411

Cell Transplantation, Vol. 25, pp. 411–424, 2016 0963-6897/16 $90.00 + .00Printed in the USA. All rights reserved. DOI: http://dx.doi.org/10.3727/096368915X688137Copyright Ó 2016 Cognizant, LLC. E-ISSN 1555-3892 www.cognizantcommunication.com

Received February 17, 2015; final acceptance May 6, 2015. Online prepub date: May 7, 2015.1These authors provided equal contribution to this work.2These authors provided equal contribution to this work.Address correspondence to Ann M. Parr, M.D., Ph.D., Assistant Professor, Department of Neurosurgery, University of Minnesota, D427 Mayo Bldg., MMC96, 420 Delaware St. SE, Minneapolis, MN 55455, USA. Tel: +1-612-624-6666; Fax: +1-612-624-0644; E-mail: amparr@umn.edu

Directed Differentiation of Oligodendrocyte Progenitor Cells From Mouse Induced Pluripotent Stem Cells

Dino Terzic,*†1 Jacob R. Maxon,*1 Leah Krevitt,* Christina DiBartolomeo,* Tarini Goyal,* Walter C. Low,*† James R. Dutton,*2 and Ann M. Parr*†2

*Stem Cell Institute, University of Minnesota, Minneapolis, MN, USA†Department of Neurosurgery, University of Minnesota, Minneapolis, MN, USA

Several neurological disorders, such as multiple sclerosis, the leukodystrophies, and traumatic injury, result in loss of myelin in the central nervous system (CNS). These disorders may benefit from cell-based therapies that prevent further demyelination or are able to restore lost myelin. One potential therapeutic strategy for these disorders is the manufacture of oligodendrocyte progenitor cells (OPCs) by the directed differentiation of pluri-potent stem cells, including induced pluripotent stem cells (iPSCs). It has been proposed that OPCs could be transplanted into demyelinated or dysmyelinated regions of the CNS, where they would migrate to the area of injury before terminally differentiating into myelinating oligodendrocytes. OPCs derived from mouse iPSCs are particularly useful for modeling this therapeutic approach and for studying the biology of oligodendrocyte progenitors because of the availability of mouse models of neurological disorders associated with myelin defi-ciency. Moreover, the utility of miPSC-derived OPCs would be significantly enhanced by the adoption of a consistent, reproducible differentiation protocol that allows OPCs derived from different cell lines to be robustly characterized and compared. Here we describe a standardized, defined protocol that reliably directs the differen-tiation of miPSCs to generate high yields of OPCs that are capable of maturing into oligodendrocytes.

Key words: Oligodendrocyte progenitor cells (OPCs); Induced pluripotent stem cells (iPSCs); Differentiation; Transplantation; Myelin

INTRODUCTION

A number of neurodegenerative disorders, includ-ing multiple sclerosis, leukodystrophies, and injury to the central nervous system (CNS), lead to myelin loss in the CNS and currently lack any definitive treatment options (2,12,13,16,19). Transplanting progenitor cells that will become oligodendrocytes capable of remyelina-tion into the regions of demyelination has been indicated as a viable treatment option for restoring functionality in these disorders.

Oligodendrocytes perform crucial structural and functional roles in the CNS, producing lipid-rich myelin sheaths that segmentally envelop axons. Myelin provides insulation and organizes the distribution of axonal volt-age-gated ion channels, enabling efficient conduction of signals throughout the neuraxis (11). Oligodendrocytes also aid in maintaining axonal stability and play a protec-tive role during inflammatory events. During embryogen-esis, oligodendrocytes are generated from neuroepithelial

precursor cells known as oligodendrocyte progenitor cells (OPCs). In both rodents and humans, two sepa-rate progenitor cell populations in the developing fore-brain contribute to the OPCs that differentiate to form all the oligodendrocytes in the adult brain [reviewed in (3,22,26)]. In the developing spinal cord, the majority of the OPCs are generated in the ventral pre-motoneu-ron (pMN) domain. Cells in this domain initially gener-ate motoneurons before a neurogenic/gliogenic switch occurs, and the same domain then produces OPCs that migrate to populate the entire spinal cord. Later in devel-opment, a more dorsal region of the spinal cord also pro-duces OPCs that contribute ~10% of the oligodendrocytes in the cord (4,5,15). Some OPCs are considered to persist into adulthood and may aid in oligodendrocyte turnover and minor injury repair. Regardless of origin, embryonic OPC populations migrate extensively before differentiat-ing to generate functional, myelinating oligodendrocytes (40). These characteristics make OPCs ideal candidates

412 TERzIC ET AL.

for cell therapy strategies targeting missing, defective, or damaged oligodendrocytes to restore myelin lost due to trauma or disease, as the cells can be manufactured and transplanted as precursor cells that will then migrate to the site of action.

In vitro directed derivation of OPCs has been described from a variety of stem cell populations, including mouse embryonic stem cells (ESCs) (18), mouse induced pluri-potent stem cells (iPSCs) (7,23,39), mouse epiblast cells (24), human ESCs (17,19,20,25,32,35,36), and human iPSCs (10,41). Although these protocols aim to gener-ate populations of OPCs with similar properties, the differentiation strategies employed are diverse, and the quality and quantity of the OPCs generated is variable. Furthermore, the reproducibility of many of these proto-cols remains to be established (1).

Advances in cellular reprogramming now allow iPSCs to be generated from somatic cells by the tran-sient forced expression of defined transcription factors (28,29,37,38). iPSCs can offer several advantages over other stem cell types as a source for generating OPCs for potential clinical use or research. In addition to generat-ing an unlimited supply of pluripotent stem cells, this technology enables the creation of patient-specific stem cells, allowing the potential for autologous transplanta-tion of differentiated cells. This would avoid complica-tions from host rejection of the transplanted cells and the expense and side effects of antirejection immune drug therapies. iPSCs can also be derived from patients with genetic disorders affecting oligodendrocytes or myelina-tion, or from well-characterized transgenic mouse mod-els, providing valuable models of disease and tools for drug screening.

