The PAX3-FKHR fusion protein is present in a majority of alveolar rhabdomyosarcomas associated with increased aggressiveness and poor prognosis. To better understand the molecular pathogenesis of PAX3-FKHR, we carried out the first, unbiased genome-wide identification of PAX3-FKHR binding sites and associated target genes in alveolar rhabdomyosarcoma. The data shows that PAX3-FKHR binds to the same sites as PAX3 at both MYF5 and MYOD enhancers. The genome-wide analysis reveals that the PAX3-FKHR sites are (a) mostly distal to transcription start sites, (b) conserved, (c) enriched for PAX3 motifs, and (d) strongly associated with genes overexpressed in PAX3-FKHR–positive rhabdomyosarcoma cells and tumors. There is little evidence in our data set for PAX3-FKHR binding at the promoter sequences. The genome-wide analysis further illustrates a strong association between PAX3 and E-box motifs in these binding sites, suggestive of a common coregulation for many target genes. We also provide the first direct evidence that FGFR4 and IGF1R are the targets for PAX3-FKHR. The map of PAX3-FKHR binding sites provides a framework for understanding the pathogenic roles of PAX3-FKHR, as well as its molecular targets to allow a systematic evaluation of agents against this aggressive rhabdomyosarcoma. Cancer Res; 70(16); 6497–508. ©2010 AACR.

The paired box transcription factor (PAX) family consists of nine highly related members with critical roles in organogenesis and tissue specification during development, including skeletal muscle, pancreas, central nervous system, kidney, and immune system (1). In various cell model systems, PAX genes have been implicated in the maintenance of the multipotent state, the initiation of differentiation programs, cell migration, proliferation, and survival. Mutations and gene rearrangement of PAX genes are frequently associated with human diseases and cancers (24). In particular, PAX3 and PAX7 are required during early embryogenesis for the development of the neural crest and skeletal muscle (2, 5), and gene rearrangements with FKHR are frequently implicated with rhabdomyosarcoma (6).

Rhabdomyosarcoma is the most common soft tissue sarcoma in children (7), and consists of alveolar (ARMS) and embryonal subtypes. The majority of ARMS are associated with specific chromosome translocations that result in PAX3-FKHR (FOXO1A) or PAX7-FKHR fusion protein, consisting of the NH2-terminal PAX3/PAX7 DNA binding domain and the COOH-terminal FKHR transactivation domain (8, 9). The fusion protein was suggested as a more potent transactivator than PAX3 in vitro (10, 11) and in vivo (12). Conditional activation of PAX3-FKHR was shown to collaborate with the inactivation of Ink4a/Arf or p53 in facilitating the development of murine ARMS (13). The PAX3-FKHR translocation is not only the most common translocation in ARMS but is also associated with increased treatment failure and mortality rate (14). Thus, it is important to identify the direct transcription targets of PAX3-FKHR. Nearly all published studies on PAX3-FKHR targets and their cis-acting sequences focused on promoter elements immediately adjacent to the transcription start sites, even though the role of PAX3-FKHR binding sites as a promoter has not been verified. In fact, for the best-studied targets of PAX3, the myogenic determination genes, MYF5 and MYOD, the cis-acting sequences responsible for PAX3 regulation are distal enhancers (1518). As the identification of the direct effectors of PAX3-FKHR might have pivotal roles in delineating its molecular pathogenic mechanism and in identifying new targets for therapy, a genome-wide analysis of the direct targets of PAX3-FKHR is the next necessary step.

To elucidate the targets of PAX3-FKHR, chromatin immunoprecipitation and DNA sequencing (ChIP-seq) were performed with a novel antibody specific for PAX3-FKHR fusion proteins. Our results show that the majority of binding sites are conserved distal regulatory elements, enriched for PAX3 and E-box/MYF motifs, and are strongly associated with genes upregulated by PAX3-FKHR. The data set of PAX3-FKHR binding sites has a high degree of accuracy, as shown in the examples of MYF5 and MYOD enhancers, with a resolution sufficient for rapid identification and confirmation of specific recognition sequences. The identification of the PAX3-FKHR binding sites and associated genes enables the rapid identification and validation of its targets, which would enhance our understanding of myogenic development, the molecular pathogenesis of rhabdomyosarcoma, and the evaluation of targeted agents for this cancer.

Cell lines and reagents

All human rhabdomyosarcoma cell lines, along with an Rh1 Ewing's sarcoma cell line, were maintained in RPMI 1640 with 10% fetal bovine serum. PAX3-FKHR-HA and PAX3 and PAX3-FKHR expression plasmids were kindly provided by Drs. Michael Anderson and Fred Barr. Small hairpin RNA (shRNA) plasmids targeting FKHR were obtained from Open Biosystems with the target sequences: s2, GCTTAGACTGTGACATGGAAT; s4, CAGGACAATAAGTCGAGTTAT. Lentiviruses for shRNA vectors were generated. For RNAi experiments, cells were infected with the lentiviruses for 3 days with at least 50% of the cells infected.

Quantitative PCR analysis

Total and immunoprecipitated DNA were used for quantitative PCR (qPCR) on glyceraldehyde-3-phosphate dehydrogenase promoter and MYF5 enhancer using SYBR Green PCR kit (Qiagen); the primers used are listed in Supplementary Table S7. For quantitative expression analysis, we used cells (3 days after infection with lentiviruses) carrying scramble or shRNA vectors against the FKHR portion of PAX-FKHR. Total RNA was isolated with Trizol (Invitrogen) and cDNA was generated using SuperScript III (Invitrogen). qPCR analysis was performed in triplicate with SYBR Green PCR kit, with primers listed in Supplementary Table S7 to determine the relative levels of specific transcripts, normalized against that of glyceraldehyde-3-phosphate dehydrogenase. All qPCR assays were reproduced at least three times.

Antibodies and immunoblots

Cell lysate preparations and immunoblots were done as previously described (19). A monoclonal antibody against PAX3-FKHR (PFM2) was generated against a fusion region peptide (Supplementary Methods). Ascites fluid was obtained for PFM2 and the IgG fraction was isolated. Antibodies against PAX3, FKHR, MYCN, MET, ERK, and AKT, as well as actin, were from Cell Signaling Technology; the antibody against IGF1R-β (C20) was from Santa Cruz Biotechnology.

Luciferase reporter assay for PAX3-FKHR enhancer

PAX3-FKHR binding sites were amplified from human genomic DNA using the primers listed in Supplementary Table S7. They were cloned into the SalI and BamHI sites of the pGL3-promoter vector (Promega), which contains an SV40 promoter upstream of the firefly luciferase reporter gene. Mutant enhancers were generated with a site-directed mutagenesis kit (Stratagene) with primers listed in Supplementary Table S7. All mutations were confirmed by DNA sequencing. Enhancer assays were performed with a Dual Luciferase Assay System (Promega) to adjust for differences in transfection efficiency. Reporter cotransfection experiments were done with reporter to activator plasmid ratios of 1:1. All reporter assays were reproduced at least three times.

