Abstract
Studies investigating the role of bradykinin in disease states such as hypertension, sepsis, and asthma have been confounded by difficulties in measuring the concentration of this short-lived peptide. The purpose of this study was to determine a stable metabolite of bradykinin in the systemic circulation of humans. Bradykinin (containing trace concentrations of [3H]bradykinin) was administered i.v. into three human volunteers in increasing amounts up to a maintenance rate of 200 ng/kg/min until a total dose of 1 mg was given. Metabolic products were purified and identified by HPLC and by electrospray ionization mass spectrometry. Infused bradykinin was rapidly degraded, such that no exogenous bradykinin was detected in venous plasma sampled during infusion. BK1-5 (Arg-Pro-Pro-Gly-Phe), the 1-to-5 amino acid fragment of bradykinin, was identified as a major stable plasma metabolite of bradykinin. Plasma concentrations of BK1-5 correlated with dose of bradykinin infused and concentrations at the end of bradykinin infusion were 1510 to 4600 fmol/ml of blood. BK1-5 was cleared from blood with a terminal half-life of 86 to 101 min. Thus, in humans, bradykinin is rapidly degraded in vivo to BK1-5, a stable metabolite. Measurement of this metabolite could provide a tool to assess pathophysiologic and pharmacologic alterations in systemic bradykinin generation associated with human disease.
Bradykinin is a vasoactive nonapeptide that has cardioprotective effects. Bradykinin promotes vasodilatation via bradykinin B2 subtype receptors, stimulating the endothelial production of nitric oxide, prostaglandin I2, and endothelium-derived hyperpolarizing factor (Vanhoutte, 1989). Bradykinin induces the release of tissue-type plasminogen activator in vitro (Emeis and Tranquille, 1992) and in humans (Brown et al., 1999); inhibits thrombin-induced platelet activation (Hasan et al., 1996); and contributes to ischemic preconditioning in animal models (Parratt et al., 1997). Bradykinin indirectly exerts antiproliferative effects on vascular smooth muscle via nitric oxide production (Ritchie et al., 1998). Inhibition of bradykinin degradation contributes to the acute blood pressure-lowering effects of angiotensin-converting enzyme (ACE) inhibitors (Gainer et al., 1998) and this mechanism also is thought to contribute to favorable cardiovascular consequences seen with chronic administration of this class of drugs (Linz et al., 1995). In addition, alterations in the kallikrein-kinin system are hypothesized to play a role in the pathophysiology of such disease states as hypertension, insulin resistance, sepsis, arthritis, and asthma (Margolius, 1995; Kaplan et al., 1997).
Despite the prominent role of the kallikrein-kinin system in the regulation of vascular tone and inflammation, studies in humans have been limited by difficulties in accurately measuring bradykinin concentrations (Pellacani et al., 1992; Margolius, 1995). Bradykinin is rapidly degraded by enzymes such as ACE (kininase II, EC 3.4.15.1), carboxypeptidase N (kininase I, EC 3.4.17.3), neutral endopeptidase (EC3.4.24.11), and aminopeptidase P (EC 3.4.13.19; Bhoola et al., 1992; Fig. 1). The reported half-life of bradykinin in vivo is 17 s (Ferreira and Vane, 1967). Early studies using radioimmunoassay methodologies lacked sufficient specificity to distinguish bradykinin from its precursors and metabolites (Goodfriend and Odya, 1979). In later studies with more specific antibodies, measured kinin levels have depended on the particular antiserum used in the assay (Bönner et al., 1987). In addition, low circulating concentrations of bradykinin in the presence of substantial amounts of its precursor kininogen and both kinin-generating and -degrading enzymes can lead to artifactual changes in bradykinin concentrations during blood sampling. Due to these confounding variables, the reported range of normal bradykinin plasma concentrations has varied over several orders of magnitude (Bönner et al., 1987; Pellacani et al., 1992) and the relationship between disease states and bradykinin concentrations has been difficult to define.
