Converging data from our individual laboratories support the notion that brain cancers (gliomas) are complex and dynamic ecosystems composed of several non-neoplastic cell types critical for glioma formation and maintenance. Specifically, we have shown that one of these non-neoplastic cell types (microglia) in the tumor microenvironment is a critical determinant of glioma cell growth and invasion. To more completely define the molecular mechanisms underlying tumor microenvironment regulation of glioma behavior, we have assembled a highly interactive team of researchers, including one early stage investigator focused on the tumor microenvironment (Dolores Hambardzumyan), one senior scientist expert in microglia-glioma interactions (Helmut Kettenmann), and two established physician-scientists (Eric Holland and David Gutmann) whose laboratories employ genetically engineered mouse models (GEMMs) to evaluate the dynamic interactions between non-neoplastic and neoplastic cells in glioma. Based on experimental findings from each of our individual laboratories, we hypothesize neoplastic glia (glioma cells) recruit and alter the function of resident brain microglia to create specialized tumor-associated microglia, that elaborate molecules that both create a permissive stroma ("stromagenesis") (Aim 1) and activate astrocytes (Aim 2). These "reactive" astrocytes support the creation and maintenance of the perivascular niche (gliomagens) (Aim 3), which provides the proper microenvironment for cancer stem cells - the treatment-resistant population of glioma cells (Aim 4). Collectively, this cross-disciplinary initiative capitalizes on converging lines of evidence that conceptualize gliomas as dynamic ecosystems composed of neoplastic and non-neoplastic cells and leverages expertise in mouse modeling, glioma biology, tumor microenvironment, microglia function, and translational medicine to identify new therapeutic targets for treating these deadly brain cancers.