Article Stochastic Interaction between Neural Activity and Molecular Cues in the Formation of Topographic Maps Highlights Authors d HeterozygousIslet2-EphA3knockinmicedisplayvariationin MelindaT.Owens,DavidA.Feldheim, visualmaporganization MichaelP.Stryker,JasonW.Triplett d Variationwasfoundbetweenmice,betweenhemispheres, Correspondence andwithinasinglemap [email protected] d Variationinmaporganizationisreflectedbothfunctionally In Brief andanatomically Inthisstudy,Owensetal.demonstrate d Properlycorrelatedspontaneousactivityisrequiredformap variabilityintopographicmap organizationvariation organizationinamousemodelinwhich molecular-andactivity-dependentforces arerenderednearlyequalinbalance. Thesefindingssuggestaninherent stochasticityinestablishmentof topography. Owensetal.,2015,Neuron87,1261–1273 September23,2015ª2015ElsevierInc. http://dx.doi.org/10.1016/j.neuron.2015.08.030 Neuron Article Stochastic Interaction between Neural Activity and Molecular Cues in the Formation of Topographic Maps MelindaT.Owens,1DavidA.Feldheim,2MichaelP.Stryker,1andJasonW.Triplett2,3,* 1CenterforIntegrativeNeuroscienceandDepartmentsofPhysiologyandBioengineering&TherapeuticSciences,UniversityofCalifornia, SanFrancisco,SanFrancisco,CA94143,USA 2Molecular,Cell&DevelopmentalBiology,UniversityofCalifornia,SantaCruz,SantaCruz,CA95064,USA 3CenterforNeuroscienceResearch,Children’sNationalHealthSystem,Washington,DC20010,USA *Correspondence:[email protected] http://dx.doi.org/10.1016/j.neuron.2015.08.030 SUMMARY rin-As,areexpressedincountergradientsalongthetemporal- nasal (T-N) axis of the retina and the anterior-posterior (A-P) Topographic maps in visual processing areas main- axisoftheSC.InteractionsbetweeninnervatingRGCsandcol- tain the spatial order of the visual world. Molecular licularcellsresultsintemporalRGCsterminatinganteriorly,while cues and neuronal activity both play critical roles in nasal RGCs terminate posteriorly, in the SC (Triplett, 2014). mapformation,buttheirinteractionremainsunclear. DisruptionofEphA/ephrin-Asignalingresultsintopographicer- Here, we demonstrate that when molecular- and rorsalongtheazimuthaxisintheSC,dorsallateralgeniculate nucleus (dLGN), and primary visual cortex (V1) that manifest activity-dependentcuesarerenderednearlyequalin anatomicallyasectopicterminationzones(TZs)andfunctionally force,theydrivetopographicmappingstochastically. astopographicallyincorrectreceptivefields(Brownetal.,2000; The functional and anatomical representation of Cang et al.,2005a, 2008b; Feldheim et al.,2000; Frise´net al., azimuth in the superior colliculus of heterozygous 1998;Pfeiffenbergeretal.,2006;Triplettetal.,2009). Islet2-EphA3 knockin (Isl2EphA3/+) mice is variable: In addition to molecular cues, correlated neuronal activity maps may be single, duplicated, or a combination playsacriticalroleinmapformation.Duringdevelopment,spon- of the two. This heterogeneity is not due to genetic taneouswavesofactivitypropagateacrosstheretinaandinitiate differences, since map organizations in individual corresponding waves in retinorecipient nuclei (Ackman et al., mutantanimalsoftendifferbetweencolliculi.Disrup- 2012;Meisteretal.,1991).Disruptionofthesewavesresultsin tionofspontaneouswavesofretinalactivityresulted errorsintopography,suchthattheTZsofRGCsintheSCare in uniform map organization in Isl2EphA3/+ mice, foundinapproximatelythecorrecttopographiclocationbutfail to refine to establish a discrete map (Chandrasekaran et al., demonstrating that correlated spontaneous activity 2005;McLaughlinetal.,2003).Functionally,thisisreflectedas is required for map heterogeneity. Computational anincreaseinthenumberofRGCsynapticinputstocollicular modelingreplicatesthisheterogeneity,revealingthat cellsandincreasesinthereceptivefieldsizesoftheseneurons molecular- and activity-dependent forces interact (Cangetal.,2008a;Chandrasekaranetal.,2005).Interestingly, simultaneouslyandstochasticallyduringtopographic anatomical and functional perturbations in the absence of mapformation. spontaneous activity are observed specifically in the azimuth domain, which has been attributed to the unique spatiotem- poral properties of retinal waves (Stafford et al., 2009). Similar INTRODUCTION changes are also observed in the dLGN and V1, suggesting a common role for spontaneous activity in topographic map In the visual system, the spatial relationships of objects in the formation throughout the visual system (Cang et al., 2005b; worldarefaithfullyrelayedthroughouthigherprocessingcenters Grubbetal.,2003).Together,molecular-andactivity-dependent throughtheestablishmentoftopographicmaps(CangandFeld- forces play the dominant roles in topographic