Disclosure of potential conflicts of interest is found at the end of this article.
Author contributions: S.W.: collection and/or assembly of data, manuscript writing; H.H.: collection and/or assembly of data; Y.A.: collection and/or assembly of data; C.T.: collection and/or assembly of data; M.V.: collection and/or assembly of data; C.K.: provision of study material or patients; J.R.D.: provision of study material or patients; A.A.: conception and design, financial support, collection and/or assembly of data, data analysis and interpretation, manuscript writing, final approval of manuscript. S.W. and H.H. contributed equally to this work.
First published online in STEM CELLSEXPRESS January 14, 2011.
Expression of the four transcription factors, that is, Oct4, Sox2, cMyc, and Klf4 has been shown to generate induced pluripotent stem cells (iPSCs) from many types of specialized differentiated somatic cells. It remains unclear, however, whether fully committed skeletal muscle progenitor cells (myoblasts) have the potency to undergo reprogramming to develop iPSCs in line with previously reported cases. To test this, we have isolated genetically marked myoblasts derived from satellite cell of adult mouse muscles using the Cre-loxP system (Pax7-CreER:R26R and Myf5-Cre:R26R). On infection with retroviral vectors expressing the four factors, these myoblasts gave rise to myogenic lineage tracer lacZ-positive embryonic stem cell (ESC)-like colonies. These cells expressed ESC-specific genes and were competent to differentiate into all three germ layers and germ cells, indicating the successful generation of myoblast-derived iPSCs. Continuous expression of the MyoD gene, a master transcription factor for skeletal muscle specification, inhibited this reprogramming process in myoblasts. In contrast, reprogramming myoblasts isolated from mice lacking the MyoD gene led to an increase in reprogramming efficiency. Our data also indicated that Oct4 acts as a transcriptional suppressor of MyoD gene expression through its interaction with the upstream enhancer region. Taken together, these results indicate that suppression of MyoD gene expression by Oct4 is required for the initial reprogramming step in the development of iPSCs from myoblasts. This data suggests that the skeletal muscle system provides a well-defined differentiation model to further elaborate on the effects of iPSC reprogramming in somatic cells. STEM CELLS 2011;505–516
Induced pluripotent stem cells (iPSCs) have been generated from somatic cells reprogrammed by transduction of four transcription factors, that is, Oct4, Sox2, cMyc, and Klf4 . These iPSCs have been derived from fibroblasts, B lymphocytes , natural killer T cells , stomach cells, liver cells , pancreatic β-cells , and neural stem cells . The iPSCs generated from adult tissues have been thought to mimic ESCs and hold tremendous therapeutic promise due to their ability to proliferate, differentiate, and self-renew. However, many of the molecular mechanisms of the reprogramming process, particularly tissue-specific factors limiting pluripotency, remain undefined.
Along these lines, recent work demonstrated that the efficiency of iPSC reprogramming is dependent on the differentiation status of the cell source. For example, stem cells and progenitor cells are more readily reprogrammed to iPSCs compared with differentiated cells . In addition, reprogramming mature B cells to iPSCs requires additional interruption of the transcriptional state that had maintained the cells' identity . These results suggest that the promotion of differentiation inhibits the reprogramming process. Furthermore, cells inherit a stable genetic program partly through various epigenetic markers such as DNA methylation and histone modifications. This cellular memory needs to be erased during genetic reprogramming for the cellular program to revert to that of an earlier developmental stage . However, the relationship between the transcriptional activity of the four factors and the cellular memory status remains unclear.
Adult skeletal muscle possesses extraordinary regeneration capabilities. After exercise or muscle injury, large numbers of new muscle fibers are normally formed within a week due to the expansion and differentiation of muscle satellite cells . Satellite cells are a small population of myogenic stem cells for muscle regeneration, which are normally mitotically quiescent . Following injury, satellite cells initiate proliferation to produce myogenic precursor cells, or myoblasts, to mediate the regeneration of damaged muscle tissue [9, 11]. The myoblasts undergo multiple rounds of cell division prior to terminal differentiation and formation of multinucleated myotubes by cell fusion.
Our knowledge of the molecular mechanisms regulating skeletal muscle development and differentiation has been dramatically accelerated following the discovery of the MyoD-family of master transcription factors that can convert a broad range of cells to a myogenic lineage . MyoD belongs to the myogenic regulatory factors, a group of skeletal muscle-specific basic helix-loop-helix transcription factors that play essential roles in myogenic specification and differentiation during embryogenesis and satellite cell development [13–15]. Satellite cell-derived primary myoblasts isolated from adult mice lacking the MyoD gene (MyoD−/−) display accelerated growth rates as well as delayed terminal differentiation and muscle regeneration [13, 16–19]. MyoD−/− myoblasts have been shown to display characteristics that are more primitive than wild-type myoblasts and may represent an intermediate stage between stem cells and myogenic precursor cells [13, 19].
To address whether myoblasts, or fully committed myogenic cells, can be reprogrammed to iPSCs and whether MyoD inhibits the reprogramming process of iPSC generation, we examined the induction of iPSCs from myoblasts isolated from both wild-type and MyoD−/− mice by infection with retroviral vectors containing the Oct4, Sox2, cMyc, and Klf4 genes.
