Differentiation induction of neural stem cells (NSCs) into oligodendrocytes during embryogenesis is the result of a complex interaction between local induction factors and intracellular transcription factors. At the early stage of differentiation, in particular, the helix-loop-helix transcription factors Olig1 and Olig2 have been shown to be essential for oligodendrocyte lineage determination. In view of the possible application of NSCs as a source for remyelinating cell transplants in demyelinating diseases (e.g., multiple sclerosis), in vitro procedures need to be developed to drive the oligodendrocyte differentiation process. Mere culture in medium supplemented with major embryonic oligodendrogenic induction factors, such as Sonic hedgehog, results in oligodendrocyte differentiation of only about 10% of NSCs. We previously showed that induction of Olig1 expression by gene transfection could indeed initiate the first stage of oligodendrocyte differentiation in NSCs, but appeared to be unable to generate fully mature, functional oligodendrocytes. In this study, we transfected NSCs isolated from the embryonic mouse brain with the Olig2 gene and found that the introduced overexpression of Olig2 could induce the development of fully mature oligodendrocytes expressing the transcription factor Nkx2.2 and all major myelin-specific proteins. Moreover, Olig2-transfected NSCs, in contrast to nontransfected NSCs, developed into actively remyelinating oligodendrocytes after transplantation into the corpus callo-sum of long-term cuprizonefed mice, an animal model for demyelination. Our results show that transfection of genes encoding for oligodendrogenic transcription factors can be an efficient way to induce the differentiation of NSCs into functional oligodendrocytes.
The relapsing/remitting form of multiple sclerosis (MS) is characterized by acute demyelinating episodes followed by the generation of new oligodendrocytes, remyelination, and recovery. Repeated attacks, however, eventually lead to a reduction in the capacity to generate new oligodendrocytes and consequently to a failure to induce full remyelination. The increasing load of demyelinated lesions and, particularly, the axonal degeneration that eventually accompanies the loss of myelin is responsible for the severe progressive neurological deterioration seen in patients with this form of MS . In the search for therapeutic interventions to promote remyelination, research has not only focused on stimulating the remyelination capacity of endogenous oligodendrocytes but also on the application of exogenous remyelinating cells. Various experimental studies have reported the survival, integration, and remyelinating activity of transplanted oligodendrocytes and oligodendrocyte progenitor cells (OPCs) in rat models for demyelination . Embryonic, bone marrow-derived stem cells and neural stem cells (NSCs) have been forwarded as potential sources for remyelinating cell transplants [3–12]. So far, in most studies, undifferentiated labeled stem cells were implanted, assuming that the site of demyelination would provide the proper niche for the stem cells to differentiate into remyelinating cells. For instance, Pluchino et al.  injected labeled NSCs i.v. in the experimental allergic encephalomyelitis rat model for MS. Indeed, labeled remyelinating cells were observed at the site of the lesions, but most of the beneficial effects of the NSCs were ascribed to their production of anti-inflammatory and remyelination-promoting growth factors . In addition, the transplanted undifferentiated NSCs may have given rise to astrocytes that interfere with the remyelination process. To increase and stimulate the contribution of grafted NSCs in actual remyelination, they require predifferentiation into an oligodendrocyte cell lineage before transplantation.
We previously showed that transient Olig1 expression through nonviral gene transfection could initiate the induction of oligodendrocyte differentiation in over 50% of embryonic mouse brain-derived NSCs in vitro . Olig1 is a basic-helix-loop-helix (bHLH) protein and a member of the family of Olig transcription factors. Various studies have demonstrated the crucial role of Olig1 and Olig2 in oligodendrogenesis and myelination in vivo [15–25], with a difference between brain and spinal cord. The specific temporal-spatial expression of Olig2 in the developing spinal cord (pMN domain) has been shown to induce the formation of a motoneuron-oligodendrocyte lineage [26, 27]: activation by Olig2 of the Ngn1 and Ngn2 genes leads to spinal motoneuron differentiation, whereas coactivation of Olig2 and Nkx2.2 provokes the formation of OPCs in the spinal cord. Olig1 has been thought to be involved only in promoting the survival and maturation of the developing spinal cord OPCs. On the contrary, in the brain, Olig1 has been suggested to be the key factor for the induction of cortical OPC formation: ectopic expression of Olig1 in embryonic mouse forebrain ventricles appeared to promote oligodendrocyte formation in all brain regions . Olig2 was assumed to be only involved in survival and maturation of the newly formed brain OPCs [16, 28]. In our in vitro studies  the transient expression of Olig1 could only initiate oligodendrocyte differentiation in embryonic mouse NSCs, but appeared to be insufficient for their complete maturation. Our data suggested that Olig1 overexpression in mouse embryonic brain NSCs in vitro could not induce other coactivating transcription factors (e.g., Nkx2.2) to promote full oligodendrocyte maturation. Recent studies in Olig1-null mice have shed a different light on the role of Olig1 in the developing brain and suggested a role as a central regulator of oligodendrocyte myelinogenesis: in mice lacking a functional Olig1 gene, expression of myelin-specific genes was abolished, whereas the formation of oligodendrocyte progenitors seemed to be unaffected .
