The ultimate interest of biomedical research is the study of humans and the goal is the improvement of human health. However, research on humans is severely compromised by the inability to perform invasive experimental work in order to establish the cause and effect relationships between processes. Mouse models play an important role in characterizing key aspects of the driving forces of malignant transformation and disease in humans. However, they rarely represent the genetic complexity and clinicopathologic characteristics of human disease. Xenotransplantation of human cells into mice is regularly used to study the biology of human cancer and test the efficacy of novel anti-cancer therapies in vivo. But its value is limited because scientists are currently mainly limited to aggressively growing primary tumors or cell lines that are less susceptible to xenorejection and less dependent on critical signals provided by the microenvironment. Therefore there is an unmet need to develop methods to reliably grow primary human tumor cells in mice. In the here proposed project, humanization of the microenvironment will be achieved by replacement and controlled expression of essential, but not cross-reactive, growth and survival factors in the mouse genome with their human counterparts. Furthermore, novel approaches to reduce xenorejection by myeloid cells and to generate space in the bone marrow niche complete the proposed genetic manipulations. These novel in vivo models will then be used to test and compare as a benchmark acute myeloid leukemia (AML) engraftment, and to establish models for primary chronic myeloid leukemia and myeloproliferative neoplasias, as well as multiple myeloma - diseases in which the lack of in vivo models faithfully reflecting these diseases curtails research. Successful establishment of these models would allow detailed studies of biology as well as validation of approved and experimental therapies. The development of a versatile platform allowing the study of human primary malignancies in vivo in genetically modified mice would represent a major advance for the field. In summary, this model will serve as platform for preclinical therapy testing with increased predictive value and allow studies on biology of the dynamic evolution of neoplasias with an immediate impact on clinical practice.