Predominance of cyclooxygenase 1 over cyclooxygenase 2 in the generation of proinflammatory prostaglandins in autoantibody-driven KBxN serumtransfer arthritis.код для вставкиСкачать
ARTHRITIS & RHEUMATISM Vol. 58, No. 5, May 2008, pp 1354–1365 DOI 10.1002/art.23453 © 2008, American College of Rheumatology Predominance of Cyclooxygenase 1 Over Cyclooxygenase 2 in the Generation of Proinflammatory Prostaglandins in Autoantibody-Driven K/BxN Serum–Transfer Arthritis Mei Chen,1 Eric Boilard,1 Peter A. Nigrovic,1 Patsy Clark,2 Daigen Xu,2 Garret A. FitzGerald,3 Laurent P. Audoly,2 and David M. Lee1 COX-2ⴚ/ⴚ mice as well as isoform-specific inhibitors. The relative importance of PGE2 and PGI2 (prostacyclin) was determined using mice deficient in microsomal PGE synthase 1 (mPGES-1) and in the receptors for PGI2. Results. High levels of PGE2 and 6-keto-PGF1␣ (a stable metabolite of PGI2) were detected in arthritic joint tissues, correlating strongly with the intensity of synovitis. Pharmacologic inhibition of PG synthesis prevented arthritis and ameliorated active disease. While both COX isoforms were found in inflamed joint tissues, only COX-1 contributed substantially to clinical disease; COX-1ⴚ/ⴚ mice were fully resistant to disease, whereas COX-2ⴚ/ⴚ mice remained susceptible. These findings were confirmed by isoform-specific pharmacologic inhibition. Mice lacking mPGES-1 (and therefore PGE2) developed arthritis normally, whereas mice incapable of responding to PGI2 exhibited a significantly attenuated arthritis course, confirming a role of PGI2 in this arthritis model. Conclusion. These findings challenge previous paradigms of distinct “housekeeping” versus inflammatory functions of the COX isoforms and highlight the potential pathogenic contribution of prostanoids synthesized via COX-1, in particular PGI2, to inflammatory arthritis. Objective. Prostaglandins (PGs) are found in high levels in the synovial fluid of patients with rheumatoid arthritis, and nonsteroidal blockade of these bioactive lipids plays a role in patient care. The aim of this study was to explore the relative contribution of cyclooxygenase (COX) isoforms and PG species in the autoantibody-driven K/BxN serum–transfer arthritis. Methods. The prostanoid content of arthritic ankles was assessed in ankle homogenates, and the importance of this pathway was confirmed with pharmacologic blockade. The presence of COX isoforms was assessed by Western blotting and their functional contribution was compared using COX-1ⴚ/ⴚ and Drs. Chen and Boilard’s work was supported by grants from the Arthritis Foundation. Dr. FitzGerald’s work was supported by a grant from the National Heart, Blood, and Lung Institute, NIH (P01-HL-62250). Dr. Lee’s work was supported by grants from the Arthritis Foundation, the NIH (P01-AI-065858-01), and the Cogan Family Foundation. 1 Mei Chen, MD, PhD, Eric Boilard, PhD, Peter A. Nigrovic, MD, David M. Lee, MD, PhD: Brigham and Women’s Hospital, Harvard Medical School, Boston, Massachusetts; 2Patsy Clark, MSc, Daigen Xu, PhD, Laurent P. Audoly, PhD: Merck Frosst Centre for Therapeutic Research, Kirkland, Quebec, Canada; 3Garret A. FitzGerald, MD: University of Pennsylvania School of Medicine, Philadelphia. Dr. FitzGerald has received honoraria or fees for consulting or speaking (less than $10,000 each) from Merck, Novartis, Daiichi, and NicOx. Dr. Audoly owns stock or stock options in Merck & Company, Inc. Dr. Lee has received honoraria or fees for consulting or speaking (less than $10,000 each) from Resolvyx, UCB Pharma, and Religen; owns stock or stock options in Synovex; and has received research support from MedImmune, Biogen, and Genentech. Address correspondence and reprint requests to David M. Lee, MD, PhD, Department of Medicine, Division of Rheumatology, Immunology and Allergy, Brigham and Women’s Hospital, Harvard Medical School, Boston, MA 02115. E-mail: [email protected] harvard.edu. Submitted for publication July 17, 2007; accepted in revised form February 1, 2008. Prostaglandins (PGs) are lipid mediators that, in addition to their role in numerous physiologic activities, contribute to the pathogenesis of pathologic inflammation. They are generated by conversion of arachidonic acid by the cyclooxygenase (COX) enzymes to PGH2, which is further catalyzed by distinct synthases to 5 major bioactive prostaglandins (PGE2, PGI2, PGF2␣, 1354 COX ISOFORMS AND PG SPECIES IN K/BxN SERUM–TRANSFER ARTHRITIS PGD2, and thromboxane A2) (1). Two isoforms of COX, designated COX-1 and COX-2, exist outside the brain (2,3). Classically, COX-1 is constitutively expressed in most tissues, whereas COX-2 is induced by a range of mitogenic and inflammatory stimuli. These expression patterns gave rise to the hypothesis that COX-1 provides “housekeeping” synthetic activity, while prostaglandin synthesis in inflammatory conditions is largely attributable to COX-2. However, in vivo studies have called into question this paradigm of a clear division of labor between COX-1 and COX-2, with synthetic contributions from COX-2 to healthy gastric and renal physiology and contributions from COX-1 to inflammatory states (4–10). Thus, there is an increasing appreciation that pathways of prostanoid generation in health and disease are not readily described by simple paradigms. In rheumatoid arthritis (RA) and other inflammatory joint diseases, high concentrations of prostaglandin species have been detected in synovial fluid (11,12). To assess the potential pathogenic contribution of these mediators, investigators have turned to mouse models of arthritis, in particular, collagen-induced arthritis (CIA). In this model, immunization with type II collagen in Freund’s complete adjuvant elicits a chronic inflammatory arthritis. This model exhibits a significant reliance on the COX-2 isoform for the development of anticollagen antibodies and clinical synovitis (13,14). Dissection of the role of individual prostanoid species in mouse CIA has established a requirement for PGE2, whose synthesis is reliant on microsomal PGE synthase 1 (mPGES-1) (15), and for PGI2 (prostacyclin) via its receptor (the IP receptor) (16). Further analysis of the mechanisms of the PGE2 contribution has focused on the function of receptors for PGE2, revealing a dual contribution by the EP2 and EP4 receptors in CIA (16). A variation of the mouse model of CIA, using