Safety of Anti-Inflammatory Treatment—New Ways of Thinking

Home , Licofelone, Lipoxin

Rheumatology 2004;43(Suppl. 1):i16–i20 doi:10.1093/rheumatology/keh104 Safety of anti-inﬂammatory treatment—new ways of thinking K. Brune

The development of osteoarthritis may be accompanied by increased production of leukotrienes (LTs) and prostaglandins (PGs) from arachidonic acid. These products contribute to joint damage, pain and inflammation. Cyclooxygenase (COX)-1 and Downloaded from https://academic.oup.com/rheumatology/article/43/suppl_1/i16/1788145 by guest on 27 September 2021 COX-2 are responsible for the production of PGs. Inhibition of these enzymes by non-steroidal anti-inflammatory drugs and selective COX-2 inhibitors reduces the levels of PGs, resulting in a reduction in pain and inflammation. However, this inhibition can cause alternative processing of arachidonic acid via the 5-lipoxygenase (5-LOX) pathway, resulting in increased production of proinflammatory and gastrotoxic LTs. Licofelone is a competitive inhibitor of 5-LOX, COX-1 and COX-2 that is currently being developed for the treatment of osteoarthritis. Licofelone decreases the production of both LTs and PGs, and thereby reduces inflammation and pain with low gastrotoxicity. Unlike selective COX-2 inhibitors, coadministration of licofelone and aspirin does not appear to be associated with an increase in gastrointestinal adverse events, at least under experimental conditions. Furthermore, there is evidence from animal models to suggest that licofelone may stop disease progression.

KEY WORDS: Selective COX-2 inhibitors, NSAIDs, Osteoarthritis, Licofelone, LOX/COX inhibition, Renal, Cardiovascular, Gastrointestinal.

Osteoarthritis (OA) is a complex disease that is characterized by the gastrointestinal (GI) mucosa. Of particular relevance to OA is joint pain, inflammation and stiffness caused by the degradation of the observation that increases in PGI2 and PGE2 levels result in collagen and proteoglycans in cartilage tissue. Although the exact hyperalgesia, leading to pain, and also inflammation [1–4]. cause or trigger for OA is not clear, many mediators are known to Additionally, PGs—in particular PGE2—are thought to stimulate be involved in the pathophysiology and progression of the disease. bone resorption by increasing the number of osteoclasts, which These include cartilage-degrading enzymes, cytokines and, as could contribute to the joint damage seen in OA [5]. reviewed below, leukotrienes (LTs) and prostaglandins (PGs), Many of the processes affected by PGs are also affected by LTs, which contribute to pain and inflammation in OA. including vascular homeostasis, vascular permeability, gastric vessel constriction and gastric acid secretion [1–3, 6]. For example, the cysteinyl LTs (LTC4, LTD4 and LTE4) are vasoconstrictors and Role of leukotrienes and prostaglandins in osteoarthritis play an important role in mediating vascular permeability. Their vasoconstrictive effects are antagonistic to the effects of PGs and LTs and PGs are produced by the activity of three enzymes, result in gastric vessel vasoconstriction, gastric acid secretion, namely 5-lipoxygenase (5-LOX), cyclooxygenase (COX)-1 and proinflammatory cytokine production and consequent GI damage COX-2, as part of the arachidonic acid (AA) pathway (Fig. 1) [1]. In addition, the cysteinyl LTs have powerful effects on the [1–3]. 5-LOX (together with other enzymes) converts AA to the smooth muscle of the airways, leading to bronchoconstriction [1, 3]. leukotrienes LTB4, LTC4, LTD4 and LTE4. COX-1 converts AA LTB4 and the cysteinyl LTs are potent mediators of inflamma- to, among other molecules, thromboxanes, such as TXA2, and tion, causing increased activation, recruitment, migration and PGs, such as PGD2, PGE2, PGF2 and PGI2 (prostacyclin). The adhesion of immune cells. Indeed, the levels of LTB4 and PGE2 in activity of COX-2 leads to production of a narrower spectrum of particular are higher in the joints of patients with OA than in those PGs, specifically PGE2 and PGI2. The AA pathway, its products of healthy individuals [7, 8]. Furthermore, LTB4 increases the and the enzymes mediating their formation play a major role production and release of the cytokines, tumour necrosis factor- in many aspects of human physiology, including vascular (TNF- ) and interleukin1- (IL-1 ) [9]. This is important in OA homeostasis, gastroprotection, renal homeostasis and bone because these cytokines are thought to mediate damage to cartilage formation, and pathophysiological processes, including pain and by increasing the expression of degradative enzymes and reducing inflammation in OA [1–4]. the repair of damaged cartilage by chondrocytes. Similarly to PGs have various physiological and pathophysiological effects. PGE2, LTB4 is also thought to have a role in stimulating bone For example, PGI2 and PGE2 affect vascular homeostasis because resorption in OA, and this may be owing to increased TNF- and they have potent effects on vasodilation and vascular permeability, IL-1 production [9, 10]. Although many of these observations are and PGI2 also inhibits platelet aggregation [1–3]. Furthermore, from in vitro studies, it is likely that they are pertinent, at least in both PGI2 and PGE2 are involved in modulating normal glomer- part, to human OA. ular filtration rate and blood flow. The vasodilatory properties of In addition to LTs, PGs and TXA2, which is a potent these two molecules increase mucus production and reduce acid vasoconstrictor that stimulates platelet aggregation, the AA and pepsin levels in the stomach, thereby protecting the integrity of pathway also produces lipoxins (LXA4, LXB4 and 15-epi-LXs)

Department of Experimental and Clinical Pharmacology and Toxicology, Friedrich Alexander University, Erlangen-Nuremberg, Erlangen, Germany

Submitted 29 August 2003; revised version accepted 25 November 2003. Correspondence to: K. Brune, Department of Experimental and Clinical Pharmacology and Toxicology, Friedrich Alexander University, Erlangen- Nuremberg, Postfach 35 20, 91023 Erlangen, Germany. E-mail: [email protected] i16 Rheumatology Vol. 43 Suppl. 1 ß British Society for Rheumatology 2004; all rights reserved Safety of anti-inﬂammatory treatment i17 Downloaded from https://academic.oup.com/rheumatology/article/43/suppl_1/i16/1788145 by guest on 27 September 2021

FIG. 1. Arachidonic acid metabolism to leukotrienes and prostaglandins. via 12-LOX and 15-LOX [1], and endocannabinoids via fatty acid amide hydrolase [11]. These molecules have a range of physiological functions [1]. LXA4 and LXB4 are formed as part of the inﬂammatory response, but, in contrast to LTs and PGs, have anti- inﬂammatory effects and inhibit LT-stimulated chemotaxis. Endocannabinoids are thought to function as neuromodulators and vasodilators [12] and also as mediators of antinociception [11].

Cyclooxygenase inhibition and gastrointestinal damage Non-selective non-steroidal anti-inflammatory drugs (NSAIDs), such as naproxen, inhibit both COX-1 and COX-2 in the AA pathway and are commonly used in the treatment of OA. This leads to a reduction not only in the levels of the PGs (largely produced by COX-2) that cause pain and inflammation in OA, but also in the levels of gastroprotective PGs (mainly produced by COX-1). The resulting GI damage has been well documented as a serious complication associated with the use of NSAIDs, as discussed in the previous article in this Supplement. Selective COX-2 inhibitors, such as celecoxib and rofecoxib, were developed to overcome the GI disadvantages of NSAIDs and are also used to treat the signs and symptoms of OA. These agents reduce the levels of inflammatory PGs while leaving the levels of gastroprotective PGs largely unaltered [1, 4]. However, it has emerged that there is still a risk of GI damage associated with the use of selective COX-2 inhibitors [13–15]. Accumulating evidence suggests that this may be because of an overlap in the roles of COX-1 and COX-2 in physiological and pathophysiological processes. One trial comparing a selective COX-2 inhibitor (rofecoxib) with a non-selective NSAID (naproxen) found that the incidences of confirmed GI events were 2.1 and 4.5 per 100 patient-years, respectively, leading to a relative risk of 0.5 (P < 0.001) [16]. However, the risk of GI complications with selective COX-2 inhibitor therapy is greater in patients who are elderly, taking low-dose aspirin or have a history of GI ulceration [17]. Selective COX-2 inhibitors may also delay GI ulcer healing, possibly by shifting the balance of angiogenic factors to reduce new blood vessel growth [18]. Complications arising from GI ulcers can prove fatal, and therefore inhibition of ulcer healing is an undesirable side-effect, particularly in patients with an increased risk of FIG. 2. Leukotriene levels in the gastric mucosa. Reproduced developing ulcers. from an article in Gut [19] with permission from the BMJ It has been proposed that the inhibition of one or both COX Publishing Group. enzymes, while reducing the levels of gastrotoxic PGs, may result in alternative processing of AA via the 5-LOX pathway. This secretion [1]. One study examined gastric mucosal synthesis of increases the production of cysteinyl LTs and LTB4, which LTB4 in patients taking NSAIDs (Fig. 2) and found the median contribute to GI toxicity by promoting the migration of leucocytes, level of LTB4 was 0.9 pg/mg (0.2–2.5 pg/mg), compared with breaking down the mucosal barrier and stimulating gastric acid 0.0 pg/mg (0.0–0.6 pg/mg) in patients not taking NSAIDs i18 K. Brune

