroup showed30%(p=0.131)lowerneuronaldensitycomparedto non-operated animals. On the other hand, the VEGF-infused group showed31%(p=0.843)higherdensitycomparedtothePBS-infused groupand8%(p=1.000)lowerthanthenon-operatedgroup.The observedtrendwasmaintainedinthecontralateralhemisphere. WhereasthePBS-infusedgrouppresented26%lessvesselsthan the control group,the VEGF-infusedgroup showed22% higher and9%lowerdensitythanthePBS-infusedandnon-operated con-trolgroups,respectively(p=1.000foralltheexperimentalgroups) (Fig.2c).

3.2.2. Enrichedenvironment(SC-EE)

In animals reared in enriched environment, vascular den-sity showed statistically-significant differences between the studied experimentalgroups. The PBS-infused grouppresented 34% (p=0.000) lower vascular density than the non-operated

control group, whereas the VEGF-infused group showed 56% (p=0.000) higher and 2% lower (p=1.000) density than the PBS-infused and non-operated control groups, respectively. In the contralateral cortex results were similar. The PBS-infused group showed 33% (p=0.000) lower density than the control group. On the other hand, the VEGF-infused group presented 52% (p=0.000) and 1% (p=1.000) higher vascular density than the PBS-infused and non-operated control groups, respectively (Fig.2b).

Neuronaldensityofthegranularcelllayershowedno statis-ticaldifferencesamongtheexperimentalgroupsalthoughsimilar resultscomparedtostandardrearedanimalswerefound.Inthe ipsilateralcortex,thePBS-infusedgrouppresented26%(p=0.513) lower neuronal density thanthe controlgroup,whereas in the VEGF-infused group, the observed neuronal density was 18% (p=0.989)higherand12%(p=0.998)lowerthaninthePBS-infused andnon-operatedcontrolgroups,respectively.Inthecontralateral hemisphere,thePBS-infusedandVEGFinfusedgroupsshowed13% and1%lowerdensity,respectively,comparedtothenon-operated controlgroup(p=1.000inbothcases).Finally,theVEGF-infused grouppresented14%(p=1.000)higherneuronaldensitythanthe PBS-infusedone(Fig.2d).

Table2

Effectofenrichedenvironmentonvascularandneuronaldensitiesofthedentategyrusintheipsilateral(IL)andcontralateral(CL)hemispheresofanimalsrearedinstandard

conditions(SC-SC)andinenrichedenvironment(SC-EE).Averagemeasurementsofstudiedgroups(Meanvalue±SEM),andpercentageofdifferenceandpvaluebetween

SC-EEandSC-SCgroupsofeachexperimentalgroup(Control,PBS,VEGF).

Control PBS VEGF

SC-SC SC-EE SC-SC SC-EE SC-SC SC-EE

Vasculardensity IL 19,307±726 20,452±839 18,651±649 13,352±750 20,027±634 20,896±619

CL 19,172±635 19,871±658 18,238±564 13,180±658 20,289±550 20,030±537

Neuronaldensity IL 148,552±15,315 209278±18,471 104,213±18,471 155,138±15,818 136,763±13,061 182,811±14,439

CL 149,081±13,363 165,371±17,563 110,607±17,563 143,400±14,127 134,844±13,025 163,835±13,363

Vasculardensity(SC-EEvs.SC-SC) Neuronaldensity(SC-EEvs.SC-SC)

Control PBS VEGF Control PBS VEGF

%Dif IL 5 −28 4 41 48 33 CL 3 −28 1 10 30 22 p IL 1.000 0.000 1.000 0.199 0.261 0.607 CL 1.000 0.000 1.000 1.000 1.000 1.000 

http://doc.rero.ch

Fig.3.EffectsofVEGFadministrationonvascularandneuronaldensitiesintheprimaryvisualcortex.Quantitativestudybetweendifferentexperimentalgroups(Control, PBSandVEGF)inanimalsrearedinstandardconditions(SC-SC)andinenrichedenvironment(SC-EE).Horizontalaxesshowtheipsilateralandcontralateralhemispheres. Verticalaxesshownumberofvessels(a,b)andNeuNpositivecells(c,d)permm3_{oftheprimaryvisualcortex.Mean±SEM.*p<0.05;***p<0.001.}

3.2.3. Enrichedenvironment(SC-EE)vs.standardconditions (SC-SC)

Inanimalsrearedinenrichedenvironment,vasculardensitywas 5%(p=1.000)and3%(p=1.000)higherinthenon-operatedcontrol group(ipsilateralandcontralateralhemispheres,respectively).The PBS-infused group presented statistically significant-differences comparedtoSC-SCrearedratsinbothhemispheres(28%,p=0.000 intheipsilateralcortex;28%,p=0.000inthecontralateralcortex). In theVEGF-infused group,noeffects of enriched environment werefound, as SC-EEreared ratsshowed4%(p=1.000) and 1% (p=1.000)highervasculardensitythanSC-SCrearedratsinthe ipsilateralandcontralateralcortex,respectively(Table2).

Althoughnostatisticaldifferenceswerefoundintheneuronal densityofthegranularcelllayerinthedentategyrus,ratsraisedin SC-EEconditionspresentedhigherneuronaldensityinall exper-imentalgroups.Intheipsilateralhemisphere,41%,48%and33% higherneuronaldensitywasfoundinSC-EErearedanimalsinthe non-operatedcontrols(p=0.199),thePBS-infused(p=0.261)and theVEGF-infused(p=0.607)groups,respectively.Inthe contralat-eralhemisphere,SC-EErearedratsalsoshowed10%,30%and22% higherneuronaldensityinthecontrol(p=1.000),PBS(p=1.000) andVEGF-infused(p=1.000)groups(Table2).

3.3. EffectsofVEGFadministrationinV1cortex 3.3.1. Standardconditions(SC-SC)

No significantdifferences wereobserved invascular density inanimalsrearedinstandardconditions.Intheipsilateral hemi-sphere,thePBS-infusedgroupshowed5%(p=1.000)lessvessels thannon-operatedcontrols.TheVEGF-infusedgroupshoweda5% (p=1.000)and10%(p=0.464)highervasculardensitythan non-operatedcontrolsandthePBS-infusedgroup.Inthecontralateral hemisphere,similarresultswerefound.WhereasthePBS-infused grouppresented1%(p=1.000)lessvesselsthannon-operated con-trols,theVEGF-infusedgrouppresenteda 4%(p=1.000)and 5% (p=0.998)highervascular densitycompared tothecontroland PBSgroups,respectively.Noneofthedifferenceswerestatistically significant(Fig.3a).

On the other hand, neuronal density in the ipsilateral hemisphereshowednodifferencebetweenPBS-infusedand non-operated controls (0.38%; p=0.977). The VEGF-infused group

presented5%higherneuronaldensitythanboththecontroland PBSgroups(p=0,998andp=1.000respectively).Nonetheless,the VEGFgroupshowedstatistically-significantmoreneuronsthanthe non-operatedcontrols(20%;p=0.000)andthePBS-infused(15%; p=0.030)groupinthecontralateralhemisphere.ThePBS-infused group also showed 5% more neurons than controls (p=1.000) (Fig.3c).

3.3.2. Enrichedenvironment(SC-EE)

In animals reared in enriched environment after minipump implantation, significantdifferences werefoundinthevascular densityoftheprimaryvisualcortex.Intheipsilateralhemisphere, thePBS-infusedgroupshowed40%(p=0.000)lowervascular den-sitythannon-operatedcontrolsand79%(p=0.000)lowerthanthe VEGF-infusedgroup.Thus,7%higherdensitywasobservedinthe VEGF-infusedgroupcomparedtothenon-operatedcontrolgroup, resultinginastatistically-insignificantdifference.Similarresults were foundin thecontralateral hemisphere. Whereasthe PBS-infusedgroupshowed37%(p=0.000)lowervasculardensitythan non-operatedcontrols,theVEGF-infusedgroupshoweda signif-icantlyhigherdensitythanboththecontrol(15%;p=0.014)and PBS-infusedgroups(83%;p=0.000)(Fig.3b).

Intheipsilateralcortex,lowerneuronaldensitycomparedto enriched-environmentrearedcontrolswasfoundinboththePBS andVEGF-infusedgroups,(20%,p=0.001and19%,p=0.074 respec-tively).Therefore,thePBSandVEGF-infusedgroupsshowedsimilar neuronaldensitytoeachother(2%;p=1.000).Inthecontralateral hemisphere, the PBS-infusedgroup alsoshowed a statistically-significant lower neuronal density than non-operated controls (11%; p=0.044). On the contrary,the VEGF-infused group pre-sented 3% less neurons than the control group, resulting in a non-significant difference (p=1.000). Finally, the VEGF-infused groupshowed8%(p=0.477)higherdensitythanthePBS-infused group(Fig.3d).

3.3.3. Enrichedenvironment(SC-EE)vs.standardconditions (SC-SC)

In thevascular densityof non-operatedcontrols, no statisti-caldifferenceswerefoundinratsrearedinenrichedenvironment compared to rats reared in standard conditions (p=1.000 in bothhemispheres).IntheVEGF-infusedgroup,SC-EErearedrats alsopresented 9%(ipsilateral;p=0.564) and10%(contralateral;



Table3

Effectofenrichedenvironmentonvascularandneuronaldensitiesoftheprimaryvisualcortexintheipsilateral(IL)andcontralateral(CL)hemispheresofanimalsrearedin

standardconditions(SC-SC)andinenrichedenvironment(SC-EE).Averagemeasurementsofstudiedgroups(Meanvalue±SEM),andpercentageofdifferenceandpvalue

betweenSC-EEandSC-SCgroupsofeachexperimentalgroup(Control,PBS,VEGF).

