Original Article | Published:

PAX-FKHR function as pangenes by simultaneously inducing and inhibiting myogenesis

Oncogene volume 27, pages 20042014 (27 March 2008) | Download Citation


Alveolar rhabdomyosarcomas (ARMS) escape terminal differentiation despite exhibiting a skeletal muscle phenotype. To understand the role of the ARMS-specific PAX-FKHR proteins in myogenesis, we characterized their regulation of MyoD expression and function. Reporter assays show that PAX-FKHR transactivate MyoD expression through its 258 bp core enhancer. Gel-shift assays confirm that PAX-FKHR bind to core enhancer sequences showing similarity to consensus PAX3/PAX-FKHR-binding sites. We show that while PAX3-FKHR activates the expression of endogenous MyoD and myogenin proteins in transduced NIH3T3 fibroblasts, it inhibits them from terminally differentiating as shown by low myogenin and myosin heavy chain expression, and lack of myotube formation. Attenuation of MyoD transcriptional activity via phosphorylation coupled to the lack of cell cycle arrest is the underlying mechanism for the differentiation block. Lastly, we show that fibroblast growth factor receptor signaling likely mediates the inhibition of differentiation by PAX3-FKHR. In a single experimental system we demonstrate that PAX3-FKHR can simultaneously induce myogenesis while preventing its completion. We propose a model whereby PAX-FKHR commit a yet undefined precursor cell to the myogenic lineage while at the same time inhibit terminal differentiation, thereby contributing to ARMS formation.


Rhabdomyosarcomas (RMS) are among the most common extracranial solid tumors of childhood. RMS are subdivided into two main histopathologic classes, alveolar (ARMS) and embryonal (ERMS), which differ in their clinical presentation, response to therapy and prognosis. ARMS are characterized by specific reciprocal translocations that fuse the PAX3 or PAX7 genes to FKHR (Barr, 2001). The consequence of these translocations is the expression of the PAX3-FKHR and PAX7-FKHR (collectively, PAX-FKHR) chimeric transcription factors containing the PAX DNA-binding domains and the potent FKHR transactivation domain. The functions of PAX-FKHR are modified compared to the wild-type PAX proteins due to changes in their abundance (Davis and Barr, 1997), transcriptional activity (Fredericks et al., 1995) and target gene recognition (Epstein et al., 1998). PAX-FKHR are therefore thought to contribute to the phenotype and malignancy of ARMS by aberrantly regulating PAX-specific target genes, signaling pathways and biological processes.

Studies have identified numerous PAX-FKHR downstream effects, with recurring indications that they aberrantly control cellular growth, apoptosis and differentiation. An intriguing property of these proteins is their ability to regulate myogenesis. Expression array studies have shown that PAX3-FKHR induces a myogenic expression pattern (Khan et al., 1999). Perhaps paradoxical, PAX3-FKHR can also inhibit C2C12 myoblasts and MyoD-expressing 10T1/2 fibroblasts from terminally differentiating (Epstein et al., 1995). Although a single study has not addressed these seemingly contradictory findings, they suggest that PAX-FKHR may determine the ARMS precursor cell(s) to the myogenic lineage while simultaneously inhibiting those committed cells from terminally differentiating. This is an attractive hypothesis considering ARMS exhibit a skeletal muscle phenotype yet escape terminal differentiation.

To better understand how PAX-FKHR regulate myogenesis, we studied the relationship between PAX-FKHR and one of the key regulators of myogenesis, MyoD. MyoD is a member of the myogenic regulatory factors (MRFs) that was first identified based on its ability to convert fibroblasts into myogenic cells (Davis et al., 1987). Through its interaction with the E-proteins, and by cooperating with the MEF2 proteins, MyoD regulates muscle-specific gene expression. The activity of MyoD is controlled by post-translational mechanisms including protein–protein interaction, phosphorylation, acetylation and ubiquitination (Puri and Sartorelli, 2000). MyoD expression is regulated at the 258 bp core enhancer (Goldhamer et al., 1995) and distal regulatory region (Tapscott et al., 1992).

Here we tested whether PAX-FKHR regulate MyoD expression and function, and sought to identify the mechanisms. We show that PAX-FKHR activate MyoD expression by binding to sequences within the core enhancer. We also show that PAX3-FKHR can induce the stable expression of MyoD and myogenin in NIH3T3 fibroblasts. Despite expressing these proteins, and the terminal muscle protein myosin heavy chain (MHC), PAX3-FKHR-expressing NIH3T3 populations do not terminally differentiate. Inhibition of differentiation is attributed to the attenuation of MyoD transcriptional activity through phosphorylation, and the inability to cell cycle arrest. Together, our results place PAX-FKHR into the recently described family of pangenes, which share the ability to promote lineage commitment while simultaneously preventing differentiation (Lang et al., 2005).


