Cyclooxygenase-2 inhibition prevents renal toxicity but not hypertension during sunitinib treatment

Background: Anticancer angiogenesis inhibitors cause hypertension and renal injury. Previously we observed in rats that high-dose aspirin (capable of blocking cyclooxygenase (COX)-1 and-2) was superior to low-dose aspirin (blocking COX-1 only) to prevent these side-effects during treatment with the angiogenesis inhibitor sunitinib, suggesting a role for COX-2. High-dose aspirin additionally prevented the rise in COX-derived prostacyclin (PGI(2)). Therefore, we studied the preventive effects of selective COX-2 inhibition and the hypothesized contributing role of PGI(2) during angiogenesis inhibition.Methods: Male WKY rats received vehicle, sunitinib ((SU), 14 mg/kg/day) alone or combined with COX-2 inhibition (celecoxib, 10 mg/kg/day) or a PGI(2) analogue (iloprost, 100 mu g/kg/day) for 8 days (n = 8-9 per group). Mean arterial pressure (MAP) was measured via radiotelemetry, biochemical measurements were performed via ELISA and vascular function was assessed via wire myography.Results: SU increased MAP (17 +/- 1mmHg versus 3 +/- 1mmHg after vehicle on day 4, P < 0.002), which could not be significantly blunted by celecoxib (+12 +/- 3mmHg on day 4, P = 0.247), but was temporarily attenuated by iloprost (treatment days 1 + 2 only). Urinary PGI(2) (996 +/- 112 versus 51 +/- 11ng/24h after vehicle, P < 0.001), but not circulating PGI(2) increased during SU, which remained unaffected by celecoxib and iloprost. Celecoxib reduced sunitinib-induced albuminuria (0.36 +/- 0.05 versus 0.58 +/- 0.05mg/24h after SU, P = 0.005). Wire myography demonstrated increased vasoconstriction to endothelin-1 after SU (Emax P = 0.005 versus vehicle), which remained unaffected by celecoxib or iloprost.Conclusion: Selective COX-2 inhibition ameliorates albuminuria during angiogenesis inhibition with sunitinib, which most likely acts independently of PGI(2). To combat angiogenesis inhibitor-induced hypertension, dual rather than selective COX-1/2 blockade seems preferential.