Author + information
- Received August 4, 1999
- Revision received November 8, 2000
- Accepted December 13, 2000
- Published online March 15, 2001.
- Mark C Petrie, BSc, MB, ChB, MRCP∗,
- Neal Padmanabhan, MA, BM, BCh, MRCP†,* (, )
- John E McDonald, BSc, MB, ChB, MRCP∗,†,
- Chris Hillier, BSc, PhD∗,
- John M.C Connell, MD, FRCP† and
- John J.V McMurray, BSc, MD, FRCP, FESC, FACC∗
- ↵*Reprint requests and correspondence:
Dr. Neal Padmanabhan, Medical Research Council Blood Pressure Group, Department of Medicine and Therapeutics, Western Infirmary, University of Glasgow, Glasgow, United Kingdom, G11 6NT
We sought to demonstrate non-angiotensin converting enzyme (ACE) dependent angiotensin II (AII) generating pathways in resistance arteries from patients with chronic heart failure (CHF).
Non-ACE dependent AII generation occurs in resistance arteries from normal volunteers. Inhibition of non-ACE dependent AII generation may have therapeutic potential in CHF.
Resistance arteries were dissected from gluteal biopsies from patients with coronary heart disease (CHD) and preserved left ventricular function and from patients with CHF. Using wire myography, concentration response curves to angiotensin I (AI) and AII were constructed in the presence of 1) vehicle, 2) chymostatin [an inhibitor of chymase], 3) enalaprilat, and 4) the combination of chymostatin and enalaprilat.
In resistance arteries from patients with CHD, the vasoconstrictor response to AI was not inhibited by either inhibitor alone (chymostatin [p ≥ 0.05] or enalaprilat [p ≥ 0.05]) but was significantly inhibited by the combination (p < 0.001). In arteries from patients with CHF, AI responses were inhibited by enalaprilat (p < 0.05) but not by chymostatin alone (p > 0.05). The combination of chymostatin and enalaprilat markedly inhibited the response to AI (p < 0.001) to a greater degree than enalaprilat alone (p ≤ 0.01).
Non-ACE dependent AII generating pathways exist in resistance arteries from patients with both CHF and CHD. In resistance arteries from patients with CHD, inhibition of either the ACE or chymase pathway alone has no effect on AII generation, and both pathways must be blocked before the vasoconstrictor action of AI is inhibited. In CHF, blockade of ACE results in marked inhibition of responses to AI, but this is enhanced by coinhibition of chymase. These studies suggest that full suppression of the renin-angiotensin system cannot be achieved by ACE inhibition alone and provide a rationale for developing future therapeutic strategies.
The traditional view that angiotensin II (AII) formation is solely dependent on angiotensin converting enzyme (ACE) has recently been challenged. Non-ACE dependent conversion of angiotensin I (AI) to AII has been demonstrated in homogenates of human myocardial tissue (1,2). In vitro, non-ACE dependent conversion of AI to AII is brought about by one or more serine proteases. The most important of these is thought to be chymase (3,4)as non-ACE mediated AII formation is substantially blocked by chymase inhibitors such as chymostatin (2). That these non-ACE pathways are functionally important in human blood vessels has recently been demonstrated by Voors et al. (5)in large arteries and by our group in small resistance arteries from normal human volunteers (6).
Demonstration of this dual pathway for AII generation has important potential implications for the treatment of a number of cardiovascular diseases, especially chronic heart failure (CHF). Blockade of the renin-angiotensin-aldosterone system (RAAS) improves symptoms and survival in heart failure (7,8), and greater inhibition brings about greater benefit (9). The current approach to RAAS interruption is ACE inhibition. However, the existence of a dual pathway means that that AII generation might persist in CHF, despite ACE inhibition, and raises the possibility that either the syndrome itself, or ACE inhibition, might also up-regulate the alternative pathway. It is known that treatment of CHF with an ACE inhibitor does not result in long-term suppression of AII plasma levels (10). Furthermore, in one study of patients with CHF, deterioration in left ventricular (LV) function occurred despite ACE inhibition, and this was associated with plasma concentrations of AII that not only failed to show suppression but were elevated compared with those found in age and gender-matched controls without CHF (11).
