Please note that we have stopped the regular imports of Gene Expression Omnibus (GEO) data into ArrayExpress. This may not be the latest version of this experiment.

E-GEOD-12689 - GBM invasion transcriptiome

Submitted on 4 September 2008, released on 18 May 2010, last updated on 1 May 2014
Homo sapiens
Samples (76)
Array (1)
Protocols (10)
Glioblastoma multiforme (GBM), the most advanced form of a large subset of brain tumors collectively known as glioma, is the most aggressive and invasive type of brain tumor. Patients usually experience a median survival range of 9 to 12 months. GBMs are extremely difficult to manage because of the tumor cells’ tendency to migrate and pervade into adjacent tissues. Surgical resection of GBMs generally only slows disease progression because any remaining tumor cells proceed to migrate through brain tissues and reform a new tumor mass. It is hypothesized that the invasive phenotype of these tumor cells may be attributable to unique gene expression. Stationary core and invasive rim tumor cells were collected separately by laser capture microdissection (LCM) from 19 biopsy samples. Identification of differentially expressed genes in the tumor core and invasive rim can give valuable insight to the genes and pathways potentially involved with the invasive phenotype. This information can then be used to generate possible biomarkers, diagnostic markers, or drug targets. Human glioblastoma tumor samples were obtained from patients who underwent primary therapeutic brain tumor resection. All specimens were collected and submitted to the study under institutional review board approved protocols. None of the patients had been subjected to chemotherapy or radiotherapy prior to resection, in order to avoid genetic signatures that are due to exposure to alkylating agents and/or ionizing radiation. All specimens were verified as GBMs by a neuropathologist. Tumors are embedded in OCT, then sectioned in a cryostat at -20C, to a thickness of 8-10 um and placed onto HistoGene slides. Sections to be microdissected are removed from the -80C fixed and stained with an abbreviated Hematoxylin and Eosin protocol. Two thousand tumor core and invasive cells are dissected onto separate caps using the PixCell II instrument using CapSure™ Macro LCM Caps. The CapSure™ Macro LCM Caps LCM 0211 should be used with the AutoPix instrument. Cells in the tumor core are identified by nuclear atypia and size and captured using the larger spot sizes. Tumor cells immediately adjacent to necrotic areas, cortical areas, cells with small regular nuclei, endothelial and blood cells should be avoided. Individual white matter invading GBM cells can be identified by means of their nuclear atypia and heteropyknotic staining, which is consistent with that of the cells within the tumor core. They should be microdissected using the 7.5 um laser spot size and the initial power settings recommended by the PixCell II manual.  After microdissection all harvested material should immediately be lysed on the cap by applying the lysis buffer (XB) from the PicoPure RNA Isolation Kit according to the PicoPure protocol and stored -80C. Cell populations harvested on different caps can be pooled at the time of RNA isolation. RNA integrity is varified by identification of distinct 28S and 18S ribosomal bands with an Agilent Bioanalyzer using the RNA 6000 Nano LabChip kit. Total RNA was isolated from 2000 LCM cells (to ensure at least 500 ng total RNA) using the PicoPure RNA solation Kit, following manufacturers protocol. mRNA is reverse transcribed with the RiboAmp RNA Amplification kit. 500 ng total RNA is amplified with the RiboAmp RNA Amplification kit, following manufacturer’s instructions. The total yield that can be expected falls between 30 and 60 µg copy RNA. The size of the copy RNA should be verified by gel electrophoresis. Approximately 500ng of copy RNA can be separated on a 1% agarose gel in TAE buffer. A smear of amplified material should be seen between 200 and 3000bp. Six µg amplified RNA are labeled in a RT with SuperScriptIII in the presence of dUTP Cy5 utilizing 6 µg random hexamers as primers. Universal reference RNA is amplified one round in the same manner and labeled with Cy3 dUTP. Labeled cDNA is hybridized overnight onto cDNA microarray. Following hybridization, arrays are washed, scanned and quantitated with the Axon GenePix 4000 microarray reader (Axon Instruments). Gene expression results are analyzed using GeneSpring (Silicon Genetics) software. Intensity dependent normalization is applied, where the ratio is reduced to the residual Lowess fit of the intensity versus ratio curve. The measured intensity of each gene is divided by its reference channel (signal from the universal reference RNA) in each sample. When the reference channel is below 10, the data point is considered uninformative. The ratios (sample over reference) for the tumor core experiments and invasive rim experiments are averaged and compared. Genes that are more than two-fold differentially regulated in the majority of the matched core/invasive rim sets are selected.
Experiment type
transcription profiling by array 
Brock Armstrong <>, Dominique Hoelzinger, Linsey Reavie, Michael E Berens, Tim Demuth
Investigation descriptionE-GEOD-12689.idf.txt
Sample and data relationshipE-GEOD-12689.sdrf.txt
Raw data (3),,
Processed data (1)
Array designA-AGIL-11.adf.txt
R ExpressionSetE-GEOD-12689.eSet.r