Functional dissection of homeodomain transcription factors HoxA9 and Meis1 in myeloid leukemia : their epigenetic regulation and their downstream targets
- Author(s): Wang, Gang;
- et al.
Homeodomain-containing transcriptional factors determine patterning and cellular fate during development. Upregulation of homeodomain factors including Hox-A and Meis1 was found in most of acute myeloid leukemias (AML). Hox-A and Meis1 proteins promote hematopoietic progenitor self-renewal, arrest their differentiation and cooperate to induce AML. However their upregulation mechanisms in AML, the molecular basis of their cooperation, and their downstream leukemogenesis-relevant targets are largely unknown. Here I report a Hox-A gene activation mechanism in AML harboring translocation that fuses Nucleoporin-98 (NUP98) to Nuclear-receptor-binding SET-domain-containing 1 (NSD1). I discovered that AML induced by NUP98-NSD1 was concurrent with upregulation of the Hox-A genes and Meis1. Mechanistically, NUP98-NSD1 directly bound to the Hox-A locus and induced a coordinate epigenetically modifying program via the intrinsic methyltransferase activity of NSD1, recruitment of the histone acetyltransferase CBP/ p300, and suppression of the recruitment of Polycomb-group repressor EZH2. Both the NUP98 fusion part and the NSD1 methyltransferase domain were required for leukemogenesis. To understand Hox-Meis1 cooperation, I established a leukemia stem cell model by coexpression of HoxA9 plus Meis1, which recapitulated the unique AML-initiating character and led to the identification of Meis1- associated leukemia-signature genes. Among them included CD34, FLT3, Erg1 and etc, the genes found important for stem cell biology and mutated in leukemia. To program leukemogenesis, Meis1 and Hoxa9 required binding to cofactor PBX and to DNA, Meis1 required its C-terminal transactivation domain (CTD), and Hoxa9 required its N- terminal transactivation domain (NTD). Unexpectedly, I observed that fusing the Vp16 activation domain to Meis1 (Vp16-Meis1) mimicked combined activities of Meis1 plus Hoxa9 and caused leukemia without expression of Hox genes. Interestingly, further expression of HoxA9 or HoxA7 stimulated transcription of the Meis1-related signature genes, and Vp16 activation domain replaced the functions of the Hoxa9 NTD and the Meis1 CTD. These results favored a model in which Meis-Pbx and Hox-Pbx (or Hox-Pbx-Meis) complexes co-occupy leukemia-relevant promoters and that Meis1 CTD and Hox NTD cooperate in gene activation. Furthermore, I described a myeloid-linage differentiation system utilizing estrogen-controlled HoxB8 that blocked the differentiation of neutrophil or macrophage-specific progenitors and, upon its inactivation, these progenitors synchronously differentiated into functional mature cells of pre-defined lineages