In recent years, research has revealed an enormous amount of information about the genetic complexity of many types of cancer, but much more remains to be discovered. It is clear, however, that cancer cells accumulate a wide range of mutations, only some of which "drive" cancer development, and that the cells of a single tumour – and even the subset of cells responsible for tumour propagation – may be genetically diverse.

Mel Greaves of the Institute of Cancer Research, Sutton, UK, and his co-workers chose to investigate the genetic architecture of cellular sub-clones in childhood acute lymphoblastic leukaemia (ALL).

The genetic events involved in the development of the B-cell form of this disease are fairly well known; the fusion of the ETV6 and RUNX1 genes is an initiating event, often occurring before birth. Greaves and his team selected 60 cases of childhood ALL bearing this gene fusion; 30 of these also had deletions of either or both the PAX5 and CDKN2A genes in at least 10% of cells.

In each case, 200 leukaemia cells bearing the founder ETV6-RUNX1 mutation were evaluated separately for these mutations, other deletions and copy number variations, and extra copies of chromosome 21q. The number and relative frequency of subclones (groups of cells bearing the same set of mutations) was recorded in each case.

The results revealed a wide range of complex genetic architectures. The least complex samples contained three genetically distinct subclones, but most samples contained more and some as many as eight. There seemed to be no preferential order for the development of mutations and copy number alterations after the initiating event, and it was clear that copy number alterations in the same gene could develop more than once in the same case.

Any description of the genetic complexity of a tumour is a snapshot obtained at a single time point. Using the same technology, Greaves genotyped ALL cells from one patient at two points seven months apart and observed a marked change in clonal architecture. Cells observed at the first time point, during the pre-malignant phase of the disease, had many fewer CDKN2A deletions than those at the second. Studies of five patients in the PAX5 / CDKN2A deletion subset found that the genetic architecture of ALL also differed between diagnosis and relapse.

These results indicate, although not conclusively, that the cells that are responsible for leukaemia propagation are also genetically heterogeneous and consist of several sub-clones. Greaves and co-workers tested this by transplanting leukaemia cells into immuno-deficient NOD/SCID IL2Rgnull mice, and then re-transplanting some of the expanded populations of leukaemia cells into further identical mice. Regenerated leukaemia cells were then harvested for genetic analysis. Genetically distinct sub-clones were observed in all samples from both primary and secondary transplants, with the same primary sample giving rise to the same sub-clones when transplanted into different mice. The sub-clones also varied in regeneration potential.

Many studies have shown genetic diversity in sub-clones within a single tumour sample, and these results match well with the widely held evolutionary model of cancer development at the cellular level. These results, however, are the first to indicate diversity within the cancer-propagating cells in individual patients. If they are replicated in other leukaemia subtypes and in other cancers, this very diversity may make the development of targeted therapies for cancer even more complex and uncertain.

Reference

Anderson, K., Lutz, C., van Delft, F.W. and 10 others (2010) Genetic variegation of clonal architecture and propagating cells in leukaemia Nature, published online ahead of print 15 December 2010 doi:10.1038/nature09650