mapping along heim,2013).Theretina’sprojectiontothesuperiorcolliculus(SC) theazimuthaxis,asdisruptionofbothforcessimultaneouslyre- hasservedasamodeltounderstandthemolecularandactivity- sults in a nearly complete loss of topography in the SC (Cang dependentforcesthatdrivemapformationfordecades(Gaze, etal.,2008b;Pfeiffenbergeretal.,2006).However,therelative 1981). Retinal ganglion cells (RGCs) project to the SC during strengthofeachoftheseforcesremainsunclear. earlypost-natal lifeandarerefinedtoafinaltopographicmap Basedoninvivostudiesingeneticallymodifiedmice,several by the end of the first post-natal week (McLaughlin et al., computationalmodelshavebeendevelopedtoexplaintherole 2003).Graded molecularcuesplayaroleintheestablishment ofmolecular-andactivity-dependentforcesintheestablishment of retinocollicular topography along the azimuth axis. Specif- of topography. The relative signaling model argues that local ically,EphAreceptortyrosinekinasesandtheirligands,theeph- comparisonsinEphAsignalingstrengthbyRGCsresultsinan Neuron87,1261–1273,September23,2015ª2015ElsevierInc. 1261 ordered distribution of terminals in the SC (Reber et al., 2004; varies depending on their location along the T-N axis of the Bevinsetal.,2011).Suchamodelissupportedbyexperiments retina.Forinstance,innasalretina,expressionofendogenous inwhichthelevelofEphAreceptorsignalingwasalteredinasub- EphAs is low, so Isl2+, EphA3-overexpressing RGCs should setofRGCs,resultingintheformationofabifurcatedmapinthe have distinct EphA signaling strengths from Isl2(cid:2) RGCs. In SC(Brownetal.,2000).Asecond,permissivearborizationmodel contrast, RGCs in the temporal retina express high levels of positsthatmapdevelopmentproceedsinastep-wisefashion, endogenousEphAs,suchthatthedifferenceinEphAsignaling in which molecular forces establish broad zones for potential strengthisminimalbetweenIsl2+andIsl2(cid:2)RGCs.Thus,onepre- synapseformation,whilelateractivity-dependentandcompeti- dictionofthispatternofEphAexpressionwouldbethatIsl2+and tive forces dictate final map organization (Grimbert and Cang, Isl2(cid:2)RGCsfromthenasalretinawouldprojecttotwodistinctre- 2012).Thismodelissupportedbyaccumulatingevidenceofa gionsoftheSC,whilethosefromthetemporalretinawouldproj- roleforEphA/ephrin-Areversesignaling(Limetal.,2008;Rashid ect to a similar region in the anterior SC. Indeed, anatomical etal.,2005;Yatesetal.,2001),axonalcompetition(Triplettetal., tracingexperimentsinIsl2EphA3/+micesupportedthishypothe- 2011),andinteraxonalsignaling(SuetterlinandDrescher,2014) sis(Brownetal.,2000).Thesedataalsopredictthattheazimuth inretinocollicularmapformation.Finally,stochasticmodelspro- representationalongtheA-PaxisoftheSCwouldbesplitinto posethatRGCsexchangesynapsesintheSCtooptimizeadhe- twoincompletemapsofspace,whereintheanteriorofapproxi- sion,drivenbyacombinationofmolecularforces,competition, matelytwo-thirdsofthemapwouldhaveafullrepresentationof and activity (Koulakov and Tsigankov, 2004; Tsigankov and centralvisualspaceandapartialrepresentationoftheperiphery, Koulakov,2006,2010).Whilethismodelreplicatestopographic whiletheposteriorone-thirdwouldhaveapartialrepresentation organizations found in control and mutant mice in vivo, clear oftheperiphery.However,thefunctionalorganizationinthatSC empiricaldatasupportingstochasticityremainselusive. inIsl2EphA3/+micehasnotbeeninvestigated. Here, we demonstrate an activity-dependent, stochastic To determine the functional topographic organization of the mechanismfortheestablishmentofretinocolliculartopography. SCinadultIsl2EphA3/+mice,weperformedintrinsicsignaloptical Utilizing intrinsic signal optical imaging in heterozygous Islet2- imaging,asdescribedpreviously(KalatskyandStryker,2003). EphA3knockinmice(Isl2EphA3/+),weshowthatmaporganization This technique allows the observation of either the azimuth or exhibitstremendousheterogeneitybothamongIsl2EphA3/+indi- elevationrepresentationofspaceatapopulationlevelinalarge viduals and even within single animals. Map heterogeneity is portionoftheSCatonce.Indeed,inWTmice,asingleazimuth alsoexhibitedanatomically,asassayedbyanterogradetracing representationandsingleelevationrepresentationareobserved ofRGCsandanalysisofIsl2+RGCprojectionpatterns.Neural intheSCcontralateraltothestimulatedeye(Figures1Dand1E). activityisrequiredforthemanifestationofheterogeneityinmap InadulthomozygousIsl2EphA3/EphA3mice,theazimuthrepresen- organization,sincedisruptionofspontaneousretinalwavesre- tationintheSCisduplicated,whiletheelevationrepresentation sulted in only a single type of map. Heterogeneity akin to that remainssingular,consistentwithourpreviousfindings(Figures observed in the Isl2EphA3/+ was