|Home | About | Journals | Submit | Contact Us | Français|
Satellite cells are myogenic precursors responsible for skeletal muscle regeneration. Satellite cells are absent in the Pax-7−/− mouse, suggesting that this transcription factor is crucial for satellite cell specification [Seale, P., Sabourin, L.A., Girgis-Gabardo, A., Mansouri, A., Gruss, P., Rudnicki, M.A., 2000. Pax7 is required for the specification of myogenic satellite cells. Cell 102, 777–786]. Analysis of Pax-7 expression in activated satellite cells unexpectedly revealed substantial heterogeneity within individual clones. Further analyses show that Pax-7 and myogenin expression are mutually exclusive during differentiation, where Pax-7 appears to be up-regulated in cells that escape differentiation and exit the cell cycle, suggesting a regulatory relationship between these two transcription factors. Indeed, overexpression of Pax-7 down-regulates MyoD, prevents myogenin induction, and blocks MyoD-induced myogenic conversion of 10T1/2 cells. Overexpression of Pax-7 also promotes cell cycle exit even in proliferation conditions. Together, these results suggest that Pax-7 may play a crucial role in allowing activated satellite cells to reacquire a quiescent, undifferentiated state. These data support the concept that satellite cell self-renewal may be a primary mechanism for replenishment of the satellite cell compartment during skeletal muscle regeneration.
Adult skeletal muscle tissue possesses a tremendous capacity for regeneration in response to acute injury or chronic disease. Because the nuclei within skeletal muscle fibers are terminally differentiated and thus thought to be incapable of participating in muscle repair, these responses are largely attributed to a distinct and small population of myogenic progenitors that are resident in adult skeletal muscle, called satellite cells (Hawke and Garry, 2001; Schultz and McCormick, 1994; Smith et al., 1994). This cell population is physically distinct from the adult myofiber as they reside between the sarcolemma and the basal lamina in a nonproliferative, quiescent state (Mauro, 1961; Schultz et al., 1978). Upon stimuli such as muscle injury, satellite cells become activated, proliferate extensively and ultimately, fuse with existing fibers, or fuse together to form new myofibers (Hawke and Garry, 2001; Schultz and McCormick, 1994; Smith et al., 1994). A critical aspect of muscle tissue regeneration is the replenishment of the satellite cell population, originally proposed to occur via self-renewal (Angello and Hauschka, 1996; Baroffio et al., 1996; Schultz et al., 1986). Recently, the capacity for satellite cell self-renewal and the lineal origin of the satellite cell population has been questioned and generated considerable controversy (Ferrari et al., 1998; Hawke and Garry, 2001; Seale and Rudnicki, 2000).
The developmental origin of satellite cells remains unclear, as do the molecular mechanisms responsible for maintaining the satellite cell population. The lack of satellite cells in the newborn Pax-7 null mouse suggests that this paired-box transcription factor is critical for satellite cell biogenesis (Seale et al., 2000). Although Pax-7−/− mice lack satellite cells in all the skeletal muscle groups examined, there appear to be no major structural or patterning defects in the differentiated muscle (Seale et al., 2000). Despite these extensive descriptive analyses of the Pax-7−/− mouse, there is no detailed information on the behavior of Pax-7 protein in activated and proliferating satellite cells. More importantly, little is known regarding potential regulatory interactions between Pax-7 and the MyoD transcription factor family. In differentiating cultures of human fetal myoblasts, Pax-7 is co-expressed with MyoD and myf-5, but is not present in cells expressing myogenin (Reimann et al., 2004). Similarly, ectopic expression of Pax-7 in 10T1/2 cells fails to induce myogenin, yet ectopic expression of Pax-7 in skeletal muscle side population cells that are CD45+/Sca1+ induces the myogenic program (Seale et al., 2004). Thus, it appears that Pax-7 may regulate myogenesis via the MyoD family of transcription factors.
To further clarify a role for Pax-7 in satellite cells, we examined the timing and distribution of Pax-7 protein expression in satellite cells on isolated adult muscle fibers, primary satellite cell-derived clonal cultures and satellite cell-derived cell lines. We found that: (1) Pax-7 is differentially expressed in satellite cells as well as in satellite cell progeny; (2) Pax-7 and myogenin expression are mutually exclusive in differentiating adult myoblasts; and (3) Pax-7 is up-regulated in a small population of cells that exit the cell cycle, yet escape differentiation, potentially representing a self-renewing satellite cell. In support of this idea, we show that overexpressed Pax-7 interferes with MyoD-dependent transcriptional activation, down-regulates MyoD expression, and prevents myogenin induction. Overexpression of Pax-7 also induces cell cycle exit even under growth conditions. The loss of MyoD protein, interruption of MyoD function, and exit from the cell cycle induce a mitotically quiescent state in cultured satellite cells qualitatively similar to that of quiescent satellite cells in vivo. Thus, our data supports a model whereby Pax-7 is required for acquisition of the satellite cell phenotype; this is consistent with the apparent loss of satellite cells seen in Pax-7−/− muscle tissue and suggests that satellite cells are capable of self-renewal.
