Author contributions: R.D.: conception and design, performed experiments, collection of data, data analysis and interpretation, manuscript writing; F.N.C.S.: conception and design, performed experiments, collection of data; A.F.: conception and design, performed experiments, collection of data; W.P.: performed experiments, collection of data; R.K.: performed experiments; M.A.R.: provision of study material; M.K.: provision of study material, data analysis and interpretation; R.C.R.P.: conception and design, financial support, data analysis and interpretation, manuscript writing, final approval of manuscript.
First published online in STEM CELLSEXPRESS March 3, 2011.
Disclosure of potential conflicts of interest is found at the end of this article.
Filipe N. C. Santos,
Department of Anesthesiology, State University of Campinas, Campinas, SP, Brazil
An effective long-term cell therapy for skeletal muscle regeneration requires donor contribution to both muscle fibers and the muscle stem cell pool. Although satellite cells have these abilities, their therapeutic potential so far has been limited due to their scarcity in adult muscle. Myogenic progenitors obtained from Pax3-engineered mouse embryonic stem (ES) cells have the ability to generate myofibers and to improve the contractility of transplanted muscles in vivo, however, whether these cells contribute to the muscle stem cell pool and are able to self-renew in vivo are still unknown. Here, we addressed this question by investigating the ability of Pax3, which plays a critical role in embryonic muscle formation, and Pax7, which is important for maintenance of the muscle satellite cell pool, to promote the derivation of self-renewing functional myogenic progenitors from ES cells. We show that Pax7, like Pax3, can drive the expansion of an ES-derived myogenic progenitor with significant muscle regenerative potential. We further demonstrate that a fraction of transplanted cells remains mononuclear, and displays key features of skeletal muscle stem cells, including satellite cell localization, response to reinjury, and contribution to muscle regeneration in secondary transplantation assays. The ability to engraft, self-renew, and respond to injury provide foundation for the future therapeutic application of ES-derived myogenic progenitors in muscle disorders. STEM CELLS 2011;29:777–790
Skeletal muscle is well recognized for its exceptional regenerative capacity. This tissue is composed of terminally differentiated specialized postmitotic contractile myofibers, myogenic precursors or myoblasts, and a pool of muscle stem cells, known as satellite cells, located beneath the basal lamina of each myofiber  that are particularly responsible for postnatal muscle growth. In mature skeletal muscle, these cells are normally in a quiescent state, being activated in response to muscle damage or disease. Activated satellite cells give rise to proliferating myoblasts, which then fuse to existing necrotic muscle fibers or together to form new myofibers with central nuclei [2–4]. Satellite cells can be defined by their position and phenotypically, by the expression of specific surface markers including M-cadherin , c-Met , CD34 [6, 7], syndecan-4 , and more importantly, Pax7, a paired box homeodomain-containing transcription factor [9, 10]. It has been demonstrated that following the activation and proliferation of satellite cells, a small subset of cells do not undergo the terminal differentiation pathway, but retain the ability to go back to a quiescent state and thus preserve the satellite cell pool [11–13].
This well-orchestrated system of maintaining muscle function throughout life can be disturbed by pathological conditions, in particular, muscular dystrophies. In the case of Duchenne muscular dystrophy (DMD), in which myofibers are dysfunctional due to the lack of dystrophin, repetitive cycles of regeneration/degeneration culminate in exhaustion of the satellite cell pool and loss of functional myofibers, which over time are replaced by fibrotic and adipose tissue [14–18]. Although no effective treatment is available at present, one attractive therapeutic approach is to use cell-based therapies to promote muscle regeneration. Although satellite cells represent the ideal cell population to be used since transplantation of these cells results not only in efficient muscle regeneration but also in engraftment of the satellite cell pool [19–23], a number of caveats limit their therapeutic application. The major challenge in applying these cells for clinical purposes is to obtain enough cells for transplantation as activation and in vitro expansion diminishes their engraftment ability . A similar problem has been observed for murine and human hematopoietic stem cells [24–26].
