The novel PAR2 ligand C391 blocks multiple PAR2 signalling pathways in vitro and in vivo

Abstract

Background and Purpose

Proteinase-activated receptor-2 (PAR2) is a GPCR linked to diverse pathologies, including acute and chronic pain. PAR2 is one of the four PARs that are activated by proteolytic cleavage of the extracellular amino terminus, resulting in an exposed, tethered peptide agonist. Several peptide and peptidomimetic agonists, with high potency and efficacy, have been developed to probe the functions of PAR2, in vitro and in vivo. However, few similarly potent and effective antagonists have been described.

Experimental Approach

We modified the peptidomimetic PAR2 agonist, 2-furoyl-LIGRLO-NH₂, to create a novel PAR2 peptidomimetic ligand, C391. C391 was evaluated for PAR2 agonist/antagonist activity to PAR2 across G_q signalling pathways using the naturally expressing PAR2 cell line 16HBE14o-. For antagonist studies, a highly potent and specific peptidomimetic agonist (2-aminothiazo-4-yl-LIGRL-NH₂) and proteinase agonist (trypsin) were used to activate PAR2. C391 was also evaluated in vivo for reduction of thermal hyperalgesia, mediated by mast cell degranulation, in mice.

Key Results

C391 is a potent and specific peptidomimetic antagonist, blocking multiple signalling pathways (G_q-dependent Ca²⁺, MAPK) induced following peptidomimetic or proteinase activation of human PAR2. In a PAR2-dependent behavioural assay in mice, C391 dose-dependently (75 μg maximum effect) blocked the thermal hyperalgesia, mediated by mast cell degranulation.

Conclusions and Implications

C391 is the first low MW antagonist to block both PAR2 Ca²⁺ and MAPK signalling pathways activated by peptidomimetics and/or proteinase activation. C391 represents a new molecular structure for PAR2 antagonism and can serve as a basis for further development for this important therapeutic target.

Tables of Links

TARGETS
GPCRsa2011
PAR2
Enzymesb2011
ERK1/2
p38 MAP kinase

LIGANDS
2-f-LIGRLO-NH2
GB88

These Tables list key protein targets and ligands in this article which are hyperlinked to corresponding entries in http://www.guidetopharmacology.org, the common portal for data from the IUPHAR/BPS Guide to PHARMACOLOGY (Pawson et al., 2014) and are permanently archived in the Concise Guide to PHARMACOLOGY 2013/14

^a,bAlexander et al., 2013a,b,).

Introduction

The proteinase-activated receptor-2 (PAR2) is one of the four members of the family of GPCRs that are activated after proteolytic cleavage of their extracellular, amino terminus (Adams et al., 2011; Ramachandran et al., 2012). The resulting ‘tethered-peptide’ sequence (ending with SLIGRL in the rodent receptor and SLIGKV in the human receptor) exposed after proteolytic cleavage activates PAR2. A variety of potent and effective peptidomimetic agonists based upon the exposed tethered sequences have been developed to PAR2 (Adams et al., 2011; Boitano et al., 2014). These compounds have been very useful in understanding the consequences of PAR2 activation across experimental models. However, the natural tethered agonist presentation for PAR2, and its corresponding access to the PAR2 binding pocket, has proved a difficult target for development of antagonists. Despite this difficulty, a number of PAR2 antagonists have been proposed (Suen et al., 2012; Yau et al., 2013). Further complicating the issue of drug development is the growing evidence for ‘biased signalling’ that can follow PAR2 agonism (Hollenberg et al., 2014) or antagonism (Goh et al., 2009; Suen et al., 2014).

PAR2 plays an important role in a variety of diseases linked to proteinase release from endogenous sources or exposure to exogenous proteinases (Ramachandran et al., 2012; Hollenberg et al., 2014). One consequence of PAR2 activation in the peripheral nervous system is sensitization of neurons responsible for transmitting noxious information to the CNS. These nociceptive neurons express PAR2, and PAR2 activation on these neurons leads to enhanced signalling via a variety of channels including the capsaicin and noxious heat receptor, TRPV1 (Dai et al., 2004). PAR2 is responsible for proteinase sensitization of TRPV1 in vivo, leading to thermal hyperalgesia. PAR2 null animals have deficits in pain sensitization in a variety of inflammatory pain models, and PAR2 activation is sufficient to induce a transition to a chronic pain state, making this receptor an important target for drug development for pathological pain (Vergnolle, 2009; Bao et al., 2014; Tillu et al., 2015). Additionally, a broad variety of pre-clinical and clinical findings link exogenous proteinases, and more specifically PAR2, to asthma (Reed and Kita, 2004; Vergnolle, 2009; Jacquet, 2011; Snelgrove et al., 2014). While there is a strong rationale for PAR2 antagonist drug discovery for these indications, few PAR2 antagonists have been described and even fewer have been demonstrated to exhibit efficacy in pre-clinical disease models (Yau et al., 2013).