Adult primary myoblasts were obtained from hind limb muscles of 10–20-week-old B6D2F1 mice. Briefly, dissected muscle was digested with collagenase type I (Worthington) for approximately 1 h. A single cell suspension was obtained after filtering the samples through a 70-μm mesh (Falcon). For clonal cultures, approximately 500 cells/cm2 were plated on collagen coated-dishes and cultured for 48 h in F12-C supplemented with 15% horse serum and 1 nM FGF-2, at 37 °C and 5% CO2. Differentiation was induced by incubating the cultures in F12-C supplemented with 2% horse serum and 60 μg/ml insulin. When required, 10 μM BrdU was added for 1 h before fixation.
Myofibers and their associated satellite cells were prepared as described previously (Cornelison and Wold, 1997), with some modifications. Briefly, muscle was dissected from adult mouse hind limbs and digested with collagenase type I (Worthington) for approximately 45 min to yield single intact fibers that were either fixed or cultured in F12-C supplemented with 15% horse serum and 1 nM FGF-2, for 48 h. When required, 10 μM BrdU was added for 2 h before fixation.
MM14 and C2C12 myoblasts were plated at 100–500 cells/cm2 and cultured for 24 h (F12-C, 15% horse serum, 500 pM FGF-2; for MM14 and DMEM/F12, 10% FBS; for C2C12), at 37°C and 5% CO2. When required, 10 μM BrdU was added for 1 h before fixation. C2C12 cultures were induced to differentiate in DMEM/F12, 5% horse serum at 37 °C and 5% CO2 for at least 48 h. When required, MM14 and C2C12 cells were induced to differentiate in the presence of 0.1 mM 1-β-D-arabinofuranosylcytosine (araC, Sigma), starting at 48 h after induction for a period of 24 h and 6 days, respectively. 10T1/2 cells were cultured in DMEM, 10% FBS, at 37°C and 5% CO2. In myogenic conversion assays, cultures were induced to differentiate in DMEM, 2% FBS for 48 h.
pcDNA3-Pax-7d (1.5 μg/well; kindly provided by Dr. M. Rudnicki) per well, was used to overexpress Pax-7 in MM14 and adult primary myoblasts, cultured in six-well plates (Corning), using the Lipofectamine 2000 transfection reagent (Invitrogen), following manufacturer’s instructions. Staining and BrdU incorporation were done at 24 and 48 h after transfection, as described above. For Pax-7 knockdown experiments, 200 nM custom siRNA duplexes (Dharmacon, kindly provided by Dr. X. Liu) were transfected in MM14 cells (or 10T1/2 cells, when required) using the Trans-messenger transfection reagent (Qiagen), following manufacturer’s instructions. The target sequences in Pax-7 mRNA were CCACATCCGTCACAAGATA (D1); GAATCAAGTTCGGGAAGAA (D2); CAAGATTCTGTGCCGATA (D3); and CATGAACCCTGTCAGCAAT (D4). Firefly Luciferase targeting siRNA (Dharmacon, kindly provided by Dr. X. Liu) was used as a negative control (ctrl: AGACGATGACGGAAAAAGA). Staining or whole cell lysates were performed 48 or 72 h after transfection, when required.
Myogenic conversion was induced in 10T1/2 cells by transfecting (Lipofectamine 2000, Invitrogen) 1 μg per well of the pRSV-MyoD expression vector, and 24 h after transfection, cells were induced to differentiate. Myogenic conversion was scored after 48 h of differentiation by quantifying myogenin staining. Transfection efficiency was quantified by counting all myogenin- and Pax7-positive cells per field. When required, 0.5, 1, 1.5, or 2 μg of pcDNA-Pax-7 vector was co-transfected along with 1 μg of MyoD or MyoDE47 (kindly provided by Dr. S. Tapscott) expression vectors. When required, pcDNA3 vector was used as carrier DNA to transfect 2 μg of total DNA in every case. As a measure of MyoD transcriptional activity, 1 μg per well of the 4Rtk-luc reporter was transfected in the absence or the presence of 0.4 μg of pRSV-MyoD and in the absence or presence of 0.2, 0.4, or 0.8 μg of pcDNA-Pax-7. The CMV-LacZ expression vector (0.025 μg) was co-transfected to normalize for transfection efficiency. A total amount of 2.5 μg of DNA was transfected for each well, using an empty pcDNA3 plasmid as carrier when required. After inducing differentiation for 48 h, whole cell lysates were collected and luciferase and β-galactosidase activities determined using the Dual-Light System (Applied Biosystems), following manufacturer’s instructions.
