Growth of new blood vessels (angiogenesis), required for all tumor growth, is stimulated by the expression of vascular endothelial growth factor (VEGF). VEGF is up-regulated in all known solid tumors but also in atherosclerosis, diabetic retinopathy, arthritis, and many other conditions. Conventional VEGF isoforms have been universally described as proangiogenic cytokines. Here, we show that an endogenous splice variant, VEGF 165 b, is expressed as protein in normal cells and tissues and is circulating in human plasma. We also present evidence for a sister family of presumably inhibitory splice variants. Moreover, these isoforms are down-regulated in prostate cancer. We also show that VEGF 165 b binds VEGF receptor 2 with the same affinity as VEGF 165

Despite production by podocytes of the proangiogenic molecule vascular endothelial growth factor-A (VEGF), the glomeruli are not sites of angiogenesis. We recently described mRNA expression of an inhibitory splice variant of VEGF (VEGF165b) in normal kidney (Bates DO, Cui TG, Doughty JM, Winkler M, Sugiono M, Shields JD, Peat D, Gillatt D, and Harper SJ. Cancer Res 62: 4123-4131, 2002). Available anti-VEGF antibodies do not distinguish stimulatory from inhibitory VEGF families. To assess the production of VEGF165 (stimulatory) and VEGF165b (inhibitory) isoforms by human podocytes, we examined both primary cultured and conditionally immortalized human podocytes using family- and isoform-specific RT-PCR. In addition, VEGF protein production was analyzed in podocytes, using isoform-specific double-strand small-interference RNAs (siRNA). RT-PCR demonstrated the production of VEGF189 mRNA by podocytes of both phenotypes. In contrast, on differentiation there was a splicing change from VEGF165 to VEGF165b mRNA. In addition, VEGF protein in the supernatant of conditionally immortalized, differentiated podocytes was reduced by VEGF165b siRNA to 20+/-11% of the level of mock-transfected cells (P < 0.01). No reduction was seen with mismatch siRNA. Moreover, there was no reduction in VEGF protein concentration in the supernatant of primary cultured, dedifferentiated human podocytes (109+/-8% of mismatch siRNA, P > 0.1). In conclusion, differentiated but not dedifferentiated human podocytes secrete significant amounts of VEGF165b protein. It is possible that this may explain the paradox of high VEGF production in the glomerulus but no angiogenesis. Furthermore, the existence of this splicing switch in relation to podocyte phenotype suggests that alternative splicing of the VEGF pre-RNA is a regulated process that is open to manipulation and therefore could be a target for novel cancer therapies.