Activation of the hedgehog pathway confers a poor prognosis in embryonal and fusion gene-negative alveolar rhabdomyosarcoma


Rhabdomyosarcoma (RMS) is the most common soft tissue sarcoma in children and comprises two major histological subtypes: alveolar rhabdomyosarcoma (ARMS) and embryonal rhabdomyosarcoma (ERMS). Seventy-five percent of ARMS harbor reciprocal chromosomal translocations leading to fusion genes of the forkhead transcription factor FOXO1 and PAX3 or PAX7. The hedgehog (Hh) pathway has been implied in tumor formation and progression of various cancers including RMS. However, whether Hh pathway activation presents a general feature of RMS or whether it is restricted to specific subgroups has not yet been addressed. Here, we report that marker genes of active Hh signaling, that is, Patched1 (Ptch1), Gli1, Gli3 and Myf5, are expressed at significantly higher levels in ERMS and fusion gene-negative ARMS compared with fusion gene-positive ARMS in two distinct cohorts of RMS patients. Consistently, Gli1 expression correlates with Ptch1 expression in ERMS and fusion gene-negative ARMS, but not in fusion gene-positive ARMS. In addition, expression levels of MyoD1 are significantly lower in ERMS and fusion gene-negative ARMS, pointing to an inverse association of Hh activation and early muscle differentiation. Moreover, Myf5 is identified as a novel excellent class predictor for RMS by receiver operating characteristic analysis. Importantly, high expression of Ptch1 or low MyoD1 expression significantly correlate with reduced cumulative survival in fusion gene-negative RMS underscoring the clinical relevance of these findings. By showing that Hh signaling is preferentially activated in specific subgroups of RMS, our study has important implications for molecular targeted therapies, such as small molecule Hh inhibitors, in RMS.


Rhabdomyosarcoma (RMS) is the most common soft tissue sarcoma in children (Dagher and Helman, 1999) and comprises two major subtypes: alveolar (ARMS) and embryonal (ERMS), which are associated with distinct genetic alterations (Dagher and Helman, 1999). ARMS is characterized by reciprocal chromosomal translocations between chromosome 2 or chromosome 1 and chromosome 13, that is, t(2;13) and t(1;13) (Merlino and Helman, 1999). These translocations lead to characteristic fusion genes that typically consist of the DNA-binding domain of the PAX3 or PAX7 genes located on chromosome 2 or chromosome 1 and the transactivation domain of the forkhead transcription factor FOXO1 gene that is located on chromosome 13 (Merlino and Helman, 1999). However, at least 25% of ARMS lack this translocation despite classical alveolar histology (Barr et al., 2002; Sorensen et al., 2002; Williamson et al., 2010). ERMS accounts for approximately two-thirds of all RMS and is typically associated with a more favorable prognosis compared with the alveolar subtype (Dagher and Helman, 1999). Recently, we have shown that fusion gene-negative ARMS are clinically and biologically similar to ERMS and distinct from ARMS with PAX3/PAX7-FOXO1 fusion genes (Williamson et al., 2010). At the molecular level, ERMS frequently shows loss of heterozygosity at chromosome 11p15 and not the chromosomal translocations that are typical for ARMS (Merlino and Helman, 1999). Despite multimodal treatment regimens and advances in combination chemotherapy, the overall survival for childhood RMS remains approximately 70% (Dantonello et al., 2009; Hayes-Jordan and Andrassy, 2009). This highlights the need to identify and validate novel molecular targets in RMS that can be exploited for cancer drug development to further improve the prognosis of these patients.

The hedgehog (Hh) pathway is an evolutionally highly conserved signaling cascade, which has a non-redundant function in the control of many developmental processes and has also been implicated to contribute to cancer formation and progression (Ruiz i Altaba et al., 2002; Pasca di Magliano and Hebrok, 2003). Within this pathway, Patched1 (Ptch1) together with smoothened (Smo) forms a receptor complex for Hh. Ptch1 normally represses Smo function in an autocrine and paracrine manner (Bijlsma et al., 2006). Binding of Hh to Ptch1 results in the inactivation of Ptch1 and consecutively in the activation of Smo. Smo positively controls transcription of Gli1 and Ptch. Therefore, the expression of both genes serves as an indicator of pathway activation (Hooper and Scott, 2005).