pathogenic anticollagen antibodies and lipopolysaccharide (LPS) (collagen antibody–induced arthritis [CAIA]), has also demonstrated significant contributions from both PGE2 (via its receptor EP4) and prostacyclin to disease (17,18). Given the dynamic regulation of COX isoform expression, we elected to reevaluate the pathogenic role of COX enzymes and downstream mediators in a model system that requires neither adjuvant nor LPS. In the K/BxN serum–transfer model of arthritis, administration of serum from arthritic K/BxN mice induces inflammatory arthritis in most recipient strains, which is mediated by IgG autoantibodies; coadministration of additional 1355 agents is not required (19–22). In the present study, we examined the contribution of prostaglandins to both the induction and the perpetuation of arthritis in this model. We found a significant elevation of prostanoids in the joint that mirrored disease progression. Genetic and pharmacologic studies demonstrated a prominent contribution of prostanoids to disease initiation and to propagation of chronic inflammation. Interestingly, we showed a particular requirement for the COX-1 isoform, whereas the COX-2 isoform was apparently dispensable. Finally, we found that disease progressed in the absence of PGE2 and that there was a partial dependence on PGI2 acting via the IP receptor. These findings provide a further counterexample to the paradigm that inflammatory prostaglandin production is dependent on COX-2 and underscore the rationale for continued examination of prostanoid pathways in human arthritis. MATERIALS AND METHODS Mice. We used mice ages 6–10 weeks for these studies. C57BL/6 mice were purchased from The Jackson Laboratory (Bar Harbor, ME). Male COX-1⫺/⫺, COX-2⫺/⫺, and wild-type (WT) mice (C57BL/6 ⫻ 129/Ola founders [7,23] intercrossed for ⬎30 generations) were purchased from Taconic (Germantown, NY). Microsomal PGE synthase 1–null mice (N5 backcross onto the C57BL/6 background) (15), PGE2 receptor EP1– and PGI receptor IP–null mice (N ⬎10 on the C57BL/6 background) were bred locally (24,25). K/BxN mice were maintained as described elsewhere (20). All procedures were approved by the Institutional Animal Care and Use Committee of the Dana-Farber Cancer Institute (Boston, MA). Drugs. The following drugs were used. Sulindac sulfone ([Z]-5-fluoro-2-methyl-1-[p-(methylsulfonyl)benzylidene]indene-3-acetic acid) was obtained from Sigma (St. Louis, MO). SC-560 (5-[4-chlorophenyl]-1-[4-methoxyphenyl]-3[trifluoromethyl]-1H-pyrazole) was purchased from Cayman Chemical (Ann Arbor, MI). MF-tricyclic (MFT; 3-[3,4-difluorophenyl]-4-[4-(methylsulfonyl)phenyl]-2-[5H]-furanone) was provided by Merck Frosst Centre for Therapeutic Research (Kirkland, Quebec, Canada) (26). These drugs were suspended and diluted in 1% methylcellulose. Medications were administered orally via gavage once daily. The doses used, 10 mg/kg for sulindac, 10 mg/kg for SC-560, and 3 mg/kg for MFT, were chosen based on previously defined pharmacokinetic profiles of these drugs in mice (26–29). A vehicle control (1% methylcellulose) was administered orally in the same volume and frequency to control mice. The biochemical activity of COX inhibitors was assessed in ex vivo analyses of whole blood using enzyme-linked immunosorbent assay measurement of thromboxane B2 (a surrogate for COX-1 activity) or LPS-stimulated PGE2 (for COX-2 activity) production according to the manufacturer’s instructions (Cayman Chemical) (30–34). Serum transfer protocol and arthritis scoring. Arthritogenic K/BxN serum was transferred to recipient mice to 1356 induce arthritis, as described previously (20). Briefly, 150 l of serum was administered intraperitoneally on experimental days 0 and 2. Clinical indices were recorded at 24–48-hour intervals. Ankle thickness was measured at the malleoli with the ankle in a fully flexed position, using spring-loaded dial calipers (Long Island Indicator Service, Hauppauge, NY). The clinical index of arthritis was graded on a scale of 0–12 as described previously (35). Measurements of prostanoids in the inflamed joints by liquid chromatography mass spectrometry (LC-MS). At selected times after injection with K/BxN serum, ankle tissues were harvested, weighed, and frozen in liquid nitrogen (18). Frozen joint tissues were pulverized using a mortar and pestle to obtain a fine powder. This powder was homogenized (Polytron PRO200 homogenizer; PRO Scientific, Oxford, CT) at 4°C in phosphate buffered saline (PBS) supplemented with 10 M indomethacin and 1⫻ Complete Protease Inhibitor mixture (Roche Applied Science, Laval, Quebec, Canada). The homogenates were subsequently sonicated on ice for 10– 30 seconds (Cole-Parmer Ultrasonic homogenizer, 50% output; Cole-Parmer, Montreal, Quebec, Canada) and centrifuged at 1,000g for 10 minutes at 4°C. Supernatants were isolated and stored at ⫺80°C until further analyses were performed. The measurement of PGE2 and 6-keto-PGF1␣ (stable breakdown product of PGI2) by LC-MS was performed as described previously (36). Briefly, samples (100 l) were protein-precipitated by the addition of 150 l of acetonitrile containing 2 ng/ml of deuterated prostanoids that served as internal standards for quantification. Samples were mixed thoroughly by pipetting, centrifuged at 1,200g for 10 minutes at 4°C, and supernatants were transferred to a new 96-well plate for analysis by LC-MS. The detection limit for LC-MS is 0.002 ng/mg of protein for PGE2 and 6-keto-PGF1␣. The concentrations for PGE2 and 6-keto-PGF1␣ were normalized relative to tissue wet weight. Isolation of protein and Western blot analysis. Ankles were harvested at the indicated times, placed in 1.5 ml of ice-cold lysis buffer (10 mM Tris, pH 7.4, 140 mM NaCl, 10 mM EDTA, 1% Triton X-100, 0.05% sodium dodecyl sulfate [SDS] and freshly added 5 mM phenylmethylsulfonyl fluoride, as well as a protease inhibitor cocktail [catalog no. P8340; Sigma]), cut into small pieces with a scalpel, and homogenized using a Brinkmann homogenizer (Brinkmann, Westbury, NY). Lysates were subsequently clarified by centrifugation (9,500g for 10 minutes at 4°C), and the protein concentration in the lysates was determined using Bradford reagent (Sigma). Typically, lysates contained 2–7 mg/ml of protein. Proteins (80 g) boiled in Laemmli sample buffer (37) were separated on 10% SDS– polyacrylamide gels and transferred