Supplemental material to this article can be found at: http://molpharm.aspetjournals.org/content/suppl/2016/02/25/mol.115.101832.DC1

1521-0111/89/5/593–605$25.00 http://dx.doi.org/10.1124/mol.115.101832 MOLECULAR PHARMACOLOGY Mol Pharmacol 89:593–605, May 2016 Copyright ª 2016 by The American Society for Pharmacology and Experimental Therapeutics

Fevipiprant (QAW039), a Slowly Dissociating CRTh2 Antagonist with the Potential for Improved Clinical Efficacy s

David A. Sykes, Michelle E. Bradley, Darren M. Riddy, Elizabeth Willard, John Reilly, Asadh Miah, Carsten Bauer, Simon J. Watson, David A. Sandham, Gerald Dubois, and Steven J. Charlton Novartis Institutes for Biomedical Research, Horsham, West Sussex, UK (D.A.S., M.E.B., D.M.R., E.W., J.R., A.M., S.W., D.A.S., G.D., S.J.C.); Novartis Institutes for Biomedical Research, Cambridge, Massachusetts (D.A.Sa., J.R.); Novartis Institutes for

Biomedical Research, Basel, Switzerland (C.B.); School of Life Sciences, Queen’s Medical Centre, University of Nottingham, Downloaded from Nottingham, UK (D.A.Sy., S.J.C.) Received September 22, 2015; accepted February 22, 2016

ABSTRACT

Here we describe the pharmacologic properties of a series of reversed by increasing concentrations of PGD2 after an initial molpharm.aspetjournals.org clinically relevant chemoattractant receptor-homologous mole- 15-minute incubation period. This behavior is consistent with cules expressed on T-helper type 2 (CRTh2) receptor antago- its relatively slow dissociation from the human CRTh2 receptor. nists, including fevipiprant (NVP-QAW039 or QAW039), which is In contrast for the other ligands tested this time-dependent currently in development for the treatment of allergic diseases. effect on maximal stimulation was fully reversed by the 15-minute [3H]-QAW039 displayed high affinity for the human CRTh2 time point, whereas QAW039’s effects persisted for .180 minutes. receptor (1.14 6 0.44 nM) expressed in Chinese hamster ovary All CRTh2 antagonists tested inhibited PGD2-stimulated human cells, the binding being reversible and competitive with the native eosinophil shape change, but importantly QAW039 retained agonist prostaglandin D2 (PGD2). The binding kinetics of QAW039 its potency in the whole-blood shape-change assay relative to determined directly using [3H]-QAW039 revealed mean kinetic on the isolated shape change assay, potentially reflective of its 7 -1 21 (kon)andoff(koff) values for QAW039 of 4.5 10 M min and relatively slower off rate from the CRTh2 receptor. QAW039 at ASPET Journals on September 30, 2021 21 0.048 minute , respectively. Importantly, the koff of QAW039 was also a potent inhibitor of PGD2-induced cytokine release (half-life 5 14.4 minutes) was .7-fold slower than the slowest in human Th2 cells. Slow CRTh2 antagonist dissociation reference compound tested, AZD-1981. In functional studies, could provide increased receptor coverage in the face of QAW039 behaved as an insurmountable antagonist of PGD2- pathologic PGD2 concentrations, which may be clinically stimulated [35S]-GTPgS activation, and its effects were not fully relevant.

Introduction Wenzel et al., 1989; Crea et al., 1992) lead to recruitment and activation of Th2 cells and stimulate eosinophil shape The chemoattractant receptor-homologous molecule ex- change and migration (Pettipher et al., 2007). PGD also – 2 pressed on T-helper type 2 cells (CRTh2) is a class A G protein promotes increased CD11b expression, which facilitates coupled receptor involved in the modulation of inflamma- cell adhesion on the vascular cell wall and movement of tory responses. CRTh2 (or DP2) is also expressed on innate eosinophils from the circulation to the site of inflammation immune cells such as eosinophils, basophils (Nagata et al., (Monneret et al., 2001; Gyles et al., 2006). Interestingly, 1999a,b), and group 2 lymphoid cells (ILC2) (Xue et al., 2014) severe asthmatics have been reported to accumulate a and is known to play a role in the pathophysiology of airway particular subtype of PGD2-producing mast cells in the disease. airway submucosa and epithelium resulting in increased Activation of CRTh2 receptors occurs via the natural ligand sputum levels of PGD2 relative to other steroid treated prostaglandin D2 (PGD2), a major product of the cyclo- asthma patients (Balzar et al., 2011). oxygenase pathway released primarily by allergen activated Stimulation of CRTh2 receptors found in the bronchial mast cells (Miadonna et al., 1990) in asthmatic and allergic epithelium has been shown to drive epithelial differentiation, rhinitis patients. High levels of PGD2 (Murray et al., 1986; suggesting that, in addition to its well-characterized role in inflammatory cell migration, CRTh2 may contribute to airway dx.doi.org/10.1124/mol.115.101832. s This article has supplemental material available at molpharm. remodeling in asthmatic patients (Stinson et al., 2015). Pre- aspetjournals.org. clinical studies suggest that CRTh2 antagonism may also

ABBREVIATIONS: CHO, Chinese hamster ovary; CRTh2, chemoattractant receptor-homologous molecule expressed on T-helper type 2; DMSO, dimethylsulfoxide; DR, dose ratio; EC50, molar concentration of an agonist that produces 50% of its maximal possible effect; Emax, maximal agonist effect; Emin, effect in absense of any agonist response; HSA, human serum albumin; IL, interleukin; Kd, equilibrium binding constant; kob, kinetic observed rate of association; koff, kinetic off rate; kon, kinetic on rate; MOA, mechanism of action; PGD2, prostaglandin D2; QAW039, 2-(2-methyl-1- 1(4-(methylsulphonyl)-2-(trifluromethyl)benzyl)-1H-pyrrolo[2,3-b]pyridine-3-yl)acetic acid (fevipiprant); TP, thromboxane A2 receptor; t½, half-life.

593 594 Sykes et al. potentially accelerate apoptosis and clearance of Th2 cells Materials and Methods from inflamed tissue, thereby promoting the resolution of Cell Culture and Membrane Preparation the underlying inflammation in allergic asthma and other related conditions such as allergic rhinitis (Xue et al., 2009). Chinese hamster ovary (CHO) cells expressing the human CRTh2 – 2 In animal models of acute allergen challenge PGD - receptor were grown to 80% 90% confluence in 500 cm cell culture 2 plates at 37°C in 5% CO . All subsequent steps were conducted at 4°C mediated effects are blocked by CRTh2 antagonists acting 2 to avoid receptor degradation. The cell culture media were removed, on CRTh2 receptors on Th2 cells and inflammatory effector and ice-cold HBS-EDTA (1 10 ml; 10 mM HEPES, 0.9% w/v NaCl, cells downstream of the allergic cascade (Uller et al., 2007). 0.2% w/v EDTA pH 7.4) was added to the cells, which were then Several selective orally available antagonists have been since scraped from the plates into a 50-ml Corning tube and subsequently described in the literature, with compounds such as setipi- centrifuged at 250g for 5 minutes to allow a pellet to form. The prant (Diamant et al., 2014), BI-671800 (Krug et al., 2014), supernatant fraction was aspirated, and 10 ml/500 cm2 tray of wash and OC-459 (Barnes et al., 2012) providing clinical proof of buffer (10 mM HEPES, 10 mM EDTA, pH 7.4) was added to the pellet, concept in asthma and allergic rhinitis. A comprehensive which was homogenized using an electrical homogenizer “ultra- review of the current status of these and other compounds, turrax” (position 6, 4 5-second bursts) and subsequently centrifuged including the dual CRTh2/thromboxane A receptor (TP) at 48,000g at 4°C (Beckman Avanti J-251 Ultracentrifuge) for 2 30 minutes. The supernatant was discarded and the pellet rehomoge- antagonist ramatroban (Sugimoto et al., 2003), is available nized and centrifuged, as described, in wash buffer. The final pellet Downloaded from (Norman 2014). was suspended in ice-cold 10 mM HEPES, 0.1 mM EDTA, pH 7.4, at a Drug-receptor interactions are traditionally character- concentration of 5–10 mg ml21. Protein concentration was determined ized in terms of potency and efficacy in preclinical screen- by the Bio-Rad Protein Assay based on the method of Bradford (1976), ing programs; however, increasing evidence suggests that using bovine serum albumin as a standard. Membrane aliquots were the time molecules reside at their cellular target can maintained at 280°C until required. provide an important indicator of their clinical performance molpharm.aspetjournals.org (Swinney, 2009). In functional assays, these slowly dissociat- Compounds ing compounds exhibit the well documented phenomenon of “hemiequilibrium” (Charlton and Vauquelin, 2010), QAW039 was prepared using procedures reported for QAV680 (Sandham et al., 2013), and all reference CRTh2 antagonists were characterized functionally as a depressed agonist re- prepared according to patent procedures referenced in Lamers et al., sponse and one that potentially can aid the selection of (2013). The preparation of [3H]-QAW039 has been previously de- molecules for further preclinical development. For these scribed (see Luu et al., 2015). [3H]-OC-459 was prepared by direct 35 reasons, we used the [ S]-GTPgS assay format to assess hydrogen isotope exchange on solid support. Further details, plus the the ability of CRTh2 antagonists to inhibit PGD2 re- structures, specific activity, and chemical purity of both radioligands, sponses over time. are provided in the Supplemental Materials.