One strategy for exploring the pathophysiologic role of a short-lived effector molecule such as bradykinin is to identify a stable metabolite that can be measured as an index of the parent molecule. Such a strategy has been used to investigate the role of prostanoids and other vascular mediators in various disease states (Falardeau et al., 1981;Brash et al., 1983). Thus, the purpose of this study was to determine a stable metabolite of systemic bradykinin in humans. To do this, we analyzed blood and urine samples for bradykinin and its metabolites after i.v. administration of [3H]bradykinin into normal humans. HPLC was used to separate bradykinin and its fragments. The identity of a stable metabolite was confirmed by liquid chromatography-mass spectroscopy (LC-MS) with an electrospray (ESI) source. The data indicate that, in humans, bradykinin is rapidly metabolized to the pentapeptide BK1-5 (Arg-Pro-Pro-Gly-Phe), a stable plasma metabolite.
Materials and Methods
Bradykinin Infusion and Sample Collection.
Written informed consent was obtained from volunteers, and the protocol was approved by the Vanderbilt University Medical Center Institutional Review Board. [2,3-prolyl-3,4-3H]Bradykinin (96.0 Ci/mmol; Amersham, Arlington Heights, IL) and unlabeled bradykinin (Sigma, St. Louis, MO) were administered i.v. to three normal Caucasian volunteers who were taking no medications (2 male, 1 female; age 39 ± 1.5 years; weight 72 ± 4 kg). Bradykinin infusion was initiated at 25 ng/kg/min and the rate doubled every 5 min until a final rate of 200 ng/kg/min was achieved. A total dose of 50 μCi of [3H]bradykinin and 1 mg of unlabeled bradykinin was administered. Purity of the infusate was confirmed by HPLC analysis. Heart rate and blood pressure were monitored continuously during the infusions. Blood samples for determination of bradykinin and metabolites were obtained during the dose escalation, before termination of infusion, and at 0.5, 1, 2, 3, 5, 7.5, 10, 12.5, 15, 20, 60, 120, and 180 min after discontinuation. Blood (5 ml) was drawn into a plastic syringe via an indwelling i.v. catheter and immediately placed in 15 ml of chilled ethanol. We and others have found this procedure to effectively limit ex vivo bradykinin production and degradation (Nasjletti et al., 1975; Hilgenfeldt et al., 1998). After 1 h at 4°C, samples were centrifuged and the supernatant stored at −70°C until analyzed.
Urine was collected for 6 h after the infusion for determination of bradykinin and metabolites. Urine collections were acidified with 10% (v/v) of 1 N HCl and stored at −20°C until analysis. Preliminary experiments showed [3H]bradykinin was stable in urine for several weeks under these conditions.
Sample Preparation.
Ethanolic plasma supernatant (5 ml) was dried under vacuum at 37°C. The residue was resuspended in 1500 μl of HPLC mobile phase and filtered through a 0.2-μm filter and injected onto the HPLC. Urine samples were extracted through 1 ml C18 Sep-Pak cartridges (Waters, Milford, MA) activated by 3 ml of methanol and 3 ml of 0.1% trifluoroacetic acid (TFA) in water. Sample (1 ml) added to 3 ml of 0.1% TFA-water was applied to the Sep-Pak, washed with 3 ml of 0.1% TFA-water, and eluted with 3 ml of 80:20 mixture of 0.1% TFA-water:acetonitirile. The eluate was dried under vacuum at 37°C, reconstituted in a final volume of 500 μl of HPLC mobile phase, and injected onto the column. When blank urine samples were spiked with [3H]bradykinin, 99% of radiolabel was recovered from the extraction process.3H-Metabolites were not available to evaluate recovery, however, 88 to 92% of total counts present in a plasma or urine sample were recovered after specimen processing on either a Sep-Pak or with evaporation and reconstitution.
HPLC Assay.
