Acrylamide alters neurotransmitter induced calcium responses in murine ESC-derived and primary neurons

Acrylamide alters neurotransmitter induced calcium responses in murine ESC-derived and primary neurons

Julia Sisnaiske

^a,

*, Vanessa Hausherr

^a

, Anne K. Krug

^b

, Bastian Zimmer

^b,c,d

, Jan G. Hengstler

^a

, Marcel Leist

^b

, Christoph van Thriel

^a

aIfADo–LeibnizResearchCenterforWorkingEnvironmentandHumanFactors,Dortmund,Germany

bUniversityofKonstanz,Doerenkamp-ZbindenChairforInVitroToxicologyandBiomedicine,Konstanz,Germany

cCenterforStemCellBiology,Sloan-KetteringInstituteforCancerResearch,NewYorkCity,USA

dDevelopmentalBiologyProgram,Sloan-KetteringInstituteforCancerResearch,NewYorkCity,USA

1. Introduction

Recent papers in stem cell biotechnology (Brustle, 2013;

Lancasteretal.,2013)haveshownthatpropertiesofsomerelatively complexhumantissuescanbeinvestigatedininvitromodels.For example,cerebralorganoidsderivedfromhumanpluripotentstem cellshavebeendevelopedtoinvestigatemicrocephaly,aneurode velopmentaldisorder(Lancasteretal.,2013).Cellulartestsystems

havebeen established(Kruget al.,2013a,b)in ordertoidentify molecularpathwaysunderlyingtoxicevents(Collinsetal.,2008;

Leistetal.,2008b)indifferenttargetorgans(e.g.liver,brain,etc.).

Withinvivotestinganexceedinglyhighnumberofanimalswouldbe neededtotestallmajorchemicalsforsafety.Therefore,thenewNRC strategy suggests to rather identify pathways of toxicity and mechanismsofactionsoftoxicantsinvitroandthentoextrapolate thedatatothehumansituation.Forinstance,manytoxiceffects relevantfortheactionofchemicalsonthedevelopingbraincanbe observedandquantifiedininvitrosystems.Thiscanbeachieved usingstemcellsdifferentiatingtowardsneuronalcells(vanThriel etal.,2012;Kadereitetal.,2012).Aspecificadvantageofthesestem cell derived in vitro systems is, that large cell batches can be distributedbetweendifferent researchgroupsand reduce inter laboratory variability. These are also good conditions for high throughputmethods.

Thenervoussystemhasaspecialrelevanceintoxicitytesting (vanThrieletal.,2012). Neuronsareparticularlysusceptibleto toxinsduetotheirconstitutionandarchitecture.Theyhavelong Keywords:

Acrylamide Neurotoxicitytesting Stemcell

3R

Neuronalfunction

ABSTRACT

Stemcell derivedspecializedcelltypesareofinterestasanalternativecellsystemtoidentifyand research neurotoxic effects and modes of action. Developmental toxicity may be studied during differentiation,whileorgan specifictoxicitymaybeassessedinfullyfunctionalcells,suchasneurons.In thisstudywetestediffullydifferentiatedneuronsderivedfrommurineembryonicstemcells(ESCN) could be used to investigate the effects of the well characterized neurotoxic model compound acrylamide(ACR)andifESCNbehavesimilartomurineprimarycorticalneurons(pCN)from16daysold embryos.WecharacterizedthedifferentiationprocessofcryopreservedESC derivedneuralprecursor cells(NPC)differentiatingtoESCN.Duringthedifferentiationprocess(days11 20)astrongincreasein calciumresponsestoglutamate,acetylcholineandGABAwereobserved.Moreover,neuronspecific markerproteins,bIII tubulin,MAP2,Tau,Rbfox3andsynaptophysinshowedsimilarpatternstopCN.In ESCNandpCNtheneuronalstructure,e.g.neurites,wasnotaffectedbylowconcentrationsofACR[0.5 1.6mM].However,24hincubationperiodswith0.5 1.6mMACRledtoareductionofacetylcholineand glutamateinducedcalciumresponses.Inconclusion,weshowthatnon cytotoxicconcentrationsofACR alterneurotransmissioninESCNaswellaspCN.

Abbreviations:ACh,acetylcholine;ACR,acrylamide;[Ca²⁺]i,intracellularcalcium concentration;CI,confidenceinterval;DoD,dayofdifferentiation;ESCN,embryonic stemcellderivedneuron(NPCatDoD20);Glu, glutamate;MAP,microtubule associatedprotein; Rbfox3, RNAbinding protein fox-1homolog=NeuN; NPC, neural precursorcell (DoD7-19);NT, neurotransmitter; pCN,primarycortical neuron;Syp,synaptophysin;Tubb3,neuronalb-III-tubulin.

* Correspondingauthor.IfADo-LeibnizResearchCenterforWorkingEnvironment andHumanFactors,Ardeystr.67,44139Dortmund,Germany.

Tel.:+4902311084414.

E-mailaddresses:julia.sisnaiske@rub.de,sisnaiske@ifado.de(J.Sisnaiske).

Konstanzer Online-Publikations-System (KOPS) URL: http://nbn-resolving.de/urn:nbn:de:bsz:352-0-264780 Erschienen in: NeuroToxicology ; 43 (2014). - S. 117-126

https://dx.doi.org/10.1016/j.neuro.2014.03.010

processes axons anddendrites,collectively termedneurites whichinnervatethewholebodyandpossessnumerousbranches.

