Investigation of signalling pathways in high grade glioma in order to predict responsiveness to tyrosine kinase inhibitors.
High grade gliomas are the most common primary brain tumour in adults and are virtually uniformly fatal. Therapies that target specific pathway proteins, such as tyrosine kinases, are transforming the treatment of many cancer types. Initial studies using tyrosine kinase inhibitors (TKls) in the treatment of high grade glioma demonstrated reduced tumour volumes in certain patient populations (Butowski and Chang, 2005, Mellinghoff et al., 2005). However, subsequent survival studies have been disappointing (De Witt Hamer, 2010). To help clarify the potential role of such agents, it has been suggested that clinical trials should seek to identify subgroups of patients most likely to respond to molecular targeting with TKls.
In this study it is hypothesised that the pathway expression profiles of high grade gliomas may predict response to EGFR I PDGFR pathway blockade with TKls. Specifically, it was determined if EGFR, EGFRvIII, PDGFR c-KIT and c-ABL expression together with PTEN expression and downstream activation of AKT and P70s6K predicted response to EGFR or PDGFR blockade with erlotinib, gefitinib or imatinib. The immunohistochemical (IHC) expression profiles of the EGFR I PDGFR pathway in formalin fixed paraffin embedded sections of high grade gliomas and in primary cultures derived from the same tumours were established. This facilitated the characterisation of changes introduced as a result of in vitro culturing. Next, response to treatment with TKls was determined for each treated primary culture using anti-proliferation toxicity assays. Response was correlated with the above expression profiles and with patient survival data.
Tumour expression profiles showed that no one protein of the EGFR 1 PDGFR pathway significantly influenced overall survival and that diagnostic grade remained the best prognostic indicator of outcome. The EGFR I PDGFR pathway profiles remained stable over time in patients with repeat biopsies. Comparison of pathway profiles between tumour tissue and their derived primary cultures demonstrated the introduction of significant genomic and protein differences as a result of culturing. Increased surface receptor tyrosine kinases were observed with an opposing decrease in activation of the downstream pathway proteins when the cells were grown in vitro. Despite these differences, the overall EGFR I
PDGFR pathway profile of primary cultures retained key signature characteristics reflective of the coordinated pathway disruption observed in high grade gliomas. This verified their use as a biologically representative model of high grade glioma in determining if biomarkers predicting response to tyrosine kinase inhibitors (TKls) could be identified in the EGFR I PDGFR pathway. It became apparent that although some cultured tumours responded to therapy with TKls, this response was not statistically related to the pathway proteins that were evaluated in this study.
EGFR 1 PDGFR in vitro pathway status was a poor predictor of in vitro responsiveness to molecular blockade using the TKls: erlotinib, gefitinib and imatinib. In conclusion the complexity and evasiveness of signalling pathways in high grade gliomas coupled with the possible emergence of alternate, deviant signalling pathways combine to minimise the likelihood that single agent pathway inhibitors will have a significant clinical benefit for glioma patients at this time.