Sca-1 negatively regulates proliferation and differentiation of muscle cells

Developmental Biology 283 (2005) 240 – 252 www.elsevier.com/locate/ydbio

Sca-1 negatively regulates proliferation and differentiation of muscle cells Patrick O. Mitchella, Todd Millsa, Roddy S. O’Connora,b, Timothy Graubertc, Elaine Dzierzakd, Grace K. Pavlatha,* a

Department of Pharmacology, Emory University School of Medicine, Room 5027, O. W. Rollins Research Center, Atlanta, GA 30322, USA b Graduate Program in Molecular and Systems Pharmacology, Emory University, Atlanta, GA 30322, USA c Department of Internal Medicine, Division of Oncology, Stem Cell Biology Section, Washington University School of Medicine, St. Louis, MO 63110, USA d Department of Cell Biology and Genetics, Erasmus University, Rotterdam, The Netherlands Received for publication 5 January 2005, revised 12 March 2005, accepted 15 April 2005 Available online 17 May 2005

Abstract Satellite cells are tissue-specific stem cells critical for skeletal muscle growth and regeneration. Upon exposure to appropriate stimuli, satellite cells produce progeny myoblasts. Heterogeneity within a population of myoblasts ensures that a subset of myoblasts readily differentiate to form myotubes, whereas other myoblasts remain undifferentiated and thus available for future muscle growth. The mechanisms that contribute to this heterogeneity in myoblasts are largely unknown. We show that satellite cells are Sca-1neg but give rise to myoblasts that are heterogeneous for sca-1 expression. The majority of myoblasts are sca-1neg, rapidly divide, and are capable of undergoing myogenic differentiation to form myotubes. In contrast, a minority population is sca-1pos, divides slower, and does not readily form myotubes. Sca-1 expression is not static but rather dynamically modulated by the microenvironment. Gain-of-function and loss-of-function experiments demonstrate that sca-1 has a functional role in regulating proliferation and differentiation of myoblasts. Myofiber size of sca-1 null muscles is altered in an age-dependent manner, with increased size observed in younger mice and decreased size in older mice. These studies reveal a novel system that reversibly modulates the myogenic behavior of myoblasts. These studies provide evidence that, rather than being a fixed property, myoblast heterogeneity can be modulated by the microenvironment. D 2005 Elsevier Inc. All rights reserved. Keywords: Ly-6A/E; Satellite cell; Myoblasts; Heterogeneity

Introduction Satellite cells are critical for muscle growth and regeneration. Satellite cells are quiescent myogenic progenitor cells that are located underneath the basal lamina of myofibers. In response to growth stimuli, satellite cells are activated, and their progeny myoblasts proliferate and subsequently differentiate and fuse to form new myofibers or fuse with existing myofibers (Hawke and Garry, 2001). Molecules important for the initiation of differentiation in myoblasts, such as the family of myogenic regulatory factors, have been extensively studied but far less is known about molecules that inhibit differentiation of myoblasts. * Corresponding author. Fax: +1 404 727 0365. E-mail address: [email protected] (G.K. Pavlath). 0012-1606/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2005.04.016

Subsets of myoblasts can remain undifferentiated in the face of differentiation-inducing stimuli. After muscle regeneration or growth, the portion of myoblasts that remain undifferentiated are thought to give rise to new satellite cells for use in future muscle growth. In vitro studies indicate that differentiated myotube cultures contain undifferentiated, quiescent ‘‘reserve’’ cells, which downregulate myoD and can be stimulated to proliferate upon addition of serum, reminiscent of satellite cells (Kitzmann et al., 1998; Yoshida et al., 1998). Myoblast populations are heterogeneous in their growth and differentiation properties both in vivo and in vitro (Heslop et al., 2000; Rantanen et al., 1995; Rouger et al., 2004; Schultz, 1996). Whether such heterogeneity is fixed or can be modulated is unknown. The molecular mechanisms responsible for such heterogeneity have been little studied.

P.O. Mitchell et al. / Developmental Biology 283 (2005) 240 – 252

The Ly-6 superfamily of proteins appears to play a role in cell differentiation and development of immune and nonimmune cells (Bamezai, 2004). These are a family of 10 –12 kD proteins that are linked to the cell surface by a GPI anchor and have cell signaling and cell adhesion properties. Stem cell antigen-1 (sca-1) is one member of the Ly6 family in mouse that has been implicated in cellular differentiation. The sca-1 protein is encoded by the strain-specific allelic genes, Ly-6E.1 (Ly-6E) and Ly-6A.2 (Ly-6A) and is also known as Ly-6A/E. Sca-1 was first described in mouse hematopoietic cells (Gumley et al., 1995) and has been used extensively to enrich for and characterize hematopoietic stem cells. Sca-1 is also expressed by a number of nonimmune cells including muscle-derived stem cells (Asakura et al., 2002; Benchaouir et al., 2004; Gussoni et al., 1999; Jankowski et al., 2001; Lee et al., 2000; Majka et al., 2003; Tamaki et al., 2002). Sca-1 is expressed in a high proportion of muscle-derived stem cells (45 – 94%). Controversy over satellite cell expression of sca-1 exists, however. Diverse functions have been proposed for sca-1 in immune and non-immune cells. Sca-1 is a negative regulator of lymphocyte proliferation as T cells overexpressing sca-1 are hypoproliferative (Henderson et al., 2002), whereas sca1 null T cells are hyperproliferative (Stanford et al., 1997). An antibody against a 66 kD protein expressed in spleen and thymus inhibits sca-1 dependent cell –cell adhesion (English et al., 2000), implicating a direct role for sca-1 in mediating cellular adhesion. Sca-1 also regulates cellular differentiation as mistimed expression of sca-1 in immature T cells leads to impairment in T cell development (Bamezai et al., 1995). Furthermore, sca-1 null mice display defects in lineage commitment and stem cell self-renewal of hematopoietic stem cells (Ito et al., 2003) and mesenchymal progenitors in the bone marrow (Bonyadi et al., 2003). Here, we examine the expression and function of sca-1 in satellite cells, primary myoblasts, and the C2C12 myogenic cell line. Satellite cells are sca-1neg, whereas primary and C2C12 myoblasts are heterogeneous for sca-1. We used several techniques to show that sca-1 is functionally important in negatively regulating growth and differentiation of muscle cells.

241

mice. Heterozygous mice harboring enhanced GFP cDNA integrated into the Sca-1 locus were provided by Dr. Timothy Graubert (Hanson et al., 2003). Heterozygous mice were bred to homozygosity to generate Sca-1 null mice. Sca-1 null mice and their wildtype littermates were used either at 2 –4 months or 1 year of age unless otherwise stated. All mice were housed under a 12:12-h light – dark cycle, and food and water were provided ad libitum. All procedures involving animals were approved by Emory University’s Institutional Animal Care and Use Committee. Collection of muscles and morphometric measurements Tibialis anterior (TA) and soleus muscles were collected and sectioned as described (Mitchell and Pavlath, 2004). All analyses and photography were performed on a Zeiss Axioplan microscope equipped with a video camera and Scion Image (v.1.63) and Adobe Photoshop\ (v.7.0) software. For the TA and soleus, anatomical markers of each muscle were used to find the same region in different samples, and these sections were subsequently used for analysis. The cross-sectional area (XSA) of individual myofibers of wildtype (n = 3– 14) and sca-1 null (n = 3– 4) mice was determined by capturing an image in the center of each section and analyzing 100– 250 myofibers within this 307,200-Am2 field. Between 330 –2800 myofibers were analyzed per genotype at each age. Determination of capillary density Transverse sections from the belly of TA muscles from 8-month-old wildtype (n = 3) and sca-1 null (n = 3) mice were stained for alkaline phosphatase (AP) using an AP chromogen kit (Biomeda Corporation) to detect capillary endothelial cells as per the manufacturer’s instructions and then counterstained with eosin (Hershey et al., 2001). Capillary density was calculated as the number of capillaries per fiber in a 307,200-Am2 field. 1100 and 1400 capillaries were counted for wildtype and sca-1 null mice, respectively. Primary myoblast isolation and cell culture

