Differentiation of Induced Pluripotent Stem Cells IntoFunctional Oligodendrocytes

MARCIN CZEPIEL,1 VEERAKUMAR BALASUBRAMANIYAN,1 WANDERT SCHAAFSMA,1 MIRJANA STANCIC,2

HARALD MIKKERS,3 CHRISTIAN HUISMAN,1 ERIK BODDEKE,1 AND SJEF COPRAY1*1Department of Neuroscience, University Medical Centre Groningen, A.Deusinglaan 1, 9713AV Groningen, The Netherlands2Department of Cell Biology, Section Membrane Cell Biology, University Medical Centre Groningen, A.Deusinglaan 1,9713AV Groningen, The Netherlands3Department of Molecular Cell Biology, Regenerative Medicine Program, Leiden University Medical Center, 2300 RC Leiden,The Netherlands

KEY WORDSpluripotency; remyelination; cell lineage; multiple sclerosis

ABSTRACTThe technology to generate autologous pluripotent stemcells (iPS cells) from almost any somatic cell type hasbrought various cell replacement therapies within clinicalresearch. Besides the challenge to optimize iPS protocols toappropriate safety and GMP levels, procedures need tobe developed to differentiate iPS cells into specific fully dif-ferentiated and functional cell types for implantationpurposes. In this article, we describe a protocol to differen-tiate mouse iPS cells into oligodendrocytes with the aim toinvestigate the feasibility of IPS stem cell-based therapy fordemyelinating disorders, such as multiple sclerosis. Ourprotocol results in the generation of oligodendrocyte precur-sor cells (OPCs) that can develop into mature, myelinatingoligodendrocytes in-vitro (co-culture with DRG neurons) aswell as in-vivo (after implantation in the demyelinatedcorpus callosum of cuprizone-treated mice). We report theimportance of complete purification of the iPS-derived OPCsuspension to prevent the contamination with teratoma-forming iPS cells. � 2011 Wiley-Liss, Inc.

INTRODUCTION

In 2006, Yamanaka et al. showed for the first timethat mouse somatic cells can be reprogrammed to a plu-ripotent state by means of forced expression of definedtranscription factors (Takahashi and Yamanaka, 2006).These reprogrammed cells, induced pluripotent stem(iPS) cells, exhibited typical embryonic stem (ES) cellcolony morphology and demonstrated endogenousexpression of pluripotency genes. The mouse IPS cellsappeared to comply to the most strict criteria for pluri-potency including teratoma formation after subcutane-ous injection and contribution to chimeric animals oninjection into blastocysts (O’Malley et al., 2009; Scheperand Copray, 2009; Yamanaka, 2008).

The subsequent successful attempts to generatehuman iPS cells revealed their potential value for theclinic (Aasen et al., 2008; Park et al., 2008b; Takahashiet al., 2007; Yu et al., 2007). The possibility to acquirepatient-specific pluripotent cells seems to be the majoradvantage of this novel technology, as the autologous

origin of iPS-derived differentiated cells enables theiruse in cell-based therapies without the risk of transplantrejection by the host immune system. Apart from cell-based therapies, the iPS technology allows the use ofdisease-specific stem cells to study cell pathogenic mech-anisms underlying many human diseases (Colman andDreesen, 2009; Park et al., 2008a; Saha and Jaenisch,2009).

With these motives, we started to examine the possi-bility to differentiate iPS cells towards a glial fate, inparticular into an oligodendrocyte cell lineage – a keycell type in demyelinating diseases such as multiplesclerosis (MS). The most disabilitating symptoms andfunctional deficits in the more progressed stages of MSare caused by axonal loss followed by retrograde neuro-nal degeneration (Tallantyre et al., 2010). It has becomeclear that axonal degeneration is not only occurring inchronic lesions but also signs of axonal injury have beendetected already in acute lesions (Filippi et al., 2003;Trapp et al., 1998). Acute lesions in the relapsing-remit-ting form of MS are caused by infiltrating reactiveT-cells that induce inflammation and cytokine-release,damaging myelin and oligodendrocytes. Due to the tem-poral character of these relapses, axonal damage can berepaired by local oligodendrocyte precursor cells (OPCs),but in most cases parallel axonal circuitry takes overthe lost axonal connections resulting in partial or com-plete functional recovery (Bjartmar et al., 2003). Thetransition of relapsing-remitting MS to the secondaryprogressive form of MS marks the boundary of the plas-ticity of the brain where axonal loss can no longer becompensated. The present therapy for MS, with immu-nomodulatory drugs such as interferon-b, glatirameracetate, mitoxantrone is predominantly directed toreduce the invasion of aggressive T-cells (Stuve et al.,2008); they do not prevent or restore axonal damage. Itis clear that efficient treatment of MS must also include

M.C. and V.B. contributed equally to this work.

Grant sponsor: Dutch Foundation MS Research, 08-637 MS.

*Correspondence to: Dr. S. Copray, Department of Neuroscience, UniversityMedical Centre Groningen, A. Deusinglaan 1, 9713AV Groningen, The Netherlands.E-mail: j.c.v.m.copray@med.umcg.nl

Received 20 April 2010; Accepted 19 January 2011

DOI 10.1002/glia.21159

Published online 24 March 2011 in Wiley Online Library (wileyonlinelibrary.com).

GLIA 59:882–892 (2011)

VV� 2011 Wiley-Liss, Inc.

strategies to protect and restore axons from the veryfirst signs of MS. Although various novel neuroprotec-tive agents may become useful, the most effective way toprotect an axon is to induce its remyelination as soon aspossible either by promoting the myelinating activity ofendogenous OPCs or by the implantation of exogenousremyelinating or remyelination-promoting cells (Martinoet al., 2010; Uccelli and Mancardi, 2010).

So far, potential clinically relevant OPCs have beenderived from neural stem cells (NSCs) (Ben-Hur et al.,2003; Einstein et al., 2007; Magalon et al., 2007; Plu-chino et al., 2003; Sher et al., 2009) or ES cells (Brustleet al., 1999; Liu et al., 2000; Nistor et al., 2005). Bothcell types have unacceptable limitations as sources forclinically relevant transplantable OPCs, related to theirrestricted availability in case of autologous NSCs andheterologous, ethically disputed origin in case of EScells. iPS cells clearly lack these drawbacks.

