Institute of Reproductive and Developmental Biology, Hammersmith Campus, London, United Kingdom
Department of Obstetrics and Gynaecology, Yong Loo Lin School of Medicine, National University of Singapore & National University Hospital, 5 Lower Kent Ridge Road, Singapore 119074. Telephone: +65-6772-4268; Fax: +65-6779-4753
Cell therapy for degenerative muscle diseases such as the muscular dystrophies requires a source of cells with the capacity to participate in the formation of new muscle fibers. We investigated the myogenic potential of human fetal mesenchymal stem cells (hfMSCs) using a variety of stimuli. The use of 5-azacytidine or steroids did not produce skeletal muscle differentiation, whereas myoblast-conditioned medium resulted in only 1%–2% of hfMSCs undergoing muscle differentiation. However, in the presence of galectin-1, 66.1% ± 5.7% of hfMSCs, but not adult bone marrow-derived mesenchymal stem cells, assumed a muscle phenotype, forming long, multinucleated fibers expressing both desmin and sarcomeric myosin via activation of muscle regulatory factors. Continuous exposure to galectin-1 resulted in more efficient muscle differentiation than pulsed exposure (62.3% vs. 39.1%; p < .001). When transplanted into regenerating murine muscle, galectin-1-exposed hfMSCs formed fourfold more human muscle fibers than nonstimulated hfMSCs (p = .008), with similar results obtained in a scid/mdx dystrophic mouse model. These data suggest that hfMSCs readily undergo muscle differentiation in response to galectin-1 through a stepwise progression similar to that which occurs during embryonic myogenesis. The high degree of myogenic conversion achieved by this method has relevance for the development of therapies for muscular dystrophies.
Skeletal muscle satellite cells have long been considered the only source of stem cells for postnatal muscle regeneration [1, 2]. However, recent reports of muscle formation from bone marrow-derived cells [3, –5] in mouse models suggested that bone marrow may also be an important source of stem cells for therapy of muscular dystrophies. Although the cell type derived from bone marrow (BM) in these studies was not characterized, several lines of evidence suggest that they may be mesenchymal stem cells (MSCs). First, after separating BM cells, transplantation of the adherent fraction into regenerating muscle resulted in higher engraftment than with the nonadherent fraction, the former containing MSCs and the latter hemopoietic stem cells (HSCs) . Next, expression of the protein dystrophin was higher following implantation into dystrophin-null muscle of nonpurified BM compared with marrow enriched for HSCs and thus depleted of MSCs . More recently, Shi et al. demonstrated that the MSCs and not the hemopoietic compartment in human BM is the predominant cell population capable of fusing with myoblasts . Finally, other groups have confirmed myogenic differentiation after MSC transplantation into regenerating muscle [7, 8].
The identification of human fetal MSCs (hfMSCs) [9, 10] raises the possibility of using these cells to repopulate diseased or traumatized muscle. We  have reported that hfMSCs can be readily isolated in first trimester fetal blood, liver, and BM. These cells self-renew extensively and differentiate into multiple mesenchymal lineages [9, 11], as well as nonmesenchymal lineages such as neurons and oligodendrocytes . hfMSCs are readily transducible with integrating vectors without affecting expansion and differentiation capabilities and are thus promising vehicles for autologous ex vivo gene therapy . Autologous use of hfMSCs may now be possible, as ultrasound-guided techniques have already developed to the extent that fetal blood can be sampled in ongoing pregnancies as early as 12 weeks of gestation with <5% loss rate , and advances in imaging and thin-gauge fetoscopy should render hfMSC harvest and intraperitoneal reinfusion feasible in the late first/early second trimester [14, 15]. Intrauterine treatment capitalizes on the immune naïveté of the developing fetus and offers the possibility of treating diseases before pathology manifests . Allogeneic therapy with fetal BM and liver-derived hfMSCs is another promising option, as hfMSCs are HLA-II-negative and inhibit lymphocyte proliferation . Recently, allogeneic hfMSCs have been transplanted in utero in a case of prenatally diagnosed skeletal dysplasia with resultant clinical benefit .
We investigated myogenic differentiation of hfMSCs with a view to eventual use as autologous or allogeneic cells for treatment of progressive degenerating muscle disease. We show that hfMSCs undergo activation of myogenic regulatory factors (MRFs), followed by extensive muscle differentiation on exposure to galectin-1 in a manner akin to embryonic myogenesis. In addition, they contribute to muscle regeneration in both a murine muscle-injury model and the dystrophin-negative scid/mdx mice and their contribution to skeletal muscle regeneration is significantly increased after prestimulation with galectin-1 prior to transplantation. Our findings suggest that hfMSCs are a potential cell source for therapy of degenerative muscular disease or traumatic muscle loss.
Materials and Methods
Fetal blood and tissue and adult BM collection was approved by the Institutional Ethics Committee in compliance with national guidelines regarding the use of fetal tissue for research . All women gave written informed consent for collection and use of fetal and human tissues. All animal procedures were approved by the local ethical review process in accordance with Home Office Project Licenses in the United Kingdom and Italy.
First trimester fetal blood was obtained by ultrasound-guided cardiac aspiration between 7 and 13 weeks of gestation before clinically indicated termination of pregnancy. Fetal gestational age was determined by crown-rump length measurement. Following the procedure, fetal BM was collected from long bones. Adult MSCs were collected from BM obtained at clinically indicated sternotomy or thoracotomy.
Fetal blood was plated at 106 nucleated cells per 100-mm dish  and cultured in D10 (10% fetal bovine serum [FBS] [Stem Cell Technologies, Vancouver, BC, Canada, http://www.stemcell.com] in Dulbecco's modified Eagle's medium [DMEM]-high glucose [Sigma-Aldrich, St. Louis, http://www.sigmaaldrich.com]) supplemented with 2 mM l-glutamine, 50 IU/ml penicillin/streptomycin (Gibco-BRL, Gaithersburg, MD, http://www.gibcobrl.com) at 37°C in 5% CO2. Single-cell suspensions of fetal BM were prepared by flushing long bones with 25-gauge needles into D10 via a 70-μm filter and plating as for fetal blood. After 3 days, nonadherent cells were removed, and the medium was replaced. Cells were trypsinized at subconfluence and used for subsequent experiments. Immunophenotyping was performed at passage 2, and hfMSCs from passage 3–10 (n = 10) were used in experiments (at 15 and 40 population doublings).
Adult BM MSCs (n = 4) were either provided by the Tulane Center for Gene Therapy or obtained by flushing marrow from ribs and cultured as for hfMSC isolation. Immunophenotyping and osteogenic and adipogenic differentiation confirmed MSC properties as previously described .
C2C12 cells  were used as positive controls for immunocytochemical staining. Terminal differentiation was induced by plating C2C12 cells at 8 × 104 cells per cm2 on 0.1 mg/ml Matrigel (BD Biosciences, San Diego, http://www.bdbiosciences.com) in 5% horse serum (HS) (Sigma-Aldrich) for 4 days.
Adipogenic and Osteogenic Differentiation
For osteogenic differentiation, hfMSCs were plated at 2 × 104 cells per cm2 in fibronectin-coated chamber slides and cultured in osteogenic medium (D10 supplemented with 10 mM β-glycerophosphate, 0.2 mM ascorbate, and 10–8 M dexamethasone (Sigma-Aldrich) for 14 days with changes of medium every 3–4 days. Evidence of osteogenic differentiation was sought from Von Kossa staining. For adipogenic differentiation, hfMSCs were plated as above and cultured in D10 supplemented with 5 μg/ml insulin, 10–6 M dexamethasone, and 60 μM indomethacin (Sigma-Aldrich) for 4 weeks with changes of medium every 3–4 days. Adipogenic differentiation was evidenced by the appearance of lipid inclusion vacuoles, which take up the neutral oil red O.
In Vitro Muscle Differentiation
hfMSCs were plated in a variety of densities of 1–4 × 104 cells per cm2 on glass or TPX slides coated with fibronectin, Matrigel, gelatin, or collagen (Stem Cell Technologies). They were then exposed to a variety of concentrations of 5-azacytidine (Sigma-Aldrich; 1–24 μM) for 6–48 hours duration in either 2% FBS or serum-free medium (defined as DMEM with 2 mM l-glutamine, 50 U/ml penicillin, 50 μg/ml streptomycin supplemented with 10 ng/ml platelet-derived growth factor-BB, and epidermal growth factor (Sigma-Aldrich) and ITS-plus (Fisher Scientific International, Hampton, NH, http://www.fisherscientific.com). In some experiments, cells received a further 24-hour exposure to 5-azacytidine 3 days later. Following 5-azacytidine exposure, cells were maintained in serum-free medium for up to 21 days. Adult BM MSCs were used as controls.
Horse Serum, Dexamethasone, and Hydrocortisone.
hfMSCs were plated at 2 × 104 cells per cm2 on fibronectin-coated plates and exposed to 5% HS, 10–7 dexamethasone, and 50 μM hydrocortisone in DMEM with 2 mM l-glutamine, 50 U/ml penicillin, and 50 μg/ml streptomycin (all from Sigma-Aldrich), and medium was changed every 3 days, as described by Zuk et al. [21, 22].
Conditioned medium was harvested from C2C12 cells, cultured at 106 cells per T-75 flask in heat-inactivated D10 medium. After 3 days of culture, the supernatant was collected and filtered (0.22-μm filter) for subsequent experiments. hfMSCs were plated at 2 × 104 cells per cm2 on fibronectin-coated slides with conditioned medium (5%–50%) and serum-free medium. Medium was changed every 3–4 days, and cells were analyzed for expression of the muscle-specific markers desmin and sarcomeric myosin (MF20) after up to 14 days in culture. Adult BM MSCs served as controls.
Galectin-1-enriched medium was derived as previously reported. Briefly, COS-1 cells were transfected with a plasmid encoding a cDNA clone of rat galectin-1. Supernatant was collected and filtered before use. The concentration of galectin-1 in the supernatant was approximately 20 μg/ml, estimated by comparing dot blots with known concentrations of recombinant galectin-1 . hfMSCs (n = 6) were plated at 2 × 104 cells per cm2 on fibronectin-coated chamber slides with varying concentrations of galectin-enriched medium (100–1,000 ng/ml galectin-1) in serum-free medium for 12 days. Identical control experiments were undertaken using medium from nontransfected COS-1 cells. Adult BM-derived MSCs were also plated under similar conditions and exposed to galectin-1.
Evidence of myogenic differentiation was sought by morphological criteria and immunostaining for desmin and sarcomeric myosin at days 0, 3, 6, 9, and 12. Myogenic conversion was assessed by counting the number of cells positive for desmin and MF20 in six randomly encountered low-powered fields (237–654 cells per field). Pax7, MyoD, and Myogenin expression was similarly assessed using immunocytochemical staining.
In Vivo Muscle Differentiation
Two immunodeficient mouse strains were used to investigate in vivo myogenesis. The first model is complement and γ-chain-null/RAG2– (c–/γ–/RAG2–) . Initially, the lower limbs (n = 9) were subjected to 18-Gray irradiation to inhibit endogenous muscle satellite cell proliferation . Three days later, the tibialis anterior muscles of both limbs were subjected to cryodamage by a 10-second application of a cryoprobe. This was repeated twice after thawing of the frozen muscle, before transplanting 5 × 105 hfMSCs in 5 μl of phosphate-buffered saline (PBS) . The right tibialis anterior was transplanted with hfMSCs prestimulated in galectin-1 for 3 days, whereas hfMSCs exposed to nontransfected COS-1 medium was transplanted on the left. Muscles were harvested after 28 days and snap-frozen in cooled isopentane, and 6-μm transverse cryosections were collected on polylysine coated slides. Immunostaining for lamins A/C, spectrin, and desmin was performed as described below.
The scid/mdx , which was derived by crossing dystrophic mdx mice  with scid mice , was used for transplantation. Two-month-old scid/mdx mice were anaesthesized with ketamine hydrochloride (80 mg/kg) and xylazine (10 mg/kg). hfMSCs with or without 3 days of pretreatment with galectin-1 were transplanted into the tibialis anterior muscle as described above (n = 3).
Immunohistochemistry and Cytochemistry
Primary antibodies used for immunochemical analysis were mouse monoclonal anti-desmin (DAKO, Glostrup, Denmark, http://www.dako.com), anti-sarcomeric myosin (MF20), anti-Pax7 (Developmental Hybridoma Bank of Iowa, Iowa City, IA ), anti-Myogenin, anti-MyoD1 (DAKO), anti-human lamins A/C, anti-human spectrin (Novocastra Ltd., Newcastle upon Tyne, U.K., http://www.novocastra.co.uk), and rabbit anti-desmin (Sigma-Aldrich). Immunophenotyping of cells was performed with mouse monoclonal anti-CD14, CD29, CD31, CD34, CD44, CD45, CD49b/d/e (BD Pharmingen, San Diego, http://www.bdbiosciences.com/pharmingen), CD105, SH3, SH4, CD106, vimentin, laminin, fibronectin, Von Willebrand factor (vWF), human leukocyte antigen (HLA)-DR, and HLA-I (DAKO, except that CD105, SH3, and SH4 were from Osiris Therapeutics, Baltimore, http://www.osiristx.com).
Cells were fixed in methanol:acetone 1:1 (vol/vol) for 5 minutes at room temperature and blocked with nonserum protein block (X0909; DAKO). For Pax7, MyoD, and Myogenin, cells were additionally incubated with 0.2% Triton X-100 to permeabilize the nuclear envelope. Thereafter, incubation with the appropriate primary antibody (desmin, 1:200; MF20, 1:10; Pax7, 1:50; MyoD, 1:100; Myogenin, 1:100) for 1 hour at room temperature, with subsequent incubations with a biotinylated secondary antibody (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com) followed by streptavidin-conjugated fluorochromes Alexafluor 488 and 594 (Invitrogen, Carlsbad, CA, http://www.invitrogen.com). Slides were then treated with 4,6-diamidino-2-phenylindole (H-1200; Vector Laboratories) and analyzed by epifluorescence microscopy (Axioscope I; Carl Zeiss, Jena, Germany, http://www.zeiss.com). Images were captured using a cooled charge-coupled device camera and reviewed in Quipps m-FISH software (Vysis, Abott Laboratories, Berkshire, U.K., http://www.vysis.com/).
Muscle sections were air-dried and blocked with 5% goat serum, 5% fetal calf serum, and papain-digested whole fragments of unlabeled secondary anti-mouse immunoglobulin  for 1 hour at room temperature, before incubation with monoclonal anti-lamins A/C (1:400), anti-spectrin (1:20), and rabbit anti-desmin (1:50) antibodies at 4°C overnight. A secondary fluorescein-conjugated goat anti-rabbit IgG (Vector Laboratories) was then used to label the desmin, and a biotinylated goat anti-mouse and streptavidin-conjugated Alexafluor 488 or 594 fluorochrome was used to label lamins A/C and spectrin.
Unstimulated hfMSCs, galectin-1-stimulated hfMSCs, and C2C12 cells were washed with ice-cold PBS, collected (by scraping), and centrifuged. Cell pellets were then resuspended in 1% SDS (Sigma-Aldrich) and heated for 5 minutes at 90°C. Resulting whole cell lysates were centrifuged at 13,000g at 4°C for 5 minutes, and the supernatant was collected. Protein concentrations were calculated by the bicinchoninic acid method (Pierce, Rockford, IL, http://www.piercenet.com). Fifty μg of protein was denatured by boiling for 5 minutes, resolved on 10% SDS-polyacrylamide gel, and transferred to Hybond P membrane (Amersham, GE Healthcare, Buckinghamshire, U.K., http://www.amersham.com/) in a semidry chamber with a three-buffer system as previously described . The membranes were blocked overnight at 4°C in blotto (5% nonfat dried milk, vol/vol, in Tris-buffered saline plus 0.1% Tween). The membranes were incubated in mouse 1:1,000 anti-desmin (DAKO) antibody in blotto for 1 hour, followed by incubation in goat anti-mouse horseradish peroxidase-conjugated antibody (DAKO) in blotto for 1 hour. After washing, chemiluminescence ECL Plus system (Amersham Pharmacia Biotech) was used for signal detection, and the membrane was subsequently exposed onto autoradiography films.
Total RNA was isolated from hfMSCs exposed to galectin-1 for various durations (0–12 days) using RNeasy kit (Qiagen, Hilden, Germany, http://www1.qiagen.com) according to the manufacturer's instructions and quantified spectrophotometrically. Complementary DNA were obtained by reverse transcriptase of 1 mg of total RNA using oligo(dT)15 primer and the reverse transcription system kit (A3500; Promega, Madison, WI, http://www.promega.com). Polymerase chain reactions (PCRs) were carried out using specific primers derived from published cDNA sequences (Invitrogen) (Table 1). Negative controls were either reverse transcriptase (RT) without enzyme or PCR with Mili-Q (Millipore, Brussels, Belgium) ultrapure water instead of cDNA. Amplification of the control RNA without reverse transcription did not generate any products in PCRs (data not shown). Following PCR, the products were electrophoresed in ethidium-bromide agarose gels photographed under UV light. 18SRNA were used as a loading control, and fetal muscle (17 weeks of gestation) was used as a positive control.
Table Table 1.. PCR primers
Parametric variables were expressed as mean and standard deviations and compared by Student's t test or one-way analysis of variance as appropriate. Nonparametric data were compared by the Wilcoxon signed rank test. A p value of <.05 was considered significant.
When 106 mononuclear cells from fetal blood or BM were plated, adherent spindle-shaped cells were seen, which rapidly formed colonies over 2–3 days. After removal of nonadherent cells, these colonies grew rapidly to subconfluence, whereby they were trypsinized and replated at low density. Characterization via immunocytochemical staining in some of the samples at passage 2 revealed a consistent immunophenotype as previously reported , which was negative for hemopoietic and endothelial markers CD34, CD45, CD14, CD31, and vWF and positive for mesenchymal markers CD105 (SH2), SH3, SH4, and vimentin and cell adhesion molecules CD29; α2, α4, and α5 integrin (CD49b, CD49d, and CD49e); CD44; CD106 (VCAM-1); laminin; and fibronectin. They did not express HLA-II and have low levels of HLA-I. In permissive media, they differentiated readily into osteoblasts and adipocytes. Comparable populations of adult BM MSCs had an immunophenotype and a differentiation capacity similar to those of hfMSCs.