Because embryonic stem (ES) cells are pluripotent and can be extensively expanded in culture while maintaining their self-renewal and multilineage differentiation potential, an ES cell-based therapy has unique advantages. A hurdle is that on in vitro differentiation of ES cells into embryoid bodies (EBs), skeletal myogenic progenitors are generated very inefficiently with only a minority of cells expressing myogenic markers. This is probably due to the scarcity of paraxial mesoderm and lack of appropriate inductive signals within EBs. Regulated expression of Pax3, a key regulator of embryonic myogenesis, can bypass this deficiency and generate a significant number of early myogenic progenitors, which are endowed with substantial regenerative capacity and the ability to restore muscle function following their engraftment in dystrophic mice . The extensive in vitro self-renewal capacity of Pax3/ES-derived cells as well as the persistence of their engraftment after systemic delivery points to the possibility of self-renewal of these cells in the recipient.
In this study, we sought to determine whether ES-derived myogenic progenitors, in addition to their efficient capacity to generate functional myofibers, also have the ability to seed the muscle stem cell pool. Because of the unique role of Pax7 in postnatal satellite cell maintenance, we directly compare muscle progenitors derived by Pax3 induction with those derived by Pax7 induction. By assessing in situ localization, the ability to respond to reinjury after primary engraftment, and the ability to engraft secondary recipients, we find that both Pax3 and Pax7 promote the derivation of a progenitor population with the ability to seed the muscle stem cell compartment.
MATERIALS AND METHODS
Generation of an Inducible Pax7 ES Cell Line
Pax7-expressing vector was provided by Michael Rudnicki . Pax7 fragment (1,599 bp) was cut from its carrying vector (pBRIT) using EcoRI and XbaI, and then subcloned into the P2lox targeting vector. The inducible Pax7 ES cell line was generated in A2Lox.cre ES cells, an improved version of A2Lox , in which cre is present at the doxycycline-inducible locus before recombination and catalyzes its replacement by the gene of interest. Inducible expression of Pax7 was assessed by Western blot using a monoclonal anti-Pax7 antibody (R&D Systems, Minneapolis, MN, www.rndsystems.com).
Growth and Differentiation of ES Cells
Mouse ES cells were maintained and differentiated as described . To induce Pax7 expression during EB differentiation, doxycyclin (Sigma, St. Louis, MO, www.sigmaaldrich.com) was added to the cultures at 1 μg/ml beginning at day 2 of EB differentiation. To assess the presence of myogenic precursors in these cultures, intact EBs were collected and plated in a 10-cm dish in the same medium (±dox). Outgrowths were then further evaluated for myogenic differentiation as described .
Fluorescence-Activated Cell Sorting (FACS) Analysis and Sorting of EB-Derived Cells
EB cells were collected after a short incubation with Trypsin or phosphate-buffered saline (PBS) without calcium and magnesium supplemented with 1 mM EDTA and 0.5% bovine serum albumin (BSA), washed twice, first with Iscove's modified Dulbecco's media (IMDM) 10% fetal bovine serum (FBS) and then with staining buffer (PBS 2% FBS), suspended in the same buffer containing 0.25 μg/106 cells of Fc block (Pharmingen, Bedford, MA, www.bdbiosciences.com). The following antibodies were added at 1 μg/106 cells in 100 μl of staining buffer, and incubated at 4°C for 30 minutes before washing with the same buffer. For platelet-derived growth factor receptor alpha (PDGFαR) and Flk-1, phycoerythrin (PE)– and allophycocyanin (APC)–conjugated antibodies were used, respectively (eBioscience, San Diego, CA, www.ebioscience.com). We used PE-conjugated anti-CD34 (clone RAM34; eBioscience), anti-Syndecan4 (clone KY/8.2; Pharmingen), and anti-CD44 (eBioscience, San Diego, CA, www.ebioscience.com) antibodies. PE-Cy7-conjugated anti-CD29 (clone KMI6; Pharmingen), APC-conjugated anti-CXCR4 (Pharmingen), and mouse anti-M-cadherin (clone 5; Pharmingen) antibodies. For secondary staining in the case of M-cadherin, we used APC-conjugated goat anti-mouse Ig (Pharmingen) at 1 μg/106 cells in 100 μl of staining buffer. Stained cells were analyzed and sorted on a FACS Aria (Becton–Dickinson, Bedford, MA, www.bdbiosciences.com) after addition of propidium iodide (Pharmingen) to exclude dead cells.
