I. Introduction

In mammals, the endothelin (ET¹) family comprises three endogenous isoforms, ET-1, ET-2, and ET-3 (Yanagisawa et al., 1988; Inoue et al., 1989). These peptides mediate their actions via two receptor types, classified as the ET_A and ET_B receptors in the first NC-IUPHAR (International Union of Pharmacology Committee on Receptor Nomenclature and Drug Classification) receptors report on nomenclature by Masaki et al. (1994). This report was compiled before the discovery of the majority of ET receptor antagonists (particularly nonpeptides) currently used in the characterization of receptors and now updated in the present review. The single proposed ET_B receptor antagonist included by Masaki et al. (1994) was subsequently shown to lack efficacy and was withdrawn by the original discoverers (Urade et al., 1994). This review reflects both peptide and nonpeptide ET_B antagonists that are the compounds of choice. A second major change is the inclusion of radioligands that are highly selective for either ET_A (¹²⁵I-PD151242,¹²⁵I-PD164333, and ³H-BQ123 (cyclo-[d-Asp-l-Pro-d-Val-l-Leu-d-Trp-]) or ET_B receptors [¹²⁵I-BQ3020 ([Ala^11,15]Ac-ET-1_(6–21)) and ¹²⁵I-IRL1620 (Suc-(Glu⁹, Ala^11,15)-ET-1_(8–21))] that have been crucial to consolidating the present classification.Masaki et al. (1994) referred to a gene encoding a third receptor (P32940) cloned from the amphibian Xenopus laevis dermal melanophores, which was reported to be ET-3 specific (ET-3 > ET-1), the so-called ET_C receptor. However, to date, molecular and ligand binding techniques have failed to identify a mammalian homolog. Since NC-IUPHAR is predominantly interested in human receptors, with an extension to mammalian receptors, this review follows the recommendation to exclude consideration of nonmammalian receptors in the classification.

ET-1 is the principal isoform in the human cardiovascular system and remains the most ubiquitous, potent, and unusually long-lasting constrictor of human vessels discovered. ET-1 is unusual among the mammalian bioactive peptides in being released from a dual secretory pathway (Russell et al., 1998). The peptide is continuously released from vascular endothelial cells by the constitutive pathway, producing intense constriction of the underlying smooth muscle and contributing to the maintenance of endogenous vascular tone (Haynes and Webb, 1994). The peptide is also released from endothelial cell-specific storage granules (Weibel-Palade bodies) in response to external physiological, or perhaps pathophysiological, stimuli producing further vasoconstriction (Russell et al., 1998). Thus, ET-1 functions as a locally released, rather than circulating, hormone, and concentrations are comparatively low in plasma and other tissues. ET-2 has been less extensively studied than other ET peptides, but it is present in human cardiovascular tissues and was as potent a vasoconstrictor as ET-1 in human arteries and veins (Maguire and Davenport, 1995). Endothelial cells do not synthesize ET-3, but the mature peptide is detectable in plasma (Matsumoto et al., 1994) and other tissues, including heart and brain. ET-3 is unique in that it is the only endogenous isoform that distinguishes between the two endothelin receptors. It has the same affinity at the ET_B receptor as ET-1 but, at physiological concentrations, has little or no affinity for the ET_A.

The only endogenous peptides with a high degree of sequence similarity to the ETs are the sarafotoxins (S6a, S6b, S6c, and S6d). This family of 21-amino acid (aa) peptides was originally discovered in the venom of a snake, Atractaspis engadensis (Takasaki et al., 1988).

II. Cloned Endothelin Receptors

Receptors can be identified by their amino acid structure and provide unambiguous evidence for expression of a gene encoding a particular type in specific cells or tissues. An increasing number of mammalian species have been studied, but only two ET receptors have been isolated and cloned (Table 1;Arai et al., 1990; Sakurai et al., 1990; Adachi et al., 1991; Lin et al., 1991; Nakamuta et al., 1991; Saito et al., 1991; Baynash et al., 1994). The deduced amino acid sequences for the two human receptors display only 59% similarity and are shown in Table2. The amino acid sequences of ET_A receptors also differ between humans and other species, for example by 9% between human and rat ET_A receptors and by 12% for the ET_B. These may contribute to differences in efficacy and potency of selective agonists and antagonists.

