Osteoarthritis - Treatment Information
Discussion of Specific Cox-2 Inhibitors
by Joan M. Bathon, M.D
- The Discovery of Two Isoforms of COX
- Differential expression of COX-1 and COX-2
- The Rationale for the Development of Specific COX-2 Inhibitors
- Does COX-2 Serve a Physiological Function(s)?
- Evaluating the COX-2 Specificity of an NSAID
- Human Studies of Selective COX-2 Inhibition
- Clinical and Treatment Implications of Selective COX-2 Inhibition
- References for Specific COX-2 Inhibitors
The Discovery of Two Isoforms of COX
In the early 1990s, evidence emerged from two lines of investigation suggesting that there were two different COX enzymes (refs. 1-3). Investigators studying cell growth discovered a new gene product that was induced in vitro and that exhibited great similarity to COX. At the same time, other investigators were discovering that COX activity could be induced by cytokines such as interleukin-1 (IL-1) and inhibited by corticosteroids. Steroids inhibited the interleukin-1 induced COX activity but not basal COX activity. These observations led to the hypothesis that there are two COX isoenzymes. One COX enzyme was theorized to be constitutively (constantly) expressed and responsible for basal PG production, while the second COX enzyme was induced by inflammatory stimuli such as interleukin-1 and suppressed by glucocorticoids (slide). Indeed, these two distinct COX genes have now been cloned and localized to different chromosomes.
COX-1, the name assigned to the constitutively expressed enyzme, is found in nearly all tissues and cells. COX-2, the inducible enzyme, is responsible for the profound increase in prostanoid synthesis and release that occurs at sites of inflammation. Not surprisingly, COX-2 expression is tightly regulated and its mRNA transcript appears to be highly unstable.
(top of section)
(top of page)
Differential Expression of COX-1 and COX-2
COX-1 and COX-2 serve identical functions in catalyzing the conversion of arachidonic acid to prostanoids. The specific prostanoid(s) generated in any given cell is not determined by whether that cells expresses COX-1 or COX-2, but by which distal enzymes in the prostanoid synthetic pathways are expressed. For example, stimulated human synovial cells synthesize small amounts of PGE2 and prostacyclin but not thromboxane or PGD or PGF2a. Following exposure to interleukin-1, synovial cells make considerably more PGE2 and prostacyclin, but they still do not synthesize PGD, TxB2 or PGF2a (slide)(ref. 4). The IL1-induced increase in PGE2 and prostacyclin is mediated through COX-2 (ref. 5).
Thus, while the species of prostanoid synthesized in a cell is dependent upon the specific distal synthetic enzyme(s) expressed, the amount synthesized is determined by the amount of COX -1 and -2 activities expressed. COX-1 is expressed in nearly all cells (except red cells) in their basal (unstimulated) states, suggesting that low levels of prostanoids are important in serving critical physiological (homeostatic) functions. COX-1 mediated production of prostaglandins in the stomach serves to protect the mucosa against the ulcerogenic effects of acid and other insults (the so-called "cytoprotective" role of prostaglandins). COX-1 mediated production of thromboxane in platelets promotes normal clotting. And COX-1 mediated synthesis of prostaglandins in the kidney appears to be responsible for maintaining renal plasma flow in the face of vasoconstriction. COX-1 levels remain relatively stable in most cells, although mild increases (2-4 fold) have been reported in response to growth factors.
COX-2 levels, in contrast, are dramatically upregulated in inflamed tissues. For example, COX-2 expression and concomitant PGE2 production are greatly enhanced in rheumatoid synovium compared to the less inflamed osteoarthritic synovium, and in animal models of inflammatory arthritis (refs. 5,6). This is undoubtedly the result of excessive production of interleukin-1, tumor necrosis factor and growth factors in the rheumatoid joint. Initially, it was thought that COX-2 was not constitutively expressed (i.e., present in the basal state) in any tissues. However, recent work has demonstrated constitutive COX-2 expression in several organs, such as kidney, brain and ovary. This may have implications for treatment (See Clinical and Treatment Implications).
