Biological therapies such as tumor necrosis factor-α inhibitors have advanced the treatment of rheumatoid arthritis, but one-third of patients do not respond to such therapy. Furthermore, these inhibitors are now usually administered in combination with conventional disease-modifying antirheumatic drugs, suggesting they have not achieved their early promise. This study investigates a novel therapeutic target, proteinase-activated receptor (PAR)-2, in joint inflammation. Intra-articular carrageenan/kaolin (C/K) injection in mice resulted in joint swelling that was associated with synovial PAR2 up-regulation. Inhibiting receptor up-regulation using small interfering RNA technology, as confirmed by immunoblotting, substantially reduced the inflammatory response in the joint. Serine proteinase-induced joint swelling was mediated primarily via PAR2 activation, since the response to exogenous application of trypsin and tryptase was absent in PAR2 knockout mice. Furthermore, serine proteinase inhibitors were effective anti-inflammatory agents in this model. Disrupting proteolytic activation of PAR2 using antiserum (B5) directed to the receptor cleavage/activation site also attenuated C/K-induced inflammation, as did the similarly targeted PAR2 monoclonal antibody SAM-11. Finally, we report the activity of a novel small molecule PAR2 antagonist, N1-3-methylbutyryl-N4-6-aminohexanoyl-piperazine (ENMD-1068), that dose dependently attenuated joint inflammation. Our findings represent a major advance in collectively identifying PAR2 as a novel target for the future treatment of arthritis.
Proteinase-activated receptor (PAR)-2 is one of a unique subfamily of G protein-coupled receptors originally cloned from murine and human sources (Nystedt et al., 1994, 1995; Bohm et al., 1996). It is activated by a novel mechanism involving the proteolysis of the receptor N terminus to expose a “tethered ligand” sequence. This proteolytically revealed N-terminal tethered ligand then binds to and activates the receptor, resulting in signaling and internalization of the receptor complex (for reviews, see Macfarlane et al., 2001; Hollenberg and Compton, 2002; Ossovskaya and Bunnett, 2004). Four members of the PAR family, PAR1, PAR2, PAR3, and PAR4, have been cloned and characterized to date (Macfarlane et al., 2001; Hollenberg and Compton, 2002; Ossovskaya and Bunnett, 2004). Unlike other members of the PAR family, where the primary activating proteinase is thrombin, PAR2 is preferentially activated by trypsin and related serine proteinases, including mast cell tryptase (Nystedt et al., 1994; Corvera et al., 1997; Molino et al., 1997). In addition to this cleavage mechanism, PAR2 can also be activated selectively by the application of synthetic peptides modeled on the sequence of the tethered ligand such as the human sequence SLIGKV-NH2, and the more potent murine variant, SLIGRL-NH2 (Blackhart et al., 1996; Hollenberg et al., 1997; Al-Ani et al., 1999; Ferrell et al., 2003).
PAR2 has been implicated in inflammatory responses, and its actions include increased vascular permeability (Kawabata et al., 1998), leukocyte infiltration (Vergnolle et al., 1999), and smooth muscle relaxation (Al-Ani et al., 1995). There has also been increasing interest in PAR2 as a mediator of nociception (Vergnolle et al., 2001a) and neurogenic inflammation (Steinhoff et al., 2000), with the latter study demonstrating coexpression of PAR2 on sensory nerves along with substance P and calcitonin gene-related peptide. Because the edema formation induced by SLIGRL-NH2 was blocked by neuropeptide antagonists, or by depleting sensory nerve endings with capsaicin, PAR2 was concluded to mediate neurogenic inflammation via release of sensory neuropeptides. Inflammation in the joint has also previously been reported to involve a neurogenic component mediated by sensory neuropeptide release (Lam and Ferrell, 1991), although it is not known whether this process involves PAR2. We recently demonstrated that administration of specific PAR2 agonists induced joint swelling and hyperemia, both cardinal signs of inflammation (Ferrell et al., 2003). Moreover, chronic joint inflammation is attenuated in PAR2 knockout mice (Ferrell et al., 2003), arguing for a critical proinflammatory role for this receptor in the joint. From a clinical perspective, the key question to be addressed in this study is whether PAR2 presents a new therapeutic target for the treatment of arthritis. An acute model of arthritis was used to establish proof of concept, assessing various approaches of preventing PAR2 activation from genomic intervention to use of a novel PAR2 antagonist (Fig. 1).