In Vitro Muscle Differentiation
To induce in vitro myogenesis, we exposed hfMSCs to a variety of stimuli.
We exposed hfMSCs (n = 10) to 1–24 μM 5-azacytidine (concentrations similar to those published on myogenic induction of rodent or rabbit MSCs [33, 34]) over varying time points (6–48 hours) and multiple exposures in various culture media (serum-free, low-serum, and 10% fetal calf serum) and substrata (glass, plastic, Matrigel, and fibronectin). We found no evidence of any myogenic differentiation but noted massive cell death. Similarly, adult BM MSCs responded in a similar fashion with death of most of the cells without any expression of myogenic markers in surviving cells.
Horse Serum, Dexamethasone, and Hydrocortisone.
Next, we exposed hfMSCs (n = 3) to a mixture of horse serum, dexamethasone, and hydrocortisone over 12 days but did not find any myotube formation or the expression of any myogenic markers.
Following this, we investigated whether conditioned media [35, –37] from myoblast cultures could induce myogenic differentiation. After 7 days of culture in C2C12 myoblast-conditioned medium (1:2–4 conditioned/serum-free medium), a small number of desmin-staining cells (1–2 positive cells or myotubes per low-powered field) were present. After 14 days, long, multinucleated, desmin-positive fibers were evident, which stained positively for the muscle markers desmin and sarcomeric myosin. However, the conversion efficiency of hfMSCs approached only 1%–2% (Fig. 1A–1D, 1G). Culture of adult BM MSCs under identical conditions failed to result in muscle differentiation. Immunostaining of control C2C12 myoblasts revealed desmin expression in all cells, but the more mature muscle-specific protein, myosin, was only expressed upon myotube formation (Fig. 1E, 1F).
Next, we examined the ability of galectin-1 to induce myogenic differentiation in hfMSCs. After exposure of fetal blood or fetal BM-derived hfMSCs to galectin-1-enriched serum-free medium (200 ng/ml galectin-1), the appearance of desmin-positive single cells at day 3, followed by the formation of desmin-positive multinucleated fibers over the next 9 days in culture was observed. The percentage of desmin-positive cells increased over time, reaching a peak after 12 days of 66.1% ± 5.7% (mean ± SD) of total cells in culture (Fig. 2A–2E, 2G), most of which were incorporated into myotubes (3–23 nuclei per fiber). The latest time point examined was 12 days, as culture beyond this time resulted in the lifting of the myotubes from the plate. Western blots confirmed the accumulation of desmin only in galectin-1-stimulated cells (Fig. 2F). In a similar manner, the expression of sarcomeric myosin (MF20) increased over time to reach a peak of 63.0% ± 7.5% (mean ± SD) (Fig. 3A–3F).
When hfMSCs were exposed to various concentrations of galectin-1, the highest proportion of desmin-positive cultures (65.9% ± 8.0%) was achieved with 200 ng/ml of galectin-1. Higher concentrations of galectin-1 (400 and 1,000 ng/ml) resulted in significant cell death with fewer desmin-positive cells (10.0% ± 3.9% and 4.5% ± 1.7%, respectively), whereas lower dilutions (100 ng/ml), fewer desmin-positive cells were seen (44.9% ± 18.2%; Fig. 2H) (p < .0001).
In a separate experiment, continuous exposure of hfMSCs to galectin-1 over 12 days resulted in the development of multinucleated desmin-positive fibers in 62.3% ± 7.8% of cells, whereas a shorter exposure of 6 days followed by removal of galectin-1, and then 6 days of culture in serum-free medium resulted in 39.1% ± 2.3% of cells positively staining for desmin (p < .001). Exposure of hfMSCs to 6 days of galectin-1 followed by immediate staining for desmin resulted in only 18.5% ± 5.2% desmin-positive cells (p < .001) (Fig. 2I). Although hfMSCs grown in serum-free media failed to proliferate in culture, the addition of galectin-1 stimulated a mitogenic response, with a doubling time of 120 ± 10 hours (mean ± SD) as compared with 30 ± 4 hours when grown in the presence of serum. This compares with a doubling time of 115 ± 7 hours for adult BM MSCs grown in D10.