A few in-vitro protocols have been developed to inducethe differentiation of ES cells into an oligodendrocyticcell lineage with a variable efficiency and stability (Bar-beri et al., 2003b; Brustle et al., 1999; Glaser et al.,2005) and more recently the first attempts to differenti-ate IPS cells into neural cell lineage have been described(Onorati et al., 2010). In this study, we have modifiedand optimized these protocols and tested their applic-ability to obtain pure suspensions of functional OPCsfrom mouse iPS cells that can differentiate and matu-rate into myelinating oligodendrocytes in-vitro andin-vivo.

MATERIALS AND METHODSCell Culture and Generation of Mouse iPS Cells

For the generation of mouse iPS cells, E14 mouse em-bryonic fibroblasts (MEFs) were used. C57BL/6 miceembryos (E14) were explanted from the anaesthetizedmother, decapitated and after removal of spinal cord,intestines, long, heart, etc. . ., the embryo body wasminced and trypsinized: cells were plated in Dulbecco’smodified Eagle’s medium (DMEM) medium (Invitrogen,

Breda, The Netherlands, www.invitrogen.com). MEFswere cultured until passage 3 before reprogramming.For retroviral particle production, Phoenix eco packag-ing cells were transfected (Fugene-HD; Roche, Almere,The Netherlands, https://www.roche-applied-science.com)separately with Oct4, Sox2, Klf4, cMyc plasmids(Addgene, Cambridge, www.addgene.org). Retroviruscontaining supernatants were collected and replacedwith the fresh medium 48 h after transfection. Mediacontaining virus and 4 lg/mL polybrene (Sigma-Aldrich,Zwijndrecht, The Netherlands, www.sigmaaldrich.com)were mixed, filtered through an 0.45-lm filter(Whatman, The Netherlands, www.whatman.com) andadded freshly to MEFs. The transduction procedure wasrepeated three times in 24 h gaps. Three days after thelast transduction, cells were trypsinized and replated onirradiated mouse embryonic fibroblasts (iMEFs) and cul-tured in ES medium. One week after replating ofinfected MEFs, the first ES-like colonies started toappear. Colonies were picked and propagated on iMEFsin ES medium (Knockout D-MEM, 15% knockout serumreplacement, 1% nonessential amino acids (all Invitro-gen, Breda, The Netherlands, www.invitrogen.com), 2mM L-Glutamine, 100 U/mL penicillin, 100 lg/mL strep-tomycin (all PAA, C€olbe, Germany, www.paa.com), 100lM 2-mercaptoethanol) supplemented with 1000 U/mLLeukemia inhibitory factor (LIF; Millipore, Amsterdam,The Netherlands, www.millipore.com). Subsequent pas-sages were done using trypsin-based dissociation. To an-alyze the differentiation potential of our iPS cells, theywere cultured as embryoid bodies (EBs) for 8 days andtransferred onto gelatine-coated coverslips for another 5days. Cells were then fixed and stained for markers ofthe three germ layers.

Directed Differentiation of iPS TowardOligodendrocytes

To induce their differentiation into an oligodendrocyticcell lineage, iPS cells were trypsinized and culturedas EBs for 8 days in EB medium [Knockout D-MEM,15% fetal calf serum (FCS; PAA), 1% nonessentialamino acids, 2-mM L-Glutamine, 100 U/mL penicillin,100 ug/mL streptomycin, 100-lM 2-mercaptoethanol]. Togenerate neural precursor cells (NPCs), 8-day-old EBswere dissociated and cultured in serum free N2 medium(DMEM/F12 (Invitrogen), 2% N2 supplement (PAA), 2mM L-Glutamine, penicillin/streptomycin, 1% sodiumpyruvate (Invitrogen) supplemented with 20 ng/mLbasic fibroblast growth factor (FGF2; Invitrogen), and20 ng/mL epidermal growth factor (EGF; Invitrogen) for7–8 days. For oligodendrocyte differentiation, iPS-derived NPCs were plated on coverslips coated withpoly-D-lysine (PDL,10 lg/mL)-laminin (5 lg/mL) andcultured in N2 medium supplemented with 10 ng/mLplatelet-derived growth factor (PDGF; PeproTech, Lon-don, UK, www.peprotechec.com) for another 4 days(for first 2 days N2 medium also contained 20 ng/mLFGF2 and 20 ng/mL EGF). To enhance the maturation

Abbreviations

DMEM Dulbecco’s modified Eagle’s mediumEB embryoid bodyEGF epidermal growth factorES embryonic stemGFAP glial fibrillary acidic proteiniMEF irradiated mouse embryonic fibroblastiPS cells induced pluripotent stem cellsMBP myelin-basic proteinMEF mouse embryonic fibroblastMS multiple sclerosisNGS normal goat serumNSC neural stem cellNT neurotrophinOPC oligodendrocyte precursor cellPDGF platelet derived growth factorPDGFR platelet derived growth factor receptorPDL poly-D-lysinePFA paraformaldehyde.

883IPS CELL-DERIVED OLIGODENDROCYTES

GLIA

of oligodendrocytes, triiodothyronine (T3) (30 ng/mL;Sigma-Aldrich) and neurotrophin 3 (NT-3) (10 ng/mL;R&D Systems, Abingdon, UK, www.rndsystems.com)were added during the final phase of differentiation (6days).

Reverse Transcription-Polymerase ChainReaction Analysis

Reverse transcription-polymerase chain reaction (RT-PCR) analysis of the changing expression of various celltype markers during the differentiation from IPS cellsinto an oligodendrocytic cell lineage was performed onmRNA collected from pooled MEFs, mouse IPS cells,from IPS-derived EBs, from cells pooled in the initialIPS-N2 stage, and from differentiating IPS cells in theN2-PDGF stage in which the formation of oligodendro-cyte precursors (OPCs) was induced. The forward andreverse primers used are indicated in Table 1. Each RT-PCR analysis was done on three samples.