Materials and Methods
Animals. Wild-type (PAR2+/+) C57BL/6J and PAR2-deficient (PAR2–/–) mice (Ferrell et al., 2003), body weight 25 to 30 g, were used. All procedures were performed in accordance with United Kingdom regulations.
Joint Inflammation. To establish proof of principle by multiple therapeutic interventions, an acute model of inflammation was used. Knee joint diameter was measured using calipers (Kroeplin GmbH, Schlüchtern, Germany) before and after inflammation was induced by 20 μl of intra-articular injection of 2% λ-carrageenan and 4% kaolin (C/K; Sigma Chemical, Poole, Dorset, UK) in saline under anesthesia (O2, N2O, 2% halothane; withdrawal reflex absent). Measurements were reproducible with coefficient of variation ≤2%.
PAR2 Antagonist Characterization. The PAR2 antagonist ENMD-1068 was designed as a disubstituted piperazine whose side chains were based upon the results of peptide antagonist screening. The antagonist activity of ENMD-1068 was characterized using a calcium mobilization assay (Kawabata et al., 1999). Cells were loaded with 5 μM Fluo-4 in Hanks' balanced salt solution assay buffer (20 mM HEPES pH 7.4, 1 mM MgSO4, 1 mM CaCl2, and 2.5 mM probenecid) for 40 min. After rinsing, cells were incubated with increasing concentrations of ENMD-1068 for 30 to 40 min before addition of PAR agonists or PAR2-activating proteinase. The antagonist was assessed using both human (SLIGKV-NH2) and murine (SLIGRL-NH2) agonist peptides.
Western Blotting. Protein extracts derived from 100 ng of tissue were prepared from inflamed (C/K) and control synovium using standard procedures, separated by SDS-polyacrylamide gel electrophoresis, and transferred onto nitrocellulose membrane (GE Healthcare, Little Chalfont, Buckinghamshire, UK) in a buffer containing 20 mM Tris, 20 mM glycine, and 20% methanol at a constant voltage of 100 V for 25 min. Residual binding sites were blocked, and the membrane was incubated with a PAR2 monoclonal primary antibody (SAM-11; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) at a 1:1000 dilution. This antibody is directed to amino acids 37 to 50 (SLIGKVDGTSHVTG) of hPAR2. After a further washing period the membrane was incubated with goat anti-mouse IgG conjugated to horseradish peroxidase (BD Biosciences, Oxford, UK) at a 1:2000 dilution for 1 h. The membrane was developed using ECL Plus, and antibody detection was conducted with HyperFilm (GE Healthcare).
B5 Antiserum. A polyclonal antiserum, B5 (Kong et al., 1997; Al-Ani et al., 1999), raised to a peptide corresponding to the rat PAR2 sequence 30GPNSKGR↓rSLIGRLDT46P-yggc, coupled to keyhole limpet hemocyanin (↓, trypsin cleavage site; yggc added for conjugation). The peptide sequence SKGRSLIGRL of the immunogen corresponds to the sequence of murine PAR2 incorporating the tethered ligand. B5 detects both the preactivated and activated forms of PAR2 (i.e., total PAR2 immunoreactivity) in various species, including rat and mouse (Wang et al., 2003). The antiserum (1:1000 dilution) was injected (20 μl) into the knee joint 5 min before administration of C/K. As a control, nonimmune rabbit serum, in the same total protein concentration, was administered in separate animals.