Prior to myogenic induction, both hfMSCs and adult BM MSCs were negative for markers of the muscle lineage, including desmin, myosin, Pax7, MyoD, and Myogenin on immunocytochemistry. In contrast, exposure of adult BM-derived MSCs to galectin-1 for up to 12 days did not result in the appearance of any desmin/myosin-positive cells or multinucleated myotubes.
Expression of Myogenic Regulatory Factors.
A time course examination of nuclear Pax7 expression using immunocytochemical staining revealed a gradual increase in the proportion of Pax7-positive hfMSCs, with 0%, 6.7%, 11.1%, 14.7%, and 24.2% of cells staining positive at 0, 3, 6, 9, and 12 days, respectively (Fig. 4F). In contrast, control C2C12 myoblast cells undergoing terminal differentiation in low serum culture were negative for Pax7. MyoD expression increased in a similar fashion, with 0%, 7.6%, 8.4%, 20.0%, and 20.4% of cells staining positive at 0, 3, 6, 9, and 12 days, respectively (Fig. 4A–4F). The expression of Myogenin also increased sequentially, with 0%, 1.9%, 5.8%, 5.5%, and 9.5% of cells being positive on day 0, 3, 6, 9, and 12 respectively (Fig. 4F). Analysis of the MRFs at the mRNA level revealed the transient upregulation of Myf5 from day 3 through day 6, Pax7 during days 3–9, and MyoD and myogenin from day 3 to day 12. Expression of both desmin and myosin heavy chain transcripts, though negative on immunocytochemical staining and Western blotting, was positive in naïve undifferentiated hfMSCs and upregulated through the 12 days of galectin-1 stimulation. Dystrophin mRNA was upregulated by day 9 of exposure to galectin-1 (Fig. 4G).
Differentiation of Engrafted Cells After In Vivo Transplantation into Injured and Dystrophic Muscle
To investigate whether hfMSCs could contribute to muscle regeneration in vivo, we used an in vivo injury model in the c–/γ–/RAG2– mouse. Initial pilot experiments revealed that intramuscular transplantation of naïve hfMSCs resulted in the formation of small numbers of human muscle fibers after 28 days (1–10 per muscle section). We then investigated whether prestimulation of hfMSCs with galectin-1 for 3 days would improve differentiation. After i.m. transplantation with galectin-1-prestimulated hfMSCs, a median of 44 (range, 20–93) human spectrin-positive muscle fibers per muscle section was found (Fig. 5A), whereas only 11 (range 3–26) human spectrin-positive fibers were seen in the contralateral muscle transplanted with nonstimulated hfMSCs after 28 days (Fig. 5B) (Wilcoxon signed ranked test, p = .008). This is against a background of 900–1,450 muscle fibers per cross section of the tibialis anterior muscle (i.e., ≤4%). Some human spectrin-positive muscle fibers displayed centrally located nuclei, indicating regenerated fibers (Fig. 5A). Double labeling with antibodies against human lamins A/C (components of the human nuclear membrane) and human spectrin indicated the human origin of these centrally-located nuclei (Fig. 5C). These spectrin-positive fibers could be traced along the tibialis anterior muscle by immunolabeling serial sections. H&E staining revealed some areas infiltrated with amorphous fibrous tissue that encased human cells revealed by lamin A/C staining (Fig. 5D). These cells of human origin did not express desmin (Fig. 5E), showing that they were nonmyogenic. Nontransplanted control muscle contained no cells staining for human lamins A/C and spectrin.
To further investigate the participation of hfMSCs in regenerating dystrophic fibers, we injected hfMSCs into the immunodeficient scid/mdx dystrophic mouse . After i.m. delivery, only galectin-1-prestimulated hfMSCs formed clusters of spectrin-positive fibers (a mean of 32.8 ± 20.2 per section) against a background of approximately 1,000 fibers per muscle fiber (Fig. 6A–6D). In contrast, none of the non-galectin-1-stimulated hfMSCs formed any human spectrin-positive muscle fibers or expressed any muscle marker such as desmin.
The recent discovery of a novel population of MSCs in the human fetus that displays multilineage differentiation in vitro  and in vivo  led us to hypothesize that hfMSCs can differentiate into muscle and participate in muscle regeneration. Although hfMSCs did not differentiate to skeletal muscle in vitro in response to 5-azacytidine or a combination of horse serum, dexamethasone, and hydrocortisone, we found that a soluble factor within myoblast-conditioned medium resulted in muscle differentiation in a small proportion of cells. As recent work in our laboratory suggested that the soluble factor responsible for myogenic conversion of dermal fibroblast was galectin-1 , we tested the ability of galectin-1 to induce myogenic differentiation in hfMSCs. We report that this resulted in significant recruitment of hfMSCs into the myogenic lineage and formation of mature myotubes. This occurred via activation of the myogenic regulatory factors in a manner reminiscent of that which occurs during embryonic myogenesis . Although hfMSCs contributed to muscle formation in a murine model of muscle regeneration, prestimulation of hfMSCs with galectin-1 for a short period of time resulted in the formation of substantially more muscle than with naïve hfMSCs.