Materials and methods Animals For isolation of primary myoblasts and single myofiber explants, adult (8 –10 weeks) C57BL/6 mice were purchased from Charles River. Sca-1 (Ly-6A.2) green fluorescent protein (GFP) transgenic mice, in a C57BL/10  CBA background, were provided by Dr. Elaine Dzierzak (Ma et al., 2002). In these transgenic mice, 8 copies of a 14 kb sca-1 expression cassette drive GFP expression under the control of Ly-6A gene regulatory sequences in numerous tissues. We refer to these mice as Sca-1/GFP transgenic

Primary cultures were prepared from hindlimb muscles of 12- to 20-week-old mice and subjected to both mechanical and enzymatic dissociation as described previously (Kastner et al., 2000; Mitchell and Pavlath, 2004; Yablonka-Reuveni et al., 1987). Cells were collected from the interface of the Percoll gradient and either lysed in TRIZOL reagent (Life Technologies) for RT-PCR analyses or resuspended in primary growth medium (Ham’s F10, 20% FBS, 5 ng/ml bFGF, 100 U/ml penicillin G, and 100 Ag/ml streptomycin) and grown on Vitrogen-100 (Cohesion)-coated tissue culture dishes in a humidified, 37-C, 5% CO2 incubator. The percentage of myogenic cells was

242

P.O. Mitchell et al. / Developmental Biology 283 (2005) 240 – 252

approximately 97% as accessed by Pax 7 staining. Myogenic differentiation was induced by culturing myoblasts on entactin-collagen IV-laminin (E-C-L) (Upstate Biotechnology)-coated plates in differentiation medium (DMEM, Insulin-Transferrin-Selenium-A supplement (GIBCO), 100 U/ml penicillin G and 100 Ag/ml streptomycin) for 4 days. In some experiments, cells were cultured in osteogenic differentiation medium (DMEM, 5% FBS, 100 U/ml penicillin G, 100 Ag/ml streptomycin, and 0 – 100 ng/ ml bone morphogenic protein-2 (BMP2, Kamiya Biomedical Company)) for 5 days. C2C12 myoblasts were maintained in C2C12 growth medium (DMEM with 4.5 mg/ml glucose, 10% FBS, 100 U/ml penicillin G, and 100 Ag/ml streptomycin) in a humidified, 37-C, 5% CO2 incubator. C2C12 cells originated from a C3H mouse (Yaffe and Saxel, 1977) and thus express the Ly-6E.1 alloantigen. Myotube formation was induced by switching the cells to myogenic differentiation medium as above. Isolation and culture of single myofiber explants Single myofibers were isolated from the gastrocnemius muscle as described by Rosenblatt et al. (1995). Briefly, the gastrocnemius was dissected and digested in DMEM containing 0.1% collagenase (type I: 400 U/ml, Worthington) at 37-C for 1 h with gentle agitation. Single myofibers were then dissociated by repeated gentle trituration of the digested muscles. Using a dissecting microscope, single myofibers were extracted individually using fire-polished pipettes and transferred serially into fresh plates containing DMEM with 4.5 mg/ml glucose, 10% FBS, 100 U/ml penicillin G, and 100 Ag/ml streptomycin. Individual single myofibers were transferred to 24-well plates pre-coated with growth-factor-reduced Matrigel (BD Biosciences), centrifuged for 40 min at 1100  g to facilitate adhesion to the Matrigel, and incubated for 0– 72 h in a humidified, 37-C, 5% CO2 incubator. Immunocytochemistry, GFP imaging, and alkaline phosphatase staining The following antibodies were used: anti-myoD (NCLMyoD1, 1:30; NovoCastra), anti-Ly-6A/E (D7 and E13 161 – 7 clones, 5 Ag/ml; BD Pharmingen), anti-embryonic myosin heavy chain (eMyHC; F.1652, neat; Developmental Studies Hybridoma Bank), anti-Pax7 (neat; Developmental Studies Hybridoma Bank), anti-c-met (1:100; Santa Cruz), anti-m-cadherin (5 Ag/ml; nanoTools), and anti-CD34 (1:20; BD Pharmingen). The appropriate isotype control antibodies were purchased from Jackson ImmunoResearch. For all immunostaining procedures, cultures of C2C12 cells, primary myoblasts, and single myofiber explants were rinsed with PBS and fixed in 2% formaldehyde for 10 min. All steps were performed at room temperature unless otherwise noted. Images were obtained using a Zeiss

Axiovert 200 M microscope and Openlab version 3.1.5 software (Improvision). Sca-1 staining alone Primary myoblasts were pre-incubated with sca-1 or isotype control antibodies 30 min prior to fixation. Endogenous peroxidase activity in cells was quenched by incubating in 3% hydrogen peroxide for 10 min. Cells were blocked in 5% donkey serum for 30 min and incubated overnight at 4-C with anti-sca-1 antibody. The tyramide signal amplification system (TSA kit, NEN) was used to visualize antibody binding according to the manufacturer’s protocol. Briefly, cultures were incubated with TNB buffer (0.1 M Tris –HCl pH 7.5, 0.15 M NaCl, and 0.5% blocking reagent) for 1 h followed by biotinconjugated donkey anti-rat F(ab V)2 fragments (1:500; Jackson ImmunoResearch) for 1 h. Cultures were washed with PBS + 0.2% Tween 20 (PBS-T) and incubated with streptavidin –horseradish peroxidase (1:200 in TNB buffer) for 30 min followed by fluorescein– tyramide (1:250 in amplification diluent) for 5 min. Nuclei were stained with 4 V, 6-diamidino-2-phenylindole, dihydrochloride (DAPI, 250 ng/ml; Sigma). eMyHC staining alone Primary myoblasts from wild-type or sca-1 null mice were plated at 2  105 cells per well of a six-well plate and switched to DMEM + 2% horse serum 8 –12 h later. Myotube formation was assessed after 3 days. Cells were fixed and blocked with TNB buffer for 1 h followed by a 2 h incubation in anti-eMyHC antibody. Cells were washed in PBS-T and incubated with biotin-conjugated goat antimouse IgG F(ab V)2 fragments (1:250 in PBS-T) for 1 h. Antibody binding was detected using the Vectastain Elite Kit (Vector Laboratories). Nuclei were stained with DAPI as above. eMyHC and sca-1 co-staining Endogenous peroxidase activity in cells was quenched as above. Cells were blocked in 5% goat serum (Sigma) in PBS-T for 30 min followed by a 1 h incubation in antieMyHC antibody. Cells were washed with PBS-T and incubated with Texas-Red goat anti-mouse IgG F(ab V)2 fragments (1:250 in PBS-T; Cappel) for 1 h. After another series of washes in PBS-T, cells were stained for sca-1, and nuclei visualized with DAPI as above. Detection of GFP in single myofiber explants and myoblasts from Sca-1/GFP transgenics To detect GFP alone, images of live single myofibers and myoblasts were obtained. In other experiments, single myofiber explants and primary myoblast cultures were coimmunostained for GFP and myoD, pax7, BrdU, c-met, or m-cadherin. Endogenous peroxidase activity in cells was quenched by incubating in 3% hydrogen peroxide for 10 min. Cells were blocked in 5% donkey serum (Sigma) for

