We previously demonstrated the substantial capacity for angiotensin (ANG) II formation in the interstitium of the dog heart in vivo. The current study tested the hypothesis that interstitial fluid (ISF) bradykinin (BK) is influenced by ANG II formation. Four microdialysis probes were inserted into the left ventricular myocardium of eight open-chest anesthetized dogs. The probe effluent was collected during four stages in each dog. Probes 1 and 3 sequentially delivered: 1) buffer; 2) ANG I (15 μM); 3) ANG II type 1 receptor antagonist (AT1-ant; irbesartan, 50 μM) or AT2-ant (PD123319, 50 μM); and 4) ANG I + AT1-ant or ANG I + AT2-ant. Probes 2 and 4 used the same protocol, substituting ANG II for ANG I in a concentration (0.5 μM) equivalent to that achieved during ANG I infusion. ISF BK levels increased 15-fold during ANG I (p < 0.001) but not during ANG II infusion. Co-infusion of selective AT1- and AT2-ants or nonselective AT-ant did not block the increase in ISF BK. ISF infusions of ANG I also produced a greater than 400-fold rise in ISF ANG-(1-7) over baseline. ISF infusion of ANG-(1-7) (10 μM) produced a 15-fold increase in ISF BK (p < 0.001). The metabolic machinery exists for the formation of BK and ANG-(1-7) in the cardiac ISF space that is not blocked by an AT receptor antagonist. The differential increase in ISF BK during ANG I and ANG-(1-7) but not during ANG II infusions suggests the possibility of decreased catabolism of ISF BK by an angiotensin-converting enzyme due to active site occupation by ANG I and ANG-(1-7).
Studies in vivo and in vitro have demonstrated that angiotensin-converting enzyme (ACE) efficiently catabolizes kinins and that ACE inhibition leads to increases in circulating or local kinin levels (Margolius, 1996). However, there is mounting evidence that both the renin-angiotensin system (RAS) and kallikrein/kinin system are activated in pathophysiologic states and that bradykinin (BK) protects against the adverse effects of angiotensin (ANG) II in these situations (Dell'Italia and Oparil, 1999). Studies in animals have demonstrated that BK increases across the coronary vascular bed during ischemia and heart failure (Lamontagne et al., 1995; Cheng et al., 1998). In addition, kallikrein and kininogen were released into the coronary sinus effluent from the isolated rat heart in response to acute volume and pressure overload (Nolly et al., 1997), suggesting an increased cardiac kinin-generating activity with acute hemodynamic stress. Indeed, the heart possesses an intact kallikrein/kinin system and, thus, is capable of de novo synthesis of kinins (Nolly et al., 1994). The RAS and kallikrein/kinin system are also functionally linked via ANG peptide breakdown products, in particular, ANG-(1-7), which has been shown to produce vasodilatation via a kinin-mediated mechanism (Brosnihan et al., 1996; Abbas et al., 1997; Ferrario et al., 1997;Oliveira et al., 1999; Roks et al., 1999).
Studies to date had demonstrated that ischemia increases kinin generation across the coronary vascular bed but not within the myocardium itself. Measurement of myocardial BK has been lacking because of the action of kinin-degrading enzymes that are released and activated in tissue assay systems. Furthermore, tissue assay systems cannot account for dynamic fluctuations of BK in vivo, especially in response to hemodynamic stresses or pharmacologic interventions. We have utilized the microdialysis technique to study the mechanisms of ANG II formation in the interstitial fluid (ISF) space of the dog heart in vivo (Dell'Italia et al., 1997). A major advantage of the microdialysis technique in studying biochemical reactions in vivo, compared with assays of tissue extracts in vitro, is that compartmentalization of enzymes is preserved within the tissue. Because the molecular weight cutoff of the dialysis membrane functions as a barrier separating small and large molecules, it can help separate peptides of interest from degrading enzymes. Indeed, a recent study in the cat demonstrated increased ISF BK in response to acute episodes of ischemia and that higher ISF BK levels in the endo- versus epicardial myocardium (Pan et al., 2000).
