Erythropoietin Stimulates Proliferation and Interferes with Differentiation of Myoblasts*

Erythropoietin (Epo) is required for the production of mature red blood cells. The requirement for Epo and its receptor (EpoR) for normal heart development and the response of vascular endothelium and cells of neural origin to Epo provide evidence that the function of Epo as a growth factor or cytokine to protect cells from apoptosis extends beyond the hematopoietic lineage. We now report that the EpoR is expressed on myoblasts and can mediate a biological response of these cells to treatment with Epo. Primary murine satellite cells and myoblast C2C12 cells, both of which express endogenous EpoR, exhibit a proliferative response to Epo and a marked decrease in terminal differentiation to form myotubes. We also observed that Epo stimulation activates Jak2/Stat5 signal transduction and increases cytoplasmic calcium, which is dependent on tyrosine phosphorylation. In erythroid progenitor cells, Epo stimulates induction of transcription factor GATA-1 and EpoR; in C2C12 cells, GATA-3 and EpoR expression are induced. The decrease in differentiation of C2C12 cells is concomitant with an increase in Myf-5 and MyoD expression and inhibition of myogenin induction during differentiation, altering the pattern of expression of the MyoD family of transcription factors during muscle differentiation. These data suggest that, rather than acting in an instructive or specific mode for differentiation, Epo can stimulate proliferation of myoblasts to expand the progenitor population during differentiation and may have a potential role in muscle development or repair.

Erythropoietin (Epo) 1 is required for the development and maturation of erythroid cells and acts to stimulate the proliferation and differentiation of erythroid progenitor cells. Mice lacking expression of erythropoietin or its receptor die in utero due to insufficient erythropoiesis in the fetal liver (1). Erythropoietin production can be induced by hypoxia and provides physiologic regulation of the red cell mass. Erythropoietin receptor is a member of the cytokine receptor superfamily characterized by a single transmembrane domain, homology in the extracellular domain that includes a WSXWS motif, and a cytoplasmic domain that does not contain a kinase motif. Binding of erythropoietin to its receptor results in receptor dimerization, increased affinity for Jak2 to the receptor's membrane proximal region, and subsequent phosphorylation of Jak2 and tyrosines on the cytoplasmic region of the receptor (2). As with other members of this superfamily such as thrombopoietin, interleukin-3, granulocyte-macrophage colony-stimulating factor, and prolactin, Jak2 is required for signaling (3,4) and Jak2 phosphorylation activates Stat5 (5,6) and other signal transduction pathways. A role for calcium has been implicated in erythropoietin activity. For example, in erythroid progenitor cells, erythropoietin activates an increase in intracellular calcium in a dose-dependent manner mediated via tyrosine phosphorylation of the erythropoietin receptor requiring the cytoplasmic tyrosine 460 (7). The ability of prolactin with retrovirally induced prolactin receptor expression to rescue late stages of erythropoiesis (8) suggests that the primary function of erythropoietin may be its activity as a viability (9,10) and proliferation factor (11) secondary to its ability to induce erythroid differentiation.
Expression of erythropoietin receptor is not restricted to the erythroid lineage and can be found on hematopoietic stem cells (12), and cells of endothelial (13,14) and neural origin (15). In select endothelium such as the chick embryo chorioallantoic membrane and mouse uterine endometrium, erythropoietin administration in vivo can stimulate neovascularization or blood vessel formation (14,16). In vascular smooth muscle cells, Epo can stimulate calcium influx and act as a vascular growth factor (17). Erythropoietin production is inducible by hypoxia, and in vivo results in increased erythropoiesis. In neuronal cells, erythropoietin provides protection from oxidative stress in culture (18), and protects neurons from ischemic damage in vivo (15). Erythropoietin is also expressed in the embryonic heart. Recent observations reveal a defect in nonhematopoietic development related to erythropoietin/erythropoietin receptor. Mice lacking erythropoietin or erythropoietin receptor expression suffer from ventricular hypoplasia and exhibit a reduction in the number of proliferating cardiac myocytes (19).