P.O. Mitchell et al. / Developmental Biology 283 (2005) 240 – 252

30 min and incubated overnight at 4-C with anti-myoD, pax7, c-met, or m-cadherin, and anti-GFP (Molecular Probes; 1:500) antibodies. Primary antibody binding for m-cadherin, myoD, and pax7 was detected with biotinconjugated donkey anti-mouse IgG F(ab V)2 fragments, and c-met was detected with biotin-conjugated donkey antirabbit (1:250; Jackson ImmunoResearch) for 1 h. The tyramide signal amplification system (TSA kit, NEN) was used to visualize antibody binding as described above. AntiGFP binding was detected using FITC-conjugated donkey anti-rabbit IgG F(ab V)2 fragments (1:250 in PBS-T; Jackson ImmunoResearch). AP staining For alkaline phosphatase staining of cells cultured in osteogenic differentiation medium, cells were fixed in 2% formaldehyde, and alkaline phosphatase activity was detected using a commercially available histochemical kit (Sigma) according to the manufacturer’s instructions. APpositive cells stain blue using this protocol. Flow cytometry Cells were trypsinized and resuspended in cold PBS containing 0.25% BSA. Cells were incubated for 30 min at 4-C in the following antibodies at a concentration of 1 Ag antibody per 106 cells: FITC- or PE-conjugated anti-sca-1 (D7 and E13 161– 7 monoclonal antibodies; BD Pharmingen) and an isotype control rat IgG2an (BD Pharmingen). Cells were washed three times in PBS containing 0.25% BSA and resuspended in PBS, 0.25% BSA, and 0.5 Ag/ml propidium iodide (Sigma) for cell sorting. All sorting was performed on a FACSVantage (Becton-Dickinson). Due to the lack of clearly defined sca-1pos C2C12 cells, the purity of the sca-1pos cells usually ranged from 85 – 95%. The purity of the sca-1neg cells, however, was higher and usually ranged from 90 –98%. Double sorting of the sca1pos cells resulted in only a small increase in purity and thus was not continued. Primary myoblasts were processed as described above for C2C12 cells but sorted on the basis of GFP expression. Data from all flow cytometric experiments were analyzed using FlowJo software (version 4.3; Tree Star). Retroviral plasmid construction and infection The control TJ66 retroviral plasmid contains a chimeric GFP-zeocin selectable marker driven by IRES-dependent translation that allows the efficiency of retroviral-mediated gene delivery to be assessed by fluorescence microscopy (Murphy et al., 2002). Sca-1 (Ly-6E.1) cDNA was generated by RT-PCR of C2C12 myoblast RNA. Primers were designed to amplify the sca-1 cDNA with XhoI and NruI restriction sites inserted into the 5 Vand 3 Vends, respectively. A 704 bp sca-1 cDNA insert was subcloned into XhoI and NruI sites in pTJ66 to generate the retroviral plasmid pPM4.

243

Sca-1 expression in pPM4 is under the control of the viral 5 V-LTR promoter. The efficiency of 2 rounds of retroviral infection was >94% as assessed by GFP expression. Overexpression of sca-1 was verified by RT-PCR and flow cytometric analyses. RT-PCR analyses of mRNA expression Total RNA was isolated using TRIZOL reagent according to the manufacturer’s instructions. The RT reaction was performed using 2.5 Ag of total RNA per sample. All PCR reactions were performed using 1 Al of the above RT reaction, and primers specific for each particular gene and designed to cross intron –exon boundaries were used to generate an amplicon in its linear range. Myogenin (sense: 5 V-agcggctgcctaaagtggagat-3 V; antisense: 5 V-ggactgaagggagtgcagattgtg-3 V; 266 bp; Genbank accession #: D90156). Sca-1 (sense: 5 V-cgagggagggagctgtgaggtt-3 V; antisense: 5V-gagggcagatgggtaagcaaagat-3V; 285 bp; Genbank accession #: NM_010738). eMyHC (sense: 5 V-gaagaagaacctggagcagacg-3 V; antisense: 5 V-agcctgcctcttgtaggacttg3 V; 222 bp; Genbank accession #: M11154). Osteocalcin (Sense: 5 V-accctggctgcgctctgtctct-3 V; antisense: 5V-gatgcgtttgtaggcggtcttca-3 V; 239 bp; Genbank accession #: L24431). 18S rRNA was used as an internal control in each sample using standard QuantumRNA 18S primers (Ambion). PCR products were resolved on 1.5% agarose gels, and the 18S rRNA amplicon was used to verify equal input of RNA to the RT reaction. Cell proliferation assays Cells were seeded into 6-well plates at a density of 1  105 cells per well. Cultures were refed 8– 12 h later with growth media containing 5-bromo-2-deoxyuridine (BrdU; 25 AM; Sigma) for 15 min (C2C12 cells) or 30 min (primary myoblasts) in a humidified, 37-C, 5% CO2 incubator. Cells were then rinsed with PBS and fixed in 2% formaldehyde for 10 min. DNA was denatured by incubating with 2 N HCl in PBS at 45-C for 1 h and subsequently neutralized with borate buffer (0.1 M borate, pH 8.5). Cells were incubated with blocking buffer (5% donkey serum and 0.25% Triton X-100 in PBS) for 1 h followed by rat anti-BrdU antibody (1:500 in blocking buffer; Harlan Sera-lab) and anti-GFP antibody (1:1000) for 1 h and washed with PBS-T. AntiBrdU binding was detected using Texas Red-conjugated donkey anti-rat IgG F(ab V)2 fragments (1:250 in PBS-T; Jackson ImmunoResearch), and anti-GFP binding was detected as above. Cell mixing experiments Sorted GFPneg (99.8% pure) and GFPpos (100% pure) primary myoblasts were infected with control or aSGC (Dhawan et al., 1991) retroviruses, respectively. After expansion, GFPneg cells were 100% negative for h-gal

244

P.O. Mitchell et al. / Developmental Biology 283 (2005) 240 – 252

activity, and GFPpos cells were 99.6% positive for h-gal activity. Cells were then mixed at 10% GFPpos/lacZpos to 90% GFPneg/lacZneg and cultured in growth media for 48 h. Cultures were stained for the presence of h-gal as described previously (Mohler and Blau, 1996). Cells were incubated with the fluorescent h-gal substrate for 90 min at 37-C. For immunolocalization of GFP, cells were fixed and stained for the presence of GFP as described above. Conversion of GFPneg cells to GFPpos cells was evident by the presence of GFPpos/lacZneg cells, whereas GFPpos cells converting to GFPneg cells were GFPneg/lacZpos. Retroviral infection had no effect on the proliferation or differentiation of these cells. In addition, no silencing of the viral lacZ gene was observed in unmixed cells over 3 weeks of culture. Retrovirally infected cells were used in experiments within 2 weeks of infection. Statistics To determine significance between two groups, comparisons were made using Student’s t tests using GraphPad Prism version 4.0 a for Macintosh (GraphPad Software, San Diego, CA). For all statistical tests, P < 0.05 was accepted for statistical significance.

Results Heterogeneous expression of sca-1 in skeletal muscle cells Several techniques were utilized to detect sca-1 expression in skeletal muscle cells. Immunocytochemical analyses indicate that sca-1 is expressed by a subset of C2C12 myoblasts (approximately 21%). In myotube cultures, sca1pos cells are embryonic myosin heavy chain (eMyHC) negative and comprise approximately 20% of the undifferentiated cells present (Fig. 1A). Upon myogenic differentiation, sca-1 mRNA expression is upregulated (Fig. 1A). Since immunocytochemistry indicates no change in the percentage of cells that are sca-1pos, the RT-PCR data suggest that sca-1 mRNA expression is upregulated in cells already expressing sca-1. Primary mouse myoblast cultures also exhibit sca-1 immunostaining. The frequency of sca1pos cells is 20% in early cultures and declines to 2– 5% depending on culture conditions and passage number. In agreement with C2C12 cultures, sca-1 expression is restricted to undifferentiated primary myoblasts, and sca-1 mRNA expression is upregulated during myogenic differentiation (Fig. 1B). Sca-1pos primary cells are also myoDpos, thus excluding the possibility that these cells are contam-

Fig. 1. Heterogeneous expression of sca-1 in skeletal muscle cells. (A) C2C12 cells and (B) primary myoblasts were analyzed for sca-1 expression either as myoblasts or myotubes. Cells were immunostained for sca-1 (green) and eMyHC (red). DAPI was used to stain nuclei (blue). Sca-1 is expressed exclusively on undifferentiated cells. RT-PCR analyses of sca-1 expression in myoblasts (Mb), and myotubes (Mt) were performed. Sca-1 mRNA is increased upon differentiation. Myogenin was used as a marker of differentiation. 18S rRNA was used as a control for RNA input. Immunocytochemical and RT-PCR images are representative of 3 independent experiments. Scale bar = 100 Am. (C) Sca-1pos (green) primary myoblasts are myoDpos (red), demonstrating that sca-1pos cells are not contaminating non-muscle cells. Scale bar = 50 Am.

