Two Novel GATA1 Mutations in Transient Abnormal Myelopoiesis of Thai Neonates with Down Syndrome

Abstract

Children with Down syndrome (DS) are 150 times more likely to develop acute myeloid leukemia (ML-DS), compared with those without. One risk factor is transient abnormal myelopoiesis (TAM). Somatic truncating GATA1 mutations are found in most TAM patients and are markers for future ML-DS. We identified two novel frameshift mutations in our seven newborns with DS and TAM: a heterozygous mutation of 17 nucleotide duplication (c.154_170 dup) and a heterozygous 9-nucleotide deletion combined with a 2-nucleotide insertion (c.150_158delins CT). Both mutations introduced a truncated GATA1 protein. Thus, neonates with DS and TAM require frequent ML-DS monitoring.

Introduction

Transient abnormal myelopoiesis (TAM) or transient leukemia of Down syndrome (TL-DS) is a unique condition, found in 5 to 10% of neonates with DS. ^{1 2} It is a preleukemic stage, presenting with an increase in peripheral blood myeloblasts. It has been challenging to develop stable TAM diagnosis criteria, especially without a definitive cutoff percentage for blast cells. ^{3 4 5} Clinical manifestations, such as hepatomegaly, splenomegaly, jaundice, pericardial effusion, pleural effusion, respiratory distress, and bleeding diathesis, are nonspecific and not always present ⁶ ; in addition, most neonates with TAM will have spontaneous resolution without treatment. However, the mortality rate for neonates with TAM is reported to be 20%. ^{3 7}

TAM is classified as either classic (clinical) or silent type. Classic type is associated with blast cells > 10% in the peripheral blood smear. Silent type is characterized by blasts < 10% but with patients having carrying a GATA1 mutation. ⁴ Both types are at risk of developing acute myeloid leukemia of DS (ML-DS), ^{2 4 8} i.e., acute megakaryoblastic leukemia or myelodysplastic syndrome. This risk is 150 times greater in children with DS versus those without, ⁹ and previous studies found neonates with DS and TAM have a 12 to 20% risk of developing ML-DS in the first 2 to 4 years of life. ^{3 4 7 10 11}

How exactly ML-DS develops is not yet clearly known, but the “three hits theory” is proposed. The first hit is having trisomy 21, which leads to a greater number of megakaryocyte–erythroid progenitors in the fetal liver. Although megakaryocytes increase, differentiation is impaired, causing dysmegakaryopoiesis. The second hit is carrying an acquired GATA1 mutation, leading to an abnormal proliferation of myeloid cells, with a gain of function mechanism. The third hit is epigenetic change, possibly arising with transcription regulation pathway mutations. ^{2 6 8 12 13 14} Thus, it is believed the acquired GATA1 mutation in TAM plays a critical role in ML-DS development.

The GATA-binding protein 1 gene ( GATA1 ) has 6 exons, and is located on the X chromosome (Xp11.23). It encodes the hematopoietic transcription factor, GATA1 protein, with 413 amino acids. Its function is differentiation and proliferation of erythroid cells and megakaryocytic lineages. ¹³ The majority of GATA1 mutations occur in exon 2 and 3, predicting a stop codon which leads to a truncation of the N-terminal GATA1 protein (GATA1s), which lacks the amino-transactivation domain. ^{11 13} GATA1 mutations in people without trisomy 21 has not been associated with leukemia but are associated with other diseases, such as Diamond–Blackfan anemia ¹⁵ and X-linked dyserythropoietic anemia with thrombocytopenia. ¹⁶ Associations between GATA1 mutations in TAM and ML-DS have been reported since 2002. ¹³ GATA1 mutations have been identified in 75 to 97% of neonates with TAM. ^{4 6 11 17} Sanger sequencing having a lower detection rate than next generation sequencing. Mutations disappear when TAM is resolved, which further compounds detection. ¹³

Obviously, the early diagnosis of ML-DS permits patients to receive appropriate management. However, ML-DS monitoring in children with DS is contingent on perceived ML-DS risk. As neonates with TAM have a higher risk of developing ML-DS, they require complete blood count review and peripheral blood smears every 3 months for the first 2 to 4 years of life, instead of the recommended yearly screening for children with DS. Nevertheless, a sole complete blood count may not be adequate as lesser blasts are apparent in silent TAM. Surprisingly, neonates with blasts > 10%, but without GATA1 mutation, appear unlikely to develop ML-DS. ^{4 6 11 17}

Previously, GATA1 mutation testing was not in the clinical guidelines for neonates with DS. More recently, the British 2018 guideline ¹⁸ recommended performing GATA1 mutation analysis for diagnosis confirmation and to aid in risk prediction for ML-DS. Therefore, we studied whether any GATA1 mutations are present in Thai newborns with both DS and clinical TAM, and if they may be at increased risk of developing ML-DS. We also reviewed the management and clinical course of these individuals.