Statistical analyses

All quantitative experiments, including qPCR for immunoprecipitated DNA, RT-qPCR for specific gene expression, and luciferase analyses were performed in triplicate. Data analyses were performed with Prism (GraphPad) and shown in charts as mean ± SEM. Its effect on gene expression patterns was tested based on the null hypothesis that the number of genes associated with binding sites within the set of differentially expressed genes was not larger than expected by chance when compared with the set of all annotated genes. P values were calculated using the hypergeometric distribution.

Genome-wide mapping of PAX3-FKHR binding sites in rhabdomyosarcoma

To identify the targets of PAX3-FKHR, a novel monoclonal antibody PFM2 was generated using a peptide corresponding to the fusion region. The antibody is specific to PAX3-FKHR (Fig. 1A). The precipitation with PFM2 in PAX3-FKHR+Rh4 cells resulted in a 20- to 40-fold enrichment of the MYF5 enhancer over the input DNA, the best known target of PAX3 (Fig. 1B). In contrast, parallel ChIP with fusion-negative RD cells gave no enhancer enrichment (Fig. 1B). Both the ChIP and input DNA were sequenced to obtain 3.96 and 4.06 million aligned sequence tags, respectively, with similar chromosome distributions (Supplementary Fig. S1, Supplementary Table S1). Although the input DNA tags were evenly distributed, distinct sequence tag clusters were identified with PFM2 ChIP DNA, as shown in chromosome 11 (Fig. 1C). Altogether, a total of 1,463 statistically significant putative PAX3-FKHR binding sites were identified in the human genome with an average size of 186 bp, at approximately one site per 2 million base pairs (GEO no. GSE19063, see Supplementary Methods for a detailed description of selection criteria and data access). The chromosome distribution of the binding sites is closely correlated with the length of the chromosome (Supplementary Fig. S1, r2= 0.83), similar to that of estrogen receptor and p53 (Supplementary Fig. S1, Supplementary Table S1; refs. 20, 21). The ChIP-seq result was highly reproducible with 80% (1,166 sites) identified in a repeat experiment (data not shown). The specificity of the binding sites was very good, as there was no overlap between the 1,463 sites in Rh4 and the 223 sites obtained from fusion-negative RD cells via ChIP-seq, which also gave 4 million aligned sequences (data not shown).

Figure 1.

ChIP-seq analysis of PAX3-FKHR binding sites in rhabdomyosarcoma cells. A, immunoblots show that PFM2 recognizes PAX3-FKHR proteins in rhabdomyosarcoma cell lines positive for the translocation (RMS13 and Rh30). B, qPCR analysis shows that PFM2 selectively immunoprecipitates the MYF5 enhancer from PAX3-FKHR+ Rh4 cells, not from the negative RD cells. C, distribution of PAX3-FKHR binding sites in chromosome 11. Peak height represents the number of sequence tags. D, a PAX3-FKHR site mapped within the MYF5 enhancer. The maximum signal aligns with a known PAX3 recognition site at this PAX3-dependent enhancer.

Figure 1.

ChIP-seq analysis of PAX3-FKHR binding sites in rhabdomyosarcoma cells. A, immunoblots show that PFM2 recognizes PAX3-FKHR proteins in rhabdomyosarcoma cell lines positive for the translocation (RMS13 and Rh30). B, qPCR analysis shows that PFM2 selectively immunoprecipitates the MYF5 enhancer from PAX3-FKHR+ Rh4 cells, not from the negative RD cells. C, distribution of PAX3-FKHR binding sites in chromosome 11. Peak height represents the number of sequence tags. D, a PAX3-FKHR site mapped within the MYF5 enhancer. The maximum signal aligns with a known PAX3 recognition site at this PAX3-dependent enhancer.

Close modal

The specificity of the binding sites was verified with the best characterized PAX3-dependent enhancer located at −58/−56 kb distal to MYF5 (Fig. 1D). The binding site identified by ChIP-seq is defined in a 108 bp region, within the previously identified 145 bp minimal PAX3-responsive MYF5 enhancer (17), demonstrating that PAX3-FKHR binds to identical sites in rhabdomyosarcoma cells as PAX3 does during development. The maximum of the tag density was located at the boxed PAX3 binding sequence AAGCATGACT, at the enhancer previously shown to be critical for PAX3-dependent activation (17, 22). This result shows that the data set might have sufficient resolution for identifying specific recognition sequences.

PAX3-FKHR binding at conserved distal regulatory elements

Analysis of the binding sites showed that the vast majority of the sites were either more than 4 kb from the transcription start sites outside of the genes (60.2%), or in the introns (32.5%). Only 0.4% of the binding sites were located within 1 kb upstream of the initiation sites (Fig. 2A). In addition, there was no binding site enrichment within either a 1 or 4 kb window upstream of the start sites (Fig. 2A). The data suggests that PAX3-FKHR binds primarily to distal sites. The enrichment analysis further revealed reduced binding sites in the exons (Fig. 2A), consistent with distinctive functions for the two. Maximal conservation analysis shows that there is a significant conservation among these PAX3-FKHR binding sites in vertebrates, in contrast to that of randomly selected sequences. The degree of conservation is similar to that of RNA Pol-II binding sites (Fig. 2B), again indicating the functionality of these PAX3-FKHR sites. The analysis of the 1,463 binding sites identifies 1,072 genes either overlapping or located proximal to them (Supplementary Table S2). Gene ontology analysis shows a bias towards developmental processes, including those of the nervous, skeletal, and muscular systems (Fig. 2C), consistent with the known roles of PAX3 in early development and its expression in those systems. Thus, the binding sites are conserved and identify potential targets of PAX3 during development.

Figure 2.

PAX3-FKHR binding sites as putative enhancers. A, distribution of PAX3-FKHR sites relating to the transcription start sites, exons, and introns. Enrichment or suppression PAX3-FKHR sites. Unit of 1 represents the average number of sites across the genome. B, degree of conservation of PAX3-FKHR sites compared with that of Pol-II ChIP-seq sites and randomly generated sequence tags. C, gene ontology analysis of genes associated with PAX3-FKHR sites with Ingenuity IPA. D, the association of PAX3-FKHR sites with genes upregulated by PAX3-FKHR in RD cells (Supplementary Table S3), and genes upregulated in a panel of PAX3/7-FKHR–positive rhabdomyosarcoma tumors (Supplementary Table S4).

Figure 2.

PAX3-FKHR binding sites as putative enhancers. A, distribution of PAX3-FKHR sites relating to the transcription start sites, exons, and introns. Enrichment or suppression PAX3-FKHR sites. Unit of 1 represents the average number of sites across the genome. B, degree of conservation of PAX3-FKHR sites compared with that of Pol-II ChIP-seq sites and randomly generated sequence tags. C, gene ontology analysis of genes associated with PAX3-FKHR sites with Ingenuity IPA. D, the association of PAX3-FKHR sites with genes upregulated by PAX3-FKHR in RD cells (Supplementary Table S3), and genes upregulated in a panel of PAX3/7-FKHR–positive rhabdomyosarcoma tumors (Supplementary Table S4).