P.O. Mitchell et al. / Developmental Biology 283 (2005) 240 – 252

inating non-muscle cells (Fig. 1C). In addition, sca-1pos cells are CD45neg and c-kitneg (data not shown). Sca-1 expression was also analyzed in primary myoblasts isolated from transgenic sca-1/GFP transgenic mice (Ma et al., 2002). GFP expression is heterogeneous in primary myoblasts (Fig. 2C). To exclude the possibility that GFPpos cells are contaminating non-muscle cells, GFPpos cells were sorted on the basis of GFP expression and immunostained for skeletal muscle markers. Counterstaining sorted GFPpos cells for myoD, pax7, c-met, and m-cadherin indicates that the GFPpos cells are indeed myogenic (Fig. 2A). The same results were obtained with sorted GFPneg cells (data not shown). Primary myoblasts from sca-1/GFP mice were also counterstained for sca-1 and analyzed by flow cytometry to determine the concordance between GFP expression and the presence of sca-1 protein. Approximately 96% of the GFPpos and GFPneg cells are also sca-1pos and sca-1neg, respectively (Fig. 2B), demonstrating excellent concordance between the two markers. Upon myogenic differentiation of primary myoblasts, GFP expression is restricted to mono-

245

nucleated cells outside myotubes (Fig. 2C), in agreement with the data presented in Fig. 1. Treatment of GFPneg myoblasts with g-interferon, a known inducer of sca-1 expression in immune cells (Khan et al., 1993), leads to induction of GFP in a subset of cells (Fig. 2D). These data indicate that the transgene is appropriately regulated in myoblasts. Thus, these transgenic cells are an appropriate model for studying sca-1 in myoblasts. Together, these data indicate that sca-1 is only expressed by a low proportion of undifferentiated skeletal muscle cells, and these sca-1pos cells also express myogenic markers. Satellite cells express Sca-1 upon activation and proliferation To determine whether satellite cells are heterogeneous for Sca-1 expression similar to proliferating myoblast cultures, single myofiber explants from wildtype mice were immunostained for sca-1 expression immediately upon isolation. Approximately two sca-1pos cells are seen per 100 single

Fig. 2. Detection of sca-1 using Sca-1/GFP transgenic mice. (A) GFPpos myoblasts from transgenic mice were sorted and immunostained for GFP (green) and myoD, pax7, c-met, or m-cad (red). DAPI was used to stain nuclei (blue). The majority of GFPpos cells are positive for these myogenic markers. (B) Myoblasts from transgenic mice were analyzed for GFP and sca-1 expression by flow cytometric analyses. Approximately 96% of GFPneg and GFPpos myoblasts are sca1neg and sca-1pos, respectively. (C) In myotube cultures, GFP expression is restricted to a subset of unfused cells. (D) GFPneg cells were sorted and cultured in the absence (vehicle) or presence (100 U/ml) of IFN-g. GFP expression is induced after 4 days of IFN-g treatment. Scale bars = 100 Am.

246

P.O. Mitchell et al. / Developmental Biology 283 (2005) 240 – 252

myofibers (Fig. 3A). No more than one sca-1pos cell is detected on a single myofiber. Single myofiber explants from sca-1/GFP transgenic mice were also analyzed for GFP expression. A low frequency of GFP expression (approximately two GFPpos cells per 100 single myofibers) is also observed in freshly isolated single myofiber explants (Fig. 3B). These myofibers were immunostained for myoD to determine the identity of the GFPpos cells. Of eight single myofibers harboring GFPpos cells, only one was myoDpos (Fig. 3D). Thus, in some instances, the rare GFPpos cells found on freshly isolated myofibers are definitively myogenic. These may be activated satellite cells or contaminating interstitial myogenic cells. In contrast, as single myofiber explants are maintained in culture over time, the frequency of myofibers with at least one GFPpos cell increases to 46% and 68% after 2 and 4 days, respectively (Figs. 3B,C). These timepoints reflect the activation and subsequent proliferation of satellite cells. Although a high percentage of myofibers give rise to GFPpos cells, the percentage of GFPpos cells associated with any one fiber is never more than ¨5%. The fact that the percentage of

myofibers with GFPpos cells increases from 2% to 68% indicates that GFP is upregulated in the descendants of GFPneg satellite cells. GFP expression is modulated differentially in disparate growth conditions To study the modulation of sca-1 expression more directly, primary myoblasts from Sca-1/GFP transgenic mice were sorted into GFPneg and GFPpos populations. GFPneg cells were infected with a control retrovirus, whereas GFPpos cells were infected with a h-galactosidase (h-gal) expressing retrovirus and the two cell populations mixed in a 1:9 ratio, respectively (Fig. 4A). After 48 h of coculture, modulation of GFP expression is evident by the presence of GFPpos/h-galneg cells (originally GFPneg/hgalneg cells) and GFPneg/h-galpos cells (originally GFPpos/ h-galpos cells) (Fig. 4B). In contrast, GFP expression is not modulated when GFPneg and GFPpos myoblasts are grown under clonal conditions (Fig. 4C). These results suggest a mechanism whereby GFPpos and GFPneg cells modulate

Fig. 3. Satellite cells express Sca-1 upon activation and proliferation. (A) Single myofibers were immunostained for sca-1 within 2 h of isolation. Approximately two sca-1pos cells are seen per 100 single myofibers (¨300 single myofibers analyzed from 5 mice). No more than one sca-1pos cell was detected on a single myofiber. Scale bar = 100 Am. (B) Single myofibers were isolated and analyzed for GFP expression 2 h post-isolation (d0), 2 days (d2), or 4 days (d4) in culture. Scale bar = 100 Am. (C) GFP expression is induced in myofiber cultures with time. Number of myofibers analyzed: d0 = 240, d2 = 183, d4 = 155. (D) Eight freshly isolated single myofibers harboring a single GFPpos cell were immunostained for the presence of myoD (red). One GFPpos cell was myoDpos, and the remaining 7 GFPpos cells were myoDneg. DAPI was used to stain nuclei (blue). Scale bar = 50 Am.

P.O. Mitchell et al. / Developmental Biology 283 (2005) 240 – 252

247

Fig. 4. GFP expression is modulated differentially in disparate growth conditions. (A) Schematic representation of the experiment shown in panel (B). GFPneg and GFPpos myoblasts were infected with control and h-gal expressing retroviruses, respectively, mixed at a ratio of 1:9 (pos to neg) and cultured in growth media for 48 h. Modulation of GFP expression is indicated by the presence of GFPpos/h-galneg cells (originally GFPneg/hgalneg cells) and GFPneg/h-galpos cells (originally GFPpos/h-galpos cells). (B) Mixtures of GFPneg/pos myoblasts as described in panel (A) were analyzed for GFP expression (green) and h-gal activity (red). Some myoblasts that were originally GFPneg/h-galneg are now GFPpos/h-galneg (left panels), and some myoblasts originally GFPpos/h-galpos are now GFPneg/h-galpos (right panels). Scale bar = 100 Am. (C) Images of GFPneg and GFPpos cells grown clonally were obtained after 5 and 12 days, respectively. GFP expression is not modulated under clonal growth conditions. Scale bar = 200 Am.

