EMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 1 Rescueofamyotrophiclateralsclerosisphenotypeina mousemodelbyintravenousAAV9-ADAR2deliveryto motorneurons TakenariYamashita,HuiLinChai,SayakaTeramoto,ShojiTsuji,KunikoShimazaki,Shin-ichi MuramatsuandShinKwak Correspondingauthor:Shin Kwak, GraduateSchoolofMedicine,UniversityofTokyo Review timeline: Submissiondate: 22 April2013 EditorialDecision: 25 April2013 Appealreceived: 25 April2013 EditorialDecision: 06 June 2013 Revisionreceived: 18 July 2013 AdditionalEditorialCorrespondence: 17 August2013 AdditionalAuthorCorrespondence: 20 August2013 EditorialDecision: 20 August2013 Revisionreceived: 23 August2013 Accepted: 23 August2013 TransactionReport: (Note:With the exception ofthe correction oftypographicalor spelling errors thatcould be a source ofambiguity, letters and reports are not edited. The originalformatting oflettersand referee reportsmaynotbe reflected in this compilation.) Editor: Céline Carret 1stEditorialDecision 25 April2013 Thank you forthesubmission ofyourmanuscript"Rescueofamyotrophiclateralsclerosis phenotypein amousemodelby intravenousAAV9-ADAR2deliverytomotorneurons." I have now had the opportunity to carefully read your paper and the related literature and Ihave also discussed itwith my colleagues.Iam afraid thatweconcluded thatthemanuscriptisnotwellsuited forpublication in EMBO MolecularMedicine and have therefore decided notto proceed with peer review. Weappreciatethatyourstudyreportson a successfulgene therapy forALS in a mouse model. RestorationofADAR2expressioninneuronsofpre-symptomatic orpost-symptomatic conADAR2- KO miceresultsinimprovedmotorfunctions(butwithoutrestoringgrippower)inanimalsand prevented motorneurondeathbynormalizingTDP-43 expression.However,theunderlying molecularpathwayonwhichthistherapyisbasedwaspreviouslyreported.Moreimportantly,gene therapy improving ALS phenotype was shown before using different genes in differentmouse models.Therefore,whilethestudywillcertainlybeofinteresttotheimmediatecommunity,given the existing literature, the conceptual advance provided by the study is not sufficient for further consideration in EMBO MolecularMedicine.EMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 2 I am sorry thatIcould notbring betternewsthistime. Appealreceived 25 April2013 Thankyouforreviewingthouroughlyourmanuscript. I would like to draw your attention in the following points: 1)OurconditionalADAR2 knockoutmiceisonly themechanisticmodelforthesporadicALSthat accountsforthe majority ofALS patients. PreviousALS modelmiceusedforgenetherapyareforSOD1-associated familialALS and SOD- associated ALS differfrom sporadic ALS in the absence ofTDP-43 pathology and lack ofADAR2 down-regulation.The majority of ALS researchers now consider thatthe two diseases have different etiology despite thatthe phenotype issimilar(therefore the disease name ALS isused forthese two differentdiseases).Therefore,thetherapeutic strategy forthese two diseasesmustbe differentand transgene of the SOD1 gene does not rescue sporadic ALS even if is rescues SOD1-associated ALS. Pleaseunderstandthatitisnotrighttosay"genetherapyimprovingALS phenotypewasshown before using differentgenes in differentmouse models" . 2)Ourstudy aimsto provideevidencethatourmethod oftheADAR2 genedelivery worksin the patientsand scoping clinicaltrialon patientswith ALS.Becausethereisno effectivetherapy for ALSand cureofpatientswith ALS,particularly sporadicALS,iswhatpeopleeagerly desire. Wedemonstratedtherapeuticeffectsourstrategyusinguniquemechanisticmousemodefor sporadic ALS,which may provide patients with hope. I appreciate itvery muchif you could reconsider the possibility of publishing our results in EMBO MolMed. 2nd EditorialDecision 06 June 2013 ThankyouforthesubmissionofyourmanuscripttoEMBO MolecularMedicine.Wehavenow heard back from thethreerefereeswhom weasked to evaluate yourmanuscript.Although the referees find the study to be of potentialinterest,they also raise a number of concerns thatneed to befully addressed in thenextfinalversion ofyourarticle. Asyouwillseefrom theenclosedreports, referee 1 makesvery detailed and clearsuggestionsto improve the conclusiveness of the data. While referees 2 and 3 are more concise, they also ask for additionalcontrolsand fullerpresentation ofthe resultsasthe findingsare currently minimally presented and the discussion is insufficient,obscuring the primary data. In our view the suggested revisions would render the manuscriptmuch more compelling and interesting to a broad readership. We therefore hope that you will be prepared to undertake the recommended experimentalrevision. Pleasenotethatthatitisourjournal'spolicytoallow onlyasingleroundofrevision,andthat acceptance orrejection ofthe manuscriptwilltherefore depend on the completenessofyour response and the satisfaction ofthe refereeswith it. I look forward to seeing a revised form of your manuscriptas soon as possible.EMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 3 ***** Reviewer'scomments***** Referee#1(GeneralRemarks): Yamashitaetal.reportedatherapeuticstudytotesttheeffectofAdar2genedelivery to conditional ADAR2KO (AR2)micewhichshowedALS-like phenotype. The authors found that a single intravenous injection of AAV9-ADAR2improvedmotorfunctioninrotarodtask,motorneuronal loss, axonal degeneration, GluA2 editing activity,and TDP-43 pathology.Itisnoteworthy thata single injection could reverse ALS-phenotypein AR2 micein regard to thetherapeuticpossibility; howeverthemanuscriptisoverallimmatureand remainsmany issuesto beaddressed. MajorCritiques 1.Theauthor'sgroup reported thatAR2 micealso showed abnormalCapermeability,activation of calpain,and TDP-43 fragmentation (Hideyamaetal.,2010,Yamashitaetal.,2012).Itis recommended to demonstrate the rescue effects on them by refilling ADAR2 expression. 2.Itisunclearwhattheratio ofFlag-ADAR2-positiveAHCswasand how much mRNA/protein of ADAR2wererecoveredaftertheAAV treatment.Theauthorspreviouslyreportedthat approximately halfofAHCsin the spinalcord were knocked outforADAR2 genein AR2 mice (Hideyama etal.,2010).This study lacks the information of AHCs with ADAR2 staining.Itis highly recommended to show theratio ofneuronsrecovered with ADAR2 expressionsby injection. ThereadersmayalsowonderwhathappenedinAHCswithout KO as well as non-motorneurons afterFlag-ADAR2overexpression.Didtheyshow twiceormoreamountofADAR2expression than WT did? If so, was excess of ADAR2 expression beneficial or harmful for AHCs? To address these questions, the precise quantification of ADAR2 expression forAHCs and otherneurons in CNSisnecessary. TheauthorsactuallypresentedmRNA levelsofAdar2mRNA inFig4A byqPCR insteadof immunohistochemistry. Although the right bar of total ADAR2 seems that AAV injection achieved twice amounts of ADAR2 mRNA as saline control, the way of measurement is not appropriate. The amountsofmRNA levelsmeasured by differentprimers,differentsetsofprimersformouse and human Adar2,arenotaddablewhen calculated by theratio to thereference gene.Itisnecessary to usethesetofprimerswhich areableto recognizeboth mouseand human ADAR2.Asin the sophisticated data oftheirprevious work (Fig2 in Hideyama etal.,2010),the single-cell-based analysisusing microlaserdisection would bethemostdesirableway to show thepreciselevelsof ADAR2togetherwiththeactivityofGluR2Q/Rsite-editing. 3.Controlmustbeinjection ofAAV9-mockorAAV9-GFPinsteadofsalineonly.Theauthorsstated in Fig 1E-F thatproliferationofboth activated astrocytesshowing increased GFAP immunoreactivity and MAC2-positiveactivated microglialcellswasdetected in theanteriorhorns ofAR2 miceinjected with AAV9- Flag-hADAR2.Thiscould bedueto AAV virusinjection.From this point of view, controlshould beAAV virusinjection. 4.Two injection groupswereapplied forthestudy:11 animalsfortheinjection atthepre- symptomatic stage and 5 forthe injection afterthe onset.Itis confusing since Fig 2A,C,D,and E werefortheentiregroup butonly Fig 2B showed thegroup oftheinjection aftertheonset.Fig 2 mustbeshownseparatelyforeachgroup. 5.Theageofmiceused in Fig3-5 mustbeindicated.Itisalso necessary to specify which treatment group ofmicewasused,injection before or after the onset. 6.Although theageofmicein Fig3 isnotspecified,thenumbersofAHCs/section in Fig3C seem to beapparently discrepantfrom theirpreviouswork (Fig4C in Hideyamaetal.,2010).Theprevious workfoundthenumbersofAHCspersection lessthan 20 atthe age of2 months.Itseemsthatthe age ofmice using in Fig3 wasnearoratthe end pointthatwas9 months. 7.ThecriteriaforTDP-43 positiveAHCsin Fig 5C areuncertain.Thesignalquantification is necessary. 8.In Discussion section (P8,L7-10),theauthorsstated that"Furthermore,theexpression ofthe delivered ADAR2 prevented theprogression ofmotordysfunction and neuronaldeath by EMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 4 normalizing theCa2+ influx through AMPA receptorsand downstream moleculareventsleading to death,such ascalpain activation with resultantmislocalization ofTDP-43 in AR2 mice(Fig 3,4 and 5)".Therewasno such dataofcalcium influx orcalpain in themanuscript. MinorCritiques 1.In P5 L15 whatare"remoteregions"? 2.Wild-type should be added in Fig 3A,3C,4A,and 5C. 3.Thedescription in P6 L16,"apparently,although notsignificantly,improved"isnotscientific. 4.In P12 L12,"2.14x10^-12"mustbe"2.14x10^-12 VG/body". 5.TO-PRO-3 staining wasused fordetecting theentire AHCs in Fig5.TO-PRO-3 isbasically a nuclearstaining method.Thereason mustbementioned oritwould bebetterto bereplaced with a standard neuronal-staining marker. 6.Thereareseveralsentenceswith no spacebeforeparenthesis. Referee#2(Commentson Novelty/ModelSystem): Pleaseseereview. Referee#2(GeneralRemarks): TheauthorshavedeterminedtheabilityofADAR2expressedfrom anAAV vectortorestoreediting function in neurons in a modelforALS,in which endogenous ADAR2 knockoutoutin motor neurons.Iffully validated,thisisan importantobservation. ThecriticaldataappearinFigure2inwhichtheauthorsshow thatAAV-ADAR2expression restores rotarod performance in the mouse model.A motor task,and prevents further declinewhen administered late. In Figure 1D,Flag-ADAR2appearstobeexpressedinneuronsthatarenotCHATpositive.How widespreadistheexpression? Figure4presentthedataontheeffectofADAR2expressioninthemousemodelonGlyA2mRNA. Itwould help to show examplesoftheoriginalPCR data.Thebargraphsarehard to interpret. Specifically,whatisthemeaningofthestatement(Figure4legend)that"therelativeabundanceof hADAR2 mRNA wassignificantly higherin AR2 miceinjected with AAV9-Flag-hADAR2 than in saline injected controls"? There should be absolutely no human ADAR2 mRNA in mouse and if there is a signal, it is possible that the PCR assay is in error. In Figure 4C, there is only a modest change in editing efficiency with AAV expression ofADAR2.How can such asmallchange accountforthe big recovery ofrotarod function seen in Figure 2? Referee#3(CommentsonNovelty/ModelSystem): TheauthorsconsiderthatthedownregulationofadenosinedeaminaseactingonRNA1(ADAR2)is a generalpathogenic mechanism in Amyotrophic LateralSclerosis.TDP43 accumulation in the cytoplasm is,in theirview,related to the expression ofabnormalCa2+-permeableAMPA receptor through activation of calpain. The hypothesis is interesting but has been developed and tested only by theresearch group oftheauthorsand awaitsconfirmation from otherteams. Referee#3(GeneralRemarks): Theauthorshypothesize thatthe downregulation ofadenosine deaminase acting on RNA1 (ADAR2) is a generalpathogenic mechanism in Amyotrophic LateralSclerosis (ALS).TDP43 accumulation in the cytoplasm is,in theirview,related to the expression ofabnormalCa2+-EMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 5 permeable AMPA receptorthrough activation ofcalpain.To testfurthertheirhypothesis,they treated ADAR2 conditional KO mice wich develop a motor deficit (AR2 mice)with an AAV9 vector containing human ADAR2 gene.The treatmentprevented othe neuronallossand thelossofaxons. Thepaperisclear.Iwouldhoweverrecommendprovidinglargerandhigherqualityhistological pictures.Fig 1,E and F arebarely visible. In fig 2,a bar should indicate thatthe significantdifference was between the saline treated and the AAV treatedanimal.Infig3A,nohistologicaldifferenceisvisiblebetweensalineandAAV treated animals.The density ofmyelin fibersappearsto remain very high in the AR2 mice.A betterview of the histological sections should demonstrate themyelinloss.Infig4,CIbelievethatthenusedto compute the SEM isthe numberofsamples(4 samplesby mouse for8 and 7 individualspergroup). Thisexplainstheverylow SEM andthestatisticalsignificancebetweenthegroupsdespitethesmall difference between the means.Ibelieve the correctmethodology would be to compute mean per individual, to calculate a SEM using these (mean) individual values with n=8 and 7 respectively. I am notconvinced thatthe difference would remain significant.Fig 5,A TDP43 IHC is negative in the nuclei but there is apparently no labelling in the cytoplasm in the saline treated animals. Was this expected?Isthere no inclusion in the cytoplasm asin ALS?Could the authorsprovide a betterview ofthepathologicaland rescued anterior horn cells (larger picture,higher magnification) ? 1stRevision - authors'response 18 July 2013 In line with the comments from editors and reviewers,we have provided additionaldata to present the results more clearly with sufficient discussion in the revised manuscript. Below,weprovide responses (beginning ateach arrow) thatspecifically address each of the reviewers’ comments, whichareitalicized. ToReviewer#1: Referee#1(GeneralRemarks): Yamashitaetal.reportedatherapeuticstudytotesttheeffectofAdar2genedeliverytoconditional ADAR2KO (AR2)mice, whichshowedALS-like phenotype. The authors found that a single intravenous injection of AAV9-ADAR2improvedmotorfunction in rotarod task,motorneuronal loss, axonal degeneration, GluA2 editing activity, and TDP-43 pathology.Itisnoteworthythata single injection could reverse ALS-phenotypein AR2 micein regard to thetherapeuticpossibility; howeverthemanuscriptisoverallimmatureand remainsmanyissuesto beaddressed. MajorCritiques 1.Theauthor'sgroup reported thatAR2 micealso showed abnormalCa permeability,activation of calpain,and TDP-43 fragmentation (Hideyama etal.,2010,Yamashita etal., 2012). It is recommended to demonstrate the rescue effects on them by refilling ADAR2 expression. → (Response)Wethankthereviewerfordeepunderstandingofourwork.Wehavedemonstrated that the level of Ca2+ influx through AMPA receptors determinesthecalpain activity and cleavage ofTDP-43 (Yamashitaetal,2012,NatCommun).In response to this comment,we added figures showing TDP-43 mislocalization wasrescued in theAHCsexpressing Flag-hADAR2 in theAR2 micetreatedwithAAV9-Flag-hADAR2 in Supporting Information Fig.S9.Thedescription in the results section (page 7 line 7from the bottom) has been amended accordingly as follows. “ConsistentwiththeeffectivepreventionofthedeathofAHCs, loss or mislocalization of TDP-43 in the AHCs ofsaline-treated AR2 mice was rescued in AAV9-injected AR2 mice, and Flag- expressing AHCsexhibited predominantly nuclearTDP-43 localization (Fig 5 and Supporting Information FigS9).Indeed,the number of AHCs showing normal nuclear localization of TDP-43 wasmarkedlyincreasedintheAAV9-injected AR2 mice compared with the control AR2 mice (Fig 5C).“ 2.Itisunclearwhattheratio ofFlag-ADAR2-positiveAHCswasand how much mRNA/protein of ADAR2wererecoveredaftertheAAV treatment.Theauthors previously reported thatEMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 6 approximatelyhalfofAHCsin thespinalcord wereknocked outforADAR2 genein AR2 mice (Hideyama etal.,2010).This study lacks the information ofAHCs with ADAR2 staining.Itis highly recommended to show the ratio ofneurons recovered with ADAR2 expressions by injection.The readers may also wonder whathappened in AHCs withoutKO as wellas non-motorneuronsafter Flag-ADAR2overexpression.Didtheyshow twiceormoreamountofADAR2expressionthanWT did? Ifso,wasexcess ofADAR2 expression beneficialor harmfulfor AHCs? To address these questions,theprecisequantification ofADAR2 expression forAHCsand otherneuronsin CNS is necessary. TheauthorsactuallypresentedmRNA levelsofAdar2mRNA inFig4A byqPCR instead of immunohistochemistry. Although the right bar of total ADAR2 seems that AAV injection achieved twice amounts of ADAR2 mRNA as saline control, the way of measurement is not appropriate. The amountsofmRNA levelsmeasured bydifferentprimers,differentsetsofprimersformouse and human Adar2,arenotaddablewhen calculated bytheratio to thereferencegene.Itisnecessaryto usethesetofprimerswhich areableto recognizeboth mouseand human ADAR2.Asin the sophisticated data oftheir previouswork(Fig2 in Hideyama etal.,2010),thesingle-cell-based analysisusing microlaserdisection would bethemostdesirablewayto show thepreciselevelsof ADAR2togetherwiththeactivityofGluR2Q/R site-editing. → (Response)Wethankthereviewer for the valuable comment.In line with the comment, we added Figs.3C and 4A,and SupportingInformationFigs.S3-S5andS8.Wedeterminedthe expression levelofADAR2 mRNA using primerpairsthatrecognize both human and mouse ADAR2cDNA (SupportingInformation TableS2)andfound1.5-fold increase in the levelof ADAR2mRNA comparedtowild-type mice (Fig. 4A). The level of ADAR2 protein in the spinal cordsand brainsdid notdifferbetween AR2 mice treated with AAV9-hADAR2 and wild-type mice (Supporting Information Fig.S8).Asthisreviewerrightlypointedout,wecouldhavedefinitive information regarding the ADAR2 gene expression if we could determine the expression level of ADAR2mRNA inindividualneurons,butunfortunatelywecouldnotamplifyFlag mRNA in a single neuron levelto differentiate the AAV9-delivered and –non-delivered motorneuronsand the level of ADAR2 mRNA expression is below the level of single cell-based analysis. WehaveamendedthesentencesintheResultssection(page7line 12) as follows. “TherelativeabundanceofmouseADAR2didnotsignificantlydifferbetweentheAAV9-injected and saline-injected AR2 mouse spinal cords (Fig 4A) but AAV9-Flag-hADAR2 infection induced 1.5-fold increase in the expression levelof totalADAR2mRNA inthespinalcords(Fig4A)and brains(SupportingInformationFigS8).MessengerRNA ofbothhADAR2andcholine acetyltransferase (ChAT)wasdemonstrated in the spinalcord lysatesofAAV9-injected AR2 mice (Fig.4B),indicating thathADAR2 wasdelivered to cholinergic AHCs.Theediting efficiency atthe GluA2Q/Rsitewassignificantly higherin the AAV9-injected AR2 mice than in the control AR2 mice(Fig 4C),although the expression levelof ADAR2 protein did notsignificantly increase in the AAV9-injected mice (SupportingInformationFigS8B,C). “ AsfortoxicityofadditiveexpressionoftheexogenousADAR2geneandAAV9 infection, we added figures demonstrating lack of glial reactions in the spinal cord of wild-type mice injected with AAV9-GFP(SupportingInformationFigS4)andofAR2miceinjectedwithAAV9- Flag-hADAR2 (SupportingInformationFigS3,S5).Theseresultswereinmarkedcontrastwiththe markedproliferationofGFAP-positiveactivated astrocytesand MAC2-postiveactivated microglial cellsin the spinalcord ofAR2 mice treated with saline (SupportingInformationFig.S5).A previousreportdemonstrated thattheADAR2 transgenicmicewerephenotypically normalexcept formoderate obesity (Singh M.etal;2007,JBC),suggesting the safety of expression of the exogenousADAR2 gene.Furthermore,we found largernumbersofAHCsin the AAV-treated AR2 micecomparedtosaline-treated ones with considerable proportions of Flag-ADAR2positiveAHCs (Fig 3C and SupportingInformationFig.S3D),indicatingthattreatmentwithAAV9-Flag-ADAR2 rescued AHCs from death in AR2 mice.These results indicate thatneither AAV9 infection nor ADAR2overexpressionistoxictomotorandnon-motorneuronsinmiceandAAV9-Flag-ADAR2 treatment is beneficialto AR2 mice. Also,wehaveamendedtheResultsandDiscussionsectionsasfollows. “In addition,proliferation ofboth activated astrocytesshowing increased GFAP immunoreactivity and MAC2-positiveactivated microglialcellswasnotdetected in thespinalcords,including the regions around AAV9-infected neurons of wild-type mice injected with AAV9-GFPandofAR2 EMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 7 miceinjectedwithAAV9-Flag-hADAR2 (Fig 1E, F andSupportingInformationFigS4,S5).There wasnosignificantexpressionofFlag protein in theperipheralorgans,aspreviously reported (Iwata etal,2013). Theseresultsindicatethattheintravenously injected AAV9 vectordeliverstheADAR2 geneto neuronsand expression ofthedelivered ADAR2in neurons is not toxic without inducing abnormalglialcellreaction.” (Page 5,line 3from the bottom) “Furthermore,theexpressionofthedeliveredADAR2did notinduceany adverseeffectsin neurons orsurrounding tissuesand prevented theprogression ofmotordysfunction and neuronaldeath with restoring ADAR2-mediatedRNAeditingin AR2 mice (Fig3,4and5), which provide a mechanistic modelofsporadicALS(Hideyamaetal,2010; Yamashitaetal,2012a). ” (Page 9,line 7) 3.Controlmustbeinjection ofAAV9-mockorAAV9-GFPinsteadofsaline only.The authorsstated in Fig 1E-F thatproliferationofbothactivatedastrocytesshowingincreasedGFAP immunoreactivity and MAC2-positiveactivated microglialcellswasdetected in theanteriorhorns ofAR2 miceinjected with AAV9- Flag-hADAR2.Thiscould be due to AAV virusinjection.From this pointofview,controlshould beAAV virusinjection. → (Response)Wearesorrytoconfusethereviewersandthankthereviewer’scarefulreading.Due to an unknown reason, the word ‘not’ was spelled out from thetext(page5 line1 from thebottom; “--- activated microglialcellswas(not)detected ---“)Asdescribed in theresponseto theprevious comments,injection of AAV9-GFPtowild-type mice did not cause activation of astrocytes or microgliainthespinalcord including theregion around theGFP-positiveneurons(Supporting Information Fig.S4).Likewise,therewasnoincreaseofGFAP orMAC2immnuoreacitivityaround the Flag-positiveneuronsin AR2 micetreated with AAV9-Flag-ADAR2despitethattherewas markedglialproliferationinthespinalcordofAR2micetreatedwithsaline(Supporting Information FigS5).BecauseAAV9infectionandADAR2overexpressionisnottoxictoneurons and ADAR2 expression rescued AHCsfrom death,we hope the reviewerwill accept the use of saline-injected AR2 mice as controls. 4.Two injection groupswereapplied forthestudy:11 animalsfortheinjection atthepre- symptomatic stage and 5 for the injection after the onset.Itis confusing since Fig 2A,C,D,and E were forthe entire group butonly Fig 2B showed the group ofthe injection afterthe onset.Fig 2 mustbeshownseparatelyforeachgroup. → (Response)Tofacilitatereader’sunderstanding,wepresentedseparatelytherotarodperformance ofAR2 micetreated in the pre-symptomatic stage and those treated in the post-symptomatic stage in SupportingInformationFig.S6. 5.Theageofmiceused in Fig3-5 mustbeindicated.Itisalso necessaryto specifywhich treatment group ofmicewasused,injection before or after the onset. → (Response)Accordingtothereviewer’scomment,weamendedfigurelegendsofFig.3-5 and added description aboutthe mice used forimmunohistochemistry and biochemicalanalysesin the MaterialsandMethodssectionasfollows. “AR2miceinjectedwithAAV9-hADAR2 vectorsbeforetheinitiation ofmotordysfunction were used forimmunohistochemistry and biochemicalanalysesat36 weeksofage(n = 5-8).“(Page15, line 7) 6.Although theageofmicein Fig3 isnotspecified,thenumbersofAHCs/sectioninFig3Cseem to beapparentlydiscrepantfrom theirpreviouswork(Fig4C in Hideyama etal.,2010).Theprevious workfoundthenumbersofAHCspersectionlessthan20attheageof2months.Itseemsthatthe ageofmiceusing in Fig3 was near or at the end point that was 9 months. → (Response)ThedifferenceinthenumbersofAHCspersectionbetweenthisreportandour previousonemay bedueto thedifferencein thecriteriaforAHCs.AsdescribedintheMaterials and Methodssection,we counted large anterior horn cells (AHCs) with diameters larger than 20 µm in spinal cord sections stained with TO-PRO-3.On theotherhand,wecounted AHCslarger than 20 µm indiameteranddouble-immunopositive for ADAR2 and SMI-32 using DAB as a colourEMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 8 developerin ourpreviousreport(Hideyama etal.,2010). The reviewer will admit that consistent results were obtained using the method adopted in this study 7.Thecriteria forTDP-43 positiveAHCsin Fig 5C areuncertain.Thesignalquantification is necessary. → (Response)TherewereTDP-43-positiveand –negativeAHCs(Merge/TO-PRO-3 in Fig.5A). WedesignatedanAHCasTDP-43-negaivewhen wecould notdetectTDP-43 immunofluorescence in the AHC even in the highest sensitivity level of fluorescencemicroscopy and counted these AHCsasTDP-43(-) in Fig.5C. 8.In Discussion section (P8,L7-10),theauthorsstated that"Furthermore,theexpression ofthe delivered ADAR2 prevented theprogression ofmotordysfunction and neuronaldeath by normalizing theCa2+ influxthrough AMPA receptorsand downstream moleculareventsleading to death,such ascalpain activation with resultantmislocalization ofTDP-43 in AR2 mice(Fig 3,4 and 5)".Therewasno such data ofcalcium influxorcalpain in themanuscript. → (Response)In line with the reviewer’s appropriate comment,we changed the sentence to describe the summary of the present results as follows, “Furthermore,theexpressionofthedeliveredADAR2didnotinduceanyadverseeffectsinneurons orsurrounding tissuesand prevented theprogression ofmotordysfunction and neuronaldeath with restoring ADAR2-mediatedRNAeditingin AR2 mice (Fig3,4and5),” (Page9,line7) In addition,we inserted the mechanistic discussion on how ADAR2 delivery rescued TDP-43 mislocalizationtopage10,line9,asfollows. “Therefore,itislikelythatADAR2deliverynormalizedthesubcellularlocalizationofTDP-43 in the AHCs by reducing Ca2+ influx through AMPA receptors and the resulting calpain activation.” MinorCritiques 1.In P5 L15 whatare"remoteregions"? → (Response)Wehaveamended‘‘remoteregions’’to‘‘remote regions(ipsilateralhemisphere 2.0-2.5 mm posteriorto theinjection site)’’ (page 5 line 16). 2.Wild-type should be added in Fig 3A, 3C, 4A, and 5C. → (Response)Wehaveaddeddataonwild-type mice in Figs. 3A, 3C, 4A and 5C. 3.Thedescription in P6 L16,"apparently,although notsignificantly,improved"isnotscientific. → (Response)Wehave amended “apparently,although notsignificantly,improved’’to ‘‘AAV9- injected AR2 mice exhibited higher spontaneous locomotor activity than saline-injected AR2 mice, butthedifferencebetween thetwo groupsdid notreach statisticalsignificance(Fig 2B).Thegrip powerand body weightdid notdifferbetween theAAV9-injected and saline-injected AR2 mice (Fig 2C,D).’’ (Page6 line5 from thebottom ) . 4.In P12 L12,"2.14x10^-12"mustbe"2.14x10^-12 VG/body". → (Response)Weappreciateyourappropriatecommentandhaveamended "2.14x10^-12"to "2.14 x 1012 vg/body". 5.TO-PRO-3 staining wasused fordetecting theentireAHCsin Fig5.TO-PRO-3 isbasicallya nuclearstaining method.Thereason mustbementioned oritwould bebetterto bereplaced with a standard neuronal-staining marker. → (Response)Wethankthereviewerforthiscomment.BecauseTO-PRO-3 stainsnotonly the nucleusbutthecytoplasm ofneuronsin thefixed sectionsofmousebrainsand spinalcords,to avoid readers’ misunderstanding,wehavechangedthesentenceasfollows. (Figure 5 legends) ‘‘TO-PRO-3 wasused asacellmarker’’.EMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 9 6.Thereareseveralsentenceswith no spacebeforeparenthesis. → (Response)Wethankthereviewerforthiscommentandhad amended thesentences. ToReviewer#2: Referee#2(GeneralRemarks): TheauthorshavedeterminedtheabilityofADAR2expressedfrom anAAV vectortorestoreediting function in neurons in a model for ALS, in which endogenous ADAR2 knockout out inmotor neurons.Iffullyvalidated,thisisan importantobservation. ThecriticaldataappearinFigure2inwhichtheauthorsshow thatAAV-ADAR2expression restores rotarod performance in the mouse model.A motor task,and prevents further decline when administered late. In Figure 1D,Flag-ADAR2appearstobeexpressedinneuronsthatarenotCHATpositive.How widespreadistheexpression? → (Response)Wethankthereviewer’sdeepunderstandingandvaluablecomments.Becausewe used theSYN1 promoter,expression ofthe gene delivered with the AAV9 vectorisuniversalin the centralneurons(Iwata etal,2013 SciReport).In line with this suggestion,we indicated the expression ofFlag-ADAR2inthespinalcordinSupportingInformationFig.S3A andC. Figure4presentthedataontheeffectofADAR2expressioninthemousemodelonGlyA2mRNA.It wouldhelptoshow examplesoftheoriginalPCRdata.Thebargraphsarehardtointerpret. Specifically,whatisthemeaning ofthestatement(Figure4 legend) that"the relative abundance of hADAR2 mRNA wassignificantlyhigherin AR2 miceinjected with AAV9-Flag-hADAR2 than in saline injected controls"? There should be absolutely no human ADAR2 mRNA in mouse and if there is a signal, it is possible that the PCRassayisinerror.InFigure4C,thereisonlyamodest change in editing efficiency with AAV expression ofADAR2.How can such a smallchange account for the big recovery of rotarod function seen in Figure 2? → (Response)WequantifiedADAR2mRNAusing alightcycler480 (RocheInstruments),and the results were presented as computerized ones using internalstandards for reference (Sawadaetal, 2009; Yamashitaetal,2012c). Therefore, we could not provide “original PCR data”. Asthereviewerrightlypointedout,nosignalforhumanADAR2mRNA shouldbe detectablein thesamplesfrom saline-injected AR2.Actually,however,very low levelsofsignal could notbe eliminated from reasonsresiding in quantitative PCR system.To facilitate the readers’ understanding,weadded thedataillustrating theupregulation oftotalADAR2 (using primerpairs that recognize both human and mouse ADAR2 cDNA) to Fig. 4A and amended the figure legend as follows. “TherelativeabundanceofmouseADAR2mRNA didnotsignificantly differbetween AR2 mice injected with AAV9-Flag-hADAR2 (AAV)and thoseinjected with saline (Saline;n=5 foreach group).hADAR2mRNA wasexpressedatasignificantlevelinAAV.Therelativeabundanceoftotal (human and mouse) ADAR2mRNA was1.5-fold higherin AAV (n=4) than in Saline (n=4).**p < 0.01 (Mann–WhitneyU-test).Allerrorbarsrepresentthe s.e.m.” AsfortheresultspresentedinFig.4C,wewouldlikeattractthereviewer’sattentionto the fact that the number of remaining AHCs is about 30% higher in the AAV group than in the saline group (Fig.3C and Supplementary Information Fig. S3D). Because the majority of AHCs that expressed unedited GluA2 had already disappeared atthe age ofexamination,editing efficiency was compared between the remaining AHCsthatexpressADAR2.Therefore,apparently smalldecrease in the expression level of unedited GluA2 may result from the significant increase in the expression level of ADAR2 with a big difference in neuronal functions. Larger number and a higher ADAR2 activity in the remaining AHCsofAAV-treated AR2 mice likely resultin betterbehaviouraland morphologicalchangesthaninsaline-treated AR2 mice. EMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 10 ToReviewer#3: TheauthorsconsiderthatthedownregulationofadenosinedeaminaseactingonRNA1(ADAR2)is a generalpathogenicmechanism in AmyotrophicLateralSclerosis.TDP43 accumulation in the cytoplasm is,in theirview,related to the expression ofabnormalCa2+-permeableAMPA receptor through activation of calpain. The hypothesis is interesting but has been developed and tested only bytheresearch group oftheauthorsand awaitsconfirmation from otherteams. → (Response)Weappreciateyourcommentandconfirmationstudyofourhypothesisbyother teams is highly welcome. Referee#3(GeneralRemarks): TheauthorshypothesizethatthedownregulationofadenosinedeaminaseactingonRNA1(ADAR2) is a general pathogenic mechanism in Amyotrophic Lateral Sclerosis (ALS). TDP43 accumulation in the cytoplasm is, in their view, related to the expression of abnormalCa2+-permeableAMPA receptor through activation ofcalpain.To testfurther their hypothesis,they treated ADAR2 conditionalKO mice whichdevelop a motordeficit(AR2 mice)with an AAV9 vectorcontaining human ADAR2 gene.Thetreatmentprevented otherneuronallossand thelossofaxons. Thepaperisclear.Iwouldhoweverrecommendprovidinglargerandhigherqualityhistological pictures.Fig 1,E and F arebarelyvisible. → (Response)Wethankthereviewerforthecomment.Wehaveaddedbetterhistologicalpictures in SupportingInformationFigS5. In fig 2,a bar should indicate thatthe significantdifference was between the saline treated and the AAV treatedanimal. → (Response)WethankthereviewerforthecommentandhaveamendedFig2and figurelegend to indicate that the difference was between the saline- and AAV-treated groups. In fig 3 A,no histologicaldifference is visible between saline and AAV treated animals.The density ofmyelin fibres appearsto remain veryhigh in theAR2 mice. A better view of the histological sections should demonstrate the myelin loss. → (Response)Wethankthereviewerforthecommentandhaveaddedadditionalexamplesthat demonstrateaxonaldegeneration in theventralrootsofthesaline-treated but not of the AAV-treated AR2miceinSupportingInformationFigS7tofacilitatereaders’understanding. In fig 4,C I believe thatthe n used to compute the SEM is the number ofsamples (4 samples by mousefor8and7individualspergroup).ThisexplainstheverylowSEM andthestatistical significance between the groups despite the smalldifference between the means.Ibelieve the correctmethodology would be to compute mean perindividual,to calculate a SEM using these (mean) individualvalues with n=8 and 7 respectively.I am notconvinced thatthe difference would remain significant. → (Response) Wealsotestedthedifferencebetweenthegroupsandfoundthedifferencewas significant(Fig.4C legend).As described in the response to the previous comments,the AHCs expressing abundantunedited GluA2 had disappeared atthe time ofexamination and wecompared remaining AHCs thatexpress ADAR2 between the two groups.An increase of editing efficiency in the AHCs in the AAV-treated group reflects the increase in the ADAR2 level resulting from additionalexpression ofthe delivered ADAR2gene. Fig5,A TDP43 IHC isnegativein thenucleibutthereisapparentlyno labelling in thecytoplasm in the saline treated animals. Wasthis expected? Is there no inclusion in the cytoplasm as in ALS? Couldtheauthorsprovideabetterview