《细胞生物学》课程教学资源（文献资料）Oct4-Induced Pluripotency in Adult Neural Stem Cells

CellOct4-lnduced Pluripotencyin Adult Neural Stem CellsJeongBeomKim,1VittorioSebastiano,1GuangmingWu,1MarcosJ.Araizo-Bravo,1PhilippSasse,2LucaGentile,1Kinarm Ko,1David Ruau,3Mathias Ehrich,Dirk van den Boom,4Johann Meyer,5Karin Hbner,Christof Bernemann,ClaudiaOrtmeier,1MartinZenke,3BerndK.Fleischmann,2HolmZaehres,1andHansR.Scholer1,1Department of Cell and Developmental Biology,Max Planck Institutefor Molecular Biomedicine,Rontgenstrasse 20, 48149 Munster,NRW, Germany2Institute of Physiology I, Life&Brain Center, University of Bonn,53105 Bonn, NRW, Germany3institutefor Biomedical Engineering,Department of CellBiology,RWTH Aachen University Medical School, Pauwelsstrasse352074Aachen,NRW,Germany4SEQUENOMInc.,3595JohnHopkinsCourt,SanDiego,CA92121,USA5Hannover Medical School,DepartmentofExperimental Hematology,Carl-Neuberg-Strasse 1,30625 Hannover,GermanyCorrespondence:office@mpi-muenster.mpg.deDOI10.1016/j.cell.2009.01.023SUMMARYand GC stem cells exhibit intriguing similarities. Several keyfactors required forpluripotency arealso expressed in primor-ThefourtranscriptionfactorsOct4,Sox2,Kif4,anddial germ and spermatogonial stem cells.However, when itcomes to comparing them to the repertoire of factors in thec-Myccan inducepluripotencyin.mouseandhumansoma, very few actually stand out. Based on its expressionfibroblasts.We previously described direct reprog-profile, the transcription factor Oct4 was from the beginningramming of adult mouse neural stem cells (NsCs)considered to be a key regulator during mouse embryogenesis.by Oct4 and either Kif4 or c-Myc.NSCs endoge-Oct4 is expressed in the pluripotent cells of an embryo and innously express Sox2,c-Myc,andKif4as wellascell lines derived thereof, as well as in the GC lineage.Embry-several intermediate reprogramming markers.Hereonic stemcells (ESCs)derived from preimplantation embryoswe reportthatexogenous expression ofthe germ-can easily integrate into the germline. This has previouslyline-specifictranscriptionfactorOct4is sufficienttobeenshownformousepreimplantationembryos(Boianiandgenerate pluripotent stem cells from adult mouseScholer,2005).In contrast,Oct4 isdownregulated in thethreeNsCs.These one-factor induced pluripotent stem somatic lineages, but only during gametogenesis around thecells (1F iPS)are similar to embryonic stem cellsinitiation of male and female meiosis (Pesce et al., 1998).Oct4 is re-expressed in unfertilized oocytes after birth andin vitro and in vivo.Not onlycanthese cells can beOct4 protein can be detected until the final stages of oocyteefficientlydifferentiatedintoNSCs,cardiomyocytes,maturationand germ cells in vitro,but they are also capable ofLoss of Oct4 function early in development causes cells ofteratoma formationand germline transmissionthe preimplantation embryo to acquire a trophectodermal fateinvivo.OurresultsdemonstratethatOct4isrequired(Nicholsetal.,1998),whereaslossintheGClineageleadstoandsufficienttodirectlyreprogramNsCstopluripo-apoptosis of primordial germ cells (PGCs) (Kehler et al.,tency.2004).These studies of the developing embryo were comple-mented by studies of ESCs in which the Oct4 function wasINTRODUCTIONabolished (Niwa et al., 2000). ESCs are especially interestingin this context, as they are derived from the inner cell massFor centuries, embryologists have been intrigued by theof theblastocystand thus represent a developmentalstagedistinction between germ cells (GCs) and soma (McLaren,following the establishment of thepluripotentfounder popula-1981; Pesce et al.,1998; Weismann, 1892).Any attempt totion. While loss of Oct4 function results in differentiation intoestablishsuch adistinctioneventuallyneedsto addressthetrophectodermal cells,overexpression ofOct4 in ESC leadsquestion, which genetic program specifically defines the plurip-to differentiation along the mesodermal and primitive endo-otent founder population that gives rise to all cell lineages ofdermal lineages.the body as well as the GC lineage and ensures the flow ofThe early developmental phenotype suggests that Oct4 ispart of a regulatory system that maintains pluripotency.geneticinformationfromonegenerationtothenext.Onewayof establishing a distinction between germline and soma is toA precise level of Oct4 is required to maintain pluripotency indetermine the regulatoryfactors expressed in both the pluripo-ES cells, which is in agreement with this hypothesis. The factthat stem cells of the GC lineage are unipotent despiteexpress-tent cells and stem cells of the GC lineage but not in somaticcells.Interestingly,thegeneticprograms ofpluripotent cellsingOct4,albeitatlowerlevels,showsthatitsmerepresenceCell 136,411-419, February 6,2009@2009 Elsevier Inc.411

Oct4-Induced Pluripotency in Adult Neural Stem Cells Jeong Beom Kim,1 Vittorio Sebastiano,1 Guangming Wu,1 Marcos J. Arau´ zo-Bravo,1 Philipp Sasse,2 Luca Gentile,1 Kinarm Ko,1 David Ruau,3 Mathias Ehrich,4 Dirk van den Boom,4 Johann Meyer,5 Karin Hu¨ bner,1 Christof Bernemann,1 Claudia Ortmeier,1 Martin Zenke,3 Bernd K. Fleischmann,2 Holm Zaehres,1 and Hans R. Scho¨ ler1, * 1Department of Cell and Developmental Biology, Max Planck Institute for Molecular Biomedicine, Ro¨ ntgenstrasse 20, 48149 Mu¨ nster, NRW, Germany 2Institute of Physiology I, Life & Brain Center, University of Bonn, 53105 Bonn, NRW, Germany 3Institute for Biomedical Engineering, Department of Cell Biology, RWTH Aachen University Medical School, Pauwelsstrasse 30, 52074 Aachen, NRW, Germany 4SEQUENOM Inc., 3595 John Hopkins Court, San Diego, CA 92121, USA 5Hannover Medical School, Department of Experimental Hematology, Carl-Neuberg-Strasse 1, 30625 