Loss-of-Function and Gain-of-Function Consequences of GATA2 Disease Mutations

The master regulator of hematopoiesis GATA2 controls generation and function of hematopoietic stem and progenitor cells, and heterozygous GATA2 mutations create a predisposition to develop immunodeficiency, myelodysplasia, and acute myeloid leukemia (Spinner et al. Blood, 2014; Dickinson et al. Blood, 2014; Churpek and Bresnick J. Clin. Invest. 2019). Although mechanisms that trigger the transition of a non-pathogenic GATA2 mutation into overt pathology are enigmatic, a paradigm has arisen in which GATA2 mutations are considered to be loss-of-function. We developed a genetic rescue assay to quantify the function of wild type GATA2 and GATA2 disease mutants when expressed at near-physiological levels in primary progenitor cells and demonstrated that GATA2 disease mutations abrogate certain biological and molecular activities, while enabling others (Katsumura et al., 2018, PNAS). We isolated lineage-negative (Lin^-) or Lin^-Kit⁺ cells from fetal liver of mice with a homozygous mutation of the Gata2 -77 enhancer, which downregulates Gata2 expression by ~80%. The mutant progenitor cells are largely defective in erythroid, megakaryocytic and granulocytic differentiation and exhibit a predominant monocytic differentiation fate (Johnson et al., 2015, Science Adv.). We compared GATA2 and GATA2 disease mutant activities in the rescue system using a colony formation assay. GATA2, R307W mutant (in N-finger) and T354M mutant (in DNA-binding C-finger) rescued myeloid colony formation and promoted granulocyte proliferation. Surprisingly, R307W and T354M induced more CFU-GM than GATA2. GATA2 and R307W, but not T354M, rescued BFU-E. These data indicated that GATA2 disease mutations were not strictly inhibitory, and in certain contexts, mutant activities exceeded that of GATA2.

To extend these results, we subjected -77^+/+ or -77^-/- Lin^- cells to a short-term ex vivo liquid culture, expressed GATA2, R307W, or T354M and used RNA-seq to elucidate progenitor cell transcriptomes. While -77^+/+ Lin^- cells generate erythroid and myeloid cells, -77^-/- Lin^- cells are competent for myeloid, but not erythroid, differentiation. Comparison of -77^+/+ and -77^-/- cell transcriptomes revealed 3064 differentially expressed genes (>2-fold). 1824 genes were >2-fold higher in -77^+/+ cells, and 1240 genes were >2-fold higher in -77^-/- cells. GATA2 expression in -77^-/- cells activated 834 genes >2-fold and repressed 503 genes >2-fold. 60-65% of these genes overlapped with genes differentially expressed between -77^+/+ cells and -77^-/- cells. R307W expression activated 661 genes >2-fold and repressed 523 genes >2-fold. T354M expression activated 468 genes >2-fold and repressed 575 genes >2-fold. The genes regulated by mutants included GATA2-regulated genes and certain genes that were not GATA2-regulated. Multiple genes were hypersensitive to the mutants, relative to GATA2, and the mutants ectopically regulated certain genes. However, R307W and T354M did not universally regulate an identical gene cohort. For example, both R307W and T354M activated Ncam1, Nrg4, and Mpo more strongly than GATA2. R307W, but not T354M, activated Ear2 and Ces1d more strongly than GATA2. By contrast, T354M, but not R307W, activated Ctsg, Epx, and Rab38 more strongly than GATA2. Both R307W and T354M repressed macrophage genes similarly to GATA2, but they lacked the capacity to activate mast cell genes, differing from GATA2.

To elucidate molecular mechanisms underlying GATA2 mutant activities, we leveraged our prior discovery that p38 or ERK kinases induce multi-site GATA2 phosphorylation (Katsumura et al. Blood. 2017). We tested whether these kinases mediate the ectopic transcriptional regulatory activity of GATA2 disease mutants. p38 inhibition attenuated aberrant regulation of Ear2 and Ces1d by R307W (p < 0.05), and mutation of S192 to S192A decreased R307W-induced CFU-GM formation by 49% (p < 0.05). In aggregate, these results indicate that GATA2 disease mutants exert context-dependent activities to regulate transcription and differentiation, activities can be signal-dependent and certain activities are distinct from GATA2. It is attractive to consider the pathogenic consequences of GATA2 disease mutant gain-of-function activities, and an important implication is GATA2 mutation-associated hematologic diseases might not solely reflect haploinsufficiency.