Thymic Precursor Cells Generate Acute Myeloid Leukemia in NUP98-PHF23/NUP98-HOXD13 Double Transgenic Mice

Introduction: Oncogenic fusion genes have been identified in a large number of hematologic malignancies. Using Vav regulatory elements, we previously developed transgenic mice that expressed either a NUP98-PHF23 (NP23) or NUP98-HOXD13 (NHD13) fusion in the hematopoietic compartment. Both NP23 and NHD13 mice develop a wide variety of leukemias, including myeloid, erythroid, megakaryocytic and lymphoid, with an age of onset typically between 9 and 14 months, but rarely before 6 months of age.

Results: NP23-NHD13 double transgenic mice were generated by interbreeding NP23 and NHD13 single transgenic mice. Surprisingly, 100% of the NP23-NHD13 double transgenic mice showed rapid onset of acute myeloid leukemia (AML) within three months. In contrast, none of the single transgenic mice had developed leukemia by 3 months of age. The leukemias were characterized by extraordinarily high WBC (>500,000/uL) and almost complete replacement of the thymus with Mac1+/Gr1+ myeloblasts, with the percent of malignant myeloid cells in the thymus often higher than the percent in the bone marrow or spleen. These findings led to the intriguing hypothesis that the AML generated in NP23-NHD13 mice arose in the thymus, as opposed to the bone marrow.

We harvested cells from the thymus of a NP23-NHD13 mouse that had been invaded by Mac1+/Gr1+ AML cells, and transplanted unfractionated cells (10E06 per recipient) as well as residual CD4-/CD8- double negative (DN) thymocytes (10E05 per recipient; these thymocytes were negative for Mac1 and Gr1 as well as CD4 and CD8) into sub-lethally(600 rds) irradiated recipients. All mice (transplanted with unfractionated thymocytes or DN cells from same leukemic mouse) developed Mac1+/Gr1+ AML within 26 days, indicating that the AML was aggressive and transplantable, and could be transmitted by DN thymocytes. The observation that there was no difference in onset of AML transmitted by the DN thymocytes or Mac1+/Gr1+ AML suggested that the DN thymocytes may be the cell of origin for this leukemia. To rule out the possibility that the leukemia in this experiment was transmitted by rare, contaminant Mac1+/Gr1+ cells, we repeated the experiment, twice, using DN thymocytes from 4-5 wk old mice with no signs of leukemia. In these experiments, the DN thymocytes again transmitted a Mac1+/Gr1+ AML. Finally, we further fractioned DN thymocytes into DN1, DN2, DN3 and DN4 populations; the recipients of DN1 and DN2 populations developed a Mac1+/Gr1+ AML, whereas the DN3 and DN4 recipients were healthy with no signs of engraftment in the peripheral blood.

To gain further insight into this phenomenon, DN thymocytes from non-leukemic NP23-NHD13 mice were co-cultured on an OP9 stromal layer, which has been shown to support myeloid, B, and T lymphocyte (when supported with Notch1 ligands) differentiation in vitro. DN thymocytes from non-leukemic NP23-NHD13 mice co-cultured on a OP9 stromal layer showed a markedly enhanced ability to differentiate into Mac1+/Gr1+ myeloid lineage cells compared to WT (56% vs 1.4 % at day 10). However, this "wave" of myeloid differentiation was extinguished by 26 days. The NP23-NHD13 cells lost expression of Mac1 and Gr1; an expanded immunophenotype was CD4^-, CD8^-, CD25-, CD44⁺, Thy-1.2⁺ and cKit⁺, consistent with a self-renewing DN1 thymocyte. These cells were transplanted; all recipients were anemic, and demonstrated engraftment of NP23-NHD13 myeloid cells as well as a less prominent (7.85%-38.5%) population of CD71+/Ter119+ erythroid cells in the BM and spleen. There was no evidence of T-cell differentiation of the transplanted NP23-NHD13 cells. Further thymocyte fractionation experiments using the OP9 co-culture system demonstrated that myeloid differentiation potential resided in the DN1 population.

Conclusions: Taken together, these results indicate that NP23-NHD13 thymic progenitors are potently leukemogenic, and retain myeloid and erythroid differentiation potential.