Angiotensin-Converting Enzyme–Independent Contraction to Angiotensin I in Human Resistance Arteries
Background—In vitro studies of myocardial tissue suggest that angiotensin II (Ang II) may be generated by both ACE and chymase. A similar dual pathway may exist in the vasculature. We studied the effects of ACE and chymase inhibitors on the contractile response to angiotensin I (Ang I) in human resistance arteries to investigate ACE-independent generation of Ang II.
Methods and Results—Subcutaneous resistance arteries (250 to 350 μm) were obtained from gluteal biopsies from volunteers and New Zealand White rabbits and mounted on a wire myograph. Contractile ability was tested with high-potassium depolarization and norepinephrine 10 μmol/L and endothelial integrity by relaxation to acetylcholine 3 μmol/L. Cumulative concentration-response curves were constructed for Ang I in the presence of enalaprilat 1 μmol/L, chymostatin 10 μmol/L, or both inhibitors together. In the rabbit, enalaprilat completely inhibited the Ang I response. In human vessels, enalaprilat or chymostatin alone had no effect, but the combination of enalaprilat and chymostatin almost completely inhibited the response to Ang I.
Conclusions—A dual pathway for Ang II generation exists in human resistance arteries, mediated by ACE and a chymostatin-sensitive enzyme, probably chymase. We confirm that a marked species difference exists in the mechanism of Ang II generation between the human and the rabbit. More efficacious suppression of the renin-angiotensin system may require development of novel enzyme inhibitors or combinations of currently available drugs.
Angiotensin-converting enzyme inhibitors (ACEIs) play a major role in the treatment of chronic heart failure (CHF), renal failure, and hypertension. During chronic therapy with these drugs, plasma angiotensin (Ang) II concentrations are frequently not suppressed, and deterioration in patients with CHF has been correlated with this escape from inhibition.1 2 Recently, the ELITE investigators reported a survival advantage with the Ang II type 1 receptor antagonist losartan over conventional ACE inhibition in elderly patients with CHF, suggesting that continuing generation of Ang II during ACEI therapy might be deleterious.3 The inability of ACEIs to fully block the generation of Ang II may arise from the existence of alternative serine protease pathways that are able to cleave Ang I. The ability of enzymes such as kallikrein, tonin, and cathepsin G to generate Ang II under the appropriate in vitro conditions is well known.4
Animal studies have identified ACE-independent Ang II generation in a variety of species and preparations: chymostatin-sensitive Ang II generation (CAGE) has been observed in canine mesenteric arteries and monkey pulmonary and mesenteric arteries.5 6 CAGE has also been observed in extracts from canine myocardium, but it seems to be less important when assessed across the intact myocardial bed.7 There is also a marked species difference in the mechanisms of Ang II formation.8 Thus, local Ang II generation across the isolated perfused rat heart is very sensitive to ACE inhibition, and Ang II formation in rodent vessels is almost totally ACE-dependent.8 9
In homogenized human myocardial tissue, Ang II generation is mediated by both ACE and chymase, with the latter pathway predominant.10 This finding, however, has been challenged in studies of Ang II generation across the intact myocardial bed.11 Vascular ACE-independent Ang II generation has also been observed. Thus, 60% of the Ang I–mediated contraction in human gastroepiploic arteries is ACE-independent, and the contraction of human cutaneous resistance arteries to porcine tetradecapeptide renin substrate is inhibited by combinations of captopril and the serine protease inhibitor aprotinin (Trasylol) or captopril and the renin inhibitor H261 but not by captopril alone, again suggesting that the non-ACE component of Ang II generation is mediated by a serine protease.8 12 However, no studies have looked directly at Ang II generation from Ang I in human resistance arteries.
These experiments were designed to investigate further the existence of non-ACE pathways for Ang II generation in human resistance arteries from healthy subjects compared with those from a control species, in this case the New Zealand White rabbit.
Human Ang I, bradykinin (BK), norepinephrine, and acetylcholine were purchased from Sigma-Aldrich. Trasylol was purchased from Bayer PLC. Chymostatin was purchased from Bachem, and enalaprilat was a gift from Merck, Sharp, and Dohme Ltd. CH 5450 was a gift from Ferring Research Laboratories. Experiments were performed on a Mulvany-Halpern 4-channel myograph (JP Trading).
