Ludwig Institute for Cancer Research Department of Medicine, Cancer Center and Center for Molecular Genetics University of California at San Diego La Jolla, California, USA wcavenee@ucsd.edu
A poorly understood cancer of childhood has just become less mysterious. A new animal model identifies three pathways that underlie the progression to rhabdomyosarcoma (pages 1276−1280).
Concerted efforts over the past two decades have led to an understanding of the genetic underpinnings of many human cancers. Left largely behind in this effort have been rare tumors with poorly understood heredity. These include childhood tumors with the undistinguished appearance of cells that are small, round and poorly differentiated. Rhabdomyosarcoma is one such tumor, and is the most common sarcoma of the soft tissues in children (although still relatively uncommon at 5 per million per year)1. In this issue, Sharp et al.2 describe a mouse model that, at a very young age, produces rhabodymyosarcomas at almost 100% frequency. This finding should accelerate research into these rare but devastating tumors.
Therapies for rhabdomyosarcoma and related tumors vary depending on the subtype. Currently, placing these tumors in a category can be difficult, so the hope is that an understanding of their genetic basis could help in classification. Therapies often involve extensive surgeries and lengthy rounds of combination chemotherapies. Although cooperative intergroup clinical trials have led to moderate success in treatment, it is still often ineffective and many children still die.
The name rhabdomyosarcoma derives from the occasional presence in the tumor of cells that resemble the precursors of striated muscle, called rhabdomyoblasts. This apparently incomplete or abortive striated-muscle differentiation has led to the idea that such tumors arise from muscle precursors. It also suggests that they might arise through an oncogenic process whereby transformation is superimposed upon incomplete or arrested tissue development. Despite these notions, relatively little is known about the etiology of this disease and current therapies are lengthy, toxic and not necessarily targeted to the genetic features of the tumor.
Genetic analyses of sporadically occurring rhabdomyosarcoma tumors have pinpointed several common alterations (Fig. 1). The list is long but includes an alteration uncommon in other tumors, the expression of myogenic master regulatory factors MYOD1 and myogenin3. Other mutations can affect the retinoblastoma pathway (through retinoblastoma mutations or amplifications of cell cycle−dependant kinase 4 (CDK4)4. The p53 pathway is also often altered (through p53 mutations and mouse double-minute 2 (MDM2) homolog amplifications)5,
6. Both the retinoblastoma and p53 pathways are affected by deletions in the gene encoding INK4a/ARF (cyclin-dependant kinase inhibitor 2a/alternative reading frame); these deletions are common in rhabdomyosarcomas. Other genetic lesions can increase expression of the c-MET receptor tyrosine kinase oncogene7 and the muscle-specific proteins desmin and -actin3. Curiously, mice constructed to mimic rhabdomyosarcoma-associated alterations singly do not develop these tumors at any substantial frequency.
Figure 1. Genetic pathways altered in rhabdomyosarcomas.
Symbols in purple designate the central regulatory players for rhabdomyosarcoma, c-Met, pRB and p53. Simultaneous disruption of the c-Met, pRB and p53 pathways severely aggravate deregulation of myogenic growth and differentiation. A subtype of rhabdomyosarcoma carries chromosomal translocations that fuse the gene encoding FKHR with genes for either PAX3 or PAX7 (proteins that regulate c-MET expression). Depicted modes of inhibition or stimulation of these pathways are either direct or indirect
In order to test rhabdomyosarcoma-associated genes in combination, the authors crossed two mouse strains that each had been developed to carry only one alteration. One strain expressed the ligand of the c-MET receptor, HGF/SF (hepatocyte growth factor/scatter factor), relatively broadly throughout the mouse8. The c-MET oncogene was originally isolated from a chemically treated osteogenic sarcoma cell line (another of the small round-cell tumors, mainly occurring in adolescence). Activation of the c-MET signaling pathway occurs in several types of human cancers6, and affects cell motility, proliferation and survival in experimental systems6. Activation of c-MET can occur through overexpression or activating mutations of the c-MET receptor itself or by autocrine expression of HGF/SF. Transgenic HGF/SF mice develop tumors rarely, and these are mostly melanomas, although a few rhabdomyosarcomas also develop.