We have previously demonstrated that ISF ANG II levels of the dog heart are 50-fold higher than in plasma and do not change after combined systemic infusion of ANG I and ACE inhibitor (Dell'Italia et al., 1997), thereby demonstrating a compartmentalization of ANG II formation/degradation in the cardiac interstitium and intravascular space. Furthermore, infusion of ANG I (15 μM) into the cardiac ISF space resulted in a 100-fold increase in ISF ANG II (0.5 μM) levels, which was accomplished by the combined action of ACE and chymase, whereas intravascular ANG II levels were unaffected (Wei et al., 1999). Thus, the metabolic machinery for conversion of ANG I to ANG II is substantial in the cardiac ISF space and separate from the intravascular process. Accordingly, in the current investigation we tested the hypothesis that ISF BK levels are influenced by ANG II formation in the cardiac ISF space of the dog heart in vivo.
Materials and Methods
Eighteen adult mongrel dogs (25 to 30 kg) were utilized in this study. Each dog was screened to rule outEhrlichia canis et platys and Dirofilaria immitis. Dogs underwent general anesthesia using pentobarbital (50 mg/kg initial bolus injection followed by ∼50 mg/h) and had mechanical ventilation with a Harvard ventilator. The heart was exposed and suspended in a pericardial cradle through a median sternotomy. An 8 French sheath was inserted into the right carotid artery and positioned in the ascending aorta. Descending thoracic aortic pressure was continuously monitored using a 4 French microtip Millar catheter (Millar Instruments, Houston, TX) inserted through a femoral artery. A 7 French catheter with multiple side holes was inserted into the coronary sinus through the left jugular vein. A Doppler coronary flow probe (Transonic Systems Inc., Ithaca, NY) was placed around the left anterior descending artery distal to the takeoff of the diagonal branch for coronary flow measurement during all drug infusions. Four microdialysis probes were inserted into the left ventricular (LV) myocardium in the region perfused by the left anterior descending coronary artery at the base, mid, and apical regions of the anterior wall of the LV as previously described (Dell'Italia et al., 1997; Wei et al., 1999). This study was approved by the Animal Resource Program at the University of Alabama at Birmingham.
Each microdialysis probe (Clirans, Terumo Corporation, Tokyo, Japan) contains a semipermeable membrane with a molecular cutoff of 35 kDa and an inner diameter of 200 μm. The membrane is joined to methyl deactivated silica capillary tubing (outside diameter 0.17 mm), and the inflow capillary tube of each probe is connected via the larger deactivated silica tube to a gas-tight glass syringe filled with normal saline and perfused with a precision infusion syringe pump (BAS, West Lafayette, IN) at a rate of 2.5 μl/min. The effluent, or dialysate, is collected from the outflow silica tube in small plastic tubes containing 10 μl of cold ethanol for BK, 10 μl of acetic acid (2.5 M) for ANG-(1-7) and frozen (−80°C) until biochemical analysis.
Cardiac microdialysis is based on the principle that, as the dialysate solution passes through the microdialysis fiber, diffusion occurs between the fluid within the fiber and the ISF surrounding the fiber (Van Wylen et al., 1990). However, at flow rates utilized in the microdialysis experiments in vivo, it is possible that complete equilibration does not occur between normal saline within the fiber and the cardiac ISF in the vicinity of the fiber. Therefore, we performed in vitro experiments to estimate recovery from our microdialysis probes using the method described by Van Wylen et al. (1990). Recovery (determined by comparing the concentration in the dialysis probe effluent with that of the medium, i.e., the percentage of recovery) depends primarily on the perfusion rate through the dialysis fiber. We perfused microdialysis probes (n = 5) with isotonic saline at rates of 0.5, 1.0, and 2.5 μl/min in a beaker containing a bathing medium of isotonic saline (maintained at 37.5°C) before and during the addition of [3H]ANG II (49.2 Ci/mmol; PerkinElmer Life Sciences, Boston, MA) and BK at a concentration of 1 mCi/ml. At a rate equivalent to that employed in the in vivo experiments (2.5 μl/min) the percentage of recovery was 17 ± 2%. This recovery value was used in the final calculation of ISF values and thus represents an estimate of ISF levels since diffusional exchange may differ between a beaker and the beating heart.
