| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Article |
Division of Cellular and Molecular Pharmacology, Department of Experimental Medical Science, Lund University, Lund, Sweden (L.M.F.L.-L.); Centre de Recherche, Centre Hospitalier Universitaire de Quebec, Quebec, Canada (F.M.); Institute of Biochemistry II, Johann Wolfgang Goethe University School of Medicine, Frankfurt, Germany (W.M.-E.); Departments of Medicinal Chemistry and Neuroscience, Merck Research Laboratories, West Point, Pennsylvania (D.J.P.); and Department of Medicine, Veterans Affairs Medical Center and University of California, San Diego, California (B.L.Z.)
Abstract
Abstract I. A Short History of Kinins and Their Receptors II. Pharmacological Classification of Kinin Receptor Subtypes A. Peptide Agonists B. Peptide Antagonists C. Nonpeptide Ligands 1. B2 Receptor Agonists. 2. B2 Receptor Antagonists. 3. B1 Receptor Antagonists. III. Structural Aspects of Kinin Receptors and Their Genes A. Organization and Structure of the Receptor Genes B. Receptors and Their Post-Translational Modifications 1. Glycosylation. 2. Disulfide Bridging. 3. Acylation. 4. Phosphorylation. C. Agonist and Antagonist Binding Sites in the Receptors 1. B2 Receptor Agonists. 2. B2 Receptor Antagonists. 3. B1 Receptor Agonists. 4. B1 Receptor Antagonists. D. Evolutionary Aspects of Kinin Receptors IV. Molecular and Cellular Aspects of Kinin Receptor Signaling and Regulation A. Agonist-Dependent and -Independent Mechanisms of Receptor Activation B. Receptor Cellular Signaling Pathways C. Protein-Protein Interactions in Receptor Signaling D. Receptor Desensitization E. Cellular Distribution and Trafficking of Receptors V. Long-Term Regulation of Kinin Receptors by Proinflammatory Factors A. Postisolation Induction of the B1 Receptor B. Proinflammatory Cytokines and Growth Factors C. Agonists D. Ras and B2 Receptors E. Regulatory Elements in the Gene Promoters F. mRNA Stability VI. Distribution and Pathophysiological Function of Kinin Receptors A. Circulation and Renal Function B. Inflammation C. Pain and Neurology D. Diabetes VII. Kinin Receptors and Human Disease A. Cardiovascular Disease 1. Left Ventricular Hypertrophy and Cardiomyopathy. 2. Vascular Tone. 3. Hypertension. B. Renal Disease C. Airway Disease D. Neurological Disease E. Cancer F. Other Disease States G. Caveats VIII. Kinin Receptors and Drug Development A. Pain B. Cardiovascular Function C. Airway Function D. Cancer IX. Epilogue
Kinins are proinflammatory peptides that mediate numerous vascular and pain responses to tissue injury. Two pharmacologically distinct kinin receptor subtypes have been identified and characterized for these peptides, which are named B1 and B2 and belong to the rhodopsin family of G protein-coupled receptors. The B2 receptor mediates the action of bradykinin (BK) and lysyl-bradykinin (Lys-BK), the first set of bioactive kinins formed in response to injury from kininogen precursors through the actions of plasma and tissue kallikreins, whereas the B1 receptor mediates the action of des-Arg9-BK and Lys-des-Arg9-BK, the second set of bioactive kinins formed through the actions of carboxypeptidases on BK and Lys-BK, respectively. The B2 receptor is ubiquitous and constitutively expressed, whereas the B1 receptor is expressed at a very low level in healthy tissues but induced following injury by various proinflammatory cytokines such as interleukin-1
. Both receptors act through G
q to stimulate phospholipase C followed by phosphoinositide hydrolysis and intracellular free Ca2+ mobilization and through Gi to inhibit adenylate cyclase and stimulate the mitogen-activated protein kinase pathways. The use of mice lacking each receptor gene and various specific peptidic and nonpeptidic antagonists have implicated both B1 and B2 receptors as potential therapeutic targets in several pathophysiological events related to inflammation such as pain, sepsis, allergic asthma, rhinitis, and edema, as well as diabetes and cancer. This review is a comprehensive presentation of our current understanding of these receptors in terms of molecular and cell biology, physiology, pharmacology, and involvement in human disease and drug development.
I. A Short History of Kinins and Their Receptors
Kinin receptors have come a long way since Rocha e Silva made some speculations about their dual nature in his visionary book on "Kinin Hormones" (Rocha e Silva, 1970
). He concluded in 1970 that "... it is perhaps a little early to deduce any convincing configuration for the bradykinin receptors from the data obtained by studying the biological actions of bradykinin analogs and derivatives" (Rocha e Silva, 1970). Indeed, it took almost a quarter of a century before the first kinin receptor cDNA was cloned (McEachern et al., 1991), and more than a dozen years later the three-dimensional structure of a kinin receptor is still elusive.
Tracing back the history of kinin receptors inevitably leads to the early work of the 1930's when Werle and Frey described kallikrein, the kinin-producing enzyme from pancreas (Kraut et al., 1930), and identified the biological actions of lysyl-bradykinin (Lys-BK1) or kallidin, then called substance DK (Werle and Grunz, 1939). In the 1940's, the kinins came into the limelight with the identification of bradykinin (BK) by Rocha e Silva and his colleagues (Rocha e Silva et al., 1949). At the end of the 1950's, BK had been isolated in significant quantities from plasma (Elliott et al., 1959), and finally its primary structure was solved in a dramatic race between the groups of Boissonnas and Elliott (Boissonnas et al., 1960; Elliott et al., 1960a,b). Interestingly, BK was the first biologically active peptide to be assembled by the then newly developed solid phase peptide synthesis (Merrifield, 1964). By the end of the 1960's, hundreds of different kinin derivatives had been synthesized and analyzed for their biological effects in vitro and in vivo (Frey et al., 1967).