In the present study, we transfected mouse embryonic NSCs with the Olig2 gene and compared its oligodendrocyte differentiation induction potential with that after Olig1 and Nkx2.2 gene transfection. Moreover, we examined the remyelinating capacity of the Olig2-transfected NSCs after transplantation in the cuprizone demyelination mouse model [30, 31]. Long-term feeding (>12 weeks) of 0.2% of the copper chelator cuprizone induces complete demyelination of the corpus callosum with an almost complete depletion of the pool of endogenous OPCs . It has been shown to be an appropriate model to study remyelination by exogenous remyelinating cells in the absence of inflammation .
Materials and Methods
Neural Stem Cell Isolation and Culture
NSCs were obtained from C57/BL6 mouse embryos at embryonic day (E)14. To that end, the whole embryonic brain was dissected and dissociated as described previously. Briefly, the brain was cut into small pieces, incubated with 0.05% trypsin-EDTA for 30 minutes at room temperature, and mechanically triturated through a 26-gauge needle. Cells were cultured in serum-free Neurobasal medium (Invitrogen, Breda, The Netherlands, http://www.invitrogen.com) supplemented with human recombinant epidermal growth factor (EGF, 20 ng/ml; Invitrogen), basic fibroblast growth factor (bFGF, 20 ng/ml; Invitrogen), B27 (Invitrogen), penicillin-streptomycin 1% (Sigma-Aldrich, Zwijndrecht, The Netherlands, http://www.sigmaaldrich.com), l-glutamine 1%, glutamax 1% in T25 (Nunc, Roskilde, Denmark, http://www.nalgenunc.com) in culture flasks in a humidified 5% CO2/95% air incubator at 37°C. Within 5–7 days, the cells grew as free-floating neurospheres and were passaged after mechanical dissociation through a fire-polished Pasteur pipette 3 days after formation. After two passages, neurospheres were used for the differentiation induction experiments. These neurospheres were carefully dissociated (after a short exposure of [5 minutes] to 0.1 mmol EGTA) through a fire-polished Pasteur pipette and subsequently cultured in poly-l-lysine/laminin-coated chamberslides (Nunc), at approximately 104 cells per well. To study in vitro differentiation induction of these NSCs into oligodendrocytes, they were cultured in the oligodendrocyte-specific Sato medium, consisting of Dulbecco's Modified Eagle's medium (DMEM) with additives of glutamax 1 (1%), l-glutamine (1%), penicillin-streptomycin (1%), putrescine (16 μg/ml; Sigma), thyroxin (T4, 400 ng/ml; Sigma), triiodothyroxin (T3, 400 ng/ml; Sigma), progesterone (6.2 ng/ml; Sigma), sodium selenite (5 ng/ml; Sigma), bovine serum albumin (BSA, 100 μg/ml; Sigma), insulin (5 μg/ml; Sigma), and transferrin (50 μg/ml; Sigma). During the first 2 days of culture, the Sato medium was supplemented with Sonic hedgehog (Shh, 100 ng/ml; R&D Systems, Abington, U.K., http://www.rndsystems.com), fibroblast growth factor 2 (FGF-2, 10 ng/ml; R&D Systems), and platelet-derived growth factor (PDGF-α, 10 ng/ml; R&D Systems).
A nucleofector (Amaxa GmbH, Cologna, Germany, http://www.amaxa.com) was used for gene transfection of the NSCs using an electroporation protocol specifically designed for the trans-fection of embryonic mouse NSCs by the manufacturer. Neurospheres were carefully dissociated, and approximately 5 × 106 NSCs were transfected with 10 μg of the expression vector. The plasmids used for transfection were the pIRES expression vector containing the Olig2 gene (972 bp) or the Olig1 gene (735 bp) (both kindly donated by Dr. J. Cai, Louisville, KY) or the pcDNA3 expression vector containing the Nkx2.2 gene (822 bp, kindly donated by Dr. D. Rowitch, Boston, MA). NSCs transfected with the gene encoding for green fluorescent protein (GFP) served as a (transfection) control. Following transfection, the NSCs were kept overnight at 37°C in proliferation medium according to the manufacturer's protocol. The transfection efficiency of this procedure for mouse NSCs has been demonstrated to be 60%–80% under optimal conditions, and this nonviral transfection results in the transient expression of the transfected gene lasting up to 12 days . Next, the cells were plated in poly-l-lysine/laminin-coated chamberslides (Nunc), at approximately 104 cells per well. To study differentiation induction of these NSCs into oligodendrocytes, they were cultured in Sato medium, as described above. During the first 2 days of culture, the Sato medium was supplemented with Shh (100 ng/ml; R&D Systems), FGF-2 (10 ng/ml; R&D Systems), and PDGF-α (10 ng/ml; R&D Systems).
After various culture periods of up to 10 days, cell cultures were fixed with 4% paraformaldehyde (PFA) and immunostained to identify the various stages of differentiation of the developing neural cell types. The following antibodies were used to identify neuronal cells: anti-microtubule-associated protein 2 (MAP-2) (1:500, catalog no. AB5622; Chemicon, Hampshire, U.K.), antineuronal-specific nuclear protein (NeuN) (1:200, catalog no. MAB377; Chemicon), and anti-β-tubulin III (1:200, catalog no. T8660; Sigma). To identify astrocytes we used anti-glial fibrillary acidic protein (GFAP) (1:200, catalog no. MAB3402; Chemicon). To identify oligodendrocytes at different developmental stages, the following antibodies were used: anti-O4 (1:300, catalog no. MAB345; Chemicon), anti-A2B5 (1:200, catalog no. MAB312R; Chemicon), anti-PDGFRα (1:200, catalog no. SC338; Santa Cruz Biotechnology Inc., Santa Cruz, CA, http://www.scbt.com), anti-myelin basic protein (MBP) (1:500, catalog no. AB980; Chemicon), anti-galactosidase C (GAL-C) (1:500, catalog no. MAB342; Chemicon), anti-myelin/oligodendrocyte-specific protein (1:200, catalog no. MAB328; Chemicon), anti-Olig1 (1:200, catalog no. AB5991; Chemicon), anti-Olig2 (1:2,000, kindly donated by Dr. H. Take-bayashi [28, 34]), and anti-Nkx2.2 (1:200; Developmental Studies Hybridoma Bank, Iowa City, IA, http://www.uiowa.edu/∼dshbwww). Undifferentiated NSCs were identified with anti-nestin antibody (1:200, catalog no. MAB353; Chemicon). Immunolabeled cells were detected using fluorescently labeled secondary antibodies (Jackson Immunoresearch Laboratories, West Grove, PA, http://www.jacksonimmuno.com). To identify individual viable cells, Hoechst nuclear staining was used.
