Functional and Biochemical Analysis of Angiotensin II–Forming Pathways in the Human Heart
Blockade of the renin-angiotensin system by inhibition of angiotensin-converting enzyme (ACE) is beneficial for the treatment of hypertension and congestive heart failure. However, it is unclear how complete the blockade by ACE inhibitors is and if there is continuing angiotensin II (Ang II) formation during chronic treatment with ACE inhibitors. Indeed chymase, a serine protease, which is able to form angiotensin II from angiotensin I (Ang I) and cannot be blocked by ACE inhibitors, has been shown to be present in human heart. The goal of the present study was to evaluate the extent of renin-angiotensin system blockade and the Ang II–forming pathways in cardiac tissue of patients chronically treated with ACE inhibitors or in patients without ACE inhibition therapy. Our studies indicate an incomplete ACE inhibition in human heart tissue after chronic ACE inhibitor therapy. Moreover, ACE contributes only a small portion to the total Ang I conversion, as shown in biochemical studies in ventricular and coronary homogenates or functionally as Ang I contractions in isolated rings of coronary arteries. A serine protease was responsible for the majority of Ang II production in both the membrane preparation and Ang I–induced contractions of isolated coronary arteries. In humans, the serine protease pathway is likely to play an important role in cardiac Ang II formation. Thus, drugs such as renin inhibitors and Ang II receptor blockers might be able to induce a more complete blockade of the renin-angiotensin system, providing a more efficacious therapy.
It has been well documented that the RAS plays an important role in cardiovascular homeostasis, including blood pressure, mineral balance, and tissue remodeling.1 2 Blockade of the RAS with either ACE inhibitors or angiotensin receptor blockers is a beneficial therapeutic approach for the treatment of hypertension, heart failure, and left ventricular dysfunction after myocardial infarction.3 4 5 6 7
However, studies have shown that complete blockade of the RAS is not achieved with ACE inhibitors. It has been reported that plasma Ang II levels return to normal despite ACE inhibitor therapy.8 9 With equihypotensive doses, the renin inhibitor remikiren had more extensive renal effects than the ACE inhibitors lisinopril and losartan in sodium-depleted guinea pigs.10 These results seem to be confirmed in human volunteers under sodium restriction, in whom renin inhibitors showed larger renal effects over ACE inhibitors.11 Furthermore, an additive blood pressure–lowering effect was reported in hypertensive patients when an Ang II receptor blocker was added to an ACE inhibitor.12
One of the reasons that ACE inhibitors may not reach a total blockade of the RAS may be the presence of chymase, a serine protease that is able to form Ang II from Ang I but is not blocked by ACE inhibitors.13 Human heart chymase has been isolated, cloned, and expressed.14 Kinetic studies showed that chymase is an Ang II–forming enzyme as efficient as ACE.15 Chymase is distributed in human tissue, such as blood vessels, lung, kidney, and liver.16 Chymase-like immunoreactivity is localized in the cardiac interstitium, and several cell types, including cardiac mast cells and endothelial cells, are sites of chymase biosynthesis and storage.17 However, chymase, in contrast to ACE, is not present in plasma.13
Up to now, there have been conflicting results on the physiological importance of non–ACE-dependent Ang II formation. One of the first studies was in 1979; in that study, Cornish et al18 found that the angiotensin vasoconstriction of the blood vessels from a hamster cheek pouch was only partially inhibited by ACE inhibitors but was completely inhibited by Ang II antiserum or by an Ang II receptor antagonist. In biochemical studies using human left ventricular membranes, it was shown that ACE inhibitors could block only 10% to 20% of the Ang I conversion.13 19 The remaining activity was blocked by a serine protease inhibitor, pointing to chymase as an ACE-bypassing enzyme. In pharmacological studies, ACE inhibitors could not block >40% of the contraction in human gastroepiploic arteries or in human detrusor muscle induced by Ang I.20 21 In vivo, the exercise-induced increase of plasma Ang II could not be prevented by ACE inhibition but was reduced by a serine protease inhibitor.22 In contrast with these results, it was recently found that ACE inhibitors could reduce the Ang I conversion in membranes from human left ventricle by 90%.23 In the same study, it was shown that ACE inhibition could also decrease Ang I conversion in vivo across the human heart by 90%.
The goal of the present study was to evaluate the extent of ACE inhibition in cardiac tissue of patients chronically treated with ACE inhibitors. Furthermore, the contribution of ACE and chymase in the Ang I conversion capacity was evaluated in the left ventricular membranes and in human coronary arteries. For the purpose of clarifying conflicting published results, different possible technical artifacts were investigated: dependence on substrate and protein concentration, membrane preparation techniques, and the effects of iodination of Ang I.
The present results clearly show that the contrasting previously published results are due to different tissue preparations. Under our more physiological conditions, a serine protease is the major enzymatic pathway in the formation of Ang II in human heart.
Materials and Methods
Human hearts were obtained from 14 male patients undergoing heart transplantation. Indication for transplantation was cardiomyopathy in 8 patients (mean age, 54±3 years), coronary artery disease in 5 patients (mean age, 58±3 years), and congenital heart disease in 1 patient (age, 25 years). There were also donor hearts obtained from 3 male patients (mean age, 22±3 years) that were not able to be used for transplantation. Most of the patients undergoing cardiac transplantation were under treatment with diuretics and digoxin; all were under ACE inhibitor treatment for a period of at least 3 months before transplantation, except the patient with coronary heart disease: 6 patients received captopril (dose range, from 25 to 75 mg daily; mean daily dose, 44.7±20.3 mg), 6 patients received enalapril (dose range, from 15 to 40 mg daily; mean daily dose, 27.5±12.5 mg), and 1 patient received ramipril (10 mg).
