Comparison of the cyclooxygenase-1 inhibitory properties of nonsteroidal anti-inflammatory drugs (NSAIDs) and selective COX-2 inhibitors, using sensitive microsomal and platelet assays.

Two forms of cyclooxygenase (COX) activity are involved in the synthesis of prostaglandins, prostacyclins, and thromboxanes in mammalian cells. There is now convincing evidence, obtained with a number of structurally distinct inhibitors, that selective COX-2 inhibitors possess anti-inflammatory effects with an improved gastrointestinal tolerability compared with conventional nonsteroidal anti-inflammatory drugs (NSAIDs) affecting both COX-1 and COX-2. As more selective COX-2 inhibitors are being developed, assays with a high degree of sensitivity to inhibition are needed to compare the relative effects of compounds on COX-1 activity. In the present report, we describe a sensitive assay for the inhibition of human COX-1 based on the production of prostaglandin E2 by microsomes from U937 cells incubated with a subsaturating concentration of arachidonic acid. More than 45 NSAIDs and selective COX-2 inhibitors were tested in this assay. IC50 values ranged from 1 nM for flunixin and flurbiprofen to about 200-500 microM for salicylate and acetaminophen. Potent and nonselective NSAIDs such as sulindac sulfide, diclofenac, and indomethacin showed IC50 values of < 20 nM. Among the compounds that have been reported to show selectivity for COX-2, the rank order of potency against COX-1 was DuP 697 > SC-58451 > celecoxib > nimesulide-meloxicam-piroxicam-NS-398-RS-57067 > SC-57666 > SC-58125 > flosulide > etodolac > L-745,337 > DFU-T-614, with IC50 values ranging from 7 nM to 17 microM. A good correlation was obtained between the IC50 values for the inhibition of microsomal COX-1 and both the inhibition of TXB2 production by Ca2+ ionophore challenged platelets and the inhibition of prostaglandin E2 production by CHO cells stably expressing human COX-1. However, the microsomal assay was more sensitive to inhibition than cell-based assays and allowed the detection of inhibitory effects on COX-1 for all NSAIDs and selective COX-2 inhibitors examined with discrimination of their potency under conditions of limited availability of arachidonic acid.