Genomic Markers for Glioblastoma Multiforme, Revealed by Array CGH Analysis

Abstract

Aims: Despite of intensive studies of glioblastoma multiforme, still no unified concept for the most important specific molecular alterations exists for this tumor type. The method of array CGH has great potential for molecular characterization of glioblastoma. The aim of our study was to determine the type, frequency and fine mapping of unbalanced genomic changes and to suggest candidate genes for the emergence and development of brain tumors. Study Design: Ten tumor samples were collected from patients with glioblastoma multiforme after taking informed consent. Histological examination was done to confirm the presence of tumor cells in more than 75% of the samples. DNA was isolated from each tumor sample. Place and Duration of Study: The material was collected in Department of Neurosurgery (Medical University Sofia) and processed for analysis in Department of Medical genetics (Medical University Sofia) between June 2010 and December 2011. Methodology: We used the method of array CGH with BAC clones, covering the entire genome, for investigation of copy number changes in tumors. We applied specific software microarray analysis (Blue Fuse, Blue Gnome, Cambridge, UK). For fine mapping of the most significant aberrations and to identify possible candidate genes: 1) we identified BAC clones with aberrations of high amplitude (ratio T / H> 0.5 for gains and <-0.5 for losses); 2) we select only those clones that have a frequency of aberrations more than 30% and have at least one adjacent clone with the same aberration. Doing this we also determined the smallest regions of overlap (SRO) of aberrations that occur in at least 50% of tumors. Results: Trisomy 7 (70%) and monosomy 10 (80%) were the most common big aberrations in tumors. Regional aberrations across all chromosomes were characterized in details. The most frequent were: amplification of 1q43-1q44 (50-70%), deletion 1p36 (60%), gains 9p11-9p13 (50%), deletion 18q22 (50%), gains of 20q11-q13 (50%), loss of 22q12 (50%), deletion Xp21 (50%). By determination of smallest regions of overlap and based on their function, we suggested potential candidate-oncogenes (RGS7, CDK5, OPN3, CDK5RAP1, PTPN1) and tumor-suppressor genes (NF2 and OSM). Conclusion: Our study provide with basis for further studies in which the role of identified candidate-genes will be validated by other molecular genetics methods and at other levels - transcriptional and protein. This will lead to significant advances in knowledge of glioblastoma multiforme and suggestion of new more effective molecular-based prognostic and therapeutic indicators.