For the functional studies, coronary arteries from 4 patients with cardiomyopathy and 4 patients with coronary artery disease were used.
The hearts were transported in ice-cooled Krebs-Henseleit solution and reached our laboratories within 2 to 10 hours after harvesting. All the arteries were dissected free. The arteries with the least coronary artery disease were used for the functional studies and tested immediately. For the biochemical studies, the heart tissue and remaining coronary arteries were frozen and stored at −80°C before analysis.
Processing of Heart Tissue and Coronary Arteries
Membranes were prepared at 4°C in a manner similar to one previously described,13 but care was taken to ensure minimal handling in order not to lose ACE and/or chymase. Pieces of left ventricular myocardium (mean±SD, 1.13±0.13 g; n=14) or coronary arteries (mean±SD, 1.01 g±.02; n=13) were dissected, minced, and homogenized in 2.5 mL of 50 mmol/L potassium phosphate, pH 7.4, using a Polytron homogenizer (Heidolph GmbH). The homogenates were centrifuged at 40 000g for 30 minutes at 4°C. The supernatants were discarded, and the pellets were washed twice with 0.1 mol/L sodium phosphate, pH 7.4, including 150 mmol/L sodium chloride (buffer A). Finally, the pellets were taken up in 4 mL of buffer A and divided in two aliquots, the protein concentration was measured according to the method of Lowry et al,24 and the aliquots were frozen at −80°C. A total of 11.0±1.9 mg protein/g wet wt was extracted. For the coronary arteries, a total of 4.4±0.7 mg protein/g wet wt was extracted.
Solubilization of Cardiac Ventricular Membranes
Solubilization and dialysis were achieved under conditions described by Zisman et al.23 Aliquots of the crude homogenate were centrifuged at 25 000g for 30 minutes at 4°C. The pellets were resuspended in solubilization buffer consisting of 0.01 mol/L HEPES buffer, pH 7.5, containing 0.3 mol/L KCl and 0.6% Triton X-100 and incubated for 4 hours at 4°C. The samples were centrifuged for 1 hour at 25 000g and 4°C and finally dialyzed at 4°C against 0.01 mol/L HEPES buffer, pH 8.1, containing 0.3 mol/L KCl, 0.02% Triton X-100, and 20 μmol/L zinc sulfate.
Measurement of Cardiac ACE Activity
Heart tissue ACE was estimated by measuring the release of [14C]hippuric acid from the synthetic substrate [14C]Hip-His-Leu according to the method of Cushman and Cheung.25 The assay was conducted in 50 mmol/L HEPES buffer (pH 8.3) containing 300 mmol/L sodium chloride and 20 μmol/L ZnSO4. From the homogenate, 20 μL was added to 20 μL of [14C]Hip-His-Leu (30 nCi per assay tube) and 50 μL of assay buffer and incubated at 37°C for 1 hour. The reaction was stopped by the addition of 100 μL of 0.2 mol/L sulfuric acid, and the hippuric acid was extracted by 1 mL of ethyl acetate. The radioactivity of 500 μL of the extract was measured in a beta scintillation counter. The homogenates were assayed undiluted and resulted in a 5% consumption of the substrate [14C]Hip-His-Leu.
In order to evaluate the percentage of ACE activity inhibited by the ACE inhibitor therapy, the ACE inhibitor was removed by exhaustive dialysis. Aliquots (300 μL) of the homogenates were dialyzed for 5 days at 4°C against 2 L of 2 mmol/L Tris-HCl buffer, pH 8.0, including 100 mmol/L NaCl and 10 μmol/L EDTA. The buffer was changed after 1 and 3 days. As positive control for the completion of the dialysis, a separate aliquot (300 μL), spiked with the ACE inhibitor cilazaprilat (1 μmol/L), was dialyzed.
Measurement of Total Cardiac Ang II–Forming Activities
The conversion of both Ang I and [125I]Ang I to Ang II and [125I]Ang II, respectively, was estimated in the absence (maximal generation) and presence (fractional conversion) of inhibitors of ACE or chymase. The procedure was similar to that described previously.13 The assay was run at physiological pH (pH 7.4) and salt concentration (150 mmol/L NaCl) in 0.1 mol/L sodium phosphate. The prototypical incubation mix consisted of (1) 50 μL of heart homogenates diluted 1:20 in assay buffer (7.7±4.8 μg protein, n=14) or coronary artery homogenate diluted 1:32 in assay buffer (4.4±.7 μg protein, n=13), (2) 50 μL of labeled [125I]Ang I (≈550 000 cpm; final concentration, 125 fmol or 1.0 nmol/L) or cold Ang I (final concentration, 1 nmol/L), and (3) 25 μL of assay buffer containing the ACE inhibitor captopril (final concentration, 100 μmol/L) or the chymase inhibitors chymostatin (final concentration, 100 μmol/L) or SBTI (final concentration, 100 μg/mL). After 30 minutes of incubation at 37°C, the reaction was stopped by the addition of 1 mL ice-cold ethanol. To evaluate the conversion of [125I]Ang I, 20 μL of the ethanol solution was added to 500 μL HPLC sample buffer (0.1 mol/L imidazole, 0.12 mmol/L NaN3, and 1% BSA, pH 6.5); 400 μL of this mixture was injected into the HPLC. To evaluate the conversion of cold Ang I, the total ethanol mixture was evaporated, the residue was taken up in 400 μL RIA assay buffer, and Ang II was measured as described for the HPLC fractions.
