Clinical significance of preleukemic somatic GATA1 mutations in children with Down syndrome

ABSTRACT: Children with Down syndrome (DS) have a high risk of GATA1-associated myeloid leukemia (ML-DS) before age 4 years. Somatic N-terminal GATA1 mutations (GATA1s) are necessary, but not sufficient, for ML-DS, but their significance at birth for individual babies and whether mutations occur after birth is unclear. To address these questions, we performed a prospective study of newborns with DS using next-generation sequencing-based GATA1 mutation analysis, with hematologic and clinical evaluation and follow-up for the window of ML-DS risk. Of 450 neonates with DS, 113 (25%) had GATA1s mutations, among whom 20/113 (17.7%) had multiple mutations and 59 (52%) were clinically silent. Variant allele frequency (VAF) varied from 0.3% to 89%. VAF positively correlated (P < .0001) with the percent blasts, leukocytes, dyserythropoiesis and dysmegakaryopoiesis scores, and clinical disease severity, and negatively with hemoglobin, although only 4/113 were anemic. GATA1s mutations were detected from 28 weeks gestation; the highest frequency (45%) was at 34 to 35 weeks, whereas mutation frequency in early fetal samples (<20 weeks) was <4% (2/57). GATA1s clones (VAF, percent blasts) fell rapidly postnatally, becoming undetectable by 6 months, except in neonates who developed ML-DS. Of 110 surviving neonates, 7 (6.4%) developed ML-DS at a median age of 17.5 months. GATA1s clone size at birth was the only predictor of ML-DS. No neonates lacking GATA1s mutations acquired mutations after birth or developed ML-DS. Taken together, the fetal environment is essential for GATA1s mutation selection and expansion of GATA1s clones. Rates of leukemic transformation of GATA1s clones detected at birth are low, but clones that persist >6 months transformed.