The objective of this study was to determine whether or not non-ACE dependent AI to AII conversion occurs in the resistance arteries of patients with CHF who receive long-term treatment with ACE inhibitors. We also studied whether or not non-ACE dependent AII generation occurs in resistance arteries from patients with coronary heart disease (CHD) and preserved ventricular function who have not been treated with ACE inhibitors.
All patients with renal failure (creatinine >200 μmol/l) and diabetes mellitus were excluded. Written informed consent was obtained from each patient, and all protocols were approved by the local committee on medical ethics.
Patients with CHF
Ambulatory patients with New York Heart Association class II/III CHF were studied. All were on long-term (>3 months) ACE inhibitor and diuretic treatment. The etiology of CHF was CHD in all cases, and each patient had an echocardiographic LV ejection fraction (LVEF) <40% (Simpson’s biplane method). All patients had suffered a previous myocardial infarction. The patient’s usual medication (including ACE inhibitor therapy) was taken on the study morning. These patients underwent gluteal biopsy and study of subcutaneous resistance arteries.
Patients with CHD
Patients with chronic stable angina attending outpatient clinics were studied. All patients had preserved LV systolic function, determined as an echocardiographic LVEF ≥40% (Simpson’s biplane method), and none was treated with an ACE inhibitor. Patients underwent gluteal biopsy and study of subcutaneous resistance arteries.
Human AI, AII, bradykinin (BK), norepinephrine (NE) and acetylcholine (ACh) were purchased from Sigma (Poole, United Kingdom). Chymostatin was purchased from Bachem (Safron-Walden, United Kingdom), and enalaprilat was a gift from Merck, Sharp and Dohme Ltd. The studies were performed on a Mulvany-Halpern four-channel wire myograph (J.P. Trading, Aarhus, Denmark).
Blood sampling, biopsy procedure and artery preparation
After 15 min supine rest, blood was drawn from a cannula in an antecubital vein for estimation of blood chemistry and serum cholesterol. Subcutaneous gluteal biopsies were then obtained from each patient under local anesthesia (1% lidocaine) by the method previously described (12). Dissected tissue was placed immediately into cold 0.9% NaCl and then transferred to cold Kreb’s solution (composition in mM: NaCl 118.4, KCl 4.7, MgSO4.H2O 1.2, KH2PO41.2. Na HCO324.9, CaCl22.5, glucose 11.1, EDTA 0.023, which gives a pH of 7.4 when gassed with a 5% CO2/95% O2mixture). Where possible, four resistance arteries approximately 2 mm in length were dissected free of fat. Dissected arteries were stored in Kreb’s solution overnight at 4°C. Approximately 24 h after the biopsy they were mounted on two 40 μm diameter stainless steel wires in a four-channel myograph in which the wires are attached to a force transducer and micrometer, respectively. The temperature was raised to 37°C, and a gas mixture (composition above) was bubbled in for the duration of the experiment. Dissection of resistance arteries from gluteal biopsies and myography protocols were performed by an operator who was blind to the patient type (i.e., CHF or CHD).
Myography protocol: ACE and chymase inhibition
After a rest period of 30 min, each artery was stretched at 1-min intervals to determine the passive exponential wall tension-internal circumference (L) relationship. From the Laplace equation, where P = T/r (P is the effective pressure, T is the wall tension and r is the internal radius), the equivalent circumference (L100) for a transmural pressure of 100 mm Hg was calculated for each vessel by an iterative computer method. Each vessel was then set to the normalized internal diameter, L1= 0.9 × L100/π, at which contraction is thought to be optimal (13).
After the above normalization procedure, the arteries were exposed twice to KPSS (Kreb’s solution with KCl substituted for NaCl on an equimolar basis) and once to NE 10 μM. After a plateau contraction had been attained with NE, ACh 3 μM was added to the bath in order to stimulate endothelium-dependent vasodilation. Vessels that were unable to contract to either KPSS or NE, or that showed no relaxation to ACh (and were, therefore, considered to have no functionally intact endothelium), were discarded.
Vessels were then incubated for a further 30 min in either Kreb’s solution alone (vehicle, vessel 1) or with enalaprilat 10−6M (vessel 2), chymostatin 10−5M (vessel 3) or both enalaprilat 10−6M and chymostatin 10−5M (vessel 4), respectively. Cumulative concentration response (contraction) curves were then obtained with AI (10−11M to 3 × 10−6M). After exposure to AI, vessels were washed with Kreb’s solution to reestablish baseline, and inhibitors were again added maintaining the same relationship between vessel and inhibitor. Responses to AI are expressed as the percent contraction to that elicited by 123 mM KCl (second exposure) in that vessel.
