Cancer genetics

Evolution after tumour spread

A genetic study of brain cancers in mice and humans reveals distinct mutations in primary tumours and their metastases, suggesting that the two disease 'compartments' may require different treatments.

The spread of a primary tumour to secondary sites in the body is a key step in the development of many cancers, and treatment of these secondary metastatic tumours represents one of the foremost challenges in oncology. In an article published on Nature's website today, Wu et al.1 describe two new mouse strains that serve as models of metastasis in the childhood brain cancer medulloblastoma. In the mice, primary and metastatic tumours seem to occupy two genetically distinct 'compartments', which arise from divergent DNA-sequence mutations that occur after metastasis. The authors also detect similar differences in human medulloblastoma tumours — a finding that may influence the development of anticancer therapies.

Wu et al. used an experimental system called Sleeping Beauty mutagenesis2,3 to introduce random genetic mutations into cerebellar progenitor cells in the developing brains of two strains of mice. These two new strains were derived by breeding the existing Tp53mut and Ptch+/− strains, which are predisposed to brain tumours1,4, with a strain that expresses the Sleeping Beauty mutagen in cerebellar progenitors. This system leaves a unique genetic 'footprint' at each mutation site, which allows mutated genes to be identified by DNA sequencing. Such mutagenesis experiments are used to identify genes in which mutations frequently arise, because the likelihood of these being involved in tumour development is reasoned to be above average2,3.

As in mouse models of other cancer types2,3, Wu and colleagues' Sleeping Beauty mutagenesis accelerated the development of medulloblastoma in both mouse strains. The authors identified a range of new and established cancer-related genes that had mutations, including some that have previously been implicated in medulloblastoma. They also observed that, following mutagenesis, mice of both strains developed metastases around a type of brain tissue called the leptomeninges, in patterns that are reminiscent of metastatic human medulloblastoma5. The two mouse models thus provided an opportunity to track mutations present in the primary and metastatic disease, and to investigate their genetic provenance.

Wu et al. found that there were, in general, only a few mutations common to primary and metastatic tumours from the same mouse, but that the mutations in different metastases from the same mouse tended to be more similar to each other. Moreover, certain mutations observed in metastases were detected at only low levels within the primary tumour, and some mutations were unique to one or the other tumour type. The authors conclude that their findings are consistent with a model in which metastases originate from rare cells in the primary tumour, and that, following metastasis, additional mutations accumulate independently — both in the primary tumour (post-dispersion events) and in metastases (post-metastasis events) (Fig. 1).

Figure 1: A bi-compartmental genetic model of cancer metastasis.

By analysing the tumours from two strains of mice that model the brain cancer medulloblastoma, Wu et al.1 found differences in the DNA-sequence mutations present in primary and metastatic tumours. They propose that rare cells in the primary tumour that are capable of metastasizing disperse to other sites in the brain, where they form metastases. The cells of the primary tumour and the metastases then continue to accumulate mutations, generating two distinct genetic compartments.

Turning our attention away from mice, an obvious question is whether primary and metastatic tumours in the human disease also show 'bi-compartmental' genetics. Approximately 30% of patients with medulloblastoma already have metastases when they are first diagnosed, and this is associated with a poor prognosis5. However, few previous studies have compared the biology of human primary tumours with their associated metastases, mainly because metastases are not routinely biopsied. Despite the limited sample availability, Wu et al.1 show initial evidence of differing genetics in primary and metastatic tumours from seven human patients.

Further investigation is required to establish whether the authors' findings are broadly relevant to human medulloblastoma. The human disease exhibits5 more complex patterns of metastases than are observed in mice, and is classified into four molecular subgroups (WNT, SHH, Group 3 and Group 4), which each display distinct biological and clinical characteristics6. The Ptch+/− mice used by Wu et al.1 develop SHH-associated medulloblastomas4; similar mutagenesis-driven approaches using existing mouse models of other medulloblastoma disease groups, such as WNT7, might prove informative.

Perhaps the most urgent question arising from this study1 is whether the genetic differences between the two disease compartments lead to distinct biological features that make them respond differently to treatment. In mice, these compartments remain genomically characterized entities, the biological and therapeutic importance of which is untested. In humans, clinical-trial data show5 that primary and metastatic sites respond similarly to current therapies (with cure achieved at both sites) in around 60% of children with metastatic disease, but a more objective assessment of treatment response is confounded by the fact that primary tumours are mostly removed by surgery prior to treatment. Wu et al. provide initial evidence that the tumour compartments may respond differently to current therapies in certain patients, but they rightly caution that these effects could also relate to clinical factors such as radiotherapy being delivered at different intensities to different tumour sites.

Some of the mutations identified by Wu and colleagues' experiments may also reveal biological processes or pathways that could offer drug targets for the improved treatment of primary tumours, metastases, or both. The new mouse strains provide excellent models in which to test this possibility. The multitude and variety of mutations described by Wu et al.1 are noteworthy, but the next challenge is to determine which of them can drive tumour development, which are therapeutically relevant, and which occur at sufficient frequency in the human disease to warrant their pursuit as potential targets. The authors justifiably reason that targets that are common to primary tumours and metastases, in both humans and mice, are those most attractive for further development. However, only one cellular pathway, insulin-dependent signalling, meets these criteria on the basis of their current data.

Providing answers to all these questions will require further biological investigation across species, as well as clinical studies. An additional challenge is posed by the fact that there are fewer than 700 cases of medulloblastoma per year in Europe. More routine biopsy and characterization of human metastases will be essential, and the impetus and ethical justification for such a fundamental change to clinical practice will, at least in part, come from experimental studies such as those presented here. Time will tell whether this tale of Sleeping Beauty and mice develops into a clinically relevant human paradigm.


  1. 1

    Wu, X. et al. Nature (2012).

  2. 2

    Collier, L. S., Carlson, C. M., Ravimohan, S., Dupuy, A. J. & Largaespada, D. A. Nature 436, 272–276 (2005).

    CAS  ADS  Article  Google Scholar 

  3. 3

    Copeland, N. G. & Jenkins, N. A. Nature Rev. Cancer 10, 696–706 (2010).

    CAS  Article  Google Scholar 

  4. 4

    Goodrich, L. V., Milenković, L., Higgins, K. M. & Scott, M. P. Science 277, 1109–1113 (1997).

    CAS  Article  Google Scholar 

  5. 5

    Pizer, B. L. & Clifford, S. C. Br. J. Neurosurg. 23, 364–375 (2009).

    Article  Google Scholar 

  6. 6

    Taylor, M. D. et al. Acta Neuropathol. (2011).

  7. 7

    Gibson, P. et al. Nature 468, 1095–1099 (2010).

    CAS  ADS  Article  Google Scholar 

Download references

Author information



Corresponding author

Correspondence to Steven C. Clifford.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Clifford, S. Evolution after tumour spread. Nature 482, 481–482 (2012).

Download citation


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing