Department of Biochemistry and Molecular Medicine, School of Medicine, Sacramento, California, USA
Institute for Pediatric Regenerative Medicine, Shriners Hospitals for Children Northern California, Sacramento, California, USA
Correspondence: Wenbin Deng, Ph.D., Department of Biochemistry and Molecular Medicine, UC Davis School of Medicine, Institute for Pediatric Regenerative Medicine, Shriners Hospitals for Children Northern California, 2425 Stockton Blvd., Sacramento, California 95817, USA. Telephone: 916-453-2287; Fax: 916-453-2288; e-mail: email@example.com
Author contributions: P.J.: conception and design, collection and assembly of data, data analysis and interpretation, and manuscript writing; C.C., collection and assembly of data and data analysis and interpretation; X.-B.L., V.S., W.L., and D.H.F.: collection and assembly of data; Y.L.: conception and design and provision of study material; D.E.P.: conception and design; R.A.L.: provision of study material; W.D.: conception and design, financial support, data analysis and interpretation, manuscript writing, and final approval of manuscript.
Pluripotent stem cells (PSCs) have been differentiated into oligodendroglial progenitor cells (OPCs), providing promising cell replacement therapies for many central nervous system disorders. Studies from rodents have shown that brain OPCs express a variety of ion channels, and that a subset of brain OPCs express voltage-gated sodium channel (NaV), mediating the spiking properties of OPCs. However, it is unclear whether PSC-derived OPCs exhibit electrophysiological properties similar to brain OPCs and the role of NaV in the functional maturation of OPCs is unknown. Here, using a mouse embryonic stem cell (mESC) green fluorescent protein (GFP)-Olig2 knockin reporter line, we demonstrated that unlike brain OPCs, all the GFP+/Olig2+ mESC-derived OPCs (mESC-OPCs) did not express functional NaV and failed to generate spikes (hence termed “nonspiking mESC-OPCs”), while expressing the delayed rectifier and inactivating potassium currents. By ectopically expressing NaV1.2 α subunit via viral transduction, we successfully generated mESC-OPCs with spiking properties (termed “spiking mESC-OPCs”). After transplantation into the spinal cord and brain of myelin-deficient shiverer mice, the spiking mESC-OPCs demonstrated better capability in differentiating into myelin basic protein expressing oligodendrocytes and in myelinating axons in vivo than the nonspiking mESC-OPCs. Thus, by generating spiking and nonspiking mESC-OPCs, this study reveals a novel function of NaV in OPCs in their functional maturation and myelination, and sheds new light on ways to effectively develop PSC-derived OPCs for future clinical applications. Stem Cells2013;31:2620–2631
Oligodendroglial progenitor cells (OPCs), also known as “NG2 glia” due to the expression of the membrane proteoglycan NG2 , mainly generate oligodendrocytes in the developing and mature central nervous system (CNS) [1, 2]. Electrophysiological recordings from both in situ brain slice and in vitro cell culture have shown that voltage-gated ion channels are expressed in rodent CNS OPCs, and that ion channel expression is developmentally regulated [3-5]. Voltage-gated potassium currents (IK), mainly delayed rectifier potassium current (IKD) and inactivating A-type potassium current (IKA), are expressed in OPCs [3, 4, 6]. When OPCs mature into oligodendrocytes, another type of IK, inward rectifier potassium current (IKir) is expressed . Accumulating evidence has shown that voltage-gated sodium channel (NaV)-mediated current (INa) is also expressed in OPCs [3-6, 8-11]. Moreover, studies have also shown that a subset of OPCs can fire action potentials upon depolarization and that this spiking property relies on the expression of NaV [9, 10]. In oligodendrocyte lineage cells, the INa expression is specific to OPCs. When OPCs mature into oligodendrocytes, INa is not observed . Despite the subdivision of spiking and nonspiking OPCs, the role of the expression of functional NaV in OPC development remains unclear.
Pluripotent stem cells (PSCs) have been successfully differentiated into OPCs [12-16] for potential regenerative therapies for oligodendrocyte injury-related CNS disorders, such as spinal cord injury [17, 18] and multiple sclerosis . However, very few studies have explored the functional properties of PSC-derived OPCs, particularly their electrophysiological properties. Here, we first differentiated green fluorescent protein (GFP)-Olig2 mouse embryonic stem cells (mESCs), in which GFP was inserted into the Olig2 locus and thus GFP expression mirrored endogenous Olig2 expression , into GFP+/Olig2+ OPCs (mESC-OPCs) by the treatment of small molecules retinoic acid (RA) and purmorphamine [21, 22]. We further showed that IKD and IKA were expressed in GFP+ mESC-OPCs. However, unlike in rodent CNS OPCs, the INa could not be detected in mESC-OPCs. By ectopically expressing Nav1.2 α subunit, the mESC-OPCs started to express INa and acquired spiking properties. In this study, we thus refer the mESC-OPCs with and without the expression of INa as spiking and nonspiking mESC-OPCs, respectively. The generation of nonspiking mESC-OPCs and engineered spiking mESC-OPCs thus provides us with a powerful tool to explore the functional roles of INa in the OPCs. Using in vitro coculture with neurons and in vivo transplantation into shiverer mice, we demonstrated that spiking mESC-OPCs had better capability of maturating into myelin basic protein (MBP) positive oligodendrocytes and myelinating axons than nonspiking mESC-OPCs. Overall, by providing the insights into the function of NaV and engineered spiking activities in OPC maturation and myelination, this study demonstrates the need for applying ion channel physiology not only to the differentiation of stem cells into functional glial precursor cells but also more importantly to future clinical application of stem cells.