replicated by a computational 1F and 1G) (Triplett et al., 2009). Interestingly, while heterozy- model when the EphA signaling factor was set appropriately. gous Isl2EphA3/+ animals had functional maps of similar size Smallvariationsinthisfactorresultedinconsistentmaporganiza- and strength to WT and Isl2EphA3/EphA3 mice (Figures S1A– tions,mimickingtheconsistencyfoundinwild-type(WT)andho- S1C),weobservedheterogeneityintheorganizationofazimuth mozygousIsl2EphA3/EphA3animals.Takentogether,theseresults representations(Figures1H–1O,n=16).Theserepresentations suggestthatactivity-dependentandmolecularforcesarenearly fellintothreequalitativecategories:(1)single(Figure1H),similar equalinbalanceduringtheestablishmentoftopographyandthat tothoseofWTanimals,(2)doubled(Figure1J),similartothose localtopographycanbedrivenbyeitherinastochasticmanner. seen in Isl2EphA3/EphA3 mice, and (3) partially doubled and partiallysingle(Figure1L),hereafterreferredtoasmixed-type. RESULTS Importantly,thedifferencesobservedwerenotduetovariability inresponsiveness,aseachofthetypesoccupiedsimilarareasof HeterogeneityofMapOrganizationinIsl2EphA3/+Mice theSC(FigureS1D). One of the most elegant investigations of the role of EphA To quantitatively categorize azimuth map organizations in signalingintopographicmapformationinvolvedthegeneration Isl2EphA3/+mice,weplottedtheazimuthphaseasafunctionof oftheIslet2-EphA3knockinmouseline(Isl2EphA3)(Brownetal., positionalongtheA-Paxisatthreedifferentisoelevationpoints 2000). Inthesemice, anEphA3cDNAwas insertedinto the30 inthemap(Figure1N).Wethendeterminedtwoindicestobetter UTR of the Islet2 locus. This manipulation produced a salt describetherelationshipsbetweentheazimuthrepresentations andpepperarrangementoftwopopulationsofRGCs:(1)Isl2(cid:2) at these distinct elevations (see Experimental Procedures). RGCs that express endogenous levels of EphA receptors and Briefly, wefirstgenerated a‘‘linear fitindex,’’ whichdescribes (2)Isl2+RGCsthatexpressendogenousEphAsandexogenous thegoodnessoffittoalinearregressionfortheazimuthphase EphA3. In homozygous Isl2EphA3/EphA3 mice, the difference in plotateachelevationmeasured.Next,wedeterminedan‘‘intra- total EphA signaling between Isl2(cid:2) and Isl2+ RGCs is large, mapcorrelationindex,’’whichdescribesthefidelityofmaporga- resultinginthesegregationofthesetwopopulationsintheSC nizationbetweeneachazimuthphaseplotwithinasinglemap. and the formation of two coherent anatomical and functional Byplottingthelinearfitindexasafunctionoftheintramapcor- mapsofspace(Brownetal.,2000;Triplettetal.,2009).Inhetero- relation index, we found that individual maps fell into distinct zygous Isl2EphA3/+ mice, however, the difference in EphA sig- clusters, corresponding to single maps of Isl2+/+ mice, double naling strength between populations of Isl2+ and Isl2(cid:2) RGCs maps of Isl2EphA3/EphA3 mice, and mixed maps of Isl2EphA3/+ 1262 Neuron87,1261–1273,September23,2015ª2015ElsevierInc. Figure1. HeterogeneityintheFunctionalAzimuthRepresentationintheSCofIsl2EphA3/+Mice (A–C)Schematicsofintrinsicsignalopticalimagingparadigm.Driftingvertical(azimuth,B)andhorizontal(elevation,C)arepresentedonascreenplaced25cm fromananesthetizedmouse.RegionsoftheSCrespondingtodifferentregionsofthevisualfieldarepseudo-coloredasindicated. (DandE)Azimuth(D)andelevation(E)representationsinawild-typeSCrevealasinglecoherentmapineachcase(n=4). (FandG)ImagingofSCresponsesinanIsl2EphA3/EphA3mouserevealacompleteduplicationofazimuth(F)andadiscontinuityattheborderofthesemapsinthe elevationrepresentation(G)(n=5). (H–M)ImagingofSCresponsesinanIsl2EphA3/+revealsavarietyoforganizationsintheazimuthrepresentations:single(HandI),double(JandK),andmixed-type (LandM)(n=16). (N)Phaseplotsoftheazimuthrepresentationalongtheanterior-posterioraxisoftheSCatthreeisoelevationpointsfromIsl2EphA3/+miceexhibitingasinglemap organization(toprow),doubleorganization(middlerow),andmixedorganization(bottomrow).Bluedashedlinesrepresentthebestlinearfitforeachphaseplot. (O)Maporganizationsplottedbytwoquantitativeindices.Isl2+/+maps(blackdots)fallwithintheregiondeterminedtobesingle,Isl2EphA3/EphA3maps(reddots) fallwithinthedoublezone,andIsl2EphA3/+maps(greendots)aredistributedacrosssingle,double,andmixedregionsoftheplot.Scalebar,0.5mm;A,anterior; L,lateral. that fell between the single and double clusters (Figure 1O). the maps of space to be coherent (Figure 1G; Triplett et al., Basedonthisanalysis,wefoundthat25.0%(4/16)ofIsl2EphA3/+ 2009). This discontinuity is present only in the mice that have mapsweresingle,18.8%(3/16)weredouble,and56.3%(9/16) doubled or partially doubled maps of azimuth and is absent weremixedtype. in the Isl2EphA3/+ mice with single maps (Figure 1M). Taken Althoughthemapsofazimuthdisplaythemostobviousdiffer- together,thesedatademonstrateheterogeneityintheorganiza- ences,theelevationmapsdifferaswell.InIsl2EphA3/EphA3mice, tionofmapsintheIsl2EphA3/+SCandsuggestastochasticpro- the elevation maps display a discontinuity that allows each of cessmayunderliemapformation. Neuron87,1261–1273,September23,2015ª2015ElsevierInc. 