Comparison of human ETA and ETB receptor signalling via G-protein and β-arrestin pathways

Life Sciences 91 (2012) 544–549

Contents lists available at SciVerse ScienceDirect

Life Sciences journal homepage: www.elsevier.com/locate/lifescie

Comparison of human ETA and ETB receptor signalling via G-protein and β-arrestin pathways Janet J. Maguire a,⁎, Rhoda E. Kuc a, Victoria R. Pell a, Andrew Green b, Mike Brown b, Sanj Kumar b, Tom Wehrman c, Elizabeth Quinn c, Anthony P. Davenport a a b c

Clinical Pharmacology Unit, University of Cambridge, Level 6 ACCI, Box 110 Addenbrooke's Hospital, Cambridge, CB2 0QQ, UK DiscoveRx Corporation Ltd, Birmingham, UK DiscoveRx Corporation, Fremont, CA, USA

a r t i c l e

i n f o

Article history: Received 2 November 2011 Accepted 8 March 2012 Keywords: ETA ETB Endothelin β-arrestin recruitment Vasoconstriction Human pharmacology Human left ventricle Sitaxentan Ambrisentan BQ123 BQ788 Biased agonist Biased antagonist

a b s t r a c t Aims: To determine the pharmacology of ETA- and ETB-mediated β-arrestin recruitment and compare this to established human pharmacology of these receptors to identify evidence for endothelin receptor biased signalling and pathway specific blockade by antagonists. Main methods: The ability of ET-1, ET-2, ET-3, sarafotoxin 6b and sarafotoxin 6c to activate ETA and ETBmediated β-arrestin recruitment was determined in CHO-K1 cells. Affinities were obtained for ETA selective (BQ123, sitaxentan, ambrisentan), ETB selective (BQ788) and mixed (bosentan) antagonists using ET-1 and compared to affinities obtained in competition experiments in human heart and by Schild analysis in human saphenous vein. Agonist dependence of affinities was compared for BQ123 and BQ788 in the ETA and ETB β-arrestin assays respectively. Key findings: For β-arrestin recruitment, order of potency was as expected for the ETA (ET-1 ≥ ET-2 >> ET-3) and ETB (ET-1 = ET-2 = ET-3) receptors. However, at the ETA receptor sarafotoxin 6b and ET-3 were partial agonists. Antagonism of ET peptides by selective and mixed antagonists appeared non-competitive. BQ123, but not BQ788, exhibited agonist-dependent affinities. Bosentan was significantly more effective an inhibitor of β-arrestin recruitment mediated by ETA compared to the ETB receptor. In the ETA vasoconstrictor assay, ET-1, ET-2 and S6b were equipotent, full agonists and antagonists tested behaved in a competitive manner, although affinities were lower than predicted from the competition binding experiments in left ventricle. Significance: These data suggest that the pharmacology of ETA and ETB receptors linked to G-protein- and βarrestin mediated responses was different and bosentan appeared to show bias, preferentially blocking ETA mediated β-arrestin recruitment. © 2012 Elsevier Inc. All rights reserved.

Introduction The endothelin (ET) antagonists, bosentan (Dingemanse and van Giersbergen, 2004) and ambrisentan (Cheng, 2008) are clinically used in the treatment of pulmonary arterial hypertension. Bosentan reportedly has similar affinity for both of the endothelin receptor subtypes (Clozel et al., 1994), ETA and ETB, whereas ambrisentan is an ETA selective compound (Vatter and Seifert, 2006) although the degree of ETA selectivity reported in the literature varied depending on whether it was determined using animal or human, cloned or native receptor systems (Maguire and Davenport, 2012; Palmer, 2009; Vatter et al., 2002). A third antagonist, sitaxentan (Wu et al., 1997), is much more ETA selective than ambrisentan. The optimal receptor selectivity for ⁎ Corresponding author at: Clinical Pharmacology Unit, Level 6 ACCI, Box 110 Addenbrooke's Hospital, Cambridge, CB2 0QQ, UK. Tel.: + 44 1223 762579; fax: + 44 1223762564. E-mail address: [email protected] (J.J. Maguire). 0024-3205/$ – see front matter © 2012 Elsevier Inc. All rights reserved. doi:10.1016/j.lfs.2012.03.021

treatment of PAH remains debatable (Davie et al., 2009) but an ETA selective compound has the advantage of leaving unaffected potentially beneficial vasodilator (Takayanagi et al., 1991) and clearing ETB receptors (Johnström et al., 2005; Kelland et al., 2010) that are present, for example on endothelial cells (Bacon et al., 1996; Gardiner et al., 1994) and expressed at particularly high densities in the lung and kidney (Johnström et al., 2002, 2006). Endothelin peptides mediate their cellular actions via interaction with the two G protein-coupled receptors (GPCRs), ETA and ETB (Arai et al., 1990; Sakurai et al., 1990). Agonist binding activates canonical G-protein-linked intracellular signalling cascades with subsequent receptor phosphorylation leading to recruitment of the scaffolding protein β-arrestin which may result in receptor desensitisation and internalisation (Dai and Galligan, 2006; Drake et al., 2006). It is now known that GPCRs can couple promiscuously to multiple G protein classes but may also activate signalling pathways such as the MAP kinases in a G protein-independent manner via complexes of the receptor with β-arrestin 1 or 2 (Schulte and Levy, 2007). Ligand-specific pathway

J.J. Maguire et al. / Life Sciences 91 (2012) 544–549

modulation (biased agonism/biased antagonism) may have important therapeutic application (Whalen et al., 2011; Zheng et al., 2010), but little is known about the potential for pathway bias of ET peptides and ET receptor antagonists. Our aim was to determine the ability of ET peptides and related agonists to recruit β-arrestin via ETA and ETB receptors and to compare the ability of well characterised ET receptor selective and nonselective antagonists to block these responses. These data were then compared to the human cardiovascular pharmacology of theses agonists and antagonists determined by radioligand binding in human left ventricle that expresses both ET receptor subtypes (Molenaar et al., 1993) and in functional experiments in saphenous vein in which vasoconstriction is mediated via the ETA receptor (Maguire and Davenport, 1995). Materials and methods Human tissue Anonymised human cardiovascular tissue samples were used in this study with local ethical approval (REC 05/Q0104/142). The samples were obtained from Papworth Hospital Research Tissue Bank (Cambridgeshire 1 Research Ethics Committee reference 08/H0304/ 56, samples collected with written informed patient consent). Unless otherwise stated n-values refer to the number of patients from whom tissue was obtained. Competition binding assays in human left ventricle Competition binding experiments were carried out as previously described (Davenport and Kuc, 2005). Briefly, sections of human heart (10 μm) were pre-incubated in HEPES buffer and then with [ 125I]ET-1 (0.1 nM) for 2 h at room temperature in the presence of increasing concentrations of antagonists (1 pM to 100 μM). Non-specific binding was determined in the presence of 1 μM of unlabelled ET-1. Affinities (KD) of antagonists for ETA and ETB receptors were determined using the non-linear curve fitting programme KELL (Biosoft, http://Biosoft.com). β-Arrestin recruitment assay Initial saturation binding experiments were performed, as previously described (Davenport and Kuc, 2005), to measure level of receptor density in the CHO-K1 cells expressing either ETA or ETB receptors. Briefly, frozen cells were thawed and plated as per the manufacturer's instructions. Cells were re-suspended in HEPES buffer and incubated for 90 minutes with increasing concentrations of [ 125I]ET-1 (8 pM– 4 nM) at room temperature, non-specific binding was determined using 1 μM ET-1. Binding equilibrium was broken by centrifugation (20,000 ×g for 10 min), pellets were washed in Tris–HCl buffer (pH 7.4 at 4 °C), re-centrifuged and pellets counted using a gamma counter. Protein determination carried out using BioRad DC assay. Data were analysed, using the non-linear curve fitting programme KELL (Biosoft, http://Biosoft.com), to determine receptor affinity (KD) and receptor density in fmol mg− 1 protein (BMAX). Experiments were carried out using the DiscoveRx PathHunter® eXpress β-Arrestin GPCR Assays (EDNRA and EDNRB) according to the manufacturer's instructions. Briefly, CHO-K1 cells expressing either human ETA or human ETB were incubated with antagonists or vehicle for 60 min at 37 °C. Increasing concentrations of ET and related peptides were added and cells incubated for further 90 min at 37 °C. Detection reagent solution was added and cells incubated for 2 hr at room temperature. The resulting chemiluminescent signal was measured and agonist concentration–response curves (CRC) in the absence and presence of antagonists, measured as relative light units, were normalised to the maximum response obtained to ET-1 in the same experiment. Resulting agonist CRC data were fitted to a 4 parameter logistic equation using the curve fitting program FigSys

545

(Biosoft UK). Values of EC50 and EMAX were determined for agonists and compared to control responses to ET-1. In the ETA and, to a lesser extent, ETB assays all antagonists produced a concentration dependent reduction in the maximum response of the agonist concentration– response curves and therefore affinities for the receptors were determined using the method described by Kenakin (1993). Human saphenous vein vasoconstriction functional assay Experiments were carried out as previously described (Maguire, 2002; Maguire et al., 1995). Briefly, 4 mm rings of endotheliumdenuded human saphenous vein were set up for isometric force recordings. Cumulative CRCs were constructed to ET and related peptides (10− 10 to 10− 6 M), once the maximum response to each agonist had been achieved the curve was terminated by the addition of 100 mM KCl to determine the maximal possible contractile response for each preparation. In additional experiments, ET-1 CRCs were repeated in the absence and presence of bosentan (0.3 μM–30 μM), ambrisentan (30 nM–3 μM) or sitaxentan (30 nM–3 μM). Vein preparations were incubated with antagonists for 40 minutes prior to addition of ET-1. Agonist data were expressed as a % of the terminal KCl response and were analysed using the non-linear iterative curve-fitting programme FigSys (Biosoft) to give values of pD2 (−log10 of the agonist concentration producing 50% of the maximum response) and EMAX(maximum response to an agonist expressed as a %KCl). Antagonist affinities (KB) were determined by Schild analysis (Arunlakshana and Schild, 1959). If not significantly different from one, Schild slopes were constrained to one to determine value of KB. Materials ET-1, ET-2, ET-3, sarafotoxin 6b (S6b) and sarafotoxin 6c (S6c) were from The Peptide Institute (Osaka, Japan). BQ123 (cyclo[D-AspL-Pro-D-Val-L-Leu-D-Trp-]) was from Novabiochem, BQ788 (N-cis2,6- dimethylpiperidinocarbonyl-γL-MeLeu-D-Trp(CooMe)-D-Nle-ONa) from Neosystems. Bosentan was from Parke-Davis, sitaxentan from Pfizer Inc., and ambrisentan from Gilead Sciences Inc. [125I]-ET-1 (2200 Ci/ mmol) was from PerkinElmer NEN® Radiochemicals. Results Affinity and selectivity of endothelin receptor antagonists determined in human left ventricle BQ123 (ETA KD 0.73 ± 0.22 nM, ETB KD 24.3 ± 2.0 μM, ETA:ETB selectivity 33288) ambrisentan (ETA KD 0.28 ± 0.23 nM, ETB KD 0.25 ± 0.05 μM, ETA:ETB selectivity 893) and sitaxentan (ETA KD 1.65 ± 0.80 nM, ETB KD 327 ± 134 μM, ETA:ETB selectivity 198,182) were confirmed as ETA selective antagonists with subnanomolar affinities for this subtype in human left ventricle. This compares to our previously published data showing that bosentan does not distinguish between human ETA and ETB receptors (one site fit 77.6 ± 7.9 nM, non-selective (Peter and Davenport, 1996) and that BQ788 is highly ETB selective (ETA KD 1.01 ± 0.2 μM, ETB KD 9.81 ± 1.32 nM, ETA:ETB selectivity 0.01) (Russell and Davenport, 1996) in this tissue. ETA mediated β-arrestin recruitment Receptor density (BMAX) for ETA CHO-K1 cells was 30.4 ± 4.4 fmol mg− 1 protein (n = 3) with affinity for [125I]ET-1 of KD =0.24 ± 0.05 nM. In the absence of agonists, BQ123 (0.3–10 μM), ambrisentan (0.3– 30 μM) and sitaxsentan (0.01–10 μM) had no effect on basal activity in this assay. Bosentan had no effect at 3 and 30 μM, but produced a small increase in response (b10% of the maximum response to ET-1) at 300 μM.

546

J.J. Maguire et al. / Life Sciences 91 (2012) 544–549

The order of potency of agonists was ET-1 > ET-2 = S6b >> ET-3, with S6c and ET-1(1–31) inactive (Fig. 1A). Interestingly, ET-2 and particularly ET-3 and S6b elicited maximum responses that were significantly less than that of ET-1 (Fig. 1A, Table 1). Responses to all agonists were antagonised by BQ123 (0.3 μM–1 μM) in an apparently non-competitive manner, with a BQ123 concentration-dependent reduction in maximum agonist response (Fig. 2). However, BQ123 appeared to produce agonist dependent affinities that ranged from 0.3 nM to 30 nM, being least effective as an antagonist of ET-2 in this assay (Table 2). All three non-peptide antagonists tested, bosentan, ambrisentan and sitaxentan (Fig. 3 (A–C), produced rightward shifts if the ET-1 curve with concentration-dependent decreases in maximum response resulting in affinity estimates for the antagonism of ET-1 in the low nanomolar range (Table 3).

Table 1 Potency and efficacy of endothelins and related peptides in the ETA and ETB β-arrestin recruitment assays. ETA

ET-1 ET-2 ET-3 S6b S6c BQ3020

ETB

pD2

EMAX (% ET-1)

n

pD2

EMAX (% ET-1)

n

9.38 ± 0.09 8.75 ± 0.05⁎ 7.31 ± 0.05⁎ 8.83 ± 0.03⁎

100 ± 2 87 ± 3⁎ 31 ± 1⁎ 45 ± 2⁎

No response No response

– –

11 3 6 9 3 3

9.04 ± 0.10 8.55 ± 0.03⁎ 9.44 ± 0.09 9.11 ± 0.08 9.10 ± 0.04 8.94 ± 0.11

100 ± 2 105 ± 2 83 ± 4 99 ± 5 96 ± 5 82 ± 7

12 3 3 7 3 3

Values are mean ± s.e. mean. n = Number of replicates. *Significantly different from ET-1 control (P b 0.05, Student's 2-tailed t-test).

for the non-peptide antagonist ambrisentan which had a KB determined for antagonism of ET-1 mediated vasoconstriction 800 fold lower than the affinity determined in the competition experiment (Table 4).

ETB mediated β-arrestin recruitment Receptor density (BMAX) for ETB CHO-K1 cells was 156.6 ± 33.8 fmol mg− 1 protein (n = 3) with affinity for [125I]ET-1 of KD = 0.37 ± 0.12 nM. In the absence of agonists, BQ788 (0.1–100 μM), ambrisentan (0.1– 300 μM) and bosentan (0.1–300 μM) had no effect on basal activity in this assay. Agonist order of potency in the ETB assay was ET-1 = ET-3 = S6b = S6c = BQ3020 > ET-2 (Fig. 1B). In contrast to the ETA assay maximum responses to all agonists were more than 80% of the maximum response to ET-1 (Table 1). BQ788 (1 μM) produced a reduction in maximum response to all agonists tested (Fig. 3F) and therefore affinity was determined as for the ETA assay. The affinity for BQ788 was comparable for all agonists tested of between 0.5 nM and 2 nM (Table 2). The affinities of bosentan and ambrisentan to block ET-1 mediated β-arrestin recruitment in the ETB assay were two orders of magnitude less than for BQ788 with both antagonists exhibiting ETA selectivity for β-arrestin recruitment (Table 3, Fig. 3D and E).