Materials and Methods
Maintenance of mESCs
The mouse ESC line GFP-Olig2 was obtained from the American Type Culture Collection and maintained using standard mESC culture methods as described in our previous studies [21, 22]. In brief, the mESCs were grown at an optimal density that required routine passaging every 3 days on irradiated mouse embryonic fibroblast feeder layers (GlobalStem, Rockville, MD, http://www.globalstem.com/). The culture medium was Dulbecco's modified Eagle's medium (DMEM; GIBCO, Grand Island, NY, http://www.invitrogen.com/site/us/en/home/brands/Gibco.html) supplemented with 20% fetal bovine serum (GIBCO), 2 mM l-glutamine (GIBCO), 1 mM sodium pyruvate (GIBCO), 0.1 mM β-mercaptoethanol (GIBCO), 0.1 mM nonessential amino acids (NEAA; GIBCO), and 1,000 U/ml leukemia inhibitory factor (Millipore, Billerica, MA, http://www.millipore.com/).
Differentiation of mESCs
Neural differentiation of mESCs was initiated using our published protocol [21, 22]. Briefly, mESC colonies were trypsinized into single cells and suspended in differentiation medium to form embryoid bodies (EB). The differentiation medium consisted of α-minimal essential medium (α-MEM) supplemented with 20% knockout serum replacement, 1 mM sodium pyruvate, 0.1 mM NEAA, and 0.1 mM β-mercaptoethanol (all from GIBCO). As shown in Figure 1A, from day 4 to day 7, RA (0.2 μM; Sigma, St. Louis, MO, http://www.sigmaaldrich.com/united-states.html) and purmorphamine (Pur, 1 μM; Cayman Chemical, Ann Arbor, MI, https://www.caymanchem.com/app/template/Home.vm) were included in differentiation or N2 medium. The N2 medium was added from day 6 to day 7 and consisted of α-MEM supplemented with 1× N2 (GIBCO), 1 mM sodium pyruvate (GIBCO), 0.1 mM NEAA, and 0.1 mM β-mercaptoethanol. At day 8, EBs were disaggregated in TrypLE (Invitrogen, Grand Island, NY, http://www.invitrogen.com/site/us/en/home.html) and plated on poly(l-ornithine) (0.002%; Sigma)/fibronectin (10 μg/ml; Millipore) or poly(l-ornithine) (0.002%)/laminin (50 μg/ml; Sigma)-coated dishes in OPC medium that consisted of Dulbecco's modified Eagle's medium with F-12 (DMEM/F-12; Hyclone, Logan, UT, http://www.bioresearchonline.com/ecommcenter/hyclone), 1× N2, 1× B27, Pur (1 μM), platelet-derived growth factor-AA (PDGF-AA, 10 ng/ml; Invitrogen), and fibroblast growth factor-2 (FGF-2, 20 ng/ml; Millipore). For further differentiation, PDGF-AA and FGF-2 were excluded from OPC medium and 3,3′5′-triido-l-thyronine (T3, 30 ng/ml; Sigma) was added to differentiate OPCs to oligodendrocytes. For testing the effects of extracellular factors on NaV expression, heregulin (10 ng/ml; Peprotech, Rocky Hill, NJ, http://www.peprotech.com/en-US) or laminin (5 μg/ml) was supplemented to the OPC medium.
Fluorescence-Activated Cell Sorting
We used fluorescence-activated cell sorting (FACS) to estimate the percentage of GFP+ cells and purify OPC populations. Single-cell suspensions were prepared the same way as during passage of mESC-OPCs . Data acquisition and sorting of Olig2-GFP+ cells were performed in a BD FACSCalibur (BD Bioscience, San Jose, CA, http://www.bdbiosciences.com/home.jsp). Acquired data were subsequently analyzed by calculating proportions/percentages and average intensities using the FlowJo software (Treestar, Inc., Ashland, OR, http://www.treestar.com/).
Ectopic expression of NaV1.2 α subunit in mESC-OPCs was achieved using BacMam Na+ channel expression kit (Invitrogen). When the cells reached 70% confluence, the virus with SCN2A or control virus with GFP gene (Invitrogen) was added to the cultures according to the manufacturer's detailed protocol and our previous report .
Highly purified mouse cortex OPCs were routinely obtained using previously published protocols [23-25]. Primary hippocampal neurons were prepared from C57BL/6 mouse fetuses (Charles River, Hollister, CA, http://www.criver.com/) at embryonic day 15. The hippocampi were digested with trypsin and dissociated by trituration in DMEM/F-12 and 10% fetal bovine serum. The hippocampal neurons were plated onto poly(d-lysine) (50 μg/ml; Millipore) coated 12 mm coverslips. At day 1 in vitro (DIV 1), the medium was replaced with neuron culture medium consisted of Neurobasal medium, 1× B27 (Gibco). Cytosine arabinose (1 μM, Sigma) was added to the neurons from DIV 3 to DIV 7 to eliminate proliferating cells. At DIV 7, the nonspiking and spiking (24 hours after viral infection) mESC-OPCs were plated onto the neurons at 2 × 104 cells per square centimeter in the coculture medium consisted of 50% neuron culture medium and 50% OPC medium without FGF2 and PDGF-AA . The cells were either patch-clamp recorded at 2 days after coculture or immunostained for MBP at 10 days after coculture.
