Foreword

Top Next References

A literature review on kinin pharmacology that appeared in this journal in 1980 was based on a systematic effort to define two receptor types for bradykinin- (BK ^b) related peptides, and proposed a receptor nomenclature (B₁, B₂) (Regoli and Barabé, 1980). Since then, that historical paper has been frequently cited and the nomenclature, widely adopted. The physiologically prominent B₂ receptor (B₂R) subtype has certainly been the subject of more intensive efforts in drug development, structure-function studies, and physiological investigations. However, the B₁ receptor (B₁R), activated by a class of kinin metabolites (des-Arg⁹-BK and Lys-des-Arg⁹-BK), is now a defined molecular entity (fig. 1) and its characterization, including its rapid up-regulation in tissue injury, has emerged as an important area of investigation within the study of the kallikrein-kinin system. Indeed, a major interest in the B₁R is its inducible character, an unusual feature for a G-protein coupled receptor.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Primary structure and putative domains of the human B1R. The amino acid sequence (one letter code) starts at the upper left, an extracellular domain. Asterisks represent putative sites of N-glycosylation. A putative disulfide bond between Cys residues of the second and third extracellular domain is also represented. Residues in bold are identical in all four species for which the B1R has been sequenced (human, rabbit, mouse, rat). The underlined C-terminal sequence, a part of a sequence extension found only in the human molecule, has been the immunogen for the production of anti-B1R antibodies. Numbered residues are referred to in the description of mutant or chimeric constructions (see text). Data from Hess et al. (1996), GenBank file no. U66107, MacNeil et al. (1995), and Menke et al. (1994). Sequence variants for human B1R are reported: R146 (Bastian et al., 1997), R246, and S259 (GenBank file no. U22346).

We will present the findings on the kinin B₁R with emphasis on contemporary molecular research and, when possible, on the human molecules, cells, tissues, and subjects. We do not aim at an exhaustive or historical coverage of early work on B₁Rs, which was essentially based on functional approaches and structure-activity investigations of peptide analogs; the reader is referred to earlier reviews for coverage of these aspects (Marceau, 1995; Marceau and Regoli, 1991; Regoli and Barabé, 1980). Finally, the field of B₁R research cannot be isolated from pharmacological and molecular studies of the BK B₂Rs or from other components of the kallikrein-kinin system. A comparative and parallel coverage will be used in several sections of the present text.

	I. Introduction: The B1 Receptors in the Kallikrein-Kinin System

Top Previous Next References

Knowledge concerning the molecular elements of the kallikrein-kinin system is being constantly refined (table 1; reviewed by Kaplan et al., 1997; Margolius, 1995). A single human kininogen gene codes for the hepatic production of both high (HMW) and low molecular weight (LMW) kininogen via alternative splicing. Kininogens are multi-domain proteins that include the BK sequence, but also domains capable of binding Ca²⁺, binding to the cell surface or inhibiting the cysteine proteinases. Kallikreins are the enzymes that cleave kininogens and liberate kinins (defined as BK-related peptides). Plasma prekallikrein is cofactor of coagulation that is involved in a reciprocal activation reaction with the Hageman factor in the contact system. Plasma kallikrein releases BK (the nonapeptide H-Arg-Pro-Pro-Gly-Phe-Ser-Pro-Phe-Arg-OH) from HMW-kininogen (Margolius et al., 1995). There is also evidence that HMW-kininogen adsorbed on extracellular (M) proteins of Streptococcus pyogenes can release BK in the presence of prekallikrein, thus extending the physiopathological applications of the contact system to sepsis (Ben Nasr et al., 1997).

Tissue kallikreins are represented in humans by three related enzymes (hK1 to hK3). Tissue kallikrein (hK1) is widely expressed in glandular and duct cells, as well as in neutrophils (Wu et al., 1993), cultured vascular smooth muscle cells (SMCs), colonic goblet cells, and renal distal tubules and collecting ducts (Chen et al., 1995a, 1995b; Nolly et al., 1993; Oza et al., 1990). Tissue kallikreins are synthesized as zymogens and little is known of their production, storage, or release. Pure tissue kallikreins are not equally capable of generating biologically active kinins: hK1, the classical tissue kallikrein, is highly effective, whereas hK3 (also called prostate-specific antigen) is inactive, hK2 being of intermediate potency (Deperthes et al., 1997). The exclusive prostatic localization of hK2 and hK3 is currently being challenged, as they are found in other cell types (e.g., a human breast cancer line expresses hK2; Hsieh et al., 1997). Although hK1 is preformed in the human kidney, further synthesis of this protease occurs in this organ after trauma or immunopathological insult (Cumming et al., 1994). Human tissue kallikrein preferentially releases the decapeptide Lys-BK (kallidin) from kininogens, as found in the nasal secretions of subjects with allergic or viral rhinitis (Nacleiro et al., 1988; Proud et al., 1983).

The metabolism of kinins will be briefly summarized here, because B₁Rs are selectively and exclusively stimulated by one class of kinin metabolites (fig. 2). BK exhibits a remarkably short half-life in the blood plasma in vitro (10 to 50 sec, depending on species; Décarie et al., 1996b) or in vivo. The great susceptibility of these peptides to hydrolysis accounts for this rapid disposal. BK and Lys-BK are predominantly metabolized in the circulation by the angiotensin I converting enzyme (ACE; EC 4.4.15.1; kininase II; Erdös, 1990). ACE removes the C-terminal dipeptide from BK or Lys-BK, which leads to their complete inactivation (K_m of about 1 µM for this reaction). ACE is predominantly a surface enzyme located on the luminal membrane of endothelial cells, an observation which explains the extensive pulmonary inactivation of kinins. The proximal nephron epithelium is also rich in ACE (Erdös, 1990). ACE eventually cleaves further its primary metabolite, BK_1-7, into shorter fragments (e.g., BK_1-5).

View larger version (10K): [in this window] [in a new window]   Fig. 2.   Proposed cleavage sites for Lys-BK and Lys-des-Arg9-BK by the kininases. Abbreviations: AmM, aminopeptidase M; AmP, aminopeptidase P; NEP, neutral endopeptidase 24.11; ACE, angiotensin I converting enzyme; CPN, carboxypeptidase N; CPM, carboxypeptidase M.

An aminopeptidase P (aminoacyl proline aminopeptidase; EC 3.4.11.9) which cleaves the Arg¹-Pro² bond, may also contribute to the pulmonary inactivation of BK (Ryan et al., 1994; Ward, 1991). Another form of membrane-bound peptidase, the "neutral" endopeptidase (E.C. 3.4.24.11), cleaves two bonds in BK (Gly⁴-Phe⁵ and Pro⁷-Phe⁸; Zolfaghari et al., 1989). Endopeptidase 24.11 is notably located at a high concentration in the proximal nephron epithelium and is also present on the membrane of some leukocytes, but barely detectable in the vasculature or plasma (Erdös, 1990; Ward, 1991).

Kininase I activity is of particular importance, because it is believed to generate the physiological agonists of B₁Rs, the des-Arg⁹-kinins, from native kinins. This activity is composed of the arginine carboxypeptidases, carboxypeptidase N from plasma (EC 3.4.17.3), and carboxypeptidase M, predominantly membrane-bound and widely distributed, including in the microvasculature (table 1; Erdös, 1990; Ward, 1991). These are enzymes of rather low affinity (K_m 20 to 50 µM for BK and carboxypeptidase N) and broad specificity, digesting several other peptides generated by trypsin-like enzymes, such as the anaphylatoxins derived from complement. The relative role of arginine carboxypeptidases (kininase I) in the metabolism of kinins is discussed below.

