Identification of cooperating genetic events in acute leukemia

The genetic alterations associated with acute leukemia can be divided into two functional groups. The class I mutations enhance cellular proliferation and survival by constitutive activation of mainly protein tyrosine kinases signaling pathways. In contrast, the class II mutations frequently involve transcriptional regulators of normal hematopoietic differentiation, and result in a block of hematopoietic cell maturation and/or aberrant self-renewal capacity. However, expression of the class II mutations in the murine hematopoietic system often leads to myelodysplasic changes and acute leukemia after a long latency, suggesting that collaboration of additional genetic alterations might be required.
To identify potential cooperating genetic events facilitating induction of acute leukemia by the class II mutations, we modeled acute myelogenous leukemia (AML) in mouse by retrovirally expressing the MLL/ENL and MOZ/TIF2 fusion genes in the bone marrow. The genetic background of leukemic mice (F1 hybrid mice resulting from crossbreeding the FVB/N and 129/s1 strains) allowed us to perform a genomewide polymorphism analysis screening for loss of heterozygosity (LOH) that is frequently found in blasts from AML patients. However, a simple sequence length polymorphism (SSLP) based allelotyping as well as a mouse 5K single nucleotide polymorphism (SNP) array analysis did not showed any LOH, indicating that large scale of LOH might be a rare event in our mouse leukemia models.
Recent studies have shown that genomic insertion of retrovirus could influence the expression of adjacent genes and therefore contribute to oncogenic transformation of hematopoietic cells. The retroviral integration tagging approach has been widely used for seeking new proto-oncogenes or tumor suppressor genes, and especially for identifying collaborative events in tumor models that already harbored an initiating oncogenic event. To test the hypothesis that the integration of the provirus could act as cooperating events in our mouse leukemia models, we characterized the retroviral integration sites from 21 leukemic mice induced by retroviral expression of MLL/ENL and MOZ/TIF2 fusion genes. Sixty-six integration flanking genes were identified, and most of them have been previously linked to tumorigenesis. Further determination of their expression levels demonstrated that integration flanking genes like Tcf7, Tnfrsf1, Mn1 and Lhx2 were up-regulated, whereas Pur , Ppp2r5c, Runx3, Socs1 and Prdm2 were down-regulated. Interestingly, in a MLL/ENL leukemic mouse carrying the integration adjacent to the meningioma 1 (Mn1) gene, the clone harboring the Mn1 integration prevailed over other co-existing clones harboring different integrations. Moreover, an in vitro cellular proliferation assay showed that the overexpression of MN1 significantly enhanced proliferation and self-renewal capacity of primary bone marrow cells. These findings suggested that MN1 possesses leukemic transforming potential and might functionally collaborate with the MLL/ENL fusion in the development of the acute leukemia. In order to experimentally address this hypothesis, we performed a series of bone marrow transplant experiments. Indeed, co-expression of MN1 with MLL/ENL enhanced in vivo disease development, and resulted in a significantly reduced latency for induction of an aggressive acute leukemia than expression of MN1 or MLL/ENL alone. In addition, co-expression of MN1 increased the granulocytemacrophage progenitor (GMP) cell population expressing Gr1/Mac1, Cd34 and c-Kit with leukemia-initiating properties as shown in secondary transplantation experiments. As MN1 has been previously proposed to exert its function as a transcriptional co-activator, we also aimed to identify the potential downstream target genes by transient MN1 expression in primary bone marrow cells. Gene expression profiling experiments revealed a series of genes with known roles in normal or malignant hematopoiesis such as CD34, FLT3, HLF, and DLK1 that were upregulated in MN1 overexpressing murine leukemias, as well as pediatric acute leukemias with high MN1 levels. We also determined the MN1 levels in a large panel of pediatric acute leukemias. High MN1 expression levels were observed in 50 of 87 samples: high MN1 levels were found in a large proportion of B-cell ALL cases and in most infant leukemias that carry MLL fusions and are of B-cell origin. Additionally, siRNA-mediated MN1 knockdown resulted in cell cycle arrest and impaired clonogenic growth of human leukemia cell lines with high MN1 levels but not in cells with low (or undetectable) MN1 levels, suggesting the aberrant expression of MN1 contributes to malignant cellular proliferation, and the inhibition of MN1 could represent a new therapeutic approach. Taken together, while searching for cooperative genetic alterations in murine leukemias, we found that elevated levels of MN1 oncogene can act as a functional collaborator in MLL/ENL (and probably other class II mutations) induced leukemia through a distinct genetic program that increases the leukemia stem cell pool. In addition, we also demonstrated for the first time that high MN1 levels are found in a significant fraction of childhood acute leukemias, and important for proliferation of the leukemic cells.