Labeling of ES cells
iPax7 and iPax3 ES cells were labeled with a lentiviral vector expressing green fluorescent protein (GFP) from the ubiquitin C promoter, as previously described [27, 31] or alternatively with a lentiviral vector expressing fire fly luciferase followed by internal ribosome entry site–GFP from the Ubiquitin C promoter. Briefly, vectors were cotransfected with packaging and coat protein constructs Δ8.91 and pVSVG into 293T cells using the FuGENE six transfection reagent (Roche, Basel, Switzerland, www.roche.com). Virus containing supernatant was collected 48 hours after transfection, filtered, and used for transduction of ES cells.
Real-Time Polymerase Chain Reaction Analysis
Real-time polymerase chain reaction (RT-PCR) for muscle-specific genes was performed using probe sets from Applied Biosystems (Carlsbad, CA, www.appliedbiosystems.com), as follows: Pax3 (Mm00435491_m1), Pax7 (Mm00834079_m1), Myf5 (Mm00435125_m1), MyoD (Mm00440387_m1), myogenin (Mm00446194_m), myosin heavy chain (MHC) (Mm01332493_g1), and glyceraldehyde 3-phosphate dehydrogenase (Mm99999915_g1).
All animals were handled in strict accordance with good animal practice as defined by the relevant national and/or local animal welfare bodies, and all animal work was approved by the University of Minnesota Institutional Animal Care and Use Committee. Six- to eight-week-old C57 BL/10ScSN-Dmd mdx/j (X-linked muscular dystrophy; Jackson Laboratories, Bar Harbor, ME, www.jax.org) mice were used for the in vivo studies. Before intramuscular cell transplantation, mice were injured with cardiotoxin, as previously described . For intra-arterial transplantation, following anesthetization of animals with ketamine/xylazine, the contra-lateral femoral artery (left leg) was canulated retrogradely toward the lumbar aorta and then directly injected with cells as previously described . For immunosuppression, mdx mice received a daily dose of 5 mg/kg FK 506 (Tacrolimus; Sigma) intraperitoneally (IP) from the day before cell injection until the time of euthanasia. For the analyses of luciferase-labeled cells in vivo, we used 4- to 6-weeks male nonobese diabetic/severely combined immunodeficient (NOD/SCID) mice from Jackson labs. Briefly, 24 hours after cardiotoxin damage into both tibilalis anterior muscles, luciferase-labeled myogenic progenitors from each cell line (1 × 106 cells per 10 μl PBS) were injected into left tibialis anterior (TA) muscles, whereas the right leg received same volume of PBS as negative control.
For capturing bioluminescence, a Xenogen' IVIS-100 system was used. Briefly, 15 minutes after intraperitoneal injection of luciferin in 100 μl of PBS (1.5 mg/kg, GBT Biotechnology, St. Louis, MO, www.goldbio.com), images were acquired for 2 minutes and total flux (p/s) was used for subsequent data analysis. Images were recorded for 4 weeks after transplantation. Following a second round of cardiotoxin damage, images were captured for 2 more weeks.
Immunofluorescence Staining of Cultured Cells and Tissue Sections
Engrafted muscles were frozen in isopentane cooled in liquid nitrogen. Serial 8- to 12-μm cryosections were collected. For immunofluorescence staining, cells cultured on slides and tissue cryosections were fixed using cold acetone or 4% paraformaldehyde, permeabilized with 0.3% Triton X-100 (Sigma) for 15 minutes and blocked with 3% BSA or mouse on mouse basic kit blocking reagent (Vector Labs, Burlingame, CA, www.vectorlabs.com) for 1 hour, and then incubated with primary antibodies including GFP (1:250, Abcam, Cambridge, MA, www.abcam.com), Pax7 (1:20, Hybridoma bank, Iowa City, IA, dshb.biology.uiowa.edu), Myf5 (1:400, SCBT, Santa Cruz, CA, www.scbt.com), myogenin (1:250, clone F5D, BD Biosciences), MyoD (1:250, clone MoAb 5.8A, BD Biosciences), MHC (1:20) and embryonic MHC (1:20, clone F1.652) (both from Developmental Studies Hybridoma Bank), M-cadherin (1:250, Pharmingen), Ki67 (1:500, Abcam), and dystrophin (1:250, Abcam). Alexa fluor 555 goat-anti-rabbit or mouse, 488 goat-anti-chicken, and 647 goat-anti- rabbit (all from Molecular Probes, Carlsbad, CA, www.invitrogen.com) were used as secondary antibodies. All secondary antibodies were diluted for 1:750 and incubated for 45 minutes at room temperature. 4,6-Diamidino-2-phenylindole (1:5,000; Fluka, St. Louis, MO, www.sigmaaldrich.com) was used to counter-stain nuclei.