The metabolism of des-Arg⁹-kinins differs from that of native kinins in several important points. First, des-Arg⁹-BK does not react with arginine carboxypeptidases, as it is devoid of the corresponding C-terminal residue (Drapeau et al., 1991a). Second, ACE cleaves a C-terminal tripeptide from B₁R agonists, yielding BK_1-5, but at a slower rate, and with much less affinity (K_m 130 to 240 µM) than the removal of a C-terminal dipeptide from BK (Drapeau et al., 1991a). The C-terminal position of Phe⁸ in des-Arg⁹-BK may protect the Pro⁷-Phe⁸ bond from the endopeptidase 24.11; this requires confirmation.

Aminopeptidase M (EC 3.4.11.2), present in plasma, hydrolyzes Lys-BK into BK, and Lys-des-Arg⁹-BK into des-Arg⁹-BK (Proud et al., 1987; Sheik and Kaplan, 1989). This reaction is pharmacologically neutral for the B₂R agonists BK and Lys-BK, because these peptides exhibit similar potencies. However, it represents a relative inactivation for the B₁R agonists, as des-Arg⁹-BK is an agonist of lesser affinity for the B₁Rs than Lys-des-Arg⁹-BK in some species (see below).

The conversion effectiveness of native kinin sequences (BK, Lys-BK) into their respective des-Arg⁹ metabolites is a question of considerable interest for the physiological role of B₁Rs, because the other competing metabolic pathways lead to fragments inactive on both B₁ and B₂Rs (Regoli and Barabé, 1980; see also below). Novel analytical techniques for BK- and des-Arg⁹-BK-like immunoreactive peptides have shown that ACE accounts for more than 75% of the hydrolysis of exogenous synthetic BK (500 nM) in the plasma of humans, rabbits, or rats (Décarie et al., 1996b), and more than 40% in cardiac membrane preparations derived from the same species (Blais et al., 1997b). The formation of des-Arg⁹-BK is minor in these species (0.9 to 3.4% of BK converted), much less than previous estimates based on high concentrations of exogenous BK as a substrate (Marceau et al., 1981; Proud et al., 1987). Blocking the competing ACE pathway with enalaprilat improved the conversion in all species (Blais et al., 1997b; Décarie et al., 1996b). Thus, when considering the relative contribution of several kininases, the mere existence of measurable des-Arg⁹-kinins in vivo could be questioned. However, in the plasma or cardiac membrane preparations of the four cited species, the half-life of exogenous des-Arg⁹-BK was 4- to 12-fold longer than that of BK under the same experimental conditions (Blais et al., 1997b; Décarie et al., 1996b). If this is an accurate representation of the situation in vivo, it may explain why the in vivo concentration of immunoreactive des-Arg⁹-BK is consistently higher than that of immunoreactive BK. Simply put, des-Arg⁹-BK, although not effectively produced, has a much greater capacity to accumulate than BK.

Kinins are very difficult to measure accurately, in part because the sampled blood contains all the necessary components to generate and destroy these peptides in vitro. For instance, the contact of glass in test tubes activates the plasma kallikrein. Current methods for measuring kinin concentrations rely on antibody-based techniques (radioimmunoassays, enzyme immunoassays) preceded by complex extraction procedures (Décarie et al., 1994; Odya et al., 1983; Raymond et al., 1995). In venous blood sampled from normal humans, the average concentration of immunoreactive des-Arg⁹-BK (204 pg/ml) is higher than that of BK (67 pg/ml) (Odya et al., 1983). These values are not significantly changed inpatients with essential hypertension or under ACE blockade with enalapril (Odya et al., 1983). The urine of these individuals also contains both BK and des-Arg⁹-BK immunoreactivities that do not differ between groups. The authors point out that venous blood or urine measurements may not reflect the status of the kinins at the tissue level, because of metabolism by multiple pathways. In the arterial blood of anesthetized rabbits, captopril treatment increases the BK immunoreactive concentration, but does not influence the concentration of des-Arg⁹-BK. The latter could be selectively and importantly increased by treatment with bacterial lipopolysaccharide (LPS), and even more so in animals pretreated with captopril (Raymond et al., 1995). Species-dependent differences in the metabolism and the rate of activation of the kallikrein-kinin system may be important parameters which determine the relative concentrations of BK and des-Arg⁹-BK. Local concentrations of kinins may generally be of greater interest, as the kinins are autacoids rather than hormones. Interestingly, both BK and des-Arg⁹-BK immunoreactivities increase in inflammatory lesions caused by carrageenan in the rat (Burch and DeHaas, 1990; Décarie et al., 1996a).

The discussion above is centered on the C-terminal structure of kinins, so important for receptor subtype selectivity (see below). However, various limitations are apparent in the analytical approach for des-Arg⁹-kinins. The antibodies efficient to discriminate the C-terminal structure do not differentiate between kinin homologues that differ in their N-terminus (e.g., BK, Lys-BK, Ile-Ser-BK; Décarie et al., 1994; Raymond et al., 1995). This represents a major limitation of our analytical capabilities, as Lys-des-Arg⁹-BK is the most potent B₁R agonist in humans, des-Arg⁹-BK being relatively ineffective (see below, Section II.A.). Combining the antibody-based method with chromatographic separation of homolog peptides may be helpful to discriminate among them (Décarie et al., 1994), although with a probable loss of sensitivity. An unambiguous analytical method for measuring Lys-des-Arg⁹-BK still has to be invented. An additional problem is that the Pro (3) residue in either BK or des-Arg⁹-BK may rather be hydroxyproline in a sizable proportion of the kinin molecules, caused by a partial posttranslational modification of high molecular weight kininogen in the human species (Schlüter et al., 1997). Although this substitution does not adversely affect receptor affinity (Regoli et al., 1996), its effect on metabolism or antigenicity is unknown. Finally, an inflammation-induced increase of immunoreactive des-Arg⁹-BK, as in the case of acute edema produced by carrageenan in the rat paw, does not necessarily predict an important role for B₁Rs, as these receptors are apparently absent in this acute model that dissociates the agonist from the response (Décarie et al., 1996a).

	II. Seven Criteria to Classify Kinin Receptors into the B1 and B2 Subtypes

Top Previous Next References

Kinins, positively charged peptides, influence tissues and cells by stimulating membrane receptors. Two types of receptors are currently completely defined in several species using a series of seven criteria, detailed in this chapter, although very significant species-specific pharmacological differences are documented. The first three criteria are pharmacological as defined by Schild (1973). The other ones are of more biochemical and molecular nature. Binding (table 2) and biochemical experiments based on cloned receptor subtypes have been valued in our coverage, because they offer the ultimate proofs regarding the behavior of definite molecular entities.