We now report on the expression of EpoR on primary satellite cells isolated from skeletal muscle and on myoblast C2C12 cells, although EpoR expression was not previously detected on analysis of whole skeletal muscle (20,21). Stimulation with Epo enhances proliferation, and reduces differentiation and fusion of these cells into myotubes. During differentiation Epo stimulation affects the production of the MyoD transcription factors, increases the expression of MyoD, associated with a less differentiated state, and suppresses or delays the increase expression of myogenin, usually expressed at high levels in myotubes. Myoblasts, but not myotubes, respond to Epo with an increase in cytoplasmic calcium. The expression of EpoR on myoblasts and their response to Epo support the hypotheses that the role of EpoR extends beyond erythropoiesis and that Epo and EpoR play a role in the proliferation and differentiation of muscle cells.

EXPERIMENTAL PROCEDURES
Generation of Transgenic Mice-Human EpoR proximal promoter fragments extending 5Ј up to Ϫ1778 bp and 3Ј to the ATG start site of translation were used to drive expression of a lacZ reporter gene (21). For generation of transgenic mice, the EpoR promoter/reporter gene construct was excised from plasmid, isolated, purified, and used for oocyte injection as described previously. Embryos from established lines were harvested from pregnant mice and stained for ␤-galactosidase activity (21). Endogenous EpoR expression was determined by RT-PCR of RNA isolated from embryonic tissue as described previously (20,21).
Cell Culture and Reagents-The human erythroid OCIM1 cell line was maintained in ␣-minimal essential medium containing 10% fetal bovine serum (FBS) (22). HeLa cells were cultured in Eagle's modified essential medium supplemented with 10% FBS. The murine C2C12 myoblast cell line was propagated in Dulbecco's modified Eagle's medium supplemented with 10% FBS (growth medium) in a humidified atmosphere of 5% CO 2 at 37°C (23,24). To induce differentiation, myoblasts were allowed to grow to approximately 80% confluence and then switched to differentiation medium (Dulbecco's modified Eagle's medium with 2% FBS). When Epo was added to the cell cultures, it was used at a final concentration of 1 unit/ml unless otherwise indicated. Primary satellite cells were isolated from freshly harvested mouse skeletal muscle as described (25) with the following modification. After dissection and mincing of muscle fragments, tissue was dissociated using 0.24% trypsin (Life Technologies, Inc., Gaithersburg, MD) and 0.1% collagenase (Roche Molecular Biochemicals, Indianapolis, IN). Cells were cultured in F-10 medium with 20% FBS, 0.5% chick extract (Life Technologies, Inc.), and penicillin/streptomycin (Life Technologies, Inc.). For proliferation assays, 10 6 C2C12 cells were seeded onto 100-mm plates, trypsinized at specific time points, diluted, and counted with a hemocytometer. Epo was added (up to 10 units/ml) as indicated under "Results." In separate experiments, fetal bovine serum (0 -20%) was added at different concentrations with and without Epo (1 unit/ml). To assay myogenic differentiation, 1 ϫ 10 6 cells were plated on 100-mm dishes in growth medium. 24 h later, the culture was changed to differentiation medium. Plates were grown in duplicate, and half received Epo at 1 unit/ml unless otherwise indicated.
Immunofluorescence-Cells grown on collagen-coated slides (Biocoat, Becton Dickinson, Bedford, MA) were fixed for 1 h in 4% paraformaldehyde, then rinsed twice for 10 min with 1ϫ PBS. Cultures were then blocked with 10% normal serum (rabbit or goat) in PBS for 30 min at room temperature. The cells were then incubated with primary antibody (EpoR or MF-20) for 2 h in 1.5% normal serum and washed three times in PBS. The primary antibodies EpoR (Santa Cruz Laboratories, Santa Cruz, CA) or MF-20 hybridoma specific for myosin heavy chain (26) were diluted 1:2000 and 1:4, respectively. Fluorescein-labeled antimouse and rhodamine-labeled anti-mouse for EpoR and MF-20, respectively, were used as secondary antibodies.
mRNA Analysis-Messenger RNA was extracted and purified from myoblasts at indicated time points using an oligo(dT)-cellulose spin column (Amersham Pharmacia Biotech) and used for Northern blot analysis. For RT-PCR, reverse transcription of isolated mRNA and amplification of cDNA were performed using Superscript II reverse transcriptase (RT) (Life Technologies, Inc.). 60 ng of mRNA were incubated with RT for 1 h at 42°C. PCR was used to assess expression of EpoR, GATA-3, Myf-5, myogenin, and S16. Amplification was carried out under the following conditions: 20 mM Tris, pH 8.3, 50 mM KCl, 2.0 mM MgCl 2 , 0.2 mM dNTPs, 0.2 mM specific upstream and downstream primers (Table I), 5% RT reaction mix, and 5 units/ml Taq DNA polymerase (Promega, Madison, WI). The PCR cycles were 94°C for 5 min for 1 cycle, 94°C for 1 min, 54°-68°C (template-dependent) for 2 min, 72°C for 3 min for 35 cycles, followed by a 7-min elongation at 72°C.