OPCs derived from mouse iPSCs provide a power-ful resource for researchers. Using iPSC intermediates, OPCs can be generated from various natural and trans-genic animal models of dysmyelination and CNS disease and injury. Additionally, iPSC technology enables use of various reporter mouse strains to label transplanted cells or to monitor the differentiation and maturation of OPCs both in vivo and in vitro. However, the utility of OPCs derived by the directed differentiation of pluripotent ESCs and iPSCs is greatly reduced by the wide variation in the final cell populations generated from diverse pro-tocols in different laboratories. Here we report a repro-ducible differentiation protocol that yields a high-purity population of OPCs from mouse iPSC lines that can ter-minally differentiate into oligodendrocytes both in vitro and in vivo.

MATERIALS AND METHODS

Mouse Induced Pluripotent Stem Cells (miPSCs)

The miPSC lines used in this study had been gener-ated previously in our laboratory essentially as described

in Greder et al. (14). Lines UMN-3F10 and UMN-JBL6 were derived using three reprogramming factors octamer-binding transcription factor-4 (Oct4), sex-determining region y-box 2 (Sox2), and Kruppel-like factor-4 (Klf4), and derivation of UMN-JG2 utilized c-Myc in addition to these factors. All three lines were derived from primary cultures of E13.5 mouse embryonic fibroblasts, UMN-3F10 and UMN-JG2 from Oct4:CreER mTmG transgenic mouse (14) and UMN-JBL6 from C57BL/6J mice (Jackson Laboratories, Bar Harbor, ME, USA). The miPSC lines were previously characterized as pluripotent using immu-nohistochemistry and qRT-PCR for pluripotency factors; teratoma analysis; differentiation in embryoid bodies; and, for UMN-3F10, methylation analysis. The miPSC lines were maintained as previously described (14,30). The miPSCs were cultured on irradiated mouse embryonic fibroblasts in miPSC medium: knockout DMEM with 4.5 g/ L d-glucose and sodium pyruvate (Gibco, Grand Island, NY, USA), 10% knockout serum replacement (Gibco), 10% fetal bovine serum (HyClone, Logan, UT, USA), 1× MEM nonessential amino acids (Gibco), 1× GlutaMAX (Gibco), 0.1 mM 2-mercaptoethanol (Life Technologies, Grand Island, NY, USA), 1× penicillin/streptomycin (Corning Cellgro, Manassas, VA, USA), and ESGRO-LIF (1,000 U/ml, Millipore, Billerica, MA, USA). UMN-3F10 cells were constitutively labeled with transmembrane GFP by supplementing their miPSC medium with 10 nM tamoxifen (Sigma-Aldrich, St. Louis, MO, USA). Cells were incubated at 37°C in 5% CO

2.

Differentiation of miPSCs to Oligodendrocyte Progenitor Cells (OPCs)

Undifferentiated miPSCs were dissociated from the irradiated mouse embryonic fibroblasts using TrypLE Express Enzyme (1×, no phenol red) (Life Technologies), washed, and resuspended in miPSC medium. The irradi-ated mouse embryonic fibroblasts were partially depleted from resuspended miPSCs by incubation in a T-75 flask (CytoOne, Orlando, FL, USA) for 30 min. The miPSCs were counted and resuspended in KSR medium [minimal essential media (Corning Cellgro), 20% knockout serum replacement, 1 mM sodium pyruvate (Lonza, Basel, Switzerland), 1× MEM nonessential amino acids, 0.1 mM 2-mercaptoethanol, and 1× penicillin/streptomycin]. miPSCs were aliquoted into ultra-low attachment surface six-well plates (Corning) at a density of 500,000 miPSCs/well in 3 ml of KSR medium/well to form cell aggregates overnight. The day that the miPSCs were resuspended for aggregation was denoted day 0 (d0).

Throughout all stages of differentiation, cells were incubated in a 37°C and 5% CO

2 environment. On days 1

to 5, the aggregates were cultured in KSR medium with daily media changes. On day 4, retinoic acid (Sigma-Aldrich) was added to the KSR medium, with a final

DIFFERENTIATION OF OPCs FROM miPSCs 413

concentration 0.2 mM. On day 5, both retinoic acid and purmorphamine (Cayman Chemical, Ann Arbor, MI, USA) were added to the KSR medium, with final con-centrations of 0.2 mM and 1 mM, respectively.

On days 6 and 7, the aggregates were cultured in N-2 medium, with daily medium changes. N-2 medium consists of 1× minimum essential media, 1× N-2 supplement (Gibco), 1 mM sodium pyruvate, 1× MEM nonessential amino acids, 55 mM 2-mercaptoethanol, 1× penicillin/streptomycin, 0.2 mM retinoic acid, and 1 mM purmorphamine.

Importantly, at each media change from days 1 to 7, the cell aggregates were centrifuged at 179 × g for 1 min and gently resuspended in 3 ml of medium per well of aggregates in order to maintain aggregate morphology and to prevent clumping between aggregates.

On day 8, the aggregates were plated onto flat-bot-tom six-well plates (CytoOne) coated with poly-l-or-nithine and fibronectin. To coat the plates, the six-well plates were incubated at 37°C in 5% CO

2 with 1 ml/

well of 0.01% poly-l-ornithine (Sigma-Aldrich) over-night. On the day of aggregate plating, the poly-l-or-nithine was removed, and the wells were washed twice with sterile cell culture-grade water (Corning Cellgro). A fibronectin (Sigma-Aldrich) coating was then applied to each well by quickly applying and removing 1 ml of 50 mg/ml fibronectin, diluted with sterile 1× phosphate-buffered saline (PBS) without calcium and magnesium (Corning Cellgro). To plate the aggregates, they were resuspended in OPC expansion medium at a density of 150 to 300 aggregates/ml of OPC expansion medium. The aggregates were then distributed onto the poly-l- ornithine/fibronectin-coated six-well plates at a density of 1 ml of aggregate suspension/well. An additional 1 ml of OPC expansion medium was then added to each well. Last, the six-well plates containing the aggregates were centrifuged at 100 × g for 1 min. The OPC expansion medium consisted of DMEM/F12 without phenol red or HEPES, with l-glutamine (Hyclone), 1× B-27 supple-ment, serum free (Life Technologies), 1× N-2 supplement (Life Technologies), 20 ng/ml fibroblast growth factor-2 (FGF-2) (R&D Systems, Minneapolis, MN, USA), 20 ng/ ml rhPDGF-AA (Sigma-Aldrich), 100 ng/ml rhShh (R&D Systems), and 1× penicillin/streptomycin. After plating, medium was changed every other day. The OPCs were grown to confluence and passaged at a 1:2 ratio using Accutase (Millipore).