Responses to BK and ACh
After 30 min, a cumulative concentration response (contraction) curve was constructed to NE, then to ACh and finally to BK, with further washes and incubation periods between curves. Responses to BK and ACh are expressed as percent change to precontraction with NE.
Responses to AII
The above protocol was repeated in arteries from different patients studying the effect of these inhibitors on cumulative concentration response curves to AII in place of AI. Responses to AI are expressed as the percent contraction to that elicited by 123 mM KCl (second exposure) in that vessel.
Effect of AII type-1 receptor blockade (with losartan)
The effect of losartan (10−6M) on the contractile response to AI was studied in resistance arteries taken from patients with CHD. The vessels were prepared as described above but were pre-incubated with losartan rather than enalaprilat or chymostatin before construction of an AI concentration response curve.
A maximal response was not observed to AI in arteries in the presence of the combination of enalaprilat and chymostatin. This prevented calculation of EC50 (concentration for half-maximum response) or comparisons of maximum responses. Instead, the threshold concentration for the response to AI (defined as the first concentration at which a response was detected) was identified for each curve (5). These values were expressed as the negative logarithm (to ease computation) and entered into a one-way analysis of variance (ANOVA) with a Bonferroni correction for multiple comparisons (GraphPad Prism, GraphPad Software, Inc., San Diego, California). Responses to BK and ACh were expressed by the pD2 (negative logarithm of the EC50) and maximum response (% vasodilation). Responses to BK and ACh in vessels 2, 3 and 4 were then compared with vehicle by one-way ANOVA with a Bonferroni correction for multiple comparisons. A similar analysis was performed for responses to AII. All data in the text are expressed as mean (± SD), and all data in graphical form are expressed as mean ± SE, unless stated otherwise. Although pD2 values were used when comparing curves, the equivalent EC50 values (or threshold concentrations) are stated in the text to ease comprehension.
The clinical characteristics of patients taking part in the study are given in Table 1. Baseline characteristics of patients with CHD and CHF differed in only two aspects (except their previously defined differences in LV function and ACE inhibitor and diuretic prescriptions). First, more patients with CHF were treated with digoxin, and more patients with CHD were treated with beta-adrenergic blocking agents. Second, the patients with CHF were older than the patients with CHD (p = 0.004).
Characteristics of small resistance arteries from patients with CHF and CHD
There were no significant differences between the normalized diameters of the arteries from the different patient groups. Similarly, there were no differences in contractile responsiveness or endothelium-dependent vasodilator responsiveness to KPSS, NE and ACh, respectively, between experimental groups (unpaired ttest).
AI concentration response curves: effect of enalaprilat and chymostatin
Responses to AI in arteries from patients with CHF are shown in Figure 1. Angiotensin I induced a dose-dependent contraction in subcutaneous resistance arteries from CHF patients (n = 6) with a threshold concentration of 5.03 (3.82) × 10−10M and a maximum at 0.1 μM in the absence of inhibitors. In the presence of enalaprilat (n = 6), the threshold concentration for AI was 2.44 (3.86) × 10−8M, representing a fifty-fold increase compared with control. In the presence of chymostatin (n = 6), the threshold was 0.56 (1.20) × 10−8M. However, in the presence of the combination of enalaprilat and chymostatin (n = 8), the threshold was 0.77 (1.07) × 10−6M, representing a 1,500-fold increase compared with control. Thus, enalaprilat alone induced a significant shift to the right of the dose-response curve to AI (p < 0.05), but no inhibition was observed with chymostatin alone (p > 0.05) compared with vehicle. The combination of enalaprilat and chymostatin significantly inhibited the response to AI (p < 0.001) compared with vehicle. The shift to the right of the dose-response curve observed in the presence of the combination of enalaprilat and chymostatin was significantly greater than that obtained with enalaprilat alone (p < 0.01).