In vitro smooth muscle preparations assume a particular importance in the history of kinin receptor classification. The first hint of the receptor heterogeneity was the atypical order of agonist potency found for BK and its fragment on the isolated rabbit aorta: the fragment without the C-terminal arginine residue, des-Arg⁹-BK, was more potent than the intact nonapeptide (Regoli et al., 1977). A few other preparations are worth mentioning at this point, as they are useful to compare potency orders obtained in binding assays with cloned receptors in each species: comprehensive data are available for the contractility mediated by B₁Rs in the human umbilical vein (Gobeil et al., 1996b), the longitudinal muscle of the rat ileum (Meini et al., 1996), and the mouse stomach (Allogho et al., 1995). These preparations were further used to determine the potency and intrinsic activity of antagonists and the postisolation regulation of B₁Rs (see below). Of the preparations mentioned above, the rabbit aorta and the rat ileum respond to kinins only via B₁R, whereas the human umbilical vein and mouse stomach possess both receptor subtypes, a situation that complicates pharmacological analyses.

The potency order of agonists, as established in separate binding assays (table 2), confirms that the removal of the C-terminal arginine is essential for high affinity on the B₁R and detrimental for affinity toward the B₂R. This holds true for Lys-BK (kallidin) and Lys-des-Arg⁹-BK (des-Arg¹⁰-kallidin), which have higher affinities for the human and rabbit B₁Rs compared with BK and des-Arg⁹-BK, respectively (table 2). Met-Lys-BK, resulting from a different cleavage of the kininogen sequence, does not significantly differ from Lys-BK, being a relatively selective B₂R agonist in the human and rabbit (table 2). The only natural kinin sequence with a subnanomolar affinity for human and rabbit B₁Rs is Lys-des-Arg⁹-BK, a fact of particular significance that may suggest that the B₁R belongs to the "tissue" kallikrein-kinin system, jointly with tissue kallikrein that generates Lys-BK (kallidin) from low molecular weight kininogen. This does not invalidate the experimental use of des-Arg⁹-BK in certain animal species as a tool of high specificity for B₁Rs; however, this peptide has a low absolute affinity for the human B₁R. This is reflected by the low relative potency of functional responses induced by des-Arg⁹-BK as compared with those elicited by Lys-des-Arg⁹-BK, in the human umbilical vein assay (Gobeil et al., 1997), or in cells expressing recombinant human B₁Rs (MacNeil et al., 1997).

Limited functional data based on porcine renal vein contractility show that this B₁R, not yet cloned, is also preferentially stimulated by Lys-des-Arg⁹-BK (pD₂ of 8.4, versus 7.2 for des-Arg⁹-BK; Rizzi et al., 1997). By contrast, Lys-des-Arg⁹-BK has no decisive advantage over des-Arg⁹-BK in small rodents such as mouse (table 2) or rat. The sequenced kininogen genes in small rodents differ from the human ones, as the amino acids preceding the BK sequence are arginine, or Ile-Ser in T-kininogen, not lysine (Hess et al., 1996; Takano et al., 1997). In view of this recent discovery the precise sequence released by tissue kallikreins in rats and mice becomes an analytical issue of interest. Only BK has been isolated from rat urine and other experimental systems based on rat tissue kallikrein, but not Arg-BK, Ile-Ser-BK or Lys-BK (Hagiwara et al., 1995). A form of coevolution of the B₁R with the kininogen genes may have occurred in rodents to maintain the B₁R functional by conferring a subnanomolar affinity to des-Arg⁹-BK, which is reflected by the high potency of this peptide (superior or equal to Lys-des-Arg⁹-BK in functional assays based on rat or mouse smooth muscle contractility; Allogho et al., 1995; Meini et al., 1996; or cultured rat SMCs, where des-Arg⁹-BK exerts its effect with an EC₅₀ of about 300 pM; Dixon and Dennis, 1997).

Thus, the B₁R 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 or des(Phe⁸, Arg⁹)-BK (the primary metabolite generated by 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 B₁R, as these sequences are transformed into their respective des-Arg metabolites (discussed by Marceau and Regoli, 1991). Even a binding assay based on the cloned mouse B₁R seems to be prone to this error, as the apparent potency of BK in a competition with the ligand [³H]Lys-des-Arg⁹-BK is decreased by a factor of 5 in the presence of the arginine carboxypeptidase inhibitor MERGETPA (Pesquero et al., 1996). This was not the case for des-Arg⁹-BK. Thus, intact BK and Lys-BK may exhibit a very high selectivity toward B₂R.

Most published binding studies to human or animal B₁Rs have used the agonist ligand [³H]Lys-des-Arg⁹-BK (table 2). [¹²⁵I]Tyr-Gly-Lys-Aca-Lys-des-Arg⁹-BK is an alternative high affinity B₁R ligand which can be also used (Levesque et al., 1995a). This agonist incorporates a N-terminal extension which essentially acts as a spacer between iodotyrosine and the Lys-des-Arg⁹-BK sequence.

A synthetic agonist modeled on Lys-des-Arg⁹-BK incorporates modifications to improve resistance to metabolism: Sar-[D-Phe⁸]des-Arg⁹-BK is a high affinity B₁ receptor agonist of high selectivity that is completely resistant to blood aminopeptidase, kininases I and II (angiotensin I converting enzyme) and kidney neutral endopeptidase (Drapeau et al., 1991b, 1993). In a binding assay to rabbit B₁ receptors, this analog was 7-fold more potent than des-Arg⁹-BK, but 11-fold less potent than Lys-des-Arg⁹-BK (Levesque et al., 1995a). It is very potent in inducing contraction of isolated rabbit aorta (fig. 3D). Contractility data from the human umbilical veins also suggest an intermediate potency between des-Arg⁹-BK and Lys-des-Arg⁹-BK in the human B₁R (Gobeil et al., 1997). However, Sar-[D-Phe⁸]des-Arg⁹-BK is equipotent to Lys-des-Arg⁹-BK as an hypotensive agent in LPS-pretreated rabbits (Drapeau et al., 1991b) and the hypotensive episodes are prolonged (fig. 3C). The analog is inert in rabbits not treated with LPS, caused by the lack of cardiovascular B₁Rs in these animals (see below). The hypotensive effect of Sar-[D-Phe⁸]des-Arg⁹-BK is antagonized by Lys-[Leu⁸]des-Arg⁹-BK, a B₁R antagonist, and its analogues, but not by a B₂R antagonist Hoe 140 (Drapeau et al., 1993). A formal pharmacokinetic approach showed that Sar-[D-Phe⁸]des-Arg⁹-BK has a longer half-life than Lys-des-Arg⁹-BK in the rabbit circulation (Audet et al., 1997), probably accounting for the decisive advantage of the modified analog in vivo. Sar-[D-Phe⁸]des-Arg⁹-BK is also more potent than des-Arg⁹-BK as an algesic agent in a rat model of chronic inflammation (Davis and Perkins, 1994a), and may be superior to any natural sequence for demonstrating the LPS induction of hypotensive responses mediated by B₁Rs in this species (Nicolau et al., 1996). Sar-[D-Phe⁸]des-Arg⁹-BK appears to be a valid tool across different species to demonstrate the presence of a B₁R population. The replacement of a central tetrapeptide (Pro²-Phe⁵) in Lys-des-Arg⁹-BK by alkyl spacers, optimally 12-aminododecanoic acid, is another reported approach to modify a natural B₁R agonist (Tancredi et al., 1997). In this series of compounds, the best one retained 10% of Lys-des-Arg⁹-BK potency in the rat ileum contractility assay.

View larger version (26K): [in this window] [in a new window]   Fig. 3.   The rabbit treated with lipopolysaccharide (LPS), an animal model for the study of the regulation of the B1R. (A) Hypotensive effect of BK, Lys-BK and their des-Arg9-BK fragments in anesthetized rabbits. The endotoxin-treated animals (dotted lines) were injected with 10 µg of E. coli LPS 5 h before anesthesia. Peptides were injected intra-arterially as boluses. Abcissa: dose of peptide (ng per 1.5 to 2 kg animal) ; ordinate: fall of the mean arterial pressure, mmHg. Points are means of 5 to 18 determinations; the asterisk indicates value significantly changed by the LPS treatment. Reprinted in a modified form from Marceau and Regoli (1991), p. 40, by permission from Marcel Dekker, Inc. (B) B1R mRNA concentration in rabbit hearts relative to those of a housekeeping gene, GAPDH. Animals were untreated (controls) or injected with 25 µg/kg LPS 1 or 3 h before they were killed. Values are means ± SEM of triplicate determinations (with permission from Marceau et al., 1997, and the NRC Research Press). (C) Hemodynamic changes induced by an intraarterial bolus injection of the B1R agonist Sar-[D-Phe8]des-Arg9-BK (750 ng/kg) in anesthetized rabbits previously treated with LPS (30 µg/kg i.v. 5 h before anesthesia). CO, cardiac output reading derived from an electromagnetic flowmeter placed on the thoracic aorta; MABP, mean arterial blood pressure; TRP, calculated total peripheral resistance. The asterisk indicates value significantly different from baseline. Modified, from Audet et al. (1997), with permission from ASPET. (D) Initial responsiveness of aortic rings derived from control (saline vehicle i.v.) or LPS-treated rabbits (25 µg/kg, i.v. 8 h before they were killed). The cumulative concentration-effect curve of the B1R agonist Sar-[D-Phe8]des-Arg9-BK was established after a short in vitro incubation (45 min) to minimize the influence of isolation and tissue incubation on the responses. Values (means ± SEM of eight determinations from four animals in each group) are expressed as a percent of an internal contractile standard, the maximal effect of phenylephrine, established in each tissue. The maximal effects of the kinin differed significantly between groups (J.-F. Larrivée, D. Bachvarov, and F. Marceau, unpublished).

The development of kinin receptor antagonists has been pursued for more than two decades (Stewart, 1995). Selective antagonists have been crucial to define kinin receptor subtypes. The prototype of the kinin B₁R antagonists was [Leu⁸]des-Arg⁹-BK (Regoli et al., 1977). The B₁ nomenclature was latter applied to the rabbit aortic preparation in which the receptors were initially defined by both a typical order of agonist potency and by these antagonists. However, it became clear that these antagonists were not active in the most common contractile bioassays for BK (e.g., rat uterus, guinea pig ileum) or in vivo (hypotension) (Regoli and Barabé, 1980). The B₂R was not well defined up to Schild's criteria until 1985, when the first generation of antagonists based on [D-Phe⁷]BK were produced (Vavrek and Stewart, 1985). Since then, more efforts have been invested into the development of B₂R antagonists, relative to B₁R antagonists. Hoe 140 (icatibant; D-Arg[Hyp³, Thi⁵, D-Tic⁷, Oic⁸]-BK; abbreviations for unusual residues defined in footnote h to table 2, previous page; Hock et al., 1991) is a representative of the "second generation" B₂R antagonists, as it has several hours duration of action in animal models. Hoe 140 also exhibits a very high affinity (table 2) and is a competitive antagonist in most species with the noticeable exception of the rabbit (nonequilibrium insurmountable antagonism; Bachvarov et al., 1995). The effect of Hoe 140 on the human B₁R has probably been underestimated in the binding study quoted in table 2, because other functional or binding experiments rather indicate a K_i around 100 to 400 nM (Aramori et al., 1997; Bastian et al., 1997; Gobeil et al., 1996b; Zuzack et al., 1996). However, in practical terms, this peptide still exhibits a high selectivity toward B₂Rs. Hoe 140 has been used in many studies, particularly to assess the role of kinins in various models of experimental pathology. NPC 17731 is a B₂R antagonist that is chemically and pharmacologically similar to Hoe 140 (table 2). The first nonpeptide kinin antagonists, the phosphonium WIN 64338 and the quinoline FR 173657, are selective B₂R antagonists (Aramori et al., 1997; Sawutz et al., 1994; table 2). The latter compound is orally bioavailable, an important step toward the clinical evaluation of such antagonists.

Thus far, nonpeptide B₁R antagonists have not been yet reported. Structure-activity relationships for early peptide antagonists of the B₁R are discussed elsewhere (Regoli and Barabé, 1980). In summary, the substitution of Phe⁸ in des-Arg⁹-BK by a residue with an aliphatic (Ala, Ile, Leu, D-Leu, norleucine) or saturated cyclic hydrocarbon chain (cyclohexylalanine) produced antagonists of variable potency. Early experiments have also shown that substituting Pro⁷ (e.g., with Ala or Gly) can produce partial agonists or antagonists (Regoli and Barabé, 1980). Substitutions at this position were introduced in some of the more recent B₁R antagonists. The N-terminal extension of [Leu⁸]-des-Arg⁹-BK by a Lys residue confers a higher affinity in the species which have a similar relationships for agonists: the rabbit and human (table 2). Thus, Lys-[Leu⁸]des-Arg⁹-BK is the optimal B₁R antagonist based on natural amino acids. This family of peptides exhibits a competitive behavior and an almost ideal selectivity for the B₁R subtype (table 2). A tritiated form of Lys-[Leu⁸]des-Arg⁹-BK has been used as a ligand for the human B₁R in a recent study (Bastian et al., 1997), and binding competition data were very similar to those obtained with the agonist ligand [³H]Lys-des-Arg⁹-BK. Thus, the antagonist activity of [Leu⁸]des-Arg⁹-BK and its homologues is a reasonable basis for the pharmacological definition of the kinin B₁R.

Improving the stability and selectivity of B₁R antagonists is critically important for physiopathological investigations and therapeutic applications. Peptides generally suffer from many disadvantages as potential therapeutic agents, such as metabolic instability, poor oral absorption, rapid elimination, short duration of action and, in some cases, partial agonist activity. A metabolically stable peptide B₁R antagonist, Ac-Lys-[MeAla⁶, Leu⁸]des-Arg⁹-BK, has been elaborated by successive modifications of the high affinity prototype, Lys-[Leu⁸]des-Arg⁹-BK, with the main purpose of documenting the metabolic pathways of this class of agents (Drapeau et al., 1993). The N-methyl-alanine residue in position 6 removes recognition sites for ACE and neutral endopeptidase (fig. 2). Thus, although the novel antagonist peptide was not very potent (pA₂ of 6.5 in the rabbit aorta; K_i of 670 nM on cloned human B₁R, table 2), it showed resistance to ACE, neutral endopeptidase and also aminopeptidases, and exhibited an increased duration of action (15-30 min) as compared with the reference compound, against cardiovascular responses induced by an exogenous B₁R receptor agonist in LPS-pretreated rabbits. Ac-Lys[D-Nal⁷, Ile⁸]des-Arg⁹-BK (also referred to as R-715) has been designed with a similar rationale, and proved to be partially resistant to ACE, although retaining a high affinity for the human and rabbit B₁R (pA₂ of 8.5 in the rabbit aorta; Gobeil et al., 1996a; table 2).

Recently, metabolically stable B₁ and mixed B₁ + B₂ antagonists of high affinity and stability have been synthesized. B9858 [Lys-Lys-[Hyp³, Igl⁵, D-Igl⁷, Oic⁸]des-Arg⁹-BK] belongs to this series and has some selectivity for B₁Rs (table 2; Gera et al., 1996; Gobeil et al., 1997). This peptide retains the Lys⁰ residue, a favorable feature for binding to human B₁Rs (table 2). In binding assays, B9858 is more potent on the human B₁Rs than on its murine homologues: it is functionally highly active as a competitive antagonist against Lys-des-Arg⁹-BK-induced contractility of the human umbilical vein (pA₂ 9.2) and in the rabbit aorta (pA₂ 8.4; Gobeil et al., 1997). In the anesthetized dog, representing a hemodynamic system coexpressing B₁ and B₂Rs (see below, Section III.C.), 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).

Optimal peptide B₂R antagonists retain both Arg¹ and Arg⁹ residues, and B₁R agonists or antagonists typically lack Arg⁹. However, some novel antagonists with backbones constrained by nonnatural residues are somewhat more promiscuous in their selectivity. The des-Arg⁹ fragment of Hoe 140 has been studied by some investigators (table 2). This peptide may not be a metabolite of Hoe 140. It exhibits increased affinity toward the B₁R in all species studied, but retains fairly high residual antagonistic effects on B₂Rs in functional in vitro (Rhaleb et al., 1992) or in vivo studies (rat blood pressure, Lagneux and Ribuot, 1997). Similarly, NPC 18565 is the des-Arg⁹ fragment of NPC 17731, but exhibits a better selectivity and affinity for the human B₁R than Hoe 140 des-Arg (table 2). However, recently produced antagonists retaining Arg⁹, such as B9430 (D-Arg-[Hyp³, Igl⁵, D-Igl⁷, Oic⁸ ]BK) represent combined B₁ and B₂R antagonists, although the corresponding des-Arg⁹ fragment has a large selectivity toward B₁Rs (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 effects of 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 B₂R molecules are sufficiently similar to be antagonized by a single drug, a fact not appreciated until recently. Conformational analyses of B-9430 and B-9858 have been performed (Sejbal et al., 1997). The peptides exhibit no observable secondary structure in aqueous solution. However, some common and divergent elements emerge under certain experimental conditions, such as a type II -turn involving residues 2 to 5 common to both B₁ and B₂ antagonists. Such efforts may lead to the discovery of a universal "pharmacophore" for antagonizing kinin receptors.

Partial agonists are often discovered upon development of antagonist drugs. For example [Leu⁸]des-Arg⁹-BK, the prototype B₁R antagonist, may exhibit fairly high partial agonist behavior in some species, especially in the rats and mice. This has practical implications, because one of the most exciting emerging therapeutic applications of B₁R antagonists, analgesia, is commonly based on models involving these species (see below, Section IV.C.). The molecular structure of both the drug and the receptor are critically important for the partial agonist behavior of certain antagonists, as shown in a study involving the cloned human or murine B₁Rs transiently expressed in a cell line that constitutively produces the calcium sensitive photoprotein aequorin (MacNeil et al., 1997). [Leu⁸]des-Arg⁹-BK or Lys-[Leu⁸]des-Arg⁹-BK increased intracellular calcium concentrations in cells expressing the murine receptor (about 40% of the maximal activity of the full agonist des-Arg⁹-BK), whereas B9858 and R-715 were not active in this respect. None of these peptides acted as agonists on cloned human B₁Rs. Functional studies based on mouse or rat isolated tissues that express B₁Rs have confirmed that R-715 and des-Arg-Hoe 140 possess no or very low intrinsic stimulant activities in these species (Meini et al., 1996; Teater and Cuthbert, 1997; Allogho et al., 1995). However, partial agonist behavior of [Leu⁸]des-Arg⁹-BK or Lys-[Leu⁸]des-Arg⁹-BK was observed in some preparations (e.g., in the mouse isolated stomach, the rat isolated colon; Allogho et al., 1995; Teater and Cuthbert, 1997). This does not necessarily indicate the existence of B₁R subtypes that could be differentiated using a partial agonist behavior of some antagonists. If a pharmacological effect requires the stimulation of a very large proportion of the receptors on each cell, ligands with reduced efficacy are more likely to behave as antagonists (discussed by Leslie, 1987). This model may be applied quite well to the interaction of [Leu⁸]des-Arg⁹-BK on rat B₁Rs, as the maximum contractile (agonist) effect of [Leu⁸]des-Arg⁹-BK, relative to that of des-Arg⁹-BK, increases as a function of incubation time in the rat longitudinal ileum (i.e., the intrinsic activity is variable and increases; Meini et al., 1996). This is one of the smooth muscle preparations that exhibit in vitro up-regulation of B₁Rs upon isolation (see below, Section III.A.).

A partial agonist behavior may explain why the analgesic effect of [Leu⁸]des-Arg⁹-BK in rats and mice may be limited to a relatively narrow window of dosage (around 30 nmol/kg), as higher doses may exacerbate pain perception (Rupniak et al., 1997). No existing peptide antagonist seems ideal in term of affinity or selectivity to fully characterize the analgesic potential of B₁R blockade in small rodents (Rupniak et al., 1997).

Cross-desensitization is a valid criterion to distinguish substances acting on the same receptors (Schild, 1973). This implicitly refers to adaptation mechanisms present in, or closely associated with the receptor molecules. Thus, a "fading" of a response caused by the depletion of intracellular stores of an ion or some other desensitizing mechanisms distal to the receptors cannot be linked to receptor selectivity and would not qualify in exploiting this approach for receptor classification. In rat mesangial cells, a line that coexpresses both B₁ and B₂Rs, it has been possible to down-regulate a functional response (acute increase of [Ca²⁺]_i) to BK without depressing that to des-Arg⁹-BK, and vice versa (Bascands et al., 1993). An inositol trisphosphate formation or calcium response to des-Arg⁹-BK has been demonstrated upon BK desensitization in bovine endothelial cells or rabbit mesenteric artery SMCs, but the inverse was not readily demonstrable, because B₁R-mediated responses were not as tachyphylactic (Mathis et al., 1996; Smith et al., 1995;).

The two types of kinin receptors differ significantly in regard to receptor-mediated ligand internalization: exposure of cloned human B₂R to BK results in a rapid receptor-mediated ligand internalization and the sequestration of both the receptor and the G protein subunit in caveolae, accompanied by a profound loss of surface receptor binding (Austin et al., 1997; de Weerd and Leeb-Lundberg, 1997). In contrast, the activation of the cloned B₁R is not associated with ligand-induced receptor internalization at 37°C (Austin et al., 1997), and exhibits much less desensitization (further discussed below, Section III.D.). Experiments using truncated or chimeric human B₁ and B₂Rs in which the cytoplasmic carboxy terminus is either missing, or exchanged between the two receptor subtypes have indicated that the cytoplasmic carboxy terminus of the human B₂R contains sequences responsible for the receptor-mediated ligand internalization and receptor sequestration, which are not present in the C-terminal cytoplasmic domain of the B₁R subtype (Faussner et al., 1996), a domain of low interspecies conservation (fig. 1). Replacement from Cys³²⁴ of the cytoplasmic carboxy terminus of the B₂R with that of the B₁R has greatly reduced ligand-receptor complex internalization, whereas replacement from Cys³³⁰ of the cytoplasmic carboxy terminus of the B₁R (see fig. 1) with the B₂R counterpart has lead to a striking increase (~80% within 10 min) of internalization. Another study has confirmed the important role of the carboxy tail in B₂R internalization, identifying also other structural determinants (Prado et al., 1997). Thus, the consistent differences in signaling persistence between B₁ and B₂Rs appear to have a definite molecular basis and are of great potential interest to understand the relative importance of kinin receptor types in physiopathology. Whether other molecular mechanisms of desensitization are functionally important for kinin receptors (temporary loss of coupling of the G proteins, transcriptional suppression) is yet to be determined.