Close modal

Inspection of the PAX3-FKHR sites provides a direct link between PAX3/PAX3-FKHR and the genes implicated in skeletal development, muscular development, and rhabdomyosarcoma (Table 1). These genes are involved in growth, survival and oncogenesis, cell migration and tumor metastasis, as well as being important for myogenic differentiation. All their binding sites were re-confirmed in the repeat experiment. Despite their roles in muscle development and cancer, many were not previously known to be direct targets of either PAX3 or PAX3-FKHR, and in most cases, the PAX3-FKHR–dependent regulatory elements were not defined.

Table 1.

Fold of induction was normalized against that of the vector reporter without an activator

(A) Selected genes with PAX3-FKHR binding sites
IDEntrez gene nameLocationTypeDrugs
Self-renewal, growth, survival, oncogenesis 
ALK Anaplastic lymphoma receptor tyrosine kinase Plasma membrane Receptor kinase PF-02341066 
FGFR2 Fibroblast growth factor receptor 2 Plasma membrane Receptor kinase TKI-258, AZD2171, brivanib, PHA-739358 
FGFR4 Fibroblast growth factor receptor 4 Plasma membrane Receptor kinase  
IGF1R Insulin-like growth factor 1 receptor Plasma membrane Receptor kinase CP-751871, AMG479, R1507, MK-0646, AVE1642 
KDR Kinase insert domain receptor, VEGFR2 Plasma membrane Receptor kinase Bevacizumab, sunitinib, vatalanib, sorafenib, vandetanib 
MEIS1 Meis homeobox 1 Nucleus Transcription factor  
MYCN v-myc myelocytomatosis viral related oncogene, neuroblastoma-derived Nucleus Transcription factor  
SPRY1 Sprouty homologue 1, antagonist of FGF signaling (Drosophila) Plasma membrane FGFR4 ligand  
     
Migration, metastasis 
CXCR4 Chemokine (C-X-C motif) receptor 4 Plasma membrane GPCR JM 3100 
MET met proto-oncogene (hepatocyte growth factor receptor) Plasma membrane Receptor kinase PF-02341066 
     
Myogenic differentiation 
EYA2 Eyes absent homologue 2 (Drosophila) Nucleus Phosphatase  
EYA4 Eyes absent homologue 4 (Drosophila) Cytoplasm Phosphatase  
MEOX1 Mesenchyme homeobox 1 Nucleus Transcription factor  
MEOX2 Mesenchyme homeobox 2 Nucleus Transcription factor  
MYF5/6 Myogenic factor 6 (herculin) Nucleus Transcription factor  
MYOD1 Myogenic differentiation 1 Nucleus Transcription factor  
PRRX1 Paired related homeobox 1 Nucleus Transcription factor  
     
Other potential targets 
ABAT 4-Aminobutyrate aminotransferase Cytoplasm Enzyme Valproic acid 
ADRA1D Adrenergic, α-1D-, receptor Plasma membrane GPCR Paliperidone, risperidone, antazoline, acetaminophen, epinephrine, acetaminophen 
ADRA2A Adrenergic, α-2A-, receptor Plasma membrane   
ADRA2C Adrenergic, α-2C-, receptor Plasma membrane   
DHFR Dihydrofolate reductase Unknown Enzyme Pyrimethamine, trimethoprim, iclaprim, methotrexate 
(B) Verification of PAX3-FKHR binding site via enhancer assay 
Gene Enhancer no. Binding intensity Fold induction by activators Rel. Location Dist. Tx Start bp Dist. Tx End bp 
None PAX3 PAX3-FKHR 
Vector None NA 1.0 1.0 0.9 NA NA NA 
ALK High 2.6 10.8 19.5 3rd Intron −262,835 464,601 
FGFR4 Low 1.5 1.9 3.7 3′ downstream −8,913 −839 
FGFR4 High 1.9 5.7 8.3 3′ downstream −16,273 −8,199 
IGF1R Med 1.8 4.0 4.8 2nd Intron −136,375 171,486 
MET Low 3.9 10.2 7.4 3rd Intron −35,329 61,711 
MET High 1.3 2.5 3.8 18th Intron −83,277 13,763 
MYCN Med 0.9 3.8 5.4 3′ downstream −96,819 −92,786 
MEOX1 Med 1.2 2.4 5.4 5′ upstream 994 20,274 
MEOX2 High 0.8 0.8 3.8 1 intron −798 73,218 
PRRX1 Med 0.6 2.1 18.2 1st intron −4,226 67,742 
(A) Selected genes with PAX3-FKHR binding sites
IDEntrez gene nameLocationTypeDrugs
Self-renewal, growth, survival, oncogenesis 
ALK Anaplastic lymphoma receptor tyrosine kinase Plasma membrane Receptor kinase PF-02341066 
FGFR2 Fibroblast growth factor receptor 2 Plasma membrane Receptor kinase TKI-258, AZD2171, brivanib, PHA-739358 
FGFR4 Fibroblast growth factor receptor 4 Plasma membrane Receptor kinase  
IGF1R Insulin-like growth factor 1 receptor Plasma membrane Receptor kinase CP-751871, AMG479, R1507, MK-0646, AVE1642 
KDR Kinase insert domain receptor, VEGFR2 Plasma membrane Receptor kinase Bevacizumab, sunitinib, vatalanib, sorafenib, vandetanib 
MEIS1 Meis homeobox 1 Nucleus Transcription factor  
MYCN v-myc myelocytomatosis viral related oncogene, neuroblastoma-derived Nucleus Transcription factor  
SPRY1 Sprouty homologue 1, antagonist of FGF signaling (Drosophila) Plasma membrane FGFR4 ligand  
     
Migration, metastasis 
CXCR4 Chemokine (C-X-C motif) receptor 4 Plasma membrane GPCR JM 3100 
MET met proto-oncogene (hepatocyte growth factor receptor) Plasma membrane Receptor kinase PF-02341066 
     
Myogenic differentiation 
EYA2 Eyes absent homologue 2 (Drosophila) Nucleus Phosphatase  
EYA4 Eyes absent homologue 4 (Drosophila) Cytoplasm Phosphatase  
MEOX1 Mesenchyme homeobox 1 Nucleus Transcription factor  
MEOX2 Mesenchyme homeobox 2 Nucleus Transcription factor  
MYF5/6 Myogenic factor 6 (herculin) Nucleus Transcription factor  
MYOD1 Myogenic differentiation 1 Nucleus Transcription factor  
PRRX1 Paired related homeobox 1 Nucleus Transcription factor  
     