Hannover, Germany *Correspondence: office@mpi-muenster.mpg.de DOI 10.1016/j.cell.2009.01.023 SUMMARY The four transcription factors Oct4, Sox2, Klf4, and c-Myc can induce pluripotency in mouse and human fibroblasts. We previously described direct reprog￾ramming of adult mouse neural stem cells (NSCs) by Oct4 and either Klf4 or c-Myc. NSCs endoge￾nously express Sox2, c-Myc, and Klf4 as well as several intermediate reprogramming markers. Here we report that exogenous expression of the germ￾line-specific transcription factor Oct4 is sufficient to generate pluripotent stem cells from adult mouse NSCs. These one-factor induced pluripotent stem cells (1F iPS) are similar to embryonic stem cells in vitro and in vivo. Not only can these cells can be efficiently differentiated into NSCs, cardiomyocytes, and germ cells in vitro, but they are also capable of teratoma formation and germline transmission in vivo. Our results demonstrate that Oct4 is required and sufficient to directly reprogram NSCs to pluripo￾tency. INTRODUCTION For centuries, embryologists have been intrigued by the distinction between germ cells (GCs) and soma (McLaren, 1981; Pesce et al., 1998; Weismann, 1892). Any attempt to establish such a distinction eventually needs to address the question, which genetic program specifically defines the plurip￾otent founder population that gives rise to all cell lineages of the body as well as the GC lineage and ensures the flow of genetic information from one generation to the next. One way of establishing a distinction between germline and soma is to determine the regulatory factors expressed in both the pluripo￾tent cells and stem cells of the GC lineage but not in somatic cells. Interestingly, the genetic programs of pluripotent cells and GC stem cells exhibit intriguing similarities. Several key factors required for pluripotency are also expressed in primor￾dial germ and spermatogonial stem cells. However, when it comes to comparing them to the repertoire of factors in the soma, very few actually stand out. Based on its expression profile, the transcription factor Oct4 was from the beginning considered to be a key regulator during mouse embryogenesis. Oct4 is expressed in the pluripotent cells of an embryo and in cell lines derived thereof, as well as in the GC lineage. Embry￾onic stem cells (ESCs) derived from preimplantation embryos can easily integrate into the germline. This has previously been shown for mouse preimplantation embryos (Boiani and Scho¨ ler, 2005). In contrast, Oct4 is downregulated in the three somatic lineages, but only during gametogenesis around the initiation of male and female meiosis (Pesce et al., 1998). Oct4 is re-expressed in unfertilized oocytes after birth and Oct4 protein can be detected until the final stages of oocyte maturation. Loss of Oct4 function early in development causes cells of the preimplantation embryo to acquire a trophectodermal fate (Nichols et al., 1998), whereas loss in the GC lineage leads to apoptosis of primordial germ cells (PGCs) (Kehler et al., 2004). These studies of the developing embryo were comple￾mented by studies of ESCs in which the Oct4 function was abolished (Niwa et al., 2000). ESCs are especially interesting in this context, as they are derived from the inner cell mass of the blastocyst and thus represent a developmental stage following the establishment of the pluripotent founder popula￾tion. While loss of Oct4 function results in differentiation into trophectodermal cells, overexpression of Oct4 in ESC leads to differentiation along the mesodermal and primitive endo￾dermal lineages. The early developmental phenotype suggests that Oct4 is part of a regulatory system that maintains pluripotency. A precise level of Oct4 is required to maintain pluripotency in ES cells, which is in agreement with this hypothesis. The fact that stem cells of the GC lineage are unipotent despite express￾ing Oct4, albeit at lower levels, shows that its mere presence Cell 136, 411–419, February 6, 2009 ª2009 Elsevier Inc. 411

CellGFP+ colonies on day 7 (Kim et al., 2008). In contrast todoes not result in pluripotency. However, we hypothesize thatthe presence of Oct4 augments the induction of pluripotency.NSCs, GFP+cells are first observed in MEFs on day7 postAs shown for PGCs, Oct4-expressing cells can give rise toinfection of 4F (Okita et al., 2007). We therefore consideredpluripotent cells (termed embryonic germ cells)when culturedthe possibility that NSCs might represent a more advancedin a simple cocktail of three growth factors (Matsui et al.,stage in the reprogramming process and that terminally1992;Resnicketal.,1992)differentiated cells have to go through a stage like this in orderRecently,Yamanaka and colleagues have shown that theto subsequently acquire a pluripotent state. In this scenarioexpression of AP and SSEA-1 in NSCs marks a quasi-genetic program of somatic cells can be induced to acquireintermediate state, during which NSCs are reprogrammedpluripotency by overexpression of specific transcription factors.Initially,generationof pluripotent cells was achievedbyectopicearlierandmore efficiently thanMEFs.expressionoffourfactors(Oct4,Sox2,Klf4,andc-Myc)frommouseandhuman somaticcells(Lowry etal.,2008;MaheraliipS Cells Generated fromAdultNeural Stem Cells withet al.,2007:Meissner et al..2007:Okita etal.,2007.2008:Transduction of Oct4attheMolecularLevel AreSimilarParketal..2008:Stadtfeldetal.,2008:Takahashietal..2007:to Embryonic Stem CellsTakahashi and Yamanaka, 2006; Wernig et al.,2007; Yu et al.Inourpreviousstudy,NSCscouldbe reprogrammedwith three2007). These ESC-like induced pluripotent stem (iPS)cellsfactor (3F)and two-factor (2F)combinations (Kim et al., 2008)havecharacteristicssimilartoESCsMoreover, in the 2F combinations c-Myc and KIf4 could beThemolecularmechanismsunderlyingreprogrammingarestillreplaced with one another, although c-Myc was less efficientthan Kif4 in inducing the formation of GFP+ colonies (3-4 weeks,far from being completely understood, which is largely due toseveralvariables(numberofexogenousfactorsrequired,hetero-instead of2-3weeks).Next,wesoughttodefineconditionsinwhich NSCs can be reprogrammed bytransduction of Oct4geneityofreprogrammingtargetcells)withinthesystem.Weattempted to simplify this scenariobychoosing NSCs as a cellalone.Astimehasbeenanissuewhenusing variousfactormodel for reprogramming with different combinations of tran-combinations,weprolongedtheculturingofOct4-infectedcellsscription factors (Kim et al., 2008). The combination of theIn the current study,we succeeded in generating three one-specificfactorsneededtoobtainiPScellsvaries,thoughOct4factor(1F)iPSclones(clones2.3,and4)fromfiveGFP+coloniescan apparentlynotbe replacedby other factors.Recently,weof Oct4-infected NSCs within 4-5 weeks in culture (Figure 1A).demonstrated that Oct4 and Kif4 are sufficient to induceThe iPS cells expressed Oct4-GFP, SSEA-1,and AP and werepluripotency in NSCs (Kim et al., 2008). By omitting Kif4, wemorphologically indistinguishable from mouse ESCs underhaveestablishedconditionsthatdemonstratethatOct4isnotmouse ESC culture conditions(Figure1B).Theestimatedonly essential but also sufficient to induce pluripotency inreprogramming efficiency for the one-factor approach can beNSCs.Our study shows for thefirst time that just one transcrip-calculatedas 0.014%,which is 10-fold lowerthan our two-factortionfactorcanconvertsomaticcellsintopluripotentcells,whichapproach(Kim etal.,2008)and similartoreprogrammingMEFscan differentiate into derivatives of all three germ layers and intowithfourfactors (Okita etal.,2007).functional GCs.We next characterized the1F iPS at the molecular level.Allthree clones expressed typical ESC marker genes (Figure 2A)RESULTSandtheendogenousgenesOct4,Sox2,c-Myc,andKif4,similartomouseESCs(Figure2B).TheOct4transgenewascompletelyNeural StemCells EndogenouslyExpress APsilencedafterpassage 5(FigureS2).Todeterminethe numberofandSSEA-1viral Oct4transgene integrations,weperformed Southern blotIn our previous study we utilized Oct4-GFP expression asanalysis on ESCs,1F,2F,and 4FiPS cells.The1FiPS clones dis-a reprogramming marker,and we observed that the timeplayed 2 and 5 integrations (Figure S3A), while 2F or 4F iPS celldisplayed 7 integrations (Figure S3B).We performed Southernrequiredfor reprogramming was affectedbyboth the number(4, 3, or 2) and the composition of the pluripotency-inducingblotanalysisfromthreeindependent subclones of clone2factors(Oct4,Sox2,Kif4,andc-Myc)(Kim et al.,2008).Other(Figure S3B). The same integration pattem confirmed that nostudies examined the reprogramming timing in more detail bycrosscontamination occurred during the expansion of thesesubclones. Thus, 1F iPS could be generated with less viral inte-reprogrammingmousefibroblastsusingthedoxycycline (dox)-inducible system (Brambrink et al., 2008; Wernig et al., 2007).grations of theOct4transgenethan2For4F iPS cells.WealscThese studies demonstrated the sequential expression ofconfirmed integration of the Oct4 transgene by polymerasechain reaction (PCR):all 1F iPS clones contained only the Oct4markergenes,suchas alkalinephosphatase(AP),stage-specific embryonic antigen-1 (SSEA-1), and Oct4 or Nanog,transgene (Figure S4)during the reprogramming process.Thus, expression ofWe next performed DNA methylation analysis to assess theSSEA-1 represents an intermediate stage,which precedesepigenetic status of the Oct4,Sox2,and Nanog promoters withexpression ofOct4orNanogand which in turn is requiredbeforeDNAisolatedfrom ESCs,NSCs,2FiPScells,and1FiPSclonesacellcanbefullyreprogrammedtoapluripotentstate.Interest-(Figure2C).Sox2promotermaintainsthedemethylatedstatusiningly,NSCs endogenouslyexpress SSEA-1and also showAPallpopulations(ESCs:10%,2FiPS:3%,1FiPS:5%,NSCs:3%activity (Figure S1available online)(Capela and Temple,2002;NSC1F:3%).Oct4andNanogpromoters were highly demethy-Peh et al., 2008). Following infection of NSCs with the fourlated inESCs (13%,9%)and2F iPScells (13%,8%)andretainedfactors (4F), the cells give rise to GFp+ cells on day 4 andpartial methylation in 1F iPS (31%, 27%). 1F iPS-derived NSCs412Cell 136,411419, February6,20092009 Elsevier Inc

does not result in pluripotency. However, we hypothesize that the presence of Oct4 augments the induction of pluripotency. As shown for PGCs, Oct4-expressing cells can give rise to pluripotent cells (termed embryonic germ cells) when cultured in a simple cocktail of three growth factors (Matsui et al., 1992; Resnick et al., 1992). Recently, Yamanaka and colleagues have shown that the genetic program of somatic cells can be induced to acquire pluripotency by overexpression of specific transcription factors. Initially, generation of pluripotent cells was achieved by ectopic expression of four factors (Oct4, Sox2, Klf4, and c-Myc) from mouse and human somatic cells (Lowry et al., 2008; Maherali et al., 2007; Meissner et al., 2007; Okita et al., 2007, 2008; Park et al., 2008; Stadtfeld et al., 2008; Takahashi et al., 2007; Takahashi and Yamanaka, 2006; Wernig et al., 2007; Yu et al., 2007). These ESC-like induced pluripotent stem (iPS) cells have characteristics similar to ESCs. The molecular mechanisms underlying reprogramming are still far from being completely understood, which is largely due to several variables (number of exogenous factors required, hetero￾geneity of reprogramming target