GFP expression and hence, sca-1, by interacting with each other, either through cell – cell contacts or secreted factors. Sca-1neg cells readily differentiate into myotubes C2C12 myoblasts were sorted for sca-1neg and sca-1high cells using flow cytometry (Fig. 5A). Equal numbers of sorted cells were plated and cultured in myogenic DM. Myotubes formed by sca-1neg cells are more abundant and larger in size compared to unsorted C2C12 myoblasts (Fig. 5B). In contrast, myotube formation by sca-1pos cells is rarely observed. Any myotubes most likely are due to the presence of contaminating sca-1neg cells. In addition, GFPpos and GFPneg primary myoblasts from sca-1/GFP transgenic mice were sorted and induced to differentiate. As with C2C12 cells, GFPpos cells fail to form myotubes (Fig. 5C), even when plated at higher density or for longer periods of time (data not shown). The failure to form myotubes is not due to GFPpos cells being non-myogenic (Fig. 2A). To determine whether the lack of myotube

Fig. 5. Sca-1neg cells readily grow and differentiate into myotubes. C2C12 myoblasts were sorted on the basis of sca-1 expression. (A) Post-sort analyses of sca-1neg and sca-1pos cells. Inset: RT-PCR analyses of sca-1 expression in sca-1neg and sca-1pos cells immediately post-sort. 18S rRNA was used as a control for RNA input. (B) Equal numbers of sorted or unsorted C2C12 cells were maintained in myogenic differentiation media to induce myotube formation. Myotube formation in sca-1neg cells is more extensive than in unsorted C2C12 cells. Myotube formation is impaired in sca-1pos cells. Scale bar = 100 Am. (C) GFPneg and GFPpos primary myoblasts were sorted and induced to differentiate. GFPneg myoblasts form myotubes, whereas GFPpos cells do not. Scale bar = 100 Am. (D) RNA was harvested from sca-1neg and sca-1pos cells in myogenic differentiation media and subjected to RT-PCR analyses of myogenin and eMyHC mRNA expression. For both myogenin and eMyHC, sca-1pos cells exhibit decreased expression compared to sca-1neg cells. 18S rRNA was used as a control for RNA input. Data are representative of 3 independent experiments. (E) The proliferation of GFPneg and GFPpos primary myoblasts was analyzed by BrdU incorporation. Scale bar = 100 Am. (F) The percentage of GFPpos cells that incorporate BrdU is ¨50% that of GFPneg cells. Data are mean T SE; n = 3; *P < 0.05.

248

P.O. Mitchell et al. / Developmental Biology 283 (2005) 240 – 252

formation is due to a fusion defect, RT-PCR was performed for markers of myogenic differentiation that are expressed before myotube fusion. Myogenin is a transcription factor first expressed when cells commit to differentiation, whereas eMyHC is a later marker. Sca-1pos cells express less myogenin and eMyHC mRNAs, suggesting that the sca-1pos cells do not form myotubes as a result of a defect early in differentiation (Fig. 5D). Thus, sca-1 expression defines a subpopulation of muscle cells that are restricted in their capacity to differentiate and form myotubes. GFP pos and GFP neg cells have different growth properties The earlier appearance of clones and growth of GFPneg cells in Fig. 4 suggested that GFPneg and GFPpos cells have different proliferation rates. BrdU labeling studies showed that GFPpos cells proliferate approximately half as much as GFPneg cells (Figs. 5E,F). Furthermore, GFPpos cells grow in compact colonies, whereas GFPneg cells grow in more dispersed colonies (Fig. 4C). Increased expression of osteogenic markers in GFP pos cells Since sca-1pos C2C12 cells and GFPpos primary myoblasts do not readily express markers of myogenic differentiation and fuse, we hypothesized that these cells may instead differentiate into other lineages. Sorted GFPneg and GFPpos myoblasts were cultured in osteogenic DM in the presence and absence of BMP2 and either fixed and stained for alkaline phosphatase (AP) activity or analyzed via RTPCR for osteocalcin (OC) mRNA. AP is a metabolic enzyme involved in mineral deposition, whereas OC is an extracellular matrix protein. Without BMP2, AP is only observed in GFPpos myoblasts (Fig. 6A). However, AP is induced in GFPneg myoblasts with 100 ng/ml BMP2, albeit at much lower levels than GFPpos myoblasts. Even without BMP2, GFPpos myoblasts express OC, a marker of mature osteoblasts, whereas the GFPneg myoblasts do not (Fig. 6B). Adding BMP2 to the osteogenic DM induces expression of OC in GFPneg myoblasts and increases further the expression of OC in GFPpos myoblasts. The same results were obtained using C2C12 cells (data not shown). Therefore, GFPpos myoblasts are not generally defective in differentiation but display a specific inability to undergo myogenic differentiation. Sca-1 regulates growth and myogenic differentiation To determine if sca-1 is merely a surface marker or whether it has a functional role, sca-1 was overexpressed using a retroviral expression vector in sca-1neg and sca-1pos sorted C2C12 cells. Infection with the sca-1 retrovirus results in a 4-fold increase in sca-1 mRNA (data not shown) and a 6-fold increase in the expression of sca-1 protein (Fig. 7A). Overexpression of sca-1 decreases the number of BrdUpos cells by ¨13% (Fig. 7B). This modest decrease by

Fig. 6. Increased expression of osteogenic markers in sca-1pos cells. Sorted GFPneg and GFPpos primary myoblasts were cultured in osteogenic differentiation media T 0 – 100 ng/ml BMP2 and then analyzed for alkaline phosphatase (AP) activity and osteocalcin (OC) mRNA levels. (A) In the absence of BMP2, AP activity (dark stain) is observed only in GFPpos cells. Treatment with 100 ng/ml BMP2 increases AP activity most in GFPpos cells, in contrast to GFPneg cells, where AP activity is low. (B) RT-PCR analyses of OC mRNA expression reveal increased expression in GFPpos cells, even in the absence of BMP-2. Data shown are representative of 3 independent experiments. The same results were obtained with sorted sca1neg and sca-1pos C2C12 cells (data not shown). Scale bar = 100 Am.

sca-1 overexpression likely reflects the synergistic effects of sca-1 with other proteins in mediating its effects on cellular growth. Overexpression of sca-1 in sorted C2C12 sca-1neg cells attenuates both the number and the size of the myotubes formed (Fig. 7C). In contrast to the gain-offunction experiments, primary myoblasts from sca-1 null mice have a ¨1.5-fold increase in the number of cells that are BrdUpos compared to wildtype controls (Fig. 7D). These sca-1 null myoblasts are not impaired in their ability to form myotubes (Figs. 7E,G). However, the percentage of undifferentiated eMyHCneg cells is decreased in cultures of sca-1 null myoblasts (Figs. 7F,G). These data indicate that sca-1 is, not only a marker of a subpopulation of myogenic cells, but also serves a functional role by regulating growth and differentiation. Sca-1pos cells are important for muscle homeostasis in vivo To determine if sca-1pos myoblasts serve a physiologic role in vivo, myofiber size of sca-1 null muscles was analyzed and compared to wildtype (Fig. 7H). In 2- to 4month-old sca-1 null TA muscles, the mean cross-sectional area is increased 19% (wt: 2299 Am2; null: 2728 Am2) and is reflected by a shift of the myofiber frequency distribution to

