Distinct Roles of GATA2 in Development and Evolution of CBFB-MYH11 AML

Introduction. Core binding factor acute myeloid leukemia (CBF-AML) is caused by the dysfunction of a heterodimeric protein complex composed of the transcription factor RUNX1 and its partner CBFb. An inversion of chromosome 16 generates a fusion between CBFB and MYH11. The encoded fusion protein, CBFb-smooth muscle myosin heavy chain (SMMHC), contributes to the pathogenesis of CBF-AML. Our previous reports suggest that CBFB-MYH11 contributes to leukemogenesis by up-regulation of genes such as Gata2, which is an essential hematopoietic transcription factor. On the other hand, we recently identified recurrent monoallelic deletions of GATA2 on chromosome 3 in relapsed CBF-AML patients (Sood et al., Leukemia 30:501-504, 2016). From these findings we propose two hypotheses; 1) up-regulation of GATA2 contributes to leukemogenesis by CBFB-MYH11 in the initiation phase; 2) GATA2 deficiency contributes to the relapse of CBF-AML.

Methods. Two datasets (GSE19194 and GSE102388) from microarray and RNA-Seq were used to determine Gata2 expression level in Cbfb-MYH11 preleukemic murine hematopoietic cells. Cbfb-MYH11 conditional knock-in (Cbfb^+/56M), Gata2 conditional knockout (Gata2^+/f), and Mx1-Cre transgenic mice were crossed to generate Gata2^+/fCbfb^+/56MMx1-Cre mice. Mice were injected with pIpC to induce the expression of Cbfb-MYH11 and/or knockout of Gata2 through Cre-recombinase activation. For transplantation assays, spleen cells obtained from leukemic mice were injected into irradiated recipient mice through tail vein. For in vitro colony forming assays, colonies were counted after 10 days in culture. Cell apoptosis was determined by Annexin V and 7AAD staining.

Results. To test the first hypothesis, we determined the expression level of Gata2 in preleukemic cells in the Cbfb-MYH11 expressing mice. Data from both microarray and RNA-seq experiments revealed that Gata2 was highly expressed in the preleukemic hematopoietic cells of the Cbfb-MYH11 mice, as compared to those of the WT mice, and this finding was confirmed by qRT-PCR. Based on published ChIP-seq data, Gata2 is likely a direct transcriptional target of CBFb-SMMHC. Next, we determined the impact of Gata2 deficiency on leukemogenesis by Cbfb-MYH11. qRT-PCR showed reduced Gata2 expression in bone marrow cells from Gata2^+/fCbfb^+/56MMx1-Cre mice 12 days after pIpC injection (0.029±0.0092 vs 0.076±0.014; p=0.0089). Colony forming ability was decreased for the pre-leukemic bone marrow cells in Gata2^+/fCbfb^+/56MMx1-Cre mice when compared to Cbfb^+/56MMx1-Cre mice (mean 37.2±6.35 vs. 74.23±8.335; p=0.0002). In addition, the Gata2^+/fCbfb^+/56MMx1-Cre mice had a smaller abnormal myeloid population in the bone marrow, which is capable of inducing leukemia, when compared with Cbfb^+/56MMx1-Cre mice (mean 0.43±0.14% vs. 1.42±0.34%; p=0.0092). Most significantly, Gata2^+/fCbfb^+/56MMx1-Cre mice developed leukemia with a much longer latency than Cbfb^+/56MMx1-Cre mice (median survival 215 days vs 125 days; p=0.0007).

To test hypothesis 2, we compared the phenotype of the end stage mice for each genotype. Gata2^+/fCbfb^+/56MMx1-Cre mice had higher WBC count in peripheral blood than Cbfb^+/56MMx1-Cre mice (mean 92,000±20,429 cells/ul vs. 35,644±12,001 cells/ul; p=0.0243), which is a poor prognostic marker in human leukemia. Leukemic cells from Gata2^+/fCbfb^+/56MMx1-Cre mice also had lower percentage of Annexin V positive cells than Cbfb^+/56MMx1-Cre mice in short term culture (31.0±7.1 vs. 68.9±6.5%; p=0.0117). More importantly, upon transplantation, the recipient mice transplanted with Gata2^+/fCbfb^+/56MMx1-Cre leukemia cells developed leukemia much faster than recipient mice transplanted with equal numbers of Cbfb^+/56MMx1-Cre leukemia cells (median survival 35.5 vs. 91.0 days; p<0.0001).

Conclusions. Our findings suggest that Gata2 plays important but distinct roles in two different stages of Cbfb-MYH11 leukemia. Reduction of Gata2 activity delays leukemia development in primary Cbfb-MYH11 knockin mice, while contributing to a more aggressive phenotype in leukemic phase as shown in primary leukemic mice and transplanted recipients, which may be correlated with leukemia relapse in human patients. We are analyzing data from whole exome sequencing and RNA-seq to understand the mechanism underlying the observed phenotypes, and the findings will be presented at the annual meeting.