Targeting constitutively activated FLT3 (FLT3-ITD) by tyrosine kinase inhibition (TKI) in acute myeloid leukemia (AML) leads to clearance of blasts in the periphery but not in the bone marrow, suggesting a protective effect of the marrow niche on leukemic stem cells (LSC). We have previously shown that interaction of CD34+FLT3-ITD+ LSC with stromal niche cells mimicking the bone marrow environment specifically protects these cells from the effects of TKI and confers a growth advantage to FLT3-ITD+ leukemic stem/progenitor cells over normal ones (Parmar et al, Cancer Research 2011). To study human FLT3-ITD+ LSC in vivo in the context of the bone marrow niche, we aimed to establish a xenogeneic NOD/SCID mouse model of human FLT3-ITD+AML.
Human CD34+ enriched cord blood cells were transduced with a pWPI lentivirus containing a VSV-pseudotyped SIN/LTR vector with eGFP and full length human FLT3 cDNA harboring a 30 bp length internal tandem duplication (FLT3-ITD) or empty vector control. Transduction efficiency ranged between 1–4.4% for FLT3-ITD and 1.3–18% for vector control. Sub-lethally irradiated NOD/SCID mice were then transplanted with 1 × 104 – 7 × 104 unselected or GFP-sorted CD34+FLT3-ITD+ cells. Acute leukemia developed in 7/9 animals after a median latency of 85 days (range 70–168), with involvement of peripheral blood, bone marrow, spleen and liver. Three mice developed acute lymphoblastic leukemia (ALL) whereas the remaining mice showed signs of AML. In contrast, mice receiving empty pWPI vector-transduced human CD34+ cord blood cells (n = 8) all remained healthy during the observation period of 28 weeks and, in 4/8 animals, normal human CD34+cells could be recovered from the bone marrow (human engraftment range 0.3–35.5%). Leukemic mice exhibited hepatomegaly and splenomegaly with an average 10-fold increase in spleen weight, 2-fold increase in spleen length and 2.7-fold increase in liver weight compared to control mice. In mice that developed ALL, lymph node enlargement was also noted. Whole bone marrow, spleen and liver cells from primary mice were re-transplanted and were able to reproduce acute leukemia in all secondary (n=10/10) and tertiary mice (n=11/11) with a median latency of 25 and 20 days, respectively (p<0.01).
Surprisingly, detailed immunophenotypical and immunohistochemical analysis revealed all leukemias to be of murine origin. Leukemic cells stained positively for murine CD45.1 antigen but negatively for human CD45. However, a small population of human CD34+CD45+ cells (range 1–7%) was continuously detectable in the bone marrow of primary, secondary and tertiary transplanted leukemic mice. Accordingly, human FLT3-ITD was detectable by PCR specific for human FLT3 up to the third serial transplantation. Viral integration site analysis by LM-PCR on genomic DNA isolated from spleens of leukemic mice revealed lentiviral integration into the human genome, excluding the possibility of in vivo viral shuttling from human cord blood CD34+ cells to mouse hematopoietic stem cells. Moreover, multicolor fluorescent in situ hybridization (M-FISH) on metaphases generated from peripheral blood lymphocytes revealed only murine chromosomes, also ruling out the possibility of fusion between human and mouse cells. To further characterize these murine leukemias, we performed array CGH on murine spleen gDNA of four immunophenotypically different mice. All mice showed recurrent clonal chromosomal aberrations frequently found in AML. Surprisingly, we have no evidence for the presence of FLT3-ITD in these murine leukemias, suggesting that CD34+FLT3-ITD+ stem cells can trigger development of acute leukemia. We propose that leukemogenesis may mechanistically be related to the host microenvironment and that the bone marrow niche in NOD/SCID mice is susceptible to modulation by the FLT3-ITD oncogene.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.