In collaboration with Dr. Donald Small (Johns Hopkins University), we have crossed mice that express a NUP98-HOXD13 (NHD13) transgene and develop myelodysplastic syndrome (MDS) with mice that express a FLT3 internal tandem duplication and develop a myeloproliferative disesase (MPD). Both of these abnormalities were initially identified in patients with acute myeloid leukemia (AML), underscoring the clinical relevance of this experiment. Although very few of the NHD13 or FLT3 only mice developed AML, 100% of the mice with both NHD13 and FLT3 alleles devleoped AML, demonstrating that these two genetic lesions collaborated with one another quite powerfully. A careful study of these mice has revealed a surprising deletion that is highly specific for leukemic mice. Parts of this study have been presented in abstract form in Dec. 2009, and a manuscript describing the findings in detail is in preparation. Retroviral insertional mutagenesis (RIM) has proven to be a valuable whole-genome screen for the identification of genes involved in malignant transformation. Common insertion sites (CIS) are regions that have retroviral integrations in more than one leukemic sample, and have been biologically selected as dominant clones, and are thought to harbor genes important for malignant transformation. We previously described the use of RIM to identify a number of genes which collaborate with a NUP98-HOXD13 to produce a leukemic clone. We have used the same approach to identify genes that collaborate with a CALM-AF10 fusion to produce acute leukemia. We identified 19 common insertion sites, including Zeb2, Nf1, Mn1, Evi1, Ift57, Mpl, Plag1, Kras, Erg, Vav1, and Gata1. Of note, over 25% of the mice had retroviral integrations near Zeb2, a transcriptional co-repressor in the TGF-beta signaling pathway, leading to over-expression of the Zeb2-transcript. 91% of mice with Zeb2 insertions developed B-lineage ALL suggesting that Zeb2 activation promotes the transformation of CALM-AF10 hematopoietic precursors toward B-lineage leukemias. Over half of the mice with Zeb2 integrations also had Nf1 integrations, suggesting cooperativity among CALM-AF10, Zeb2 and Ras pathway mutations. We then searched for Nras, Kras, and Ptpn11 point mutations in this series of CALM-AF10 leukemic mice infected with the replication competent retrovirus . Three mutations were identified, all of which occurred in mice with Zeb2 integrations, consistent with the hypothesis that Zeb2 and Ras pathway activation promotes B-lineage leukemic transformation in concert with CALM-AF10. In addition, targeted """"""""knock-down"""""""" of Zeb2 led to decreased proliferation of cells with a CALM-AF10 fusion. This data was published in 2010. The prior experiments used either gene targeting or RIM to experimentally induce mutations that we suspect are leukemogenic in combination with an NHD13 or CALM-AF10 transgene. As a complementary approach, we searched for spontaneous (ie, not induced by RIM or gene targeting) mutations that might collaborate with the NHD13 or CALM-AF10 fusions. We searched for mutations of Runx1, Npm1, Tp53, Flt3, Kit, Nras, Kras, and Cbl. We thought this was an important study to help validate murine AML models, because there are no examples of spontaneous N/Kras mutations associated with murine AML. We studied 22-26 mice with each transgene. We found no mutations of Runx1, Npm1, Tp53, or Kit. 25-30 % of both NHD13 and CALM-AF10 mice had Nras or Kras mutations. Intriguingly, almost 30% of the CALM-AF10 mice had Flt3 mutations, inlcuding length mutations, whereas none of the NHD13 mice had Flt3 mutations. One potential explanation for the lack of Flt3 mutations in the NHD13 mice is as follows. Both the NHD13 and CALM-AF10 fusions lead to overexpression of HOXA-cluster genes, specifically HOXA7/9/10, in the bone marrow of clinically healthy (pre-leukemic) mice. However, one distinction is that the CALM-AF10 mice overexpress Meis1, whereas the NHD13 mice downregulate Meis1. Therefore, the lack of Flt3 mutations in the NHD13 mice may be due to the lack of Meis1 expression in the NHD13 mice, as Meis1 is reported to drive Flt3 expression. Given that Nras, Kras, and Flt3 mutations have been shown to enhance proliferation, these results support a working hypothesis that predicts AML cells have one mutation which impairs differentiation (such as NHD13 or CALM-AF10), and a second, complementary mutation which enhances proliferation or inhibit apoptosis. We have used gene expression arrays to compare and contrast genes that are differentially expressed between WT BM, NHD13 leukemias, and CALM-AF10 leukemias. Intriguingly, we find that NMYC, a gene not previously implicated in AML, is markedly overexpressed in a subset of CALM-AF10 leukemias. Preliminary experiments suggest that simultaneous overexpression of the CALM-AF10 and NMYC proteins in vitro lead to increased proliferation and immortalization of murine BM cells. In addition, we find that a subset of NHD13 leukemias spontaneously upregulate Meis1, and Flt3. These studies support recently published data which suggests that Flt3 is downstream of Meis1;we are currently searching for flt3 mutations in this subset of samples. An abstract describing these findings was presented in 2010, and a manuscript describing the findings is in preparation. In collaboration with Dr. Paul Meltzer, we have used multiplex PCR and deep sequencing to identify mutations in candidate genes in a set of 152 mouse leukemias. The genotypes of the leukemias are NHD13, CALM-AF10, NUP98-PHF23, and Lin28b;24 genes commonly mutated in AML and/or T-ALL, including Nras, Kras, Tp53, Notch1, Runx1, Kit, and Flt3 are assayed. The initial samples have been run and the data acquired;the data is currently being assembled and interpreted.
Showing the most recent 10 out of 23 publications
