Inhibition of the angiotensin converting enzyme by perindoprilat and release of nitric oxide

A]H 1995; 8:15-65

Inhibition of the Angiotensin Converting Enzyme by Perindoprilat and Release of Nitric Oxide Barnabas Desta, Paul M. Vanhoutte, and Chantal M. Boulanger

Experiments were designed to investigate the mechanism underlying the endothelium-dependent relaxations to perindoprilat, a converting enzyme inhibitor, in canine coronary arteries previously exposed to bradykinin. Rings suspended in organ chambers were exposed to bradykinin for 3 min and washed extensively for 150 min. In rings previously exposed to the peptide, bradykinin induced relaxations which were augmented in the presence of perindoprilat; this response was not affected by indomethacin, but nitro-t-arginine induced a rightward shift of the relaxation to the peptide without affecting its maximal effect. In canine coronary arteries previously exposed to the peptide, perindoprilat caused endothelium-dependent relaxations (lCso = 7.83), which had been observed previously at concentrations where the converting enzyme inhibitor did not augment the response to bradykinin. Carboxypeptidase B, but not

aprotinin, impaired the relaxation to perindoprilat, suggesting a contribution of bradykinin. The relaxation to perindoprilat was not affected by the B1 antagonist Leu8 -des-Arg9 -bradykinin. However, the bradykinin Bz antagonist HOE-140 displayed a noncompetitive antagonism against the response to perindoprilat. The response to the converting enzyme inhibitor was not affected by indomethacin but was impaired significantly by nitro-t-arginine. The present findings suggest that in canine coronary arteries previously exposed to bradykinin, the relaxation to perindoprilat is mediated mainly by endothelium-derived nitric oxide. In addition, the response to perindoprilat may be due to factors other than just protection of bound bradykinin from degradation. Am J Hypertens 1995;8:15-65

he response of the vascular smooth muscle to vasoactive agents is modulated by the release of relaxing factors from endothelial cells. 1-4 The best-characterized relaxing factors are prostacyclin and nitric oxide, which is produced by the metabolism of L-arginine by the enda-

thelial nitric oxide synthase. A third relaxing substance of unknown nature (endothelium-derived hyperpolarizing factor) is produced by endothelial cells and causes endothelium-dependent hyperpolarization of the smooth muscle. s Angiotensin-converting enzyme inhibitors are used for the treatment of hypertensive patients with high and low plasma renin activity. Their bloodpressure-lowering effect likely results both from the prevention of the formation of active angiotensin II from inactive angiotensin I and from the concommitant inhibition of the degradation of locally produced bradykinin by converting enzyme. 6 Indeed, in isolated blood vessels, angiotensin converting enzyme inhibitors such as perindoprilat augment the endo-

T

From the Center for Experimental Therapeutics, Baylor College of Medicine, Houston, Texas. Paul M. Vanhoutte's present address is Institut de Recherches Intemationales Servier, 6 place des Pleiades, 92415 Courbevoie, France. This work was supported in part by Grant HL 35614 from the National Institutes of Health. Address correspondence and reprint requests to Chantal M. Boulanger, PhD, Center for Experimental Therapeutics, Room 826E, Baylor College of Medicine, Houston, Texas, 77030.

© 1995 by the American Journal of Hypertension, Ltd.

KEY WORDS: Nitric oxide, carboxypeptidase, HOE140, Bz receptors, Ht receptors.