PAX-FKHR activate MyoD transcription through its 258 bp core enhancer

To test whether PAX-FKHR regulate MyoD transcription, reporter assays were performed (Figure 1a). −24CAT alone had 30-fold higher activity than −2.5CAT in C2F3 myoblasts (Figure 1b). When −24CAT was co-transfected with PAX3-FKHR, activity increased 2.8-fold. We hypothesized that PAX-FKHR likely regulate MyoD through the core enhancer, and tested this using enhancer-containing reporter constructs (F3/−2.5CAT, 258/−2.5CAT). Similar to −24CAT, activity was 2.3- and 3-fold higher when co-transfected with PAX3-FKHR. Similarly, PAX7-FKHR increased 258/−2.5CAT activity 2.9-fold. To determine whether the paired domain (PD), homeodomain (HD) and transactivation domain were required, mutated PAX3-FKHR constructs were tested. Co-transfection of 258/−2.5CAT with each of the mutants resulted in chloramphenicol acetyltransferase (CAT) activities that were similar to 258/−2.5CAT alone (Figure 1c). To determine the influence of the cellular background, reporter assays were done in non-myogenic NIH3T3 fibroblasts (Figure 1d). When 258/−2.5CAT was co-transfected with PAX3-FKHR or PAX7-FKHR, CAT activity again increased. With increasing PAX3-FKHR, activation occurred in a dose-dependent manner (Figure 1e). Lastly, when PAX3 and PAX3-FKHR were co-transfected in equal molar amounts, transactivation of 258/−2.5CAT by PAX3-FKHR was competed to a third of its activity (Figure 1f).

Figure 1
Figure 1

PAX-FKHR activate MyoD transcription. (a) Diagram of the MyoD-CAT reporter constructs. PoCAT was used to subtract non-specific activity. (b, d) Reporter constructs were assayed in the absence (white bar) or presence of PAX3-FKHR (black bar) or PAX7-FKHR (gray bar). (c) 258/−2.5CAT was assayed in the presence of wild-type (P3F) or mutated PAX3-FKHR. (e) 258/−2.5CAT was assayed in the presence of increasing amounts of PAX3-FKHR. (f) 258/−2.5CAT was assayed in the presence of PAX3-FKHR, PAX3 or both. Activity is relative to −24CAT alone (b, d), 258/−2.5CAT alone (e, f) or 258/−2.5CAT plus PAX3-FKHR (c). The mean+s.d. from three experiments carried out in triplicate are shown. CAT, chloramphenicol acetyltransferase.

PAX3 and PAX-FKHR bind the MyoD core enhancer

Analysis of the core enhancer identified several candidate PD- and HD-binding sites (Supplementary Figure S1). To test for PAX3 and PAX-FKHR binding, electrophoretic mobility shift assays (EMSA) were carried out (Figure 2a). The proteins were synthesized by in vitro transcription/translation (ivtt), and HA-tagged. Western blot analysis confirmed that the ivtt lysates contained equal protein amounts (Figure 2b). Because PAX3-FKHR has a lower binding affinity than PAX3, we first localized the binding region with PAX3. Using the entire 258 bp core enhancer (probe 1), a DNA–protein complex was formed (Figure 2c, lane 2). Smaller overlapping probes 2, 3 and 4 were then tested, with PAX3 forming a complex with probes 3 and 4 (lanes 6 and 8). Assuming the binding region is located within the 64 bp overlap between probes 3 and 4, non-overlapping probes 5 and 6 were tested. Only probe 5 (lane 10) formed a complex. The complex between PAX3 and probe 5 was specific, since HA (Figure 2d, lane 3) but not His antibody (lane 4) blocked formation. The complex was competed in a dose-dependent manner by unlabeled probe 5 (lanes 5 and 7) but not probe 6 (lanes 6 and 8). Although weaker, PAX3-FKHR formed a complex with probe 5 (Figure 2e, lane 2) that was similarly blocked (lane 3) and competed (lanes 5 and 7). Competition was also seen with unlabeled PRS9 oligonucleotide (lane 9), a variant of the e5 sequence that contains PD- and HD-binding sites (Chalepakis et al., 1991). Equivalent results were seen with PAX7-FKHR (Figure 2f).

Figure 2
Figure 2

Binding to the core enhancer. (a) Diagram of the core enhancer probes. (b) Western blot analysis of ivtt lysates using HA and PAX3 antibodies. (c) EMSA using probes 1–6 and PAX3. Arrows indicate DNA–PAX3 complexes with probes 1 (arrow a), 3 (arrow b), 4 (arrow c) and 5 (arrow d). (d–f) EMSA using probe 5 and PAX3, PAX3-FKHR or PAX7-FKHR, respectively. Supershifts were carried out with HA or His antibodies; competitions were carried out with unlabeled probes 5, 6 or PRS9 (500-/1000-fold molar excess). ivtt, in vitro transcription/translation; EMSA, electrophoretic mobility shift assays.