We recently reported the discovery of QAV680, a potent at ASPET Journals on September 30, 2021 selective CRTh2 receptor antagonist, suitable for clinical Common Procedures Applying to [3H]-QAW039 and [3H]- testing in allergic diseases (Sandham et al., 2013). QAV680 OC-459 Radioligand Binding Experiments dosed twice or four times daily was shown to reduce nasal symptom scores after allergen challenge in patients with All radioligand experiments were conducted in 96 deep-well plates in assay binding buffer (Hanks’ balanced salt solution containing allergic rhinitis (Erpenbeck et al., 2014a,b). A follow-up 10 mM HEPES, 1% dimethylsufoxide (DMSO), and 0.02% pluronic compound was consequently sought with improved duration acid, pH 7.4) at 37°C. In all cases, nonspecific binding was determined of action (DOA) giving the potential for once-daily dosing in the presence of 3 mM AZD-1981, a selective CRTh2 antagonist. After to increase patient compliance. This optimization resulted the indicated incubation period, bound and free radioligand were in the discovery of 2-(2-methyl-1-1(4-(methylsulphonyl)-2- separated by rapid vacuum filtration using a FilterMate Cell (trifluromethyl)benzyl)-1H-pyrrolo[2,3-b]pyridine-3-yl)acetic Harvester (PerkinElmer, Beaconsfield, UK) onto 96-well GF/B filter acid (QAW039, fevipiprant), which is currently in clinical plates pretreated with assay buffer and rapidly washed three times development for asthma and atopic dermatitis. with ice-cold wash buffer (20 mM HEPES containing 1 mM MgCl2 Previous work has demonstrated different modes of action pH 7.4). After drying (.4 hours), 40 ml of Microscint 20 (PerkinElmer) (MoA) of antagonists of CRTh2 receptors, for example, studies was added to each well and radioactivity quantified using single photon counting on a TopCount microplate scintillation counter have suggested a noncompetitive MoA for the small-molecule (PerkinElmer). Aliquots of [3H]-QAW039 and [3H]-OC-459 were also antagonist AZD-1981 (Schmidt et al., 2013), whereas struc- quantified accurately to determine how much radioactivity was added turally similar compounds such as OC-459 and ramatroban to each well using liquid scintillation spectrometry on a Hidex 300SL have been shown to be competitive (Pettipher et al., 2012; scintillation counter (LabLogic, Sheffield, UK). In all experiments, Uller et al., 2007; Mathiesen et al., 2006). A complicating total binding never exceeded more than 10% of that added, therefore factor or explanation could be the phenomenon of hemiequilib- limiting significant depletion of the free radioligand concentration rium resulting from slow dissociation, which leads to in- (Carter et al., 2007). To ensure the experiment was at the described surmountable antagonism under certain conditions and can temperature, all components were prewarmed for $2 hours. be inadvertently mistaken for noncompetitive inhibition. The aim of the present study was to determine the binding [3H]-QAW039 and [3H]-OC-459 Saturation Binding characteristics of QAW039 and a cohort of clinically relevant Binding was performed with a range of concentrations of [3H]-QAW039 reference antagonists at the human CRTh2 receptor and to and [3H]-OC-459 to construct saturation binding curves, as described explore the influence of kinetics on compound MoA by in- by Sykes et al., (2010). CHO-CRTh2 cell membranes (0.5 mg well21 vestigating the effect of these ligands on PGD2-stimulated respectively) were incubated in 96-deep well plates at 37°C in assay- CRTh2 responses in both recombinant and relevant inflam- binding buffer with gentle agitation for 3 hours to ensure equilibrium matory cell types. was reached. Owing to the low concentrations of [3H]-QAW039 used, Fevipiprant, a Slowly Dissociating CRTh2 Antagonist 595 the assay volume was increased to 1.5 ml to avoid significant depletion [35S]-GTPgS Functional Binding Assay of the radioligand and unlabeled competitors. A range of concentrations of test compound were prepared in 100% DMSO, and 2.5 ml was added to a 96-well white OptiPlate. To each 3 [ H]-QAW039 Competition Binding well, 200 ml of assay buffer (20 mM HEPES, 10 mM MgCl2, 100 mM NaCl, and 1 mM EDTA) containing 0.1% human serum albumin To obtain affinity estimates for unlabeled QAW039 and the other 21 3 (HSA), 30 mgml saponin (added fresh on the day of experimen- unlabeled CRTh2 reference compounds, [ H]-QAW039 displacement 2 tation), 50 mgml 1 CHO-CRTh2 membranes, 3.7 mM guanosine experiments were performed in a total reaction volume of 0.5 ml. 2 59-diphosphate sodium salt (GDP), 2.5 mg ml 1 WGA PVT SPA [3H]-QAW039 was used at a concentration of approximately 1 nM such beads, and 50 mlof[35S]-guanosine 59-(g-thio)triphosphate that the calculated total binding never exceeded more than 10% of ([35S]-GTPgS) at a concentration of 300 pM was added and incubated that added and therefore avoided ligand depletion. [3H]-QAW039 was at room temperature with gentle agitation for 2 hours. After this incubated in the presence of the indicated concentration of unlabeled incubation, 2.5 ml of increasing concentrations of PGD prepared at CRTh2 antagonist and CHO cell membranes expressing the CRTh2 2 100 concentration in 100% DMSO was added and incubated at room receptor (0.5 mg well21 respectively) at 37°C in assay binding buffer temperature with gentle agitation for 15, 60, 120, and 180 minutes. containing GTPgS (30mM) in 96-deep well plates with gentle agitation Eleven concentrations of PGD were studied in total. After this, the for 2.5 hours to ensure equilibrium was reached. In a separate study 2 plates were centrifuged for 3 minutes at 1000g and bound [35S]-GTPgS designed to assess the MoA of QAW039, increasing concentrations of 3 quantified using single-photon counting on a TopCount microplate

[ H]-QAW039 (0.5, 1, 3, 10, 30 K ) were incubated against increasing Downloaded from d scintillation counter (PerkinElmer). concentrations of unlabeled QAW039 (homologous binding) and in- creasing concentrations of the native agonist PGD2 plus selected CRTh2 antagonists all in the presence of GTPgS (30 mM). Isolated Eosinophil Shape-Change Assay Granulocytes were isolated from the blood of healthy normal 3 volunteers. The granulocytes were incubated with PGD2 for 5 minutes Determination of the Binding Kinetics of [ H]-QAW039 at at 37°C. The reaction was terminated by addition of CellFixTM, and

Human CRTh2 Receptors data were acquired on FACSCalibur flow cytometer, first with FSC/SSC molpharm.aspetjournals.org

To accurately determine the kinetic on (kon) and off (koff) rates, the plots and then FSC/FL-2 to distinguish eosinophils from neutrophils. observed rate of association (kob) was calculated with at least three Data were fitted to a sigmoidal curve (Origin 7 or XL fit) to give a 3 6 5 different concentrations of [ H]-QAW039 (0.3, 1, and 3 Ki). The typical EC50 for PGD2 of 0.72 0.11 nM (n 3, data not shown). appropriate concentration of [3H]-QAW039 was incubated with Antagonist curves were determined as follows: aliquots of granulocyte CRTh2 CHO-cell membranes (0.5 mg well21) in assay binding buffer in assay buffer containing 1% dimethylsulfoxide (DMSO) were treated with gentle agitation (final assay volume 500 ml). The binding of with aliquots of 10 increasing test compound concentrations, and the 3 [ H]-QAW039 to the CRTh2 receptor was initiated with the addition mixture was incubated for 5 minutes at 37°C. Agonist (EC70 concen- of CHO cell membranes, and bound and free [3H]-QAW039 were tration) was added, and the mixture was incubated for a further separated at the indicated time points to construct association kinetic 5 minutes. The reaction was terminated and activity measured as curves. Care was taken to ensure that equilibrium was reached for already described. Data were fitted to a sigmoidal curve (Origin 7 or at ASPET Journals on September 30, 2021 each concentration before the experiment was terminated. After XL Fit) to give an IC50 value. incubation, bound was separated from free by rapid filtration, plates were left to dry, and radioactivity was quantified (all as previously Whole-Blood Eosinophil Shape-Change Assay described). Blood was collected into sterile 3.8% (w/v) trisodium citrate from healthy atopic volunteers and used immediately with minimum 3 Determination of the Dissociation Rate (koff)of[H]- disturbance. Agonist curves were determined as follows: aliquots of QAW039 and [3H]-OC-459 at the Human CRTh2 Receptor whole blood (80 ml) were incubated in assay buffer containing 1% DMSO for 5 minutes at 37°C. Increasing concentrations of PGD were A single concentration of [3H]-QAW039 (2.5 nM) and [3H]-OC-459 2 added, and the mixture was incubated with gentle shaking at 37°C for (5 nM) was allowed to fully associate with human CRTh2 CHO cell 5 minutes. The reaction was terminated by the addition of ice-cold membranes (0.5 mg well21) in a final reaction volume of 500 ml. To CellFixTM (250 ml), followed by ammonium chloride lysis solution construct dissociation kinetic curves, an excess of AZD1981 (3 mM) was (1.5 ml). After incubation on ice for 40–60 minutes, data were acquired added at various time intervals to the reaction mixture to initiate on FACSCalibur flow cytometer, first with FSC/SSC plots and then dissociation of the radioligands. Dissociation was monitored for up to FSC/FL-2 to distinguish eosinophils from neutrophils. Data were 150 minutes until [3H]-QAW039 was fully dissociated from the human fitted to a sigmoidal curve (Origin 7 or XL fit) to give a typical EC for CRTh2 receptor. Bound and free [3H]-QAW039 and [3H]-OC-459 were 50 PGD of 0.94 6 0.18 nM (n 5 3, data not shown). Antagonist curves separated at the indicated time points by rapid filtration, plates were 2 were determined as follows: aliquots of whole blood (80 ml) and test left to dry, and radioactivity was quantified (all as previously compound (10 concentrations) were incubated in assay buffer contain- described). ing 1% DMSO for 5 minute at 37°C. Agonist (EC70 concentration in 10 ml in assay buffer) was added, and the mixture was incubated with Competition Kinetics at Human CRTh2 receptors gentle shaking at 37°C for 5 minutes. The reaction was terminated and samples measured as described herein. The kinetic parameters of unlabeled QAW039 and the other CRTh2 antagonists were assessed using the methodology of Sykes et al., (2010). An experimentally calculated concentration of [3H]-QAW039 Th2 Cell Cytokine-Release Assay (∼2.5 nM, a concentration that avoided ligand depletion in this assay CRTh21 Th2 cells isolated by either MACS or MoFlo (see Supple- volume) was incubated with the indicated concentration(s) of un- mental Materials) were counted and seeded in flat-bottom 96-well labeled competitor in a final volume of assay binding buffer of 500 ml. plates at a concentration of 100,000 cells/well. Cells were seeded in Each time point was conducted on the same 96-deep-well plates X-VIVO 15 medium plus 10% human AB serum and 1% penicillin/ incubated with constant gentle agitation. The experiment was initiated streptomycin. QAW039 was applied to the cells at a range of by the addition of CRTh2 CHO cell membranes (0.5 mgwell21) concentrations between 1000 and 0.3 nM and incubated with the cells containing GTPgS (30 mM), and the plates were harvested (as (37°C, 5% CO2) for 30 minutes. PGD2 then was added to the cells at a previously described) at the indicated time points. final concentration of 200 nM. Unstimulated cells were included in the 596 Sykes et al. assay and cells treated with medium alone; PGD2 at 200 nM 6 Fixing the preceding parameters allowed the following to be 21 21 vehicle (0.01% DMSO) was also included as a control. Once all calculated: k3 5 association rate of unlabeled ligand (M min ), k4 5 additions had been made, the plate was incubated for a further dissociation rate of unlabeled ligand (min21). 3 3 6 hours at 37°C, 5% CO2, after which the plate was centrifuged at Dissociation of [ H]-QAW039 and [ H]-OC-459 was fitted to a one- 400g for 10 minutes, and the supernatant was harvested into 96-well phase monoexponential decay function to estimate the dissociation plates and stored at -20°C before cytokine analysis was carried out. rate of QAW039 and OC-459 directly. 35 The supernatant samples were analyzed for levels of the Th2 PGD2 stimulated [ S]-GTPgS binding data were fitted as sigmoidal cytokines interleukin (IL)-13, IL-5, and IL-4 and also IL-12 (as a (variable slope) curves using the four-parameter logistic equation, and the negative control) using the Mesoscale Discovery human IL-13 issue negative logarithm to base 10 of the molar concentration of an antagonist culture assay kit plus detection antibodies for human IL-4, IL-5, and that makes it necessary to double the concentration of the agonist required IL-12 (total). Briefly, 25 ml of each Calibrator (standard) at a range of to elicit the original submaximal response obtained in the absence of 10,000–2.4 pg/ml was applied to the plate in duplicate and 25 mlof antagonist (pA2). pA2 values were estimated directly from Schild analy- neat test sample in triplicate. The plate was then incubated at room sis after normalization of data to the control maximal response (see temperature for 2 hours with shaking at 800 rpm. Then 25 mlofthe Supplemental Fig. 3 and Supplemental Table 1). Specifically, data were detection antibody solution mix (each antibody diluted 1 in 50 in fitted using a four-parameter logistic equation in Prism 6.0 as described antibody dilution buffer) was added to each well and incubated for a already [each curve was fit to its own value of n, the molar concentration of further 2 hours at room temperature with shaking at 800 rpm. The an agonist that produces 50% of its maximal possible effect (EC50), effect plate was then washed three times in phosphate-buffered saline in the absence of any agonist response (Emin) and maximal agonist effect Downloaded from containing 0.05% Tween-20 and 150 mlof2 Read buffer T was (Emax)] and to a simpler model (where a common Emin, Emax,andn values added to each well. The plate was analyzed on the SECTOR Imager, were used). When the F statistic indicated that a significantly better fit and the results for each cytokine were analyzed using Discovery was obtained with the simple model, dose ratios (DRs) were calculated for

