ABSTRACT
One of the first links between COX-2 and angiogenesis was established during studies of the anti-tumor effects of existing COX inhibitors. In 1997, Seed et al observed that non-selective COX inhibition with diclofenac suppressed the growth of COX-2 positive colon tumor cells in vivo by blocking angiogenesis (Seed et al. 1997). Later in vivo studies revealed that selective COX-2 inhibitors block angiogenesis, an effect which was reversed by addition of a TP agonist (Daniel et al. 1999). Treatment with NSAIDs such as sulindac inhibited angiogenesis in tumor stroma in vivo. This was characterised by an inhibition in cell growth, as well as PGE2 and VEGF production (Segawa et al. 2009). Treatment with both NSAIDs and selective COX-2 inhibitors significantly reduced microvessel density in vivo, while expression of both VEGF and FGF-1 was also reduced. Microvessel density (MVD) was also positively correlated with tumor volume, lending support to this hypothesis (Wu et al. 2005). This reduction in MVD was associated increased tumor endothelial cell apoptosis (Raut et al. 2004; Wu et al. 2005). Furthermore, the selective COX-2 inhibitor, celecoxib, also significantly inhibited both PGE2 and TXB2 production in vivo (Leahy et al. 2002). In an implantation tumor model of colon cancer, treatment with the selective COX-2 inhibitor, celecoxib, reduced PGE2 levels and angiogenesis (CD-34 staining for microvessels). Treatment with high concentrations of celecoxib was also associated with significantly reduced levels of VEGF and MMP-2, while microvessel density in tumor tissue was strongly associated with PGE2, VEGF and MMP-2 levels (Wang et al. 2008). The selective COX-2 inhibitor, rofecoxib, was also shown to negatively regulate angiogenesis, although no effects on apoptosis or metastasis were observed (Fenwick et al. 2003). Selective COX-1 inhibition has also been investigated as an antiangiogenic approach for cancer treatment, although studies are limited. In a mouse model of ovarian cancer, the effects of non-selective COX inhibition with ibuprofen on angiogenic parameters were similar to those observed following selective COX-1 inhibition (Li et al. 2009). A number of in vitro and in vivo reports indicate anti-angiogenic effects following the inhibition of COX-derived prostanoids (Table 1). Selective inhibition of COX-2, or non-selective inhibition of cyclooxgenases or EP 1/2, significantly reduced the migratory and invasive potential of breast cancer cells (Rozic et al. 2001). Treatment of human intestinal microvascular endothelial cells (HIMEC) with the HDAC inhibitor sodium butyrate, inhibited angiogenic parameters as well as levels of COX-2, PGE2 and PGI2 (Ogawa et al. 2003). However, an angiogenic role for PGI2 has also been identified using a murine corneal model. PGI2 analogs such as iloprost and carboxyprostacyclin were able to act on nuclear PPARs to induce angiogenesis in vivo, an effect which was associated with increased VEGF expression (Pola et al. 2004). Interestingly,