0895-7061/95/$9.50 0895-7061 (95)OOO26-L

2S

DESTA ET AL

thelium-dependent relaxation evoked by bradykinin, a response which is mediated by activation of the endothelial bradykinin B2 receptors and the subsequent release of nitric oxide and hyperpolarizing factor. 3,7-11 In the canine coronary artery, the converting enzyme inhibitor trandolaprilat evoked potent endothelium-dependent relaxations provided the preparations were exposed previously to bradykinin but not to acetylcholine 12; preliminary findings suggest that the same is true for perindoprilat. The present experiments were designed to investigate the mechanism of release of endothelial mediators by perindoprilat. MATERIALS AND METHODS Organ Chambers Experiments The experiments were performed on the left circumflex or the left anterior descending coronary artery of mongrel dogs (15 to 30 kg; either sex). All procedures followed were in accordance with the guidelines of the Animal Protocol Review Committee of Baylor College of Medicine. The dogs were anesthetized with pentobarbital sodium (15 to 30 mg/kg, intravenously). The hearts were excised and placed into cold modified KrebsRinger bicarbonate solution of the following composition (mmoUL): NaCl 118.3; KCl 4.7; MgS0 4 1.2; KH2P04 1.2; CaCl2 2.5; NaHC03 25.0; edetate calcium disodium, 0.026; and glucose 11.1 (control solution). The arteries were cleaned of adherent connective tissue and cut into rings (4 to 5 mm long). In some preparations, the endothelium was removed by gently rubbing the intimal surface with a small forceps. In the remaining rings, care was taken not to touch the inner surface of the blood vessel. The rings were suspended in organ chambers, which contained 25 mL of control solution (37'C) aerated with a mixture of 95% O2 and 5% CO2 (pH 7.4); they were connected to force transducers (Statham Universal UC2 or Grass FT 03C, Quincy, MA) to record changes in isometric force. Before the actual experiments began, the preparations were stretched progressively and exposed to KCl (45 mmoUL) at each level of stretch, until the optimal point of the lengthactive tension relationship was reached. After this procedure, the rings were allowed to equilibrate for 45 minutes. All rings were then exposed to KCl (60 mmollL) to determine their maximal responsiveness. Then, the presence or absence of endothelial cells was confirmed by the presence or absence of responses to either bradykinin (10- 8 to 10- 6 moUL) or acetylcholine (10 -6 mol/L), respectively. The responses were monitored for 3 min. Preparations were considered to be with endothelium when the relaxation to either bradykinin or acetylcholine was greater than 95%. Rings without endothelium were included in these results only if no relaxation to

AJH-MAY1995-VOL. 8, NO.5, PART 2

bradykinin or acetylcholine (both 10- 6 mollL) was observed during contractions to prostaglandin F2a · Unless otherwise stated, the preparations were exposed to bradykinin (10- 6 moUL) prior to investigating the effects of perindoprilat or bradykinin. The preparations were washed extensively (at least 10 times during 100 min) and were equilibrated for 45 min in control solution or in the presence of blockers. Then they were exposed to increasing concentrations of perindoprilat (10- 11 to 10- 4 moUL) during contraction to prostaglandin F2a (2 to 6 x 10- 6 moUL, to match the level of contraction between preparations and to reach about 80% of the response to 60 mollL KCI). Chemicals The following drugs were used: acetylcholine HCI, bradykinin, indomethacin, Leu8-DesArg9 -bradykinin, (Sigma Chemical, St. Louis, MO); aprotinin and DL-2-mercaptomethyl-3-guanidinoethylthiopropanoic acid (mergepta; Calbiochem, La Jolla, CA); nitro-L-arginine (Aldrich Chemical Company, Milwaukee, WI), perindoprilat (Institut de Recherches Internationales Servier, Courbevoie, France); and HOE-140 (Hoechst-Roussel Laboratories, Somerville, NJ). Drug concentrations are expressed as final molar concentrations in the bath solution. All drugs were prepared daily and dissolved in distilled water except indomethacin (which was dissolved in 10 mL distilled water containing 5 x 10- 3 mollL Na2C03 and was sonicated before use). The bradykinin solution was prepared by diluting frozen aliquots of a stock solution of the peptide (-20°C; 10- 3 mollL) into distilled water. Statistical Analysis Experiments were performed on rings from the same animals studied in parallel; n represents the number of dogs from which coronary arteries were obtained. For each preparation, the relaxation is expressed as percent inhibition of the contraction to prostaglandin F2a (2 to 6 x 10- 6 moIlL). ICoo represents the negative logarithm of the concentration of perindoprilat inducing 50% inhibition of the contraction to prostaglandin F2a · Results are given as means ± SEM. Statistical evaluation was done by Student's t test for paired observations. When more than two means were compared, a two-way analysis of variance was performed. Means were considered to be statistically significantly different when P < .05. RESULTS Bradykinin In canine coronary artery with (but not in those without) endothelium, bradykinin caused relaxations in the presence of mergepta (10- 5 mollL), which were augmented by perindoprilat (10- 7 to 10- 6 mollL; Figure 1). The IC50 of bradykinin increased significantly from 8.92 ± 0.19 to 9.93 ± 0.136 and 10.11 j: 0.18 in the presence of 10- 7 and 10-