Moreover,theytransmitinformationaselectricalsignalsthatare transformedintochemicalsignalsatthesynapses.Withthisunique connectivitythey formhugenetworkscharacterized byactivity dependentdynamicsaswellasinhibitoryandexcitatorypathways (e.g.Bearetal.,2008;Kandeletal.,2000).Allthesefeaturesrender theneuronalnetworkacomplexsystem,vulnerabletotoxicevents.

Accordingly,neurotoxicitytestingisanimportantaspectforhuman riskassessment(Rohlmanetal.,2008).

Inadditiontotheiruseinclinicalandbasicresearch,stemcell derivedneuronsareagoodalternativesystemtoassessneurotoxic effectsandmodesofaction(KlemmandSchrattenholz,2004).Due totheirpotentialtodifferentiate,embryonicstemcells(ESC)provide unlimitedaccesstoallcelltypesofthebody,andtheyofferthe possibilitytoobservetoxiceffectsduringdevelopment(Leistetal., 2008a).Differentcellsystemsarealreadyusedtoinvestigatethe differentiation and maturation of neurons, and possible factors disturbingtheseprocesses,forexampleneurotoxins(e.g.Fritsche etal.,2005;Gassmannetal.,2010;HuangandSchneider,2004;

Tammetal.,2006;Tatenoetal.,2004;Zimmeretal.,2011,2012).Up tonow,thesecellsystemshavehardlybeenusedtoanalyzetoxic effectsonfullydifferentiatedneurons,yet(Visanetal.,2012;Zeng etal.,2006);especiallynot,whentoxinsaffectneuronalfunction.

Such effects have been demonstrated using electrophysiological methodsinculturesofprimaryneuronsfromrodents(Defranchi etal.,2011;Stettetal.,2003;vanVlietetal.,2007).Duetothe extensive use of functional primary neuronal cells fromrodent species in neurotoxicology (Howard et al., 2005; Schmuck and Schluter,1996)wethereforewantedtocharacterizefullydifferen tiatedmouseembryonicstemcellderivedneurons(ESCN)ofthe CGR8lineandassesstheirsuitabilityforfunctionalneurotoxicity testingincomparisontomurineprimarycorticalneurons(pCN).For thiscomparisonacrylamide(ACR)wasusedasapositivecontrol compound.ACRisawater solublemonomer,whichcanbefoundin food, tobacco smoke (Tareke et al., 2000) and is widely used occupationally, for example in the production of plastics and cosmetics, in wastewater treatment or as a soil conditioner (Friedman,2003).ACRisaneurotoxinandleadstoperipheraland centralneuropathiesandthedegenerationofneurites(Spencerand Schaumburg,1975;SuzukiandPfaff,1973).While manystudies haveshownanimpairmentoftheneuronalstructure,thereisalso someevidencethatACRmaydisturbtheneuronalfunction,too:In 1996 Croftonet al. observedan impaired motoractivity inrats already 2h after ACR exposure without any neuropathological damage. Tunnel workers, exposed to ACR, showed disturbed neuronal function, but no effects on structural integrity of the nervoussystem(Goffengetal.,2008).Recently,theaffinityofACRto cysteine thiolate groups located in active sites of synaptic compartments has been discussed as molecular mechanism of ACRneurotoxicity(LoPachinandGavin,2012).Thesefindingsand mechanisticconsiderationsofGoffeng,Croftonandcolleaguesledus tothehypothesisthatanacuteexposuretolowconcentrationsof ACRratherimpairsneuronalfunctionthanneuronalstructure.To testthishypothesisweanalyzedchangesofintracellularcalciumin response to different neurotransmitters (NT) utilizing both cell types,ESCNandpCN.Here,wereportthatshort termACRexposure indeedreducesmeanresponseamplitudesofdifferentNT.Similar resultswereobtainedforESCNandpCN.

2. Materialsandmethods

2.1. Dissectionandcultureofprimarycorticalneurons

Allexperimentswereconductedinaccordancewithnational lawsfortheuseofanimalsinresearchandapprovedbythelocal

ethical committee. Isolation of pCN from C57BL/6N mice at embryonicday16wasconductedasdescribedinDinhetal.(2013;

seesuppl.material).Afterremovingthemeninges,corticeswere mincedandtransferredtofreshHBSSbufferwith0.125%trypsin (PANBiotech) and incubated for 10min at 378C. After adding 0.25mg/ml soybean trypsin inhibitor (Life Technologies) and 0.01%DNase(Sigma Aldrich),corticesweretrituratedwiththree fire polished Pasteur pipettes 15 20 times withdecreasing tip diameters.Subsequently,cellswerecentrifuged(200gfor5min) andthecellpelletwasresuspendedinNeurobasalmedium(Life Technologies) containing1B 27 serum free supplement (Life Technologies),0.1%Gentamycin(PANBiotech)and0.5mMstabi lizedGlutamine (PANBiotech). pCNwereplated at a densityof 60,000cells/cm² on 0.01% polyLlysine (Sigma)coated surfaces and cultivated at 378C and 5% CO2. Half of the medium was changedeveryfourthday.FortoxicitytestspCNwereanalyzed after4daysinvitro(DIV4).OnthatdaypCNwerestill ableto respondtoGABAincalciumimagingexperimentsandweremore comparabletoESCN.