Distinct mutations in key components of the Hh signaling cascade, for example, mutations in Ptch1, have been identified in human cancers that result in pathological activation of the Hh pathway (Pasca di Magliano and Hebrok, 2003; Scales and de Sauvage, 2009). In addition to genetic alterations, ligand-dependent stimulation of Hh signaling has recently been shown to promote tumor cell growth in different types of human cancers (Berman et al., 2002, 2003; Thayer et al., 2003; Watkins et al., 2003; Karhadkar et al., 2004; Scales and de Sauvage, 2009). In RMS, aberrant activation of the Hh cascade has been detected in primary specimens and has been attributed to genetic inactivation of Ptch1 or of suppressor of fused (SUFU), like Ptch1, a negative regulator of Hh signaling, or to Gli1 amplification (Roberts et al., 1989; Khatib et al., 1993; Ragazzini et al., 2004; Tostar et al., 2006; Oue et al., 2010). Ptch1 heterozygous mice have been shown to develop RMS resembling the embryonal subtype (Hahn et al., 1998). Furthermore, overexpression of Ptch1 and Gli1 mRNA has been reported in RMS (Tostar et al., 2006).

Small molecule inhibitors of the Hh pathway have recently been shown to exhibit antitumor activity in pre-clinical mouse models of Hh-driven cancers as well as in pilot clinical studies (Romer et al., 2004; Rubin and de Sauvage, 2006; Rudin et al., 2009; Von Hoff et al., 2009). It will therefore be important to identify those cancers or subgroups of patients that harbor constitutively elevated Hh signaling, as they will likely be susceptible to pathway inhibition. As far as RMS is concerned, it is currently unknown whether or not abnormal Hh activation universally occurs in all RMS patients or whether it is restricted to specific subtypes of RMS. In addition, the question as to whether abnormal Hh pathway activation bears prognostic impact in RMS has not yet been answered. In this study, we therefore investigated Hh pathway activation in distinct subgroups of RMS.


Hh pathway activation in RMS subtypes

To explore whether the Hh pathway is differentially activated in ARMS or ERMS, we analyzed gene expression of Gli1 and Ptch1 as markers of Hh pathway activation (Scales and de Sauvage, 2009). As Gli3 is essential for Gli1 expression in the somites during muscle formation (McDermott et al., 2005), Gli3 was analyzed as well. As we previously reported that Hh activation interferes with muscle differentiation (Koleva et al., 2005), we also determined expression levels of markers of muscle differentiation, that is, MyoD1 and Myf5 as early markers of muscle differentiation and Myh1 as a late marker (Tomczak et al., 2004). Furthermore, Myf5 is regulated by Hh signaling (Gustafsson et al., 2002; Gerber et al., 2007). Initially, we assessed expression profiles of these markers in a group of 103 RMS patients that we recently characterized (cohort 1) (Williamson et al., 2010). In addition, we analyzed a previously published separate group of 132 RMS patients to validate our results (cohort 2) (Davicioni et al., 2006).

Analysis of gene expression profiles in the two major histological subgroups of patients, that is, ERMS and ARMS, revealed significantly higher expression levels of Gli1, Gli3 and Myf5 in ERMS in both cohorts of RMS patients, and significant Ptch1 overexpression in ERMS of cohort 2 (Figure 1, Supplementary Figure 1 and Table 1). In contrast, MyoD1 was expressed at significantly lower levels in ERMS in both cohorts (Figure 1, Supplementary Figure 1 and Table 1). When the fusion gene status of tumors was taken into account, significantly higher levels of Gli1, Gli3, Ptch1 and Myf5 and significantly lower levels of MyoD1 were detected in ERMS compared with fusion gene-positive ARMS in both cohorts (Figure 1, Supplementary Figure 1 and Table 1). This was also revealed by a comparison between fusion gene-negative and fusion gene-positive ARMS in both cohorts (Figure 1, Supplementary Figure 1 and Table 1). In contrast, no differential gene expression of Gli1, Gli3, Ptch1, Myf5 or MyoD1 was observed when ERMS was compared with fusion gene-negative ARMS (Figure 1, Supplementary Figure 1 and Table 1). Expression levels of Myh1 displayed no significant difference between the distinct subgroups (Figure 1 and Table 1). To confirm these results, we also analyzed an independent cohort by a complementary approach using real-time PCR analysis. Patients characteristics of cohort 3 are summarized in Table 2. Similarly, we found significantly higher Myf5 expression levels in embryonal compared with alveolar tumors (Supplementary Table 1). In addition, expression levels of Gli3 were higher in ERMS than in ARMS, although this did not reach statistical significance (Supplementary Table 1). In summary, these data showed that in general ERMS and fusion gene-negative ARMS highly express Gli1, Gli3, Ptch1 and Myf5, whereas fusion gene-positive ARMS do not. In addition, MyoD expression is lower in the former cohorts. Thus, we identify subgroups of RMS cases that exhibit high Hh pathway activity, that is, ERMS and fusion gene-negative ARMS.