Reverse Transcription-Polymerase Chain Reaction
Total RNA was prepared from cells using RNeasy Mini Kit (Qiagen, Valencia, CA, http://www.qiagen.com/). Single-stranded cDNA was synthesized from ∼1 μg of total RNA using random hexamers and SuperScript reverse transcription (Invitrogen) according to the manufacturer's protocols, followed by polymerase chain reaction (PCR) amplification with gene-specific primers using Biolase DNA polymerase (Bioline, Taunton, MA, https://www.bioline.com/h_usa.asp). Primers, annealing temperatures, and product sizes are given in Supporting Information Table S1. The cDNA (1 μl) was replaced by sterile nuclease-free water for negative control in each pair of primers, and no significant band was observed in the negative control. The reaction was conducted using the following protocol: initial denaturing of the template for 5 minutes at 94°C followed by 32 repeating cycles of denaturing for 1 minute at 94°C, annealing for 1 minute, extension for 1 minute at 72°C, and a final elongation at 72°C for 7 minutes. The PCR products were separated by a 1% agarose gel electrophoresis and visualized by ethidium bromide staining.
Whole-cell voltage-clamp recordings were performed at room temperature. Pipette electrodes (TW120F-6; World Precision Instruments, Sarasota, FL, http://www.wpiinc.com/index.php) were fabricated using either a Sutter P-97 horizontal puller or a Narishige PC-10 vertical puller and had final tip resistances of 3–5 MΩ. Membrane currents were measured using a patch-clamp amplifier (Axon 200B, Molecular Device, Sunnyvale, CA, http://www.moleculardevices.com/), sampled, and analyzed using a Digidata 1440A interface and a personal computer with Clampex and Clampfit software (Version 10, Molecular Device). In most experiments, 70%–90% series resistance was compensated. All recordings were performed in external solution consisting of (in mM) NaCl 150, KCl 5, CaCl2 2, MgCl2 1, HEPES 10, and glucose 10, pH adjusted to 7.4 with Tris-base. The osmolarity of the solutions was adjusted to 310–320 mosM/l with sucrose. The internal solution contained (in mM) NaCl 30, KCl 120, MgCl2 1, CaCl2 0.5, HEPES 10, EGTA 5, and MgATP 2, pH adjusted to 7.2 with Tris-base. The osmolarity of the solutions was adjusted to 300–310 mosM/l with sucrose. For recording Na+ currents, K+ in internal solution was replaced with equimolar Cs+. The inward voltage-gated Na+ current was identified by their sensitivity to specific blocker NaV channel blocker, tetrodotoxin (TTX, 1 μM; Abcam Biochemicals, Cambridge, MA, http://www.ascentscientific.com/). For recording K+ currents, 1 μM TTX was added to the perfusing external solution. Current densities (pA/pF) were calculated as peak outward currents divided by the cell capacitance. Spontaneous postsynaptic currents (sPSCs) were recorded from GFP+ mESC-OPCs in the coculture. Combined administration of (2R)-amino-5-phosphonopentanoate (APV) (50 μM; Tocris, Minneapolis, MN, http://www.tocris.com/) and 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) (10 μM; Tocris) was used to block glutamatergic neurotransmission.
All animal experiments were performed following protocols approved by the Animal Care and Use Committee at the University of California, Davis. The shiverer mice at 4–6-week old (Jackson Laboratory, Bar Harbor, ME, http://www.jax.org/) were used for transplantation study. Both spiking and nonspiking mESC-OPCs were suspended at a final concentration of 100,000 cells per microliter in PBS (Invitrogen). Under surgical anesthesia, shiverer mice were placed in a stereotaxic apparatus (David Kopf Instruments, Tujunga, CA, http://www.kopfinstruments.com/) and received 2 μl of cell suspension either injected into the spinal column between T12 and T13 or injected into the brain corpus callosum (anteroposterior: 0.5 mm, lateral: 2 mm, dorsoventral: 2 mm with reference to bregma). For transplantation into the spinal cord, the procedure was performed according to a previous study . Control animals received injections of 2 μl PBS. Starting from 1 day before surgery, all animals received daily injections of 10 mg/kg cyclosporine (Millipore) for the duration of the study. Two or three weeks after transplantation or control injection, animals were deeply anesthetized and transcardially perfused with 4% paraformaldehyde in 0.1 M PBS, pH 7.4. The spinal cords and brains were removed and postfixed for 2 hours in the same fixative, and then placed in 30% sucrose in PBS at 4°C for 24 hours. Transverse sections were cut at 16-μm thickness using a cryostat, then processed for immunohistochemistry, and viewed under a Nikon Eclipse C1 confocal laser-scanning microscope.
Immunostaining and Cell Counting
Cells were cultured on 12-mm glass coverslips (Deckglaser, Germany) and fixed in 4% paraformaldehyde in PBS for 15 minutes at room temperature. After washing with PBS, cells were permeablized for 10 minutes in 0.1% Triton X-100, and incubated for 1 hour in blocking buffer (5% goat or donkey serum and 1% bovine serum albumin) to reduce nonspecific binding. Cells then were incubated in primary antibodies diluted in blocking buffer at 4°C overnight. Appropriate fluorescence-conjugated secondary antibodies were used for single and double labeling. For staining of NG2 and A2B5 antigens, live cells were incubated with primary antibodies for 15 minutes, and then fixed in 4% paraformaldehyde. After permeablization and incubation in blocking buffer as described above, the cells were exposed to appropriate secondary antibodies. For immunostaining on the spinal cord or brain sections of shiverer mice, sections were blocked and incubated with the primary antibodies overnight at 4°C, followed by incubation of appropriate fluorescence-conjugated secondary antibodies. All antibodies were tested for cross-activity and nonspecific immunoreactivity.