The structures of the mature receptors have been deduced from the nucleotide sequences of the cDNAs. The encoded proteins contain seven stretches of 20 to 27 hydrophobic aa residues in both receptors, consistent with both subtypes belonging to the seven-transmembrane (7TM) domain , G protein-coupled rhodopsin-type receptor superfamily. Both receptors have an N-terminal signal sequence, which is rare among heptahelical receptors, with a relatively long extracellular N-terminal portion preceding the first transmembrane domain. There are two separate ligand interaction subdomains on each endothelin receptor. The extracellular loops, particularly between TM 4 to 6, determine selectivity.

At present, there is no justification for further types beyond the current classification into ET_A and ET_B in mammalian tissue. Functional studies have suggested that PD142893 [Ac-(beta-phynyl)d-Phe-l-Leu-l-Asp-l-lle-l-lle-l-Trp], a hexapeptide antagonist, can block the vasodilator actions of ET-1 at endothelial ET_B receptors but not constrictor responses mediated by ET_B smooth muscle receptors (Warner et al., 1993; Douglas et al., 1995). However, in the ET_B receptor gene knockout mouse, both the PD142893-sensitive vasodilator response and the PD142893-resistant contractile response to the ET_B agonist sarafotoxin S6c were completely absent. These results indicate that the pharmacologically heterogeneous responses to S6c are mediated by ET_B receptors derived from the same gene (Mizuguchi et al., 1997). In agreement, a very detailed binding study (including PD142893) was unable to distinguish between ET_B receptors expressed by human isolated endothelial cells compared with smooth muscle cells in culture (Flynn et al., 1998). Furthermore, in human tissue, both ET_A- and ET_B-selective radiolabeled ligands bound with a single affinity and Hill slopes close to unity (Molenaar et al., 1992; Davenport et al., 1994, 1998;Davenport, 1997). Similarly, competition studies using unlabeled ligands provided no evidence for further subtypes (Peter and Davenport, 1996; Russell and Davenport, 1996).

III. Mammalian Splice Variants of Endothelin_A and Endothelin_B Receptors

Alternative splice variants of ET receptors have been reported but to date these variants either show little or no change in binding characteristics and their physiological or pathophysiological significance is unclear. The following is intended to be a guide only because the field has not developed sufficiently with unequivocal quantitative evidence for significant expression and function in native tissues rather than artificial cell lines, to make any firm recommendation for classification.

The existence of alternative splice variants of the ET_B receptor in human and porcine tissue has been reported. A variant human ET_B receptor that results in a 10-aa increase in the length of the second cytoplasmic domain has been described (Shyamala et al., 1994). Messenger RNA measured by reverse transcription-polymerase chain reaction in a limited number of human tissues was found only in low abundance in human brain (which expresses one of the highest densities of ET_B receptors) as well as the heart, lung, and placenta but was not detected in other species tested (bovine, porcine, and rat). The increase in amino acids did not result in any change in either ligand affinities or signal transduction (cAMP and inositol phosphate turnover), and the physiological importance of this variant receptor is unclear.

Elshourbagy et al. (1996) discovered a second splice variant from a human placental library. Analysis indicated that the deduced polypeptide was identical to the native ET_Bsequence except that the 42 aa of the intracellular carboxy terminus of the former was replaced with an alternative 36-aa sequence, bearing no significant homology with other known proteins. Northern blot analysis indicated an mRNA species of 2.7 kilobases, which was expressed in all of a limited number of human tissues tested (lung, placenta, kidney, and skeletal muscle) in addition to mRNA encoding the native ET_B receptor. However, mRNA encoding the variant was not particularly abundant. The relative ratio of each individual variant mRNA was less than 10% of the total ET_BmRNA, with the intriguing exception of skeletal muscle where it represented more than 40%. Two cell types were also examined, endothelial and smooth muscle cells, but only mRNA encoding the native receptor was detected. The cloned variant receptors expressed in COS cells displayed similar binding properties for ET peptides compared with expressed native receptors in the same cells, indicating unsurprisingly that the splice variant had little or no effect on ligand binding. However, functional studies showed that ET-stimulated inositol phosphate accumulation in expressed native receptors was abolished in cells transfected with the splice variant. These data suggest the difference in the amino acid sequences between the two receptors may alter functional coupling in the variant receptor.

Nambi et al. (2000) detected a novel cDNA from another species, porcine cerebellum, that was predicted to encode an ET_Breceptor also with alternate splicing of the carboxy terminus, resulting in a deduced polypeptide of 429 aa, 14 residues shorter than the wild-type receptor. The relative abundance of mRNA encoding the splice variant compared with the wild-type receptor was not reported, but mRNA was detected in ET_B-rich tissues including porcine lung, kidney, and cerebellum. However, the splice variant did not alter the binding of radiolabeled ET-1 or functional coupling when expressed in COS cells. The lack of effect on inositol phosphate accumulation is in marked contrast to the human variant (see above) previously described by this group. Combined with the lack of sequence similarity between the human (38 aa) and porcine (29 aa) carboxy terminal splice variants, it is not clear whether the porcine variant is a homolog of the human or whether these are distinct splice variants.