(P < 0.001) [19]. The increased levels of LTB4 were associated with In addition to the role of LTs and PGs in GI damage/protection, GI damage. there is evidence to suggest that the adhesion of leucocytes to As mentioned earlier, in addition to their role in GI homeostasis, mesenteric venules could play a role in the development of NSAID- PGs have a role in renal function and vascular homeostasis. It is induced GI damage [28, 29]. This could be caused by leucocyte perhaps not surprising, therefore, that the modulation of PG levels chemotaxis due to increased LT levels following a shunt in AA through the activity of selective COX-2 inhibitors and non- metabolism to the 5-LOX pathway. A study was conducted to selective NSAIDs is associated with adverse renal effects such as compare the adherence of leucocytes to mesenteric venules in rats ﬂuid and electrolyte disturbances and blood pressure elevation treated with indomethacin or licofelone [30]. Within 60 min of [20–23]. Moreover, selective COX-2 inhibitors may interfere with indomethacin administration, there was a substantial increase in blood pressure control in hypertension via variable mechanisms leucocyte adhesion. This increase continued over the course of the [24, 25]. Furthermore, selective COX-2 inhibitors have been study (120 min), and was statistically signiﬁcant when compared associated with thromboembolic complications such as myocardial with the control (P < 0.05). Conversely, licofelone was associated Downloaded from https://academic.oup.com/rheumatology/article/43/suppl_1/i16/1788145 by guest on 27 September 2021 infarction [16]. Consequently, selective COX-2 inhibitors should be with only a small increase in adhesion, which did not increase used with caution in patients with a history of, or at risk of, further after 90 min. cardiovascular or renal disease [22].