The use of 5-azacytidine for myogenic differentiation of MSCs was described in immortalized rodent and rabbit MSCs [33, 34] via stochastic hypomethylation of random DNA residues , which in turn results in the activation of myogenic regulatory factors. Although there are no reports of successful myogenic conversion of adult BM-derived MSCs with 5-azacytidine in humans, this approach produced low levels of myogenic conversion of synovial membrane-derived human MSCs . Notwithstanding this, high-efficiency conversion has been achieved with a subpopulation of human BM-derived stem cells called multipotent adult progenitor cells, although there was no confirmation that these generated myotubes . Our finding of massive cell death in both hfMSCs and adult MSCs with 5-azacytidine is consistent with experience in human adult BM-derived MSCs, where there is a paucity of published literature to demonstrate myogenic differentiation. The use of a combination of horse serum, hydrocortisone, and dexamethasone was first described by Zuk et al. in adiposal-derived MSCs [21, 22] and later used by Gang et al. in umbilical cord blood-derived MSCs  for induction of myogenic differentiation. This, however, did not result in any myogenic differentiation in hfMSCs. Although the use of proprietary skeletal muscle differentiation medium (Promocell) has been reported to induce muscle differentiation in adipose-derived MSCs , we have not tested this in hfMSCs. Exposure to myoblast-conditioned media resulted in a small proportion of hfMSCs, but not adult BM MSCs, acquiring muscle-specific markers and, unlike dermal fibroblasts, fusion to form myotubes [35, 36].
Galectin-1, a 14–15-kDa lectin, belongs to a family of animal β-galactoside-binding proteins that are highly conserved during evolution [45, –47]. It is secreted by myoblasts in culture  and has a diverse variety of biological activities . Under different circumstances, it may act as a mitogen, an inhibitor of cell proliferation, a promoter of apoptosis , and, as we have previously shown, it is implicated in skeletal muscle determination [23, 51]. Interestingly, gene profiling showed that galectin-1 is one of the most commonly expressed genes and proteins in adult bone marrow-derived MSCs [52, 53]. As our previous work has implicated galectin-1 to be the soluble factor responsible for myogenic conversion of dermal fibroblasts [23, 54], we extended this work by exposing hfMSCs to galectin-1 to find efficient myogenic conversion with formation of mature myotubes in vitro. This contrasted with dermal fibroblasts, which did not form mature myotubes in vitro. Continuous galectin-1 exposure over 12 days resulted in significantly higher proportion of hfMSCs entering myogenic differentiation than cells exposed for 6 days followed by the removal of galectin-1 and subsequent culture for 6 days under serum-free conditions. This suggested that either cells entering myogenic differentiation have a proliferative advantage over others, or that continued galectin-1 stimulation resulted in a higher proportion of cells recruited into the myogenic lineage. Interestingly, exposure of hfMSCs to 6 days of galectin-1 followed by immediate staining for desmin resulted in a much lower proportion of desmin-positive cells, suggesting that although the presence of galectin-1 is important for continued recruitment of noncommitted hfMSCs into the myogenic lineage, hfMSCs that have already initiated myogenic differentiation can still divide and proceed to terminal differentiation without continued galectin-1 stimulation. Galectin-1 also has a mitogenic effect on hfMSCs.
The level of in vitro myogenic conversion and myotube development we report here when hfMSCs were grown in galectin-1-enriched medium is considerably greater than previously reported in human MSC populations [8, 21, 41, 55] or even murine myoblast cultures . de Bari et al. found only rare myogenic differentiation and little myotube formation with their human synovial membrane-derived MSCs , whereas Reyes et al., in reporting that up to 80% of human multipotent adult progenitor cells acquire muscle markers on exposure to 5-azacytidine, still did not find significant myotube formation . Although Gang et al. recently demonstrated muscle differentiation by expression of myosin heavy chain in up to 56% of umbilical cord blood-derived MSCs after culture with horse serum, dexamethasone, and hydrocortisone, there was no morphological evidence of myotube formation . Barberi et al. demonstrated that 10% of mesenchymal progenitor cells derived from human embryonic stem cells could fuse with C2C12 cells in coculture . Recently, Dezawa et al. described a multistage myogenic induction program resulting in up to 40% of adult BM MSCs fusing to form myotubes, with clonal populations achieving an 89% fusion index . However, this was accomplished via the transfection of a Notch 1 intracellular domain gene with selection via neomycin resistance within the plasmid, which may affect subsequent cell function/metabolism both in vitro [58, 59] and in vivo .