Control PBS VEGF

SC-SC SC-EE SC-SC SC-EE SC-SC SC-EE

Vascular density IL 25,940±798 27,677±798 24,600±788 16,496±724 27,153±872 29,572±758 CL 26,432±794 26,591±850 26,124±863 16,623±765 27,404±850 30,497±716 Neuronal density IL 68,281±2465 87,284±2662 68,547±2772 69,523±3170 71,966±2732 70,787±2903 CL 68,505±2111 86,998±2369 71,650±2629 77,157±2338 82,571±2297 83,573±2102

Vasculardensity(SC-EEvs.SC-SC) Neuronaldensity(SC-EEvs.SC-SC)

Control PBS VEGF Control PBS VEGF

%Dif IL 2 −33 9 28 1 −1

CL 0.006 −36 10 27 7 1

p IL 1.000 0.000 0.564 0.001 1.000 1.000

CL 1.000 0.000 0.179 0.000 1.000 1.000

p=0.179)highervasculardensity.Incontrast,inthePBS-infused groupSC-EErearedratsshowedstatisticallylowervasculardensity comparedtoSC-SCrearedratsinbothhemispheres(33%inthe ipsi-lateralhemisphere,p=0.000;36%inthecontralateralhemisphere, p=0.000)(Table3).

Neuronal density of non-operated controls presented a statistically-significantincreaseinratsrearedinSC-EEconditions (28%intheipsilateralhemisphere,p=0.001;27%inthe contralat-eralhemisphere, p=0.000). No effects ofenriched environment were found in the PBS and VEGF-infused groups. In the PBS-infused group,SC-EE reared ratsshowed1%(p= 1.000) and 7% (p=1.000)higherneuronaldensitycomparedtoSC-SCrearedrats intheipsilateralandcontralateralhemispheresrespectively.The VEGF-infusedgrouppresented1%lower(p=1.000)and1%higher (p=1.000)neuronaldensityinratsrearedinanenriched environ-ment,thisbeingastatistically-insignificantdifference(Table3). 3.4. VEGFco-expressioninDG

VEGFexpressionwassimilarinthedentategyrusofall exper-imentalgroupsandmostoftheVEGF-positivecells co-localized withGFAP (Fig.4a-c).VEGF/NeuN immunofluorescencedidnot showdifferencesamongtheexperimentalgroups,althoughsome VEGF-positiveneuronswerefound(Fig.4d–f).Noco-localization withLEAwasobserved(Fig.4g–i).Therefore,VEGFexpression cor-respondstoastrocytesinthedentategyrusatP46.

3.5. VEGFco-expressioninV1cortex

Double immunofluorescenceof VEGF/GFAPshowed a signif-icantly lowerexpression of both antigens in non-operated rats comparedtothePBSand VEGFgroups,higherexpressionbeing observedinthelatter(Fig.5a–c).MostoftheVEGF-positivecells correspondedtoastrocytes,astheyco-localizedwithGFAP.In addi-tion,VEGF/NeuNimmunolabellingconfirmedthatco-localization withNeuNpositivecellsislowerinbothinfusedgroups(PBSand VEGFgroups)comparedtothecontrolgroup(Fig.5d–f).No co-localization withLEA wasobserved in anyof theexperimental groups(Fig.5g–i).Thus,themajorityofVEGF expression corre-spondstoastrocytesintheprimaryvisualcortexatP46.

4. Discussion

4.1. EffectsofVEGFadministration

VEGF administration during the critical period of the visual cortexshowsdifferentresponsesdependingontherearing envi-ronment.Angiogenic,neuroprotectiveandneurogenicproperties

havebeendescribedforVEGFinthebrain,aswellasitsinfluence overhippocampus-dependentmemory[12,14,15].

Ourresultsrevealedanimprovedtrendduringthelearning pro-cessinducedbyVEGFadministrationcomparedtonon-operated and PBS-infusedanimalsraisedinstandardconditions although nobeneficialeffectswerefoundinlong-termspatialmemoryand reversal learning. Since thedentate gyrusis theprimary affer-entintothehippocampus,wecarriedoutaquantitativeanalysis of vascular andneuronal densities ofDG.Resultsdidnot show statistically-significantdifferencesinvascularandneuronal den-sities of the dentate gyrusin the studied experimentalgroups rearedinstandardconditions.Nevertheless,intheVEGF-infused group,anincreaseinneuronaldensityofthegranularcelllayer ofDGwasobserved.Recently,ithasbeenshownthatVEGF pro-moteshippocampus-dependentmemory,increasingthesynaptic strength,independentlyofitseffectsonneurogenesisand angio-genesis[16],whichcouldexplainthelackofneurovascularchanges foundinthedentategyrus.

Inanimalsrearedinenrichedenvironment,dataalsoshoweda significantimprovementduringlearningacquisitioninthe VEGF-infused group compared to the PBS-infused and non-operated control groups.Moreover,whereas PBS-infusionimpaired long-term spatial memory, VEGF administration counteracted this negativeeffectandshowedanimprovedreversaltestaswell.It hasbeenpreviously describedthat VEGFimprovescognitionin rodentsandourdatarevealedsimilarresultswhenadministered intracorticallyindevelopingrats.Furthermore,quantitativeresults showedthat PBS-infusioninducedanegativeeffect onvascular densityinthedentategyrus,whichwasrevertedwithVEGF admin-istrationinenrichedanimals.Consistentwithpreviousstudiesthat demonstratedthatangiogenesiswascriticaltonormallearningand memory[32],ourdataalsosuggestthatthevascularcomponentof DGcouldbedeterminant,asincreasedvasculardensitywasfound intheVEGF-infusedgroup.VEGFcanactivateneurogenesisby tar-getingneuronsdirectlyorprotectingthemindirectlythroughits angiogeniceffect[14,33].Nonetheless,inadditiontoaugmenting thenumberofnewborncellsinthebrain,newlyformedvesselsare neededfortheirmetabolicandtrophicdemands[34].Thus,several studieshaveindicatedthatneurogenesisisnotsufficientbyitself formemoryenhancement[16,35,36].Ourresultssuggestthat vas-cularizationisrelatedtonormallearningalthoughVEGFcouldalso increaseconnectivitybetweenmatureneuronstoenhancelearning aswell.

