Sequence of Genetic Events in ETV6-RUNX1 Positive B Precursor ALL: Insights from Identical Twins with Concordant Leukaemia

Abstract

Monozygotic twin pairs with concordant ALL have provided unique insights into the molecular pathogenesis and natural history of childhood leukaemia. Data from twin pair studies and neonatal blood spot screening indicate that ETV6-RUNX1 usually arises as an early or initiating pre-natal event. Its consequence appears to be the generation of a clinically silent or covert but persistent pre-leukaemic clone. Conversion to overt, clinical ALL then requires the acquisition of one or more additional genetic lesions that functionally complement ETV6-RUNX1, often including deletions of the non-rearranged ETV6 allele. Recent genome wide single nucleotide polymorphism (SNP) array based studies have revealed considerably more genetic complexity than previously suspected, with ETV6-RUNX1 cases having an average of 6 (range 1–21) genomic losses or gains (Mullighan et al., Nature 2007, 446: 758). It is however unclear from these descriptive screens or audits when these multiple changes arise in relation to the presumed initiating gene fusion and what functional contribution they make. We have used a series of identical twin pairs with ETV6-RUNX1 positive B precursor ALL to test the proposition that, as we reported previously for ETV6 deletion (Maia et al., Blood 2001, 98: 478), all presumed functional or ‘driver’ genomic changes are post-natal in origin and therefore secondary to ETV6-RUNX1 fusion. If this were to be correct then we anticipated that genomic deletions and gains should be different or distinct within each twin pair. We used 250K Sty and 250K Nsp Affymetrix SNP mapping arrays on 5 pairs of identical twins concordant for ETV6-RUNX1 gene fusion positive ALL. We identified copy number variation using the “in-house” Genome Orientated Laboratory File v2.2.9 software package. The SNP array was performed using leukaemic DNA compared to matched remission DNA for 4 out of 5 cases. The fifth case was compared to a pool of remission DNA. The total number of genetic aberrations found was 51 (excluding T cell receptor and immunoglobulin rearrangements): 36 of these lesions were deletions (mean = 7.2) and 15 amplifications. The commonest aberration, found in 8 out of the 10 children, was a deletion on 12p13.2 involving the ETV6 gene. This was discordant in all cases, consistent with our previous reports using microsatellite markers. Other aberrations included deletions of PAX5, CDKN1B, CDKN2A and CD200/BTLA. The status of these, and other, presumed ‘drivers’ of leukaemogenesis were always different when diagnostic DNA of twins, within a pair, were compared i.e. either the genetic change was absent in one but present in the other, or the alteration was present in both but had distinct genomic boundaries. However in 2 of 5 twin pairs concordant, identical lesions were detected. These were idiosyncratic or very rare genomic changes in ALL and were either in gene sparse regions or involved loci with no known or likely contribution to B cell regulation or leukaemogenesis (e.g. CRYGD). We consider the most likely explanation for these shared genetic events in twin cases is that they arise simultaneously with (or immediately prior to) ETV6-RUNX1 fusion, and in the same incipient pre-ALL stem cell, as collateral damage or ‘passenger’ mutations. These data indicate that the common and presumed ‘driver’ genetic changes that accompany ETV6-RUNX1 in ALL are all secondary to gene fusion and most probably post-natal in origin. It remains to be established whether they contribute at all to the sustained pre-leukaemic state and whether they arise independently of each other and sequentially or as a timed suite or bolus perhaps proximate to diagnosis.