Responses to AI in arteries from patients with CHD are shown in Figure 2. Angiotensin I induced a dose-dependent contraction in subcutaneous resistance arteries from patients with CHD (group 1, n = 6) with a threshold concentration of 2.76 (3.36) × 10−9M and a maximum response at 0.1 μM in the absence of any inhibitors. The thresholds in the presence of enalaprilat (n = 6), chymostatin (n = 6) and combination (n = 8) were, respectively, 2.20 (3.85) × 10−8M, 1.32 (1.33) × 10−9M and 2.90 (2.70) × 10−7M. Thus, no inhibition was observed in the presence of chymostatin alone (p > 0.05) compared with vehicle. In contrast with the response in vessels from CHF, no inhibition of the response to AI was observed in the presence of enalaprilat alone (p > 0.05) compared with vehicle. However, there was a very significant shift in the dose-response curve to the right when enalaprilat and chymostatin were combined (p < 0.001).
BK concentration response curves
In the arteries of patients with CHF, neither enalaprilat, chymostatin or the combination changed the maximum response to BK (p > 0.05). However, there was a trend toward potentiation of the response to BK in the presence of enalaprilat alone and enalaprilat in combination with chymostatin (Table 2). This did not reach significance (p > 0.05).
A similar situation was observed in the arteries of patients with CHD. Thus, there was a trend toward potentiation in the presence of both enalaprilat and the combination of enalaprilat and chymostatin (Table 2). Again, this did not reach significance. There was no change in the maximum response to BK in the presence of any inhibitor.
ACh concentration response curves
There was no effect of either inhibitor, alone or in combination, on responses to ACh in arteries of patients with CHF and CHD. No difference was observed between the response to ACh in the absence of inhibitors in arteries from patients with CHF compared with those from patients with CHD.
AII concentration response curves
No significant inhibition of the AII response was observed with either inhibitor alone, or in combination, in arteries from patients with CHF and CHD (data not shown).
Study of the effect of AII receptor blockade on responses to AI
In the presence of losartan, the response of subcutaneous resistance arteries to AI was completely abolished (n = 7, data not shown).
In this study we have been able to confirm that a non-ACE pathway that generates AI from AII exists in human small subcutaneous resistance arteries. Our findings support those of Voors et al. (5)in internal mammary arteries taken from patients with CHD and also extend our previous findings in small resistance arteries from healthy volunteers to patients with CHD (6). More importantly, however, we now report on patients with CHF receiving long-term treatment with ACE inhibitors—a group in whom intense and persistent blockade of the renin-angiotensin system is thought to be therapeutically desirable.
Our findings in small resistance arteries from patients with CHD are qualitatively very similar to our earlier observations in younger healthy subjects. Collectively, the available data in healthy volunteers and patients with CHD show that neither an ACE inhibitor alone, nor a chymase inhibitor alone, substantially inhibits the conversion of AI to AII in small arteries, medium-sized arteries or veins (5,6). Only the combination of an ACE inhibitor and a chymase inhibitor effectively blocks AI to AII conversion. These findings imply that both the limbs of the dual pathway for conversion of AI to AII have a high capacity and that substrate can normally be rapidly and completely shunted through one limb, should the other be blocked.
Responses to AII, BK and ACh
It is important to note that the response of small resistance vessels to AII was not affected by enalaprilat, chymostatin or the combination of the two. Thus, there is no evidence that potentiation of the action of vasodilator substances by any of these inhibitors indirectly influences the response to AI. We have previously demonstrated that enalaprilat potentiates the response to BK in resistance arteries from normal subjects (6). In these studies we were able to confirm that, in both CHF and CHD subjects, there was a trend toward potentiation of the dilator response to BK in the presence of enalaprilat—confirming that ACE is involved in the degradation of this peptide and that enalaprilat can inhibit this. There was no evidence that chymostatin had any effect on responses to BK.
We also showed that vessels from patients with CHF and CHD exhibited a vasodilator response to ACh. There was no apparent difference between the patient groups, suggesting that endothelial function in CHF does not differ from CHD. However, in the absence of a normal control group, it is not possible to make any inference about vascular endothelial function in these patients with cardiovascular disease compared with healthy subjects. Our results do raise the possibility that previous reports of endothelial dysfunction in CHF may, in fact, reflect endothelial dysfunction in CHD rather than in CHF per se.