Here, we describe a novel PAR2 antagonist based upon structures from previously described peptidomimetic PAR2 agonists. This compound, C391 [(2S)-6-amino-N-((S)-1-(((S)-1-amino-3-(4-hydroxyphenyl)-1-oxopropan-2-yl)amino)-4-methyl-1-oxopentan-2-yl)-2-((3S)-1-(furan-2-carbonyl)-3-(3-methylbutanamido)-4,7-dioxooctahydro-8H-pyrazino[1,2-a]pyrimidin-8-yl)hexanamide], attenuates PAR₂ signalling via Ca²⁺ and MAPK pathways and is efficacious in vivo. C391 is structurally distinct from all previously described PAR2 antagonists and therefore represents a new lead for further PAR2 antagonist discovery.

Methods

Compound syntheses

Compound C391 (Figure 1) was synthesized by solid phase methodology using Fmoc strategy on a base labile Sheppard’s hydroxymethyl-benzoic acid linker (Atherton et al., 1981). The azabicycloalkane scaffold was cyclized on the resin by diluted formic acid, followed by basic cleavage (0.1 M NaOH) of the product from the resin. The crude product was purified by preparative HPLC on a 22 × 250 mm reverse-phase column. The structure was confirmed by NMR and high-resolution mass spectroscopy (HR-MS).

Figure 1.

Structural comparison of PAR2 ligands 2-furoyl-LIGRLO-NH₂ and C391. The chemical structures for (A) the known PAR2 agonist 2-furoyl-LIGRLO-NH₂ and (B) the novel peptidomimetic compound 391 (C391) are shown; colour coding denotes predicted surrogates between the two structures. The 2-furoyl-amide heterocycle is shown in red. The Leu residue in peptide A (blue) is mimicked by part of the azabicycloalkane scaffold in C391. The Ile residue (violet) is substituted with isobutyl side-decoration of azabicycloalkane in C391. The positively charged Arg (green) is substituted with a Lysine-like group in C391. The Leu-Tyr C-terminal dipeptide on C391 is not derived from peptide A.

PAR2 activating peptidomimetics, 2-aminothiazo-4-yl-Leu-Ile-Gly-Arg-Leu-NH₂ (2-at-LIGRL-NH₂), 2-furoyl-Leu-Ile-Gly-Arg-Leu-Orn(Aloc)-NH₂ (2-f-LIGRLO-NH₂) and compound GB88 (5-isoxazoyl-Cha-Ile-spiroindane-1,4piperidine) were synthesized as previously described (Barry et al., 2010; Flynn et al., 2011; Hoffman et al., 2012). All compounds were purified by preparative HPLC and the structures were confirmed by HR-MS and HPLC.

Tissue culture

16HBE14o- (immortalized human bronchial epithelial cell line) cells, an SV40 transformed human bronchial epithelial cell line (Gruenert et al., 1995), were obtained through the California Pacific Medical Center Research Institute (San Francisco, CA, USA). The cells were maintained and expanded as previously described (Flynn et al., 2013). Cell lines were expanded in tissue culture flasks prior to transfer to glass coverslips or 96-well plates for experiments. Cultureware was coated initially with a matrix coating solution (MCS: 88% LHC basal medium, 10% BSA, 1% bovine collagen type I and 1% human fibronectin) and incubated for 2 h at 37°C, after which the coating solution was removed and allowed to dry for at least 1 h. 16HBE14o- cells were plated onto the cultureware at a concentration of 1 × 10⁵ cells·cm⁻² and grown in Eagle’s minimal essential medium supplemented with 10% FBS, 2 mM glutamax, penicillin and streptomycin (growth medium) at 37°C in a 5% CO₂ atmosphere. Growth medium was replaced every other day until the cells reached proper cell density for experimentation.