PAX3 and PAX-FKHR bind the MyoD core enhancer via their HD and PD

To test whether binding affinities for probe 5 recapitulate differences with PRS9, a side-by-side EMSA was performed (Figure 3a). PAX3 formed complexes with both that were of comparable size and intensity (lanes 2 and 6). The same was true for PAX-FKHR (lanes 3, 4, 7 and 8). Binding differences were therefore comparable. When aligned with PRS9, probe 5 has a nearly identical PD-binding sequence (Figure 3b). Instead of GTTAC, probe 5 has GTGAC. Compared to the PRS9 HD-binding site (ATTA), probe 5 has CTTA. Based on this, the requirement of either element for binding needed to be shown. Mutations in the HD (mutHD)- and PD (mutPD)-binding sites in probe 5 were introduced, as well as 3′ (mut3′), which should not affect binding (Figure 3b). Using these probes in EMSA with PAX3 (Figure 3c) or PAX3-FKHR (Figure 3d), DNA–protein complexes formed only with mut3′ (lanes 8). These data suggest that the CTTA motif is necessary for binding. EMSA using mutated PAX3-FKHR proteins corroborated these results (Figure 3e). While ΔTAD bound to probe 5 (lane 5), BU35 and ΔHD did not. The PD and HD are therefore indispensable for core enhancer binding. To demonstrate physiologic relevance to ARMS, chromatin immunoprecipitation was performed. As shown in Figure 3f, PAX3-FKHR and/or PAX3 bind to the MyoD core enhancer in the ARMS cell line RH30.

Figure 3
Figure 3

Binding requires the HD and PD. (a) EMSA using probes 5 or PRS9 and PAX3 (white arrowhead), PAX3-FKHR or PAX7-FKHR (black arrowhead). (b) Sequence comparison between PRS9 and probe 5. Asterisks indicate identical nucleotides. HD- and PD-binding sites are boxed. Mutations are shown as bold-lowercase-underlined. (c, d) EMSA using probes 5, mutHD, mutPD or mut3' and PAX3 or PAX3-FKHR, respectively. (e) EMSA using probe 5 and wild-type or mutated PAX3-FKHR. Complexes with PAX3-FKHR (arrowhead) and ΔTAD (arrow) are indicated. (f) Chromatin immunoprecipitation was carried out using no antibody (−) or equal amounts of non-specific goat IgG or Pax3 antibody. PCR was carried out using primers to the MyoD core enhancer or upstream region. ivtt, in vitro transcription/translation; IP, immunoprecipitation; IgG, immunoglobulin G; PCR, polymerase chain reaction; PD, paired domain; HD, homeodomain; EMSA, electrophoretic mobility shift assays.

PAX3-FKHR induces endogenous MRF expression

We next created NIH3T3 populations expressing HA-tagged PAX3-FKHR (or PAX3) to ask whether PAX3-FKHR activates stable MRF expression. Four populations each were created, with PAX3-FKHR and PAX3 functionality confirmed by reporter assay (Figure 4a). All of the PAX3-FKHR populations (3T3-P3F) expressed MyoD and myogenin; Myf5 was expressed in two populations (P3F1 and P3F3) with the highest PAX3-FKHR (Figure 4b). While none of the PAX3 populations (3T3-P) expressed MRF proteins (Supplementary Figure S2a), MyoD and myogenin mRNA was detected, although 500- to 1000-fold lower than in 3T3-P3F1 (Figure 4c). To test whether MyoD induction is independent of the cellular background, PAX3-FKHR (or PAX3) populations were created in the embryonal kidney cell line 293A. Despite expressing PAX3-FKHR (or PAX3), none of the populations expressed MyoD protein (Supplementary Figure S2b). Therefore, the cellular background limits the induction of MyoD by PAX-FKHR.

Figure 4
Figure 4

PAX3-FKHR activates MRF expression in NIH3T3. (a) PAX/PAX-FKHR responsive pTK-PRS9 reporter assays in representative populations. Activity is relative to 3T3-V1. Shown are the mean+s.d. from three experiments carried out in duplicate. (b) Western blot analysis of 3T3-V and 3T3-P3F populations. PAX3-FKHR was detected with an HA antibody. (c) PAX3 (dark gray bar), PAX3-FKHR (light gray bar), MyoD (black bar) and myogenin (white bar) expression in 3T3-P populations was determined by qRT–PCR. Expression is shown relative to the highest expressing population. Error bars represent the s.e.m. from triplicate reactions. CAT, chloramphenicol acetyltransferase; qRT–PCR, quantitative reverse transcription–polymerase chain reaction; MRF, myogenic regulatory factors.

PAX3-FKHR-expressing NIH3T3 populations do not terminally differentiate

We next tested whether the 3T3-P3F populations terminally differentiate (Figure 5a). As controls, four MyoD-expressing NIH3T3 populations were created (3T3-MyoD). As expected, C2F3 myoblasts formed extensive multinucleated myotubes expressing MHC in differentiation media (DM). 3T3-MyoD1 also formed multinucleated MHC-positive myotubes (see inset). In contrast, 3T3-P3F1 only rarely exhibited single cells expressing MHC; even MHC-positive cells lying together remained mononucleated (see inset). These differences were also reflected in MHC mRNA expression (Figure 5b). 3T3-P3F1 and 3T3-MyoD1 in growth media (GM) had comparably low expression. Upon switching to DM, MHC increased 850-fold in 3T3-MyoD1 while only 35-fold in 3T3-P3F1.

Figure 5
Figure 5

Incomplete differentiation in 3T3-P3F. (a) MHC immunofluorescence in representative populations. (b) MHC expression was determined by qRT–PCR and shown relative to 3T3-MyoD1 (GM). Error bars represent the s.e.m. from triplicate reactions. (c) Western blot analysis of representative populations. PAX3-FKHR was detected with a PAX3 antibody. qRT–PCR, quantitative reverse transcription–polymerase chain reaction; MHC, myosin heavy chain; GM, growth media; DM, differentiation media.