cells) within the system. We attempted to simplify this scenario by choosing NSCs as a cell model for reprogramming with different combinations of tran￾scription factors (Kim et al., 2008). The combination of the specific factors needed to obtain iPS cells varies, though Oct4 can apparently not be replaced by other factors. Recently, we demonstrated that Oct4 and Klf4 are sufficient to induce pluripotency in NSCs (Kim et al., 2008). By omitting Klf4, we have established conditions that demonstrate that Oct4 is not only essential but also sufficient to induce pluripotency in NSCs. Our study shows for the first time that just one transcrip￾tion factor can convert somatic cells into pluripotent cells, which can differentiate into derivatives of all three germ layers and into functional GCs. RESULTS Neural Stem Cells Endogenously Express AP and SSEA-1 In our previous study we utilized Oct4-GFP expression as a reprogramming marker, and we observed that the time required for reprogramming was affected by both the number (4, 3, or 2) and the composition of the pluripotency-inducing factors (Oct4, Sox2, Klf4, and c-Myc) (Kim et al., 2008). Other studies examined the reprogramming timing in more detail by reprogramming mouse fibroblasts using the doxycycline (dox)- inducible system (Brambrink et al., 2008; Wernig et al., 2007). These studies demonstrated the sequential expression of marker genes, such as alkaline phosphatase (AP), stage￾specific embryonic antigen-1 (SSEA-1), and Oct4 or Nanog, during the reprogramming process. Thus, expression of SSEA-1 represents an intermediate stage, which precedes expression of Oct4 or Nanog and which in turn is required before a cell can be fully reprogrammed to a pluripotent state. Interest￾ingly, NSCs endogenously express SSEA-1 and also show AP activity (Figure S1 available online) (Capela and Temple, 2002; Peh et al., 2008). Following infection of NSCs with the four factors (4F), the cells give rise to GFP+ cells on day 4 and GFP+ colonies on day 7 (Kim et al., 2008). In contrast to NSCs, GFP+ cells are first observed in MEFs on day 7 post￾infection of 4F (Okita et al., 2007). We therefore considered the possibility that NSCs might represent a more advanced stage in the reprogramming process and that terminally differentiated cells have to go through a stage like this in order to subsequently acquire a pluripotent state. In this scenario, expression of AP and SSEA-1 in NSCs marks a quasi￾intermediate state, during which NSCs are reprogrammed earlier and more efficiently than MEFs. iPS Cells Generated from Adult Neural Stem Cells with Transduction of Oct4 at the Molecular Level Are Similar to Embryonic Stem Cells In our previous study, NSCs could be reprogrammed with three￾factor (3F) and two-factor (2F) combinations (Kim et al., 2008). Moreover, in the 2F combinations c-Myc and Klf4 could be replaced with one another, although c-Myc was less efficient than Klf4 in inducing the formation of GFP+ colonies (3–4 weeks, instead of 2–3 weeks). Next, we sought to define conditions in which NSCs can be reprogrammed by transduction of Oct4 alone. As time has been an issue when using various factor combinations, we prolonged the culturing of Oct4-infected cells. In the current study, we succeeded in generating three one￾factor (1F) iPS clones (clones 2, 3, and 4) from five GFP+ colonies of Oct4-infected NSCs within 4–5 weeks in culture (Figure 1A). The iPS cells expressed Oct4-GFP, SSEA-1, and AP and were morphologically indistinguishable from mouse ESCs under mouse ESC culture conditions (Figure 1B). The estimated reprogramming efficiency for the one-factor approach can be calculated as 0.014%, which is 10-fold lower than our two-factor approach (Kim et al., 2008) and similar to reprogramming MEFs with four factors (Okita et al., 2007). We next characterized the 1F iPS at the molecular level. All three clones expressed typical ESC marker genes (Figure 2A) and the endogenous genes Oct4, Sox2, c-Myc, and Klf4, similar to mouse ESCs (Figure 2B). The Oct4 transgene was completely silenced after passage 5 (Figure S2). To determine the number of viral Oct4 transgene integrations, we performed Southern blot analysis on ESCs, 1F, 2F, and 4F iPS cells. The 1F iPS clones dis￾played 2 and 5 integrations (Figure S3A), while 2F or 4F iPS cell displayed 7 integrations (Figure S3B). We performed Southern blot analysis from three independent subclones of clone 2 (Figure S3B). The same integration pattern confirmed that no crosscontamination occurred during the expansion of these subclones. Thus, 1F iPS could be generated with less viral inte￾grations of the Oct4 transgene than 2F or 4F iPS cells. We also confirmed integration of the Oct4 transgene by polymerase chain reaction (PCR): all 1F iPS clones contained only the Oct4 transgene (Figure S4). We next performed DNA methylation analysis to assess the epigenetic status of the Oct4, Sox2, and Nanog promoters with DNA isolated from ESCs, NSCs, 2F iPS cells, and 1F iPS clones (Figure 2C). Sox2 promoter maintains the demethylated status in all populations (ESCs: 10%, 2F iPS: 3%, 1F iPS: 5%, NSCs: 3%, NSC 1F: 3%). Oct4 and Nanog promoters were highly demethy￾lated in ESCs (13%, 9%) and 2F iPS cells (13%, 8%) and retained partial methylation in 1F iPS (31%, 27%). 1F iPS-derived NSCs 412 Cell 136, 411–419, February 6, 2009 ª2009 Elsevier Inc