Terminal Differentiation of OPCs to Mature Oligodendrocytes

For terminal differentiation, six-well plates were coated with a substrate of 50 mg/ml poly-l-ornithine hydrobromide (Sigma-Aldrich) and 20 mg/ml laminin (Life Technologies). The plates were first coated with 1 ml poly-l-ornithine hyd-robromide/well and incubated at room temperature for 1 h.

Next, the cells were coated with 1 ml laminin/well and incu-bated at room temperature for 1 h. After passage, cells were then cultured in OPC expansion medium for 1 week.

Confluent OPCs were passaged at a density of 1,000 cells per well and cultured in OPC expansion medium for 1 week, with medium changes every other day. The cells were then cultured in oligodendrocyte terminal dif-ferentiation medium [adapted from Wang et al. (41)], with medium changes every other day. This medium consisted of DMEM/F12 without phenol red or HEPES, 0.25× N1 medium supplement (Sigma-Aldrich), 0.5× N-2 supple-ment, 1.25× B-27 supplement, serum free, 30 ng/ml 3,3¢,5-triiodo-l-thyronine sodium salt (Sigma-Aldrich), 50 ng/ml biotin (Sigma-Aldrich), 0.5 mM dibutyryl cAMP (Sigma-Aldrich), 2.5 ng/ml recombinant human platelet- derived growth factor (PDGF-AA) (Peprotech, Rocky Hill, NJ, USA), 5 ng recombinant human BDNF (Peprotech), 2.5 ng/ml recombinant human insulin-like growth factor-1 (IGF-1) (Peprotech), 2.5 ng/ml NT3 (Peprotech), 0.5× penicillin/streptomycin, and 27.5 mM 2-mercaptoethanol.

Immunocytochemistry

Cultured cells were fixed using formalin (10%; Fisher Scientific, Pittsburgh, PA, USA) for 20 min at room temperature and then washed three times with PBS. When staining for intracellular markers, cells were permeabilized using 1% Tween 20 (Calbiochem, Billerica, MA, USA) in PBS for 5 min. Next, the cells were incubated with blocking buffer [1% bovine serum albumin (BSA) (Sigma-Aldrich) in PBS-T (PBS with 0.1% Tween 20)] for 30 min. Cells were incubated over-night at 4°C with primary antibodies diluted in block-ing buffer. After washing three times in PBS-T, the cells were then incubated with secondary antibodies diluted in blocking buffer for 2 h at room temperature before final washing. For surface antigen detection, the perme-abilization step was excluded, and antigen retrieval was performed (L.A.B. Solution, 24310; Polysciences Inc., Warrington, PA, USA) according to the manufacturer’s instructions before incubation with primary antibody, and Tween 20 was omitted from blocking and washing solutions. Nuclear contrast was performed with Hoechst nuclear stain (diluted 1:1,000 in PBS) (Thermo Scientific, Waltham, MA, USA) for 5 min. The same procedures performed without primary antibodies were used as con-trols. The primary and secondary antibodies used in this study are listed in Table 1.

Fluorescence Microscopy

Direct fluorescence imaging of live and fixed cells was performed using a DMI 6000B inverted microscope (Leica, Wetzlar, Germany). Images were captured using a Retiga 2000R camera (QImaging, Surrey, BC, Canada) using IP Lab software (BD Biosciences, San Jose, CA,

414 TERzIC ET AL.

USA). eGFP and tdTomato expression was detected directly in live and formalin-fixed cells.

Immunocytochemistry Analysis

For OPC differentiations from each miPS cell line, the percentage of cells expressing a cell marker was deter-mined by capturing images of at least five random fields of immunostained cells at 200× magnification. The num-ber of cells demonstrating positive immunostaining in each field was counted and divided by the total number of cells in that field (determined by Hoechst staining) to determine the percentage of cells positive for a specific cell marker. The average of the five fields was calculated,

and counting was repeated for at least three separate dif-ferentiations for each cell line. All cell counting was done using ImageJ (NIH, Bethesda, MD, USA).

Total RNA Extraction and Analysis

Total RNA was prepared from aggregates and cultured cells using the RNeasy MiniKit (Qiagen, Venlo, Limburg). cDNA was generated by reverse transcription from 500 ng total RNA using SuperScript III Reverse Transcriptase and oligo(dT)20 primers (Invitrogen, Waltham, MA, USA). Quantitative gene expression was analyzed in an Eppendorf LightCycler using predesigned primers (IDT). See Table 2 for a list of primers used. Samples from at

Table 1. Primary and Secondary Antibodies

Vendor Catalog Number Host Species Isotype Dilution

Primary antibodyA2B5 Millipore MAB312 Mouse IgM 1:250GFAP Millipore MAB360 Mouse IgG1 1:400Hoechst 33342 Thermo Scientific 62249 N/A N/A 1:1,000MBP Abcam ab7349 Rat IgG2a 1:250NG2 Millipore AB532 Rabbit IgG 1:100Olig2 Millipore AB9610 Rabbit N/A 1:250PDGFRa Santa Cruz

Biotechnologysc-338 Rabbit IgG 1:400

b-tubulin III Sigma-Aldrich T8660 Mouse IgG2b 1:400Secondary antibody

Alexa Fluor 555 F(ab¢)2 fragment of goat anti-mouse

Invitrogen A-21425 Goat IgG (H + L) 1:1,000

Alexa Fluor 555 donkey anti-rabbit

Invitrogen A-31572 Donkey IgG (H + L) 1:1,000

Alexa Fluor 555 goat anti-mouse Invitrogen A-21426 Goat IgM (μ chain) 1:1,000Alexa Fluor 633 goat anti-mouse Invitrogen A-21046 Goat IgM (μ chain) 1:1,000

Table 2. qRT-PCR Primers

Target Gene Vendor Assay Name RefSeqNumber

APC IDT Mm.PT.51.6788653 NM_007462CNP IDT Mm.PT.53a.8652225 NM_009923GAPDH IDT Mm.PT.39a.1 NM_008084GFAP IDT Mm.PT.56a.6609337.g NM_001131020MBP IDT Mm.PT.51.15932524.g NM_001025245Nanog IDT Mm.PT.56a.23510265 NM_028016Nestin IDT Mm.PT.56a.6955776 NM_016701(1)Nkx2.2 IDT Mm.PT.56a.43580979 NM_001077632Oct4 (Pou5f1) IDT Mm.PT.51.7439100.g NM_013633(2)Olig1 IDT Mm.PT.51.8783870.g NM_016968Olig2 IDT Mm.PT.51.6174618 NM_016967Pax6 IDT Mm.PT.56a.32331261.g NM_001244201PDGFR-a IDT Mm.PT.51.17026963 NM_011058Sox10 IDT Mm.PT.56a.913114 NM_011437Syn1 IDT Mm.PT.51.13184129 NM_013680Tubb3 IDT Mm.PT.58.32393592 NM_023279(1)

DIFFERENTIATION OF OPCs FROM miPSCs 415

least three differentiation replicates were obtained from each cell line, and results were normalized with respect to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) expression. Standard error was indicated with error bars.