On the other hand, the absence of VEGF effects in animals reared in standard conditions but not in enriched ones could be due to enhanced experience after minipump implantation, as theseanimalswere rearedin anenriched environment only after thesurgery. Interaction ofexperience withtime sensitive



Fig.4. VEGF/GFAP,VEGF/NeuNandVEGF/LEAco-expressioninthedentategyrus.ConfocalimagesofdoubleimmunolabellingofVEGFandGFAP(a–c),VEGFandNeuN(d–f),

andVEGFandLEA(g–i)todetermineifVEGFpositivitycorrespondstoastrocytes,neuronsorendothelialcellsrespectively.BluecorrespondstoHoeschst33258positive

nuclei.Scalebar=100␮m(a–i)and20␮m(a_–i_).

injury-inducedchangeshasbeendescribedasshapingbrain reor-ganizationinalong-lastingmanner,includingareasproximalto theinjurythatmightbeparticularlyimportantforfunctional out-come[37].Therefore,ourresultssuggestthatVEGFinfusionmight aidrecoveryfromminipumpimplantation-induced-lesions. 4.2. DG-V1cortexconnection

WhereasmoststudiestodetermineVEGFfunctioninlearning and memory used intracerebroventricular [8,38,39] or intra-hippocampal administration [40], in ourwork, VEGF has been administeredimmediatelyanterior totheprimary visualcortex andadirectarrivaltotheDGcouldbediscarded.Thevisual cor-texprovidesa crucialsensoryinputtothehippocampusandis

akeycomponentforthecreationofvisuospatialmemories[41]. Diversestudieshavesuggestedthatexperience-dependent synap-ticplasticityoccursintheadultprimaryvisualcortex,whichcould constitute akey element inneocortico-hippocampal transfer of information[41,42].Thisrelationshipbetweenlearningandvisual corticalchangeshasalsobeenobservedinrodents[43,44].In addi-tion to the hippocampusand theparahippocampal region, the enthorinalcortex[45]andpostrhinalcortex[46]alsohavea recip-rocalconnectionwiththevisualsystem.Ontheotherhand,the retrosplenialcortexhasalsobeendemonstratedtocontributeto memoryandnavigation,andbeinganatomicallyclosetoV1,adirect effectofVEGFshouldnotbediscarded[47].

Our datademonstratethat effectsof VEGFadministration in theDGarealsopresentinthevisualcortex,neartothecannula



Fig.5.VEGF/GFAP,VEGF/NeuNandVEGF/LEAco-expressionintheprimaryvisualcortex.ConfocalimagesofdoubleimmunolabellingofVEGFandGFAP(a–c),VEGFand

NeuN(d–f),andVEGFandLEA(g–i)todetermineifVEGFpositivitycorrespondstoastrocytes,neuronsorendothelialcellsrespectively.BluecorrespondstoHoeschst33258

positivenuclei.Scalebar=200␮m(a–i)and20␮m(a_–i_).

placement.Asinthedentategyrus,noeffectofVEGF administra-tionwasobservedintheneuronalandvascular densitiesofthe primaryvisualcortexinstandardconditions.Previousstudiesin ourgroupdemonstratedthat minipumpimplantation produced negativeeffectsintheV1cortexthatwererevertedafter1weekof VEGFadministration[48,49].Inthiscase,wecarriedoutachronic VEGForvehicleadministrationandthereforeachronicminipump implantationduringfourweeks.Thistimelapseisenoughtorepair thesurroundingtissue,thusrecoveringfrominjuriesinducedby minipump implantation.In addition,focalsmalllesionssuchas theoneproducedbycannulaimplantationcanberepairedeasily [50].

Incontrast,inanimalsrearedinenrichedenvironment,the PBS-infusedgroupshowedasignificantlylowervasculardensity,similar

tothatobservedinDG.VEGFinfusioncounteractedthisnegative effectonmicrovascularizationandreachednon-operatedcontrol values.ThePBS-infusiongroupalsopresentedasignificantlylower neuronaldensitycomparedtonon-operatedcontrols.VEGF admin-istration did notrevert totheneuronal density in the primary visualcortexasboththePBSandVEGF-infusedgroupspresented values similartothestandard-rearedcontrolgroup.Ithasbeen proposedthatthereisaceilingeffectofVEGFandFGF-2for induc-ingchangesonneurovascularcomponentsafterinjury[39].This factcouldexplainthatVEGFactsonlyifthestudiedparametersare lowerthanphysiologicalvalues.