Figure 1

Gene expression in RMS subgroups. Expression of Gli1, Gli3, Ptch1, Myf5, MyoD1 and Myh1 was analyzed by affymetrix HGU133plus2 chip in patients of cohort 1. Data are depicted by box plots; the line inside each box denotes median, boxes 25th and 75th percentiles, error bars minimum and maximum, dots and crosses refer to individual cases. ARMS Neg: fusion gene-negative ARMS; RMS P3F: PAX3/FOXO1-positive RMS; and RMS P7F: PAX7/FOXO1-positive RMS. A full colour version of this figure is available at the Oncogene journal online.

Table 1 Comparison of gene expression in RMS subgroups
Table 2 Patients characteristics (cohort 3)

Next, we correlated expression levels of Gli1, Gli3, Ptch1, Myf5, MyoD1 and Myh1 to identify potential co-regulation of individual genes. As shown in Table 2, Gli1 correlated with Ptch1 expression, when the overall RMS sample set was analyzed. When RMS subgroups were assessed, this correlation was detected in ERMS and in fusion gene-negative ARMS (Table 3). In contrast, no correlation between Gli1 and Ptch1 expression was found when ARMS or fusion gene-positive ARMS were analyzed separately (Table 3) or when Myf5 and Gli3 expression were correlated (data not shown). This points to the co-regulation of Gli1 and Ptch1 expression in specific subsets of RMS, that is, in ERMS and fusion gene-negative ARMS, and underscores the activation of the Hh pathway in these subgroups.

Table 3 Correlation of Gli1 and Ptch1 expression in RMS subgroups

Classification of RMS by markers of Hh pathway activity

We next asked which of the analyzed marker genes could serve as class predictor for any subgroup of RMS, either based on histology (ERMS or ARMS) or on the fusion gene status (fusion gene negative versus fusion gene positive). To address this question, we analyzed the area under the receiver operating characteristic (ROC) curve. Interestingly, Ptch1, Gli3, Myf5 and MyoD1 were identified as class predictors with acceptable (0.7<ROC<0.8) or excellent (0.8<ROC<0.9) discrimination for fusion gene-negative versus -positive cases in both cohorts (Table 4). These four genes were also identified as acceptable or excellent class predictors in cohort 2 (and except Ptch1 also in cohort 1), when the prediction was carried out for histology, that is, ERMS versus ARMS (Table 4). Interestingly, a better discrimination for these genes was achieved when the fusion gene status rather than histology was taken into consideration (Table 4).

Table 4 ROC analysis

Furthermore, we analyzed class predictive effects of Myf5 in a distinct group of 28 patients diagnosed with RMS to verify these observations in an additional patient cohort, that is, cohort 3. Figure 2 depicts the cross-validated prediction probability for ERMS based on log Myf5 expression determined by quantitative reverse transcription–PCR. Importantly, all but one ARMS and ERMS cases were correctly classified by means of Myf5 expression (Figure 2). This set of data confirms in an independent cohort of RMS patient that Myf5 serves as an excellent class predictor for RMS.

Figure 2

Classification of RMS by Myf5 expression. Class predictive effects of Myf5 were analyzed in cohort 3 by forward selection and score selection in parallel within logistic regression models. Shown is the cross-validated prediction probability for ERMS based on log Myf5 expression (P=0.0183). ERMS cases (blue asterisks) showed gene expression of 0.3 and larger, whereas ARMS cases (red dots) showed gene expression of lower than 0.3, with 0.30 being the optimal cut point (95% confidence interval: 0.26–0.34). The concordance between true and predicted classification was 97.5%. Solid line: prediction probability for ERMS based on Myf5 expression; dotted line: 95% confidence region of the optimal cutting point. A full colour version of this figure is available at the Oncogene journal online.

