Abstract
The relative anti-inflammatory activities of the immunomodulators HR325 and leflunomide, or its active metabolite A77 1726, were examined by determining potencies in vitro on prostaglandin endoperoxide H synthase (PGHS) and in vivo in rat air pouch inflammation. Nonsteroidal anti-inflammatory drugs (NSAIDs) were used as comparators. HR325 was more potent than A77 1726 as an inhibitor of PGHS in guinea pig polymorphonuclear leukocytes (IC50 = 415 and 4400 nM, respectively) and on isolated ovine PGHS-1 (IC50 = 64 and 742 μM) and PGHS-2 (IC50 = 100 and 2766 μM). In vivo, in rat carrageenan air pouch inflammation, HR325 but not leflunomide at 25 mg/kg inhibited accumulation of leukocytes (48%) and PGE2 (61%). HR325 was also more potent than A77 1726 against human peripheral blood mononuclear cell PGHS-1 [IC50 = 1.6 and 25.6 μM (thromboxane B2production) or 1.1 and 8 μM (PGE2 production)] and lipopolysaccharide-induced PGHS-2 in human adherent peripheral blood mononuclear cells (IC50 = 435 nM and 9.5 μM) and peripheral blood polymorphonuclear leukocytes (IC50 = 91 nM and 3.2 μM). HR325 had low PGHS-2 selectivity in the human (2.5–12-fold) and was a more potent PGHS-2 inhibitor than naproxen, ibuprofen and piroxicam (2–8-fold). Assays using endogenous arachidonic acid as substrate yielded IC50 values for NSAIDs that were in general markedly lower than those published for assays using 10 μM substrate. With this approach, piroxicam had reasonable activity on human PGHS-2 (IC50 = 260–290 nM). Only NS398 and flufenamic acid were PGHS-2 selective in the human (90–330-fold and 37–60-fold, respectively); the other NSAIDs were either PGHS-1-selective (naproxen, ibuprofen, flurbiprofen and indomethacin) or nonselective (piroxicam and diclofenac). Inclusion of 10% human plasma reduced HR325 potency against PGHS-1 in human peripheral blood mononuclear cells approximately 32-fold (IC50 = 36 μM). Plasma protein binding further reduced HR325 potency (IC50 = 164 μM) and minimized the difference between HR325 and A77 1726 (IC50 = 292 μM) in a whole blood PGHS assay. Whether the greater activity against human PGHS-2 would allow HR325 to exhibit NSAID-like therapeutic effects in humans remains unclear.
Leflunomide (N-(4-trifluoromethylphenyl)-5-methylisoxazol-4-carboxamide) is under phase III evaluation for treatment of rheumatoid arthritis and has shown clinical efficacy in patients with long-standing arthritis (Mladenovic et al., 1995). It is effective in animal models of autoimmune disease (review by Bartlett et al., 1991) and prolongs allograft survival (Schorlemmer et al., 1993;McCheseney et al., 1994). The primary leflunomide metabolite, A77 1726, inhibits cell proliferation (Bartlett et al., 1991) and is responsible for the immunosuppressive and disease-modifying effects of the drug (Bartlett et al., 1989). These effects are mediated at least in part by inhibition of dihydroorotate dehydrogenase, a mitochondrial enzyme involved inde novo pyrimidine biosynthesis (Williamson et al., 1995; Cherwinski et al., 1995).
HR325, evaluated in a phase II clinical study for rheumatoid arthritis, is a close structural analogue of A77 1726, and an equipotent inhibitor of dihydroorotate dehydrogenase (Williamson et al., 1995;Kuo et al., 1996). In addition to immunosuppressive and antiproliferative effects, leflunomide and HR325 demonstrate anti-inflammatory activity (Bartlett et al., 1993; Hambletonet al., 1992, respectively), and the current study was designed to assess the relative potencies of the compounds as anti-inflammatory agents.
The anti-inflammatory properties of NSAIDs are mediated by inhibition of the cyclooxygenase activity of PGHS (Vane, 1971). This is a key regulatory enzyme in the cascade leading from arachidonic acid to the inflammatory mediators, prostaglandins and thromboxanes (reviewed bySmith, 1989). PGHS exists in constitutively expressed (PGHS-1) and inducible (PGHS-2) forms. Inhibition of the former, together with uncoupling of oxidative phosphorylation, is believed to be responsible for the GI side effects associated with NSAID therapy (Meade et al., 1993; Mitchell et al., 1994; Somasundaram et al., 1995), whereas PGHS-2 is the probable therapeutic target for the drugs (Futaki et al., 1993; 1994; Meade et al., 1993; Mitchell et al., 1994).