ofthepathologicaland rescued anteriorhorn cells(larger picture,highermagnification)? EMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 11 → (Response)WehaveaddedSupportingInformationFig.S9toshow theTDP-43 pathology in AR2micemoreclearlyandhow itwasrescuedbyAAV-ADAR2.Asthereviewerpredicted,afew motor neurons exhibit multiple TDP-43-positiveaggregatesin thecytoplasm (Supporting Information Fig.S9B).TheseneuronslacknuclearTDP-43 immunoreactivity aswell,displaying abnormalview similarto ALS motorneurons. Also,wehaveamendedtheresults section (Page 7 line 7from the bottom) as follows. “Consistentwith theeffectiveprevention ofthedeath ofAHCs,loss or mislocalization of TDP-43 in the AHCs of saline-treated AR2 mice was rescued in AAV9-injected AR2 mice, and Flag- expressing AHCsexhibited predominantly nuclearTDP-43 localization (Fig 5 and Supporting Information FigS9).” Wearelookingforwardtohearingfavourable decision on thisrevised manuscript. AdditionalEditorialCorrespondence 17 August2013 Thankyouforresubmittingyourmanuscriptforourconsideration.AsIwouldliketo makeaninformeddecisiononyourmanuscript,Ineedclarificationfrom yourpart beforeto proceed. I have now received the two sets of reviews I asked for and I am afraid thatreferee 1 is still not satisfied (please see below) and remains concerned about the similar ADAR2mRNA expressionreportedbetweensalinecontrolandWT(Figure4A).We agree with the referee thatthisresultisindeed confusing asa difference needsto be shown to validate thatthe gene therapy approach using a single systemic delivery fully works. Assuch,Iwouldappreciateifyoucouldletmeknow whetheryouwouldbeprepared to convincingly address this issue for the paper to be accepted and how you would proposeto do so (using theproposed experimentorsomething elseto sameeffect). Pleaseseebelow thereportprovidedbyreferee1: Theauthorsaddedsomeexperimentstoanswerthecriticsraisedbythereviewer and themanuscriptwas improved partially.Nevertheless,there were stillseveral issues with regard to the validation of experimental models, which made the entire manuscriptunconvincing.Thereforethereviewerrecommendstheauthorstosend the manuscript to a morespecificjournalaftercarefulrevision. 1. Itwas appreciable thatthe authors used the primers thatrecognize both mouse and human ADAR2 mRNA forgeneexpression analysisin Figure4A.However,there wasnodifferenceoftotalADAR2mRNAexpression between salinecontroland WT (Figure 4A,right),which means there was no reduction ofADAR2 mRNA levels in saline controlofADAR2 KO mice compared to WT control.Itis quite confusing for readers and two possibilities can explain the results;there wasnotenough reduction ofADAR2 mRNA in the ADAR2 KO mice used in the study for some reason, orthemethods/technologyused in Fig 4 werenotgood enough to confirm endogenous/exogenousgene expression levelsofADAR2.Iguessitwasbecause the entire tissue homogenization process masked motor neuron-specific knock-out ofADAR2 attheventralhorn astheprotein levelsofADAR2 wasalso comparablein Fig.S8.SincethevirtueofthismodelisthewholebodyadministrationofAAV- ADAR2bya"singleintravenousinjection",theauthorsshould validateexpression of delivered genein thetargeted neuronsprecisely.Therefore,in thefirstreview commentsthe authorswere recommended to confirm ADAR2 mRNA levelsby single cellcapture-based method whichhadbeenusedintheirpreviousstudy(Hideyama etal.,2010)ifimmunohistochemistry wasnotapplicable.Unfortunately,the authors EMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 12 just mentioned "level of ADAR2 mRNA expression is below the level of single cell- based analysis",although theysuccessfully showed the data ofADAR2 mRNA changesin single cellscaptured by lasermicro dissection (Fig2 in Hideyama etal., 2010). 2. ThewayofTDP-43 positiveAHCsmeasurementwasstillsubjectiveand not acceptable.Theyshould useobjectivecriteria for it, such as certain cut-offsignal intensity. AdditionalAuthorCorrespondence 20 August2013 Thankyouforthee-mailregardingoursubmittedmanuscript.BelowIam writingtheresponseto the comments from Reviewer 1. Thecriticsraised by thereviewer1 seem to bethree-fold. 1)Thereason why thereisno differencein theADAR2 expression levelbetween thewild-type mice and saline controlAR2 mice. 2)Lack ofevidenceofexpression ofAAV-delivered ADAR2 to motorneuronsaftersinglesystemic injection of AAV9 vector. 3)Validation ofTDP-43 immunohistochemicalpositivity. Response: 1)Aswehaveresponded to thepreviouscomments,qPCR forADAR2 in asingleneuron levelis notpossible.Wetried to validatetheqPCR forADADR2 in themousespinalcord and laser- captured motorneurons,butfound thatspinalcord tissue largerthan 1 mm-thick axial slice was necessary forreliableresults.Thismeansthatthousandsormoreofmotorneuronsarenecessary. Wetriedtodemonstratethe reduction in ADAR2 mRNA expression in AR2 mice compared to wild- type mice by qPCR but were unsuccessful. This may be due to the fact that less than 50% of reduction in the number of motor neurons may be masked in the remaining motor and non-motor neuronsin theventralhorns,asreviewer1 rightly predicted. Althoughthereviewer1statedthatwepreviouslyreportedtheresultsofqPCRforADAR2ina single celllevelin Fig.2 ofHideyama etal.2010 J Neurosci,whatwe presented there was the ordinary PCR forADAR2 on pooled mousemotorneuron lysatesbutnottheqPCR data.We presented qPCR forADAR2 on pooled human motorneuron lysatesin Fig.2 ofHideyamaetal., 2012 NeurobiolDis,buttheexpression levelofADAR2 mRNA wasmorethan one-orderhigherin human motorneuronsthan in mousemotorneurons.Anotherthing oneshould takeinto accountwas that the majority of ADAR2-lacking motor neurons had been already disappeared and the majority ofremaining motorneuronswerethoseexpressing ADAR2 in the control AR2 mice at the age of examination.Therefore,itwould be technically difficultto differentiate 10-15% differenceof ADAR2mRNA expressionleveleveninthepooledmotorneuronsbetweenthesalinecontrolAR2 miceandwild-type mice as stated above.We cannottellthe difference in the expression levelof ADAR2mRNA bymeansofordinaryPCR.Theremaybesomemisunderstanding. Wepresentedthesignificantreductioninthenumberofmotorneuronsandaxonsintheventralroot ofsalinecontrolAR2micecomparedtowild-type mice in Fig. 3, which was consistent with the results in the originalreport(Hideyama etal.,2010 J Neurosci) indicating thatlack of ADAR2 induced death in the motor neurons in the AR2 mice. 2)To fillthedefectofquantitative data on ADAR2 expression, we presented the GluA2 Q/R site- editing in a single motorneuron levelin Fig.4C.Because GluA2 Q/R site-editing isspecifically catalyzed by ADAR2,an increase in the editing efficiency in the AAV group compared to the saline controlAR2 mice indicate thatexogenousADAR2 issuccessfully delivered to and expressed in the motorneurons.ThismethodforADAR2activityinasinglecellwasthesameasweadoptedinFig. 