Reversed phase HPLC was performed on a C18 column (4.6 × 250 mm, 5-μm particle size; Alltech, Deerfield, IL) with a linear gradient that separates bradykinin and the BK1-5, BK1-6, BK1-7, and des-Arg9BK metabolites. Mobile phase A was 0.1% TFA-water. Mobile phase B was 0.1% TFA in a 90:10 mixture of acetonitrile:water. The gradient of 90% A:10% B to 65% A:35% B was run over 7 min at a rate of 0.7 ml/min with a Hitachi HPLC system (Tokyo, Japan). Isocratic conditions were maintained at 65% A:35% B for the remainder of the run. A photodiode array UV detector (Hitachi) was used to detect peaks corresponding to standards of bradykinin and metabolites (Sigma). Experimental samples were collected in aliquots with a fraction collector and radioactivity determined by liquid scintillation counting (Packard Instruments, Downers Grove, IL). Radioactive metabolites were identified based on coelution with known standards. Identity was confirmed by LC-ESI-MS.
LC-ESI-MS.
The identity of bradykinin and metabolites in HPLC aliquots were determined with a FinniganMAT TSQ 7000 series triple quadrupole mass spectrometer system (San Jose, CA) in line with a Waters 2690 liquid chromatography system. Liquid chromatography on an Eclipse XBD-C18 column (2.1 × 50 mm, 5 μm; Hewlett-Packard, Palo Alto, CA) used mobile phase A (0.5% acetic acid in a 90:10 mixture of water:methanol) and mobile phase B (0.5% acetic acid in methanol) in a linear gradient of 100% A:0% B to 35% A:65% B over 4 min at a flow rate of 0.250 ml/min. For the MS analysis, the ESI source voltage was 4.00 kV with a capillary lens potential and temperature of 13 V and 200°C, respectively. The tube lens voltage was 108 V. Metabolites were detected by MS-MS. Samples were monitored for the molecular ions of bradykinin (m/z 531, [M + 2H]2+) and BK1-5 (m/z 573, [M + H]+), which then underwent collision-induced dissociation (CID) with a voltage offset of −34 eV (laboratory frame of reference) to produce a spectrum of daughter ions. For quantification, the predominant daughter ions of bradykinin (m/z 70) and BK1-5 (m/z417) were monitored and compared with known concentrations of coanalyzed internal standards ([2H8-Phe5]- bradykinin and [13C2,15N-Gly4]BK1-5, both custom synthesized by Dr. James I. Elliot, Yale University, New Haven, CT).
Results
Response to Bradykinin Infusion.
All three volunteers tolerated the bradykinin infusion without serious side effects. Mean arterial pressure decreased from 101 ± 25 (mean ± S.D.) to 86 ± 15 mm Hg at the 200 ng/kg/min maintenance rate, where it remained constant during the infusion. Heart rate increased transiently after initiation of the 200-ng/kg/min dose but returned to baseline (70 ± 8 to 82 ± 29 to 69 ± 8 beats per minute). All volunteers experienced facial flushing and one volunteer reported a metallic taste during the infusion. Another volunteer experienced profuse, watery diarrhea ∼1 h after discontinuation of infusion.
Metabolic Products of Bradykinin in Blood of Bradykinin-Infused Volunteers.
The infusion of [3H]bradykinin resulted in two peaks of radioactivity that were detected by HPLC in blood samples obtained before termination of bradykinin infusion and for 2 h thereafter. A representative sample is shown in Fig.2. A radiolabeled peak coeluted with the BK1-5 standard at 18 to 20 min (denoted by ∗ in Fig. 2). Material coeluting with the radiolabel was analyzed by ESI-MS and the total ion current chromatogram obtained is shown in Fig.3A. The parent ion ([M + H]+) for BK1-5 is m/z 573 (Dikler et al., 1997) and, as is evident in Fig. 3B, them/z 573 ion current chromatogram contains a single peak at 7.45 min. For comparison, Fig. 3C shows them/z 531 ion current chromatogram of the bradykinin parent ([M + 2H]2+), which has clear separation from the metabolite. The m/z 573 compound was subsequently analyzed by LC-ESI-MS-MS (Fig.4). The predominant daughter ions of chemically pure BK1-5 represent fragmentation between peptide bonds along the backbone (Fig. 4A). N-Terminal daughter ions are denoted “bn”, whereas C-terminal daughter ions are labeled “yn ” with the subscript indicating the number of residues in the fragment ion (Papayannopoulos, 1995). Some daughter ions have retained a molecule of water or have lost an amino group (17 Da). The CID spectrum obtained demonstrated such ions at m/z 426 (b4 + H2O), 417 (y4), 409 ([b4 + H2O]-17), 320 (y3), and 237 (b2-17;Dikler et al., 1997). Figure 4B shows the analysis of the endogenously derived metabolite shown in Fig. 3B. The predominant daughter ions of the unknown (Fig. 4B) are identical with those of the BK1-5 standard (Fig. 4A) and include the expected m/z 426, 417, 409, 320, and 237 ions.