In the 1970's, Regoli and coworkers set the early landmarks in the molecular characterization of kinin receptors by pointing to the existence of two types of kinin receptors, B1 and B2, which differ in their pharmacological profiles as well as in their expression patterns (Regoli et al., 1977, 1978; Drouin et al., 1979) (Table 1 and Fig. 1). The development of specific agonists and antagonists to the B1 receptor by the same group was instrumental in establishing the dualism of kinin receptors in mammals (Regoli et al., 1977; Regoli and Barabé, 1980). At the same time, the first steps toward the solubilization and purification of the receptors were made by Goodfriend and coworkers (Odya and Goodfriend, 1979). In the 1980's, a breakthrough came from the development of the first bradykinin analogs with antagonist activity at the B2 receptor (Vavrek and Stewart, 1985).
|
|
The identification of the B2 receptor as a G protein-coupled receptor (GPCR), which signals through the phospholipase C pathway leading to inositol 3-phosphate (IP3) formation and intracellular Ca2+ mobilization and through the phospholipase A2 pathway resulting in arachidonic acid release, was another landmark discovery that paved the way for the identification of the many intracellular signaling routes triggered by the kinin receptors (Higashida et al., 1986; Burch and Axelrod, 1987). At the same time, the role of BK as a potent stimulator of endothelial nitric-oxide synthase (eNOS) via the B2 receptor signaling pathway was recognized (Palmer et al., 1987).
The search for kinin receptors culminated in the 1990s with the cloning of the rat B2 receptor cDNA by Jarnagin and coworkers (McEachern et al., 1991). Soon after, Hess and coworkers succeeded in cloning the first B1 receptor (Menke et al., 1994), and this was followed by the identification of the first nonmammalian kinin receptor, the chick ornithokinin receptor (Schroeder et al., 1997). More recently, the advent of the first nonpeptidic B2 receptor antagonists has opened new possibilities for the application of kinin receptor blockers in various diseases (Asano et al., 1997). Eventually the targeted ablation of the genes for the B2 (Borkowski et al., 1995) and B1 receptor (Pesquero et al., 2000) in mice has helped to define more precisely the (patho)physiological roles of kinin receptors. This goal was further accomplished with the most recent advent of double knockout mice lacking both kinin receptor genes rendering these animals resistant to LPS-induced septic shock (Cayla et al., 2002b). Clearly, routine clinical applications of kinin receptor antagonists in various disease states are to be expected soon.
II. Pharmacological Classification of Kinin Receptor Subtypes
Thousands of peptide kinin analogs have been synthesized and pharmacologically evaluated. Kininogens are defined as circulating proteins that contain the BK sequence, but Lys-BK is also a cleavage product of human kininogen (Fig. 2). These sequences and their metabolites without the C-terminal arginine residue (des-Arg9) assume particular importance as natural ligands of the two human receptor (R) subtypes (B1, B2; Table 2). The B2 receptor has a high affinity for "native" kinins (those generated by either plasma or tissue kallikreins), BK and kallidin, in all mammalian species. Kallidin has itself a significant affinity at the human and rabbit B1 receptors, but not at the mouse B1 receptor (Table 2). The most discriminative structural determinant for high affinity at all mammalian B1 receptors is the removal of the C-terminal arginine, which at the same time is detrimental for affinity toward the B2 receptor. This holds true for Lys-BK and Lys-des-Arg9-BK, which have higher affinities for the human and rabbit B1 receptors compared with BK and des-Arg9-BK, respectively (Table 2, Fig. 2). The only natural kinin sequence with a subnanomolar affinity for the human and rabbit B1 receptors is Lys-des-Arg9-BK, a fact of particular significance that may suggest that the B1 receptor belongs to the "tissue" kallikrein-kinin system, jointly with tissue kallikrein that generates Lys-BK from low molecular weight kininogen (L-kininogen). If one considers a more distantly related mammalian species, the pig, functional data (porcine renal vein contractility) show that the B1 receptor is also preferentially stimulated by Lys-des-Arg9-BK, 16-fold more potent than des-Arg9-BK (Rizzi et al., 1997). Thus, an order of agonist potency widely valid for mammalian kinin receptors could be: B1, Lys-des-Arg9-BK > Lys-BK 
des-Arg9-BK >> BK; B2, BK Lys-BK >> des-Arg9-BK and Lys-des-Arg9-BK.
|
|
However, the pharmacological profile varies significantly for this set of four natural peptides in the Muridae rodents (the laboratory rat and mouse; Ni et al., 1998a; Table 2). Lys-des-Arg9-BK has no decisive advantage over des-Arg9-BK in these species. Indeed, a form of coevolution of the B1 receptor with the kininogen genes may have occurred to maintain the B1 receptor function by conferring a subnanomolar affinity to des-Arg9-BK since the kininogen genes in rodents code for an arginine instead of a lysine residue N-terminal to the BK sequence or Ile-Ser in T-kininogen (Hess et al., 1996; Takano et al., 1997). The high potency of des-Arg9-BK is specific to these rodents (superior or equal to Lys-des-Arg9-BK in functional assays; as reviewed by Marceau et al., 1998).
Thus, the B1 receptor is obviously specialized across species to respond to different kinin metabolites, either des-Arg9-BK or Lys-des-Arg9-BK, generated by arginine carboxypeptidases, such as carboxypeptidase N and M. No other kinin fragment seems to retain pharmacological activity. For instance, des-Arg1-BK (generated by aminopeptidase P) or des(Phe8, Arg9)-BK [the primary metabolite generated by kininase II or angiotensin-converting enzyme (ACE)] do not retain significant activity on either known receptor type (Regoli and Barabé, 1980). The presence of arginine carboxypeptidases in functional systems frequently distorts the potency estimates for BK or Lys-BK on the B1 receptor, as these ligands are transformed into their respective des-Arg9 metabolites (discussed by Marceau et al., 1998). Thus, intact BK may exhibit a very high selectivity toward B2, and this makes [3H]BK an indisputably selective B2 receptor ligand.
