The formal demonstration of leukemia stem cells (LSC) relies on the ability of primary human leukemias to engraft in transplanted xenohosts with the subsequent demonstration of serially transplantable tumors to establish the presence of a self-renewing population. Several immunodeficient and transgenic mouse strains have been utilized as xenohosts for human AML (e.g. NOD, NOD/SCID, NOD/SCID,β2m); however, robust engraftment frequently requires that the mouse recipients be pre-treated with antibodies against mouse immune cells or that leukemic blasts and/or mouse hosts are treated with exogenous hematopoietic growth factors before and/or during transplantation. While such treatments result in a higher percentage of engrafting AML’s, the overall effect on leukemic blasts, particularly the leukemic stem cell population, is unknown. In order to avoid these confounding factors as well as to potentially develop a more robust model of xenotransplantation, we have evaluated the use of the RAG2/gamma common chain double knockout mouse model to xenotransplant human AML’s. Using primary human samples (n=32, patient age range 1–86, median 55), we have determined that newborn pups intravenously injected via the facial vein efficiently engraft human AML’s without the use of additional growth factors or depleting antibody therapy. Among the AML’s tested, 65.6% (21/32) engrafted in mice as assessed by >5% human chimerism in either the peripheral blood, spleen, or bone marrow at 12 weeks post-transplant. In contrast to other xenotransplantation mouse models, circulating leukemic cells were typically identified in the circulation by 4 weeks post-transplant and were present in stable or increasing numbers in the circulation until engraftment was assessed (12 weeks) or until recipient mouse demise. Engrafting AML’s were composed of both CD34+ and CD34- AML’s and included all WHO/FAB subtypes except for acute promyelocytic leukemia. These AML’s included de novo and relapsed AML, as well as therapy-related AML and AML arising from pre-existing MDS or CMPD. While the number of samples tested was limited, there was no statistically significant association between engraftment and WBC at diagnosis, cytogenetic risk group, MLL or FLT3 status, age, or achievement of CR following induction therapy; however, among CD34+ AML’s the percentage of putative leukemic stem cells (defined as CD34+CD38- leukemic blasts), showed a statistically significant correlation with engraftment (student T-test p < 0.05). All CD34+AML’s containing >18% CD34+/CD38− leukemic blasts engrafted in the transplanted hosts. Overall, these findings demonstrate that the RAG2/gamma common chain double knockout mouse is a robust system for studying human AML and will be a useful model for future functional studies of human AML LSC’s.

Disclosures: Dr. Weissman - Consulting for Cellerant, Inc. and StemCells, Inc.; Dr. Weissman - Stockholder in Amgen, Cellerant, Inc., and StemCells, Inc.; Dr. Weissman - Former Scientific Advisory Board member for Amgen.

Author notes


Corresponding author