Other potential targets 
ABAT 4-Aminobutyrate aminotransferase Cytoplasm Enzyme Valproic acid 
ADRA1D Adrenergic, α-1D-, receptor Plasma membrane GPCR Paliperidone, risperidone, antazoline, acetaminophen, epinephrine, acetaminophen 
ADRA2A Adrenergic, α-2A-, receptor Plasma membrane   
ADRA2C Adrenergic, α-2C-, receptor Plasma membrane   
DHFR Dihydrofolate reductase Unknown Enzyme Pyrimethamine, trimethoprim, iclaprim, methotrexate 
(B) Verification of PAX3-FKHR binding site via enhancer assay 
Gene Enhancer no. Binding intensity Fold induction by activators Rel. Location Dist. Tx Start bp Dist. Tx End bp 
None PAX3 PAX3-FKHR 
Vector None NA 1.0 1.0 0.9 NA NA NA 
ALK High 2.6 10.8 19.5 3rd Intron −262,835 464,601 
FGFR4 Low 1.5 1.9 3.7 3′ downstream −8,913 −839 
FGFR4 High 1.9 5.7 8.3 3′ downstream −16,273 −8,199 
IGF1R Med 1.8 4.0 4.8 2nd Intron −136,375 171,486 
MET Low 3.9 10.2 7.4 3rd Intron −35,329 61,711 
MET High 1.3 2.5 3.8 18th Intron −83,277 13,763 
MYCN Med 0.9 3.8 5.4 3′ downstream −96,819 −92,786 
MEOX1 Med 1.2 2.4 5.4 5′ upstream 994 20,274 
MEOX2 High 0.8 0.8 3.8 1 intron −798 73,218 
PRRX1 Med 0.6 2.1 18.2 1st intron −4,226 67,742 

PAX3-FKHR sites associated with genes upregulated by PAX3-FKHR in rhabdomyosarcoma

To establish the association between the binding sites and genes regulated by PAX3-FKHR, correlative analysis was performed with the genes upregulated and downregulated by PAX3-FKHR when introduced to the RD cells via retroviral transduction (23). There is a 7-fold enrichment of PAX3-FKHR binding sites in the genes upregulated by PAX3-FKHR at 37.7%; whereas no enrichment of PAX3-FKHR sites was seen for genes downregulated by PAX3-FKHR at 5% (Fig. 2D, Supplementary Table S3). Hypergeometric distribution analysis shows that the association between genes with PAX3-FKHR sites and those upregulated by PAX3-FKHR is statistically significant (P < 0.0001). In addition to the association with syngeneic rhabdomyosarcoma cells with exogenously introduced PAX3-FKHR, there is a strong correlation between the binding sites and the upregulated genes in a panel of 160 rhabdomyosarcoma tumors (24). Of the 363 genes with increased expression in PAX3/7-FKHR–positive rhabdomyosarcoma, 24.2% have the binding sites (Fig. 2D, Supplementary Table S4). In contrast, of the 122 genes with decreased expression, only 5.7% have the binding sites. Again, hypergeometric distribution analysis shows that such an association is significant (P < 0.0001). Thus, PAX3-FKHR sites are strongly associated with genes upregulated by PAX3-FKHR in vitro in rhabdomyosarcoma cells and in rhabdomyosarcoma tumors with PAX3/7-FKHR translocations.

PAX3-FKHR binds MYOD core enhancer at a newly defined PAX3 recognition sequence and regulates other limb development genes

MYOD is the master controller of skeletal muscle development (25). PAX3 regulation of MYOD is mediated via the MYOD core enhancer (15, 26). Our genome-wide analysis of PAX3-FKHR sites shows a binding site in a region of 161 bp, located within the previously mapped 258-bp MYOD core enhancer (Fig. 3A). Interestingly, sequence examination at the binding signal peak identified a novel PAX3 recognition sequence, AACCCGTGAC (box). To confirm the role of this PAX3 recognition sequence, an enhancer reporter was constructed with this MYOD enhancer and a mutant was then derived by site-directed mutagenesis to inactivate the PAX3 recognition sequence. Enhancer analysis showed that this particular PAX3 recognition sequence was responsible for PAX3- and PAX3-FKHR–dependent activation of the core enhancer (Fig. 3A). Thus, our data enabled the identification of a novel recognition sequence necessary for PAX3- and PAX3-HKFR–mediated activation of MYOD core enhancer. In addition to MYF5 and MYOD, we identified novel PAX3-FKHR binding sites in several additional transcription factors important for muscle and limb development, and potentially for ARMS, including homeobox genes MEOX1, MEOX2, and PRRX1 (Table 1). Selected sites were validated by enhancer assays, site-directed mutagenesis, and shRNA depletion of PAX3-FKHR (Supplementary Figs. S2 and S3).

Figure 3.

PAX3 recognition sequence in the MYOD core enhancer and the coenrichment of PAX3 and E-box motifs in the binding sites. A, the identification of a novel PAX3 recognition sequence at the MYOD core enhancer and site-directed mutagenesis study of the site in transfected A204 rhabdomyosarcoma cells. B, enrichment of transcription factor motifs in PAX3-FKHR sites and clustering of the top enriched motifs. C, results of a de novo motif search in which 9 of the top 10 motifs were shown to be related to either E-box or PAX3 motifs (see Supplementary Table S6 for details). D, coenrichment of PAX3 and E-box motifs among PAX3-FKHR sites. Either PAX3B or LBP1 was anchored, and the relative enrichment of other sites within 100 bp of the anchor.

Figure 3.

PAX3 recognition sequence in the MYOD core enhancer and the coenrichment of PAX3 and E-box motifs in the binding sites. A, the identification of a novel PAX3 recognition sequence at the MYOD core enhancer and site-directed mutagenesis study of the site in transfected A204 rhabdomyosarcoma cells. B, enrichment of transcription factor motifs in PAX3-FKHR sites and clustering of the top enriched motifs. C, results of a de novo motif search in which 9 of the top 10 motifs were shown to be related to either E-box or PAX3 motifs (see Supplementary Table S6 for details). D, coenrichment of PAX3 and E-box motifs among PAX3-FKHR sites. Either PAX3B or LBP1 was anchored, and the relative enrichment of other sites within 100 bp of the anchor.

Close modal

Genome-wide coenrichment of PAX3 and E-box motifs within the PAX3-FKHR site

There is limited knowledge on how PAX3 functions in regulating gene expression. It is a weak transactivator and might require other elements in regulating gene expression. Genetic hierarchy analysis of skeletal muscle differentiation reveals that MYOD expression requires both PAX3 and MYF5 (27). Analysis of the MYOD core enhancer shows several E-box motifs that are required for its expression in the skeletal muscle lineage (26). We asked if the presence of both PAX3 and E-box motifs was a more common property of PAX3-FKHR binding sites. Transcription factor motif analysis of PAX3-FKHR sites showed that PAX3_B was the most enriched motif at 21.7-fold, together with several related paired box motifs (Supplementary Table S5; Fig. 3B). Interestingly, several other motifs were also enriched: AP4 by 12.9-fold and LBP1 by 11.1-fold, sharing a common CAGCTG E-box motif. Thus, motif analysis of PAX3-FKHR sites reveals a genome-wide enrichment of PAX3 and E-box motifs. To confirm the results, de novo motif analysis of the ChIP-Seq data was performed with Discriminating Motif Enumerator (DME) (28). Of the top 10 output motifs, at least 5 were PAX3 or highly related motifs, and 4 others were E-box–related motifs using the TRANSFAC database (Fig. 3C; Supplementary Table S6). Thus, motif and de novo analysis of PAX3-FKHR sites revealed a genome-wide selective enrichment of PAX3 and E-box motifs.