Here we report that HR325 is a more potent PGHS inhibitor than A77 1726, in vivo in the rat and in vitro in guinea pig PMNLs and on isolated ovine enzymes. In addition, the ability of the compounds to inhibit both constitutive and induced human enzymes in isolated peripheral blood cells and the constitutive enzyme in a whole blood assay are examined to determine whether HR325 could be expected to be a more potent anti-inflammatory agent in the human. In all cases, activities are compared with those of a range of NSAIDs.
Materials and Methods
Materials.
The following drugs and chemicals were used in this study: A23187, diclofenac, flufenamic acid, flurbiprofen, ibuprofen, indomethacin, piroxicam, dextran, percoll, RPMI 1640, fetal calf serum, nitro blue tetrazolium tablets, 5-bromo-4-chloro-3-indolyl phosphate, monoclonal anti-rabbit IgG alkaline phosphatase conjugate, prestained SDS-PAGE molecular weight markers, nonfat milk, Tween-20, glutamine, penicillin/streptomycin/amphotericin B solution, dexamethasone, hematin, arachidonic acid, LPS (from E. coliserotype 0111:B4), sodium caseinate, carrageenan λ, Tris, DMSO, heparin and calcium chloride (Sigma, Dorset, UK), aspirin and phenidone (Aldrich, Dorset, UK) and acacia oil and Tween 80 BDH (Lutterworth, U.K.), chloroform, ethyl acetate, hexane, methanol and phenol (Fisons, Leicestershire, UK), HPTLC plates (Merck, Leicestershire, UK), ram seminal vesicle PGHS-1 (19300 U/ml, ≥95% pure), sheep placental PGHS-2 (7161 U/ml, ∼70% pure) and rabbit polyclonal anti-human PGHS-2 peptide antiserum (Cayman Chemical Company, Nottingham, UK), rabbit polyclonal anti-PGHS-1 peptide antiserum (Oxford Biomedical Research Inc., Oxford, MI), [125I]PGE2 RIA kits (Du Pont NEN, Hertfordshire, UK) and [I-14C] arachidonic acid (2.2 GBq/mmol, 1.85 MBq/ml) (Amersham, Buckinghamshire, UK). HR 325, A77 1726, naproxen and tenidap were synthesised in the Chemistry Department of Hoechst Marion Roussel Limited, and leflunomide was prepared in the Chemistry Department of Hoechst AG, Werk Kalle-Albert, Wiesbaden, Germany.
Test compounds.
All test compounds were prepared over a range of concentrations (usually five or six in DMSO and were directly diluted into the assay (final DMSO concentration 0.1–1%). Concentrations generally increased in full-log or half-log unit intervals, with each concentration assayed at least in duplicate.
Inhibition of PGHS activity in guinea pig peritoneal PMNLs.
Casein (2%)-elicited peritoneal PMNLs were obtained from male guinea pigs (350–450 g, Dunkin Hartley strain, Interfauna) and assayed for PGHS activity largely as described (Harvey and Osborne, 1983). Briefly, the PMNLs were collected and washed in PBS/heparin (12.5 U/ml) and were resuspended at 0.7 to 1 × 107 cells/ml in assay buffer (15 mM Tris-HCl, 134 mM NaCl, 5 mM glucose, pH 7.4). The cells (920 μl) were incubated with test compound for 5 min at 37°C. [1-14C] arachidonic acid was added (3.4 kBq; 1.5 μM), followed 1 min later by 10 mM CaCl2 and 10 μM A23187 (final concentrations) in a total assay volume of 1 ml. Incubations were terminated after a further 12 min with 100 μl 0.2 M citric acid. Samples were extracted with ethyl acetate and separated by silica gel HPTLC using a chloroform/methanol/acetic acid/water (90:9:1.0:0.65) solvent system. Radioactivity profiles were monitored with a TLC plate scanner (Berthold LB 2482), and peaks were identified by comparison with authentic radiolabeled products. TXB2 and HHT were used as the measured PGHS products. Incubations in the presence of 10 μM piroxicam were used to define maximum PGHS inhibition, residual counts in the peak region in these incubations being subtracted from the peak counts.
Inhibition of isolated ovine PGHS-1 and -2.