Therefore,theconnectionsdescribedabovecouldexplainthe factthatchangesinDGarereflectedintheprimaryvisualcortex andviceversa.

4.3. Enrichedenvironmentvs.standardconditions

InadditiontoVEGFadministration,anenrichedenvironmentby itselfpresentsdirecteffectsonstudiedneurovascularcomponents andonthelearningprocess.Exposuretoanenrichedenvironment hasbeenshown toinducemorphological changes inthe brain, bothinnormalandpathologicalconditions[51,52].These morpho-logicalchangesincludestructuralreorganization,especiallyinthe visualcortex[53],whichisassociatedwithimprovedlearningand memoryandenhancedneuralplasticity[54].

As hasbeenobserved beforein the adultbrain [53,54], our resultshaveshownthatenrichmentexposureuntilP46inthe non-operatedcontrolgroupenhancesthelearningprocessandreversal learning comparedtostandard-rearedrats.Enrichment-induced improvement of learning hasbeen related to increased synap-tic plasticityorhippocampalneurogenesis,among othereffects [9,55,56].Evenifnodifferencesinvasculardensitywereobserved, ourresultsshowedthatenrichmentinducedanenhancementin thegranularcelllayerofthedentategyrus.Althoughithasbeen reportedthatenrichmentinducesasurvival-promotingeffectin thedentategyrus[9],thePBSandVEGF-infusedgroupsdidnot show any improvement during the learning process. Nonethe-less, the enriched environment enhances reversal learning in VEGF-infusedanimalswhereasinPBS-infusedones,itimpairsthe performance. Infact, recentstudieshave suggestedthat neuro-genesisisnottheuniquesupporteroftheenhancementinduced by enrichment [35,36] since newly proliferated neurons need 3–4weekstobefunctionallyincorporatedintoexistinglearning networks [57,58]. In agreementwith these studies,our results showed that in the VEGF and PBS-infused groups, rearing in enrichedenvironmentalsoinducedanincreaseinneuronal den-sityofthegranularcelllayerin thedentategyrusthatwasnot accompaniedbyanimprovementduringthelearningprocess.

On theotherhand,previousstudieshavedemonstratedthat rearing in an enriched environment from birth to adulthood increasesvascular densityintheratvisualcortex,and thatthis is correlatedwithanincreaseinVEGF expression[25].In addi-tion, it was observed that exposure to EE from P18 onwards increasedmicrovasculatureatP25 [48].Nevertheless,rearingin EEduringtheentirecriticalperiod(fromP18toP46)inducedan increasein neuronaldensity.Enrichmentelicitsneuroprotective responsesasitinduceschangesingenesrelatedto neuroprotec-tion[59],andexpressionofangioglioneurinssuchasBDNF[60] andVEGF[25]areincreasedaswell.Therefore,enrichment elic-itsneurorescuereffectsifappliedthroughoutthecriticalperiod.In contrast,nobeneficialeffectswerefoundintheV1ofthePBSand VEGF-infusedgroups,probablyduetotheminipump implantation-induced lesion. Enrichment-induced effects are different when appliedinnormalorpathologicalsituations[61–63].

4.4. VEGFco-expressioninV1andDG

Inthedevelopingbrain,VEGFisinitiallyproducedbyneurons. AtP13neuronalexpressionofVEGFstartstodiminishand astro-cyticexpressionbecomesmoreevidentuntillocalizationofVEGF switchesfrombeingpredominantlyneuronaltoglialatP24[25]. However,inthehypoxicbrain,highlevelsofneuronalandglial VEGFaremaintaineduntilP33[63].OurresultsshowedthatVEGF immunonoreactivityco-localizedwithastrocytesandneuronsin theV1cortexalthoughmostVEGF-positivecellscorrespondedto astrocytesinallexperimentalgroups.

Ontheotherhand,whereasourresultsshowedthatVEGFis expressedinastrocytesandneuronsofthegranularcelllayerand thepolymorphiclayerofthedentategyrus,ashasbeenpreviously reported[12],noco-localizationwithendothelialcellshasbeen found.Moreover,mostVEGFexpressioncorrespondstoastrocytes.

Someauthorshaveproposedthatastrocytesareanimportantniche for neurogenesis[64],and thatastrocytes-derivedVEGF expres-sioncouldinduceanincreaseinangiogenesisastheyarelocalized adjacenttocerebralvessels[12].NeuronalexpressionofVEGFhas alsobeenproposedtoplayanimportantroleintheregulationof neurogenesis[12].

5. Conclusions

VEGF infusion as well as enriched environment induces neurovascular and cognitive effects in developing rats. VEGF administration produces an enhancement during the learning process and actsasanangiogenic factorin ordertocounteract minipumpimplantation-induceddamage,revealingthatDG vas-cularization is critical for normal learning. On the other hand, enrichedenvironmentactsontheneuronalcomponentofDGand V1cortex.Moreover,resultsshowedlearningenhancementonly innon-operatedrats,leadingtotheconclusionthatinadditionto neurogenesis,vascularizationplaysapivotalroleforlearningand memory.