In order to evaluate the influence of substrate concentration, [125I]Ang I was combined with 1 μmol/L of cold Ang I. Reduction in substrate concentration was reached by simply adding less counts to the incubation mix but adding increasing volumes of the incubates to the imidazole buffer, allowing for enough measurable counts in the fractions.
RIA Measurement of Ang II
The polyclonal antibody used for the RIA (Ang II–AS 923) was raised in rabbits. It was very sensitive, with an IC50 value of 5.5±0.31 fmol per assay tube (n=8), and extremely specific for Ang II, with essentially no cross-reactivities against other angiotensin peptides and fragments: Ang II, 100%; Ang I, 0.37±0.10%; Ang III, <0.02%; Ang (2-10), <0.02%; Ang (3-8), <0.02%; Ang (4-8), <0.02%; and Ang (5-8), <0.02%.
To estimate the Ang II generation in the incubates, they were evaporated, taken up directly in 400 μL of the RIA buffer (0.1 mol/L imidazole, pH 6.5, including 1% human serum albumin), and put for 10 minutes in the ultrasonification bath for dissolution. Finally, antiserum in RIA buffer (50 μL) was added, and the sample was preincubated at 4°C overnight before [125I]Ang II in RIA buffer (5000 cpm in 50 μL) was added and incubated for another 24 hours at 4°C. The incubation was stopped by adding 1 mL of charcoal suspension (14 mg/mL charcoal/1.4 mg/mL dextran T70) in RIA buffer without human serum albumin and centrifuged (2000g), and the supernatant was counted.
HPLC Separation of Angiotensins
The HPLC technique, using the reverse-phase resin Nucleosil C18 (7 μm) and a column (100×4 mm), was similar to that described by Nussberger et al.26 Elution was achieved by a linear gradient of methanol from 25% to 50% for 15 minutes in 0.085% phosphoric acid with a flow of 1 mL/min. Fractions of 0.5 mL were collected. Under these conditions, the peptides could be well separated with the following retention times: [125I]fragments, 2 minutes; Ang II, 7 minutes; [125I]Ang II, 9 minutes; Ang I, 10 minutes; and [125I]Ang I, 12 minutes. The recovery for the chromatographed [125I]Ang I and [125I]Ang II was 94.7±2.1% (n=4) and 87.5±2.0% (n=4), respectively. For comparison, the recoveries for the cold peptides (as measured by RIA) were similar: Ang I, 98.2±3.3% (n=4); Ang II, 86.8±4.9% (n=4).
Functional Studies Using Coronary Arteries
Midsections of LAD or CX were harvested, freed from connective tissue, and cut into rings. Care was taken not to disrupt the endothelium. Rings (5 mm long) from midsections of the LAD or CX were suspended under 2-g resting force in 10-mL organ chambers containing gassed (95% O2/5% CO2) and warmed (37°C) Krebs-Henseleit solution of the following composition (mmol/L): NaCl 115, KCl 4.7, MgSO4 1.2, KHPO4 1.5, NaHCO3 25, CaCl2 2.5, and glucose 10.0. Rings were connected to force transducers (Swena), and isometric force was recorded (Hellig 112 037, Hellig AG). The artery sections were equilibrated for ≈1.5 hours or until stable. After the equilibration period, rings were contracted using 0.1 mmol/L ACh. ACh is a known contractor of human coronary arteries in the concentration used.27 After a maximal contraction was obtained, the baths were washed, and the tension was adjusted until a stable 2-g resting force was reached. The arteries were allowed to equilibrate for a minimum of 45 minutes before another ACh contraction was made. Then, ACh was washed out, and the rings were again equilibrated for 45 minutes before the addition of Ang I or the chymase-specific substrate.28
For each coronary artery, the evaluation was performed with 12 rings in parallel. After the contraction with ACh and equilibration, the rings were preincubated with Krebs' buffer (control, n=4), cilazaprilat (100 μmol/L, n=2), chymostatin (100 μmol/L, n=2), chymostatin and cilazaprilat together (n=2), and losartan (1 μmol/L, n=2). The baths were allowed to incubate for 20 minutes, and then 1 μmol/L of Ang I or SUB was added.
Chymostatin (N-[-Na-carbonyl-Cpd-X-Phe-al]-Phe), human Ang I (hypertensin I, Asp-Arg-Val-Tyr-Ile-His-Pro-Phe-His-Leu), and human Ang II (Asp-Arg-Val-Tyr-Ile-His-Pro-Phe) were from Sigma. ACh chloride was from Fluka Chemie AG. Cilazaprilat, polyclonal antibody (Ang II–AS 923), losartan, and SUB Ang I-Pro-d-Ala were from F. Hoffmann-La Roche Ltd. Captopril was from ICN. Synthetic ACE substrate [14C]Hip-His-Leu and Ang I were from NEN.
All results are expressed as mean±SEM or, where noted, mean±SD. Means were compared by ANOVA. Values of P<.05 were considered to be statistically significant.