Immunocytochemical Characterization

For immunocytochemistry stainings, cells were cul-tured on coverslips coated with gelatin (in case of EBs),PDL-laminin (for oligodendrocyte differentiation) or onirradiated MEFs (undifferentiated iPS cells). EBs oncoverslips and cell cultures, at every third day from thestart of oligodendrocyte differentiation, were fixed with4% paraformaldehyde (PFA), washed with PBS andincubated in PBS containing 2% FCS and 5% normalgoat serum (NGS) for 1 h to block unspecific antibodybinding sites. Cells were then incubated with primaryantibodies for 1 h at room temperature, or at 4�C over-night. After the washing step, cells were incubated withappropriate secondary antibody and Hoechst for 1 h atroom temperature. Coverslips were than mounted andexamined under the fluorescent microscope. Antibodiesused to identify IPS cells and their derivatives: SSEA1,Oct4, GATA-4 (sc-21702, sc-5279, sc-25310 respectively;

all Santa Cruz Biotechnology, Heidelberg, Germany,www.scbt.com), Nanog, Sox2, Beta III Tubulin (ab80892,ab15830, ab7751, respectively; all Abcam, Cambridge,UK, www.abcam.com), Nestin, NG2, MBP and NF(mab353, ab5320, ab980, mab1615, respectively; allChemicon, Millipore, Amsterdam, The Netherlands,www.millipore.com), Desmin (m0760; DAKO, Heverlee,The Netherlands, www.dakobv.nl), RIP (DevelopmentalStudies Hybridoma Bank, Iowa, http://dshb.biology.uiowa.edu/).

Co-culture of iPS-Derived OPCs withDRG Neurons

To examine the myelinating capacity of IPS-derivedOPCs in-vitro, we co-cultured them with rat DRG neu-rons. DRG neurons were dissected from 15-day-oldWistar rat embryo’s and digested in papain (1.2 U/mL,Sigma), L-cysteine (0.24 mg/mL, Sigma) and DNaseI (40 mg/mL, Roche) for 1 h at 37�C. The dissociatedcells were plated at a density of 60,000 cells per 13-mmcoverslips (VWR) pre-coated with PLL (5 lg/mL, Sigma)and growth factor reduced matrigel (1:40 dilution; BDBioscience). DRG neurons were cultured 21 days inDMEM (Gibco, supplemented with 10% FCS, Bodinco; L-glutamine and PenStrep, Invitrogen) in the presence ofnerve growth factor (100 ng/mL, Serotec). Cells werepulsed four times for 2 days with fluorodeoxyuridine (10lM, Sigma) to remove contaminating proliferating cells,in particular fibroblasts and Schwann cells. Purity ofthe DRG culture was microscopically confirmed. Subse-quently, approximately 50,000 iPS-derived OPCs (seedifferentiation protocol) were seeded onto coverslips con-taining the DRG neurons with extensive axonaloutgrowth. The following day medium was changed toN2 with 30 ng/mL T3 and 10 ng/mL NT3. OPCs wereco-cultured with the DRG neurons for 14 days with me-dium change every second day. After that period, cellswere fixed and subjected to immunocytochemical stain-ing for myelin-basic protein (MBP), neurofilament, andparanodin protein.

TABLE 1. Primers Used in Reverse Transcription-Polymerase Chain Reaction Analysis

Forward primer Reverse primer

Ascl1 AACAAGAAGATGAGCAAGGTGG TCTGGGCTAAGAGGGTCGTAGBlbp ACATGAAAGCTCTGGGCGTG TCTCCATCCAACCGAACCACGFAP GGCTCGTGTGGATTTGGAGA GACTCCAGATCGCAGGTCAAGGlast ACAGGAATGGCGGCCCTAG CTCCATGGCCTCTGACACGTMBP GCTTCTTTAGCGGTGACAGGG TGGAGGTGGTGTTCGAGGTGMusashi-1 GAAGGGCTGCGCGAATACTT TCTGTGCCTGTTGGTGGTTTTNanog CAGGTGTTTGAGGGTAGCTC CGGTTCATCATGGTACAGTCNestin CAGCAACTGGCACACCTCAA TGCATTTTTAGGATAGGGAGCCTNG2 ACCTGCTCTACCGGGTGGT CTGGCACACTGGCTAGGAGGNkx2.2 CTCCAATACTCCCTGCACGG GATTGTGCTCCCAGGCCTGNotch1 CCTGGAGACCAAGAAGTTCCG CAGAAACTGCTGCATGTAACGGOct4 TCTTTCCACCAGGCCCCCGGCTC TGCGGGCGGACATGGGGAGATCCOlig1 CTCTTCCACCGCATCCCTTC GATGTAGTTGCGGGCGAGCOlig2 TGGCTTCAAGTCATCTTCCTCC TCGCTCACCAGTCGCTTCATOtx2 TCCAGGGTGCAGGTATGGTTT GGACGCTGGGCTCCAGATAGPax6 GGCAATCGGAGGGAGTAAGC TCTGCCCGTTCAACATCCTTPDGFRa ACCAGACCCAGACATGGCC AAGACGGCACAGGTCACCACSox2 TAGAGCTAG ACTCCGGGCGATGA TTGCCTTAAACAAGACCACGAAATlx TCAAGAGGAGCATTCGAAGGAA CGAGAGTCTGCCGTTCAGGA

884 CZEPIEL ET AL.

GLIA

IPS-Derived OPC Implantation in CuprizoneMouse Model

To examine the in-vivo myelinating capacity of IPS-derived OPCs, we implanted them in the demyelinatedcorpus callosum of cuprizone-fed mice. C57Bl/6 micewere put on a diet of 0.2% (w/w) cuprizone (SigmaAldrich, Zwijndrecht, The Netherlands), a copper chela-tor. This diet leads to selective oligodendrocyte death fol-lowed by demyelination of axons mainly in the corpus cal-losum (Kipp et al., 2009; Matsushima and Morell, 2001).During short-term exposure (<6 weeks), dying oligoden-drocytes are replaced by oligodendrocyte precursors cells(OPCs) located in the corpus callosum and adjacent tis-sues. However, feeding cuprizone diet for a longer term(>12 weeks) results in depletion of the pool of endoge-nous OPCs in the corpus callosum and finally leads to itscomplete demyelination (Mason et al., 2004). This long-term cuprizone-diet model with the demyelinated corpuscallosum axons has been shown adequate to study theremyelination potential of implanted OPCs.