Discussion In human blood vessels, from a range of vascular beds, the ETA receptor predominates on medial smooth muscle cells and mediates vasoconstriction (Lipa et al., 1999; Maguire and Davenport, 1995; Nilsson et al., 1997; Spieker et al., 2003). In the ETA mediated recruitment of β-arrestin assay we have found that the order of potency of endogenous endothelins and related sarafotoxins is as previously obtained for constriction of human isolated blood vessels, such as saphenous vein (Maguire and Davenport, 1995). One striking difference between these two assays was the observation that in the β-arrestin assay compared to ET-1 other agonists, particularly ET-3 and S6b, appeared to be partial agonists. However, we have previously reported that in saphenous vein, although maximum vasoconstrictor responses to ET-1 and S6b were comparable; an additional response to ET-1

Human saphenous vein vasoconstriction functional assay

A Response (% ET-1 Control)

ET-1, ET-2 and S6b were equipotent constrictors of human saphenous veins with comparable maximum responses (ET-1 pD2 8.57 ± 0.06, EMAX 84± 2, n = 98; ET-2 pD2 8.15 ± 0.20, EMAX 85± 5, n = 8; S6b pD2 8.09 ± 0.09, EMAX 89 ± 2, n = 9). ET-3 was much less potent than these with the concentration–response curve incomplete at 1 μM. S6c (n = 9) and BQ3020 (n = 6) were essentially inactive over the concentration range tested. BQ123, bosentan, ambrisentan and sitaxsentan produced concentration-dependent, parallel rightward shifts of the ET-1 CRC without affecting maximum response. Schild analysis confirmed that antagonism at the concentrations tested appeared competitive, as Schild slopes were close to unity. Affinities determined in the functional assay were lower than predicted from the competition binding experiment for all antagonists, but particularly

100

100

80

80

60

60

40

40 10 8 6 0 20 40 60 80 100

20

10

ET-1 ET-2 ET-3 S6b S6c ET-1(1-31)

100 80 60

100 80

40

40

20

20

0

ET-1 ET-2 ET-3 S6b S6c BQ3020

60

0 10

8

6

-Log10 [Agonist] M

10

8

6

-Log10 [Agonist] M

Fig. 1. Concentration–response curves to endothelin, sarafotoxin and related peptides in the (A) ETA and (B) ETB β-arrestin recruitment assays.

20

8

6

10

-Log10 [ET-1] M

Response (% ET-1 Control)

Response (%ET-1 Maximum)

B

ET-2 +1µM BQ123

0

0

C A

B

ET-1 +0.3µM BQ123 + 1µM BQ123

D

ET-3 +0.3µM BQ123

100

100

80

80

60

60

40

40

20

20

0

8

6

-Log10 [ET-2] M S6b +0.3µM BQ123 + 1µM BQ123

0 10

8

6

-Log10 [ET-3] M

10

8

6

-Log10 [S6b] M

Fig. 2. Antagonism of (A) ET-1, (B) ET-2, (C) ET-3 and (D) S6b mediated β-arrestin recruitment via the ETA receptor by 0.3 μM (■) and 1 μM (▲) BQ123.

J.J. Maguire et al. / Life Sciences 91 (2012) 544–549 Table 2 Affinity (KB) estimates of subtype selective peptide antagonists BQ123 and BQ788 determined in ETA and ETB β-arrestin recruitment assays, respectively.

ET-1 ET-2 ET-3 S6b S6c BQ3020

ETA Assay BQ123 (KB)

n

ETB Assay BQ788 (KB)

n

8.36 nM 32.8 nM 1.05 nM 1.35 nM ND ND

6 3 3 6

2.11 nM 1.33 nM 1.49 nM 0.66 nM 0.76 nM 0.51 nM

3 3 3 3 3 3

Table 3 Affinity of non-peptide endothelin antagonists determined against ET-1 in ETA and ETB β-arrestin recruitment assays.