These differences are likely explained by MyoD and myogenin expression. MyoD was relatively constant in 3T3-MyoD1 and C2F3, seen as a doublet containing equal amounts of the faster and slower migrating forms (Figure 5c). In 3T3-P3F1 (DM), however, MyoD was present almost entirely as the slower migrating form. Myogenin was expressed equally in 3T3-P3F1 and 3T3-MyoD1 (GM). Upon differentiation, myogenin was induced in 3T3-MyoD1 and C2F3 accompanied by strong MHC expression, while myogenin in 3T3-P3F1 was markedly less and MHC was barely expressed. These differences were seen in all 3T3-P3F and 3T3-MyoD populations (shown for MyoD in Supplementary Figure S3). As reference, RH30 (GM) expressed comparable PAX3-FKHR to 3T3-P3F1, and mostly the slower migrating MyoD form. The ARMS cell lines JR and RH28 gave similar results (data not shown). Myogenin was comparable to 3T3-MyoD1 and C2F3 (DM), but MHC was barely expressed. Therefore, despite determining cells to the myogenic lineage, PAX3-FKHR simultaneously inhibits them (quantitatively) from terminally differentiating.

Attenuation of MyoD transcriptional activity by phosphorylation

Because MyoD and myogenin can be inhibited by phosphorylation (Li et al., 1992), we set out to show that MyoD phosphorylation in 3T3-P3F (DM) inhibits its transcriptional activity. As shown in Figure 6a, MyoD expression patterns following immunoprecipitation and western blot analysis were identical when MyoD was examined by western blot alone. Consistent with Liu et al. (1998), MyoD in RH30 (GM) exists almost entirely in the phosphorylated form. Next, equal aliquots of the same immunoprecipitation were treated with (+) or without (−) calf intestinal alkaline phosphatase (CIP), and detected by western blot analysis (Figure 6b). Absence of the slower migrating form in the CIP-treated samples proved that MyoD was phosphorylated.

Figure 6
Figure 6

MyoD phosphorylation with attenuated transcriptional activity in 3T3-P3F1. (a, b) MyoD was immunoprecipitated, treated (+) or not treated (−) with CIP (b only), and detected by MyoD western blot analysis. (c) 4RE-Luc activity in representative populations. Activity is relative to 3T3-V1 (GM). The mean+s.d. from three experiments carried out in triplicate are shown. CIP, calf intestinal alkaline phosphatase; DM, differentiation media; GM, growth media.

MyoD transcriptional activity was tested by reporter assay (Figure 6c). 4RE-Luc contains four tandem copies of the right E-box from the muscle creatine kinase enhancer upstream of the thymidine kinase promoter and luciferase gene (Lu et al., 2000), and is therefore responsive to all MRFs. Reporter activity was constant in all populations (GM). Upon differentiation, activity increased 8.5-fold in 3T3-MyoD1 while only 1.8-fold in 3T3-P3F1. Reduced activity in 3T3-P3F1 was not due to mislocalization of MyoD or myogenin, since both were expressed in the nucleus (Supplementary Figure S4). These results suggest that 3T3-P3F escapes differentiation due to attenuated MyoD (MRF) transcriptional activity.

Ectopic MyoD does not relieve the block in terminal differentiation

Because MyoD expression in 3T3-P3F is less than in 3T3-MyoD, it may be that insufficient MyoD is present to drive terminal differentiation. To test this, MyoD was ectopically expressed in 3T3-P3F1 and 3T3-V1 so as to compare differentiation in populations expressing equivalent MyoD. Polyclonal populations transduced with vector (−BP) or MyoD were created. Despite expressing equivalent MyoD, V1-MyoD expressed equal amounts of both MyoD forms in DM while P3F1-MyoD expressed predominantly phosphorylated MyoD (Figure 7a). Myogenin was induced to similar levels in both populations (DM). MHC was highly expressed in V1-MyoD (DM), in both multinucleated myotubes (see inset) and mononucleated cells (Figure 7b). In contrast, P3F1-BP and P3F1-MyoD expressed relatively little MHC in few mononucleated cells. These differences were also reflected in MyoD (MRF) transcriptional activity (Figure 7c). Therefore, ectopic MyoD did not elicit terminal differentiation.

Figure 7
Figure 7

Effects of ectopic MyoD in 3T3-P3F1. (a) Western blot analysis of the double-transduced populations. PAX3-FKHR was detected with an HA antibody. Note that MyoD in P3F1-BP (GM and DM) is undetectable due to the short autorad exposure. In a longer exposure, however, MyoD is shown to be predominantly phosphorylated in DM compared to GM. (b) MHC immunofluorescence in populations grown in DM. (c) 4RE-Luc activity in populations grown in DM. Activity is relative to V1-BP. The mean+s.d. from five experiments carried out in triplicate are shown. GM, growth media; DM, differentiation media; MHC, myosin heavy chain.