Single Myofiber Isolation and Staining
For single myofiber staining, we used a method previously described .
Muscle Digestion for FACS
After harvesting, muscles were minced into small pieces and digested using 0.2% collagenase type I in IMDM for 1 hour at 37°C with frequent triturating by pipette. The digested muscles were then filtered, washed, and stained with specific antibodies as described above.
Sorting of GFP+ Cells for In Vitro Studies and Secondary Transplantation
One month after transplantation of ES-derived myogenic progenitors mdx mice (1 × 106 cells/TA, n = 10 for each cell line), engrafted TA muscles were processed as described above and sorted for GFP+ cells. For in vitro studies, sorted cells were cultured under proliferative myogenic medium containing basic fibroblast growth factor (bFGF) and chicken extract in the presence or absence of Dox for a period of 2 weeks, at which point, dox-induced cultures were switched to myogenic differentiation conditions. For secondary transplantations, 5,000 GFP+ sorted cells per 10 μl PBS were injected into left TAs of secondary mdx mice, whereas right TAs received PBS (n = 4 mice for each cell line). One month later, mice were analyzed for dystrophin and GFP expression.
Muscle Preparation for Mechanical Studies
For the measurement of contractile properties, mice were anesthetized with avertin (250 mg/kg i.p.) and intact TA muscles were dissected and placed in an experimental organ bath filled with mammalian Ringer solution containing (mM): NaCl 120.5; NaHCO3 20.4; glucose 10; KCl 4.8; CaCl2 1.6; MgSO4 1.2; NaH2PO4 1.2; pyruvate 1.0, adjusted to pH 7.4. The chamber was perfused continuously with 95% O2–5% CO2 and maintained at a temperature of 25°C. The muscles were stimulated by an electric field generated between two platinum electrodes placed longitudinally on either side of the muscle (square wave pulses 25 V, 0.2 milliseconds in duration, 150 Hz). Muscles were adjusted to the optimum length (Lo) for the development of isometric twitch force and a 5 minute-recovery period was allowed between stimulations. Optimal muscle length (Lo) and stimulation voltage (25 V) were determined from micromanipulation of muscle length and a series of twitch contractions that produced maximum isometric twitch force. For measuring fatigue time, muscles were stimulated for 1 minute and the time for force to decline to 30% of Fo was measured. In brief, after determination of optimal muscle length (Lo) and measurement of maximum isometric tetanic force, total muscle fiber cross-sectional area (CSA) was calculated by dividing muscle mass (milligram) by the product of fiber length (millimeter) and 1.06 mg/mm3, the density of mammalian skeletal muscle. Specific force (sFo) was determined by normalizing maximum isometric tetanic force to CSA.
Differences between samples were assessed by using the Student's two-tailed t test for independent samples.