2.2. CultureanddifferentiationofmurineCGR8stemcells

ForcultureofCGR8cellsonlyplasticwarewithNunclonDelta (Nunc) surface orHCl etched glass wasused for enhancedcell attachment.Incubation of glasswith HClremovedany organic residuesfora betterbindingofpolyL lysineorpolyL ornithin.

Undifferentiated pluripotent stem cells of the murine ESC line CGR8 were routinely cultured in Glasgow’s modified Eagles medium(GMEM),supplementedwith10%heatinactivatedfetal bovine serum (FBS; PAA), 2mM stabilized glutamine, 100mM nonessentialaminoacids,50mM

b

mercaptoethanol(LifeTech nologies), 2mM sodium pyruvate and 1,000U/ml leukemia inhibitory factor (Chemicon) on 0.1% gelatine at 378C and 5%

CO2asdescribedinZimmeretal.(2011).Tostartthedifferentia tionprocesstowardsneurallinage,themediumwaschangedto N2/B 27medium(DMEM/F 12:Neurobasal1:1,0.5N2,0.5B 27 (Life Technologies), 0.05mg/ml BSA (Calbiochem), 2mM stabilized glutamine, 0.75

m

^{l/ml Insulin (9.5 10.5mg/ml, Sig}

ma Aldrich),0.1mM

b

mercaptoethanol(YingandSmith,2003), andcellswerere platedongelatinecoatedsurfacesat10⁴cells/

cm². After seven days of differentiation (DoD7) cells were harvested and frozen in liquid nitrogen for shipment toIfADo.

Forfurther differentiation,cells werequicklythawedina 378C water bath, transferred into fresh N2/B 27 medium, and after centrifugation at 300g for 3min, plated on poly^Lornithin/

laminin (10

m

^{g/ml each,} Sigma Aldrich) coated surfaces at 15,000cells/cm².Firstexchangeofmediumwasperformedafter 3daysandsubsequentlyeveryotherday.Inthefollowing,weuse theabbreviationNPC(neural precursorcells) fordifferentiating cells from DoD7 until DoD19 and ESCN for fully differentiated embryonicstemcellderivedneuronsatDoD20.Afterthisdaythere wasnofurtherchangedetectablewithrespecttotheproteinand mRNAanalysisofneuronalmarkersofZimmeretal.(2011).

2.3. Immunofluorescence

ForimmunocytochemistryNPCandpCNwereplatedonpolyL

ornithin/lamininorpolyLlysinecoated14mm,HCl etchedglass coverslips,respectively.Cells werefixed with4%paraformalde hyde(Roti Histofix,CarlRoth)for15minfollowedbypermeabi lization of cell membranes for 10min with 0.1% Triton X 100 in PBS (phosphate buffered saline). Nonspecific binding sites wereblockedwith5%normaldonkeyserum(dianova)inPBSfor 1h. Fixedcells wereincubatedwithprimaryantibodiesfor 2h at roomtemperature:polyclonal rabbit anti

b

III tubulin(Cov ance,1:2,000),monoclonalmouseanti nestin(Millipore,1:100),

monoclonal mouse anti Rbfox3 (Millipore, 1:200), monoclonal mouseanti synaptophysin(SantaCruz,1:100),polyclonalrabbit anti Tau (abcam, 1:200) and polyclonal chicken anti MAP2 (abcam, 1:200). After the following incubation with fluoro chrome labeled secondary antibodies for 30min (donkey anti chickenIgYcy3,donkeyanti rabbitIgGdylight488,donkeyanti mouseIgGcy3,1:500,dianova),nucleiwerestainedwithDAPI(4⁰, 6 diamidino 2 phenylindole,LifeTechnologies)for30min.Cover slipsweremounted tomicroscopicslides withFluorPreserve^TM Reagent (Calbiochem). Immunostainings were analyzed with a LeicaDMI6000Bmicroscope,amonochromeCCD cameraDFC360 FXandLeicaLASAF600Softwarewithidenticalsettingsforboth celltypes.

2.4. Cellviabilityassay

ToassesscytotoxicconcentrationsofACRcellsweregrownon 24wellplasticplatesandtreatedwithdifferentconcentrationsof the toxin. After 24h cell viability was determined using the CellTiter Blue¹ Cell ViabilityAssay(Promega) according tothe manual:2h(ESCN)or4h(pCN)beforetheendofexposure,cells wereincubatedalongwith1/5oftheCellTiter Blue¹compoundat 378C.Subsequently,100

m

^{lmediaofeachwellwere}transferredto a black 96 well plate to measure fluorescence. Fluorometric analysisinvolvedexcitationat540nmandemissiondetectionat 595nmwithanInfinite200Proanalyzer(Tecan).

2.5. Quantificationofneuritearea

Neurite area was measured as previously described in Stiegler et al. (2011) with minor changes as detailed below:

NPCandpCN werefixed andimmunostained for

b

III tubulin (Tubb3)andRbfox3asdescribedunder‘‘Immunofluorescence’’.

Afterstaining,1mlPBSwasfilledintoeachwell(24wellplate format),andanArray ScanVIHCSReader(Cellomics)equipped with a Hamamatsu ORCA ER camera was used to image the plates.Fivefieldsperwellwereimagedintwochannelsusinga 10objective(22pixelbinning).Excitation/emissionwave lengthsof 36550nm/53545nmwereusedtodetectRbfox3 in channel 1 and 47440nm/53545nm to detect Tubb3 in channel2.Neuronalnuclei wereidentified byRbfox3stainingas objects in channel 1 according to their size, area, shape, and intensity.Thenuclearoutlineswereexpandedineachdirection,to defineavirtualcellsomaarea(VCSA).AllTubb3 positivepixelsof thefieldweredefinedasneuronalcellularstructures(NCS).Inan automaticcalculation,the VCSAs,definedinthe Rbfox3channel, wereusedasafilterintheTubb3channelandsubtractedfromthe NCS.Theremainingpixels(NCS VCSA)intheTubb3channelwere defined as neurite area. NPC and pCN were incubated with differentconcentrationsofACRonDoD19orafter3daysinvitro, respectively.Theeffectsonneuriteswereobservedbymeasuring theneuriteareaafter 24and48h ofexposure.Bymarking only neuronal nuclei with anti Rbfox3 antibody and neuronal cell bodiesandneuriteswith anti Tubb3,we excludednon neuronal cellsfromneuritequantificationmeasurements.Twoindependent experimentswereperformedwithfourreplicateseachwithESCN and3independentexperimentswith4replicateseachwithpCN.

2.6. Calciumimaging

Calciumplaysanimportantroleintheprocessofneurotrans mission and can be visualized using the calcium imaging technique(Grynkiewicz et al., 1985).Cellswere cultivatedon HCl etched14mmcoatedglasscoverslipsin24wellplates.At certaintimepointsofdifferentiationorexposure,3

m

^{MFura 2}

AM(Tocris) wereaddedtotheculturemediumandcellswere

incubatedfor 45minat 378Cand5% CO2.Afterwashing with extracellularbuffer(140mMNaCl,5mMKCl,2mMCaCl2,1mM MgCl2,10mM HEPES, pH 7.4,osmolarity: 290 320 mosmol/l) cover slips were transferred to the imaging chamber with extracellular buffer. Cells were imaged after 10min to allow removal of intracellular acetoxymethyl groups. For functional characterization live cell video microscopy wasperformed on aninvertedmicroscope(OlympusIX81).Forintermittentexcita tionwithlightof340and380nm(50mseach)aXenonarclamp was used together with a filter wheel (MT 10) and images weretakenwitha monochromecharged coupled device(CCD) camera(Hamamatsu,MegaView)andCell^Rsoftware.Relative changes of intracellular calcium concentrations ([Ca²⁺]_i) were calculated as a ratio of the emitted light (510nm) of both excitation wavelengths (f 340/380). During the experiment, cells were constantly superfused with extracellular buffer usingan 8 in 1perfusionsystem(Harvard Apparatus), with8 independent reservoirs for test solutions. NT were applied in randomized protocols to reduce the impact of repeated stimulationofthesame cells for5switha concentrationof 100

m

^{Meach.Attheendofeach}measurementa5sstimuluswith high potassiumbuffer (10mM NaCl, 45mM KCl, 2mM CaCl2, 1mM MgCl2,10mMHEPES,pH7.4,osmolarity: 290 320mos mol/l)wasgiventoverifytheviabilityofthecellsandtoidentify neurons as only the activation of voltage gated potassium channelsofhealthyneuronsleadstoanincreaseof[Ca²⁺]_i.The increaseof[Ca²⁺]iwasdefinedasaresponsewhentheamplitude (maximumratio after stimulationminus baseline)was higher thanthemeanofthebaseline+4thestandarddeviationofthe baseline.Onlycellsthatrespondedtohigh KClbufferstimulation were used for data analysis and the mean amplitude of all respondingcellswascalculated.FunctionalityofNPCandESCN was assessed in three independent experiments with three technicalreplicatesperexperiment.DoD11,DoD15,DoD18and DoD20 were analyzed, respectively. Toxicity tests were per formedusingaLeicaDMI6000microscopeasdescribedabove.

Therefore, NPC and pCN were incubated with different ACR concentrations for 24h on DoD19 or after 3 days in vitro, respectively,andstimulatedfor5sinrandomizedprotocolswith Glu(50

m

^M),acetylcholine(ACh,100

m

^M)andGABA(50

m

^M)as

well as high KCl buffer at the end of each measurement to identifyviableneuronalcells.

2.7. Statistics

AllstatisticalanalyseswereperformedusingSPSSVersion21 (IBMCorp.)andPrism5(GraphPadSoftware).General(GLM)or generalized linearmodels (GENLIN) wereused toanalyze the effect of the applied concentrations and stability of possible effects across the biological replicates. The concentration was treated as fixed effect while the replicates were treated as random effects. Accordingto the levels of measurement F or Wald Chi² values were used to determine the significance of the treatment effect (overall effect). To account for different variancesinthedifferenttreatmentgroupsDunnett T3post hoc tests were used to compare the respective control condition withthe variousACRconcentrations. Pairwisecomparisons to the control condition using the Wald type 95% confidence interval were used to identify significant differences for the numberof responders(binary variable).Statistical significance wassetatp<0.05andmultiplecomparisonswereadjustedby Bonferronicorrection.TheEC50valuesandtheir95%confidence intervals were calculated by fitting symmetrical sigmoidal functionstothe normalizeddata (i.e. expressed aspercentage of controls). In the text ‘‘n’’ means the number of biological replicatesthatcontainedseveral technicalreplicates.

3. Results

3.1. SimilarmarkerexpressionbyESCNandpCN

During the course of differentiation NPC were stained with differentneuronspecificantibodiesandcomparedtopCN(Fig.1A).