ACE Activity in Heart Membranes: Dissociation of ACE Inhibitors With Dialysis
Because the hearts reached our laboratory up to 10 hours after harvesting and all patients had different medication regimes, the dissociation of ACE inhibitors was evaluated. Left ventricular material was homogenized and divided into two aliquots. The second aliquot was kept for the other studies, and the first aliquot was solubilized, according to the method of Zisman et al.23 The solubilized material was separated into four parts, and saline (control), cilazaprilat, captopril, or enalaprilat was added. The samples were dialyzed for 120 hours, with sampling at intervals (0, 2, 4, 8, 28, 72, and 120 hours) for ACE activity measurements (Fig 1⇓). The activity in the control sample was stable (92, 85, 95, 90, 102, 101, 81 pmol/mL homogenate per minute), the captopril disassociated fairly rapidly (15, 50, 109, 101, 89, 82, 88 pmol/mL homogenate per minute), and ACE activity was back to control levels within 4 hours. For cilazaprilat and enalaprilat, it took the full 120 hours to reach control ACE activity levels (3, 0, 8, 1, 4, 56, and 92 pmol/mL homogenate per minute and 3, 5, 4, 1, 22, 63, and 91 pmol/mL homogenate per minute, respectively). It should be stressed that these were extremely rigorous conditions, including homogenization and solubilization of the samples. In the samples without ACE inhibition, there was no increase in ACE activity levels during the 120 hours, showing that the solubilization processes had disassociated the ACE inhibitors given during patient therapy.
ACE Inhibition in Heart Membranes: Effects of Chronic Treatment with Captopril and Enalapril
The nonsolubilized homogenates were separated into two groups. One group was from tissue obtained from patients that had undergone captopril therapy. The second group was from tissue obtained from patients that had undergone enalapril therapy. Each homogenate was divided in half: one half remained as the control (nonspiked), and the other was treated with 1 μmol/L cilazaprilat (spiked). All the homogenates were exhaustively dialyzed for 5 days at 4°C. ACE activity was measured in the conventional way by release of [14C]hippuric acid from the synthetic substrate [14C]Hip-His-Leu at pH 8.3, which was well measurable and could be totally blocked by 1 μmol/L cilazaprilat (Fig 2⇓). The ACE activities in the captopril group increased from 18.8±4.6 to 45.2±5.0 pmol/mg wet wt per minute (n=5, P<.001) and from 0.04±2.21 to 43.5±11.4 pmol/mg wet wt per minute (n=5, P<.001) in the non–cilazaprilat-spiked and the spiked aliquots, respectively. In the enalapril group, the ACE activities increased from 31.9±5.9 to 49.4±5.4 pmol/mg wet wt per minute (n=5, P<.001) and from 0.24±0.94 to 40.2±6.5 pmol/mg wet wt per minute (n=5, P<.001) in the non–cilazaprilat-spiked and the spiked aliquots, respectively. Thus, ACE activity in all samples reached the same levels after dialysis.
The similar ACE activities of the nonspiked versus spiked aliquots after dialysis suggests that all the added cilazaprilat and the ACE inhibitors from the patient therapy had been removed; dialysis was complete. These data indicate that after transport and deep freezing, there is 58% ACE inhibition in the captopril group and 35% ACE inhibition in the enalapril group (comparison of the “nonspiked” before and after dialysis). This suggests that the left ventricular enzyme was not totally blocked by chronic ACE inhibitor pharmacotherapy.
Protein Dependence on the Processing of [125I]Ang I
The conversion of [125I]Ang I to [125I]Ang II and fragments was assessed using various protein concentrations in one heart homogenate (data not shown). The protein concentrations used were 5, 10, 20, 40, 80, and 160 μg per assay tube. Each of the above concentrations of protein was incubated with 1 nmol/L [125I]Ang I. Protein concentrations >10 μg per assay tube led to substrate depletion, which is considered prohibitive for enzyme kinetics, and concentrations <5 μg per assay tube did not produce enough Ang II to be adequately analyzed. Thus, all further experiments were performed at a protein concentration of ≈5 μg per assay tube, which consumed 40% of the substrate to create 25% [125I]Ang II and 10% 125I fragments.
Effects of Protease Inhibitors on the Processing of [125I]Ang I in Left Ventricular Tissue and Coronary Arteries
Aliquots of left ventricular tissue or coronary arteries were incubated with 1 nmol/L [125I]Ang I in the presence or absence of the ACE inhibitor captopril (100 μmol/L) and the serine protease inhibitors chymostatin (100 μmol/L) and SBTI (100 μg/mL) (Fig 3⇓). The total [125I]Ang II formation in the absence of any inhibitor for the ventricular tissue that had prior ACE inhibitor therapy was 1.5±0.18 fmol/mg wet wt per minute (n=10) and could not be significantly inhibited by captopril (1.6±0.22 fmol/mg wet wt per minute) in this conversion assay. Only the two serine protease inhibitors significantly inhibited the [125I]Ang II formation to (0.2±0.02 fmol/mg wet wt per minute) for chymostatin (P<.001) and to (0.2±0.03 fmol/mg wet wt per minute) for SBTI (P<.001). Likewise, the tissue from four hearts that had no history of ACE inhibitor therapy gave similar results, with 1.6±0.44 fmol/mg wet wt per minute for control, 1.4±0.34 fmol/mg wet wt per minute for captopril, 0.2±0.05 fmol/mg wet wt per minute for chymostatin, and 0.3±0.05 fmol/mg wet wt per minute for SBTI. The coronary arteries from 11 patients with previous ACE inhibitor therapy and 2 patients without prior therapy were also analyzed. The values for the ACE-inhibited groups were 2.18±0.28 fmol/mg wet wt per minute for control, 2.29±0.43 fmol/mg wet wt per minute for captopril, 0.14±0.02 fmol/mg wet wt per minute for chymostatin (P<.001), and 0.13±0.02 fmol/mg wet wt per minute for SBTI (P<.001); the values for the two coronary arteries that had no ACE inhibitor therapy were 2.84±1.34 fmol/mg wet wt per minute for control, 2.28±1.72 fmol/mg wet wt per minute for captopril, 0.11±0.05 fmol/mg wet wt per minute for chymostatin, and 0.17±0.04 fmol/mg wet wt per minute for SBTI. Thus, the contribution of [125I]Ang I conversion by ACE was only a minor one.