Pax7 Induces Myogenesis in EB-Derived Cells
To examine the effect of Pax7 expression on the myogenic differentiation from ES cells, we generated an inducible Pax7 (iPax7) ES cell line using a system similar to that previously described for Pax3 . After cre-mediated integration into a doxycycline-inducible locus, regulated expression of Pax7 in iPax7 ES clones was confirmed by Western blot and immunostaining (Fig. 1A, 1B). Activation of the myogenic program was observed when dox was added to the culture medium but not in the absence (Fig. 1B). As observed with Pax3 , sorting for the PDGFαR+Flk-1− cell fraction on day 5 EBs resulted in a cell population enriched for skeletal myogenic progenitors (Fig. 1C, 1D). A difference between Pax3 and Pax7 is that expansion of Pax7-induced PDGFαR+Flk-1−-derived monolayers required addition of bFGF to the culture medium (Fig. 1E), as observed for satellite cell proliferation , whereas proliferation of Pax3-induced cells did not . Under these proliferation conditions, a majority of Pax7-induced cells expressed Pax7 (95.5% ± 2.3%), Myf5 (79.1% ± 6.8%), and MyoD (46.7 ± 1.2%). MHC, a marker of terminal muscle differentiation, was also present but to a lesser extent (21.3% ± 2.4%) (Fig. 1F, upper panel). This profile changed when cells were exposed to differentiation conditions, which consisted of low-glucose Dulbecco's modified Eagle's medium (DMEM) containing 2% horse serum, and withdrawal of dox and bFGF. Under these conditions, most cells expressed MHC (84.4 ± 5.3%) (Fig. 1F, lower panel), and exhibited a typical and uniform morphology of multinucleated myotubes. The majority of the Pax7-induced ES-derived muscle progenitors expressed the surface markers CD44, CXCR4, M-cadherin, CD29, Syndecan-4, Sca-1, and CD34 (Fig. 1G). These two latter antigens were barely or not detected in Pax3-induced cultures .
To assess comparatively the potential of Pax3 and Pax7 on the activation of the myogenic program, we performed gene expression analyses on proliferating monolayers obtained from Pax3- and Pax7-induced EB-derived PDGFαR+Flk-1− cells side-by-side. As expected, both Pax3 and Pax7 efficiently activated the myogenic program as evidenced by the upregulation of Myf5, MyoD, and myogenin (Supporting Information Fig. S1). In this system, there is no cross induction between Pax3 and Pax7, as Pax3 expression is undetectable in Pax7-induced cultures and vice versa (Supporting Information Fig. S1).
Next, we compared the in vivo regenerative potential of Pax7-induced ES-derived myogenic progenitors to their Pax3-induced counterparts by transplanting them into cardiotoxin-injured mdx mice, a mouse model for Duchenne's muscular dystrophy characterized by the deficiency of dystrophin. To track transplanted cells in vivo, prior to injection, cells were labeled with GFP by lentiviral transduction. Control TA muscles were injected with PBS (contralateral leg). A group of mice were transplanted with iPax3-derived cells as positive control. Mice were treated with daily IP injection of an immunosuppressive agent (tacrolimus) to prevent rejection of nonisogenic cells. Four weeks following transplantation, TAs were harvested and analyzed for assessment of engraftment. Immunostaining showed the presence of dystrophin+ myofibers in muscle sections from mice injected with Pax3- (Fig. 2A, middle panel) or Pax7-induced myogenic progenitors transplanted via intramuscular (Fig. 2A, right panel; Fig. 2B) or intra-arterial administration routes (Supporting Information Fig. S2). Cryosections from PBS-injected mice presented only sporadic revertant dystrophin+ myofibers (Fig. 2A, left panel). Quantification of dystrophin+ myofibers in mdx mice transplanted via intramuscular revealed similar levels of primary engraftment between Pax3- and Pax7-derived cells (15.4% ± 2.09% and 14.1% ± 1.31%, respectively) (Fig. 2C). No tumors were detected within 3 months after cell transplantation (data not shown).
In terms of contractile parameters of treated muscles, Pax7-induced myogenic progenitors demonstrated similar potential to Pax3, as shown by superior isometric tetanic force (Fig. 2D), increased absolute (Fig. 2E), and specific force (Fig. 2F) when compared with their respective contralateral PBS-injected leg. Fatigue time (Fig. 2G, 2H), weight (Fig. 2I), and CSA (Fig. 2J) were not affected, as previously observed for Pax3 following intramuscular cell transplantation .