Human Subcutaneous Resistance Arteries
Nonsmoking healthy male volunteers with no history of cardiovascular disease, hypertension, diabetes, or renal impairment attended the Research Unit at the Western Infirmary, Glasgow. All gave written informed consent to participate in the study, which was approved by the Hospital Ethics Review Committee. After 10 minutes’ rest, blood pressure was recorded in the right arm, with the patient sitting, with a standard mercury sphygmomanometer (Accoson). Venous blood was drawn for measurement of electrolytes, random glucose, and cholesterol. A subcutaneous gluteal fat biopsy was then performed under local anesthesia with 1% lignocaine: an ellipse of skin 1.5×0.5 cm was excised and adherent fat dissected free to a depth of ≈1.5 cm.13 This was placed immediately into cold 0.9% NaCl solution and then transferred to cold Krebs solution (composition in mmol/L: NaCl 118.4, KCl 4.7, MgSO4 · H2O 1.2, KH2PO4 1.2, NaHCO3 24.9, CaCl2 2.5, glucose 11.1, and EDTA 0.023, which gives a pH of 7.4 when aerated with a 5% CO2/95% O2 mixture). When possible, 4 resistance arteries ≈2 mm long were dissected and mounted on two 40-μm-diameter stainless steel wires in a 4-channel myograph. The temperature was raised to 37°C, and the medium was aerated (gas mixture as above) continuously.
Rabbit Cutaneous Artery Preparation
Experiments were carried out in male New Zealand White rabbits weighing 3.0 to 3.5 kg. These animals were sham-operated controls for a rabbit coronary artery ligation model of heart failure and had normal left ventricular function as determined by echocardiogram. They were killed by an overdose of pentobarbital into the ear vein, and a flap of skin from the area overlying the gluteal muscles was removed. Connective tissue was cleared from above the network of cutaneous vessels, and resistance arteries were dissected free and placed into cold Krebs solution as described above.
After a rest period of 30 minutes, a normalization procedure was followed for each artery to determine the normalized internal diameter (ID), L0, at which contraction is thought to be optimal, and the vessel was set to that diameter.13 Vessels were regarded as resistance arteries if the normalized ID was <500 μm, according to the definition of Mulvany.14
After the normalization procedure, the arteries were exposed twice to KPSS (Krebs solution with KCl substituted for NaCl on an equimolar basis) and once to norepinephrine 10 μmol/L. After a plateau contraction had been attained with norepinephrine, acetylcholine 3 μmol/L was added to stimulate endothelium-dependent vasodilatation. Vessels that were unable to contract to either KPSS or norepinephrine or showed no relaxation to acetylcholine (and were therefore considered to have no functionally intact endothelium) were discarded.
Arteries were then incubated for a further 30 minutes in either Krebs solution alone (control, vessel 1), enalaprilat 1 μmol/L (vessel 2), chymostatin 10 μmol/L (vessel 3), or both enalaprilat 1 μmol/L and chymostatin 10 μmol/L (vessel 4). A cumulative concentration-response curve was then performed to Ang I from 0.01 nmol/L to 3 μmol/L. Arteries were exposed to each concentration of Ang I for 4 minutes or until a peak response at that concentration had been reached, whichever was sooner, because it was observed that tachyphylaxis developed even during a single dose-response curve. Owing to tachyphylaxis, vessels could not be exposed to Ang I more than once. In preliminary experiments, concentration-response curves were performed to Ang I in the presence of losartan 1 μmol/L to confirm that the contractile response to Ang I was mediated through the action of locally generated Ang II at the Ang II type 1 receptor.
In a subset of vessels, responses to BK were investigated. Following the dose-response curve to Ang I, baseline was reestablished by washing with Krebs’ solution. Vessel 2 was incubated for 30 minutes in the presence of enalaprilat 1 μmol/L, vessel 1 remaining the control. Precontraction was then established with norepinephrine 10 μmol/L, and once a stable plateau had been reached, vessels were exposed to BK 0.01 nmol/L to 3 μmol/L.
The protocol was essentially the same for rabbit arteries, the difference lying in the inhibitors used. Rabbit arteries were incubated with either enalaprilat 1 μmol/L, Trasylol 100 U/mL, or CH5450 10 μmol/L, the latter being a peptide derivative that specifically inhibits chymase.15 The combination of enalaprilat and chymase inhibitor was not used because the inhibition achieved with the former was so marked (Figure 2⇓). Rabbit vessels showed no relaxation to BK.