We asked whether the enrichment was due to the colocalization of these two motifs adjacent to each other. When PAX3_B was used as the anchor, both LBP1 and AP4 motifs (E-box) were enriched by 7- to 8-fold within 100 bp of the anchor (Fig. 3D). Neither the PAX3_B site nor controls (CREBP1 or PPAR-α) were enriched adjacent to the PAX3_B anchor. Similarly, when LBP1 was used as the anchor, only the PAX3_B motif was enriched by 14-fold. Thus, the genome-wide analysis of PAX3 and E-box motifs shows a strong colocalization at PAX3-FKHR binding sites. These observations suggest a broad role for PAX3 and MYF/E-box proteins in regulating the expression of many targets in addition to MYOD.

FGFR4 is a direct target of PAX3-FKHR

The fibroblast growth factor (FGF) signaling pathway has been implicated in vivo in promoting cell proliferation and differentiation during myogenesis (29). FGFR4 is expressed during myogenic differentiation under the control of PAX3 via a downstream enhancer (30). The importance of this pathway has recently been emphasized by the identification of activating mutations of FGFR4 in 7% of rhabdomyosarcomas (31), which promotes metastasis in the xenograft mouse. The expression level of FGFR4 was associated with ARMS and poor survival (31). Our results revealed two binding sites downstream of FGFR4 (Fig. 4A). Both putative FGFR4 enhancer sequences were cloned into the enhancer reporter and used for enhancer analysis. The results showed that the enhancer activities were detected in PAX3-FKHR+Rh4 and Rh30 cells, but not in the negative Rh18 cells (Fig. 4B). The distal FGFR4.E2 gave significantly greater activity than that of FGFR4.E1. In cotransfection experiments into RD cells, PAX3-FKHR was able to induce both E1 and E2 activity (Fig. 4B), of which E2 is a stronger enhancer.

Figure 4.

Fine-mapping of a PAX3-FKHR–dependent enhancer for FGFR4 gene. A, ChIP-seq results identified two PAX3-FKHR binding sites downstream of FGFR4. The peak of the FGFR4.E2 binding signal is associated with a conserved PAX3 sequence. B, enhancer activity for both binding sites in PAX3-FKHR (P3F)–positive and –negative cells, and with cotransfection PAX3 or PAX3-FKHR expression plasmid in RD cells. C, site-directed mutagenesis of PAX3 recognition sequence FGFR4.E2 enhancer resulted in a loss of transactivation by PAX3 or PAX3-FKHR in RD cells. D, RT-qPCR data showing downregulation of PAX3-FKHR with a lentiviral shRNA (s2) vector led to reduced ALK and FGFR4 mRNA in Rh4.

Figure 4.

Fine-mapping of a PAX3-FKHR–dependent enhancer for FGFR4 gene. A, ChIP-seq results identified two PAX3-FKHR binding sites downstream of FGFR4. The peak of the FGFR4.E2 binding signal is associated with a conserved PAX3 sequence. B, enhancer activity for both binding sites in PAX3-FKHR (P3F)–positive and –negative cells, and with cotransfection PAX3 or PAX3-FKHR expression plasmid in RD cells. C, site-directed mutagenesis of PAX3 recognition sequence FGFR4.E2 enhancer resulted in a loss of transactivation by PAX3 or PAX3-FKHR in RD cells. D, RT-qPCR data showing downregulation of PAX3-FKHR with a lentiviral shRNA (s2) vector led to reduced ALK and FGFR4 mRNA in Rh4.

Close modal

The examination of FGFR4.E2 reveals PAX3 and E-box recognition sequences that are both conserved in vertebrates (Fig. 4A). A mutant in the putative PAX3 recognition site resulted in the complete inactivation of PAX3-FKHR–dependent induction (Fig. 4C). The data show that the PAX3-FKHR binding site downstream of FGFR4 is a PAX3-FKHR–dependent enhancer.

To determine if the expression of FGFR4 in rhabdomyosarcoma cells is dependent on PAX3-FKHR, Rh4 cells were infected with a lentivirus containing a shRNA (s2) targeting PAX3-FKHR. RNA was isolated and analyzed for PAX3-FKHR and FGFR4 mRNA levels. The results showed that both PAX3-FKHR and FGFR4 mRNA levels were reduced with this shRNA (P3F.s2, Fig. 4D). Thus, PAX3-FKHR is responsible for the elevated FGFR4 mRNA levels in a fusion-positive rhabdomyosarcoma cell line.

It is worth noting that our study might provide more accurate data on PAX3-FKHR–dependent regulation for MET. In contrast with a previous report showing that the MET promoter was regulated by PAX3 (32), our data identified two PAX3-FKHR binding sites in the introns of MET. Enhancer reporter assays showed that both were activated by PAX3 and PAX3-FKHR (Table 1; Supplementary Fig. S4). As these sequences are selectively bound by PAX3-FKHR in rhabdomyosarcoma cells, the results indicate that MET is regulated by PAX3 or PAX3-FKHR via intronic enhancers.

Identification of PAX3-FKHR–dependent enhancers at ALK and MYCN

Of the genes containing proximal or internal binding sites for PAX3-FKHR, Ingenuity Pathway analysis (Ingenuity Systems) revealed the strongest disease association with developmental, skeletal, and muscular disorders, among which MYCN was implicated (Supplementary Fig. S5). In addition to amplification in neuroblastomas, MYCN was shown to be amplified in ∼20% of both rhabdomyosarcomas associated with adverse outcome (33). However, its expression was significantly higher in ARMS than in embryonal rhabdomyosarcoma. Moreover, an increase in MYCN expression was associated with worse prognosis only in ARMS, not embryonal rhabdomyosarcoma. Our ChIP-seq data showed a binding site located at 96.8 kb downstream of the transcription start site that could serve as a PAX3-FKHR–dependent enhancer (Table 1). Furthermore, protein analysis in a panel of rhabdomyosarcoma cell lines showed the highest levels of MYCN only in PAX3-FKHR fusion–positive rhabdomyosarcoma cells (Fig. 5C), and it is downregulated with a shRNA targeting PAX3-FKHR (Fig. 5D), suggesting a role of PAX3-FKHR in the expression of MYCN in rhabdomyosarcoma.

Figure 5.

IGF1R as a target for PAX3-FKHR transactivation. A, identification of a PAX3-FKHR binding site in the 2nd intron of IGF1R. ChIP-seq results of input and PFM2 immunoprecipitated DNA from two experiments. B, PAX3 and PAX3-FKHR–dependent activation of the IGF1R enhancer in cotransfected RD cells. C, immunoblots of IGF1R, MYCN, MET, and PAX3-FKHR in a panel of rhabdomyosarcoma cell lines. D, shRNA lentiviral vectors against PAX3-FKHR results in the downregulation of IGF1R, MYCN, and MET in Rh4.

Figure 5.

IGF1R as a target for PAX3-FKHR transactivation. A, identification of a PAX3-FKHR binding site in the 2nd intron of IGF1R. ChIP-seq results of input and PFM2 immunoprecipitated DNA from two experiments. B, PAX3 and PAX3-FKHR–dependent activation of the IGF1R enhancer in cotransfected RD cells. C, immunoblots of IGF1R, MYCN, MET, and PAX3-FKHR in a panel of rhabdomyosarcoma cell lines. D, shRNA lentiviral vectors against PAX3-FKHR results in the downregulation of IGF1R, MYCN, and MET in Rh4.