In vitro physiological response with real-time cell analysis

16HBE14o- cells on E-plates in growth medium and in a 37°C, 5% CO₂ incubator were monitored for the establishment of, and changes in, impedance overnight every 15 min using the xCELLigence™ real time cell analyser (RTCA; ACEA Biosciences, San Diego, CA, USA) (Atienza et al., 2005; Flynn et al., 2013). When cells reached appropriate impedance the next day, and prior to the experiment, the RTCA was moved to room air and temperature where full growth medium was replaced with 100 μL of modified Hank’s balanced salt solution (HBSS) pre-warmed to 37°C. The RTCA was then allowed to reach room temperature (45 min) prior to ligand addition. Each well was then supplemented with 100 μL of HBSS containing appropriate concentrations of 2-at-LIGRL-NH₂ and/or C391 ligands in quadruplicate. Additional wells were used for vehicle controls. Relative impedance in each well was monitored every 30 s over 4 h. Traces were normalized for experimental start time and baselined to ligand-free-treated cells as described (Flynn et al., 2013).

Intracellular Ca²⁺ concentration ([Ca²⁺]_i) signalling

[Ca²⁺]_i was measured using digital imaging microscopy as previously described (Flynn et al., 2013), with slight modifications. Coverslip cultures were washed with HBSS and loaded for 45 min in 5 μM fura 2-AM in HBSS. Fura 2 fluorescence was observed on an Olympus IX70 microscope (Waltham, MA, USA) with a 40× oil immersion objective after alternating excitation between 340 and 380 nm by a 75 W Xenon lamp linked to a Delta Ram V illuminator (PTI, London, Ontario, Canada) and a gel optic line. Images were captured using an Evolve camera (Roper Scientific, Tucson, AZ, USA) under EasyRatioPro software (PTI). [Ca²⁺]_i for each individual cell in the field of view was calculated from captured images by ratiometric analysis using equations originally published in Grynkiewicz et al. (1985). A minimum of one ratio per second was calculated for all experiments <10 min. For 20 min experiments presented in the Supporting Information, one ratio per 3 seconds were captured over the first 15 min, and one ratio per second over the final 5 min. Line drawings representing [Ca²⁺]_i traces over time are mean [Ca²⁺]_i of all cells within a field of view (80–110) and are representative of at least four experiments. Ca²⁺ response graphs are the percentage of cells that increase [Ca²⁺]_i above 150 nM within the 3 min experimental time frame that included agonist addition. Inhibition experiments included a 2 min pre-exposure to antagonists prior to the addition of agonist/antagonist combination, unless otherwise noted. Agonists included the PAR2-specific peptidomimetic 2-at-LIGRL-NH₂ (Flynn et al., 2011) and porcine pancreas trypsin (5 nM; Sigma-Aldrich T-0303, St. Louis, MO, USA). Ca²⁺ concentration–response curves and IC₅₀ values were calculated from Ca²⁺ response experiments using GraphPad Prism software (GraphPad Software Inc., San Diego, CA, USA).

Detection of MAPK signalling

In-cell Western (ICW) studies were performed as previously described (Flynn et al., 2011) with modification to assess antagonistic responses. 16HBE14o- cells were grown to confluence on black-walled 96-well plates pre-coated with MCS. Cells were exposed to a range of concentrations of C391 from 300 nM to 100 μM for 2 min at room temperature. Cells were then exposed to PAR2-specific peptidomimetics (2.5 μM 2-at-LIGRL-NH₂) in the presence of C391 for an additional 5 min, fixed, incubated in primary antibody (rabbit anti-pERK; Cell Signaling Technologies, Danvers, MA, USA), secondary antibody (Invitrogen, Life Technologies, Grand Island, NY, USA), nuclear stain DRAQ5 (Cell Signaling Technologies) and imaged on an Odyssey Scanner (LI-COR, Lincoln, NE, USA) as per manufacturers’ instructions. For Western blot experiments, 16HBE14o- cells were plated on MCS-coated 6-well plates. Once cells reached confluence, the culture medium was removed, cells were pre-treated with C391 for 1 min and then exposed to PAR2 agonist with C391 for an additional 5 min. Agonist/antagonist was removed, and cells were lysed in a modified radioimmune precipitation assay buffer [50 mM Tris (pH 7.4), 150 mM NaCl, 1 mM EDTA (pH 8.0), 1% Triton X-100]. Lysates were sonicated for 10 min at 4°C, centrifuged for 15 min at 14 000× g and stored at −80°C until blot analysis. Both activated ERK1/2 [phosphorylated ERK (pERK)] and non-activated ERK1/2 (total ERK; tERK), were assayed by standard SDS-PAGE (antibodies from Cell Signaling Technology).