The formation of 125I fragments was not significantly different in any of the four conditions (control, captopril, chymostatin, and SBTI) and for all the groups in the ventricular tissue and coronary groups (data not shown). Therefore, neither ACE nor chymase is involved in the fragment formation. The fragments are most likely formed from the substrate [125I]Ang I and not the product [125I]Ang II, since the decreased [125I]Ang II formation by the serine protease inhibitors does not alter the fragment formation.
Correlation Between the Conversion of Cold and Iodinated Ang I
The potential difference in the processing of Ang I and [125I]Ang I was investigated. Productions of cold and labeled Ang II were similar in 10 heart homogenates (Fig 4⇓) and were well correlated (correlation factor, .787). This confirms that the use of [125I]Ang I in our assay system is valid. It also confirms that the measured radioactivity that coelutes from HPLC at the same retention time as commercially available [125I]Ang II corresponds indeed to the generated [125I]Ang II.
Effects of Different Substrate Concentrations on the Conversion of [125I]Ang I by ACE and Chymase
The labeled substrate was incubated at different concentrations of [125I]Ang II (1000, 1, and 0.01 nmol/L) with membranes from four failing hearts in the presence or absence of vehicle, captopril (100 μmol/L), chymostatin (100 μmol/L), and SBTI (100 μg/mL) (Fig 5⇓). The percent conversions of [125I]Ang I using the concentrations 1000, 1, and 0.01 nmol/L, respectively, were 26.0±3.0%, 28.3±3.9%, and 27.3±3.8% (control), 24.7±3.9%, 23.9±3.1%, and 22.7±2.9% (captopril), 2.3±0.5%, 4.9±0.7%, and 3.3±0.8% (chymostatin), and 4.0±0.9%, 7.0±0.8%, and 3.9±0.5% (SBTI). Thus, the inhibition pattern of [125I]Ang II generation is not influenced by the variation in the concentration of [125I]Ang I over five orders of magnitude. This is important, since many published biochemical and pharmacological studies have used various Ang I concentrations far from the physiological level of 10 pmol/L.
Effects of Solubilization and Dialysis on the [125I]Ang II–Generating Activity From Patients in Heart Failure
Since there have been several published methodologies for the preparation of ventricular membranes, the effects of different membrane preparations on [125I]Ang I processing were analyzed (Table⇓). Aliquots of the previously analyzed homogenates from 10 failing hearts were dialyzed without or with previous solubilization, and the [125I]Ang II generation was compared with the one by the unprocessed membranes. Solubilization was achieved using 0.6% Triton X-100, as described by Zisman et al.23 The total generation of [125I]Ang II without the addition of inhibitors dropped from 198±24 fmol/mg per minute to 88±15 fmol/mg per minute after dialysis and then to 55±6 fmol/mg per minute after a combined dialysis/solubilization. The reason was the major loss of chymase during handling of the membranes. The activity measured in presence of captopril, representing the non-ACE conversion activity, which is presumably chymase, dropped from 203±25 to 60±11 fmol/mg per minute after dialysis and then to 11±1 fmol/mg per minute after a combined solubilization/dialysis. In contrast, the nonchymase activity, assumed to be ACE (chymostatin column), dropped also from 21±4 to 11±1 fmol/mg per minute but was stabilized by solubilization, which rendered 25±2 fmol/mg per minute. The fluctuation in the activity of ACE (chymostatin column) is due to combined loss of enzymatic activity during handling and gain by the disassociation of the ACE inhibitor from the enzyme during solubilization.
Effects of Protease Inhibitors on the Ang I–Induced Contraction of Coronary Arteries
The pharmacological response to the administration of Ang I and the SUB [Pro11, d-Ala12]Ang I at 1 μmol/L was investigated in isolated coronary arteries. Fig 6A⇓ represents an original tracing showing the Ang I–induced contraction and inhibitions by the AT1 blocker losartan (1 μmol/L), the ACE inhibitor cilazaprilat (100 μmol/L), and the serine protease inhibitor chymostatin (100 μmol/L). In this tracing, losartan and chymostatin completely blocked the pressor response to Ang I, whereas cilazaprilat was ineffective. To confirm that chymase was in fact the Ang II–generating enzyme, the arteries were shown to be contracted by SUB. This contraction could be completely inhibited by chymostatin (Fig 6B⇓).
Coronary arteries from 8 patients in heart failure chronically treated with ACE inhibitors were contracted using Ang I in the absence and presence of RAS blockers (Fig 7⇓). Ang I at 1 μmol/L induced an average contraction that was of a magnitude similar to the one induced by 10 μmol/L ACh (4.1±0.58 versus 4.1±0.31 g). Cilazaprilat did not block this response (4.03±0.78 g), but chymostatin blocked it by 78% (0.9±0.26 g); the combination of cilazaprilat and chymostatin, by 97% (0.12±0.12 g); and losartan, by 100%.