PAX3-FKHR-expressing NIH3T3 populations do not withdraw from the cell cycle

Because differentiation requires cell cycle arrest, DNA synthesis and cell cycle distribution were compared. 3T3-MyoD1 withdrew from the cell cycle, as the percentage of 5-bromodeoxyuridine (BrdU)-positive cells decreased to 5% in DM (Figure 8a). In contrast, 3T3-P3F1 continued to proliferate in DM, with 34% of the cells undergoing DNA synthesis. These differences were also present in the double-transduced populations.

Figure 8
Figure 8

PAX3-FKHR-expressing NIH3T3 populations do not cell cycle arrest. (a) BrdU incorporation and cell cycle distribution. Values were from triplicate (top) or duplicate (bottom) determinations and shown as the mean±s.d. (b, c) Western blot analysis of cell cycle regulators. BrdU, 5-bromodeoxyuridine; DM, differentiation media; GM, growth media.

We next examined whether the failure to terminally differentiate correlated with altered cell cycle protein expression (Figure 8b). C2F3 myoblasts were included for comparison, giving results similar to previous studies. In C2F3 (DM), cyclin D3, p21 and p27 (data not shown) increased; cyclin D1, cyclin E and cyclin A decreased. Rb went from hyperphosphorylated (GM) to hypophosphorylated (DM). Cyclin D2, cdk1, cdk2 and cdk4 were relatively unchanged (data not shown). While p21 and cyclin D3 increased in 3T3-MyoD1 (DM), they were lower in 3T3-P3F1. Rb went from hyper- and hypophosphorylated (GM) to hypophosphorylated (DM) in 3T3-MyoD1. Rb was hyper- and hypophosphorylated in 3T3-P3F1 regardless of growth condition. In DM, cyclin D1 decreased in 3T3-P3F1 but not in 3T3-MyoD1, cyclin E was elevated in both populations, and cyclin A decreased in 3T3-MyoD1 while remaining elevated in 3T3-P3F1. No significant differences were seen for cyclin D2, cdk1, cdk2, cdk4 or p27 (data not shown).

Several of these differences were also seen in the double-transduced populations (Figure 8c). p21 increased in V1-MyoD (DM); Rb went from hyper- and hypophosphorylated (GM) to hypophosphorylated (DM). In contrast, p21 was lower in P3F1-MyoD (DM), and Rb was hyper- and hypophosphorylated. Cyclin D3 increased similarly in V1-MyoD and P3F1-MyoD (DM). Cyclin A was strong in both P3F1 populations (DM), while barely detectable in V1 populations. No significant differences were found for the other proteins mentioned (data not shown). In summary, ectopic MyoD did not induce cell cycle arrest.

Fibroblast growth factor receptor signaling likely mediates the differentiation block

We recently identified a PAX-FKHR expression signature in primary ARMS that suggests these proteins inhibit myogenic differentiation (Davicioni et al., 2006). Of interest was the identification of fibroblast growth factor receptor (FGFR)4 in the signature, since FGFR signaling can negatively regulate MyoD and myogenin. To test whether FGFR might mediate the differentiation block in 3T3-P3F, FGFR expression was examined (Figure 9a). FGFR1, FGFR2 and FGFR4, but not FGFR3 (data not shown), were expressed in all populations. Expression of each increased in 3T3-P3F (DM), with FGFR4 showing the greatest increase. Total expression was also significantly higher than in 3T3-MyoD. Furthermore, increased FGFR expression in 3T3-P3F (DM) was accompanied by continued or increased expression of several of their ligands (that is, fibroblast growth factor (FGF)1, FGF2, FGF8 and FGF9) (data not shown).

Figure 9
Figure 9

Enhanced differentiation of 3T3-P3F1 in serum-free media. (a) FGFR expression was determined by qRT–PCR. Expression is shown relative to the highest expressing population. (b) Western blot analysis of 3T3-P3F1. (c) MHC expression was determined by qRT–PCR. Expression is shown relative to GM. Error bars in (a, c) represent the s.e.m. from triplicate reactions. (d) 3T3-P3F1 was grown as in (c), and MHC expression determined by immunofluorescence. The arrow (phase contrast) indicates a double-nucleated cell. bFGF, basic fibroblast growth factor; DM, differentiation media; FGFR, fibroblast growth factor receptor; GM, growth media; ITS, media containing only insulin, transferrin and sodium selenite; MHC, myosin heavy chain; qRT–PCR, quantitative reverse transcription–polymerase chain reaction.

We next assessed 3T3-P3F1 differentiation in serum-free media (ITS; media containing only insulin, transferrin, sodium selenite), therefore devoid of FGFs. Compared to DM, myogenin and MHC increased in ITS (Figure 9b). While MyoD remained predominantly phosphorylated in ITS, there was more unphosphorylated MyoD that correlated with higher MyoD (MRF) transcriptional activity (data not shown). Addition of 10 ng ml−1 basic fibroblast growth factor (bFGF) to ITS led to reduced myogenin and MHC expression, with little change in MyoD. These effects were bFGF dose dependent (Figure 9c). Growth in ITS also led to phenotypic changes; elongated cells with double-nuclei expressing MHC were seen (Figure 9d, arrow). This was also inhibited in a dose-dependent manner. While cells grown in ITS containing 1 ng ml−1 bFGF contained double-nucleated MHC-expressing cells, those in 10 or 20 ng ml−1 bFGF contained fewer and only mononucleated cells. Together, these results suggest that FGFR signaling contributes to the differentiation block in 3T3-P3F.


