Background: Glioblastoma multiforme (GBM) is the most common and most malignant form of primary brain cancer. The standard of care for GBM patients is surgical resection followed by radiation (XRT) and temozolomide (TMZ). Populations of cells within GBMs that have properties similar to neural stem cells are thought to be the source of tumor cell self-renewal. These cancer stem cells (CSCs) may also be particularly resistant to therapy, although support for this important concept is limited. Additionally, widespread epigenetic changes and their ensuing dysregulation of gene networks are known to contribute to GBM formation and to therapeutic response. In particular, the widespread deacetylation of histones occurs in cancers with poor prognosis. Histone deactylase inhibitors (HDACi) are in clinical trials for GBM patients and one particular agent, vorinostat, has modest effects as a single agent in recurrent GBM. We asked whether primary-derived CSC cultures might respond differently than patient-matched non-CSC cultures to both standard-of-care and experimental epigenetic therapy.

Experimental Design: We established primary cultures of CSC as adherent glioma neural stem cell (GNS) and patient-matched non-CSC monolayer cultures in serum (ML) from each 10 primary glioblastomas. To test response to standard glioma therapy, we treated early passage cultures with physiological doses of TMZ (5-50 uM), with or without XRT (2Gy). To examine the relative efficacy of HDACi, the paired cultures were treated with physiological doses of vorinostat (1 −2 uM). Response to therapy was assayed by cell cycle analysis on a BD FACSCalibur.

Results: The majority (90%) of patient-derived primary cultures, including GNS and ML cell lines, are resistant to physiological doses of TMZ, similar to the general lack of response in most but not all GBM patients. When treated with TMZ and XRT alone or in combination, one particular GNS exhibited a response to XRT and the combination with TMZ, but not TMZ alone, while the patient-matched ML culture did not exhibit sensitivity to XRT, contrary to expectations. Lastly, treatment with vorinostat resulted in G2 arrest and increased cell death for 90% of cultures. However, the response observed between patient-matched GNS and ML cultures was not uniform across the panel of GBMs, with some GNS lines being more sensitive and others more resistant compared to ML cultures.

Conclusions: Most primary tumor-derived cultures are resistant to TMZ, irrespective of their enrichment for CSC. Furthermore, one GNS line was more sensitive to XRT than patient-matched non-CSC line suggesting that not all glioma CSCs are resistant to therapy as previously thought. The difference in response to vorinostat between GNS and non-CSC cultures reinforces this finding. Our study demonstrates that early passage primary gliomas cultures can be used to assess drug response in a timely manner. If the in vitro response mirrors in vivo response, this patient-specific model system might be beneficial for making clinical decisions on a patient-specific basis.

Citation Information: Clin Cancer Res 2010;16(14 Suppl):B38.