Glioblastoma (GBM) is the most prevalent form of malignant brain cancer and its aggressive nature makes it incurable. Current therapies, including surgery when possible, chemotherapy, and radiotherapy, are debilitating strategies that are not curative. We have developed mouse models of GBM based on targeted mutation of three tumor suppressor genes that are frequently mutated in GBM: Pten, p53, and NF1. These mouse strains develop GBM with 100 percent penetrance and the histopathology is highly related to that observed in human GBM. The fact that all mice develop GBM in predictable fashion has allowed detailed investigation into the earliest events and progression of these tumors.
Cell of Origin
Our studies have pointed to adult neural stem and progenitor cells as the most favorable and likely sources for development of GBM. Comparative analysis of tumor development over time based on targeting to more versus less differentiated cells supports the idea that accumulation of a minimal number of mutations to cause tumor development occurs in more primitive cells rather than terminally differentiated cells.
Cancer Stem Cells
We have pursued the question of how tumors grow and recur following chemotherapy. We find that the spontaneous GBM tumors in our mouse strains develop in a hierarchical manner with a Cancer Stem Cells (CSC) residing at the apex. We have used transgenic tools to visualize these cells and to develop strategies for their functional characterization within tumors and in isolation. We can inhibit tumor growth or transplantation by specifically targeting CSC. Moreover, after chemotherapy-mediated killing of the rapidly dividing spontaneous tumor cells, we have demonstrated that the more quiescent CSCs are spared and serve as the source of tumor recurrence. More recent advances permit us to use fluorescence activated sorting to greatly enrich for this CSC population and to move forward toward a detailed genomic characterization in comparison to wild type adult neural stem cells and progenitors.
GBM Subclassification
Genomic profiling of human GBM has resulted in a classification of tumors subtypes based on the expression profiles, mutational configurations, etc. These in silico stratifications have not been fully validated in functional analyses. We find that our GBM mouse models have relatedness to all human GBM subtypes. Our studies point to an alternative and complementary method for tumor categorization. We find that identical GBM relevant mutations (Pten, P53 and NF1) in distinct adult CNS lineage progenitors initiate distinguishable GBM phenotypes. These studies suggest that phenotype is not only “driver-mutation” dependent but also cell-of-origin dependent. These studies can complement genomic profiling to functionally stratify GBM subtypes.
Alcantara Llaguno, S.R., Wang, Z., Sun, D., Chen, J., Kim, E., Xu, J., Hatanpaa, K.J., Raisanen, J.M., Burns, D.K., Johnson, J.E., and Parada, L.F. 2015. “Adult Lineage Restricted CNS Progenitors Specify Distinct Glioblastoma Subtypes.” Cancer Cell, in press.
High Throughput Screens
We have exploited the ability to culture and pool genetically similar tumors for the purpose of small compound and siRNA high throughput screens on low passage cells. After validation, these efforts have resulted in the identification of small compounds that hold promise for highly specific toxicity on GBM cells but not on normally dividing cells such as primary mouse embryo fibroblasts. We seek to identify the molecular targets and mechanisms of activity for your compounds and to functionally validate the novel genes emerging from our screens.