Satellite Cell Engraftment from Pax3- and Pax7-Induced EB-Derived Cells
To assess to what extent Pax3- and Pax7-induced cells seed the satellite cell compartment and are endowed with the ability to self-renew, we performed in situ analyses of single myofibers, characterization of the donor-derived mononuclear fraction by FACS and ex vivo expansion, secondary transplantation, and response to reinjury (Fig. 3A). As discussed, satellite cells are located beneath the basal lamina closely juxtaposed to the muscle fiber and express Pax7 (Fig. 3B, right panel) [1, 9, 10]. Additional markers to identify these cells, include CD34 [6, 7], Syndecan-4 , and M-cadherin (Fig. 3B, left panel) [33, 34]. Pax3- and Pax7-induced myogenic progenitors that had been labeled with GFP were transplanted into TA muscles of mdx mice (total of four mice for each cell line). After 1 month, muscles were harvested, digested using collagenase, and single myofibers were then analyzed by immunostaining using antibodies to GFP, Pax7, and M-cadherin. As observed in Figure 3C, distinct levels of chimerism could be observed among single myofibers isolated from transplanted muscles, which varied from no chimerism to high levels of GFP. Costaining of GFP with Pax7 and M-cadherin revealed the presence of donor-derived satellite cells in single myofibers isolated from mdx mice that had been transplanted with either Pax3- (Fig. 3D) or Pax7-induced myogenic progenitors (Fig. 3E). These results were confirmed by staining muscle cryosections of transplanted mice, in which donor-derived Pax7+ cells were also detected at the satellite cell position for both Pax3- and Pax7-engrafted muscles (Fig. 4A).
To address the question whether engrafted satellite cells are in a dormant or activated state, we stained transplanted muscles with Ki67, which identifies cycling cells . Although we could easily detect Ki67+ cells in both control and transplanted muscles (Fig. 4B, red arrows), none of these cells were positive for Pax7 (Fig. 4B, white arrows), suggesting that donor-derived GFP+Pax7+, as recipient GFP-Pax7+ cells, are in a quiescent state 1 month after the transplantation. It is important to note that induction was not maintained following cell transplantation, and therefore, results reflect the nature of injected cells. In addition to our in situ results, a set of transplanted mdx mice (two mice for each cell line) was assessed by FACS analyses. Although GFP expression was undetectable in the mononuclear cell (MNC) fraction of control nontransplanted mice (Supporting Information Figs. S3A, S3B; left), FACS analyses of pooled MNCs isolated from mdx mice that had been transplanted in both TAs with Pax3- or Pax7-induced ES-derived myogenic progenitors revealed a very distinctive population of GFP+ cells (Supporting Information Figs. S3A-S3B, right). Consistently, GFP+ cells expressed several surface markers that have been associated with satellite cells, including CXCR4 , M-cadherin , CD34 [6, 7], and Sca-1 (Supporting Information Figs. S3A-S3B, far right panels). Although the latter has being associated with activated satellite cells , our results with Ki67 staining do not support the hypothesis that engrafted satellite cells are in an activated state as they are not cycling (Fig. 4B). This discrepancy is probably due to the heterogeneity within the GFP+ mononuclear fraction, which encompasses not only donor-derived quiescent satellite cells but also myoblasts and interstitial cells, among others, which are possibly labeled by Sca-1.
To further assess whether ES-derived myogenic progenitors can contribute to the satellite cell compartment, we alternatively used Pax7−/− mice as primary recipients, which are devoid of satellite cells [9, 10]. Although these studies were challenged by the extremely low viability of this strain (most of these mice died between 2 and 4 weeks of age), we succeeded in analyzing several Pax7−/− mice that had been transplanted with Pax3 (n = 4)- or Pax7 (n = 2)-myogenic cells at 4 weeks of age. The results from these transplantation experiments confirmed the presence of donor-derived Pax7+ satellite cells in engrafted mice 1 month after injection (Supporting Information Fig. S4).