The cellular distribution and fluorescence levels of the marker proteinsfoundonDoD20inESCNwereverysimilartothatofpCNat 4daysinvitro.AtDoD11NPCwerepositivefor

b

III tubulin(Tubb3) andnestin.AtDoD20thenumberofTubb3 positivecellsincreased and nestin positive cells were still present but with weaker fluorescence.SomecellswithDAPI positivenucleishowedneither Fig.1.StructuralcharacterizationofNPCandESCNincomparisontomousepCN.ForanalysisoftheneuralcharacterofNPCandESCNimmunostainingsofthesecellsat differentDoDwerecomparedtothoseofpCNatDIV4andpicturesweretakenwiththesamesettings.Thestartingpopulation(NPC)wasnegativeformostneuronalmarker proteins.TheESCN,resultingfroma20daydifferentiationprocesswerepositiveforTubb3,Rbfox3,Tau,MAP2,SypandweaklyfornestinandcomparabletopCN.Boxes indicateaxon-likestructureswithweakornoMAP2,butastrongTausignal(A).ThesestructuresaremagnifiedinB:singlechannelsoftheESCNneuriteareshownontheleft thoseofthepCNneuriteontheright.

Tubb3,nornestin staining. AtDoD11 NPC showed no or weak fluorescenceforRbfox3(NeuN),Tauorsynaptophysin(Syp),but werepositiveforMAP2.AtDoD20ahighnumberofRbfox3positive cellswasobserved.TauandSypwerepresentinthemajorityofthese cells.WedetectedneuritesofESCNbeingstronglystainedforTau butmuchweakerforMAP2(Fig.1AboxesandB).InpCN stainings wefoundneuritesshowingonlyfluorescencesignalsforTaubutnot forMAP2(Fig.1B).

3.2. NPCandESCNrespondtovariousNT

Clear differentiation associated increases in the number of respondingcellsandmeanresponseamplitudeswereobtainedfor Glu,AChandGABA(Fig.2).Whenthecellswerestimulatedwith Glu on DoD11, 67% (46/56cells) showed a response to this stimulus.Thenumberofrespondingcellsincreasedsignificantly until DoD20 to 100% (286/286cells). In the beginning of the differentiationprocess(DoD11)wefoundnocellbeingactivated byaserotoninnoradopaminestimulusand0.8%(1/56cells)by ATP.Duringthefollowingdaysthesepercentagesdidnotchange significantly.However,slightincreaseswerefoundforthesethree NT. At DoD11 we observed that 39% (31/56cells) of the cells responded with an increase in [Ca²⁺]i when applying GABA.

ThenumberofthesecellsincreasedsignificantlyatDoD18with 82%(153/183cells)andDoD20with86%(250/286cells).Alsothe relative response amplitudes for Glu, ACh and GABAincreased significantlyduringdifferentiation:AfteraGlu stimulusthemean amplitude increased significantly from 0.070.03at DoD11 to 0.320.04atDoD20.AChevokedameanamplitudeof0.030.01at DoD11andoneof0.080.02atDoD20.AfteraGABAstimulusthe meanamplitudeatDoD11reached0.020.004andwassignificantly higheratDoD18(0.080.01)andDoD20(0.100.03).Wecouldnot observeanysignificantdifference inthemeanamplitudeforATP, dopamineorserotonin.RepresentativeresponsecurvesofNPCon differentDoDandESCNareshownwithtwoexemplarytraceseach (Fig.3A).Thesetracesillustratetheincreaseintheresponsepeaks withproceedingdifferentiation.Singletraceanalysisalsoshowsthat responsebehaviorsofdifferentcellsvaried(Fig.3A).

3.3. CytotoxicityandneuritedegenerationbyexposuretoACRinESCN andpCN

After24hACRexposurethecellviabilityassayrevealedanEC50

of5.7mM(CI:4.9 6.4mM)forpCN.ThelowestACRconcentration withasignificanteffectoncellviabilityinpCNwas1.6mM.Here the cell viability wasreduced to 80% compared to the control condition(p=0.02).InESCNanEC50of7.6mM(CI:3.7 11.5mM) wasobservedandnoneoftheinvestigatedconcentrationshada significant effecton cellviability. The highest concentration of Fig.2. Functional characterizations of NPC/ESCN.Cells werestimulated –in

randomizedprotocols–for5swithdifferentNT(100m^{M)incalciumimaging} experimentstoinvestigatefunctionalexcitability.Attheendofeachmeasurement a5sKClstimuluswasgiventodemonstrateviabilityofthecellsandtoidentify neuronalcelltypes.Thenumberofrespondingcells(A)andresponseamplitudes(B) ofn=3(3replicateseach)areshownasmeanSEM.(numberofcellsanalyzed:

DoD11=56, DoD15=173, DoD18=183, DoD20=286; **p0.01, ***p0.001, Dunnett-T3post-hoctests).

Fig.3.ResponsebehaviorofNPC/ESCNafterstimulationwithNT.Cellswerestimulatedfor5swithdifferentNT(100m^{M)incalciumimaging}experiments.Theorderof stimuliwasrandomized.Attheendofeachmeasurementa5sKClstimuluswasappliedtoidentifyneuronalcellsandtotestviabilityofthecells.Exemplarytracesoftwo singlecellsperDoDillustratedifferentresponsebehaviorsandresponsesofthesecellstodifferentstimuli(A).Falsecolorimagesofresting(left)andstimulated(right)ESCN.

Bluecolorindicatesalow[Ca²⁺]iandpinkahigh[Ca²⁺]i(B).

5mMinducedareductionto85%bytrend.Firstsignificanteffects onneuriteareawereobservedwhenESCNorpCNwereexposedto 5mMACRfor48h.Comparedtothecontrolcondition,theneurite area of ESCN decreased by 80% after 48h of ACR incubation (Fig.4A).WhenpCNwereincubatedalongwith5mMACRfor48h the neurite area was reduced by 69% (Fig. 4B). There was no significanteffectafter24hACRexposure.However, theneurite areaofESCNwasreducedafterincubationwith5mMbyaclear trend.

3.4. ACRincubationaltersNTresponsebehaviorinESCNandpCN

BecauseNPC/ESCNrespondedtoabroad spectrumofNTwe askedwhetherthiscelltypecouldbeusedtoexploretheeffectsof ACR on NT induced calcium responses as an indicator of neurotransmission. We found that ACR affected the response amplitudeatthelowestconcentrationof0.5mM.Similareffectsof ACRwereobservedforbothcelltypeswiththestrongestinfluence for Glu. We could not detect any significant difference in the numberofrespondingESCN,exceptfor5mMACR.Here,aclear declineinthenumberofrespondingcellsforACh (from78to26%

responding cells; Fig. 5A) and GABA stimuli (from 85 to 45%

responding cells; Fig.5B) wasobserved. The numberof GABA responsive pCN reduced to23% (Fig. 5D) afterincubation with 5mMACR.Adecreaseto54%Glu responsivecellswasobserved afterexposureto5mMACRinpCN,andto67%afterexposureto 1.6mMACR(Fig.5E).AChwasnotinvestigatedinpCNasthiscell typeisnotresponsibletoAChduetoitscorticalorigin.Thebasal calcium level of both cell types was not affected by all concentrations of ACR (Fig. 5F and G). After ACR exposure the mean amplitudes of ESCN for ACh and Glu stimuli were significantly decreased in a concentration dependent manner (Fig.6AandC):at0.5mMACRtheACh amplitudewasreducedto 71%, at1.6mMto57% andat 5mMACR to26% ofthecontrol condition.TheGlu inducedresponseamplitudesweredecreased to66%afterincubationwith0.5mMACR,to62%with1.6mMand to47%with5mMACR(Fig.6B).Theexposurewith1.6mMACRled toareductionofthemeanamplitudeto67%afteraKCl stimulus and with5mM ACR to59% (Fig.6D). There wasno significant changeintheKCl inducedmeanamplitudeafterincubationwith 0.5mMACR.StimulatingtheESCNwithGABAafterexposureto 0.5mM ACR led to a 59% higher mean response amplitude comparedtothecontrolcells.InpCNweobservedaconcentration dependent reductionofGlu andKCl inducedmeanamplitudes (Fig.6FandG):afterexposureto0.5mMACRthemeanGlu evoked amplitudewasdecreasedto71%,afterexposureto1.6mMto53%

andafterexposureto5mMto26%incomparisontothecontrol cells.TheKCl inducedmeanamplitudewasdecreasedto81%after

incubationwith0.5mMACR,to65%(p0.001)afterincubation with1.6mMandto30%afterincubationwith5mMACR.When pCNwereincubatedwith5mMACRthemeanamplitudeinduced byGABAwasdecreasedto30%.A26%increaseofthemeanGABA induced response amplitude after exposure to 0.5mM ACR, as observedwithESCN,couldbedetectedwithpCNbytrend.Anacute exposuretoACRitselfdidnotchange[Ca²⁺]i(Fig.6H).

4. Discussion

Recently a mouse ESC based in vitro system has been establishedthat allows differentiationto neuronal cells (ESCN) (Zimmeretal.,2011).ESCNhavebeenshowntoexpressneuron specificmarkersandtobeelectrophysiologicallyexcitable.Inthe presentstudywedemonstratethatESCNgainneuronalfunction alityduringthe differentiationprocess.Moreover, we illustrate thatACRnotonlycausesstructuralalterations,butdecreasesNT induced response amplitudes at even lower concentrations.

SimilarresultswereobtainedforESCNandpCN.

Fig.4.NeurotoxiceffectsofACRonneuritesofESCNandpCN.Cellswereincubated withdifferentconcentrationsofACRfor24h(whitebars)or48h(greybars) respectively.NeuriteareaofESCNwascalculatedofn=2with4replicatesateach timepoint (A)andofpCN forn=3with 4replicates ateachtimepoint(B).

(meanSEM;*p0,05;**p0.005;***p0.001,Dunnett-T3post-hoctests).

Fig.5.Effectsof24hexposuretoACRonthenumberofcellsrespondingtoNT- stimuli.ESCNandpCNwereincubatedwithdifferentconcentrationsofACRfor 24h. The number of responsive cells was determined in calcium imaging experimentsforACh(A),GABA(B,D)andGlu(C,E).OnlyKClresponsivecells wereanalyzed(=100%).ShownaremeanSEMofn=2(2measurementseach)for ESCNandn=3(1–2measurementseach)forpCN(numberofESCNanalyzed:Ctr.=96, 0.5mM=54, 1.6mM=93, 5mM=41; number of pCN analyzed: Ctr.=150, 0.5mM=107,1.6mM=153,5mM=93;*p0.05;**p0.005;***p0.001,Wald- Chi-Squaretest)TherewasnosignificantchangeinbasiccalciumlevelsafterACR exposure(F,G;Dunnett-T3post-hoctests).