The structural features of the B₁R defined upon the expression cloning of the receptor cDNA confirmed its belonging to the superfamily of G protein-coupled receptors, exhibiting a typical 7 transmembrane domain architecture (fig. 1). Recently, it has been shown that CHO cells stably transfected with human B₁Rs cause Gq/11 and, to a lesser extent, Gi_1,2 to bind GTP upon stimulation with Lys-des-Arg⁹-BK (Austin et al., 1997). Thus, the identity and relative importance of G protein subtypes linked to the B₁R are similar to the ones coupled to B₂R (e.g., see de Weerd and Leeb-Lundberg, 1997). Both B₁ and B₂ receptors are primarily linked to polyphosphoinositide phospholipase C (PLC) activation. Specifically, activation of the naturally regulated B₁R is linked to increased phosphatidylinositol turnover in all systems studied up to now (reviewed by Marceau, 1995; see also Smith et al., 1995; Butt et al., 1995; see application to rabbit aortic SMCs, fig. 4A). Human B₁Rs stably expressed in Chinese hamster ovary (CHO) or 293 cells increase the turnover of inositol-phosphates when stimulated with Lys-des-Arg⁹-BK (Austin et al., 1997; Bastian et al., 1997). This effect was not prevented by pertussis toxin treatment in CHO cells, suggesting that the Gq/11 protein may be the more important than Gi_1,2 for coupling the B₁R to PLC (-isoform; Austin et al., 1997).

View larger version (46K): [in this window] [in a new window]   Fig. 4.   The rabbit cultured arterial smooth muscle cell (SMC), a model for the regulated cellular effects of the B1R. Middle: immunofluorescence pattern for -actin in rabbit aortic SMCs, passage 1. (A) Inositol phosphates (IP) extracted from rabbit aortic SMCs 5 min after exposure to des-Arg9-BK (1.7 µM) and pretreated or not with human recombinant IL-1 (5 ng/ml, 20 h before challenge). The asterisk indicates value significantly different from the corresponding control; the dagger indicates value significantly different from the corresponding one without IL-1 treatment. (Modified from Levesque et al., 1993, with permission.) (B) The B1R Bmax is significantly (asterisk) increased in these cells by the 16- to 20-h IL-1 treatment in a proportion very similar to the effect of the cytokine on IP hydrolysis. Nevertheless, the baseline population of B1Rs is important under the culture conditions employed (Levesque et al., 1995a). (C) Persistent [Ca2+]i signaling in a single representative SMC derived from the rabbit mesenteric artery and exposed to increasing concentrations of des-Arg9-BK (DBK). The signal (Fura-2 fluorescence) is composed of oscillating peaks, of an increase of the baseline, or both, depending on the individual cell and the agonist concentration. From Mathis et al. (1996), with permission of the authors, editor, and ASPET. (D) Effects of kinins on [3H]-thymidine incorporation into rabbit aortic SMCs in cells preteated with IL-1 (0.25 ng/ml, 24 h) and diclofenac (500 nM, 4 days). Both types of treatments are required to record a significant (asterisk) incorporation. Modified from Levesque et al. (1995b), with permission. Figures 4A and 4D were used with permission from the British Journal of Pharmacology; 109:1254-1262,1993; and 116:1673-1679,1995, respectively.

Calcium signaling was instrumental in the expression cloning of the B₁R (Menke et al., 1994). A transient increase in [Ca²⁺]_i (largely dependent on intracellular stores) is conceivably mediated by inositol 1,4,5-trisphosphate, one of the direct products of PLC, and this has been observed in response to stimulation of naturally expressed B₁Rs in animal or human cells (Bascands et al., 1993; Bkaily et al., 1997; Marsh and Hill, 1994; Mathis et al., 1996; Smith et al., 1995; fig. 4C) or cloned human or murine B₁R expressed in heterologous systems (Austin et al., 1997; Bastian et al., 1997; MacNeil et al., 1997). The sustained or oscillating phase of intracellular calcium increase is relatively more dependent on calcium influx by largely uncharacterized mechanisms (Bascands et al., 1993; partial blockade by nickel: Mathis et al., 1996; Smith et al., 1995). There is preliminary evidence for the activation of multiple types of calcium channels by the B₁R agonists in vascular SMCs (L, T, and R types, the latter being relatively more important in the sustained phase, Bkaily et al., 1997).

The moderate stimulatory effect of B₁R agonists on DNA synthesis in rabbit vascular SMCs (fig. 4D) is apparently dependent on another product of PLC, diacylglycerol, that activates a protein kinase C (PKC) (Levesque et al., 1995b). The SMC contractility mediated by B₁Rs (in the rabbit aorta) probably involves the cooperation of PKC and Ca²⁺ of both intra- and extracellular sources (Levesque et al., 1993). Filamin translocation from the membrane to the cytosol is another example of a Ca²⁺-dependent response elicited by either B₁ or B₂R stimulation in cultured bovine ECs (Wang et al., 1997). In this case, PKC action attenuates the response by limiting changes in calcium concentrations.

Thus, the identity of second-messengers is of limited value for the classification of the kinin receptors caused by common G protein subtypes and downstream elements. The most interesting differences observed so far may rather relate to the temporal patterns in systems that naturally coexpress both receptors. In bovine endothelial cells, rat mesangial cells, and rabbit mesenteric SMCs, the rise in [Ca²⁺]_i is more persistent and less tachyphylactic if elicited by B₁R stimulation, as opposed to B₂ activation, (Bascands et al., 1993; Mathis et al., 1996; Smith et al., 1995). Similarly, the effect of the B₁R stimulation on PLC activation is similarly more persistent than that the B₂R, in separate transfected CHO cell lines expressing analogous densities of each receptor type (Austin et al., 1997). In addition, these comparable cell lines were used to show that B₁Rs mediate much less receptor-ligand complex internalization than B₂Rs (Austin et al., 1997), a fact that could explain the persistence of B₁R signaling. Structural determinants for B₂R internalization not shared by the B₁R, are discussed above (Section II.C.) and may be the molecular basis of these temporal differences.

The kinin receptors might be involved in other transduction mechanisms, different from those described above. For instance, both B₁ and B₂Rs are equally capable of suppressing platelet-derived growth factor- (PDGF) induced DNA synthesis in rat arterial SMCs (Dixon and Dennis, 1997). The mechanism of this suppression is still obscure, as the effects were not dependent on PLC, prostaglandins (PGs), or cyclic adenosine 3',5'-monophosphate (AMP), or correlated with the extent of PLC activation (the B₁R being relatively ineffective in this respect; Dixon and Dennis, 1997).

Secondarily released mediators formed by Ca²⁺ dependent enzymes, such as endothelial nitric oxide synthase and cytosolic phospholipase A₂, may account for the production of nitric oxide (Drummond and Cocks, 1995a; Pruneau et al., 1996) and eicosanoids (Levesque et al., 1995b) mediated by B₁Rs. The secondary mediators are also shared with B₂Rs in many experimental systems, and are of considerable importance for the in vivo pharmacology of B₁Rs (see below, Section IV.). NO and PGs act in an autocrine or paracrine manner, thus extending the signaling mechanisms that kinins may activate in receptive tissues (notably to cyclic guanosine 3',5'-monophosphate and cyclic AMP).

The two kinin receptor subtypes have been further defined on molecular basis after the recent isolations and characterizations of their 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 B₂R subtype (McEachern et al., 1991). Subsequently, the human B₂R has been cloned from the fibroblast cell line CCD-16Lu by PCR and cDNA screening (Hess et al., 1992), its genomic organization studied (Kammerer et al., 1995; Ma et al., 1994) and its localization mapped to chromosome 14q32 (Ma et al., 1994; Powell et al., 1993). The murine and rabbit B₂R genes were also characterized (Bachvarov, et al. 1995; McIntyre et al., 1993). The deduced amino acid sequences of the B₂Rs from the four species studies have shown extensive similarity (80 to 84%), consistent with being orthologs of the same gene.

The B₁R became a defined molecular entity since the cloning and sequencing of the corresponding human cDNA. That the human embryonic cell line IMR-90 expresses B₁Rs has been known for some time, and receptor stimulation leads to metabolic effects such as collagen and DNA synthesis (Goldstein and Wall, 1984). Menke et al. (1994) observed that these cells bind [³H]Lys-des-Arg⁹-BK and that the binding is vigorously up-regulated by pretreating the cells with IL-1, consistent with current regulation studies (see below, Section III.). These stimulated cells were an enriched source for the isolation of mRNA further used for the expression cloning of the human B₁R cDNA. Different molecular weight mRNA fractions were consecutively injected in X. laevis oocytes and the photoprotein aequorin was used as an indicator of the ability of the B₁R agonist Lys-des-Arg⁹-BK to mediate Ca²⁺ mobilization in the injected oocytes. A 1307-bp clone that included a 1059 nucleotide open reading frame encoding a 353-amino acid protein was thus isolated and proposed to encode the B₁R (Menke et al., 1994; predicted primary structure, fig. 1). This discovery was followed by the full analysis of the cDNA of the B₁R (1.4 kb), considerably smaller than that coding for B₂R (over 4 kb), and by the determination of the corresponding genomic organization and chromosomal localization (14q32 between markers D14S265 and D14S267) (Bachvarov et al. 1996, 1998b; Chai et al., 1996; Yang and Polgar, 1996; fig. 5). The rabbit and mouse homologues of the B₁R gene were also isolated and characterized (MacNeil et al., 1995; Pesquero et al., 1996) as there is 68 to 78% homology between the B₁Rs of the three species studied. A rat nucleotide sequence with a similar level of homology is proposed to encode the rat B₁R (GenBank sequence U66107), but its pharmacological profile is not reported yet. The cloned human B₁R also exhibits the seven transmembrane structure typical for G protein coupled receptors, although its amino acid sequence identity with the B₂R type of BK receptor is only about 36%. Nevertheless, they are the closest relatives based on the degree of similarity, followed by angiotensin receptors.

View larger version (17K): [in this window] [in a new window]   Fig. 5.   Schematic structure of the human B1 receptor gene (Bachvarov et al., 1996; Yang and Polgar, 1996). The hatched region represents the uninterrupted coding sequence. Intron 1 length has been estimated to be 6 to 7 kb. Motifs of functional significance are indicated (start and stop codons, the polyadenylation signal, the TATA box in the 5' flanking sequence, an important NF-B-like site). Two polymorphic sites that can be studied by restriction length polymorphism are indicated (AciI, TaqI). Some of the fragments tested for promoter activity in constructions involving a reporter gene are indicated by numbered lines terminated with an arrowhead pointing to the right (sense orientation), or the left (antisense orientation) : (1) a 1.9-kb fragment 5' of the coding sequence and spanning exon 2 (Chai et al., 1996); (2) a 0.87-kb fragment in the same region; (3 and 4) 0.45-kb fragments 5' of and partially overlapping exon 1; (5) a 0.92-kb fragment starting in intron 1 and spanning exon 2 (Yang and Polgar, 1996); (6) a 0.14-kb gene fragment (111/+34 relative to the transcription initiation site) containing a consensus TATA box (Ni et al., 1998); (7) a 226-nt gene fragment (139/+86); (8) the most active promoter fragment tested (735/+86); a postulated negative control region in the core promoter is located 5' relative to fragment 8 (D. R. Bachvarov, R. Drouin, M. Angers, J.-F. Larrivée, M. Bachvarova, and F. Marceau, in preparation).

The predicted intracellular carboxy tail of the human B₁R is longer than that of their rabbit, mouse or rat counterparts (fig. 1). Perhaps significantly, the only effective antiserum to human B₁R to date has been derived from a rabbit immunized with the 16-mer C-terminal peptide (underlined in fig. 1; Hess et al., 1996). These antibodies detected human recombinant B₁Rs expressed under the control of a viral promoter in transfected COS cells (immunofluorescence, Western blot; Hess et al., 1996). Four immunoreactive bands were identified in the Western blot experiments, corresponding to 39, 31, 25, and 20 kDa. Because the predicted molecular weight of the human B₁R without carbohydrate is 40.4 kDa, it is quite possible that all four bands correspond to proteolytic cleavage fragments. All the published immunohistochemistry of the B₁R (see Section IV.) can be traced back to this antiserum.

Few studies of the structure-function relationship have been performed on the B₁R, and they are always comparative to the B₂R, for which a putative binding pocket involving transmembrane (TM) domain VI and the adjacent extracellular residues has been characterized (discussed by Leeb et al., 1997). The similarity between the two receptor subtypes is significantly higher in these regions. Leeb et al. (1997) have constructed chimeric human B₁ and B₂Rs in which TM-VI domains have been exchanged and have shown that TM-VI in both receptor subtypes are important for the ability of these receptors to discriminate between their subtype-selective agonists. Substitution of B₁ TM-VI into the B₂R dramatically reduced the affinity of the B₂R-selective agonist BK. This affinity was fully restored when two residues, Tyr²⁵⁹ and Ala²⁶³ near the extracellular surface of the chimeric B₁R-TM-VI (see fig. 1) were replaced with the corresponding residues in the wild-type B₂R, which are Phe²⁵⁹ and Thr²⁶³. Conversely, substitution of the B₂ TM-VI into the B₁R reduced the high affinity binding of the B₁-selective agonist Lys-des-Arg⁹-BK. Interestingly, the latter chimera retained a functional response ([Ca²⁺]_i increase) to a high concentration of Lys-des-Arg⁹-BK and to Lys-BK, the latter peptide confirming its status as a B₁ and B₂R agonist. TM-VI is probably less important for ligand docking in the B₁R than in the B₂R. Recent experiments have focused on TM-III exchange between the two kinin receptors, as this domain is presumably facing the TM-VI docking domain (Fathy et al., 1997; L. M. F. Leeb-Lundberg, personal communication). Introduction of B₁R TM-III into the B₂R sequences decreased the affinity of the B₂R selective agonist BK, whereas the substitution of B₂R TM-III into the B₁R decreased the affinity of the selective ligand Lys-des-Arg⁹-BK. High affinity for BK was restored when Lys¹¹⁸ in B₁R TM-III (numbering as in B₁R) was replaced with the corresponding wild-type B₂R residue (Ser¹¹¹). These observations and complementary analyses suggest that Lys¹¹⁸ in the B₁R structure repels the positive charge of the ligand C-terminal Arg⁹, thus providing a possible explanation for the fundamental feature of the B₁R pharmacological profile: selectivity for des-Arg⁹-BK homologues. These types of experiments further show the surprising structural compatibility of the two types of human kinin receptors, already suggested by the recent development of promiscuous antagonist ligands (see above, Section II.B.).