Close modal

Although originally identified in translocations in lymphoma, activating mutations in ALK were also recently associated with neuroblastoma (3437). Our ChIP-seq data again showed a strong PAX3-FKHR site in the third intron of ALK, which was a very potent PAX3-FKHR–dependent enhancer (Table 1). Infecting Rh4 cells with a lentivirus containing a shRNA (s2) for PAX3-FKHR led to a significant reduction of ALK mRNA (Fig. 4D). Thus, ALK expression could also be regulated by PAX3-FKHR.

PAX3-FKHR directly upregulates IGF1R, which might serve as a biomarker for anti-IGF1R therapy

PAX3 is known to be important for the proliferation and survival of muscle progenitor cells (38, 39). Although it is not known how PAX3 accomplishes these functions, it is possible that PAX3-FKHR targets genes important for growth and survival, and for the development of rhabdomyosarcoma. Ingenuity Pathway analysis of PAX3-FKHR targets also revealed IGF1R in the developmental disorder network (Supplementary Fig. S5). The IGF1R has a PAX3-FKHR binding site in the second intron (Fig. 5A). When cloned into the enhancer reporter, it exhibited specific enhancer activity that was dependent on either PAX3 or PAX3-FKHR (Fig. 5B). Analysis of IGF1R protein levels in rhabdomyosarcoma cell lines showed an apparent correlation with PAX3-FKHR fusion proteins, in which all four cell lines with the fusion protein had high levels of IGF1R (Fig. 5C). To show the requirement of PAX3-FKHR for elevated IGF1R expression, Rh4 cells were infected with two shRNA lentiviral vectors targeting PAX3-FKHR. The results showed that in addition to downregulating PAX3-FKHR, these RNAi vectors also reduced IGF1R protein levels, similar to that of MET (Fig. 5D), whereas the levels of ERK and AKT were constant. The results indicate that PAX3-FKHR could directly upregulate IGF1R.

In this study, we present a high-resolution whole genome map of PAX3-FKHR binding sites in ARMS. To our knowledge, this is the first such map for a PAX protein. Understanding how PAX3-FKHR acts to drive ARMS tumorigenesis depends on developing a comprehensive catalog of the genes which are directly regulated by this aberrant oncogenic transcription factor. Previous expression array analysis of PAX3-FKHR–regulated genes provided information on the genes that were differentially expressed (4042). However, that approach has significant limitations due to its inability to identify direct targets and to link the genes with specific cis-acting sequences for follow-up investigations. Promoter analysis, although used frequently for the study of putative PAX3-FKHR–regulated genes, is entirely constrained to sequences immediately adjacent to the transcription start sites. Development studies of PAX3 targets reveal that PAX3 regulates myogenic initiating genes via distal enhancers (1518, 30). Although our PAX3-FKHR binding site results show concise binding on the PAX3-dependent enhancers of MYF5 and MYOD, which were identified with transgenic animal models, our data failed to provide support for nearly all previous reports on the characterization of PAX3-FKHR–mediated gene regulation as they mostly focused on limited promoter sequences (4345). In one instance, we showed two intronic enhancers for MET which were not at the previously described MET promoter. PAX3 motifs are rather degenerate and our data shows that only ∼1 in 1,000 of the predicted PAX3 motifs in the genome were actually bound by PAX3-FKHR (Supplementary Table S5). We believe that the traditional plasmid-based report assay is necessary, but not sufficient, to establish the roles of the regulatory sequences, whereas the binding of PAX3-FKHR to these sequences in vivo is essential. Overall, in our genome-wide analysis, only 0.4% of the binding sites are within 1 kb upstream of the transcription start site. The binding sites are conserved distal elements, enriched for PAX3 motifs, present in genes implicated in the developmental processes, and associated with genes induced by PAX3-FKHR in rhabdomyosarcoma cells and tumors. Our whole genome map of PAX3-FKHR binding sites nominates numerous candidates for a role in the transforming activity of ARMS, and strongly suggests that the action of this fusion oncogene is achieved through a network of pathways affecting myoblast differentiation, migration, proliferation, and survival (Table 1). Thus, a map of PAX3-FKHR binding sites should provide the framework for future studies of its downstream targets and pathogenesis mechanisms.

Our study also provides a genomic overview in understanding PAX3 or PAX3-FKHR–mediated gene regulation. It was previously established that MYF5 (E-box protein) and PAX3 were genetically required for proper MYOD expression (27). Our genome-wide analysis of PAX3-FKHR sites shows the colocalization of PAX3 and E-box motifs at these sites, suggesting the potential coregulation of many target genes by PAX3 and MYF-like E-box proteins. The roles of such coregulation in development and cancer are questions for future investigations.

The mechanism of PAX3-FKHR action in ARMS could also be viewed as a form of aberrant muscle development. Sustained expression of other developmental transcription factors as a result of PAX3-FKHR is likely to be an important aspect of ARMS pathogenesis. In addition to the myogenic E-box factors, we also identified a number of homeobox genes as PAX3-FKHR targets. Our data showed that PAX3-FKHR binds to the enhancers of MEOX1, MEOX2, and PRRX1, and is responsible for their expression. Mesenchyme homeobox genes 1 and 2 (MEOX1 and MEOX2) are expressed in somites that give rise to connective tissue, cartilage, muscle, and the axial skeleton. Both genes have been implicated in axial skeleton and limb muscle development (46). PAX3 and MEOX2 were coexpressed in most migrating myoblasts and MEOX2 was absent from limbs in homozygous Pax3Sp mice (47). Similarly, PRRX1 is implicated in limb development. Mice homozygous for mutant Prrx1 died soon after birth and exhibited defects in skeletogenesis (48). PRRX1 was expressed at a significantly reduced level in homozygous Pax3Sp mice (49). Thus, our work identified new developmental transcription factors driven by PAX3-FKHR in ARMS (and likely by PAX3 in normal development). Determining which of these transcriptional regulators are important in establishing the ARMS phenotype and how they act are important questions for future research.

Identification of the genes and pathways directly regulated by PAX3-FKHR provides an opportunity to better understand the molecular pathogenesis of ARMS, to mine new therapeutic targets, and to provide biomarkers for the development of novel agents. In this connection, our study identifies ALK, FGFR4, IGF1R, and MYCN as direct targets for PAX3-FKHR, which may have important roles in the pathogenesis of rhabdomyosarcomas. Previous studies showed that MYCN amplification and overexpression were associated with adverse prognosis in ARMS (33) and was downstream of PAX3-FKHR (50). Our data identifies a PAX3-FKHR–dependent regulatory sequence in MYCN. In addition, our data indicates ALK as a target of PAX3-FKHR in rhabdomyosarcoma, opening the possibility of follow-up investigations on the roles of ALK in rhabdomyosarcoma. Finally, the comprehensive map of PAX3-FKHR targets provided by this study will facilitate the search for therapeutic approaches for rhabdomyosarcoma. Although it is currently not possible to target PAX3-FKHR itself, selectively targeting one or more critical proteins regulated by PAX3-FKHR such as IGF1R or FGFR4 might improve the outcome of patients with this aggressive cancer. Our results provide the first evidence of the regulation of FGFR4 by PAX3-FKHR in ARMS, which was recently suggested to contribute to in vivo tumor growth and lung metastasis (31). Our data further shows that PAX3-FKHR regulates IGF1R expression in rhabdomyosarcoma. We previously showed that its elevated expression was directly correlated with rhabdomyosarcoma sensitivities to a therapeutic antibody against IGF1R, and inhibiting IGF1R resulted in a profound reduction of AKT signaling both in vitro and in vivo (19). Thus, the systematic identification of PAX3-FKHR targets provides a blueprint for the evaluation of targeted therapeutics against signaling pathways driven by the essential causative genetic aberration in this tumor.

