Thrombosis Is Reduced by Inhibition of COX-1, but Unaffected by Inhibition of COX-2, in an Acute Model of Platelet Activation in the Mouse

Background: Clinical use of selective inhibitors of cyclooxygenase (COX)-2 appears associated with increased risk of thrombotic events. This is often hypothesised to reflect reduction in anti-thrombotic prostanoids, notably PGI2, formed by COX-2 present within endothelial cells. However, whether COX-2 is actually expressed to any significant extent within endothelial cells is controversial. Here we have tested the effects of acute inhibition of COX on platelet reactivity using a functional in vivo approach in mice. Methodology/Principal Findings: A non-lethal model of platelet-driven thromboembolism in the mouse was used to assess the effects of aspirin (7 days orally as control) diclofenac (1 mg.kg(-1), i.v.) and parecoxib (0.5 mg.kg(-1), i.v.) on thrombus formation induced by collagen or the thromboxane (TX) A(2)-mimetic, U46619. The COX inhibitory profiles of the drugs were confirmed in mouse tissues ex vivo. Collagen and U46619 caused in vivo thrombus formation with the former, but not latter, sensitive to oral dosing with aspirin. Diclofenac inhibited COX-1 and COX-2 ex vivo and reduced thrombus formation in response to collagen, but not U46619. Parecoxib inhibited only COX-2 and had no effect upon thrombus formation caused by either agonist. Conclusions/Significance: Inhibition of COX-1 by diclofenac or aspirin reduced thrombus formation induced by collagen, which is partly dependent upon platelet-derived TXA(2), but not that induced by U46619, which is independent of platelet TXA(2). These results are consistent with the model demonstrating the effects of COX-1 inhibition in platelets, but provide no support for the hypothesis that acute inhibition of COX-2 in the circulation increases thrombosis.