There is no genetic support for the existence of B₁ or B₂R subtypes in any given species, although fairly large differences of pharmacological profiles exist from one species to another. Experiments based on cloned and highly homologous receptors expressed in heterologous systems show conclusively that these pharmacological discrepancies are species differences (table 2).

Perhaps it can be postulated that the ultimate form of receptor antagonism is the targeted disruption ("knockout") of receptor genes by homologous recombination. In mammalian species this technology, based on manipulations of embryonic stem cells, is currently restricted to the mouse. The B₂R knockout mice fail to respond to BK (smooth muscle, afferent nerve stimulation, etc.), indicating that no other B₂R subtype or variant responsive to that peptide exists in this species (Borkowski et al., 1995). These animals develop normally and may not exhibit important hemodynamic changes. However, if overloaded with dietary NaCl, they develop severe hypertension with end organ damage (Alfie et al., 1996). The relative importance of renal deficit in B₂R knockout mice is an important novel aspect of the kallikrein-kinin system, further exemplified by a polymorphism in the human B₁R gene (see below, Section III.C.). There is ample evidence that cytokine or postinjury regulation of the B₁R gene expression is not disrupted in B₂R knockout mice (Asonganyi et al., 1996; Cuthbert et al., 1996; Rupniak et al., 1997; Seabrook et al., 1997; see Section IV.). Such results are important, because both kinin receptor genes lie close to each other on human chromosome 14. This clustering may not represent a functionally coordinated locus comparable to those of globins or serum albumin homologues, for instance.

The B₁R knockout mice have been produced quite recently (Bader et al., 1997; J. B. Pesquero, personal communication). These animals develop normally, are normotensive, but fail to functionally respond to des-Arg⁹-BK (e.g., contractility of the mouse isolated stomach). Further, basal and LPS-induced expression of B₁R mRNA is absent (RT-PCR, reverse transcriptase polymerase chain reaction). Further experiments will be needed to exploit this murine strain in different physiopathological conditions. In this context, the creation of a double B₁/B₂R knockout strain is extremely intriguing.

The transcriptional activity of the receptor genes is regulated by both tissue-specific and physiological factors. This may be relevant for receptor classification, and refers to the promoter function of distinct genes. Although there is some evidence that B₂R expression is transcriptionally regulated in cultured cells (e.g., modest positive regulation by cyclic AMP, PDGF and IL-1; Bathon et al., 1992; Dixon, 1994; Dixon et al., 1996; Schmidlin et al., 1997; suppression by tumor necrosis factor-; Sawutz et al., 1992), these receptors are constitutively expressed in a wide variety of tissues. One of the few systematic efforts to document B₂R regulation concerns a sexual dimorphism of B₂R expression in several organs of the rat, the females expressing more of the corresponding mRNA (detected using RT-PCR; Madeddu et al., 1997). Ovariectomy reduced the hypotensive effect of BK and estrogen treatment restored the hemodynamic effect of the kinin; however, hemodynamic responses to some agonists for other receptors followed a similar pattern (Madeddu et al., 1997). Based on hemodynamic responses to des-Arg⁹-BK, the B₁Rs were not regulated by estrogen in these studies. In some chronic inflammation models, B₂R function may be depressed (several examples will be given in Section IV.), but it is not known whether this is caused by transcriptional repression or ligand-mediated down-regulation. In contrast, a large body of evidence shows that the B₁R is generally absent from normal tissues and animals (with some exceptions), but is rapidly induced after certain types of injuries in many species (Marceau, 1995). Molecular and genetic techniques have recently confirmed this receptor up-regulation. The regulatory mechanisms of gene expression are a recent focus of the research on B₁Rs, which is covered in detail in the next section.

	III. Immunological and Molecular Analysis of B1 Receptor Regulation by Tissue Injury

Top Previous Next References

Smooth muscle contractility measured using organ bath methodology has been historically important to discover and characterize the B₁Rs. The distinct pharmacological profile of the kinin B₁R was initially derived from the analysis of kinin-induced contractility of the rabbit aorta (Regoli et al., 1977). On this preparation, exogenous kinins exert a contractile effect that develops from an initial null level but increases in magnitude as a function of the in vitro incubation time (Bouthillier et al., 1987). In this sense, the rabbit isolated aorta is a model for many other smooth muscle preparations derived from normal animals and used to illustrate the phenomenon of postisolation B₁R induction. Many such systems from various animal species are reviewed elsewhere (Marceau, 1995). To cite just a few recently documented examples, the rat portal and pig renal veins also demonstrate this phenomenon based on contractility (Campos and Calixto, 1994; Rizzi et al., 1997). It is now clear that nonmuscle cellular elements can also exhibit the postisolation induction of B₁R: the vascular endothelial and colonic epithelial cells acquire such a responsiveness in vitro from a null or low initial level in the bovine or porcine isolated coronary artery and rat isolated colon, respectively (Drummond and Cox, 1995a; Pruneau et al., 1996; Teater and Cuthbert, 1997). In such preparations, the de novo synthesis of B₁R is supported by the selectivity of the regulated behavior; similar responses mediated by other receptor types (including B₂Rs in some preparations) are much more stable as a function of time. Characteristically, the use of metabolic inhibitors also support the postisolation B₁R induction. Inhibition of RNA synthesis with actinomycin D, of protein synthesis with cycloheximide, anisomycin or puromycin, or of protein maturation with brefeldin A or tunicamycin, specifically prevent the development of a responsiveness to exogenous B₁R agonists without displaying toxic effects on other types of responses, or an acute inhibitory effect on responses to kinin that had been allowed to develop without metabolic inhibitors (Deblois et al., 1991; Drummond and Cox, 1995a; Pruneau et al., 1996; Teater and Cuthbert, 1997). Brefeldin A is a selective inhibitor of the translocation of newly synthesized proteins from the endoplasmic reticulum to the Golgi apparatus. A membrane receptor with transmembrane domain(s) would follow such a relatively slow maturation process and accordingly, the drug completely and selectively prevented des-Arg⁹-BK-induced contraction in the rabbit isolated aorta (Audet et al., 1994). The B₁R contains up to three putative but conserved N-glycosylation sites (fig. 1), which may be of importance for receptor assembly, stability or function. This is supported by the partial inhibitory effect of tunicamycin on the development of responses to des-Arg⁹-BK in rings of rabbit aorta (Audet et al., 1994). The biochemical effects of some of the metabolic inhibitors have been validated in assays involving the incorporation of [³⁵S]methionine (Deblois et al., 1991