Adaptation to aspirin Development of licofelone—a 5-LOX, COX-1 and COX-2 inhibitor Hypertension and cardiovascular disease are common in elderly patients with OA, and aspirin is often used for cardiovascular As COX inhibitors are associated with adverse GI (and possibly prophylaxis by these patients [31–34]. Aspirin covalently modifies renal and cardiovascular) events, there is a clear, unmet medical the active site of COXs, preventing access to AA [35]. While this need for drugs that provide the pain and inflammation relief seen prevents the action of COX-1, the activity of COX-2 is modified, with these agents, but without the associated adverse events. As resulting in the production of an intermediate that is converted by described previously, these adverse GI events may be ascribed to a 5-LOX to 15-epi-lipoxin A4, also known as aspirin-triggered shunt in AA metabolism to the 5-LOX pathway. One possible lipoxin (ATL) [34]. At low doses of aspirin, the production of solution to this problem would therefore be a drug that could ATL limits leucocyte recruitment and activation and mediates the inhibit 5-LOX, COX-1 and COX-2, leading to decreased produc- anti-adhesive—and hence cardioprotective—effects of aspirin. The tion of LTs and PGs. Such a drug would have the potential to activity of aspirin also results, via the action of 5-LOX, in increased provide anti-inflammatory and analgesic effects with a reduction in levels of LTB4, which, as described earlier, is a potent leucocyte the GI side-effects associated with the use of COX enzyme activator and chemoattractant. At higher doses of aspirin, inhibitors [26]. Licofelone, a 5-LOX, COX-1 and COX-2 inhibitor, gastrotoxic levels of LTB4 are produced [36]. Over time, however, was developed for the treatment of OA with this rationale in mind. the GI mucosa shows increased resistance to the damage caused by Prior to the development of licofelone, several attempts had LTB4, owing to the formation of ATL, in a process formerly been made to develop LOX/COX inhibitors. The LOX and COX termed adaptation [37]. enzymes are redox-active and contain an iron ion at their active Importantly, the use of COX-2 inhibitors prevents the produc- site. Inhibitors developed prior to licofelone (such as tebufelone) tion of ATL, and therefore prevents adaptation. As indicated were redox-active and, in addition to inhibiting LOX and COX earlier, studies have shown that coadministration of aspirin and a enzymes, also inhibited redox enzyme systems in the liver, resulting selective COX-2 inhibitor increases the risk of GI damage [38]. in hepatotoxicity [26]. Licofelone is not redox-active and conse- This is likely to be a result of an imbalance between the quently is not hepatotoxic. Indeed, licofelone is an AA substrate gastroprotective effects of ATL and the gastrotoxic effects of analogue and is a near-perfect match for the horseshoe conforma- LTB [39]. With the use of licofelone, however, the adaptive tion of AA. Furthermore, licofelone is a competitive inhibitor of 4 response to aspirin does not appear to be affected, as the inhibition 5-LOX, COX-1 and COX-2, and inhibits all three enzymes in the of 5-LOX, COX-1 and COX-2 reduces the production of both same concentration range, leading to a decrease in LT and PG LTB and ATL. In one animal study, adaptation to aspirin was production [26, 27]. It is important to note, however, that 4 prevented by cotreatment with celecoxib, indomethacin or keto- licofelone does not inhibit 12-LOX or 15-LOX, and therefore rolac (P < 0.01) (Fig. 3) [39]. In contrast, licofelone cotreatment does not compromise the production of the lipoxins, which have anti-inflammatory effects [1]. Studies in animal systems have did not affect adaptation. Indeed, a single dose of licofelone on the shown that licofelone has analgesic, anti-inflammatory, anti- last day of the study did not exacerbate the GI damage caused by pyretic, antiplatelet and antibronchoconstrictory activities [26]. aspirin, whereas a single administration of a COX inhibitor was associated with GI damage (Fig. 4). As OA patients commonly use aspirin, the GI toxicity due to coadministration of aspirin and a COX inhibitor may have important clinical consequences. Evidence of gastrointestinal safety with licofelone The GI tolerability of licofelone compared with NSAIDs has been investigated in a rat model by examining the levels of the Cyclooxygenase inhibition and cardiovascular safety gastroprotective, proinflammatory prostaglandin, PGE2, and the gastrotoxic, proinflammatory leukotriene LTB4 [27]. Licofelone As discussed, selective COX-2 inhibitors, in addition to having 100 mg/kg significantly decreased PGE2 levels to a similar extent potential GI and renal side-effects, also have potential cardiovas- as the non-selective NSAIDs, diclofenac and indomethacin cular side-effects. The activity of COX-2 in the endothelium results (P 0.001). However, increased levels of LTB4 in the non-selective in the production of cardioprotective PGI2, which inhibits platelet NSAID groups indicated a shunt of AA metabolism to the 5-LOX aggregation, leucocyte activation and adhesion, and accumulation pathway. Licofelone prevented the shunt by inhibiting 5-LOX, and of cholesterol in vascular cells [4]. These effects are antagonized, prevented an increase in LTB4 levels [27]. In this model, LTB4 however, by the activity of COX-1, which is expressed in platelets. levels in the licofelone group were equivalent to those in the control COX-1 produces TXA2, which causes vasoconstriction and platelet group (2.5 Æ 0.4 pg/mg protein), but in the diclofenac and aggregation [1]. It is therefore possible that the selective inhibition indomethacin groups LTB4 levels were elevated (9.2 Æ 2.3 and of COX-2 might lead to an imbalance in the levels of PGI2 and 8.8 Æ 1.6 pg/mg protein, respectively; P < 0.001). TXA2, and possibly to an increase in thromboembolic events [1]. Safety of anti-inflammatory treatment i19