Bosentan Ambrisentan Sitaxsentan

Response (%ET-1 Maximum)

B

ET-1 +0.03µM Ambrisentan +0.3µM Ambrisentan +3µM Ambisentan

ETB Assay (pKB)

n

ETA:ETB selectivity

8 8 9

6.15 ± 0.01 5.90 ± 0.23 ND

6 9

73 677 –

C

100

100

100

80

80

80

60

60

60

40

40

40

20

20

20

0

0 10

8

6

D

8

6

10

-Log10 [ET-1] M

E

ET-1 +3µM Bosentan +30µM Bosentan

ET-1 +3µM + 30µM + 300µM + 1000µM Ambrisentan

F

80

80

80

60

60

60

40

40

40

20

20

20

0 8

6

-Log10 [ET-1] M

6

ET-1 +1µM BQ788

100

10

8

-Log10 [ET-1] M

100

0

ET-1 +0.03µM Sitaxentan +0.3µM Sitaxentan +3µM Sitaxentan

0 10

-Log10 [ET-1] M

Response (%ET-1 Maximum)

n

(desensitisation) and may target the receptor for subsequent internalisation and trafficking (see Fergusson 2001). It has been shown that, following agonist activation, the ETB receptor is rapidly desensitised by phosphorylation (Cramer et al., 1998), internalised (Wu-Wong et al., 1995) and targeted to the late endosomal/lysosomal pathway for degradation (Bremnes et al., 2000; Oksche et al., 2000), a mechanism that effectively removes endothelin peptides, for example, from the plasma. Efficacious recruitment of β-arrestin by agonists activating the ETB receptor, as observed in this study, is therefore consistent with the hypothesis that this subtype acts as a clearing receptor (Johnström et al., 2005; Kelland et al., 2010). We found that ET-1 was equally potent in the ETA-mediated β-arrestin recruitment ass ay as the ETB. However, whilst the ETA subtype is also internalised following activation, this occurs more slowly than for the ETB receptor (Cramer et al., 1998) and it has been demonstrated that ET-1 remains associated with the receptor for up to 60 minutes after internalisation (Wang et al., 2000). This has been proposed to explain in part the long-lasting vascular responses to ET-1 mediated by the ETA receptor. In the endothelin field there are many reports of agonist-specific effects of ETA receptor antagonists. For example, we and others have previously demonstrated that BQ123 is a more potent inhibitor of S6bmediated vascular responses than ET-1 (Bodelsson and Stjernquist, 1993; Maguire et al., 1996; Salom et al., 1993). Data from our present

could be obtained on a maximal response to S6b but no additional response to S6b could be achieved in the presence of a maximal response to ET-1 (Maguire et al., 1996). In the same study, in crossdesensitisation experiments, S6b had no effect in veins rendered insensitive to ET-1 by repeated administration whereas veins insensitive to S6b responded to ET-1. One explanation for these observations is that S6b is a partial agonist at the ETA receptor but that this is more apparent in the β-arrestin assay because this assay provides a direct measure of β-arrestin binding to the receptor (one to one stoichiometry) compared to the more amplified vasoconstrictor assay. In the ETB β-arrestin recruitment assay the order of potency of the endothelins and related peptides was as expected for this subtype (Sakurai et al., 1990) and, in contrast to the ETA assay, all peptides tested appeared to be full agonists. We carried out saturation binding assays in both the ETA and ETB expressing CHO-K1 cells and confirmed that receptor densities in both cell lines was in the same range as expressed in native human tissues, for example human heart (91.7 ± 3.2 fmol mg− 1 protein, (Peter and Davenport, 1996)) and coronary artery (20.6 ± 2.1 fmol mg− 1 protein (Bacon et al., 1996)). Recruitment of β-arrestin, following agonist binding to a GPCR, results in the uncoupling of the receptor from the associated G-protein

ET-1 +0.3µM Bosentan + 3µM Bosentan + 30µM Bosentan

ETA Assay (pKB) 8.01 ± 0.02 8.73 ± 0.08 8.38 ± 0.07

n = number of replicates. ND = not determined.

n = number of replicates. ND = not determined.

A

547

100

0 10

8

6

-Log10 [ET-1] M

10

8

6

-Log10 [ET-1] M

Fig. 3. Concentration-dependent inhibition of ET-1 responses in the ETA mediated β-arrestin assay by (A) bosentan, (B) ambrisentan and (C) sitaxentan and in the ETB-mediated β-arrestin assay by (D) bosentan, (E) ambrisentan and (F) BQ788.

548

J.J. Maguire et al. / Life Sciences 91 (2012) 544–549

Table 4 Antagonist affinities, ETA KB, determined against ET-1 mediated vasoconstriction in human saphenous vein by Schild analysis compared to ETA KD obtained in competition binding experiments using human left ventricle.

Bosentan Ambrisentan Sitaxsentan BQ123

KB⁎

n

KD

n

KB/KD

1.59 μM 220 nM 64.6 nM 141.3 nM

4 6 8 8

77.6 ± 7.9 nM 0.28 ± 0.23 nM 1.65 ± 0.80 nM 0.73 ± 0.22 nM

3 4 4 3

21 785 39 194

*Slopes of Schild regressions not significantly different from one. Therefore, KB values were calculated from Schild regressions with slopes constrained to unity.