The present results demonstrate that the blockade of RAS in human cardiac tissue is incomplete by chronic ACE inhibitor pharmacotherapy and that non–ACE-dependent Ang I conversion plays an important role in the generation of Ang II. In left ventricular heart and coronary artery homogenates, a serine protease, probably chymase, is the major enzyme that converts Ang I to Ang II. Moreover, this chymase-specific conversion is also shown to play a major functional role in the Ang I–induced contraction of coronary arteries in the organ bath.
ACE is present in plasma and many tissues, including the heart.29 The beneficial effects of ACE inhibitors in cardiac diseases suggest a crucial role of ACE in cardiac pathology.30 Indeed, it was found that the ACE gene is linked to ischemic heart attacks.31 Specifically, there is good evidence that the local intracardiac RAS mediates the Ang-induced effects in normal and failing hearts and that the ACE inhibitors may exert their effects locally.32 However, the real mode of action of the ACE inhibitors is not completely understood. When cardiac ACE activity was measured in patients with or without chronic pharmacotherapy with the ACE inhibitor captopril, no inhibition was found when analyzed in vitro.13 Also, in heart failure patients, inadequate reduction of Ang II might be the cause of the persistent deterioration seen despite continuous ACE inhibitor therapy.33 Finally, it is worth mentioning that there seem to be major differences between the effects of different ACE inhibitors in heart failure in experimental models34 as well as in humans,35 showing their individual potentials of tissue (cardiac) ACE inhibition. Thus, the question may be raised whether ACE is blocked sufficiently during ACE inhibitor pharmacotherapy and whether this therapy could be improved.
We have found that chronic patient therapy with three different ACE inhibitors induced a measurable yet incomplete inhibition of left ventricular ACE when measured using the ACE methodology ([14C]Hip-His-Leu conversion at pH 8.3). Complete inhibition could be achieved by adding 1 μmol/L cilazaprilat to the homogenate (spiking), thus validating the selectivity of the assay. These data are in contrast to a study done by Urata et al13 showing that chronic ACE inhibitor therapy did not significantly alter left or right ventricular ACE activity levels. The present study and that of Urata et al use similar methodologies; the difference in the findings of the two studies may be explained by the approach. Our conclusion is derived from comparing the ACE activity levels of the same samples before and after dialysis. In contrast, Urata et al compared cardiac ACE activities in homogenates from ACE inhibitor–treated versus nontreated patients.
The subsequent question is whether the cardiac ACE inhibition is underestimated due to dissociation of the ACE-inhibitor complex in vitro. This could have occurred during the relatively long lag time between harvesting of the hearts, transportation time, and sample preparation (between 2 and 10 hours) or during the assay incubation.36 Indeed, relatively fast off-rates were reported for different ACE inhibitors from papillary muscles at 37°C.36 However, this is different from our conditions, since the hearts were transported at 4°C, a temperature at which the off-rates of high-affinity complexes are known to be slow. Indeed, dialysis experiments at 4°C even under vigorous conditions (solubilized homogenates dialyzed in presence of EDTA) showed relatively slow off-rates. Samples spiked with captopril showed the fastest recovery of ACE activity within 2 hours, whereas the activity of the samples spiked with enalaprilat and cilazaprilat recovered only after 5 days. It is worth noticing that the samples of patients receiving captopril, which should be most susceptible to the dissociation artifacts, seemed to be more inhibited than the samples of the enalapril patients. Thus, the dissociation artifacts seem to be less of a problem than anticipated, and the measured ACE inhibition in vivo may have been only moderate.
Our biochemical analyses on the total Ang I conversion in left ventricular tissue and coronary arteries showed, however, only a minute contribution by ACE but a major contribution by a serine protease, which may be chymase. These analyses confirm already published results13 but are in contrast to others.23 Since handling could produce loss of enzymatic activity, we used minimum handling in our sample preparation. Indeed, the controversial results published by Zisman et al23 seem to be explainable with a major loss of Ang I conversion capacity, mainly chymase, during their sample preparation (solubilization and dialysis, Table). The Ang I conversion rates that we found after this procedure reached almost exactly the numbers reported by Zisman et al (55 versus 50 fmol [125I]Ang II formed per milligram protein per hour). It is interesting to note that the processing of Ang I in the conversion assay under our conditions is four orders of magnitude lower than the processing of the artificial substrate [14C]Hip-His-Leu in the ACE assay (compare Fig 2⇑ with Fig 3⇑). Indeed, this may be the reason why Ang I conversion by ACE cannot be measured at all.
Since our results are astonishing in light of the effectiveness of ACE inhibitor therapy, we have performed a series of experiments to investigate and rule out potential artifacts that could have led to an underestimation of ACE and or chymase-like activity: (1) Ang I conversion was estimated in the present study under the most physiological conditions (pH 7.4, 150 mmol/L NaCl). (2) We have taken special care to find the appropriate protein (enzyme) working concentration that would not lead to substrate depletion and that thus allows a correct interpretation of the results. (3) We have verified that the results have not been altered by the iodination of Ang I. Conversion of [125I]Ang I was found to be similar to the one of cold Ang I. (4) The inhibition pattern of [125I]Ang I conversion was found to be similar between 10 pmol/L and 1 μmol/L of [125I]Ang I and indicates the relevance of the results created under nonphysiological high Ang I concentrations also.