No potential conflicts of interest were disclosed.

We thank Drs. Fred Barr and Michael Anderson for PAX3 and PAX3-FKHR plasmids, Andrew Smith for assistance with DME analysis, and Zhigang Kang and Sean Davis for technical assistance.

Grant support: The Intramural Research Program of the NIH, National Cancer Institute, and in part with federal funds from the NIH, under contract N01-CO-12400. The contents of this publication does not necessarily reflect the views or policies of the Department of Health and Human Services, nor does mention of trade names, commercial products, or organizations imply endorsement by the U.S. Government.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1
Blake
JA
,
Thomas
M
,
Thompson
JA
,
White
R
,
Ziman
M
. 
Perplexing Pax: from puzzle to paradigm
.
Dev Dyn
2008
;
23710
:
2791
803
.
2
Buckingham
M
,
Relaix
F
. 
The role of Pax genes in the development of tissues and organs: Pax3 and Pax7 regulate muscle progenitor cell functions
.
Annu Rev Cell Dev Biol
2007
;
23
:
645
73
.
3
Robson
EJ
,
He
SJ
,
Eccles
MR
. 
A PANorama of PAX genes in cancer and development
.
Nat Rev Cancer
2006
;
61
:
52
62
.
4
Lang
D
,
Powell
SK
,
Plummer
RS
,
Young
KP
,
Ruggeri
BA
. 
PAX genes: roles in development, pathophysiology, and cancer
.
Biochem Pharmacol
2007
;
731
:
1
14
.
5
Buckingham
M
. 
Skeletal muscle progenitor cells and the role of Pax genes
.
C R Biol
2007
;
3306-7
:
530
3
.
6
Barr
FG
. 
Gene fusions involving PAX and FOX family members in alveolar rhabdomyosarcoma
.
Oncogene
2001
;
2040
:
5736
46
.
7
Kramer
S
,
Meadows
AT
,
Jarrett
P
,
Evans
AE
. 
Incidence of childhood cancer: experience of a decade in a population-based registry
.
J Natl Cancer Inst
1983
;
701
:
49
55
.
8
Galili
N
,
Davis
RJ
,
Fredericks
WJ
, et al
. 
Fusion of a fork head domain gene to PAX3 in the solid tumour alveolar rhabdomyosarcoma
.
Nat Genet
1993
;
53
:
230
5
.
9
Barr
FG
,
Galili
N
,
Holick
J
,
Biegel
JA
,
Rovera
G
,
Emanuel
BS
. 
Rearrangement of the PAX3 paired box gene in the paediatric solid tumour alveolar rhabdomyosarcoma
.
Nat Genet
1993
;
32
:
113
7
.
10
Fredericks
WJ
,
Galili
N
,
Mukhopadhyay
S
, et al
. 
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
1995
;
153
:
1522
35
.
11
Bennicelli
JL
,
Advani
S
,
Schafer
BW
,
Barr
FG
. 
PAX3 and PAX7 exhibit conserved cis-acting transcription repression domains and utilize a common gain of function mechanism in alveolar rhabdomyosarcoma
.
Oncogene
1999
;
1830
:
4348
56
.
12
Relaix
F
,
Polimeni
M
,
Rocancourt
D
,
Ponzetto
C
,
Schafer
BW
,
Buckingham
M
. 
The transcriptional activator PAX3-FKHR rescues the defects of Pax3 mutant mice but induces a myogenic gain-of-function phenotype with ligand-independent activation of Met signaling in vivo
.
Genes Dev
2003
;
1723
:
2950
65
.
13
Keller
C
,
Arenkiel
BR
,
Coffin
CM
,
El-Bardeesy
N
,
DePinho
RA
,
Capecchi
MR
. 
Alveolar rhabdomyosarcomas in conditional Pax3:Fkhr mice: cooperativity of Ink4a/ARF and Trp53 loss of function
.
Genes Dev
2004
;
1821
:
2614
26
.
14
Sorensen
PH
,
Lynch
JC
,
Qualman
SJ
, et al
. 
PAX3-FKHR and PAX7-FKHR gene fusions are prognostic indicators in alveolar rhabdomyosarcoma: a report from the Children's Oncology Group
.
J Clin Oncol
2002
;
2011
:
2672
9
.
15
Goldhamer
DJ
,
Faerman
A
,
Shani
M
,
Emerson
CP
 Jr.
Regulatory elements that control the lineage-specific expression of myoD
.
Science
1992
;
256
:
538
42
.
16
Goldhamer
DJ
,
Brunk
BP
,
Faerman
A
,
King
A
,
Shani
M
,
Emerson
CP
 Jr.
Embryonic activation of the myoD gene is regulated by a highly conserved distal control element
.
Development
1995
;
1213
:
637
49
.
17
Bajard
L
,
Relaix
F
,
Lagha
M
,
Rocancourt
D
,
Daubas
P
,
Buckingham
ME
. 
A novel genetic hierarchy functions during hypaxial myogenesis: Pax3 directly activates Myf5 in muscle progenitor cells in the limb
.
Genes Dev
2006
;
2017
:
2450
64
.
18
Hadchouel
J
,
Tajbakhsh
S
,
Primig
M
, et al
. 
Modular long-range regulation of Myf5 reveals unexpected heterogeneity between skeletal muscles in the mouse embryo
.
Development
2000
;
12720
:
4455
67
.
19
Cao
L
,
Yu
Y
,
Darko
I
, et al
. 
Addiction to elevated insulin-like growth factor I receptor and initial modulation of the AKT pathway define the responsiveness of rhabdomyosarcoma to the targeting antibody
.
Cancer Res
2008
;
6819
:
8039
48
.
20
Carroll
JS
,
Meyer
CA
,
Song
J
, et al
. 
Genome-wide analysis of estrogen receptor binding sites
.
Nat Genet
2006
;
3811
:
1289
97
.
21
Wei
CL
,
Wu
Q
,
Vega
VB
, et al
. 
A global map of p53 transcription-factor binding sites in the human genome
.
Cell
2006
;
1241
:
207
19
.
22
Buchberger
A
,
Freitag
D
,
Arnold
HH
. 
A homeo-paired domain-binding motif directs Myf5 expression in progenitor cells of limb muscle
.
Development
2007
;
1346
:
1171
80
.
23
Davicioni
E
,
Finckenstein
FG
,
Shahbazian
V
,
Buckley
JD
,
Triche
TJ
,
Anderson
MJ
. 
Identification of a PAX-FKHR gene expression signature that defines molecular classes and determines the prognosis of alveolar rhabdomyosarcomas
.
Cancer Res
2006
;
6614
:
6936
46
.
24
Davicioni