MATERIALS AND METHODS
MyoD−/− mice  were kindly provided by Dr. Michael A. Rudnicki. B6;129S4-Pou5f1tm2(EGFP)Jae/J (Oct4EGFP/+ or Oct4-GFP) mice  were purchased from Jackson Laboratory (Bar Harbor, ME, www.jax.org). MyoD−/−:Oct4-GFP mice were established by crossing MyoD−/− mice with Oct4-GFP mice. MyoD−/−:Oct4-GFP mice and their litter mates MyoD+/+:Oct4-GFP mice were used for isolation of MyoD−/− and wild-type primary myoblasts, respectively. Myf5B6;129S4-Myf5tm3(cre)Sor/J (Myf5Cre/+ or Myf5-Cre) mice  and 129S-Gt(ROSA)26Sortm1Sor/J (ROSA26 reporter or R26R) mice  were also purchased from Jackson Laboratory. Pax7CreERp/+ (Pax7-CreER) mice were obtained from Dr. Charles Keller . Myf5-Cre:R26R mice or Pax7-CreER:R26R mice were established by crossing Myf5-Cre or Pax7-CreER and R26R mice, respectively. Genotyping to detect the mutated alleles of MyoD−/−, Oct4-GFP, Myf5-Cre, and R26R mice was performed by polymerase chain reaction (PCR) using primers described on the website of Jackson Laboratory. Genotyping for Pax7-CreER mice was described previously . B6;129-Gt(ROSA)26Sortm1(rtTA*M2)Jae, Col1a1tm2(tetO−Pou5f1)Jae/J (iOct4) mice were also purchased from Jackson Laboratory . Nod:Scid immunodeficient mice and ICR (CD1) mice were purchased from Charles River (Wilmington, MA, www.criver.com). The Institutional Animal Care and Use Committee for the University of Minnesota approved all experimental protocols. Care was taken to minimize the number of animals used, as well as any pain and suffering.
Myoblast Isolation and Culture
Quiescent satellite cells were isolated from the hind limb skeletal muscle of 1- to 2-month-old adult mice after muscle digestion with collagenase type B and dispase II (Roche Applied Science, Basel, Switzerland, www.roche-applied-science.com) [26–28]. Fluorescence-activated cell sorting (FACS) was performed on a FACS Aria sorter equipped with triple lasers (BD Biosciences, San Jose, CA, www.bdbiosciences.com) as described previously [28, 29]. The following antibodies were used for FACS sorting: Integrin α7 (MBL International, Woburn, MA, www.mblintl.com) with Alexa 488-labeled anti-mouse IgG, biotin-labeled Integrin β1 with allophycocyanin (APC)-labeled avidin, phycoerythrin (PE)-labeled CD45, PE-labeled CD31, and PE-labeled Sca-1 (all from BD Biosciences). Mouse and rat normal IgG (BD Biosciences) were used in the control experiment. Sorting gates were strictly defined based on single antibody-stained control cells, as well as the forward scatter (FSC) and side scatter (SSC) patterns of satellite cells. After the FSC/SSC gating, the triple-negative cells for CD45-PE, CD31-PE, and Sca-1-PE were gated out. Finally, double-positive cells for Integrin α7-Alexa 488 and Integrin β1-APC were sorted to enrich for satellite cells. Sorted satellite cells were cultured on collagen-coated dishes in myoblast growth medium consisting of HAM's F-10 medium (Invitrogen, Carlsbad, CA, www.invitrogen.com) supplemented with 20% fetal bovine serum (FBS; Fisher Scientific, Pittsburgh, PA, www.fishersci.com), penicillin/streptomycin (50 μg/ml; Invitrogen), and 5 ng/ml basic fibroblast growth factor (R&D Systems, Minneapolis, MN, www.rndsystems.com). Differentiation medium, Dulbecco's modified eagle medium (DMEM; Invitrogen) supplemented with 5% horse serum (Invitrogen), and penicillin/streptomycin (50 μg/ml; Invitrogen), was used to induce myogenic differentiation. For genetic lineage tracing, Pax7-CreER:R26R mice were given intraperitoneal (I.P.) injections of tamoxifen (Sigma Aldrich, St. Louis, MO, www.sigmaaldrich.com; 300 μl, 10 mg/ml, diluted in corn oil [Sigma-Aldrich]) daily for 3 days. Two days after the third injection, hind limbs were harvested for myoblast isolation . For inducible Oct4 expression, doxycycline (Sigma-Aldrich) was administrated at a concentration of 2 μg/ml. To ensure that the muscle cells retained physiological characteristics, all experiments were performed using cells that had been passaged between 2 and 4 times.
Preparation for Murine Embryonic Fibroblasts, Tail Tip Fibroblasts, and Muscle-Derived Fibroblasts
For murine embryonic fibroblast (MEF) isolation, uteri isolated from 13.5-day-pregnant mice were dissected as described previously  and γ-irradiated for feeder cell preparation. To establish tail tip fibroblasts (TTFs), the tails from adult mice were used as described previously . For isolation of muscle-derived fibroblasts, hind limb skeletal muscle derived from adult mice was digested with collagenase type B and dispase II (Roche Applied Science, Basel, Switzerland, www.roche-applied-science.com) as described previously . Dissociated cells were cultured on tissue culture plates (noncollagen coated) to obtain MEFs, TTFs, and muscle-derived fibroblasts and were maintained in DMEM supplemented with 10% FBS.
