Evolutionary Trajectory of Hyperdiploid ALL: Insights From Monozygotic Twins

Abstract 28

Identical twins have provided substantial insight into the natural history, developmental timing and sequence of genetic events in ALL. Much of these data are focussed on ALL with ETV6-RUNX1 fusions. More limited data exists for the most prevalent subset of childhood ALL with hyperdiploidy though retrospective screening from cord blood (Maia et al. Genes Chromosomes and Cancer, 2004) provide evidence that hyperdiploidy itself can be an early, possibly initiating and pre-natal event.

We have now explored the genetic complexity of hyperdiploid ALL using four pairs of monozygotic twins, three concordant and one discordant for ALL. In parallel with our prior studies on ETV6-RUNX1 ALL (Bateman et al. Blood, 2010), we hypothesized that hyperdiploidy itself would be a consistent pre-natal event and all other potential ‘driver’ mutations would be secondary and probably post-natal. If correct, this twin pair should share identical hyperdiploid karyotypes but have different copy number alterations (CNA) or other secondary mutations (e.g. in FLT3 or RAS).

We used Affymetrix SNP 6.0 mapping arrays to identify CNA using the Genome Orientated Laboratory File (GOLF) v3.2 software package and dCHIP. The SNP array was performed using leukaemic DNA compared to matched remission DNA for all cases. The mutation screen included the activating D835 FLT3, NRAS exon 2 (codons 12 and 13) exon 3 (codon 61), KRAS exon 2 (codons 12 and 13) and exon 3 (codon 61), PTPN11 (exon 3 and exon 13). In the discordant twin pair, Fluorescence In-Situ Hybridisation (FISH) was used to interrogate the pre-leukaemic cells of the unaffected twin for the CNA found in the corresponding affected twin.

In the identical twin pairs concordant for hyperdiploid ALL, the whole chromosome gains were identical within each twin pair apart from an additional but sub-clonal gain of chromosome 10 in one twin of twin pair set 2. The median number of whole chromosome gains was 56 (Range 55 to 59). The total number of CNA found was 6 (Range 0 to 4 per case) (excluding T cell receptor and immunoglobulin rearrangements). These included genes previously found to be recurrently altered in hyperdiploid ALL such as PAX5, CDKN2A/2B and ADD3. All the CNA abnormalities detected were different within each twin pair. The mutation screen revealed an NRAS exon 2 codon 13, GGT to GAT (glycine to aspartic acid) mutation in one twin of twin pair set 1, which was absent in leukaemic DNA of the co-twin.

In the twin set discordant for ALL, the CNA at 19p13.3 (incl TCF3 gene) found in the leukaemic DNA of twin 1 were absent from the hyperdiploid pre-leukaemic cells of the unaffected twin 2. These pre-leukaemic cells were found at a consistent frequency between 0.0012% and 0.002% (over a period of 13 months).

These data accord with the view that hyperdiploidy is a pre-natal and probably initiating event in childhood B cell precursor ALL and that the set of additional ‘driver’ genetic changes including CNA and sequence based mutations are all secondary. As all of the latter are single twin restricted, it is likely, though not formally proven, that they occur post-natally and perhaps more proximal to disease diagnosis. These data extend our understanding of the history of B cell precursor ALL. Whole genome sequencing is required to complete the picture of essential genetic events and their timing.