Figure1 outlines the experimental protocol. The probe effluent was collected during four experimental stages in each dog. In each dog (n = 8) probes 1 and 3 were sequentially infused with either: 1) buffer; 2) ANG I (15 μM); 3) ANG II type 1 receptor antagonist (AT1-ant, irbesartan, 50 μM; n = 4) or ANG II type 2 receptor antagonist (AT2-ant, PD123319, 50 μM;n = 4); or 4) ANG I + AT1-ant (50 μM; n = 4) or ANG I + AT2-ant (50 μM; n = 4). In probes 2 and 4, the protocol was the same, except that ANG I was replaced by ANG II (0.5 μM). Infusions into probes 1 and 3 and probes 2 and 4 were performed simultaneously in each dog.
We have previously found that exogenous infusion of ANG I (15 μM) into the ISF space of the dog heart results in an increase in ISF ANG II to a level of 0.5 μM, which is approximately 50-fold higher than baseline levels (Wei et al., 1999). Therefore, we infused a similar dose of ANG II (0.5 μM) in the current investigation to achieve approximately equimolar levels of ANG II, either indirectly via ANG I or directly via ANG II infusions into the ISF. Blood samples for measurement of BK and ANG-(1-7) concentrations in plasma were collected from the aorta and coronary sinus after each infusion period. Heart rate and systemic arterial pressure were continuously recorded and the left anterior descending coronary artery blood flow was recorded at 1-min intervals throughout the protocol.
Plasma and ISF Bradykinin and ANG-(1-7) Concentrations.
Plasma and ISF BK concentrations were determined using a standard radioimmunoassay kit (Phoenix Pharmaceuticals, Inc., Belmont, View, CA). Figure 2 demonstrates that there is 100% cross-reactivity between BK and Lys-BK, whereas there is no cross-reactivity between BK and ANG peptides including ANG I, ANG II, ANG III, and ANG-(1-7) in our assay procedure. For BK collections, ISF was collected immediately in an iced Eppendorf tube containing 95% ethanol. Plasma and ISF ANG-(1-7) concentrations were determined from blood collected in a cocktail of protease inhibitors as previously described (Senanayake et al., 1994; Nakamoto et al., 1995). Plasma was extracted using Sep-Pak columns (Waters Associates, Milford, MA). High-pressure liquid chromatography separation was performed on an LKB 2150 gradient system equipped with a Nova Pak C18 column (3.9 × 150 mm, Waters Associates). Angiotensin peptides were quantified by radioimmunoassay from the fractions obtained in a single high-pressure liquid chromatography run. Recoveries of radiolabeled angiotensin added to the plasma sample and followed through the extraction were 92% (n = 23). Plasma samples were corrected for recovery. For ISF ANG-(1-7) assays, a Tris buffer with 0.1% bovine serum albumin was used. ANG-(1-7) was measured using the antibody previously described (Senanayake et al., 1994; Nakamoto et al., 1995). The minimum detectable level of the assay was 2.5 pg per tube for ANG-(1-7). The intra-assay coefficient of variation averaged 8% for ANG-(1-7).
All data are presented as mean ± standard error. Analysis of variance with Newman-Keuls post hoc comparison was used to assess differences in hemodynamics, BK, and ANG-(1-7) levels at baseline and during the three infusions (Fig. 1). Ap value less than 0.05 was required for significance.
Infusions of ANG I and ANG II into the ISF did not change mean arterial pressure (from 115 ± 2 mm Hg to 122 ± 3 mm Hg) or heart rate (from 125 ± 2 bpm to 124 ± 2 bpm). In addition, left anterior descending coronary artery blood flow did not change from a baseline value of 20.1 ± 0.6 ml/min during the four phases of the protocol.
ISF and Plasma Bradykinin and ANG-(1-7) Concentrations.
Infusion of ANG I into the ISF resulted in a 15-fold increase in ISF BK (p < 0.01) that was not blocked by administration of either an AT1 or AT2antagonist (n = 8 dogs) (Fig.3, A and C). In contrast, infusion of either ANG II, ANG II combined with the AT1-antagonist, or ANG II in combination with the AT2-antagonist had no effect on ISF BK (Fig. 3, B and D). Therefore, we tested whether the increase in BK during ISF infusion of ANG I could be blocked by the nonspecific AT antagonist, sarcosine isoleucine ANG II (50 μM), in a subset of two dogs (four probes per dog). The increase in ANG I-mediated BK release was not blocked by sarcosine isoleucine ANG II.