1263 Figure 2. Bilateral Heterogeneity of Azi- muthRepresentationinOpposingCollicular HemispheresofIsl2EphA3/+Mice (A) Schematics of imaging paradigm to evaluate azimuthandelevationrepresentationsineachSC hemisphereofasinglemouse. (BandC)Azimuthandelevationrepresentations intheleftandrightSCoftwodifferentIsl2EphA3/+ micerevealsheterogeneitybetweenhemispheres (n=7). (D)MaporganizationsofIsl2EphA3/+miceinwhich bothSCswereimagedplottedbytwoquantitative indices.Eachpairofcoloreddotsrepresentsthe twoSCsofanindividualIsl2EphA3/+mouse.Scale bar,0.5mm;A,anterior;L,lateral;M,medial. imuth representation was singular in the right hemisphere but mixed type in the lefthemisphere.Thesedatasuggestthat heterogeneity in map organization is not due to subtle changes in EphA3 expres- sion level due to genetic background. Despite this, the possibility remains that subtle differences in EphA3 expression from the Isl2 locus between eyes may cause the heterogeneity observed be- tweencolliculiofthesameanimal.How- ever, our observations that different regions of a single SC are organized differently in a substantial portion of Isl2EphA3/+micesuggeststhatgeneticdif- ferencesplayaminimalroleinthehetero- geneityobserved. HeterogeneityofMapOrganization inIsl2EphA3/+MiceIsNotdueto SuppressionofVisualResponses Anotherpossibleexplanationforthehet- erogeneity observed in individual func- tional maps is that some of the retinal inputs into the SC are being functionally suppressed.Visualmapsinprimaryvisual cortexareabletosuppressaberrantinput HeterogeneityofMapOrganizationinIsl2EphA3/+MiceIs fromthelateralgeniculatenucleusinthe‘‘Midwestern’’varietyof NotduetoGeneticVariation Siamesecat(KaasandGuillery,1973),buttheextenttowhich One potential source of heterogeneity in map organization mapsinSCcansuppressretinalinputisunknown.Toaddress observed in Isl2EphA3/+ mice could be differences in genetic thepossibilityoffunctionalsuppression,weimagedtheSCsof backgroundbetweenmice.Ifthiswerethecase,thentheorga- Isl2EphA3/+micewhileshowingthemspatiallyrestrictedstimuli. nizationofthevisualmapshouldbethesameinbothcolliculiof These stimuli would only excite a small part of the retina at anindividualmouse.Todeterminewhetherthiswerethecase, once, potentially activating ectopic termination sites that are weimagedthecolliculiofbothhemispheresinseveralIsl2EphA3/+ normally suppressed when full-field stimuli are presented. We mice. Strikingly, we often found different azimuth representa- focused on Isl2EphA3/+ mice with mixed-type maps, since we tionsintherightandleftSCsinIsl2EphA3/+miceinwhichbothcol- would be able to compare in the same SC areas of activation liculiwereimaged(Figures2Band2C).Indeed,wefoundthatin thatgiveadoubledmapandareasofactivationthatgiveasingle fiveofsevenmice,theorganizationofmapsintheleftandright map.Isl2EphA3/+micewereshownaseriesofstimuliconsisting SCfellintodistinctcategorieswhenplottingthelinearfitindexas ofshortbarsthatsubtendedonly5(cid:3) ofvisualangleinazimuth a function of the intramap correlation index (Figure 2D). For (Figure3B).TheareasinSCthateachbaractivatedwerecom- example,Figure2CdepictsanIsl2EphA3/+animalinwhichtheaz- bined, and this map was compared to the map derived from 1264 Neuron87,1261–1273,September23,2015ª2015ElsevierInc. Figure3. LackofFunctionalSuppressionin theSCofIsl2EphA3/+Mice (A and B) Schematics of full-field (A) and spatially restricted (B) stimuli used to determine whethersubsetsofSCregionsmaybefunctionally suppressed. (C–F)AzimuthrepresentationsfromtheSCoftwo Isl2EphA3/+micederivedfromtheuseoffull-field stimulus(CandE)orspatiallyrestrictedstimuli(D andF). observed the greatest degree of hetero- geneity. In some colliculi, Isl2-GFP+ RGCsappearedtobeevenlydistributed over the posterior SC (Figure 4A, right hemisphere), while in others, large re- gions appeared devoid of Isl2-GFP+ RGC innervation (Figure 4A, left hemi- sphere, and Figure 4C). In several in- stances we observed dense islands of Isl2-GFP+ innervation (Figures 4A, 4A0 and 4A00, arrows), suggesting that clus- teringoftheseterminalsoccursinasto- chastic manner. To quantitatively asses theheterogeneityofIsl2-GFP+RGCspro- jectionsintheSCofIsl2EphA3/+mice,we comparedtheintensityofGFP+terminals alongtheA-Paxisatthreepositionsalong