The primary antibodies used in this study were listed in Supporting Information Table S2. 4′,6-diamindino-2-phenylindole, dihydrochloride (DAPI, 1:100; Invitrogen) was used to identify the nuclei. Slides or coverslips were mounted with the antifade Fluoromount-G medium (Southern Biotechnology, Birmingham, AL, http://www.southernbiotech.com/). Images were captured using a Nikon Eclipse C1 confocal laser-scanning microscope.
The cells were counted with the ImageJ software (NIH, http://rsbweb.nih.gov/ij/). For cultured cells, at least five fields of each coverslip were chosen randomly and three coverslips in each group were counted. For brain sections, an average of three sections was counted per mouse spanning 160 μm along the antero-posterior axis.
Transmission Electron Microscopy
Mice were perfused and fixed overnight in Karnovsky's fixative (5% glutaraldehyde + 4% formaldehyde in 0.08 M phosphate buffer). The brain samples were postfixed in 2% osmium tetroxide in 0.1 M cacodylate buffer. Samples were subsequently dehydrated gradually from 50% to absolute ethanol, placed in propylene oxide, and then embedded in epon. Semi-thin (1 μm) sections were stained with toluidine blue to aid in orientation of white matter tracts. Ultrathin sections (70 nm) were cut for regions of interest using an ultramicrotome (Leica, Buffalo Grove, IL, http://www.leica-microsystems.com/). Ultrathin sections were collected on Formva-coated single slot copper grids. Ultrathin sections were stained for uranyl acetate and lead citrate and examined in a Philips CM120 Electron Microscope at 80 kV. Low magnification images were taken to view the axon distributions, high magnification images were obtained to show the myelin sheath layers for g-ratio calculation. Images were acquired via a high resolution CCD camera (Gatan, Pleasanton, CA, http://www.gatan.com/) and processed in DigitalMicrograph (Gatan). Images were imported to Adobe Photoshop for adjusting the brightness and contrast and composing figures.
For quantification of myelinated axons, at least 10 low magnification images (×5,000) from three animals (n = 3) were used in control group and groups received spiking and nonspiking mESC-OPCs. Each image covered approximately 180 μm2 region. All the axon profiles (myelinated and unmyelinated axons) in the region were identified based on the ultrastructural features such as neurofilaments and mitochondria, in most case no synaptic contacts . The percentage of myelinated axons (contain at least two myelin sheath layers) was calculated as following: Percentage of myelinated axon = number of myelinated axons/total number of axons × 100%. Quantification of g-ratio was carried out from animals received spiking and nonspiking mESC-OPCs. The myelin sheaths were clearly identified in animals received both types of mESC-OPCs. For each group, at least 20 myelinated axon profiles were identified from three animals (n = 3) and photographed at ×17,500 to ×20,000. Digital photographs showing myelinated axons were imported to Image J (NIH) for g-ratio analysis which was calculated by dividing the diameter of the axon by the diameter of the entire myelinated caliber.
All data represent means ± SEM. Student's t tests were performed to evaluate the statistical significance of difference between two groups. ANOVA followed by Tukey post hoc test was performed for multiple comparisons. Values of p < .05 were considered statistically significant.
Pur- and RA-Induced Differentiation of GFP-Olig2 mESCs Recapitulates the In Vivo Mouse Embryonic Development and Efficiently Generates GFP+/Olig2+ OPCs at Day 30
In the existing protocols [13, 22, 27] for deriving OPCs from mESCs, the first step usually involves the induction Olig2 expression in the mESC-derived neural progenitor cells, because Olig2 is required for oligodendrocyte development in mice . Previous studies show that treatment of RA and Shh efficiently induces neural differentiation of mESCs, which recapitulates in vivo mouse embryonic development at the spinal pMN domain that Olig2+ progenitors sequentially generate motoneurons and OPCs at the early and late stages, respectively [13, 27, 29, 30]. In order to examine whether the differentiation process induced by adding RA and purmorphamine (Pur), a small-molecule activator of the Shh signaling pathway , recapitulates in vivo mouse embryonic development, we induced GFP-Olig2 mESCs differentiation (Fig. 1A) and assessed the composition of the cell populations at the differentiation days 10, 20, and 30. As shown in Figure 1B, 1C, the expression of NG2 and another marker for OPCs, A2B5,  at day 10 was detected at low percentage (7.9% ± 1.7% and 4.9% ± 0.7% for NG2+ and A2B5+, respectively). To examine the percentage of neurons, the neuronal marker βIII-tubulin and the motoneuron lineage marker HB9 were stained. At day 10, βIII-tubulin- and HB9-positive cells were detected at the rate of 12.7% ± 2.0% and 8.2% ± 1.3%, respectively (Fig. 1B, 1C). At day 20 in culture, the percentage of NG-2 and A2B5-positive cells showed a dramatic increase (Fig. 1B, 1C, 54.0% ± 4.2% for NG2+ and 40.1% ± 8.3% for A2B5+). However, after three to four passages, the percentage of βIII-tubulin- and HB9-positive cells decreased at day 20—0.8% ± 0.5% and 0.6% ± 0.3%, respectively, suggesting the decreased capability of generating motoneurons from the Olig2+ progenitors at day 20. At day 30, the percentage of NG2- and A2B5-positive cells reached 81.8% ± 1.3% and 82.8% ± 2.4%, respectively (Fig. 1B, 1C), and βIII-tubulin- and HB9-positive cells were not observed (Fig. 1C). Throughout the differentiation period, no glial fibrillary acidic protein (GFAP)-positive astrocytes were observed (Fig. 1B, 1C). Thus, despite replacing Shh with Pur, mESC differentiation process recapitulates the mouse embryonic development. Moreover, as shown in Figure 1A, 1D, at day 30, 96.4% ± 1.3% of GFP-positive cells were NG2 positive, and 94.8% ± 1.5% of NG2-positive cells were GFP positive. Thus, GFP-positive cells coexpressed Olig2 and NG2, which are the hallmarks of OPCs [33, 34]. The GFP expression also overlapped with A2B5 (Fig. 1A, 1D, 97.0% ± 1.3% of GFP-positive cells were A2B5 positive, and 88.3% ± 4.0% of A2B5-positive cells were GFP positive). Therefore, using this differentiation protocol, we could obtain OPCs from GFP-Olig2 mESCs with a high efficiency in 30 days. We also double-stained cells at days 10 and 20 for Olig2 and GFP, and found that approximately 98% of Olig2 expression colocalized with GFP, indicating that GFP fluorescence reliably marked Olig2 expression in this paradigm (Supporting Information Fig. S1).