Cell Transplantation

Homozygous shiverer mice, strain C3Fe.SWV-Mbpshi/J (6), were obtained from Jackson Laboratories. Mice were anesthetized with ketamine–xylazine (Phoenix, St. Joseph, MO, USA; AnaSed Injection, Shenandoah, IA, USA), and a burr hole was used as the access point for the

stereotactic injection of cells. Cells were injected 0.4 mm anterior and 2.0 mm lateral to bregma, into the corpus callosum and striatum (depths of 1.5 and 2.0 mm, respec-tively). The injected cells were at day 25 in culture at the time of injection. Injections consisted of 10 μl of cell sus-pension, containing 25,000 cells/μl of DMEM/F12 with-out HEPES or phenol red, with l-glutamine. The animals were sacrificed at 14 days and 28 days postinjection. All animal protocols and experiments were approved by the Institutional Animal Care and Use Committee of the University of Minnesota.

Figure 1. Directed differentiation protocol for generating oligodendrocyte progenitor cells from pluripotent mouse iPSCs. Mouse iPSCs from the UMN-3F10 line were subjected to our protocol to direct the generation of Olig2-expressing OPCs. Stage 1 undifferentiated UMN-3F10 mouse iPSCs; Stage 2 neural aggregates; Stage 3 (day 8) early OPCs; Stage 4 OPC expansion (scale bars: 100 μm).

416 TERzIC ET AL.

Immunohistochemical Analysis of Transplanted Cells

Transplanted mice were perfused transcardially with 4% paraformaldehyde (Fisher Scientific). The brains were removed and placed in 4% paraformaldehyde at 4°C for 1 day. They were then stored in 30% sucrose at 4°C until embedding in OCT (Sakura Finetek U.S.A., Torrance, CA, USA) and freezing at −80°C. The brains were sec-tioned using a Leica CM 3050S cryostat at a thickness of 10 mm per section. Transplanted cells were identified by eGFP expression under fluorescence microscopy.

RESULTS

Generation of OPCs From miPSCs

Our protocol directs the differentiation of miPSCs into a population of Olig2- and PDGFRa-positive cells that can be terminally differentiated into MBP-positive cells. Figure 1 illustrates the progression of the differ-entiation protocol using miPSC line UMN-3F10, which expresses transmembrane-bound eGFP (14). Neural induc-tion is conducted following aggregation of miPSCs in

serum-free media (Fig. 1C, D) to produce cells express-ing Pax6. Patterning of the neural population mim-icking a ventral spinal cord phenotype is initiated by sequential addition of retinoic acid and then retinoic acid with purmorphamine, a small molecule agonist of the smoothened receptor that activates the hedgehog-signaling pathway (34). The induction is completed in N-2 medium. On day 8, aggregates are plated onto poly-l-ornithine- and fibronectin-coated dishes and cultured in OPC medium. Within 2 days, bipolar cells, expressing the nuclear located Olig2 transcription fac-tor, migrate from the plated aggregates (Fig. 1E, F) and are expanded in serum-free media containing FGF2, PDGFAA, and Shh.

Quantitative Gene Expression Analysis During OPC Differentiation

Using miPSC line UMN-3F10, we assessed the sequential expression of key genes at discrete stages of the OPC differentiation protocol using quantitative

Figure 2. Quantitative analysis of gene expression during directed differentiation of mouse iPSCs into OPCs. Mouse iPSC line UMN-3F10 was directed to generate OPCs, and mRNA expression was analyzed at successive time points. Gene expression of key genes on specific days of the differentiation protocol was normalized to GAPDH.

DIFFERENTIATION OF OPCs FROM miPSCs 417

Figure 3. Immunocytochemical analysis of multiple iPSC lines differentiated to OPCs. Three mouse iPSC lines were subjected to our protocol, and immunocytochemistry was used to characterize the cells at d25. (A) UMN-3F10 cells demonstrating morphology and phenotype. (B) UMN-3F10 eGFP and Olig2. (C) UMN-JBL6 Olig2 (green) Hoechst (blue). (D) UMN-JG2 Olig1 (red) Hoechst (blue). (E) UMN-JG2 A2B5 (blue) Olig2 (red). (F) UMN-3F10 A2B5 (red) Hoechst (blue). (G) UMN-JG2 PDGFR (red) Hoechst (blue). (H) UMN-3F10 NG2 (red) Hoechst (blue).

418 TERzIC ET AL.

RT-PCR (Fig. 2). In the early stages of the differentiation protocol, the expression of the pluripotency gene Nanog decreases. Pax6 expression is rapidly upregulated fol-lowing cell aggregation and is detected by day 4 and increases to maximal expression on d20. The upregula-tion of Pax6 expression is followed by the expression of Nestin, another gene associated with early neuronal identity. The expression of genes linked to early OPC identity, Olig2, Nkx2.2, PDGFRa, and Olig1, was first measured at the end of the neural patterning stage, and the level of expression of these genes continued to increase during the expansion of the OPC population. The expression of the genes characteristic of mature oligodendrocytes, such as myelin basic protein (MBP) and adenomatous polyposis coli (APC) (not shown), was minimal throughout the protocol. Minimal expres-sion of Tubb3 (b-tubulin III), expressed in immature and mature neurons was detected; expression of Synapsin1, a gene associated with neuronal synaptic vesicles, was not detected (data not shown).

