Targeted Deep Sequencing of GATA1 more Sensitively Identifies Transient Myeloproliferative Disorder in Neonates with Down Syndrome

Down syndrome (DS) is caused by a constitutional trisomy 21 and up to 10% of neonates with DS develop a temporary accumulation of leukemic blasts in the peripheral blood, called transient myeloproliferative disorder (TMD). In most cases, TMD is a self-limiting disorder, however 20% of children develop serious symptoms and treatment with low dose cytarabine is then necessary. TMD is caused by a mutation in the GATA1 gene, occurring in cells with trisomy 21. Around 20% of children with TMD will develop a myeloid leukemia (ML) before the age of 4 years. It is a megakaryoblastic leukemia characterized by a GATA1 mutation identical to the mutation these children had in their TMD blast cells. The primary aim of this study was to define a population based frequency of TMD in DS neonates.

In this nation-wide study we prospectively screened 368 peripheral blood (PB) samples from DS neonates (< 4 weeks old) in the Netherlands between January 1^st 2008 and January 1^st 2013 for TMD using morphology and immunophenotyping (IPT) (Dutch Trial Registry number NTR1667). TMD was defined as >5% myeloblasts in the PB. Patients with less than 5% of cells in the PB consistent with TMD using IPT were also included. PB samples of TMD positive patients were tested for GATA1 exon 2 mutations using Sanger sequencing. Patients with TMD were followed at least until the age of 4 years. TMD positive patients without a GATA1 exon 2 mutation using Sanger sequencing and TMD negative patients were retrospectively screened for GATA1 exon 2 and 3 mutations using targeted deep sequencing. Using this technique we could detect mutations with a frequency of 0.1%.

In the prospective part of the study, TMD was diagnosed in 45 of the 368 cases screened (12.2%). In 41 of 45 neonates with TMD, GATA1 mutation analysis was performed. In 4 cases no GATA1 PCR was performed. In 28 cases a GATA1 exon 2 mutation was detected (28/41=68%), while in 13 cases no GATA1 mutation was detected.

It was recently shown that there is a subgroup of DS neonates who do not have TMD by conventional criteria, but when more sensitively screened do harbour a GATA1 mutation (Roberts et al, Blood 2013). Therefore using targeted deep sequencing we retrospectively screened the 13 TMD cases without a GATA1 mutation by Sanger sequencing and 2 of the 4 cases that were not screened at TMD diagnosis. No GATA1 exon 3 mutations were detected. In 9 out of 15 patients a GATA1 mutation was detected using targeted deep sequencing. Both cases not prospectively Sanger sequenced were positive. 7 of 13 TMD samples tested negative using Sanger sequencing, using targeted deep sequencing were positive for a GATA1 mutation (median mutational frequency 6,8%). Overall, 43 TMD patient samples were screened using Sanger sequencing and/or targeted deep sequencing and 38 of these tested GATA1 positive (88%).

In addition, we retrospectively screened 319 of the 323 peripheral blood samples of the TMD negative DS neonates for GATA1 using targeted deep sequencing. No GATA1 exon 3 mutations were detected. In 11 out of 319 TMD negative cases we detected a GATA1 exon 2 mutation with a median mutational frequency of 0,8%.

14 children with DS born in the Netherlands between 1-1-2008 and 31-12-2013 were diagnosed with ML DS (follow-up until 31st of dec 2016). Of these 14 children, 7 were screened neonatally and in 6 cases TMD was diagnosed. Therefore 6 out of 45 TMD patients identified using conventional screening developed ML DS (13.3%). Out of the 11 'subclinical' TMD cases only detected by targeted deep sequencing GATA1 screening, one patient was later diagnosed with ML DS (1/11=9.1%). The patient's neonatal blood sample harboured two different GATA1 exon 2 mutations (14 base pair (bp) insertion in 0,24% of cells and a 1 bp insertion in 2,6% of cells). In the ML DS cells of this patient the identical 1 bp exon 2 insertion in GATA1 was detected.

In conclusion, conventional screening for TMD using immunophenotyping en morphology identified TMD in 12% of the screened DS neonates. An additional 3.4% can be detected using targeted deep sequencing. This is clinically relevant as these children are at risk of developing ML DS. Therefore, for future clinical studies we are developing targeted deep sequencing methods screening for GATA 1 in neonatal DS PB samples.