Synthetic peptides incorporate modifications to improve resistance to metabolism. In labradimil (RMP-7, [Hyp3, Thi5, 4-Me-Tyr8 
(CH2-NH)Arg9]-BK), the peptide bond between the two C-terminal residues has been replaced by a nonhydrolyzable bond, resulting in complete resistance to kininases I (arginine carboxypeptidases) and II (ACE) (Table 2; Doctrow et al., 1994). Labradimil is a selective B2 receptor agonist developed to transiently open the blood-brain barrier to optimize the chemotherapy of intracranial tumors (Emerich et al., 2001). This compound has been used in clinical trials. An earlier version of a kininase-resistant BK analog with a similar design, [Phe8(CH2–NH)Arg9]BK, has been described (Drapeau et al., 1988). It permanently promotes the endocytosis of a fluorescent form of the rabbit B2 receptor in live cells, whereas the effect of BK is transient due to its degradation in serum-containing medium (Marceau et al., 2002). Sar-[D-Phe8]des-Arg9-BK is a B1 receptor agonist of high selectivity that is completely resistant to blood aminopeptidase, kininases I and II (ACE), and kidney neutral endopeptidase (Table 2; Drapeau et al., 1991, 1993). It does not exhibit a very high affinity (intermediate between des-Arg9-BK and Lys-des-Arg9-BK) but its prolonged duration of action is proven by its hypotensive effect in endotoxin-pretreated rabbits (Drapeau et al., 1991) and its efficacy as an angiogenesis-promoting agent in mice with an arterial occlusion (Emanueli et al., 2002).
Efforts to develop receptor antagonists by modifying the BK sequence date back to the 1970's (see Stewart, 1995 for review). The first family of compounds capable of antagonizing BK and des-Arg9-BK with specificity was based on the prototype [Leu8]des-Arg9-BK (Fig. 2 and Table 2; Regoli et al., 1977). The same group of investigators had previously developed angiotensin II antagonists by replacing the aromatic and C-terminal Phe residue with an aliphatic residue (Regoli et al., 1974). The B1 receptor nomenclature was later applied to the rabbit aortic preparation in which the kinin receptors were defined for the first time. This was based on both a typical order of agonist potency, with des-Arg9 fragment more potent than the native sequence, and by the affinity of antagonists modeled on [Leu8]des-Arg9-BK, thus satisfying the two first Schild criteria for receptor classification (Schild, 1973). However, these antagonists were not active in established bioassays for BK such as contraction of the rat uterus and guinea pig ileum in vitro or production of hypotension in vivo (Regoli and Barabé, 1980). Early compounds such as [Leu8]des-Arg9-BK and Lys-[Leu8]des-Arg9-BK had a complete selectivity, being devoid of affinity for the B2 receptor (Table 2). Structure-activity relationships for peptide antagonists of the B1 receptor are discussed elsewhere (Regoli and Barabé, 1980; Marceau et al., 1998; Gobeil et al., 1999). [Leu8]des-Arg9-BK, the prototype B1 receptor antagonist, may exhibit fairly high partial agonist behavior in some species, especially in the rats and mice (reviewed elsewhere, Marceau et al., 1998). This has practical implications because one of the emerging therapeutic applications of B1 receptor antagonists, analgesia, is commonly evaluated using behavioral models involving these species.
The B2 receptor was not well defined until 1985 when the first generation of antagonists based on [D-Phe7]BK were produced (Vavrek and Stewart, 1985). In these early compounds (e.g., [Thi5,8, D-Phe7]BK), the transition to an antagonist/partial agonist was caused by a structural constraint on the peptide backbone introduced by the non-natural conformation in D-Phe7. The added rigidity at this position in subsequent peptide antagonists such as icatibant (Hoe 140, D-Arg-[Hyp3,Thi5, D-Tic7, Oic8]-BK; Fig. 2 and Table 2; Hock et al., 1991) and NPC17731(D-Arg[Hyp3, D-HypE(transpropyl)7, Oic8]-BK) (Trifilieff et al., 1993) showed that the spatial orientation of the C-terminal region of the peptide molecule is critical for antagonism. Icatibant exhibits a high affinity, no residual agonist activity in most mammalian species and an impressive resistance to peptidases; the combinations of these properties contribute to its prolonged duration of action (several hours) in animal models. In practical terms, this peptide exhibits a high selectivity toward the B2 receptor, although its affinity for the human B1 receptor is measurable (Table 2). Icatibant has been used extensively (about 700 publications indexed in Medline) to assess the role of kinins in various models of physiological adaptation or experimental pathology and occasionally in humans (Groves et al., 1995; Akbary et al., 1996; Dear et al., 1996; Gainer et al., 1998; Squire et al., 2000; Witherow et al., 2001; Hornig et al., 2003).
Optimal peptide B2 receptor antagonists retain both Arg1 and Arg9 residues, and B1 receptor agonists or antagonists typically lack Arg9. However, more recent antagonists with backbones constrained by non-natural residues are somewhat more promiscuous in their selectivity. The des-Arg fragment of icatibant (des-Arg10-Hoe 140) has been promoted as a B1 receptor antagonist by some investigators, although it is a mixed antagonist (Table 2). This peptide exhibits increased affinity toward the B1 receptor relative to icatibant in all species studied, but retains fairly high residual antagonistic effects on the B2 receptor in functional studies (Rhaleb et al., 1992; Lagneux and Ribuot, 1997). Recently produced antagonists retaining Arg9, such as B9430 (D-Arg-[Hyp3, Igl5, D-Igl7, Oic8]BK) represent combined B1 and B2 receptor antagonists, although the corresponding des-Arg9 fragment has substantial selectivity toward the B1 receptor (Burkard et al., 1996). B9430 has been deliberately exploited as a dual kinin receptor antagonist in the blood pressure assay of the dog where it can abolish, at certain doses, hypotensive responses to either BK or des-Arg9-BK without affecting the responses to several other unrelated agonists (Stewart et al., 1996). The development of this compound not only demonstrates that a "polypharmaceutic" approach covering both receptor types is possible, but also that the structures of the B1 and B2 receptor molecules are sufficiently compatible to accommodate a universal antagonist "pharmacophore".
An example of a peptidic B1 receptor antagonist of a modern design is B9858 (Lys-Lys-[Hyp3, Igl5, D-Igl7, Oic8]des-Arg9-BK); it exhibits a fairly high selectivity for the B1 receptor (Table 2; Gera et al., 1996; Gobeil et al., 1999). This peptide retains the Lys0 residue, a favorable feature for binding to the human B1 receptor (Table 2). In binding assays, B9858 is more potent at the human B1 receptor than at its murine ortholog; it is highly active as a competitive antagonist against Lys-des-Arg9-BK-induced contractility of the human umbilical vein (pA2 9.2) and is potent, but insurmountable, at the rabbit B1 receptor (Larrivée et al., 2000). In the anesthetized dog representing a hemodynamic system with mixed B1 and B2 receptor responses (see below), B9858 is reported to be 20 times more potent than Lys-[Leu8]des-Arg9-BK in antagonizing des-Arg9-BK-induced hypotension, and the antagonist effect lasts for more than 4 h (versus about 15 min for Lys-[Leu8]des-Arg9-BK; Stewart et al., 1996).
Numerous structural experiments have been conducted with peptide kinin antagonists, such as producing pseudopeptides that retain limited structural elements of the parent peptide (Chakravarty et al., 1995; Galoppini et al., 1999; Bedos et al., 2000), or the production of dimers of antagonists. An early implementation of the latter idea was deltibant (CP-0127), the homodimer of D-Arg0[Hyp3, Thi5, Cys6, D-Phe7, Leu8]BK coupled through the side chains of Cys6 by the linker bis-succinimidohexane. This B2 receptor antagonist of rather low potency had nevertheless a good metabolic stability and reached clinical trials for sepsis and head trauma but with mixed results (Fein et al., 1997; Marmarou et al., 1999). The production of such antagonists of molecular weight (
2.5 kDa) higher than that of the antagonist peptide and of high solubility in water goes against common drug design wisdom unless special effects are sought. A recent avatar of the dimer approach is B9870 (also called CU201, two molecules of D-Arg-[Hyp3, Igl5, D-Igl7, Oic8]BK linked at their N terminus by the suberimidyl moiety, Stewart et al., 1997). In tumor-derived cell lines, B9870 antagonized BK-induced calcium signaling but also directly blocked cell proliferation and activated c-Jun kinase and caspase-3 (Chan et al., 2002a). The monomeric component of B9870 was essentially devoid of the "biased agonist" properties of the dimer. The B9870 dimer has been called a biased agonist since it activates only a subset of the signal transduction pathways activated by the reference agonist. Whether all effects of B9870 are truly B2 receptor-mediated is currently unknown.
The peptide analogs of BK described above have been extremely important in developing an understanding of the roles of kinins and their receptors in physiology and pathophysiology. Recently, that understanding has led to the conjecture that B1 or B2 receptor modulators could represent novel therapeutic agents across a wide range of human conditions and diseases such as pain, inflammation, stroke, asthma, rhinitis, endotoxic shock, and cancer. Development of nonpeptide agonists and antagonists offers the greatest opportunity for new medicines since peptides are generally poor for oral bioavailability and brain penetration. Numerous laboratories are now engaged in research activities to develop and clinically test novel B1 and B2 receptor modulators (Altamura et al., 1999; Heitsch, 2002; Sharma and Al-Dhalmawi, 2003). A key issue that has faced development of such compounds for either B1 or B2 receptors has been species selectivity of their pharmacology. In some cases, the human specificity of certain structural classes has limited the preclinical evaluation of efficacy, and conversely, rodent specific compounds are not useful for therapeutic development (Hess et al., 1994, 2002; Jones et al., 1999; Burgess et al., 2000; Marceau et al., 2003). Many of these efforts have yet to result in published articles describing the compounds, but the patent literature bears ample evidence for these activities (Bock and Longmore, 2000; Heitsch, 2002; Bock et al., 2003). No doubt the therapeutic potential of these approaches will become evident in the near future.
1. B2 Receptor Agonists.
FR190997 is a very interesting example of a nonpeptidyl structure exhibiting agonist activity at the B2 receptor (Fig. 2; Aramori et al., 1997a). FR190997 was derived chemically from, and bears close structural similarity to, a series of nonpeptide B2 receptor antagonists (see FR173657, Fig. 2). Other examples of GPCR ligands exist where very slight structural modifications, such as addition or deletion of a simple methylene group, change antagonists to agonists at cholecystokinin (Aquino et al., 1996) and AT1 receptors (Perlman et al., 1995), not withstanding the well known close structural similarity of the opiate ligands morphine and naloxone. In the case of FR190997, potent B2 receptor agonist activity is demonstrated in vitro with a level of efficacy approaching BK itself, and in vivo FR190997 produces hypotension in rats (Aramori et al., 1997a). Although compounds like FR190997 may have therapeutic utility as antihypertensive or as cardioprotective agents, appropriate clinical testing remains. Concerns over tachyphylaxis (Mathis et al., 1996; Cuthbert, 1999) or potential adverse effects of a BK mimic such as proinflammatory responses demonstrated with FR190997 in animals (Hayashi et al., 2001) may limit considerations for therapeutic development.
2. B2 Receptor Antagonists.
The earliest example of the successful synthesis of a nonpeptide B2 receptor antagonist is WIN64338(Fig. 2) (Salvino et al., 1993). As a prototype, WIN64338had somewhat limited affinity for the human B2 receptor (Ki, 64 nM), and in vivo studies have not been published (Sawutz et al., 1994). WIN64338 however, was clearly an important advance for the field and an early pharmacological tool.
Subsequently, researchers at Fujisawa described new structurally distinct nonpeptide B2 receptor antagonists (Aramori et al., 1997b; Abe et al., 1998b), such as FR173657 (Fig. 2) with substantially higher B2 receptor affinity and demonstrated selectivity versus the B1 receptor. Moreover, these compounds including FR173657 and FR184280 (Abe et al., 1998a) were the first to demonstrate oral activity in vivo, in a guinea pig model of bronchoconstriction and various models of edema, inflammation, and pain (Asano et al., 1999). FR173657 in particular has proven to be a valuable tool to investigate the physiology and pathophysiology of BK and has been studied by a number of laboratories.
Another series of selective nonpeptide B2 receptor antagonists have emerged from the Fournier laboratories exemplified by LF 16-0687 (Fig. 2; Pruneau et al., 1999) as a high affinity antagonist across species and a potent antagonist against BK-induced hypotension and edema. LF 16-0687 also was shown to reduce cerebral edema and improve neurological outcome in a rodent model of head trauma (Kaplanski et al., 2002) implicating BK in this pathology.
A striking example of the species selectivity that can emerge during the development of B2 receptor antagonists is with bradyzide (Fig. 2; Burgess et al., 2000), a potent antagonist that displays >500 times higher affinity for the rat versus human B2 receptor. Bradyzide is highly selective for the B2 receptor over the B1 receptor and other receptors, is orally active, and shows antihyperalgesic effects in a rat model of inflammatory pain. Interestingly, working from the thiosemicarbazide scaffold of bradyzide, Dziadulewicz et al. (2002) succeeded in obtaining subnanomolar affinity for the human receptor while maintaining affinity for the rat receptor in Compound 14e (Fig. 2) and other analogs. Like bradyzide, Compound 14e exhibited oral activity to block inflammatory hyperalgesia in rats.
Researchers from Johnson and Johnson have recently published their early efforts in the development of B2 receptor antagonists (Lee et al., 2003) as exemplified by Compound 38 (Fig. 2; Youngman et al., 2003). FR173657 served as a starting template for this chemical series. Compound 38 is a high affinity antagonist (Ki, 4 nM) at the human B2 receptor. These compounds are analgesic after oral dosing in a rodent model of chemical-induced irritation.
It is clear that the development of nonpeptide B1 and B2 receptor modulators as potential new therapeutics, particularly antagonists, is currently an area of considerable effort. At present, no articles have been published that disclose clinical results with nonpeptide BK antagonists, but it is likely that compounds will emerge soon that are suitable for human clinical trials.
3. B1 Receptor Antagonists.
Recently, the results of research into the development of B1 receptor antagonists as novel therapeutic agents have begun to bear fruit. In 1999, researchers from Pharmacopeia published on a compound PS020990 as a potent and competitive B1 receptor antagonist (Horlick et al., 1999). PS020990 was derived from compounds discovered from the screening of a combinatorial chemistry library and is the first nonpeptide B1 receptor antagonist described in the literature. Although the structure was not initially disclosed, subsequent publications have identified PS020990 as shown in Fig. 2 (Anonymous, 2001). PS020990 exhibits high affinity for the human B1 receptor (Ki, 6 nM) and is an appropriately potent B1 receptor antagonist in vitro. To date, no results have been published on the in vivo properties of PS020990.
More recently, researchers from Merck Research Laboratories have disclosed two additional nonpeptide antagonists with divergent structures. Wood et al. (2003) described the in vitro and in vivo properties of a series of benzodiazepine-based compounds. The preferred structure from this series, Compound 12 (Fig. 2), exhibited high affinity for the human B1 receptor (Ki, 0.6 nM) and appropriate functional antagonist potency in vitro. Species differences are well known for both B1 and B2 receptors (Hess at al., 1994, 2002; Jones et al., 1999), so it was of interest that Compound 12 showed equally high affinity (Ki, 0.9 nM) and antagonist potency at the rat B1 receptor in vitro. This affinity allowed for its testing in a rat model of carrageenan-induced hyperalgesia where Compound 12 exhibited significant analgesic activity. A second series, the so-called dihydroquinoxalinone series, is represented by Compound 11 (Fig. 2) (Su et al., 2003). Compound 11 was derived from a micromolar affinity lead detected in a high throughput screening effort, and optimization efforts led to extraordinary affinity for the human B1 receptor (Ki, 34 pM) producing a highly potent antagonist. Unlike the benzodiazepine-based Compound 12 described above, Compound 11 shows considerably lower affinity for the rat B1 receptor (Ki, 62 nM), but high affinity for the rabbit B1 receptor (Ki, 50 pM) allowing for demonstration of robust antihyperalgesic activity in this species. Like the benzodiazepine Compound 12, the dihydroquinoxalinone antagonist exhibited high B1 receptor selectivity versus the B2 receptor and limited to no activity at a host of other enzyme, GPCR, and transporter targets. These representatives of both structural series, however, had suboptimal pharmacokinetic profiles that precluded further development. However, both Compounds 11 and 12 serve as structures for further optimization work, and their analgesic activity across rat and rabbit species further substantiate the role of B1 receptor mechanisms to mediate pain (see below).
Researchers from Sanofi-Synthelabo recently published on a novel nonpeptidyl B1 receptor antagonist, SSR240612 (Fig. 2) (Gougat et al., 2004). SSR240612 is a high affinity, competitive B1 receptor antagonist with considerable selectivity against the B2 receptor and a variety of other receptor and enzyme targets. Of interest, SSR240612 is a potent B1 receptor antagonist across several species and exhibits oral antihyperalgesic and anti-inflammatory activities in mice and rats and is currently in preclinical development for inflammatory and neurogenic pain. Other efforts to develop nonpeptide B1 receptor antagonists are evident at Novartis, Bayer, and Fournier based on their patent publications (Bock et al., 2003), but the status of any clinical progress is not known.
III. Structural Aspects of Kinin Receptors and Their Genes
A. Organization and Structure of the Receptor Genes
A BK receptor was first cloned from rat uterus, using a Xenopus oocyte expression assay and shown to have the pharmacological profile of a B2 receptor subtype (McEachern et al., 1991). Subsequently, orthologs have been isolated and sequenced in several mammalian species, including the human (Table 3; Hess et al., 1992). The B2 receptor protein structure is typical of that of a GPCR consisting of a single polypeptide chain that spans the membrane seven times, with the amino terminus (N-terminal domain) being extracellular and the carboxy terminus (C-terminal domain) being intracellular, and with three extracellular loops (EL-1–3) and three intracellular loops (IL-1–3) (Fig. 3). Three consensus sites for N-linked glycosylation are found in extracellular domains. Furthermore, the protein contains motifs such as DRY and NPXXY partially embedded in cytosolic receptor domains that are common to most rhodopsin family GPCR (Fredriksson et al., 2003), and the C-terminal tail contains serines and threonines that are putative phosphorylation sites and cysteines that are putative sites for acylation.
|
|
The three-exon structure of the gene (BDKRB2) has been determined, as well as its localization on the human chromosome 14 (Fig. 4; Powell et al., 1993; Kammerer et al., 1995). The coding sequence of the human B2 receptor was initially believed to be intronless in the gene exon 3 where a plausible initiation codon was found; the recombinant receptor based on this initiation codon is fully functional. However, another methionine codon in exon 2 is the likely initiation codon and extends the amino terminal sequence (first extracellular domain) by 27 residues (proved notably by sequencing the immunoprecipitated receptor from human HF15 cells; AbdAlla et al., 1996a). However, the exon 3 methionine has generally been used in the numbering of residues in the B2 receptor. In Fig. 3, upstream residues are labeled with negative numbers.
|
The expression cloning of the human B1 receptor was achieved later (Menke et al., 1994), and its orthologs have been sequenced in several mammalian species (Table 3). The B1 receptor is homologous to the B2 receptor (36% identity at the amino acid sequence level) and also possesses three consensus sites for N-linked glycosylation in extracellular domains, DRY and NPXXY motifs, and putative sites for phosphorylation and acylation (Fig. 3). The three-exon structure of the human B1 receptor gene (BDKRB1) has been determined with the protein sequence being encoded by exon 3 exclusively (Fig. 4; Bachvarov et al., 1996; Yang and Polgar, 1996).
The two genes corresponding to the kinin receptors are clustered in tandem in a compact locus, the B2 receptor gene being located 5' relative to the B1 receptor gene separated by only 12 kb (Cayla et al., 2002a; http://www.ncbi.nlm.nih.gov/mapview/, Fig. 4). Species-specific alterations of this scheme, such as the deletion of B1 receptor exon 2 in the rat and mouse genome and the possible existence of an alternatively spliced exon 2b in the B2 receptor, have been discussed (Cayla et al., 2002a). This locus is on chromosome 14 in the human (14q32). The mRNA coding for the B2 receptor is large compared with that of the B1 receptor (approximately 4 and 1.4 kb, respectively, depending on the species). The large 3'-untranslated region of the B2 receptor accounts for most of the difference (Fig. 4).
Boels and Schaller (2003) claimed that a human orphan GPCR, GPR100 (GenBank accession number AY170828), is an additional receptor for BK, only distantly related to the known kinin receptors. The cloned receptor was said to confer calcium transients to cells stimulated with BK only if the promiscuous G16 protein is cotransfected. The pharmacological profile was that of a B2 receptor because the stimulation was common to BK and Lys-BK, whereas the des-Arg9 fragments of these peptides were inactive. No antagonists were tested in this study. These findings were provocative since the B2 receptor gene knockout mouse has lost all responses to exogenous BK (Borkowski et al., 1995), failing to support the concept of additional B2 receptor subtypes. GPR100 (also called GPCR142 or relaxin-3 receptor-2) has also been shown to bind relaxin-3 and to inhibit adenylate cyclase upon stimulation with this peptide (Liu et al., 2003). In a subsequent study by Meini et al. (2004a), human GPR100 was compared with human B2 receptor in CHO cells in the absence of G16 cotransfection. It was shown that GPR100 is a high affinity relaxin-3 receptor, which is coupled to Gi-mediated inhibition of adenylate cyclase but not to Gq-mediated stimulation of phospholipase C and intracellular calcium mobilization. Interestingly, BK stimulates this receptor with relatively lower potency, whereas the nonpeptide B2 receptor agonist FR190997 was completely devoid of effect. The BK response is antagonized by icatibant supporting an evolutionary relationship between GPR100 and B2 receptor binding sites.
Two receptors cloned from the chicken and the zebrafish are more related to the mammalian B2 receptor than to the B1 receptor and are stimulated by sequences related to BK (ornithokinin = RPPGFSPLR, trout BK = RRPPGWSPLR), but not by BK itself (Schroeder et al., 1997; Dunér et al., 2002). The peptide icatibant, which acts as an antagonist of BK at mammalian B2 receptors, acts as a full agonist at the ornithokinin receptor (Table 2). The fish receptor gene is located in a chromosomal region syntenic to the human kinin receptor gene region with implications for the evolution of the kallikreinkinin system.
The targeted disruption ("knockout") of each kinin receptor gene by homologous recombination has been reported. The B2 receptor knockout mice fail to respond to BK using assays such as smooth muscle contraction and afferent nerve stimulation (Borkowski et al., 1995). These animals are fertile and apparently healthy, although they may not develop and age in an entirely normal manner (see below). The B1 receptor knockout mice have been produced more recently (Pesquero et al., 2000). These animals develop normally, are normotensive, but fail to respond to des-Arg9-BK (e.g., contractility of the mouse-isolated stomach). The creation of a double B1/B2 receptor knockout strain is not a trivial crossing experiment since both genes lie in the same chromosomal locus (see above), and the probability of recombination is very low at such a small genetic distance. Numerous physiological and pathological analyses based on either knockout strain will be reported below.
B. Receptors and Their Post-Translational Modifications
Kinin receptors undergo multiple post-translational modifications including glycosylation and disulfide bridge (cystine) formation in their extracellular domains as well as acylation and phosphorylation of their intracellular domains. Most of these modifications, including phosphorylation, can occur in a ligand-independent manner, although phosphorylation and possibly even acylation (palmitoylation) appear to be primarily regulated through receptor activation.
1. Glycosylation. All known mammalian and nonmammalian kinin receptors expose several (two to four) consensus sites of the Asn-Xaa-Ser/Thr type for N-glycosylation on their predicted extracellular domains, and most of them are clustered on their N-terminal domains (Fig. 3). For mammalian B2 receptors, conclusive evidence has been presented that they are sialoglycoproteins, and it appears likely that nonmammalian B2 receptors as well as mammalian B1 receptors fall into the same category. In general, glycosylation is thought to increase the hydrophilic nature of the extracellular portions of the receptors, however, glycosylation may also affect ligand affinity, receptor oligomerization, efficient G protein coupling, receptor degradation, maturation (folding, stabilization), and/or intracellular trafficking.
Human and rabbit B1 receptors expose three consensus sites for N-glycosylation in their extracellular domains (Fig. 3; Menke et al., 1994; MacNeil et al., 1995). The presence of multiple bands of the recombinantly expressed human B1 receptor that is strongly reminiscent of the patterns of the partially glycosylated B2 receptor has pointed to the possibility that the B1 receptor is also prone to N-glycosylation (Blaukat et al., 1999a). Treatment of the HA-tagged human B1 receptor expressed in HEK293 cells with N-glycosidase F resulted in the conversion of the receptor from a heterogeneous species migrating at 35 to 45 kDa to a relatively homogeneous species migrating at 37 kDa. Furthermore, a receptor in which the N-terminal domain had been truncated to remove putative N-glycosylation sites also migrated as a homogeneous species (Kang et al., 2005). In addition, studies using the N-glycosylation inhibitor tunicamycin have revealed a profound influence on des-Arg9-BK-induced responses in the isolated human umbilical vein (Sardi et al., 1999). Likewise tunicamycin caused a significant, although incomplete, reduction of des-Arg9-BK responses in the isolated rabbit aorta (Audet et al., 1994). The precise molecular mechanisms underlying the observed effects remain to be elucidated.
The protein sequences of the rat (McEachern et al., 1991) and human B2 receptors (Hess et al., 1992) predict three potential N-glycosylation sites for each of these orthologs. Two of the sites are present in the N-terminal domain of the receptors, whereas a single site is present in the second extracellular loop, EL-2 (3). Nonmammalian B2 receptors differ from most mammalian B2 receptors in that they lack a predicted N-glycosylation site in EL-2 (zebrafish B2 receptor) (Dunér et al., 2002) or have two additional sites in their N-terminal domain (ornithokinin receptor) (Schroeder et al., 1997).
A careful study of the rat B2 receptor by Yaqoob et al. (1995) provided conclusive evidence that all three predicted sites are indeed glycosylated, most likely through complex-type oligosaccharide side chains containing terminal sialic acid residues. The same authors were unable to find evidence for O-glycosylation of the rat B2 receptor. Removal of the N-linked carbohydrates from the native receptor (81 kDa) by endo--N-acetylglucosaminidase yielded a core protein of 42 to 44 kDa, indicating that the difference in the apparent molecular sizes is not due to dimerization of the unmodified receptors but truly reflects receptor glycosylation (Yaqoob et al., 1995). This notion is further corroborated by the findings from overexpression systems where multiple bands for the human B2 receptor were observed that were sensitive to N-glycosidase F treatment (Blaukat et al., 1999a). The complex carbohydrate chains with their terminal sialic acid residues may also contribute to the apparent isoelectric point of 4.5 to 4.7 for the mature rat B2 receptor, which is significantly lower than the predicted pI of 7.0 calculated from its primary structure. Importantly, glycosylation does not appear to affect ligand-receptor interactions, as deglycosylation failed to change the affinity of the rat B2 receptor for the agonist BK as well as for the antagonist icatibant (Yaqoob et al., 1995).
2. Disulfide Bridging.
The predicted protein structures of mammalian kinin receptors identify four cysteine residues in the extracellular domains of both B1 and B2 receptors, one each being located in the N-terminal domain, EL-1, EL-2, and EL-3, respectively (Fig. 3). Of the nonmammalian kinin receptors, the chick receptor shows the same distribution pattern of cysteine residues, whereas the zebrafish B2 receptor has a single cysteine each in EL-1 and EL-2 and lacks cysteines in the N-terminal domain and EL-3, respectively. Notably, cysteine in EL-2 and EL-3 are conserved in most members of the GPCR superfamily, and the crystal structure of rhodopsin has revealed that they form a disulfide loop which most likely stabilizes the correctly folded conformation of the receptor (Palczewski et al., 2000).
To date, no experimental data are available for the kinin receptors that would prove the presence of disulfide bonds in their extracellular domains. However, a careful study by Leeb-Lundberg and coworkers addressing the availability of free thiol residues in wild-type and mutant B2 receptors has provided indirect evidence for the presence of a disulfide bond between Cys103 (EL-1) and Cys184 (EL-2) (Herzig and Leeb-Lundberg, 1995; Herzig et al., 1996). They showed that Cys20 (N-terminal domain) and Cys277 (EL-3) of the wild-type B2 receptor were accessible for heterobifunctional cross-linkers that bind to free thiol groups and to the amino group of BK, whereas Cys103 (EL-1 domain) and Cys184 (EL-2) were not. The authors concluded that residues Cys103 and Cys184 likely form a disulfide bridge, whereas the former two residues, i.e., Cys20 and Cys277, do not (Herzig et al., 1996). Of note, application of dithiothreitol had little or no effect on the binding of BK to the B2 receptor in both bovine myometrial membranes and intact rat myometrial cells suggesting that, even though disulfide bond formation between extracellular cysteines in the B2 receptor may be crucial during insertion of the receptor into the membrane to form a proper binding site, reducing some of these bonds in receptors already expressed in the membrane does not destroy their agonist binding site (Herzig et al., 1996).
3. Acylation.
As with many other GPCRs, the kinin receptors may expose cysteine residue(s) in their C-terminal portion facing the cytosol (Fig. 3). These cysteines are often modified by palmitoylation, and due to the insertion of the fatty acid chain into the membrane, an extra loop is created which may modulate G protein coupling, internalization, resensitization, and/or intracellular trafficking of the receptors. All the known mammalian B2 receptors contain two cysteine residues in their tail region, separated by four amino acid residues (CxxxGC). The nonmammalian B2-like receptors from chick and zebrafish have a single cysteine residue in their cytosolic tail region. Mammalian B1 receptors fall into two main groups regarding cysteines for putative palmitoylation (Marceau et al., 1998); the rodent B1 receptor with an unusually short C-terminal tail of 11 residues is devoid of cysteines and the nonrodent receptors contain one or two cysteines in their C-terminal tail.
Mass spectrometric analysis of the rat B2 receptor has revealed that the proximal Cys326 residue present in the hexapeptide CRKGGC is indeed palmitoylated (Soskic et al., 1999). Site-directed mutagenesis studies of the human B2 receptor demonstrated that both the proximal (Cys324) and the distal (Cys329) cysteines present in the linear sequence of CQKGGC are subject to palmitoylation based on serine mutants, indicating that the two cysteines may be used alternatively by the corresponding acyl transferase (Pizard et al., 2001). At present, the acylation status of the B1 receptor has not yet been addressed experimentally; however, given the fact that the rodent B1 receptor lacks a cysteine residue in their tail region, it is conceivable that palmitolyation is dispensable for proper function of B1 receptors in general.