GFP+ mESC-OPCs Differentiate into Oligodendrocytes But Show Different Electrophysiological Properties from Brain-Derived OPCs
We next purified mESC-OPCs using FACS based on GFP fluorescence. The FACS analysis showed the percentage of GFP+ cells increased from 49.1% ± 4.0% at day 10 to 80.3% ± 1.3% at day 30 (Fig. 2A, 2B, n = 4, p < .001). The purified day 30 GFP+ mESC-OPCs could be expanded in the OPC medium and could be subjected to the frozen/thaw cycles without losing their capability of proliferation and maturation. Unless otherwise mentioned, the FACS-purified day 30 mESC-OPCs within passages 1–6 were used in the following experiments. As shown in Figure 2C, the purified day 30 mESC-OPCs matured into O4+ and MBP+ oligodendrocytes when cultured in the presence of thyroid hormone T3, which promotes OPCs differentiation into oligodendrocytes . The inactivating inward currents, which resembled the IKir, were also detected in the GFP+ cells cultured in the presence of T3, but not in the bipolar mESC-OPCs (Supporting Information Fig. S2B). These properties thus showed that the mESC-OPCs were capable of differentiating to mature oligodendrocytes in vitro.
To further characterize the properties of mESC-OPCs, we compared the electrophysiological properties of the mESC-OPCs with those of primary cultures of mouse brain OPCs. We recorded 96 GFP+ mESC-OPCs by whole-cell patch-clamp recordings. The mESC-OPCs had an average capacitance of 9.7 ± 0.8 pF and particularly a high resistance of 1.1 ± 0.1 GΩ, a characteristic of OPCs . Next, we tried to record inward currents from mESC-OPCs using Cs+ internal solution which could completely block the outward potassium currents. Unexpectedly, no inward currents could be recorded from any mESC-OPCs, while inward INa could be observed from cultured brain OPCs and were sensitive to 1 μM TTX (Fig. 2D). When held at −60 mV under current-clamp, no spikes could be stimulated from mESC-OPCs even though the membrane potential was depolarized to 0 mV, while slight spikes were observed from cultured brain OPCs (Fig. 2D), which is consistent with previous recordings from rodent OPCs in culture . We also recorded the mESC-OPCs cultured for 50 days, and no INa was observed, suggesting that the lack of INa was unlikely due to a development delay in in vitro culture. The same results were also observed from the mESC-OPCs generated by adding Shh peptide instead of Pur during the differentiation process, suggesting these differences were not resulted from replacing Shh peptide with Pur. To compare the outward potassium currents in mESC-OPCs and cultured brain OPCs, we then recorded the whole-cell voltage-gated ionic currents in the presence of 1 μM TTX with a prepulse to −80 mV. The delayed rectifier potassium current (IKD) was isolated using a prepulse to +20 mV, which inactivated the currents carried by outward inactivating A-type potassium current (IKA). To reveal IKA, the currents elicited following a prepulse to +20 mV were digitally subtracted from those elicited following prepulse to −80 mV. Both IKD and IKA components were detected in mESC-OPCs and cultured brain OPC (Fig. 2E). I–V relationships of IKD and IKA were shown in Figure 2F, 2G, respectively. Notably, the current densities of IKD and IKA were much larger in mESC-OPCs than those in the cultured brain OPCs.
Nonspiking mESC-OPCs Acquire Spiking Properties After Ectopic Expression of NaV1.2 Subunit
To explore whether NaV subunit transcripts could be detected in the mESC-OPCs, we performed PCR on cells at differentiation day 10 and FACS-purified day 30 mESC-OPCs. As shown in Figure 3A, consistent with the immunostaining results showing that neurons existed at day 10 (Fig. 1A), multiple NaV transcripts were identified from day 10 unpurified cells, such as transcripts that encode CNS (SCN2A and SCN3A) and dorsal root ganglion (SCN10A and SCN11A) NaV α subunits. Some other transcripts encoding skeletal (SCN4A) and cardiac (SCN5A, SCN7A) NaV α subunits were also detected. None of β subunit transcripts were detected in day 10 cells. In day 30 purified GFP+ mESC-OPCs, of all sodium channel genes (10 α and 4 β subunits) studied, only the SCN1B transcripts, which encode the NaV β1 auxiliary subunit, were detected. Consistent with the immunostaining results (Fig. 1A and Supporting Information Fig. S1), the Olig2 transcripts were detectable at day 10 but strongly expressed at day 30, and the NG2 transcripts were nearly undetectable at day 10 but highly expressed at day 30. Thus, due to the lack of expression of pore-forming NaV α subunit transcripts, the day 30 mESC-OPCs did not show any detectable INa (Fig. 2D).