Cheng et al. (1993) identified cDNA from rat brain, which they described as producing a receptor protein with four amino acid substitutions that displayed equal affinity for the three ET isoforms. However, Cheng et al. (1993) probably described the correct rat ET_B sequence, correcting a sequencing error in the previously deduced sequence of Sakurai et al. (1990), for the following reasons. The Cheng et al. (1993) sequence has 3 extra bases in a 9-base span, which corrects a pair of adjacent frameshifts in the Sakurai et al. sequence, making the DNA sequence identical to the mouse sequence (Hosoda et al., 1994) in the same region and matching 3 of 4 amino acids in the human sequence as opposed to 0 of 4 with the Sakurai et al. sequence. Cheng et al. (1993) also report a different sequence in the 5′-untranslated region, which could be an alternative first exon, reflecting transcription initiating from an alternative promoter. It is also possible, although less likely, that one of the 5′-untranslated region sequences is an artifact, such as a chimeric cDNA or sequencing assembly error.

The human ET_A receptor gene has been proposed to give rise to at least three alternatively spliced ET_A receptor transcripts corresponding to deletion of exon 3 (producing a protein with two membrane-spanning domains), exon 4 (producing a protein with three membrane-spanning domains), and exon 3 plus exon 4 (producing a protein lacking the third and fourth domain) (Miyamoto et al., 1996; Bourgeois et al., 1997). Although alternative transcripts were identified in human tissues including lung, aorta, and atrium, the truncated receptors when expressed in COS cell lines did not bind ET-1 (Miyamoto et al., 1996), and a physiological role remains unclear. Intriguingly, mRNA encoding the putative truncated receptor with the deletion of exon 3 plus 4 was more abundant than the wild type in human melanoma cell lines and melanoma tissue (Zhang et al., 1998).

IV. Physiological Role of Receptors

Endothelin receptors are widely expressed in all tissues, which is consistent with the physiological role of endothelins as ubiquitous endothelium-derived vasoactive peptides, contributing to the maintenance of vascular tone. In humans, ET_Areceptors predominate on the smooth muscle of blood vessels, and the low density of ET_B receptors (<15%) also present on the smooth muscle contributes little to vasoconstriction in either normal or diseased tissue (Maguire and Davenport, 1995). ET_B receptors are the principal type in the kidney, localizing to nonvascular tissues. Evidence is emerging that the ET_B receptor functions as a “clearing receptor” to remove ET from the circulation. ET_B receptors localized to the single layer of endothelial cells that line all blood vessels, may play a role in the release of endothelium-derived relaxing factors, such as nitric oxide and prostanoids (Warner et al., 1989), where all three isoforms have a similar potency (de Nucci et al., 1988). Although ET_A receptors present on smooth muscle cells are mainly responsible for contraction throughout the human vasculature, the situation in animals is more complex since the relative contribution from activating constrictor ET_Breceptors can vary, depending on the species and vascular bed. In some blood vessels, such as the rabbit saphenous vein, rabbit jugular vein, rat renal vascular bed, and porcine pulmonary vein, ET_B receptors mediate vasoconstriction. In other vessels, ET-1 is thought to mediate vasoconstriction by activating both receptors.

Receptors are also localized to nonvascular structures, such as epithelial cells, as well as occurring in the central nervous system on glia and neurones. Endothelin stimulates proliferation in a number of different cell types, including smooth muscle cells (mainly via the ET_A subtype) or astrocytes (ET_B). In most of these cells, ET is thought to be comitogenic, potentiating the actions of other growth factors such as platelet-derived growth factor.

V. Endogenous and Synthetic Agonists

ET receptors continue to be classified (Table3; Davenport, 2000) according to their rank order of potency for the endogenous ET isoforms. A selective ET_A receptor agonist has not been discovered.

Sarafotoxin S6c is one of the most widely used ET_B-selective agonists, displaying over 200,000-fold selectivity in the rat (William et al., 1991), although the peptide is much less selective in human tissues, perhaps reflecting species differences in the receptors (Russell and Davenport, 1996). [Ala^1,3,11,15]ET-1 (Saeki et al., 1991), the linear analog of ET-1 in which the disulfide bridges have been removed by substitution of Ala for Cys residues, is ET_B-selective. The truncated linear synthetic analogs BQ3020 and IRL1620 are the most widely used selective synthetic agonists to characterize ET_B receptors. The compounds cause endothelium-dependent vasodilatation in preparations such as porcine pulmonary artery, which is consistent with ET_B receptor-mediated release of relaxing factors from the endothelium.

