Expression microarray analysis of brain tumors: what have we learned so far.

This review covers the emerging field of expression microarray technology as applied to human brain tumors. Dual and single color techniques are described and contrasted, and the importance of proper handling of the starting material is emphasized. Difficulties with data interpretation are described and current approaches to cluster analysis reviewed. Microarray studies of general importance or specifically pertaining to brain tumors, published in the initial few years of this technology, are summarized. Although this technology is still in its infancy, microarray has distinguished prognostic groups within medulloblastomas and separated medulloblastomas from morphologically identical supratentorial PNETs. Differential expression of a number of genes previously known to be involved in the pathogenesis of brain tumors has been confirmed. These genes include EGFR, VEGF, transcription factor AP-2, insulin growth factor binding proteins 3 and 5, matrix metalloproteinases, tissue inhibitors of matrix metalloproteinases, CD44, basic fibroblast growth factor, and cathepsin H. Finally, novel roles for a few genes, including insulin growth factor binding protein 2 and apolipoprotein D, have been revealed for the first time by expression microarrays.