Materials and Methods

All Thai neonates with DS with clinical TAM were enrolled: our definition of clinical TAM was blasts > 10% in the peripheral blood smear. This study was done at Thammasat University Hospital, from June 2017 to October 2018 and approved by the Human Research Ethics Committee of Thammasat University No 1 (Faculty of Medicine; MTU-EC-SA-6–070/60).

Clinical history and management for each participant was reviewed, and 3 mL of blood was collected with genomic DNA isolated from peripheral leukocytes by Gentra Puregene blood kit (Qiagen, Germantown, Maryland, United States). The oligonucleotides primers were designed using software Primer3Plus ( http://www.bioinformatics.nl/cgi-bin/primer3plus/primer3plus.cgi ). The forward and reverse primer sequences for polymerase chain reaction (PCR) amplification are shown in Supplementary Table S1 (available in the online version). PCR for each exon of GATA1 was performed, starting with denaturation at 95°C for 3 minutes. Then there were 35 cycles of denaturation at 95°C for 30 seconds, annealing at 58°C for 30 seconds, and elongating at 72°C for 30 seconds. This is followed by extended elongation at 72°C of 10 minutes. The PCR products were visualized by electrophoresis on 2% agarose gel. The products were then sequenced by capillary electrophoresis (Applied Biosystems ABI 3730; Massachusetts, United States). Sequence analysis was performed by using the National Center for Biotechnology Information reference sequence, NM_002049. All variants and predicted amino acid changes were categorized according to the nomenclature of the Human Genome Variation Society (HGVS).

Flow cytometry studies were performed, according to a predefined study protocol. Specimens were stained by incubation for 15 minutes in the dark with the following fluorescence conjugated monoclonal antibodies: CD2, CD3, CD5, CD7, CD10, CD11b, CD11c, CD13, CD14, CD15, CD19, CD22, CD33, CD34, CD56, CD64, CD117, HLA-DR, MPO, and TdT. Flow cytometric acquisition was done using a BC Cytomics FC500 (Beckman-Coulter, Brea, California, United States) which was equipped for 5-color analysis. Data analysis was performed using CXP software (Beckman-Coulter).

Clinical Presentation

Patient 1

A male newborn was transferred to the neonatal intensive care unit (NICU) for respiratory distress due to early neonatal sepsis 1 hour after birth. He had atrial septal defect (ASD), ventricular septal defect (VSD), patent ductus arteriosus (PDA), mild coarctation of aorta, and pulmonary hypertension (PHT). His bone marrow study displayed large blast cells, fine nuclear chromatin, and a deep blue cytoplasm with knobs: erythroblasts or megakaryoblasts suspected. However, bone marrow flow cytometry was unable to identify any subpopulations of erythroblasts or megakaryoblasts due to insufficient markers ( Supplementary Fig. S1 , available in the online version). Furthermore, immunohistochemical staining of clot marrow tissue was unsatisfactory.

He was treated for persistence of thrombocytopenia and bleeding diathesis. His abnormal myelopoiesis was resolved, but he remained hospitalized for Klebsiella pneumonia and Escherichia coli septicemia, congestive heart failure, pulmonary hypertension, severe tracheomalacia, and hypocalcemic seizure until he was 6 months old, after all comorbid conditions were controlled.

Patient 2

A female newborn was transferred to the NICU at 15 minutes of life for cyanosis, due to moderate to severe PHT. Her bone marrow flow cytometry was interpreted as abnormal myeloid proliferation ( Supplementary Fig. S1 , available in the online version). She received treatment for persistence of thrombocytopenia and was discharged at 6 weeks after clipping of the ductus arteriosus; her idiopathic neonatal hepatitis resolved as well.

Patient 3

A male newborn received neonatal resuscitation after birth and was then transferred to the NICU for respiratory distress. His initial vital signs were unstable, and further investigations revealed acute liver failure, PDA 3.5 mm with severe PHT, acute kidney injury with left ureteropelvic junction obstruction, and coagulopathy. The treatment included total blood exchange transfusion and chemotherapy; however, he progressed to multiple organ failure and passed away.

Patient 4

A male newborn was transferred to the NICU due to respiratory distress and desaturation, caused by pericardial effusion. He also presented with petechial rash, and a left congenital cataract. His clinical signs were transient and spontaneously resolved. His pericardial effusion was negative for all viral studies; he received treatment by corticosteroids with resolution 1 week later.

Patient 5

A female newborn was transferred to the NICU at 15 days old with complicated congenital pneumonia and congestive heart failure from 6 mm of perimembranous VSD. She also had neonatal jaundice and acute kidney injury. She was treated because of the persistence of thrombocytopenia and required platelet transfusion. She was followed up with regularly until 14 months old. She developed persistent bicytopenia (anemia and thrombocytopenia), but her bone marrow study revealed no evidence of leukemia.