theM-Laxis(Figures4Cand4D,bottom row). As a control, we measured the in- tensity along the same M-L positions of all RGC terminals, which were labeled conventionalfull-fieldstimuli(Figures3C–3F).Ineverycase(n= by intraocular injection of fluorescently tagged Cholera Toxin 4),themapobtainedfromspatiallyrestrictedstimulimatchedthe subunitB(Figures4Band4D,toprow).Plottingfluorescencein- mapsobtainedwithfull-fieldstimulationverywell.Theseresults tensityasafunctionofthepositionalongtheA-Paxis,wefound confirmthereliabilityofourimagingmethodanddonotsupport thatthedistributionofallRGCswassimilarateachM-Lposition functionalsuppressionasthesourceofheterogeneityobserved acrosscolliculi(Figure4D,toprow).Incontrastthedistributionof inIsl2EphA3/+mice. Isl2-GFP+terminalsshowedmorevariability(Figure4D,bottom row).Toquantifythis,wecalculatedthepairwisecorrelationco- FunctionalHeterogeneityofMapOrganizationin efficients between intensity plots for both Isl2-GFP+ and all Isl2EphA3/+MiceReflectsAnatomicalHeterogeneity RGCs for each SC and determined the average to generate a Anotherpossibleinterpretationofthesedataisthatthehetero- correlation index (n = 9 SCs for all RGCs and n = 12 SCs for geneityintheorganizationofintrinsicopticalsignalmapsreflects Isl2-GFP+RGCs).WefoundthatateachM-Lposition,thecorre- alterations in collicular neuron functionality rather than altered lation index of Isl2-GFP+ RGC projections was significantly projectionpatternsofRGCs.Asaninitialtestofthispossibility, reducedcomparedtothatofallRGCs(Figure4E).Thesedata weexaminedtheprojectionpatternsofIsl2+RGCsusingatrans- demonstrate heterogeneity in the projection patterns of Isl2- genicmouseinwhichGFPisexpressedinthesoma,dendrites, GFP+RGCsubtypesinIsl2EphA3/+mice. andaxons(Triplett etal.,2014).Consistentwithourfunctional Tofurtheranalyzepossiblevariationintheanatomicalorgani- data, we observed heterogeneity in the organization of Isl2- zationinIsl2EphA3/+mice,wetracedRGCprojectionsbyfocalin- GFP+RGC projections inthe SCof Isl2EphA3/+mice(Figure 4). jectionofDiIintothenasalretina,sincetheposteriorSCiswhere As expected, we found the densest Isl2-GFP+ innervation in theobservedheterogeneityisexpressedfunctionally.Incontrast theanteriorSCandareadilyobservable‘‘boundary,’’presum- topreviousreports(Bevinsetal.,2011;Brownetal.,2000;Reber ably at the border of a duplicated map (Figure 4A, dotted line etal.,2004),wefoundthatlabelednasalaxonsinIsl2EphA3/+mice in Figures 4A0 and 4A00). Interestingly, the shape and precise exhibitedvariabilityintheorganizationoftheirterminations(Fig- locationofthis‘‘boundary’’variedbetweencolliculiofdifferent ure5).Consistentwithpreviousstudies,wefoundthatasingle mice and between collicular hemispheres of the same mouse injectioninnasalretinaofadultIsl2EphA3/+miceresultedintwo (Figure 4A0 versus 4A00). Posterior to the ‘‘boundary,’’ we termination zones (TZs) in the SC (55.6%, 10/18) (Figures 5A Neuron87,1261–1273,September23,2015ª2015ElsevierInc. 1265 Figure 4. Heterogeneity in the Projection PatternsofIsl2+RGCsinIsl2EphA3/+Mice (A–A00)DorsalviewsofIsl2-GFPprojectionstothe SC(whitedashedregions)inwholemountofan adultIsl2EphA3/+micerevealdistinctorganizations (A).Enlargedviewsofregionsofinterest(A0 and A00) indicated in (A). Dashed lines indicate the ‘‘border’’ between regions of high and low Isl2- GFP+RGCinnervation.Arrowsindicateclustersof Isl2-GFP+RGCterminals. (B and C) Dorsal views of the projections of all RGCs(B)andIsl2-GFP+RGCs(C)totheleftSCof an adult Isl2EphA3/+ mouse. Dashed purple and greenlinesarelocatedatthe25th,50th,and75th percentileofthemedial-lateralaxisoftheSC. (D)Plotsofthefluorescenceintensitysignalofall RGCs(toprow)andIsl2-GFP+RGCs(bottomrow) fromthreerepresentativeIsl2EphA3/+SCs.Intensity wasplottedasafunctionofpositionalongtheA-P axis at three different parasagittal planes of the M-Laxis,indicatedbydashedlinesin(B)and(C). (E) Quantification of the average correlation indices at each parasagittal plane for all RGCs (blackbars,n=8)andIsl2-GFPRGCs(graybars, n=11).Dataarepresentedasaverage±SEM. Scalebar,0.5mm;L,lateral;P,posterior;OT,optic tract;M,medial;A,anterior;**p<0.01versusAll RGCs;***p<0.001versusAllRGCs. force that would drive map organization to a duplicated state in the Isl2EphA3/+ SC; thus, a counter-balancing force of equalmagnitudemustexistinthedevel- oping SC to drive map organization to the single and mixed-type states. One possible counter-balancing force is correlated spontaneous retinal waves and5G);however,inasubstantialproportionofcases,asingle during early post-natal development, which have been shown injection in nasal retina resulted in one TZ (33.3%, 6/18) or a