Generation and Culture of iPSCs
The pMX-based retroviral vectors encoding the mouse cDNAs of Oct4, Sox2, cMyc, and Klf4  (Addgene, Cambridge, MA, www.addgene.org) were kindly provided by Dr. Toshio Kitamura . The pMX-based retroviral vectors for MyoD expression were constructed by insertion of a mouse MyoD cDNA fragment into the EcoRI site of the pMX vectors. These pMX-based retroviral vectors were separately cotransfected by packaging defective helper plasmids into PlatE cells  using Lipofectamine 2000 Reagent (Invitrogen). Forty-eight hours after transfection, virus supernatants were harvested. Low-passage (2–3 passages) myoblasts or fibroblasts were seeded at a density of 2–5 × 105 cells per 6-cm plate 24 hours prior to viral infection. Myoblasts or fibroblasts were incubated with supernatants of retroviral vectors containing the four factors (Oct4, Sox2, cMyc, and Klf4) or five factors (Oct4, Sox2, cMyc, Klf4, and MyoD) with 10 μg/ml of polybrene (Millipore, Bedford, MA, www.millipore.com) for exactly 3 hours. Then, the medium was replaced with myoblast growth medium and cultured for 3 days. The infected cells were replated onto MEF feeder cells and cultured on tissue culture dishes (BD Biosciences, San Jose, CA, www.bdbiosciences.com) coated with 0.1% gelatin (Sigma-Aldrich) in ESC medium containing Knockout DMEM (Invitrogen) supplemented with 20% Knockout serum replacement (Invitrogen), 0.1 mM 2-mercaptoethanol (Sigma-Aldrich), 0.1 mM nonessential amino acids (Invitrogen), penicillin/streptomycin (50 μg/ml; Invitrogen), and 5,000 U/ml leukemia inhibitory factor (Millipore). Two weeks after subculture, Oct4-GFP+ and ESC-like colonies were subcloned onto MEF feeder cells as iPSC colonies. Generated iPSCs were maintained on the MEF feeder cells in ES medium.
Histochemical Staining and Immunostaining
The cultured cells and embryos were stained for β-galactosidase expression as described previously [33, 34]. Alkaline phosphatase (AP) staining was performed as described previously . For immunostaining, cells were fixed with 2% paraformaldehyde (PFA) and incubated with primary antibodies followed by secondary antibodies as described previously . The following primary antibodies were used: anti-Nanog (Bethyl Laboratories, Montgomery, TX, www.bethyl.com), anti-Oct4 (Santa Cruz, Santa Cruz, CA, www.scbt.com), anti-Sox2 (R&D Systems), anti-cMyc (Abcam), anti-Klf4 (Santa Cruz), anti-stage specific embryonic antigen-1 (SSEA-1) (Developmental Studies Hybridoma Bank, Iowa City, IA, dshb.biology.uiowa.edu), anti-MyoD (BD Pharmingen, San Diego, CA, www.bdbiosciences.com), anti-sarcomeric myosin heavy chain (MHC; Developmental Studies Hybridoma Bank), and anti-Pax7 (Developmental Studies Hybridoma Bank). The following secondary antibodies were used: Alexa 488-conjugated anti-mouse IgG, Alexa 594-conjugated anti-mouse IgM, and Alexa 594-conjugated anti-rabbit IgG (all from Invitrogen). 4′,6-diamidino-2-phenylindole (DAPI) (Sigma-Aldrich) was used for counter staining of nuclei. Fluorescence images were captured by a DP-1 digital camera attached to a CKX31 inverted florescence microscope with ×20 and ×40 LUCPlanFLN objectives and a BX51 fluorescence microscope with ×20 and ×40 UPlanFLN objectives (all from Olympus America, Center Valley, PA, www.olympusamerica.com). Photoshop CS2 (Adobe Systems, San Jose, CA, www.adobe.com) was used for image processing.
Genomic PCR and RT-PCR
Preparation of genomic DNA was carried out as reported previously . Total RNA was isolated from cells by TRIZOL (Invitrogen). Purified total RNA was reverse-transcribed (Roche Applied Science: Transcriptor First Strand cDNA Synthesis Kit) and 20–35 cycles were performed (Eppendorf Thermal Cycler, Hauppauge, NY, www.eppendorf.com) using the gene-specific primer pairs described in Supporting Information Fig. S1. Optimal PCR cycles for each pair were determined by several different amplifications of the PCR products.
Embryoid Body Formation
The formation of embryoid bodies was carried out as described previously . Briefly, undifferentiated iPSCs were cultured in hanging drops for 3 days at a density of 1,600 cells per 20 microliter of differentiation medium, which consisted of DMEM supplemented with penicillin/streptomycin (Invitrogen), 0.1 mM nonessential amino acids (Invitrogen), 0.1 mM 2-mercaptonethanol (Sigma-Aldrich), 5% horse serum (Invitrogen), and 10% FBS (Fisher Scientific). The embryoid bodies were transferred to suspension cultures for an additional 3 days. Finally, the embryoid bodies were plated into differentiation medium on 48-well plates coated with Matrigel (all from BD Biosciences). The medium was changed every 3 days.