to polyvinylidene difluoride membranes. The membranes were blocked for 30 minutes at room temperature with 5% milk proteins in Tris buffered saline–Tween (TBST; 190 mM NaCl, 0.05% Tween 20, 25 mM Tris, pH 7.6), washed in TBST, and incubated with anti– COX-1 (1:400 dilution), anti–COX-2 (1:400 dilution) (both from Cayman Chemical), or anti-GAPDH (1:5,000 dilution) (Abcam, Cambridge, MA) for 18 hours at 4°C. Membranes were washed 6 times for 5 minutes each in TBST and incubated for 1 hour with horseradish peroxidase– CHEN ET AL conjugated donkey anti-rabbit or donkey anti-mouse IgG (1:15,000 dilution; Jackson ImmunoResearch, West Grove, PA). After 6 washes with TBST, the membranes were developed using Western Lightning chemiluminescence reagent (PerkinElmer, Boston, MA). For protein band densitometry, a MultiImage Light Cabinet (Alpha Innotech, San Leandro, CA) was used to capture images, and spot densitometry was performed using ChemiImager 4400 software (Alpha Innotech). Histologic examination. For histomorphometric analysis, ankle tissues were fixed for 24 hours in 4% paraformaldehyde in PBS and decalcified for 48–72 hours with modified Kristensen’s solution (38). Tissues were then dehydrated, embedded in paraffin, sectioned at 5 m thickness and stained with hematoxylin and eosin. Histologic scoring of inflammation, cartilage erosion, and bone erosion was performed as described previously (39,40). Briefly, inflammation was scored on a scale of 0–5, where 0 ⫽ normal, 1 ⫽ minimal infiltration of inflammatory cells, 2 ⫽ mild infiltration, 3 ⫽ moderate infiltration, 4 ⫽ marked infiltration, and 5 ⫽ severe infiltration. Bone erosion was scored on a scale of 0–5, where 0 ⫽ normal, 1 ⫽ small areas of resorption, 2 ⫽ more numerous areas of resorption, 3 ⫽ obvious resorption of trabecular and cortical bone, 4 ⫽ full-thickness defects in the cortical bone and marked trabecular bone loss, and 5 ⫽ full-thickness defects in the cortical bone and marked trabecular bone loss, with distortion of the profile of the remaining cortical surface. Cartilage erosion was scored on a scale of 0–5, where 0 ⫽ no cartilage injury, 1 ⫽ synovial adherence to margins of cartilage in fewer than 3 sites, 2 ⫽ synovial adherence to margins of cartilage in 3 or more sites, 3 ⫽ synovial adherence to cartilage not limited to margins, but no full-thickness injury (damage does not extend beyond the tidemark), 4 ⫽ full-thickness injury in fewer than 3 sites, and 5 ⫽ full-thickness injury in 3 or more sites. Statistical analysis. Results are presented as the mean ⫾ SEM. The statistical significance for comparisons between groups was determined using Student’s unpaired 2-tailed t-test or two-way analysis of variance, followed by Bonferroni correction using the Prism software package (version 4.00; GraphPad Software, San Diego, CA). P values less than 0.05 were considered significant. RESULTS PGs in joint inflammation induced by K/BxN serum transfer. To understand the potential contribution of prostanoids to K/BxN arthritis, we started by assaying specific PGs in arthritic, chronically inflamed joint tissue from K/BxN mice. As shown in Figure 1, high levels of PGE2 and the PGI2 metabolite 6-keto-PGF1␣ were evident in joint tissues from K/BxN mice with chronic arthritis. We then proceeded to examine the kinetics of prostanoid elevations after arthritis induction via K/BxN serum transfer. There was a close temporal association between the development of joint inflammation (Figure 1C) and the elevation of tissue COX ISOFORMS AND PG SPECIES IN K/BxN SERUM–TRANSFER ARTHRITIS 1357 Figure 1. Prostaglandin levels in joint tissues from mice with K/BxN serum–transfer arthritis. Concentrations of A, prostaglandin E2 (PGE2) and B, 6-keto-PGF1␣ in chronically inflamed joint tissues from 10-week-old K/BxN mice or in joint tissues from wild-type (WT) mice were measured by liquid chromatography mass spectrometry on days 0, 4, and 8 after administration of arthritogenic K/BxN serum (n ⫽ 13 mice per group and per time point). Differences in PGE2 levels were statistically significant at P ⬍ 0.01 for WT mice on day 8 versus day 0 and for K/BxN mice versus WT mice on day 0; differences in 6-keto-PGF1␣ levels were statistically significant at P ⬍ 0.001 for WT mice on day 4 and on day 8 versus day 0 and for K/BxN mice versus WT mice on day 0. C, Clinical index of arthritis after administration of arthritogenic K/BxN serum to WT mice (n ⫽ 13 per time point). Differences were statistically significant at P ⬍ 0.001 on day 4 and on day 8 versus day 0. Values are the mean and SEM of pooled data from 3 individual experiments. P values were determined by Student’s unpaired 2-tailed t-test. PGE2 (Figure 1A) and 6-keto-PGF1␣ (Figure 1B) levels. Thus, the kinetics of prostanoid generation are consistent with the participation of these molecules in joint inflammation. Amelioration of K/BxN serum–transfer arthritis by pharmacologic inhibition of prostaglandin synthesis. Having demonstrated the presence of prostanoids concurrent with joint inflammation, we next examined the functional contribution of this pathway to disease induction and perpetuation. To this end, we administered oral doses of sulindac, a potent inhibitor of both COX isoforms (27,41), or vehicle control to WT mice and induced arthritis via passive transfer of K/BxN serum. As shown in Figure 2, initiation of sulindac prior to arthritis induction substantially prevented joint inflammation in this model (Figure 2A). Furthermore, administration of sulindac to mice with established joint inflammation rapidly reduced clinical signs of arthritis (Figure 2B). These clinical changes corresponded to a clear decrease in inflammation as well as joint injury, as assessed by histologic scoring (Figures 2C and D). As anticipated, PGE2 (Figure 2E) and 6keto-PGF1␣ (Figure 2F) production was inhibited in joint tissues from the mice administered sulindac. Thus, prostanoids contribute to both the initiation and the perpetuation of joint inflammation in this model. Role of COX-1, but not COX-2, as an obligate participant in arthritis. Western blot analysis of joint lysates demonstrated increased levels of COX-1 and COX-2 after disease induction (by day 4); these increased levels were maintained during the course of the disease (Figures 3A and B). The specificity of the COX-1 staining was confirmed by the disappearance