Immunohistochemistry. To confirm PAR2 selectivity, cryopreserved brains from PAR2+/+ and PAR2–/– mice were analyzed using B5 since all four PARs have been detected in the central nervous system (Striggow et al., 2001). Sections (6 μm) were acetone fixed, treated to quench endogenous peroxidase, and then incubated with blocking buffer. Sections were incubated with B5 at 1:1000 (12 h; 4°C). Endogenous biotin was blocked using an Avidin/Biotin blocking kit (Vector Laboratories, Peterborough, UK). Biotinylated secondary antibody (Santa Cruz Biotechnology, Inc.) was applied for 30 min and then incubated with peroxidase-conjugated streptavidin for 30 min. Antigen-antibody complexes were visualized using 3,3′-diaminobenzadine. Slides were counterstained with Mayer's hematoxylin and mounted in di-n-butyl phthalate polystyrene xylene.
Small Interfering RNA. Small interfering RNA (siRNA) sequences were designed as described previously (Elbashir et al., 2001). A sequence 5′-AA(N19) UU-3′ was selected from the open reading frame of the cDNA sequence of the PAR2 gene ∼100 nucleotides downstream of the start codon. This sequence, siRNAPAR2 (AAU GGC AUG GCC CUC UGG AUC), and a scrambled sequence derived from it, siRNACON (AAG UCG UGA CGC UCC GUG CAU), were synthesized (Dharmacon Research, Inc., Lafayette, CO). Both sequences were checked (www.ncbi.nlm.nih.gov/BLAST) to ensure that only the PAR2 gene was targeted by the siRNAPAR2 and that the siRNACON sequence was nonspecific. Either siRNAPAR2 or siRNACON was injected i.p. at a dose of 2 nmol/day for 3 days before C/K administration.
In Vitro Studies. NCTC2544 cells, expressing hPAR2 (clone G) were prelabeled with 1μCi/ml [3H]myoinositol for 24 h in serum-free medium. Cells were preincubated with 10 mM LiCl for 30 min and stimulated for a further 40 min with trypsin (50 nM) or the PAR2 agonist SLIGKV-NH2 (100 μM). Reactions were terminated and [3H]inositol phosphates extracted using 1 ml of methanol/0.5 ml of chloroform. Water-soluble inositol phosphates were separated from phospholipids by addition of 0.5 ml of chloroform and 0.8 ml of H2O followed by centrifugation. Total inositol phosphates were then assayed by ion exchange chromatography on Dowex-1 columns (Plevin et al., 1990) and measured by liquid scintillation counting.
Drugs. Trypsin (catalog no. T1426), soybean trypsin inhibitor (catalog no. T9003), thrombin (catalog no. T6634), and nonimmune rabbit serum were obtained from Sigma Chemical. β-Tryptase was expressed in Pichia pastoris (Niles et al., 1998) and purified to apparent homogeneity (specific activity >90% of the theoretical value). 4-Amidinophenylpyruvic acid (APPA) was a gift from J. Stürzebecher (Germany). SLIGKV-NH2 was obtained from Neosystem (Strasbourg, France).
Statistics. Data are expressed as mean ± S.E.M., and comparisons were performed using two-tailed t test or one- or two-way repeated measures ANOVA, with Bonferroni post hoc testing (SigmaStat; SPSS Inc., Chicago, IL). Swelling was evaluated as change in joint diameter as percentage of preinjection values.
PAR2 Inhibition at the Genomic Level. Intra-articular injection of C/K resulted in progressive knee joint swelling in PAR2+/+ mice, a response significantly (p = 0.014; two-way ANOVA; n = 4–6) attenuated in PAR2–/– mice (Fig. 2A). The involvement of PAR2 in acute joint inflammation was further highlighted by immunoblot analysis of PAR2+/+ mouse synovium, which showed a marked up-regulation of PAR2 protein expression in the inflamed joint in a time-dependent manner (Fig. 2A, inset). Immunoblot analysis also revealed an up-regulation of several glycosylated PAR2 species (in the 50- to 100-kDa range) known to be synthesized by cells or tissues, in agreement with the findings of Compton et al. (2002).
