Introduction

Top Abstract Introduction Methods Results Discussion References

NSAIDs alleviate the hyperalgesic symptoms associated with inflammation by inhibiting the COX enzyme and the resultant inhibition of prostaglandin synthesis from arachidonic acid (Vane, 1971). However, these drugs have been associated with GI side effects (Whittle et al., 1980, De Witt and Smith, 1988, Fries et al., 1989) thought to result from inhibition of constitutive cyclooxygenase products in the mucosa of the GI tract (Main and Whittle, 1975). Prostaglandins are believed to play a role in maintaining the integrity of the mucosal lining (Robert et al., 1967, Dajani et al., 1975, Kauffman and Grossman, 1978, Boughton-Smith and Whittle, 1981). Recently, a second isoform of COX-2 has been discovered (Xie et al., 1991, Kujubu et al., 1991, O'Banion et al., 1991, Hla and Neilson, 1992, Xie et al., 1992) and shown to be induced during an inflammatory event (Maier et al., 1990, O'Banion et al., 1992). The constitutive isozyme (COX-1) is also expressed during an inflammatory reaction; however, not to the same degree as the induced isozyme (COX-2) (Sano et al., 1992, Mitchell et al., 1993, Meade et al., 1993, Crofford et al., 1994). Because COX-2 is induced during inflammation and COX-1 is thought to be responsible for GI protection, it is conceivable that a selective inhibitor of COX-2 may provide antiinflammatory effects without causing deleterious GI side effects (Masferrer et al., 1994); however, this concept has not yet been clinically proven.

Although nonselective or COX-1 selective NSAIDs are therapeutically effective, most of these drugs have other pharmacological activities (Walker et al., 1976, Siegel et al., 1979, Abramson and Weissman, 1989) which may contribute to their antiinflammatory efficacy. An agent that selectively inhibits the inducible COX-2 isozyme may alleviate the symptoms of inflammation without causing the gastrointestinal side effects, although it may not prevent the progression of the inflammatory disorder. The clinical antiinflammatory efficacy of specific selective COX-2 inhibitors has yet to be determined. This study evaluated several of the newer in-vitro COX-2 selective inhibitors (SC-58125, nimesulide and nabumetone) in a canine model of local inflammation induced by carrageenan into a subcutaneous chamber. The oral effects of these selective COX-2 inhibitors were compared to those of standard NSAIDs (aspirin, indomethacin and tenidap) that inhibit both isoforms of cyclooxygenase. In addition, several compounds which do inhibit chemotactic responses were evaluated in this system, including a glucocorticoid (dexamethasone), a 5-LO inhibitor (zileuton) and a new dual COX-2 selective/5-LO inhibitor (RWJ 63556). Inhibition of leukocyte influx into the chamber was used as an indicator of antiinflammatory activity. Selective inhibition of prostanoid production in the inflammatory exudate and peripheral blood was used as an indicator of COX-1/COX-2 enzyme selectivity, because COX-2 should be induced during an inflammatory response and prostanoid production in the blood should be indicative of the constitutive COX-1 response. Additionally, 5-LO activity was evaluated by quantitating LTB₄ production in the exudate and the blood.

	Methods

Top Abstract Introduction Methods Results Discussion References

In Vitro COX-1, COX-2, 5-LO Assays

Specific COX-1 and 5-LO activity was assayed using methods previously described (Argentieri et al., 1994) Briefly, rat basophilic leukemia (RBL-1) cells (ATCC 1378-CRL, American Type Cell Culture, Rockville, MD) were used to evaluate COX-1 and 5-LO activity. A cell-free homogenate was prepared by subjecting the cells to a polytron followed by centrifugation at 9000 × g. The CaCl₂ -dependent production of cyclooxygenase and 5-LO products from ¹⁴C-arachidonic acid (NEN, Boston, MA) in the 9000 × g supernatant was monitored in the presence of vehicle or drug. Products were isolated by acidification and extraction, followed by thin layer chromatography. Radioactive areas corresponding to cyclooxygenase and 5-lipoxygenase products, PGD₂ and 5-HETE, respectively, were quantitated using a BIOSCAN imaging scanner 200 (BIOSCAN, Washington, DC). The data were expressed as percent inhibition of products as compared to vehicle treatment and IC₅₀ values were calculated from curve fit equations derived using the linear least-squares regression method (Cricket Graph III, Computer Associates International, Islandia, NY).