Hh pathway activation correlates with poor outcome in fusion gene-negative RMS

Finally, we explored whether gene expression of the selected marker genes correlates with patients survival. To address this question, we analyzed all fusion gene-negative RMS, as fusion gene-positive status already has a strong negative impact on prognosis. Interestingly, we found that high expression of Ptch1 correlated with a reduced overall survival compared with low Ptch1 expression in both cohorts (Figure 3a and Supplementary Figure 2A). Also, survival was significantly decreased in patients with low MyoD1 expression in both cohorts (Figure 3b and Supplementary Figure 2B).

Figure 3

Correlation of Ptch1 and MyoD1 expression with survival in fusion gene-negative RMS. Expression levels (high or low based on best split value) of Ptch1 (a) and MyoD1 (b) in cohort 1 were correlated with overall survival in cases with fusion gene-negative RMS. Shown is the 10-year cumulative survival. The number of subjects is as follows: Ptch1 high: 7, Ptch1 low: 52 (a); MyoD1 low: 8, MyoD1 high: 51 (b). A full colour version of this figure is available at the Oncogene journal online.


In this study, we show for the first time that Hh pathway activation is a characteristic feature of specific subgroups of RMS patients defined by histology and molecular genetics, that is, ERMS and fusion gene-negative ARMS. This conclusion is supported by data showing that markers of Hh pathway activation, that is, Ptch1, Gli1, Gli3 and Myf5, are expressed at significantly higher levels in ERMS and fusion gene-negative ARMS compared with fusion gene-positive ARMS. Consistent with Hh pathway activation, Gli1 expression correlates with Ptch1 expression in ERMS and fusion gene-negative ARMS, but not in fusion gene-positive ARMS. The clinical relevance of Hh pathway activation in these RMS subtypes is underscored by data showing that Hh activation correlates with reduced survival in fusion gene-negative RMS, indicating that activation of the Hh pathway in the group of fusion gene-negative tumors is associated with more aggressive tumors. However, as the analyses were carried out in a retrospective manner with a relatively small number of patients, the findings on the prognostic impact await confirmation in a larger prospective clinical trial.

Gli1 and Ptch1 are pivotal markers of ongoing active Hh signaling (Ruiz i Altaba et al., 2002; Pasca di Magliano and Hebrok, 2003). Although both genes have previously been shown to be expressed in human RMS at higher levels than in normal muscle (Tostar et al., 2006), the question as to whether Hh signaling is differentially activated in distinct subsets of RMS has not been previously addressed. Thus, the novelty of our study resides in the identification of specific subgroups of RMS patients, that is, ERMS and fusion gene-negative ARMS, that exhibit Hh pathway activation. Although Gli3 has been reported to be essential for Gli1 expression in the somites during muscle formation (McDermott et al., 2005), we found no correlation between RNA expression levels of these genes in the RMS samples analyzed. This may indicate that Gli1 expression depends on cellular Gli3 levels predominately in a developmental context, implying a higher level of non-redundancy under these conditions.

Moreover, our findings showing an inverse association between Gli1 and MyoD1 expression confirm our previous findings that the Hh pathway negatively regulates myogenesis and early muscle differentiation in muscle progenitor cells (Koleva et al., 2005). It is also consistent with a study of Gerber et al. (2007), who showed that Gli1 directly suppresses MyoD-mediated transcriptional activation and thus inhibits myoblast differentiation. This direct biochemical link between Gli1 and MyoD might also explain why we found a correlation between Gli1 and MyoD1 expression, but not for Myh1.

