Introduction: Time to acute myeloid leukemia (AML) development is not assessable in patients, but a longer time to relapse after apparent remission is observed in patients with favorable risk as compared to adverse risk AML, suggesting differences in in vivo leukemia kinetics. Repopulation of immunodeficient mice remains the primary method to functionally assess human AML cells but published data report engraftment of ~40 to 66% of AML cases, mostly confined to FLT3-mutated intermediate or poor risk subtypes. Here we hypothesized that subsets of AML (e.g. of favorable risk) take longer time to induce detectable leukemia in xenotransplant assays, and that therefore extending post-transplant follow-up (beyond 10 to 16 weeks as used in previous studies) can enhance engraftment efficiency.

Methods: 2 adverse, 11 intermediate I/II, 4 favorable risk AML and 2 APL were transplanted into NOD/SCID/IL2Rγnull (NSG) mice and followed up to 1 year. Briefly, peripheral blood (PB) mononuclear cells (PBMCs) were purified by MACS or FACS and 0.4-1x106 blasts transplanted via tail vein (18/19 AML) or intrafemorally (2/19 AML) in non- (4/19) or sublethally irradiated (15/19) mice. Animals were monitored for disease signs and by routine monthly bone marrow (BM) punctures. Experiments were terminated at sickness or >1% human engraftment and mice analyzed by whole body histopathology, multi-parameter flow cytometry of PB, BM and organs and next generation sequencing (NGS) of xenogeneic leukemic cells. Colony-forming (CFU) capacity and CD34+(CD38-) cell percentages were correlated with molecular risk stratification and engraftment.

Results: Extending follow-up beyond standard analysis end-points improved engraftment and permitted leukemogenesis in 18/19 (~95%) of AML cases, including favorable risk subtypes transplanted without prior irradiation. Indeed, only 7/19 (~37%, termed standard engrafters) of cases engrafted at time-points defined in previous studies, while 11/19 (~58%, termed long latency engrafters) engrafted later than 16 weeks. Importantly, routine BM punctures showed no evidence of leukemic cells at standard analysis time-points in long-latency engrafters, clearly indicating that engraftment with these samples would have been missed using standard protocols.

Consistent with our hypothesis, all favorable risk AML were long latency engrafters. Time to engraftment (22±9 weeks) and mouse survival (23±9 weeks) were influenced by the AML molecular risk group, but not by in vitro CFU or CD34+(CD38-) cell percentage. Engraftment was confirmed by whole body histopathology and multi-parameter flow cytometry showing conserved immune phenotypes as compared with corresponding pre-transplant samples. Importantly, both standard (n=2) and long latency engrafters (n=5), showed high conservation of genetic signatures in xenogeneic leukemic cells, with no signs of clonal selection upon engraftment. In one mouse-derived sample, a de novo occurring ASLX1 mutation was detected at low allelic burden (4%), while in another a low allelic burden NRAS mutation was lost, suggesting that genetic evolution can occur, but however is not a frequent event.

Finally, we applied this model to investigate previously understudied favorable risk AML with inv(16). In spite extended follow-up, engraftment was only observed from CD34+ putative leukemic stem cells (LSCs) - which also in microarrays showed enhanced expression of stem cell genes - but not from CD34- inv(16) AML blasts. These data reinforce the notion of a strict hierarchy in AML and suggest CD34 expression as an LSC marker in this AML subtype.

Conclusions: In sum, we present a model that enables in vivo studies of AML subtypes previously considered non-engraftable (such as favorable risk AML), using widely accessible and well-characterized NSG mice. We show that in vivo xenotransplantation of human AML cells in NSG mice faithfully recapitulates human disease since xenogeneic leukemic cells (1) retain the phenotypic, genetic and functional leukemia-initiating properties of the corresponding pre-transplant AML samples, (2) follow disease kinetics and mortality induction in mice according to molecular risk groups established in humans and (3), importantly, can persist in animals over several months at undetectable levels without losing disease-initiating properties, thus mimicking the clinical course of AML in humans.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.