The possibility that the effects of ANG I on ISF BK were associated with increases in ANG-(1-7) generation was evaluated in additional experiments. Four dogs (four microdialysis probes per dog) received ANG I and ANG II plus AT1-ant (n = 2) or AT2-ant (n = 2) as outlined in Fig. 2 and dialysate was collected for ANG-(1-7). Figure4 shows that the infusion of ANG I produced a 400-fold increase in ISF levels of ANG-(1-7). In contrast, infusion of ANG II was associated with a smaller 60-fold rise in ISF ANG-(1-7). The increases in ISF ANG-(1-7) resulting from the infusion of either ANG I or ANG II were not prevented by coadministration of either AT1 or AT2antagonists.
Since there were substantial increases in ANG-(1-7) during ANG I infusions, we next tested the hypothesis that infusion of ANG-(1-7) into the ISF was associated with an increase in ISF BK. Accordingly, ISF in two dogs (three probes per dog) was perfused with: 1) buffer and 2) ANG-(1-7) at doses of 0.1, 1.0, and 10 μM in each probe, respectively. There was a dose-dependent increase in ISF BK levels in response to ISF infusion of ANG-(1-7) (Fig.5A). In an additional two dogs (three probes per dog), we infused AT2-antagonist (PD123319, 50 μM) alone and in combination with ANG-(1-7). These experiments demonstrated that ISF infusion of AT2-antagonist did not affect ISF BK during ANG-(1-7) infusions (Fig. 5B). Finally, plasma BK levels (Fig.6) and ANG-(1-7) levels (data not shown) from the aorta and coronary sinus did not change throughout all phases of the protocols.
The current study demonstrates that the metabolic machinery exists for the formation of BK and ANG-(1-7) during infusion of ANG I into the cardiac ISF space. Selective AT1/AT2 antagonists or a nonselective AT receptor antagonist did not block this process. ISF BK increased during ANG I and ANG-(1-7) and not during ANG II infusions. These results suggest that ANG peptides may modulate ISF BK by active site occupation of ACE in ANG I conversion or in ANG-(1-7) catabolism.
The increase in ISF BK during ANG I infusion cannot be attributed to changes in cardiac hemodynamics or regional ischemia because blood pressure and coronary blood flow did not change throughout the experiments. A potential for activation of tissue kallikrein due to the presence of intravascular catheters and intramyocardial microdialysis probes could produce an increase in ISF BK. However, this possibility may be excluded because the rise in ISF BK occurred only in response to the infusion of ANG I but not ANG II. In addition, ISF BK levels returned to baseline levels after discontinuation of ISF infusion of ANG I. Thus, the transient and differential changes in ISF BK provide strong evidence against an artifact due to the presence of dialysis probes in the myocardial tissue. Other potential mechanisms by which kinins can increase in the ISF include changes in clearance from the ISF and/or uptake from the circulation. Both coronary sinus and aortic plasma BK levels were unchanged over the entire protocol, suggesting that the increase in ISF BK was not a product of increased uptake from the circulation or decreased clearance from the cardiac interstitium.
Evidence for a direct relationship between RAS activation and BK formation has been demonstrated in the systemic vasculature and in the kidney via activation of the AT2 receptor. Studies in spontaneously hypertensive rats and AT2 transgenic mice have shown that the AT2 receptor mediates vasodilatation by a kinin-mediated mechanism (Gohlke et al., 1998; Tsutsumi et al., 1999). Utilizing renal microdialysis in the conscious transgenic mouse, low sodium intake resulted in greater ISF cGMP and BK kidney levels in wild-type versus AT2 knockout mice (Siragy et al., 1998). In the rat after experimentally induced myocardial infarction, the beneficial effects of AT1antagonist were blocked by concomitant treatment with AT2 receptor antagonist or a BK2 receptor antagonist, suggesting that part of the therapeutic effects of AT1 antagonist on LV remodeling was mediated by a kinin-AT2 receptor mechanism (Liu et al., 1997). However, a major difference in these and our studies is the prior activation of the RAS by sodium deprivation or myocardial infarction, which may in turn have important differential effects on expression and function of the AT2receptor compared with the normal state in our dogs. Accordingly, ISF infusion of ANG I into the cardiac interstitium of our dogs led to an increase in ISF BK, but ANG II did not. Furthermore, the increase in ISF BK during ANG I infusion was not blocked by AT receptor antagonist. Since ANG I itself has no direct physiologic receptor, other indirect mechanisms may explain the observed increase in ISF BK.