Acknowledgements

ThisstudyhasbeensupportedbySAIOTEK(IndustryDept., Gov-ernmentoftheBasqueCountry),IT/491/10(BasqueGovernment) andUFI11/32(UPV/EHU).N.Ortuzarissupportedbyapredoctoral grant (FPU)fromtheSpanishMinistryofEducationand a post-doctoralgrantoftheUniversityoftheBasqueCountry(UPV/EHU). Confocal microscopyanalysiswasperformedatthe“Serviciode MicroscopiaAnalíticaydeAltaResoluciónenBiomedicina”(High ResolutionAnalyticalMicroscopyServicesforBiomedicine)atthe UniversityoftheBasqueCountry(UPV/EHU).

AppendixA. Supplementarydata

Supplementarydataassociatedwiththisarticlecanbefound,in theonlineversion

References

[1]ZacchignaS,LambrechtsD,CarmelietP.Neurovascularsignallingdefectsin

neurodegeneration.NatureReviewsNeuroscience2008;9:169–81.

[2] Argando ˜naEG,BengoetxeaH,OrtuzarN,BulnesS,Rico-BarrioI,LafuenteJV.

VascularEndothelialGrowthFactor:Adaptativechangesinthe

Neurogliovas-cularunit.CurrentNeurovascularResearch2012;9:72–81.

[3]FerraraN,GerberHP,LeCouterJ.ThebiologyofVEGFanditsreceptors.Nature

Medicine2003;9:669–76.

[4]Jin K, Zhu Y, Sun Y, Mao XO, Xie L, Greenberg DA. Vascular

endothe-lial growth factor (VEGF) stimulates neurogenesisin vitro and in vivo.

ProceedingsoftheNationalAcademyofSciencesoftheUnitedStatesofAmerica

2002;99:11946–50.

[5]RosensteinJM,KrumJM.NewrolesforVEGFinnervoustissue–beyondblood

vessels.ExperimentalNeurology2004;187:246–53.

[6]Storkebaum E, Lambrechts D, Carmeliet P. VEGF: once regarded as a

specific angiogenic factor, now implicated in neuroprotection. BioEssays

2004;26:943–54.

[7]KayaD,Gürsoy-OzdemirY,YemisciM,TuncerN,AktanS,DalkaraT.VEGF

pro-tectsbrainagainstfocalischemiawithoutincreasingblood-brainpermeability

whenadministeredintracerebroventricularly.JournalofCerebralBloodFlow

andMetabolism2005;25:1111–8.

[8]SunY,JinK,ChildsJT,XieL,MaoXL,GreenbergDA.Vascularendothelialgrowth

factor-B(VEGFB)stimulatesneurogenesis:evidencefromknockoutmiceand

growthfactoradministration.DevelopmentalBiology2006;289:329–35.

[9]KempermannG,KuhnHG,GageFH.Morehippocampalneuronsinadultmice

livinginanenrichedenvironment.Nature1997;386:493–5.

[10]Louissaint Jr A, Rao S, Leventhal C, Goldman SA. Coordinated

interac-tionofneurogenesisandangiogenesisintheadultsongbirdbrain.Neuron

2002;34:945–60.

[11]Greenberg DA, Jin K. From angiogenesis to neuropathology. Nature

2005;438:954–9.

[12] Cao L,JiaoX,ZuzgaDS,LiuY,FongDM,YoungD,etal.VEGFlinks

hip-pocampalactivitywithneurogenesis,learningandmemory.NatureGenetics

2004;36:827–35.



[13]DuringMJ,CaoL.VEGF,amediatoroftheeffectofexperienceonhippocampal

neurogenesis.CurrentAlzheimerResearch2006;3:29–33.

[14]PlaschkeK,StaubJ,ErnstE,MartiHH.VEGFoverexpressionimprovesmice

cog-nitiveabilitiesafterunilateralcommoncarotidarteryocclusion.Experimental

Neurology2008;214:285–92.

[15]KimBW,ChoiM,KimYS,ParkH,LeeHR,YunCO,etal.Vascularendothelial

growthfactor(VEGF)signalling regulateshippocampalneuronsby

eleva-tionofintracellularcalciumandactivationofcalcium/calmodulinprotein

kinaseIIandmammaliantargetofrapamycin.CellularSignalling2008;20:

714–25.

[16]LichtT,GoshenI,AvitalA,KreiselT,ZubedatS,EavriR,etal.Reversible

modu-lationsofneuronalplasticitybyVEGF.ProceedingsoftheNationalAcademyof

SciencesoftheUnitedStatesofAmerica2011;108:5081–6.

[17] RosenzweigMR,BennetEL,Hebert M,MorimotoH.Socialgrouping

can-notaccountforcerebraleffectsofenrichedenvironments.BrainResearch

1978;153:563–76.

[18]NithianantharajahJ, HannanAJ.Theneurobiologyof brainandcognitive

reserve:Mentalandphysicalactivityasmodulatorsofbraindisorders.Progress

inNeurobiology2009;89:369–82.

[19] Martínez-CuéC,RuedaN,GarcíaE,DavissonMT,SchmidtC,FlórezJ.Behavioral,

cognitiveandbiochemicalresponsestodifferentenvironmentalconditionsin

maleTs65Dnmice,amodelofDownsyndrome.BehaviouralBrainResearch

2005;163:174–85.