Workbench software. IC50 values were calculated using GraphPad construction of Schild plots. Data were then fitted to the Gaddum-Schild Prism v4.0 software. equation in Prism 6.0 to obtain an estimation of the Schild slope. If the Schild slope did not differ significantly from unity (F test), the data were Interpretation of Data refit to a slope of 1.0 to derive pA2 values. Corrected pA2 estimates were molpharm.aspetjournals.org derived by correcting for % bound to HSA (0.1%) using the method of Data Analysis. As the amount of radioactivity varied slightly for Toutain and Bousquet-Melou (2002). The validity of this correction is each experiment (,5%), data are shown graphically as the mean 6 borne out of the fact that plasma-protein binding (Table 3) and HSA S.D. for individual representative experiments, whereas all values binding are directly correlated for the CRTh2 antagonists tested in this 6 2 reported in the text and tables are mean S.E.M. for the indicated study (P , 0.0001, r 5 0.97 data not shown). For a full explanation of this number of experiments. All experiments were analyzed by non- conversion and the effects of protein on apparent receptor affinity, see linear regression using Prism 6.0 (GraphPad Software, San Diego, Blakeley et al. (2015). PGD2 did not reach equilibrium with QAW039 in U.S.A.). Competition displacement binding data were fitted to thetimeframeoftheGTPgS assay, so EC30 estimates were taken, from sigmoidal (variable slope) curves using a four-parameter logistic which we were able to estimate log DR-1 values to derive a Schild plot equation (eq. 1): and obtain a pA2 value. Estimated pA2 values along with corrected at ASPET Journals on September 30, 2021 estimated pA2 estimates are shown in Supplemental Table 1. ð 2 Þ: Y 5 Bottom 1 ðTop 2 BottomÞ= 1 1 10 logEC50 X Hillcoefficient (1) Functional GTPgS data for QAW039 were fitted according to a com- bined operational and hemiequilibrium model for competitive antagonism

IC50 values obtained from the inhibition curves were converted to Ki under nonequilibrium conditions (Kenakin, 2009; Mould et al., 2014; Riddy values using the method of Cheng and Prusoff (1973). [3H]-QAW039 et al., 2015) to derive kinetic parameters for QAW039 as in eq. 11–14: association data were fitted as follows to a global fitting model using ½B=K GraphPad Prism 6.0 to simultaneously calculate kon and koff using 5 B a ð½ = 1 ½ = 1 Þ (11) the eq. 2, where kob equals the observed rate of association: B KB A KA 1

5 ½ × 1 ½B=KB kob Radioligand kon koff (2) b 5 (12) ð½B=KB 1 1Þ Association and dissociation rates for unlabeled antagonists were ð½B=K 1 ½A=K 1 1Þ calculated using eq. 3–10 described by Motulsky and Mahan (1984): g 5 B A (13) ð½A=KA 1 1Þ K 5 k ½L 1 k (3) A 1 2 ½ = ð 2 ð ×ð 2 2 koff ×g×tÞ 1 × 2 koff ×g×tÞÞ× × 5 A KA 1 a 1 e b e t Em ; Y 2 koff ×g×t 2 koff ×g×t (14) KB 5 k3½I 1 k4 (4) ½A=KAðð1 2 ða×ð1 2 e Þ 1 b×e ÞÞ×t 1 1Þ 1 1 qÀffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiÁ 2 18 2 1 4×k1 ×k3 ×L×I×10 where [A] and [B] represent the concentrations of PGD and S 5 ðKA 2 KBÞ (5) 2 QAW039, respectively; KA and KB represent the respective equilib- KF 5 0:5×ðKA 1 KB 1 SÞ (6) rium dissociation constants; koff is the dissociation rate constant for the antagonist (min21); t is the assay incubation time (min); t is the KS 5 0:5×ðKA 1 KB 2 SÞ (7) operational efficacy of PGD2; and Em is the maximal system response. All parameters were shared across all data sets except t, which were DIFF 5 KF 2 KS (8) fixed to the assay incubation time. In all cases, potency and affinity 2 9 Bmax×K1×L×10 values were estimated as logarithms (Christopoulos, 1998). Data Q 5 (9) DIFF shown are the mean 6 S.E.M. (see Supplemental Fig. 4).   k4 × DIFF k4 2 KF ð 2 × Þ k4 2 Ks ð 2 × Þ Y 5 Q × 1 ×exp KF X 2 ×exp KS X (10) KF×KS KF Ks Results where X is the time (min), Y is the specific binding (cpm), k1 5 kon 3 3 3 [ H]-QAW039, k2 5 koff [ H]-QAW039, L is the concentration of [ H]- QAW039 and Related CRTh2 Compounds. The struc- QAW039 used (nM), Bmax is the total binding (cpm), and I is the ture of QAW039 and related CRTh2 antagonists and the concentration unlabeled antagonist (nM). agonist indomethacin are shown in Fig. 1. QAW039 is a Fevipiprant, a Slowly Dissociating CRTh2 Antagonist 597

Fig. 1. Structures of the selective CRTh2 antagonist QAW039 and reference CRTh2 ligands. Downloaded from molpharm.aspetjournals.org

triuoromethyl derivative of the previously described 7-azaindole- A and B. The expression level of the CHO-CRTh2 in the CHO 3-acetic acid QAV680 (Sandham et al., 2013). The measured cell line and the corresponding equilibrium dissociation 3 3 lipophilicity (logD7.4) of these compounds is lower than other constant (Kd)of[H]-QAW039 and [ H]-OC-459 determined described CRTh2 antagonists (Table 3), which likely contrib- from saturation binding studies are shown in Table 1. 3 utes to their lower plasma-protein binding. Determination of the Dissociation Rate (koff)of[H]- Characterization of CHO-CRTh2-Expressing Cell QAW039 and [3H]-OC-459 at the Human CRTh2 Re-

Line. Specific [3H]-QAW039 and [3H]-OC-459 binding to ceptor. A single concentration of [3H]-QAW039 (2.5 nM) and at ASPET Journals on September 30, 2021 human CRTh2 receptors in CHO membranes was saturable [3H]-OC-459 (5 nM) was allowed to fully associate with the and best described by the interaction of the radioligand with a human CRTh2 receptor. An excess of AZD-1981 (3 mM) was single population of high-affinity sites. A representative then added to the reaction mixture to initiate dissociation of saturation binding curve for [3H]-QAW039 and [3H]-OC-459 [3H]-QAW039 and [3H]-OC-459. Dissociation was monitored binding to the human CHO-CRTh2 receptor is shown in Fig. 2, until [3H]-QAW039, and [3H]-OC-459 was fully dissociated

Fig. 2. Characterization of the CRTh2 radioligands [3H]-QAW039 and [3H]-OC-459. Saturation binding of [3H]-QAW039 (A) and [3H]-OC-459 (B) to human CRTh2 receptors expressed in membranes from CHO cells. Dissociation of [3H]-QAW039 (C) and [3H]-OC-459 (D) from human CRTh2 receptors expressed in membranes from CHO cells. Data are shown as mean 6 S.E.M. and are representative of three separate experiments performed in triplicate. 598 Sykes et al.

Fig. 3. Competition binding between [3H]-QAW039 and CRTh2 receptor ligands for human CRTh2 receptors expressed in membranes from CHO cells. Displacement of [3H]-QAW039 (1 nM) by increasing concentrations of CRTh2 antagonist (A) QAW039, QAV680, AZD-1981, OC-459, and BI-671800 and (B) setipiprant and ramatroban plus the CRTh2 ago- nists PGD2 and indomethacin. Data are presented as the mean range from a representative of three experiments performed in duplicate.

from the human CRTh2 receptor (Fig. 1, C and D). The data agreement, with values estimated from the global fit de- 21 were fitted to a monoexponential decay function to estimate a scribed herein (linear fit mean koff 5 0.044 6 0.004 min and 2 2

7 1 1 Downloaded from koff value for QAW039 and OC-459 directly. A koff value of kon 5 4.86 6 0.53 10 M min ). 21 0.050 6 0.006 minute [half-life (t1/2) 5 14.2 minutes] was Determination of the Kinetic Parameters of QAW039 determined for the slowly dissociating [3H]-QAW039 and a and Competitor Compounds at Human CRTh2 Receptors. 21 value of 2.80 6 0.53 minute (t1/2 5 0.28 minutes) for the Using the equations described in Materials and Methods and more rapidly dissociating [3H]-OC-459 binding to human the kinetic values obtained for [3H]-QAW039, as detailed CRTh2 receptors. already, the association and dissociation rates for unlabeled

Determination of the Equilibrium Binding Affinity QAW039, OC-459, and the other CRTh2 antagonists for the molpharm.aspetjournals.org Constant of CRTh2 Antagonists and Agonists at CRTh2 human CRTh2 (Figs. 4, C and D; Table 2; and Supplemental Receptors. IC50 values determined from competition bind- Fig. 1, A–E, respectively) receptor were calculated. The kinetic ing experiments were converted to equilibrium binding con- on (kon) and off rates (koff) determined for QAW039 using the stants (pKi) by using the equation of Cheng and Prusoff (1973) Motulsky-Mahan method were not significantly different from 3 and using the Kd value of [ H]-QAW039 as described in the values calculated from the global kinetic analysis, sug- Materials and Methods. The competition binding curves gesting that [3H]-QAW039 is suitable for the determination of showing displacement of [3H]-QAW039 by QAW039 and the kinetic parameters of unlabeled CRTh2 antagonists. To vali- other CRTh2 compounds tested are shown in Fig. 3, A and B, date the rate constants, the kinetically derived dissociation

and the affinity values (pKi) for CRTh2 antagonist and agonist constant (Kd)values(koff/kon) were compared with the dissoci- at ASPET Journals on September 30, 2021 binding to human CRTh2 receptors are summarized in ation constant (Ki) obtained from equilibrium competition Table 2. binding experiments. Good correlation (r2 5 0.95, P , 0.001, Determination of the Kinetic Parameters of [3H]-QAW039 data not shown) was found between these two values, indicating at the Human CRTh2 Receptor. To calculate the kinetic on that the kinetic parameters were accurate. (kon) and off rates (koff) of the unlabeled QAW039 and the other QAW039 is a Competitive Antagonist of the Native CRTh2 antagonists, the kinetic parameters of the radioligand, CRTh2 Receptor Agonist PGD2. To explore the nature of [3H]-QAW039 were first determined. A family of association the interaction of QAW039 with the CRTh2 receptor we used kinetic curves was constructed using a range of [3H]-QAW039 the technique of competitive binding. Using this technique, we concentrations. Each association curve was monitored until have explored the inhibitory potency of unlabeled QAW039, equilibrium so that Ymax was reached (Fig. 4A). From this type AZD-1981, OC-459, and the native agonist PGD2 over a range 3 of analysis, it is possible to estimate both kon and koff values for of concentrations of [ H]-QAW039 designed to saturate the QAW039 by fitting the data to a global kinetic model. A kon receptor. At the lowest radioligand concentration used, ∼0.3 nM, 3 7 21 21 value for [ H]-QAW039 of 4.52 6 0.45 10 M min and a the pIC50 value for displacement by QAW039 itself was 9.05 6 21 koff value of 0.048 6 0.005 min was determined. From the koff 0.10 (Fig. 5A). As expected, a linear increase in the pIC50 value value, a t1/2 of dissociation (0.693/koff) of 14.4 minutes was was obtained as the radioligand concentration was increased 3 calculated for [ H]-QAW039. Kob increased in a linear manner to the maximum concentration used in the study, ∼30 nM 3 with [ H]-QAW039 concentration (Fig. 4B), and the mean (∼30-fold greater than Kd from saturation binding) (Fig. 5E). kinetic parameters estimated from these plots were in good This same pattern of behavior was mimicked by the other

TABLE 1 3 3 Saturation binding and dissociation kinetic rate constant (koff) data for [ H]-QAW039 and [ H]-OC-459 binding to human CRTh2 receptors Data are mean 6 S.E.M. from at least three separate experiments.

Equilibrium Radioligand Receptor Expression Level Dissociation Kinetic Rate Constant koff Dissociation Constant Kd

pmol mg21 min21 nM [3H]-QAW039 6.26 6 0.51 0.050 6 0.006 1.14 6 0.44 [3H]-OC-459 6.44 6 0.06 2.80 6 0.53 2.44 6 0.58 Fevipiprant, a Slowly Dissociating CRTh2 Antagonist 599

TABLE 2 Kinetic and equilibrium binding parameters of unlabeled QAW039 and other CRTh2 ligands interacting with the human CRTh2 receptor Data are mean 6 S.E.M. from at least three separate experiments.

