We have previously identified thromboxane synthase as an important regulator of glioma cell migration. Inhibitors of this enzyme abrogate cell motility and induce apoptosis. However, the formation rate of thromboxanes is indirectly dependent on the activity of cyclo-oxygenase, which represents the rate-limiting step in the synthesis of prostaglandins and thromboxanes. In this study we have analyzed the expression of COX-1 and COX-2 in glioma cell lines and biopsies of glial tumors. In normal glia no expression of both COX isoforms was present, however, reactive astrocytes and glial tumors of all grades demonstrated expression of both COX-1 and COX-2. In contrast to inhibitors of thromboxane synthase, selective and non-selective cyclo-oxygenase inhibitors did not block cell motility. Specific COX-2 inhibitors resulted in growth inhibition and induction of intracellular DNA fragmentation indicative of apoptotic cell death. Treatment of glioma cells with thromboxane synthase inhibitors had a synergistic effect on induction of apoptosis by camptothecin, whereas COX inhibitors had not. Furthermore, combined treatment using COX-2 inhibitors and specific thromboxane synthase inhibitors did not show a synergistic increase of apoptosis. These data indicate that COX inhibitors and thromboxane synthase inhibitors influence apoptosis in glioma cells through different pathways. We hypothesize that, in contrast to the COX-2 inhibitors, thromboxane synthase inhibitors block the invasive phenotype of glioma cells and therefore increase the pro-apoptotic disposition of the cells and increase the susceptibility to induced apoptosis. This effect may be independent of prostaglandin synthesis controlled by cyclo-oxygenases.

Cancer formation and progression is a complex process determined by several mechanisms that promote cell growth, invasiveness, neo-angiogenesis, and render neoplastic cells resistant to apoptosis. The tumor suppressor p53 and the proto-oncogenic factor ets-1 are important regulators of such mechanisms. While it is well established that p53 and ets-1 influence various aspects of cell behavior by regulating the transcription of specific genes, little is known about the functional relationship between these transcription factors. We found that the gene encoding thromboxane synthase (TXSA), which we recently identified as a factor promoting invasion and resistance to apoptosis in gliomas, is a novel target gene for both p53 and ets-1. We demonstrate that p53 and ets-1 coregulate TXSA in an antagonistic and interrelated manner, with ets-1 being a potent transcriptional activator and p53 inhibiting ets-1-dependent transcription. Negative interference with ets-1 transcription requires functional p53 and is lost in mutant p53 proteins. We show that ets-1 and p53 associate physically in vitro and in vivo and that their interaction, rather than a direct binding of p53 to the TXSA promoter, is required for transcriptional repression of TXSA by wild-type p53. An important implication of our findings is that the loss of p53-mediated negative control over ets-1-dependent transcription may lead to the acquisition of an invasive phenotype in tumor cells.

Previous studies have demonstrated that inhibitors of the arachidonic acid metabolism block migration and sensitise human glioma cells to treatment inducing apoptosis. This paradigm may provide a new concept for anti-invasive treatment strategies targeting invasive glioma cells. However, the effect of such treatment on other cellular elements in glial tumours such as endothelial cells is unknown. In this study we have analysed the expression of metabolites of the arachidonic acid pathway in endothelial cells in vitro and in vivo and we have assessed the influence of inhibitors of this pathway on motility, capillary like tube formation, and apoptosis in human endothelial cells. Human endothelial cells (HUVEC) in culture showed expression for thromboxane synthase and both isoforms of cyclo-oxygenase, COX-1 and COX-2. Immunostaining demonstrated low levels of COX-1 expression in capillaries and larger vessels of normal brain and moderately elevated levels of this enzyme in small vessels of brain tumours of various grades. Both thromboxane synthase and COX-2 expression was limited to endothelial cells found in anaplastic gliomas and glioblastomas. Thromboxane synthase inhibitors strongly decreased endothelial cell migration in HUVEC in vitro and capillary like tube formation was strongly inhibited by the compound at a similar dose range. The non-selective cyclo-oxygenase inhibitor ASA and the selective COX-2 inhibitor sulindac only had a minor effect on endothelial cell migration, however, the COX-2 inhibitor sulindac showed a synergistic effect with the thromboxane synthase inhibitor. Thromboxane synthase inhibitors induced apoptosis in endothelial cells as demonstrated by intracellular histone-complexed DNA fragmentation. These data suggest that inhibitors of thromboxane synthase influence migration and apoptosis in both human glioma cells and human endothelial cells. An anti-invasive treatment strategy using this class of compounds may therefore not only sensitise glioma cells to conventional treatments inducing apoptosis but may also be supported by an anti-angiogenic effect.

Combining the results obtained from our human and mouse studies, it has to be postulated that host factors other than the sensitivity to TAM of the individual cell, determine the efficacy of TAM-treatment in vivo.