P.O. Mitchell et al. / Developmental Biology 283 (2005) 240 – 252

249

Fig. 7. Sca-1 plays a functional role in growth and myogenic differentiation. C2C12 myoblasts were sorted on the basis of sca-1 expression and infected with control (ctrl) or sca-1 retrovirus (RV). Forty-eight hours post-infection, cells were analyzed by RT-PCR and flow cytometry for sca-1 expression by BrdU incorporation assays to determine the number of proliferating cells or switched to myogenic differentiation media. (A) Flow cytometric analyses demonstrate that sca-1 protein is increased 6-fold after infection with sca-1 RV. (B) Overexpression of sca-1 decreases proliferation of C2C12 myoblasts. Analyses of BrdU incorporation in sca-1 RV-infected cells reveal a ¨13% decrease in the number of BrdUpos cells. (C) Overexpression of sca-1 attenuates myotube formation in sca-1neg C2C12 cells and has no effect on sca-1pos cells. (D) The percentage of BrdUpos cells in sca-1 null primary myoblasts (KO) is increased compared to wildtype controls (WT). Data shown are fold increase in the percentage of BrdUpos cells relative to wildtype controls. (E) The percentage of nuclei within eMyHCpos myotubes with 2 nuclei is not significantly different between WT and KO cultures. (F) Sca-1 KO cultures have ¨50% fewer undifferentiated cells compared to WT cultures. (G) Primary myoblasts from sca-1 KO mice are not impaired in their ability to form myotubes (top panels). Fewer undifferentiated cells (represented by red arrows) are present in myotube cultures from KO as compared to WT mice (bottom panels). (H) Hematoxylin and eosin stained muscle sections from 2 – 4-month-old (TA) and 1-year-old (soleus) wildtype and sca-1 null mice. (I) Young sca-1 null mice (2 – 4 months) display increased myofiber size in TA muscles compared to wildtype. (J,K) Older sca-1 null mice (1 year) display decreased myofiber size in both the TA (I) and soleus (J) muscles. Data shown in panels (A – G) are from 3 independent experiments. Scale bar = 50 Am. *P < 0.05.

250

P.O. Mitchell et al. / Developmental Biology 283 (2005) 240 – 252

the right relative to wildtype (Fig. 7I). In contrast, in 1-yearold sca-1 null TA muscles, the mean cross-sectional area is decreased by 13% (wt: 2680 Am2; null: 2362 Am2), and the frequency distribution is shifted to the left relative to wildtype (Fig. 7J). A similar trend is observed in Sca-1 null soleus muscles with mean cross-sectional area reduced by 21% (wt: 1751 Am2; null: 1400 Am2), and the frequency distribution shifted to the left relative to wildtype (Fig. 7K). No difference is observed either in the number of myofibers or in the proportion of Type I and Type II myofibers between wildtype and null muscles at any age (data not shown). The decrease in myofiber size seen in older muscles may be the result of alterations in either the number or function of satellite cells. The number of CD34pos cells underneath the basal lamina of older muscles was analyzed and did not differ between wildtype and sca-1 null muscles (data not shown). These data suggest that the function of satellite cells in sca-1 null muscles is impaired with age. We were unable to test directly the properties of sca-1 null myoblasts from older mice due to the limiting number of animals available. The absence of sca-1 in capillaries, which normally express sca-1 in muscle (Zammit and Beauchamp, 2001), may contribute to some of the effects on decreased myofiber size. However, no difference in the number of capillaries in muscle is observed (wildtype: 1.49 T 0.077; sca-1 null: 1.39 T 0.058 capillaries/myofiber).

Discussion Our work is the first to describe a functional role for sca1 in primary myoblasts, where it negatively regulates growth and myogenic differentiation. Satellite cells do not express Sca-1 but give rise to myoblast populations that are heterogeneous for sca-1 expression. We show that the majority of myoblasts are sca-1neg, rapidly divide, and are capable of undergoing myogenic differentiation to form myotubes. In contrast, a minority population is sca-1pos, divides slower, and does not readily form myotubes. Sca1pos cells express a number of myogenic markers, therefore, they are not contaminating non-muscle cells. Sca-1 expression in myoblasts is not static but rather dynamically modulated by the microenvironment. Myofiber size of sca-1 null muscles is altered in an age-dependent manner, with increased size observed in younger animals and decreased size in older animals compared to control. Conflicting reports exist in the literature regarding the expression of sca-1 by satellite cells or primary myoblasts (Asakura et al., 2002; Dominov et al., 2001; Epting et al., 2004; Nicole et al., 2003; Qu-Petersen et al., 2002; Sherwood et al., 2004; Zammit and Beauchamp, 2001). We observed that omission of Triton X-100 before incubating with sca-1 antibody is critical for the detection of sca-1 in primary myoblasts (data not shown). The mechanism by which Triton X-100 abolishes immunodetection of sca-1 is unclear. Those groups who have reported detecting sca-1 in

satellite cells or primary myoblasts did not include Triton X100 before incubating cells with sca-1 antibody (Dominov et al., 2001; Epting et al., 2004; Qu-Petersen et al., 2002). Sca-1 expression is modulated differentially depending on growth conditions. Using myofibers derived from sca-1/ GFP mice, a high proportion of single myofiber explants give rise to GFPpos cells in time in culture, indicating that transcriptional activity of the sca-1 gene is increased in a subset of cells as satellite cells activate and proliferate. The potential significance of sca-1 expression in a subset of myoblasts at a time of rapid expansion of other myoblast populations is discussed below. The few sca-1pos cells observed in freshly isolated single myofibers might represent satellite cells that were activated in vivo due to a growth stimulus. Cell mixing experiments in bulk cultures demonstrated that GFP expression is modulated both positively and negatively, whereas GFP expression under clonal conditions is not modulated. The mechanisms responsible for sca-1 modulation in myoblasts are unknown, but we are currently investigating the role of cell – cell contacts vs. secreted factors in regulating sca-1 expression. Gain-of-function and loss-of-function experiments demonstrate decreased or increased myoblast proliferation, respectively. These results indicate that sca-1 is directly responsible, in part, for the growth properties of sca-1pos cells. Our results are consistent with studies in transgenic T cells (Henderson et al., 2002; Stanford et al., 1997). Sca-1 null mice have hyperproliferative T cells (Ito et al., 2003), whereas overexpression of sca-1 in T cells leads to hypoproliferation (Bonyadi et al., 2003) in response to antigens and mitogens. The significance of the lower growth rate of sca-1pos muscle cells is unknown, however, it may limit proliferation of a subset of myoblasts allowing preservation of replicative capacity. We demonstrate a further role for sca-1 in regulating myogenic differentiation of myoblasts. Interestingly, sca-1 can regulate differentiation of other cell types such as immune cells (Bamezai et al., 1995; Ito et al., 2003). Our RT-PCR analyses show that sca-1 mRNA expression increases upon myogenic differentiation in agreement with Shen et al. (2003). The percentage of sca-1pos cells does not decrease upon myogenic differentiation, thus the increase in sca-1 expression is not due to an increase in the number of sca-1pos cells. The significance of an increase in expression of sca-1 with differentiation is unknown, but Sca-1 may function to maintain a population of cells in an undifferentiated state, similar to the maintenance of satellite cells in vivo after a bout of muscle growth. In support of this hypothesis, overexpression of sca-1 in sca-1neg C2C12 cells decreases the number and size of myotubes. In contrast, fewer undifferentiated cells are present in sca-1 null myotube cultures. This may be due to an increased ability of Sca-1 null myoblasts to differentiate. Together, these data indicate that sca-1 plays an important role in determining which myogenic cells are capable of differentiating and