Quantitative Real-time RT-PCR-Quantitative real-time PCR assay was carried out with the use of gene-specific primers (Table II) and fluorescently labeled TaqMan probes or SYBR Green dye (Molecular Probes, Inc., Eugene, OR) in a 7700 Sequence Detector (PE Applied Biosystems, Foster City, CA). The TaqMan probe is designed to span exon junctions in order to prevent the amplification of any contaminating genomic DNA. Specific probes were used for EpoR, GATA-3, and S16 (Table III). Plasmid containing the cDNA of interest was used as template to generate a standard curve. S16 was used as an internal control.
Measurements of the Cytosolic Calcium Concentration-Cytosolic calcium concentration ([Ca 2ϩ ] c ) measurements were performed on 2-4day-old myoblasts or 7-9-day-old myotubes loaded with the fluorescent dye Fura-2. Cells grown on gelatin-coated glass coverslips were washed three times with control physiological salt solution (PSS; composition, in mM: NaCl 145, KCl 5, MgCl 2 1, HEPES 5, glucose 10 and CaCl 2 1.2, pH 7.4 at 37°C). Cells were incubated in the dark for 45 min at room temperature with Fura-2/AM (5 M) (Molecular Probes Inc., Eugene, OR) in PSS buffer containing 0.01% Pluronic F-127 (Molecular Probes Inc.) and 0.1% Me 2 SO. Cells grown on coverslips were washed six times with PSS, transferred to a superfusion chamber thermostated at 37°C, and superfused with PSS buffer. The chamber was mounted onto the stage of an inverted epifluorescence microscope (Nikon Diaphot, Tokyo, Japan). The cells were excited at alternating wavelengths of 340 and 380 nm with a rotating filterwheel, and emission was monitored at 510 nm with a PhoCal cell fluorescence analyzer (Life Science Resources, Cambridge, United Kingdom). The ratio R of the emitted light (510 nm) at 340 and 380 was determined and calibration was performed by treating the cells with CaCl 2 (6 mM) and ionomycin (10 M) to obtain R max , followed by the addition of EGTA (40 mM) to estimate R min . Background fluorescence, obtained by quenching the fluorescence with MnCl 2 (10 mM), was subtracted from all measurements. Results are given as [Ca 2ϩ ] c calculated as described (27). Cells were stimulated by Epo (Amgen Inc., Thousands Oaks, CA), carbachol (carbamoylcholine chloride, Fluka, Switzerland), tyrphostin A51, or incubated in absence of extracellular calcium (buffer composition, in mM: NaCl 145, KCl 5, MgCl 2 1, HEPES 5, and EGTA 3, pH 7.4 at 37°C).