Galectin-1-induced myogenic differentiation is accompanied by the sequential expression of early, intermediate, and late MRFs (Myf5 and Pax7, MyoD, and Myogenin, respectively) differentiating in a stepwise manner into cells bearing a satellite cell immunophenotype (Pax7+) and finally into multinucleated mature muscle fibers in a manner reminiscent of that which occurs during embryonic myogenesis and skeletal muscle regeneration. This is in keeping with the in vivo finding of LaBarge et al. that bone marrow-derived cells undergo recruitment to muscle satellite cells and finally mature myofibers in a stepwise manner . The finding that both desmin and myosin heavy chain was amplified by RT-PCR but the protein was not detected by immunocytochemical staining (desmin and myosin) or Western blot (desmin) in the naïve unstimulated hfMSCs shows the importance of looking for protein as well as gene expression, as small amounts of message may not have given rise to functional amounts of protein. This mirrors findings by Seshi et al. in adult bone marrow-derived MSCs, where both heterogeneous as well as clonal MSC populations were found to express multiple lineage genes, including those of the myogenic lineage such as myosin and dystrophin [62, 63].
To elucidate the role of hfMSCs in muscle repair, we used a well-established murine muscle injury model in which both soluble factors and cell-cell contact within regenerating muscle would influence myogenic specification of transplanted cells. Transplantation of galectin-1-prestimulated hfMSCs rather than naïve hfMSCs resulted in a fourfold increase in muscle fiber formation, suggesting that although local environmental cues of a regenerating muscle environment stimulate myogenesis in hfMSCs, the use of galectin-1 pre-stimulation potentiates the contribution of hfMSCs to skeletal muscle regeneration. The majority of transplanted cells within the muscle, however, did not undergo muscle differentiation and remained as nonmyogenic, desmin-negative cells. hfMSCs behaved in a similar fashion when transplanted into dystrophic scid/mdx mice muscle, with only galectin-1-prestimulated cells contributing to muscle regeneration.
Although both Carmargo et al.  and Corbel et al.  recently showed muscle formation in mouse models from the hemopoietic compartment of the bone marrow in mice, neither group completely ruled out MSC participation. There have been few reports of myogenic differentiation of human MSCs in vitro apart from MSCs derived from non-BM sources such as adipose tissue , synovial membrane , or umbilical cord blood . We have demonstrated that a well-characterized population of human MSCs from fetal blood and BM undergoes myogenic differentiation with the formation of mature myotubes when stimulated with a soluble factor found in myoblast-conditioned medium. We speculated that the compound responsible is galectin-1 and confirmed this by showing far more impressive myogenic conversion of hfMSCs in vitro following galectin-1 stimulation. Our speculation is strengthened by the fact that prestimulation of hfMSCs with galectin-1 results in a fourfold increase in their contribution to muscle regeneration in injured murine muscle, confirming that galectin-1 is a crucial factor in increasing the myogenic conversion of hfMSCs to skeletal muscle in vivo as well as in vitro.
In summary, we have demonstrated efficient in vitro muscle differentiation with myotube formation of hfMSCs in response to galectin-1 in vitro and an increase in contribution to muscle regeneration in both an injury and dystrophic murine model after galectin-1 prestimulation. Although the mechanism by which galectin-1 switches on the myogenic regulatory factors and hence induces myogenic differentiation remains unknown, we suggest that hfMSCs may be a promising source of cells for the autologous or allogeneic treatment of genetic/degenerative muscle diseases and serve as an important model of myogenic cell specification/differentiation. Further investigations are now under way to increase the level of hfMSCs contribution to muscle regeneration and to elucidate the mechanism by which galectin-1 myogenic differentiation occurs in hfMSCs.
K.O. received salary support from Action Medical Research, U.K. N.K. was funded by a research training fellowship from the Wellcome Trust, U.K. J.E.M. received funding from the Muscular Dystrophy Campaign and the MRC, U.K. D.W. received funding for the galectin-1 work from the Muscular Dystrophy Campaign.
We thank Prof. Terence Partridge for his helpful suggestions and advice, Dr. Briana Cloke for help with western blotting, and Karimah Brimah and Janine Ehrhardt for help with immunohistochemistry. This study was funded by the Institute of Obstetrics and Gynaecology Trust.