At 12 weeks after the start of the cuprizone diet, micewere divided into three groups one group (n 5 6)received a sorted suspension of IPS-derived OPCs (seedifferentiation protocol), a second group (n 5 3) wasgiven an unsorted suspension of IPS-derived OPCs, anda third group (n 5 2) was injected with an unsorted sus-pension of cells taken at the beginning of the differentia-tion protocol containing still undifferentiated IPS cells.One day before implantation, IPS-derived OPCs weredetached from the PLO-Laminin substrate with accutaseand subjected to cell sorting for the presence of PDGFre-ceptor alpha. Briefly, cells were first incubated with antiCD16/32 antibody (eBioscience; 14-0161) for 20 min. toblock unspecific antibody binding with Fc fragment andthan incubated with PE-conjugated anti-PDGF receptoralpha antibody (eBioscience; 12-1401) for additional 30min. Proper isotype controls were used. Cells weresorted into the PLO-Laminin coated cell culture platesand kept overnight in N2 medium with FGF2, EGF,PDGF. Finally, the cells to be implanted were exposed tothe N2 medium containing 10 lg/mL of the fluorescentdye rhodamine-dextrane (MiniRuby; Molecular probes;D3312) which was taken up via endocytoses by allOPCs. The exposure time of the cells to the mini-Rubycontaining medium was confined to 1 h to limit theuptake, since longer exposure has been shown to fill upcells to an extent that may interfere with normal func-tioning. On the day of transplantation, cells were har-vested with accutase, counted, and resuspended in PBSat concentration of 25,000 cells/lL. Four microliters ofcell suspension was injected into corpus callosum ofC57BL/6 mice using the following stereotactic co-ordi-nates (in reference to Bregma point): 10.98 mm (anteri-oposterior axis), 21.75 mm (lateromedial axis), 22.25mm (vertical axis) (Copray et al., 2006; Sher et al.,2009), suspensions of 100,000 cells in 4 lL PBS wereslowly injected into the corpus callosum of ketamine-anesthetized mice using a 10 lL Hamilton injectionsyringe (22s/200/3) (Hamilton, Reno, Nevada, 80365).

Every injection was done within a standardized time-window, i.e., 5 min injection time and 2 min depositionrest, before needle retraction to prevent potential varia-tion in the effect of shearing forces. After the implantationof the IPS-OPCs, mice were taken off the cuprizone dietand put back on normal diet to avoid degeneration of theimplanted OPCs by cuprizone. Mice receiving intracere-bral cell suspensions containing undifferentiated IPS cellswere terminated and perfusion fixated within 2 weekssince they all developed massive teratomas at the site ofinjection. Mice with implantation of sorted and unsortedIPS derived OPCs were perfusion fixated at 2, 4, and 24weeks (long-term check for teratoma formation).

Immunohistochemical Analyses ofCell Implants

Mice were perfused transcardially with 4% PFA underisoflurane anaesthesia. Brains were excised and sec-tioned on a cryostat for immunohistochemical analysisof the cell implants. Implanted cells were identified bytheir inclusion of mini-Ruby. To analyze the differentia-tion of the implanted cells, the following primary anti-bodies directed against oligodendrocyte-, neuron- orastrocyte-markers were used: for astrocytes, anti glialfibrillary acidic protein (GFAP) (Chemicon, mab3402);for neurons, anti microtubule-associated protein (MAP2)(Chemicon, ab5622); for oligodendrocytes, anti-RIP(Chemicon, mab1580), anti-platelet derived growth fac-tor receptor alpha (PDGFR-a) (Santa Cruz Biotechnol-ogy, Santa Cruz, Ca) anti-NG2 (Chemicon, ab5320), andanti-MBP (Chemicon, ab980). Subsequently, variousfluorescent secondary antibodies were used to visualizethe specific primary immunoreaction product in singleand double immunohistochemical stainings.

RESULTSIPS Cell Induction

Using retroviral transduction of MEFs with the Yama-naka pluripotency factors, we were able to generate iPScells that resembled ES cells in terms of morphology(Fig. 1A), alkaline phosphatase activity (Fig. 1A) and en-dogenous expression of pluripotency markers at mRNAlevel (Fig. 1C). Immunocytochemical stainings alsorevealed the presence of pluripotency-associatedmarkers such as Nanog, SSEA1, Oct4, and Sox2 in thesecells at protein level (Fig. 1D–G).

We then tested the in-vitro differentiation capacity ofour iPS cells. For this, cells cultured as EBs (Fig. 2A)for 8 days were seeded on gelatin-coated coverslips andfive days later stained for the markers of three germlayers. Immunostainings showed that iPS-derived EBscan robustly differentiate into cells of endodermal(GATA-4), mesodermal (Desmin) and ectodermal (bIII-tubulin) origin (Fig. 2B–D). Moreover, we were able toobserve contracting cardiomyocytes in floating as well asin seeded EBs (data not shown). Expression of lineage

885IPS CELL-DERIVED OLIGODENDROCYTES

GLIA

Fig. 1. (A) Phase contrast morphology of mouse iPS cell colonies.(B) Alkaline phosphatase staining of iPS colonies. (C) PCR showingendogenous expression of pluripotency-associated genes. (D–G) Immu-nocytochemical stainings of mouse iPS cells for pluripotency-related

genes: Nanog (D-D00), SSEA1 (E-E00), Oct4 (F-F00), and Sox2 (G-G00).[Color figure can be viewed in the online issue, which is available atwileyonlinelibrary.com.]

886 CZEPIEL ET AL.

GLIA

specific markers was also evident in EBs formed fromboth iPS and ES cells at mRNA level, but not detectablein undifferentiated cells and MEFs used as a control(Fig. 2E). Additional evidence for the true pluripotentnature of our IPS cells was obtained after their intrace-rebral implantation (see below).

