Project 3: Inhibition of radiation-induced phenotype conversion in glioblastoma


Radiation therapy (RT) is currently still one of the most effective treatment modalities against glioblastoma (GBM). However, while RT cures patients from many other cancers, all GBM eventually recur and are ultimately fatal. This data is in sharp contrast to the moderate radiosensitivity of GBM cells in vitro and in vivo, and the relatively high total radiation doses given to GBM patients clinically, thus indicating that RT failure in GBM is not easily explained by the intrinsic radioresistance of GBM cells. Recent experimental data support the view that GBM are organized hierarchically with a small number of radiation-resistant GBM-initiating cells (GICs) capable of re-growing a tumor and giving rise to all lineages of differentiated cells found in GBM, while their progeny lack these features.

According to the cancer stem cell hypothesis, all tumor-initiating cells have to be eliminated to achieve cure. A competing model of GBM is based on the "clonal evolution" model and assumes that all cells in a tumor can stochastically acquire a cancer stem cell phenotype. This model is supported by recent studies demonstrating acquisition of stem cell traits by non-tumorigenic GBM cells under hypoxia, at low pH, upon nitric oxide exposure, and our recent report of radiation-induced generation of induced tumor initiating cells. Since each clinical radiation fraction of typically 2 Gy kills only a portion of the total cellular tumor mass, our data suggest that elimination of GICs alone will be insufficient for GBM cure unless non-tumorigenic GBM cells are eliminated in parallel, and phenotype conversion is prevented.

Using an imaging system for cell populations enriched for tumor-initiating cells, we have screened chemical libraries of over 80,000 compounds and identified classes of compounds, including dopamine receptor antagonists, that interfere with radiation-induced phenotype conversion, eliminate non-stem GBM cells, and prolong survival in a mouse model of GBM. Results from the proposed studies could have a wider impact on cancer, as these principles may apply not only to GBM, but to many other solid tumors as well.