Cultured iPSCs (1 × 106 cells) on MEF feeder cells were subcutaneously injected into 2- to 3-month-old Nod:Scid immunodeficient mice. Three weeks after injection, teratomas were fixed with 4% PFA and frozen in O.C.T.-compound (Fisher Scientific). Cryosections were stained with H&E solution (Sigma-Aldrich).
Aggregation of iPSCs with Zona-Free Embryos
Eight-cell stage embryos were flushed out from 2.5 dpc pregnant CD1 female mice. Clumps of iPSCs (2–20 cells per clump) were selected after a brief period of trypsinization. These selected iPSC clumps and zona-free morula embryos were transferred into micro drops of KSOM medium (Millipore) supplemented with 10% FBS under mineral oil (Fisher Scientific). After overnight culture at 37°C in 5% CO2 incubator, these aggregated blastocysts were transferred into uterine horns of 2.5 dpc pseudopregnant recipient CD1 female mice or into oviducts of 0.5 dpc pseudopregnant recipient CD1 female mice .
Luciferase Reporter Assays
The firefly luciferase reporter genes 4R-SV-luciferase [39, 40], proximal regulatory region (PRR)-luciferase, distal regulatory region (DRR)-PRR-luciferase , and F3-2.5-luciferase , kindly provided by Dr. David J. Goldhamer, were used in this study. pRL-TK (Promega, Madison, WI,www.promega.com) was used as an internal control. Myoblasts were transfected with expression vectors for MyoD (pMX-MyoD), Oct4 (pMX-Oct4), empty vectors (pMXs), and the luciferase reporter genes using Lipofectamine 2000 (Invitrogen). Cells were harvested at 48 hours after transfection. Luciferase activity was measured with a plate reader (LD400; Beckman Coulter, Brea, CA, www.beckmancoulter.com) using a dual luciferase reporter assay system (Promega).
Chromatin Immunoprecipitation Assay
Chromatin immunoprecipitation (ChIP) was performed on iOct4 myoblast cultures using a ChIP assay kit (Millipore) and the manufacturer's instructions . Myoblasts infected with empty control pMXs or pMX-MyoD retrovirus vectors were harvested at 24 hours after treatment with or without doxycycline. Immunoprecipitation was done overnight at 4°C with control rabbit IgG (BD Biosciences), anti-MyoD (Santa Cruz), and anti-Oct4 antibodies (Santa Cruz). The PCR primers, core enhancer region (CER)-IP, DRR-IP, PRR-IP, and immunoglobulin heavy chain (IgH)-IP, used for this assay are described in Supporting Information Fig. S1.
All data are expressed as means ± SEM. Statistical significance between groups was analyzed by Student's t test. At least three independent experiments were performed and values denote mean ± SEM (n = 3). Asterisk and double asterisk indicate experimental pairs where differences between the compared values were statistically significant (*, p < .05 and **, p < .01, respectively).
Generation of Myoblast-Derived iPSCs
To address whether myoblasts or fully committed myogenic cells can be reprogrammed to iPSCs, we examined the induction of iPSCs from primary myoblasts isolated from adult mice. Initially, we isolated and expanded primary myoblasts from adult mice (1–2 months age) carrying Oct4-GFP transgene  by FACS and expanded them as described previously . In addition, to confirm the myogenic origin of the myoblast-derived iPSCs, we used genetic myogenic lineage tracing. We crossed Pax7-CreER or Myf5-Cre mice with ROSA26R (R26R) reporter mice  to generate Pax7-CreER:R26R or Myf5-Cre:R26R compound mice (Fig. 1A). Pax7-CreER mice carry the CreER gene introduced into the Pax7 gene locus . Myf5-Cre mice carry the Cre recombinase gene introduced into the Myf5 gene locus . Tamoxifen was administered to induce gene recombination in Pax7-CreER:R26R mice. Following tamoxifen treatment Pax7-CreER:R26R mice and Myf5-Cre:R26R mice permanently express lacZ in Myf5+ and Pax7+ myogenic cells and their progeny [23, 44]. Immunofluorescent detection in the isolated myoblasts with anti-Pax7, anti-MyoD, and anti-desmin antibodies clearly showed that more than 98% of isolated cells were Pax7+, MyoD+, and/or desmin+ myoblasts (Fig. 1B). The myoblasts promptly differentiated into MHC+ multinuclear myotubes in high cell density or under differentiation medium (data not shown). Primary myoblasts and fibroblasts were purified from Pax7-CreER:R26R and Myf5-Cre:R26R mice. X-gal staining clearly showed lacZ expression in the myoblasts but not in the fibroblasts isolated from Pax7-CreER:R26R and Myf5-Cre:R26R mice (Fig. 1C).
