Cell-Specific Gene Deletion Reveals the Antithrombotic Function of COX1 and Explains the Vascular COX1/Prostacyclin Paradox

Endothelial cells (ECs) and platelets, which respectively produce antithrombotic prostacyclin and prothrombotic thromboxane A(2), both express COX1 (cyclooxygenase1). Consequently, there has been no way to delineate any antithrombotic role for COX1-derived prostacyclin from the prothrombotic effects of platelet COX1. By contrast, an antithrombotic role for COX2, which is absent in platelets, is straightforward to demonstrate. This has resulted in an incomplete understanding of the relative importance of COX1 versus COX2 in prostacyclin production and antithrombotic protection in vivo. OBJECTIVE: We sought to identify the role, if any, of COX1-derived prostacyclin in antithrombotic protection in vivo and compare this to the established protective role of COX2. METHODS AND RESULTS: We developed vascular-specific COX1 knockout mice and studied them alongside endothelial-specific COX2 knockout mice. COX1 immunoreactivity and prostacyclin production were primarily associated with the endothelial layer of aortae; freshly isolated aortic ECs released >10-fold more prostacyclin than smooth muscle cells. Moreover, aortic prostacyclin production, the ability of aortic rings to inhibit platelet aggregation and plasma prostacyclin levels were reduced when COX1 was knocked out in ECs but not in smooth muscle cells. When thrombosis was measured in vivo after FeCl(3) carotid artery injury, endothelial COX1 deletion accelerated thrombosis to a similar extent as prostacyclin receptor blockade. However, this effect was lost when COX1 was deleted from both ECs and platelets. Deletion of COX2 from ECs also resulted in a prothrombotic phenotype that was independent of local vascular prostacyclin production. CONCLUSIONS: These data demonstrate for the first time that, in healthy animals, endothelial COX1 provides an essential antithrombotic tone, which is masked when COX1 activity is lost in both ECs and platelets. These results help us define a new 2-component paradigm wherein thrombotic tone is regulated by both COX1 and COX2 through complementary but mechanistically distinct pathways.