The other mutant mouse strain created by Sharp et al. carried a germline knockout of the INK4a/ARF locus9. This locus encodes two unrelated proteins with tumor suppressor function, p16 and p19. p16 regulates the retinoblastoma pathway through modulation of CDK4 and CDK6 and p19 regulates p53 checkpoint function. The retinoblostoma and p53 pathways are among the most frequent targets in human cancer, and often alterations occur through mutations in p16 and the human homolog of p19 (p14). So it's not surprising that INK4a/ARF-knockout mice are prone to cancer. They develop mainly hematopoietic tumors, some melanomas and fibrosarcomas and, rarely, rhabdomyosarcomas. The tumors in either HGF/SF-transgenic or INK4a/ARF-knockout mice resemble those that develop in adult humans more than those in children, and both strains have a relatively long average lifespan over a 16-month interval.
The surprise came from mice that expressed the HGF/SF transgene and, in addition, were deficient for INK4a/ARF. The doubly mutant mice developed rhabdomyosarcomas with nearly a 100% frequency and died rapidly within 4−5 months. The mouse tumors had several features in common with their human counterparts: some muscle striations, stabilized p53 and expression of desmin, -actin, MYOD1 and Pax7. Moreover, myoblast precursor cells appeared to differentiate poorly, suggestive of the arrested or abortive differentiation of human rhabdomyosarcomas: While proliferating well in culture, these precursors were unable to fuse to form myotubes. Mice with this genetic combination represent the most faithful and predictable animal model of rhabdomyosarcoma yet.
The precise mechanisms by which this particular combination of mutations causes the selective and unusual development of rhabdomyosarcomas are unclear. However, it is tempting to speculate that it is the consequence of dysregulated proliferation and survival of rhabdomyoblasts coupled to a loss of orderly cell-cycle control. This combination could have the effect of allowing precursor cells to proliferate for a longer period of time before terminal differentiation, thereby increasing the target pool for transforming events. Because not all rhabdomyoblasts in these mice transform, it is clear that other mutations layered onto this background drive tumor progression. This raises several questions.
Would any proliferation advantage coupled to a cell-cycle defect lead to the development of rhabdomyosarcomas? Perhaps it is the combination of specific deregulated signaling and the rhabdomyoblast cell milieu which elicits the tumor type. The early tumor development typical of childhood cancer may also result from the coupling of specific deregulating mutations in early stages of tissue development. It will also be important to determine whether other genetic lesions seen in these tumors are actually secondary in human rhabdomyosarcoma or if they can be primary initiators under some circumstances. It is possible that different combinations of mutations give rise to the several histological and genetic subtypes of rhabdomyosarcoma in humans. Finally, the stage of muscle development at which transformation occurs might affect phenotype, tumor location or tumor aggressiveness.
Testing of preclinical candidate therapies should now be enhanced for these tumors. Perhaps most interesting is the ability to directly dissect the relationship of muscle development and tumorigenesis. It may be asking too much of the present set of mouse strains to answer the numerous outstanding questions about rhabdomyosarcoma and related cancers. But clearly, the work has just begun.
Sharp, R. et al. Synergism between INK4a/ARF inactivation and aberrant HGF/SF signaling in rhabdomyosarcomagenesis. Nature Med.8, 1276–1280 (2002). | Article | PubMed | ISI | ChemPort |
Tonin, P.T. et al. Muscle-specific gene expression in rhabdomyosarcomas and stages of human fetal skeletal muscle development. Cancer Res.51, 5100–5106 (1991). | PubMed | ISI | ChemPort |
Maelandsmo, G.M. et al. Homozygous deletion frequency and expression levels of the CDKN2A gene in human sarcomas—relationship to amplification and mRNA levels. Br. J. Cancer72, 393–398 (1995). | PubMed | ISI | ChemPort |
Mulligan, L.M. et al. Mechanisms of p53 loss in human sarcomas. Proc. Natl. Acad. Sci. USA87, 5863–5867 (1990). | PubMed | ChemPort |
Merlino, G. & Helman, L.J. Rhabdomyosarcoma—working out the pathways. Oncogene12, 1697–1705 (1996). | PubMed |
Ferracini, R. et al. Retrogenic expression of the MET proto-oncogene correlates with the invasive phenotype of human rhabdomyosarcoma. Oncogene12, 1697–1705 (1996). | PubMed | ISI | ChemPort |
Takayama, H. et al. Diverse tumorigenesis associated with aberrant development in mice overexpressing hepatocyte growth factor/scatter factor. Proc. Natl. Acad. Sci. USA94, 701–706 (1997). | Article | PubMed | ChemPort |
Serrano, M. et al. Role of the INK4a locus in tumor suppression and cell mortality. Cell85, 27–37 (1996). | Article | PubMed | ISI | ChemPort |
-actin
