GATA2 is a New Predisposition Gene for Familial Myelodysplastic Syndrome (MDS) and Acute Myeloid Leukemia (AML)

Abstract

Abstract Abstract LBA-3 Mutations in the transcription factor genes, RUNX1 and CEBPA, can lead to an autosomal dominant familial predisposition to MDS/AML. Using a candidate gene approach, we have detected domain specific heterozygous mutations in the GATA2 gene in 4 MDS/AML families which predispose to MDS/AML. The same novel heterozygous T354M missense mutation was observed in 3 families and a 355delT mutation in 1 family, all with multigenerational transmission of MDS and/or MDS/AML. Importantly, these genetic variants segregate with all affected members in each of the families. The 2 mutated threonine residues are in 5 consecutive highly conserved threonine residues at the DNA-binding, protein-protein interacting second zinc finger (ZF2) of GATA2. Neither these mutations, nor any other variants in the GATA2 coding sequence, were seen in a population screen of 695 normal individuals. Haplotype analysis suggests that the T354M mutation has multiple ancestral origins. While mutations in RUNX1 and CEBPA, can also lead to familial predisposition to MDS/AML, these patients with GATA2 mutations are unique in that there is no obvious pre-MDS or pre-leukaemic phenotype such as thrombocytopenia (RUNX1) and eosinophilia (CEBPA) in predisposed carriers. Most patients in these families have had a rapid disease course “appearing out of the blue” leading to death, with a variety of ages of onset from teenagers to early 40s. Yet remarkably, there are still asymptomatic carriers in their 60s. One of these carriers, and his 2 children, has had bone marrow prophylactically stored over 15 years ago in case of disease onset. No pathogenic GATA2 coding sequence changes were found in 268 sporadic MDS/AML patient samples. Additionally, GATA2 mutations were not found in germline samples from 35 other families predisposed to AML and various other hematological malignancies. Both the T354M and 355delT mutants appear to localize appropriately to the nucleus and maintain at least some DNA binding in electrophoretic mobility shift assays. We used the known murine Gata3 ZF2 structure bound to DNA to model the effects of the observed mutations and demonstrated that the T354 residue does not contact DNA but makes polar contact with the adjacent threonines, and via its amino group, with C349 which coordinates the zinc atom. Replacement of the T354 side-chain with the bulky methionine moiety may affect the zinc contacts and is predicted to alter the overall structure of this ZF2. In contrast, 355delT will shorten the conserved threonine string which is predicted to impact on the orientation and position of L359 which directly contacts DNA. Thus, 355delT is likely to have an effect on DNA binding. Luciferase reporter assays indicate that T354M and 355delT greatly reduce the transactivation ability of GATA2 on multiple response elements, impacting on downstream target genes such as RUNX1 and CD34. Of note, T354M shows a markedly lesser synergistic effect than wildtype (WT) GATA2 with PU.1 on the CSF1R promoter. Competition assays show that these mutations may be acting in a dominant negative fashion in some biological contexts. In stable promyelocytic HL-60 cell lines expressing regulatable GATA2 (WT or T354M), T354M allows proliferation to proceed even under stimulus to differentiate with all-trans retinoic acid. Microarray studies indicated that the down regulation of proapoptotic BCL-xS by T354M, but not WT, may be responsible for this phenotype. GATA2 is considered to be a hematopoietic “stemness” gene, highly expressed in haematopoietic stem cells and is required for megakaryocyte and mast cell production. GATA2 is down regulated during myeloid differentiation and forced overexpression prevents such differentiation. Discovery of GATA2 mutants in MDS/AML predisposed families provides new tools for probing the mechanism of GATA2 induced leukemogenesis, and possibly also for clarifying its role in maintenance of stemness. Our findings highlight the power of investigating familial predispositions to cancer identifying specific mutations with unique biological effects. They have immediate implications for diagnostic genetic testing, and longer term therapeutic implications through identification of drugable biological pathways such as apoptosis. The poor outcome associated with these mutations may suggest that an aggressive strategy is appropriate in the treatment of affected individuals in families found to be carrying GATA2 mutations. Disclosures: No relevant conflicts of interest to declare.