Malignant glioma formation is associated with characteristic genetic alterations, although epigenetic mechanisms may contribute in tumorigenesis. Until recently, our knowledge has mainly been based on chromosomal and molecular studies performed in the last two decades. This has increased tremendously with the advent of new technologies, in particular expression arrays for simultaneous analysis of thousands of genes. Consequently, gene therapy of gliomas may aim at molecular interference with 'gain of function' genes (oncogenes) or replacement of 'loss of function' genes (tumor suppressor genes). Such approaches require transgene expression in whole tumor cell populations (if not other mechanisms come into play) which cannot be achieved with current vector systems. Hence other strategies have been pursued which may be independent of genes actually involved in tumorigenesis. Microbial genes (e.g. herpes simplex virus thymidine kinase) may be transferred into the tumors allowing for prodrug activation (e.g. ganciclovir). Furthermore, cytokines or other immunomodulatory genes may be used for vaccination purposes which frequently involves ex vivo transfection of autologous tumor cells with such genes. These approaches proved promising in preclinical studies performed in cell culture and different inbred rodent models. A considerable number of clinical trials have been initiated based on these approaches. Although most therapeutic strategies proved safe, clinical responses fell short of expectations raised by preclinical results. This, to a large extent, has to be attributed to a lag in the development of efficient vector systems. Although much effort has been put into this area of research, neuro-oncologists are still in await of a vector system allowing for selective and efficient tumor cell transduction. This has led to increased interest in distinct but related strategies, e.g. oncolytic viruses or direct intra-tumoral delivery of anti-sense oligonucleotides.