Animal welfare and ethical statement

All animal care and experimental protocols were approved by the Institutional Animal Care and Use Committee of The University of Arizona. These procedure comply with the ARRIVE guidelines (Kilkenny et al., 2010; McGrath et al., 2010). All experimental procedures were as humane as possible. A total of 20 animals were used in the experiments described here.

In vivo thermal hyperalgesia

Male ICR mice (Harlan, Houston, TX, USA) weighing 25–30 g were used for these studies. Mice were placed in Plexiglas boxes for habituation and then baseline thermal latencies were assessed using the method of Hargreaves et al. (1988). Mice received intraplantar injections of 0.2 μg Compound 48/80 (Sigma-Aldrich) (Vergnolle et al., 2001) with or without indicated concentrations of C391 in a total volume of 25 μL through a 31 gauge needle. Thermal latencies were then assessed at the indicated time points over a total of 180 min. Experimental measurements were performed by observers unaware of the treatments and animals were randomized to groups (3–8 mice) by drawing numbers.

Data analysis

All statistical analyses were performed with GraphPad Prism software (GraphPad Software Inc.). Curve fitting was carried out using non-linear regression with variable slope. Multivariate comparisons were performed with a two-way anova with Tukey’s or Bonferroni multiple comparisons post test, as appropriate for the individual experiment. Pairwise comparisons were performed with a two-tailed Student’s t-test. A value of P < 0.05 was used to establish a significant difference between samples. Data in the figures are shown as mean ± SEM, unless otherwise noted.

Results

C391 is a weak, partial PAR2 agonist in RTCA and Ca²⁺ signalling assays

We modified the peptidomimetic agonist, 2-f-LIGRLO-NH₂ (McGuire et al., 2004), substituting Leu₂-Ile₃-Gly₄ with an azabicycloalkane scaffold and Orn₆ with a Tyr, to create a novel peptidomimetic structure, C391 (Figure 1A, B). To evaluate the signalling effects of C391 on PAR2, we first examined its ability to activate signalling in PAR2-expressing 16HBE14o- cells using the highly sensitive xCELLigence real time cell analyser (RTCA) and Ca²⁺ imaging signalling assays (Flynn et al., 2011; Boitano et al., 2014). In the RTCA assay, C391 displayed a weak, partial agonist response at the highest concentration tested, with no detected responses at concentrations below 10 μM and an E_max that was <50% of the maximum response of an EC₇₅ concentration of the potent agonist 2-at-LIGRL-NH₂ (1 μM; Figure 2A). The RTCA EC₅₀ for partial activation by C391 was 17.9 μM with a 95% confidence interval (CI) of 6.2–51 μM. Similarly, in the highly sensitive Ca²⁺ imaging signalling assay, C391 induced only minimal [Ca²⁺]_i changes following addition of ≤10 μM concentrations. However, at higher concentrations, C391 induced partial activation of Ca²⁺ signalling (Figure 2B, D–G), with a reduced output compared with an EC₇₅ concentration of 2-at-LIGRL-NH₂ (2 μM; Figure 2B, H–K). At E_max (≥30 μM), C391 could only activate 33% of cells, whereas full agonists reach 100% activation. A calculated Ca²⁺ signalling EC₅₀ for partial activation by C391 was 19.6 μM (CI: 13.9–27.6 μM). C391 activity was eliminated following PAR2 desensitization with high concentrations (50 μM) of the potent agonist 2-at-LIGRL-NH₂ (Figure 2C). Such changes are consistent with specific, partial agonism of PAR2 at high concentrations of C391 (Boitano et al., 2014).

Figure 2.

C391 is a weak PAR2-dependent agonist. (A) RTCA traces of PAR2 ligands on 16HBE14o- cells. C391 displays minimal activation except at high concentrations. For comparison, EC₇₅ and EC₅₀ responses for the potent agonist 2-at-LIGRL-NH₂ are shown. Traces are mean values from four wells from a single E-plate and are representative of at least three plates. (B) Maximum Ca²⁺ signalling responses for C391 and EC₇₅ Ca²⁺ signalling response for 2-at-LIGRL-NH₂ are shown for comparison. Traces are mean values from typical experiments that include >80 cells and are plotted with their SEM. (C) PAR2 desensitization assay using 50 μM 2-at-LIGRL-NH₂ demonstrates PAR2 specificity of the C391 partial agonism response. (D–K) [Ca²⁺]_i colour maps for C391 (D–G) and 2-at-LIGRL-NH₂ (H–K) are overlaid onto a DIC image of 16HBE14o- cells for a typical ligand addition experiment. Data demonstrate partial agonism of PAR2 at high concentrations of C391.

