Abstract 5356: In search for mutually exclusive marker signatures for glioma initiating cells (GICs) and non-GICs, respectively

Abstract

Glioblastoma (or WHO grade IV glioma) is the most common and most malignant primary brain tumor, in adults. Despite gradual optimization of the conventional treatment protocol, the prognosis for these patients is still poor. The tumors are very heterogeneous and division into four main groups has been suggested. Although intra-tumor heterogeneity has long been regarded a major cause for treatment failure the extent of glioma intra- tumor heterogeneity at the cellular level is not well known. Heterogeneity can arise by two parallel but interdependent mechanisms. (i) Sequential accumulation of mutations as a consequence of genetic instability. (ii) Hierarchical organization of glioma cells. High grade gliomas contain subpopulations of tumor cells with stem-like properties with a unique ability to reconstitute the original tumor when xenografted and have been referred to as glioma initiating cells (GICs). GICs resist conventional treatment and the intense interest is based on the assumption that accurate targeting of these cells will cure glioma patients. Separation and direct comparision of GICs and non-GICs is difficult since universal and exclusive GIC (or non-GIC) markers are missing. GICs are instead defined by functional criteria, self-renewal under neural-stem cell (NSC) culture conditions, multipotency and most importantly, ability to reconstitute the morphology (and expression pattern) of the original tumor when xenografted to immunocompromised mice. In vitro culturing under NSC conditions (as spheres or adherent to laminin) strongly selects for GICs. Conversely, the tumor forming ability of human glioma cells have repeatedly been observed to be repressed by serum or BMP exposure which also slow or cease proliferation in NSC and induce expression of markers associated with glial/neuronal differentiation. Thus, the malignant phenotype of glioma cells can be modulated by external factors. We are here setting up a marker independent separation method based on functional properties of GICs and non-GICs, respectively. Glioma cells exposed to serum, i.e. an induced mixed population of GICs and non-GICs, were transferred to NSC and probed with CMFDA to follow proliferation rates. After 12-14 days in NSC culture, two subpopulations of U2987MG cells were visible, one highly proliferative (HP) and one low proliferative (LP). Immunostainings for stemness and glial lineage markers identified GFAP when combined with S100 or MAP-2 as robust markers for the LP group of cells. The HP and LP subgroups as well as a GFAP+ clone with LP properties were orthotopically xenografted to SCID mice. HP cells gave rise to big, high grade tumors whereas injection of the GFAP+ clone with LP properties resulted in very small non-invasive tumors. Our preliminary data thus suggests that this is a useful approach. We are now analyzing these subgroups using transcriptome sequencing. 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 5356. doi:1538-7445.AM2012-5356