CellANext,we attempted todetermine whether 1F iPS could differ-Passage 1Passage 0entiateintolineage-committedpopulationsofNSCs,cardiomyo-cytes, or PGCs. We derived NSCs from 1F iPS in serum-freeadherentmonolayerculturesaccordingtoapublishedprotocol(Ying et al.,2003).Interestingly,these NSC 1F were very similarto the donor NSCs. NSC 1F can maintain their self-renewalcapacity,even after repeated passaging. RT-PCR analysisrevealed NSC-specific gene expression similar to donor NSCsbut no expression of pluripotency markers such as Oct4 orNanog(Figure3A).NSC 1F appeared highly homogenous,andimmunocytochemicalstainingconfirmedtheuniformexpressionBOct4-GFPoftheNSCmarkerNestin (Figure3B).Moreover, theseNSC 1FPhasewere capable of differentiating into neurons, oligodendrocytes,and astrocytes after prolonged time in culture (passage 20)(Figures 3B and S6).Themodal chromosome number of 40was maintained throughout reprogramming and derivation ofNSCs (Figure3C).Our resultsdemonstrate that 1F iPS canbedifferentiated efficiently into multipotent NSCs that have self-renewing capacity with global gene expression profiles similarto donorNSCs (Figure 3D).APDuring EB differentiation,weobserved PECAM-positive cellsSSEA-1(amarkerofmatureendothelialcells)withvessel-likestructuresand beating a-actinin-positive cardiomyocytes (Figures 4Aand 4B). Intracellular recordings from the beating areas demon-strated both atrial- (Figure 4C, upper) and ventricular-like(Figure 4C, lower) cardiac action potentials. Chronotropicregulation was found to be intact (Figure 4D), indicating thepresence offunctional cardiomyocytes.PreviousreportshavedemonstratedtheinvitrogenerationofGCs frommurine ESCs (Geijsen et al.,2004; Hubner et al.,2003)Figure1.GenerationofOne-FactoriPSCellsfromAdultNSCsbyTo further verify the pluripotency of 1F iPS, we attempted toRetroviralTransductionofOct4differentiate 1F iPS into GCs in vitro. Putative GC cultures(A) Morphology and Oct4 promoter-driven GFP (Oct4-GFP) expression (inset)showedmeioticcompetenceasdemonstratedbyexpressionin reprogrammed one-factor (1F) iPS without feeders (passage 0) on day 35of Sycp3protein (Figures 5A and 5B) and were FACS sortedpost-infection and 1F iPS (clone 2) grown on feeders (passage 1). Scalefor GFP-positive (GFP*) or GFP-negative (GFP-) cells represen-bars,100μm(B) Phase contrast image showing ESC-like morphology (upper left) of 1F iPStative of earlyand late postmigratory stages, respectively.Cell(clone 2) on feeders. Colonies express Oct4-GFP (green, upper right) andfractions were analyzed using quantitative real-time PCR.Thestained positive for SSEA-1 (red, lower left) and for AP (lower right). Abbrevia-GC characterof in vitrodifferentiated1F iPSwasshown by thetions: SSEA-1, stage-specific embryonic antigen-1; AP, alkaline phosphatase.upregulation of the specific markers Blimp1,Stella,Fragilis,Scale bars, 200 μm.and Oct4 (Figure 5C). Importantly, in vitro differentiated cellsshowed expression of Gdf9, an oocyte-specific marker,(NSC 1F) were highly methylated at those loci (68%, 85%)suggesting that 1F iPS, although male, could differentiate intosimilar to donor NSCs (70%, 80%). The methylation dataoogonia-like cells as has been shown for EsCs (Hubner et al.,suggest slight differencesbetween1F iPS and ESCs.Scatter2003).These results demonstrate that 1F iPS can developplots of the global gene expression profiles obtained fromeven in vitro into GCs.The capacity of 1F iPS to differentiatecDNAmicroarraysdemonstratedthat1FiPS exhibitadistributioninto lineage-committed cell populations of different germ layerspattem of gene expression comparable to ESCs, hencevalidatestheirpluripotentstatus.completelydifferentfromNSCs(Figures2Dand2E)1F iPs Cells Have aninVivoDevelopmental Potential1FiPSCan BeDifferentiated intoThreeGerm LayersSimilartoEmbryonic StemCellsincludingNsCs,Cardiomyocytes,and GCs InVitroTo investigate the in vivo developmental potential of 1F iPS,To investigatethedevelopmental potential of 1F iPS in vitro, wewe performed teratoma formation and chimeras-contributiondetermined whether 1F iPS could differentiate into three germassays.First, wedetermined that1FiPSwere capable offorminglayers by embryonic body (EB)differentiation.EBs derivedteratomasuponsubcutaneousinoculationintonudemice.Thefrom all three 1F iPS clones expressed markers of the threeteratomas contained tissues of all three germ layers includinggerm layers including GATA-4 (endoderm), Brachyury (meso-neural rosette (ectoderm),cuboidal epithelium (endoderm),andderm), and MAP2 (ectoderm) as determined by RT-PCRmuscle (mesoderm) (Figures 6A and S5B). These data reveal(Figure S5A).that 1F iPS possess multilineage potential in vivo.Cell 136,411-419, February 6,2009@2009Elsevier Inc.413

(NSC 1F) were highly methylated at those loci (68%, 85%), similar to donor NSCs (70%, 80%). The methylation data suggest slight differences between 1F iPS and ESCs. Scatter plots of the global gene expression profiles obtained from cDNA microarrays demonstrated that 1F iPS exhibit a distribution pattern of gene expression comparable to ESCs, hence completely different from NSCs (Figures 2D and 2E). 1F iPS Can Be Differentiated into Three Germ Layers including NSCs, Cardiomyocytes, and GCs In Vitro To investigate the developmental potential of 1F iPS in vitro, we determined whether 1F iPS could differentiate into three germ layers by embryonic body (EB) differentiation. EBs derived from all three 1F iPS clones expressed markers of the three germ layers including GATA-4 (endoderm), Brachyury (meso￾derm), and MAP2 (ectoderm) as determined by RT-PCR (Figure S5A). Next, we attempted to determine whether 1F iPS could differ￾entiate into lineage-committed populations of NSCs, cardiomyo￾cytes, or PGCs. We derived NSCs from 1F iPS in serum-free adherent monolayer cultures according to a published protocol (Ying et al., 2003). Interestingly, these NSC 1F were very similar to the donor NSCs. NSC 1F can maintain their self-renewal capacity, even after