C391 is a potent antagonist across signalling assays

Having determined that C391 binds PAR2 at lower concentrations without readily inducing signalling via this receptor, we sought to determine if C391 could function as a novel PAR2 antagonist. C391 was initially evaluated for PAR2 antagonist activity using the Ca²⁺ signalling assay. In these experiments, a concentration inhibition curve for C391 was constructed against an EC₇₅ concentration of 2-at-LIGRL-NH₂ for Ca²⁺ activation in 16HBE14o- cells. A 2 min pretreatment of 16HBE14o- cells with C391 potently inhibited peptidomimetic-induced PAR2 Ca²⁺ signalling (Figure 3; IC₅₀ = 1.30 μM, CI: 277 nM–6.07 μM), with a maximal effect observed at 3 μM. When the C391 concentration was increased to 10 μM, strong antagonism was still apparent. However, at 30 μM, C391 failed to block the effect of 2 μM 2-at-LIGRL-NH₂. This was likely to be due to the partial agonist activity observed at higher C391 concentrations, as described above.

Figure 3.

C391 is a potent, PAR2-dependent Ca²⁺ signalling antagonist. (A) PAR2 antagonism experimental protocol is illustrated. [Ca²⁺]_i is monitored throughout. Arrows above the axis correspond to changes in imaging medium. C391 at appropriate concentration in HBSS is introduced at 20 s. HBSS containing identical C391 concentration supplemented with an EC₇₅ concentration of 2-at-LIGRL-NH₂ is introduced at 140 s. Arrowheads under the axis correspond to Ca²⁺ maps presented in C–N. (B) Percentage of cells with Ca²⁺ response between 120–300 s of the experiment are shown for each C391 concentration. Data points (n ≥ 4) are plotted ± SEM and shown with a non-linear regression curve. Because the 30 μM concentration caused significant Ca²⁺ response, it was excluded from the curve fit. (C–N) Typical [Ca²⁺]_i over time in a group of cells treated as described in (A) with 3 μM (C–F), 10 μM (G–J), or 30 μM (H–K) C391. Colours on individual maps correspond to [Ca²⁺]_i indicated by bar in (C), (G), (K). Maps are overlaid onto DIC image of cells within the experiment to observe individual cell variability. Panels correspond to 120, 180, 240 and 300 s as indicated in (A). C391 blocks PAR2-dependent Ca²⁺ response at 3–10 μM. However, 30 μM C391 induced Ca²⁺ response and is ineffective as a PAR2-Ca²⁺ signalling antagonist.

PAR2 activates MAPK signalling through multiple pathways; this can be measured via ERK 1/2 phosphorylation (pERK). We explored whether C391 antagonized MAPK signalling using the ICW technique. As in the Ca²⁺ signalling assays, 16HBE14o- cells were pre-incubated for 2 min with a range of concentrations of C391 (300 nM–100 μM) at room temperature, followed by C391 supplemented with 2.5 μM 2-at-LIGRL-NH₂ for 5 min (Figure 4A). C391 alone did not alter pERK at any of the concentrations tested (not shown). In the presence of 2-at-LIGRL-NH_2, C391 effectively blocked pERK signalling with an IC₅₀ of 14.3 μM (CI: 13.5–15.1 μM). Here, we did not observe a loss of antagonist effect at higher concentrations of C391 (30 and 100 μM), as we did in the Ca²⁺ signalling assay. To confirm the ICW results with an independent assay, we used traditional immunoblotting. As observed earlier, a full block of 2-at-LIGRL-NH₂ (2.5 μM)-induced ERK 1/2 phosphorylation was seen at higher concentrations of C391 (Figure 4B). Also as seen in the ICW, we did not observe MAPK activation by C391 in the absence of PAR2 agonist (not shown). We conclude that C391 blocked PAR2-induced MAPK signalling even at those concentrations where partial agonist activity was observed in the Ca²⁺ signalling assay.

Figure 4.

