Acute Myeloid Leukemia in Mice Associated with Retrovirally-Mediated Overexpression of HOXB4

Abstract

Numerous studies have established that retrovirally-mediated overexpression of the homeobox gene HOXB4 in murine bone marrow cells (BMCs) enhances expansion of hematopoietic stem cells (HSCs) in vivo and in vitro without causing hematopoietic abnormalities or impairing the ability of HSCs to give rise to differentiated progeny. More recently, however, retrovirally-mediated overexpression of HOXB4 in BMCs of large animals, such as dogs and macaques, has been implicated in leukemogenesis, but there are no prior reports of such a phenomenon occurring in mice transplanted with HOXB4-transduced BMCs. We report here two independent occurrences of acute myeloid leukemia (AML) in mice that received transplants of BMCs transduced with a HOXB4-overexpressing retrovirus. 5-FU treated BMCs were transduced with a MSCV-HOXB4-IRES-GFP retrovirus and transplanted into lethally irradiated mice. Two of the 54 mice (#1869 & #2050), transplanted with HOXB4-transduced BMCs, developed AML at 4 and 9.5 months, respectively, following transplantation. None of the 90 control mice, transplanted with BMCs transduced with the parent MSCV-IRES-GFP retrovirus, developed leukemia. Leukemia 1869 (L1) manifested marked leukocytosis, thrombocytopenia and mild anemia. Results of morphological and immunophenotyping analyses of bone marrow were consistent with AML. Leukemia 2050 (L2) also showed marked leukocytosis and significant anemia, but a near normal platelet count. Morphological analysis revealed a myelomonocytic phenotype, while immunophenotying showed the presence of multi-lineage cell surface markers, including myeloid (CD11b, Gr-1), erythroid (Ter119), and B lymphoid (B220) markers. A small percentage of cells that expressed the stem cell markers c-kit and Sca1 were identified in both leukemias. Both leukemias could be passaged by serial transplantation of leukemic bone marrow or spleen cells into wild type recipient mice. Analysis of retroviral integration sites identified ten separate integration sites in L1 and two in L2. This correlated with 6 fold higher levels of HOXB4 mRNA in L1 BMCs compared to L2 BMCs, as assessed by northern blot analysis that showed undetectable levels of HOXB4 mRNA in RNA of control BMCs. Some of the integration sites occurred in introns of genes known to be involved in leukemogenesis and resulted in aberrant expression of these genes. In L1, one of the retroviral integration sites occured in intron 1 of the N-ras gene and was associated with markedly increased levels of its mRNA in RNA of leukemic BMCs (15–20 fold higher levels compared to that in control BMCs). Another integration site in L1 was in intron 2 of the Lmo2 gene and was associated with a mild (~2 fold) increase in its mRNA level in leukemic cells. In L2, one of the retroviral integration sites was in intron 1 of the Prdm16 gene and was associated with markedly increased levels of its mRNA in leukemic cells. The second retroviral integration site in L2 was located ~50 kb upstream of the Notch1 gene and there was ~2 fold increase in its mRNA in BMCs of leukemic mice. We speculate that overexpression of HOXB4 acts in a cooperative or combinatorial fashion with overexpression of proto-oncogenes at the retroviral integration sites to induce AML in these mice. To test this hypothesis, we are testing various culture conditions to allow viable growth of the leukemic cells ex vivo and assess the effects of knockdown of HOXB4 mRNA accumulation on growth and differentiation of the cells in vitro, as well as on their ability to transmit leukemia in vivo. We are also characterizing the nature of the cells in the total leukemic cell population that are associated with in vivo leukemia-initiating activity. This is the first report of AML in mice associated with retrovirally-mediated overexpression of HOXB4.