toplayacriticalroleintopographicmapping,actingtocluster ‘‘smeared’’ TZ (11.1%, 2/18) (Figures 5A–5C). In addition to axonalterminalsofneighboringRGCsandrefinethereceptive examining the TZ pattern of labeled RGCs in these mice, we fields of individual collicular neurons (Chandrasekaran et al., alsoperformedintrinsicsignalopticalimagingintheSCofthe 2005;McLaughlinetal.,2003).Wehypothesizedthatthecorre- samemouseinordertomakedirectcomparisonsofanatomical lated activity generated by spontaneous waves provides a andfunctionalorganizations.Asexpected,functionalmaporga- counter-balancing force strong enough to overcome the eph- nizations reflected anatomical organizations on the whole. rin-A/EphAmolecularforceintheIsl2EphA3/+SC.Totestthispos- Indeed, a c2 test in which single maps were assigned an ex- sibility,weimagedtheSCsofmiceinwhichthenormalpatternof pectedvalueof1TZ,doublemaps2TZs,andmixedmap1.5 spontaneousretinalwavesweredisrupted.Inmicelackingthe TZs, the probability of observing the expected number of TZs b2subunitofthenicotinicacetylcholinereceptor(b2(cid:2)/(cid:2)),retinal isveryhigh(c2 =0.963,c2 =0.987,c2 =0.975). waves have altered spatiotemporal properties, resulting in single double mixed These data suggest that the anatomical organization of nasal abnormal topographic refinement in the SC (Chandrasekaran RGC inputs in the SC of Isl2EphA3/+ mice are heterogeneously etal.,2005;McLaughlinetal.,2003;Staffordetal.,2009).We organizedinasimilarmannertothatobservedforthefunctional crossedIsl2EphA3/+miceintotheb2(cid:2)/(cid:2)backgroundtodetermine representation. theroleofspontaneousretinalwavesinthegenerationofhetero- geneity in topographic organization. While we found that both SpontaneousRetinalActivityRequiredfor the azimuth and elevation representations of Isl2EphA3/+/b2(cid:2)/(cid:2) Heterogeneity mice hadsignificantly decreased peak amplitudes incompari- Ourfindings thusfarsuggestthatEphAsignaling in Isl2EphA3/+ son to those of Isl2EphA3/+/b2+/(cid:2) littermates (Figure S2), we miceisatalevelthatproducesinstabilityandheterogeneityin wereabletoobtaininterpretablemaps(Figure6).Insupportof maporganizationintheposteriorSC.Severallinesofevidence ourhypothesis,wefoundthatinIsl2EphA3/+/b2(cid:2)/(cid:2)mice,retino- suggest that ephrin-A/EphA molecular interactions provide a collicular maps were similar across SCs, all resembling single 1266 Neuron87,1261–1273,September23,2015ª2015ElsevierInc. Figure 5. Correlation of Anatomical and FunctionalMapOrganizationsinIsl2EphA3/+ Mice (A–C)DorsalviewsoftheSCsofIsl2EphA3/+micein whichthenasalretinawasfocallylabeledwithDiI revealsone(A),two(B),orasmeared(C)termi- nationzone(TZ)(n=18). (D–F)Correspondingfunctionalrepresentationsof azimuthtotheIsl2EphA3/+micelabeledin(A)–(C). (G) Map organizations of traced and imaged Isl2EphA3/+SCsplottedbytwoquantitativeindices. Axonlabelingresultinginone(blackcircles),two (red circles), or smeared (green circles) TZs is indicated.Scalebar,0.5mm;A,anterior;L,lateral. studies have revealed this possibility. Indeed,previouscomputationalmodeling ofSCdevelopmenthasbeenusedtovali- date the results of anatomical tracing in Isl2EphA3/EphA3andIsl2EphA3/+mice(Bev- ins et al., 2011; Grimbert and Cang, 2012;KoulakovandTsigankov,2004;Tsi- gankov and Koulakov, 2010). Here, we modifyapreviouslydescribedmathemat- ical model to replicate in silico the vari- abilityweobservedinvivo.Inourmodel, weassumethatallRGCsinitiallyproject to the posterior-most SC, regardless of their cell body location along the T-N axisoftheretina,aspreviouslydescribed (SimonandO’Leary,1992).Indeed,thisis necessary for a stochastic resolution to the push and pull forces of activity and molecules we observe in Isl2EphA3/+ mice. However, recent models have suggested that in the Isl2EphA3/+ and Isl2EphA3/EphA3 mouse models, Isl2+ RGCsareunabletoprojecttotheposte- rior-mostSCduetothehighexpression level of EphA receptors (Grimbert and Cang, 2012). To resolve this issue, we mapsbothqualitativelyandquantitatively(Figures6B,6E,and examinedIsl2-GFP+RGCprojectionpatternsduringthedevel- 6F).Thissurprisingresultmaybeexplainedbytheknownalter- opment of retinocollicular topography. In contrast to previous ationsinspatiotemporalpropertiesofwavesintheb2(cid:2)/(cid:2)mouse models, we find Isl2-GFP+ fibers in the posterior-most SC at (Stafford et al., 2009; see Discussion). Importantly, Isl2EphA3/+/ earlydevelopmentalages(post-natalday2and4[P2andP4]) b2+/(cid:2) littermates exhibited map heterogeneity, as 50.0% (2/4) inIsl2EphA3/+mice(FiguresS3AandS3B).Indeed,evenatages hadsinglemapsand50.0%(2/4)hadmixed-typemaps(Figures inwhichrefinementhasbegunandIsl2-GFP+fibersshowacon- 