Primary myoblasts, isolated fibers, and cultured cells were fixed in 4% paraformaldehyde for 20 min. Primary antibodies and dilutions used were as follows: Mouse monoclonal anti-Pax7 (Hybridoma Bank, Iowa University) at 1:5 (cell culture supernatant); mouse monoclonal anti-BrdU (BMB) at 1:10 (cell culture supernatant); chick polyclonal anti-syndecan-4 extracellular domain (Cornelison et al., in press) at 1:1000; rabbit polyclonal anti-MyoD (Santa Cruz) at 1:30; rabbit polyclonal anti-myogenin (Santa Cruz) at 1:30; rabbit polyclonal anti-c-met (Santa Cruz) at 1:30. When the rabbit polyclonal anti-myogenin (Santa Cruz) was used, low levels of perinuclear staining were observed in all nondifferentiated myoblasts. This staining was not detected when using a monoclonal antimyogenin antibody (F5D, Hybridoma Bank); it was also absent in all nonmyogenic cells (data not shown). Both antibodies stained the cell nucleus of differentiating myogenic cells. In Western blots, the two antibodies recognized indistinguishable patterns of proteins (data not shown). The polyclonal myogenin antibody was used to allow co-staining of myogenin and Pax-7 within a single cell. Secondary antibodies were conjugated with Alexa 594, Alexa 488, Cascade Blue (Molecular Probes) or AMCA (Jackson ImmunoResearch). Vectashield (Vector Laboratories) was used as mounting media in every case. When the anti-syndecan-4 antibody was used on fibers, a specific blocking step was perform with Blokhen II (Aves Labs) at 1:10 for 1 h before blocking with 5% horse serum for double labeling. For BrdU detection, a DNA denaturation step was included by incubating cell cultures or fibers in 2 N HCl for 1 h or 4 N HCl for 30 min, respectively, at room temperature before antibody incubations.
Micrographs were taken from a Nikon (Eclipse E800) epifluorescence microscope, using Slidebook v3.0 acquisition software (Intelligent Imaging Innovations Inc.) coupled to a Cooke Sensicam digital camera. Digital deconvolution for single plane images (no neighbors) was applied to acquired images (Slidebook v3.0) before figure preparation.
Whole 10T1/2 cell extracts were obtained by incubating cells in 300–500 μl of Passive Lysis Buffer (Promega), for 15 min at room temperature. Lysates were cleared by spinning at 13,000 rpm for 10 min, at 4°C. Total protein (30–50 μg) was loaded onto 8% SDS-PAGE gels and transferred onto PVDF membranes (Millipore) for Western blotting. Primary antibodies and dilutions used were monoclonal anti-MyoD1 (NCL, Novocastra), at 1:100; monoclonal anti-Pax-7 (Hybridoma Bank, Iowa University), at 1:10 (cell culture supernatant); monoclonal anti-αTubulin (DM1A, Sigma), at 1:100. Anti-mouse or anti-rabbit HRP conjugated secondary antibodies (Promega) were used at 1:5000 and HRP activity was visualized using the Western Lightning chemiluminescence system (PerkinElmer).
Pax-7 is reported to be essential for muscle satellite cell specification in mice (Seale et al., 2000), yet the molecular mechanisms involved are unclear. To gain a better understanding of the Pax-7 requirement for satellite cell specification, we first studied the appearance and subcellular localization of Pax-7 protein in satellite cells on intact myofibers. Surprisingly, an examination of Pax-7 expression by immunofluorescence showed substantial heterogeneity in Pax-7 levels, with Pax-7 protein undetectable in 30% of satellite cells associated with freshly harvested myofibers (Fig. 1A). This unexpected result suggests that Pax-7 may not be a universal marker for satellite cells, and raises the possibility that the requirement of Pax-7 for satellite cell specification may be transient. Although Pax-7 was not present in all satellite cells (identified by syndecan-4 immunofluorescence), Pax-7 protein expression was restricted to this subset of cells and not present in other cell types within the muscle (data not shown). Heterogeneous Pax-7 expression persisted after 48 h of myofiber culture (Fig. 1B, compare upper and middle panels). Furthermore, this heterogeneity was also detected within daughter cells of the same clone (Fig. 1B, lower panel), strongly suggesting that Pax-7 is differentially expressed in activated satellite cells. The heterogeneity observed for Pax-7 does not correlate with detectable differences in MyoD staining early after activation (Fig. 1C), where MyoD staining within a clone of satellite cells is uniform and Pax-7 staining is highly heterogeneous.
To assess a role for Pax-7 in satellite cells, we established dispersed cultures of satellite cells and asked if heterogeneity in Pax-7 protein levels was maintained in the absence of the myofiber. Primary satellite cells from an adult mouse were plated at clonal density and cultured for 48 h. Clonal cultures of satellite cells remain syndecan-4-positive (Fig. 2A), and, in agreement with the observations from isolated fibers, heterogeneity in the levels of Pax-7 protein detected by immunofluorescence is also present (Fig. 2A). These results confirm that differential Pax-7 expression is maintained after satellite cell isolation and culture.
We then compared Pax-7 expression in two commonly used mouse adult myoblast cell lines (MM14 and C2C12). Consistent with the results observed in myofiber cultures and in primary satellite cell cultures, C2C12 cells are highly heterogeneous for Pax-7 and include a substantial number of cells that do not express detectable Pax-7 protein (Fig. 2B). In contrast, clonal cultures of MM14 myoblasts are homogeneous for both Pax-7 and MyoD (Fig. 2B). Quantification of Pax-7 and MyoD expression in clonal cultures further illustrates the differences between C2C12 and MM14 cells: 39 ± 14% of C2C12 cells are Pax-7-positive and 48 ± 16% are MyoD-positive out of 212 cells counted, while 95 ± 2% of MM14 cells are MyoD-positive and 98.7 ± 2.3% are Pax-7-positive out of 167 cells counted. Because the presence or absence of MyoD does not appear to correlate with heterogeneous levels of Pax-7 protein in proliferating myoblasts, we asked if myoblasts committed to differentiate and thus expressing myogenin maintained heterogeneous levels of Pax-7 protein.
When clonal cultures of satellite cells are induced to differentiate, myogenin is rapidly up-regulated (Fig. 3A). Notably, Pax-7 staining is not detectable in cells that are positive for myogenin staining, suggesting that the presence of both proteins may be mutually exclusive (Fig. 3A). Similar behavior was observed in MM14 cells (data not shown). Loss of Pax-7 protein appears to be established early in differentiation-committed cells before cell alignment or cell fusion takes place. Although Pax-7 and myogenin are also mutually exclusive in C2C12 clonal cultures, we observed a subpopulation of mononuclear cells with up-regulated Pax-7 (Fig. 3B). Surprisingly, in contrast to what was observed in proliferating cultures, clones of differentiating C2C12 cells do exhibit a consistent Pax-7-dependent characteristic: All myogenin-positive cells are Pax-7 negative, while all cells not expressing myogenin appear to be Pax-7 positive (Fig. 3B, 2 days differentiation). As differentiation proceeds, these expression patterns are maintained (Fig. 3B, 4 days differentiation). Intriguingly, in clonal cultures, we observed a gradient of Pax-7 expression that was high near differentiation foci (myogenin-positive), and low in peripheral cells located at the edges of the clones (Fig. 3B). These differences were evident early during differentiation (Fig. 3B, 2 days) and maintained later as cell fusion proceeded (Fig. 3B, 4 days). Interestingly, a similar increase in Pax-7 expression has been documented during differentiation of human myoblasts (Reimann et al., 2004).
Together, these results highlight the mutual exclusivity between Pax-7 and myogenin expression, a phenomenon that appears to be established early during cell differentiation. The up-regulation of Pax-7 protein in cells that do not appear to be committed to terminal differentiation raises the possibility that dynamic regulation of Pax-7 may play a role in maintaining a population of myoblasts that appear to arise via self-renewal. Highly resembling the “reserve population” described for C2C12 cells (Yoshida et al., 1998), these cells may remain capable of additional expansion and differentiation.
To directly test whether Pax-7 is involved in maintaining a potentially self-renewing myoblast population, we asked if overexpression of Pax-7 would alter the expression of the MyoD family of transcription factors, which are not detectably expressed in satellite cells until 3–12 h following satellite cell activation (Cornelison and Wold, 1997). Overexpression of Pax-7 greatly reduced or eliminated MyoD protein as determined by immunofluorescence in cultured satellite cells and in MM14 cultures (Fig. 4A). The endogenous Pax-7 protein (untransfected cells) is not detectable due to the short exposure times used for detection of overexpressed Pax-7 (Fig. 4A, Pax-7-negative cells). Of 452 Pax-7-positive satellite cells counted, 91 ± 2% were negative for MyoD. Both MM14 cells and satellite cells exhibited similar staining patterns. Furthermore, we could not detect nuclear myogenin protein in Pax-7-positive transfected cells induced to differentiate for 24 h (Fig. 4B). Of 231 cells counted in two independent experiments, 99.9 ± 0.2% of cells overexpressing Pax-7 were negative for nuclear myogenin staining.
Pax-7-mediated down-regulation of MyoD and myogenin protein suggests that Pax-7 may act to inhibit MyoD function in myoblasts. To directly test this possibility, we asked if ectopic expression of Pax-7 would prevent MyoD-induced myogenic conversion of 10T1/2 cells. Transfection of 10T1/2 cells with a MyoD expression vector efficiently converts these cells into myoblasts (compare myogenic conversion to transfection efficiency; Fig. 5A) as determined by the presence of myogenin protein 72 h after transfection. Inclusion of increasing levels of a Pax-7 expression vector caused a dose-dependent decrease in the MyoD-dependent conversion of 10T1/2 cells (Fig. 5A). The loss of myogenic conversion cannot be attributed to a decrease in MyoD protein in co-transfected cells, as MyoD expression remained similar regardless of the amount of Pax-7 present (Fig. 5A). The decrease in myogenic conversion is also unlikely to arise from a loss of cells because the percentage of transfected cells was similar under all conditions (Fig. 5A). To determine if Pax-7 was acting at the level of muscle gene transcription or later in myogenic differentiation, we assayed for expression from a MyoD reporter containing four tandem E-box (MyoD binding site) repeats driving the luciferase reporter (4Rtk-luc) (Weintraub et al., 1990) as a function of the level of Pax-7. When the ratio of the Pax-7 to MyoD expression vectors was 1:2, the relative activity of the reporter was similar to MyoD alone (Fig 5B). Increasing the ratio of Pax-7 to MyoD to 2:1 reduced reporter activity to background levels (Fig. 5B). Together, these data suggest that ectopic expression of Pax-7 interferes with MyoD function and the ability of MyoD to induce muscle-specific gene expression. A number of mechanisms could be postulated to account for this. One possibility would be that Pax-7 alters subcellular localization of MyoD protein. When co-transfected cells were stained for MyoD and Pax-7, we found that MyoD remained localized to the nucleus even in cells transfected with the highest amount of Pax-7 expression plasmid (Fig. 5C). Thus, we suggest that a more likely mechanism for the inhibition of MyoD function by high levels of Pax-7 involves the interference with MyoD-dependent muscle gene transcription. Inhibition of MyoD-dependent transcription could occur either by interference with the function of the MyoD protein or directly by competition for MyoD DNA binding sites. To distinguish between these mechanisms, we asked if Pax-7 could overcome the effect of a tethered MyoD-E47 dimer. Co-expression of MyoDE47 alone or at a 1:1 or 1:2 ratio of the Pax-7 expression plasmid into 10T1/2 cells resulted in myogenic conversions that were indistinguishable (Fig. 5D). Thus, Pax-7 does not appear to be capable of overcoming MyoDE47-induced myogenic conversion, suggesting that Pax-7 indirectly inhibits wild-type MyoD function.
