Abstract 3301: c-Myc modulates the dynamic transition between tumor initiating states in glioblastoma

Abstract

Abstract Introduction: The cancer stem cell hypothesis states that elite subpopulations of cells within a tumor exclusively exhibit the ability to initiate and maintain tumors. Recent evidence indeed suggests that glioblastoma is driven by a subset of Tumor Initiating (TI) cells characterized by their sufficiency to form tumors as well as their unlimited potential for self-renewal. During expansion, daughter cells of TI cells variably lose these properties. The transition between TI proficiency states remains poorly understood. Methods: To model differing TI states, we screened glioblastoma lines that exhibited varied in vivo tumorigenicity when cultured in differing conditions. The transition between these cell states was characterized by transcriptome profiling, and potential modulators of this transition were tested by forced gene expression. Results: We identified both primary and long-term passaged lines that exhibited increased in vivo tumorigenicity when cultured in neurosphere conditions (pro-TI state) relative to serum conditions. Cells cultured in the pro-TI state exhibited increased expression of neural stem cell markers, including Nestin, Musashi, Olig2, and Sox2. These cells also exhibited increased xenograft tumor growth with 10-fold fewer injected cells, and decreased murine survival. Individual cells within even a long-term passaged line showed a spectrum of TI proficiency as measured by rates of neurosphere formation from single cells. This spectrum showed hybrid properties between those predicted by the cancer stem cell hypothesis, which states that only elite cells give rise to tumors, and the opposing view that tumor formation is a stochastic process. Although several subclones showed an absolute loss of TI proficiency, those that maintained it varied in proficiency to an extent significantly higher than that predicted by a pure elite cell model (p < 0.05 using F-test for homoscedasticity or equality of variances). Transcriptome profiling of subclones differing in TI proficiency identified a signature previously implicated in glioblastoma “stem” hierarchy determination. Bioinformatic analysis of this signature revealed an over-representation of targets of c-myc (17.8% of TI proficiency genes vs. 12.3% of all genes, chi square p < 0.0001). c-Myc was more highly expressed among subclones showing high TI proficiency, and forced c-myc expression increased TI proficiency as measured by increased growth in soft agar. Conclusion: Analogous to induced pluripotency (iPS), the transition between differing TI states in glioblastoma appeared driven by a combination of elite and stochastic processes. TI proficiency may best be explained using a threshold model, in which factors including c-myc expression can increase the probability that individual cells can give rise to tumors. Even well passaged cell lines can serve as models to better understand these transition states. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr 3301. doi:10.1158/1538-7445.AM2011-3301