Using another genomic approach, siRNAPAR2 was administered i.p. to PAR2+/+ mice 3 days before induction of acute joint inflammation. This down-regulated PAR2 as confirmed by immunoblot analysis (Fig. 2B, inset), with a corresponding significant (p < 0.0001; two-way ANOVA; n = 8–11) reduction of joint swelling (Fig. 2B). siRNACON had minimal effect on PAR2 expression relative to siRNAPAR2 treatment (Fig. 2B, inset). In contrast to siRNA PAR2 treatment, siRNACON did not reduce joint swelling (p = 0.09), with the two siRNA treatments differing significantly (p < 0.0001; Bonferroni; n = 8–11).
PAR2 Inhibition by Proteinase Suppression. Intraarticular injection of trypsin (210 μmol) in PAR2+/+ mice caused a rapid increase in knee joint swelling, peaking at approximately 4 h and slowly subsiding thereafter (Fig. 3A). This effect is specific for PAR2 as the same procedure repeated in PAR2–/– mice was ineffective. Likewise, intraarticular injection of human β-tryptase (5 μg) also resulted in knee joint swelling over a comparable time course (Fig. 3B). Again, this effect is specific for PAR2 because β-tryptase had no effect in PAR2–/– mice. In both cases, the responses in PAR2+/+ mice were dose-dependent (data not shown). To investigate the ability of trypsin to activate PAR2, further experiments were performed in NCTC2544 cells expressing human PAR2. PAR2 activation by trypsin (50 nM) in these cells, mimicked by the PAR2-activating peptide SLIGKV-NH2, resulted in an increase in the intracellular levels of inositol-1,4,5-trisphosphate, an effect that was significantly (p < 0.0001; Bonferroni; n = 3) and dose dependently inhibited by soybean trypsin inhibitor (SBTI; Fig. 3C). The effect of SBTI did not involve receptor blockade, since administration of the human PAR2-activating peptide SLIGKV-NH2 resulted in PAR2 activation and an elevation of inositol-1,4,5-trisphosphate irrespective of the SBTI dose. The proinflammatory effects of trypsin and β-tryptase suggested that such serine proteinases might participate in C/K-induced joint inflammation. This possibility was therefore investigated using serine proteinase inhibitors. Preadministration of soybean trypsin inhibitor (5 mg), which can block a number of trypsin family members, significantly (p < 0.0001; Bonferroni; n = 4–5) inhibited C/K-induced joint swelling. Likewise, APPA (5 μg), a tryptase inhibitor (Stürzebecher et al., 1992; Sommerhoff et al., 2000) that can also inhibit other serine proteinases, was even more effective (Fig. 3D) at attenuating the inflammatory response triggered by C/K (p < 0.0001; Bonferroni; n = 4–5).
PAR2 Inhibition by Occluding the Tethered Ligand. To first check the specificity of B5 for PAR2, we examined the effect of this antiserum on PAR1 activation. Thrombin is known to activate PAR1, -3, and -4, but not PAR2, which is preferentially activated by other serine proteinases, such as trypsin. We therefore examined the effect of intra-articular injection of thrombin (20 units) on joint swelling. Thrombin induced modest swelling (7.5 ± 2%), but this was not significantly (p = 0.57; two-way ANOVA; n = 4) inhibited by prior administration of B5 (Fig. 4A), demonstrating pharmacologically the specificity of B5 for PAR2. This result was confirmed further by immunohistochemical analysis of PAR2 expression using B5 in brain as a control tissue known to express all four PARs (Striggow et al., 2001). Immunohistochemical analysis revealed PAR2 expression in neurons of the PAR2+/+ mouse, but no staining was evident in the PAR2–/– mouse (Fig. 4B), despite the presence of PAR1, PAR3, and PAR4 in the latter.
