by ecancer reporter Clare Sansom

Medulloblastoma is both the most common malignant brain tumour and the second most common cancer type in children. Its main treatments are surgery followed by radiation and/or chemotherapy, and in many cases this is curative although with a serious risk of damage to the developing nervous system. 

However, the tumour can disseminate from the cerebellum, where it arises, to coat the brain and spinal cord, and these cases are typically fatal.

Little is known about the mechanisms through which the cells of this tumour disseminate through the cerebrospinal fluid, and its metastases have generally been assumed to be genetically similar to the primary tumour. 

 Now, however, a large group of researchers led by Michael Taylor from the Hospital for Sick Children, University of Toronto, Toronto, Canada have used mouse models to show that medulloblastoma metastases in an individual differ from the primary tumour but are genetically similar to each other. Preliminary data from human tumours also suggested that this pattern also occurs in children.

The patched-1-heterozygous (Ptch^+/-) mouse model is predisposed to medulloblastoma, with 30% of animals developing the tumour by 8 months of age.

Addition of a transposon to this mouse model produced mice with greatly increased incidence of medulloblastoma and in which the tumours disseminated through the cerebrospinal fluid in an identical way to that seen in human children. 

A similar pattern was seen with transposon initiation in a mouse model with mutant P53.

The researchers used PCR and 454 sequencing to identify sites of transposon insertion in the genomes of primary medulloblastomas and matched metastases in both these mouse models, and found that a large number of these occurred in genes that have been suggested as likely oncogenes or tumour suppressor genes in the human disease. 

The researchers concluded that these mouse models resembled human medulloblastoma in both genetics and pathology.

Strikingly, the researchers found that there was a very low degree of overlap – typically about 9% – between the locations of common transposon insertion sites in the primary tumours and those in metastases from the same animals. Primary tumour samples from each animal were usually genetically identical, however, as were metastases.

This is consistent with a model in which all metastases arise from a subclone of the original tumour. Both insertions occurring in the metastases but not in the primary tumour and those occurring in the primary tumour but not the metastases were observed; genes involved in the cytoskeleton were more commonly mutated in the metastases.

This finding, if replicated in human children, would have important implications for the development of targeted therapy.

To test this hypothesis further, the researchers reviewed all cases of metastatic medullo-blastoma treated at the Hospital for Sick Children, Ontario, Canada during the previous decade. In 11 of 19 of these cases a significant difference in response to radiotherapy was observed between the primary tumour and the metastases. 

Tissue from seven matched human primary and metastatic medulloblastomas was then examined for copy number aberrations. 

The results were similar to those in the mouse model, with a different pattern of genetic aberrations observed in the metastases from the primary tumour in each child. Genetic differences between metastases in a single individual were also occasionally observed.

Further analysis using interphase fluorescence in situ hybridization (FISH) showed the known medulloblastoma oncogene MYC to be amplified in both primary and metastatic tumour tissue, whereas the oncogene MYCN was amplified in primary tumours only. 

Analysis of promoter methylation showed epigenetic changes as well as genetic ones, to differ between the primary tumour and the metastases in each individual. Furthermore, preliminary pathway analysis suggested insulin signaling as a potential therapeutic target for both primary and metastatic tumours. 

Taken together, these results indicate that there are significant genetic differences between primary and metastatic medulloblastoma cells in human children as well as in a validated mouse model of medulloblastoma. Therefore, it is very likely that any drugs designed to target only the better studied and more tractable primary tumours will not affect the more virulent metastatic disease.

Reference

Wu, X., Northcott, P.A., Dubuc, A. and 29 others (2012). Clonal selection drives genetic divergence of metastatic medulloblastoma. Nature, published online ahead of print 16 February 2012. doi:10.1038/nature10825