Significance of COX-1 Inhibition by NSAIDs

For decades, NSAIDs have been the cornerstone of pharmacologic management of arthritic and rheumatologic illnesses. NSAIDs are generally well tolerated, but they have tissue-specific toxicity. GI intolerance and GI bleeding were recognized early during NSAID use and have been persistent features of NSAID therapy for nearly a century. Prospective studies have shown significant risk of serious gastrointestinal complications and mortality associated with NSAID use, (20-26) which results in about 16,500 mortalities annually in the United States.(27) Although individual nonselective NSAIDs vary in their relative inhibition of COX-1 and COX-2, their toxicity is rather uniform. GI mucosal injury is believed to result from local and systemic events. Inhibition of COX-1-mediated prostaglandin leads to decreased mucus and bicarbonate, lowered mucosal blood flow, and inhibition of epithelial proliferation.(27) Additional side effects of blocking COX-1 include inhibition of platelet aggregation and increased bleeding, which contribute to GI consequences. NSAIDs also have renal effects and can result in fluid retention.(28)

Role of COX-2 in Arthropathy

COX-2 and inflammatory arthritis

The molecular biology of COX-2 regulation is consistent with observations that COX-2 expression increases in response to inflammatory stimuli, duress, and tissue repair.(3) Prostaglandins are clearly influential in the pathogenesis of arthritic disorders. Therefore, the relative expression of COX enzymes in arthritic tissues may offer clues to the potential therapeutic benefit of COX-2 inhibition.

In synovial tissues, the regulation of COX-2 transcription is under the influence of a number of cytokines abundant during arthritic inflammation, including IL-1 and TNF.(29) IL-1 enhanced de novo COX-2 transcripts but not COX-1 transcripts in synovial explants from patients with RA.(30) In addition, COX-2 mRNA is upregulated in the cellular response to fluid shear stress in the joint.(31) The effect of COX-2-selective inhibitors has been examined in rheumatoid synoviocytes and found to prevent PGE2 production in response to IL-1 and TNF.(32,33) In animal models of inflammatory arthritis, COX-2 synovial expression increased markedly, paralleling amplified PGE2 levels. Furthermore, pharmacologic inhibition of COX-2 abrogated inflammation in these models.(34,35) In humans, COX-1 levels are similar in normal synovium and that from patients with OA or RA. In synovia of OA and RA patients, however, significant upregulation of COX-2 transcription and expression occurs (Figure 1).(12,36-38)

Figure 1. Immunohistologic staining of COX-1 (A, B) and COX-2 (C, D). Samples are the same synovial tissues from OA (A, C) and RA (B, D). Positive immunoreactivity of COX-1 is seen in the synovial lining cells in OA (A) and RA (B), and COX-2 expression is seen to be intense in inflammatory cells from OA (C) and RA (D) (high power field, 400). (From Lee et al with permission.)(36)

COX-2 and nitric oxide

The nitric oxide (NO) and COX pathways share a number of potentially significant similarities.

Briefly, both enzymes are induced in tandem in inflammatory settings.(39) Cartilage explants from patients with OA or RA produce NO ex vivo, as do synoviocytes and chondrocytes.(40) IL-1 can also stimulate inducible nitric oxide synthase (NOS) pathways.(40) NO can substantially induce prostaglandin production via upregulation of COX- 2.(39) On the other hand, addition of an NOS inhibitor augments PGE2 production in OA cartilage explants, suggesting that NO may inhibit PGE2 release.(12) NO has detrimental effects on chondrocytes, and can inhibit collagen and proteoglycan synthesis. NO can activate MMPs, resulting in cartilage degradation. Finally, NO triggers chondrocyte apoptosis, a process enhanced by PGE2, and specific inhibition of COX-2 blocks NO-mediated chondrocyte apoptosis.(41)

COX-2 is emerging as a pivotal enzyme in the inflammation and tissue damage that occurs in the arthritic joint. Intensified expression of COX-2 but not COX-1 in rheumatoid tissues suggests an "Achilles' heel" in the prostaglandin-mediated biologic events because PGE2 and its downstream effects can be blocked with COX-2 inhibitors. It is this rationale that has provided the basis for the development and use of COX-2-selective inhibitors in clinical practice.

COX-2-selective inhibitors

Following cloning and characterization of COX- 2, it was clear that structural differences could be exploited for the development of selective inhibitors (42). The determination of selectivity, however, has only recently been formally addressed.

Conventional NSAIDs vary in their relative inhibition of COX-1 and COX-2 enzymes, and the reported ratio of COX-1 to COX-2 specificities for a specific agent can vary by up to 100-fold.(28) The International Consensus Meeting on the Mode of Action of COX-2 Inhibition (ICMMAC) brought together experts in rheumatology, gastroenterology, and pharmacology to assess the significance of differential inhibition of COX-1 and COX-2.(28) ICMMAC suggests that a drug be considered COX-2-selective if it inhibits COX-2 but not COX-1 across the entire therapeutic dose range based on whole blood assays. The panel concluded that, according to these criteria, with the exception of rofecoxib and celecoxib, all NSAIDs available in 1999 inhibit both isoenzymes and are COX-nonspecific.(28)

The clinical implications of even a small degree of COX-1 inhibition are unknown. Therefore, ICMMAC recommended that agents that preferentially inhibit COX-2 (based on a COX-1/COX-2 IC50 ratio) be considered nonselective if there is evidence that they may inhibit COX-1 at therapeutic concentrations. From a clinical perspective, the pivotal criteria for COX selectivity are safety and efficacy as demonstrated by large clinical trials in generalizable groups of patients.

Clinical Application of COX-2-Selective Inhibitors

Rofecoxib and celecoxib have been for some time the only available COX-2-selective inhibitors approved by the US Food and Drug Administration (recently, valdecoxib was approved for use in OA, RA, and menstrual pain). Rofecoxib and celecoxib are prescribed widely in the United States, and the use of COX- 2-selective inhibitors is now included in the current American College of Rheumatology treatment guidelines for OA.(43) Both of these coxibs lack clinically relevant COX-1 inhibition at or above therapeutic levels, though rofecoxib is about 30 times more selective for COX-2 than celecoxib. Both result in improved GI safety, and each has efficacy equivalent to that of nonselective NSAIDs. An additional agent, meloxicam, has recently been approved for use in the United States and exhibits a high degree of specificity for COX-2 but also inhibits COX-1 at a low dosage of 7.5 mg/day.(44) Studies of inhibition of serum thromboxane B2 show that celecoxib at single doses of 100 mg and 400 mg (but not 800 mg), and rofecoxib at doses of 12.5 mg and 25 mg do not inhibit COX-1 to a significant degree compared with placebo; meloxicam (15 mg) and ibuprofen (800 mg) both resulted in significant COX-1 inhibition.(44,45)

Detailed discussions of the efficacy of coxibs as analgesics, in the treatment of OA and RA, and of their GI safety are presented in this supplement. The cardiovascular and renal side effect profiles of coxibs have received much attention, and these issues are also discussed in detail.

Bingham CO III. Development and clinical application of COX-2-selective inhibitors for the treatment of osteoarthritis and rheumatoid arthritis. Cleve Clin J Med 2002;69:SI5-12.