P.O. Mitchell et al. / Developmental Biology 283 (2005) 240 – 252

fusing at any given time, suggesting that sca-1 expression may be an indicator of the myogenic capacity of the cell. A recent study by Epting et al. (2004) used both antibodies and antisense against sca-1 and described a role for sca-1 in C2C12 myoblasts that is partially contradictory to our data reported here. In the absence of sca-1, we both observe enhanced cell proliferation. However, myotube formation is inhibited in their study, whereas we observe that sca-1 null and sca-1neg muscle cells differentiate well. This discrepancy is not just due to use of a cell line vs. primary myoblasts as in our hands we obtain consistent results with C2C12 cells and primary myoblasts. More likely the use of antibodies and antisense in C2C12 cultures that are heterogeneous for sca-1 expression account for the differences. Why these techniques would inhibit myotube formation by cells that do not express sca-1 is unclear. The downstream signaling cascades that result from Sca1 activation have been studied mostly in lymphocytes. The paradox in these studies is how sca-1, a GPI-linked molecule that lacks transmembrane and cytoplasmic regions, transduces intracellular signals. The current hypothesis is that sca-1 interacts with a ligand, thereby recruiting molecules into lipid rafts that themselves have intracellular signaling domains (Pflugh et al., 2002). Thus, sca-1 may affect signaling by remodeling the protein composition of lipid rafts. The effects of sca-1 on growth and differentiation defined here in primary myoblasts and C2C12 cells may be greater or lesser depending on the specific molecular context in which it is expressed in different myoblast populations. Identifying the ligand for muscle sca-1 is critical for understanding how sca-1 functions to regulate growth and differentiation of myoblasts. Different signaling pathways are likely to regulate sca-1’s effects on growth and differentiation as these two cellular processes in myogenesis are rarely modulated by a single signal. The signal cascades activated by engagement of sca-1 by its ligand in myoblasts rather than by an antibody (Epting et al., 2004) need to be elucidated and may shed light on how sca-1 regulates myoblast proliferation and differentiation. The existence of sca-1pos myoblasts appears to be important for tissue homeostasis as indicated by analyses of myofiber size in sca-1 null muscles. An increase in myofiber size in young muscles could stem from the enhanced proliferation and differentiation demonstrated in sca-1 null myoblasts isolated from young animals. However, with time, the enhanced proliferation of myoblasts in young animals may lead to their exhaustion resulting in the observed decrease in myofiber size in older muscles. Loss of hematopoietic stem cell function has been attributed to excessive proliferation (Cheng et al., 2000; Hock et al., 2004). We observe no decrease in the number of CD34pos satellite cells with age in sca-1 null muscles, however, we cannot determine if the 20% of satellite cells that do not express any known myogenic markers (Beauchamp et al., 2000) are affected in older sca-1 null animals. Loss of sca-1 in capillaries may contribute also to the observed phenotype

251

in older muscles, but we observe no difference in the number of capillaries. Loss of sca-1 in other cells could contribute in part to the observed decrease in myofiber size with age. A defect in capillaries or other cell types is unlikely to account for the enhanced myofiber size in young muscles, however. We propose that sca-1 expression serves as a protective mechanism to maintain an adequate pool of functional satellite cells in muscle tissue throughout the life of an organism. Satellite cells are normally quiescent but in response to growth stimuli are activated to proliferate and undergo differentiation/fusion. A proportion of satellite cells must remain undifferentiated such that the satellite cell pool is restored and available to participate in future muscle growth. Modulation of sca-1 expression may alter the properties of the satellite cell pool and lead to long-term consequences for muscle growth. Our identification of sca-1 as a critical regulator of muscle cell growth and differentiation may lead to new strategies for manipulating satellite cells in disease, repair, and aging and define novel therapeutic targets for enhancing muscle growth.

Acknowledgments We thank Cristina Bush and Erik Kline for initial experiments involving sca-1 immunostaining and osteogenic differentiation, Robert Karaffa for assistance with flow cytometry, and Dr. Thomas Rando for critical reading of the manuscript. This work was supported by National Institute of Health grants AR47314, AR48884, and DE13040 to G.K. Pavlath.

References Asakura, A., Seale, P., Girgis-Gabardo, A., Rudnicki, M.A., 2002. Myogenic specification of side population cells in skeletal muscle. J. Cell Biol. 159, 123 – 134. Bamezai, A., 2004. Mouse Ly-6 proteins and their extended family: markers of cell differentiation and regulators of cell signaling. Arch. Immunol. Ther. Exp. (Warsz) 52, 255 – 266. Bamezai, A., Palliser, D., Berezovskaya, A., McGrew, J., Higgins, K., Lacy, E., Rock, K.L., 1995. Regulated expression of Ly-6A.2 is important for T cell development. J. Immunol. 154, 4233 – 4239. Beauchamp, J.R., Heslop, L., Yu, D.S., Tajbakhsh, S., Kelly, R.G., Wernig, A., Buckingham, M.E., Partridge, T.A., Zammit, P.S., 2000. Expression of CD34 and Myf5 defines the majority of quiescent adult skeletal muscle satellite cells. J. Cell Biol. 151, 1221 – 1234. Benchaouir, R., Rameau, P., Decraene, C., Dreyfus, P., Israeli, D., Pietu, G., Danos, O., Garcia, L., 2004. Evidence for a resident subset of cells with SP phenotype in the C2C12 myogenic line: a tool to explore muscle stem cell biology. Exp. Cell Res. 294, 254 – 268. Bonyadi, M., Waldman, S.D., Liu, D., Aubin, J.E., Grynpas, M.D., Stanford, W.L., 2003. Mesenchymal progenitor self-renewal deficiency leads to age-dependent osteoporosis in Sca-1/Ly-6A null mice. Proc. Natl. Acad. Sci. U. S. A. 100, 5840 – 5845. Cheng, T., Rodrigues, N., Shen, H., Yang, Y., Dombkowski, D., Sykes, M., Scadden, D.T., 2000. Hematopoietic stem cell quiescence maintained by p21cip1/waf1. Science 287, 1804 – 1808.

