Abstract 489: Micro-RNA signature of glioblastoma multiforme invasion: An in vivo study in patient-tumor derived orthotopic xenograft mouse models

Abstract

Abstract Glioblastoma multiforme (GBM) is the most malignant brain tumor that occurs both in adults and children. One of biologic features that make GBM particularly difficult to treat is its diffuse invasion into normal brain tissue. Efforts of developing new targeted therapies against GBM invasion are hampered by poor understanding of the invasion process and biological differences between the tumor core cells and those in the invasive front, because the invasive GBM cells are rarely available for comparative analysis. To understand the molecular mechanisms of GBM invasion, we isolated invasive cells that have migrated into normal mouse brain tissue from a panel of four (1 adult and 3 pediatric) patient-tumor derived orthotopic xenograft mouse models and compared their microRNA expression profiles with that of the matched tumor core cells using Taqman MicroRNA array, a real-time PCR-based assay that profiles 768 microRNAs utilizing a normal human cerebral tissue as reference. To test our hypothesis that the invasive GBM cells depend on a common set of microRNAs to migrate into normal brain tissue, we focused our data analysis on the microRNAs that are shared in at least 3 of the 4 GBM models. From the total of 322 microRNAs that were found to be expressed in at least 3 of the four models (159 were present in all four models), we identified 7 microRNAs that were significantly upregulated in the invasive cells in all four models, and an additional 49 in three of the four models. This data set included miR-223 and miR-138, which have been previously associated with tumor invasion. A total of 16 microRNAs were also found to be decreased in the invasive GBM cells, including miR-107 and miR-18b that have been previously described in multiple human tumors. Given the critical roles of cancer stem cells in tumor initiation and progression, we also compared the neurosphere forming efficiency between the invasive and tumor core subpopulations. In both of the models analyzed to date, a significantly increased neurosphere forming efficiency was observed in the invasive cells. To determine if and to what extent that the microRNA signature of GBM invasion is preserved in the neurospheres, we compared the microRNA profiles between neurospheres and their corresponding freshly isolated xenograft cells of the invasive front and tumor core. The correlation coefficients ranged from 0.58 to 0.84(0.725 ± 0.136). Some of the dysregulated microRNAs originally found in vivo were preserved in the cultured neurospheres. In conclusion, we have established a novel in vivo model system that allows for isolation of paired and biologically accurate invasive and tumor core GBM cells in vivo in mouse brains. Using this system, we have identified a novel subset of microRNAs that are significantly altered in the invasive cells. This group of microRNAs warrants further evaluation as potential therapeutic targets. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 489. doi:1538-7445.AM2012-489