Implantation in Cuprizone Mouse Model for MS
C57/BL6 mice were put on a diet of 0.2% (w/w) cuprizone (Sigma), a copper chelator. This diet leads to oligodendrocyte death and subsequent demyelination in some regions of the brain, but most prominently in the corpus callosum [30, 31]. Initially, endogenous OPCs in the corpus callosum and those migrating in from the environment replace the lost callosal oligodendrocytes. However, long-term feeding of cuprizone (>12 weeks) eventually leads to complete demyelination of the corpus callosum and an almost complete exhaustion of the endogenous OPC pool in the corpus callosum and the wider surroundings . The bundle of denuded axons in the corpus callosum has been shown to be an ideal substrate to test the in vivo remyelination capacity of exogenous OPCs  without the interfering contribution of endogenous OPCs. To test the remyelinating capacity of NSC-derived oligodendrocytes in vivo, suspensions of 30,000 cells in 3 μl of phosphate-buffered saline (PBS) were injected into the cuprizone-treated corpus callosum of ketamine-anesthetized mice using the following stereotactic coordinates in reference to the bregma: +0.98 mm (anteroposterior axis), − 1.75 mm (lateromedial axis), −2.25 mm (vertical axis) . To enable identification of the grafted cells afterward in (immuno)-histological sections, cells were prelabeled in vitro with mini-Ruby (Molecular Probes Inc., Eugene, OR, http://probes.invitrogen.com) 2 hours before implantation. Three groups of mice receiving cell implants were analyzed: cell implants containing nontransfected NSCs (n = 10), cell implants containing Olig2-transfected NSCs (n = 12), and implants without cells containing only PBS (sham-operated) (n = 4). Moreover, the contralateral unoperated sides of the corpus callosum of all groups were analyzed as controls. At 5 and 14 days postimplantation, mice were perfused transcardially with prerinse containing 0.9% NaCl and 1% heparin, followed by 200 ml of 4% PFA in 0.1 M PBS, pH 7.4. After explantation, the brains were sectioned on a cryostat for (immuno)-histochemical analysis of the corpus callosum. The same antibodies as described above were used. In addition, sections were stained with Luxol Fast Blue/cresyl fast violet for myelin evaluation . In addition to those used for (immuno)-histological examination at the light microscopic level, a number of brains were processed for electron microscopic analysis. To that end, 100-μm sections of the corpus callosum area were cut on a Vibratome tissue slicer (Warner Instruments, Hamden, CT, http://www.warneronline.com) and postfixed for 1 hour in 1% osmium tetroxide (w/v) in 0.1 M sodium acetate cacodylate buffer, pH 7.6. The sections were dehydrated and embedded in Epon 812. Semithin (1-μm) sections were stained with toluidine blue to check for the presence of relevant areas before cutting ultrathin sections (60 nm) that were counterstained with uranyl acetate and lead citrate and examined using a Philips CM-100 electron microscope (Philips, Eindhoven, The Netherlands, http://www.medicalphilips.com). Electron microscopic sections were used to quantify de- and remyelination of the corpus callosum. To that end, for each mouse, photomicrographs (at a magnification of × 3,400) were made of a standardized cross-sectional area (0.1 × 0.1 mm) through the corpus callosum on both sides (i.e., the operated and unoperated sides) 1.75 mm lateral to the midline. All myelinated and demyelinated axons present in this area were counted. The number of (re)myelinated axons was expressed as a percentage of the total number of axons present in the area.
All animal experiments were carried out in accordance with the National Institutes of Health Guide for the Care and Use of Laboratory Animals.
Oligodendrocyte Differentiation Induction In Vitro
Preceding the gene transfection and differentiation experiments, immunostaining for the oligodendrogenic transcription factors Olig1, Olig2, and Nkx2.2 was performed on undifferentiated mouse embryonic brain NSCs. Whereas no expression of Olig1 and Nkx2.2 could be detected in the NSCs, almost all (> 95%) of the NSCs expressed low levels of Olig2. Differentiation of nontransfected or GFP-transfected mouse embryonic NSCs in the oligodendrocyte-specific growth factor-enriched Sato medium resulted in the differentiation of the NSCs into astrocytes (65%), neurons (26%), and, in approximately 10%, early-stage oligodendrocytes, as demonstrated by their staining for O4 (Fig. 1B). Transfection of NSCs with the Olig2 gene using nucleofection, inducing a transient overexpression of Olig2, resulted in a greater number of NSCs differentiating into an oligodendrocyte cell lineage, up to 25% (Fig. 1C–1G). After 7 days in culture, all of these oligodendrocytes developed into fully mature oligodendrocytes, with elaborate myelin extensions staining positive for major oligodendrocytic proteins, such as O4, Gal-C, myelin oligodendrocyte glycoprotein (MOG), and MBP (Fig. 1C–1F). Immunostaining for the transcription factor Nkx2.2 revealed that only the fully mature oligodendrocytes developing after Olig2 transfection expressed Nkx2.2 (Fig. 1D–D″). Neither Nkx2.2 gene transfection nor transfection with the Olig1 gene in the NSCs was able to generate the differentiation of fully mature (i.e., MOG- or MBP-positive) oligodendrocytes. As also previously shown by us , Olig1 gene transfection with nucleofection could only initiate oligodendrocyte differentiation, but the primitive immature precursors did not develop into mature oligodendrocytes and did not express the Nkx2.2 transcription factor. To verify the in vivo (re)myelinating capacity of the Olig2-induced NSC-derived oligodendro-cytes, transplantation experiments were performed in a de-myelination mouse model.