We next sought to generate functional INa in the day 30 mESC-OPCs. SCN2A gene, which encodes NaV1.2 α subunit, is expressed in the CNS OPCs and is the most significantly changed sodium channel during the maturation of OPCs to preoligodendrocytes [5, 36]. Thus, we transduced the mESC-OPCs with baculovirus carrying the gene SCN2A. As shown in Figure 3B, the expression of SCN2A gene transcript was detected from 1 to 3 days after viral infection, and then gradually deceased over time in culture and was undetectable at 8 days after viral infection. The high efficiency of viral infection was demonstrated by the observation that the majority of the GFP+ cells were positive for NaV1.2 staining (Fig. 3C). The GFP+ cells remained positive for A2B5, indicating their unchanged identity as OPCs (Fig. 3C). Patch-clamp recordings performed from 1 to 3 days after viral infection also showed that 94.1% of mESC-OPCs (16 of 17) expressed inward currents and all the cells (17 of 17) continued to express outward potassium currents (Fig. 3D). The sensitivity to TTX and I–V relationship of the inward currents indicated their nature of INa (Fig. 3D, 3E). Under current-clamp recording, when the membrane potentials were held at −60 mV, these transduced cells generated single spikes that grew in size as the depolarization was gradually increased (Fig. 3F). When held at −100 mV, the spike threshold was more negative and spike magnitudes were increased. Some of the cells (17.6%, 3 of 17) held at −100 mV fired all-or-none action potentials (Fig. 3F).
We next attempted to search for extracellular factors including growth factors and morphogens that have been implicated in OPC differentiation conditions to see which would induce expression of Nav in the cells and drive the mESC-OPCs to electrical activity. Thus, we examined whether different culture conditions could trigger and maintain the expression of INa in the mESC-OPCs. First, Pur, FGF2, and PDGF were routinely added into the OPC culture medium. However, these mESC-OPCs did not show functional INa, indicating that long-term culture in the presence of the morphogen Shh agonist Pur and the growth factors FGF2 and PDGF did not promote INa expression. As neuregulins critically regulate the survival, mitosis, migration, and differentiation of oligodendrocyte lineage cells [37-39], we next tested the addition of heregulin, a key member of neuregulin-1 family, to the OPC culture medium and cultured the mESC-OPCs for 7–10 days. However, no INa was detected in these mESC-OPCs. In addition, as previous studies have shown that the interaction between axonal laminin and oligodendroglial integrin contributes importantly to the formation of myelin sheath , and laminin modulates the expression of sodium channel-mediated currents in CNS progenitor cells  and epithelial cells , we then cultured the mESC-OPCs on laminin-coated coverslips in the OPC medium containing laminin. Interestingly, after 7–10 days, 7.7% of the mESC-OPCs (1 in 13) expressed TTX-sensitive INa (Fig. 3G). Although the amplitude of INa was small, slight spikes were observed (Fig. 3H), indicating that laminin was capable of inducing electrical activity in mESC-OPCs, but did not work robustly.
Spiking mESC-OPCs Receive Synaptic Input from Neurons and Show Superior Capability of Maturation into Oligodendrocytes In Vitro
Spiking and nonspiking OPCs have been identified in the rodent brain. However, the functional differences between these two subtypes of OPCs are largely unclear. To address this question, we used the mESC-OPCs with distinct electrophysiological properties, that is, spiking and nonspiking mESC-OPCs, to perform comparison studies. To achieve optimum comparisons, we used the spiking mESC-OPCs generated by viral transduction of NaV1.2 α subunit in these experiments. We first cocultured the spiking and nonspiking mESC-OPCs with primary mouse hippocampal neurons. Two days after coculture, we patch clamped both mESC-OPCs identified by GFP fluorescence (Fig. 4A). We observed that approximately 40% (6 of 15) of the spiking mESC-OPCs showed INa after 2 days in coculture and no INa was detected in any nonspiking mESC-OPCs. We also found that no spontaneous postsynaptic currents (sPSCs) were detected in any nonspiking mESC-OPCs cocultured with the neurons. Interestingly, sPSCs were recorded from some of the spiking mESC-OPCs (13.3%, 2 of 15) and the sPSCs were completely blocked by APV and CNQX (Fig. 4B), indicating that the sPSCs were glutamatergic excitatory neurotransmissions. This observation is consistent with and in support of a previous report showing that spiking OPCs, but not nonspiking OPCs, receive synaptic input from axons . We further asked whether the expression of INa in mESC-OPCs could affect their maturation in the coculture. Ten days after coculture, we triple stained the cells with GFP, MBP, and neuronal marker MAP2. As shown in Figure 4C, there were many MAP2+ neurons in the culture and the presence of neurons largely promoted the maturation of both spiking and nonspiking mESC-OPCs into MBP+/GFP+ oligodendrocytes that were readily distinguished from endogenous MBP+ but GFP− (negative) oligodendrocytes in the primary culture. Quantitative analysis showed that the percentage of MBP+/GFP+ cells over total GFP+ cells was significantly higher in the coculture with spiking mESC-OPCs than that in the coculture with nonspiking mESC-OPCs (Fig. 4C, 4D, 10.4% ± 3.9% and 23.4% ± 9.6% for coculture with nonspiking and spiking mESC-OPCs, respectively; p < .05, n = 4). We also examined the coculture of neurons with mESC-OPCs infected with control virus. Similar results to mESC-OPCs without viral infection were obtained, indicating that the observation was not resulted from the viral infection.
