PO34MICRORNA INTRATUMOUR HETEROGENEITY IN GLIOBLASTOMA

Abstract

INTRODUCTION: It is now clear that intratumour heterogeneity exists in many cancers, with clear evidence of differential gene expression temporo-spatially in glioblastoma multiforme (GBM). MicroRNA (miRNA) profiling studies have identified risk-stratifying signatures in GBM, but it has not been clarified whether microRNA profiles vary in a similar fashion to gene expression within individual GBM tumours. METHODS: miScript Human Brain Cancer miRNA PCR Array was used to assess 84 miRNA's expression levels between distinct regions within primary GBMs. Primary tumour specimens (n = 2) were sampled in 4 distinct regions including enhancing and necrotic core, invasive margin and medial rim. miRNA expression was analysed with quantitative RT-PCR. Gene targets for analysis were selected with miRtarget gene prediction software, and protein levels were validated through immunohistochemistry (IHC). RESULTS: A large degree of intratumour heterogeneity was apparent in the samples. Among the most differentially expressed miRNAs were miR-137, miR-182-5p, miR-183-5p and miR-218-5p. Conversely, miR-129-5p, miR-15b-5p and miR-185-5p were the most consistently expressed miRNAs across regions. Homology existed between individuals with 8 common miRNAs differentially expressed between both central and edge regions. Protein levels of CDK6 and c-KIT nuclear accumulation were significantly elevated at tumour edges relative to core samples as predicted by microRNA profiling. CONCLUSION: The miRNA intratumour heterogeneity in GBM could produce significant profiling variation depending on spatial location of the sample within the tumour. Therefore, assessment of GBM miRNA signatures requires selective sampling from multiple distinct tumour regions.