Contractile responses were expressed as an increase in active effective pressure (kPa), calculated as an increase in isometric tension above resting divided by the normalized internal radius. When possible, agonist potency in the presence or absence of inhibitor was expressed in terms of EC10, EC25, and EC50 values, these being the concentrations required to produce 10%, 25%, and 50% of the maximum response, respectively. To calculate these values, responses were averaged at each concentration of agonist. Because some responses did not achieve a clear maximum even at the highest concentration of agonist, EC values could not be calculated for all curves. Thus, for the purposes of comparison, the area under the curve (AUC) was calculated for each concentration-response curve by a simple trapezoidal rule. This gave a dimensionless numerical value expressing the contractile response. Responses to Ang I for each inhibitor were then compared as a group with control responses, and statistical testing was performed with a Mann-Whitney U test. AUC was also calculated for responses to BK, and because a matched control was available for all experimental curves, statistical testing was performed with a Wilcoxon rank sum test. All analyses were performed with SPSS software, and differences were considered significant at a level of P<0.05.
Clinical characteristics of the volunteers are given in Table 1⇓. Volunteers were normotensive and had normal biochemical characteristics. In total, 16 biopsies were performed, but it was not possible to find 4 vessels from each. Hence, concentration-response curves to Ang I alone (control) were performed in 15 arteries from separate patients, in the presence of enalaprilat in 12 arteries, chymostatin in 6 arteries, and the combination of chymostatin and enalaprilat in 7 arteries. The sizes of the vessels in each group are given in Table 2⇓, as are the responses of the vessels in each group to the second exposure to KPSS, norepinephrine, and acetylcholine. There were no significant differences between control and experimental groups in terms of vessel diameter, contractile ability, or endothelium-dependent vasodilatation to acetylcholine.
Table 3⇓ shows similar data from rabbit vessels. There are no significant differences between experimental and control groups in any of these parameters. It was a consistent finding that rabbit vessels contracted more powerfully than human vessels and displayed less relaxation to acetylcholine.
Ang I generated a concentration-response curve in human and rabbit cutaneous arteries (Figures 1⇓ and 2⇓). In human vessels, the contractile threshold to Ang I was 0.3 nmol/L, with a maximum response at 0.1 μmol/L. In rabbit vessels, enalaprilat 10−6 mol/L inhibited this response markedly, but no inhibition was observed with either CH 5450 or Trasylol. No inhibition was observed with either enalaprilat or chymostatin alone in human vessels. However, the combination of enalaprilat and chymostatin caused a clear inhibition of the response. In preliminary experiments, losartan 10−6 mol/L completely abolished the response to Ang I (data not shown).
Table 4⇓ shows the effects of each inhibitor on the EC10, EC25, and EC50 for Ang I. As noted above, it was not possible to express the response to Ang I in the presence of enalaprilat and chymostatin combined in human vessels or in the presence of enalaprilat alone in rabbit vessels as an EC value. Therefore, the AUC was calculated to compare the responses to Ang I in the presence of each inhibitor with control. In human vessels, this gave a value of 5.11±1.02 for Ang I alone, 4.01±0.99 for Ang I plus enalaprilat, 5.52±2.30 for Ang I plus chymostatin, and 0.82±0.55 for Ang I plus enalaprilat and chymostatin (all data mean±SD). By this method, it is clear that neither enalaprilat nor chymostatin alone had a significant inhibitory effect in human vessels (P=0.486 and P=0.712, respectively). However, the inhibition observed in the presence of both chymostatin and enalaprilat was highly significant (P=0.006). In rabbit vessels, AUC values were 10.53±2.35 for Ang I alone, 10.20±1.89 for Ang I plus Trasylol, 7.23±1.27 for Ang I plus CH 5450, and 0.95±0.48 for Ang I plus enalaprilat. Thus, enalaprilat inhibited the response to Ang I (P=0.001), but neither CH 5450 nor Trasylol had any effect (P=0.463 and P=0.906, respectively).
Responses to BK (n=8) in human vessels are shown in Figure 3⇓. In vessels precontracted with norepinephrine, BK elicited a dose-dependent relaxation, which was potentiated by enalaprilat, increasing the degree of relaxation obtained at low concentrations of BK without affecting the maximum response. The effect of enalaprilat on the potency of BK is demonstrated in Table 5⇓. AUC calculated for BK responses gave values of 185.92±33.02 for control and 125.41±33.30 for vessels treated with enalaprilat. These values were significantly different (P=0.012).