2A ofHideyamaetal,2010,JNeurosci.Wealso presented that the AAV vector was delivered to motorneuronsbydemonstratingFlagexpressioninChAT-positivelargeneuronsin theanterior horn ofthespinalcord in Fig.1D and moreADAR2-positiveAHCswereremained in AAV-treated AR2micethaninthesalinecontrolAR2 mice in Fig.S3D.We believe thatthese data sufficiently indicate that exogenously delivered ADAR2 was successfully expressed in the motor neurons by EMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 13 single systemic injection ofthe AAV vector. 3)AsfortheTDP-43 immunohistochemistry,in line with the reviewer 1's comment, we adopted the criteria thatthe positivity isabove 3-fold of background.Forreviewer1's convenience,the results of quantification areattached and themethod used isindicated in theMariterialsand Methodssection. Theresultsdidnotchangeafteradoptingthecriteria. Webelievewecouldappropriatelyrespondtoalltheconcernsraisedbyreviewer1'sandtheeditors. 3rd EditorialDecision 20 August2013 ThankyouforthesubmissionofyourrevisedmanuscripttoEMBO MolecularMedicineandyour furthercorrespondence. Wehavenowreceivedtheenclosedreportsfromtherefereesthatwereaskedtore-assessit.Asyou willseefrom thepastedcommentsbelow andfollowingourrecente-mailexchange,Iam happyto let you know that the reviewers are now supportive and that we will be able to accept your manuscriptpendingthefollowingfinalamendments: - Pleasemodifythefinaltextofthemanuscripttoremoveanybluetext. - In lightof the documentprovided to referee 1,would you agree to incorporate this data within the supplementalinformation and therefore change the textaccordingly? Similarly,as you may know wewillpublishthereferees'reportsalong with yourreply to theircommentsonline and will incorporate our latest discussion (see below). Please do let us know immediately if you would object. Pleasesubmityourrevisedmanuscriptassoonaspossible. I look forward to hearing from yousoon. ***** Reviewer'scomments***** Referee#1(GeneralRemarks): Theauthorsaddedsomeexperimentstoanswerthecriticsraisedbythereviewerandthemanuscript wasimprovedpartially.Nevertheless,therewerestillseveralissueswithregardtothevalidationof experimentalmodels,which made the entire manuscriptunconvincing.Thereforethereviewer recommends the authors to send the manuscriptfor more specific journalafter carefulrevision. 1.Itwasappreciablethattheauthorsused theprimersthatrecognizeboth mouseand human ADAR2mRNA forgeneexpression analysisin Figure4A.However,therewasno differenceof total ADAR2 mRNA expression between saline control and WT (Figure 4A, right), which means there was no reduction of ADAR2 mRNA levels in saline control of ADAR2 KO mice compared to WTcontrol.Itis quite confusing for readers and two possibilities can explain the results; there was notenough reduction ofADAR2 mRNA in theADAR2 KO miceused in thestudy forsomereason, orthemethods/technology used in Fig 4 werenotgood enough to confirm endogenous/exogenous geneexpression levelsofADAR2.Iguessitwasbecausetheentiretissuehomogenization process maskedmotorneuron-specific knock-outofADAR2 attheventralhorn astheprotein levelsof ADAR2wasalsocomparableinFig.S8.Sincethevirtue of this model is the whole body administration ofAAV-ADAR2bya"singleintravenousinjection",theauthorsshouldvalidate expression ofdelivered gene in the targeted neuronsprecisely.Therefore,in the firstreview commentsthe authorswasrecommendedtoconfirm ADAR2mRNAlevelsbysinglecellcapture- based method which had been used in theirpreviousstudy (Hideyamaetal.,2010)if immunohistochemistry was not applicable. Unfortunately, the authors just mentioned "level of ADAR2mRNA expression isbelow thelevelofsinglecell-based analysis",although they EMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 14 successfully showed the data ofADAR2 mRNA changes in single cells captured by lasermicro dissection (Fig2 in Hideyamaetal.,2010). 2.Theway ofTDP-43 positiveAHCsmeasurementwasstill subjective and not acceptable. They should use objective criteria forit,such as certain cut-offsignalintensity. Referee#2(CommentsonNovelty/ModelSystem): Themodelsystem isadequateandwasdevelopedbytheauthors. Referee#2(GeneralRemarks): Theauthorshaveaddressedmyconcerns.Themedicalquestionposedhereishighlysignificant although ithasbeen pursued almostexclusively by the authors.The data supporttheirconclusions although there isclearly substantialscope forfurtherinvestigation of this modeland the underlying medicalproblem. 2nd Revision - authors'response 23 August2013 Wearegratefultohaveheardyourfavourable decision with appropriatecommentsofthereviewers. Wecarefullyrevisedourmanuscript,figuresand supplementary figuresand tablesin linewith reviewers’ comments as described bellow.In addition,we made some changes for the refinementof style.The totalnumberofwords in the abstractis 173 and the totalnumberofwords in the maintext is 7893. The revised manuscript is with 29 references, 5 figures, 10 SupportingInformationfigures and 2 SupportingInformationtables. In line with the comments from editors and reviewers,we have provided additionaldata to present the results more clearly with sufficientdiscussion in the revised manuscript.Below,weprovide responses (beginning ateach arrow) thatspecifically address each of the reviewers’ comments, whichareitalicized. ToReviewer#1: Referee#1(GeneralRemarks): Theauthorsadded someexperimentsto answerthecriticsraised bythereviewerand the manuscriptwasimprovedpartially.Nevertheless,therewerestillseveralissueswithregardtothe validation ofexperimentalmodels,which made the entire manuscriptunconvincing.Therefore the reviewer recommends the authors to send the manuscriptfor more specific journalafter careful revision. 1.Itwasappreciablethattheauthorsused theprimersthatrecognizeboth mouseand human ADAR2mRNA forgeneexpressionanalysis in Figure 4A. However, there was no difference of total ADAR2mRNA expressionbetweensalinecontrolandWT(Figure4A,right),whichmeansthere wasnoreductionofADAR2mRNAlevelsinsalinecontrolofADAR2KO micecomparedtoWT control.Itisquiteconfusing forreadersand two possibilitiescan explain theresults;therewasnot enough reduction ofADAR2 mRNA in the ADAR2 KO mice used in the study forsome reason,orthe methods/technologyusedinFig4werenotgoodenoughtoconfirm endogenous/exogenous gene expression levelsofADAR2.Iguessitwasbecause the entire tissue homogenization processmasked motorneuron-specific knock-outofADAR2 attheventralhorn astheprotein levelsofADAR2 was also comparablein Fig.S8.Sincethevirtue ofthismodelisthewholebodyadministration ofAAV- ADAR2bya"singleintravenousinjection",the authorsshould validateexpression ofdelivered gene in the targeted neurons precisely. Therefore, in the first review comments the authors was recommended to confirm ADAR2 mRNA levels by single cellcapture-based method which had been EMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 15 used in theirpreviousstudy(Hideyama etal.,2010)ifimmunohistochemistrywasnotapplicable. Unfortunately,theauthorsjustmentioned,"levelofADAR2 mRNA expression is below the levelof single cell-based analysis",although theysuccessfullyshowed thedata ofADAR2 mRNA changesin single cells captured by laser micro dissection (Fig2 in Hideyama etal.,2010). → (Response)Thepointsofcriticsraised by the reviewer1 could be summarized asthe followings. 