Transplantation in Cuprizone Mouse Model for MS
Feeding mice with 0.2% cuprizone during a period of over 12 weeks caused complete demyelination of the corpus callosum and an almost complete exhaustion of the endogenous OPC pool in the corpus callosum (Fig. 2). Between 12 and 15 weeks after starting the 0.2% cuprizone diet, mice received stereotactical implantation of a suspension of mini-Ruby-labeled NSCs (30,000 in 3 μl) in the striatum, just below the (demyelinated) corpus callosum (Fig. 3A). From that day on, mice were put back on normal food. Within 5 days after implantation, the mini-Ruby-labeled cells had migrated into the demyelinated corpus callosum and dispersed over the entire, predominantly ipsilateral, corpus (Fig. 3B). Only sporadically did NSCs migrate ventrally into the striatum toward the ventricle or dorsally into the cortex (data not shown). Although a vast majority of the grafted NSCs died in the first week after transplantation, approximately 10% could still be detected 12 days postimplantation (Fig. 3C–3E). Most of these cells remained in an undifferentiated state, whereas a few NSCs differentiated into astrocytes and a few, particularly those that had migrated into the cortex, differentiated into neurons. Grafted NSCs that were transfected with the Olig2 gene 24 hours previously developed into oligo-dendrocytes and contributed to the remyelination of the denuded axons in the corpus callosum (Fig. 3F). Double immunostaining revealed that the mini-Ruby-labeled Olig2-transfected NSCs expressed oligodendrocyte-specific proteins such as Gal-C, MOG, and MBP (Fig. 3G, 3H); most of these cells appeared to be surrounded by GFAP-positive astrocytes (Fig. 3I). At the electron microscopic level, in contrast to sham-operated mice and mice that received nontransfected NSCs (Fig. 4B, 4C), newly remyelinated axons could be observed near the transplantation site of the Olig2-transfected NSCs (Fig. 4D, 4G). Quantification of the percentage of remyelinated versus demy-elinated axons in the corpus callosum demonstrated significantly greater myelination in comparison with the nontreated site and with the nontransfected NSC grafts (Fig. 4H). Electron microscopic analysis further revealed that the remyelinating grafted Olig2-transfected NSCs appeared to be in close contact with astrocytes (Fig. 5).
Our results show that the induction of overexpression of Olig2 in mouse embryonic NSCs by nonviral gene transfection, in contrast to Olig1 and Nkx2.2 transfection, can promote their differentiation into mature oligodendrocytes in oligodendrocyte-specific, enriched Sato medium. Moreover, in contrast to non-transfected NSCs, NSCs transfected with the Olig2 gene appear to develop into functional oligodendrocytes after implantation and to remyelinate denuded axons in the corpus callosum of cuprizone-fed mice.