Equilibrium CRTh2 Ligand kon koff t1/2 Kinetic pKd pKi

M-1min21 min21 min QAW039 6.27 6 0.93 107 0.061 6 0.006 12.04 8.99 6 0.06 8.98 6 0.03 QAV680 4.80 6 2.14 107 0.66 6 0.18 1.29 7.82 6 0.04 7.54 6 0.09 AZD-1981 3.01 6 0.76 108 0.77 6 0.23 1.26 7.58 6 0.02 8.49 6 0.03 OC-459 9.50 6 5.14 108 1.83 6 0.32 0.41 8.61 6 0.17 8.35 6 0.13 BI-671800 2.47 6 0.36 108 1.22 6 0.11 0.58 8.30 6 0.01 8.14 6 0.20 Ramatroban 1.89 6 0.66 107 0.80 6 0.22 1.10 7.33 6 0.05 7.45 6 0.02 Setipiprant 1.87 6 0.45 108 0.95 6 0.15 0.76 8.28 6 0.04 8.35 6 0.06 PGD2 nd nd nd nd 6.41 6 0.08 Indomethacin nd nd nd nd 5.74 6 0.12

nd, Not determined. Downloaded from

35 structurally related CRTh2 antagonists and PGD2 which were antagonist (Fig. 6). [ S]-GTPgS binding to CHO-hCRTh2 investigated in parallel. The difference in pIC50 values for membranes was measured at different time intervals after unlabeled PGD2 were precisely in line with those predicted by incubation with the endogenous CRTh2 agonist PGD2. the Cheng-Prusoff relationship (Cheng and Prusoff, 1973), QAW039 produced a clear reduction in the Emax of PGD2 using the Kd value for QAW039 obtained from saturation stimulated GTPgS binding after a 15-minute incubation molpharm.aspetjournals.org binding (Table 1), suggesting that unlabeled PGD2 displaced period, with PGD2 (Fig. 6A) demonstrating its ability to the radioligand competitively. produce insurmountable antagonism in a functionally rele- Determination of the Effect CRTh2 Antagonists on vant second-messenger reporter assay. The incubation was 35 PGD2 Stimulated [ S]-GTPgS Binding to Human extended to 180 minutes to allow the PGD2 to reach equilib- CRTh2 Receptors. To investigate the potential for rium with the receptor and fully activate a second-messenger QAW039 and the other CRTh2 antagonists to display in- response. The inhibitory effect of QAW039 was readily re- 35 surmountable behavior in a [ S]-GTPgS binding assay, PGD2 versible with time; however, the inhibitory effects of QAW039 concentration response curves were constructed in the were not fully reversed even after a 180-minute incubation presence of increasing concentrations of pre-equilibrated period with PGD2 (see Supplemental Fig. 2 and Supplemental at ASPET Journals on September 30, 2021

Fig. 4. Direct determination of [3H]-QAW039 kinetic binding parameters plus competition kinetic binding to determine the kinetic rate constants of unlabeled compounds. (A) Association of [3H]-QAW039 to the human CRTh2 receptor expressed in membranes from CHO cells. A family of association 3 kinetic curves was constructed using a range of [ H]-QAW039 concentrations. The kon and koff values for QAW039 were determined by fitting the data to a global kinetic association model. (B) Plot of ligand concentration verses kob. Binding followed a simple law of mass action model, kob increasing in a linear manner with radioligand concentration. [3H]-QAW039 competition kinetic curves in the presence of unlabeled QAW039 (C) and OC-459 (D). CHO-CRTh2 membranes were incubated with ∼2.5 nM [3H]-QAW039 and increasing concentrations of unlabeled competitor. Plates were incubated at 37°C for the indicated time points and nonspecific binding levels were determined in the presence of 3 mM AZD-1981. Data were fitted to the equations described in Materials and Methods to calculate kon and koff values for the unlabeled ligands; these are summarized in Table 2. Data are presented as mean 6 range from representative three or more experiments performed in duplicate. 600 Sykes et al.

Fig. 5. Determination of the competitive mode of action of QAW039 with PGD2 and the selected other CRTh2 antagonists. Effect of increasing concentrations of [3H]-QAW039 on the degree of inhibition observed with increasing concentrations of Downloaded from (A) QAW039, (B) PGD2, (C) AZD-1981, and (D) OC-459 at human CRTh2 receptors expressed in membranes from CHO cells. (E) Linear increase in the compound pIC50 values with increasing [3H]-QAW039 con- centrations. Data are shown as mean 6 range and are representative of three separate experiments performed in duplicate. molpharm.aspetjournals.org at ASPET Journals on September 30, 2021

Fig. 3). Also apparent was a reduction in the basal accumu- Supplemental Fig. 3G); however, Schild plots of the lower lation of GTPgS binding in the presence of increasing concentrations of QAW039, QAV680, AZD-1981, OC-459, BI- concentrations of QAW039 and the other CRTh2 antagonists 671800, and ramatroban (,10 mM, i.e., below the saturating (Fig. 6). This finding demonstrates that these compounds protein concentration) tested resulted in Schild slopes not behave as inverse agonists of CRTh2 mediated GTPgS significantly different from 1.0 and pA2 values and in line accumulation. In contrast to QAW039, the more rapidly with expected pKi estimates after correction for HSA bind- dissociating CRTh2 antagonists tested did not produce any ing, further suggesting that these compounds are behaving noticeable depression of the PGD2 maximal response after competitively with PGD2 (see Table 2, Supplemental Fig. 3, 15 minutes of incubation (Fig. 6, B–G). PGD2-stimulated and Supplemental Table 1). Although PGD2 did not fully [35S]-GTPgS binding in the presence of these more rapidly reach equilibrium with QAW039 in the 180-minute time dissociating CRTh2 compounds after a 180-minute incuba- frame of the GTPgS assay, we previously determined it to be tion period is shown for comparison with QAW039 (see competitive with PGD2 based on competitive binding data; Supplemental Fig. 2 and 3, B–G, respectively). hence, we have estimated DR-1 values to derive a Schild plot 35 It is not possible to accurately derive pKB values directly and obtain an estimated pA2 value. Thus, the [ S]-GTPgS from the GTPgS binding experiments since the compounds are competition assay data are entirely consistent with the bound to plasma proteins (e.g., HSA; Table 3), resulting in radioligand competition binding analysis and suggest that irregular shifts of the PGD2 concentration response curves at these molecules interact with the same orthosteric binding concentrations of compound that come close to saturating the pocket as PGD2. protein (0.1% HSA or 15 mM) present in the assay buffer. Once When a comparison of CRTh2 antagonist inhibitory potency HSA gradually becomes saturated by the increasing concen- was made at the 15-minute time point in the [35S]-GTPgS trations of antagonist, the Schild slopes become gradually assay, after the addition of a maximally effective concentra- steeper as more antagonist becomes free in solution and tion of PGD2 (1mM), the superior potency of QAW039 com- available to bind the CRTh2 receptor (i.e., deviation from pared with the other CRTh2 compounds tested in this study constant % bound; see Blakeley et al., 2015). As would be was clearly visible (Fig. 7). This result suggests that the predicted, this effect is most pronounced for the low CRTh2 kinetic off rate potentially could play a role in driving the affinity, highly protein-bound compound setipiprant (see superior potency of this compound. Fevipiprant, a Slowly Dissociating CRTh2 Antagonist 601 Downloaded from Fig. 6. Effect of CRTh2 antagonists on PGD2- stimulated GTPgS accumulation. Insurmountable behavior of (A) QAW039 at human CRTh2 receptors expressed in membranes from CHO cells after a 15-minute incubation period with the agonist PGD2. Surmountable behavior of (B) QAV680, (C) AZD-1981, (D) OC-459, (E) BI-671800, (F) ramatroban, and (G) setipiprant after a 15-minute incubation period. Data are shown as mean 6 S.E.M. and are representative of molpharm.aspetjournals.org three separate experiments performed in singlet. at ASPET Journals on September 30, 2021

The effect of time on the potency of CRTh2 antagonists to compared with the full 180-minute incubation period (Fig. inhibit PGD2 (1mM)-stimulated GTPgS accumulation is 8A), reflecting its very slow dissociation from the CRTh2 shown in Fig. 8. QAW039 is .10-fold more effective at receptor in the face of high concentrations of PGD2.Theother inhibiting the PGD2 response after a 15-minute incubation CRTh2 antagonist examined are equieffective at all time points,

TABLE 3 CRTh2 antagonist human whole-blood (WBSC) and isolated cellular (ISC) shape-change potency values plus protein binding and physicochemical descriptors clogP, logD7.4, and logKIAM Data are mean 6 S.E.M. for three or more experiments.

% Human Plasma CRTh2 Antagonist cLogP Log D7.4alogKIAM7.4b Protein Unboundc WBSC IC50 ISC IC50

nM nM QAW039 2.26 0.65 1.12 7.8 0.44 6 0.19 0.40 6 0.04 QAV680 1.15 20.34 0.55 21.4 31.0 6 1.0 5.0 6 1.0 AZD-1981 4.79 1.19 1.77 1.8 13.10 6 4.03 2.21 6 0.87 OC-459 4.57 1.32 1.91 0.5 422.5 6 155.9 6.4 6 1.4 BI-671800 4.58 1.65 1.95 0.4 104.9 6 35.3 1.87 6 0.88 Ramatroban 3.97 1.34 1.89 2.7 195 6 75.4 11.4 6 1.9 Setipiprant 4.28 1.35 1.88 0.4 .500 10.7 6 1.7

aDetermined by high-performance liquid chromatography (HPLC) column method (Kerns et al., 2003). bDetermined by HPLC column method (Valko et al., 2000). cDetermined by equilibrium dialysis (Banker et al., 2003). 602 Sykes et al.