Responses to AI in arteries from patients with CHF and CHD
In contrast with the responses seen in arteries from CHD patients, the response to AI in arteries from patients with CHF proved to be quite different. Contrary to our original expectation, we did not find “up-regulation” of the chymostatin-sensitive pathway in these subjects and instead found that the small resistance arteries from these patients were more sensitive to ACE inhibition than those from healthy volunteers or patients with CHD. It is unlikely that this finding is due to the minor age and therapy differences between CHF and CHD patients. There are, however, at least two potential explanations.
The first possibility is that the chymostatin-sensitive non-ACE pathway is actually down-regulated in patients with CHF treated with an ACE inhibitor. Why this should occur is not clear, though it could occur if the ACE pathway is up-regulated. The second possibility, that ACE activity is induced, could also explain our findings in CHF, and it is known that both CHF and ACE inhibitor treatment, experimentally, can induce ACE. Even if this were the case, however, AI should be diverted through the alternative limb of the dual pathway when an ACE inhibitor is present. Our findings, therefore, imply that the chymase pathway is down-regulated alone or in conjunction with induction of ACE in CHF.
Our findings differ from those obtained by Wolny et al. (14)who studied the responses to AI in coronary arteries from patients with CHF treated with ACE inhibitors. In these vessels, cilazaprilat was unable to inhibit the response to AI; but chymostatin inhibited it by 78%, and the combination of chymostatin and cilazaprilat inhibited the response by 97%. Thus, it would appear that, in coronary arteries, the chymase pathway is predominant though the incremental inhibition obtained by combining chymostatin and cilazaprilat suggests that some of the conversion of AI to AII is mediated by ACE. It is possible that the results obtained by Wolny et al. reflect a difference in the physiology of coronary arteries compared with resistance arteries. It is possible that the relative contribution of ACE and chymase to AII generation varies between subcutaneous resistance arteries and coronary arteries. Furthermore, the effect of the inhibitors was only studied at 1 μM AI, which is supraphysiological, and not over a whole concentration range. It is, therefore, difficult to make a direct comparison between their results and ours.
One limitation of our study is that we are not able to distinguish between the effect of ACE inhibition and the presence of CHF on our findings. This question could be resolved in two ways: one approach might be to study arteries obtained from ACE inhibitor naive patients with CHF and a second would be to study arteries from patients with CHD before and after ACE-inhibition. Despite this limitation, however, we feel that our findings are important since they suggest that physiological escape from ACE inhibition can occur in a relevant clinical situation.
Even though the chymase pathway may be less active in CHF, we did find an incremental and statistically significant inhibition of AI to AII conversion with the combination of chymostatin and enalaprilat compared with enalaprilat alone in these patients. This finding provides a rationale for using drugs that inhibit the action of AII directly in the treatment of patients with CHD and CHF; an alternative, and potentially worthwhile approach, would be the development of agents that inhibit non-ACE AII generating pathways.
☆ Dr. Petrie was the recipient of a British Heart Foundation Junior Research Fellowship (FS/97031:1997), and Dr. Padmanabhan was the recipient of a Wellcome Trust Junior Research Fellowship. Also supported by the Medical Research Council (Programme Grants held by J.J.V.M. and J.M.C.C.) and the Chief Scientist Office of the Scottish Executive (project grant held jointly by J.J.V.M., J.M.C.C, N.P. and M.C.P.).
- angiotensin converting enzyme
- angiotensin I
- angiotensin II
- coronary heart disease
- chronic heart failure
- Kreb’s solution with KCl substituted for NaCl on an equimolar basis
- left ventricle, left ventricular
- left ventricular ejection fraction
- renin-angiotensin-aldosterone system
- Received August 4, 1999.
- Revision received November 8, 2000.
- Accepted December 13, 2000.
- American College of Cardiology
- Urata H.,
- Healy B.,
- Stewart R.W.,
- Bumpus F.M.,
- Husain A.
- Urata H.,
- Kinoshita A.,
- Misono K.S.,
- Bumpus F.M.,
- Husain A.
- Kinoshita A.,
- Urata H.,
- Bumpus F.M.,
- Husain A.
- Padmanabhan N.,
- Jardine A.G.,
- McGrath J.C.,
- Connell J.C.
- Cleland J.G.F.,
- Massie B.M.,
- Packer M.,
- et al.
- Rousseau M.F.,
- Konstam M.A.,
- Benedict C.R.,
- et al.
- Wolny A.,
- Clozel J.P.,
- Rein J.,
- et al.