Investigation of mechanisms regulating leukemogenesis using mouse xenograft models of human acute myeloid leukemia

Acute myeloid leukemia (AML) is a malignant neoplasia of the blood system and can occur in people at all ages but more frequently affects people older than 65 years. Although the outlook for patients with AML has improved over the past decades, still more than half of young adult and about 90% of elderly patients die from their disease. The main obstacles to cure are refractoriness to initial induction treatment and, more frequently, relapse after apparent remission. From a cellular perspective, relapse is thought to occur from rare cell populations of so-called “leukemia stem cells” (LSCs) which share molecular and functional features with their healthy counterparts, the hematopoietic stem cells (HSCs). Because they underlie complex regulatory mechanisms, particularly also involving other cell types and niches, LSCs need to be studied in vivo. For human LSCs, such studies are confined to xenotransplantation models, which usually are performed in immunosuppressed mice. In this thesis, a better understanding of human AML was built up by (1) establishing and improving a stable xenotransplantation model, (2) by identifying a novel LSC marker based on the concept of immune escape and (3) by unraveling novel non-cell autonomous pro-leukemogenic mechanisms involving BM niche and healthy hematopoietic stem and progenitor cells (HSPCs) modulation by CDX2 expressing leukemic cells. Lastly, we use an ex vivo 3-dimensional niche surrogate for expansion of healthy HSPCs. Firstly, an improved transplantation protocol of human AML cells into immunosuppressed NOD/SCID/IL2Rγnull (NSG) mice that mimics the clinical course of the disease in patients was established. In this model, the latency of AML cell engraftment in mice depends on molecular risk groups established in patients and xenogeneic leukemic cells show conserved genetic and phenotypic features. Most importantly, the model enables the engraftment of favorable risk AML subtypes previously considered non-engraftable in NSG mice, opening up new perspectives for in vivo studies on these disease subtypes. We further optimized this model by observing that transplantation at night or under enhanced catecholamine activity favors homing and adhesion of leukemic cells to BM niches, thereby shortening time-to-leukemia in vivo. Next, this model was used for mechanistic in vivo studies on leukemia initiation. In close collaboration with the immunology research group led by Prof. Helmut Salih (University of Tuebingen, Germany), NKG2D-associated im-mune privilege was identified as a feature of human AML LSCs. These findings offer a novel method for LSC isolation in several subtypes of AML and demonstrate in functional assays an unrecognized and targetable mechanism for sensitizing LSCs to immune control. Furthermore, our studies on the transcription factor CDX2, which is expressed in >80% of AML but not detectable in healthy HSPCs, unraveled a novel non-cell-autonomous role by which CDX2 promotes leukemogenesis, namely via DKK1 secretion to outcompete resident healthy HSPCs, which indicates treatment with WNT agonists as potential strategy to treat incipient AML (e.g. AML at minimal residual disease stage). Last, as part of a collaborative project with the research group led by Prof. Ivan Martin (Department of Biomedicine, University of Basel), an ex vivo BM niche surrogate system was analyzed for its potential to expand and analyze healthy cord-blood derived HSPCs in vitro, which could also be adapted to recapitulate pathological situations as leukemia. This could represent a pow-erful tool with wide range of applications, from the identification of factors deregulating niche or blood functions, to the screening of drugs to predict pa-tient-specific response to defined treatments without using xenotransplanta-tion models