Evaluation of specific COX-2 activity of a compound was performed using a whole cell assay with ECV-304 (human, endothelial, umbilical cord) cells (ATCC # 1998-CRL, American Type Cell Culture, Rockville, MD). These cells were cultured in Media 199/10% bovine serum albumin (BioWhitaker, Walkersville, MD) at 37°C and 5% CO₂, then trypsinized and plated at a density of 9 × 10⁴ cells per well of a 96 well plate before assay. Approximately 28 hr later, 50 µg/ml PMA (Sigma, St. Louis, MO) and 2 µM ionomycin (Sigma, St. Louis, MO) (final concentrations) were added to each well. Cells were incubated in the presence of vehicle or drug for 18 hr. PGE₂ production was determined via RIA (PerSeptive Biosystems, Framingham, MA) after the addition of 30 µM arachidonic acid. The data were expressed as percent inhibition of products as compared to vehicle treatment and IC₅₀ values were calculated from curve fit equations derived using the linear least-squares regression method (Cricket Graph III, Computer Associates International, Islandia, NY).

Carrageenan-Induced Inflammation in Dogs

Surgical implantation of subcutaneous chambers. Seventy-two beagle dogs of either sex, approximately 1 yr of age and weighing 8 to 13 kg were used in this study. Purpose-bred dogs were purchased from Marshall Farms (North Rose, NY) and housed in an American Association for Accreditation of Laboratory Animal Care (AAALAC) accredited facility. All procedures conformed to the requirements of the Animal Welfare Act and were conducted according to the United States Department of Health and Human Services Guide for the Care and Use of Laboratory Animals. The study was approved by the Institutional Animal Care and Use Committee. The dogs were housed singly in 12-sq.ft. stainless steel cages and were fed Purina (St. Louis, MO) High Density Canine Diet once daily. Water was provided ad libitum. The rooms were maintained at 64 to 84°F and 30 to 70% humidity with a 12-hr diurnal cycle.

Carrageenan-induced inflammation. After a postoperative recovery period of at least 1 mo, a 1-ml sample of exudate from each dog was aspirated from within the chamber using a 1-ml syringe and 20-gauge 1-inch needle inserted through one of the perforations in the ball. The area was cleansed with 70% ethanol before each sample was collected. Immediately after obtaining the sample, 1.5 ml of 0.33% sterile carrageenan lambda (Sigma Chemical Co., St. Louis, MO) was injected into the chamber. Samples of the exudates were obtained again at 5 and 24 hr post carrageenan challenge and are indicative of the inflammatory response. Dilutions of the samples were made in saline and analyzed for leukocyte count using a Coulter Sample Stand and Multisizer II (Coulter Corp, Miami, FL). Samples of exudate (200 µl) were also extracted in ethanol (1 ml) for measurement of eicosanoid content. The precipitate was centrifuged at 4°C (6 min, 12,000 rpm) and the supernatant was stored at -20°C until analyzed for the presence of PGE₂, TxB₂ and LTB₄ by enzyme-linked immunosorbant assays (ELISA Technologies, Lexington, KY). The exudate within the chambers was evacuated after the last sample was obtained and the dogs were allowed a recovery period of at least 4 wk between carrageenan challenges. The dogs were used repeatedly for up to 1 yr and they served as their own controls when measuring treatment effects on the inflammatory response elicited by carrageenan. Compounds were administered orally in gelatin capsules immediately after carrageenan challenge when various treatments were evaluated.

FACScan determination of differential cell populations in fluid exudate. Aliquots (0.5 ml) of chamber exudate were treated with the Whole Blood Erythrocyte Lysing Kit (R&D Systems, Minneapolis, MN) according to the manufacturer's protocol. Briefly, 2 ml lysing buffer were added to each fluid sample. After vigorous mixing, the samples were incubated at room temperature for 10 min and then centrifuged at 300 × g for 5 min. The supernatants were removed and the cells were washed with 2 ml wash buffer, followed by centrifugation at 300 × g for 5 min. The pellets were resuspended in 1 ml of wash buffer with 100 ml of 10X fixative. The chamber cells were then analyzed on a FACScan flow cytometer (Becton Dickinson, San Jose, CA). Differential cell percentages were determined using the LYSYS II program (Becton Dickinson, San Jose, CA).

Western blot analysis of COX-2 protein produced by granulocytes during an inflammatory response. Chamber exudate was centrifuged at 200 × g for 10 min to pellet cells. The cells were washed with PBS and centrifuged again. The supernatants were discarded and the pellets were frozen at -20°C until further processing. Upon thawing, cell pellets were resuspended in 0.5 ml lysis buffer (50 mM Tris, 150 mM NaCl, 0.1% Triton X-100, pH 7.5 with 1 mM phenylmethylsulfonyl fluoride) and incubated on ice for 30 min. Samples were centrifuged for 15 min at 10,000 × g at 4°C. The supernatant was then incubated with 6 µl rabbit antihuman COX-2 polyclonal antibody (Cayman, Ann Arbor, MI) with mixing for 1 hr at 4°C. Protein-A Sepharose was then added to the sample and again the sample was mixed for 1 hr. Samples were centrifuged for 5 sec and the supernatant was discarded. The pellets were then washed two times with 1 ml lysis buffer containing 0.1% Triton X-100 and 1 time with lysis buffer without Triton X-100, discarding the supernatant after each wash. The Sepharose pellet was resuspended in 30 µl PAGE loading buffer, boiled for 5 min and run on a 10% polyacrylamide SDS gel under reducing conditions. The proteins were blotted onto Immobilon PVDF membrane (Millipore, Bedford, MA) and probed with COX-2 antibody (Cayman, Ann Arbor, MI) at a 1:1000 dilution. Proteins were visualized by ECL Western blotting (Amersham, Arlington Heights, IL) according to manufacturer's procedure.

	Results

Top Abstract Introduction Methods Results Discussion References

Specific enzyme effects of compounds (table 1) were evaluated in-vitro on COX-1, COX-2 and 5-LO as described in "Methods." Aspirin, indomethacin and tenidap demonstrated preferential inhibition of COX-1, which is consistent with the activity of standard NSAIDs. In the specific enzyme systems, nimesulide and SC-58125 are COX-2 selective inhibitors (Magni, 1993, Bevilacqua and Magni, 1993, Davis and Brogden, 1994, Isakson et al., 1994, Seibert et al., 1994). None of these cycooxygenase inhibitors exhibited 5-LO inhibition. Zileuton was the only 5-LO selective compound tested, although RWJ-63556 (N-[5-(4-fluorophenoxy)thien-2-yl] methanesulfonamide), demonstrated marked selectivity for COX-2 (IC50 COX-2 = 1.86 µM, IC50 COX-1 > 10 µM) and additionally potently inhibited 5-LO (IC50 = 0.13 µM).

These compounds were evaluated for their oral activity on carrageenan-induced inflammation in s.c. chambers in dogs. Carrageenan administration produces a significant influx of leukocytes into the exudate which peaks by 24 hr (fig. 1) and is maintained over a period of 3 or more days. Leukocyte counts usually begin to return to initial baseline values by day 6 (data not shown). The increase in leukocyte count is reproducible, as evidenced by repeated control responses measured in 31 dogs. Control responses were evaluated twice in each dog with 2 to 8 mo and one to four treatment responses administered between each control experiment (fig. 2). Flow cytometry analysis of the samples revealed a shift in the fluid cell populations from primarily lymphocytes at time 0 to predominantly granulocytes by 24 hr postchallenge (fig. 3). Fluid smears, examined microscopically, showed that the granulocytes that infiltrate the chamber were primarily neutrophils but also included a low percentage of macrophages and mast cells. The patency of the implanted chambers was very good, as only 3 of the 72 animals required the removal of the chamber due to infection or severe responses to the administration of carrageenan.

View larger version (14K): [in this window] [in a new window]   Fig. 1.   Leukocyte influx into subcutaneous chambers after carrageenan challenge. Values represent the mean ± SEM (n = 72). *Significant changes from baseline levels, P < .0001 using one-tailed one-sided t test.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Reproducibility of the control response to carrageenan. Values represent the mean ± S.E.M. (n = 31). Control responses were evaluated twice with 2 to 8 mo and one to four treatment responses administered between each control experiment. *Significant from corresponding point, P < .05 using Student's t test.

View larger version (35K): [in this window] [in a new window]   Fig. 3.   FACScan analysis of leukocyte populations in chamber fluid samples removed before and 24-hr postcarrageenan challenge in an individual dog. Sample size = 10,000 events.

Because cells involved in the inflammatory process release eicosanoid metabolites of arachidonic acid (Higgins and Lees, 1984, Jagels and Hugli, 1994), the inflammatory response was further characterized using ELISA to measure changes in the concentration of these products in the inflammatory exudate. The cyclooxygenase products, PGE₂ and TxB₂, increased along with granulocytic infiltration over the course of the study. LTB₄ concentration was increased at 5 hr, but fell below baseline levels by 24 hr (fig. 4), thus assessment of 5-LO inhibition could only be made at 5 hr postchallenge. The inflammatory response should be accompanied by production of COX-2 protein. Western blot analysis using antihuman COX-2 antibody and lysed exudate cells from the dogs was performed to determine if the cyclooxygenase products released during the reaction were due to the upregulation of the inducible COX-2 gene. COX-2 protein was expressed over the 24-hr experimental period. This protein was not evident in 0-hr samples or in samples where the dogs were treated with dexamethasone (3 mg/kg, P.O.) (fig. 5). Studies have shown that dexamethasone inhibits the expression of COX-2 induced by a variety of stimuli (Masferrer et al., 1990, Masferrer et al., 1992).

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Changes in eicosanoid concentration in exudate during the inflammatory response. Each bar represents the mean change from baseline control ± S.E.M. (n = 72). *Significant (P < .05), **Significant (P < .0001) from baseline using one-tailed one-sample t test.

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Western blot analysis using antihuman COX-2 antibody with lysed cells collected in exudates from a single control dog and from another treated with dexamethasone (3 mg/kg, p.o.).

The oral activity and selectivity of standard antiinflammatory agents were evaluated in this model. The compounds were administered in gelatin capsules at the time of carrageenan challenge (0 hr). The effects of indomethacin (3 mg/kg, p.o.), which inhibits both COX-1 and COX-2 in vitro, are illustrated in figure 6. Granulocyte infiltration and ex vivo PGE₂ production in whole blood were inhibited more than 60%, although the PGE₂ concentration in the exudate was inhibited more than 90% throughout the course of the study, demonstrating indomethacin's lack of selectivity for COX isozymes. In comparison, nimesulide (10 mg/kg, p.o.), a compound more specific for COX-2 in vitro (Magni, 1993, Bevilacqua and Magni, 1993, Davis and Brogden, 1994), attenuated the early increase in cell infiltration although having no inhibitory activity at 24 hr, but continued to inhibit PGE₂ production in the exudate. The lack of activity of nimesulide on COX-1 is evident by its lack of inhibitory activity on the ex-vivo production of PGE₂ in blood (fig. 6). Dose/response curves were generated to obtain ED₅₀s from linear regression. The activities of a number of compounds were compared on each parameter evaluated (table 2). In this model, all of the more COX-1 selective inhibitors tested demonstrated ex-vivo activity on prostaglandin production in the blood. Although indomethacin and tenidap were effective inhibitors of cell infiltration and PGE₂ production in the exudate, aspirin inhibited only the prostaglandin levels but had no effect on cell migration. The corticosteroid, dexamethasone, effectively inhibited the inflammatory components measured in this model without altering eicosanoid production ex vivo. Zileuton, a 5-LO inhibitor, also inhibited the influx of granulocytes and PGE₂ present in the fluid at 24 hr. The selective COX-2 inhibitors, nimesulide and SC-58125, reduced both cell influx and PGE₂ production throughout the 24-hr period with no effect on ex vivo eicosanoid production. The metabolite of nabumetone has been reported to be a COX-2 inhibitor (De Witt et al., 1993, Friedel et al., 1993), thus nabumetone acts in a similar fashion to nimesulide and SC-58125 in vivo. The dual COX-2 selective/5-LO inhibitor, RWJ 63556, is a potent inhibitor of leukocyte influx over a period of 24 hr and maintains its in vitro profile on eicosanoids.