There was a substantial capacity for ANG-(1-7) formation in the ISF space of the heart. ANG I and ANG II are processed into ANG-(1-7) by a diversity of endopeptidases (Welches et al., 1993). These enzymes, combined with ACE, represent the major degradative pathways of kinins in the heart (Welches et al., 1993; Blais et al., 1997; Kokkonen et al., 1999). In addition to the hydrolysis of BK and ANG I, there is evidence that ACE also participates in the metabolism of ANG-(1-7) with an efficiency greater than that for ANG I but equal to that of BK (Chappell et al., 1998). Moreover, ANG-(1-7) is an endogenous inhibitor of the C-terminal active site of ACE with an IC50of 0.65 μM (Li et al., 1997). This concentration is similar to the ISF ANG-(1-7) levels achieved during exogenous ANG I infusion (0.2–0.3 μM), but was 5-fold lower than the ANG-(1-7) concentrations achieved during ANG II infusions. Indeed we demonstrated a progressive increase in ISF BK with increasing ISF infusion concentrations of ANG-(1-7) (Fig. 5). Taken together, ANG I and ANG-(1-7) are substrates for ACE; the lower ANG-(1-7) concentrations produced by ANG II infusions in the absence of ANG I could account for the failure of ANG II infusion alone to increase ISF BK.
In the intravascular space ACE is plentiful and bound to endothelial cells with its catalytic sites exposed to the vessel lumen (Johnston et al., 1994). Our previous work supports the presence of ACE in the cardiac interstituim of the dog (Wei et al., 1999). However, ACE activity in the dog myocardium is extremely low when compared with the kidney (150-fold higher) and lung (40-fold higher) (Su et al., 1999). The pharmacologic doses of ANG I and ANG-(1-7) utilized in the current investigation could indeed saturate ACE in the cardiac interstitium surrounding the microdialysis probes. We have previously shown that captopril in doses of 0.1, 1.0, and 10 mM produced similar inhibition of ANG II formation (83%) (Wei et al., 1999). In a subset of two dogs, we found a 2.5-fold increase of BK with captopril (2.5 mM) and a further 10-fold increase in ISF BK during infusion of ANG I and captopril, suggesting an additional interaction that may involve ANG I and endopeptidases in the heart.
Taken together, ANG I binding to the active site of ACE and neutral endopeptidases, combined with the formation of large amounts of ANG-(1-7) with its inhibitory effects on ACE, could provide a mechanism for the increase of ISF BK. Interestingly, we have previously demonstrated high concentrations of cardiac ISF ANG I and II that could reach even higher levels during RAS blockade and during pathophysiologic conditions (Dell'Italia et al., 1997; Wei et al., 1999). Whether ANG peptide concentrations and their interaction with ACE and neutral endopeptidases play an important role in regulating BK levels in the cardiac interstitium requires further investigation.
We thank Joan Durand (University of Alabama at Birmingham) and Margaret B. King (Wake Forest University Baptist Medical Center) for their performance of the radioimmunoassays and Pamala Gibson for her expert secretarial skills in preparing the manuscript.
This study was supported by the Office of Research and Development, Medical Service, Department of Veterans Affairs (L.J.D.), National Heart, Lung and Blood Institute Grants RO1 HL54816 and HL6707 (L.J.D.), Mechanisms of Hypertension and Cardiovascular Diseases (2T32HL07457-16A1), and Fellowship 9820134V (C.C.W.) from the Southeast affiliate of the American Heart Association.
- interstitial fluid
- ANG type
- ANG II type 1 receptor antagonist
- AT2 ant
- ANG II type 2 receptor antagonist
- angiotensin-converting enzyme
- renin-angiotensin system
- left ventricular or left ventricle
- beats per minute
- Received August 22, 2001.
- Accepted October 10, 2001.
- U.S. Government