Joel Ellis Wilson and Dr. Dan Simmons, Chemistry and Biochemistry

Nonsteroidal anti-inflammatory drugs (NSAIDs) exhibit their effect by inhibiting the cyclooxygenase (COX) enzymes. However, inhibition of both COX-1 and COX-2 by the administration of NSAIDs to both humans and animals often leads to negative side effects, most commonly gastric toxicity, including ulcers, which may progress to perforation with risk of sepsis. Functional COX-1 and COX-2 are both necessary to maintain the integrity of the gastrointestinal lining. Fueled by the discovery of COX-2 and the elucidation of its biological properties, new NSAIDs have been produced that selectively inhibit only COX-2. This new class of inhibitors effectively reduces pain and inflammation with decreased side effects. COX-2 selective drugs such as celecoxib and refecoxib are now widely used to treat rheumatoid arthritis and osteoarthritis in humans.

Experts estimate that as many as 20% of the nearly 53 million dogs in the United States suffer to some degree from arthritis, making dogs the second largest target population for COX-2 selective inhibitors. The negative side effects associated with NSAIDs are even more difficult to manage in canines as mild clinical signs are easily overlooked. In addition, it has been reported that canines are even less tolerant of NSAIDs than humans and more at risk for NSAID-induced gastropathy. While many research efforts have sought to identify COX-2 selective inhibitors for use in humans, canines and other animals have not been studied as extensively. The aim of the current study was to evaluate the canine cyclooxygenase system to determine if the function and expression of canine COX-1 and COX-2 is similar to humans. In addition, we studied thirteen NSAIDs using a canine whole blood assay system to determine which currently available NSAIDs are most COX-2 selective in canines.

Our evaluation of the reported canine COX sequences with those encoding human COX-1 and COX-2 suggest that critical residues in the canine COX active sites are identical with their human homologs. COX-1 mRNA and protein expression in canines was found to be similar to that known in other mammalian species. Nearly all of the canine tissues studied express both COX-1 mRNA transcripts and COX-1 protein. Of particular interest was expression of COX-1 in the gastrointestinal tract, since prostaglandin synthesis serves to protect the integrity of the mucosal surfaces of these organs. NSAIDs in dogs exhibit a high degree of GI toxicity in comparison with humans. We determined that the COX-1 present in these tissues is functionally active and likely is the primary COX isoenzyme that protects the GI tract in canines as in other mammalian species.

COX-2 expression was detected at the RNA level in all tissues, without concomitant expression of COX-2 protein. The finding of COX-2 mRNA but not COX-2 protein in the eleven canine tissues examined presents an interesting phenomenon. These results suggest that either the expression of COX-2 is translationally blocked or post-transcriptionally regulated in canines. Post-transcriptional regulation may result in a high turnover rate of the COX-2 protein. A striking and unexpected finding was that canine cerebral cortex expressed two COX-1 mRNAs, one of approximately 2.7 kb and the other of 2.2 kb. The identity of the this new 2.2 kb mRNA is of interest both with regard to the role it may play in nociception as well as other potential neuronally based functions of COX-1 (e.g. Alzheimer’s disease and fever).

We showed, in earlier studies, that the use of NSAIDs to measure COX inhibition using TPA-stimulated canine kidney CCL34 cells is highly reproducible and inhibits COX activity in a dose dependent fashion. However, these studies were performed using only cell culture media as a drug delivery vehicle. Using the William Harvey Research Institute whole blood and modified whole blood assays, we surveyed the effect of thirteen NSAIDs on canine COX activity. In these assays, COX-1 inhibition in platelets was measured, as was COX-2 inhibition in TPA-stimulated CCL34 cells. These assays have been reported to be superior to other types of whole blood assays because measurement of both COX-1 and COX-2 inhibition can be determined in parallel.

Our study of the inhibition of cyclooxygenases by thirteen NSAIDs reveals that four of these NSAIDs (indomethacin, peroxicam, diclofenac, and meclofenamic acid) exhibited less than a 5- fold preferential inhibition of COX-2. Four drugs (carprofen, tolfenamic acid, nimesulide, and etodolac) exhibited greater than 5-fold selective inhibtion of COX-2. The known COX-2 selective inhibitor, NS398, exhibited 18-fold selective inhibition of COX-2. Another known COX-2 selective inhibitor, nimesulide, exhibited 8-fold COX-2 selectivity. Of the drugs tested, NS398 is known to be the most COX-2 selective in humans; nimesulide and etodolac are known to be preferential COX-2 inhibitors. In our study, carprofen exhibited 5.5-fold preferential inhibition toward COX-2, but less than these three compounds.

An unexpected finding was that tolfenamic acid, which has not been previously shown to exhibit COX-2 selectivity, showed 14-fold selectivity toward canine COX-2. Also of interest was the finding that the L enantiomer of etodolac, which is significantly less potent in inhibiting COXs than D-etodolac, was the most COX-2 selective of all drugs tested, showing almost 50-fold selectivity towards COX-2. In contrast, etodolac-D showed slight COX-1 preferential inhibition, while the enanteomeric mix of etodolac exhibited COX-2 preferential inhibition that was in between the two values for the purified enantiomers. In vivo, etodolac is metabolized within two hours, leaving almost entirely the L enantiomer, which may account for the known COX-2 selectivity and gastric sparing properties of this drug. In humans, ketoprofen and aspirin are both COX-1 preferential and known to induce ulceration. Similarly, in this study, we found ketoprofen, aspirin and etodolac-D to be COX-1 preferential.

One of the goals of current pharmacological research is the determination of a COX-2 selective inhibitor for use in canines, which represent the second largest target population for these drugs. Our studies confirm that the COX-1/COX-2 hypothesis applies also to canines and that the synthesis of highly selective, non-toxic COX-2 selective inhibitors will prove useful as gastric-sparing anti-inflammatory agents in these animals.