The assay method of Futaki et al. (1994) was adapted. Preliminary studies examined substrate and enzyme concentrations, assay time course and PGE2 extraction efficiency (results not shown). The final optimized assay was as follows: enzyme preparation (2 U) was incubated with test compound for 2 min at 37°C in 100 mM Tris-HCl pH 8.0 containing 1 μM hematin, 2 mM phenol and 0.1% DMSO. Arachidonic acid was added (10 μM final, total assay volume 0.5 ml), and the incubation was continued for a further 10 min before termination with 2 ml hexane/ethyl acetate (2:1). The samples were centrifuged at 1200 × g for 10 min at 4°C, and the organic phase was discarded after freezing the aqueous phase on dry ice/propanol. The extraction procedure was repeated twice, and the samples were assayed for PGE2 by RIA.
Inhibitory effects in a rat air pouch model of inflammation.
The model of Hambleton and Miller (1989) was used. Briefly, male Wistar rats (Interfauna: 160–180 g, 6–7 weeks old) were lightly anesthatised with halothane/oxygen, and dorsal air pouches were formed by s.c. injection of 20 ml of air; after 3 days, the air pouches were reinflated with 10 ml of air. Inflammation was induced in 6-day-old air pouches by injection of carrageenan λ (10 mg in 1 ml of saline); negative controls were sham-injected. After 6 hr the rats were sacrificed by asphyxiation in carbon dioxide, and pouch contents were collected in a 2-ml saline/heparin (10 U/ml)/phenidone (0.1 mg/kg) wash. Leukocytes were counted, and PGE2 concentrations were determined by RIA (after centrifugation at 130 × g for 8 min at 4°C to remove cell debris). Drugs were administered p.o. (10 and 25 mg/kg) in an aqueous solution of 5% acacia oil and 0.01% Tween 80 (10 ml/kg) 1 hr before carrageenan administration. The compounds are both markedly immunosuppressive in delayed type hypersensitivity assays at these doses in the rat (Bartlett and Schleyerbach, 1985; Kuoet al., 1996). Statistical analysis was carried out using Student’s t test.
Human peripheral blood cell preparation.
Erythrocytes were removed from human leukocyte concentrate (Regional Blood Transfusion Centre, Southmead, Bristol) by sedimentation in 1% dextran in PBS for 15 to 30 min at room temperature. The supernatant was centrifuged (350 × g, 10 min, 4°C), and the resulting pellets were washed once in PBS and resuspended in isotonic 45% percoll. The cells were then separated by centrifugation (350 × g, 30 min, 7°C) through isotonic 58.5% and 74.25% percoll layers. PMNLs were collected from the interface of the 58.5% and 74.25% layers, and PBMCs from the 45% and 58.5% interface. The cells were washed twice in PBS before resuspension in RPMI 1640 for PGHS-1 studies or, for PGHS-2 studies, in RPMI 1640 containing 5% fetal calf serum, 2 mM glutamine and 100 U/ml penicillin, 0.1 mg/ml streptomycin and 0.25 μg/ml amphotericin B (complete medium).
Inhibition of constitutive PGHS in human PBMCs.
Two types of assays were used. The first, using [1-14C] arachidonic acid as substrate, was identical to that on guinea pig PMNLs except that the PBMCs were diluted to 1 × 107 cells/ml, incubation with test compound was for 10 min and 100 μM piroxicam was used to define background incorporation into TXB2 and HHT peak regions.
The second assay used endogenous arachidonic acid as substrate. The PBMCs were diluted to 5 × 106 cells/ml with assay buffer (15 mM Tris-HCl, 134 mM NaCl, 5 mM glucose, pH 7.4), and 970 μl were incubated with test compound for 10 min at 37°C before CaCl2 (10 mM) and A23187 (1 μM) were added (in a dose-response study, ionophore increased PGE2 production dose-dependently with a plateau at 0.3 μM A23187; data not shown). The tubes were incubated for a further 10 min before termination with indomethacin (100 μM final) and cooling on ice (in a time course study, PGE2 increased linearly to 299 pg/106cells over a 15-min incubation). Supernatants, collected after centrifugation at 1250 × g for 10 min, were assayed for PGE2 by RIA.
In one study, 10% human plasma was included in the cell incubation, and 50 μM A23187 was used.
Induction of PGHS in isolated human peripheral blood cells.