Furthermore, our analysis identifies Myf5 as an excellent class predictor for RMS based on the fusion gene status or on histology. Myf5 has been shown to be a downstream target of Gli1 during muscle development, at least in the somite epaxial muscle progenitors in mice (Gustafsson et al., 2002). In addition, its expression is dependent on Gli3 expression during embryonic muscle formation (McDermott et al., 2005). Myf5 overexpression was also recently reported in zebrafish model of ERMS (Langenau et al., 2007), underscoring the relevance of Myf5 in these tumors. Besides Myf5, Gli3, Ptch1 and MyoD1 also qualified as class predictors in this study. Previously, MyoD1 has been reported to distinguish undifferentiated embryonal sarcoma of the liver from biliary tract RMS (Nicol et al., 2007). Together with gene expression signatures and histology (Davicioni et al., 2006, 2009; Wachtel et al., 2006), class predictors may add to the molecular classification and diagnosis of RMS. Future clinical correlative studies are warranted to evaluate the relevance of these genes in class prediction. It is interesting to note that we consistently found that gene expression profiles of fusion gene-negative ARMS closely resembled that of ERMS. This is in line with recent reports that ARMS cases lacking PAX3/7-FOXO1 chromosomal translocations are indistinguishable in their gene expression profiles from ERMS (Davicioni et al., 2009; Williamson et al., 2010).

Our findings are expected to have important implications for the development of molecular targeted therapies in RMS. First, Ptch heterozygous mice, which develop embryonal subtype-like RMS (Hahn et al., 1998; Kappler et al., 2004), likely present a suitable model for the pre-clinical evaluation of Hh antagonists, as the gene expression pattern of Hh signaling and myogenic markers identified in this study in ERMS and fusion gene-negative ARMS is similar to the profile that we previously described in Ptch heterozygous mice (Kappler et al., 2004; AZ and HH, unpublished data). Second, our study identifies specific subgroups of RMS that harbor constitutive Hh activation, which may occur in a cell-intrinsic manner, implying that they are prime candidates for future studies with small molecule inhibitors of the Hh pathway. Recently, Hh antagonists showed antitumor activity in pilot clinical studies in Hh-driven cancers, for example, medulloblastoma and basal cell carcinoma (Rudin et al., 2009; Von Hoff et al., 2009). Our results suggest that pre-selected RMS patients with Hh pathway activation may also benefit from Hh antagonist-based treatment protocols, a hypothesis that warrants further investigation.

Materials and methods


Cohort 1 (n=103) (Williamson et al., 2010) and cohort 2 (n=132) (Davicioni et al., 2006) of RMS patients were described previously. Patients’ characteristics of cohort 3 are summarized in Table 4. Diagnoses were defined according to international classication guidelines as reported previously (Davicioni et al., 2006; Williamson et al., 2010).

RNA isolation and reverse transcription–PCR

Total RNA was isolated using Trizol reagent (Invitrogen, Carlsbad, CA, USA). Reverse transcription of each sample RNA was conducted with random hexamers and SuperScriptII reverse transcriptase (Invitrogen, Karlsruhe, Germany). Expression of genes was quantified using SYBR-green-based assays on an ABI PRISM 7900HT Sequence Detection System (Applied Biosystems Inc., Foster City, CA, USA). All primer pairs used are exon spanning and were as follows: hsGLI3F.1, 5′-IndexTermGCCAGCGCAGCC CCT-3′IndexTerm; AT/hsGLI3R.1, 5′-IndexTermCGGCCTGGCTGACAGCCT-3′; and hsMYF5F.2, 5′-IndexTermAGAACTACTATAGCCTGCCGG-3′; GA/hsMYF5R.2, 5′-IndexTermATCTGTGGCATATACATTTGATACATCA-3′ for amplification of Gli3 and Myf5, respectively. Amplification of 18S rRNA (18S-fwd, 5′-IndexTermCGCAAATTACCCACTCCCG-3′/18S-rev2, 5′-IndexTermTTCCAATTACAGGGCCTCGAA-3′) was performed to standardize the amount of sample RNA. All sample probes were reverse transcribed in two independent experiments and measured in triplicates. For calibration of the data, Human Skeletal Muscle PCR-Ready cDNA from Ambion (Austin, TX, USA) was amplified.

Microarray analysis

The expression profiling of the two data sets was quality evaluated as described previously in Missiaglia et al. (2009) and Williamson et al. (2010) and normalized by the gcRMA function, which is a modified Robust Multiarray Average that takes into account probe sequence information for the background adjustment. Microarray analysis was performed by HGU133plus2 arrays (cohort 1) or by U133A expression arrays (cohort 2). For each of the genes, the probeset with the highest median express level was selected for further analysis. Non-parametric Wilcoxon rank test was employed to evaluate difference between subtypes, whereas Pearson's product moment correlation coefficient was used to assess correlation between gene expressions. All P-values were adjusted using Bonferroni correction in order to minimize multitesting error. ROC curves were generated using the ROC library. Discrimination was classified into outstanding (0.9<ROC<1.0), excellent (0.8<ROC<0.9) and acceptable (0.7<ROC<0.8). All analyses were performed in R 2.9.0.