goal in the treatment of OA is therefore the development of agents that specifically target the disease process and reduce cartilage damage. As described earlier, LTs and PGs are involved in the pathogenesis of OA, therefore inhibiting the production of these molecules could potentially slow disease progression. The effect of licofelone on cartilage lesions was examined in an OA dog model [40]. The size and grade of the cartilage lesions at the macroscopic level were reduced in the two groups of dogs treated with licofelone (2.5 and 5.0 mg/kg/day), compared with the control group. Indeed, the severity of cartilage lesions at the histological level was also decreased. Of particular importance were the observations that PGE2 levels in the synovial fluid and Downloaded from https://academic.oup.com/rheumatology/article/43/suppl_1/i16/1788145 by guest on 27 September 2021 LTB4 levels in the synovial membrane were reduced. Furthermore, there was reduced production of the important proinflammatory cytokine IL-1 and the cartilage-degrading enzyme collagenase-1 in the synovial membrane and cartilage, respectively. This indicates that licofelone may prevent OA progression. While the results of this study are promising, trials are needed to determine whether licofelone can arrest disease activity in humans with OA.

Conclusions FIG. 3. Adaptation to aspirin when administered alone or in LTs and PGs are important pathogenic mediators that contribute combination with COX inhibitors or licofelone. Reproduced to the pain and inflammation of OA. Current therapeutic from an article in the FASEB Journal [39] with permission. interventions rely upon the inhibition of COX enzymes to reduce the levels of PGs, and therefore pain and inflammation. However, this mode of action can result in the shunting of AA metabolism to produce LTs via 5-LOX, increasing the potential for GI damage. The observed gastrotoxicity of NSAIDs is a limiting factor in their use in treating the symptoms of OA. Selective COX-2 inhibitors have an improved GI tolerability profile compared with NSAIDs; however, their interference with the adaptation to aspirin (which is commonly used in OA patients) and their potential renal and cardiovascular toxicity can limit their prolonged use. It has been proposed that a drug capable of inhibiting 5-LOX, COX-1 and COX-2 would possess analgesic and anti-inflammatory properties while having a more favourable toxicity profile compared with current treatments. This proposal has led to the design of licofelone, a competitive inhibitor of 5-LOX, COX-1 and COX-2. Licofelone has been shown in experimental models to have excellent GI tolerability. Furthermore, the ability of licofelone to inhibit both LT and PG synthesis has led to the suggestion that this drug could offer cardiovascular tolerability advantages over selective COX-2 inhibitors. Of additional interest are the preliminary data from animal models indicating that licofelone can prevent OA disease progression, a factor not addressed by current therapeutic options for OA.

This supplement was supported by an unrestricted grant from Merckle GmbH.

FIG. 4. Adaptation to aspirin when administered alone or in Acknowledgements combination with a single dose of a COX inhibitor or licofelone. Adapted from [39]. The author thanks Thomson Gardiner-Caldwell London for its editorial support in the preparation of this article. In theory, these thromboembolic events are unlikely to occur with licofelone because inhibition of COX-1 and COX-2, as well as 5-LOX, would theoretically prevent an imbalance between TXA2 Key messages and PGI2 [1]. Furthermore, licofelone is not known to affect the levels of endocannabinoids, which have vasodilatory properties Current OA therapies reduce pain and [12]. Studies are required to further investigate the potential for inflammation by inhibiting COX en- cardiovascular side-effects with selective COX-2 inhibitors and the zymes, but can cause GI damage. cardiovascular safety of licofelone. Licofelone is a competitive inhibitor of 5-LOX, COX-1 and COX-2, developed for the treatment of OA. Licofelone and progression of osteoarthritis Rheumatology LOX/COX inhibition reduces pain and inflammation and is associated with low The drugs currently used to treat OA reduce inflammation and gastrotoxicity. relieve pain, but do not act directly on the disease process. A major i20 K. Brune