All incubations were at 37°C and 5% CO2. PMNLs (2.5 × 106 cells/ml) were incubated in complete medium with 100 μM aspirin for 1 hr. The cells were washed four times and resuspended to 2.5 × 106 cells/ml before overnight incubation with 100 μg/ml LPS in complete medium (LPS up to 100 μg/ml caused a dose-related increase in PGE2 production in a preliminary study). The cells were washed once in RPMI 1640 and resuspended at 2.5 × 106 cells/ml.
PBMCs (1 × 107 cells/ml) were incubated with 100 μM aspirin in Petri dishes in complete medium (10 ml/dish) for 1 hr. Nonadherent cells and aspirin were removed by washing four times, and the adherent cells were treated overnight with 10 μg/ml LPS in complete medium. The cells were resuspended by scraping and were washed twice in RPMI 1640 before resuspension at 3 × 105 to 1 × 106 cells/ml.
The effects of 2 μM dexamethasone added in buffer at the start of the LPS induction were evaluated for both PBMCs and PMNLs in preliminary studies.
Inhibition of induced PGHS in isolated human peripheral blood cells.
Cells (1.25 × 106 PMNLs or 1.5 to 5 × 105 adherent PBMCs) were incubated with test compounds in RPMI 1640 (total volume 500 μl) at 37°C for 10 min (PMNLs) or 2 min (adherent PBMCs). A23187 was added; stimulation with ionophore was not required to produce detectable levels of PGE2, but it did enhance production in a dose-dependent manner (data not shown), and 10 μM was used as standard for PMNLs and 1 μM for adherent PBMCs. After a further 10 min, incubations were terminated with ibuprofen (100 μM final) and cooling on ice. Cells were pelleted by centrifugation (300 × g, 4°C, 5 min), and supernatants were assayed for PGE2 by RIA.
Inhibition of human whole blood PGHS activity.
Fresh blood was collected from healthy volunteers using heparinized safety monovettes (Sarstedt), and 989 μl was incubated with test compound for 10 min at 37°C. Incubations were continued for a further 20 min after the addition of 50 μM A23187 (in a preliminary study, PGE2 production increased in a dose-dependent manner with increasing calcium ionophore until it reached a plateau at 30 μM ionophore; data not shown). Finally, incubations were cooled on ice for 5 min, and plasma was collected by centrifugation (1250 ×g, 4°C, 10 min). PGE2 was assayed within 4 days of storage at −80°C.
PGE2 radioimmunoassay and data analysis.
The PGE2 RIA was carried out using a commercial assay kit according to the manufacturer’s instructions. Briefly, 100 μl of PGE2 standard (1–250 pg) or unknown sample (diluted as required) was incubated with 100 μl of [125I] PGE2 tracer and 100 μl of anti-PGE2 antibody for 20 to 24 hr at 4°C. Antibody bound PGE2 was removed by precipitation with 1 ml of 16% polyethylene glycol precipitating reagent and centrifugation at 2000 × g for 30 min at 4°C. The pellets were counted for radioactivity.
All data analysis was performed using the regression analysis software Grafit 3.0 (Erithacus software). Normalized percent bound values were calculated for PGE2 standards, and a standard curve was prepared by fitting the data to a 4-parameter logistic equation. The PGE2 concentrations in unknowns were determined using the calculated parameters of the fit. The PGE2 concentration in nonstimulated cells was subtracted from that in control and drug-treated cells. IC50 values for inhibition of PGE2 generation were then determined by fitting all the data in a dose-response curve to the equation defining an IC50.
Electrophoresis and Western blotting.
PBMCs were incubated in the presence or absence of LPS for 5 or 18 hr as described in the section on induction of PGHS; dexamethasone was included in one of the 18-hr incubations. One Petri dish was used per treatment, and each dish contained approximately 3 × 106 adherent cells. At the required time-point, adherent cells and supernatants were collected and separated by centrifugation at 320 × g for 10 min. The supernatants (for PGE2 RIA) and cell pellets, resuspended in 10 ml of PBS, were stored at −80°C until required.
The cells were thawed, sonicated (3 × 10 sec on maximum power using a Microson Ultrasonic Cell Disruptor) and centrifuged at 440 × g for 10 min to remove nuclei and cell debris. Supernatants were centrifuged at 125,000 × g for 30 minutes, and the resulting membrane pellets were resuspended in SDS-PAGE sample buffer and subjected to discontinuous SDS-PAGE (Laemmli, 1970). Electrophoretic transfer onto Immobilon P was carried out at 0.8 mA/cm2 (Williamson et al., 1995).