In Vitro Myogenic Potential of Donor-Derived MNCs
To probe the nature of these GFP+ cells, we reisolated donor-derived GFP+ cells from the muscle mononuclear fraction of mdx mice that had been transplanted with Pax3- and Pax7-induced ES-derived myogenic progenitors, and examined their myogenic potential before as well as following their ex vivo expansion with and without dox induction for a period of 2 weeks. The myogenic nature of GFP+ mononuclear cells, from both Pax3- and Pax7-transplanted mice, was confirmed by the ability of freshly sorted engrafted cells to differentiate in vitro into MHC+ multinucleated myotubes (Fig. 5A). When sorted cells were subjected to in vitro expansion, outgrowth of a myogenic progenitor population was observed for both Pax3 and Pax7 in the presence and absence of dox (Fig. 5B-5D). Although by morphology there were no obvious differences (Fig. 5B), dox-induced ex vivo cultured cells were more proliferative than cells cultured in the absence of dox (Fig. 5C), which on the other hand, contained more myotubes. These observations were corroborated by RT-PCR, which revealed that GFP+ MNCs expanded in the absence of dox expressed myogenin, dystrophin, and MHC but much lower levels of Pax3, Pax7, and Myf-5 than their Pax3- and Pax7-induced counterparts maintained in the presence of dox (Fig. 5D). When subjected to myogenic differentiation conditions, low-glucose DMEM containing 2% horse serum, and withdrawal of dox and bFGF, both Pax3- and Pax7-derived cell preparations expressed high levels of MyoD, myogenin, dystrophin, and MHC (Fig. 5D). FACS analysis revealed that dox-treated cells showed greater expression of CXCR4 and M-cadherin (Supporting Information Fig. S5). Taken these data together, we conclude that the GFP+ mononuclear cells are myogenic.
Regenerative Potential of Engrafted GFP+ MNCs
To test for self-renewal in vivo, we performed secondary transplantations into CTX-injured mdx mice. Therefore, we performed secondary transplantations directly from engrafted GFP+ cells obtained from primary recipient mdx mice that had previously received 106 ES-derived cells per TA 1 month earlier. 5,000 GFP+ cells from iPax3 and iPax7 primary recipients were sorted by FACS and transplanted into the TA muscle of secondary mice (four mdx mice each for iPax3 and iPax7). The contralateral leg of each mouse was injected with PBS as control. One month later, the presence of dystrophin+GFP+ myofibers was assessed in secondary mice for both iPax3- and iPax7 ES-derived primary GFP+ cells (Fig. 6A). Quantification of dystrophin+ myofibers indicated significant levels of secondary engraftment with both Pax3- and Pax7-derived mononuclear cells (10.9% ± 0.9% and 10.4% ± 1.5%, respectively) (Fig. 6A, 6B). Additionally, donor-derived GFP+Pax7+ cells were detected at the satellite cell position for both Pax3- and Pax7-secondary transplantations (Fig. 6C). These results demonstrate the presence of a donor-derived self-renewing cell within the mononuclear fraction of mice transplanted with ES-derived myogenic progenitors. First, we transplanted the ex vivo cultured GFP+ mononuclear cells described earlier, but these gave very low levels of engraftment (data not shown). Loss of engraftability following ex vivo culture of satellite cells has been noted previously .
To investigate whether ES-derived engrafted cells have the ability to respond to new rounds of injury, we transplanted luciferase/GFP-labeled iPax3- and iPax7-ES-derived progenitors into CTX-injured NOD/SCID recipients, which 4 weeks later were subjected to a second round of CTX injury. Luciferase activity tended to increase gradually and reach a plateau during the first month after transplantation of Pax3-induced cells, whereas signal from the Pax7-induced ES-derived cells tended to plateau earlier and decline (Fig. 7A). Reinjury caused an initial reduction of bioluminescence, but engrafted cells in both arms proliferated as evidenced by the overall higher levels of Luciferase activity 2 weeks after injury (Fig. 7A), suggesting that ES-derived engrafted cells are able to respond to reinjury. It is important to note that these mice were not irradiated, and thus, transplanted cells were competing with recipient satellite cells during regeneration. We also performed histological staining for embryonic myosin heavy chain (embMHC), a marker of newly regenerated myofibers (Fig. 7B). The presence of GFP+embMHC+ myofibers, reflecting donor-derived newly formed myofibers, confirmed the ability of engrafted cells to respond to new rounds of injury. We also detected GFP−embMHC+ myofibers, representing host-derived newly formed myofibers, indicating that ES-derived engrafted cells competed with recipient satellite cells to promote the regeneration of reinjured muscles, as mice were not subjected to irradiation. We also observed GFP+embMHC− myofibers which represent fibers that were present before the reinjury (Fig. 7B).