Isolated Pax7-CreER:R26R, Myf5-Cre:R26R and Oct4-GFP  myoblasts were infected with pMX-retroviral vectors carrying four transcription factors (Oct4, Sox2, cMyc, and Klf4), as described previously . As a control, muscle-derived fibroblasts and tail tip fibroblasts were also used for infection with retroviral vectors carrying these four transcription factors. ESC-like colonies were normally observed by day 10–14 after infection for these myoblasts (Fig. 1D) and fibroblasts (data not shown). ESC-like colonies were picked up and expanded to obtain myoblast-derived iPSCs. We observed that these myoblast-derived iPSC colonies generated from Oct4-GFP mice clearly expressed green fluorescent protein (GFP) (Fig. 1D). ESC-like iPSC colonies generated from both Pax7-CreER:R26R and Myf5-Cre:R26R myoblasts showed that myoblast-derived iPSC colonies from both compound mice expressed the lacZ lineage tracer (Fig. 1D), confirming that iPSCs generated from primary myoblasts had a myogenic origin. In agreement with previous findings , these iPSC colonies strongly expressed alkaline phosphatase and SSEA1 at the cell membrane and Nanog in the nucleus (Fig. 1D, 1E). In addition, RT-PCR experiments showed that all nine independent myoblast-derived iPSC colonies expressed ESC-specific genes such as endogenous Oct4 and Nanog at a similar level as ESCs (Supporting Information Figs. S2A and S3A). Furthermore, expression of the myogenic markers MyoD and Pax7 was lost, both of which were expressed in the parent myoblasts (Supporting Information Figs. S2A and S3A). Genomic PCR showed that these nine iPSC colonies retained infected retroviral vector DNA for Oct4, Sox2, cMyc, and Klf4 (Supporting Information Fig. S2B). By contrast, exogenous expression from these retroviral vectors was markedly reduced (Supporting Information Fig. S2C). Thus, retroviral vector DNA silencing occurred during iPSC generation, as reported previously . The efficiency of iPSC generation from myoblasts was 0.0019% ± 0.0005% (n = 4). By contrast, the efficiency of iPSC generation from muscle-derived fibroblasts was 0.0047% ± 0.0009% (n = 3). These efficiencies are comparable with the efficiencies associated with iPSCs derived from other murine cell sources [1, 45, 46]. These results clearly indicated that myoblasts were successfully reprogrammed into iPSCs.
Multilineage Differentiation Potential of Myoblast-Derived iPSCs
To determine the differentiation potential of the myoblast-derived iPSCs, we induced differentiation using embryoid body formation method over 15 days. RT-PCR for differentiation-specific gene expression showed that marker genes for endoderm, ectoderm, and mesoderm were detected during embryoid body formation (Fig. 1F).
Next, to examine the developmental potential of the myoblast-derived iPSCs in vivo, we subcutaneously injected the iPSCs into Nod:Scid immunodeficient mice. Three weeks after injection, the iPSCs produced teratomas containing primitive tissues originating from all three germ layers including keratinized epithelium (ectoderm), cuboidal epithelium (endoderm), and skeletal muscle (mesoderm; Fig. 2A). Taken together, both in vitro embryoid body culture and in vivo teratoma formation indicated that myoblast-derived iPSCs are able to differentiate into all three germ layers.
Developmental Potential of Myoblast-Derived iPSCs
To examine pluripotency of the myoblast-derived iPSCs, we aggregated the iPSCs with mouse embryos (CD1 strain) at the eight-cell stage. iPSCs derived from Oct4-GFP myoblasts were able to integrate into the GFP+ inner cell mass of the blastocysts (Fig. 2B). In addition, the presence of GFP+ cells in the gonads of 13.5 dpc embryos (Fig. 2C) indicated that myoblast-derived iPSCs contributed to germ cells, demonstrating their germ cell differentiation potential. We also aggregated iPSCs derived from Myf5-Cre:R26R myoblasts with CD1 embryos. We detected widespread lacZ expression in the whole body of 12.5 dpc chimeric embryos (Fig. 2D). In addition, genomic DNA PCR confirmed that these iPSCs contributed to a wide variety of fetal tissues originating from all three germ layers (Fig. 2E). We also noticed that the iPSCs generated from both Oct4-GFP (data not shown) and Myf5-Cre:R26R (Fig. 2D) myoblasts were capable of contributing to the formation of chimeric mice and germ line transmitted offspring. These data strongly indicate that the iPSCs reprogrammed from myoblasts possess phenotypical and functional characteristics of pluripotent stem cells.