The highest levels of endogenous Pax-7 were observed in small numbers of mononuclear cells retained in C2C12 and satellite cells cultures that had been induced to differentiate. Because these cultures are maintained in low serum, we predicted that the cells with high Pax-7 levels might be quiescent and in a G0 state. To test if high levels of Pax-7 are associated with cell cycle withdrawal, we performed a short BrdU pulse in asynchronously cycling MM14 cells that were transfected with a Pax-7 expression vector. We chose to test the effects of Pax-7 on cell cycle using the MM14 cell line since proliferating MM14 cells express relatively uniform levels of Pax-7 protein (see Fig. 2). In asynchronously cycling MM14 cells, a 1 h BrdU pulse before fixation yields readily visible incorporation into the DNA and labels the expected approximately 50% of the population (Figs. 6A,B). However, transfected cells overexpressing Pax-7 protein do not detectably incorporate BrdU (Fig. 6A). Lack of BrdU incorporation was not associated with cell toxicity and programmed cell death (data not shown). Quantification reveals at least an 80% reduction in BrdU incorporation when compared to mock-transfected control cells (Fig. 6B). Surprisingly, Pax-7 overexpression in MM14 cells induces cell cycle exit without commitment to differentiation. To provide a more stringent test of this observation and an independent confirmation of these results, we treated differentiated MM14 and C2C12 cultures with 1-β-D-arabinofur-anosylcytosine (araC) to eliminate any cycling cells from the population. Typically, this approach is used to enrich C2C12 cultures for myotubes and eliminate “differentiation-defective” cells (Hinterberger and Barald, 1990). Following araC treatment for 72 h in MM14 cells and 6 days in C2C12 cultures, the cells were fixed and stained for myogenin and Pax-7 proteins. Upon araC treatment, we observed three categories of cell types ordered on their relative representation in the cultures: (1) terminally differentiated cells fused into myotubes that are myogenin positive and Pax-7 negative; (2) mononuclear differentiated cells that are myogenin-positive and Pax-7-negative; and (3) mononuclear cells with undetectable myogenin levels and high levels of Pax-7 (Fig. 6C). Thus, in myogenic cell lines, quiescent, undifferentiated cells appear to be uniquely marked with high levels of Pax-7 expression, consistent with a role for Pax-7 in exit from the cell cycle without commitment to terminal differentiation.
If these observations are indicative of the behavior of satellite cells in vivo, we predict that overexpression of Pax-7 in satellite cells on intact myofibers would down-regulate MyoD protein and prevent proliferation. Consistent with this hypothesis, we find that satellite cells on intact myofibers transfected with a Pax-7 expression vector do not detectably express MyoD (Fig 6D, upper panels), whereas those transfected with a control vector are MyoD-positive (Fig 6D, lower panels). Moreover, while control satellite cells cultured on intact myofibers for 72 h typically undergo two cell divisions and appear in clones (Cornelison and Wold, 1997; Cornelison et al., 2000, 2001), satellite cells transfected with the Pax-7 plasmid always occur as single cells (Fig. 6D), suggesting that these cells do not proliferate. We were unable to obtain sufficient numbers of transfected satellite cells for statistical analysis due to the extremely low efficiency of transfection (≤2%, data not shown), yet all satellite cells transfected with the Pax-7 expression vector have undetectable MyoD protein (Fig. 6D). Thus, overexpression of Pax-7 in satellite cells on intact myofibers and in myogenic cell lines consistently down-regulates MyoD, prevents myogenin induction, and induces exit from the cell cycle, a state that mimics quiescent adult satellite cells in normal muscle tissue.
Up-regulation of Pax-7 appears to inhibit both MyoD and myogenin expression and function and promotes cell cycle exit, inducing a mitotically quiescent, undifferentiated state. In addition, Pax-7 and myogenin proteins appear mutually exclusive in muscle cell lines and satellite cells. As elevated levels of Pax-7 appear associated with the prevention of differentiation, we asked if loss of Pax-7 promoted differentiation or affected the kinetics of differentiation. To accomplish this, we used a small interfering RNA (siRNA)-based approach to reduce the levels of Pax-7 protein in proliferating MM14 cells, which will normally express uniform levels of Pax-7. If the siRNA-induced loss of Pax-7 promotes myogenesis, muscle cells would be forced to exit the cell cycle and differentiate, inducing myogenin and down-regulating Pax-7. As we would not be able to distinguish between loss of Pax-7 occurring during normal differentiation or as a result of siRNA transfection, the ability of the siRNAs to reduce Pax-7 was initially tested in 10T1/2 cells ectopically expressing Pax-7. Cultures of 10T1/2 cells transfected with constant levels of a Pax-7 expression vector, a control siRNA, four different Pax-7 siRNAs, and the pool of all four Pax-7 siRNAs were harvested and the level of Pax-7 present determined by Western blot analysis. All of the Pax-7 siRNAs reduced Pax-7 protein levels in 10T1/2 cells expressing ectopic Pax-7 with two of them (D2 and D3) and the pooled siRNAs substantially reducing Pax-7 protein (Fig. 7A). These data confirm that these siRNAs are capable of reducing or eliminating Pax-7 protein expression. MM14 cells were co-transfected with the pooled siRNAs and a β-galactosidase expression vector to mark transfected cells, cultured for 48 h, and either grown or induced to differentiate for an additional 24 h. Immunofluorescence analysis reveals that transfected cells (β-galactosidase-positive) fail to detectably express Pax-7, while in cells transfected with a control siRNA, Pax-7 is clearly present (Fig. 7B). When siRNA-transfected cells were examined for precocious induction of myogenin protein, we found no apparent differences between control and Pax-7 siRNA-transfected cells (Fig. 7C, upper panel). Similarly, siRNA-mediated reduction of Pax-7 in differentiated cultures had no apparent effect on cell morphology, myogenin expression and localization, or the kinetics of differentiation when compared to control siRNA transfected cells (Fig. 7C, lower panel). Thus, the loss of Pax-7 protein in MM14 cells does not appear to either induce differentiation or affect the endogenous differentiation program, suggesting that the proposed role for Pax-7 during satellite cell renewal may be transient, functioning in a subset of the activated satellite cell population. These results are consistent with the described Pax-7−/− phenotypes where muscle formation appears to be normal despite the absence of satellite cells.