4.1. ImmunostainingconfirmssimilarmarkerlocalizationinESCN andpCN

Immunocytochemistry analysis clearly demonstrates that differentiatedESCNpossessimportantcharacteristicsofneurons.

ComparedtothepCNfrommouseembryosthecellulardistribu tionand appearance ofthese proteins weresimilar in thetwo investigated cell types. Rbfox3 is expressed exclusively in the nucleiofpost mitoticneurons(Mullenetal.,1992)andsignalsof thismarkercouldbedetectedinNPC andESCN nucleiaswell.

Moreover,we foundneurites withstrong Tau and weak MAP2 signalsinESCN.InmatureneuronsMAP2isabsentinaxonsand locatedonlysomato dendritically,whereasTauismainlypresent inaxons(BernhardtandMatus,1984;Binderetal.,1985;Dotti etal.,1987; KosikandFinch,1987;MandellandBanker,1995).

Inthedevelopingneuron,TausegregatesintoaxonsandMAP2into dendritesrightafteraxogenesis (Matus,1990). Therefore,ESCN seemtobestillintheprocessofaxogenesiswhereaspCNalready finishedthisprocess.ESCNalsoexpresspresynapticstructuresas shown by the positive signal for Syp, a protein of presynaptic

vesicles(WiedenmannandFranke,1985)andinvolvedincalcium dependentNT release(Rehmetal.,1986).Besidesneurons,there are alsofew non neuronal cells which wereidentifiedby their stainednucleiandbeingnegativeforanyneuronalmarker.Someof thesecellswerepositivefornestin.Similarcelltypescouldalsobe observedinpCN cultures.Asnestinisalsoexpressedinglialcells (Schmidt KastnerandHumpel,2002),thesefindingsareconsistent withtheresultsof Zimmer etal. (2011),who foundastrocyte markersinNPCcultures,suchasgfapandglutaminesynthetase,as wellasmarkersforasmallsubpopulationofcellsshowingsmooth musclefeatures.

4.2. ESCNarefunctionalneurons

ESCNexpressvariousfunctionalNTreceptorswhichmediatean increase in the membrane potential, as confirmed by specific stimulationincalciumimagingexperiments.Acrossthecourseof differentiationthenumberofrespondingcellsincreasedformostof the NT and also the mean amplitude increased over time.

Theseresultsargueforanongoingdifferentiationprocesstowards Fig.6.Effectsof24hexposuretoACRonNT-evokedresponseamplitudesofESCNandpCN.CellswereincubatedwithdifferentACR-concentrationsfor24h.Themean responseamplitudesSEMofACh-(A),GABA-(B,E),Glu-(C,F)andKCl-(D,G)stimuliareshownofn=2(2measurementseach)forESCN(A–D)andofn=3(1–2measurements each)forpCN(E–G);(numberofESCNanalyzed:Ctr.=96,0.5mM=54,1.6mM=93,5mM=41;numberofpCNanalyzed:Ctr.=150,0.5mM=107,1.6mM=153,5mM=93;

*p0.05;**p0.005;***p0.001,Dunnett-T3post-hoctests).Exemplarytracesoftwocells(H)demonstratethatneithera4minstimulationwith1.6mMACRnora7min stimulationwith50mMACRchanged[Ca²⁺]i.

aneuronalphenotype.TheNTserotoninanddopamineshowedan excitatoryeffectinourexperiments.Thissuggeststhat5 HT3and 5 HT2receptorsandthoseoftheD1familyareexpressedinNPCand ESCNbecause onlythese receptorsevokeanexcitatory receptor potentialandanincreaseof[Ca²⁺]i(Maricqetal.,1991;Nicholsand Nichols,2008;UndieandFriedman,1990).Throughoutthewhole differentiation process also GABA stimulation showed a strong excitatoryeffect.Therefore,NPCandESCNexpressGABAAreceptors thataretheonlyGABA receptors accompaniedwithanaltered chloridehomeostasisduringmaturation leadingtoresponsesin calciumimagingexperiments.Thisreceptorhasadepolarizingeffect in cortical progenitors and immature neurons (Ben Ari, 2002;

LoTurcoetal.,1995)suchasthepCNanalyzedinthisstudy.Thus,we suggestthatourESCNareaculturewithcellsofdifferentmaturation stages,similartopCN.

The strongest increase in response amplitudes during the differentiationprocesswasobservedforGlu.Gluneurotransmission canbemediatedbyvariousreceptorsubtypes,includingionotropic NMDA,AMPAandkainatereceptorsaswellasmetabotropicGlu receptors(HollmannandHeinemann,1994).Fromapreviousstudy itisknownthatESCNexpressNMDAandAMPAreceptors(Zimmer etal.,2011).However,whichspecificreceptorsubtypeisresponsible fortheobservedcalciumresponsebyGluremainstobedetermined byspecificinhibitorsandproteinexpressionanalysis.

Cells used and measured in calcium imaging experiments represent a random sample of the entire NPC/ESCN culture.

Theobservedculturesectionswerechosenduetothemorphology ofthecells,preferringneuronalphenotypes.Accordingtosingle cell traces these sections usually contained cells of different differentiationstages,as notallresponseamplitudes, especially thosetoKClandGlu,wereequal.Thisobservationisinlinewith thefindings of theimmunostainingexperiments: atDoD20 we foundfullydifferentiatedneuronswithstrongRbfox3signalsas wellascellsthatshowedweakerRbfox3signalsorthatexpressed NestinalongwithTubb3.