MyoD Suppresses Reprogramming of Myoblasts
As described earlier, the efficiency of myoblast reprogramming to iPSCs was relatively low compared with the iPSCs generated from fibroblasts. We hypothesized that myogenic differentiation status may inhibit the molecular process of iPSC reprogramming by the four factors. Previous studies indicated that MyoD is a potent myogenic master transcription factor that converts many nonmyogenic cells to a myogenic lineage [12, 47]. To examine whether MyoD expression affected the reprogramming efficiency of myoblasts, we created MyoD+/−:Oct4-GFP mice by crossing mice carrying MyoD null alleles (MyoD−/−) and Oct4-GFP mice. Next, MyoD+/−:Oct4-GFP mice were crossed with each other to generate MyoD null (MyoD−/−:Oct4-GFP) and the control MyoD+/+:Oct4-GFP mice. Isolated MyoD−/−:Oct4-GFP myoblasts displayed elongated shapes compared with the rounded control MyoD+/+:Oct4-GFP myoblasts (Fig. 3A) . These MyoD−/−:Oct4-GFP myoblasts clearly expressed the myogenic marker Pax7 but not MyoD. Both MyoD+/+:Oct4-GFP and MyoD−/−:Oct4-GFP myoblasts were used to generate iPSCs by infection with retroviral vectors (Oct4, Sox2, cMyc, and Klf4). Two weeks after infection, both MyoD+/+:Oct4-GFP and MyoD−/−:Oct4-GFP myoblasts generated ESC-like colonies that expressed GFP, alkaline phosphatase, and SSEA1 proteins, along with ESC-specific proteins such as endogenous Oct4 and Nanog (Fig. 3B and Supporting Information Fig. S3B). Interestingly, the efficiency of iPSC generation from MyoD−/−:Oct4-GFP myoblasts was significantly higher than that of MyoD+/+:Oct4-GFP myoblasts (0.0060% ± 0.0009% [n = 4] vs. 0.0025% ± 0.0004% [n = 5], p < .01) (Fig. 3C). The efficiency for the MyoD−/−:Oct4-GFP myoblast-derived iPSCs was comparable with that of MyoD+/+:Oct4-GFP fibroblasts (0.0060% ± 0.0009% [n = 4] vs. 0.0073% ± 0.0010% [n = 3]). The infection efficiencies of retroviruses were not different between these three cell types (Supporting Information Fig. S4A). These results suggested that MyoD could be a negative regulator of iPSC generation.
To investigate whether forced expression of MyoD can inhibit iPSC generation, we infected myoblasts with a retroviral vector for MyoD expression together with the four factors (Oct4, Sox2, cMyc, and Klf4). Two weeks after infection, we counted the number of GFP+ ESC-like colonies generated from control MyoD+/+:Oct4-GFP myoblasts, MyoD−/−:Oct4-GFP myoblasts, and MyoD+/+:Oct4-GFP fibroblasts following infection with or without MyoD retroviral vector (Fig. 3C). Clearly, forced expression of MyoD significantly reduced efficiency of iPSC generation in these cell types. These results strongly suggest that MyoD expression inhibits the myoblast reprogramming process and the downregulation or inactivation of MyoD may be essential for the initial step in reprogramming myoblasts to iPSCs.
Oct4 Suppresses MyoD Gene Expression in Myoblasts
To examine which of the four factors may alter MyoD gene expression in myoblasts, we infected wild-type myoblasts with a single retroviral vector expressing Oct4, Sox2, cMyc, or Klf4, instead of all four factors together. Two days after infection, immunostaining, and RT-PCR experiments clearly showed that most of the myoblasts expressed the transgenes and their proteins in their nuclei (Fig. 4A, 4B). Interestingly, ectopic expression of Oct4 significantly reduced MyoD and Pax7 gene expression in the myoblasts (Fig. 4B). Ectopic expression of Sox2 also reduced the expression of MyoD and Pax7 to some extent. cMyc or Klf4 did not downregulate MyoD or Pax7. These results strongly suggest that Oct4 is the primary factor in the downregulation of MyoD gene expression during iPSC generation.
Therefore, we further characterized Oct4-expressing myoblasts by RT-PCR. Importantly, ectopic expression of Oct4 upregulated many ESC-specific genes including endogenous Oct4, Sox2, and Nanog (Fig. 4C), suggesting that Oct4 alone can partially reprogram myoblasts. Similar upregulation of ESC-specific genes was observed in Oct4-expressing fibroblasts (Supporting Information Fig. S4B). However, infection of retroviral Oct4 expression vector alone could not induce iPSCs from myoblasts and fibroblasts (unpublished observation). Forced expression of MyoD led to the abrogation of the ESC-specific gene expression that was induced by ectopic expression of Oct4. By contrast, myogenic marker genes such as endogenous MyoD and Pax7 were maintained in these cells (Fig. 4C), indicating that MyoD dominantly regulates myogenic status even though Oct4 is expressed.
To further characterize the effects of ectopic expression of Oct4 in myoblasts, we used a doxycycline-dependent Oct4 gene expression system (Tet-ON system) . Primary myoblasts and tail tip fibroblasts were isolated from adult induced Oct4 (iOct4) mice. Immunostaining and RT-PCR experiments clearly showed that the Oct4 expression was quickly induced by doxycycline administration within 6 hours in iOct4 myoblasts (Fig. 5A, 5B). Similar to myoblasts and fibroblasts infected with retrovirus Oct4 expression vector, both iOct4 myoblasts and iOct4 fibroblasts also showed downregulation of MyoD gene expression and upregulation of ESC-specific gene expression within 48 hours after treatment with doxycycline (Supporting Information Fig. S4B). Interestingly, MyoD gene expression was quickly downregulated by doxycycline-induced Oct4 expression within 6 hours (Fig. 5B), indicating that MyoD gene expression may be directly suppressed by Oct4. Oct4 also suppressed Pax7 gene expression but the downregulation was not as drastic compared with MyoD. Immunostaining also showed that both MyoD and Pax7 proteins were downregulated by ectopic expression of Oct4 (Fig. 5C, 5D). However, removal of doxycycline reversed this downregulation leading to a quick recovery of MyoD and Pax7 expression. Forced expression of MyoD could overcome Oct4-dependent downregulation of MyoD and Pax7 (Fig. 5C, 5D).