RESULTS
EpoR Expression in Myoblasts-Transgenic mice produced from the lacZ reporter gene driven by the human EpoR pro-  moter (21) provided evidence that EpoR may be expressed in muscle progenitor cells or developing muscle (Fig. 1). Transgene expression was observed in fetal liver, the site of fetal hematopoiesis. Other sites of transgene expression included regions associated with developing muscle and expression of MyoD and/or Myf-5 such as the visceral arches, the proximal forelimb, and intercostal area. The facial and forelimb staining pattern were observed in embryos constructed from a promoter fragment as short as Ϫ150 bp 5Ј, and expression of the endogenous EpoR in these areas was confirmed by tissue dissection and RT-PCR. Examinations of isolated myoblasts show that they also express EpoR and are responsive to Epo stimulation. The murine C2C12 myoblast cell line (23,24) can be maintained in a proliferating state by culture in growth medium (containing 10% serum) and can be induced to differentiate and fuse into myotubes in differentiation medium (containing 2% serum). To examine EpoR expression in C2C12 cells, RNA was isolated and the presence of EpoR transcripts determined by Northern blot analysis. A specific 32 P-labeled probe was prepared by random labeling of a full-length 1.7-kilobase pair murine EpoR cDNA fragment. This probe showed specific hybridization with mRNA isolated from C2C12 cells corresponding to an EpoR transcript similar in size to that detected in the control RNA from erythroid cells ( Fig. 2A). Satellite cells, muscle progenitor cells, make up only a small fraction of 2 to 7% of the total population (28) of the adult skeletal muscle and exhibit the potential for repair and maintenance of skeletal muscle (29). To directly determine the expression of EpoR on these muscle progenitor cells, primary satellite cells were isolated from adult mouse skeletal muscle and cultured in vitro. These primary cells were analyzed within five passages. RNA was isolated and used for RT-PCR analysis (Fig. 2B). Expression of EpoR transcripts was readily detected in the satellite cell cultures using specific primer pairs (Table I). The presence of Myf-5, an early MyoD transcription factor, was used to confirm the cell type of these less differentiated myocytes. As a positive control, primers specific for mouse S16 protein were used. The low number of satellite cells or myoblasts in adult skeletal muscle would account for the lack of EpoR detection in our earlier studies of adult skeletal muscle (21).
For analysis of EpoR protein, whole cell extracts were prepared from C2C12 cells and primary satellite cell cultures. Western blot analysis showed specific binding of antibodies directed against the mouse EpoR to a band corresponding to the size of the EpoR control (Fig. 2C), confirming that the EpoR transcripts are functional and translated into EpoR protein.
For immunohistochemistry, myoblast cells were cultured in growth medium on collagen-coated slides and harvested for staining (Fig. 2D). The mouse EpoR antibody was used as primary antibody followed by fluorescein-conjugated secondary antibody. Staining of C2C12 cells and of primary satellite cells confirmed the production of EpoR protein in these cells. The antibody staining was specific and could be blocked by specific EpoR polypeptide.
Epo Stimulation of Myoblast Proliferation-Myoblasts expressing EpoR exhibited a mitogenic response to Epo. When Epo was added to the growth medium, enhancement of proliferation was observed (Fig. 3). The proliferative effect of Epo was even more apparent in differentiation medium (Fig. 3, B and D). These cultures reached confluence 48 h or more before the untreated C2C12 cells. In the absence of serum, cell death occurred within 48 h. Epo could postpone but not prevent cell death. Epo concentrations up to 100 units/ml were added to C2C12 cultures, and the proliferative effect of Epo on myoblasts was dose-dependent (Fig. 3, A and B). The optimum Epo concentration for proliferation was at 1 unit/ml for cells cultured in growth medium and in differentiation medium, a concentration used to culture primary erythroid progenitor cells (30). These results indicate that Epo induces a proliferative response in C2C12 myoblasts cultured under either growth or differentiation conditions. Primary satellite cells harvested from adult mouse skeletal muscle were also responsive to Epo and exhibited a marked effect on proliferation and differentiation. As with C2C12 myoblasts, addition of Epo to the culture medium increased proliferation of the satellite cells and extended the ability to passage these cells. As observed for C2C12 cells, the effect on proliferation was more pronounced when the cells were cultured in differentiation medium (Fig. 3E), where proliferation without Epo stimulation was markedly reduced.
