OS10.5.A Modeling immunocompetent tumor microenvironment in glioblastoma patient-derived orthotopic xenografts

Abstract

Abstract Background To date, glioblastoma (GBM) remains a fatal disease, with a median overall survival of roughly over a year. There is a crucial need of new treatment options, yet most clinical trials have failed partly due to the lack of predictive preclinical model systems. Currently, most patient-derived preclinical models suffer from the reduction or absence of immune system components, which represents a bottleneck for adequate immunotherapy testing. Humanized mice offer new opportunities here, since they rebuild an adaptive human immune system in a NSG mouse. Derivation of glioblastoma patient-derived orthotopic xenografts (PDOXs) in humanized mice appears thus as a promising tool for testing new treatment strategies targeting the tumor microenvironment (TME). Material and Methods We derived PDOXs through intracranial implantation of GBM primary organoids in different immunocompromised mouse strains (Nude, NOD/SCID, NSG). To introduce back the adaptive human immune system, GBM PDOXs were further derived in human CD34+ hematopoietic stem cell-engrafted NSG (HU-CD34+) mice. We applied single-cell RNA-sequencing, multicolor flow cytometry, immunohistochemical analyses and functional studies to examine the heterogeneous TME in a cohort of GBM PDOX models. We further interrogated the contribution and crosstalk between the human and mouse components constituting the brain TME in HU-CD34+ PDOXs. Results We show that glioma PDOXs can be derived in mice of different background including Nude, NOD-SCID, NSG and HU-CD34+ mice. Mouse-derived TME created in PDOX models contains tumor-associated macrophages (TAMs) known as major immuno-suppressive components of human GBM tumors. We further show that PDOXs derived in HU-CD34+ NSG mice present human CD45+ immune cells in the bone marrow and blood. Interestingly, we detect an influx of human immune cells in tumors developed in the mouse brain, which interact with the brain-derived immunosuppressive TME of mouse origin. Conclusion We here provide a thorough characterization of the heterogeneous brain TME created in GBM PDOX models. We show that human GBM can instruct mouse-derived brain cells towards immune-suppressive TME. The missing adaptive immune component can be introduced by derivation of GBM PDOXs in humanized mice. Such immunocompetent in vivo models will be important for testing novel therapies targeting different immune components in GBM.