Survival analysis

Gene expression based on best split value of Ptch1 or MyoD1 was correlated with cumulative survival in fusion gene-negative cases. The correlation with overall survival was assessed using log-rank test.

Class prediction

Class predictive effects of Myf5 were analyzed in cohort 3 by forward selection and score selection in parallel within logistic regression models. Allowing main effects and two-way interactions only, the best fitting model by both strategies contains effect parameters for log (Myf5) expression.


  1. Barr FG, Qualman SJ, Macris MH, Melnyk N, Lawlor ER, Strzelecki DM et al. (2002). Genetic heterogeneity in the alveolar rhabdomyosarcoma subset without typical gene fusions. Cancer Res 62: 4704–4710.

    CAS  PubMed  Google Scholar 

  2. Berman DM, Karhadkar SS, Hallahan AR, Pritchard JI, Eberhart CG, Watkins DN et al. (2002). Medulloblastoma growth inhibition by hedgehog pathway blockade. Science 297: 1559–1561.

    CAS  Article  Google Scholar 

  3. Berman DM, Karhadkar SS, Maitra A, Montes De Oca R, Gerstenblith MR, Briggs K et al. (2003). Widespread requirement for Hedgehog ligand stimulation in growth of digestive tract tumours. Nature 425: 846–851.

    CAS  Article  Google Scholar 

  4. Bijlsma MF, Spek CA, Zivkovic D, van de Water S, Rezaee F, Peppelenbosch MP . (2006). Repression of smoothened by patched-dependent (pro-)vitamin D3 secretion. PLoS Biol 4: e232.

    Article  Google Scholar 

  5. Dagher R, Helman L . (1999). Rhabdomyosarcoma: an overview. Oncologist 4: 34–44.

    CAS  PubMed  Google Scholar 

  6. Dantonello TM, Int-Veen C, Harms D, Leuschner I, Schmidt BF, Herbst M et al. (2009). Cooperative trial CWS-91 for localized soft tissue sarcoma in children, adolescents, and young adults. J Clin Oncol 27: 1446–1455.

    CAS  Article  Google Scholar 

  7. Davicioni E, Anderson MJ, Finckenstein FG, Lynch JC, Qualman SJ, Shimada H et al. (2009). Molecular classification of rhabdomyosarcoma—genotypic and phenotypic determinants of diagnosis: a report from the Children's Oncology Group. Am J Pathol 174: 550–564.

    CAS  Article  Google Scholar 

  8. Davicioni E, Finckenstein FG, Shahbazian V, Buckley JD, Triche TJ, Anderson MJ . (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.

    CAS  Article  Google Scholar 

  9. Gerber AN, Wilson CW, Li YJ, Chuang PT . (2007). The hedgehog regulated oncogenes Gli1 and Gli2 block myoblast differentiation by inhibiting MyoD-mediated transcriptional activation. Oncogene 26: 1122–1136.

    CAS  Article  Google Scholar 

  10. Gustafsson MK, Pan H, Pinney DF, Liu Y, Lewandowski A, Epstein DJ et al. (2002). Myf5 is a direct target of long-range Shh signaling and Gli regulation for muscle specification. Genes Dev 16: 114–126.

    CAS  Article  Google Scholar 

  11. Hahn H, Wojnowski L, Zimmer AM, Hall J, Miller G, Zimmer A . (1998). Rhabdomyosarcomas and radiation hypersensitivity in a mouse model of Gorlin syndrome. Nat Med 4: 619–622.

    CAS  Article  Google Scholar 

  12. Hayes-Jordan A, Andrassy R . (2009). Rhabdomyosarcoma in children. Curr Opin Pediatr 21: 373–378.

    Article  Google Scholar 

  13. Hooper JE, Scott MP . (2005). Communicating with hedgehogs. Nat Rev Mol Cell Biol 6: 306–317.

    CAS  Article  Google Scholar 

  14. Kappler R, Bauer R, Calzada-Wack J, Rosemann M, Hemmerlein B, Hahn H . (2004). Profiling the molecular difference between patched- and p53-dependent rhabdomyosarcoma. Oncogene 23: 8785–8795.