study provide evidence that the affinity of BQ123 as an antagonist of ETA-mediated β-arrestin recruitment was agonist-dependent, consistent with an allosteric mode of action of this compound [44] (Sokolowsky, 1993) whereas in the ETB assay BQ788 did not distinguish between the different agonists. Whether non-peptide antagonists also show characteristics consistent with allosteric modulation of endothelin receptor function requires further investigation. The possibility that ETA antagonists may act as allosteric modulators at this receptor has been discussed in a recent review (De Mey et al., 2011). Of particular interest was our observation that BQ123 was least effective in antagonising the response to ET-2. There is emerging evidence that ET-2 may have distinct physiological/pathophysiological roles to ET-1 (Ling et al., in press) for example in ovulation (Al-Alem et al., 2007; Bridges et al., 2010) and cancer (Grimshaw et al., 2002; Tanese et al., 2010). It may therefore be important to determine to what extent endothelin antagonists designed for clinical use are effective blockers of both ET-1 and ET-2 function. In the ETA β-arrestin assay we observed that similar to BQ123, the non-peptide antagonists, ambrisentan, sitaxsentan and bosentan, produced rightward shifts of the concentration–response curves to ET-1 with an antagonists concentration-dependent reduction in ET-1 maximum response, consistent with a non-competitive mode of action. This contrasted with the apparent competitive actions of these antagonists determined against ET-1 contraction in human isolated saphenous vein. A possible explanation is that hemi-equilibrium conditions are achieved in the β-arrestin assay. The reduction in maximum agonist response in the presence of increasing concentrations of antagonist was less pronounced in the ETB assay and this may result from the higher density of receptors in this assay compared to the ETA assay resulting in greater ETB receptor reserve and preservation of maximum response over a wider antagonist concentration range. The established ETA selectivity of ambrisentan and sitaxentan was confirmed in the human ETA and ETB β-arrestin assays however, bosentan, which does not distinguish between the two receptor subtypes expressed in human heart, was an ETA selective antagonist in the β-arrestin assay with an affinity comparable to that predicted by the human heart binding data. It is now recognised that in addition to a role in GPCR desensitisation and internalisation, β-arrestin may also activate GPCR-mediated signalling pathways such as the MAP kinase cascade (Daake et al., 1998; Ferguson, 2001). ET-1 has been shown to stimulate pulmonary, but not aortic, smooth muscle cell migration by activating ERK1/2 MAP kinase (Meoli and White, 2010) and this effect of ET-1 may be amplified by the bone morphogenetic system (Yamanaka et al., 2010) and contribute to the remodelling of pulmonary arteries that occurs in pulmonary arterial hypertension. Our data suggest that bosentan exhibits ETA bias for β-arrestin recruitment. Whether preferential blockade of ETA mediated β-arrestinstimulated ERK1/2/ signalling contributes to the clinical benefit of bosentan in patients with PAH remains to be investigated. Conclusion The relative potency of endothelin and related peptide in the ETA and ETB mediated β-arrestin recruitment assays described in this study was as expected for the two receptor subtypes. However, this

study revealed that at the ETA receptor, compared to ET-1 all other agonists tested were partial agonists. Differences in the predicted pharmacology of antagonists were also revealed in the β-arrestin assays. Specifically, the peptide antagonist BQ123 behaved as an allosteric modulator of agonist activity in the ETA assay, exhibiting agonist-dependent affinities. The non-peptide compound bosentan was a potent ETA-selective antagonist in the β-arrestin assays in contrast to its non-selective profile determined in competition binding assays in human heart that possesses both ETA and ETB receptors. The significance of these data requires further understanding of the potential contribution of β-arrestin mediated signalling to endothelin receptor function in diseases such as pulmonary arterial hypertension. Conflict of interest statement Andrew Green, Mike Brown, Sanj Kumar, Tom Wehrman and Elizabeth Quinn are employees of DiscoveRx Corporation.

Acknowledgements JJM was supported by the British Heart Foundation (PG/09/050/ 27734). The authors would like to thank Pfizer for providing sitaxentan and Gilead Sciences for ambrisentan. We thank the Pulmonary Hypertension Association UK for financial support. This study was supported in part by the NIHR Cambridge Biomedical Research Centre. References Al-Alem L, Bridges PJ, Su W, Gong MC, Iglarz M, Ko C. Endothelin-2 induces oviductal contraction via endothelin receptor subtype A in rats. J Endocrinol 2007;193:383–91. Arai H, Hori S, Aramori I, Ohkubo H, Nakanishi S. Cloning and expression of a cDNA encoding an endothelin receptor. Nature 1990;348:730–2. Arunlakshana O, Schild HO. Some quantitative uses of drug antagonists. Br J Pharmacol 1959;14:48–58. Bacon CR, Cary NR, Davenport AP. Endothelin peptide and receptors in human atherosclerotic coronary artery and aorta. Circ Res 1996;79:794–801. Bodelsson G, Stjernquist M. Characterization of endothelin receptors and localization of 125I-endothelin-1 binding sites in human umbilical artery. Eur J Pharmacol 1993;249:299–305. Bremnes T, Paasche JD, Mehlum A, Sandberg C, Bremnes B, Attramadal H. Regulation and intracellular trafficking pathways of the endothelin receptors. J Biol Chem 2000;275:17596–604. Bridges PJ, Jo M, Al Alem L, Na G, Su W, Gong MC, et al. Production and binding of endothelin-2 (EDN2) in the rat ovary: endothelin receptor subtype A (EDNRA)mediated contraction. Reprod Fertil Dev 2010;22:780–7. Cheng JW. Ambrisentan for the management of pulmonary arterial hypertension. Clin Ther 2008;30:825–33. Clozel M, Breu V, Gray GA, Kalina B, Löffler BM, Burri K, et al. Pharmacological characterization of bosentan, a new potent orally active non-peptide endothelin receptor antagonist. J Pharmacol Exp Ther 1994;270:228–35. Cramer H, Müller-Esterl W, Schroeder C. Subtype-specific endothelin-A and endothelin-B receptor desensitisation correlates with differential receptor phosphorylation. J Cardiovasc Pharmacol 1998(31 Suppl. 1):S203–6. Daake Y, Luttrell LM, Ahn S, Della Rocca GJ, Ferguson SSG, Caron MG, et al. Essential role for G protein-coupled receptor endocytosis in the activation of mitogen-activated protein kinase. J Biol Chem 1998;273:685–8. Dai X, Galligan JJ. Differential trafficking and desensitisation of human ET(A) and ET(B) receptors expressed in HEK 293 cells. Exp Biol Med (Maywood) 2006;231:746–51. Davenport AP, Kuc RE. Radioligand-binding and molecular-imaging techniques for the quantitative analysis of established and emerging orphan receptor systems. Methods Mol Biol 2005;306:93-120. Davie NJ, Schermuly RT, Weissmann N, Grimminger F, Ghofrani HA. The science of endothelin-1 and endothelin receptor antagonists in the management of pulmonary arterial hypertension: current understanding and future studies. Eur J Clin Invest 2009(39 Suppl. 2):38–49. De Mey JG, Compeer MG, Lemkens P, Meens MJ. ETA-receptor antagonists or allosteric modulators? Trends Pharmacol Sci 2011;32:345–51. Dingemanse J, van Giersbergen PL. Clinical pharmacology of bosentan, a dual endothelin receptor antagonist. Clin Pharmacokinet 2004;43:1089–115. Drake MT, Shenoy SK, Lefkowitz RJ. Trafficking of G protein-coupled receptors. Circ Res 2006;99:570–82. Ferguson SS. Evolving concepts in G protein-coupled receptor endocytosis: the role in receptor desensitization and signalling. Pharmacol Rev 2001;53:1-24. Gardiner SM, Kemp PA, March JE, Bennett T, Davenport AP, Edvinsson L. Effects of an ET1-receptor antagonist, FR139317, on regional haemodynamic responses to endothelin-1 and [Ala11,15]Ac-endothelin-1(6–21) in conscious rats. Br J Pharmacol 1994;112(2):477–86. Grimshaw MJ, Wilson JL, Balkwill FR. Endothelin-2 is a macrophage chemoattractant: Implications for macrophage distribution in tumors. Eur J Immunol 2002;32:2393–400.

J.J. Maguire et al. / Life Sciences 91 (2012) 544–549 Johnström P, Harris NG, Fryer TD, Barret O, Clark JC, Pickard JD, et al. (18)F-Endothelin1, a positron emission tomography (PET) radioligand for the endothelin receptor system: radiosynthesis and in vivo imaging using MicroPET. Clin Sci (Lond) 2002;103(Suppl. 48):4S–8S. Johnström P, Fryer TD, Richards HK, Barret O, Clark JC, Pickard JD, et al. Positron emission tomography using 18F-labelled endothelin-1 reveals prevention of binding to cardiac receptors owing to tissue-specific clearance by ETB receptors in vivo. Br J Pharmacol 2005;144(1):115–22. Johnström P, Rudd JH, Richards HK, Fryer TD, Clark JC, Weissberg PL, et al. Imaging endothelin ET(B) receptors using [18F]-BQ3020: in vitro characterization and positron emission tomography (MicroPET). Exp Biol Med (Maywood) 2006;231:736–40. Kelland NF, Kuc RE, MacLean DL, Azfer A, Bagnall AJ, Gray GA, et al. Endothelial cell-specific ETB receptor knockout autoradiographic and histological characterisation and crucial role in the clearance of endothelin-1. Can J Physiol Pharmacol 2010;88:644–51. Kenakin TP. Allotopic, non-competitive, and irreversible antagonism. In: Kenakin TP, editor. Pharmacologic analysis of drug-receptor interaction. New York: Raven Press; 1993. p. 323–43. Ling L, Maguire JJ, Davenport AP. Endothelin-2, the Forgotten Isoform: Emerging Role in the Cardiovascular System, Ovarian Development, Immunology and Cancer. Br J Pharmacol, in press, doi:10.1111/j.1476-5381.2011.01786.x [Electronic publication ahead of print]. Lipa JE, Neligan PC, Perreault TM, Baribeau J, Levine RH, Knwlton RJ, et al. Vasoconstriction of endothelin-1 in human skin: role of ETA and ETB receptors. Am J Physiol 1999;276:H359–67. Maguire JJ. In: Maguire JJ, Davenport AP, editors. Endothelin-converting enzyme activity in vascular smooth muscle preparations in vitro. Peptide research protocols: endothelin New Jersey: Humana Press; 2002. p. 165–77. Maguire JJ, Davenport AP. ETA receptor-mediated constrictor responses to endothelin peptides in human blood vessels in vitro. Br J Pharmacol 1995;115:191–7. Maguire JJ, Davenport AP. Importance of sub-type selectivity for endothelin receptor antagonists in the human vasculature. In: Handler C, Abraham D, Dashwood M, Coghlan G, editors. Translational vascular medicine: pathogenesis, diagnosis, and treatment. Springer-Verlag London Limited; 2012. p. 151–72. Maguire JJ, Kuc RE, Rous BA, Davenport AP. Failure of BQ123, a more potent antagonist of sarafotoxin 6b than of endothelin-1, to distinguish between these agonists in binding experiments. Br J Pharmacol 1996;118:335–42. Meoli DF, White RJ. Endothelin-1 induces pulmonary but not aortic smooth muscle migration by activating ERK1/2 MAP kinase. Can J Physiol Pharmacol 2010;88:830–9. Molenaar P, O'Reilly G, Sharkey A, Kuc RE, Harding DP, Plumpton C, et al. Characterization and localization of endothelin receptor subtypes in the human atrioventricular conducting system and myocardium. Circ Res 1993;72:526–38. Nilsson T, Cantera L, Adner M, Edvinsson L. Presence of contractile endothelin-A and dilatory endothelin-B receptors in human cerebral arteries. Neurosurgery 1997;40: 346–51. Oksche A, Boese G, Horstmeyer A, Furkert J, Beyermann M, Bienert M, et al. Late endosomal/lysosomal targeting and lack of recycling of the ligand-occupied endothelin B receptor. Mol Pharmacol 2000;57:1104–13.

549

Palmer MJ. Endothelin receptor antagonists: status and learning 20 years on. Prog Med Chem 2009;47:203–37. Peter MG, Davenport AP. Characterization of the endothelin receptor selective agonist, BQ3020 and antagonists BQ123, FR139317, BQ788, 50235, Ro462005 and bosentan in the heart. Br J Pharmacol 1996;117:455–62. Russell FD, Davenport AP. Characterization of the binding of endothelin ETB selective ligands in human and rat heart. Br J Pharmacol 1996;119:631–6. Sakurai T, Yanagisawa M, Takuwa Y, Miyazaki H, Kimura S, Goto K, et al. Cloning of a cDNA encoding a non-isopeptide-selective subtype of the endothelin receptor. Nature 1990;348:732–5. Salom JB, Torregrosa G, Barberá MD, Jover T, Alborch E. Endothelin receptors mediating contraction in goat cerebral arteries. Br J Pharmacol 1993;109:826–30. Schulte G, Levy FO. Novel aspects of G-protein-coupled receptor signalling-different ways to achieve specificity. Acta Physiol (Oxf) 2007;190:33–8. Sokolowsky M. BQ-123 identifies heterogeneity and allosteric interactions at the rat heart endothelin receptor. Biochem Biophys Res Commun 1993;196:32–8. Spieker LE, Lüscher TF, Noll G. ETA receptors mediate vasoconstriction of large conduit arteries during reduced flow in humans. J Cardiovasc Pharmacol 2003;42:315–8. Takayanagi R, Kitazumi K, Takasaki C, Ohnaka K, Aimoto S, Tasaka L, et al. Presence of non-selective type of endothelin receptor on vascular endothelium and its linkage to vasodilatation. FEBS Lett 1991;282:103–6. Tanese K, Fukuma M, Ishiko A, Sakamoto M. Endothelin-2 is upregulated in basal cell carcinoma under control of hedgehog signaling pathway. Biochem Biophys Res Commun 2010;391:486–91. Vatter H, Seifert V. Ambrisentan, a non-peptide endothelin receptor antagonist. Cardiovasc Drug Rev. 2006;24(1):63–76. Vatter H, Zimmermann M, Jung C, Weyrauch E, Lang J, Seifert V. Effect of the novel endothelin(A) receptor antagonist LU 208075 on contraction and relaxation of isolated rat basilar artery. Clin Sci (Lond) 2002(Suppl. 48):408S–13S. Wang J, Chiou WJ, Gagne GD, Wu-Wong JR. Internalization of type-A endothelin receptor. J Cardiovasc Pharmacol 2000;36(5 Suppl. 1):S61–5. Whalen EJ, Rajagopal S, Lefkowitz RJ. Therapeutic potential of β-arrestin and G protein biased agonists. Trends Mol Med 2011;17:126–39. Wu C, Chan MF, Stavros F, Raju B, Okun I, Mong S, et al. Discovery of TBC11251, a potent, long acting, orally active endothelin-A selective antagonist. J Med Chem 1997;40: 1690–7. Wu-Wong JR, Chiou WJ, Magnuson SR, Opgenorth TJ. Endothelin receptor in human astrocytoma U37 3MG cells: binding, dissociation, receptor internalization. J Pharmacol Exp Ther 1995;274:499–507. Yamanaka R, Otsuka F, Nakamura K, Yamashita M, Otani H, Matsumoto Y, et al. Involvement of the bone morphogenetic protein system in endothelin- and aldosteroneinduced cell proliferation of pulmonary arterial smooth muscle cells isolated from human patients with pulmonary arterial hypertension. Hypertens Res 2010;33: 435–45. Zheng H, Loh HH, Law PY. Agonist-selective signaling of G protein-coupled receptor mechanisms and implication. IUBMB Life 2010;62:112–9.