The Immobilon membrane was incubated overnight in PBS containing 0.25% Tween-20 and 3% nonfat milk and then was washed four times before incubation for 1 hr with primary antibody (anti-PGHS-1 at 1-in-100 dilution, anti-PGHS-2 at 1-in-2000 dilution). The wash step was repeated, and the membrane was incubated with secondary antibody for 1 hr (alkaline phosphatase-conjugated mouse anti-rabbit IgG at 1-in-100,000 dilution). After washing, the blots were developed using nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate according to the manufacturer’s instructions. All washes were in PBS/Tween-20; antibody dilutions were in PBS/Tween-20/nonfat milk.
Results
Effects of test compounds on PGHS from guinea pig peritoneal PMNLs.
All test compounds inhibited guinea pig PGHS activity as evidenced by reductions in TXB2 or HHT production, and both products were affected with similar potency by each compound (table1). Although HR325 was a more effective inhibitor than A77 1726 (11- to 21-fold), it was much weaker (>17-fold for TXB2 production) than all the NSAIDs except ibuprofen, with which it was broadly equipotent.
Isolated ovine PGHS-1 and -2.
IC50 values for PGHS-1 and -2 inhibition are shown in table 2. A77 1726 was an extremely weak inhibitor of both enzymes, and HR325 was approximately 12-fold more potent for PGHS-1 (IC50 = 64 μM) and about 26-fold more potent for PGHS-2 (IC50 = 101 μM). In both cases, the potency of HR325 was within the range seen with the tested NSAIDs. All compounds were, however, much weaker inhibitors of the isolated enzymes than in the guinea pig PMNL assay (table 1) and the assays on human cells discussed below. For the isolated enzyme assay, the PGE2 is formed nonenzymatically from PGH2. If this were rate limiting, then the IC50 values for PGE2 generation would significantly overestimate the true IC50 values for PGHS inhibition. However, in preliminary studies, the rate of PGE2 formation increased linearly with enzyme concentration, which indicates that the rate of PGE2formation was directly proportional to PGHS activity and that the conversion of PGH2 to PGE2 was not rate-limiting. The possibility of species differences or loss of ovine enzyme integrity during isolation cannot be entirely discounted. However, Mitchell et al.(1994) reported similar potencies for ibuprofen (5–15 μM for COX1, 200–800 μM for COX2) and aspirin (2–44 μM for COX1, 0.3–1 mM for COX2) on isolated ovine enzymes and broken or whole cells assayed at high arachidonate concentrations (7 μM for ovine enzymes and 30 μM on whole cells). The disparity in drug potencies between isolated enzymes and whole cell assays can be accounted for at least in part by the competitive nature of the inhibition of PGHS activity by the NSAIDs (and presumably HR325 and A77 1726) and by the use of a high substrate concentration (10 μM) in the isolated enzyme assays.
Anti-inflammatory effects in the rat air pouch.
In vivo HR325 inhibited accumulation of leukocytes and PGE2 when dosed at 25 mg/kg but had no significant effect at 10 mg/kg (table 3). Whereas leflunomide was inactive at both doses, the reference compounds tenidap (25 mg/kg) and indomethacin (10 mg/kg) markedly inhibited both parameters. The inhibitory effects of indomethacin have been reported previously (Ohuchi et al., 1982; Hirasawa et al., 1987), and the effects of tenidap are consistent with its being a potent PGHS inhibitor in the rat (Proudmann and McMillian, 1991).
Effects of test compounds on PGHS-1 in human PBMCs.
The potencies of test compounds as inhibitors of constitutive human PGHS were determined in two types of assays on PBMCs. One approach, which was similar to that used for guinea pig PMNLs, used [14C] arachidonic acid as substrate with HHT and TXB2 as the measured products. The second approach involved the use of endogenous arachidonic acid as substrate with PGE2 determined by radioimmunoassay.
Test compounds inhibited constitutive PGHS in a dose-dependent manner, and potencies were similar in assays using endogenous or exogenous arachidonic acid as substrate and, in the latter assay, with TXB2 or HHT as the measured product (table4). The PGHS-2-selective compound NS398 had low inhibitory activity, which identified the constitutive enzyme as PGHS-1. A77 1726 was an extremely weak inhibitor of human PGHS-1 (table4). Although HR325 was more potent than A77 1726 (∼7- to 20-fold), it was weaker than the compounds with which it was compared, though piroxicam and flufenamic acid were only ∼2- to 5-fold more active.