    CAS  Article  Google Scholar 

  15. Karhadkar SS, Bova GS, Abdallah N, Dhara S, Gardner D, Maitra A et al. (2004). Hedgehog signalling in prostate regeneration, neoplasia and metastasis. Nature 431: 707–712.

    CAS  Article  Google Scholar 

  16. Khatib ZA, Matsushime H, Valentine M, Shapiro DN, Sherr CJ, Look AT . (1993). Coamplification of the CDK4 gene with MDM2 and GLI in human sarcomas. Cancer Res 53: 5535–5541.

    CAS  PubMed  Google Scholar 

  17. Koleva M, Kappler R, Vogler M, Herwig A, Fulda S, Hahn H . (2005). Pleiotropic effects of sonic hedgehog on muscle satellite cells. Cell Mol Life Sci 62: 1863–1870.

    CAS  Article  Google Scholar 

  18. Langenau DM, Keefe MD, Storer NY, Guyon JR, Kutok JL, Le X et al. (2007). Effects of RAS on the genesis of embryonal rhabdomyosarcoma. Genes Dev 21: 1382–1395.

    CAS  Article  Google Scholar 

  19. McDermott A, Gustafsson M, Elsam T, Hui CC, Emerson Jr CP, Borycki AG . (2005). Gli2 and Gli3 have redundant and context-dependent function in skeletal muscle formation. Development 132: 345–357.

    CAS  Article  Google Scholar 

  20. Merlino G, Helman LJ . (1999). Rhabdomyosarcoma—working out the pathways. Oncogene 18: 5340–5348.

    CAS  Article  Google Scholar 

  21. Missiaglia E, Selfe J, Hamdi M, Williamson D, Schaaf G, Fang C et al. (2009). Genomic imbalances in rhabdomyosarcoma cell lines affect expression of genes frequently altered in primary tumors: an approach to identify candidate genes involved in tumor development. Genes Chromosomes Cancer 48: 455–467.

    CAS  Article  Google Scholar 

  22. Nicol K, Savell V, Moore J, Teot L, Spunt SL, Qualman S . (2007). Distinguishing undifferentiated embryonal sarcoma of the liver from biliary tract rhabdomyosarcoma: a Children's Oncology Group study. Pediatr Dev Pathol 10: 89–97.

    Article  Google Scholar 

  23. Oue T, Yoneda A, Uehara S, Yamanaka H, Fukuzawa M . (2010). Increased expression of the hedgehog signaling pathway in pediatric solid malignancies. J Pediatr Surg 45: 387–392.

    Article  Google Scholar 

  24. Pasca di Magliano M, Hebrok M . (2003). Hedgehog signalling in cancer formation and maintenance. Nat Rev Cancer 3: 903–911.

    Article  Google Scholar 

  25. Ragazzini P, Gamberi G, Pazzaglia L, Serra M, Magagnoli G, Ponticelli F et al. (2004). Amplification of CDK4, MDM2, SAS and GLI genes in leiomyosarcoma, alveolar and embryonal rhabdomyosarcoma. Histol Histopathol 19: 401–411.

    CAS  PubMed  Google Scholar 

  26. Roberts WM, Douglass EC, Peiper SC, Houghton PJ, Look AT . (1989). Amplification of the gli gene in childhood sarcomas. Cancer Res 49: 5407–5413.

    CAS  PubMed  Google Scholar 

  27. Romer JT, Kimura H, Magdaleno S, Sasai K, Fuller C, Baines H et al. (2004). Suppression of the Shh pathway using a small molecule inhibitor eliminates medulloblastoma in Ptc1(+/−)p53(−/−) mice. Cancer Cell 6: 229–240.

    CAS  Article  Google Scholar 

  28. Rubin LL, de Sauvage FJ . (2006). Targeting the Hedgehog pathway in cancer. Nat Rev Drug Discov 5: 1026–1033.

    CAS  Article  Google Scholar 

  29. Rudin CM, Hann CL, Laterra J, Yauch RL, Callahan CA, Fu L et al. (2009). Treatment of medulloblastoma with hedgehog pathway inhibitor GDC-0449. N Engl J Med 361: 1173–1178.