A second peak was unretained on the reversed phase HPLC column (denoted by + in Fig. 2). Although this peak was not definitively identified, several lines of evidence suggest it represents the single residue [3H]Pro from a hydrolyzed [3H]bradykinin molecule. First, it did not coelute with standards for Arg-Pro-Pro, Pro-Pro or other amino-terminal metabolic fragments of bradykinin that included the3H-labeled Pro2 and Pro3 residues. Second, proline standard coeluted with the radiolabeled peak in the void volume, whereas dipeptides and larger peptides are retained on this and similar C18 HPLC systems (Sheikh and Kaplan, 1986a,b).
Subsequently, we explored the pharmacokinetics of BK1-5 in vivo. Decay of [3H]BK1-5 in plasma at the end of the infusion is shown in Fig. 5. BK1-5 concentrations were calculated from the specific activity of [3H]bradykinin in the infusate and confirmed by LC-ESI-MS. The peak concentrations of BK1-5 were 1510, 1920, and 4610 fmol/ml of blood. We were able to perform decay calculations in two volunteers. The initial half-life of decay (t1/2α) for BK1-5 in the two volunteers was 1.3 and 2.1 min and terminal half-life of decay (t1/2β) was 86 and 101 min (Fig. 5). Blood samples also were obtained for metabolites during the dose-escalation phase of the bradykinin infusion. BK1-5 concentrations could be quantified by LC-ESI-MS and increased with bradykinin infusion rate (Fig. 6).
Using scintillation counting, no [3H]bradykinin was detected in blood samples collected during bradykinin infusion. The limits of detection of bradykinin for scintillation counting are <1000 fmol/ml. However, with more sensitive LC-ESI-MS, we could detect bradykinin in blood obtained during the 200-ng/kg/min infusion (Fig.3C).
Metabolic Products of Bradykinin in Urine.
Recovery of radiolabel in the urine ranged from 6 to 42% of total activity infused in the three volunteers. Analysis of urine samples by HPLC revealed a single, early [3H] peak in the void volume consistent with [3H]proline and similar to that seen in plasma samples. No radioactivity corresponding to BK1-5 or bradykinin was detected.
Discussion
We have identified BK1-5 as the major, stable plasma peptide metabolite of systemic bradykinin in humans. The metabolite was detected initially by HPLC coelution. Aliquots from HPLC containing the in vivo metabolite were subsequently analyzed by LC-ESI-MS and contained the predicted molecular ion for BK1-5,m/z 573. Significantly, the CID daughter spectrum of this m/z 573 ion matched that of a chemically synthesized BK1-5 standard, confirming the identity of this metabolite. The daughter ions of BK1-5 determined in this study correspond to those previously reported for the pentapeptide (Dikler et al., 1997).
Data identifying BK1-5 as the primary metabolite of bradykinin support the extensive role of the vascular endothelium, where ACE and other kininases are primarily localized, in the catabolism of bradykinin. In this study, there was virtually complete removal of [3H]bradykinin from the circulation after passing through the pulmonary and systemic circuit before sampling. The lack of measurable [3H]bradykinin in venous samples suggests that the half-life of bradykinin in vivo is considerably shorter than that reported in vitro in human plasma (30 s to 60 min, depending on bradykinin concentration; Shima et al., 1992;Decarie et al., 1996). This is consistent with studies with isolated rat lung models that suggest ≥99% metabolism of bradykinin with one pass through the pulmonary vascular bed (Ryan et al., 1994; Prechel et al., 1995).