[20]BennettEL,RosenzweigMR,DiamondMC.Ratbrain:effectsofenvironmental

enrichmentonwetanddryweights.Science1969;163:825–6.

[21]DiamondMC,InghamCA,JohnsonRE,BennettEL,RosenzweigMR.Effectsof

environmentalmorphologyofratcerebralcortexandhippocampus.Journalof

Neurobiology1976;7:75–86.

[22]GreenoughWT, Volkmar FR. Patternof dendritic branching in occipital

cortexof rats reared incomplex environments. Experimental Neurology

1973;40:491–504.

[23] RosenzweigMR,BennettEL,DiamondMC,WuSY,SlagleRW,SaffranE.

Influ-encesofenvironmentalcomplexityandvisualstimulationondevelopmentof

occipitalcortexinrat.BrainResearch1969;14:427–45.

[24]FahertyCJ,KerleyD,SmeyneRJ.AGolgi-Coxmorphologicalanalysisofneuronal

changesinducedbyenvironmentalenrichment.BrainResearchDevelopmental

BrainResearch2003;141:55–61.

[25]BengoetxeaH,Argando ˜naEG,LafuenteJV.Effectsofvisualexperienceon

vas-cularendothelialgrowthfactorexpressionduringthepostnataldevelopment

oftheratvisualcortex.CerebralCortex2008;18:1630–9.

[26]IckesBR,PhamTM,SandersLA,AlbeckDS,MohammedAH,GranholmAC.

Long-termenvironmentalenrichmentleadstoregionalincreasesinneurotrophin

levelsinratbrain.ExperimentalNeurology2000;164:45–52.

[27] PhamTM,WinbladB,GranholmAC,MohammedAH.Environmental

influ-encesonbrainneurotrophinsinrats.PharmacologyBiochemistryandBehavior

2002;73:167–75.

[28]CanceddaL,PutignanoE,SaleA,ViegiA,BerardiN,MaffeiL.Accelerationof

visualsystemdevelopmentbyenvironmentalenrichment.Journalof

Neuro-science2004;24:4840–8.

[29]HooksBM,ChenC.Criticalperiodsinthevisualsystem:changingviewsfora

modelofexperience-dependentplasticity.Neuron2007;56:312–26.

[30] FagioliniM,PizzorussoT,BerardiN,DomeniciL,MaffeiL.Functional

post-nataldevelopmentoftheratprimaryvisualcortexandtheroleofvisual

experience:darkrearingandmonoculardeprivation.VisionResearch1994;34:

709–20.

[31]FagioliniM,HenschTK.Inhibitorythresholdforcritical-periodactivationin

primaryvisualcortex.Nature2000;404:183–6.

[32]KerrAL,SteuerEL,PochtarevV,SwainRA.Angiogenesisbutnot

neurogen-esisiscriticalfornormallearningandmemoryacquisition.Neuroscience

2010;171:214–26.

[33]Marti HJ, Bernaudin M, Bellail A, Schoch H, Euler M, Petit E, et al.

Hypoxia-induced vascular endothelialgrowth factorexpression precedes

neovascularizationafter cerebralischemia.AmericanJournalofPathology

2000;156:965–76.

[34]PalmerTD,WillhoiteAR,GageFH.Vascularnicheforadulthippocampal

neu-rogenesis.JournalofComparativeNeurology2000;425:479–94.

[35]MeshiD,DrewMR,SaxeM,AnsorgeMS,DavidD,SantarelliL,etal.Hippocampal

neurogenesisisnotrequiredforbehaviouraleffectsofenvironmental

enrich-ment.NatureNeuroscience2006;9:729–31.

[36]ShorsTJ,TownsendDA,ZhaoM,KozorovitskiyY,GouldE.Neurogenesismay

relatetosomebutnotalltypesofhippocampal-dependentlearning.

Hip-pocampus2002;12:578–84.

[37]Jones TA, Jefferson SC. Reflections of experience-expectant development

in repair on the adult damaged brain. Developmental Psychobiology

2011;53:466–75.

[38]WangY,GalvanV,GorostizaO,AtaieM,JinK,GreenbergDA.Vascular

endothe-lialgrowthfactorimprovesrecoveryofsensorimotorandcognitivedeficits

afterfocalcerebralischemiaintherat.BrainResearch2006;1115:186–93.

[39]Thau-ZuchmanO,ShohamiE,AlexandrovichAG,LekerRR.Combinationof

vascularendothelialandfibroblastgrowthfactor2forinductionof

neuro-genesisandangiogenesisaftertraumaticbraininjury.JournalofMolecular

Neuroscience2012;47:166–72.

[40] PatiS,OrsiSA,MooreAN,DashPK.Intra-hippocampaladministrationofthe

VEGFreceptorblockerPTK787/ZK222584impairslong-termmemory.Brain

Research2009;1256:85–91.

[41]TsanovM,Manahan-VaughanD. Synapticplasticityfromvisualcortexto

hippocampus:systemsintegrationinspatialinformationprocessing.

Neuro-scientist2008;14:584–97.