Inclusion of 10% human plasma in the assay reduced the potencies of HR325 (IC50 = 36 μM), A77 1726 (IC50 = 100 μM) and naproxen (IC50 = 759 nM) without markedly affecting piroxicam activity (IC50 = 251 nM).
Characterization of LPS-induced PGHS activity in human peripheral blood cells.
PMNLs did not produce detectable levels of PGE2 before stimulation with LPS. However, constitutive PGHS that may have been present at undetectable levels was irreversibly inactivated by a 1-hr aspirin treatment before the induction of PGHS-2 (Roth et al., 1975; DeWitt et al., 1990). On stimulation with A23187 (10 μM), the LPS-induced PMNLs produced 91 ± 25 pg PGE2/106 cells (n = 5). Dexamethasone (2 μM), included in the LPS induction step, inhibited subsequent PGE2 production by 56% and 44% in replicate experiments.
In contrast to PMNLs, PBMCs stimulated with A23187 produced detectable levels of PGE2 in the absence of LPS induction (fig.1). This was markedly reduced (one experiment) or completely abolished (replicate experiment) after treatment with 100 μM aspirin for 1 hr at 37°C (followed by overnight incubation in the absence of aspirin or LPS). Inclusion of LPS (10 μg/ml) in the incubation induced PGHS activity, and PGE2 production was further enhanced by ionophore stimulation (fig. 1). As with the PMNLs, dexamethasone (2 μM) inhibited PGHS induction. This treatment also abolished residual PGE2 generation in aspirin-treated noninduced cells, which suggests that it reflected adhesion-induced PGHS-2 activity as has been observed in murine peritoneal macrophages (Tordjman et al., 1995). LPS-induced ionophore-stimulated PBMCs produced 5.6 ± 4.9 ng PGE2/106cells, (n = 13).
Adherent PBMCs exhibited time-dependent induction of an anti-PGHS-2 detectable Mr 72-kDa doublet, which was markedly stronger in LPS-treated cells (fig. 2A). The time course for induction and the relative levels of the induced protein were reflected in the amount of PGE2 produced during the induction period (fig. 2). PGHS-2 was not detectable in cells induced with LPS in the presence of dexamethasone, and the cells produced only very low levels of PGE2 (fig. 2). No anti-PGHS-1 detectable bands were induced in PBMCs under any conditions tested (data not shown).
Effects of test compounds on LPS-induced human PGHS activity.
NS398 displayed potent inhibition of PGE2 generation in LPS-induced cells (table 5), a result that confirms the identity of the induced enzyme as PGHS-2 in both PMNLs and adherent PBMCs. The potency of HR325 was within the range of the NSAIDs, and it was more active (22- to 35-fold) than A77 1726, which was the weakest inhibitor tested (table 5).
Effects of test compounds on PGHS activity in whole blood.
Potencies of all the tested compounds were markedly lower in the whole blood assay than in assays on isolated cells (table 6). Dose responses for HR325 and A77 1726 were incomplete, with only 50% to 55% inhibition achieved at the highest drug concentrations. A77 1726 had extremely weak activity (IC50 ∼292 μM), and HR325 was only slightly more potent (IC50 ∼164 μM). Both compounds were much weaker inhibitors than all the tested NSAIDs.
Discussion
A77 1726 and HR325 exhibit both anti-inflammatory and immunomodulatory properties (Bartlett et al., 1993;Hambleton et al., 1992), and the drugs are equipotent inhibitors of dihydroorotate dehydrogenase, a key mediator of their immunomodulatory functions (Williamson et al., 1995;Cherwinski et al., 1995; Kuo et al., 1996). The current studies were designed to compare the anti-inflammatory activities of the compounds in vivo and to determine their potencies as PGHS inhibitors in vitro.
HR325 inhibited PGHS activity of guinea pig PMNLs more potently than did A77 1726 (table 1). Similarly, it inhibited isolated ovine PGHS-1 and PGHS-2 more strongly than did A77 1726 and with a potency within the range of the tested NSAIDs. The in vitro activity of HR325 was expressed as in vivo anti-inflammatory activity at a 25-mg/kg dose, as evinced by inhibition of cellular infiltration and PGE2 production in rat carrageenan-induced air pouch inflammation; the compound was inactive, however, at 10 mg/kg (table3). Leflunomide displayed no in vivo anti-inflammatory activity at the 25-mg/kg dose, a finding consistent with the much weaker inhibition of PGHS by A77 1726 (tables 1 and 2) and with published data (Bartlett et al., 1989). In contrast, both HR325 and leflunomide are immunologically active at 10 mg/kg (Bartlett and Schleyerbach 1985; Popovic and Bartlett, 1986; Kuo et al., 1996). The lack of effect of leflunomide in the current study is not related to a delay in release of the active metabolite A77 1726, because this occurs very rapidly in the digestive tract (Brazelton and Morris, 1996).