Our results demonstrate that in vitro blockade of ACE is insufficient to prevent the contraction of human subcutaneous resistance arteries to Ang I. Because our preliminary experiments showed that this response is completely blocked by losartan, we assume that contraction to Ang I represents the effect of Ang II generated locally and acting on the Ang II type I receptor. Thus, treatment with an ACEI appears to be unable to prevent conversion of Ang I to Ang II in human resistance arteries. In the rabbit, in contrast, ACEI administration fully prevented Ang I–induced contraction.
Because ACE is identical to kininase II, the enzyme responsible for the degradation of kinins, it has been suggested that potentiation of BK may be partly responsible for the actions of ACEI. Thus, plasma kinin concentrations are increased in humans by quinapril, and ACEI-induced coronary artery vasodilation in dogs and humans has been shown to be mediated by BK.16 17 18 We investigated the effect of enalaprilat on the response to BK. As expected, enalaprilat clearly potentiated the response to BK, with a left shift in the dose-response curve and a 5-fold reduction in the EC50.
Previous studies have shown that serine proteases provide an alternative pathway for the conversion of Ang I to Ang II. In specific tissues (notably the myocardium), this alternative pathway may be predominant. Urata et al10 observed that 80% of Ang II generation in human heart homogenates was ACE-independent and identified the enzyme responsible to be human heart chymase (HHC). This enzyme has now been cloned and sequenced.19 It is remarkably specific to the conversion of Ang I to Ang II, has a high catalytic rate but no kininase activity, and thus should not influence responses to BK.20 Immunohistochemistry and Northern blot analysis localize the enzyme to secretory granules within cardiac mast cells, the cardiac interstitium, and vascular adventitia. It is not found in human plasma.21 Our finding that chymostatin alone was ineffective in blocking the response to Ang I suggests that there is a dual pathway for Ang II generation in human resistance arteries, mediated by ACE and a chymostatin-sensitive enzyme, which is probably HHC. These pathways appear to be equally able to generate Ang II, so that significant inhibition of the response to Ang I was achieved only by combining inhibitors of both ACE and chymase.
The comparison with rabbit vessels is of interest: here, enalaprilat alone inhibited the response to Ang I to the same degree as combination treatment in human vessels, but a chymase inhibitor and a broad serine protease inhibitor had no effect. Thus, a species difference exists in the response to Ang I, and it would appear that the local production of Ang II in rabbit subcutaneous arteries is ACE-dependent. Moreover, these rabbit vessels consistently failed to vasodilate when challenged with BK. Such species differences are clearly of importance in choosing suitable animal models of CHF, hypertension, or chronic renal failure, in which ACE is an important therapeutic target.
It is possible that these in vitro experiments exaggerate the importance of alternative Ang II–generating pathways. In humans, as in dogs, a dichotomy exists, with chymase appearing to be responsible for the majority of Ang II generation in homogenized myocardial tissue in vitro7 10 but ACE being the predominant enzyme in vivo.11 This problem has been noted before and raises the possibility that the importance of non-ACE pathways may be exaggerated in in vitro preparations that expose Ang I to an enzyme that is normally hidden from it.22 In vivo, Ang I will be exposed to both circulating and endothelial ACE and thus may be unavailable for conversion by chymase. Furthermore, in our experiments, drugs are added to the solution in the organ bath and therefore delivered to the vessel abluminally, perhaps exposing adventitial and interstitial chymase to an unphysiologically high concentration of Ang I. It is also possible that the apparent lack of effect of enalaprilat might be due to a methodological problem, such as failure of the drug to reach adequate tissue concentrations. However, enalaprilat was clearly effective in rabbit vessels, and the marked potentiation of the response to BK in human arteries suggests that tissue ACE was inhibited.
Why, then, is redundancy observed in the final limb of the Ang II–generating pathway? One possibility is that Ang II generated locally by non-ACE pathways has a different role from Ang II generated by circulating and endothelial ACE. Ang II is known to regulate cellular hypertrophy and hyperplasia and contribute to vascular remodeling.23 Thus, the contraction of the vessels to Ang I generated by chymase could be a epiphenomenon of another function of the enzyme. This is consistent with the observation that ACEI reduces restenosis in rodent models of balloon angioplasty but not in larger animals, in which chymase can be shown to be induced after vascular injury.24 25 It is also consistent with the differences in the Ang II–generating pathways observed between species.
A number of approaches could be used to resolve these questions. In the technique of perfusion myography, vessels are cannulated at both ends and then perfused at a controlled pressure, allowing a distinction to be made between drugs applied luminally or abluminally.26 In addition, in vivo studies using forearm occlusion plethysmography to study the effects of ACEI on the contraction to Ang I would give an indication of global vascular bed Ang I–Ang II conversion but would not give information about specific parts of the vascular tree. Unfortunately, no specific inhibitors of chymase are currently available for in vivo use.