(top of section)
(top of page)
The Rationale for the Development of Specific COX-2 Inhibitors
Currently available NSAIDs are nonselective - that is, they inhibit both COX-1 and COX-2. It is likely that inhibition of COX-2 is responsible for the anti-inflammatory effects of these drugs, while inhibition of COX-1 is responsible for the recognized toxicities of NSAIDs, including: a) peptic ulcers and the associated risks of bleeding, perforation and obstruction; b) prolonged bleeding time; and, c) renal insufficiency . NSAIDs that would selectively inhibit COX-2 are thus highly desirable since inflamed tissues could be targeted without disturbing the homeostatic functions of prostaglandins in noninflamed organs. Theoretically, then, selective COX-2 inhibition should preserve the anti-inflammatory efficacy without causing the associated toxicities of NSAIDs.
(top of section)
(top of page)
Does COX-2 serve a physiological function(s)? Does COX-1 serve an inflammatory function(s)?
Low levels of COX-2 have been observed in some non-inflamed tissues, suggesting that COX-2, like COX-1, may also serve some normal physiological functions (slide) (ref. 7). In the kidney, COX-2 expression was enhanced in the macula densa in response to sodium restriction (a physiological, not inflammatory, stress), suggesting that COX-2 plays a role a role in salt and volume regulation. Interestingly, mice in which COX-2 was "knocked out" have a moderately high rate of still births due to severe renal dysplasia, indicating that COX-2 is also critical for development of the kidney (ref. 8). COX-2 also plays a role in ovulation and fertility, as evidenced by its enhanced expression in the ovary prior to ovulation, and in the uterus prior to parturition. Furthermore, COX-2 knockout mice that survive exhibit markedly decreased fertility.
The COX-2 isoform has also been observed under basal conditions in rat brain in the hippocampus, pyramidal cells and amygdala (ref. 7). Its function in the brain is not yet defined but is likely to include temperature control. Interestingly, epidemiological studies have demonstrated that individuals taking chronic NSAIDs have a lower incidence of Alzheimer’s disease than controls. The exact function of COX-2 in the brain remains to be elucidated.
In summary, COX-2 appears to play a role in the development of the kidney, in salt and water regulation by the kidney, and in ovulation and parturition, and perhaps in an unknown brain function(s).
That COX-1 mediates important physiological functions is now well established, but it remains to be determined whether it contributes at all to inflammatory responses. The mechanisms of many animal models of inflammation are not well defined, making interpretation of currently available data difficult. In COX-1 knock out mice, however, homozygous mutant animals had a reduced inflammatory response to the application of exogenous arachidonic acid to the ear (ref. 8).
Surprisingly, these COX-1 knockout mice did not exhibit spontaneous gastric ulcers or bleeding, as expected.
(top of section)
(top of page)
Evaluating the COX-2 Specificity of an NSAID
The in vitro potency of a drug is reflected by its IC50 value. This is the concentration at which the drug achieves 50% of its maximal inhibition of COX, and is usually expressed in molarity. The lower the IC50 value, the more potent the drug is. It has been proposed that the COX-2 selectivity of an NSAID can be defined by the ratio of the IC50 for COX-2 divided by the IC50 for COX-1 (IC50 COX-2/IC50 COX-1). NSAIDs with values < 1 would be said to have selectivity for COX-2; this means that a smaller dose of the drug is needed to inhibit COX-2 than is needed to inhibit COX-1. The smaller the number, the higher the selectivity for COX-2.
Conversely, compounds with ratios > 1 would be said to be COX-1 selective, while those with ratios equal to 1 are nonselective. Although a relatively simplistic concept, published data on any individual NSAID showed widely varying ratios (ref. 9). This is because a variety of assays, cell types, and experimental conditions were utilized from study to study,and also because COX-2 kinetics are different from those of COX-1 (ref. 10). Thus, the validity of this ratio has been called into question.
Currently, the more accepted definition of COX-2 selectivity is by ex vivo assay in whole blood. In these assays, COX-1 activity is measured by the release of thromboxane B2 (metabolite of thromboxane A2) from platelets during clotting. COX-2 is measured by release of prostacyclin from monocytes in response to an inflammatory stimulus such as LPS. Specific COX-2 inhibitors should inhibit prostacyclin, but not thromboxane B2, release over a range of assay times and drug doses. Using these assays, celecoxib (Celebrex™) and rofecoxib (VIOXX™), but not traditional NSAIDs, are considered to be COX-2 selective.