Determination of the Effect of QAW039 on Cytokine Release from PGD2 Stimulated CD41 Th2 Cells. The chronic inflammatory response characteristically observed in allergic asthma occurs by the selective accumulation of Th2 lymphocytes, which further potentiate the inflammatory response by releasing Th2 cytokines such as IL-5 and IL-13. The localized increase in the release of Th2 cytokines at the primary site of inflammation causes activation of Th2 cells, eosinophils, and basophils. Here we have shown that QAW039 is a potent inhibitor of PGD2-induced release of IL-5 and IL-13 Fig. 7. Functional consequences of hemiequilibrium. Effect of CRTh2 from human CD41 Th2 lymphocytes ( Table 4). The potency of antagonists acting at human CRTh2 receptors expressed in membranes from CHO cells on % depression of a maximally effective concentration of QAW039 in this cellular functional assay system is 15- to 35 PGD2 (1mM) in the [ S]-GTPgS assay at the 15-minute time point. 20-fold greater than the previous lead compound QAV680. Antagonist potency is expressed as a ratio of concentration added over compound affinity corrected for % HSA bound. Data are shown as mean 6 S.E.M. from three separate experiments performed in singlet. Discussion It is potentially advantageous when developing antago- Downloaded from demonstrating that kinetics does not play any role in de- nists of G protein–coupled receptors to identify compounds termining their potency under these experimental conditions that exhibit slow dissociation from the receptor (Tummino (Fig. 8, B–G). and Copeland 2008). CRTh2 antagonists exhibiting such An estimate of the affinity and dissociation rate of the behavior may display greater clinical efficacy as a conse- CRTh2 antagonist QAW039 from functional [35S]-GTPgS quence of prolonged receptor blockade, which will effectively data after a 60-minute incubation was obtained by fitting reduce the recruitment of proinflammatory eosinophils into molpharm.aspetjournals.org data according to an operational hemiequilibrium model for existing areas of inflammation. Such a profile could be competitive antagonism first described by Kenakin (2009) particularly important in the asthmatic lung, where local and recently used practically to estimate the dissociation concentrations of PGD2 are likely to vary depending on the rates of orexin-2 antagonists and muscarinic M1 antagonists numbers of activated mast cells, especially given that PGD2 in functional assays (Mould et al., 2014; Riddy et al., 2015) release occurs in a matter of minutes after IgE activation (see Supplemental Fig. 4 and Supplemental Table 2). A pA2 (Kawata et al., 1995). Importantly, the transit time of mixed value for QAW039 of 3.2 6 0.7 nM was estimated, which is leukocytes across the pulmonary circulation has recently very much in line with affinity estimates obtained from been quantified and is thought to be similar to that of kinetic association experiments performed at room temper- neutrophils measured precisely at 14.2 seconds, only margin- at ASPET Journals on September 30, 2021 ature (Kd 1.39 6 0.23 nM; see Supplemental Fig. 5 and ally slower than that reported for RBCs (Summers et al., Supplemental Table 3) and the pA2 obtained from the Schild 2014). This implies that sustained CRTh2 blockade will analysis (see Supplemental Table 1). Importantly, the koff protect against eosinophil recruitment as they transit 21 value for QAW039 of 0.005min was also within 3-fold of through areas of high PGD2 concentrationinthelung. that estimated in kinetic association binding experiments Clinical data generated with QAW039 in eosinophilic asth- performed at room temperature (QAW039 koff 5 0.012 6 matics show a reduction in both sputum and tissue bronchial 0.001 min21; see Supplemental Fig. 5 and Supplemental biopsy eosinophils, supporting this hypothesis (Berair et al., Table 3), demonstrating the potential for the GTPgS assay to 2015). Prolonged CRTh2 receptor blockade therefore repre- identify very slowly dissociating compounds. In contrast, it sents a new and potentially powerful approach for the was not possible to fit other ligands to the operational treatment of allergic asthma. hemiequilibrium model as their effects were fully reversible The binding properties of several clinically relevant CRTh2 within the 15-minute incubation period. Confirmation that antagonists, including the recently developed QAW039, were [3H]-OC-459 was relatively rapidly dissociating when pro- determined at the human CRTh2 receptor in assay buffer filed at room temperature comes from kinetic association containing physiologic sodium ion concentration at 37°C. binding experiments (see Supplemental Fig. 5 and Supple- Previous studies designed to assess the kinetics of unlabeled mental Table 3). CRTh2 antagonists have used nonphysiologic assay condi- Determination of the Effect of QAW039 on PGD2 tions, experiments being conducted at room temperature in Stimulated Whole-Blood Eosinophil Shape-Change. the absence of salt (Gervais et al., 2011). Findings from these Migration of eosinophils from the microcirculation into sites early kinetic studies are therefore likely to overestimate the of inflammation occurs as a consequence of cytoskeletal receptor target coverage since it has been well documented rearrangements in a process known as eosinophil shape that compound kinetics is highly dependent on temperature change. Purified human eosinophils undergo shape change on and sodium ion concentration(Sykesetal.,2012).This exposure to PGD2 in response to activation of CRTh2 and phenomenon was also apparent in the current study, with the a similar response is seen in whole blood. This study con- off rates of QAW039 and OC-459 being up to 5-fold slower firms the inhibitory effect of the tested CRTh2 antagonists in when measured at room temperature compared with 37°C. both the isolated and whole-blood shape-change assays. Other researchers have attempted to estimate dissociation 35 QAW039 was the most potent compound tested at inhibiting t1/2s indirectly from functional [ S]-GTPgS assays performed PGD2 stimulated eosinophil shape change, with compara- at room temperature by measuring responses to PGD2 (100 mM) ble potencies in both the isolated and whole blood shape- at fixed time periods following a 60-minute preincubation change assays (Table 3). period with test compounds, allowing antagonist t1/2 values to Fevipiprant, a Slowly Dissociating CRTh2 Antagonist 603 Downloaded from Fig. 8. Effects of hemiequilibrium on CRTh2 antagonist potency. The effect of time on the inhibitory potency of CRTh2 antagonists (A) QAW039, (B) QAV680, (C) AZD-1981, (D) OC-459, (E) BI-671800, (F) ramatroban, and (G) setipiprant in the [35S]-GTPgS binding assay after stimula-

tion of the CRTh2 receptor expressed in molpharm.aspetjournals.org membranes from CHO cells with a maximal effective concentration of PGD2 (1 mM). Data are shown as mean 6 S.E.M. and of three separate experiments performed in singlet. at ASPET Journals on September 30, 2021

3 3 be estimated by fitting an exponential decay curve to the PGD2 In saturation binding experiments, [ H]-QAW039 and [ H]-OC- responses observed over time (Andrés et al., 2014). Although 459 display nM affinity for the human CRTh2 receptor expressed these results enabled a rank order of residence time to be in CHO cells. The mean kinetic kon and koff values determined established, the use of an accumulation assay mean the for [3H]-QAW039 were 4.5 107 M-1min21 and 0.048 min21, dissociation t½ values are likely to be overestimated. respectively, producing a t1/2 value of 12 minutes and a kinetic Kd value of 1.06 6 0.06 nM, which is in excellent agreement with the equilibrium Kd value obtained in the saturation binding experi- TABLE 4 ments described here. In contrast, despite displaying comparable ∼ Comparison of potencies of CRTh2 antagonists QAV680 and QAW039 for nM affinity, OC-459 was 30 faster at dissociating from the inhibition of PGD2-induced cytokine release in human CD4+ Th2 CRTh2 receptor (t1/2 5 0.41 minute). To put this in context, .50% lymphocytes of OC-459 could potentially be displaced from the CRTh2 recep- 6 Data are mean S.E.M. from at least three separate experiments. tor in the time it takes for eosinophils to pass through the asth- Inhibition of IL-5 Inhibition of IL-13 matic lung, whereas in direct contrast, ,1.2% of QAW039 will CRTh2 Antagonist Release IC Release IC 50 50 likely become displaced in the same time frame. nM nM The binding kinetics of unlabeled QAW039 and the other QAW039 2.56 6 1.24 1.40 6 0.30 related CRTh2 antagonists including OC-459 were also de- 6 6 QAV680 57 28 29 5 termined indirectly using [3H]-QAW039 as the tracer through 604 Sykes et al. application of a competition binding model first described by Supplemental Fig. 3), support the findings of homologous Motulsky and Mahan (1984). The kinetic Kd values generated binding studies (see Fig. 5) and demonstrate that QAW039 indirectly were in agreement with the Ki values generated by behaves as a simple competitive antagonist of PGD2 func- equilibrium competition kinetic binding, validating use of the tional responses. current methods. The pharmacologic properties of QAW039 To investigate the profile QAW039 at the cellular level, we are a significant improvement on those of our original lead chose two relevant inflammatory cell types. CRTh2-mediated compound, QAV680, and the other CRTh2 antagonists exam- shape change in eosinophils was used to profile QAW039 in ined in this study. In particular, QAW039 is significantly whole blood and represents a physiologically relevant envi- slower to dissociate from the CRTh2 receptor (.9-fold com- ronment. The comparable IC50 values for QAW039 in the pared with QAV680), which potentially affords this compound whole blood and isolated shape-change assays are consistent increased efficacy and an improved duration of action at the with its lower plasma-protein binding and its relatively slow CRTh2 receptor due to sustained blockade of the receptor in dissociation kinetics that drive its increased potency (Fig. 8A). the face of increasing concentrations of PGD2. This phenom- Other compounds showed large discrepancies between these enon of insurmountability is highlighted in the GTPgS assay, two assay formats, which are likely the result of higher where QAW039, which exhibits the slowest dissociation rate plasma-protein binding coupled with rapid dissociation from constant, also produced the most pronounced suppression of the CRTh2 receptor. In the case of QAV680, the reduced Downloaded from the maximal response to PGD2 in the GTPgS binding assay. protein binding does not compensate for the faster dissociation The slower the compound koff, the greater the % inhibition kinetics. These data demonstrate that QAW039 is highly achieved in the GTPgS binding assay for any given multiple of potent in whole-blood systems, with the IC50 value obtained the compound Ki added. QAW039 displays a time-dependent consistent with the affinity values calculated from radioligand inhibitory effect in the GTPgS binding assay, its potency being experiments. In a further disease-relevant cellular context, highest at early time points (Fig.8A),characteristicofits the potency of QAW039 in the isolated Th2 cell cytokine slow dissociation from the CRTh2 receptor. This effect may inhibition assay is consistent with its CRTh2 receptor affinity, molpharm.aspetjournals.org contribute to QAW039’s increased potency in the eosinophil and, as with eosinophil assay readouts, this represents an whole-blood shape-change assay relative to the more rap- improved potency compared with QAV680. idly dissociating reference compounds. In summary, in these human functional studies across The detailed pharmacologic analysis presented in this study diverse cellular systems, we have determined that the potency confirms the competitive nature of QAW039 and structurally of QAW039 is cell type and cell function independent. related ligands binding to the human CRTh2 receptor. In Importantly, the pharmacologic profile determined for particular, the observation that the pIC50 for PGD2 displace- QAW039 suggests it should competitively inhibit other dis- ment of [3H]-QAW039 binding (to the CRTh2 receptor) in- ease relevant CRTh2-mediated responses in human cells, creased in a linear fashion with equivalent pA2 values at including Th2 cell cytokine production, PGD2-mediated Th2 at ASPET Journals on September 30, 2021 increasing radioligand concentration demonstrates that cell apoptosis, basophil chemotaxis, and eosinophil activation. QAW039 behaves as a simple competitive antagonist. Simi- To our knowledge, this is the first study to demonstrate in both larly, PGD2 itself and the two other reference CRTh2 antag- binding and functional studies the simple competitive antag- onists examined, OC-459 and AZD-1981, appeared fully onist nature of clinically relevant CRTh2 antagonists, in- competitive, with increasing concentrations of QAW039 dem- cluding the slowly dissociating compound QAW039, and onstrating that all these molecules occupy the same CRTh2 represents an important step forward in our understanding receptor binding pocket. of the pharmacology of low-molecular-weight CRTh2 antago- Limited information exists on the MoA of other CRTh2 nists in competition with their native agonist PGD2. QAW039 antagonists, but previous publications using functional assays (fevipiprant) is under ongoing clinical study as a novel oral have highlighted potential insurmountable antagonist pro- therapy for allergic diseases. files (Mathiesen et al., 2006; Gervais et al., 2011). Both publications attribute this compound profile to slow receptor Authorship Contributions dissociation kinetics whereby equilibrium between PGD2 and Participated in research design: Sykes, Dubois, Charlton. receptor is not achieved in the time frame of their experi- Conducted experiments: Sykes, Riddy, Bradley, Willard, Reilly, ments. In contrast, AZD-1981 has been shown to exhibit a Miah. noncompetitive binding profile, despite possessing rapid Contributed new reagents or analytic tools: Bauer, Watson. dissociation kinetics (Schmidt et al., 2013), as demonstrated Performed data analysis: Sykes, Riddy, Bradley, Willard, Reilly, Miah. using an agonist ([3H]-PGD ) binding assay and could poten- 2 Wrote or contributed to writing on manuscript: Sykes, Sandham, tially result from stabilization of the uncoupled form of the Charlton. CRTh2 receptor. In functional studies performed at room temperature and in References salt containing buffer, QAW039 behaves as an insurmount- Andrés M, Buil MA, Calbet M, Casado O, Castro J, Eastwood PR, Eichhorn P, Ferrer M, Forns P, and Moreno I et al. (2014) Structure-activity relationships (SAR) and able antagonist of PGD2-stimulated GTPgS activation after a structure-kinetic relationships (SKR) of pyrrolopiperidinone acetic acids as CRTh2 15-minute agonist equilibration time. This insurmountable antagonists. Bioorg Med Chem Lett 24:5111–5117. ’ Balzar S, Fajt ML, Comhair SA, Erzurum SC, Bleecker E, Busse WW, Castro M, behavior is entirely consistent with QAW039 s relatively slow Gaston B, Israel E, and Schwartz LB et al. (2011) Mast cell phenotype, location, dissociation (t1/2 5 14.4 minutes) from the human CRTh2 and activation in severe asthma. Data from the Severe Asthma Research Program. Am J Respir Crit Care Med 183:299–309. receptor, but it does not result from a noncompetitive in- Banker MJ, Clark TH, and Williams JA (2003) Development and validation of a 96- teraction. Thus, the time-dependent concentration-dependent well equilibrium dialysis apparatus for measuring plasma protein binding. J Pharm Sci 92:967–974. rightward shifts observed with QAW039 in the human GTPgS Barnes N, Pavord I, Chuchalin A, Bell J, Hunter M, Lewis T, Parker D, Payton M, Collins assay, coupled with the Schild analysis (Fig. 8A; and see LP, and Pettipher R et al. (2012) A randomized, double-blind, placebo-controlled Fevipiprant, a Slowly Dissociating CRTh2 Antagonist 605

study of the CRTH2 antagonist OC000459 in moderate persistent asthma. Clin Exp Nagata K, Hirai H, Tanaka K, Ogawa K, Aso T, Sugamura K, Nakamura M, Allergy 42:38–48. and Takano S (1999b) CRTH2, an orphan receptor of T-helper-2-cells, is expressed Berair R, Gonem S, Singapuri A, Hartley R, Laurencin M, Bacher G, Holzhauer B, on basophils and eosinophils and responds to mast cell-derived factor(s). FEBS Lett Bourne M, Mistry V, Pavord I, Mansur AH, Wardlaw A, Siddiqui S, Kay R, 459:195–199. and Brightling CE. Effect of QAW039, an oral prostaglandin D2 receptor (DP2/ Nagata K, Tanaka K, Ogawa K, Kemmotsu K, Imai T, Yoshie O, Abe H, Tada K, CRTh2) antagonist, upon sputum and bronchial eosinophilic inflammation and Nakamura M, and Sugamura K et al. (1999a) Selective expression of a novel clinical outcomes in treatment-resistant asthma: a phase 2a randomized placebo- surface molecule by human Th2 cells in vivo. J Immunol 162:1278–1286. controlled trial. Am. J. Respir. Crit. Care. Med., 2015, 191, A636. Norman P (2014) Update on the status of DP2 receptor antagonists; from proof of Blakeley D, Sykes DA, Ensor P, Bertran E, Aston PJ, and Charlton SJ (2015) Sim- concept through clinical failures to promising new drugs. Expert Opin Investig ulating the influence of plasma protein on measured receptor affinity in bio- Drugs 23:55–66. chemical assays reveals the utility of Schild analysis for estimating compound Pettipher R, Hansel TT, and Armer R (2007) Antagonism of the prostaglandin D2 affinity for plasma proteins. Br J Pharmacol 172:5037–5049. receptors DP1 and CRTH2 as an approach to treat allergic diseases. Nat Rev Drug Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram Discov 6:313–325. quantities of protein utilizing the principle of protein-dye binding. Anal Biochem Pettipher R, Vinall SL, Xue L, Speight G, Townsend ER, Gazi L, Whelan CJ, Armer 72:248–254. RE, Payton MA, and Hunter MG (2012) Pharmacologic profile of OC000459, a Carter CM, Leighton-Davies JR, and Charlton SJ (2007) Miniaturized receptor potent, selective, and orally active D prostanoid receptor 2 antagonist that inhibits binding assays: complications arising from ligand depletion. J Biomol Screen 12: mast cell-dependent activation of T helper 2 lymphocytes and eosinophils. 255–266. J Pharmacol Exp Therap 340:473–482. Charlton SJ and Vauquelin G (2010) Elusive equilibrium: the challenge of inter- Riddy DM, Valant C, Rueda P, Charman WN, Sexton PM, Summers RJ, Christopoulos preting receptor pharmacology using calcium assays. Br J Pharmacol 161: A, and Langmead CJ (2015) Label-free kinetics: exploiting functional hemi- 1250–1265. equilibrium to derive rate constants for muscarinic receptor antagonists. Mol Phar- Cheng Y and Prusoff WH (1973) Relationship between the inhibition constant (K1) macol 88:779–790. and the concentration of inhibitor which causes 50 per cent inhibition (I50) of an Sandham DA, Arnold N, Aschauer H, Bala K, Barker L, Brown L, Brown Z, Budd D,

enzymatic reaction. Biochem Pharmacol 22:3099–3108. Cox B, and Docx C et al. (2013) Discovery and characterization of NVP-QAV680, a Downloaded from Christopoulos A (1998) Assessing the distribution of parameters in models of ligand- potent and selective CRTh2 receptor antagonist suitable for clinical testing in receptor interaction: to log or not to log. Trends Pharmacol Sci 19:351–357. allergic diseases. Bioorg Med Chem 21:6582–6591. Crea AE, Nakhosteen JA, and Lee TH (1992) Mediator concentrations in broncho- Schmidt JA, Bell FM, Akam E, Marshall C, Dainty IA, Heinemann A, Dougall IG, alveolar lavage fluid of patients with mild asymptomatic bronchial asthma. Eur Respir Bonnert RV, and Sargent CA (2013) Biochemical and pharmacological character- J 5:190–195. ization of AZD1981, an orally available selective DP2 antagonist in clinical de- Diamant Z, Sidharta PN, Singh D, O’Connor BJ, Zuiker R, Leaker BR, Silkey M, velopment for asthma. Br J Pharmacol 168:1626–1638. and Dingemanse J (2014) Setipiprant, a selective CRTH2 antagonist, reduces allergen- Stinson SE, Amrani Y, and Brightling CE (2015) D prostanoid receptor 2 (chemo- induced airway responses in allergic asthmatics. Clin Exp Allergy 44:1044–1052. attractant receptor-homologous molecule expressed on TH2 cells) protein expres-

Erpenbeck VJ, Salapatek AM, Patel P, Sanni J, Dubois G, Zhou P, Perry S, sion in asthmatic patients and its effects on bronchial epithelial cells. J Allergy molpharm.aspetjournals.org and Larbig M. Efficacy and safety of 10 day treatment with the CRTh2 antagonist Clin Immunol 135:395–406. QAV680 evaluated in patients with allergic rhinitis in an environmental exposure Sugimoto H, Shichijo M, Iino T, Manabe Y, Watanabe A, Shimazaki M, Gantner F, chamber. Eur. Resp. J. 2014, 44, suppl 58 (poster P4075). and Bacon KB (2003) An orally bioavailable small molecule antagonist of CRTH2, Erpenbeck VJ, Salapatek AM, Patel P, Carew D, Dubois G, Perry S, and Larbig M ramatroban (BAY u3405), inhibits prostaglandin D2-induced eosinophil migration QAV680, cetirizine and their combination reduce nasal symptoms in patients with in vitro. J Pharmacol Exp Ther 305:347–352. intermittent allergic rhinitis in an Environmental Exposure Chamber. Eur. Resp. Summers C, Singh NR, White JF, Mackenzie IM, Johnston A, Solanki C, Balan KK, J. 2014, 44, suppl 58 (poster P4069). Peters AM, and Chilvers ER (2014) Pulmonary retention of primed neutrophils: a Gervais FG, Sawyer N, Stocco R, Hamel M, Krawczyk C, Sillaots S, Denis D, Wong E, novel protective host response, which is impaired in the acute respiratory distress Wang Z, and Gallant M et al. (2011) Pharmacological characterization of MK-7246, syndrome. Thorax 69:623–629. a potent and selective CRTH2 (chemoattractant receptor-homologous molecule Sykes DA, Dowling MR, and Charlton SJ (2010) Measuring receptor target coverage: expressed on T-helper type 2 cells) antagonist. Mol Pharmacol 79:69–76. a radioligand competition binding protocol for assessing the association and dis- Gyles SL, Xue L, Townsend ER, Wettey F, and Pettipher R (2006) A dominant role for sociation rates of unlabeled compounds. Curr Protoc Pharmacol. Chapter 9: Unit chemoattractant receptor-homologous molecule expressed on T helper type 2 (Th2) 9.14. cells (CRTH2) in mediating chemotaxis of CRTH21 CD41 Th2 lymphocytes in Sykes DA, Dowling MR, Leighton-Davies J, Kent TC, Fawcett L, Renard E, Trifilieff at ASPET Journals on September 30, 2021 response to mast cell supernatants. Immunology 119:362–368. A, and Charlton SJ (2012) The Influence of receptor kinetics on the onset and Kawata R, Reddy ST, Wolner B, and Herschman HR (1995) Prostaglandin synthase 1 duration of action and the therapeutic index of NVA237 and tiotropium. J Phar- and prostaglandin synthase 2 both participate in activation-induced prostaglandin macol Exp Ther 343:520–528. D2 production in mast cells. J Immunol 155:818–825. Swinney DC (2009) The role of binding kinetics in therapeutically useful drug action. Kerns EH, Di L, Petusky S, Kleintop T, Huryn D, McConnell O, and Carter G (2003) Curr Opin Drug Discov Devel 12:31–39. Pharmaceutical profiling method for lipophilicity and integrity using liquid Toutain PL and Bousquet-Melou A (2002) Free drug fraction vs free drug concen- chromatography-mass spectrometry. J Chromatogr B Analyt Technol Biomed Life tration: a matter of frequent confusion. J Vet Pharmacol Ther 25:460–463. Sci 791:381–388. Tummino PJ and Copeland RA (2008) Residence time of receptor-ligand complexes Kenakin TP (2009) Orthosteric drug antagonism, in A Pharmacology, 3rd ed. and its effect on biological function. Biochemistry 47:5481–5492. (Kenakin T, ed) pp 105–127, Elsevier Academic Press, London. Uller L, Mathiesen JM, Alenmyr L, Korsgren M, Ulven T, Högberg T, Andersson G, Persson Krug N, Gupta A, Badorrek P, Koenen R, Mueller M, Pivovarova A, Hilbert J, Wetzel CG, and Kostenis E (2007) Antagonism of the prostaglandin D2 receptor CRTH2 atten- K, Hohlfeld JM, and Wood C (2014) Efficacy of the oral chemoattractant receptor uates asthma pathology in mouse eosinophilic airway inflammation. Respir Res 8:16. homologous molecule on TH2 cells antagonist BI 671800 in patients with seasonal Valko K, Du CM, Bevan CD, Reynolds DP, and Abraham MH (2000) Rapid- allergic rhinitis. J Allergy Clin Immunol 133:414–419. gradient HPLC method for measuring drug interactions with immobilized ar- Lamers C, Flesch D, Schubert-Zsilavecz M, and Merk D (2013) Novel prostaglandin tificial membrane: comparison with other lipophilicity measures. JPharmSci receptor modulators: a patent review (2002 - 2012) - part I: non-EP receptor 89:1085–1096. modulators. Expert Opin Ther Pat 23:47–77. Wenzel SE, Westcott JY, Smith HR, and Larsen GL (1989) Spectrum of prostanoid Luu VT, Goujon JY, Meisterhans C, Frommherz M, and Bauer C (2015) Synthesis of release after bronchoalveolar allergen challenge in atopic asthmatics and in control a high specific activity methyl sulfone tritium isotopologue of fevipiprant (NVP- groups. An alteration in the ratio of bronchoconstrictive to bronchoprotective me- QAW039). J Labelled Comp Radiopharm 58:188–195. diators. Am Rev Respir Dis 139:450–457. Mathiesen JM, Christopoulos A, Ulven T, Royer JF, Campillo M, Heinemann A, Xue L, Barrow A, and Pettipher R (2009) Interaction between prostaglandin D and Pardo L, and Kostenis E (2006) On the mechanism of interaction of potent sur- chemoattractant receptor-homologous molecule expressed on Th2 cells mediates mountable and insurmountable antagonists with the prostaglandin D2 receptor cytokine production by Th2 lymphocytes in response to activated mast cells. Clin CRTH2. Mol Pharmacol 69:1441–1453. Exp Immunol 156:126–133. Miadonna A, Tedeschi A, Brasca C, Folco G, Sala A, and Murphy RC (1990) Mediator Xue L, Salimi M, Panse I, Mjösberg JM, McKenzie AN, Spits H, Klenerman P, release after endobronchial antigen challenge in patients with respiratory allergy. and Ogg G (2014) Prostaglandin D2 activates group 2 innate lymphoid cells J Allergy Clin Immunol 85:906–913. through chemoattractant receptor-homologous molecule expressed on TH2 cells. Monneret G, Gravel S, Diamond M, Rokach J, and Powell WS (2001) Prostaglandin J Allergy Clin Immunol 133:1184–1194. D2 is a potent chemoattractant for human eosinophils that acts via a novel DP receptor. Blood 98:1942–1948. Motulsky HJ and Mahan LC (1984) The kinetics of competitive radioligand binding Addresses correspondence to: Dr. David Sandham, Global Discovery predicted by the law of mass action. Mol Pharmacol 25:1–9. Chemistry, Novartis Institutes of Biomedical Research, 100 Technology Mould R, Brown J, Marshall FH, and Langmead CJ (2014) Binding kinetics differenti- Square, Cambridge MA 02139. E-mail: [email protected] or Prof. ates functional antagonism of orexin-2 receptor ligands. Br J Pharmacol 171:351–363. Steven Charlton, School of Life Sciences, Queen’s Medical Centre, University Murray JJ, Tonnel AB, Brash AR, Roberts LJ, 2nd, Gosset P, Workman R, Capron A, of Nottingham, University of Nottingham, UK. E-mail: StevenCharlton@ and Oates JA (1986) Release of prostaglandin D2 into human airways during acute nottingham.ac.uk antigen challenge. N Engl J Med 315:800–804.

Supplemental data and methods MOLPHARM/2015/101832

QAW039 a slowly dissociating CRTh2 antagonist with the potential for improved clinical efficacy

David A Sykes, Michelle E Bradley, Darren M Riddy, Elizabeth Willard, John Reilly, Asadh

Miah, Carsten Bauer, Simon J Watson, David A Sandham, Gerald Dubois and Steven J

Charlton

Supplemental Figure 1

A B

[Q A V 6 8 0 ] : n M [A Z D -1 9 8 1 ] : n M 2 5 0 0 2 5 0 0

0 0

c

c

i

i

f

f i i 2 0

2 0 0 0 2 0 0 0 2 .7 8

c

c

)

)

e

e M M 6 0

p 8 .3 3

p

P

P

s

s

C

C 1 5 0 0 1 8 0 1 5 0 0 9

9 2 5

(

(

3

3

g

g

0

0

n

n

i

i W

W 1 0 0 0 1 0 0 0

d

d

A

A

n

n

i

i

Q

Q

-

-

b b ]

] 5 0 0 5 0 0

H H

3 3

[ [

0 0 0 2 0 4 0 6 0 0 2 0 4 0 6 0 8 0 1 0 0 T im e (m in ) T im e (m in ) C D

[B I-6 7 1 8 0 0 ] : n M [R a m a tr o b a n ] : n M 2 5 0 0 1 2 5 0

0 0

c

c

i

i

f

f i i 5 .5 5 1 0 0 0 9 .3

2 0 0 0 c

c

)

)

e

e M

M 1 1 8 p

p 1 6 .6 7

P

P

s

s

C 7 5 0 C 1 5 0 0

9 3 5 4 (

9 5 0

(

3

3

g

g

0

0

n

n

i

i W

W 1 0 0 0 5 0 0

d

d

A

A

n

n

i

i

Q

Q

-

-

b

b ]

] 5 0 0 2 5 0

H

H

3

3

[ [

0 0 0 2 0 4 0 6 0 0 2 0 4 0 6 0 8 0 1 0 0 T im e (m in ) T im e (m in )

E

[S e tip ip r a n t] : n M 1 2 5 0

0

c

i f

i 1 0 0 0 9 .3

c

) e

M 2 8

p

P

s

C 7 5 0

9 8 4

(

3

g

0

n i

W 5 0 0

d

A

n

i

Q

- b

] 2 5 0

H

3 [

0 0 2 0 4 0 6 0 8 0 1 0 0 T im e (m in )

1

Supplemental Figure 1. Competition kinetic binding to determine the kinetic rate constants of unlabeled compounds. Competition kinetic binding curves for [3H]-QAW039 binding to CRTh2 receptors expressed in membranes from CHO cells in the presence of increasing concentrations of unlabeled (A) QAV680, (B) AZD-1981, (C) BI-671800, (D) ramatroban and (E) sepipiprant. Plates were incubated at 37oC for the time points indicated.

Data are shown as mean ± range and are representative of 3 separate experiments performed in duplicate.

2

Supplemental Figure 2.

A B

1 8 0 m in P G D 2 in c u b a tio n 1 8 0 m in P G D 2 in c u b a tio n

)

) l

[Q A W 0 3 9 ] : n M l [Q A V 6 8 0 ] : n M

o

o r

1 5 0 r 1 5 0 t c o n tro l t

n c o n tro l

n

o

o c

3 c 6

1 0 0 1 0 0

% %

( 9

( 18

d

5 0 30 d 5 0 60

n

n

u u

o 90

o 1 8 0

b b

0

0

3 0 0 6 0 0

S

S

 

P 9 0 0 P 1 8 0 0

T -5 0

T -5 0

G 3 0 0 0

G 6 0 0 0

]

]

S

S 5

-1 0 0 5 -1 0 0

3

3

[

[

-1 1 -1 0 -9 -8 -7 -6 -5 -4 -1 1 -1 0 -9 -8 -7 -6 -5 -4 [P G D ] (L o g M ) 2 [P G D 2 ] (L o g M )

C D

1 8 0 m in P G D 2 in c u b a tio n 1 8 0 m in P G D 2 in c u b a tio n

)

) l

l [A Z D -1 9 8 1 ] : n M [O C -4 5 9 ] : n M

o

o r

r 1 5 0 1 5 0 t

t c o n tro l c o n tro l

n

n

o

o c

c 9 9

1 0 0 1 0 0

%

% (

( 27 27

d

d 5 0 90 5 0 90

n

n

u

u o

o 2 7 0 2 7 0

b

b

0 0

9 0 0 9 0 0

S

S



 P P 2 7 0 0 2 7 0 0

T -5 0

T -5 0 G

G 9 0 0 0 9 0 0 0

]

]

S

S 5

5 -1 0 0 -1 0 0

3

3

[

[

-1 1 -1 0 -9 -8 -7 -6 -5 -4 -1 1 -1 0 -9 -8 -7 -6 -5 -4

[P G D 2 ] (L o g M ) [P G D 2 ] (L o g M )

E F

1 8 0 m in P G D 2 in c u b a tio n 1 8 0 m in P G D 2 in c u b a tio n

)

) l

l [B I-6 7 1 8 0 0 ] : n M [R a m a tro b a n ] : n M

o

o r

r 1 5 0 1 5 0 t

t c o n tro l c o n tro l

n

n

o

o c

c 30 60

1 0 0 1 0 0

%

% (

( 90 1 8 0

d

d 5 0 3 0 0 5 0 6 0 0

n

n

u

u o

o 9 0 0 1 8 0 0

b

b

0 0

3 0 0 0 6 0 0 0

S S

  P

P 9 0 0 0 1 8 0 0 0 T

T -5 0 -5 0 G

G 3 0 0 0 0 6 0 0 0 0

]

]

S S 5

5 -1 0 0 -1 0 0

3 3

[

[

-1 1 -1 0 -9 -8 -7 -6 -5 -4 -1 1 -1 0 -9 -8 -7 -6 -5 -4

[P G D 2 ] (L o g M ) [P G D 2 ] (L o g M )

G

1 8 0 m in P G D 2 in c u b a tio n )

l [S e tip ip ra n t] : n M o

r 1 5 0 t

n c o n tro l o

c 1 0 0

1 0 0 %

( 3 0 0

d 5 0 1 0 0 0

n u

o 3 0 0 0 b

0

1 0 0 0 0

S 

P 3 0 0 0 0

T -5 0

G

] S

5 -1 0 0

3 [

-1 1 -1 0 -9 -8 -7 -6 -5 -4

[P G D 2 ] (L o g M )

Supplemental Figure 2. Effect of CRTh2 antagonists on PGD2 stimulated GTPS accumulation. Insurmountable behavior in a [35S]-GTPS binding assay of (A) QAW039 at

3

human CRTh2 receptors expressed in membranes from CHO cells following a 180 min incubation period with the agonist PGD2. Surmountable behavior of (B) QAV680, (C)

AZD-1981, (D) OC-459, (E) BI-671800, (F) ramatroban and (G) setipiprant following a 180 min incubation period. Data are shown as mean ± range and are representative of 3 separate experiments performed in singlet.

4

Supplemental Figure 3

A

[Q A W 0 3 9 ] : n M 4

e 1 5 0

s 0 n

o 3

p 3

s

) e

9 1

r - 1 0 0

2 2

3 0 R

D

D

G (

9 0

P

g

l 1 a

5 0 3 0 0 o

l m

i 9 0 0 x

a 0

3 0 0 0 m

-8 -7 -6 -5

% 0 -1 [Q A W 0 3 9 ] : lo g M -1 0 -9 -8 -7 -6 -5 -4

[P G D 2 ] : lo g M

B

[Q A V 6 8 0 ] : n M 3

e 1 5 0

s 0

n o

p 6 2 s

e 1 8

r )

1 0 0

2 1

6 0 -

D 1 R

G 1 8 0

D

P

(

l

a 6 0 0

5 0 g 0

m

o i

1 8 0 0 l -8 -7 -6 -5 -4 x a 6 0 0 0

m -1 [Q A V 6 8 0 ] : lo g M

% 0 -1 0 -9 -8 -7 -6 -5 -4 -2 [P G D 2 ] : lo g M

C [A Z D -1 9 8 1 ] : n M

e 1 5 0

s 0 4

n o

p 9

s e

r 2 7 3

1 0 0 2

9 0 )

D

1 G

2 7 0 - 2

P

R

l a

5 0 9 0 0 D

(

m i

2 7 0 0 g 1

x

o a

9 0 0 0 l

m

% 0 0 -1 0 -9 -8 -7 -6 -5 -4 -7 -6 -5 -4 [P G D ] : lo g M 2 -1 [A Z D -1 9 8 1 ] : lo g M

D [O C -4 5 9 ] : n M

e 1 5 0 4

s 0

n o

p 9

s 3 e

r 2 7

1 0 0 )

2 9 0

1

D -

G 2 7 0 2

R

P

l

D a

5 0 9 0 0 (

m i

2 7 0 0 g 1

x

o a

9 0 0 0 l

m

% 0 0 -1 0 -9 -8 -7 -6 -5 -4 -7 -6 -5 -4 [P G D ] : lo g M 2 -1 [O C 4 5 9 ] : lo g M

5

E

[B I-6 7 1 8 0 0 ] : n M 3

e 1 5 0

s 0

n o

p 3 0 2

s )

e 9 0

r

1 0 0 1 -

2 3 0 0

D

R G

9 0 0 D 1

P

(

l a

5 0 3 0 0 0 g

m

o i

9 0 0 0 l x

a 0

3 0 0 0 0 m -7 -6 -5 -4

% 0 -1 0 -9 -8 -7 -6 -5 -4 [B I-6 7 1 8 0 0 ] : lo g M -1 [P G D 2 ] : lo g M F [R a m a tro b a n ] : n M 2 .0 e 1 5 0

s 0

n o

p 6 0

s 1 .5 e

r 1 8 0

1 0 0

) 2

6 0 0 1 D

- 1 .0 G

1 8 0 0 R

P

D

l

( a

5 0 6 0 0 0 g

m 0 .5

i 1 8 0 0 0

o

x l

a 6 0 0 0 0 m 0 .0

% 0 -1 0 -9 -8 -7 -6 -5 -4 -7 -6 -5 -4

[P G D 2 ] : lo g M -0 .5 [r a m a tr o b a n ] : lo g M G

[S e tip ip ra n t] : n M 6

e 1 5 0

s 0 n

o S lo p e = 1 .0

p 6 0

s 4

e )

r 1 8 0

1 0 0 1 -

2 6 0 0

D

R G

1 8 0 0 D 2

P

(

l a

5 0 6 0 0 0 g

o

m l

i 1 8 0 0 0 x

a 0

6 0 0 0 0 m

-7 -6 -5 -4 -3

% 0 -1 0 -9 -8 -7 -6 -5 -4 [s e tip ip r a n t] : lo g M -2 [P G D 2 ] : lo g M

Supplemental Figure 3. Effect of increasing concentrations of CRTh2 antagonist (A) QAW039,

(B) QAV680, (C) AZD-1981, (D) OC-459, (E) BI-671800, (F) ramatroban and (G) setipiprant on PGD2 stimulated [35S]-GTPS accumulation in CRTh2 expressing CHO-cell membranes following a 180min incubation. Data were fitted according to a four parameter logistic equation and values for the minimum asymptote and hill slopes were shared across all of the data sets.

Estimated pA2 values along with corrected pA2 values are shown in Supplemental Table 1. Data are

Mean ± SEM of 3 experiments performed in singlet.

6

Supplemental Figure 4

Q A W 0 3 9

) 4 0 0 0

0

m p

c + 3 n M ( 3 0 0 0

d + 1 0 n M n

u + 3 0 n M

o b

2 0 0 0 + 1 0 0 n M S

 + 3 0 0 n M P

T 1 0 0 0

G

]

S

5 3 [ 0 1 0 -1 0 1 0 -9 1 0 -8 1 0 -7 1 0 -6 1 0 -5 1 0 -4

[P G D 2 ] : M

Supplemental Figure 4. Estimation of the dissociation rate of the CRTh2 antagonist QAW039 from functional [35S]-GTPS data, obtained in CRTh2 expressing membranes from CHO cells following a 60min incubation with PGD2. Data were fitted according to an operational hemi- equilibrium model for competitive antagonism (Kenakin, 2009 and Mould et al., 2014). Data shown as the Mean ± SEM of 3 experiments performed in singlet.

7

Supplemental Figure 5.

A

3

2 5 0 0 [ [ H ]-Q A W 0 3 9 ] : n M c

i 0 .4 6 f

i 2 0 0 0 c

) 0 .9 2

e

M

p P

s 1 .8 2

C 1 5 0 0

9

(

3 3 .6 4

g

0

n i

W 1 0 0 0

d

A

n

i

Q

- b

] 5 0 0

H

3 [

0 0 5 0 1 0 0 1 5 0 2 0 0 T im e (m in )

B 3 [ [ H ]-O C -4 5 9 ] : n M

2 5 0 0

c 2 .7 2

i

f

i )

c 2 0 0 0 1 .3 5

M

e p

P 0 .6 8

s

C

(

9 1 5 0 0 0 .3 4

5

g

4

n

i C

d 1 0 0 0

O

n

-

i

]

b H

3 5 0 0 [

0 0 1 0 2 0 3 0 T im e (m in )

Supplemental Figure 5. Association of [3H]-QAW039 and [3H]-OC-459 at room temperature to the human CRTh2receptor expressed in membranes from CHO cells. A family of association

3 kinetic curves was constructed using a range of [ H]-QAW039 concentrations. Both kon and koff values for QAW039 and OC-459 were estimated by fitting the data to a global kinetic model. Associated kinetic parameters are shown in Supplemental Table 3. Data are representative of 3 experiments performed in duplicate.

8

Supplemental Table 1. Fitted affinity estimates for CRTh2 antagonism of PGD2 stimulated [35S]-

GTPS accumulation. pA2 values were estimated directly from the Schild analysis and corrected pA2 estimates were derived by correcting for % bound to HSA using the method of Toutain and Bousquet-

Melou (2002) as outlined in Blakeley et al., 2015. Data are mean from at least 3 separate experiments.

CRTh2

Schild slope pA2 Corrected pA2 antagonist

QAW039 0.98 ± 0.04 8.54 8.64

QAV680 1.0 ± 0.04 7.63 7.67

AZD-1981 1.01 ± 0.06 7.98 8.33

OC-459 1.03 ± 0.05 7.66 8.44

BI-671800 1.08 ± 0.03 7.09 7.89

Ramatroban 1.00 ± 0.05 6.86 7.11

Setipiprant 1.58 ± 0.09 6.27 nd

Supplemental Table 2. Fitted parameter estimates for QAW039 mediated antagonism of PGD2 stimulated [35S]-GTPS accumulation according to an operational hemi-equilibrium model for competitive antagonism (Kenakin, 2009). Data are mean ± s.e. mean from at least 3 separate experiments.

Time

Parameter 60min

QAW039 Kb (nM) 3.2 ± 0.7

PGD2 (KA) 23.3 ± 5.0

Dissociation constant (koff) 0.005 ± 0.001

Tau 1.66 ± 0.81

EMax 8795 ± 2376

9

Supplemental Table 3. Kinetic parameters of [3H]-QAW039 and OC-459 at the human

CRTh2 receptor measured at room temperature. Data are Mean ± SEM from at least 3 separate experiments.

-1 -1 -1 CRTh2 kon (M min ) koff (min ) t1/2 (min) Kinetic pKd

antagonist

[3H]-QAW039 9.24 ± 0.68 x106 0.012 ± 0.001 57.1 8.87 ± 0.07

[3H]-OC-459 1.78 ± 0.56 x108 0.41 ± 0.06 2.1 8.63 ± 0.03

10

Supplemental method

Isolation, culture and expansion of CRTh2+ Th2 lymphocytes.

Isolation of CD4+ T cells from human blood. 500 mL of peripheral blood was collected from healthy adult volunteers. Blood was collected into 50 mL Falcon tubes containing 1 mL of anti-coagulant (20mg/mL EDTA in PBS) per 10 mL of blood. Blood was diluted with 12.5 mL PBS per 20 mL of blood and layered onto Ficoll (15 mL Ficoll per 50 mL Falcon tube).

PBMCs were isolated by density gradient centrifugation, 300 xg for 30 minutes at 18˚C. The

‘buffy coat’ containing PBMCs was recovered and washed twice (300 xg at 4oC for 7 minutes) in PBS. CD4+ lymphocytes were isolated by negative selection using the MACS

CD4 Isolation kit II (Miltenyi Biotec) according to the manufacturer’s instructions. Briefly, to

107 cells, 40 μL of buffer and 10 μL of Biotin-Antibody Cocktail were added, mixed and incubated at 4˚C for 10 minutes. A further 30 μL of buffer and 20 μL of Anti-Biotin

Microbeads were then added to the cells and incubated at 4oC for 15 minutes. Cells were washed (300 xg for 10 minutes), resuspended in 500 μL of buffer per 108 cells and applied to the ‘primed’ LS column. The untouched CD4+ cell fraction was collected and put through a second LS column. The purity of the CD4+ cell population isolated was greater than 90 %.

Culture of CD4+ T cells prior to isolation of CRTh2+ cells Freshly isolated CD4+ T cells were resuspended at 2 x 106 per mL in X-VIVO 15 medium plus 10 % human AB serum and

1% Penicillin/Streptomycin. Cells were cultured in T75 flasks, with 20 mL of cells per flask.

Recombinant human IL-2 was added at a concentration of 30 ng/mL and rhIL-4 was added at

100 ng/mL. These cytokines were replenished on day 3 of the culture period by centrifuging the cells at 300 xg for 5 minutes then carefully removing half the medium and adding back the same volume of new medium plus the relevant concentration of cytokine. Cells were

+ cultured for a total of 7 days at 37˚C, 5 % CO2 prior to isolating the CRTh2 cells

11

Isolating CRTh2+ T cells using MACS. CRTh2+ T cells were isolated from the CD4+ population after 7 days in culture with rhIL-4 and rhIL-2. Initially, dead cells in the CD4+ population were removed prior to isolating CRTh2+ cells in order to try to increase the purity.

This was carried out using Dead Cell Removal Microbeads (Miltenyi Biotec), which bind to dead or dying cells and are then retained on an LS column, with the live cell population flowing through the column. Then the CD294 (CRTh2) Microbead kit (Miltenyi Biotec) was used to isolate CRTh2+ cells by positive selection. Briefly, to 107 cells 80 μL of MACS buffer was added plus 20 μL of FcR Blocking Reagent and 10 μL CD294 (CRTH2)-PE antibody. Cells were then incubated at 4˚C for 10 minutes, washed once in MACS buffer

(300 xg for 10 minutes) and then resuspended in 80 μL buffer per 107 cells plus 20 μL of

Anti-PE Microbeads. Cells were incubated for a further 15 minutes at 4˚C, then washed once in MACS buffer. Cells were resuspended at 108 cells in 500 μL buffer (usually 2 -3 mL) before being added to a ‘primed’ LS column. The CRTh2+ fraction was eluted off the column and analysed. The cells were counted and seeded at 1 x 106 per mL in a 12 well plate.

Medium used was X-VIVO 15 medium plus 10 % human AB serum with rhIL-2 added at 30 ng/mL. The cells were then activated and expanded in culture as described below.

Isolating CRTH2+ T cells by MoFlo.CRTH2+ T cells were isolated from the CD4+ population after 7 days in culture with rhIL-4and rhIL-2. Cells were counted and resuspended in 80 μL MACS buffer for every 107 cells. Cells were stained with anti-human CRTh2 biotinylated antibody (Miltenyi Biotec). To every 107 cells, 20 μL of FcR Block and 10 μL of the antibody was added and incubated on ice for 15 minutes. Cells were washed once at

300xg for 10 minutes, resuspended in 90 μL of MACS buffer and 10 μL of a fluorochrome- conjugated streptavidin antibody was added (conjugated. to either PE-Cy5 or APC). Cells were incubated on ice for 25 minutes, then washed once at 300xg for 10 minutes. For isolation using the MoFlo, cells were resuspended at approximately 20 x 106 cells per mL in

12

MACS buffer and the CRTH2+ cells were collected, then activated and expanded using the

Miltenyi Biotec T cell Activation/Expansion kit as described below

Activation and Expansion of CRTH2+ T cells using the T cell Activation/Expansion kit. For every 5 x 106 CRTh2+ cells, 25μL of the MACSiBead particles in the kit that had previously been ‘loaded’ with human antibodies against CD2, CD3 and CD28 was added. The cells plus

MACSiBeads are then cultured at 2.5 x 106/mL in a 24 well plate at 1mL/well for 3 days at

37˚C, 5 % CO2. After 3 days, the beads were removed using the MACSiMAG separator

(Miltenyi Biotec) and the cells were then further cultured at a density of 1 x 106/mL with 30 ng/mL rhIL-2. Cells were used in the assay between days 7 and 12 after the MACSiBeads had been removed. CRTh2 expression was measured by FACS prior to using the cells in the assay – on day 7 after bead removal. A sample of cells was incubated at 4˚C for 20 minutes with either 10μL of rat IgG2a Alexa-Fluor 647 isotype control antibody or anti-human

CRTh2 Alexa-Fluor 647 antibody. Cells were then washed and analysed by FACSCaliber.

Preparation of [3H9]OC459 and ([3H3]QAW039

3 3 3 3 [ H9]OC459 ((5-Fluoro-2-[ H3]methyl-3-[3,4,6,8- H4]quinolin-2-yl-[ H2]methyl-indol-1-yl)- acetic acid, 1) was prepared at International Isotope Clearing House, Inc. following patent literature for direct hydrogen isotope exchange on solid support (Zolotarev et al., 1991). Detailed experimental procedures were not disclosed by the contractor. After re-purification 1 was released as a trifluoro acetate salt (radiochemical purity 96.4% HPLC, specific activity

3.3TBq/mmol. Tritiated Fevipipant ([3H3]QAW039, 2, radiochemical purity 98.2% HPLC, specific activity 3.1TBq/mmol) was prepared by the Isotope Laboratory of NIBR Basel as disclosed in detail elsewhere (Luu et al., 2015). The positions of (partial) tritium labelling were determined by 3H-NMR. The structures of 1 and 2 are shown in the Supplemental Figure 6.

13

Supplemental Figure 6. Structures and positions of tritium labelling in [3H9]OC459 (1) and

[3H3]QAW039 (2).

References:

Luu VT, Goujon JY, Meisterhans C, Frommherz M, Bauer C (2015). Synthesis of a high specific activity methyl sulfone tritium isotopologue of fevipiprant (NVP-QAW039). J Labelled Comp

Radiopharm. 58(5):188-95.

Zolotarev JA, Zaitsev DA, Tatur VJ, Myasoedov NF (1991) Inventors and Assignees. Method for preparing biologically active organic compounds labelled with hydrogen isotope. US patent

5,026,909. Jun 25 1991.

Acknowledgements

The co-author would like to thank Dr. Ines Rodriguez for the design of the radio-synthesis of [3H9]-

OC459 and Dr. Albrecht Glaenzel for the release analysis of [3H9]-OC459.

14