Dataset Information

MELK Leucine Zipper Kinase is a Key Differentiator between Glioblastoma Multiforme and Pilocytic Astrocytomas

ABSTRACT: Purpose: Gene expression analyses comparing tumors of different behaviors have been used for the identification of genes that have a role in the growth and maintenance of tumors, with the aim of both understanding the molecular basis of tumorigenesis and identifying novel potential therapeutic targets. We have here used this approach to identify the underlying gene expression differences between highly invasive glioblastoma multiforme and typically benign pilocytic astrocytomas. Experimental Design: We performed cDNA microarray analyses comparing astrocytomas of polar grade: grade I, pilocytic astrocytoma (PA), the non-invasive and most frequent pediatric tumor, with grade IV, the highly infiltrative glioblastoma multiforme (GBM) Results: Despite the significant clinical and pathological differences between the two tumor types, only 63 genes were found to exhibit two-fold or greater over-expression in GBM as compared with PA. A functional classification of the GBM over-expressed genes indicated that more than 50% are related to the regulation of the cell cycle and mitosis. Microarray data was validated by RQ-PCR testing six over-expressed genes in GBM related to PA: MELK, AUKB, ASPM, PRC1, IL13RA2 and KIAA0101. The differential expression was confirmed for all six genes, showing at least a 5 fold increase in the average expression levels in GBM compared to PA. Of the over expressed genes, that which exhibited the most statistically significant difference is Maternal Embryonic Leucine zipper Kinase (MELK). We undertook a more detailed investigation of the expression of this serine/threonine kinase gene in astrocytomas in the light of its role in the regulation of multipotent neural progenitor proliferation and previous suggestions that this may be a oncogenic target of general importance. In an examination of more than 100 tumors of the central nervous system we found progressively higher expression of MELK in GBM with astrocytoma grade and a noteworthy uniformity of high level expression in GBM. This latter feature probably explains the lack of observed association of MELK expression with survival of GBM patients. Over-expression at a similar level to GBM was also observed in medulloblastoma. We found neither gene methylation nor amplification to be a factor in MELK expression but were able to demonstrate that MELK knockdown in malignant astrocytoma cell lines caused a reduction in proliferation and anchorage-independent growth in in vitro assays. Conclusions: Our results indicate that GBM and PA differ by the expression of surprisingly few genes. Among them, MELK may play a key role in differentiating these two tumor types and represent an important therapeutic target for the management of the most frequent brain tumors in adult and children. Keywords: Comparizon of CNS tumors of different grades.

ORGANISM(S): Homo sapiens

PROVIDER: GSE7330 | GEO | 2008/03/10

SECONDARY ACCESSION(S): PRJNA98051

REPOSITORIES: GEO

ACCESS DATA

Similar Datasets

Project description:The different clinical behavior of low and high grade gliomas and the chance to develop novel selective agents that specifically target tumor-associated proteins in gliomas stimulate the research of molecules playing a role in glioma progression. Gene expression profiling using microarrays allows the study at the same time of the expression patterns of thousands of genes in tumor cells. In the present study microarrays with about 20,000 genes have been employed to discover the gene expression profile in 39 glial neoplasias (28 glioblastomas (GBM) and 11 low grade gliomas, namely 4 oligodendrogliomas, 5 pilocytic astrocytomas (PA), 2 fibrillary astrocytomas (FA)). Unsupervised classification through hierarchical cluster analysis identified 2 groups of tumours: one group mainly composed of low grade malignant tumours (10 low grade gliomas and 3 GBM), the other one constituted by GBM, with the exception of one low grade case. The nearest shrunken centroid classification method was used to identify genes useful to classify and best characterize high and low grade gliomas. [Tibshirani et al. 2002]; This procedure selected 9 genes as most informative for the classification task: ; Among them 7 genes were overexpressed in low grade gliomas, but underexpressed in GBM; on the contrary 2 genes were overexpressed in GBM, but underexpressed in low grade tumours; Forty five tumors were immunostained for IGFBP-2 . 81,5% GBM resulted immunopositive. On the contrary only one low grade glioma was positive. Gene expression profiling and immunohistochemistry suggest that IGFBP-2 may play a role in glioma progression. IGFBP-2 appears to be a novel immunohistochemical marker of malignancy in glial tumours and probably is the basis for targeted chemotherapy. Experiment Overall Design: We retrieved for this study 185 randomly selected cases of gliomas. All the cases had been received unfixed. A sample of neoplastic tissue had been flash frozen over liquid nitrogen and stored at -80Â°C into the frozen tissue bank of the Section of Pathology of Bellaria Hospital between 1990 and 2002. The remaining (specular) tissues were formalin fixed and paraffin embedded for routine histologic diagnosis. Experiment Overall Design: In the present investigation, only samples with rRNA 28S/18S ratio > 1.0 measured by Bioanalyzer 2100 (Agilent)and no evidence of ribosomal degradation were included. Some cases were excluded as the tissue samples were composed only by necrosis; other cases did not show neoplastic proliferation. Therefore only 39 glial neoplasias were suitable for the study: i.e. 28 GBMs and 11 low grade CNS tumors, namely 4 oligodendrogliomas (OL), 5 pilocytic astrocytomas (PA), 2 fibrillary astrocytoma (FA) showed a good quality of RNA. Experiment Overall Design: All the tumors were re-staged and graded according to 2000 WHO criteria at time of gene expression analysis by two pathologists (GM and VE). Experiment Overall Design: 39 glial neoplasias were labeled by Cy5. Common reference pool consisted of a mixture of total RNA from the low grade CNS tumor group (4 OL, 5 PA, 2 FA) labeled by Cy3. A series of the most informative genes were confirmed by RT-qPCR. Experiment Overall Design:

Dataset Information

MELK Leucine Zipper Kinase is a Key Differentiator between Glioblastoma Multiforme and Pilocytic Astrocytomas

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure