The natural history of transient abnormal myelopoiesis in neonates and children with Down syndrome

Abstract

Children with Down syndrome (DS) have a markedly increased risk of acute myeloid leukaemia (ML-DS) during the first 5 years of life compared to children without DS. Many children with ML-DS have a preceding history of a neonatal pre-leukaemic disorder known as Transient Abnormal Myelopoiesis (TAM). Both TAM and ML-DS harbour the same acquired N-terminal GATA1 mutation(s) indicating they are clonally-linked conditions. Currently, there is a lack of clear clinical, haematological and molecular diagnostic criteria for TAM. My project aimed to (i) precisely define TAM (ii) document the natural history of TAM and identify the population at risk of developing ML-DS and (iii) characterise blast cells in TAM in order to establish whether particular blast cell sub-populations might be associated with a higher frequency of subsequent transformation to ML-DS and whether it is possible to distinguish between blast cells which do and do not carry GATA1 mutations. To address these aims I analysed data from 382 DS neonates recruited to the Oxford Imperial Down Syndrome Cohort Study (OIDSCS); a prospective, population-based study to systematically examine clinical findings, haematological indices, blood cell morphology and GATA1 mutation status in neonates with DS. Using standard methods (PCR of GATA1 exons 2/3 followed by dHPLC and direct sequencing) GATA1 mutation(s) were detected in 45/382 neonates (11.8%). Of these 35/45 neonates (77.8%) had clinical and/or haematological evidence of TAM (blasts >10%) and 10/45 (22.2%) were clinically and haematologically 'silent'. Neonates with clinically and haematologically silent TAM could not be distinguished from DS neonates where no GATA1 mutation(s) were detectable by standard methods. Using next-generation-sequencing (NGS) an additional 27/337 (8.0%) DS neonates had one or more small mutant GATA1 clones. These 27 neonates were clinically, haematologically and molecularly 'silent' and would be missed unless highly sensitive methodology is employed. To investigate the natural history of the evolution of TAM and both types of silent TAM to ML-DS, I studied all of the DS children in the cohort who have progressed to ML-DS (n=5). Four of these children had overt TAM and one had clinically and haematologically silent TAM. Compared to DS children who did not progress to ML-DS, these 5 children were more likely to have persistent hepatomegaly, thrombocytopenia and additional cytogenetic abnormalities at time of disease progression.

I studied the immunophenotypic profile of blasts in TAM and was unable to identify a TAM-specific immunophenotypic population which DS neonates without GATA1 mutation(s) did not have. However, in neonates with GATA1 mutation(s) the CD45dimCD34+/-CD33+/-CD36+CD7+ blast sub-population was significantly expanded as compared to neonates without GATA1 mutation(s). These studies describe for the first time the full range of clinical and haematological features associated with a mutant GATA1 clone in DS neonates. As there were no reliable clinical, haematological or biological features (such as blast cell properties) which can identify neonates with GATA1 mutation, TAM is accurately defined as a neonate with DS and the presence of one or more GATA1 mutation(s) determined by a sensitive screening method such as NGS. This definition of TAM would identify all neonates at risk of transformation to ML-DS... Follow up of the whole cohort until the age of 5 years will be necessary to fully evaluate the significance of very small mutant GATA1 clones. A precise diagnosis of TAM may help plan treatment to better manage TAM and prevent ML-DS by eradicating mutant GATA1 clones.