AJH-MAY 1995-VOL. 8, NO.5, PART 2

ENDOTHELIUM·DEPENDENT RELAXATION TO PERINDOPRILAT

35

moUL perindoprilat, respectively (n = 5). Lower concentrations of the converting-enzyme inhibitor did not augment the response to bradykinin (ICso at 10- 8 moUL perindoprilat: 9.38 ± 0.22). The response to the peptide was not affected by indomethacin (10- 5 moll L) (data not shown). Nitro-L-arginine (NLA; 10- 4 moUL) shifted the concentration response curve of bradykinin to the right, with a significant change in ICso (9.62 ± 0.19 in control v 8.53 ± 0.29; n = 6) without inhibiting the maximal relaxation to the peptide (control: 100%; NLA: 89.7 ± 7.3%; n = 6).

z

o

§

20

w

a::

-12

-10

-11

-8

-9

-7

BRADYKININ, (logM) FIGURE 1. Effects of increasing concentrations of perindoprilat on relaxations evoked by bradykinin in canine coronary arteries with endothelium. All experiments were performed in the presence of mergepta (10- 5 mol/L), in control conditions (e), or in the presence of perindoprilat 10- 8 moUL (.), 10- 7 moUL (J.), or 10- 6 moUL (.) (n = 5). The relaxation is expressed as percent inhibition of the contraction evoked by prostaglandin F20t (PGF2ot). Data are given as mean ± SEM. The asterisks denote aSignificant effect of the converting enzyme inhibitor.

ENDOTHELIUM, •

o

••

Perindoprilat In preparations previously exposed to bradykinin, perindoprilat evoked endotheliumdependent relaxations (Figure 2). The ICso of perindoprilat was 8.30 ± 0.15 (n = 6). The response to the converting enzyme inhibitor was reproducible with time in the same preparations; however, the relaxation evoked by perindoprilat (10- 6 mollL) was significantly reduced, from 88.5 ± 11.5% to 48.4 ± 26.7% (n = 4) during the second contraction response curve to the converting enzyme inhibitor. In rings with endothelium that had been exposed to acetylcholine (10- 6 moUL), the subsequent effect of the converting enzyme inhibitor was not significantly different from that observed in rings without endothelium (n = 6). Perindoprilat also induced significant endothelium-dependent relaxation of preparations previously exposed to bradykinin at concentrations lower than 10- 6 moUL (10- 8 and 10- 7 mollL) (Figure

EXPOSURE TO:

•

WIth

br.dykinin 10 4 M

• 0

EXPOSURE TO BRADYKININ:

without

•

.cetylcholin.10 4 M

~ 100

•

...Cl

.li ...

0

0

z

~

«a: z

~

z

0

~ 0

60

«a: z

~

0

.z U

~

41l

0

20

z 0

0

~

~ W

a:

10 4 M

Cl.

80

0

U

10"M

•

Cl

Cl. ~

100

10 4 M

...

0~

n0 8

~

§w

0

·9

-8

·7

-8

PERINOOPRILAT (100M)

a:

...8

-9

-8

·7

-8

PERINOOPRILAT, (109M)

FIGURE 2. Left, Endothelium-dependent relaxation to perindoprilat in canine coronary arteries. Response to perindoprilat in rings with (closed symbols) and without (open symbols) endothelium to canine coronary arteries which had been exposed to either acetylcholine (.; 10- 6 mol/L> or bradykinin (e; 10- 6 mol/L). Right, Effect of previous exposure to increasing concentrations of bradykinin on the relaxation to perindoprilat. Rings with endothelium were briefly exposed to either 10- 8 moUL (.),10- 7 moUL ('f) or 10- 6 moUL (e) bradykinin and then washed extensively before examining the response to the converting enzyme inhibitor. The relaxation is expressed as 1!"~e~t inh/~ition of the concentration evoked by prostaglandin F2u (PGF2u)' Data are given as mean ± SEM. The asterisks denote a Significant difference between rings exposed briefly to bradykinin and those exposed to acetycholine.