Having established the specificity of B5 antiserum for PAR2, experiments were then performed to determine whether prevention of receptor cleavage and activation by the revealed tethered ligand using B5 would attenuate the proinflammatory role of PAR2 in this acute model of joint inflammation. Prior intra-articular injection of the B5 antiserum substantially reduced the acute inflammatory response to C/K (Fig. 4C), the 4-fold reduction being highly significant compared with C/K alone (p < 0.0001; Bonferroni; n = 4–6). Prior administration of nonimmune serum did not significantly attenuate the inflammatory response (p = 0.3; Bonferroni; n = 4–6). The attenuated response in the presence of the B5 antiserum was also significantly different (p < 0001; Bonferroni; n = 6) from inflammation observed in the presence of nonimmune serum (Fig. 4C). SAM-11, a monoclonal antibody to human PAR2 that targets the residues SLIGXXD in common with B5, also significantly (p < 0.0001; two-way ANOVA; n = 4) and dose dependently attenuated joint swelling (Fig. 4D), with both doses of antibody differing from vehicle treatment (p < 0.0001; Bonferroni; n = 4) and from each other (p < 0.002; Bonferroni; n = 4). Together with the B5 studies, these results suggest that proteinase activation of PAR2 via its revealed tethered ligand is necessary for its proinflammatory role.
PAR2 Inhibition by Receptor Antagonism. ENMD-1068 is a novel selective PAR2 antagonist, based on a disubstituted piperazine (Fig. 5A), whose selectivity was confirmed both in vitro and in vivo. Figure 5B shows a representative inhibition study using murine Lewis lung carcinoma (LLC) cells stimulated with 200 μM SLIGKV-NH2. ENMD-1068 dose dependently inhibited calcium signaling, this being complete at 5 mM. Similar inhibition was observed on both human and murine cells stimulated with either human or murine agonist peptide (data not shown). Proteolytically activated PAR2 studies were performed using trypsin, and Fig. 5C shows complete inhibition of trypsin signaling in murine LLC cells at 5 mM. ENMD-1068 had no effect on the proteolytic activity of trypsin (data not shown). Inhibition of trypsin signaling was also observed in human cell lines (data not shown). ENMD-1068 specificity was demonstrated by assessing signaling using the PAR1 agonist peptide TFLLRN (Fig. 5D) and ATP (data not shown) in the presence of increasing concentrations of ENMD-1068. In all cases, no inhibition of calcium signaling was observed with ENMD-1068. Finally, ENMD-1068 did not inhibit murine platelet aggregation stimulated by the addition of thrombin, demonstrating that up to 5 mM ENMD-1068 has no inhibitory activity against thrombin-mediated PAR3 and PAR4 signaling (data not shown).
Having demonstrated the in vitro selectivity of ENMD-1068, we next confirmed its selectivity in vivo. The observation that C/K-induced swelling was not completely abolished in PAR2–/– mice (Fig. 2A), presented an opportunity to assess specificity by administering the antagonist before induction of inflammation in these mice. ENMD-1068 (4 mg i.p.) in PAR2–/– mice had no significant effect across 24 h (p = 0.544; two-way ANOVA; n = 3–6) compared with vehicle-treated PAR2–/– mice (Fig. 5E), confirming ENMD-1068 is not influencing inflammatory pathways other than those mediated via PAR2. The specificity of this compound was further confirmed (Fig. 5E) by the observation that ENMD-1068 (4 mg i.p.) had no effect on thrombin-mediated knee joint swelling over 48 h (p = 0.99; two-way ANOVA; n = 4–5).
The culmination of this study is the key finding that joint inflammation was dose dependently attenuated by prior i.p. administration of ENMD-1068 (Fig. 6), this effect being highly significant (p < 0.0001; two-way ANOVA; n = 5–6). Compared with i.p. injection of vehicle, joint swelling was significantly reduced by ENMD-1068 at both the 1- and 4-mg doses (p < 0.0001 in both cases; Bonferroni; n = 5–6), which also differed from each other (p = 0.0065; Bonferroni; n = 5–6).
Investigation of the pathophysiological roles for PAR2 has been limited to date by the lack of a selective receptor antagonist. The current study presents the first evidence demonstrating the inhibition of acute joint inflammation by multiple strategies targeting PAR2 (Fig. 1), including the use of a novel PAR2 antagonist.