To assess the clinical relevance of inhibition of PGHS by HR325, we performed assays using human enzymes. PGHS-1 activity was measured in PBMC, the identity of the activity being confirmed by the low potency of inhibition by the PGHS-2-selective compound NS398 (Futakiet al., 1994) and also by the constitutive nature of the activity. PGHS-2 activity was measured in LPS-induced PMNLs and adherent PBMCs. The identity of the enzyme responsible for the activity was confirmed first by its sensitivity to inhibition by NS398 (table5), second by the LPS-driven induction of a protein recognized by an anti-PGHS-2 antibody in PBMC (fig. 2) and finally by the dexamethasone-sensitivity of the induction of both PGHS activity (figs.1 and 2) and anti-PGHS-2 reactive protein (fig. 2: O’Banion et al., 1991; Kujubu and Herschman 1992; Lee et al., 1992). We failed to detect induced PGHS-2 protein in LPS-induced PMNLs (data not shown), a finding that is consistent with a previous report (Patrignani et al., 1994) and that probably reflected the low level of induced enzyme activity in these cells. Despite this, PMNL-derived PGE2 is likely to be important in the early phase of the inflammatory response and during acute inflammation when PMNLs predominate at the inflammatory site (Mackay et al., 1985; Hambleton and Miller, 1989; Martin et al., 1994).
HR325 inhibited both human PGHS-1 and human PGHS-2 more potently than did A77 1726 (tables 4 and 5). Furthermore, although its potency against PGHS-1 was below the range of the tested NSAIDs, it was slightly more potent as a PGHS-2 inhibitor than naproxen, ibuprofen and piroxicam (2- to 8-fold) and it displayed some selectivity for PGHS-2 (3- to 12-fold).
Because of the competitive nature of inhibition of PGHS by a large number of NSAIDs, assays using endogenous arachidonate have been proposed as the most accurate assessment of inhibitory activity (Laneuville et al., 1994; Smith et al., 1995). To our knowledge, this is the first report of potencies of a range of NSAIDs against human enzymes using such an approach. The IC50 values are thus not directly comparable with those in earlier studies, where high substrate concentrations (10–30 μM) were used, forcing IC50 values into the micromolar range (Meadeet al., 1993; Mitchell et al., 1994; Barnettet al., 1994; Laneuville et al., 1994; Gierseet al., 1995; Grossman et al., 1995). Moreover, the use of high arachidonic acid concentrations (10 μM) has been proposed, by Laneuville et al. (1994) and Smith et al. (1995), to account for the paradox between the therapeutic efficacy of piroxicam and its lack of potency as a human PGHS-2 inhibitor (IC50 > 500 μM and IC50 > 100 μM, Laneuville et al., 1994; Gierse et al., 1995; respectively). The much lower IC50 values for piroxicam on human PGHS-2 that are reported in Table 5 (260–290 nM) are consistent with this suggestion. In addition, the NSAIDs show high serum binding (Williams et al., 1994), and the potencies of all but piroxicam were markedly reduced in the whole blood assay (table6). As a result, piroxicam ranked as one of the most potent whole blood inhibitors despite its comparatively low potency in the isolated cell assays. In accordance with this, inclusion of human plasma in the PGHS-1 assay reduced the potency of naproxen (the IC50value went from 28 nM to 759 nM) while having no effect on the potency of piroxicam (the IC50 value went from 209 nM to 251 nM). The apparently lower serum binding of piroxicam is likely to be an important factor in defining its relative therapeutic value.
The degree of selectivity of the NSAIDs for the isozymes can be compared with earlier reports. Of the tested NSAIDs, only NS398 and flufenamic acid were PGHS-2-selective (tables 4 and 5); diclofenac and piroxicam were largely nonselective, and the other NSAIDs displayed higher potencies against PGHS-1 than against PGHS-2. With the exception of piroxicam and flufenamic acid, these profiles are in agreement with earlier reports (Futaki et al., 1994; Meade et al., 1993; Mitchell et al., 1994; Barnett et al., 1994; Laneuville et al., 1994; Gierse et al., 1995); the exceptions may be related to the use of endogenous arachidonic acid in the current studies.