4. Phosphorylation.
On repeated or prolonged agonist stimulation, GPCR signaling is attenuated by negative feedback loops. Agonist-induced reversible phosphorylation, typically of serine and threonine residues located in the C-terminal domains, plays an important role in the desensitization of GPCRs (Fig. 3). Phosphorylation may also affect endocytosis, recycling, trafficking, or even G protein coupling of the receptors. Both mammalian and nonmammalian B2 receptors expose clusters of Ser/Thr residues in the tail region, and their ligand-dependent phosphorylation has been extensively studied for human and rat B2 receptors (Blaukat, 2002). Mammalian B1 receptors fall into two categories regarding putative serines and threonines for phosphorylation. The rodent receptors with a truncated tail region contain only two threonine residues but not a single serine residue, and the nonrodent B1 receptors contain a cluster of at least six Ser/Thr residues in their C terminus (Marceau et al., 1998; Hess et al., 2002).
The correlation between agonist-induced receptor phosphorylation and desensitization has been firmly established both for endogenous and recombinant B2 receptors (Blaukat et al., 1996, 2001). The precise mapping of the relevant phosphorylation sites revealed three major sites, i.e., Ser339, Ser346, and Ser348 and two minor sites, i.e., Thr342 and Thr345, forming a cluster in the C-terminal region of the human B2 receptor which spans over 10 residues (S339MGTLRTSIS348) (Figs. 3 and 5; see Section IV. for the functional consequences of mutating these residues in the B2 receptor and B1/B2 receptor chimeras). These studies revealed that tandem phosphorylation of two serine residues, i.e., Ser346/Ser348 or Ser339/Ser346 in the human B2 receptor is necessary and sufficient for desensitization of BK-triggered phospholipase C activation (Blaukat et al., 2001). Mass spectrometry of the unstimulated rat B2 receptor had also demonstrated tandem phosphorylation, although with varying combinations, e.g., Ser346/Ser348, Thr342/Ser348, and Ser331/Ser339 (numbering according to the human B2 receptor) (Soskic et al., 1999).
|
Tyrosine phosphorylation has also been reported for human and rat B2 receptors. Early studies claimed the tyrosine phosphorylation of the human B2 receptor (Jong et al., 1993), and mass spectrometric analysis of unstimulated rat B2 receptors demonstrated phosphorylation of Tyr131 (human residue Tyr129) present in the second intracellular loop (IL-2) and of Tyr322 (human residue Tyr320) in proximity to the identified palmitoylation site in the C-terminal domain (Soskic et al., 1999). However, ligand-dependent tyrosine phosphorylation of the B2 receptor has not yet been shown. Therefore, it is possible that the receptor becomes phosphorylated on tyrosine residues through homologous B2 receptor-mediated or through nonreceptor-regulated heterologous mechanisms (Blaukat et al., 2000). Accordingly, the effects seen for mutations of Tyr131 and Tyr322 of the rat B2 receptor on the signaling properties of the receptor apparently do not correlate with phosphorylation of these residues (Prado et al., 1997).
Unlike the phosphorylation of B2 receptors, ligand-induced phosphorylation of B1 receptors was undetectable in the absence or presence of an agonist, even by the most sensitive radiolabeling techniques (Blaukat et al., 1999a). In line with these findings, the signaling patterns of the two types of kinin receptors differ with respect to their duration; B2 receptor-mediated signaling is transient, whereas B1 receptor-induced signaling is sustained (Mathis et al., 1996; Faussner et al., 1998) (see Section IV.D.).
C. Agonist and Antagonist Binding Sites in the Receptors
1. B2 Receptor Agonists.
Several methods have been used to map the agonist and antagonist ligand binding sites in B2 and B1 receptors including site-directed and chimeric B2 and B1 receptor mutagenesis, ligand receptor cross-linking, and probing with antipeptide antibodies against extracellular receptor domains and antiidiotypic antibodies. These studies have led to the identification of several residues that appear to face the ligand binding sites of the B2 and B1 receptors (Table 4). The binding energy for BK and related peptide agonists in the B2 receptor seems to be provided by nonionic interactions contributed by residues located throughout the peptide as well as by ionic interactions provided by the side chain of Arg1 or the N terminus (Regoli and Barabé, 1980; Tancredi et al., 1997). An early model of BK bound to the rat B2 receptor based on structural homology modeling with bacteriorhodopsin, molecular modeling, and systematic conformational searching methods led to the identification of two candidate residues, Asp268 and Asp286 (human residues Asp266 and Asp284) in EL-3 near the extracellular ends of TM-6 and TM-7, respectively, with which BK may interact (Fig. 3) (Kyle et al., 1994). It was proposed that these aspartates interact electrostatically with the N-terminal amino group, the guanidinyl side chain, or both on Arg1 in BK, a residue absolutely critical for function (Regoli and Barabé, 1980). Alanine mutation of these aspartate residues in the rat receptor confirmed their critical role in BK binding (Novotny et al., 1994, Jarnagin et al., 1996). Conclusive evidence for this orientation of the receptor-bound BK was gained from chemical cross-linking of the N termini of bound [3H]BK and [125I-Tyr8]BK to Cys277 in EL-3 (Herzig and Leeb-Lundberg, 1995; Herzig et al., 1996) and Lys172 near TM-4 (AbdAlla et al., 1996b), respectively, in the human B2 receptor (Fig. 3). Indeed, the close proximity of the BK N terminus (3 Å) to both Cys277 in EL-3 and Cys20 in the receptor N terminus provided direct support for the helical bundle GPCR structure (Herzig et al., 1996). Inhibition of BK binding by antipeptide antibodies specific for the N-terminal halves of EL-3 and EL-2 provided further evidence for this orientation (AbdAlla et al., 1996c). The sensitivity of the human B1 receptor to the presence and absence of L-Lys at the N-terminal end of the peptide ligand was also used to orient the bound peptide agonists in the human B2 receptor (Fathy et al., 2000). To do so, a chimeric construct was made in which B1 receptor EL-3 was substituted in the B2 receptor. This substitution resulted in a drop in the affinity of BK to that for the B1 receptor. Addition of an L-Lys at the N terminus of the peptide to make Lys-BK, which is critical for high affinity binding to B1 receptors, fully restored the affinity of the peptide.
|
Early mutagenesis studies in the rat B2 receptor also identified Thr265 and Phe261 (human residues Thr263 and Phe259), located one and two helical turns proximal to Asp266 in TM-6, as critical BK binding residues (Fig. 3; Nardone and Hogan, 1994; Jarnagin et al., 1996). Thr265 
Ala and Phe261 Val mutations yielded approximately 2000- and 1000-fold drops, respectively, in BK binding affinity to this receptor. Critical roles for
