Accurate genetically engineered mouse (GEM) disease models are potentially very powerful tools for disease mechanistic studies and drug tests. We used this tool to study glioblastoma (GBM), a very challenging disease clinically. In a previous study, we generated inducible genetically engineered mouse models by targeting pathway aberrations [retinoblastoma protein (RB), K-Ras and phosphatase and tensin homolog (PTEN)] to astrocytes that are frequently dysregulated in human astrocytomas. Each genetically engineered mouse line is a conditional floxed-stop or floxed-exon line, in which the activation of mutation is dependent on the presence of active Cre enzyme. We crossed those mice to GFAP-CreER mouse line, so that the mutations are restricted in the glial fibrillary acidic protein (GFAP) expressing astrocytic cells. The mutations are inducible at arbitrary time points because they are latent until mice are intraperitoneally injected with 4-hydroxytamoxifen (4-OHT) to activate Cre activity. This approach has resulted in stepwise progression of GBM. In brief, genetically engineered mice with aberrant retinoblastoma protein pathway alone developed low-grade astrocytomas (grade II). Neither constitutive K-Ras activation nor PTEN inactivation alone produced detectable brain pathology. Genetically engineered mice harboring both inactivated retinoblastoma protein and constitutively active K-Ras developed high-grade astrocytomas (grade III);subsequent PTEN inactivation produced tumors with histopathological features of glioblastom (grade IV). To determine the role in the astrocytoma/glioblastoma genesis of epidermal growth factor receptor (EGFR), a frequent mutation and a hot therapeutic target in glioblastoma treatments, we firstly investigated whether EGFR is activated in glioblastoma development in our models. Particularly, we examined the model that has both K-Ras mutation and Rb inactivation. We have harvested the brains of mice at different time points after induction. After analyses of those samples, we have found that EGFR signal is activated by multiple methods, including immunohistochemistry, western blot and real time reverse transcriptase PCR. We found that, other than EGFR itself, the level of EGFR ligands were also increased in the tumors by using real time reverse transcriptase PCR. We confirmed that EGFR signal was activated in the mouse model that has both K-Ras mutation and Rb inactivation. In the meanwhile, we tested if EGFR signal was required for the tumorigenesis in this model by genetically knocking out the EGFR gene. We have crossed the above genetically engineered mouse model to EGFR conditional knockout mouse line. If EGFR is required for the tumorigenesis, then tumor development was expected to be inhibited to a certain degree. However, our results showed that tumors actually progressed to a more advanced stage upon the loss of EGFR, suggesting some compensatory mechanism contributing to the tumorigenesis upon EGFR loss. To explore the underlying mechanism, we have compared the EGFR signal and related receptor tyrosine kinase signals between the tumors with and without EGFR. To do this, we used both brain tissues and tumor cell cultures. We have generated multiple tumor cell lines on tumors with EGFR wild type, EGFR heterogeneous deletion and EGFR homogeneous deletion. We have done some testing on the culture tumor cells. All of the results support the possibility of a compensatory mechanism of tumorigenesis upon EGFR loss. We also have collected a bank of astrocytoma/glioblastoma samples for the non-bias micro array experiment, which will help us understand the underlying mechanisms. 47% of the overall budget was spent on establishing the new laboratory.
|Song, Yurong; Zhang, Qian; Kutlu, Burak et al. (2013) Evolutionary etiology of high-grade astrocytomas. Proc Natl Acad Sci U S A 110:17933-8|