E
,
Anderson
MJ
,
Finckenstein
FG
, et al
. 
Molecular classification of rhabdomyosarcoma-genotypic and phenotypic determinants of diagnosis: a report from the Children's Oncology Group
.
Am J Pathol
2009
;
1742
:
550
64
.
25
Tapscott
SJ
,
Weintraub
H
. 
MyoD and the regulation of myogenesis by helix-loop-helix proteins
.
J Clin Invest
1991
;
874
:
1133
8
.
26
Kucharczuk
KL
,
Love
CM
,
Dougherty
NM
,
Goldhamer
DJ
. 
Fine-scale transgenic mapping of the MyoD core enhancer: MyoD is regulated by distinct but overlapping mechanisms in myotomal and non-myotomal muscle lineages
.
Development
1999
;
1269
:
1957
65
.
27
Tajbakhsh
S
,
Rocancourt
D
,
Cossu
G
,
Buckingham
M
. 
Redefining the genetic hierarchies controlling skeletal myogenesis: Pax-3 and Myf-5 act upstream of MyoD
.
Cell
1997
;
891
:
127
38
.
28
Smith
AD
,
Sumazin
P
,
Zhang
MQ
. 
Identifying tissue-selective transcription factor binding sites in vertebrate promoters
.
Proc Natl Acad Sci U S A
2005
;
1025
:
1560
5
.
29
Buckingham
M
. 
Myogenic progenitor cells and skeletal myogenesis in vertebrates
.
Curr Opin Genet Dev
2006
;
165
:
525
32
.
30
Lagha
M
,
Kormish
JD
,
Rocancourt
D
, et al
. 
Pax3 regulation of FGF signaling affects the progression of embryonic progenitor cells into the myogenic program
.
Genes Dev
2008
;
2213
:
1828
37
.
31
Taylor
JG
,
Cheuk
AT
,
Tsang
PS
, et al
. 
Identification of FGFR4-activating mutations in human rhabdomyosarcomas that promote metastasis in xenotransplanted models
.
J Clin Invest
2009
;
11911
:
3395
407
.
32
Epstein
JA
,
Shapiro
DN
,
Cheng
J
,
Lam
PY
,
Maas
RL
. 
Pax3 modulates expression of the c-Met receptor during limb muscle development
.
Proc Natl Acad Sci U S A
1996
;
939
:
4213
8
.
33
Williamson
D
,
Lu
YJ
,
Gordon
T
, et al
. 
Relationship between MYCN copy number and expression in rhabdomyosarcomas and correlation with adverse prognosis in the alveolar subtype
.
J Clin Oncol
2005
;
234
:
880
8
.
34
Chen
Y
,
Takita
J
,
Choi
YL
, et al
. 
Oncogenic mutations of ALK kinase in neuroblastoma
.
Nature
2008
;
455
:
971
4
.
35
George
RE
,
Sanda
T
,
Hanna
M
, et al
. 
Activating mutations in ALK provide a therapeutic target in neuroblastoma
.
Nature
2008
;
455
:
975
8
.
36
Janoueix-Lerosey
I
,
Lequin
D
,
Brugieres
L
, et al
. 
Somatic and germline activating mutations of the ALK kinase receptor in neuroblastoma
.
Nature
2008
;
455
:
967
70
.
37
Mosse
YP
,
Laudenslager
M
,
Longo
L
, et al
. 
Identification of ALK as a major familial neuroblastoma predisposition gene
.
Nature
2008
;
455
:
930
5
.
38
Relaix
F
,
Montarras
D
,
Zaffran
S
, et al
. 
Pax3 and Pax7 have distinct and overlapping functions in adult muscle progenitor cells
.
J Cell Biol
2006
;
1721
:
91
102
.
39
Relaix
F
,
Rocancourt
D
,
Mansouri
A
,
Buckingham
M
. 
A Pax3/Pax7-dependent population of skeletal muscle progenitor cells
.
Nature
2005
;
435
:
948
53
.
40
Khan
J
,
Bittner
ML
,
Saal
LH
, et al
. 
cDNA microarrays detect activation of a myogenic transcription program by the PAX3-FKHR fusion oncogene
.
Proc Natl Acad Sci U S A
1999
;
9623
:
13264
9
.
41
Barber
TD
,
Barber
MC
,
Tomescu
O
,
Barr
FG
,
Ruben
S
,
Friedman
TB
. 
Identification of target genes regulated by PAX3 and PAX3-FKHR in embryogenesis and alveolar rhabdomyosarcoma
.
Genomics
2002
;
793
:
278
84
.
42
Begum
S
,
Emami
N
,
Cheung
A
,
Wilkins
O
,
Der
S
,
Hamel
PA
. 
Cell-type-specific regulation of distinct sets of gene targets by Pax3 and Pax3/FKHR
.
Oncogene
2005
;
2411
:
1860
72
.
43
Epstein
JA
,
Song
B
,
Lakkis
M
,
Wang
C
. 
Tumor-specific PAX3-FKHR transcription factor, but not PAX3, activates the platelet-derived growth factor α receptor
.
Mol Cell Biol
1998
;
187
:
4118
30
.
44
Tarnowski
M
,
Grymula
K
,
Reca
R
, et al
. 
Regulation of expression of stromal-derived factor-1 receptors: CXCR4 and CXCR7 in human rhabdomyosarcomas
.
Mol Cancer Res
2010
;
81
:
1
14
.
45
Zhang
L
,
Wang
C
. 
Identification of a new class of PAX3-FKHR target promoters: a role of the Pax3 paired box DNA binding domain
.
Oncogene
2007
;
2611
:
1595
605
.
46
Candia
AF
,
Hu
J
,
Crosby
J
, et al
. 
Mox-1 and Mox-2 define a novel homeobox gene subfamily and are differentially expressed during early mesodermal patterning in mouse embryos
.
Development
1992
;
1164
:
1123
36
.
47
Mankoo
BS
,
Collins
NS
,
Ashby
P
, et al
. 
Mox2 is a component of the genetic hierarchy controlling limb muscle development
.
Nature
1999
;
400
:
69
73
.
48
Martin
JF
,
Bradley
A
,
Olson
EN
. 
The paired-like homeo box gene MHox is required for early events of skeletogenesis in multiple lineages
.
Genes Dev
1995
;
910
:
1237
49
.
49
Conway
SJ
,
Henderson
DJ
,
Copp
AJ
. 
Pax3 is required for cardiac neural crest migration in the mouse: evidence from the splotch (Sp2H) mutant
.
Development
1997
;
1242
:
505
14
.
50
Mercado
GE
,
Xia
SJ
,
Zhang
C
, et al
. 
Identification of PAX3-FKHR-regulated genes differentially expressed between alveolar and embryonal rhabdomyosarcoma: focus on MYCN as a biologically relevant target
.
Genes Chromosomes Cancer
2008
;
476
:
510
20
.

Supplementary data