of the 70-kd COX-1 band in immunoblots of ankle lysates prepared from COX-1⫺/⫺ mice and following preincubation of the primary antibody with a COX-1–blocking peptide (murine amino acids 274–288; Cayman Chemical) (data not shown). We next used a genetic approach to examine the functional contribution of the COX isoforms. We found that COX-1⫺/⫺ mice were remarkably resistant to the development of K/BxN serum–induced inflammatory arthritis, whereas COX-2⫺/⫺ mice showed no diminu- 1358 CHEN ET AL Figure 2. Prevention and treatment of K/BxN serum–induced arthritis by inhibition of cyclooxygenase. A and B, C57BL/6J mice (n ⫽ 15 per group) were given oral doses of sulindac or vehicle control (1% methylcellulose) beginning on day –2 before administration of arthritogenic K/BxN serum (A) or day 6 after administration of arthritogenic K/BxN serum (B). Differences were statistically significant at P ⬍ 0.001 for pretreated mice versus vehicle-treated controls and at P ⬍ 0.01 for treated mice versus vehicle-treated controls. C and D, Histomorphometric quantification of inflammation, bone erosion, and cartilage erosion was performed on day 14 after K/BxN serum transfer in pretreated mice (C) and treated mice (D). Differences were statistically significant at P ⬍ 0.001 for pretreated mice and treated mice versus vehicle-treated controls. E and F, Levels of prostaglandin E2 (PGE2) (E) and 6-keto-PGF1␣ (F) were measured in joint tissues from pretreated and treated mice. Differences in PGE2 and 6-keto-PGF1␣ levels were statistically significant at P ⬍ 0.001 for pretreated mice and for treated mice versus vehicle-treated controls. Values are the mean ⫾ SEM of pooled data from 3 experiments. P values in A and B were determined by two-way analysis of variance; those in C–F were determined by Student’s unpaired 2-tailed t-test. tion in disease activity (Figures 3C and D). We measured tissue levels of PGE2 (Figure 3E) and 6-ketoPGF1␣ (Figure 3F) to confirm that COX-1⫺/⫺ mice (with intact COX-2) lacked elevations in tissue prostanoids, and we found levels consistent with those in healthy joint tissues. Histomorphometric examination of inflammation, bone erosion, and cartilage erosion in joint tissues from these mice confirmed the clinical findings (Figures 3G–L). Whereas WT and COX-2⫺/⫺ mice demonstrated synovial hyperplasia, leukocytic infiltration, and the presence of synovial erosion into bone and cartilage (Figures 3G, J, and K), COX-1⫺/⫺ joint tissues retained a normal appearance, with little evi- dence of these inflammatory changes (Figure 3H). These data implicate COX-1 in the development of arthritis after K/BxN serum transfer, whereas COX-2 appears to be dispensable. Effects of COX-1 and COX-2 inhibitors on K/BxN serum–induced arthritis. Given the divergence of our findings with those documented in other arthritis models, we used COX isoform–specific pharmacologic inhibition to confirm the substantial COX-1 contribution to arthritis induction as well as to examine the role of these isoforms in the perpetuation of K/BxN serum– transfer arthritis. Indeed, we found that administration of SC-560, a highly selective oral inhibitor of COX-1 (28), both prevented the development of disease when COX ISOFORMS AND PG SPECIES IN K/BxN SERUM–TRANSFER ARTHRITIS Figure 3. Expression and function of cyclooxygenase (COX) isoforms in K/BxN serum–transfer arthritis. A and B, Levels of COX-1 and COX-2 protein in mouse ankle joints. C57BL/6J mice (n ⫽ 5 mice per time point) were administered arthritogenic K/BxN serum, and ankle tissues were harvested at the indicated times. Western blots of COX-1, COX-2, and GAPDH protein in ankle lysates (A) (representative of 5 independent experiments) and densitometric quantification of the ratio of each COX protein to GAPDH (B) were performed. Expression levels were normalized relative to day 0, which was arbitrarily assigned a ratio of 1. Levels of both COX isoforms increased after K/BxN serum transfer. C and D, Clinical index of arthritis in COX-1⫺/⫺ (C) and COX-2⫺/⫺ (D) mice and their wild-type (WT) controls (n ⫽ 10–12 mice per group). Differences were significant at P ⬍ 0.001 for COX-1⫺/⫺ mice versus WT mice. E and F, Production of prostaglandin E2 (PGE2) (E) and 6-keto-PGF1␣ (F) in joint tissues from COX-1⫺/⫺ mice 2 weeks after K/BxN serum transfer. Differences in PGE2 and 6-keto-PGF1␣ levels were statistically significant at P ⬍ 0.001 for COX-1⫺/⫺ mice versus WT mice. G, H, J, and K, Histologic features of arthritis on day 14 after K/BxN serum transfer in COX-1 WT (G), COX-1⫺/⫺ (H), COX-2 WT (J), and COX-2⫺/⫺ (K) mice. Ca ⫽ cartilage; Bn ⫽ Bone; S ⫽ synovium. Bar ⫽ 100 m. I and L, Histomorphometric quantification of arthritis on day 14 after K/BxN serum transfer in COX-1⫺/⫺ (I) and COX-2⫺/⫺ (L) mice and their WT controls. Differences were significant at P ⬍ 0.001 for COX-1⫺/⫺ mice versus WT mice. Values in B–F, I, and L are the mean ⫾ SEM of pooled data from 3 experiments. P values in C and D were determined by two-way analysis of variance; those in E, F, I, and L were determined by Student’s unpaired 2-tailed t-test. 1359 1360 CHEN ET AL Figure 4. Effects of selective pharmacologic inhibition of cyclooxygenase (COX) on K/BxN serum–induced arthritis. A–D, C57BL/6J mice (n ⫽ 15 per group) were given oral doses of the COX-1 inhibitor SC-560 (10 mg/kg) (A and B) or the COX-2 inhibitor MF-tricyclic (MFT; 3 mg/kg) (C and D) beginning on day –2 before administration of arthritogenic K/BxN serum (A and C) or day 6 after administration of arthritogenic K/BxN serum (B and D), and the clinical index of arthritis was monitored. Differences were statistically significant at P ⬍ 0.001 for SC-560– pretreated mice versus vehicle-treated controls and at P ⬍ 0.01 for SC-560–treated mice versus vehicle-treated controls. Values are the mean ⫾ SEM. E and F, Ex vivo effect of a single 10-mg/kg oral dose of SC-560 or a single 3-mg/kg oral dose of MFT on COX-1 (E) and COX-2 (F) activity in whole blood, as determined by measuring serum levels of thromboxane B2 (TXB2) (E) in coagulated blood and plasma levels of prostaglandin E2 (PGE2) (F) in lipopolysaccharide-treated blood from C57BL/6J mice (n ⫽ 10 per group) at 2 hours after oral dosing. Values are the mean and SEM of pooled data from 2 experiments. Differences in TXB2 levels were statistically significant at P ⬍ 0.001 for SC-560–treated mice