Identification of human juvenile chronic myelogenous leukemia stem cells capable of initiating the disease in primary and secondary SCID mice.

Most juvenile chronic myelogenous leukemia (JCML) cells have limited long-term proliferative capacity, and only a minority of immature cells give rise to colonies in semisolid cultures. Clonogenic JCML progenitors cannot be maintained in culture because they differentiate, and within a few weeks the leukemic clone is lost. This makes it difficult to identify the cell that initiates and maintains the disease in patients. To determine the proliferative capacity of JCML cells in vivo, bone marrow (BM), peripheral blood, or spleen cells from eight patients with JCML either at diagnosis or during treatment were transplanted into sublethally irradiated severe combined immune deficient (SCID) mice. JCML cells from all patients homed to the murine BM and proliferated extensively in response to exogenous stimulation with granulocyte-macrophage colony-stimulating factor. Within a few weeks, highly engrafted mice became ill and cachectic due to infiltration of leukemic cells and secretion of tumor necrosis factor-alpha. Murine BM, spleen, and liver were infiltrated with leukemic blasts, and typical JCML colony-forming progenitors could be recovered. Kinetic experiments demonstrated that only a small minority of transplanted cells homed to the murine BM, and that these cells initiated and maintained the disease in vivo by extensive proliferation and differentiation. To characterize the cell-surface phenotype of the JCML initiating cell (JCML-IC), JCML blood or spleen cells were fractionated on the basis of CD34/CD38 marker expression and transplanted into SCID mice. Only immature CD34+ cells could initiate the disease, while mature CD34- cells did not engraft. Within the CD34+ compartment, there was enrichment for JCML-ICs by immature cells with a CD34+/CD38- stem-cell-like phenotype. Mice transplanted with more mature CD34+/CD38+ populations that also contained clonogenic JCML progenitors were poorly engrafted. These results indicate that the JCML-IC is an earlier stage of development than clonogenic JCML progenitors. Additional evidence that the JCML-IC has stem-cell properties comes from secondary transplant experiments that test the self-renewal capacity. The JCML-IC from all three patients tested could successfully reinitiate the disease in secondary murine recipients. Thus, we have developed a functional in vivo model that replicates many aspects of human JCML, and have used this model to identify and characterize JCML-ICs and their stem-cell properties.