4.3. ACRimpairsneuronalfunctionaftershort termexposure

The moststriking result of our experiments is that ACR at concentrationsof0.5mMreducesthemeanresponseamplitudeof Glu induced increase of cytosolic calcium in ESCN and pCN.

Thisreductionisnotaccompaniedbyareductionofthefractionof respondingcells.Recently,amechanismhasbeendiscussedhow chemical reactions with glutamatergic NMDA receptors may reducetheirfunctionality.TheseresultssuggestthatACRmight modulate the affinity of glutamatergic receptors towardstheir ligand especially at low concentrations when voltage gated potassium channels were not or less affected. ACR has an electrophilicdoublebond(Friedman,2003).Ithasbeensuggested that it forms covalent bondswithsoft nucleophiles (e.g.SH of cysteine)insynapticcompartments(LoPachinandGavin,2012).

ThisinteractioniswelldocumentedforNO(LoPachinandBarber, 2006).Therefore,ACRmay,similarlytothephysiologicalNO,also bind tocysteine residues (e.g.Cys 79 and Cys 208 atthe NR1 receptor subunit). In contrast to NO, binding by ACR may be irreversible,therebydecreasingtheaffinitytotheligand.Further experimentsshouldelucidatethismodeofaction.

Thereisalreadyconvincingevidencefrominvivostudiesinrats and humansthatACRimpairs neuronalfunction(Croftonetal., 1996;Goffengetal.,2008).However,theeffectsofthisneurotoxin onNT receptorsinneuronshavenotyetbeenstudied.Researchon ACR and its effects on neuronal function are based on ligand bindingassaysorthesecretionandre uptakeofNT(Agrawaletal., 1981;Bondyetal.,1981;LoPachinetal.,2004).Moreover,Nordin Andersson et al. (2003) reported increased ACh evoked, IP3

mediatedresponseamplitudesinhumanSH SY5Yneuroblastoma

cells after 72h ACR exposure. They suggested that an ACR inducedincreasedbasal[Ca²⁺]iresultedinanelevateduptakeof Ca²⁺ into intracellular stores which in turn led to increased amplitudes.Inourstudywecouldnotobservethiselevationof basal Ca²⁺ levels and detected a concentration dependent decrease of ACh induced response amplitudes. A study of Lappe Siefke et al. (2009) reported an increase of GABAA

receptors in Purkinje cells of mice with cerebellar ataxia together with an increased inhibitory postsynaptic current in thesecells.ACRexposurecausescerebellarataxiaaccompanied by an abnormal morphology of GABAergic Purkinje cells (Shi etal.,2012).ThismodeofactiononGABAergicneurotransmis sion mightbereflected in the increase ofGABA evoked mean response amplitudes after incubation with 0.5mM ACR and thus,theESCNsmightbeasuitableinvitrosystemtofollowup thisinvivo observationofShietal. (2012).

Comparedtopreviousinvitrostudiesinvestigatingtheeffects ofACRonneurites(Hardelaufetal.,2011;Martensonetal.,1995;

Nordin Anderssonetal.,1998,2003)asanindicatorofstructural neurotoxicity,ESCNaswellaspCNappeartobemoreresistantto ACR.UsingtheneuroblastomacelllineSH SY5Y,Hardelaufetal.

(2011)reportedthat1mMACRclearlycompromisedanetworkof neurites. Another study (Nordin Andersson et al., 2003) also investigatedneuritedegenerationinSH SY5Ycellsandobserved neurite degenerative effects already at 0.25mM after 24h.

ThehighercelldensityintheNPC/ESCNculturesaswellasthe heterogeneitywithrespecttothecelltypemightexplainsuch differences.However,therelativeresistanceofESCNtoneurite effectsofACRagreeswithdatafrompCNaswellasfromhuman LUHMES cells that also did not indicate specific neurite degeneration(Kruget al.,2013a,b).Bothcell types pCN and ESCN showedsimilarresultsafterexposuretoACR.ESCNmaybe a useful tool to analyze neurotoxic effects and may replace primary neurons in future experiments. ESCN are easy to differentiateandpre differentiatedcellscanbestoredinlarge batchesinliquidnitrogenandusedfordifferentexperimentsto reduce variance. ESCN provide a variety of neuron specific functionsandthereforecanbeusedtoanalyzeabroadspectrum ofeffects.

5. Conclusion

Inthepresent studywehave shownthat mouseESCNhave similar functional properties as pCN from 16 day old mouse embryos. Immunostaining illustrated also similar patterns of neuronalmarkers.AspecificadvantageofESCNisthattheycanbe establishedfromcryopreservedprecursorcells(NPC).Therefore, ESCNmaybeagoodalternativecellsystemtoprimaryneuronsto studyneurotoxiceffectsandeffectmodesinvitro.

Conflictofinterest

Theauthorsdeclarethattherearenoconflictsofinterest.

Transparencydocument

TheTransparencydocumentassociatedwiththisarticlecanbe foundintheonlineversion.

Acknowledgments

WegratefullythankLiudmilaEfremovafortechnicalassistance with the Cellomics setup and Heidrun Leisner for the pre differentiationof NPC.The project wasfunded by grants from EU(FP7 ESNATS),DFG(RTG1331),BMBF(NEURITOX717/13(ML),