Retroviral insertional mutagenesis (RIM) has proven to be a valuable whole-genome screen for the identification of genes involved in malignant transformation. Since the survival curve for nUP98-HOXD13 (NHD13) transgenic mice suggested that additional mutations were required for leukemic transformation, we used RIM to identify genes that might collaborate with NHD13. Newborn NHD13 and WT littermates were infected with the MOL4070LTR retrovirus. Onset of leukemia was accelerated, suggesting a synergistic effect between the NHD13 transgene and genes at the retroviral insertion sites20. We cloned 279 individual insertion sites from 31 mouse leukemias. Common insertion sites (CIS) are regions that have retroviral integrations in more than one leukemic sample, have been biologically selected as dominant clones, and are thought to harbor genes important for malignant transformation. The two most frequent CIS in the study were Meis1 and Mn1. Of note, Meis1 is a well-known CIS, and has been shown to collaborate with HOX and NUP98-HOX fusion genes during malignant transformation. Mn1, not previously identified as a CIS, is involved in human leukemia through fusion with the ETV6 gene, and has recently been shown to have prognostic importance for patients with AML. Other CIS identified were Gata2, Erg and Epor. Erg has also been shown to be of prognostic importance in AML. Of note, Erg has been identified as part of a Meis 1 signature, composed of genes activated by Meis156, and Erg integrations were mutually exclusive of Meis1 integrations. In addition, we identified a CIS that was more than 100 kb from the nearest coding gene, but within 20 kb of the miR29a/miR29b1 microRNA locus. Both of these miRNA were upregulated, demonstrating that RIM can target miRNA loci as well as protein-coding loci. As a complement to the whole genome RIM screen described above, we re-sequenced candidate genes to identify spontaneous mutations that might collaborate with genetically engineered mutations. Because mutations in the NOTCH1 heterodimerization (HD) and PEST domains are common in patients with pre-T LBL60, we assayed mouse pre-T LBL and identified mutations in 33/41 SCL/LMO1, 3/6 OLIG2/LMO1, and 3/9 NHD13 pre-T LBLs. All of the PEST domain mutations resulted in frameshifts, and were clustered in two mutation hotspots; all the HD mutations were missense mutations61. In contrast to NOTCH1 mutations in pre-T LBL patients, no mouse samples had mutations in both the HD and PEST domain of Notch1. In an analogous fashion, we searched for spontaneous Runx1, Npm1, Tp53, Flt3, Kit, Nras, Kras, and Cbl mutations in NHD13 mouse AML. 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 mouse AML samples and found seven Nras or Kras activating mutations, an absence of Npm1, p53, Runx1, Kit and Flt3 mutations, and a single Cbl mutation. Since Runx1 and Npm1 mutations are thought to be similar to NHD13 and impair differentiation, it was not surprising that we found no mutations in these genes. The absence of Kit mutations was also not surprising, as Kit mutations have been most strongly associated with AMLs that have translocations involving core binding factor (CBF) genes. However, we were somewhat surprised to find no Flt3 mutations, as more than 25% of patients with AML show either FLT3 length mutations or kinase domain mutations63. Given that Nras, Kras, and Cbl mutations have been shown to enhance proliferation, these results support a working hypothesis that predicts AML cells have one mutation which impairs differentiation, and a second, complementary mutation which enhances proliferation or inhibit apoptosis. Ongoing and planned experiments include using RIM to identify genes that collaborate with CALM-AF10 to cause leukemia, to re-sequence candidate genes from CALM-AF10 leukemia samples, and to use gene expression arrays to identify genes and pathways that are upregulated in bone marrow from CALM-AF10 mice compared to bone marrow from clinically healthy CALM-AF10 mice. We will infect IL3-dependent, non-malignant NHD13 cell lines with a lentiviral siRNA library, and determine if down regulation of specific gene(s) results in cells that are IL3-independent and malignant.
| Novak, Rachel L; Harper, David P; Caudell, David et al. (2012) Gene expression profiling and candidate gene resequencing identifies pathways and mutations important for malignant transformation caused by leukemogenic fusion genes. Exp Hematol 40:1016-27 |
| Beachy, Sarah H; Aplan, Peter D (2010) Mouse models of myelodysplastic syndromes. Hematol Oncol Clin North Am 24:361-75 |