It is astonishing to note that the main Ang I conversion capacity consists of a chymase-like enzyme that is even more abundant in coronary arteries than in the left ventricular tissue. In both tissue preparations, there was no apparent difference between arteries from patients with or without ACE inhibitor therapy. Since ACE is found mainly in the endothelium that lines the vessel walls,37 the vasculature is considered to be ACE-driven Ang I–converting machinery. Indeed, studies in humans have shown that ACE inhibitors are able to totally prevent the conversion of Ang I across all peripheral beds, including the coronary circulation.23 38 This is in line with the localization of both conversion enzymes. Even though both enzymes can be found in endothelial cells, chymase is secreted toward the basolateral site, whereas ACE is located on the luminal surface.17 39 Chymase is more prevalent in the interstitium in the medial and the adventitial region of the vessels,17 where it may contribute little to the blood Ang II levels but much to functional effects. Our contraction studies with the coronary arteries produced results similar to the biochemical studies and showed that the chymase-like activity is also able to induce Ang I–induced contractions. That these contractions could also be induced by the SUB [Pro11, d-Ala12]Ang I further supports the suggestion that the activity stems really from chymase itself.
Our results are in line with published studies showing that ACE inhibitors could not block >40% contraction of human gastroepiploic arteries21 or the positive inotropic effects in human detrusor muscle20 when the effects were induced by Ang I. A critical issue remains that the mast cells, a major site of chymase production, may be disrupted by the tissue preparation and might release chymase. Thus, chymase-related Ang I conversion may be overrated in vitro and be less significant in vivo. This problem is more likely to occur in tissue homogenates then in isolated ring preparations. Nonetheless, the isolated ring preparations showed results consistent with the homogenate preparations.
In addition, several studies pointed to functional chymase effects in vivo. In healthy volunteers, exercise induced Ang II increases in plasma; this can be significantly suppressed by the serine-protease inhibitor nafamostat.22 In patients with peripheral vascular disease, nafamostat significantly increases maximal walking distance and improved subjective symptoms.40 Infusion of the SUB [Pro11,d-Ala12]Ang I in conscious baboons produces hemodynamic and left ventricular functional changes consistent with arterial vasoconstriction. The ACE inhibitor captopril cannot inhibit these effects.28 The Ang II increase in the dog coronary sinus after coronary ligation was not inhibited by ACE inhibitors but by the two serine-protease inhibitors chymostatin and aprotinin.41
In general, as discussed before, the special tissular and cellular localization of chymase suggests that it may contribute to the tissular rather than the blood Ang II. In line with this hypothesis are the studies that could not detect a contribution of chymase to Ang II generation when biochemical conversion was measured in the circulatory compartment. These contrast with the functional studies, in which the chymase contribution became obvious. If tissular Ang II production is influenced by chymase, then processes like proliferation and fibrosis that are shown to be triggered by Ang II42 should be prime processes influenced by chymase. Indeed, large inductions of chymase mRNA could be shown upon vascular injury in dog carotid artery.43 This may prevent the antiproliferative effects of ACE inhibitors after balloon injury of carotid arteries in swine, monkeys, or humans44 45 ; these effects have previously been seen in rats46 and guinea pigs.47 In agreement with this hypothesis is the finding that an AT1 receptor antagonist, in contrast to an ACE inhibitor, can prevent the neointimal formation after vascular injury in dogs.48 The difference between the species may be explained by a species-different role of the serine protease system and needs to be investigated further.
In conclusion, the present results clearly show the role of a serine protease pathway for the formation of Ang II in human cardiac tissue. Our results give a strong rationale for future research on drugs that block the RAS more completely than ACE inhibitors.
Selected Abbreviations and Acronyms
|Ang (combination form)||=||angiotensin|
|CX||=||circumflex coronary artery|
|LAD||=||left anterior descending coronary artery|
|SBTI||=||soybean trypsin inhibitor|
We thank the staff of the Cardiovascular Surgery Department, University Hospital Zu¨rich, for their cooperation in collecting cardiovascular tissue. We also thank Ursula Wolfgang for her excellent technical assistance and the extra-long work days that were needed to complete these experiments.
- Received October 22, 1996.
- Accepted November 18, 1996.
Peach MJ. Renin-angiotensin system: biochemistry and mechanisms of action. Physiol Rev. 1977;57:313-317.
Pfeffer MA, Braunwald E, Moye LA, Basta L, Brown EJ Jr, Cuddy TE, Davis BR, Geltman EM, Goldman S, Flaker GC. Effect of captopril on mortality and morbidity in patients with left ventricular dysfunction after myocardial infarction: results of the survival and ventricular enlargement trial: the SAVE investigators. N Engl J Med. 1992;327:669-677.
Mento PF, Wilkes BM. Plasma angiotensins and blood pressure during converting enzyme inhibition. Hypertension. 1987;9:III42-III48.
Hollenberg NK, Fisher ND. Renal circulation and blockade of the renin-angiotensin system: is angiotensin-converting enzyme inhibition the last word? Hypertension. 1995;26:602-609.
Azizi M, Chatellier G, Guyene TT, Murieta Geoffroy D, Menard J. Additive effects of combined angiotensin-converting enzyme inhibition and angiotensin II antagonism on blood pressure and renin release in sodium-depleted normotensives. Circulation. 1995;92:825-834.
Urata H, Healy B, Stewart RW, Bumpus FM, Husain A. Angiotensin II–forming pathways in normal and failing human hearts. Circ Res. 1990;66:883-890.
