CANCER AND CHIMERIC PROTEIN IN CHILDREN
Three-quarters of children with ALL have a translocation that generates a chimeric protein from the ETV6 and RUNX1 genes (Fig. 1). However, each child is unique in the location of the breakpoints giving rise to the chimeric gene, revealing that the break has occurred in one cell in the fetus that has then multiplied. Although the incidence of ETV6-RUNX1 fusions is ~1%, only about 0.01% of children get leukaemia – clearly indicating that the fusion is a ‘promoter’ event following which further disruptions are required (deletion of ETV6 is common as a second event to ETV6-RUNX1 fusion). This is consistent with the disease having a protracted latency (that is, it may appear at any time up to about 14 years of age) and that transgenic mice with ETV6-RUNX1 or RUNX1-RUNX1TI don’t get leukaemias.
1. The ETV6-RUNX1 translocation in childhood ALL
Analysis of the mutational pattern in individual cells reveals that they differ- in other words: that a tumour is made up of different clones, albeit that has all been driven by the same initiating event (ETV6-RUNX1 fusion). These ‘genetic snapshots’ reveal extraordinary diversity in mutational patterns: for example, a clone may acquire an extra copy of a gene and then subsequently lose it (Fig. 2). This picture of genomic plasticity can only be visualised by studying single tumour cells, to which the leukaemias particularly lend themselves. The approach has revealed what Mel Greaves has described as the ‘clonal architecture’ of tumours whereby individual clones are present in different proportions, each having its own capacity for self-renewal. In other words, cancer development is not linear but branching and that there is no such thing as a ‘leukaemia stem cell’ – merely diverse populations of cancer propagating cells.
2. Differing proportions of genetically distinct cells from a patient with leukaemia caused by the ETV6-RUNX1 translocation.
These conclusions are consistent with several examples of monozygous twins, where both carry the same initiating mutation but only one has developed ALL (i.e. they are discordant). The normal twin has the initiating fusion gene (ETV6-RUNX1) but of the copy number changes in the afflicted child. As happens in the game of genetic roulette, the unaffected twin has failed to acquire the mutation(s) necessary to render the ‘driver’ effect fully tumourigenic.
It is technically difficult to apply this method of single cell analysis to solid tumours but FISH has shown that, within a malignant brain tumour, combinations of three different RTKs (EGFR, MET, PDGFRA) may be amplified in patterns that vary between individual cells in the same tumour. For example, EGFR and MET amplification in the same cell or of cells with mutually exclusive EGFR or MET amplification. Despite the heterogeneity of a tumour, subclones with specific RTK amplification(s) shared an early mutational event (e.g. CDKN2A deletion and/or TP53 (aka P53) mutation), indicating their descent from a common precursor.
The alternative approach of ‘single nucleus sequencing’ utilises DNA amplification followed by next-generation sequencing and is sufficiently sensitive to assess copy number alterations in the genomes of single cells. Application of this method to cells from different regions of a primary breast tumour revealed that growth of a single clone had led to the release of one cell that seeded a metastasis and that, in terms of copy number alterations, the metastasis had not subsequently undergone significant evolution.
Broadly speaking, molecular analysis is beginning to confirm that tumours are indeed heterogeneous, that is they are comprised of multiple clones. The independent modulation of mutational profiles and the consequent diverse proliferation rates may permit individual clones to emerge as dominant. In other words, as we might have suspected all along, cancers are a form of dynamic Darwinism.