Skeletal muscle regeneration requires the activation and proliferation of muscle satellite cells, which eventually fuse into preexisting myofibers, or fuse to form new myofibers repairing muscle damage (Schultz and McCormick, 1994). Whether existing satellite cells are capable of self-renewal, or as yet an uncharacterized stem cell pool replenishes the satellite cell compartment is unknown and a subject of current controversy (Angello and Hauschka, 1996; Hawke and Garry, 2001; Schultz and McCormick, 1994; Seale and Rudnicki, 2000). The paired box transcription factor Pax-7 has been postulated to be necessary for satellite cell specification, based on dramatically reduced numbers of satellite cells in Pax-7−/− mice (Seale et al., 2000). Whether Pax-7 is required only for specification of satellite cells during development or whether Pax-7 is involved in skeletal muscle regeneration is not yet known. Although Pax-7 mRNA appears to be present in all satellite cells (Seale et al., 2000), we observed that Pax-7 protein levels are highly heterogeneous in quiescent satellite cells, as well as in activated and proliferating satellite cells. A similar degree of heterogeneity was seen in proliferating C2C12 cells yet Pax-7 protein levels in proliferating MM14 cells appeared relatively homogeneous. The degree of Pax-7 heterogeneity in satellite cells and C2C12 cells was unexpectedly large with Pax-7 protein undetectable in some cells and at apparently high levels in others. This heterogeneity did not appear to be dependent on the cell lineage as a similar degree of Pax-7 heterogeneity was observed both within and among cell clones. We found these results surprising, given the presumed requirement of Pax-7 for satellite cell specification and surmised that Pax-7 may also play a functional role in regulating satellite cell proliferation and differentiation.
Despite the variations in Pax-7 levels in proliferating satellite cells and satellite cell-derived cell lines, all myogenic cells expressing myogenin were consistently Pax-7 negative, supporting the conclusion that expression of Pax-7 and myogenin is mutually exclusive. Loss of Pax-7 likely occurs as myogenin is induced since both mono-nucleated and multinucleated myogenin-positive cells were Pax-7 negative. Consistent with this observation, we found that ectopic expression of Pax-7 prevented myogenic differentiation and the induction of myogenin protein. Unexpectedly, we also observed that MyoD was down-regulated in cells expressing high levels of Pax-7. Thus, Pax-7 appears to function as an inhibitor of MyoD and myogenin, preventing the expression of either protein in cells overexpressing Pax-7. If Pax-7 is preventing MyoD function directly via regulation of MyoD protein levels, then we would predict that ectopically expressed Pax-7 would down-regulate MyoD. The 10T1/2 cells, which readily commit to muscle differentiation when MyoD is ectopically expressed (Davis et al., 1987; Weintraub et al., 1989), fail to undergo myogenesis when Pax-7 is ectopically expressed with MyoD. Since MyoD protein is present, it is thus possible that Pax-7 inhibition of MyoD function occurs indirectly via changes in subcellular localization of MyoD or by inhibiting MyoD-dependent transcription. Co-expression of a Pax-7 encoding plasmid at a twofold molar excess over a MyoD plasmid completely eliminated both the ability of MyoD to induce myogenesis and transcription from a MyoD-dependent promoter, yet the subcellular localization of MyoD was unaffected. Therefore, Pax-7 appears to block MyoD function by preventing induction of MyoD-dependent muscle gene expression. Notably, high levels of Pax-7 appeared to reduce the levels of MyoD in 10T1/2 cells transfected with the highest amounts of plasmids encoding Pax-7 (data not shown). Similar to other transcription factors, MyoD has a short half-life and is degraded by an ubiquitin-dependent pathway involving the proteosome (Abu Hatoum et al., 1998). Since ubiquitin-mediated degradation of MyoD is inhibited by the DNA target sequence for MyoD, and is dependent on the formation of MyoD·DNA complexes (Abu Hatoum et al., 1998), it is tempting to speculate that high levels of Pax-7 impair MyoD binding to its DNA target sequences. However, the inability of Pax-7 to overcome the effects of the MyoDE47 tethered dimer suggest that the inhibition of MyoD may not involve