Mouse Versus Human Extrinsic Cues Dictate Transformation Potential In BCR-ABL/BMI1-Induced Leukemia In Humanized Xenograft Models

The NSG [NOD/Lt-scid/IL2Rγnull] xenotransplantation mouse model is currently the model of choice to evaluate human hematopoietic engraftment and to study development of human leukemia. Indeed, we have previously shown that co-expression of BCR-ABL together with the polycomb repression complex 1 (PRC1) member BMI1 in human cord blood (CB) derived CD34+ cells was sufficient to induce a serially transplantable lymphoid leukemia (Rizo et al., Blood 2010). This leukemia was characterized by high levels of CD34+/CD19+/CD20-/IgM-/CD33-/CD15- lymphoid blasts in the bone marrow and a high degree of infiltration of blasts in spleen and liver. Clonal analysis revealed that similar clones gave rise to leukemia in primary and secondary recipients. Although in vivo no myeloid leukemias were observed, in vitro both lymphoid as well as myeloid immortalized long-term cultures could readily be established, in line with phenotypes observed in chronic myeloid leukemia patients whereby a chronic myeloid phase can egress into a myeloid or lymphoid blast crisis. It is very plausible that differences between murine and human hematopoietic stem cell niches underlie these observed differences. Human engraftment in NSG mice is typically lymphoid biased, and since many growth factors and cytokines are species-specific it is clear that the murine niche is not ideal to evaluate human hematopoietic engraftment and leukemic transformation potential.

In our current study we have evaluated the in vivo leukemic transformation potential of human CB derived CD34+ cells expressing BCR-ABL and/or BMI1 in NSG mice in which scaffolds coated with culture-expanded human mesenchymal stromal cells (MSCs) were implanted subcutaneously 8 weeks prior to injection of transduced cells, to allow the development of a humanized niche containing mineralized bone-matrix, osteoblasts, stromal cells, as well as appropriate vascularization (Groen et al., Blood 2012). BCR-ABL/BMI1 transduced human CB derived CD34+ cells or primary blast crisis CML patient cells were injected either intravenously or directly into the humanized scaffolds, and leukemia development was evaluated.

Our data indicate that in a humanized niche, in contrast to a murine niche, BCR-ABL was sufficient to induce leukemia as a single hit without overexpression of exogenous BMI1. Furthermore, both ALL as well as erythro/myeloid leukemias could be induced. The ALL could be transplanted to secondary recipients and besides the lymphoid marker CD19, the cells also expressed CD33 and CD15, but not CD11b or GPA. These data are in sharp contrast to results obtained in xenograft mouse models without human niches, where BCR-ABL expression alone in human cells was not sufficient to induce leukemia, and secondary hits such as BMI1 were essential.

Efficient engraftment of a blast-crisis CML patient sample was also observed in the human niche model, whereby the immature blast-like phenotype was maintained in the human scaffold niche, while more differentiated cells were observed in the mouse bone marrow niche. In vitro, long-term self-renewing cultures could readily be established with cells retrieved from the human scaffold niche of these leukemic mice, while no long-term cultures could be initiated with cells retrieved from the murine bone marrow niche, from the same mouse. These data indicate that a human niche is required to maintain appropriate in vivo self-renewal of human BC CML cells. Interestingly, the endogenous BMI1 levels were significantly higher in cells retrieved from the human scaffold niche as compared to the mouse BM niche.

In conclusion, our data indicate that BCR-ABL transformed cells needs secondary event such as over expression of oncogene like BMI1 for its full transformation potential, most likely to overcome or repress oncogene-induced senescence. The mouse environment is not able to provide these secondary events in human cells, whereas the human niche is able to provide signals that together with BCR-ABL are sufficient to fully transform human cells in xenograft models.