Bioluminescent Imaging of Human Leukemic Stem Cell Engraftment.

Previously, we found that candidate leukemic stem cells (LSC) involved in chronic myelogenous leukemic (CML) progression to myeloid blast crisis (BC) shared phenotypic characteristics with granulocyte-macrophage progenitors (GMP). However, CML GMP had activated a key self-renewal gene - beta-catenin. Aberrant in vitro self-renewal capacity could be specifically inhibited with axin - a potent beta-catenin antagonist (Jamieson et al, New Engl J Med2004;351:657–67). In order to determine whether these candidate LSC had enhanced in vivo self-renewal potential, we FACS-sorted hematopoietic stem cells (HSC), common myeloid progenitors (CMP), GMP, megakaryocyte-erythroid progenitors (MEP), CD34+CD38+ cells and blasts (lineage-positive cells) from advanced phase CML versus normal bone marrow or cord blood and transplanted them intrahepatically into newborn T, B and NK cell deficient (RAG2−−-γ−/−) mice (Traggiai et al. Science2004;304:104–7). Engraftment of human (ge;1%) CD45, CD19, CD3, and CD14-positive cells in the hematopoietic organs including bone marrow, liver, spleen and thymus of recipient animals was analyzed by FACS and compared with non-transplanted controls. In seven transplantation experiments performed with normal cord blood or bone marrow (n=24 mice), populations enriched for HSC, showed evidence of long-term engraftment, while committed progenitors including GMP did not. Conversely, in six experiments with myeloid BC CML (n=28 mice), GMP gave rise to long-term engraftment (7 of 11 mice) more frequently than HSC (2 of 6 mice) and blasts seldom engrafted (2 of 7 mice). These results suggested that LSC were enriched within the GMP fraction of myeloid BC CML. Subsequently, bioluminescent imaging (IVIS 100) was employed in order to track the kinetics of normal versus LSC engraftment more precisely. In 7 experiments involving normal marrow or cord blood (n=28 mice) and 3 experiments with advanced phase CML (n=18 mice), HSC, progenitor and blast (Lin+) populations were transduced with a lentiviral luciferase GFP vector and transplanted intrahepatically into newborn RAG2−/−γ−/− mice. Engraftment was monitored by weekly bioluminescent imaging as well as by tail vein bleeds to detect GFP expression. When mice were sacrificed, human engraftment in the liver, spleen, bone marrow and thymus was assessed by FACS analysis and sorted human CD45+ cells were transplanted into secondary recipients (n=4 experiments). In primary bioluminescent transplantation studies, CML HSC, CMP and GMP engrafted. Normal HSC demonstrated serial engraftment potential while more committed normal progenitors such as CMP, GMP and MEP did not. In contrast, CML blast crisis GMP demonstrated serial (2o and 3o) engraftment potential suggesting that they had gained the capacity to self-renew in vivo and thus, behaved like LSC. Hence, bioluminescent imaging of LSC engraftment in a highly immunocompromised mouse model can be used to detect LSC and may be utilized for pre-clinical evaluation of the effects of molecularly targeted therapy on LSC. Figure 1. Bioluminescent imaging was performed with the aid of a Xenogen™. IVIS 100 imaging system at 9 weeks post-transplant. Upper: RAG2−/γ0−/− mouse transplanted with no cells. Lower: Bioluminescence of 2° human CD45+GFP+ cells derived from mice transplanted with CML blast crisis GMP (mouse 1) or normal HSC (mouse 2) were compared with mice transplanted with primary normal Figure 1. Bioluminescent imaging was performed with the aid of a Xenogen™. IVIS 100 imaging system at 9 weeks post-transplant. Upper: RAG2−/γ0−/− mouse transplanted with no cells. Lower: Bioluminescence of 2° human CD45+GFP+ cells derived from mice transplanted with CML blast crisis GMP (mouse 1) or normal HSC (mouse 2) were compared with mice transplanted with primary normal