competition for MyoD-binding sites on DNA but could involve competition for proteins necessary for MyoD-dependent transcription. Thus, Pax-7-mediated inhibition of MyoD·DNA complex formation could account for the failure of MyoD to activate muscle-specific reporters as well as the loss of MyoD protein. This hypothesis requires that the absolute levels of Pax-7 protein are critical; in most myoblasts, Pax-7 and MyoD are both present, with typically low levels of Pax-7. In agreement with this idea, untransfected satellite cells and C2C12 cells with the highest levels of Pax-7 typically have low or undetectable levels of MyoD protein (not shown). Although quiescent satellite cells do not detectably express myogenic transcription factors (Cornelison and Wold, 1997), proliferating satellite cells, primary myoblasts, and skeletal muscle cell lines typically express either myf-5, MyoD, or both. Therefore, we asked if high levels of Pax-7 protein affected proliferation of these cells.
In addition to inhibition and down-regulation of MyoD, overexpression of Pax-7 reduced BrdU incorporation into DNA by at least 80%, compared to myoblasts that express endogenous levels of Pax-7. The high levels of Pax-7 expressed could artifactually affect MyoD function and not represent an event occurring in normal satellite cells. However, we and others have consistently observed a small number of myogenin-negative mononucleated cells present in cultures of C2C12 cells and primary satellite cells subjected to long-term differentiation conditions (unpublished data). We observed that these cells express very high levels of Pax-7 compared to myoblasts in asynchronously proliferating cultures. A similar population of cells have been named “reserve” cells and are capable of providing differentiated myotubes and additional reserve cells upon subsequent replating (Yoshida et al., 1998). These “reserve” cells exhibit other characteristics consistent with a self-renewing population of skeletal muscle satellite cells. They appear to cycle slowly as they are resistant to araC treatment (Yoshida et al., 1998). When both MM14 and C2C12 cells are maintained under differentiation conditions for extended periods and treated with araC, a small population of myogenin-negative cells remain that express very high levels of Pax-7. These cells bear a striking resemblance to normal resident satellite cells, as they are mitotically quiescent and do not appear to express MyoD family members.
We have formulated a model that incorporates a requirement of Pax-7 for satellite cell biogenesis but not for satellite cell specification per se (Fig 8). We postulate that one mechanism that could allow for satellite cell self-renewal might involve up-regulation of Pax-7 (Fig. 8). A prediction of our model is that the satellite cells require Pax-7 to down-regulate MyoD and become quiescent. It is not yet understood what regulates the insertion of the satellite cell between the basal lamina of the muscle fiber and the fiber itself. Future experiments are aimed at further testing this model and identifying the molecular mechanisms regulating Pax-7 expression and Pax-7-mediated regulation of MyoD.
A great deal of controversy has arisen recently regarding the lineal origin of satellite cells and their capability for self-renewal. Although not inconsistent with our model, several groups postulate that satellite cells arise from lineages other than skeletal muscle (Dreyfus et al., 2004; Goldring et al., 2002; LaBarge and Blau, 2002; Seale and Rudnicki, 2000), yet the frequency of external cell participation in muscle formation and satellite cell replenishment is extremely rare and thus inconsistent with a primary role in muscle regeneration. Our work presented here provides a possible mechanism involving Pax-7 that allow satellite cells to exit the cell cycle, down-regulate MyoD, and prevent myogenin induction, phenotypes characteristic of the quiescent satellite cell. Thus, it may be possible for satellite cells to self-renew from cycling myoblasts or a subpopulation of satellite cells.
The authors acknowledge Dr. M. Rudnicki, Dr. S. Tapscott, and Dr. X. Liu for kindly providing the Pax-7 expression vector, the MyoD-E47 tethered-dimer expression vector, and the Pax-7-targeting and control siRNAs, respectively. This work was supported by grants from the NIH (AR39467) and the MDA to Bradley B. Olwin.