Silent TAM keeps quiet Free

In this issue of Blood, Elliott et al¹ describe the results of a longitudinal study to monitor the natural history of transient abnormal myelopoiesis (TAM) in children with Down syndrome.

TAM is a myeloproliferative disorder that occurs exclusively in newborn children with Down syndrome. Children with TAM exhibit elevated white blood cell counts, abnormal megakaryopoiesis, and abnormal erythropoiesis. Affected children can experience liver injury as a consequence of leukocyte infiltration, but the disease is usually self-limited.^2,3 TAM initiation requires simultaneous trisomy 21 and a somatic mutation of the GATA1 gene, such that it encodes an N-terminally truncated, “short” protein (GATA1s).^4,5 TAM originates primarily, if not exclusively, from fetal hematopoietic progenitors.^6,7,GATA1s cooperates with microRNAs encoded on chromosome 21, including mir-99a, mir-125b, and mir-155, to reprogram fetal progenitors and enable clonal expansion.^6,8 Over time, most TAM clones lose self-renewal capacity and disappear, perhaps as they transition from fetal to postnatal or adult transcriptional states. However, ∼20% of patients with TAM acquire cooperating mutations that lead to fully transformed myeloid leukemia associated with Down syndrome (ML-DS).³ This progression risk necessitates longitudinal screening of children with TAM, even after resolution of the initial leukocytosis.

Screening strategies for TAM and ML-DS have been complicated by recent molecular data showing that TAM is far more common, based on GATA1s profiling, than was previously appreciated based on clinical observations alone.^9,10 For example, a study by Roberts et al, published in 2013, used next-generation sequencing (NGS) to screen a cohort of newborn children with Down syndrome for presence of GATA1s mutations.⁹ Among patients who underwent NGS screening, ∼15% had clinically overt TAM, all with an associated GATA1s mutation. Of the remaining children who did not have clinically overt TAM, ∼20% still had a detectable GATA1s mutation. These children were deemed to have “silent TAM,” raising questions as to whether silent TAM adheres to the same natural history as clinically overt TAM and whether there is a role for longitudinal GATA1s surveillance. In other words, in lieu of clinical or morphologic evidence of TAM, should we screen patients longitudinally by NGS to ensure that GATA1s clones disappear? Furthermore, to what extent, and at what age, does persistence of a GATA1s clone predict eventual progression to ML-DS?

To address these questions, Elliott et al evaluated 450 newborn patients with Down syndrome for GATA1s mutations and then periodically monitored GATA1s variant allele frequencies for up to 4 years.¹ The goal was to resolve the natural history of silent TAM and inform monitoring strategies for all children with a detectable GATA1s at birth. The study incorporated assessments of several clinical parameters, including blood counts, presence/absence of congenital anomalies, birth complications, and progression to ML-DS. Approximately 25% of the patients had a detectable GATA1s mutation at birth. Of these, approximately half of the patients met the criteria for clinically overt TAM (blasts >10%), whereas the other half met the criteria for silent TAM (blasts <10%). There were no associations between TAM (overt or silent) and congenital anomalies. GATA1s variant allele frequencies correlated with blast percentages and abnormal blood morphology (eg, nucleated red blood cells). The most striking findings, however, related to the prognostic relationship between GATA1s allele frequencies at birth, blast percentages at birth, and progression to ML-DS.

The study makes 4 key points that will shape our understanding of TAM and its evolution to ML-DS. First, the absence of a GATA1s clone at birth (with sensitivity down to 0.3%) strongly indicates that a child with Down syndrome is not at risk for progression to ML-DS, because none of the 325 patients who met this criterion in this study progressed. In addition to the clinical implications, this observation has a biological implication in that it reinforces the concept that clinically relevant GATA1s clones are fetal derived. Second, failure to clear a GATA1s mutation by the age of 6 months is highly prognostic for progression to ML-DS, because all patients who met this criterion ultimately developed ML-DS. The converse is not true; patients who have a detectable GATA1s at birth who clear it at age 6 months still remain at risk for ML-DS. Third, almost all patients who develop ML-DS will have had overt TAM at birth. In this study, only 1 patient with silent TAM ultimately developed ML-DS. This implies that blast count at birth offers a robust and cost-efficient measurement for predicting ML-DS risk, as opposed to more costly mutation profiling. Fourth, progression to ML-DS is preceded by falling platelet counts, again affording a cost-efficient screening method.

These findings simplify the approach to ML-DS monitoring for children with silent and overt TAM. Silent TAM carries only minimal risk for progression to ML-DS and thus does not require molecular monitoring. Clinically overt TAM carries higher risk for transformation but falling platelet counts offer an adequate proxy for impending ML-DS progression. Thus, GATA1s measurements at the age of 6 months may help physicians and families anticipate eventual transformation, but in countries with only limited access to NGS, platelet count monitoring should suffice.

These data do raise the question of why some GATA1s clones disappear with age while others persist for an extended period of time before ML-DS progression. One possibility is that persistent GATA1s clones have already acquired one or more secondary mutations that will ultimately lead to ML-DS. The authors tested for secondary mutations in several infants who ultimately developed ML-DS, but only at birth. Cooperating mutations were not evident at that early time point, but it is possible that they emerge by 6 months after birth to sustain the GATA1s mutant clone. Expanded mutation profiling in children with a persistent GATA1s clone at 6 months of age may ultimately uncover mechanisms that sustain these clones before transformation.

Conflict-of-interest disclosure: J.A.M. declares no competing financial interests.