Our demonstration of the existence of a dual pathway for Ang II generation in human resistance vessels ex vivo is of physiological and therapeutic significance. In addition to the studies showing failure of Ang II suppression on long-term ACEI therapy, 2 studies have shown that the exercise-induced rise in Ang II levels measured in the veins draining the human leg could not be prevented by captopril.27 28 However, nafamostat, a serine protease inhibitor, did attenuate this increase.26 In ischemic conditions, Ang II generation may be mediated by a serine protease, and Nafamostat has also been shown to increase lower-limb blood flow and exercise capacity in patients with peripheral vascular disease.29 Thus, continuing generation of Ang II despite ACEI therapy, mediated by serine proteases, may be of pathophysiological significance, and efficient blockade of the renin-angiotensin system may be better achieved with Ang II receptor antagonists or by a combination of these agents with ACEIs (thus combining the potential benefit of kinin potentiation with Ang II receptor blockade). Chymase itself is a potential target for pharmacological intervention.
Our data indicate that Ang II generation in human subcutaneous arteries is mediated by a dual enzymatic pathway and provide a strong basis for further investigation of these mechanisms and for novel therapeutic strategies to facilitate blockade of the renin-angiotensin system in humans.
We acknowledge the support of the Wellcome Trust in the form of a Junior Training Fellowship to Dr Padmanabhan.
- Received August 24, 1998.
- Revision received March 24, 1999.
- Accepted March 30, 1999.
- Copyright © 1999 by American Heart Association
Rousseau MF, Konstam MA, Benedict CR, Donckier J, Galanti L, Melin J, Kinan D, Ahn S, Ketelslegers J-M, Pouleur H. Progression of left ventricular dysfunction secondary to coronary artery disease, sustained neurohumoral activation and effects of ibopamine therapy during long term therapy with angiotensin-converting enzyme inhibitor. Am J Cardiol. 1994;73:488–493.
Pitt B, Segal R, Martinez FA, Meurers G, Cowley AJ, Thomas I, Deedwania PC, Ney DE, Snavely DB, Chang PI, on behalf of ELITE Study Investigators. Randomised trial of losartan versus captopril in patients over 65 with heart failure (Evaluation of Losartan in the Elderly Study, ELITE). Lancet. 1997;349:747–752.
Arakawa K. Serine protease angiotensin II systems. J Hypertens. 1996;14(suppl):S3–S7.
Balcells E, Meng QC, Hageman GR, Palmer RW, Durand JN, Dell’Italia LJ. Angiotensin II formation in dog heart is mediated by different pathways in vivo and in vitro. Am J Physiol. 1996;271:H417–H421.
Lindpaintner K, Jin MW, Niedermaier N, Wilhelm MJ, Ganten D. Cardiac angiotensinogen and its local activation in the isolated perfused beating heart. Circ Res. 1990;67:564–573.
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.
Zisman LS, Abraham WT, Meixell GE, Vamvakias BN, Quaife RA, Lowes BD, Roden RL, Peacock SJ, Groves BN, Raynolds MV, Bristow MR, Perryman MB. Angiotensin II formation in the intact human heart. J Clin Invest. 1995;95:1490–1498.
Mulvany MJ, Aalkjaer C. Structure and function of small arteries. Physiol Rev. 1990;4:922–961.
Sudhir K, Chou TM, Hutchison SJ, Chatterjee K. Coronary vasodilation induced by angiotensin-converting enzyme inhibition in vivo: differential contribution of nitric oxide and bradykinin in conductance and resistance arteries. Circulation. 1996;93:1734–1739.
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. 1990;266:17173–17179.
Kinoshita A, Urata H, Bumpus FM, Husain A. Multiple determinants for the high substrate specificity of an angiotensin II-forming chymase in the human heart. J Biol Chem. 1991;266:19192–19197.
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.
Wolny A, Clozel JP, Rein J, Mory P, Vogt P, Turino M, Kiowski W, Fischli W. Functional and biochemical analysis of angiotensin II-forming pathways in the human heart. Circ Res. 1997;80:219–227.
Berk BC, Vekstein V, Gordon HM, Tsuda T. Angiotensin II-stimulated protein synthesis in cultured vascular smooth muscle cells. Hypertension. 1989;13:305–314.
Clozel JS, 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(suppl II):II-55–II-59.