Oligodendrocyte Differentiation ofMouse IPS Cells

Our main goal was to investigate whether the iPScells that we generated can be differentiated towards anoligodendrocyte cell lineage. To examine this, we appliedthe differentiation protocol described in Materials andMethods (Fig. 3A) and checked the cells for expressionof markers typical of oligodendrocyte development at dif-ferent points of differentiation. Selective differentiationand survival of neural lineage committed cells wasachieved by culturing enzymatically dissociated 8-dayold EBs in N2 medium in the presence of FGF2 andEGF. We observed that many (nonneural) cells, forwhich the culture conditions apparently were not suita-ble, died at this stage of differentiation. The majority ofthe surviving cells were nestin-positive neural precur-sor/stem cells (Fig. 3B); they began to proliferate exten-sively and could be propagated under these conditionsfor at least several passages. After replating these neu-ral precursor cells on PDL-Laminin-coated coverslips

and culturing in medium containing PDGF, we couldobserve a clear change in cell morphology. At this stage,multipolar, NG2-positive (Fig. 3E) oligodendrocyte pre-cursors (OPCs) appeared. The differentiation and matu-ration capacity of these OPCs toward mature oligoden-drocytes was verified by continuing cell culture in me-dium supplemented with T3 and NT-3. The subsequentchange of cell morphology was apparent (Fig. 3F) andafter 6 days of treatment, we could identify fully devel-oped, MBP-expressing oligodendrocytes (Fig. 3F,G)amidst some astrocytes (Fig. 3B) and a few neurons(Fig. 3C). Eventually, at the end of this differentiationprocedure, MBP-positive oligodendrocytes constituted18% of the total cell population. We attempted to furtherpurify the final suspension of oligodendrocytes by per-forming FACs cell sorting for the PDGFreceptor alphaon the suspension of OPCs immediately after theirappearance at day 16 of the IPS differentiation protocol.This purification was not complete since apparently afraction (varying from 3% to 20%) of other PDGFrecep-tor alpha expressing cell types (presumably A2B5 cellsand astrocytes) contaminated the final suspension ofIPS-derived oligodendrocytes.

Using RT-PCR, we further documented and confirmedthe changing mRNA expression for cell specific markersduring our differentiation protocol from mouse IPS cells,toward embryoid body (EB) formation and subsequentlyduring the different culture periods in neural precursorcell selective N2 media toward oligodendrocyte precur-

Fig. 2. Spontaneous in-vitro differentiation of iPS-derived embryoidbodies (A). Immunocytochemistry indicates the presence of cellsof endodermal (GATA4) (B), mesodermal (Desmin) (C), and ectodermal(b-IIItubulin) (D) origin. (E) RT-PCR shows expression of GATA6 (endo-derm), Brachyury (mesoderm), and Map2 (ectoderm) in differentiating

but not in undifferentiated ES and iPS cells. (EBs—8-day-old embryoidbodies; EBs diff.—8-days-old embryoid bodies cultured on gelatine foranother 5 days). [Color figure can be viewed in the online issue, whichis available at wileyonlinelibrary.com.]

887IPS CELL-DERIVED OLIGODENDROCYTES

GLIA

sors (Fig. 4). The RT-PCR analysis (all performed intriplo) shows the silencing of the pluripotency factorsNanog and Oct4 during the transition into neural stem/precursor cells, whereas Sox2 apparently remainsexpressed up to the stage of OPCs. The appearance ofNSCs during the N2 phase becomes evident by the(increase in) nestin expression and Musashi1. The sub-sequent differentiation into neural cell types is accompa-nied by the expression of early neural factors such asPax6, Glast, Blbp, Ascl1 (5Mash1), Tlx, Nkx2.1. andGFAP. The final transition into an oligodendrocytic celllineage is accompanied by the strong expression of Olig1and Olig2 (in fact re-expression, since both transcriptionfactors appear to be expressed in undifferentiated IPScells), the increase in expression of PDGFRa and NG2and the appearance of MBP expression. According to ourRT-PCR, MEFs demonstrated the mRNA expression ofthe glutamate transporter Glast, which was downregu-lated during the reprogramming into IPS cells and up-regulated again during subsequent differentiation intoneural cell types. In addition, also proteoglycan NG2appears to be a component of the membrane of MEFs,downregulated during IPS formation and up regulatedagain when differentiating into OPCs. The ubiquitousrole of Notch1 signaling is reflected in its expression in

most cell types during the IPS cell reprogramming anddifferentiation procedure, although the highest expres-sion was clearly found in the NSCs and the OPCs whichappear in the N2-PDGF phase (Fig. 4).

Myelinating Activity of IPS-Derived OPCs

In-vitro

Next, we examined whether our IPS derived-OPCswere able to develop into functional oligodendrocytesand were able to form proper myelin around nude axons.To that purpose, we co-cultured the IPS-derived OPCswith a pure, 2 weeks-old culture of DRG neurons. Twoweeks after plating of IPS-derived OPCs, we observedextensive myelin formation by the IPS-derived oligoden-drocytes (Fig. 5) by the expression of MBP and parano-din, a node of Ranvier protein. Our (co)culture periodappeared to be too short for proper node of Ranvier for-mation, since clear aggregations of paranodin arounddeveloping nodes of Ranvier was not observed. ControlDRG cultures, without addition of IPS-derived OPCs,did not show any sign of myelination activity and noMBP staining was observed (data not shown).

Fig. 3. (A). Scheme of the mouse IPS-oligodendrocyte differentiationprocedure. Immunocytochemical analysis reveals the different celltypes generated during this procedure: (B) iPS-derived nestin-positiveneural precursor cells (NPCs), (C) b-III tubulin-positive neurons, (D)

GFAP-positive astrocytes, (E) NG2-positive oligodendrocyte precursorcells, (F) MBP-positive oligodendrocytes, and (G) MBP/RIP double la-beled positive oligodendrocytes. [Color figure can be viewed in theonline issue, which is available at wileyonlinelibrary.com.]