BK1-5 represents the degradation product of two successive cleavages of bradykinin by ACE (kininase II). Previous human studies have demonstrated the importance of ACE in the in vivo degradation of bradykinin in humans, where ACE inhibition markedly potentiates the systemic hypotensive effects of bradykinin (Bönner et al., 1992;Brown et al., 1996). In vitro studies suggest that BK1-5 is the major metabolic product of ACE in the bradykinin catabolic pathway (Sheikh and Kaplan, 1986b; Shima et al., 1992). BK1-5 also has been detected in nasal secretions of patients with allergic rhinitis (Majima et al., 1996). The terminal half-life of exogenous BK1-5 in rabbits is 24 min (Hasan et al., 1999). This study extends these observations, establishing that BK1-5 is an important in vivo metabolite of bradykinin in the human circulatory system with a half-life of ∼90 min in the volunteers reported herein. Significantly, in this study, plasma BK1-5 concentrations increased in proportion to the mass of bradykinin administered systemically. The prolonged half-life of BK1-5 in comparison to the almost instantaneous elimination of bradykinin suggests BK1-5 provides a unique, long-lasting in vivo marker of activity of the kallikrein-kinin system.
Studies exploring the bioactivity of BK1-5 are limited. In one report, BK1-5 prevented thrombin-induced platelet activation in vitro at supraphysiologic doses (Hasan et al., 1996). A separate in vivo study reported that BK1-5 (600 μM) had comparable efficacy to aspirin (4.6 mg/kg) in preventing occlusion in a canine coronary thrombosis model (Hasan et al., 1999).
In this study, a second radiolabeled metabolite in plasma was tentatively identified as proline. This probably represents further cleavage of the BK1-5 molecule by other endothelial peptidases such as aminopeptidase P and dipeptidyl-peptidase IV (Fig. 1; Simmons and Orawski, 1992). Aminopeptidase P has been identified in the pulmonary vascular endothelium (Simmons and Orawski, 1992) and accounts for 30% of [3H]bradykinin degradation in isolated rat lung perfusion experiments (Prechel et al., 1995).
Components of the kallikrein-kinin system have been identified in human urine (Margolius et al., 1971; Hial et al., 1976; Abe et al., 1981; Hilgenfeldt et al., 1995; Saito et al., 1995). However, we report that no [3H]bradykinin was detected in the urine after systemic administration, consistent with animal perfusion experiments demonstrating near total intrarenal degradation of the peptide (Nasjletti et al., 1975; Carone et al., 1976). Although bradykinin and BK1-5 were found in the urine by LC-ESI-MS in this study (data not shown), the lack of detectable [3H]bradykinin or [3H]BK1-5 suggests that the source of these urinary kinins was renal. Thus, measurement of urine kinin concentrations does not reflect the systemic activity of the human kallikrein-kinin system.
In conclusion, we report that BK1-5 is the major, stable plasma peptide metabolite of bradykinin in humans. This study is important because it lays groundwork for the development of specific and sensitive MS methods to accurately measure BK1-5 in human plasma as an index of systemic bradykinin production. Such a method will allow for further elucidation of the role of bradykinin in the pathophysiology of important human diseases, such as hypertension, diabetes, and sepsis.
Footnotes
-
Send reprint requests to: Nancy J. Brown, M.D., Division of Clinical Pharmacology, Vanderbilt University Medical Center, 560 MRB-1, Nashville, TN 37232-6602. E-mail:nancy.brown{at}mcmail.vanderbilt.edu
-
↵1 This study was supported by National Institutes of Health Grants HL56963, GM07569, DK48831, GM42056, GM15431, DK26657, CA68485, CA77839, and RR00095. J.D.M. is the recipient of a Burroughs Wellcome Fund Clinical Scientist Award in Translational Research.
- Abbreviations:
- ACE
- angiotensin-converting enzyme
- LC
- liquid chromatography
- MS
- mass spectrometry
- ESI
- electrospray ionization
- BK1-5
- Arg-Pro-Pro-Gly-Phe
- TFA
- trifluoroacetic acid
- CID
- collision-induced dissociation
- Received February 3, 2000.
- Accepted March 21, 2000.
- The American Society for Pharmacology and Experimental Therapeutics