View larger version (23K): [in this window] [in a new window]   Fig. 6.   Mean percent inhibition of control responses to carrageenan by indomethacin (3 mg/kg, p.o., n = 7) or nimesulide (10 mg/kg, p.o., n = 3). The granulocyte influx is indicative of an inflammatory reaction, although elevated fluid PGE2 and ex vivo PGE2 production are indicative of COX-2 and COX-1 responses, respectively. Each dog served as its own control.

	Discussion

Top Abstract Introduction Methods Results Discussion References

The method developed for this study was adapted from other models of inflammation in rabbits, dogs and horses (Higgins et al., 1987, Lees et al., 1987, Clemons et al., 1992, Zech et al., 1993, McKellar et al., 1994). Many characteristics of the inflammatory process can be evaluated in this model, which imposes minimal distress on the animals and is reproducible over time. Carrageenan is reported to initiate inflammation locally by activation of the complement pathway and by a selective cytopathic effect on macrophages which become activated to release degradative lysosomal enzymes. The response elicited in this model is characterized by neutrophil infiltration into the exudate over 24 hr. A variety of vasoactive and chemotactic inflammatory mediators (prostaglandins, leukotrienes) are also involved in the carrageenan-induced response (Weissman et al., 1980, Higgins and Lees, 1984, Hansen et al. 1992, Jagels and Hugli, 1994, Leirisalo-Repo, 1994,, Siminiak et al., 1995, Fujishima and Aikawa, 1995). Western blot analysis of cells obtained from exudates indicated that COX-2 expression is up-regulated during the inflammatory response in this model and is inhibited by dexamethasone treatment. This analysis suggests that there is some degree of cross-reactivity of the human COX-2 antibody between human and canine proteins. Therefore, in this model, the elevated prostanoid levels observed in the exudate during the inflammatory response are used as an indicator of an inducible COX-2 response, although the ex vivo production of prostanoids in the blood is considered a constitutive COX-1 mediated event. Enhanced LTB₄ concentration was only evident in the 5-hr samples in this model although cell numbers continued to increase over 24 hr. It may be possible for leukocyte infiltration to occur during 5-LO inhibition because carrageenan exerts its effects by activating the complement pathway, thus starting a cascade of events that produce several chemotactic factors such as C5a, chemokines and LTB₄. However, inhibition of chemotaxis due to 5-LO products during an inflammatory response does appear an effective approach to inhibit inflammation, as evidenced by the effects of zileuton. The decrease in LTB₄ levels observed at 24 hr postchallenge may be attributed to the high numbers of neutrophils that invade the chamber because these cells have been reported to catabolize LTB₄ into less biologically active products (Shak and Goldstein, 1985).

Standard NSAIDs inhibited cell influx in this model at clinically effective doses, as demonstrated by the inhibitory activity of indomethacin and tenidap. These compounds are COX-1 inhibitors as evidenced by their inhibition of ex vivo prostanoid production. Newer COX-2 selective compounds also inhibited leukocyte influx and demonstrate COX-2 selectivity by preferentially inhibiting prostanoid (PGE₂ and TxB₂) production in the exudate during the inflammatory response versus the ex-vivo production in peripheral blood. Dexamethasone is a potent antiinflammatory agent in this model without demonstrating COX-1 activity in the ex-vivo blood assay. This is consistent with the proposed COX-2 activity of glucocorticoids (Masferrer et al., 1990, 1992).

Although this system appears to be capable of detecting antiinflammatory effects on leukocyte infiltration and COX-1/COX-2 selectivity of cyclooxygenase inhibitors, it also indicates that 5-lipoxygenase products may play an important role in the inflammatory response. Zileuton, a selective 5-LO inhibitor, is a potent antiinflammatory agent in this system and reduces PGE₂ concentration in the exudate at 24 hr. Because LTB₄ is a potent chemotactic factor derived from the 5-LO pathway, this effect on PGE₂ production might be explained by the decreased number of cells present to produce mediators or that zileuton may exert its effects through additional mechanisms. The lack of an antiinflammatory effect on leukocyte influx by aspirin, although inhibiting PGE₂ production (table 2) is not surprising because COX inhibitors do not exert any effects on chemotactic factors such as C5a, IL-8 or LTB₄. This may also support the suggestion that NSAIDs may not exert their antiinflammatory effects by cyclooxygenase inhibition alone and that cyclooxygenase inhibition does not always result in attenuation of the inflammatory process (Walker et al., 1976, Siegel et al., 1979, Abramson and Weissman, 1989). Selective, although not necessarily specific, COX-2 inhibitors such as SC-58125 and nimesulide produced an antiinflammatory effect on granulocyte influx at 5 hr (ED₅₀s = 0.4 and 15.8 mg/kg, P.O., respectively), although demonstrating little activity at 24 hr. However, in both cases the inhibition of the COX-2 regulated PGE₂ production in exudate was more pronounced at 24 hr when the inhibition of cell influx was minimal. Neither compound was a potent inhibitor of PGE₂ production in blood. However, both compounds produced a decrease in LTB₄ production in exudate at 5 hr, corresponding to their antiinflammatory activity. These data suggest a possible association between inhibition of leukotriene production and the inhibition of cell influx and that selective inhibitors of COX-2 do not appear to exert their effect through COX inhibition alone, because the inhibition of COX-2 does not appear to correlate with the antiinflammatory activity in this model. Indeed, many antiinflammatory agents exert their activity through more than one mechanism, such as cytokine modulation and antihistaminic activity (Griswold, et al., 1993, Rossoni et al., 1993) and it may be that COX-2 inhibition alone will not inhibit an inflammatory response.

In this model, RWJ 63556 is a potent, orally active COX-2 selective/5-LO inhibitor and maintains its antiinflammatory and LTB₄ inhibitory effects throughout the entire 24-hr experimental duration. This compound is an aryl methane sulfanamide which is structurally related to nimesulide, a known COX-2 selective inhibitor. It is possible that it may derive 5-LO activity by forming a 6 membered iron chelating ring among the sulphur of the thiaphene, one of the oxygen atoms on the sulfonamide moiety and the metal ion on the active site of 5-LO. It may be advantageous for a compound to have both activities, because prostaglandins have been implicated in enhancement of LTB₄ mediated inflammation (Ferreira and Vane, 1979, Raud et al., 1989, Hedqvist et al., 1990) and the data obtained in this study suggests that effective antiinflammatory compounds elicit their effects on cell influx during times of reduced LTB₄ concentration in the exudate. Inhibition of PGE₂ alone in the exudate does not appear to correlate with inhibition of cell influx. This correlation appears to hold for zileuton, nabumetone, RWJ 63556 and dexamethasone. Only tenidap appears to deviate from these observations. Tenidap inhibits leukocyte infiltration (ED₅₀ < 3 mg/kg, P.O.) although inhibiting PGE₂ production in both the exudate and the blood, demonstrating little COX-2 selectivity, with essentially no inhibition of LTB₄ production. Perhaps the reported cytokine modulating activities of tenidap (Otterness et al. 1991, Breedveld, 1994, Conti et al., 1994, Wylie et al., 1995, Madhok, 1995) account for its antiinflammatory activity by inhibiting the activities of chemotactic cytokines. The other classes of agents evaluated in this model (glucocorticoids, COX-1/COX-2 inhibitors, 5-LO inhibitors) maintain their antiinflammatory activity throughout the experimental period when accompanied by decreased LTB₄ levels in the exudate.

RWJ 63556, an orally active dual COX-2 selective/5-LO inhibitor, produces significant antiinflammatory activity in this dog model of local inflammation and may represent a new class of antiinflammatory compounds which are needed for the effective treatment of inflammatory diseases, such as rheumatoid arthritis, without causing deleterious GI side effects. A dual COX-2 selective/5-LO inhibitor may provide symptomatic relief of hyperalgesia and inhibit the cell infiltration that leads to tissue damage at sites of inflammation. Additionally, a new in vivo model of local inflammation is described which yields information about many aspects of the response although minimizing distress and the number of animals required to evaluate new antiinflammatory therapies.