versus vehicle-treated controls; differences in PGE2 levels were statistically significant at P ⬍ 0.001 for MFT-treated mice versus vehicle-treated controls. P values in A–D were determined by two-way analysis of variance; those in E and F were determined by Student’s unpaired 2-tailed t-test. administered prior to arthritogenic serum and rapidly diminished the clinical signs of arthritis when used to treat established disease (Figures 4A and B). In contrast, administration of the selective COX-2 inhibitor MFT (26,42) resulted in no discernible decrease in arthritis activity when used either in the prevention or the treatment of disease at doses that fully inhibited LPSinduced PGE2 in peripheral blood leukocytes (Figures 4C, D, and F). Substantial contribution of PGI2, but not PGE2, to K/BxN serum–induced arthritis. Since our studies demonstrated increasing levels of PGE2 and 6-ketoPGF1␣ in arthritic joint tissues and the lack of PGE2 and 6-keto-PGF1␣ production in COX-1⫺/⫺ mice, we explored the contribution of the prostanoid species PGE2 and PGI2 in K/BxN serum–induced arthritis. We used mPGES-1–deficient mice (15) to assess the involvement of PGE2. The mPGES-1⫺/⫺ mice developed arthri- COX ISOFORMS AND PG SPECIES IN K/BxN SERUM–TRANSFER ARTHRITIS 1361 Figure 5. Dispensability of prostaglandin E2 (PGE2) in K/BxN serum–induced arthritis. A, Clinical index of arthritis in mPGES-1⫺/⫺ and wild-type (WT) control mice after administration of arthritogenic K/BxN serum (n ⫽ 15 mice per group). B and C, Concentration of PGE2 (B) and 6-keto-PGF1␣ (C) in arthritic joint tissues from mPGES-1⫺/⫺ and WT control mice (n ⫽ 10 per group). Differences in PGE2 levels were significant at P ⬍ 0.001 and differences in 6-keto-PGF1␣ levels were significant at P ⬍ 0.05 for mPGES-1⫺/⫺ mice versus WT mice. D and E, Histologic features of ankle tissues from WT (D) and mPGES-1⫺/⫺ (E) mice on day 14 after K/BxN serum transfer. Ca ⫽ cartilage; Bn ⫽ bone; S ⫽ synovium. Bar ⫽ 100 m. F, Histomorphometric quantification of arthritis on day 14 after K/BxN serum transfer in mPGES-1⫺/⫺ and WT control mice. Values in A–C and F are the mean ⫾ SEM of pooled data from 3 experiments. P values were determined by Student’s unpaired 2-tailed t-test. tis that was clinically and histologically indistinguishable from that in WT controls (Figures 5A, D, E, and F). Since previous studies demonstrated the presence of at least 3 distinct PGE2 synthases (43,44), we quantified joint tissue levels of PGE2 and 6-keto-PGF1␣ in mPGES-1⫺/⫺ mice and confirmed a lack of significant PGE2 production, while the 6-keto-PGF1␣ levels were intact (Figures 5B and C). Thus, while mPGES-1 remains the primary source of synovial PGE2 production in our experimental system, PGE2 itself appears to be dispensable for the initiation and perpetuation of arthritis in the K/BxN model. Having found no discernible contribution from PGE2, we examined the role of PGI2 by using mice deficient in the PGI2 receptor IP. In these experiments, we found a significant, albeit partial, decrease in clinical arthritis in IP⫺/⫺ mice, to 31% of the level in WT mice (Figure 6A). Histomorphometric analyses again confirmed the clinical findings, with reductions in mean scores for inflammation, bone erosion, and cartilage erosion of 59%, 61%, and 77%, respectively, in IP⫺/⫺ mice (Figure 6B). Since high concentrations of PGI2 may activate the EP1 receptor (45), we assessed a potential in vivo contribution from PGI2 via EP1 by examining arthritis in EP1-deficient mice. As shown in Figures 6C and D, we find no amelioration of clinical or histologic arthritis activity in this strain. These data indicate that a substantial proportion of the prostanoid contribution to joint inflammation in K/BxN serum– transfer arthritis can be accounted for by an interaction of PGI2 and its receptor IP. DISCUSSION Our studies revealed a striking contribution of prostaglandins to autoantibody-driven joint inflammation in the K/BxN serum–transfer model. Perhaps most surprising was the apparent reliance on the COX-1 1362 CHEN ET AL Figure 6. Contribution of prostaglandin I2 (PGI2) to K/BxN serum–induced arthritis. A, Clinical index of arthritis in IP⫺/⫺ and wild-type (WT) mice after administration of arthritogenic K/BxN serum. B, Histomorphometric quantification of arthritis on day 14 after K/BxN serum transfer in IP⫺/⫺ and WT mice (n ⫽ 15 per group). Differences were significant at P ⬍ 0.001 for IP⫺/⫺ mice versus WT mice. C, Clinical index of arthritis in EP1⫺/⫺ and WT mice after establishment of arthritis. D, Histomorphometric quantification of arthritis on day 14 after K/BxN serum transfer in EP1⫺/⫺ and WT mice (n ⫽ 12–15 per group). Values are the mean ⫾ SEM of pooled data from 3 experiments. P values in A and C were determined by two-way analysis of variance; those in B and D were determined by Student’s unpaired 2-tailed t-test. isoform and the dispensability of COX-2 for both the initiation and the perpetuation of arthritis. This conclusion is based on our findings from experiments in genetically deficient animals as well as through isoformspecific pharmacologic inhibition in normal mice. While our results do not exclude a contribution from COX-2, we showed that COX-1 is an essential and sufficient source of arthritogenic prostaglandins in both normal and inflamed murine joints. Our findings thus provide a counterpoint to the paradigm that COX-1 contributes only to “housekeeping” prostaglandin synthesis, while prostaglandins generated under inflammatory conditions reflect the activity of the highly inducible COX-2 isoform. Furthermore, we found that a significant proportion of the arthritogenic activity of prostaglandins in K/BxN serum–transfer arthritis may be attributed to the action of PGI2 (prostacyclin) via its receptor, the IP receptor. Potent proinflammatory activities of prostacyclin have only recently been reported in other models of arthritis (16,18), and its role has been explored in a limited number of other disease states (46). The demonstrated functional contribution in K/BxN arthritis provides added rationale for further examination of the predominant synthetic sources of PGI2 as well as the proinflammatory effector functions elicited in target cells in the joint. Furthermore, our observations raise the possibility that suppression of COX-1–derived PGI2 may contribute to antiinflammatory