Ectopic Expression of Oct4 Downregulated MyoD Expression in iOct4 Myoblasts
Therefore, we examined terminal differentiation of Oct4-expressing myoblasts after serum withdrawal. Clearly, MyoD and myogenin expression was downregulated by ectopic expression of Oct4 in differentiated muscle cells. In addition, MHC+ myogenic differentiation was markedly suppressed by ectopic expression of Oct4 (Fig. 6A–6C). When Oct4 expression was eliminated, myogenic differentiation was recovered. In addition, forced expression of MyoD could induce myogenic differentiation even though Oct4 was expressed (Fig. 6A–6C). Taken together, these results strongly suggest that Oct4 is a suppressor for MyoD gene expression and myogenic differentiation. In addition, forced expression of MyoD can overcome Oct4-mediated inhibition of myogenic differentiation.
Next, to determine whether Oct4 directly downregulates MyoD gene expression through suppression of MyoD gene enhancer activities, we tested the luciferase activities of MyoD enhancer-driven luciferase and 4R-SV-luciferase reporter genes. MyoD contains three transcriptional regulatory elements located upstream of the MyoD gene, termed the PRR, the DRR , and the F3 region containing the CER (Fig. 7A) . These three regions are essential for proper MyoD gene expression during muscle development [33, 41, 48, 49]. The 4R-SV-luciferase reporter gene contains 4 E-boxes, MyoD-binding sites, derived from the muscle creatine kinase (MCK) gene enhancer region [39, 40]. Cotransfection experiments demonstrated that Oct4 suppressed DRR-PRR-luciferase and F3-2.5-luciferase activities but did not suppress PRR-luciferase or 4R-SV-luciferase activities in myoblasts (Fig. 7B). The failed suppression of 4R-SV-luciferase activity by Oct4 indicated that Oct4 does not inhibit the MyoD-mediated transcriptional regulation by direct suppression of MyoD protein function or competition for the target DNA binding. Taken together, Oct4 is able to downregulate MyoD gene expression by suppression of MyoD enhancer (DRR and CER) activities possibly by directly binding to these regions . Next, we examined whether Oct4 binds directly to the upstream enhancer region sites of MyoD. A ChIP assay clearly demonstrated that Oct4 binds to the upstream enhancer region sites containing the CER or DRR but not to the sites containing the PRR. MyoD binding to the upstream region sites containing the CER, DRR, or PRR was also detected as described in previous work using ChIP-Seq . MyoD binding to the upstream region sites was abolished by Oct4 expression due to downregulation of MyoD expression (Fig. 7C). However, in forced MyoD-expressing myoblasts, MyoD was still able to bind to the upstream region sites even when ectopic Oct4 expression was induced, indicating that Oct4 did not prevent the binding of MyoD to the upstream region sites. Thus, Oct4 may act as a repressor for MyoD gene expression through direct binding to the MyoD upstream enhancer region sites, possibly interacting with other negative regulator(s). Therefore, during myoblast-derived iPSC generation, MyoD gene expression may be directly suppressed by Oct4 as an initial step of the reprogramming process.
The first important finding in this study was that iPSCs can be successfully derived from MyoD-positive primary myoblasts, fully committed myogenic cells, by retroviral infection of four factors (Oct4, Sox2, cMyc, and Klf4). These myoblast-derived iPSCs display similar characteristics to ESCs. We detected ESC-specific gene expression and proteins in myoblast-derived iPSCs. In addition, myoblast-derived iPSCs formed embryoid bodies and teratomas, and contributed to the production of chimeric mice and the offspring, demonstrating their potential to develop into all three germ layers (ectoderm, endoderm, and mesoderm) as well as into germ cells. These results strongly indicated that the myoblast-derived iPSCs possess pluripotent differentiation potential, consistent with previous studies that cell reprogramming is not restricted to specific cell types [1, 2, 4, 5, 51, 52]. Recent work demonstrates that freshly isolated skeletal muscle precursors (SMPs) can be reprogrammed to iPSCs using doxycycline-inducible lentiviral vectors expressing Oct4, Sox2, cMyc, and Klf4 . SMPs are a muscle stem cell population expressing Pax7 but negative for MyoD . Our genetic lineage tracing studies using Cre recombinase Pax7-CreER:R26R and Myf5-Cre:R26R mice demonstrated that the iPSCs generated from primary myoblasts are truly myogenic in origin as they express myogenic markers Pax7 and Myf5 prior to reprogramming. Therefore, to our knowledge, this study is the first showing that fully committed myogenic cells can be reprogrammed to iPSCs.
The second important finding in this study was that MyoD expression prevents the reprogramming of myoblasts into iPSCs. MyoD is a myogenic master transcription factor whose expression alone can program many nonmuscle cells to myogenic lineage [12, 47]. Infection with MyoD expression vector significantly reduced the efficiency of iPSC generation in myoblasts and fibroblasts. In contrast, the efficiency of iPSC production is higher in myoblasts derived from MyoD−/− mice than myoblasts derived from wild-type mice. These results strongly indicate that suppression of MyoD gene expression in myoblasts is required for the initial step of reprogramming myoblasts to iPSCs. Similar results were obtained from B cell-derived iPSC reprogramming in which B cell-specific transcription factor Pax5 inhibits reprogramming .