PAR2 has previously been reported to mediate acute inflammatory responses by neurogenic (Steinhoff et al., 2000) and non-neurogenic mechanisms (Damiano et al., 1999). Several lines of evidence are presented in the current murine study to demonstrate a role for PAR2 in acute knee joint inflammation. First, the inflammatory response in the acute model of arthritis was substantially attenuated in PAR2-deficient mice, supporting a key role for this receptor in inflammation. Second, the expression of PAR2 protein in murine articular tissue was substantially up-regulated during acute joint inflammation. Preventing such up-regulation by post-transcriptional gene silencing of PAR2, using siRNA designed against the receptor sequence, resulted in inhibition of joint inflammation, supporting the findings of the PAR2–/– murine studies. The significant reduction of joint swelling in the PAR2–/– animals suggested that preventing PAR2 activation would have significant anti-inflammatory potential. Indeed, inhibition of joint inflammation by preventing PAR2 activation using 1) proteinase inhibitors, 2) antibodies targeted against PAR2, and 3) an antagonist for the ligand binding domain on PAR2 together provide proof of principle that PAR2 plays a key role in joint inflammation via receptor up-regulation, proteinase cleavage, and endogenous activation by the revealed tethered ligand.
Post-transcriptional gene silencing using siRNA technology proved an effective means of reducing synovial expression of PAR2, thereby inhibiting joint inflammation (Fig. 2B). Although siRNA sequences have been administered in vivo with success previously (Soutschek et al., 2004), this study is the first to show both that intraperitoneal administration of siRNA can be effective in reducing levels of targeted proteins within joint tissues and that such technology can be applied to down-regulate PAR2 expression specifically, with potential therapeutic value. Further work is required to delineate which cells express PAR2 in the inflamed joint and whether the siRNA-induced reduction in expression was in the resident articular and/or infiltrating inflammatory cells.
Trypsin is well recognized as a proteinase activator of PAR2, as corroborated in the present study on murine synovium. Significantly, trypsin as well as β-tryptase were able to mimic C/K-induced joint swelling, an effect that was abrogated by inhibitors of these proteolytic enzymes. It was also significant that neither trypsin nor β-tryptase caused joint swelling in PAR2 null animals, a result that is entirely in keeping with a lack of effect of these proteinases in other inflammatory models using PAR2-deficient animals (Vergnolle et al., 2001b). An endogenous source of trypsin-like enzymes in the murine joint has not yet been established, although it has been shown in the rat lung that PAR2 colocalizes immunohistochemically with an unidentified trypsin family member in epithelial and endothelial cells (Cocks and Moffatt, 2000). Extrapancreatic trypsin can also be found in a number of sites in humans, such as in human dermal endothelial cells, wherein trypsinogen mRNA has been detected (Shpacovitch et al., 2002). Furthermore, members of the trypsin family are now known to be up-regulated in the setting of a murine model of infectious colitis, which accounts for the inflammatory response mediated by PAR2 (Hansen et al., 2005). Mast cells are found in large numbers in the inflamed human rheumatoid synovium (Woolley and Tetlow, 2000) and can be implicated in PAR2-mediated inflammation in humans since these cells release serine proteinases such as tryptase. Tryptase, an activator of PAR2 (Corvera et al., 1997; Molino et al., 1997), is known to be present in human mast cells and is thus an endogenous candidate for PAR2 cleavage/activation within the joint in humans. Murine connective tissue mast cells express mouse mast cell proteinase-6 and -7 (MMCP-6 and -7), which have approximately 75% amino acid sequence identity with human mast cell tryptases (Reynolds et al., 1991), but it cannot be excluded that other serine proteinases may have contributed to the inflammation generated by the C/K stimulus. Remarkably, both soybean trypsin inhibitor, which can potentially block a number of the murine trypsins, and APPA, another serine proteinase inhibitor that can target various serine proteinases, including human tryptase, were effective in attenuating C/K-induced joint inflammation. Thus, proteinase inhibitors themselves may represent an effective therapeutic modality for arthritis. Interestingly, the joint swelling induced by intra-articular application of either trypsin or tryptase was effectively abolished in PAR2–/– mice (Fig. 3, A and B), arguing that the inflammatory actions of both potent proteinases seem to be mediated primarily via PAR2 activation. In this context, both proteinases provide valuable experimental tools for cleavage-mediated endogenous activation of the receptor. However, other serine proteinases such as human leukocyte elastase and cathepsin G (Uehara et al., 2003) may also play a PAR2-activating role in the setting of arthritis.
