Use of CALM-AF10 and NUP98-HOXD13 fusions to Identify Oncogenic Pathways That Collaborate with HOXA9 Overexpression

Abstract 2499

Acute myeloid leukemias (AMLs) are thought to result from a series of complementary mutations that affect several vital pathways, including differentiation, self-renewal, apoptosis, and proliferation. Several mutations associated with AML (MLL fusions, NUP98 fusions, CALM-AF10 fusion, and NPM1 mutations) are thought to impair hematopoietic differentiation by dysregulation of target genes; one of the most consistently dysregulated target genes is HOXA9. Mice that express a CALM-AF10 (CA10) or NUP98-HOXD13 (NHD13) fusion gene overexpress HOXA9 and develop AML with a delayed onset and incomplete penetrance, suggesting the need for complementary mutations. We have used three techniques including candidate gene resequencing, retroviral insertional mutagenesis (RIM), and gene expression profiling to compare and contrast the complementary mutations that occur in these mice. The frequency of Ras pathway mutations is similar for both CA10 (20%) and NHD13 (32%) leukemic mice, and previously reported RIM studies identified common integration sites near Mn1 and Evi1 in both models, suggesting that dysregulation of similar collaborative pathways can lead to leukemia in both models. However, there were significant differences between these models. The aforementioned RIM studies identified Zeb2 as the most frequent insertion site for CA10 leukemias, suggesting that Zeb2 collaborated with CA10, whereas no Zeb2 insertions were found in the NHD13 RIM study. Another distinction between the NHD13 and CA10 leukemias is the expression of the HOXA9 co-regulatory factor Meis1. Meis1 is markedly upregulated in bone marrow from clinically healthy, pre-leukemic CA10 mice and CA10 leukemic tissues compared to wildtype bone marrow. Conversely, Meis1 expression was decreased in bone marrow from clinically healthy NHD13 mice compared to wildtype, and only occasionally overexpressed in NHD13 leukemias. Interestingly, RIM analysis of NHD13 mice identified Meis1 as the most frequent insertion site, suggesting that Meis1 overexpression can collaborate with NHD13 during leukemic transformation. Flt3, which is frequently mutated in patients with AML, has been shown to be transcriptionally activated by Meis1. Of note, Flt3 is consistently overexpressed in CA10 leukemias but is overexpressed only in those NHD13 leukemias that have upregulated Meis1. Intriguingly, the frequency of Flt3 activating mutations in CA10 leukemias (which overexpress Flt3) was 25%, but none were found in the NHD13 leukemias (which do not overexpress Flt3). Lastly, using gene expression arrays, N-myc was identified as one of the most differentially overexpressed mRNAs, suggesting that it may cooperate with CA10 during leukemic transformation. Furthermore, N-myc has recently been shown to promote AML in mice, and our preliminary experiments suggest that N-myc and CA10 collaborate and lead to increased cell number and replating potential in bone marrow colony formation assays. In conclusion, we have taken several genetic approaches to identify key similarities and differences in the biology of two leukemic models (CA10 and NHD13), both of which overexpress HOXA9.