    CAS  Article  Google Scholar 

  30. Ruiz i Altaba A, Sanchez P, Dahmane N . (2002). Gli and hedgehog in cancer: tumours, embryos and stem cells. Nat Rev Cancer 2: 361–372.

    CAS  Article  Google Scholar 

  31. Scales SJ, de Sauvage FJ . (2009). Mechanisms of Hedgehog pathway activation in cancer and implications for therapy. Trends Pharmacol Sci 30: 303–312.

    CAS  Article  Google Scholar 

  32. Sorensen PH, Lynch JC, Qualman SJ, Tirabosco R, Lim JF, Maurer HM et al. (2002). PAX3-FKHR and PAX7-FKHR gene fusions are prognostic indicators in alveolar rhabdomyosarcoma: a report from the children's oncology group. J Clin Oncol 20: 2672–2679.

    CAS  Article  Google Scholar 

  33. Thayer SP, di Magliano MP, Heiser PW, Nielsen CM, Roberts DJ, Lauwers GY et al. (2003). Hedgehog is an early and late mediator of pancreatic cancer tumorigenesis. Nature 425: 851–856.

    CAS  Article  Google Scholar 

  34. Tomczak KK, Marinescu VD, Ramoni MF, Sanoudou D, Montanaro F, Han M et al. (2004). Expression profiling and identification of novel genes involved in myogenic differentiation. FASEB J 18: 403–405.

    CAS  Article  Google Scholar 

  35. Tostar U, Malm CJ, Meis-Kindblom JM, Kindblom LG, Toftgard R, Unden AB . (2006). Deregulation of the hedgehog signalling pathway: a possible role for the PTCH and SUFU genes in human rhabdomyoma and rhabdomyosarcoma development. J Pathol 208: 17–25.

    CAS  Article  Google Scholar 

  36. Von Hoff DD, LoRusso PM, Rudin CM, Reddy JC, Yauch RL, Tibes R et al. (2009). Inhibition of the hedgehog pathway in advanced basal-cell carcinoma. N Engl J Med 361: 1164–1172.

    CAS  Article  Google Scholar 

  37. Wachtel M, Runge T, Leuschner I, Stegmaier S, Koscielniak E, Treuner J et al. (2006). Subtype and prognostic classification of rhabdomyosarcoma by immunohistochemistry. J Clin Oncol 24: 816–822.

    CAS  Article  Google Scholar 

  38. Watkins DN, Berman DM, Burkholder SG, Wang B, Beachy PA, Baylin SB . (2003). Hedgehog signalling within airway epithelial progenitors and in small-cell lung cancer. Nature 422: 313–317.

    CAS  Article  Google Scholar 

  39. Williamson D, Missiaglia E, de Reynies A, Pierron G, Thuille B, Palenzuela G et al. (2010). Fusion gene-negative alveolar rhabdomyosarcoma is clinically and molecularly indistinguishable from embryonal rhabdomyosarcoma. J Clin Oncol 28: 2151–2158.

    Article  Google Scholar 

Download references


This work has been supported by grants from the Deutsche Forschungsgemeinschaft, the Bundesministerium für Bildung und Forschung, the European Community (ApopTrain, APO-SYS) and IAP6/18 (to SF), and from the Wilhelm-Sander-Stiftung 2003.112.3 (to HH). We would like to thank the Chris Lucas Trust for their support (to EM) and the Children's Cancer and Leukemia Group for providing tumor material and clinical information. La Ligue Nationale Contre Le Cancer in collaboration with Olivier Delattre kindly supported the expression profiling of tumors.

Author information



Corresponding author

Correspondence to S Fulda.

Ethics declarations

Competing interests

The authors declare no conflict of interest.

Additional information

Supplementary Information accompanies the paper on the Oncogene website

Supplementary information

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Zibat, A., Missiaglia, E., Rosenberger, A. et al. Activation of the hedgehog pathway confers a poor prognosis in embryonal and fusion gene-negative alveolar rhabdomyosarcoma. Oncogene 29, 6323–6330 (2010).

Download citation


  • rhabdomyosarcoma
  • hedgehog
  • ERMS
  • ARMS

Further reading


Quick links