We also investigated the mechanism of MyoD gene suppression in myoblasts following retroviral infection of the four factors. Among the four factors, we demonstrated that Oct4 is the primary inhibitor of MyoD gene expression. Ectopic expression of Oct4 in myoblasts immediately shut down MyoD gene expression within 6 hours, possibly through direct transcriptional repression of MyoD enhancers. Our ChIP assay demonstrated that Oct4 is able to bind to the upstream enhancer region sites of the MyoD gene. However, Oct4 did not prevent MyoD from binding to the upstream enhancer region sites of the MyoD gene. Thus, Oct4 may function as a repressor for MyoD gene expression through direct binding to the upstream enhancer region sites, possibly interacting with other negative regulator(s). Recent work demonstrates that ectopic expression of Oct4 suppresses Pax7 gene expression by binding to the upstream site of the Pax7 gene in C2C12 myoblasts . In addition, Oct4 can bind to the upstream sites of Cdx2 and Cldn4, trophectoderm-specific genes, and suppress their expression . More recently, Dax1, a molecule interacting with Oct4, has been identified as a negative regulator of Oct4 in ESCs . Therefore, Oct4-interacting factors such as Dax1 may negatively regulate MyoD gene expression. During myogenesis, Oct4 expression has never been detected by immunostaining or Oct4-GFP transgene expression (unpublished observation). Recent high-throughput DNA sequencing (ChIP-Seq) demonstrates that binding scores for Oct4 at the MyoD or Pax7 locus in ESCs are very low or absent . In contrast, Oct4 binds at the transcriptionally inactive Myf5 locus in ESCs . Currently, there is no data for ChIP-mediated Oct4-binding profiles in muscle cells. One possibility is that Oct4-related transcription factor(s) containing a POU domain and/or a homeodomain such as Hoxa-11 and Hoxa-13 , which are expressed in myogenic lineages can negatively regulate MyoD gene expression by binding to these enhancer regions during muscle development and regeneration.
Recent work demonstrated that the origin of iPSCs affects DNA methylation profiles, thereby altering gene expression and differentiation capacity of the iPSCs [53, 61]. For example, iPSCs derived from blood cells maintain epigenetic memory as residual DNA methylation signatures that are characteristic of their somatic tissue of origin . In addition, these blood-derived iPSCs possess a higher hematopoietic cell differentiation efficiency while restricting alternative cell fates. Therefore, the differentiation potential to specific cell types may differ broadly among iPSCs generated from different sources. In addition, our data showed that MyoD expression is easily suppressed by Oct4 in myoblasts. Nevertheless, MyoD−/− myoblasts display higher efficiency for iPSC generation than wild-type myoblasts, indicating that wild-type myoblasts resist iPSC reprogramming. This discrepancy may be due to differences in gene expression profiles  or epigenetic status, including methylation profiles, between wild-type and of MyoD−/− myoblasts. Currently, we are investigating whether myoblast-derived iPSCs possess any epigenetic memory of myogenic status which might contribute to a higher myogenic differentiation potential than that of nonmuscle cell-derived iPSCs.
Muscle-derived stem cell transplantation is a potential therapeutic application for muscle degenerative diseases including muscular dystrophy . Myoblasts have been used for this purpose clinically and experimentally [19, 64]. A current limitation of myoblast transplantation is that myoblasts expanded ex vivo lose self-renewal activity and do not efficiently contribute to the muscle stem cell pool in the recipient muscle [65, 66]. Factors that reprogram myoblasts to a more dedifferentiated stem cell-like population might maintain self-renewal activity and efficiently contribute to the stem cell pool after cell transplantation [19, 67]. Therefore, understanding the molecular mechanisms underlying the production of myoblast-derived iPSCs may lead to new avenues for the efficient stem cell transplantation of genetically modified myoblasts. Importantly, efficiency of direct skeletal muscle differentiation of ESCs is still very poor . The possibility exists that myoblast-derived iPSCs may maintain epigenetic memory of myogenic status, which can contribute to higher myogenic differentiation potential. Therefore, the use of myoblast-derived iPSCs can be regarded as a potential cell therapy for muscular dystrophy.
We successfully generated iPSCs from myoblasts and proved that these myoblast-derived iPSCs are truly myogenic in origin. Our findings suggest that suppression of MyoD gene expression by Oct4 is required for the initial reprogramming step in the induction of pluripotent stem cells from myoblasts. In addition, the skeletal muscle system is a well-established differentiation model, allowing a more detailed study of somatic cell iPSC reprogramming.
We thank Drs. Michael A. Rudnicki for providing MyoD mutant mice, David J. Goldhamer for providing the luciferase reporter genes, Toshio Kitamura for providing the pMXs vector and PlatE cells, Nobuaki Kikyo for providing MEFs and ESCs, Genya Gekker for technical assistance in the Stem Cell Institute FACS facility, Jesse L. Mull for critical reading of the manuscript, Lucas Greder and Tetsuya Tani for technical assistance. This work was supported by grants from the Muscular Dystrophy Association (MDA), the Nash Avery Foundation, and Gregory Marzolf Jr. Award.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.