The inhibition of human PGHS-2 and a rat in vivo model of inflammation by HR325 could be indicative of anti-inflammatory properties in the human. However, A77 1726 (Lucien et al., 1995) and HR325 (S. Gadher, C.M. Yea and R.A. Williamson, unpublished observations), show extremely high levels of serum protein binding. Consequently, inclusion of 10% human plasma in the PGHS-1 assay increased the IC50 value for HR325 from 1 μM to 36 μM and for A77 1726 from 8 μM to 100 μM. Potencies were lower still in the whole blood assay (table 6), increasing the IC50 values for both HR325 (164 μM) and A77 1726 (292 μM) and almost abolishing the differential activity of the compounds seen in the isolated cell assays. Whether the higher potency of inhibition of human PGHS-2 by HR325 would be manifest as therapeutic anti-inflammatory activity is unclear. It is noteworthy, however, that HR325 exerts anti-inflammatory effects in rat air pouch inflammation despite being similarly highly protein-bound in rat serum.
A77 1726 at concentrations up to 22 μM has previously been shown to have no inhibitory effects on either PGE2 generation by sheep seminal vesicle and rabbit mucosal preparations or aggregation of human platelets (Weithmann et al., 1994). In addition, LPS-induced PGHS activity of human PMNL was not inhibited by 100 μM A77 1726. However, these determinations were made in the presence of high concentrations of exogenous arachidonate (0.7–8 μM), human plasma or 5% fetal bovine serum, all of which would diminish the effectiveness of this very weak PGHS inhibitor. Indeed, we found A77 1726 to be only very weakly active, even at concentrations in excess of those previously used, either in assays in which 10 μM arachidonate was present (ovine enzymes) or in the whole blood assay, and noin vivo anti-inflammatory activity of leflunomide was detected.
Finally, indomethacin, flurbiprofen, diclofenac and tenidap display extremely high relative potencies for both ovine and human PGHS. Although this probably reflects time-dependent inhibition in the case of the first three compounds (Laneuville et al., 1994), inhibition of PBMC PGHS-2 by tenidap displayed no such time dependence (data not shown).
In conclusion, we have shown that HR325 inhibits both PGHS-1 and PGHS-2 from both animal and human sources more potently than does A77 1726 and, furthermore, that HR325 displays human PGHS-2 selectivity in common with only two of the other compounds tested. Whether thisin vitro activity will be expressed as therapeutic, anti-inflammatory activity in the human remains to be elucidated, particularly in view of the effects of serum protein binding on drug potency.
Acknowledgments
We are very grateful to C. Nightingale for technical assistance and to D. P. Kay, C. Hidden, Dr. I. R. Ager, Dr. S. S. Matharu, G. Danswan and Dr. F. J. Kammerer for preparation of drugs. We would also like to thank G. Danswan for photographic assistance.
Footnotes
-
Send reprint requests to: E. Ruuth, Domaine Therapeutic Immunology, Roussel UCLAF, 102/111 Route de Noisy, Romainville, France.
-
↵1 Present address: Ferring Research Institute, Chilworth Research Centre, Southampton, SO16 7NP, UK.
-
↵2 Present address: Enzyme Pharmacology Unit, Glaxo Wellcome, Medicines Research Centre, Gunnels Wood Road, Stevenage, Herts, SG1 2NY, UK.
- Abbreviations:
- DMSO
- dimethylsulfoxide
- DTH
- delayed-type hypersensitivity
- HHT
- 12-hydroxyheptadecatrienoic acid
- HPTLC
- high-performance thin-layer chromatography
- LPS
- lipopolysaccharide
- NSAID
- nonsteroidal anti-inflammatory drug
- PBMC
- peripheral blood mononuclear cell
- PBS
- phosphate-buffered saline
- PGHS
- prostaglandin endoperoxide H synthase
- PMNL
- polymorphonuclear leukocyte
- RIA
- radioimmunoassay
- TXB2
- thromboxane B2
- SDS-PAGE
- sodium dodecyl sulphate-poly acrylamide gel electrophoresis
- Received September 24, 1996.
- Accepted February 26, 1997.
- The American Society for Pharmacology and Experimental Therapeutics