Patient 6

A male newborn was transferred to the NICU due to respiratory distress at 30 minutes after birth. He had a heart murmur with abdominal distension. Further investigations revealed ASD, VSD, PDA with moderate PHT, hypothyroid, and duodenal atresia. His circulating blast cells disappeared spontaneously without any treatment and no bone marrow flow study was performed.

Patient 7

A female newborn was transferred to the NICU due to persistent cyanosis since birth, which was caused by hyperviscosity from hyperleukocytosis. Further investigations revealed atrioventricular septal defect, tetralogy of Fallot with pulmonic stenosis and moderate PHT. She passed away due to multidrug-resistant Acinetobacter baumannii septicemia.

Results

All seven neonates were diagnosed as having clinical TAM; their DS was genetically confirmed by the karyotype of trisomy 21. Five of seven patients went into remission, but two (patients 3 and 7) succumbed. Five out of seven patients received cytosine-arabinoside (ara-C) chemotherapy; however, only three patients received complete therapy for 5 days with the other two (patients 3 and 7) receiving it for only 1 and 3 days, respectively, before passing away. One (patient 5) developed myelodysplastic syndrome at the age of 14 months. Patient clinical course summary is shown in Table 1 .

Table 1. Newborns with Down syndrome and transient abnormal myelopoiesis: clinical data.

	Patient number
	1	2	3	4	5	6	7
Sex	Male	Female	Male	Male	Female	Male	Female
Birth (wk)	Term	Preterm (36)	Preterm (36)	Term (38)	Preterm (35)	Preterm (36)	Preterm (35)
Maternal age (y)	34	28	32	34	30	33	33
Birthweight (kg; %tile)	3.2 (P10–50),	2.1 (P10–25)	2.6 (P10–50)	2.6 (P10–50)	2.5 (P25–50)	1.6 (<P3)	2.5 (P50)
Length (cm; %tile)	52 (P50–90)	43 (P10)	43 (P3)	45 (P3)	48 (P50)	44 (P3–10)	44 (P10–50)
HC (cm; %tile)	32 (P3)	28.5 (<P10)	30.5 (P3–10)	31 (P3)	30 (P3)	31 (P10)	31 (P10–50)
Apgar's score (at 1 and 5 min)	9, 10	8, 10	5, 8	8, 8		8, 10	5, 5
Complications (prenatal)	None	None	Polyhydramnios, hepatosplenomegaly and left pyelectasis	Thick meconium-stained amniotic fluid	None	None	Gestational DM
Age of presentation (d)	3	1	1	1	28	1	1
Hepatosplenomegaly	Present	Present	Present	Present	Present	Present	Present
Hb (g/dL)/Hct (%)	18/58	17/52	10/36	18/57	9.7/28	17/55	11/37
WBC (mm 3 )	17,300	40,200	259,900	16,900	8,100	64,500	141,300
Blasts (%)	28	24	79	24	19	65	93
Platelets (mm 3 )	52,000	49,000	774,090	63,000	32,000	225,000	202,000
Myeloblasts in BM (%)	15	20	NA	40	41	NA	NA
Immunophenotyping by flow cytometry for abnormal myelopoiesis	NA	CD34, CD45, and CD117	CD45, CD34, MPO, cCD33, CD117 and CD38	CD34, MPO, cCD33, CD117, and CD38	CD34, CD7, cCD33 and CD117	NA	CD45 (weak), CD34, cCD33, CD38, CD117, CD7 and CD56
Ara-C treatment	D12–16	D5–9	D7	No	D40–44	No	D3–5
Outcome (d)	Remission	Remission	Deceased	Remission	Remission	Remission	Deceased
	30	37	8	5	50 (developed MPS at 14 mo)	30	6
GATA1 mutation	Positive	Negative	Negative	Negative	Negative	Negative	Positive
Blood collection for GATA1 analysis (d)	5	14	2	14	420	2	1

Abbreviations: %tile, percentile; Ara-C, cytarabine; Hb, hemoglobin; Hct, hematocrit; MB, bone marrow; MPS, myelodysplastic syndrome; NA, not applicable; WBC, white blood cells.

Two heterozygous frameshift mutations in exon 2 of GATA1 were identified in patients 1 and 7 (as shown in Fig. 1 ). The first mutation was 17 nucleotides duplicated from position 154 to 170 (c.154_170 dup) which predicted the amino acid change at codon 58 from alanine to glutamine, and introduced a shift of the following 84 amino acids before a stop codon (p.Ala58Glnfs*85). Another mutation was deletion of nine nucleotides combined with insertion of two nucleotides from position 150 to 158 (c.150_158delins CT), which predicted the amino acid change at codon 52 from threonine to proline and introduced a shift of the following 82 amino acids before a stop codon (p.Thr52Profs*83). Both led to truncated GATA1 protein (GATA1s). These frameshift mutations were novel but predictable pathogenic mutations. The other 5 patients were not identified as having any mutations in GATA1 exon 2 to 6. The present and previous mutations are shown in Supplementary Fig. S2 (available in the online version).