VI. Radiolabeled Agonists

Most studies characterizing and localizing ET receptors use¹²⁵I-ET-1, directly labeled via the Tyr¹³ (Table 3). This ligand binds with the same affinity to both ET_A and ET_B receptors and is stable under nonphysiological binding conditions with little or no degradation of labeled ET-1 being detected. ¹²⁵I-ET-2,¹²⁵I-vasoactive intestinal contractor (the murine isoform of ET-2), and ¹²⁵I-sarafotoxin 6b have been labeled and used in saturation assays where they also bind to both receptors (Davenport and Morton, 1991; Maguire et al., 1996).

ET-3 can be labeled at Tyr⁶, Tyr¹³, and Tyr¹⁴. Tyr⁶ is generally used, as it is more difficult to separate ¹²⁵I-ET-3 labeled at the latter two Tyr residues, although all three ET-3 ligands have similar affinities. The selectivity of ET-3 for ET_B versus ET_A receptors is often only about two orders and it is difficult to precisely delineate the two receptors using this labeled peptide in saturation assays. ET_Breceptors are usually characterized using¹²⁵I-BQ3020 (Ihara et al., 1992b; Molenaar et al., 1992), which binds with subnanomolar affinity to the ET_B receptor, with at least 1500-fold selectivity for this receptor over the ET_A. Alternatively, the truncated analog ¹²⁵I-IRL1620 can also be used, particularly in animal tissues (Watakabe et al., 1992).

VII. Antagonists

Antagonists are currently classified as either ET_A--selective, ET_B-selective, or mixed antagonists that display similar affinity for both receptors. The most highly selective peptide antagonists (4 to 5 orders of selectivity) for the ET_A receptors are the cyclic pentapeptide BQ123 (Ihara et al., 1992a) and the modified linear peptide FR139317 (N-[(hexahydro-1-azepinyl)carbonyl]l-Leu(1-Me)d-Trp-3(2-pyridyl)-d-Ala;Aramori et al., 1993). Unlike peptide antagonists, many nonpeptide ET_A receptor-selective antagonists have oral bioavailability and some may cross the blood-brain barrier. The majority are more potent, with pA₂ values of up to 10 compared with 7 or 8 for BQ123 or FR139317, but are less selective, and plasma binding may also be significant in vivo.

VIII. Radiolabeled Endothelin_A Selective Antagonists

¹²⁵I-PD151242 is widely used to characterize and localize ET_A receptors. This linear tetrapeptide analog of FR139317, binds with subnanomolar affinity to the ET_A receptor and has about 10,000-fold selectivity for this receptor in human and animal tissues. A nonpeptide ET_A-selective ligand has also been developed,¹²⁵I-PD164333 (Davenport et al., 1998) with comparable affinity as well as a tritiated ligand,³H-BQ123 (Ihara et al., 1995). The above ligands are available commercially either as catalog items or custom syntheses.

IX. Endothelin_B Selective Antagonists

A more limited number of peptide (e.g., BQ788) and nonpeptide (e.g., A192621) ET_B antagonists have been developed, reflecting the lack of clinical need for this type of compound. They are less potent than ET_Aantagonists and display lower selectivity (usually only 1 to 2 orders of magnitude) for the ET_B receptor (Table 3).

X. Endothelin_A/Endothelin_B Antagonists

The distinction between antagonists that are ET_A-selective and those that block both ET_A and ET_B receptors is not precise but generally the former display greater than 100-fold selectivity for the ET_A subtype, and the latter less than 100-fold. These compounds are seldom reported as having equal affinity for both receptors, and this should be taken into consideration in experimental designs. Nonpeptide compounds included bosentan (RO470203, Tracleer; Actelion, San Francisco, CA) (Clozel et al., 1994), SB209670 (Elliott et al., 1994), SB217242 (enrasentan; Ohlstein et al., 1996), and RO610612 (tezosentan; Clozel et al., 1999). Plasma binding may also be significant in vivo.

XI. Conclusions

In humans, ET peptides mediate their actions via only two receptor types, classified as ET_A and ET_B. There is no strong evidence to support the existence of further receptors in mammals. Further research is required to establish whether any of the potential splice variants in the ET_B receptor have a physiological or pathophysiological role.