Many studies have demonstrated that the bHLH transcription factors Olig1 and Olig2 are critically involved in the formation of oligodendrocytes and in myelination during embryogenesis [15–19, 20–25]. Whereas Olig2 appears to be required for the initiation of oligodendrogenesis in the spinal cord, Olig1 is suggested to contribute more to oligodendrocyte differentiation in the brain . In addition, Lu et al.  showed that ectopic expression of Olig1 in cultures of rat embryonic cortical NSCs can induce the formation of OPCs. Also, in our studies, we were able to initiate oligodendrocyte differentiation in vitro by inducing the expression of Olig1 via gene transfection in NSCs isolated from the embryonic mouse brain, cells that do not express this transcription factor in the undifferentiated state. However, this ectopic expression of Olig1 in vitro could not accomplish the formation of fully mature oligodendrocytes in the NSCs and appeared unable to recruit the downstream transcription factor Nkx2.2 that is essential for full oligodendrocyte maturation. Despite the fact that we cultured the Olig1-transfected NSCs in an oligodendrocyte-specific medium enriched with known oligodendrogenic induction factors such as Shh and FGF-2, we were clearly unable to mimic the in vivo conditions of the embryonic brain. Recent observations in Olig1-null mice suggest that the Olig1 transcription factor plays a much more important role in myelinogenesis . In contrast to Olig1, similar experiments with ectopic expression of Olig2 in cultures of embryonic spinal cord appear unable to show the induction of oligodendrocyte differentiation [24, 25]. This has been explained by the fact that Olig2 requires the collaboration of Nkx2.2 for oligodendrocyte formation . Indeed, only coelectroporation (and not separate electroporation) of Olig2 and Nkx2.2 in the developing chick spinal cord appeared to be able to promote the differentiation and maturation of ectopic oligo-dendrocytes . These studies and our experiments with embryonic mouse brain NSCs seem to indicate that there are considerable differences between the activities of Olig1 and Olig2 in brain- and spinal cord-derived NSCs as far as their transcriptional regulation of oligodendrocyte differentiation is concerned. Apparently, the standard levels of Olig2 present in most of our undifferentiated embryonic mouse brain NSCs were insufficient to induce oligodendrocyte differentiation during culture in the Shh/FGF-2-enriched oligodendrocyte-specific Sato medium. Only the portion of NSCs in which we could induce a sufficient level of Olig2 overexpression by gene transfection became early-stage O4-positive OPCs, started to express Nkx2.2, and developed into fully mature oligodendrocytes during culture in the enriched Sato medium. These oligodendrocytes expressed all major lineage-specific markers. Induction of ectopic Nkx2.2 expression by gene transfection in these NSCs did not promote oligodendrocyte formation. However, cotransfection experiments of Olig2 and Nkx2.2 may eventually result in an even higher number of NSCs that differentiate into oligo-dendrocytes than those transfected with only Olig2. Experiments in human olfactory epithelial-derived progenitor cells by Zhang et al.  have shown that only cotransfection of Olig2 plus Nkx2.2 (but not separate transfections) could induce oligodendrocyte differentiation in vitro. It is difficult to compare these human cells and the basic culture conditions applied in the experiments of Zhang et al.  with the mouse embryonic brain NSCs and the enriched “oligodendrogenic” culture conditions we used. In that respect, evidence has been provided for interspecies differences between human and rodent neural stem cells as far as their capacity to generate oligodendrocytes in vitro is concerned . The fact that oligodendrocyte differentiation in human olfactory cells completely lacked a normal O4 progenitor stage suggests that at least other transcription regulatory processes are involved; it has been suggested that, in these cells, Nkx2.2 may have directly activated MBP gene expression in collaboration with Olig2 without expressing O4.
For testing the functionality and, in particular, the myelinating capacity of NSC-derived oligodendrocytes in vitro, experimental systems have been employed using coculture with dorsal root ganglion neurons. However, myelination is a complicated process that requires a specific in vivo environment that can hardly be provided by culture conditions. Therefore, we used the cuprizone demyelination model [30, 31] to test whether Olig2-transfected NSCs could develop into functional remyelinating oligodendrocytes in vivo. In a recent paper, Mason et al.  described that long-term feeding (> 12 weeks) of 0.2% cuprizone led to complete demyelination of the corpus callosum and a depletion of endogenous OPCs. They have shown this to be a most suitable animal model to study the in vivo remyelination of denuded axons by exogenous oligodendrocytes . Our results show that Olig2-transfected NSCs, in contrast to nontransfected NSCs, can develop into functional remyelinating oligodendrocytes after intracerebral implantation. Electron microscopic analysis showed that these remyelinating oligodendrocytes interact with astrocytes in accordance with observations by others that the interaction between astrocytes and oligodendrocytes may be essential for remyelination to occur .
In conclusion, our experiments show that introduction of Olig2 may be used to differentiate NSCs into an oligodendrocyte cell lineage. Application of this approach to bone marrow-, blood-, or skin-derived NSCs may enlarge the potential of these cells as an autologous accessible source for cell replacement therapy in demyelinating diseases.
We gratefully acknowledge the technical assistance of I. Manting-Otter and the preparation of the cuprizone feed by Dr. G.J. van Dijk. This research was supported by a grant of the Hersen-stichting Nederland, 12F04.51 and the Dutch Foundation MS Research, MS 04–554MS.
The authors indicate no potential conflicts of interest.