C391 is a potent PAR2-dependent MAPK signalling antagonist. (A) Using the results of the ICW assay, the graph shows a concentration–response curve for a 2 min pre-incubation of C391 followed by a 5 min incubation of C391 with 2.5 μM 2-at-LIGRL-NH₂. Individual points are means from n = 4 and plotted ± SEM. C391 effectively blocks 2-at-LIGRL-NH₂-induced phosphorylation of ERK 1/2. (B) Immunoblot comparison of 16HBE14o- cell lysates following a 5 min exposure to vehicle control, 2.5 μM 2-at-LIGRL-NH₂ or a 1 min pre-incubation of 100 μM C391 followed by a 5 min exposure to 2.5 μM 2-at-LIGRL-NH₂ with 100 μM C391. The data confirms C391 antagonism of 2-at-LIGRL-NH₂-induced pERK.

C391 blocks proteinase-activated PAR2

Although peptidomimetics provide potent and specific PAR2 activation, PAR2 is naturally activated in vivo via endogenous or exogenous trypsin-like serine proteinases (Ramachandran et al., 2012). Thus, we tested if C391 could block proteinase-activated PAR2 on airway epithelial cells using trypsin. We exposed 16HBE14o- cells to sufficient trypsin to induce Ca²⁺ responses in >50% of the cells (54.4 ± 12.1%, n = 6; Figure 5A–D). Similar to that observed with the peptidomimetics, a 2 min pre-incubation with 3 μM C391 limited trypsin-induced Ca²⁺ responses to 11.6 ± 4.3% (n = 5; Figure 5E–H). This finding demonstrates that C391 attenuates PAR2 signalling induced by proteinases, likely indicating that C391 interferes with binding of the exposed tethered ligand following proteolytic cleavage of the receptor.

Figure 5.

C391 antagonizes trypsin proteinase signalling at PAR2. (A–D) Changes in [Ca²⁺]_i are shown as colour maps over time in response to trypsin. Individual panels are from 20 s before trypsin addition and at 60 s intervals. (E–H) The same experiment as (A) but with a 2 min pre-incubation with 3 μM C391 and continued 3 μM C391 treatment during trypsin exposure. C391 blocks proteinase-induced Ca²⁺ changes.

C391 inhibits proteinase-induced pain response in vivo

PAR2 is an important drug target for pain where the receptor is thought to regulate nociceptor excitability in response to release of endogenous proteinases associated with inflammation or other painful pathologies. Compound 48/80 induces mast cell degranulation in vivo, causing a PAR2-dependent thermal hyperalgesia (Vergnolle et al., 2001). Hence, we tested whether compound 48/80-induced thermal hyperalgesia in mice could be blocked by C391 (Figure 6). Compound 48/80 (0.2 μg) was injected into the hindpaw of mice with or without C391 at doses ranging from 7.5 to 75 μg. Higher doses (25 and 75 μg) of C391 effectively attenuated compound 48/80-induced thermal hyperalgesia. These data confirm that C391 is a PAR2 antagonist capable of blocking proteinase-induced responses in vivo.

Figure 6.

C391 antagonizes proteinase-dependent thermal hyperalgesia in vivo. Mice received intraplantar injections of Compound 48/80 (0.2 μg) with or without the indicated doses of C391 and thermal sensitivity was measured over the ensuing 3 h time period. Thermal latencies are shown in (A) and quantified AUC data in (B). *P < 0.05, two-way anova with Bonferroni post hoc test. n = 3–8 per group.

Discussion and conclusions

We have produced a novel peptidomimetic PAR2 ligand that blocks peptidomimetic and proteinase-induced PAR2-dependent G_q and downstream MAPK signalling in vitro as well as in vivo pain hypersensitivity. C391 demonstrated similar antagonist IC₅₀ values in both Ca²⁺ and MAPK assays against the human receptor. Interestingly, C391 showed partial agonist activity in the Ca²⁺ assay at high concentrations but lacked this agonist activity in a MAPK assay, suggesting that it may not trigger alternate signalling pathways such as β-arrestin-dependent signalling. In vivo, C391 blocked mast cell degranulation-induced thermal hyperalgesia in mice, an effect that had previously been attributed to PAR2 activation using PAR2 null mice (Vergnolle et al., 2001). C391 is the first PAR2 antagonist to demonstrate micromolar efficacy across signalling pathways without MAPK activation at concentrations <100 μM. Because C391 is structurally distinct from previously described PAR2 antagonists, this work elucidates a new molecular entity for PAR2 antagonist drug discovery.