4S

AIH-MAY 1995-VOL. 8, NO.5, PART 2

DESTA ET AL

2). If Des-Arg9-bradykinin (10- 6 mollL, a preferential 8 1 agonist) was used instead of bradykinin, perindoprilat caused minimal relaxations of coronary arteries with endothelium (data not shown). Aprotinin (100 U/mL, an inhibitor of vascular kallikreins) did not affect significantly either the ICso (control: 8.34 ± 0.20; with aprotinin: 8.58 ± 0.19; n = 6) or the maximal relaxation (control: 93.1 ± 5.5%; with aprotinin: 98.7 ± 1.3; n = 6) to perindoprilat. However, incubation of the preparations for 20 min with carboxypeptidase B (5 U/mL, to degrade kinins) prior the addition of perindoprilat abolished the response to the converting enzyme inhibitor without altering that evoked by bradykinin (10- 6 mollL) (Figure 3). The relaxation to perindoprilat was examined during contraction to prostaglandin F2a in rings of coronary artery with endothelium, under control conditions and in the presence of either nitro-L-arginine (10- 4 mollL, to prevent the formation of nitric oxide) or indomethacin (10- 5 mollL, to prevent the formation of vasoactive prostanoids). Nitro-L-arginine, but not indomethacin, significantly impaired the endothelium-dependent relaxations to perindoprilat (Figure 4). The response to perindoprilat was compared under control conditions and in the presence of increasing concentrations of the bradykinin B2 antagonist HOE~

1001.----..-===i~~~---fi--o--Q.______O I

u. C) Q.

Z

*

80

*

*

o

i= ~

0::

...Z

60

g 40 .".

Z

o

~

UJ

20

0

0::

·10

·9

-8 ·7 ·6 PERINOOPRILAT, (log M)

FIGURE 3. Effect of carboxypeptidase B (0, 5 U/mL, 20-min exposure) on relaxations evoked by perindoprilat in preparations with endothelium previously exposed to bradykinin (10- 6 mol/L). Perindoprilat evoked full relaxation under control conditions (e). The relaxation is expressed as percent inhibition of the contraction evoked by prostaglandin Fz", (PGF 2 ",). Data shown as mean :t SEM. The asterisks denote a significant effect of carboxypeptidase B.

~

100

N U.

C) Q.

Z

80

0

~

0

...

6

0 0

40

< 0:: Z

0~

Z

0

20

~

< X

<

-oJ

W

0

0::

-10

-9

-8

-7

~

PERINOOPRILAT, (logM) FIGURE 4. Relaxation to perindoprilnt in rings if canine coronary artery with endothelium, exposed previously to bradykinin (10- 6 mol/L). Experiments were performed under control conditions (e) and in the presence of either indomethacin (10- 5 mol/L; 0) or nitro-L-arginine (10- 4 mal/L; 0). Responses to perindoprilat and bradykinin are expressed as percent inhibition of the contraction evoked by prostaglandin F2", (PGF2",). Data are given as mean :t SEM. The asterisks denote a significant effect of the inhibitors.

140 (10- 11 to 10- 9 mollL) or the bradykinin 8 1 receptor antagonist, Leu8-Des-Arg9-bradykinin (10- 7 moll L). The antagonists were not present during the exposure of the preparations to the peptide at the beginning of the experiment. HOE-140 induced a significant rightward shift of the concentration-response curve to perindoprilat, and significantly depressed the maximal relaxation to the converting enzyme inhibitor (Figure 5). The bradykinin 8 1 receptor antagonist (Leu1f-Des-Arg9-bradykinin; 10- 7 mollL) did not significantly affect the response to perindoprilat (data not shown). SIN.1 Perindoprilat (10- 6 mollL) did not affect significantly the endothelium-independent relaxations to SIN-1 in canine coronary arteries without endothelium (ICso control: 7.96 ± 0.07; ICso in the presence of perindopriIat: 7.93 ± 0.10; n = 5). DISCUSSION The present study shows that perindoprilat induces endothelium-dependent relaxations in the isolated canine coronary artery when the preparations have been exposed briefly to bradykinin, but not acetylcholine, prior to the experiment, and confirms the observations made with trandolaprilat. 12 Although the preparations were washed extensively to remove

ENDOTHELIUM.DEPENDENT RELAXATION TO PERINDOPRILAT

A]H-MAY1995-VOL. 8, NO.5, PART 2

tl

N

100

U. C) Q.

0

I-

80

Z 0

t=

U

60

~

I-

Z 0

40

U 0~

Z t= 0

~

....J

"=6

20

0

UJ

0::

-11

-10

-9

-8

-7

-6

PERINDOPRILAT, (log M) FIGURE 5. Effect of HOE-140, a preferential bradykinin Bz receptor antagonist, on the relaxation evoked by perindoprilat in rings of coronary artery with endothelium exposed previously to bradykinin (10- 6 mollL), The response to perindoprilat was examined under control conditions (e) and in the presence of increasing concentration of HOE-140 (.: 10- 10 moUL, +: 10- 9 mollL, and 'Y: 10 - 8 mollL). Relaxation is expressed as percent inhibition of the contraction evoked by prostaglandin FZat (FGF2at ). Data are shown as mean ± SEM. The asterisks denote a significant effect of HOE-140.

the peptide, a likely explanation is that the endothelium-dependent relaxation to the converting enzyme inhibitor is mediated by amounts of bradykinin remaining bound to the preparations, and protected from degradation by angiotensin converting enzyme. Indeed, the response to perindoprilat was abolished by carboxypeptidase B, which removes carboxyterminal arginine or lysine from peptides and therefore inactivates bradykinin. In addition, the potency of the response (to judge from the ICso) to perindoprilat increased with the concentration of bradykinin used to first stimulate the preparations. In the canine coronary artery, the endotheliumdependent relaxation evoked by bradykinin is mediated by the release of both nitric oxide and endothelium-dependent hyperpolarizing factor following activation of bradykinin B2 receptors. 7,lO The response to the peptide is potentiated by the converting enzyme inhibitor perindoprilat8 (this study). In canine coronary arteries, perindoprilat augments the level of cyclic GMP caused by activation of the soluble guanylate cyclase by nitric oxide generated upon stimulation with bradykinin,1O without affecting the response of the smooth muscle to nitric oxide8 (this study). These data show that perindoprilat augments

S5

the release of nitric oxide evoked by bradykinin, most likely by preventing the degradation of the peptide by angiotensin converting enzyme; perindoprilat also increases the endothelium-dependent hyperpolarizations caused by bradykinin. 10•13 However, the angiotensin converting enzyme inhibitor fails to directly affect the vascular tone, except in the presence of bradykinin exogenously added to the organ chamber. 8 •1o,14 If perindoprilat were to induce endotheliumdependent relaxations by protecting endogenously bound bradykinin from degradation, one could assume that the response to the converting enzyme inhibitor is mediated by both nitric oxide and endothelium-dependent hyperpolarizing factor and is caused by the activation of bradykinin B2 receptors. However, it appears that perindoprilat preferentially releases nitric oxide from the canine coronary artery. This interpretation is based on the fact that nitro-Larginine, an inhibitor of nitric oxide synthase, markedly decreases the maximal relaxation evoked by perindoprilat while it minimally affects that induced by bradykinin. Therefore, if endogenously bound bradykinin were to mediate the response to perindoprilat, its concentration at the level of the receptors would have to be below the threshold concentration of the peptide triggering the release of endothelium-derived hyperpolarizing factor. The fact that HOE-140, a preferential bradykinin B2 receptor antagonist, displays a noncompetitive antagonism towards the relaxation induced by perindoprilat may be explained best by the indirect agonistic properties of the converting enzyme inhibitor. IS In addition, the activation of bradykinin B1 receptors can be ruled out since Leu8 Des-Arg9 -bradykinin did not affect the response to perindoprilat. However, the interpretation that the relaxation evoked by perindoprilat is mediated by bradykinin bound to the tissues is difficult to reconcile with the fact that the converting enzyme inhibitor induces endothelium-dependent relaxations at concentrations where it does not augment the response to exogenous bradykinin. Indeed, perindoprilat (10- 8 mollL) caused about 60% relaxation, but failed to significantly augment the response to cumulative addition of bradykinin (10- 12 to 10- 7 mollL) in the same preparations. Interestingly, it has been proposed, in the rabbit coronary microcirculation, that the converting enzyme inhibitor ramiprilat may directly amplify the action of bradykinin at the receptor level rather than protecting the peptide from degradation. 16 In summary, although the present study does not allow further speculation concerning the exact mechanism underlying the release of endothelium-derived nitric oxide by perindoprilat, the results suggest that the endothelium-dependent relaxations evoked by the

AJH-MAY 1995-VOL. 8, NO.5, PART 2

6S DESTA ET At

coronary enzyme inhibitor may not be mediated only by the protection of endogenously bound bradykinin from degradation.

9.

ACKNOWLEDGMENTS We thank Drs. K. J. Morrison, T. Scott-Burden, and J-V. Mombouli for stimulating discussion, Ms. Cheila McCoo and Mr. DeWayne Coney for valuable technical assistance, and Ms. Melody Whitus for secretarial assistance.

REFERENCES 1. Furchgott RF, Zawadzki JV: The obligatory role of endothelial cells in the relaxation of arterial smooth muscle by acetylcholine. Nature 1980;288:373-376. 2. Furchgott RF, Vanhoutte PM: Endothelium-derived relaxing and contracting factors. FASEB J 1990;3:20072018. 3. Luscher TF, Vanhoutte PM: The Endothelium: Modulator of Cardiovascular Function. Baton Rouge, CRC Press 1990, pp 1-228. 4. Moncada S, Palmer RMJ, Higgs EA: Nitric oxide: physiology, pathophysiology and pharmacology. Pharmacol Rev 1991;43:109-142. 5. Nagao T, Vanhoutte PM: Endothelium-derived hyperpolarizing factor. Am J Resp Cell Mol Bioi 1993;8:1-6. 6. Vanhoutte PM, Auch-Schwelk W, Biondi ML, et al: Why are converting enzyme inhibitors vasodilators? Br J Clin PharmacoI1989;28:955-104S. 7. Vanhoutte PM, Boulanger CM, Vidal M, Mombouli J-V: Endothelium-derived mediators and the renin angiotensin system, in Robertson JIS, Nichols MG, (eds) The Renin-Angiotensin Systems. London, Gower Medical Publishing 1993, pp 29.1-29.15. 8. Kerth PA, Vanhoutte PM: Effects of perindoprilat on endothelium-dependent relaxation and contractions

10.

11.

12.

13.

14.

15.

16.

in isolated blood vessels. Am J Hypertens 1991;4: 2265-2345. Mombouli J-V, Nephtali M, Vanhoutte PM: Effects of the converting enzyme inhibitor cilazaprilat on endothelium-dependent responses. Hypertension 1991; 18(suppl II): II-22-II-29. Mombouli }V, Iliano S, Nagao T, et at: Potentiation of endothelium-dependent relaxations to bradykinin by angiotensin I converting enzyme inhibitors in canine coronary artery involves both endothelium-derived relaxing and hyperpolarizing factor. Circ Res 1992;71: 137-144. Schini VB, Boulanger C, Regoli D, Vanhoutte PM: Bradykinin stimulates the production of cyclic GMP via activation of B2 kinin receptors in cultured porcine aortic endothelial cells. J Pharmacol Exp Ther 1990; 252:1197-1201. Desta B, Nakashima M, Kirchengast M, et al: Previous exposure to bradykinin unmasks an endotheliumdependent relaxation to the converting enzyme inhibitor trandolaprilat in isolated coronary arteries. J Pharmacol Exp Ther 1995;272:885-891. Nakashima M, Mombouli JV, Vanhoutte PM: Endothelium-dependent hyperpolarization caused by bradykinin in human coronary arteries. J Clin Invest 1993;92:2867-2871. Vidal M, Joly G, Mombouli JV, et al: The endothelium and vascular effects of the angiotensin converting enzyme inhibitor trandolaprilat. J Cardiovasc Pharmacol 1994;23(suppl 4):51-55. Black JW, Jenkinson DH, Kenakin TP: Antagonism of an indirectly acting agonist: block by propranolol and sotalol of the action of tyramine on rat heart. Eur J Pharmacol 1980;65:1-10. Hecker M, P6rsti I, Bara AT, Busse R: Potentiation by ACE inhibitors of the dilator response to bradykinin in the coronary microcirculation: interaction at the receptor level. Br J PharmacoI1994;111:238-244.