Abstract A08: Neurosphere culture captures the clinical and molecular diversity of glioblastoma tumors

Abstract

Abstract Glioblastoma (GBM), a WHO grade IV astrocytoma, is the most prevalent and aggressive primary central nervous system tumor, characterized by poor response to standard post-resection radiation and cytotoxic therapy, resulting in dismal prognosis, with a 2 year survival rate around 20%. Culturing dissociated cells from fresh GBM specimens in medium formulated for the selection and expansion of neural stem cells (NSC) from mammalian brain favors the growth of floating multicellular spheroids, referred to as neurospheres (NS), enriched in cancer stem-like cells. NSs have become a frequently used model for in vitro studies, and for implant in immunocompromised mice for the development of orthotopic GBM patient-derived xenografts (PDXs). These success rates in establishing long term neurosphere cultures from GBMs has been reported to be between 30 and 70%. It is important to assess any possible bias which might result if there were differences in the likelihood of GBM tumors forming neurospheres based on clinical characteristics or molecular profile. The goal of this study is to determine to what extent GBM transcriptional subtypes and driver genomic alterations are represented in neurosphere cultures and to identify clinical correlates contributing to the success (NS(+)) or failure (NS(-)) to obtain long term neurosphere cultures from fresh GBM tissue. Samples from 145 fresh GBM surgical specimens containing viable non-necrotic tissue were dissociated and subjected to neurosphere culture for a minimum of 2 months. Of these, 103 were newly diagnosed, 39 recurrent, and 3 secondary GBMs, which progress from a lower grade tumor. The rate of neurosphere formation was not different between newly diagnosed (43%) and recurrent (44%) tumors (Fisher's Exact Test (FET), p>0.999). Among 9 pairs of matched newly diagnosed NS(+) and recurrent tumors, 6 retained NS(+) status upon recurrence and 3 switched to NP(-). The 3 secondary GBM samples, which progressed from lower grade tumors, were NS(-). For newly diagnosed GBMs, NS formation was significantly correlated with shorter time to progression (log-rank p=0.0147), but not with overall patient survival (log-rank p=0.163). There is a tendency for NS(+) tumors to have an increased risk of death, after adjusting for age at diagnosis (HR: 1.49, 95% CI: (0.95, 2.34); log-rank p=0.0830). Patient age at diagnosis, gender and race did not influence the proportion of NS(+) tumors . Mutations in IDH1 and promoter methylation leading to silencing of the MGMT gene have both been presented as biomarkers of less aggressive disease. The 6 GBMs harboring a mutation in IDH1 were NP(-). MGMT promoter methylation status did not correlate with neurosphere formation. Thirty three of the newly diagnosed GBMs were profiled by the TCGA GBM project. Neurospheres were established for 15 (45.4%) of these tumor samples. NP(+) tumors were distributed among all 4 transcriptional subtypes, with classical and non-GCIMP proneural tumors showing the highest success rate. Since most individual genomic mutations and copy number variations (CNVs) occur at a low frequency in GBMs, we investigated the representation of genes most frequently altered in a subset of the neurospheres, by exome and low pass whole genome DNA sequencing. Among the genes altered in the panel were: CDKN2A, EGFR, PTEN, TP53, CDK4, PDGFRA, NF1, PIK3CA, PIK3R1, RB1, MDM2, MDM4, ATRX, TERT, STAG2, MET, HDAC9 and MYC. Only 6 GBMs harboring IDH1 mutation were included, and this was the only genomic abnormality occurring in over 4% of GBMs that was not represented in the neurosphere panel profiled. We conclude that culturing neurospheres is a valid strategy to capture the clinical and molecular diversity associated with GBM tumors, but further comparisons of gene expression, epigenetic regulation, and other propagation methods for NS(-) and NS(+) tumors are granted. Citation Format: Laila M. Poisson, Tom Mikkelsen, Ana C. deCarvalho. Neurosphere culture captures the clinical and molecular diversity of glioblastoma tumors. [abstract]. In: Proceedings of the AACR Special Conference: Patient-Derived Cancer Models: Present and Future Applications from Basic Science to the Clinic; Feb 11-14, 2016; New Orleans, LA. Philadelphia (PA): AACR; Clin Cancer Res 2016;22(16_Suppl):Abstract nr A08.