6Aand6F).Takentogether,thesedatasuggestthatcorrelated centrationintheanteriorSC(P6andP8),asubsetoffiberscan spontaneousretinalactivityactsasacounterbalancetomolec- be found in the posterior SC (Figures S3C and S3D). Taken ularforcestoinfluencethelocalorganizationoftheretinocollic- together with our modeling results, these data suggest that ularmap. Isl2+andIsl2(cid:2)RGCaxonsareintermingledduringdevelopment ofretinocolliculartopographyinIsl2EphA3/+mice,allowingfora InSilicoModelPredictsMapHeterogeneityofMap competition between molecular and activity-dependent forces OrganizationinIsl2EphA3/+Mice thatresolvesitselfinastochasticmanner. Thepresenceofmixed-typemapsandheterogeneityobserved Tobetterdescribethestochasticnatureoftopographicmap in Isl2EphA3/+ mice suggests a stochastic process in map for- formation,wemodifiedthemodelproposedbyTsigankovand mationdrivenbyinteractionsbetweenmolecular-andactivity- Koulakov, in which map formation depends on graded chem- dependentforces.However,neitherexperimentalnormodeling ical cues and activity-dependent refinement (Tsigankov and Neuron87,1261–1273,September23,2015ª2015ElsevierInc. 1267 Koulakov,2006,2010).TheSCisrepresentedbya100by100 matrix to simulate the termination sites for RGC axons (Fig- ure7A).ThechemicalcuesalongtheN-Taxisofthemapsimu- latethegradedexpressionofEphAandephrin-Ainbothretina andSC(seeExperimentalProceduresfordetaileddescriptions). The exogenous EphA3 in the retinas of Isl2EphA3/EphA3 and Isl2EphA3/+ mice is simulated by adding a fixed factor, DR, to the‘‘chemicalcue’’profileofeveryotherRGC;theDRusedfor modelingIsl2EphA3/EphA3miceistwiceasmuchtheoneusedto model Isl2EphA3/+ mice to represent the effect of possessing two Isl2-EphA3 alleles as opposed to one. Initially, the map is random, but it undergoes a series of optimization steps. At each step, two randomly chosen RGCs are exchanged with theprobabilityp=1/[1+exp(4*DE)],whereDE=DE +DE chem act isthechangeinadhesiveenergyduetotheexchange.Ingeneral, DE is minimized when the RGC expressing the highest chem amountsofreceptorprojecttoplacesintheSCexpressingthe lowest amounts of ligand and vice versa, which makes sense giventherepulsivenatureoftheinteractionbetweentheEphAs andephrin-As.DE isminimizedwhenRGCsthatarecloseto act eachotherintheretinaprojectclosetoeachotherintheSC.Af- ter107steps,orderlyandorganizedtopographicmapsappear. WhenDRissettozero,simulatingmapformationinWTanimals, themapsarealwayssingle(Figure7B).WhenDRissettoahigh number (0.70), representing conditions in the Isl2EphA3/EphA3 mouse,themapsarealwaysdoubled(Figure7C).However,at the intermediate value of DR = 0.35, different simulation runs give maps of divergent structures, reminiscent of the maps seen in Isl2EphA3/+ mice (Figure 7D). Some runs give rise to mapsthatarealmostentirelysingle,ordoubled,andthemajority ofmapsaremixed. Previous studies of topographic map formation hint that changes in the sum of EphA in the retina results in alterations ofretinocollicular maporganization (Bevinsetal.,2011;Reber etal.,2004).Todeterminethesensitivityofouroptimalitymodel, wevariedthevalueofDRaroundtheintermediatevaluethatre- sultedinheterogenousoutcomes.Wefoundthatincreasingor decreasingthevalueofDRby±0.05(±16.6%)eliminatedhet- erogeneity and consistently resulted in fully doubled or single maps, respectively (Figure 7E). These data indicate that the in silico model utilized here is able to replicate key aspects of mapformation,includingtheheterogeneityobservedinvivoin Isl2EphA3/+ mice. Further, the sensitivity of map organization to smallchangesinDRtakentogetherwithourfindingthatsponta- neousretinalactivityisrequiredforheterogeneityprovidestrong supportthatmolecular-andactivity-dependentforcesarenearly Figure6. DisruptionofRetinalWavesEliminatesHeterogeneityin equallyinbalanceduringtheformationoftopographicmaps. theAzimuthRepresentationintheSCofIsl2EphA3/+Mice (A and B) Intrinsic signal optical imaging of the azimuth representation in DISCUSSION Isl2EphA3/+/b2+/(cid:2)(n=4)(A)andIsl2EphA3/+/b2(cid:2)/(cid:2)(n=4)(B). (CandD)Intrinsicsignalopticalimagingoftheelevationrepresentationin Theretinocollicularprojectionhasservedasamodeltounder- Isl2EphA3/+/b2+/(cid:2)(C)andIsl2EphA3/+/b2(cid:2)/(cid:2)(D). stand the mechanisms by which topographic projections be- (E)Phaseplotsoftheazimuthrepresentationalongtheanterior-posterioraxis oftheSCatthreeisoelevationpointsfromanIsl2EphA3/+/b2(cid:2)/(cid:2)SC. tween brain regions are established. Important roles for both (F)MaporganizationsofIsl2EphA3/+/b2+/(cid:2)(darkbluedots)andIsl2EphA3/+/b2(cid:2)/(cid:2) molecular cues and activity-dependent forces have been well (lightbluedots)SCsplottedbytwoquantitativeindices.Scalebar,0.5mm;A, established,yethowtheseforcesinteracttomediatemapfor- anterior;L,lateral. mation remains unresolved. Here we demonstrate that local topographic map formation proceeds stochastically, driven by equally balanced molecular- and activity-dependent forces. 1268 Neuron87,1261–1273,September23,2015ª2015ElsevierInc. Figure7. ReplicationofHeterogeneityina Mathematical Model of Retinocollicular Mapping (A)Representationsoftheretinaasa1003100 matrix,withcorrespondingcolor-codingbasedon positionalongtheD-VandT-Naxes. (BandC)Resultsofsimulationsofthemodelrun with thevalueforexogenous EphA3expression (DR) set to 0.00 (for wild-type; B) or 0.70 (for Isl2EphA3/EphA3; C) reveal predicted organizations of projections along the A-P and L-M axes of theSC. (D) Representative examples of three distinct simulationsofthemodelwhenDRwassetto0.35 (forIsl2EphA3/+mice)revealsaheterogeneityinthe organizationofinputsalongtheA-Paxis. (E) Minor changes in DR around 0.35 yielded consistent map organizations that were singular (0.30) or doubled (0.40). D, dorsal; V, ventral; T, temporal; N, nasal; A, anterior; P, posterior; L,lateral;M,medial. RGCs, there was a single TZ in the SC, whereas previous studies reported that all nasal tracings resulted in two TZs. Such a discrepancy may be due to the small number of tracings carried out in previous studies; only 5 were done in thenasalretina(definedas%40%ofthe Thesefindingsnotonlyresolvetheissueoftherelativerolesof N-Taxis)(Brownetal.,2000).Inotherstudies,tracingsofnasal molecularcuesandactivityintheestablishmentoftopography, RGCs in Isl2EphA3/+ mice with altered EphA4 or EphA5 levels butalsoshowanovelstochasticityintheestablishmentofretino- consistently showed two termination zones (Bevins et al., collicularconnectivity. 2011;Reberetal.,2004).However,asthemathematicalmodel we propose here suggests, altering the sum of the EphA HeterogeneityofRetinocollicularMapOrganizationin signaling force even slightly can have dramatic effects on the Isl2EphA3/+HeterozygousKnockinMice consistencyoftheorganizationoftheresultingcollicularmap. Previous investigations of retinocollicular mapping in the Inadditiontotheheterogeneityofmaporganizationsbetween Isl2EphA3knockinmicepointedtowardarelativesignalingmodel colliculi of ls2EphA3/+ mice, the presence of mixed-type maps ofmapformation.InheterozygousIsl2EphA3/+mice,anatomical challenges current models of map formation. Importantly, to tracing suggested that Isl2+ and Isl2(cid:2)RGCs originating in the ourknowledge,neithertherelativesignalingnorthepermissive nasalretinaprojectedtodistinctlocationsintheposteriorSC, arborization models predict the presence of either heteroge- while those from temporal retina ‘‘collapsed’’ into a single neous map organizations between Isl2EphA3/+ SC or multiple map(Brownetal.,2000).Indeed,elegantexperimentsmanipu- maporganizationswithinasingleSC.Incontrast,thestochastic lating the expression level of other retinal EphA receptors optimization model presented here is able to replicate not confirmthismodel,andmathematicalmodelswerederivedto only the heterogeneity of map organizations observed in the describe these data and predict other findings (Bevins et al., Isl2EphA3/+SC,butalsomixed-typemaporganizationsinasingle 2011;Reberetal.,2004).Incontrast,ourfunctionalexperiments SC.Thisisanimportantdistinction,astheobservationofmixed- revealextensiveheterogeneityintheorganizationofretinalin- type maps in the Isl2EphA3/+ SC suggests that the forces of putstotheSCinIsl2EphA3/+mice.Onepotentialexplanationof molecular cues and spontaneous activity are nearly equally thisdifferenceisthattheanatomicaltracingsinpreviousstudies in balance and that the local organization of retinal inputs canonlysampleasmallportionofthemapatatime,whileour is determined stochastically. Incorporating this novel aspect functionalimagingtechniqueallowsustomonitormaporgani- intofuturemodelsofretinocollicularmapformationisaneces- zationasawhole.Thus,examiningmaporganizationatapop- sary step toward the development of accurate computational ulation level may reveal heterogeneity that anatomical tracing simulations. experimentscouldnot.However,wefoundsignificanthetero- geneityintheanatomicalprojectionpatternofDiI-labelednasal Activity-DependentForces,ratherthanGenetic RGCsinIsl2EphA3/+micethatwereconsistentwiththehetero- Modifiers,DriveHeterogeneity geneous functional organization of the collicular map. In Our mathematical model of map formation revealed that 33.3% (6/18) of Isl2EphA3/+ animals in which we traced nasal subtle changes in the expression level of EphAs results in Neuron87,1261–1273,September23,2015ª2015ElsevierInc. 1269
Description: