Glioblastoma multiforme is the most malignant of the primary brain tumors with grave prognosis. Recent studies have identified so-called "cancer stem cells" in glioblastomas and other tumors. These cells have the capacity to proliferate and establish tu mors in laboratory animals after injection of only a small number of cells. Furthermore, cancer stem cells are resistant to radiation- and chemotherapy, which may explain why many tumors regress significantly with current treatment regimens, only to recu r later with more malignant properties. What is the origin of these cancer stem cells? Do they originate from more differentiated cancer cells that acquire stem cell-like properties? Or, do they originate from the normal neural stem cell population pre sent in the brain? Recent studies indicate that endogenous neural stem cells are attracted in large numbers to experimental glioma in mouse brain tumor models, and inter-mix with glioma cells. This raises the possibility that tumor-causing genes may end up in normal stem cells, transforming them into ?cancer stem cells?, which then drives cancer growth. We propose to explore whether such a normal-to-cancer stem cell conversion occurs. Determining the origins of cancer stem cells will facilitate develo pment of more effective and targeted brain tumor therapies.