Spiking mESC-OPCs Show Better Capacity of Maturation and Myelination After Transplantation into the shiverer Mouse
The shiverer mouse is an autosomal recessive mutant that lacks a functional gene for MBP, an essential component of myelin required for formation of compact mature myelin, and thus displays abnormal and poor CNS myelination [43-45]. The shiverer mice are widely used to study the myelinating capability of exogenously grafted cells . To further examine the capacity of maturation and myelination of the nonspiking and spiking mESC-OPCs in vivo, we first transplanted the mESC-OPCs into the dorsal column spinal cord of shiverer mice. Three weeks after transplantation, the control animals without receiving any grafted cells and animals received transplantation of mESCs-OPCs were euthanized and processed for immunostaining. No tumor formation or overgrowth of the transplanted cells was observed. As shown in Figure 5A, the majority of the GFP+ cells were found in proximity to the injection site. We noticed that although many transplanted spiking and nonspiking mESC-OPCs matured into MBP+ oligodendrocytes, the transplanted cells remained round-shaped and did not form long and “railway track-like” myelin segments  (Fig. 5A), similar to the results obtained by a recent report after transplanting mESC-OPCs into a rat spinal cord injury model . As shown in Figure 5C, there were some GFP+ cells from both cell transplantation groups showing the OPC marker NG2 (Fig. 5B). We also found that very few GFP+ cells (<1%) overlapped with GFAP staining, and there was no significant difference between nonspiking and spiking mESC-OPC groups. Notably, the percentage of GFP+ cells that were MBP+ in the spiking mESC-OPC group was significantly higher than that in the nonspiking mESC-OPC group (Fig. 5C, 88.1% ± 2.5% and 77.2% ± 3.3% for spiking and nonspiking mESC-OPC groups, respectively, p < .05, n = 6 for each group), indicating that spiking mESC-OPCs had better capability of maturing into MBP-expressing oligodendrocytes than the nonspiking OPCs in vivo.
To promote formation of myelin segments and further compare the capability of myelination between spiking and nonspiking mESC-OPCs, we transplanted the mESC-OPCs into the corpus callosum where the bundles of axon fibers are highly enriched. Two weeks after transplantation, the animals were euthanized and processed for immunostaining. We found that only some short myelin segments were observed from the animals received nonspiking mESC-OPCs (Fig. 6A), while long and railway track-like myelin segments were observed from the animals received spiking mESC-OPCs (Fig. 6B). We also noticed that the MBP+ myelin segments did not colabeled with GFP staining but the cell bodies did weakly express GFP, which could be due to downregulation of Olig2 expression as the mESC-OPCs initiate the myelination program . Next, we further examined the myelination from control animals and animals received spiking or nonspiking mESC-OPCs with transmission electron microscopy (TEM). As shown in Figure 6B, only the axons with loose wraps of myelin were seen in the control PBS group. The axons with compact myelin wraps were found in both spiking and nonspiking mESC-OPC groups (Fig. 6C, 6D). Quantitative analysis showed that the percentage of myelinated axons in 180 μm2 was significantly higher in spiking and nonspiking mESC-OPC groups than that in control group (Fig. 6F, 14.1% ± 4.3%, 57.6% ± 8.1%, and 48.7% ± 4.3%, for control, spiking, and nonspiking mESC-OPC groups; p < .05, n = 3 for each group). There was no significant difference in the percentage between spiking and nonspiking mESC-OPC groups, although the mean percentage of spiking group was slightly higher than that of nonspiking mESC-OPC group. Importantly, we found that the g-ratio of spiking mESC-OPC group was significantly lower than that of nonspiking mESC-OPC group (Fig. 6G, 0.80 ± 0.04 and 0.85 ± 0.03 for spiking and nonspiking mESC-OPC groups, respectively; p < .05, n = 3 for each group), indicating that the spiking mESC-OPCs had superior capability of myelination over nonspiking mESC-OPCs in vivo.
In this study, we showed that OPC differentiation of mESCs induced by small molecules RA and Pur recapitulated mouse embryonic development. However, electrophysiologically, unlike brain OPCs, the mESC-OPCs showed a lack of INa expression and were incapable of generating spikes upon stimulation. By generating the spiking and nonspiking mESC-OPCs, we revealed a novel function of NaV of OPCs in their maturation and myelination both in vitro and in vivo.