The histologic RMS subtypes exhibit varying degrees of muscle differentiation, which resemble different stages in the embryogenesis of skeletal muscle. Despite expressing myogenic markers, ARMS fail to differentiate. The ARMS cell(s)-of-origin has not been established, and whether the myogenic phenotype stems from the cell lineage and/or is a consequence of PAX-FKHR remains unclear. The aim of this study was to understand how PAX-FKHR control myogenesis.

We show that chromosomal translocation is an additional path to pangene-like deregulation of differentiation. Pangene refers to the Greek god Pan and Peter Pan, both of whom could orchestrate complex events while never growing to maturity. This term was used by Lang et al. (2005) to describe Pax3 in melanocyte stem cell differentiation. Pax3 activates Mitf expression, which drives melanogenesis, while also competing with Mitf for downstream targets thereby preventing terminal differentiation. This mechanism of simultaneous induction of commitment and maintenance of an undifferentiated state was proposed as a general paradigm for developmental and stem cell biology. Several PAX genes have been described to contribute to tumorigenesis as pangenes, including PAX3 and PAX7 (Robson et al., 2006). Our results add to the emerging role of pangenes in tumorigenesis by showing that PAX-FKHR can simultaneously initiate myogenesis and inhibit terminal differentiation. This occurs in part through MyoD.

We propose a model in which PAX-FKHR influence ARMS pathogenesis by controlling the expression and function of MyoD (Supplementary Figure S5). MyoD and myogenin are expressed in most ARMS, while Myf5 is expressed at much lower levels and less frequently. This expression pattern resembles our PAX3-FKHR-expressing populations. Khan et al. (1999) reported similar results using expression arrays. PAX3-FKHR has also been shown to transactivate myogenin directly, independent of MyoD (Zhang and Wang, 2007). Therefore, it is possible that PAX-FKHR also activate myogenin directly during ARMS formation. Of particular interest is the similarity in MyoD phosphorylation in our model system and ARMS cell lines. It is likely that MyoD phosphorylation in 3T3-P3F (DM) occurs at Threonine-115 (Thr115), as was suggested by Liu et al. (1998) for RH30. Thr115 is especially attractive since this site is phosphorylated by protein kinase C in response to FGF/FGFR signaling (Li et al., 1992). Although other phosphorylation sites have been reported, our data argue against those that induce MyoD degradation or enhance MyoD function (Puri and Sartorelli, 2000). Future studies should hopefully identify which FGFR(s) regulates this process. Complicating these studies will be the fact that these receptors (and their ligands) belong to a multigene family. Also, the fact that MyoD in 3T3-P3F1 still remained predominantly phosphorylated in ITS may suggest that autocrine signaling occurs, which cannot be bypassed by simply removing FGFs. Equally important will be to identify which FGF(s) mediates this signaling, which may not necessarily be FGF2 (as was used here). Finally, additional PAX-FKHR roles in the model include induced proliferation (Anderson et al., 2001a), malignant transformation (Lam et al., 1999) and suppression of apoptosis (Bernasconi et al., 1996). Oncogene activation and tumor suppressor loss would occur secondary to PAX-FKHR formation/expression, thereby aiding the enhanced proliferation and inhibition of differentiation.

Our results suggest that PAX-FKHR contribute to the attenuation of MyoD transcriptional activity and the inhibition of cell cycle arrest. To what extent they are interdependent will require further study. Failure of our PAX3-FKHR populations to terminally differentiate correlated with the expression of hyperphosphorylated Rb. Besides being necessary for irreversible cell cycle withdrawal, Rb is required for activation of late differentiation markers and MEF2 transcriptional function (Novitch et al., 1996, 1999). Although not shown, MEF2 induction and transcriptional activity were severely reduced in all PAX3-FKHR populations (DM).

Our study suggests that the myogenic phenotype of ARMS may reflect the induction of incomplete differentiation in a cell-of-origin that is not a committed muscle cell precursor. This mechanism fits the gain-of-function model whereby PAX-FKHR deregulate physiologic pathways of PAX3/PAX7. Keller et al. (2004) have shown that Pax3-Fkhr expression in postnatal, terminally differentiating myf-6-expressing myofibers can give rise to tumors similar to the solid variant of ARMS. Tumor formation is dependent upon p53 or Ink4a/Arf deletion, questioning the oncogenicity of PAX3-FKHR. Because the authors could not exclude the possibility that tumors arose from non-differentiating cells, it would be interesting whether their model relies on reactivation of MyoD and myogenin by PAX-FKHR. Our future studies will focus on identifying a non-muscle committed ARMS stem cell. Equally promising is the prospect of identifying and targeting pathways involved in mediating the differentiation block. Release of this block could lead to a more differentiated state, which in ERMS is directly related to better therapy response (Schmidt et al., 1986). This may be a new therapeutic approach for ARMS, which especially in its progressed stages remains a challenge.