Inhibition of Myoblast Differentiation-Proliferating C2C12 myoblasts continue to grow and divide when reseeded in growth medium (Fig. 4A). In contrast, after 48 -72 h in differentiation medium, the cells will begin to fuse giving rise to multinucleated cells, leading to myotube formation (Fig. 4G) and loss of proliferative capacity. When Epo was added at 1 unit/ml to cells cultured in growth medium, no change in cell morphology was observed (Fig. 4B). C2C12 cells cultured in differentiation medium with Epo retained their proliferative capacity and differentiated myotubes was reduced or delayed (Fig. 4, E and H). These cells could be reseeded indefinitely in differentiation medium in the presence of Epo. However, these cells retained their ability to differentiate when cultured in medium without Epo. When cells were exposed to Epo in growth medium and then cultured in differentiation medium without Epo, terminal differentiation and myotube formation was observed analogous to cells receiving no prior exposure to Epo. The ability of Epo to inhibit differentiation required exposure to Epo up to 24 h prior to culture in differentiation  medium and appeared to require the continued presence of Epo. Without preincubation with Epo in growth medium, no effect on differentiation and myotube formation was observed with Epo addition to differentiation medium. The specificity of Epo to inhibit differentiation of these myoblasts was confirmed by the addition to the culture medium of polyclonal antibody (gift of Alan D'Andrea) directed against the extracellular domain of the murine EpoR (Fig. 4, C, F, and I). Epo antibody directed to the cytoplasmic domain did not appear to be effective. When added together with Epo, the anti-EpoR antibody appeared to neutralize the ability of Epo to interfere with differentiation and reduce myotube formation in these cultures (Fig. 4I), comparable to culture conditions with the absence of Epo. These results confirm that Epo blocks differentiation of C2C12 myoblasts through interactions with its receptor. The inhibition of differentiation by Epo was also observed in primary satellite cell cultures.
Antibody specific for myosin heavy chain (MF-20) was used to identify the differentiated myocyte or myotube. C2C12 cells were grown in differentiation medium without Epo (Fig. 4J) were compared with cells grown in the presence of Epo (Fig.   4K). C2C12 myoblasts that were allowed to differentiate exhibited positive staining with the MF-20 antibody (Fig. 4L), while cells cultured in the presence of Epo did not stain with the MF-20 antibody (Fig. 4M). Differentiation of primary satellite cells was also confirmed by staining with the MF-20 antibody, and staining with the MF-20 antibody was not observed when satellite cells were cultured in differentiation media with Epo.
Increase of Intracellular Calcium by Epo-Epo stimulation of erythroblasts has been shown to stimulate an increase in their cytosolic calcium concentration [Ca 2ϩ ] c (31). We found that C2C12 cells at their myoblast stage responded also to Epo stimulation by increasing [Ca 2ϩ ] c (Fig. 5). C2C12 cells grown on coverslips were loaded with Fura-2, and changes in fluorescence ratios (340/380 nm excitation), indicative of [Ca 2ϩ ] c , were monitored. The Ca 2ϩ transients of myoblasts and of fused and differentiated myotubes were compared after exposure to Epo and carbachol, an analogue of acetylcholine known to stimulate [Ca 2ϩ ] c elevations via activation of the nicotinic acetylcholine receptor. Addition of 10 units/ml Epo caused an immediate increase in [Ca 2ϩ ] c in myoblasts (Fig. 5A). As this response was abolished when incubating the cells for 1 min with EGTA ( Fig.   FIG. 1. Transgenic mice containing human EpoR promoter and 5-untranslated region linked to the lacZ reporter gene. A, transgene expression driven by an EpoR promoter fragment extending to Ϫ150 bp 5Ј was determined at embryonic day 12.5 with staining in the region of the visceral arches (open triangle) and base of limbs (closed triangle). B, facial staining at day 13. C, In addition to facial staining and staining in base of limbs, a longer EpoR promoter fragment extending to Ϫ1778 bp 5Ј with a 3Ј BamHI fragment containing IVS-1 also provided staining in the intercostal regions (arrow). L indicates staining in the fetal liver.  1 and 2), primary satellite cells (lane 3), and HCD57 erythroleukemia cells included as a control (lane C) were analyzed by Western blot using anti-EpoR antibody. Lane 2 contains half the amount of protein used in lanes 1, 3, and C. D, for immunofluorescent analysis of myoblasts, C2C12 cells and primary satellite cells were incubated with anti-EpoR antibody followed by fluorescein-conjugated secondary antibody. 5C), it was entirely dependent on the presence of extracellular Ca 2ϩ , indicating stimulation by Epo of a Ca 2ϩ influx pathway. In contrast, fused myotubes of C2C12 cells did not respond to Epo. However, they responded to carbachol (100 M, CCh) with a rapid increase in [Ca 2ϩ ] c , which was similar in magnitude to the one caused by Epo (Fig. 5B), demonstrating that the differentiated myotubes were able to respond to a well known stimulus. Carbachol, however, had no effect on undifferentiated myoblasts (Fig. 5A). As the erythropoietin receptor is known to signal via tyrosine kinase pathways, we examined whether an inhibitor of tyrosine kinase (32) affected the calcium response to Epo in myoblasts. The calcium increase elicited by Epo was completely inhibited by tyrphostin A51 (10 M) in 50% of the cells and to a very large extent in remaining cells (Fig. 5D).
Stimulation of Phosphorylation by Epo-Treatment of C2C12 cells with Epo resulted in tyrosine phosphorylation of Jak2 and Stat5 in response to Epo. In erythroid progenitor cells, Epo binding to its receptor activates Jak2/Stat5 as well as other signal transduction pathways, and Jak2 plays a critical role in cellular response to Epo (3,4). Cellular extracts were isolated from C2C12 cells treated with and without Epo. Extracts were immunoprecipitated with an anti-phosphotyrosine antibody and Western-blotted with anti-Jak2 or anti-Stat5 antibody (Fig. 6). Bands corresponding to the molecular mass of Jak2 (130 kDa) and Stat5 (95 kDa) suggested increases in Jak2 and Stat5 phosphorylation upon Epo stimulation. Phosphorylation of these proteins was blocked in a concentration-dependent manner by tyrphostin A51, an inhibitor of tyrosine kinases. In contrast, no change in phosphorylation of Jak3 was observed (data not shown).
Modulation of MyoD Transcription Factor Family Expression by Epo-During myogenesis there is a sequential induction of the MyoD basic helix-loop-helix family of transcription factors that bind E-box motifs (CANNTG) with expression of Myf-5 and MyoD in early myoblasts, followed by expression of myogenin and loss of Myf-5 in the differentiated muscle cell. C2C12 cells were cultured with and without Epo under proliferative and differentiation-permissive conditions and mRNA isolated to determine the expression pattern of these transcription factors (Fig. 7A). RT-PCR analysis was carried out using primers specific for Myf-5, MyoD, and myogenin (Table I) with primers specific for S16 was used as control for the amount of mRNA used. In the absence of Epo, Myf-5 expression was observed only in undifferentiated C2C12 cells cultured in growth medium or during early exposure to differentiation medium. In the presence of Epo, Myf-5 expression persisted in differentiation medium. The levels of MyoD appeared to be induced earlier in the presence of Epo compared with levels observed in myoblasts grown without Epo in differentiation media. Differentiation of myoblasts into myotubes resulted in a marked induction of myogenin, whereas a more modest or delayed increase in myogenin was observed in cells cultured with Epo. These results suggest that Epo is able to alter the programmed expression of the MyoD transcription family members usually associated with induction of differentiation. The EpoR proximal promoter contains binding sites for tran-

FIG. 7. Expression analysis of EpoR and transcription factors in C2C12 and satellite cells.
A, C2C12 cells were cultured in growth medium for 48 h followed by 48 and 120 h in differentiation medium in the absence and presence of Epo (1 unit/ml). Expression of EpoR, Myf-5, MyoD, myogenin, and GATA-3 was determined by RT-PCR. S16 analysis was included as a positive control. M indicates the 100-bp ladder marker lane, and C indicates the negative control. B-F, C2C12 cells cultured in growth medium in the absence or presence of Epo (1 unit/ml) (B-F) and for 48 h in culture in differentiation medium in the absence and presence of Epo (1 unit/ml) (F) were subjected to quantitative real-time RT-PCR analysis. Results were normalized to S16. G, satellite cells were cultured in differentiation medium in the absence and presence of Epo (1 unit/ml) and subjected to quantitative real-time RT-PCR analysis. scription regulators Sp1 and the largely erythroid GATA-1 that are critical for transcription control in erythroid cells (33,34). In the development of the erythroid lineage, the transcription factor, GATA-1, and EpoR are closely linked and GATA-1 can transactivate the EpoR promoter. In myoblasts, we observed expression of another member of the GATA-like transcription factor family, GATA-3. In addition to altering expression of the MyoD transcription factor family members, stimulation of myoblast C2C12 cells by Epo induced the expression of GATA-3.