References 19. Hudson N, Balsitis M, Everitt S, Hawkey CJ. Enhanced gastric mucosal leukotriene B4 synthesis in patients taking non-steroidal anti- 1. Martel-Pelletier J, Lajeunesse D, Reboul P, Pelletier JP. Therapeutic inflammatory drugs. Gut 1993;34:742–7. role of dual inhibitors of 5-LOX and COX, selective and non-selective 20. Whelton A. COX-2-specific inhibitors and the kidney: effect on non-steroidal anti-inflammatory drugs. Ann Rheum Dis 2003;62: hypertension and oedema. J Hypertens 2002;20(Suppl. 6):S31–5. 501–9. 21. Whelton A, White WB, Bello AE, Puma JA, Fort JG. Effects of 2. Parente L. Pros and cons of selective inhibition of cyclooxygenase-2 celecoxib and rofecoxib on blood pressure and edema in patients 65 versus dual lipoxygenase/cyclooxygenase inhibition: is two better than years of age with systemic hypertension and osteoarthritis. Am J one? J Rheumatol 2001;28:2375–82. Cardiol 2002;90:959–63. 3. Bertolini A, Ottani A, Sandrini M. Selective COX-2 inhibitors and 22. Simon LS, Smolen JS, Abramson SB et al. Controversies in COX-2 dual acting anti-inflammatory drugs: critical remarks. Curr Pharm selective inhibition. J Rheumatol 2002;29:1501–10.