efficacy independently of COX-2–derived PGI2, whose suppression has been associated with cardiovascular hazard in human trials (47,48). Indeed, sparing COX-2–dependent prostacyclin production may confer cardiovascular protection to arthritis patients, a population with demonstrated increased cardiovascular risk. These results add yet another pathway to a surprisingly large number of effector mechanisms required for the full expression of arthritis in the K/BxN serum–transfer model. Thus far, lack of intact Fc receptor signaling (Fc␥RIII), complement anaphylatoxins (via C5a receptor [CD88]), neutrophils, mast cells, natural killer T cells, and the mediators interleukin-1 (IL-1; via COX ISOFORMS AND PG SPECIES IN K/BxN SERUM–TRANSFER ARTHRITIS IL-1 receptor type I), leukotriene B4 (LTB4), and prostaglandins all confer dramatic resistance to arthritis mediated by passive transfer of arthritogenic autoantibodies (35,39,49–54). A major unresolved question that arises from these observations is whether these pathways operate sequentially or in parallel in a codominant manner. Also unclear are the interactions between these pathways at the level of cellular effector responses. Further insight into these regulatory events will clarify whether autoantibody-driven arthritis proceeds as a linear series of events or whether a network of parallel inflammatory pathways conspire to propagate disease. In either case, further understanding will inform our views regarding pathogenic processes in human inflammatory arthritis. The dispensability of PGE2 and its synthetic enzyme in K/BxN arthritis was unexpected, given the substantial levels of this species in arthritic joint tissues and the previous studies documenting PGE2-stimulated effector functions on cellular populations represented within the arthritic joint (55–57). As with COX isoform dependence (13,14), these results differ from those of studies using the CIA or CAIA models (15,17,18) and represent a rare divergence in effector mechanisms between these arthritis models (58). Anti-GPI and anticollagen models share a common requirement for cytokines (IL-1 and tumor necrosis factor), IgG Fc receptors, complement, and eicosanoids. The eicosanoid requirement in both models includes intact cytosolic phospholipase A2 (59,60), the capacity to synthesize LTB4 (49,50), and an important, albeit partial, reliance on PGI2 (61,62). While the mechanistic basis for this divergence remains undefined, 2 possibilities deserve mention. The first is the potential contribution of LPS or adjuvant used in CIA and CAIA. These agents are known to stimulate the production of prostaglandins and induce the expression of COX-2 (15,63–65). Thus, it is possible that these stimuli contribute to the dominance of COX-2 in those experimental systems. Indeed, we find that the administration of LPS enables the initiation of K/BxN arthritis in COX-1⫺/⫺ animals, though whether this effect operates via COX-2 is unknown (Chen M, Lee D: unpublished observations). A second consideration is the impact of COX isoform deficiency on T cell and B cell function. Modulation of adaptive immune responses by prostanoids has been demonstrated in multiple experimental systems (66–68), and COX-2–null mice with CIA had markedly decreased levels of anticollagen antibodies (14). Thus, 1363 whether COX-2 interruption impedes the development of a lymphocyte-dependent anticollagen response, interferes with synovial inflammatory networks, or both, remains unclarified. K/BxN serum transfer proceeds in the absence of lymphocytes (20), thereby affording a focus on effector-phase arthritis mechanisms that should not be impacted by a potential COX-2 contribution to lymphocyte function. These issues notwithstanding, the findings in these arthritis models demonstrate that inflammatory arthritis can proceed via disparate pathways of prostaglandin synthesis and that multiple prostanoid species may contribute to the final common pathway of synovitis. Finally, our findings provide a rationale to reexamine the contribution of COX-1 and specific downstream prostaglandins to inflammatory arthritis in humans. While nonsteroidal antiinflammatory drugs (NSAIDs) have been used frequently for amelioration of symptoms, these agents have not been demonstrated to have disease-modifying activity. However, the extent to which prostaglandin synthesis within the RA joint is blocked by these agents has never, to our knowledge, been studied. Therefore, the magnitude of the pathogenic contribution of prostaglandins to human RA remains an open question. This consideration is particularly relevant for the COX-2–selective NSAIDs, where clinical development was focused on gastrointestinal safety and equivalence of pain relief. Indeed, both COX isozymes are expressed in the rheumatoid synovium (69,70). Our data suggest that COX-2–selective agents, even if tolerated at higher dosages because of reduced gastrointestinal toxicity, could miss important prostanoid pathways in the inflamed joint. The recent demonstration of cardiovascular toxicity from treatment with COX-2–selective NSAIDs has introduced a further limitation on the use of these drugs (71). Thus, the identification of predominant roles for COX-1 and discrete prostaglandin species by our group and others suggests that antagonism of specific prostaglandin species may be an effective therapeutic strategy if it can be achieved without the limiting toxicity that has thus far plagued the use of such inhibitors clinically. ACKNOWLEDGMENTS We are grateful to Dr. B. H. Koller (University of North Carolina, Chapel Hill, NC) for providing the mPGES1–null mice and the EP1-deficient mice (derived and maintained with support from NIH grant HL-068141 to Dr. Koller). We also acknowledge the expert histotechnical assistance of Teresa Bowman. 1364 CHEN ET AL AUTHOR CONTRIBUTIONS Dr. Lee had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. Study design. Chen, Audoly, Lee. Acquisition of data. Chen, Boilard, Clark, Xu, Lee. Analysis and interpretation of data. Chen, Boilard, Nigrovic, Lee. Manuscript preparation. Chen, Boilard, Nigrovic, FitzGerald, Audoly, Lee. Statistical analysis. Chen, Lee. REFERENCES 1. Narumiya S, Sugimoto Y, Ushikubi F. Prostanoid receptors: structures, properties, and functions. Physiol Rev 1999;79: 1193–226. 2. Morita I. Distinct functions of COX-1 and COX-2. Prostaglandins Other Lipid Mediat 2002;68-69:165–75. 