[42] HeynenAJ,BearMF.Long-termpotentiationofthalamocorticaltransmission

intheadultvisualcortexinvivo.JournalofNeuroscience2001;21:9801–13.

[43] FrenkelMY,SawtellNB,DiogoAC,YoonB,NeveRL,BearMF.Instructiveeffect

ofvisualexperienceinmousevisualcortex.Neuron2006;51:339–49.

[44] KarmarkarUR,DanY.Experience-dependentplasticityinadultvisualcortex.

Neuron2006;52:577–85.

[45] Burwell RD, Amaral DG. Cortical afferents of the perirhinal, postrhinal,

and entorhinal cortices of the rat. Journal of Comparative Neurology

1998;398:179–205.

[46]BrownMV,AggletonJP.Recognitionmemory:whataretherolesofthe

perirhi-nalcortexandhippocampus.NatureReviewsNeuroscience2001;2:51–61.

[47]VannSD,AggletonJP,MaguireEA.Whatdoestheretrosplenialcortexdo?

NatureReviewsNeuroscience2009;10:792–802.

[48]OrtuzarN,Argando ˜naEG,BengoetxeaH,LafuenteJV.Combinationof

intracor-ticallyadministeredVEGFandenvironmentalenrichmentenhancesbrain

pro-tectionindevelopingrats.JournalofNeuralTransmission2011;118:135–44.

[49]Argando ˜naEG,BengoetxeaH,BulnesS,Rico-BarrioI,OrtuzarN,Lafuente

JV.Effectof intracorticalvascularendothelialgrowth factorinfusion and

blockadeduringthecriticalperiodintheratvisualcortex.BrainResearch

2012;1473:141–54.

[50]BallantyneAO,SpilkinAM,HesselinkJ,TraunerDA.Plasticityinthedeveloping

brain:intellectual,languageandacademicfunctionsinchildrenwithischaemic

perinatalstroke.Brain2008;131:2975–85.

[51]DahlqvistP,ZhaoL,JohanssonIM,MattssonB,JohanssonBB,SecklJR,etal.

Environmentalenrichmentaltersnervegrowthfactor-inducedgeneAand

glu-cocorticoidreceptormessengerRNAexpressionaftermiddlecerebralartery

occlusioninrats.Neuroscience1999;93:527–35.

[52]JohanssonBB,BelichenkoPV.Neuronalplasticityanddendriticspines:effect

ofenvironmentalenrichmentonintactandpostischemicratbrain.Journalof

CerebralBloodFlowandMetabolism2002;22:89–96.

[53] RamponC,TangYP,GoodhouseJ,ShimizuE,KyinM,TsienJZ.Enrichment

inducesstructuralchangesandrecoveryfromnonspatialmemorydeficitsin

CA1NMDAR1-knockoutmice.NatureNeuroscience2000;3:238–44.

[54]vanPraagH,KempermannG,GageFH.Neuralconsequencesofenvironmental

enrichment.NatureReviewsNeuroscience2000;1:191–8.

[55] Nithianantharajah J, Hannan AJ. Enriched environments,

experience-dependentplasticityanddisordersofthenervoussystem.NatureReviews

Neuroscience2006;7:697–709.

[56] PizzorussoT,BerardiN,MaffeiL.Arichnessthatcures.Neuron2007;54:508–10.

[57] OverstreetLS,HentgesST,BumaschnyVF,deSouzaFS,SmartJL,Santangelo

AM,etal.Atransgenicmarkerfornewlyborngranulecellsindentategyrus.

JournalofNeuroscience2004;24:3251–9.

[58]KeeN,TeixeiraCM,WangAH,FranklandPW.Preferentialincorporationof

adult-generatedgranulecellsintospatialmemorynetworksinthedentate

gyrus.NatureNeuroscience2007;10:355–62.

[59]PabanV,ChambonC,ManriqueC,TouzetC,Alescio-LautierB.Neurotrophic

signallingmoleculesassociatedwithcholinergicdamageinyoungandaged

rats:environmentalenrichmentaspotentialtherapeuticagent.Neurobiology

ofAging2011;32:470–85.

[60]SaleA,MayaVetencourtJF,MediniP,CenniMC,BaroncelliL,DePasquale

R,etal.Environmentalenrichmentinadulthoodpromotesamblyopia

recov-ery through a reduction of intracortical inhibition. Nature Neuroscience

2007;10:679–81.

[61]KolbB.Synapticplasticityandtheorganizationofbehaviourafterearlyandlate

braininjury.CanadianJournalofExperimentalPsychology1999;53:62–76.

[62]AndersonV,Spencer-SmithM,WoodA.Dochildrenreallyrecoverbetter?

Neurobehaviouralplasticityafterearlybraininsult.Brain2011;134:2197–221.

[63]OgunsholaOO,StewartWB,MihalcikV,SolliT,MadriJA,MentLR.Neuronal

VEGFexpressioncorrelateswithangiogenesisinpostnataldevelopingratbrain.

BrainResearchDevelopmentalBrainResearch2000;119:139–53.

[64]SongH,StevensCF,GageFH.Astrogliainduceneurogenesisfromadultneural

stemcells.Nature2002;417:39–44.