Using the current small molecule-based protocol, we found that βIII-tubulin+ and HB9+ cells were detected at day 10, and over 80% of the cells expressed OPCs markers NG2 and A2B5 at day 30. This RA- and Pur-induced differentiation from GFP-Olig2 mESCs thus retains the timing observed in embryonic development, and the GFP+ mESC-OPCs provide an enriched source for studying OPC properties. Compared to previous studies on rodent brain OPCs, the mESC-OPCs obtained in this study showed both similarities and differences. Similar to brain OPCs, mESC-OPCs terminally differentiated into mature oligodendrocytes. However, under the identical culture condition, we noticed that mESC-OPCs were less efficient in generating mature oligodendrocytes than brain OPCs. Electrophysiologically, similar to brain OPCs, high input resistance and two types of potassium currents, IKD and IKA, were observed in mESC-OPCs. Consistent with previous studies [4-6, 9, 10, 50-53], we observed that brain OPCs express NaV channels. However, none of the mESC-OPCs showed INa, and the transcripts of pore-forming NaV α subunits were undetectable by PCR.
The roles of NaV channels of OPCs are largely unclear, but studies have suggested their importance in OPC functions. Activation of sodium channels on OPCs contributes to the resting membrane potential of OPCs  that might affect the proliferation of OPCs , and is required for GABA-induced migration of OPCs . Strikingly, recent studies [10, 53] show that a subset of brain OPCs can fire NaV channel-mediated action potentials. Neuronal action potentials are known to promote myelination [56, 57]. However, whether the electrical activities of OPCs could also contribute to the myelination is unclear. In this study, we did not observe significant difference in resting membrane potential and proliferation between spiking and nonspiking mESC-OPCs. Interestingly, we found that despite the lack of pore-forming NaV α subunits, the mESC-OPCs express the β1 subunit that is known to regulate the level of NaV channel expression at the plasma membrane, and function as a cell adhesion molecule to interact with extracellular matrix components [58-61]. Therefore, we ectopically expressed NaV1.2 α subunit in the mESC-OPCs by viral infection. The silent nonspiking mESC-OPCs acquired spiking property after the viral infection. Furthermore, using a coculture system, we observed that the spiking but not nonspiking mESC-OPCs received synaptic input from neurons and showed superior capability of maturing into MBP+ oligodendrocytes, compared to nonspiking mESC-OPCs. After transplantation into the shiverer mice, both spiking and nonspiking mESC-OPCs can wrap the axon and form multilayer compact myelin, suggesting that functional NaV channel expression in mESC-OPCs is not required for OPC maturation and myelination. However, the spiking mESC-OPCs showed better capability in myelinating the axons, indicated by the TEM analysis. Mechanistically, we showed that the expression of functional NaV might be required for introducing the synapse formation between spiking mESC-OPCs and neurons. The synaptic input to spiking OPCs confers them a better sensitivity to the environment [10, 62] and we propose that this synaptic input may further promote their maturation and myelination. To more definitively address the functional roles of NaV channels of OPCs, it will be necessary to generate conditional knockout mice of NaV channels in OPCs. In contrast to the previous demonstration that OPCs possess robust spiking properties at either postnatal stages or adulthood [6, 9, 10], a recent study  revealed that OPCs in NG2-dsRed transgenic mice only showed tiny Na+ spikes at the first postnatal week and became completely nonexcitable in adulthood. Further studies are required to clarify the biological functions of NaV channel of OPCs in vivo.
Previous studies have shown that without introducing NaV channel, OPCs differentiated from PSC produce substantial amount of myelin sheaths in rodents. However, these studies usually injected the PSC-derived OPCs into animals at very early age (embryos, neonate, or less than 7-day-old animal) [12, 63] or into immune-deficient animals  to avoid immune rejection and to achieve substantial myelination. In this study, we used 4–6-week-old shiverer mice that had intact immune system. Although immunosuppressant were given daily to animals, the immune rejection might still to a large degree exist, as particularly reflected by the observation that the transplanted cells in the spinal cord remained round shape and did not form railway track-like myelin sheets, similar to the results obtained by a recent report on transplanting the same mESC-OPCs into a rat spinal cord injury model . Therefore, there is much potential that could be developed to promote maturation and myelination of graft PSC-derived OPCs in a system with intact immune reaction. Here, we demonstrated that NaV channel that afforded the cells spiking properties might be one of such factors that could improve the myelinating potential of grafted OPCs. In the future, it would be interesting to expand upon this study to investigate the electrophysiological properties of human PSC-derived OPCs and to improve the myelination capacity of human OPC after transplantation. In this study, we found that, although not working robustly, laminin might help promote the expression of INa in the ESC-OPCs. It would be interesting to more exhaustively search for factors that can robustly trigger and maintain NaV expression in ESC-OPCs, which would help develop optimal differentiation conditions for deriving functional spiking OPCs from human PSCs in the future.
We have characterized electrophysiological profiles of mESC-OPCs in comparison with brain-derived OPCs. Our results show that, while primary OPCs have outward potassium currents including IKD and IKA, and TTX-sensitive inward INa, only outward potassium currents are detected in mESC-OPCs. We further demonstrate that spiking properties of mESC-OPCs can be restored, and functional maturation and myelination capability of these cells can be enhanced in vivo by ectopic expression of the NaV α subunit. Our functional characterization of mESC-OPCs and their differentiation capacity provides new insights into the mechanisms of OPC differentiation and future PSC-based therapeutic applications.
This work was in part supported by grants from National Institutes of Health (R01NS061983 and R01ES015988 to W.D.), the National Multiple Sclerosis Society (to W.D.), and Shriners Hospitals for Children (to W. D.). P.J. is a recipient of a postdoctoral fellowship from Shriners Hospitals for Children. C.C. is supported by a postdoctoral fellowship from California Institute for Regenerative Medicine. V.S. is currently affiliated with the Department of Animal Science, Cornell University, New York, NY.