repeated passaging. RT-PCR analysis revealed NSC-specific gene expression similar to donor NSCs but no expression of pluripotency markers such as Oct4 or Nanog (Figure 3A). NSC 1F appeared highly homogenous, and immunocytochemical staining confirmed the uniform expression of the NSC marker Nestin (Figure 3B). Moreover, these NSC 1F were capable of differentiating into neurons, oligodendrocytes, and astrocytes after prolonged time in culture (passage 20) (Figures 3B and S6). The modal chromosome number of 40 was maintained throughout reprogramming and derivation of NSCs (Figure 3C). Our results demonstrate that 1F iPS can be differentiated efficiently into multipotent NSCs that have self￾renewing capacity with global gene expression profiles similar to donor NSCs (Figure 3D). During EB differentiation, we observed PECAM-positive cells (a marker of mature endothelial cells) with vessel-like structures and beating a-actinin-positive cardiomyocytes (Figures 4A and 4B). Intracellular recordings from the beating areas demon￾strated both atrial- (Figure 4C, upper) and ventricular-like (Figure 4C, lower) cardiac action potentials. Chronotropic regulation was found to be intact (Figure 4D), indicating the presence of functional cardiomyocytes. Previous reports have demonstrated the in vitro generation of GCs from murine ESCs (Geijsen et al., 2004; Hu¨ bner et al., 2003). To further verify the pluripotency of 1F iPS, we attempted to differentiate 1F iPS into GCs in vitro. Putative GC cultures showed meiotic competence as demonstrated by expression of Sycp3 protein (Figures 5A and 5B) and were FACS sorted for GFP-positive (GFP+ ) or GFP-negative (GFP) cells represen￾tative of early and late postmigratory stages, respectively. Cell fractions were analyzed using quantitative real-time PCR. The GC character of in vitro differentiated 1F iPS was shown by the upregulation of the specific markers Blimp1, Stella, Fragilis, and Oct4 (Figure 5C). Importantly, in vitro differentiated cells showed expression of Gdf9, an oocyte-specific marker, suggesting that 1F iPS, although male, could differentiate into oogonia-like cells as has been shown for ESCs (Hu¨ bner et al., 2003). These results demonstrate that 1F iPS can develop even in vitro into GCs. The capacity of 1F iPS to differentiate into lineage-committed cell populations of different germ layers validates their pluripotent status. 1F iPS Cells Have an In Vivo Developmental Potential Similar to Embryonic Stem Cells To investigate the in vivo developmental potential of 1F iPS, we performed teratoma formation and chimeras-contribution assays. First, we determined that 1F iPS were capable of forming teratomas upon subcutaneous inoculation into nude mice. The teratomas contained tissues of all three germ layers including neural rosette (ectoderm), cuboidal epithelium (endoderm), and muscle (mesoderm) (Figures 6A and S5B). These data reveal that 1F iPS possess multilineage potential in vivo. Figure 1. Generation of One-Factor iPS Cells from Adult NSCs by Retroviral Transduction of Oct4 (A) Morphology and Oct4 promoter-driven GFP (Oct4-GFP) expression (inset) in reprogrammed one-factor (1F) iPS without feeders (passage 0) on day 35 post-infection and 1F iPS (clone 2) grown on feeders (passage 1). Scale bars, 100 mm. (B) Phase contrast image showing ESC-like morphology (upper left) of 1F iPS (clone 2) on feeders. Colonies express Oct4-GFP (green, upper right) and stained positive for SSEA-1 (red, lower left) and for AP (lower right). Abbrevia￾tions: SSEA-1, stage-specific embryonic antigen-1; AP, alkaline phosphatase. Scale bars, 200 mm. Cell 136, 411–419, February 6, 2009 ª2009 Elsevier Inc. 413

CellAFigure2.Characterizationof 1F iPS01FiPS(A) RT-PCR analysis of pluripotency markerC2C3 C4expression in 1F iPS clones (C2,C3,and C4) andOct4ESCs (positive control), as well as NSCs (negativeSox2KIf4Oct4NanogNanogcontrol). Primers are specific for transcripts fromSox2C-Mycthe respective endogenous locus. β-actin wasRextused as loading control.Utftoss(B) Quantitative PCR analysis of endogenousFgf4Dexpressionof thefourfactors in1FiPS clones(C2, C3, and C4). RNA levels were determined byCripto8.0.01quantitative real-time PCR using primers specificEras2forendogenous transcripts.Endogenous relativeEsg1Yexpression levels of 1F iPS on day 3 (whole popu-β-actin0.001lation) and passage 5 (clones 2, 3, and 4) werellwtEsc鄂1Fd3IIC2C3#C4HNSCcompared with those in ESCs. Transcript levelscwere normalized to β-actin levels. Shown are theOct 4Sox 2Nanogaverages with standard deviations of three inde-NSC (donor)pendentexperimentsNSC1F(C) Analysis of the DNA methylation pattern of theOct4,Sox2,andNanogpromotersinNSCsand1F iPS (C4)NSC 1F, as well as in 1F iPS (clones 4 and 2), 2F1F iPS (C2)iPS cells (clone F-4),and ESCs.The red fragments2FiPSindicate methylated CpG dinucleotides, whereasESCthe blue fragments indicate unmethylated CpGdinucleotides. Missing values are represented inCpG Unitsgray color.DE(D and E) Scatter plots of the global gene expres-sion patterns comparing1F iPS(clone 2)withLin28Oct4Lin28Kf4ESCs (D) and 1F iPS (clone 2) with NSCs (E) byOct4lanoccDNA microarrays. Black lines indicate 2-foldchanges in gene expression levels. Up-and down-C-Mycregulated genes in 1F iPS compared with ESCs or2NSCs are shown in blue or red, respectively. TheFpositions of the pluripotency genes Oct4, Nanog,60x2Sox2, c-Myc, Klf4, and Lin28 are shown in green.NanogNSCESCDISCUSSIONTofurtherassesstheirdevelopmentalpotential,1FiPSweremicroinjected intomouseblastocysts(TableS1).Aftertransferof the injected blastocysts intopseudopregnant recipients, weThisstudyhasseveralimplications.(1)Oct4alonecaninduceobservedbymeansofOct4-GFPexpressionchimeric embryopluripotency in mouse adult NSCs. 1F iPS could be differenti-and germline contribution to the genital ridge of a 13 day post-ated into lineage-committed populations includingGCsandgermline transmission.(2)iPs cells can be generated withoutcoitum (dpc)embryo(Figure6B).Histologicalanalysisrevealedthat 1F iPs contributed to the development of various organs,the oncogenic factors c-Myc and Kif4.Since reactivation ofas shown by β-galactosidase staining (Figure 6C). We further-the c-Myc virus may cause tumor development (Okita et al.,more obtained two adult chimeras from 1F iPS (clone 2) as2007),iPS cells have been generated without infection of thedetermined by coat color and PCR genotyping for the pres-c-Myc oncogene (Huangfu et al., 2008; Nakagawa et al., 2008;ence of the GFP allele,the lacZ allele, and thepMX-Oct4Wernig et al.,2008).However,Kif4, the remaining oncogenictransgene (Figures 6D and 6E, Table S1). In addition, wefactor, might cause tumor formation in offspring. (3) Reducingobserved Oct4-GFP expression in the gonads of adultthe number of factors decreases the chance of retroviral inser-chimeras, which clearly demonstrates germline contributiontional mutagenesis. Analysis of previously described iPS cells(Figure 6F). Chimeric mice were mated in order to verify germ-revealed up to 20 retroviral integrations for all four factors (Aoiet al.,2008; Wernig et al.,2007).In this study,we demonstratelinetransmissionandtwopupsobtainedthereofshowedderi-vation from 1F iPS,as revealed by genotyping for Oct4-GFPthat 1F iPS contain five integrations of only the Oct4 transgeneand viral Oct4 transgene (Figure 6G). All these data dem-(Figure S3). (4) The starting cell population of NSCs, whichonstrate that 1F iPs in vivo possess developmental potentialendogenously express Sox2, c-Myc, and Kif4 as well as APcomparabletoESCs.and SSEA-1,are a unique source for studying the mechanisms414Cell 136,411419, February 6, 2009 2009 Elsevier Inc

To further assess their developmental potential, 1F iPS were microinjected into mouse blastocysts (Table S1). After transfer of the injected blastocysts into pseudopregnant recipients, we observed by means of Oct4-GFP expression chimeric embryo and germline contribution to the genital ridge of a 13 day post￾coitum (dpc) embryo (Figure 6B). Histological analysis revealed that 1F iPS contributed to the development of various organs, as shown by b-galactosidase staining (Figure 6C). We further￾more obtained two adult chimeras from 1F iPS (clone 2) as determined by coat color and PCR genotyping for the pres￾ence of the GFP allele, the lacZ allele, and the pMX-Oct4 transgene (Figures 6D and 6E, Table S1). In addition, we observed Oct4-GFP expression in the gonads of adult chimeras, which clearly demonstrates germline contribution (Figure 6F). Chimeric mice were mated in order to verify germ￾line transmission and two pups obtained thereof showed deri￾vation from 1F iPS, as revealed by genotyping for Oct4-GFP and viral Oct4 transgene (Figure 6G). All these data dem￾onstrate that 1F iPS in vivo possess developmental potential comparable to ESCs. DISCUSSION This study has several implications. (1) Oct4 alone can induce pluripotency in mouse adult NSCs. 1F iPS could be differenti￾ated into lineage-committed populations including GCs and germline transmission. (2) iPS cells can be generated without the oncogenic factors c-Myc and Klf4. Since reactivation of the c-Myc virus may cause tumor development (Okita et al., 2007), iPS cells have been generated without infection of the c-Myc oncogene (Huangfu et al., 2008; Nakagawa et al., 2008; Wernig et al., 2008). However, Klf4, the remaining oncogenic factor, might cause tumor formation in offspring. (3) Reducing the number of factors decreases the chance of retroviral inser￾tional mutagenesis. Analysis of previously described iPS cells revealed up to 20 retroviral integrations for all four factors (Aoi et al., 2008; Wernig et al., 2007). In this study, we demonstrate that 1F iPS contain five integrations of only the Oct4 transgene (Figure S3). (4) The starting cell population of NSCs, which endogenously express Sox2, c-Myc, and Klf4 as well as AP and SSEA-1, are a unique source for studying the mechanisms Figure 2. Characterization of 1F iPS (A) RT-PCR analysis of pluripotency marker expression in 1F iPS clones (C2, C3, and C4) and ESCs (positive control), as well as NSCs (negative control). Primers are specific for transcripts from the respective endogenous locus. b-actin was used as loading control. (B) Quantitative PCR analysis of endogenous expression of the four factors in 1F iPS clones (C2, C3, and C4). RNA levels were determined by quantitative real-time PCR using primers specific for endogenous transcripts. Endogenous relative expression levels of 1F iPS on day 3 (whole popu￾lation) and passage 5 (clones 2, 3, and 4) were compared with those in ESCs. Transcript levels were normalized to b-actin levels. Shown are the averages with standard deviations of three inde￾pendent experiments. (C) Analysis of the DNA methylation pattern of the Oct4, Sox2, and Nanog promoters in NSCs and NSC 1F, as well as in 1F iPS (clones 4 and 2), 2F iPS cells (clone F-4), and ESCs. The red fragments indicate methylated CpG dinucleotides, whereas the blue fragments indicate unmethylated CpG dinucleotides. Missing values are represented in gray color. (D and E) Scatter plots of the global gene expres￾sion patterns comparing 1F iPS (clone 2) with ESCs (D) and 1F iPS (clone 2) with NSCs (E) by cDNA microarrays. Black lines indicate 2-fold changes in gene expression levels. Up- and down￾regulated genes in 1F iPS compared with ESCs or NSCs are shown in blue or red, respectively. The positions of the pluripotency genes Oct4, Nanog, Sox2, c-Myc, Klf4, and Lin28 are shown in green. 414 Cell 136, 411–419, February 6, 2009 ª2009 Elsevier Inc

CellFigure3.Differentiationof1FiPSintoNSCsABPhase(A) RT-PCR analysis showing expression of NSCNestin/DAPImarkergenes fromNSC 1F.β-actinwas usedasloading control.Oct4(B) Morphology (upper left) and Nestin expressionNanog(upper right)of NSC 1F (passage 20).NSC 1F differ-entiated into neurons (Tuj1), oligodendrocytes (O4),Sox2and astrocytes (GFAP).Nuclei were counterstainedwith DAPI (blue).Scale bars,100 μm.Pax6(C) Karyotype of NSC 1F. Metaphase spreads ofTuj1/DAPIO4/GFAP/DAPINestinundifferentiated NSC 1F with a normal set of40mousechromosomesareshownOlig2(D) Scatter plot of the global gene expressionBlbpcomparing donor NSCs with NSC 1F by cDNA mi-croarrays. Black lines indicate 2-fold changes inEmx2gene expression levels. Up- and downregulatedgenes in NSC 1F cells comparedwith NSCs areβ-actinshown in blue or red,respectivelyCDC-MycHtent cells. 1F iPS have the pluripotencySox2Oct4LOSNto differentiate into all three germ layers,as demonstrated by in vitro and in vivoanalysis.They can give rise tomultipoten-KIf4tial NSCs, cardiomyocytes with cardiacNanogaction potential and chronotropic regula-Lin28tion, and GCs as well as germline trans-mission. This study sheds light on theNSCmechanisms involved in reprogrammingofipS cell generationsincetheycanbereprogrammed byone,somatic cells to a pluripotent state.Future studies will showwhetherothersourcesofneuralstemorprogenitorcellpopula-two, three, or four factors.Strikingly,Oct4 alone is sufficient to induce pluripotency intionssuchasmouseorhumanbonemarrow-derivedmesen-chymal stem cells (Hermann et al., 2004; Lee et al., 2003; FuNSCs, which demonstrates its crucial role in the process ofet al.,2008)or dental pulp can be reprogrammed to iPS cellsreprogrammingandsupportsourhypothesisthatNSCsrepre-andwhetherexpressionofOct4canbe inducedbynonretroviralsent an intermediate state between differentiated and pluripo-Figure4.Differentiation of1F iPSinto Cardi-APECAM/DAPIRO-actinin/DAPIomyocytesIn vitro differentiation of endothelial cell and cardi-omyocytes through formation of embryoid bodies(EBs). (A) Clustered PECAM-positive (left, red)endothelial cells and (B) g-actinin-positive (right,red) cross-striated cardiomyocytes are shown.Nuclei were counterstained with Hoechst dye(blue). Scale bar, 50 μm.(C) Atrial- (upper) and ventricular- (lower) like actionpotentials recorded from cardiomyocytes withina7+11 EB.(D) Frequency of beating is enhanced by isoprena-line (ISO, 0.1 μM) and slowed down to a completehalt by the additional application of charbacholISOD(CCh, 10 μM). The CCh effect could be reversed150-by washout (bottom inserts: original traces fromCch(da)ouentime points 1-4). Horizontal dashed lines indicate324OmV lines.100150Csec1minCell136,411-419,February6,2009@2009Elsevier Inc.415

of iPS cell generation since they can be reprogrammed by one, two, three, or four factors. Strikingly, Oct4 alone is sufficient to induce pluripotency in NSCs, which demonstrates its crucial role in the process of reprogramming and supports our hypothesis that NSCs repre￾sent an intermediate state between differentiated and pluripo￾Figure 3. Differentiation of 1F iPS into NSCs (A) RT-PCR analysis showing expression of NSC marker genes from NSC 1F. b-actin was used as loading control. (B) Morphology (upper left) and Nestin expression (upper right) of NSC 1F (passage 20). NSC 1F differ￾entiated into neurons (Tuj1), oligodendrocytes (O4), and astrocytes (GFAP). Nuclei were counterstained with DAPI (blue). Scale bars, 100 mm. (C) Karyotype of NSC 1F. Metaphase spreads of undifferentiated NSC 1F with a normal set of 40 mouse chromosomes are shown. (D) Scatter plot of the global gene expression comparing donor NSCs with NSC 1F by cDNA mi￾croarrays. Black lines indicate 2-fold changes in gene expression levels. Up- and downregulated genes in NSC 1F cells compared with NSCs are shown in blue or red, respectively. Figure 4. Differentiation of 1F iPS into Cardi￾omyocytes In vitro differentiation of endothelial cell and cardi￾omyocytes through formation of embryoid bodies (EBs). (A) Clustered PECAM-positive (left, red) endothelial cells and (B) a-actinin-positive (right, red) cross-striated cardiomyocytes are shown. Nuclei were counterstained with Hoechst dye (blue). Scale bar, 50 mm. (C) Atrial- (upper) and ventricular- (lower) like action potentials recorded from cardiomyocytes within a 7+11 EB. (D) Frequency of beating is enhanced by isoprena￾line (ISO, 0.1 mM) and slowed down to a complete halt by the additional application of charbachol (CCh, 10 mM). The CCh effect could be reversed by washout (bottom inserts: original traces from time points 1–4). Horizontal dashed lines indicate 0 mV lines. tent cells. 1F iPS have the pluripotency to differentiate into all three germ layers, as demonstrated by in vitro and in vivo analysis. They can give rise to multipoten￾tial NSCs, cardiomyocytes with cardiac action potential and chronotropic regula￾tion, and GCs as well as germline trans￾mission. This study sheds light on the mechanisms involved in reprogramming somatic cells to a pluripotent state. Future studies will show whether other sources of neural stem or progenitor cell popula￾tions such as mouse or human bone marrow-derived mesen￾chymal stem cells (Hermann et al., 2004; Lee et al., 2003; Fu et al., 2008) or dental pulp can be reprogrammed to iPS cells and whether expression of Oct4 can be induced by nonretroviral Cell 136, 411–419, February 6, 2009 ª2009 Elsevier Inc. 415