Unexpected variation in leukemia stem cell frequency and genetic heterogeneity in two murine leukemia models initiated by AML1/ETO9a and CALM/AF10

Abstract

Acute myeloid leukemia (AML) is an extremely aggressive, genetically heterogeneous disorder requiring the acquisition of multiple (two to eight) genetic aberrations which act in concert and result in the malignant transformation of a hematopoietic cell [1]. Advances in molecular techniques, particularly the advent of next generation sequencing, has led to the identification of a plethora of genetic lesions and greatly increased our understanding of the molecular heterogeneity of AML [2]. However, one of the major challenges is to determine which of the identified mutations contribute to the disease process as so-called driver mutation. Further, to add to the complex genomic landscape of AML, it is now well known that heterogeneity is found not only between patients (intertumor heterogeneity) but also within individual leukemias (intratumor heterogeneity) [3]. The leukemia stem cell model provides an explanation for the functional and phenotypic heterogeneity observed within a tumor. The model posits that within the leukemic bulk, only a small subset of cells, the leukemia stem cells (LSC) possess the ability to propagate the leukemia phenotype, and are largely responsible for therapy failure and relapse in leukemia patients [4, 5]. Although a number of animal models of AML have been established, there remains a lack of clinically relevant models that recapitulate this intra and intertumor heterogeneity observed in human AML. In this study, we established two retroviral transduction, murine bone marrow transplantation leukemia models (MBMTM), in order to mimic the genetic, functional, and phenotypic heterogeneity of AML. The first model used the AML1/ETO9a (A/E9a) fusion, an alternatively spliced isoform of AML1/ETO (A/E), also known as RUNX-RUNX1T1, as a basal driver oncogene (Fig. S1a) [6]. The second model used the CALM/AF10-minimal fusion (C/A-MF), which contains the leukemogenic regions of CALM/AF10 (C/A) also known as PICALM/MLLT10 (Fig. S1b), a rare but recurring fusion gene observed in AML, acute lymphoblastic leukemia, and malignant lymphoma [7–9]. For the MBMTM, transduced bone marrow (BM) cells expressing either A/E (full length), A/E9a, C/A-MF, or the empty pMIG vector (controls), along with nontransduced BM cells, were transplanted into lethally irradiated recipient mice. Primary mice transplanted with A/E9a or C/A-MF transduced BM cells developed leukemia after a median latency of 142 and 91 days, respectively. In comparison, none of the mice transplanted with BM cells transduced with full length A/E or the empty retrovirus (MIG controls) developed leukemia (Fig. 1a, b). Postmortem analysis of the C/A-MF and A/E9a primary transplants revealed splenomegaly, leukocytosis, anemia, and infiltration of blast like cells in the spleen, liver, kidney, and lung (Figs. S2–4). Immunophenotypic analysis of the A/E9a primary transplanted mice revealed an increase in frequency of lineage low, cKit positive cells, but not myeloid or lymphoid subpopulations compared with wild-type controls (normal C57BL/6 mice; WT) (Fig. S5 and Tables S1, S2). Analysis of the C/A-MF leukemias revealed high GFP expression in 20 of the 21 primary samples, all of which These authors contributed equally: Rhea Desai, Sarvenaz Taghavi