3. Dubois RN, Abramson SB, Crofford L, Gupta RA, Simon LS, Van De Putte LB, et al. Cyclooxygenase in biology and disease. FASEB J 1998;12:1063–73. 4. Peskar BM, Maricic N, Gretzera B, Schuligoi R, Schmassmann A. Role of cyclooxygenase-2 in gastric mucosal defense. Life Sci 2001;69:2993–3003. 5. Harris RC, Breyer MD. Physiological regulation of cyclooxygenase-2 in the kidney. Am J Physiol Renal Physiol 2001;281:F1–11. 6. Rouzer CA, Kingsley PJ, Wang H, Zhang H, Morrow JD, Dey SK, et al. Cyclooxygenase-1-dependent prostaglandin synthesis modulates tumor necrosis factor-␣ secretion in lipopolysaccharidechallenged murine resident peritoneal macrophages. J Biol Chem 2004;279:34256–68. 7. Langenbach R, Morham SG, Tiano HF, Loftin CD, Ghanayem BI, Chulada PC, et al. Prostaglandin synthase 1 gene disruption in mice reduces arachidonic acid-induced inflammation and indomethacin-induced gastric ulceration. Cell 1995;83:483–92. 8. Gavett SH, Madison SL, Chulada PC, Scarborough PE, Qu W, Boyle JE, et al. Allergic lung responses are increased in prostaglandin H synthase-deficient mice. J Clin Invest 1999;104:721–32. 9. Langenbach R, Loftin CD, Lee C, Tiano H. Cyclooxygenasedeficient mice: a summary of their characteristics and susceptibilities to inflammation and carcinogenesis. Ann N Y Acad Sci 1999;889:52–61. 10. Loftin CD, Tiano HF, Langenbach R. Phenotypes of the COXdeficient mice indicate physiological and pathophysiological roles for COX-1 and COX-2. Prostaglandins Other Lipid Mediat 2002; 68-69:177–85. 11. Egg D, Gunther R, Herold M, Kerschbaumer F. Prostaglandins E2 and F2␣ concentrations in the synovial fluid in rheumatoid and traumatic knee joint diseases. Z Rheumatol 1980;39:170–5. In German. 12. Brodie MJ, Hensby CN, Parke A, Gordon D. Is prostacyclin the major pro-inflammatory prostanoid in joint fluid? Life Sci 1980; 27:603–8. 13. Ochi T, Ohkubo Y, Mutoh S. Role of cyclooxygenase-2, but not cyclooxygenase-1, on type II collagen-induced arthritis in DBA/1J mice. Biochem Pharmacol 2003;66:1055–60. 14. Myers LK, Kang AH, Postlethwaite AE, Rosloniec EF, Morham SG, Shlopov BV, et al. The genetic ablation of cyclooxygenase 2 prevents the development of autoimmune arthritis. Arthritis Rheum 2000;43:2687–93. 15. Trebino CE, Stock JL, Gibbons CP, Naiman BM, Wachtmann TS, Umland JP, et al. Impaired inflammatory and pain responses in mice lacking an inducible prostaglandin E synthase. Proc Natl Acad Sci U S A 2003;100:9044–9. 16. Honda T, Segi-Nishida E, Miyachi Y, Narumiya S. Prostacyclin-IP 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. signaling and prostaglandin E2-EP2/EP4 signaling both mediate joint inflammation in mouse collagen-induced arthritis. J Exp Med 2006;203:325–35. McCoy JM, Wicks JR, Audoly LP. The role of prostaglandin E2 receptors in the pathogenesis of rheumatoid arthritis. J Clin Invest 2002;110:651–8. Pulichino AM, Rowland S, Wu T, Clark P, Xu D, Mathieu MC, et al. Prostacyclin antagonism reduces pain and inflammation in rodent models of hyperalgesia and chronic arthritis. J Pharmacol Exp Ther 2006;319:1043–50. Kouskoff V, Korganow AS, Duchatelle V, Degott C, Benoist C, Mathis D. Organ-specific disease provoked by systemic autoimmunity. Cell 1996;87:811–22. Korganow AS, Ji H, Mangialaio S, Duchatelle V, Pelanda R, Martin T, et al. From systemic T cell self-reactivity to organspecific autoimmune disease via immunoglobulins. Immunity 1999;10:451–61. Matsumoto I, Maccioni M, Lee DM, Maurice M, Simmons B, Brenner M, et al. How antibodies to a ubiquitous cytoplasmic enzyme may provoke joint-specific autoimmune disease. Nat Immunol 2002;3:360–5. Matsumoto I, Staub A, Benoist C, Mathis D. Arthritis provoked by linked T and B cell recognition of a glycolytic enzyme. Science 1999;286:1732–5. Morham SG, Langenbach R, Loftin CD, Tiano HF, Vouloumanos N, Jennette JC, et al. Prostaglandin synthase 2 gene disruption causes severe renal pathology in the mouse. Cell 1995;83:473–82. Cheng Y, Austin SC, Rocca B, Koller BH, Coffman TM, Grosser T, et al. Role of prostacyclin in the cardiovascular response to thromboxane A2. Science 2002;296:539–41. Nataraj C, Thomas DW, Tilley SL, Nguyen MT, Mannon R, Koller BH, et al. Receptors for prostaglandin E2 that regulate cellular immune responses in the mouse. J Clin Invest 2001;108:1229–35. Black SC, Brideau C, Cirino M, Belley M, Bosquet J, Chan CC, et al. Differential effect of a selective cyclooxygenase-2 inhibitor versus indomethacin on renal blood flow in conscious volumedepleted dogs. J Cardiovasc Pharmacol 1998;32:686–94. Duggan DE, Hooke KF, Risley EA, Shen TY, Arman CG. Identification of the biologically active form of sulindac. J Pharmacol Exp Ther 1977;201:8–13. Smith CJ, Zhang Y, Koboldt CM, Muhammad J, Zweifel BS, Shaffer A, et al. Pharmacological analysis of cyclooxygenase-1 in inflammation. Proc Natl Acad Sci U S A 1998;95:13313–8. Rowland SE, Clark P, Gordon R, Mullen AK, Guay J, Dufresne L, et al. Pharmacological characterization of a selective COX-2 inhibitor MF-tricyclic, [3-(3,4-difluorophenyl)-4-(4-(methylsulfonyl) phenyl)-2-(5H)-furanone], in multiple preclinical species. Eur J Pharmacol 2007;560:216–24. Ellis EF, Oelz O, Roberts LJ II, Payne NA, Sweetman BJ, Nies AS, et al. Coronary arterial smooth muscle contraction by a substance released from platelets: evidence that it is thromboxane A2. Science 1976;193:1135–7. Patrono C, Ciabattoni G, Pugliese F, Pierucci A, Blair IA, FitzGerald GA. Estimated rate of thromboxane secretion into the circulation of normal humans. J Clin Invest 1986;77:590–4. Granstrom E, Hamberg M, Hansson G, Kindahl H. Chemical instability of 15-keto-13,14-dihydro-PGE2: the reason for low assay reliability. Prostaglandins 1980;19:933–57. Fitzpatrick FA, Aguirre R, Pike JE, Lincoln FH. The stability of 13,14-dihydro-15 keto-PGE2. Prostaglandins 1980;19:917–31. Brideau C, Kargman S, Liu S, Dallob AL, Ehrich EW, Rodger IW, et al. A human whole blood assay for clinical evaluation of biochemical efficacy of cyclooxygenase inhibitors. Inflamm Res 1996;45:68–74. Lee DM, Friend DS, Gurish MF, Benoist C, Mathis D, Brenner COX ISOFORMS AND PG SPECIES IN K/BxN SERUM–TRANSFER ARTHRITIS 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. MB. Mast cells: a cellular link between autoantibodies and inflammatory arthritis. Science 2002;297:1689–92. Guay J, Bateman K, Gordon R, Mancini J, Riendeau D. Carrageenan-induced paw edema in rat elicits a predominant prostaglandin E2 (PGE2) response in the central nervous system associated with the induction of microsomal PGE2 synthase-1. J Biol Chem 2004;279:24866–72. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970;227:680–5. Humason GL. Animal tissue techniques. 2nd ed. San Francisco: WH Freeman; 1967. Chen M, Lam BK, Kanaoka Y, Nigrovic PA, Audoly LP, Austen KF, et al. Neutrophil-derived leukotriene B4 is required for inflammatory arthritis. J Exp Med 2006;203:837–42. Pettit AR, Ji H, von Stechow D, Muller R, Goldring SR, Choi Y, et al. TRANCE/RANKL knockout mice are protected from bone erosion in a serum transfer model of arthritis. Am J Pathol 2001;159:1689–99. Shen TY, Winter CA. Chemical and biological studies on indomethacin, sulindac and their analogs. Adv Drug Res 1977;12: 90–245. Oshima M, Dinchuk JE, Kargman SL, Oshima H, Hancock B, Kwong E, et al. Suppression of intestinal polyposis in Apc⌬716 knockout mice by inhibition of cyclooxygenase 2 (COX-2). Cell 1996;87:803–9. Park JY, Pillinger MH, Abramson SB. Prostaglandin E2 synthesis and secretion: the role of PGE2 synthases. Clin Immunol 2006; 119:229–40. Kudo I, Murakami M. Prostaglandin E synthase, a terminal enzyme for prostaglandin E2 biosynthesis. J Biochem Mol Biol 2005;38:633–8. Walsh MT, Kinsella BT. Regulation of the human prostanoid TP␣ and TP␤ receptor isoforms mediated through activation of the EP1 and IP receptors. Br J Pharmacol 2000;131:601–9. Lovgren AK, Jania LA, Hartney JM, Parsons KK, Audoly LP, Fitzgerald GA, et al. COX-2-derived prostacyclin protects against bleomycin-induced pulmonary fibrosis. Am J Physiol Lung Cell Mol Physiol 2006;291:L144–56. FitzGerald GA, Patrono C. The coxibs, selective inhibitors of cyclooxygenase-2. N Engl J Med 2001;345:433–42. Grosser T, Fries S, FitzGerald GA. Biological basis for the cardiovascular consequences of COX-2 inhibition: therapeutic challenges and opportunities. J Clin Invest 2006;116:4–15. Bauldry SA, Wooten RE. Leukotriene B4 and platelet activating factor production in permeabilized human neutrophils: role of cytosolic PLA2 in LTB4 and PAF generation. Biochim Biophys Acta 1996;1303:63–73. Liberty IF, Raichel L, Hazan-Eitan Z, Pessach I, Hadad N, Schlaeffer F, et al. Cytosolic phospholipase A2 is responsible for prostaglandin E2 and leukotriene B4 formation in phagocyte-like PLB-985 cells: studies of differentiated cPLA2-deficient PLB-985 cells. J Leukoc Biol 2004;76:176–84. Ji H, Ohmura K, Mahmood U, Lee DM, Hofhuis FM, Boackle SA, et al. Arthritis critically dependent on innate immune system players. Immunity 2002;16:157–68. Feldmann M, Brennan FM, Maini RN. Role of cytokines in rheumatoid arthritis. Annu Rev Immunol 1996;14:397–440. Corr M, Crain B. The role of Fc␥R signaling in the K/B ⫻ N serum transfer model of arthritis. J Immunol 2002;169:6604–9. Kim ND, Chou RC, Seung E, Tager AM, Luster AD. A unique requirement for the leukotriene B4 receptor BLT1 for neutrophil 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 1365 recruitment in inflammatory arthritis. J Exp Med 2006;203: 829–35. Sampey AV, Monrad S, Crofford LJ. Microsomal prostaglandin E synthase-1: the inducible synthase for prostaglandin E2. Arthritis Res Ther 2005;7:114–7. Claveau D, Sirinyan M, Guay J, Gordon R, Chan CC, Bureau Y, et al. Microsomal prostaglandin E synthase-1 is a major terminal synthase that is selectively up-regulated during cyclooxygenase-2dependent prostaglandin E2 production in the rat adjuvantinduced arthritis model. J Immunol 2003;170:4738–44. Fahmi H. mPGES-1 as a novel target for arthritis. Curr Opin Rheumatol 2004;16:623–7. Monach PA, Benoist C, Mathis D. The role of antibodies in mouse models of rheumatoid arthritis, and relevance to human disease. Adv Immunol 2004;82:217–48. Hegen M, Sun L, Uozumi N, Kume K, Goad ME, NickersonNutter CL, et al. Cytosolic phospholipase A2␣-deficient mice are resistant to collagen-induced arthritis. J Exp Med 2003;197: 1297–302. Hegen M, Lamothe J, Shen MW, Stedman N, Leach MW, Collins M, et al. Cytosolic phospholipase A2␣-deficient mice are resistant to arthritis in the K/BxN serum transfer model [abstract]. Arthritis Rheum 2006;54 Suppl 9:S177. Merhi-Soussi F, Dominguez Z, Macovschi O, Dubois M, Savany A, Lagarde M, et al. Human lymphocytes stimulate prostacyclin synthesis in human umbilical vein endothelial cells: involvement of endothelial cPLA2. J Leukoc Biol 2000;68:881–9. Kan H, Ruan Y, Malik KU. Involvement of mitogen-activated protein kinase and translocation of cytosolic phospholipase A2 to the nuclear envelope in acetylcholine-induced prostacyclin synthesis in rabbit coronary endothelial cells. Mol Pharmacol 1996;50: 1139–47. Young MR, Newby M. Enhancement of Lewis lung carcinoma cell migration by prostaglandin E2 produced by macrophages. Cancer Res 1986;46:160–4. Reddy ST, Herschman HR. Ligand-induced prostaglandin synthesis requires expression of the TIS10/PGS-2 prostaglandin synthase gene in murine fibroblasts and macrophages. J Biol Chem 1994; 269:15473–80. Ikegami R, Sugimoto Y, Segi E, Katsuyama M, Karahashi H, Amano F, et al. The expression of prostaglandin E receptors EP2 and EP4 and their different regulation by lipopolysaccharide in C3H/HeN peritoneal macrophages. J Immunol 2001;166:4689–96. Chouaib S, Welte K, Mertelsmann R, Dupont B. Prostaglandin E2 acts at two distinct pathways of T lymphocyte activation: inhibition of interleukin 2 production and down-regulation of transferrin receptor expression. J Immunol 1985;135:1172–9. Harizi H, Grosset C, Gualde N. Prostaglandin E2 modulates dendritic cell function via EP2 and EP4 receptor subtypes. J Leukoc Biol 2003;73:756–63. Rocca B, Spain LM, Pure E, Langenbach R, Patrono C, FitzGerald GA. Distinct roles of prostaglandin H synthases 1 and 2 in T-cell development. J Clin Invest 1999;103:1469–77. Schonbeck U, Sukhova GK, Graber P, Coulter S, Libby P. Augmented expression of cyclooxygenase-2 in human atherosclerotic lesions. Am J Pathol 1999;155:1281–91. Crofford LJ, Wilder RL, Ristimaki AP, Sano H, Remmers EF, Epps HR, et al. Cyclooxygenase-1 and -2 expression in rheumatoid synovial tissues: effects of interleukin-1␤, phorbol ester, and corticosteroids. J Clin Invest 1994;93:1095–101. Psaty BM, Furberg CD. COX-2 inhibitors—lessons in drug safety. N Engl J Med 2005;352:1133–5.