HOXA9 Modulates Its Oncogenic Partner Meis1 To Influence Normal Hematopoiesis

While investigating the mechanism of action of the HOXA9 protein, we serendipitously identified Meis1 as a HOXA9 regulatory target. Since HOXA9 and MEIS1 play key developmental roles, are cooperating DNA binding proteins and leukemic oncoproteins, and are important for normal hematopoiesis, the regulation of Meis1 by its partner protein is of interest. Loss of Hoxa9 caused downregulation of the Meis1 mRNA and protein, while forced HOXA9 expression upregulated Meis1. Hoxa9 and Meis1 expression was correlated in hematopoietic progenitors and acute leukemias. Meis1^+/− Hoxa9^−/− deficient mice, generated to test HOXA9 regulation of endogenous Meis1, were small and had reduced bone marrow Meis1 mRNA and significant defects in fluorescence-activated cell sorting-enumerated monocytes, mature and pre/pro-B cells, and functional B-cell progenitors. These data indicate that HOXA9 modulates Meis1 during normal murine hematopoiesis. Chromatin immunoprecipitation analysis did not reveal direct binding of HOXA9 to Meis1 promoter/enhancer regions. However, Creb1 and Pknox1, whose protein products have previously been reported to induce Meis1, were shown to be direct targets of HOXA9. Loss of Hoxa9 resulted in a decrease in Creb1 and Pknox1 mRNA, and forced expression of CREB1 in Hoxa9^−/− bone marrow cells increased Meis1 mRNA almost as well as HOXA9, suggesting that CREB1 may mediate HOXA9 modulation of Meis1 expression.While the Hox homeobox genes are widely recognized as important developmental genes (26), we and others have shown that several Hox genes, and Hoxa9 in particular, are important for both normal hematopoiesis (27, 28) and leukemic transformation (25, 29). While the Hoxa9 gene plays a role in embryonic development, much of the research on this gene has focused on its role as an oncogene that is often upregulated in acute myeloid leukemias (12, 29). In an analysis of 6,817 genes, Hoxa9 was the most highly positively correlated with treatment failure in acute myeloid leukemia patients (18). Meis1 is a member of the TALE family of non-Hox homeobox genes, which was initially identified as a frequent viral integration site in myeloid leukemias arising in BXH2 mice (32). The Hoxa9 gene is also upregulated in many of the leukemias arising in the BXH2 animals (33). Forced expression of HOXA9 in murine bone marrow (BM) cells in culture results in immortalization of myeloid progenitor cells (4, 15), while transplantation of HOXA9-infected BM cells results in the eventual induction of acute myelogenous leukemia (25). In contrast, transplantation of BM cells infected with HOXA9 plus MEIS1 results in rapid development of disease (25). Both HOXA9 and MEIS1 are expressed following forced expression of the MLL oncogene (47) or in patients with MLL gene rearrangements (22).Hoxa9 is expressed in numerous tissues during development, including rib (8), limb (17), motor neuron progenitors (10), reproductive tract (9), and mammary gland (7). Hoxa9 is also expressed in normal adult BM (24, 43), and loss of Hoxa9 leads to multiple relatively mild defects in normal hematopoiesis (23, 27, 28). Retroviral expression studies have also shown that HOXA9 and MEIS1 are important for myeloid blood cell differentiation (3, 4). Despite the broad expression of Hoxa9 and other Hox genes, relatively little is known about how the HOX proteins function. An important advance was the discovery that many HOX proteins gain DNA binding specificity by forming complexes with the PBX (6, 31), MEIS1 (41), and PREP1 (2) proteins. Although HOXA9 is capable of binding DNA alone (42), it forms cooperative DNA binding complexes with MEIS1 alone (41) and in a triple complex with PBX proteins (40, 44). Despite these apparent advances, relatively few downstream targets for HOX proteins, and HOXA9 in particular (11), have been confirmed.During ongoing studies of the mechanism of action of the HOXA9 protein, we discovered that HOXA9 appeared to upregulate the Meis1 mRNA and protein. Given the numerous biological connections between HOXA9 and MEIS1, we embarked on studies to explore this pathway. Forced expression of HOXA9 in BM cells upregulated the Meis1 mRNA and protein, while loss of Hoxa9 resulted in a reduction in the Meis1 mRNA and protein. In addition, in a biological model to assess Hoxa9 modulation of Meis1, compound mutant animals that were homozygous null at the Hoxa9 locus and heterozygous at the Meis1 locus showed a significant loss of murine BM monocytes, mature B cells, and pre/pro-B-cell progenitors and an increase in orthochromatophilic erythroblasts in postnatal-day-15 mice compared to results for all controls, suggesting that HOXA9 regulates Meis1 during normal hematopoiesis. Chromatin immunoprecipitation (ChIP) analysis did not show direct binding of HOXA9 to distal or proximal Meis1 genomic regions. However, these studies, together with PCR analysis, showed that HOXA9 binds to and upregulates two genes, Creb1 and Pknox1 (the protein product is subsequently referred to as PREP1), whose protein products have previously been reported to upregulate Meis1 expression (13, 14). Addition of CREB1 to Hoxa9^−/− bone marrow cells increased Meis1 mRNA nearly as effectively as HOXA9. Taken together, our data show that HOXA9 indirectly modulates its DNA binding and oncogenic partner MEIS1 and that the DNA-binding property of HOXA9 is required for this process. We further show that Hoxa9 modulation of Mes1 is biologically important during normal hematopoiesis and that CREB1 may mediate the regulation of Meis1 by HOXA9.     

Involvement of Crosstalk between Oct4 and Meis1a in Neural Cell Fate Decision

Oct4 plays a critical role both in maintaining pluripotency and the cell fate decision of embryonic stem (ES) cells. Nonetheless, in the determination of the neuroectoderm (NE) from ES cells, the detailed regulation mechanism of the Oct4 gene expression is poorly understood. Here, we report that crosstalk between Oct4 and Meis1a, a Pbx-related homeobox protein, is required for neural differentiation of mouse P19 embryonic carcinoma (EC) cells induced by retinoic acid (RA). During neural differentiation, Oct4 expression was transiently enhanced during 6–12 h of RA addition and subsequently disappeared within 48 h. Coinciding with up-regulation of Oct4 expression, the induction of Meis1a expression was initiated and reached a plateau at 48 h, suggesting that transiently induced Oct4 activates Meis1a expression and the up-regulated Meis1a then suppresses Oct4 expression. Chromatin immunoprecipitation (ChIP) and luciferase reporter analysis showed that Oct4 enhanced Meis1a expression via direct binding to the Meis1 promoter accompanying histone H3 acetylation and appearance of 5-hydoxymethylcytosine (5hmC), while Meis1a suppressed Oct4 expression via direct association with the Oct4 promoter together with histone deacetylase 1 (HDAC1). Furthermore, ectopic Meis1a expression promoted neural differentiation via formation of large neurospheres that expressed Nestin, GLAST, BLBP and Sox1 as neural stem cell (NSC)/neural progenitor markers, whereas its down-regulation generated small neurospheres and repressed neural differentiation. Thus, these results imply that crosstalk between Oct4 and Meis1a on mutual gene expressions is essential for the determination of NE from EC cells.