888 CZEPIEL ET AL.

GLIA

In-vivo

To obtain proof of principle that IPS-derived OPCs canalso become functional oligodendrocytes in-vivo, weimplanted them in the demyelinated corpus callosum ofcuprizone-fed mice. After 14 weeks on 0.2% cuprizone,the corpus callosum becomes completely demyelinatedand devoid of endogenous OPCs yielding a proper in-vivo substrate for axonal (re) myelination for implantedOPCs. Intracerebral Injection of a cell suspension whichstill contained undifferentiated IPS cells resulted within2 weeks in the formation of a large teratoma (Fig. 6A),consisting of tissues of the three germlayers (Fig. 6B).The unsorted as well as the sorted suspension of IPS-

derived OPCs did not result in teratoma formation, evenafter a period of over 24 weeks, indicating that no undif-ferentiated IPS cells were present after our oligodendro-cyte differentiation procedure. Analysis of the fate ofthe implanted IPS-derived OPCs at 2 or 4 weeks after ste-reotactic injection in the corpus callosum by immunohis-tochemistry revealed that, as described before (Copray etal., 2006; Sher et al., 2009), over 80% of the implantedcells did not survive the procedure of stereotactic injectionand were rapidly removed by microglia. SurvivingminiRuby labeled cells, in case of the unsorted implants(containing only a low fraction of IPS-derived OPCs),were in majority undifferentiated nestin-positive NSCs(Fig. 6C). Surviving IPS-derived OPCs of the sorted graftsall developed into mature MBP-expressing oligodendro-cytes that contributed to the remyelination of the corpuscallosum axons (Fig. 6E).

DISCUSSION

There have been a few protocols described in the litera-ture that promote the generation of oligodendrocytes fromES cells, with different efficiencies (Barberi et al., 2003;Brustle et al., 1999; Glaser et al., 2005). Based on them,we composed a protocol and applied it to differentiatemouse iPS cell into an oligodendrocytic cell lineage. Ourdata show that our protocol could successfully generatemature iPS-derived oligodendrocytes that could fully my-elinate axons in culture as well as in-vivo after implanta-tion in the demyelinated corpus callosum of cuprizone-fedmice. Although the percentage of IPS-derived oligodendro-cytes was relatively low (18%), a purified suspension couldbe obtained via specific cell sorting for the OPCmembranereceptor PDGFreceptor alpha at the OPC stage. The 18%oligodendrocyte differentiation we found is comparablewith the yield of the procedure recently published byOnorati et al., but they were unable to provide evidencefor the myelinating capacity of the OPC like cells theyobtained with their approach (Onorati et al., 2010).

The procedure used by us to differentiate IPS cells to-ward an oligodendrocytic cell lineage is based on thesequential induction at proper timely intervals by appro-priate induction- and growth factors which have beenshown to play such a role during normal embryonic de-velopment (Miller, 2002; Woodruff et al., 2001). Theappearance (and disappearance) of the proper cellmarkers during the various stages of this differentiationprocess is clearly visualized in the RT-PCR analysis(Fig. 4). Obviously, the differentiation process does notaffect the cells synchronously, and so the cells pooled forRT-PCR analysis during the different stages do notall comprise a single cell stage and a single cell type:evidently, the bands identified in the PCR analysis mayinclude some mRNA contaminated by cells of a previousor even a subsequent stage. So, contaminating cells inthe OPC cell suspension pooled in the final stage of thedifferentiation procedure (culture in N2 medium plusPDGF) may account for the presence of mRNA encodingfor NSC markers (e.g., Sox2, nestin, musashi) and for

Fig. 4. RT-PCR analysis of the changing mRNA expression patternof major cell markers during the differentiation of mouse IPS cells intooligodendrocytic cell lineage. (MEF 5 mouse embryonic fibroblast; iPS5 indiced pluripotent stem cells; iPS-EBs 5 embryoid body formedfrom iPS cells; iPS N2 5 IPS-embryoid body derived cells cultured inN2 medium, induction phase into neural stem cells (NSCs); iPS N2PDGF 5 IPS-NSCs cultured in N2 medium supplemented with PDGF,induction phase of OPC cells.

889IPS CELL-DERIVED OLIGODENDROCYTES

GLIA

markers of other differentiated neural cell types, i.e.,neurons (Pax6) and astrocytes (Blbp, GFAP, and theglutamate transporter Glast). Most other markers in this

stage, however, must be considered specific OPC markers,like Olig1 and Olig2, even including Glast which also hasbeen demonstrated in developing oligodendrocytes

Fig. 5. Co-culture of IPS-derived OPCs with rat DRG neurons. (A)After 1 week, IPS-derived OPCs develop into MBP-positive oligoden-drocyte that line their extensions along the neurofilament (NF) stainedDRG-axons. (B) In the second week, the IPS-derived oligodendrocytesstart to myelinate the DRG axons in its environment. (C, detail in D).At that time, the OPCs also start to produce paranodin (red) that atthis stage is still distributed over the extensions running along the

axons. (E) Confocal microscopy revealed that in 2 weeks time, longstretches of DRG axons have been myelinated by the co-cultured IPS-derived oligodendrocytes. (F) The 2 weeks co-culture still contains earlystages of IPS-derived oligodendrocytes: here a single IPS-derived oligo-dendrocyte is depicted co-stained for the expression of paranodin, onlyvisible in the cytoplasm.

890 CZEPIEL ET AL.

GLIA

(DeSilva et al., 2009). Cell contamination of the initialsuspension of MEFs with e.g., pericytes or immune cells,may also have been responsible for the observed (weak)expression of the glutamate transporter Glast, the proteo-glycan NG2, Notch1, and the PDGF receptor, althoughspecific expression of for instance Glast and Notch1 havebeen described before in embryonic fibroblasts (Cooper etal., 1998; Ishikawa et al., 2008). The expression of theNotch1 receptor throughout the reprogramming and dif-ferentiation process seems to reflect the well-describedubiquitous role of the highly conserved Notch signalingpathway in cell fate decision, tissue patterning, andmorphogenesis in a wide variety of cell types (Yin et al.,2010; Yuan et al., 2010).

The evidence that mouse iPS cells can be differentiatedinto functional oligodendrocytes provided by this studyshould be an encouraging signal for adapting similarprocedures for human IPS research and therapy. It hadbeen shown before that OPCs can be generated in high

numbers from human ES cells by modulating Shh andFGF2 signaling pathways (Hu et al., 2009). The majoradvantage of iPS technology over the use of ES cells is thepossibility to acquire autologous oligodendrocytes, whichis fundamental in perspective of cell-based therapy in thefuture particularly for diseases with a major immunologi-cal component. There are however several issues thathave to be solved before iPS-derived cells can be safelyapplied in the clinic. One of them is the potential risk ofcontamination of undifferentiated IPS cells. We havedeveloped methods to deplete cell suspensions frommouse IPS cells via their selective destruction with sap-orin-conjugated antibodies directed against SSEA-1 (datanot shown); however, application of this purificationapproach appeared to be unnecessary in this study sinceat the end of our oligodendrocyte differentiation protocolno undifferentiated IPS cells were present in the sortedas well as in the unsorted suspensions. Furthermore, forpotential clinical application iPS cells without any

Fig. 6. (A) Injection of an undifferentiated IPS contaminated cellgraft into the corpus callosum resulted within 2 weeks in the formationof large teratomas that grew out of the brain and which on histologicalhaematoxylin-eosin staining appear to contain: (B) bone, muscle, carti-lage, and gland tissue. (C) Unsorted surviving IPS derived cells labeledwith miniRuby, in the corpus callosum (c.c.) 4 weeks after implantation

appear to consist merely of undifferentiated nestin-positive neural pre-cursors that were negative for GFAP astrocyte staining (D). (E)MiniRuby labeled cells (arrows) of the sorted IPS-derived OPC cellgraft, surviving 4 weeks after injection in the corpus callosum ofcuprizone-fed mice, all expressed MBP and contributed to the remyeli-nation of the corpus callosum axons. Bars 5 20 lm.

891IPS CELL-DERIVED OLIGODENDROCYTES

GLIA

genomic modifications should be used, meaning that non-viral, nonintegrative IPS induction protocols should befurther developed and optimized. In that respect, the lat-est developments with the use of synthetic mRNA for IPSreprogramming is very promising (Warren et al., 2010).Finally, the entire IPS induction and differentiation pro-cedure would need to be established under fully defined,xeno-free conditions.

ACKNOWLEDGMENTS

The authors gratefully acknowledge the technical as-sistance of I. Manting-Otter, F. Dijk, and M.Meijer.

REFERENCES

Aasen T, Raya A, Barrero MJ, Garreta E, Consiglio A, Gonzalez F, Vas-sena R, Bilic J, Pekarik V, Tiscornia G, Edel M, Boue S, Izpisua Bel-monte JC. 2008. Efficient and rapid generation of induced pluripotentstem cells from human keratinocytes. Nat Biotechnol 26:1276–1284.

Barberi T, Klivenyi P, Calingasan NY, Lee H, Kawamata H, Loonam K,Perrier AL, Bruses J, Rubio ME, Topf N, Tabar V, Harrison NL, BealMF, Moore MA, Studer L. 2003. Neural subtype specification of fertil-ization and nuclear transfer embryonic stem cells and application inparkinsonian mice. Nat Biotechnol 21:1200–1207.

Ben-Hur T, Einstein O, Mizrachi-Kol R, Ben-Menachem O, Reinhartz E,Karussis D, Abramsky O. 2003. Transplanted multipotential neuralprecursor cells migrate into the inflamed white matter in response toexperimental autoimmune encephalomyelitis. Glia 41:73–80.

Bjartmar C, Wujek JR, Trapp BD. 2003. Axonal loss in the pathology ofMS: Consequences for understanding the progressive phase of thedisease. J Neurol Sci 206:165–171.

Brustle O, Jones KN, Learish RD, Karram K, Choudhary K, WiestlerOD, Duncan ID, McKay RD. 1999. Embryonic stem cell-derived glialprecursors: A source of myelinating transplants. Science 285:754–756.

Colman A, Dreesen O. 2009. Pluripotent stem cells and disease model-ing. Cell Stem Cell 5:244–247.

Cooper B, Chebib M, Shen J, King NJ, Darvey IG, Kuchel PW, Roth-stein JD, Balcar VJ. 1998. Structural selectivity and molecular na-ture of L-glutamate transport in cultured human fibroblasts. ArchBiochem Biophys 353:356–364.

Copray S, Balasubramaniyan V, Levenga J, de BJ, Liem R, Boddeke E.2006. Olig2 overexpression induces the in vitro differentiation of neu-ral stem cells into mature oligodendrocytes. Stem Cells 24:1001–1010.

DeSilva TM, Kabakov AY, Goldhoff PE, Volpe JJ, Rosenberg PA. 2009.Regulation of glutamate transport in developing rat oligodendrocytes.J Neurosci 29:7898–7908.

Einstein O, Fainstein N, Vaknin I, Mizrachi-Kol R, Reihartz E, Grigor-iadis N, Lavon I, Baniyash M, Lassmann H, Ben-Hur T. 2007. Neuralprecursors attenuate autoimmune encephalomyelitis by peripheralimmunosuppression. Ann Neurol 61:209–218.

Filippi M, Bozzali M, Rovaris M, Gonen O, Kesavadas C, Ghezzi A,Martinelli V, Grossman RI, Scotti G, Comi G, Falini A. 2003. Evi-dence for widespread axonal damage at the earliest clinical stage ofmultiple sclerosis. Brain 126:433–437.

Glaser T, Perez-Bouza A, Klein K, Brustle O. 2005. Generation of purified oli-godendrocyte progenitors from embryonic stem cells. FASEB J 19:112–114.

Hu BY, Du ZW, Li XJ, Ayala M, Zhang SC. 2009. Human oligodendro-cytes from embryonic stem cells: Conserved SHH signaling networksand divergent FGF effects. Development 136:1443–1452.

Ishikawa Y, Onoyama I, Nakayama KI, Nakayama K. 2008. Notch-de-pendent cell cycle arrest and apoptosis in mouse embryonic fibro-blasts lacking Fbxw7. Oncogene 27:6164–6174.

KippM, Clarner T, Dang J, Copray S, Beyer C. 2009. The cuprizone animalmodel: New insights into an old story. Acta Neuropathol 118:723–736.

