Non-steroidal anti-inflammatory drugs (NSAIDs) remain among the most widely prescribed drugs worldwide. A great deal of research in this area has followed Vane's critical discovery in 1971 that NSAIDs such as aspirin exert their actions primarily by inhibiting the production of prostaglandins (PGs). Cyclo-oxygenase (COX) (also known as PGH synthase or prostaglandin endoperoxide synthase), the key enzyme catalyzing the biosynthesis of PGs, was purified in 1976 and cloned in 1988. A second COX gene was discovered in 1991. We now know that the two genes express two similar but distinct isoforms of the enzyme - COX-1 and COX-2. The two isoforms have similar primary protein structure (60% homology) and catalyze essentially the same reaction.

In the meantime a great deal has been learned about the metabolism, regulation, and functions of PGs, the product of COX reaction. Prostaglandins are ubiquitous fatty-acid derivatives that serve as autocrine / paracrine mediators involved in many different physiological processes in addition to their well recognized role in inflammation and immune response modulation. PGs are involved in as diverse normal processes as renal function, vasomotor tone, platelet aggregation and blood clotting, differentiation of immune cells, wound healing, nerve growth, bone metabolism, ovulation, and initiation of labor.

Figure 1 illustrates the main pathways by which PGs and other eicosanoids are produced. Tissue damage activates phospholipase A₂ (PL-A₂), which causes arachidonic acid to be split off the cell membrane phospholipids. The fate of the released arachidonic acid depends on which of several possible pathways it takes. There are two main pathways; the lipoxygenase pathway leads to the formation of leukotrienes and lipoxins, whereas the COX pathway leads to formation of prostaglandins and thromboxanes.

The two known COX isoforms are similar in size, substrate specificity, and kinetics, but vary in their expression and distribution (Figure 2). COX-1, "the good COX", is a constitutively expressed isoform that is always present in most cells throughout the body and catalyzes the formation of PGs involved in physiological, "housekeeping" processes. In the stomach, COX-1 catalyzes the synthesis of PGE₂ and PG I₂ that have cytoprotective actions and play an important role in maintaining the integrity of the gastroduodenal mucosa.

It is estimated that 25% of patients using NSAIDs experience some kind of side effect and about 5% develop serious health consequence (massive GI bleed, acute renal failure, etc.). Both COX-1 and COX-2 are inhibited to varying degrees by all currently available (conventional) NSAIDs. Studies published so far support the hypothesis that the undesirable side effects of NSAIDs such gastric erosion and renal dysfunction are due to the inhibition of COX-1, while the anti-inflammatory (therapeutic) effects are due to the inhibition of COX-2.

Inhibitory potency and selectivity of the conventional, 1^st generation NSAIDs (aspirin, diclofenac, ibuprofen, indomethacin, naprosyn, and piroxicam, etc) for COX-1 and COX-2 vary greatly. Some NSAIDs (e.g., ketoprofen) are relatively COX-1 selective, some (ibuprofen and naproxen) are essentially non-selective, while others (e.g., diclofenac) are relatively COX-2 selective. Inhibitory effects of NSAIDs on gastric PGE₂ synthesis correlate with COX-1 inhibitory potency in blood and with COX-1 selectivity, but not with COX-2 inhibitory potency. However, even COX-2 "selective" NSAIDs still had sufficient anti-COX-1 activity to cause potent inhibition of gastric PGE₂. Thus, at therapeutic concentrations, none of the currently marketed NSAIDs spare gastric COX-1 activity.

In general, NSAIDs have comparable efficacy but different safety profiles probably due to differences in their ability to inhibit COX-1 at therapeutic doses. In this respect, the different NSAIDs are evaluated by comparing their IC₅₀ values. The IC₅₀ is the drug concentration that inhibits the activity of the enzyme by 50%. Therefore, the lower the IC₅₀ value the stronger is the drug. This approach is now utilized to study the relative inhibition of the two COX isoforms by a given NSAID. The IC₅₀ values (micromoles/L) for COX-2 and COX-1 are determined is vitro and their ratio is calculated (COX-2 IC₅₀ divided by COX-1 IC₅₀) to provide a quantitative measure of the drug's selectivity. The smaller this ratio the more selective is the drug for COX-2. For example, while the ratios for indomethacin and piroxicam are similar and are about 30, the ratio for the slightly more selective meloxicam is about 0.3. Meloxicam is commercially available in some parts of the world, but not in US. However, at therapeutic doses "preferential" COX-2 inhibitors like meloxicam effectively inhibit both isoforms and therefore undesirable side effects will occur despite the preferential inhibition of COX-2. However, there are currently at least a dozen highly selective COX-2 inhibitors under development by different pharmaceutical companies. These 2nd generation NSAIDs represent a new class of drugs (truly selective COX-2 inhibitors) and are considered a major advance in the management of pain and inflammatory diseases. Of these drugs, celecoxib (SC-58635 or Celebra by Searle) is the first to pass through the FDA advisory panel and will probably be approved shortly for use in rheumatoid and osteoarthritis. Celecoxib is 375 times more selective for COX-2 relative to COX-1 and, at therapeutic doses, its plasma concentration does not reach the level required for effective COX-1 inhibition. In clinical trials celecoxib showed effective anti-inflammatory activity with virtually no gastrointestinal adverse effects compared to placebo. Other agents under developments, now dubbed "the super aspirins", may have a COX-2 selectivity several fold greater than that of celecoxib with virtually on effect on COX-1 and may therefore afford a much wider margin of safety.