Materials and methods

Cell lines and culturing

NIH3T3, 293A and 293T cells were grown in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum; C2F3 was grown in Dulbecco's modified Eagle's medium containing 20% fetal bovine serum; RH30 was grown in RPMI 1640 medium containing 10% fetal bovine serum. A complete description of all cell lines and transduced populations is found in Supplementary Table S1. For differentiation, cells were switched to Dulbecco's modified Eagle's medium with 2% horse serum or Dulbecco's modified Eagle's medium/F12 with ITS (5 μg ml−1 insulin, 5 μg ml−1 transferrin, 5 ng ml−1 sodium selenite; Roche, Indianapolis, IN, USA). Myogenesis was assessed 72 h later.

Expression constructs

Expression constructs were as described (Anderson et al., 2001b). PAX3-FKHR mutations were generated with the Chameleon or ExSite mutagenesis kits (Stratagene, La Jolla, CA, USA). BU35 contains an R56L mutation within the PD. ΔHD-C contains an in-frame deletion from Q254-M284 within the HD. ΔTAD contains a C-terminal deletion from Q776 to the stop codon. All primers are shown in Supplementary Table S2.

Reporter assays

Transfections and CAT assays were as described (Anderson et al., 2001b). Cells were transfected with 1 μg pcDNA3-lacZ, 0.2 pmol MyoD reporter construct (Goldhamer et al., 1992, 1995) and 0.6 pmol expression construct (MyoD); 0.2 pmol 258/−2.5CAT and 0.06, 0.2 or 0.6 pmol PAX3-FKHR (dose-dependence); 1 μg pcDNA3-lacZ and 1 μg pTK-PRS9 (PRS9); 1 μg 4RE-Luc and 1 ng phRL-TK (4RE-Luc). Luciferase assays were carried out using the Dual-Luciferase Assay (Promega, Madison, WI, USA), with activity normalized to Renilla luciferase activity from phRL-TK. Assays were carried out 72 h post-transfection.

EMSA and chromatin immunoprecipitation

Gel-shift assays were carried out essentially as described (Chalepakis et al., 1991). A complete description is found in the Supplementary Information. Chromatin immunoprecipitation analysis was carried out using the CHIP-IT Express kit (Active Motif, Carlsbad, CA, USA) with enzymatic-sheared chromatin from RH30 and Pax3 antibody (Abcam, Cambridge, MA, USA).

Retroviral transduction and fluorescence-activated cell sorter

PAX3-FKHR and PAX3 were expressed from MSCV-IRES-GFP, mouse MyoD from MSCV-IRES-GFP and pBabe-Puro. Viral supernatants were made in 293T as described previously (Davicioni et al., 2006). Cells were transduced and fluorescence-activated cell sorted 48–72 h later for green fluorescent protein expression.

Western blotting, immunofluorescence and qRT–PCR

Equal protein amounts (40 μg cell lysate, 2 μl ivtt lysate) were fractionated by sodium dodecyl sulfate polyacrylamide gel electrophoresis. Antibodies used were: HA (Covance, Richmond, CA, USA); PAX3 (Fredericks et al., 1995); MyoD, myogenin, cyclin D1, cyclin D3, Rb, p21 (BD Pharmingen, San Jose, CA, USA); Myf5, cyclin A, cyclin E (Santa Cruz, Santa Cruz, CA, USA); β-actin (Sigma, St Louis, MO, USA); MHC (Developmental Studies Hybridoma Bank, Iowa City, IA, USA). Immunofluorescence was carried out on cells grown on collagen-coated glass coverslips. Bound primary antibody was detected with Cy3-labeled secondary antibody. Nuclei were stained with DAPI. Quantitative reverse transcription–polymerase chain reaction (qRT–PCR) was performed as described previously (Davicioni et al., 2006).

MyoD immunoprecipitation and CIP treatment

Immunoprecipitation and CIP treatment were carried out essentially as described by Liu et al. (1998). A complete description is found in the Supplementary Information.

Cell cycle analysis

BrdU incorporation was analysed by immunofluorescence using the B44 BrdU antibody (Becton Dickinson, San Jose, CA, USA). Cells were exposed to 20 μM BrdU for 3 h before fixation. Cell cycle analysis was done by flow cytometry. Trypsinized cells were fixed in 70% ethanol, treated with 40 μg ml−1 DNase-free RNase A, stained with 40 μg ml−1 propidium iodide, and analysed by fluorescence-activated cell sorter.


  1. , , , . (2001a). PAX3-FKHR induces morphological change and enhances cellular proliferation and invasion in rhabdomyosarcoma. Am J Pathol 159: 1089–1096.

  2. , , , . (2001b). Embryonic expression of the tumor-associated PAX3-FKHR fusion protein interferes with the developmental functions of Pax3. Proc Natl Acad Sci USA 98: 1589–1594.

  3. . (2001). Gene fusions involving PAX and FOX family members in alveolar rhabdomyosarcoma. Oncogene 20: 5736–5746.

  4. , , , , . (1996). Induction of apoptosis in rhabdomyosarcoma cells through down-regulation of PAX proteins. Proc Natl Acad Sci USA 93: 13164–13169.