A key finding from this work is that C391 has both partial agonist and antagonist activity. Moreover, this partial agonist activity appears to be pathway selective, as it was observed in Ca²⁺ signalling assays but not in MAPK assays or, in all likelihood, in vivo. While the traditional occupational model does not readily account for this mix of pharmacology, it may be explained by more recent views of receptor theory such as the probabilistic theory. Biased agonism has been observed at a variety of GPCRs (Kenakin, 2004; 2014,; Onaran et al., 2014), including PAR2 (Hollenberg et al., 2014). However, examples of partial agonist/antagonist mixed activity observed herein for C391 and other PAR2 antagonists (Goh et al., 2009; Hollenberg et al., 2014; Suen et al., 2014) are less well defined (Kenakin, 2014). One possibility is that C391 is not a biased ligand; rather it is a weak partial agonist for Ca²⁺ signalling and MAPK activity with our MAPK signalling assessment lacking sufficient sensitivity to detect compound activity. This is a distinct possibility based upon the real-time resolution of digital Ca²⁺ imaging, and the well-accepted sensitivity of this assay. In this case, C391 would have a binding affinity for PAR2 far to the left of its EC₅₀ for G protein signalling activation [defined as a small intrinsic activity (tau) value in the original operational model (Black and Leff, 1983)]. Therefore, the compound would act as a neutral antagonist at most concentrations due to its weak intrinsic activity. Our in vivo findings are consistent with this being the dominant action of the compound even at very high local concentrations.

While several antagonists of PAR2 have been proposed in the literature or in patent applications (Yau et al., 2013), few have been investigated across PAR2 signalling pathways and/or for in vivo efficacy. Those antagonists that have been investigated across multiple signalling pathways include K-14585: N-[1-(2,6-dichlorophenyl)methyl]-3-(1-pyrrolidinylmethyl)-1H-indol-5-yl]aminocarbonyl}-glycinyl-L-lysinyl-L-phenylalanyl-N-benzhydrylamide (Goh et al., 2009; Kanke et al., 2009) and GB88: 5-isoxazoyl-Cha-Ile-sprioindene-1,4,-piperidine (Suen et al., 2010; 2014,; Hollenberg et al., 2014). These compounds provide baseline comparisons for the newly developed C391. K-14585 can block peptide and/or peptidomimetic-induced G_q-associated Ca²⁺ signalling following a 15 min pre-incubation in PAR2-expressing NCT2544 cells (Kanke et al., 2009; Suen et al., 2010). Notably, at 5 and 30 μM, K-14585 reduced SLIGKV-OH-induced inositol phosphate production by 65–70% in these cells. However, K-14585 was much less effective against trypsin-activated PAR2, reducing a 30 nM trypsin-induced inositol phosphate signal by a maximum of 20%. In contrast, K-14585 displayed agonistic effects on p38 MAP kinase (and to a lesser effect, ERK 1/2) in the absence of PAR2-stimulating peptides. Despite an incomplete block of Ca²⁺ signalling and an agonism of MAPK at micromolar concentrations, K-14585 blocked in situ and in vivo physiological effects attributed to PAR2 (Kanke et al., 2009). An improved peptidomimetic PAR2 antagonist, GB88, effectively blocked both peptide and proteinase-induced Ca²⁺ signalling in a variety of cell types following a 15 min pre-incubation. This compound also inhibited PAR2-associated rat paw oedema (Suen et al., 2012) and was effective after oral administration in a chronic arthritis model in rats (Lohman et al., 2012). Subsequent studies of this Ca²⁺-signalling antagonist demonstrated that it too acted as an agonist of ERK 1/2 signalling in addition to other PAR2-associated pathways (Suen et al., 2014). When tested in our more sensitive digital imaging microscopy analysis of Ca²⁺ signalling on 16HBE14o- cells under 2–3 min pre-incubation times for a direct comparison with C391 antagonism reported herein, GB88 did not block Ca²⁺ signalling at concentrations up to 60 μM (data not shown). Under conditions of a 15 min pre-incubation, as reported to be necessary for GB88 antagonism (Suen et al., 2012), GB88 did block Ca²⁺ signalling in our assay (Supporting Information Fig. S1). However, inhibiting concentrations of GB88 (30 μM, 10 μM and, to a lesser extent, 3 μM) included pre-activation of Ca²⁺ signalling in most of the 16HBE14o- cells between 2 and 6 min following initial application of GB88 (Supporting Information Fig. S1). These data suggest, at least at human PAR2, that GB88 acts as a partial or full Ca²⁺ agonist. In contrast, we found that a short pre-incubation (2 min) of 3 μM C391 was sufficient for full block of Ca²⁺ signalling induced by peptide or proteinase agonists (Figures 3 and 5). In summary, C391 differs from both K-14585 and GB88 in that it requires minimal pre-incubation for antagonism of peptide/peptidomimetic activation of PAR2-dependent Ca²⁺ signalling and it acts as an antagonist for MAPK signalling pathways, without any signs of activation, up to the 100 μM concentrations tested. It should be noted that PAR2 activation has been associated with G_i and G_12/13 to induce cAMP and RhoA signalling, respectively, and that GB88 has been shown to act as a PAR2 agonist on these pathways (Suen et al., 2014). Further studies are required to evaluate the effects, if any, of C391 on these PAR2-dependent responses.

C391 blocked Ca²⁺ and MAPK signalling induced by peptide/peptidomimetics and inhibited PAR2 activation induced by proteolytic cleavage of the N-terminus of the receptor. This is crucial because the mode of action of PAR2 activation in vivo is exposure to either exogenous or endogenous proteinases. A clinical situation where exposure to exogenous proteinases contributes to pathology is in allergy-associated asthma. Several asthma-associated allergens, including those from house dust mite, cockroach and Alternaria alternata, contain trypsin-like serine proteinases that can cleave and activate PAR2 signalling in vitro and in animal models (Schmidlin et al., 2002; Takizawa et al., 2005; Day et al., 2010; Page et al., 2010; Boitano et al., 2011; Nichols et al., 2012; Snelgrove et al., 2014). We have demonstrated here that C391 blocked trypsin-induced PAR2 signalling. Another clinical example of exposure to endogenous proteinases playing a key role in pathology is during inflammation. Proteinases secreted by immune cells during the inflammation process strongly sensitize peripheral nociceptors, causing inflammatory pain. PAR2 is central to this process, as demonstrated in PAR2^−/− mice (Vergnolle et al., 2001). Our findings with the model of thermal hyperalgesia following mast cell degranulation showed that acute antagonism of PAR2 nearly abolished this enhancement in thermal pain sensitivity. These results serve as a proof of concept for further development of PAR2 antagonists as treatments for inflammatory and possibly other types of pain, including visceral and cancer pain.

In conclusion, we have described here a novel, potent and specific antagonist of PAR2, derived by changes to the structure of existing PAR2 agonists. This compound, C391, exhibited PAR2 antagonist properties in the low micromolar range, with some indication of weak partial agonist activity emerging at mid-micromolar concentrations. Despite this weak partial agonist activity, the compound displayed antagonist activity in different signalling assays and in vivo. We conclude that C391 acts as a neutral, orthosteric antagonist of PAR2 at most concentrations, making this compound a new prototype for the further development of PAR2 antagonists.

Glossary

16HBE14o-

immortalized human bronchial epithelial cell line

2-f-LIGRLO-NH₂

2-furoyl-Leu-Ile-Gly-Arg-Leu-Orn(Aloc)-NH₂

2-at-LIGRL-NH₂

2-aminothiazo-4-yl-Leu-Ile-Gly-Arg-Leu-NH₂

[Ca²⁺]_i

intracellular free calcium ion concentration

C391

(2S)-6-amino-N-((S)-1-(((S)-1-amino-3-(4-hydroxyphenyl)-1-oxopropan-2-yl)amino)-4-methyl-1-oxopentan-2-yl)-2-((3S)-1-(furan-2-carbonyl)-3-(3-methylbutanamido)-4,7-dioxooctahydro-8H-pyrazino[1,2-a]pyrimidin-8-yl)hexanamide

GB88

5-isoxazoyl-Cha-Ile-spiroindene-1,4,-piperidine

HBSS

Hank’s balanced salt solution

HR-MS

high-resolution mass spectroscopy

ICW

in-cell Western blotting

K-14585

N-[1-(2,6-dichlorophenyl)methyl]-3-(1-pyrrolidinylmethyl)-1H-indol-5-yl]aminocarbonyl}-glycinyl-L-lysinyl-L-phenylalanyl-N-benzhydrylamide

MCS

matrix coating solution

RTCA

xCELLigence real time cell analyser

Author contributions

S. B., J. H., A. N. F., M. N. A., D. V. T., Z. Z., C. L. S., C. M. R., K. A. D. and J. V. performed the research. S. B., K. A. D., J. V. and T. J. P. designed the research study. S. B., K. A. D., J. V. and T. J. P. contributed essential reagents and tools. S. B., J. H., C. L. S., C. M. R., K. A. D. and T. J. P. analysed the data. S. B., J. H., J. V., K. A. D., T. J. P. wrote the paper.

Conflict of interest

None to declare.