It is important to remember that the Federal Drug Administration (FDA) has not yet approved the designation of a drug as "COX-2 specific". Celecoxib and rofecoxib have the same general class labeling as traditional NSAIDs. FDA approval of a drug as "COX-2 selective" is likely to depend on its demonstrated ability to reduce serious NSAID-induced GI complications such as bleeding, perforation and obstruction. Currently available data for celecoxib and rofecoxib do not include enough patients for long enough periods of treatment to make a conclusion about their long-term GI safety. However, cumulative data from multiple short-term clinical trials(see ACR meeting highlights) for each drug do suggest that each reduces the incidence of serious GI complications.(ref. 11)
(top of section)
(top of page)
Human Studies of Selective COX-2 Inhibitors
Arthritis
Celecoxib and rofecoxib have been compared to traditional NSAIDs including ibuprofen, naproxen and diclofenac for the treatment of osteoarthritis (OA) and/or rheumatoid arthritis (RA).(ref. 12-15) In these reports, celecoxib and rofecoxib in OA patients were statistically significantly better than placebo, and comparable to traditional NSAIDs, in relieving joint pain. Similarly, RA patients taking rofecoxib or celecoxib reported statistically significantly less pain and joint swelling and tenderness, than did patients receiving placebo.
Gastrointestinal Side Effects
Endoscopically proven ulcers were considerably less frequent in celecoxib- and rofecoxib- treated RA and OA patients, compared to traditional NSAIDs. For example, in patients with OA, endoscopically proven ulcers >3 mm were observed in the following frequencies after 12 weeks of treatment: placebo, 10%; 25 mg rofecoxib, 4%: 50 mg rofecoxib, 7%; ibuprofen (800 mg tid), 28%.(ref. 16) In RA patients, the incidence of endoscopically determined GI ulcers at 12 weeks was as follows: placebo, 4%; celecoxib (100 mg bid), 6%; celecoxib (200 mg bid), 4%; celecoxib (400 mg bid), 6%; and naproxen (500 mg bid), 26%.(ref. 13) With regard to non-ulcer GI symptoms such as dyspepsia, rofecoxib and celecoxib have not been convincingly shown to date to reduce these symptoms when compared to traditional NSAIDs.
Effects on Bleeding
Neither rofecoxib or celecoxib significantly interferes with platelet aggregation or prolongs bleeding time.(ref. 13) However, both have been shown to slightly prolong prothrombin time in warfarin (Coumadin™) treated patients. Consequently, caution is urged if anticoagulated patients are treated with selective COX-2 agents.
Effects on the Kidney
Both selective COX-2 inhibitors can cause pedal edema in patients with normal renal function. They have not been tested for safety in patients with compromised renal function.
(top of section)
(top of page)
Clinical and Treatment Implications of Selective COX-2 Inhibition
In summary, the studies available thus far support the hypothesis that selective COX-2 inhibition will achieve comparable efficacy, but less toxicity, than conventional nonselective NSAIDs. However, long-term GI studies are necessary before the true safety of the COX-2 agents can be definitively ascertained. The point prevalence of ulcers in patients on long-term NSAID treatment is about 20%, and the annual incidence of serious complications from these ulcers (i.e., bleeding, perforation and obstruction) is 1-4%. The critical question to be answered is whether selective COX-2 agents will significantly reduce the annual incidence of serious GI events related to ulcers, and/or reduce the associated costs and mortality rates, associated with these events.
Who might benefit from selective COX-2 inhibitors? Theoretically any patient requiring chronic NSAID therapy for management of arthritis or pain might benefit. However, specific groups of patients who are at higher risk for serious GI events from NSAIDs would be obvious candidates for COX-2 specific agents. These include the elderly, those with documented prior ulcers, and patients on concomitant steroids.
In patients who are not in the high risk category for serious GI events from NSAID-associated ulcers, cost may be an important factor in deciding whether to prescribe a selective COX-2 agents versus a conventional NSAID, since the latter are likely to be considerably cheaper. The addition of a proton pump inhibitor to dyspeptic patients on conventional NSAIDs may negate this cost advantage, although dyspepsia may also occur in patients receiving selective COX-2 agents (despite absence of ulcers!).
Can patients with chronic renal insufficiency and patients undergoing surgical procedures be treated with selective COX-2 inhibitors? The clinical trials of COX-2 inhibitors to-date excluded patients with abnormal renal function. Therefore, the safety of these agents in patients with elevated creatinines is unknown. Furthermore, as mentioned in the discussion of physiological function above, COX-2 may play a physiological role in the regulation of salt and water balance by the kidney. Therefore, further work is needed in this area before this question can be answered. However, we do know from clinical trials in patients with normal renal function that COX-2 inhibitors, while causing some pedal edema, do not cause elevation of creatinine.