Immunocytochemical Analysis of miPSC-Derived OPC Cultures

Together with gene expression analysis, immuno-cytochemistry was used to further characterize the pop-ulation of cells present on d25 (Fig. 3). To assess the reproducibility of our protocol, it was applied to three separate miPSC lines derived previously in our labo-ratory. Cells in the proliferative OPC population at d25 were typically of bi- or tripolar morphology (Fig. 3A). The transcription factor Olig2 was seen to be located in the nucleus of the d25 OPCs (Fig. 3B, C) (88.0 ± 1.3% UMN-3F10, 82.0 ± 6.1% UMN-JBL6, 47.9 ± 2.1% UMN-JG2) and was widely coexpressed in cells pos-sessing the antigen A2B5, a cell surface protein epitope characteristic of glial progenitor cells (75.7 ± 3.8% UMN-3F10, 73.0 ± 2.0% UMN-JBL6, 25.7 ± 2.0% UMN-JG2) (Fig. 3E). We were also able to detect widespread expres-sion of the cell surface antigens PDGFRa and NG2, which are more restricted to proliferating oligodendro-cyte progenitor cells than the glial progenitor marker

Figure 4. Neurons and glia represent a minority cell population in d25 cultures of miPSCs subjected to this protocol. (A, B) UMN-3F10- and UMN-JG2-derived OPCs were analyzed for the presence of neural cells expressing b-tubulin III (red). (C, D) UMN-3F10- and UMN JG2-derived cells were analyzed for GFAP-expressing cells (red) (scale bars: 10 μm).

DIFFERENTIATION OF OPCs FROM miPSCs 419

A2B5 (Fig. 3G, H). PDGFRa was present on 72.7 ± 3.7% of UMN-3F10-derived cells, 71.3 ± 1.9% of UMN-JBL6-derived cells, and 33.6 ± 1.8% of UMN-JG2-derived cells. Expression of the transcription factor Olig1 was found to be cytoplasmic in the d25 cell population (Fig. 3D).

Common contaminating cell types identified in pro-tocols reporting the differentiation of OPCs from ESCs and iPSCs are neuronal cells expressing b-tubulin III and astrocytes identified by the expression of glial fibrillary acidic protein (GFAP). We assessed if these cells were present in our d25 OPC cultures (Fig. 4). Few b-tubulin III-expressing cells were seen in OPC cultures derived from our miPSC lines (5.0 ± 2.5% UMN-3F10, 6.0 ± 1.6% UMN-JBL6, 3.3 ± 0.7% UMN-JG2). GFAP-expressing cells were also found to be present in these cultures at low frequencies (9.3 ± 3.1% UMN-3F10, 6.3 ± 2.4% UMN-JBL6, 12.0 ± 1.5% UMN-JG2). Neither Nanog nor Oct4 protein expression was detected in the d25 OPC cul-tures (not shown), indicating absence of pluripotent cells in the culture. Quantification of the immunohistochem-istry of cells generated at d25 of our protocol is shown in Figure 5.

In Vitro Terminal Differentiation of OPCs Into Oligodendrocytes Expressing MBP

Following expansion in OPC expansion medium, OPCs derived from UMN-3F10 and UMN-JBL6 cell lines were subjected to culture conditions shown previously to induce differentiation of human iPSC-derived OPCs into

oligodendrocytes in vitro (39) (Fig. 6). Oligodendrocyte cells expressing myelin basic protein were detected in cultures differentiated from both cell lines after 7 and 21 days in terminal differentiation medium (Fig. 6A–D). After 3 weeks of culture in oligodendrocyte terminal dif-ferentiation medium the iPSC-derived OPCs displayed extensively arborized and ramified phenotypes typical of in vitro differentiated oligodendrocytes (Fig. 6E, F).

Transplantation of miPSC-Derived OPC Cultures Into Adult Dysmyelinated Mice

We also examined the ability of the OPCs generated from our differentiation protocol to survive, migrate from the injection site, and mature into MBP-positive oligo-dendrocytes following transplantation into an adult ani-mal model of congenital dysmyelination. We used OPCs generated from UMN-3F10 iPSCs because they express high levels of transmembrane eGFP and are easily visi-ble in formalin-fixed cryosections of transplanted tissue without additional processing. The corpus callosum and striatum of the adult, homozygous shiverer mice were injected with 250,000 eGFP-expressing OPCs (Fig. 7). eGFP-expressing cells were found to be present in the brain tissue close to the injection sites 14 days after transplantation (Fig. 7A). In contrast, analysis of brains 28 days posttransplantation showed extensive migra-tion of eGFP-expressing cells throughout the corpus callosum and striatum, and eGFP-expressing cell pro-cesses were detected in brain tissue over 1.3 mm distant

Figure 5. Quantitative analysis from immunocytochemistry of d25 OPC populations differentiated from three separate mouse iPSC lines (UMN-3F10, UMN-JBL6, and UMN-JG2). The percent of cells expressing each antigen in cell populations of OPCs derived from each iPSC line was determined using ImageJ.

420 TERzIC ET AL.

from the position of cell injection (Fig. 7B). Multiple phenotypes of eGFP-expressing cells were observed in the brain 28 days after transplantation. Many cells were seen with single projections extending through the brain tissue, and other eGFP-expressing cells that had numerous multiple, shorter projections were also observed. Coexpression of neurofilament and eGFP was not detected (Fig. 7B, C). However, high-density areas consisting of multiple eGFP-expressing cells were observed near the neuronal tracts (Fig. 7C). Expression of MBP protein was not detected in transplanted cells in the brain tissue 14 days postinjection. However, 28 days following injection of UMN-3F10-derived OPCs, areas

of transplanted cells coexpressing MBP and eGFP were observed (Fig. 8A–C).

DISCUSSION

In this study, we sought to differentiate miPSCs into OPCs using a defined and reproducible directed differ-entiation protocol. These OPCs are capable of terminal differentiation to MBP-expressing oligodendrocytes. We focused on characterizing miPSCs as they differentiated into OPCs and then subsequently examined their termi-nal differentiation capacity. We focused on OPCs rather than fully mature oligodendrocytes so that our protocol is more relevant for models that propose transplanting OPCs

Figure 6. In vitro terminal differentiation of OPCs to MBP-positive oligodendrocytes. After passaging the UMN-3F10- and UMN-JBL6-derived OPCs to low-density conditions and expanding for 1 week in OPC expansion medium, the OPCs were cultured in oli-godendrocyte terminal differentiation medium for up to 3 weeks. Oligodendrocytes expressing myelin basic protein (MBP) (red) were observed in cultures from both cell lines after culturing in terminal differentiation medium for 1 week (A, B) and 3 weeks (C, D). After 3 weeks in terminal differentiation medium, the cells exhibited extensive ramified processes characteristic of oligodendrocytes (E, F).