The induction of EpoR and MyoD transcription factors by Epo was confirmed by real-time quantitative RT-PCR (Fig. 7, B-G). C2C12 cells cultured in the presence of Epo showed an increase of about 6-fold in EpoR, 3-fold in Myf-5, 1.5-fold in MyoD, and 2-fold in GATA-3. The induction of myogenin after 2 days of culture in differentiation media was about 80-fold in the absence of Epo (Fig. 7F). In the presence of Epo, the induction of myogenin in differentiation media was 2.4-fold lower than in the absence of Epo. Epo also altered the expression of the MyoD transcription factor family in satellite cells (Fig. 7G). Note that, as C2C12 cells cultured in growth medium approached confluence (4 -5 days), we observed an increase in myogenin expression but significantly lower than that observed in differentiating C2C12 cells. When these cells were reseeded at low density, myogenin levels dropped to base-line values. The increase in myogenin levels in the presence of Epo detected for C2C12 cells in growth medium cultured for 3 days (Fig. 7F) was comparable to levels in control cultures observed 24 -48 h later. This likely reflects the increase in proliferation of C2C12 cells by Epo and the shorter time required for the cells to become confluent. Quantitation of mRNA from satellite cells cultured in differentiation media for 2 days with Epo showed a 1.4-fold increase in EpoR and Myf-5 expression. The most dramatic change was observed for MyoD, which exhibited a 10.5fold increase in the presence of Epo. The induction of myogenin was lower by a factor of 0.57 compared with the level observed in the absence of Epo.

DISCUSSION
Epo is known for its role in hematopoiesis in the stimulation of proliferation and differentiation of erythroid progenitor cells (35). The data presented here and other investigations into the tissue distribution of EpoR indicate that it is not restricted to the erythroid lineage. Epo response has been identified on cells of endothelial and neuronal origin in vitro (36,37), and in vivo in neovascularization and uterine angiogenesis (14,16) and in protecting neurons from ischemic damage (15). This current study demonstrates that EpoR is also expressed on C2C12 myoblasts and primary satellite cells and functions, upon Epo stimulation, to maintain a proliferative state and suppress differentiation. These observations suggest that Epo may participate in muscle development or repair. We found that myoblasts express EpoR and exhibit a dose-dependent growth response to Epo capable of reducing differentiation to myotubes. This activity of Epo is another example of the interplay between growth and differentiation in the control of myogenesis (40,41). The survival effect of Epo was most striking on primary satellite cells in which no growth was observed in low serum in the absence of Epo. Maturation to myotubes results in a down-regulation of EpoR, and EpoR is not detectable in mature myotubes or adult skeletal muscle, consistent with the relatively low number of satellite cells present in the total population (28). The down-regulation of EpoR expression during differentiation/maturation is analogous to that observed during erythropoiesis and EpoR is not expressed on mature red blood cells. It is likely that this down-regulation in myoblasts is a secondary event to differentiation/maturation. Culture of myoblasts under differentiating conditions with Epo results in an increase in EpoR expression accompanying cell proliferation, as also observed for Epo-stimulated differentiating erythroid progenitor cells (42). The existence of EpoR on myoblasts and other non-erythroid cells leads us to postulate a more general role for Epo: to maintain/expand the pool of proliferating progenitor cells during differentiation, not only for hematopoietic tissue, but in other tissues such as during muscle maintenance or repair, or muscle development. Currently there are two hypotheses to explain the role of growth factors in proliferation and differentiation. In the first, a stochastic model, differentiation of stem cells along specific lineages proceeds in a predetermined fashion. The role of growth factors in this model is to provide, through enabling survival and enhancing proliferation, a sufficient pool of progenitors (43). On the other hand, the instructive model gives growth factors a direct role in the determination of cell fate (44). Studies in erythroid progenitor cells show that an overexpressed prolactin receptor stimulated by prolactin supported proliferation of erythroid progenitors, which could then differentiate appropriately supporting the stochastic model (8,45). The presence of EpoR on myoblasts and their proliferative response with reduced differentiation provides further evidence for a stochastic role for Epo and its receptor, rather than Epo activating genes specifically related to erythroid differentiation and maturation, i.e. Epo acts as a survival factor to stimulate proliferation and inhibit apoptosis rather than as a specific growth factor that activates the erythroid program. Similar trophic effects in B lymphocytes (47), fetal liver stromal cells (48), endothelial cells (36), neuronal cells (49), and cardiomyocytes (50) support this hypothesis. These results suggest that the cellular response to Epo is more general and may be related to stress response or tissue development. Epo stimulation is associated with activation of the Jak2/ Stat5 (4) and other signal transduction pathways, as well as a rapid calcium influx in human erythroblasts (51). Epo binding in hematopoietic cells induces tyrosine phosphorylation of several proteins including EpoR itself and possibly calcium channel proteins (31,52). Epo has also been shown to cause a Ca 2ϩ influx in the neuronal cell line (PC12) and vascular endothelial cells (37,53). Similarly, our results show that Epo stimulation caused a significant increase in [Ca 2ϩ ] c in myoblasts, most likely resulting from an influx. Since Epo stimulates Jak2/ Stat5 and other signal transduction pathways, the inhibition of [Ca 2ϩ ] c increase by tyrphostin A51, a tyrosine kinase inhibitor, suggests that modulation of influx is downstream of the initial signaling events and requires activation of one or more signal transduction pathways. These results support previous observations that tyrosine phosphorylation has a functional role in regulation of calcium channels by Epo, suggesting that this is may be a key mechanism for Epo modulation of gene expression. Previous studies on differentiated cardiomyocytes were unable to show an increase of Ca 2ϩ influx when cardiomyocytes were treated with physiological levels of Epo (54). This observation is consistent with our findings that in contrast to myoblasts, differentiated myotubes did not display an influx of Ca 2ϩ in response to Epo. Although EpoR (Ϫ/Ϫ) mice die presumably of severe anemia around gestational day 13.5, the importance of Epo/EpoR in cardiac development is indicated by the hypoplasia of ventricles of their heart prior to death (19). No gross morphological limb changes have been observed in EpoR (Ϫ/Ϫ) mice, and skeletal muscle defects in these mice have not yet been identified.
Myogenic basic helix-loop-helix proteins such as MyoD, Myf-5, and myogenin are required for myogenesis (55). Myogenesis is realized by the binding of these factors to E-boxes (CANNTG) in the control regions of other myogenic transcrip-tion factors (56). Myf-5 and MyoD are expressed in early and mid-myogenesis. As the cells progress toward the differentiated phenotype, myogenin is expressed late in the myogenesis pathway and, once present in high levels, signals irreversible commitment to terminal differentiation (57). Epo stimulation affects the expression pattern of muscle-specific transcription factors, but the effect of Epo on C2C12 cells and primary satellite cells is transient, as observed in interruption of myogenesis by other select growth factors such as fibroblast growth factor and transforming growth factor type ␤ (41). When Epo is removed from the medium, the cells preserve their ability to differentiate. The proliferative response, induction of Myf-5 and/or MyoD, and attenuation of the induction of myogenin expression by Epo in differentiation medium raise the possibility that Epo acts analogous to fibroblast growth factor and transforming growth factor type ␤ in altering the MyoD transcription factor pattern during differentiation (41) and implicates regulation of cyclin D1 expression (58).
Satellite cells located below the basal lamina of adult skeletal muscle comprise only a small proportion (2-7%) of the nuclei associated with a muscle fiber (28, 59) depending on age and muscle type. Although ordinarily inactive, they and their descendants have a very large potential for proliferation and exist throughout adulthood as a potential source of new myoblasts. Satellite cells become activated when muscle is injured or diseased, and will divide and differentiate into myoblasts and fuse to existing, damaged muscle fibers or form new ones (60). Epo, produced primarily in the kidney and secreted into the circulation, is a stress-responsive cytokine. Epo production is highly inducible by hypoxia. The effect of Epo on satellite cells may play a role in muscle stress response, particularly to ischemia or hypoxia. We have also observed that the Epo effect on proliferation of primary satellite cells is further induced when cultured under hypoxic conditions (data not shown).
The response of myoblasts to Epo indicates that the role of Epo as a growth factor is to maintain proliferation and prevent apoptosis (38) of stimulated progenitor cells during differentiation rather than to act in an instructive mode. Observations that muscle stem cells possibly related to the satellite cell population share similarities to hematopoietic stem cells in that they are able to repopulate to some extent both muscle and hematopoietic tissue (39,46) demonstrate a more general potential of cells in the stem cell compartment than was previously appreciated. The results reported here suggest that the similarities may extend further as stem cells differentiate toward various progenitor cell lineages providing a more general role for hematopoietic cytokines/growth factors such as Epo in stimulating proliferation and providing protection from apoptosis.