  5. , , , , , . (1991). The molecular basis of the undulated/Pax-1 mutation. Cell 66: 873–884.

  6. , , , , , . (2006). Identification of a PAX-FKHR gene expression signature that defines molecular classes and determines the prognosis of alveolar rhabdomyosarcomas. Cancer Res 66: 6936–6946.

  7. , . (1997). Fusion genes resulting from alternative chromosomal translocations are overexpressed by gene-specific mechanisms in alveolar rhabdomyosarcoma. Proc Natl Acad Sci USA 94: 8047–8051.

  8. , , . (1987). Expression of a single transfected cDNA converts fibroblasts to myoblasts. Cell 51: 987–1000.

  9. , , , , . (1995). Pax3 inhibits myogenic differentiation of cultured myoblast cells. J Biol Chem 270: 11719–11722.

  10. , , , . (1998). Tumor-specific PAX3-FKHR transcription factor, but not PAX3, activates the platelet-derived growth factor alpha receptor. Mol Cell Biol 18: 4118–4130.

  11. , , , , , et al. (1995). The PAX3-FKHR fusion protein created by the t(2;13) translocation in alveolar rhabdomyosarcomas is a more potent transcriptional activator than PAX3. Mol Cell Biol 15: 1522–1535.

  12. , , , , , . (1995). Embryonic activation of the myoD gene is regulated by a highly conserved distal control element. Development 121: 637–649.

  13. , , , . (1992). Regulatory elements that control the lineage-specific expression of myoD. Science 256: 538–542.

  14. , , , , , . (2004). Alveolar rhabdomyosarcomas in conditional Pax3:Fkhr mice: cooperativity of Ink4a/ARF and Trp53 loss of function. Genes Dev 18: 2614–2626.

  15. , , , , , et al. (1999). cDNA microarrays detect activation of a myogenic transcription program by the PAX3-FKHR fusion oncogene. Proc Natl Acad Sci USA 96: 13264–13269.

  16. , , , . (1999). The oncogenic potential of the Pax3-FKHR fusion protein requires the Pax3 homeodomain recognition helix but not the Pax3 paired-box DNA binding domain. Mol Cell Biol 19: 594–601.

  17. , , , , , et al. (2005). Pax3 functions at a nodal point in melanocyte stem cell differentiation. Nature 433: 884–887.

  18. , , , , , . (1992). FGF inactivates myogenic helix–loop–helix proteins through phosphorylation of a conserved protein kinase C site in their DNA-binding domains. Cell 71: 1181–1194.

  19. , , . (1998). Mutation of Thr115 in MyoD positively regulates function in murine fibroblasts and human rhabdomyosarcoma cells. Cell Growth Differ 9: 699–711.

  20. , , , . (2000). Regulation of skeletal myogenesis by association of the MEF2 transcription factor with class II histone deacetylases. Mol Cell 6: 233–244.

  21. , , , . (1996). Skeletal muscle cells lacking the retinoblastoma protein display defects in muscle gene expression and accumulate in S and G2 phases of the cell cycle. J Cell Biol 135: 441–456.

  22. , , , , . (1999). pRb is required for MEF2-dependent gene expression as well as cell-cycle arrest during skeletal muscle differentiation. Curr Biol 9: 449–459.

  23. , . (2000). Regulation of muscle regulatory factors by DNA-binding, interacting proteins, and post-transcriptional modifications. J Cell Physiol 185: 155–173.

  24. , , . (2006). A PANorama of PAX genes in cancer and development. Nat Rev Cancer 6: 52–62.

  25. , , , . (1986). Cellular differentiation and prognosis in embryonal rhabdomyosarcoma. A report from the Cooperative Soft Tissue Sarcoma Study 1981 (CWS 81). Virchows Arch A Pathol Anat Histopathol 409: 183–194.

  26. , , . (1992). A novel myoblast enhancer element mediates MyoD transcription. Mol Cell Biol 12: 4994–5003.

  27. , . (2007). Identification of a new class of PAX3-FKHR target promoters: a role of the Pax3 paired box DNA binding domain. Oncogene 26: 1595–1605.

Download references


We thank David Goldhamer for MyoD reporter constructs; Martyn Goulding, Eric Olsen, Andrew Lassar, Pier-Luigi Lollini, Helen Blau and Frank Rauscher for reagents; Jocelyn Montalvo and Jerry Barnhart for technical assistance; George McNamara and Shelley Nelson for helpful discussions. This work was supported by a Saban Institute Research Career Development Award (MJA).

Author information


  1. Department of Pathology, Children's Hospital Los Angeles, Los Angeles, CA, USA

    • F Graf Finckenstein
    • , V Shahbazian
    • , Y-X Ren
    •  & M J Anderson
  2. Department of Pathology, University of Southern California Keck School of Medicine, Los Angeles, CA, USA

    • E Davicioni
    •  & M J Anderson


  1. Search for F Graf Finckenstein in:

  2. Search for V Shahbazian in:

  3. Search for E Davicioni in:

  4. Search for Y-X Ren in:

  5. Search for M J Anderson in:

Corresponding author

Correspondence to M J Anderson.

Supplementary information

About this article

Publication history







Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc).

Further reading