All cancers progress via an evolutionary or Darwinian process of genetic diversification, ecological (or therapeutic) pressure and selection. A major consequence of this developmental process is that cancers are anticipated to have complex and dynamic sub-clonal, genetic architectures. The implications of this for biopsy-based prognosis and for targeted therapy are considerable.

We have explored clonal architecture in childhood acute lymphoblastic leukaemia (ALL). Using multiplexed FISH approach, we previously showed (Anderson K et al, Nature 2011, 469: 356-361) that each case had a unique, branching clonal architecture with ETV6-RUNX1 mutation at the base of the evolutionary tree. Secondary mutations (as copy number alterations (CNA)) occurred reiteratively within sub-clones of individual patients. Sub-clonal architecture, as anticipated, changed between different phases of disease, pre-leukaemic phase, primary ALL and relapse. Significantly, cancer stem cells in ALL, assayed by serial transplantation in mice mirrored clonal complexity in their variegated genetics.

Ideally, what is required to decisively interrogate clonal complexity is a high throughput, single cell method that can simultaneously detect any kind of ‘driver' mutational change including gene fusions, CNA and bp sequence alterations. A series (#58) of cases of ETV6-RUNX1-positive ALL samples have been whole genome sequenced. An average of nine non-synonymous base pair changes were identified per case, some but not all in genes known to be recurrently mutated in ALL and/or other cancers. Allele burden derived from proportion of positive sequence reads suggested many mutations were sub-clonal.

Several of these cases have now been evaluated by a new micro-fluidic PCR-based technique (Fluidigm, BioMark). This allows simultaneous detection of gene fusion sequences, CNA and bp sequence changes in hundreds of single cells. The data confirms the existence of genetic variegation of sub-clones and phylogenetic analysis, by maximum parsimony, allowed the computation of detailed clonal architecture and evolutionary trees.