DIFFERENTIATION OF OPCs FROM miPSCs 421

and not terminally differentiated oligodendrocytes as the therapeutic cell type (12). iPSC technology now allows pluripotent stem cells to be generated from virtually any mouse strain and enables the possibility of using a wide variety of transgenic and nontransgenic mouse models to study OPC/oligodendrocyte biology and disease. The utility of OPCs derived from iPSCs for any of these pur-poses is aided by reproducible, efficient differentiation protocols and outcome measurements. We determined the identity of the cells derived from our differentiation

protocol by analyzing a panel of established OPC mark-ers using qRT-PCR, immunocytochemistry, and testing in vitro and in vivo terminal differentiation.

Quantitative qRT-PCR analysis characterized gene expression changes during the conversion of the pluripo-tent iPSCs through a patterned neural precursor stage to a population of proliferating cells with the characteristic phenotype of OPCs. Initially, expression of the key pluri-potency gene Nanog decreased, while genes associated with neuroepithelium (Pax6, Nestin) and OPC identity

Figure 7. Distribution of iPSC-derived OPCs following transplantation into adult shiverer mouse brains. OPCs derived from UMN-3F10 miPSCs and cultured for 25 days were injected into the brains of adult shiverer mice, and 14 and 28 days after injection, transplanted brains were fixed, harvested, and sectioned. The distribution of transplanted OPCs or their progeny was analyzed in 10-μm sections by identifying cells expressing transmembrane-located eGFP. (A) Fourteen days after transplantation, transplanted OPCs are located near the injection site. (The inset image is a cartoon of the shiverer mouse brain, with the red rectangle representing the approximate location of injected cells.) (B) Twenty-eight days after injection, eGFP-expressing cells are present in the striatum, corpus callosum, and cortex sur-rounding the injection site. (The cell position can be identified from the inset in A by the presence of the edge of the ventricle.) (C) Twenty-eight days after injection, eGFP-expressing OPCs exhibit multiple phenotypes and are located near axon tracts (expressing NF200 in red) in the striatum of injected shiverer mouse brains. C, cortex; CC, corpus callosum; S, striatum. Scale bars: 250 μm (A, B), 100 μm (C).

422 TERzIC ET AL.

(Olig2, PDGFRa, Nkx2.2) increased (Fig. 2). Our immu-nocytochemistry data supported these findings, and after 25 days of differentiation from iPSCs, the populations from the directed differentiation of the UMN-3F10 and UMN-JBL6 lines were greater than 70% positive for Olig2 and A2B5 protein expression (Figs. 3, 5). Coexpression of these antigens has been used to indicate OPC iden-tity (27) and indicates that our protocol produces a high-yield population of OPCs. Cell populations differentiated from the UMN-3F10 and UMN-JBL6 iPSC lines were

also over 70% positive for PDGFRa expression, which has been previously used to enrich OPCs from mixed cell populations (33,41). The UMN-JG2 line resulted in a significantly lower yield of OPCs, with only 20% of the cell population derived from this line coexpressing A2B5 and Olig2 but over 30% of the cell population express-ing PDGFRa at d25. These relatively modest results compared to cell populations differentiated from UMN-3F10 and UMN-JBL6 iPSCs may result from differences in differentiation capacity inherent in some iPSC lines. However, the results still confirm that our protocol repro-ducibly creates enriched populations of OPCs in multiple miPSC lines. The immunocytochemistry data from the d25 OPC populations also demonstrated that our differ-entiation method results in an OPC population contain-ing very few contaminating neurons or GFAP-expressing astrocytes. This indicates that very low numbers of mul-tipotential neural or glial progenitors persist in the final OPC population that may confound their use for model-ing potential clinical applications.

Together, these data indicate that our protocol is more efficient than protocols previously reporting the deriva-tion of an enriched population of OPCs from miPSCs. In 2010, Tokumoto et al. (39) reported differentiating OPCs from miPSCs; however, their protocol resulted in a cell population that contained only 14.1% OPCs identified solely using A2B5 expression. We utilized other mark-ers in addition to A2B5 to identify OPCs, since the A2B5 gene is also expressed in bipotential progenitor cells able to differentiate into astrocytes or oligodendrocytes (31). Misumi et al. report a protocol that results in a cell population with <50% PDGFRa-expressing OPCs from differentiated miPSCs requiring up to 38 days of culture compared to the 25 days required for our protocol (23).

Testing the terminal differentiation capacity of stem cell-derived OPCs is important to demonstrate their abil-ity to generate mature oligodendrocytes, and we tested our cells using two separate terminal differentiation pro-tocols. In vitro terminal differentiation was used to dem-onstrate that the cells change from the typical bi- and tripolar shape of proliferating OPCs and adopt a highly arborized phenotype. Cells with a highly branched mor-phology typical of adherent in vitro grown oligodendro-cytes were observed in cultures of OPCs differentiated from UMN-3F10 and UMN-JBL6 miPSCs with some cells expressing MBP.

Terminal differentiation testing by in vivo transplanta-tion is another way to demonstrate the capacity of OPC cell populations to generate MBP-expressing oligodendro-cytes. To examine the ability of our iPSC-derived OPCs to mimic the behavior of endogenous OPCs in an in vivo system, we injected cells differentiated from eGFP-posi-tive UMN-3F10 iPSCs into the shiverer mouse brain. The shiverer mouse is a model of congenital dysmyelination

Figure 8. In vivo terminal differentiation of OPCs into MBP-positive oligodendrocytes 28 days after transplantation into adult shiverer mouse brain. OPCs derived from UMN-3F10 miPSCs were transplanted into adult shiverer brains. Twenty-eight days following transplantation, the brains were analyzed to detect myelin basic protein (MBP) expression (red) associ-ated with eGFP-expressing cells. (A) eGFP and MBP coexpres-sion is seen in multiple sites in the areas where transplanted cells are present 28 days after injection. (B, C) MBP expression is colocalized with fine projections of cells expressing eGFP.