Here, we show that myogenic progenitors obtained from differentiating embryonic stem cells not only contribute to significant functional myofiber engraftment but also seed the quiescent satellite cell compartment, as indicated by the presence of GFP+Pax7+Ki67− cells at the satellite cell position of engrafted myofibers (Fig. 4B). Importantly, engrafted cells have the ability to respond to new rounds of injury, competing with resident satellite cells for the generation of newly formed regenerating myofibers (Fig. 7B). These results are corroborated by the secondary transplantation studies, which clearly confirm the regenerative ability of donor-derived GFP+ mononuclear cells. We observed somewhat lower levels of engraftment if compared with some publications involving the transplantation of satellite cells into irradiated recipients [20, 21]. It is important to emphasize that all the transplantation experiments described here involved nonirradiated mice, and thus ES-derived progenitors had to compete with recipient's satellite cells to promote muscle regeneration. Interestingly, Boldrin and coworkers  have recently compared the transplantation of satellite cells into irradiated versus nonirradiated mdx mice, and observed that the irradiated experimental group contained substantially more donor-derived regeneration (51 ± 26 vs. 7 ± 3 dystrophin+ myofibers, respectively).
More recently two studies have claimed effective myogenic engraftment from differentiating ES cells without the use of genetic manipulation [38, 39]. One group transplanted PDGFαR+ cells, presumptive paraxial mesoderm, into nude mice , whereas the other group transplanted cells positive for SM/C-2.6, an antigen expressed by a fraction of muscle mononuclear cells, which include satellite cells, into conditioned mdx mice . Despite evidence for engraftment, neither study demonstrated that donor-derived cells are capable of substantial regeneration (dystrophin-positive myofibers) or of improving the contractile properties of engrafted muscles. Moreover, the anti-SM/C-2.6 antibody, which seems essential for the purification of putative satellite cell progenitors in these later studies  is not publically available, and the antigen is unknown, which limits its application.
Pax3 is first expressed in the presomitic mesoderm, and subsequently in the dermomyotome, preferentially at dorsomedial and ventrolateral lips, and limbs. Pax7 becomes detectable at the central region of the dermomyotome, and limbs, 24 hours after Pax3. Although Pax7 expression is maintained throughout adult life, Pax3 is mostly downregulated from day 13 onward [34, 40, 41]. The Pax3-null mutation is embryonic lethal due to a number of developmental defects including absence of limb muscles [41–43], whereas mice lacking Pax7 are viable and present no embryonic muscle defect , although their muscle satellite cells are extinguished postnatally [9, 10]. These phenotypes point to distinct spatiotemporal roles for Pax3 and Pax7 in muscle development. Our present findings indicate that despite these developmental differences, both Pax3- and Pax7-induced myogenic progenitors have the ability to self-renew in vivo as well as to produce functional myofibers in primary and secondary assays. Although the presence of donor-derived Pax7+ satellite cells has been considered the gold standard for assuring the long-term regenerative ability of a given cell preparation, recent studies suggest that Pax7 may not be as essential as previously thought for adult satellite cells . We observed that some but not all donor-derived cells associated with myofibers expressed Pax7. On the other hand, donor-derived M-cadherin+ cells were easily detected at the satellite cell position of these muscle grafts (Fig. 3). The Pax7− donor-derived cell population may represent satellite cells that have already committed to differentiate, satellite cells that for some reason, perhaps their embryonic character, do not express or need Pax7, or another type of cell. Along this line, recent evidence suggests that the stem cell pool may be comprised in part of Pax7− interstitial cells . Further work will determine whether the engrafted mononuclear cell population can be fractionated into cells of distinct functional potential; however, it is clear that these engrafted mononuclear cells have significant myogenic regenerative potential. They resemble satellite cells in their ability to respond to reinjury and their ability to engraft secondary recipients.
Although still much needs to be dissected in regard to the nature and regulation of ES-derived myogenic progenitors, the data presented here provide evidence for their long-term repopulation potential. The finding that these cells are able to improve muscle function as well as to seed the muscle stem cell compartment in primary and secondary transplanted mice indicates that a Pax3- or Pax7-engineered ES cell-based strategy may be effective for the treatment of muscular dystrophies.
We thank A. Magli for helpful discussions. The monoclonal antibody to MHC was obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by the University of Iowa. This work was supported by the Dr. Bob and Jean Smith Foundation; by Grant Number AR055299 to R.C.R.P. from NIAMS; and AG034370 to M.K. from NIA at the National Institutes of Health.
DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST
The authors indicate no potential conflicts of interest.