252

P.O. Mitchell et al. / Developmental Biology 283 (2005) 240 – 252

Dhawan, J., Pan, L.C., Pavlath, G.K., Travis, M.A., Lanctot, A.M., Blau, H.M., 1991. Systemic delivery of human growth hormone by injection of genetically engineered myoblasts. Science 254, 1509 – 1512. Dominov, J.A., Houlihan-Kawamoto, C.A., Swap, C.J., Miller, J.B., 2001. Pro- and anti-apoptotic members of the Bcl-2 family in skeletal muscle: a distinct role for Bcl-2 in later stages of myogenesis. Dev. Dyn. 220, 18 – 26. English, A., Kosoy, R., Pawlinski, R., Bamezai, A., 2000. A monoclonal antibody against the 66-kDa protein expressed in mouse spleen and thymus inhibits Ly-6A.2-dependent cell – cell adhesion. J. Immunol. 165, 3763 – 3771. Epting, C.L., Lopez, J.E., Shen, X., Liu, L., Bristow, J., Bernstein, H.S., 2004. Stem cell antigen-1 is necessary for cell-cycle withdrawal and myoblast differentiation in C2C12 cells. J. Cell Sci. 117, 6185 – 6195. Gumley, T.P., McKenzie, I.F., Sandrin, M.S., 1995. Tissue expression, structure and function of the murine Ly-6 family of molecules. Immunol. Cell Biol. 73, 277 – 296. Gussoni, E., Soneoka, Y., Strickland, C.D., Buzney, E.A., Khan, M.K., Flint, A.F., Kunkel, L.M., Mulligan, R.C., 1999. Dystrophin expression in the mdx mouse restored by stem cell transplantation. Nature 401, 390 – 394. Hanson, P., Mathews, V., Marrus, S.H., Graubert, T.A., 2003. Enhanced green fluorescent protein targeted to the Sca-1 (Ly-6A) locus in transgenic mice results in efficient marking of hematopoietic stem cells in vivo. Exp. Hematol. 31, 159 – 167. Hawke, T.J., Garry, D.J., 2001. Myogenic satellite cells: physiology to molecular biology. J. Appl. Physiol. 91, 534 – 551. Henderson, S.C., Kamdar, M.M., Bamezai, A., 2002. Ly-6A.2 expression regulates antigen-specific CD4+ T cell proliferation and cytokine production. J. Immunol. 168, 118 – 126. Hershey, J.C., Baskin, E.P., Glass, J.D., Hartman, H.A., Gilberto, D.B., Rogers, I.T., Cook, J.J., 2001. Revascularization in the rabbit hindlimb: dissociation between capillary sprouting and arteriogenesis. Cardiovasc. Res. 49, 618 – 625. Heslop, L., Morgan, J.E., Partridge, T.A., 2000. Evidence for a myogenic stem cell that is exhausted in dystrophic muscle. J. Cell Sci. 113 (Pt. 12), 2299 – 2308. Hock, H., Hamblen, M.J., Rooke, H.M., Schindler, J.W., Saleque, S., Fujiwara, Y., Orkin, S.H., 2004. Gfi-1 restricts proliferation and preserves functional integrity of haematopoietic stem cells. Nature 431, 1002 – 1007. Ito, C.Y., Li, C.Y., Bernstein, A., Dick, J.E., Stanford, W.L., 2003. Hematopoietic stem cell and progenitor defects in Sca-1/Ly-6A-null mice. Blood 101, 517 – 523. Jankowski, R.J., Haluszczak, C., Trucco, M., Huard, J., 2001. Flow cytometric characterization of myogenic cell populations obtained via the preplate technique: potential for rapid isolation of muscle-derived stem cells. Hum. Gene Ther. 12, 619 – 628. Kastner, S., Elias, M.C., Rivera, A.J., Yablonka-Reuveni, Z., 2000. Gene expression patterns of the fibroblast growth factors and their receptors during myogenesis of rat satellite cells. J. Histochem. Cytochem. 48, 1079 – 1096. Khan, K.D., Shuai, K., Lindwall, G., Maher, S.E., Darnell Jr., J.E., Bothwell, A.L., 1993. Induction of the Ly-6A/E gene by interferon alpha/beta and gamma requires a DNA element to which a tyrosinephosphorylated 91-kDa protein binds. Proc. Natl. Acad. Sci. U. S. A. 90, 6806 – 6810. Kitzmann, M., Carnac, G., Vandromme, M., Primig, M., Lamb, N.J., Fernandez, A., 1998. The muscle regulatory factors MyoD and myf-5 undergo distinct cell cycle-specific expression in muscle cells. J. Cell Biol. 142, 1447 – 1459. Lee, J.Y., Qu-Petersen, Z., Cao, B., Kimura, S., Jankowski, R., Cummins, J., Usas, A., Gates, C., Robbins, P., Wernig, A., Huard, J., 2000. Clonal isolation of muscle-derived cells capable of enhancing muscle regeneration and bone healing. J. Cell Biol. 150, 1085 – 1100. Ma, X., Robin, C., Ottersbach, K., Dzierzak, E., 2002. The Ly-6A (Sca-1)

GFP transgene is expressed in all adult mouse hematopoietic stem cells. Stem Cells 20, 514 – 521. Majka, S.M., Jackson, K.A., Kienstra, K.A., Majesky, M.W., Goodell, M.A., Hirschi, K.K., 2003. Distinct progenitor populations in skeletal muscle are bone marrow derived and exhibit different cell fates during vascular regeneration. J. Clin. Invest. 111, 71 – 79. Mitchell, P.O., Pavlath, G.K., 2004. Skeletal muscle atrophy leads to loss and dysfunction of muscle precursor cells. Am. J. Physiol.: Cell Physiol. 287, C1753 – C1762. Mohler, W.A., Blau, H.M., 1996. Gene expression and cell fusion analyzed by lacZ complementation in mammalian cells. Proc. Natl. Acad. Sci. U. S. A. 93, 12423 – 12427. Murphy, T.J., Pavlath, G.K., Wang, X., Boss, V., Abbott, K.L., Robida, A.M., Nichols, J., Xu, K., Ellington, M.L., Loss, J.R., 2002. Retroviral vectors applied to gene regulation studies. Methods Enzymol. 345, 539 – 551. Nicole, S., Desforges, B., Millet, G., Lesbordes, J., Cifuentes-Diaz, C., Vertes, D., Cao, M.L., De Backer, F., Languille, L., Roblot, N., et al., 2003. Intact satellite cells lead to remarkable protection against Smn gene defect in differentiated skeletal muscle. J. Cell Biol. 161, 571 – 582. Pflugh, D.L., Maher, S.E., Bothwell, A.L., 2002. Ly-6 superfamily members Ly-6A/E, Ly-6C, and Ly-6I recognize two potential ligands expressed by B lymphocytes. J. Immunol. 169, 5130 – 5136. Qu-Petersen, Z., Deasy, B., Jankowski, R., Ikezawa, M., Cummins, J., Pruchnic, R., Mytinger, J., Cao, B., Gates, C., Wernig, A., Huard, J., 2002. Identification of a novel population of muscle stem cells in mice: potential for muscle regeneration. J. Cell Biol. 157, 851 – 864. Rantanen, J., Hurme, T., Lukka, R., Heino, J., Kalimo, H., 1995. Satellite cell proliferation and the expression of myogenin and desmin in regenerating skeletal muscle: evidence for two different populations of satellite cells. Lab. Invest. 72, 341 – 347. Rosenblatt, J.D., Lunt, A.I., Parry, D.J., Partridge, T.A., 1995. Culturing satellite cells from living single muscle fiber explants. In Vitro Cell. Dev. Biol.: Anim. 31, 773 – 779. Rouger, K., Brault, M., Daval, N., Leroux, I., Guigand, L., Lesoeur, J., Fernandez, B., Cherel, Y., 2004. Muscle satellite cell heterogeneity: in vitro and in vivo evidences for populations that fuse differently. Cell Tissue Res. 317, 319 – 326. Schultz, E., 1996. Satellite cell proliferative compartments in growing skeletal muscles. Dev. Biol. 175, 84 – 94. Shen, X., Collier, J.M., Hlaing, M., Zhang, L., Delshad, E.H., Bristow, J., Bernstein, H.S., 2003. Genome-wide examination of myoblast cell cycle withdrawal during differentiation. Dev. Dyn. 226, 128 – 138. Sherwood, R.I., Christensen, J.L., Conboy, I.M., Conboy, M.J., Rando, T.A., Weissman, I.L., Wagers, A.J., 2004. Isolation of adult mouse myogenic progenitors: functional heterogeneity of cells within and engrafting skeletal muscle. Cell 119, 543 – 554. Stanford, W.L., Haque, S., Alexander, R., Liu, X., Latour, A.M., Snodgrass, H.R., Koller, B.H., Flood, P.M., 1997. Altered proliferative response by T lymphocytes of Ly-6A (Sca-1) null mice. J. Exp. Med. 186, 705 – 717. Tamaki, T., Akatsuka, A., Ando, K., Nakamura, Y., Matsuzawa, H., Hotta, T., Roy, R.R., Edgerton, V.R., 2002. Identification of myogenicendothelial progenitor cells in the interstitial spaces of skeletal muscle. J. Cell Biol. 157, 571 – 577. Yablonka-Reuveni, Z., Quinn, L.S., Nameroff, M., 1987. Isolation and clonal analysis of satellite cells from chicken pectoralis muscle. Dev. Biol. 119, 252 – 259. Yaffe, D., Saxel, O., 1977. Serial passaging and differentiation of myogenic cells isolated from dystrophic mouse muscle. Nature 270, 725 – 727. Yoshida, N., Yoshida, S., Koishi, K., Masuda, K., Nabeshima, Y., 1998. Cell heterogeneity upon myogenic differentiation: down-regulation of MyoD and Myf-5 generates Freserve cells_. J. Cell Sci. 111 (Pt. 6), 769 – 779. Zammit, P., Beauchamp, J., 2001. The skeletal muscle satellite cell: stem cell or son of stem cell? Differentiation 68, 193 – 204.