The absence of effect of the specific COX-2 inhibitors on platelet aggregation would suggest that they could probably be administered safely to patients up until the time of surgery. (Nonselective NSAIDs are generally discontinued 1-2 weeks prior to surgery). The long-term safety of COX-2 inhibitors in patients taking anticoagulants remains to be determined.
Will selective COX-2 inhibitors be useful for treating diseases other than arthritis and pain? Data suggest that COX-2 inhibition may reduce the prevalence of colon cancer and of Alzheimer’s disease. In fact, the FDA has just approved celecoxib for the reduction in the number of adenomatous colorectal polyps in familial adenomatous polyposis as an adjunct to usual care.
(top of section)
(top of page)
References for Specific COX-2 Inhibitors
1. Xie W, Chipman JG, Robertson DL, Erikson RL, Simmons DL. Expression of a mitogen-responsive gene encoding prostaglandin synthase is regulated by mRNA splicing. Proc Natl Acad Sci USA 88:2692-6, 1991
2. Masferrer JL, Zweifel BS, Seibert K, Needleman P. Selective regulation of cellular cyclooxygenase by dexamethasone and endotoxin in mice. J Clin Invest 86:1375-9, 1990
3. Kubuju DA, Fletcher BS, Barnum BC, Lim RW, Herschman HR. TIS10, a phorbol ester tumor promoter-inducible mRNA from Swiss 3T3 cells, encodes a novel prostaglandin synthase/cyclooxygenase homologue. J Biol Chem 266:12866-72, 1991
4. Bathon JM, Chilton FH, Hubbard WC, Towns MC, Solan NJ, and Proud D. Mechanisms of prostanoid synthesis in human synovial cells: Cytokine-peptide synergism. Inflammation 20:537-554, 1996
5. Crofford LJ, Wilder RL, Ristimaki AP, Sano H, Remmers EF, Epps HR, and Hla T. Cyclooxygenase-1 and -2 expression in rheumatoid synovial tissue. J Clin Invest 93:1095-1101, 1994
6. Anderson GD, Hauser SD, McGarity KL, Bremer ME, Isakson PC, and Gregory SA. Selective inhibition of cyclooxygenase (COX-2) reverses inflammation and expression of COX-2 and interleukin 6 in rat adjuvant arthritis. J Clin Invest 97:2672-79, 1996
7. Dubois RN, Abramson SB, Crofford L, Gupta et al. Cyclooxygenase in biology and disease. FASEB J 12:1063-1073, 1998
8. Langenbach R, Morhan SG, Tiano HF, Loftin CD, et al. Prostaglandin synthase 1 gene disruption in mice reduces arachidonic acid-induced inflammation and indomethacin induced gastric ulceration. Cell 83:483-492, 1995
9. Pairet M. Inhibition of cyclooxygenase-1 and cyclooxygenase-2. Analysis of invitro test systems and their clinical relevance. J Clin Rheum 4:S17-25, 1998 (Suppl)
10. Kurumbail RG, Stevens AM, Gierse JK, McDonald JJ, Stegeman RA, et al. Structural basis for selective inhibition of cyclooxygenase-2 by anti-inflammatory agents. Nature384:644-648, 1996.
11. Langman MJ et al. Adverse upper gastrointestinal effects of rofecoxib compared with NSAIDs. JAMA 282:1919-33, 1999.
12. Ehrich EW et al. Effect of specific COX-2 inhibition in osteoarthritis of the knee: a 6 week double blind, placebo controlled pilot study of rofecoxib, Rofecoxib Osteoarthritis Pilot Study Group. J Rheumatol 26:2438-47, 1999.
13. Simon LS, Lanza FL, Lipsky PE, Hubbard RC, Talwalker S, et al. Preliminary study of the safety and efficacy of SC-58635, a novel cyclooxygenase 2 inhibitor. Arthritis Rheum 41:1591-1602, 1998
14. Emery P et al. Celecoxib versus diclofenac in long-term management of rheumatoid arthritis: randomised double-blind comparison. Lancet 354: 2106-11, 1999.
15. Schnitzer TJ et al. The Safety profile, tolerability and effective dose range of rofecoxib in the treatment of rheumatoid arthritis. Clin Ther 21:1688-1702, 1999.
16. Laine L et al. A randomized trial comparing the effect of rofecoxib, a cyclooxygenase-2 specific inhibitor, with that of ibuprofen on the gastroduodenal mucosa of patients with osteoarthritis. Gastroenterology 117:776-783, 1999.