Fig. 1.

DNA sequence of patient 1 (top pane) showing mutation, c.154_170 dup, patient 7 (middle panel) showing mutation, c.150_158delins CT, and bottom panel showing reference sequence of this segment in exon 2 of GATA1. Black boxes indicate reference nucleotides, grey boxes indicate the extra (duplicated or inserted) sequences, and dotted grey box indicates deleted sequence.

Discussion

Our study is the first report on the GATA1 mutations in neonates with TAM in Thailand. We sequenced all translation exons in GATA1 instead of only exon 2 and 3. This was to specifically confirm all possible mutations, as some previous studies revealed a higher detection rate when exons 2 to 6 were examined. ^{11 13 19} However, we found mutations solely in exon 2; this was similar to the previous three studies, ^{11 13 19} which discovered them only at exons 2 and 3. There appears to be no benefit in performing the sequence for exons 4 to 6.

Our study reported two variants, the first one being the 17 base-pair duplicated frameshift mutation (c.154_170 dup) and the second as c.150_158delins CT. Both are novel and located in exon 2, a hotspot for acquired GATA1 mutations. The formal variant resembled two previous reports of a 20 nucleotides duplication from nucleotide position 153 (c.154_173 dup, p.Ala59Glnfs*85) in a TAM newborn, seen in Robert et al, ⁴ and in an adult non-DS patient with acute megakaryoblastic leukemia in Harigae et al. ²⁰ Kanezaki et al ¹¹ also discussed a 19 nucleotides duplication from position 155 (c.155_173 dup, p.Ala59Serfs*15) in a TAM patient. Our second variant resembled the eight nucleotides deletion from position 154 (c.154_161del, p.Thr52Argfs*13) in a patient in the study of Kanezaki et al. ¹¹

Both of our frameshift mutations are classified as pathogenic and appeared to predict GATA1s. They involved an amino acid change from the N-terminal activation domain (N-TAD) to the end, versus the normal end at 413 for GATA1. The first mutation change started at codon 58 and ended at 142, whereas the second change started at 52 and ended at 134. This phenomenon occurred with the N-TAD alteration, essential for activating target gene expression in hematopoiesis through lack of N and C terminal regions.

Our low detection rate for GATA1 mutations may be because of the small number of enrolled patients. A second factor would be that whole blood from some patients (patients 4 and 5) was collected after blast remission. Third, Sanger sequencing technique cannot detect any GATA 1 mutations if blast mutations are less than 15 to 40%. This is in contrast to next-generation sequencing which is capable of detecting mutations even if blasts are less than 10%. ²¹ Finally, recent studies ^{4 18 22} have reported a less than 100% detection rate for GATA1 mutations in neonates with TAM, although highly sensitive testing was used. Peripheral blood and bone marrow results do not differ if there are over 10 to 20% blasts present in peripheral blood. Most studies ^{4 18 19} of TAM analyzed peripheral blood, like in ours.

Ara-C is still the treatment of choice for TAM. However, as most neonates with TAM have spontaneous remission, treatment is only recommended in severe clinical conditions or for those with life-threatening symptoms, e.g., hyperviscosity, hypercellularity with blast cells counts > 100,000/mm ³ , organomegaly with respiratory compromise, hydrops fetalis, cardiac/liver/renal failure, or disseminated intravascular coagulation (DIC) with bleeding. ^{3 6 7} Ara-C was given to five of our seven (71%) patients, a higher treatment rate versus a previous study in which only 21% received chemotherapy. ³ For our team of pediatric hematology/oncology physician specialists, bleeding diathesis was a major indication for chemotherapy treatment; other experts may hold a differing opinion. Only two patients (patients 3 and 7) met the criteria of hyperleukocytosis and multiple organ failure for treatment; however, both passed away during treatment. The treatment group had higher mortality, likely because they had more severe TAM, ³ presenting with hyperleukocytosis. ⁷ Although the chemotherapy doses were low, complications may occur as were reported in 44% of the TAM patients receiving Ara-C in a previous study. ³ Before any treatment decision is made, careful weighing of pros and cons should be done.

Conclusion

Our study demonstrated two novel GATA1 mutations in Thai neonates with DS and TAM. Our data confirmed the mutations usually take place in exon 2, as previously reported. Although the detection rate was low for us, GATA1 testing is strongly recommended for neonates with TAM, specifically at the time when blast cells appear in peripheral blood. Neonates with DS with TAM and carrying GATA1 mutations must be closely monitored.