Although the B5 antiserum was originally raised against rat PAR2 (Kong et al., 1997), the sequence of the immunizing rat peptide is the same in much of its sequence as the comparable murine sequence. The specificity of the B5 antiserum for murine PAR2 was confirmed immunohistochemically by the selective positive staining in the brain of PAR2+/+ mice, but not PAR2–/– mice, which express PAR1, PAR3, and PAR4, but clearly not PAR2 in this tissue. The selectivity of B5 for PAR2 inhibition was confirmed further by the observation that the antiserum did not significantly inhibit thrombin-induced joint swelling. The effectiveness of B5 in detecting murine PAR2 protein is because of the close homology of PAR2 between rat and mouse (97% sequence identity on the DNA level; BLAST). The selectivity of SAM-11 for human PAR2 has previously been established (Molino et al., 1998), and for the present study it was also confirmed immunohistochemically in murine tissues (data not shown).
Importantly, this study provides the first demonstration that antibodies raised against the cleavage site on the tethered ligand, thereby inhibiting receptor activation, have significant anti-inflammatory action (Fig. 4, C and D). However, the therapeutic potential of this approach is limited by the need for local intra-articular administration to ensure efficacy. The small molecule antagonist for PAR2, ENMD-1068, allowed a more traditional pharmacological approach to blocking PAR2 activation and offers considerably greater potential for future anti-inflammatory therapy because this compound, when administered systemically, powerfully and dose dependently reduced knee joint swelling (Fig. 6). ENMD-1068 inhibits PAR2 selectively in vitro and in vivo (Fig. 5), although requiring relatively high concentrations, possibly reflecting low receptor affinity. This is reminiscent of peptides described previously that inhibited PAR2 activation by trypsin but again with low potency (Al-Ani et al., 2002). Even so, ENMD-1068 and analogs provide good “lead” compounds for future development of therapeutically valuable agents.
Using multiple strategies to prevent PAR2 activation (Fig. 1), including the first use of a receptor antagonist, our findings support the unifying conclusion that PAR2 is a therapeutic target in the treatment of inflammatory joint disease. This is the first study to establish such proof of concept and identifies antagonists and/or antibodies targeting PAR2 as potentially powerful anti-inflammatory agents for treatment of arthritis.
We gratefully acknowledge expert technical assistance by Marion Drew and the contribution of Vicky King (supported by Oliver Bird Rheumatism Programme) to murine immunohistochemistry using SAM-11.
- Received August 4, 2005.
- Accepted October 31, 2005.
This work was partly supported by the Arthritis Research Campaign (14521; to J.C.L. and W.R.F.) and Kowa Company, Ltd. (to W.R.F. and J.C.L.), EntreMed, Inc. (to J.C.L. and W.R.F.), and a Canadian Institutes of Health Research Proteinases and Inflammation Network Group grant (to M.D.H.).
ABBREVIATIONS: PAR, proteinase-activated receptor; C/K, carrageenan/kaolin; ENMD-1068, N1-3-methylbutyryl-N4-6-aminohexanoyl-piperazine; hPAR, human proteinase-activated receptor; siRNA, small interfering RNA; APPA, 4-amidinophenylpyruvic acid; ANOVA, analysis of variance; SBTI, soybean trypsin inhibitor; LLC, Lewis lung carcinoma.
- The American Society for Pharmacology and Experimental Therapeutics