Liu S, Qu Y, Stewart TJ, Howard MJ, Chakrabortty S, Holekamp TF,McDonald JW. 2000. Embryonic stem cells differentiate into oligoden-drocytes and myelinate in culture and after spinal cord transplanta-tion. Proc Natl Acad Sci U S A 97:6126–6131.

Magalon K, Cantarella C, Monti G, Cayre M, Durbec P. 2007. Enrichedenvironment promotes adult neural progenitor cell mobilization inmouse demyelination models. Eur J Neurosci 25:761–771.

Martino G, Franklin RJ, Van Evercooren AB, Kerr DA. 2010. Stem celltransplantation in multiple sclerosis: Current status and future pros-pects. Nat Rev Neurol 6:247–255.

Mason JL, Toews A, Hostettler JD, Morell P, Suzuki K, Goldman JE,Matsushima GK. 2004. Oligodendrocytes and progenitors become pro-gressively depleted within chronically demyelinated lesions. Am JPathol 164:1673–1682.

Matsushima GK, Morell P. 2001. The neurotoxicant, cuprizone, as amodel to study demyelination and remyelination in the centralnervous system. Brain Pathol 11:107–116.

Miller RH. 2002. Regulation of oligodendrocyte development in thevertebrate CNS. Prog Neurobiol 67:451–467.

Nistor GI, Totoiu MO, Haque N, Carpenter MK, Keirstead HS. 2005.Human embryonic stem cells differentiate into oligodendrocytes inhigh purity and myelinate after spinal cord transplantation. Glia49:385–396.

O’Malley J, Woltjen K, Kaji K. 2009. New strategies to generateinduced pluripotent stem cells. Curr Opin Biotechnol 20:516–521.

Onorati M, Camnasio S, Binetti M, Jung CB, Moretti A, Cattaneo E.2010. Neuropotent self-renewing neural stem (NS) cells derived frommouse induced pluripotent stem (iPS) cells. Mol Cell Neurosci43:287–295.

Park IH, Arora N, Huo H, Maherali N, Ahfeldt T, Shimamura A,Lensch MW, Cowan C, Hochedlinger K, Daley GQ. 2008a. Disease-specific induced pluripotent stem cells. Cell 134:877–886.

Park IH, Zhao R, West JA, Yabuuchi A, Huo H, Ince TA, Lerou PH,Lensch MW, Daley GQ. 2008b. Reprogramming of human somaticcells to pluripotency with defined factors. Nature 451:141–146.

Pluchino S, Quattrini A, Brambilla E, Gritti A, Salani G, Dina G,Galli R, Del CU, Amadio S, Bergami A, Furlan R, Comi G, Ves-covi AL, Martino G. 2003. Injection of adult neurospheres inducesrecovery in a chronic model of multiple sclerosis. Nature 422:688–694.

Saha K, Jaenisch R. 2009. Technical challenges in using human inducedpluripotent stem cells to model disease. Cell Stem Cell 5:584–595.

Scheper W, Copray S. 2009. The molecular mechanism of induced pluri-potency: A two-stage switch. Stem Cell Rev 5:204–223.

Sher F, van DG, Boddeke E, Copray S. 2009. Bioluminescence imagingof Olig2-neural stem cells reveals improved engraftment in a demye-lination mouse model. Stem Cells 27:1582–1591.

Stuve O, Bennett JL, Hemmer B, Wiendl H, Racke MK, Bar-Or A, HuW, Zivadinov R, Weber MS, Zamvil SS, Pacheco MF, Menge T, Har-tung HP, Kieseier BC, Frohman EM. 2008. Pharmacological treat-ment of early multiple sclerosis. Drugs 68:73–83.

Takahashi K, Tanabe K, Ohnuki M, Narita M, Ichisaka T, Tomoda K,Yamanaka S. 2007. Induction of pluripotent stem cells from adulthuman fibroblasts by defined factors. Cell 131:861–872.

Takahashi K, Yamanaka S. 2006. Induction of pluripotent stem cellsfrom mouse embryonic and adult fibroblast cultures by definedfactors. Cell 126:663–676.

Tallantyre EC, Bo L, Al-Rawashdeh O, Owens T, Polman CH, Lowe JS,Evangelou N. 2010. Clinico-pathological evidence that axonal lossunderlies disability in progressive multiple sclerosis. Mult Scler16:406–411.

Trapp BD, Peterson J, Ransohoff RM, Rudick R, Mork S, Bo L. 1998.Axonal transection in the lesions of multiple sclerosis. N Engl J Med338:278–285.

Uccelli A, Mancardi G. 2010. Stem cell transplantation in multiple scle-rosis. Curr Opin Neurol 23:218–225.

Warren L, Manos PD, Ahfeldt T, Loh YH, Li H, Lau F, Ebina W, Man-dal PK, Smith ZD, Meissner A, Daley GQ, Brack AS, Collins JJ,Cowan C, Schlaeger TM, Rossi DJ. 2010. Highly efficient reprogram-ming to pluripotency and directed differentiation of human cells withsynthetic modified mRNA. Cell Stem Cell 7:618–630.

Woodruff RH, Tekki-Kessaris N, Stiles CD, Rowitch DH, RichardsonWD. 2001. Oligodendrocyte development in the spinal cord and telen-cephalon: Common themes and new perspectives. Int J Dev Neurosci19:379–385.

Yamanaka S. 2008. Induction of pluripotent stem cells from mousefibroblasts by four transcription factors. Cell Prolif 41(Suppl 1):51–56.

Yin L, Velazquez OC, Liu ZJ. 2010. Notch signaling: Emerging molecu-lar targets for cancer therapy. Biochem Pharmacol 80(5):690–701.

Yu J, Vodyanik MA, Smuga-Otto K, ntosiewicz-Bourget J, Frane JL,Tian S, Nie J, Jonsdottir GA, Ruotti V, Stewart R, Slukvin II,Thomson JA. 2007. Induced pluripotent stem cell lines derived fromhuman somatic cells. Science 318:1917–1920.

Yuan JS, Kousis PC, Suliman S, Visan I, Guidos CJ. 2010. Functions ofnotch signaling in the immune system: Consensus and controversies.Annu Rev Immunol 28:343–365.

892 CZEPIEL ET AL.

GLIA