It is important to acknowledge that we do not have much experience with foetal heart enlargement in combination with hepatosplenomegaly. Previous exposure to Beckwith-Wiedemann syndrome foetuses with hepatosplenomegaly, but often with concurrent manifestations of overgrowth, an enlarged tongue, and hyperhydramnios. We also considered intrauterine infection—could it be due to an intrauterine syphilis infection? However, we noted no abnormalities in the serum treponema pallidum antibody of the pregnant women.
In this regard, we reviewed the literature related to foetal hepatosplenomegaly. TAM was brought to our attention. Children with Down syndrome have a unique susceptibility to develop ML-DS. This disorder is preceded by a transient neonatal preleukaemic syndrome known as TAM20. One possible explanation for the occurrence of early death in preterm infants who develop TAM is that the immaturity of the organs results in an inability to tolerate the blasts of TAM, leading to organ failure in the uterus16. This may also account for the intrauterine death of the foetus in Case 1. A definitive diagnosis of TAM was made on the basis of various hematological and antenatal ultrasound presentations of the two foetuses, coupled with molecular testing demonstrating a mutation in the GATA1 gene21. In addition, in case 1 we found blood cell components in the fetal liver with blast infiltrating into the tissue (Supplementary Fig. 7).
In this study of Case1, we promptly scheduled the pregnant woman for umbilical venipuncture. Amniotic fluid was also taken to test for cytomegalovirus and rubella virus. First, the cord blood QF-PCR results appeared on Day 3, suggesting a foetus with trisomy 21(Supplementary Fig. 2). However, by this time, intrauterine foetal death had already occurred. Results for amniotic fluid cytomegalovirus and rubella virus were normal. Because we noted that the non-invasive prenatal testing(NIPT) examination of the pregnant woman in early pregnancy suggested a low risk (other hospitals test and raw data is not available).The common cause of false negatives for NIPT is true fetal mosaicism22, so we took placental tissue for CNV-seq examination. The centre point of the placenta is a low proportion (6%-8%) of trisomy 21 mosiacism. This result was also verified by another method, and interphase fluorescence in situ hybridisation (FISH) analysis detected trisomy 21 signals in 5/100 (5%) uncultured placental cells. False negatives on NIPT are extremely rare and previous reports have suggested that a low percentage of placental mosiacism is a cause of false negatives23,24.
After the foetus in Case1 was delivered, the foetal face did not show any obvious facial characteristics of Down syndrome, other than a collapsed nose. Another hint is given to us through the publication by Roseman et al. of the possibility that T21 first occurs in the haematopoietic progenitor cells of the foetal liver, eventually leading to the placenta having a small amount of T21 mosiacism25, although this is still speculative. Therefore, we suspected that this might be a case of trisomy 21 mosiacism. This is because the cord blood chromosome results represent only the mesoderm of the embryo and are not representative of the other germ layers. In addition the foetus does not show any other manifestations of T21 other than TAM. However, we have been unable to confirm this hypothesis and we were unable to obtain the consent of the mother of the foetus to continue the study.
On the other hand, we also thought of the possibility of TAM in non-Down syndrome. Several cases of TAM in infants without Down syndrome have been reported in the past, and such newborns tend to have a lower mortality rate and do not present with more severe conditions, such as hydrops and multiorgan failure26. However, our cases were more severe.
Most GATA1 mutations in TAM include either frameshift or nonsense mutations within exon 2, the GATA1 mutations create an early stop codon, resulting in a short isoform of the GATA1 protein that lacks the N-terminal activation domain, which then affect the translation of the GATA1 protein 1,9. The detection of GATA1 mutations is important for the diagnosis of TAM. The haematopoietic transcription factor gene GATA1 (localised at Xp11.23) is required for the development of megakaryocytes, erythrocytes, mast cells and eosinophils, and the dominance of the GATA1 gene in combination with the gene dosage of the effect of trisomy 21 induces excessive proliferation of erythro-megakaryocytic blast cells27. Mutations in the GATA1 gene are thought to be a pathognomonic feature of all myeloproliferative disorders in children with Down syndrome, including those with TAM 28. Our study also suggests that the cause of TAM in foetuses may be due to mutations in the GATA1 gene.
In our cases, two different mutations were also identified by sequencing at exon 2: a hemizygous variation c.220G > A in Case 1 and a hemizygous variation c.49dupC in Case 2. We evaluated the gene-disease association following the ClinGen Gene-Disease Validity Standard Operating Procedures, and curated the GATA1 gene to “Definitive” grade. Both of the variations c.220G > A and c.49dupC are absent in the general population according to public databases (gnomAD, 1000 Genomes Project, and Exome Aggregation Consortium). Mutation c.220G > A has been reported in several research as pathogenic according to the ACMG guideline29–31. At the same amino acid c.220G > C (p.Val74Leu) change has been previously reported as pathogenic in the ClinVar database. The consequence of the base c.220G > A substitution was a splice mutation that changed the amino acid sequence Val74Ile 29.Also at the same amino acid c.49_50del (p.Gln17fs) and c.49C > T (p.Gln17Ter) already provide to be pathogenic in the ClinVar database. Cabelof, DC et al32 and Kanezaki R et al33 reported c.49 C > T in their case 10 and case 5, respectively. However, the c.49dupC mutation in Case 2 is a new case reported for the first time in the literature. The consequence of the c.49dupC variation was similar to the base c.49 C > T substitution, which changed the amino acid sequence Gln17Ter. This mutation was a nonsense mutation resulting in a stop codon before Met84. A variety of prediction tools (SIFT, DANN, and REVEL) were used to evaluate the possible functional impact of c.220G > C, and it is predicted to be a damaging variation by all three tools. Furthermore, various algorithms (GERP, phyloP, and phastCons) and multiple sequence alignments from the UCSC genome browser predicted that this position is conserved across multiple vertebrate species. Both of the mutations in our study, which located at the exon 2, highly possible to be disrupted the function of the protein and allow the abundant generation of truncated GATA1.
An interesting point in previous studies is that the overall survival rate of boys who develop TAM is lower and that somatic mutations in the GATA1 gene on the X chromosome may contribute to the lethality in boys, with girls still having a wild-type GATA1 gene on the other X chromosome 5,34,35. Both of our cases included male foetuses, further confirming the sex bias of TAM36.
In conclusion, we identified two rare prenatal cases of foetuses with trisomy 21 combined with TAM. The [NM_002049.4 c.220G > A (p. Val74Ile) (Case 1)] mutation and the [NM_002049.4 c.49dupC (p. Gln17ProfsTer23) (Case 2)] mutation of GATA1 were present, respectively. Further confirming the prenatal origin of the GATA1 mutation. The presence of foetal trisomy 21 in the presence of prenatal ultrasound findings such as foetal hydrops or hepatosplenomegaly should be considered.