Design 2002;9:1–11. 23. Schwartz JI, Vandormael K, Malice MP et al. Comparison of Downloaded from https://academic.oup.com/rheumatology/article/43/suppl_1/i16/1788145 by guest on 27 September 2021 4. Hinz B, Brune K. Cyclooxygenase-2–10 years later. J Pharmacol Exp rofecoxib, celecoxib, and naproxen on renal function in elderly sub- Ther 2002;300:367–75. jects receiving a normal-salt diet. Clin Pharmacol Ther 2002;72:50–61. 5. Raisz LG. Prostaglandins and bone: physiology and pathophysiology. 24. Komers R, Anderson S, Epstein M. Renal and cardiovascular effects Osteoarthritis Cartilage 1999;7:419–21. of selective cyclooxygenase-2 inhibitors. Am J Kidney Dis 6. Skelly MM, Hawkey CJ. COX-LOX inhibition: current evidence for 2001;38:1145–57. an emerging new therapy. Int J Clin Pract 2003;57:301–4. 25. Cheng HF, Harris RC. Does cyclooxygenase-2 affect blood pressure? 7. Sahap Atik O. Leukotriene B4 and prostaglandin E2-like activity in Curr Hypertens Rep 2003;5:87–92. synovial fluid in osteoarthritis. Prostaglandins Leukot Essent Fatty 26. Laufer S. Discovery and development of ML3000. Inflammopharma- Acids 1990;39:253–4. cology 2001;9:101–12. 8. Bertin P, Lapicque F, Payan E et al. Sodium naproxen: concentration 27. Tries S, Albrecht W. Mechanism of action of licofelone (ML3000), a and effect on inflammatory response mediators in human rheumatoid novel anti-inflammatory compound without the potential for causing synovial fluid. Eur J Clin Pharmacol 1994;46:3–7. gastric damage. Inflamm Res 2002;51:135. 9. He W, Pelletier JP, Martel-Pelletier J, Laufer S, Di Battista JA. 28. Asako H, Kubes P, Wallace J, Gaginella T, Wolf RE, Granger DN. Synthesis of interleukin 1beta, tumor necrosis factor-alpha, and Indomethacin-induced leukocyte adhesion in mesenteric venules: role interstitial collagenase (MMP-1) is eicosanoid dependent in human of lipoxygenase products. Am J Physiol 1992;262:G903–8. osteoarthritis synovial membrane explants: interactions with antiin- 29. Asako H, Kubes P, Wallace J, Wolf RE, Granger DN. Modulation of flammatory cytokines. J Rheumatol 2002;29:546–53. leukocyte adhesion in rat mesenteric venules by aspirin and salicylate. 10. Paredes Y, Massicotte F, Pelletier JP, Martel-Pelletier J, Laufer S, Gastroenterology 1992;103:146–52. Lajeunesse D. Study of the role of leukotriene B4 in abnormal 30. Wallace JL, Carter L, McKnight W, Tries S, Laufer S. ML 3000 function of human subchondral osteoarthritis osteoblasts. Arthritis reduces gastric prostaglandin synthesis without causing mucosal Rheum 2002;46:1804–12. injury. Eur J Pharmacol 1994;271:525–31. 11. Guhring H, Hamza M, Sergejeva M et al. A role for endocannabi- 31. Gabriel SE, Crowson CS, O’Fallon WM. Comorbidity in arthritis. noids in indomethacin-induced spinal antinociception. Eur J J Rheumatol 1999;26:2475–9. Pharmacol 2002;454:153–63. 32. Marks R, Allegrante JP. Comorbid disease profiles of adults with end- 12. Maccarrone M, Finazzi-Agro A. Endocannabinoids and their actions. stage hip osteoarthritis. Med Sci Monit 2002;8:CR305–9. Vitam Horm 2002;65:225–55. 33. Patrono C. Aspirin as an antiplatelet drug. N Engl J Med 13. Silverstein FE, Faich G, Goldstein JL et al. Gastrointestinal toxicity 1994;330:1287–94. with celecoxib vs nonsteroidal anti-inflammatory drugs for osteo- 34. Claria J, Serhan CN. Aspirin triggers previously undescribed bioactive arthritis and rheumatoid arthritis: the CLASS study: A randomized eicosanoids by human endothelial cell-leukocyte interactions. Proc controlled trial. Celecoxib Long-term Arthritis Safety Study. JAMA Natl Acad Sci USA 1995;92:9475–9. 2000;284:1247–55. 35. Lecomte M, Laneuville O, Ji C, DeWitt DL, Smith WL. Acetylation 14. Goldstein JL, Correa P, Zhao WW et al. Reduced incidence of of human prostaglandin endoperoxide synthase-2 (cyclooxygenase-2) gastroduodenal ulcers with celecoxib, a novel cyclooxygenase-2 by aspirin. J Biol Chem 1994;269:13207–15. inhibitor, compared to naproxen in patients with arthritis. Am J 36. Cryer B. Gastrointestinal safety of low-dose aspirin. Am J Manage Gastroenterol 2001;96:1019–27. Care 2002;8:S701–8. 15. Juni P, Rutjes AW, Dieppe PA. Are selective COX 2 inhibitors 37. Wallace JL. Nonsteroidal anti-inflammatory drugs and gastroenter- superior to traditional non steroidal anti-inflammatory drugs? Br opathy: the second hundred years. Gastroenterology 1997;112: Med J 2002;324:1287–8. 1000–16. 16. Bombardier C, Laine L, Reicin A et al. Comparison of upper 38. Fiorucci S, de Lima OM Jr, Mencarelli A et al. Cyclooxygenase-2-

gastrointestinal toxicity of rofecoxib and naproxen in patients with derived lipoxin A4 increases gastric resistance to aspirin-induced rheumatoid arthritis. VIGOR Study Group. N Engl J Med damage. Gastroenterology 2002;123:1598–606. 2000;343:1520–28. 39. Fiorucci S, Distrutti E, de Lima OM et al. Relative contribution of 17. Hernandez-Diaz S, Rodriguez LA. Association between nonsteroidal acetylated cyclo-oxygenase (COX)-2 and 5-lipooxygenase (LOX) in anti-inﬂammatory drugs and upper gastrointestinal tract bleeding/ regulating gastric mucosal integrity and adaptation to aspirin. FASEB perforation: an overview of epidemiologic studies published in the J 2003;17:1171–3. 1990s. Arch Intern Med 2000;160:2093–9. 40. Jovanovic DV, Fernandes JC, Martel-Pelletier J et al. In vivo dual 18. Ma L, del Soldato P, Wallace JL. Divergent effects of new inhibition of cyclooxygenase and lipoxygenase by ML-3000 reduces cyclooxygenase inhibitors on gastric ulcer healing: Shifting the the progression of experimental osteoarthritis. Arthritis Rheum angiogenic balance. Proc Natl Acad Sci USA 2002;99:13243–7. 2001;44:2320–30.