Genetics and functional characterization of GATA2: a novel cancer gene in familial leukaemia.

Abstract

We first report GATA2 mutations (heterozygous) in 4 families that are susceptible to MDS/AML (3 large families) and MDS (1 small family). Molecular analysis revealed a germline transmission of a GATA2 missense mutation (T354M) in MDS/AML families and a GATA2 deletion mutation (T355del) in MDS family. Neither germline RUNX1 nor CEBPA mutations were found in these families, in 695 non-leukemic ethnically matched controls and 268 sporadic AML samples. The mutations resided within the GATA2 zinc finger 2 domain, a critical region for DNA-binding and protein-protein interactions, but not for nuclear localization. T354M reduced DNA binding ability of GATA2; whereas, T355del bound very little, if any, to the consensus WGATAR DNA motif. T354M and T355del also significantly reduced the transactivation of GATA2 in known GATA2 responsive sequences. Moreover, co-transfection of T354M or T355del with WT reduced WT transactivation ability, suggesting that these mutants act in a dominant negative fashion. Regulatable stable promyelocytic HL- 60 cells expressing WT and mutants were generated. Forced expression of WT and T354M inhibited HL-60 cell differentiation when induced with all trans retinoic acid. However, when compared to WT, T354M enabled cell proliferation/survival while simultaneously reducing apoptosis. In contrast, T355del was a complete loss-of-function mutant. Microarray studies elucidated that both T354M and T355del significantly decreased the expression of downstream target genes. Together, our data suggest that both T354M and T355del are loss-of- function mutations with some dominant negative attributes. Recently, we and others have described GATA2 genetic lesions in other diseases. We further investigated in vitro functions of an alllelic series of GATA2 mutants representing the major disease phenotypes: MDS/AML (T354M), MDS (T355del), CML-BC (L359V), Emberger syndrome (R361L and C373R), AML-M5 and biallellic CEBPA AML (R362Q), and immunodeficiency syndrome (R398W). We showed that these GATA2 mutants (except L359V) are loss-of-function that reduce DNA binding affinity and transactivation of target genes. Nevertheless, they maintain the ability to bind to known protein binding partners. Intriguingly, T354M and C373R have an enhanced affinity for PU.1, highlighting that these mutants can influence both DNA-binding and protein-protein interaction. Preliminary transduction of Gata2 WT or mutant expression constructs into mouse whole bone marrow cells demonstrated that GATA2 mutants did not confer self-renewal capacity, but allowed specific myeloid progenitor differentiation. We further demonstrated that Gata2 is expressed in lymphatic endothelial cells and that it can bind to and transactivate a Prox1 promoter/enhancer element (PEE) region. Prox1 is required for lymphatic development and maintenance, and hence Gata2 may contribute to lymphoedema throught its action on Prox1. Intriguingly, Gata2 mutants displayed differential binding affinity to two GATA binding sites and reduced transactivation of the PEE region. Furthermore, an enhancer region 11.3kb upstream of Prox1 is activated by GATA2, FOXC2 and SOX18, but repressed by PROX1 itself suggesting that these key lymphatic TFs may cooperate to regulate Prox1 expression. In conclusion, I present the experimental work for the landmark discovery of a new MDS/AML predisposition gene. I have also characterized the molecular landscape of GATA2 mutations where each of the mutations confers specific and major effects on GATA2 function, but where there are also subtle differences between the mutants in the contexts of DNA binding and transactivation.