DIFFERENTIATION OF OPCs FROM miPSCs 423

caused by a deletion in the MBP gene (6), and the homozy-gous genotype lacks myelin basic protein. After injection of UMN-3F10 iPSC-derived OPCs, we found that many transplanted cells survived and migrated away from the site of injection. There was no evidence of neuronal dif-ferentiation of our OPCs and eGFP-expressing oligoden-drocytes derived from the transplanted cells coexpressing MBP were observed in brain sections collected 28 days after transplantation. The observation of OPC migration from the injection site is significant, since it demonstrates OPCs generated by our protocol have migratory ability, similar to endogenous OPCs (21,40). It has been suggested that successful oligodendrocyte cell transplantation treat-ments, such as that proposed for multiple sclerosis, will depend on the ability of the transplanted OPCs to migrate to the sites where remyelination is required (9). MBP is expressed solely by myelinating cells, and the MBP expression by progeny of our transplanted OPCs provides support for our protocol producing a cell population capa-ble of in vivo terminal differentiation into functional oli-godendrocytes. However, the majority of the injected cell progeny were not MBP-positive 4 weeks after injection, and therefore, an area of future study is to examine how long the transplanted OPCs retain their ability to differen-tiate into MBP-positive oligodendrocytes after injection into the central nervous system.

In conclusion, the data presented here demonstrate that our defined, reproducible, and efficient directed differen-tiation protocol consistently generates an enriched popula-tion of OPCs from miPSCs and will be valuable to groups requiring oligodendrocyte-restricted progenitor cells to study myelination diseases in mice and develop therapeu-tic applications of OPCs in injury and murine models of myelination disease. We are currently working to translate this protocol using human iPSCs to generate oligodendro-cytes for clinical remyelination treatments (8).

ACKNOWLEDGMENTS: This work was financially supported by a generous donation through the University of Minnesota Foundation, and the KL2 Scholar Program (A. M. Parr) from the Clinical and Translational Science Institute 5KL2TR113 (NIH:8UL1TR000114). The authors declare no conflicts of interest.

REFERENCESAlsanie, W. F.; Niclis, J. C.; Petratos, S. Human embryonic 1. stem cell-derived oligodendrocytes: Protocols and perspec-tives. Stem Cells Dev. 22:2459–2476; 2013.Ben-Hur, T.; Goldman, S. A. Prospects of cell therapy for 2. disorders of myelin. Ann. NY Acad. Sci. 1142:218–249; 2008.Bradl, M.; Lassmann, H. Oligodendrocytes: Biology and 3. pathology. Acta Neuropathol. 119(1):37–53; 2010.Cai, J.; Qi, Y.; Hu, X.; Tan, M.; Liu, z.; zhang, J.; Li, Q.; 4. Sander, M.; Qiu, M. Generation of oligodendrocyte precur-sor cells from mouse dorsal spinal cord independent of Nkx6 regulation and Shh signaling. Neuron 45(1):41–53; 2005.

Cameron-Curry, P.; Le Douarin, N. M. Oligodendrocyte 5. precursors originate from both the dorsal and the ventral parts of the spinal cord. Neuron 15(6):1299–1310; 1995.Chernoff, G. F. Shiverer: An autosomal recessive mutant 6. mouse with myelin deficiency. J. Hered. 72(2):128; 1981.Czepiel, M.; Balasubramaniyan, V.; Schaafsma, W.; Stancic, 7. M.; Mikkers, H.; Huisman, C.; Boddeke, E.; Copray, S. Differentiation of induced pluripotent stem cells into func-tional oligodendrocytes. Glia 59(6):882–892; 2011.Czepiel, M 8. .; Boddeke, E.; Copray, S. Human oligodendro-cytes in remyelination research. Glia 63(4):513–530; 2015.De La Pena, I.; Pabon, M.; Acosta, S.; Sanberg, P. R.; Tajiri, 9. N.; Kaneko, Y.; Borlongan, C. V. Oligodendrocytes engi-neered with migratory proteins as effective graft source for cell transplantation in multiple sclerosis. Cell Med. 6(3):123–127; 2014. Douvaras, P.; Wang, J.; zimmer, M.; Hanchuk, S.; O’Bara, 10. M. A.; Sadiq, S.; Sim, F. J.; Goldman, J.; Fossati, V. Efficient generation of myelinating oligodendrocytes from primary progressive multiple sclerosis patients by induced pluripo-tent stem cells. Stem Cell Reports 3(2):250–259; 2014.Felts, P. A.; Baker, T. A.; Smith, K. J. Conduction in 11. segmentally demyelinated mammalian central axons. J. Neurosci. 17(19):7267–7277; 1997.Franklin, R. J.; Ffrench-Constant, C. Remyelination in 12. the CNS: From biology to therapy. Nat. Rev. Neurosci. 9(11):839–855; 2008.Goldman, S. A.; Nedergaard, M.; Windrem, M. S. Glial 13. progenitor cell-based treatment and modeling of neurologi-cal disease. Science 338(6106):491–495; 2012.Greder, L. V.; Gupta, S.; Li, S.; Abedin, M. J.; Sajini, A.; 14. Segal, Y.; Slack, J. M.; Dutton, J. R. Analysis of endog-enous Oct4 activation during induced pluripotent stem cell reprogramming using an inducible Oct4 lineage label. Stem Cells 30(11):2596–2601; 2012.Gregori, N.; Proschel, C.; Noble, M.; Mayer-Proschel, M. 15. The tripotential glial-restricted precursor (GRP) cell and glial development in the spinal cord: Generation of bipo-tential oligodendrocyte-type-2 astrocyte progenitor cells and dorsal-ventral differences in GRP cell function. J. Neurosci. 22(1):248–256; 2002.Gupta, N.; Henry, R. G.; Strober, J.; Kang, S. M.; Lim, D. 16. A.; Bucci, M.; Caverzasi, E.; Gaetano, L.; Mandelli, M. L.; Ryan, T. Neural stem cell engraftment and myelination in the human brain. Sci. Transl. Med. 4(155):155ra137; 2012.Hu, B. Y.; Du, z. W.; Li, X. J.; Ayala, M.; zhang, S. C. 17. Human oligodendrocytes from embryonic stem cells: Conserved SHH signaling networks and divergent FGF effects. Development 136(9):1443–1452; 2009.Jiang, P.; Selvaraj, V.; Deng, W. Differentiation of embry-18. onic stem cells into oligodendrocyte precursors. J. Vis. Exp. 39:1960; 2010.Keirstead, H. S.; Nistor, G.; Bernal, G.; Totoiu, M.; Cloutier, 19. F.; Sharp, K.; Steward, O. Human embryonic stem cell-derived oligodendrocyte progenitor cell transplants remy-elinate and restore locomotion after spinal cord injury. J. Neurosci. 25(19):4694–4705; 2005.Li, Y.; Gautam, A.; Yang, J.; Qiu, L.; Melkoumian, z.; 20. Weber, J.; Telukuntla, L.; Srivastava, R.; Whiteley, E. M.; Brandenberger, R. Differentiation of oligodendrocyte pro-genitor cells from human embryonic stem cells on vitronec-tin-derived synthetic Peptide acrylate surface. Stem Cells Dev. 22(10):1497–1505; 2013.

424 TERzIC ET AL.

Masahira, N.; Takebayashi, H.; Ono, K.; Watanabe, 21. K.; Ding, L.; Furusho, M.; Ogawa, Y.; Nabeshima, Y.; Alvarez-Buylla, A.; Shimizu, K. Olig2-positive pro-genitors in the embryonic spinal cord give rise not only to motoneurons and oligodendrocytes, but also to a subset of astrocytes and ependymal cells. Dev. Biol. 293(2):358–369; 2006.Miller, R. H. Regulation of oligodendrocyte development 22. in the vertebrate CNS. Prog. Neurobiol. 67(6):451–467; 2002.Misumi, S.; Nishigaki, R.; Ueda, Y.; Watanabe, Y.; Shimizu, 23. Y.; Ishida, A.; Jung, C.; Hida, H. Differentiation of oligo-dendrocytes from mouse induced pluripotent stem cells without serum. Transl. Stroke Res. 4(2):149–157; 2013.Najm, F. J.; zaremba, A.; Caprariello, A. V.; Nayak, S.; 24. Freundt, E. C.; Scacheri, P. C.; Miller, R. H.; Tesar, P. J. Rapid and robust generation of functional oligodendrocyte progenitor cells from epiblast stem cells. Nat. Methods 8(11):957–962; 2011.Nistor, G. I.; Totoiu, M. O.; Haque, N.; Carpenter, M. K.; 25. Keirstead, H. S. Human embryonic stem cells differenti-ate into oligodendrocytes in high purity and myelinate after spinal cord transplantation. Glia 49(3):385–396; 2005.Noble, M.; Proschel, C.; Mayer-Proschel, M. Getting 26. a GR(i)P on oligodendrocyte development. Dev. Biol. 265(1):33–52; 2004.Ogawa, S.; Tokumoto, Y.; Miyake, J.; Nagamune, T. 27. Immunopanning selection of A2B5-positive cells increased the differentiation efficiency of induced pluripotent stem cells into oligodendrocytes. Neurosci. Lett. 489(2):79–83; 2011.Okita, K.; Nakagawa, M.; Hyenjong, H.; Ichisaka, T.; 28. Yamanaka, S. Generation of mouse induced pluripotent stem cells without viral vectors. Science 322(5903):949–953; 2008.Park, I. H.; zhao, R.; West, J. A.; Yabuuchi, A.; Huo, H.; 29. Ince, T. A.; Lerou, P. H.; Lensch, M. W.; Daley, G. Q. Reprogramming of human somatic cells to pluripotency with defined factors. Nature 451(7175):141–146; 2008.Sajini, A. A.; Greder, L. V.; Dutton, J. R.; Slack, J. M. 30. Loss of Oct4 expression during the development of murine embryoid bodies. Dev. Biol. 371(2):170–179; 2012.Schnitzer, J.; Schachner, M. Cell type specificity of a neural 31. cell surface antigen recognized by the monoclonal antibody A2B5. Cell Tissue Res. 224(3):625–636; 1982.

Sharp, J.; Frame, J.; Siegenthaler, M.; Nistor, G.; Keirstead, 32. H. S. Human embryonic stem cell-derived oligodendrocyte progenitor cell transplants improve recovery after cervical spinal cord injury. Stem Cells 28(1):152–163; 2010.Sim, F. J.; McClain, C. R.; Schanz, S. J.; Protack, T. L.; 33. Windrem, M. S.; Goldman, S. A. CD140a identifies a pop-ulation of highly myelinogenic, migration-competent and efficiently engrafting human oligodendrocyte progenitor cells. Nat. Biotechnol. 29(10):934–941; 2011.Sinha, S.; Chen, J. K. Purmorphamine activates the 34. Hedgehog pathway by targeting Smoothened. Nat. Chem. Biol. 2(1):29–30; 2006.Stacpoole, S. R.; Spitzer, S.; Bilican, B.; Compston, A.; 35. Karadottir, R.; Chandran, S.; Franklin, R. J. High yields of oligodendrocyte lineage cells from human embryonic stem cells at physiological oxygen tensions for evaluation of translational biology. Stem Cell Reports 1(5):437–450; 2013.Sundberg, M.; Hyysalo, A.; Skottman, H.; Shin, S.; Vemuri, 36. M.; Suuronen, R.; Narkilahti, S. A xeno-free culturing pro-tocol for pluripotent stem cell-derived oligodendrocyte pre-cursor cell production. Regen. Med. 6(4):449–460; 2011.Takahashi, K.; Tanabe, K.; Ohnuki, M.; Narita, M.; 37. Ichisaka, T.; Tomoda, K.; Yamanaka, S. Induction of pluri-potent stem cells from adult human fibroblasts by defined factors. Cell 131(5):861–872; 2007.Takahashi, K.; Yamanaka, S. Induction of pluripotent stem 38. cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 126(4):663–676; 2006.Tokumoto, Y.; Ogawa, S.; Nagamune, T.; Miyake, J. 39. Comparison of efficiency of terminal differentiation of oligodendrocytes from induced pluripotent stem cells versus embryonic stem cells in vitro. J. Biosci. Bioeng. 109(6):622–628; 2010.Tsai, H. H.; Frost, E.; To, V.; Robinson, S.; Ffrench-40. Constant, C.; Geertman, R.; Ransohoff, R. M.; Miller, R. H. The chemokine receptor CXCR2 controls positioning of oligodendrocyte precursors in developing spinal cord by arresting their migration. Cell 110(3):373–383; 2002.Wang, S.; Bates, J.; Li, X.; Schanz, S.; Chandler-Militello, 41. D.; Levine, C.; Maherali, N.; Studer, L.; Hochedlinger, K.; Windrem, M. Goldman, S. A. Human iPSC-derived oli-godendrocyte progenitor cells can myelinate and rescue a mouse model of congenital hypomyelination. Cell Stem Cell 12(2):252–264; 2013.