In the past 20 years, there has been little progress in identifying improved chemotherapy regimens for children with metastatic rhabdomyosarcoma. Of the two subtypes, aveolar tumours have a worse prognosis than embryonal tumours. Much of the cell and molecular biology of aveolar rhabdomyosarcoma is unknown, apart from the presence of the paired box gene 3 ( PAX3 ):forkhead ( FKHR ) t(2;13) translocation in most tumours. So Keller, Capecchi and colleagues have generated a new alveolar rhabdomyosarcoma mouse model to further investigate this rare tumour of skeletal muscle, and they present their results in two papers published in Genes and Development.

Keller and co-workers have knocked in a version of the Pax3:Fkhr fusion gene — believed to result in a gain-of-function mutation affecting muscle development — and at the same time generated the corresponding inactivation of one allele of Pax3 and one allele of Fkhr, carefully re-creating the genetic situation in the human disease. Importantly, the authors have targeted this mutation to be expressed only in terminally differentiating skeletal muscle. Their paper addressing Pax3:Fkhr expression in early development shows, through extensive use of different conditional mouse models, that the Pax3:Fkhr mutation in early precursor, embryonal and postnatal muscle stem cells seems unlikely to give rise to tumours, because of the presence of significant muscle defects during embryogenesis and the complete absence of tumour development in mice expressing Pax3:Fkhr in postnatal muscle stem cells.

The authors followed the development and growth of 228 Pax3:Fkhr mice over 29 months and only one animal developed a rhabdomyosarcoma during this time. So, to investigate what might increase the tumour incidence, the authors first examined if loss of the functional allele of Fkhr enhanced the tumorigenic process (as loss of Pax3 is known not to predispose to tumour formation), but no increase in tumour incidence was seen. Human alveolar rhabdomyosarcomas often have mutations in TP53 or CDKN2A (which encodes INK4A and ARF), so Pax3:Fkhr mice were crossed with mice that had muscle-specific loss of one allele of either gene. Again, no increase in tumour incidence was seen, so the authors started to breed homozygotes for each gene. Their results show that rhabdomyosarcomas arise more frequently in these animals only in the absence of functional p53 pathways and mostly require Pax3:Fkhr homozygosity. Importantly, these mouse aveolar rhabdomyosarcomas were largely immunohistologically similar to those found in humans.

The finding that rhabdomyosarcomas in mice most often arise from Pax3:Fkhr homozygotes and not heterozygotes differs from the human disease and might reflect that either the Pax3:Fkhr fusion gene is not completely identical to that occurring in humans, or might simply be a species-specific difference. However, these results do explain one of the controversies surrounding rhabdomyosarcoma development: that these tumours most likely arise from terminally differentiating skeletal muscle cells that seem to re-express early myogenic markers, and not from early muscle stem cells. Further characterization of this mouse model will hopefully provide novel molecular targets for the effective treatment of childhood alveolar rhabdomyosarcoma.