Urata H, Kinoshita A, Perez DM, Misono KS, Bumpus FM, Graham RM, Husain A. Cloning of the gene and cDNA for human heart chymase. J Biol Chem. 1991;266:17173-17179.
Kinoshita A, Urata H, Bumpus FM, Husain A. Multiple determinants for the high substrate specificity of an angiotensin II-forming chymase from the human heart. J Biol Chem. 1991;266:19192-19197.
Urata H, Strobel F, Ganten D. Widespread tissue distribution of human chymase. J Hypertens Suppl. 1994;12:S17-S22.
Urata H, Boehm KD, Philip A, Kinoshita A, Gabrovsek J, Bumpus FM, Husain A. Cellular localization and regional distribution of an angiotensin II-forming chymase in the heart. J Clin Invest. 1993;91:1269-1281.
Cornish KG, Joyner WL, Gilmore JP. Direct evidence for the presence of a different converting enzyme in the hamster cheek pouch. Circ Res. 1979;44:540-544.
Bumpus FM. Angiotensin I and II: some early observations made at the Cleveland Clinic Foundation and recent discoveries relative to angiotensin II formation in human heart. Hypertension. 1991;18:III-122-III-125.
Lindberg BF, Nilsson LG, Hedlund H, Stahl M, Andersson KE. Angiotensin I is converted to angiotensin II by a serine protease in human detrusor smooth muscle. Am J Physiol. 1994;266:R1861-R1867.
Zisman LS, Abraham WT, Meixell GE, Vamvakias BN, Quaife RA, Lowes BD, Roden RL, Peacock SJ, Groves BM, Raynolds MV. Angiotensin II formation in the intact human heart: predominance of the angiotensin-converting enzyme pathway. J Clin Invest. 1995;96:1490-1498.
Lowry OH, Rosebrough NJ, Farr AL, Randall RJ. Protein measurements with the Folin phenol reagent. J Biol Chem. 1951;193:265-275.
Nussberger J, Brunner DB, Waeber B, Brunner HR. Specific measurement of angiotensin metabolites and in vitro generated angiotensin II in plasma. Hypertension. 1986;8:476-482.
Toda N. Isolated human coronary arteries in response to vasoconstrictor substances. Am J Physiol. 1983;245:H937-H941.
Hoit BD, Shao Y, Kinoshita A, Gabel M, Husain A, Walsh RA. Effects of angiotensin II generated by an angiotensin converting enzyme-independent pathway on left ventricular performance in the conscious baboon. J Clin Invest. 1995;95:1519-1527.
Dostal DE, Baker KM. Evidence for a role of an intracardiac renin-angiotensin system in normal and failing hearts. Trends Cardiovasc Med. 1993;3:67-74.
Rousseau MF, Konstam MA, Benedict CR, Donckier J, Galanti L, Melin J, Kinan D, Ahn S, Ketelslegers JM, Pouleur H. Progression of left ventricular dysfunction secondary to coronary artery disease, sustained neurohormonal activation and effects of ibopamine therapy during long-term therapy with angiotensin-converting enzyme inhibitor. Am J Cardiol. 1994;73:488-493.
Hirsch AT, Talsness CE, Smith AD, Schunkert H, Ingelfinger JR, Dzau VJ. Differential effects of captopril and enalapril on tissue renin-angiotensin systems in experimental heart failure. Circulation. 1992;86:1566-1574.
Kinoshita A, Urata H, Bumpus FM, Husain A. Measurement of angiotensin I converting enzyme inhibition in the heart. Circ Res. 1993;73:51-60.
Admiraal PJ, Derkx FH, Danser AH, Pieterman H, Schalekamp MA. Metabolism and production of angiotensin I in different vascular beds in subjects with hypertension. Hypertension. 1990;15:44-55.
Sporn LA, Marder VJ, Wagner DD. Differing polarity of the constitutive and regulated secretory pathways for von Willebrand factor in endothelial cells. J Cell Biol. 1989;108:1283-1289.
Brilla CG, Maisch B, Weber KT. Renin-angiotensin system and myocardial collagen matrix in hypertensive heart disease: in vivo and in vitro studies on collagen matrix regulation. Clin Invest. 1993;5:S35-S41.
Hanson SR, Powell JS, Dodson T, Lumsden A, Kelly AB, Anderson JS, Clowes AW, Harker LA. Effects of angiotensin converting enzyme inhibition with cilazapril on intimal hyperplasia in injured arteries and vascular grafts in the baboon. Hypertension. 1991;18:II-70-II-76.
Lam JY, Lacoste L, Bourassa MG. Cilazapril and early atherosclerotic changes after balloon injury of porcine carotid arteries. Circulation. 1992;85:1542-1547.
Powell JS, Clozel JP, Muller RK, Kuhn H, Hefti F, Hosang M, Baumgartner HR. Inhibitors of angiotensin-converting enzyme prevent myointimal proliferation after vascular injury. Science. 1989;245:186-188.
Clozel JP, Hess P, Michael C, Schietinger K, Baumgartner HR. Inhibition of converting enzyme and neointima formation after vascular injury in rabbits and guinea pigs. Hypertension. 1991;18:II-55-II-59.
Okunishi H, Shiota N, Fukamizu A. Angiotensin antagonist, not ACE inhibitor, prevents neointimal formation on injured canine arteries. Hypertension. 1994;12:S132. Abstract.