Study of variables determining the engraftment and homing of human acute myeloid leukemia samples in the NSG mouse model

Acute myeloid leukemia (AML) is the most frequent hematopoietic malignancy in
adults, causing death of about 90% of elderly patients. Cytogenetic and molecular
abnormalities are used to categorize AML in favorable, intermediate and adverse risk
groups. The classification in one of these groups will directly affect clinical decisionmaking.
Although molecular criteria have significantly improved prognostication and
thus AML patient stratification and treatments, they still do not allow full risk prediction.
AML is a complex disease with a very high heterogeneity which makes it challenging
to accurately reproduce patient phenotypes in murine models. Patient-derived
xenografts can reproduce this heterogeneity but display inherent limitations including
the long-latency until AML develops in those models.
In this thesis, we aimed to study the variables that determine the in vivo
leukemogenesis in a xenotransplantation model of human AML samples into
NOD/SCID/IL2Rgnull (NSG) mice. We aimed to better understand the impact of the
characteristics of the human AML cells, as well as the effect of modulation of the bone
marrow (BM) microenvironment of the recipient mice on the homing and engraftment
of human AML cells.
Firstly, we could show that the latency of symptomatic AML induction in mice as well
as the homing of transplanted AML cells into the murine BM depends on the molecular
risk group established in patients. We also gained insights in the kinetics of disease
induction by screening for human AML cells in the mice with regular BM biopsies.
Moreover, correlation between some AML patient characteristics (expression of some
particular surface markers, remission status and Flt3 (Fms-like tyrosine kinase 3)
mutational status) and the behavior of the corresponding transplanted AML was
observed. This suggests that this mouse model accurately depicts the clinical course
of the disease and reproduces important features of human AML.
Furthermore, we observed that xenotransplantation performed at night, together with
disturbance of the circadian rhythm in NSG mice, accelerated engraftment and
enhanced homing of transplanted human and murine AML cells compared to
corresponding procedures performed in the late afternoon. Our work suggests that
observed pro-oncogenic effects are mediated by catecholamines and we could
prevent them by beta-blocker treatment. Therefore, varying the transplantation time17
point or inducing a stress response may optimize AML xenograft models and these
observations might be important for the knowledge about the tumor-initiation
processes in murine models and could possibliy be transferred to humans.