Our findings that expression of NUP98-HOXD13 (NHD13), NUP98-TOP1, NUP98-RAL1GDS, Lin28b, or Hoxa9 lead to a variety of hematopoietic neoplasms have been previously reported. In order to better understand the leukemogenicity of NUP98 fused to non-HOX genes, we generated mice that expressed a NUP98-PHF23 (NP23) fusion in hematopoietic cells. Almost 100% of these mice develop leukemia within 1 year of life; the leukemic phenotype is very broad, including T and B cell leukemias, myeloid leukemias, and erythroid leukemias. These leukemias are clonal, and frequently activate a cluster of genes within the Hoxa locus, including Hoxa7,9,10,11, and Meis1. In addition, we have identified novel genes, including Bahcc1, that are overexpressed in the NP23 leukemias; a survey of publicly available expression data revealed that BAHCC1 was overexpressed in several distinct subsets of AML. Chromatin immunoprecipitation/sequencing (ChIP-Seq) experiments demonstrated the presence of active histone marks (H3K4Me3) at the Hoxa cluster; moreover, additional ChIP-seq experiments demonstrate binding of the NP23 fusion protein at the H3K4Me3 sites, suggesting a mechanism for the leukemic transformation. Furthermore, we have identified a compound (disulfiram) that inhibits the binding of the NP23 protein to H3K4Me3 residues. Treatment of NP23 cell lines results in rapid downregulation of Hoxa cluster genes and apoptotic cell death. These studies were published in 2014. Intriguingly, a subset of NP23 mice developed a leukemia of B1 progenitor origin; a manuscript describing these findings is being prepared for submission. Lin28b transgenic mice developed an aggressive, clonal, lethal peripheral T cell lymphoma (PTCL), which is associated with release of inflammatory cytokines. The cell of origin is consistent with a T follicular helper (TFH) cell. We interrogated a publicly available gene expression database, and found that Lin28b was overexpressed (mean 7.5 fold) in patients with PTCL. A manuscript describing these features was published in 2012, and we have transferred these mice to several collaborators. Given that the PTCL in these mice was associated with diffuse, systemic inflammation, we have begun pilot experiments to determine whether chronic inflammation will accelerate the disease process. Preliminary results suggests that treatment with lipo-polysaccharide (LPS) does indeed accelerate the onset of disease. Deep sequencing studies have recently identified recurrent point mutations in ASXL1. We have generated transgenic constructs to express the mutant protein in the hematopoietic cells of mice. With respect to ASXL1, despite injections of over 200 mouse eggs, and the birth of 187 pups, no transgenic mice were identified, leading to the suspicion that expression of the mutant ASXL1 protein was embryonic lethal. We have now generated several founder mice with a revised, conditional ASXL1 mutant protein. We have recently crossed these mice to mice that express the Cre recombinase in the thymus or myeloid lineage in order to activate the conditional mutant transgene. Mice that express the mutant ASXL1 in the thymus show a marked (10 fold) reduction in total thymocytes, as well as decreased circulating T cells, demonstrating a phenotypic effect of the mutant ASXL1 transgene. Moreover, given that ASXL1 mutations frequently accompany U2AF1 mutations in patients with MDS, we have generated mice that express a U2AF1 S34F mutation, and transduced bone marrow cells from these mice with an Asxl1 mutant. The function of these cells is currently being assessed in vitro and in vivo. Given that IDH1 and IDH2 gain of function mutations are frequently associated with both MDS and AML, we have generated transgenic mice that express an IDH2 mutant in the hematopoietic compartment. Although these mice did not show a significantly increased incidence of leukemia, crossing these mice to NHD13 mice led to an increased incidence of a unique form of leukemia.
Showing the most recent 10 out of 40 publications