1) ThereasonwhythereisnodifferenceintheADAR2expression levelbetween the wild-type miceandsalinecontrolAR2mice. 2) LackofevidenceofexpressionofAAV-delivered ADAR2 to motorneuronsaftersingle systemic injection ofAAV9 vector. 1)Aswehaverespondedtothereviewer’scommentsforthe originalmanuscript,qPCR forADAR2 in a single neuron level is not possible. We tried to validate the qPCR for ADADR2 in the mouse spinalcord and laser-captured motorneurons,butfound thatspinalcord tissue largerthan 1 mm- thick axial slice was necessary forreliable results.Thismeansthatthousandsormore ofmotor neuronsarenecessary.Wetried to demonstratethereduction in ADAR2 mRNA expression in AR2 micecomparedtowild-type mice by qPCR but were unsuccessful. This may be due to the fact that less than 50% of reduction in the number of motor neurons may be masked in the remaining motor and non-motorneuronsintheventralhorns,asreviewer1rightlypredicted. Althoughthereviewer1statedthatwepreviouslyreportedtheresultsofqPCRforADAR2 in a single celllevelin Fig.2 ofHideyama etal.2010 (J Neurosci30:11917-25),whatwepresented there was the ordinary PCR for ADAR2 on pooled mouse motor neuron lysates but not the qPCR data.Wepresented qPCR forADAR2 on pooled human motor neuron lysates in Fig.2 of Hideyama etal.,2012 (NeurobiolDis45:1121-28),buttheexpression levelofADAR2 mRNA wasmorethan one-orderhigherin human motorneuronsthan in mousemotorneurons.Anotherthing one should take into account was that the majority ofADAR2-lacking motor neurons had been already disappeared and themajority ofremaining motorneuronswerethoseexpressing ADAR2 in the controlAR2 mice atthe age ofexamination.Therefore,itwould be technically difficultto differentiate10-15% differenceofADAR2 mRNA expression leveleven in thepooled motor neuronsbetween thesalinecontrolAR2 miceand wild-type mice as stated above. We cannot tell the differencein theexpression levelofADAR2 mRNA by meansofordinary PCR.Theremay besome misunderstanding. Wepresentedthesignificantreductioninthenumberofmotorneuronsandaxonsintheventral rootof saline controlAR2 mice compared to wild-type mice in Fig. 3, which was consistent with the results in the originalreport(Hideyama et al., 2010 J Neurosci) indicating that lack of ADAR2 induced death in the motor neurons in the AR2 mice. 2) TofillthedefectofquantitativedataonADAR2expression,wepresentedtheGluA2Q/R site- editing in a single motorneuron levelin Fig.4C.BecauseGluA2Q/R site-editing isspecifically catalysed by ADAR2,an increasein theediting efficiency in theAAV group compared to thesaline controlAR2 mice indicate thatexogenousADAR2 issuccessfully delivered to and expressed in the motorneurons.Thismethod forADAR2 activity in asinglecellwasthesameasweadopted in Fig. 2A ofHideyamaetal,2010,JNeurosci.Wealso presented thattheAAV vectorwasdelivered to motorneuronsbydemonstratingFlagexpressioninChAT-positivelarge neuronsin the anterior horn ofthespinalcord in Fig.1D and moreADAR2-positiveAHCswereremained in AAV-treated AR2micethaninthesalinecontrolAR2miceinSupportingInformationFig.S3D.Webelievethat these data sufficiently indicate that exogenously delivered ADAR2 wassuccessfully expressed in the motor neurons by single systemic injection of the AAV vector. Tofacilitate readers’ understanding,we amended the sentences in the Results and Discussion sections as follows,and added relevantfigure numbers in the Discussion section; “Becauserescueofdeath ofAHCslikely resultsfrom restoration ofADAR2 activity in themotor neuronsofAR2 mice(Hideyamaetal,2010),we nextinvestigated whethertheexpressionand activity ofADAR2 were increased in the motorneuronsaftersystemic injection ofAAV9- hADAR2.”(Page7 line12) “MessengerRNAofbothhADAR2andcholineacetyltransferase(ChAT)wasdemonstratedinthe spinalcord lysates ofAAV9-injected AR2 mice (Fig. 4B) and the editing efficiency atthe GluA2 Q/Rsitewassignificantly higherin the remaining motorneuronsofAAV9-injected AR2 mice than in those of the control AR2 mice(Fig 4C).These results indicated thathADAR2 was delivered to and functioned in motorneurons.However,ADAR2 protein leveldid notsignificantly differamong AAV9-injected AR2 mice, saline-injected AR2 mice and wild-type mice (SupportingInformation FigS8B,C). ThefailuretodetectthedifferencedespiteofthedifferenceintheADAR2activitywasEMBO MolecularMedicine PeerReview ProcessFile- EMM-2013-02935 © EMBO 16 presumably dueto thefactthatamodestincreasein theADAR2 expression levelwith preservation of death of 10-20% ofmotorneuronsin AAV-treated AR2 mice may be masked in ADAR2 expressed in the remaining motorand non-motorneuronsandothercellsintheanteriorhorn.” (Page7,line19) “Furthermore,theexpression ofthedelivered ADAR2 preventedthe progression of motor dysfunction and neuronaldeath with restoring ADAR2-mediatedRNAediting(Figs 2,3,4 and SupportingInformationFigS7)withoutinducinganyadverseeffectsinneuronsorsurrounding tissues (Fig1andSupportingInformationFigs S4,S5)” (Page 9,line 8). 2.ThewayofTDP-43 positiveAHCsmeasurementwasstillsubjectiveand notacceptable.They should use objective criteria for it,such as certain cut-offsignalintensity. → (Response)Wethankthereviewerforthecomment. Accordingtoreviewer’s comment,we added Supporting information Fig.S10.Asforthe TDP-43 immunohistochemistry,weadopted the criteria thatthe positivity isabove 3-fold of background.Forreviewer1’s convenience,the results ofquantification areattached and themethod used isindicated in theMaterialsand Methodssection. Theresultsdidnotchangeafteradoptingthecriteria. WehaveaddedthesentencesintheMaterialsandMethodssectionas“ThesignalintensityofTDP- 43 wasexamined using ImageJsoftware.TDP-43-positive AHCswerecountedwhenthesignal intensity wasmorethan3-fold higherthan the background”. (Page19 line6 from thebottom). ToReviewer#2: Referee#2(CommentsonNovelty/ModelSystem): Themodelsystem isadequateandwasdeveloped by the authors. Referee#2(Remarks): Theauthorshaveaddressedmyconcerns.Themedicalquestionposedhereishighlysignificant although ithasbeen pursued almostexclusivelybytheauthors.Thedata supporttheirconclusions although there is clearly substantial scope for further investigation of this model and the underlying medicalproblem. → (Response)Wethankthereviewer’sdeepunderstandingandvaluablecomments. Wearelookingforwardtohearingfavourable decision on thisrevised manuscript.