Chronic Angiotensin-Converting Enzyme Inhibition and Endothelial Function of Rat Aorta
Abstract To determine whether chronic angiotensin-converting enzyme (ACE) inhibition produces functional changes in the aorta of normotensive rats, four groups of rats were studied in parallel for 6 weeks. Group 1 orally received ramipril 10 mg/kg per day for 6 weeks; group 2, ramipril 10 mg/kg per day for 4 weeks and then a cotreatment with ramipril and β2-kinin antagonist HOE140 500 μg/kg per day SC by injection for the remaining 2 weeks; group 3, hydralazine 100 mg/kg per day PO for 6 weeks; group 4 (control), subcutaneous injections of saline solution during the last 2 of 6 weeks. In aorta isolated from group 1 the relaxations induced by bradykinin, acetylcholine, and histamine were significantly potentiated compared with those of group 4. In group 3, despite a decrease in systolic blood pressure similar to that induced by ramipril treatment, the responses to these three endothelium-dependent vasodilators were not different from those of group 4. In group 2, bradykinin-induced relaxations were completely abolished whereas acetylcholine-induced and histamine-induced relaxations were similar to those of group 4. The inhibitory effect of the endothelium on serotonin-induced contractions was significantly increased in preparations of group 1 compared with those of groups 2 through 4. Indirect measurements of nitric oxide formation such as contractions evoked by NG-monomethyl-l-arginine (L-NMMA) and aortic cGMP content were also significantly enhanced in preparations from group 1 compared with those of groups 2 and 4. Moreover, because the relaxations to nitroglycerin and nitroprusside were similar in groups 1, 2, and 4 an alteration of the guanylate cyclase activity by ramipril treatment is quite unlikely. Thus long-term treatment with ramipril potentiates the endothelium-dependent responses in the rat aorta by enhancing nitric oxide availability. This effect seems to involve an inhibition of bradykinin breakdown facilitating nitric oxide release via endothelial β2-receptors.
Chronic ACE inhibition attenuates the endothelial dysfunction in experimental models of hypertension and atherosclerosis.1 2 In these studies this protective effect seems to be independent of the hypotensive action of ACE inhibitors.1 2 Moreover, it was demonstrated recently that even in normotensive rats chronic ACE inhibition potentiates the function of the vascular endothelium.3 4 In the coronary circulation of several species this potentiation of EDRF-mediated effects protects myocardial tissue against ischemia-reperfusion–induced injury5 6 7 and is abolished by the bradykinin β2-receptor antagonist HOE140.6 7 In the rat aorta however the existence of these endothelial bradykinin β2-receptors remains controversial,3 4 and the enhanced formation of EDRF and prostacyclin after long-term ACE inhibition could be related to an increase in flow due to peripheral vasodilation.8 9 If this is true other peripheral vasodilators such as hydralazine should have comparable action on the endothelial function. The present study was designed to assess whether long-term treatment with ramipril or hydralazine affects the rat aorta vasomotricity in a similar manner. The effect of a cotreatment with ramipril and the β2-receptor kinin antagonist HOE140 on the endothelial function also was investigated.
Male Wistar rats weighing 100 to 150 g each were housed in a controlled environment and given free access to standard rat chow and tap water. All animal procedures were approved by our university animal care and use committee.
In a first set of experiments four rat groups were treated in parallel for 6 weeks. The first group received ramipril 10 mg/kg per day in drinking water. Duration of treatment and dose of ramipril were chosen according to previous methods.3 4 In our preliminary experiments smaller doses (1 mg/kg per day for 6 weeks) slightly decreased the blood pressure but did not modify the endothelium-dependent responses. In the second group ramipril 10 mg/kg per day was given for 4 of 6 weeks to meet previously established experimental conditions4 at which we know that the vascular effects of ramipril are already present. Thereafter, a cotreatment with ramipril and HOE140 (500 μg/kg per day) was administered for the last 2 of the 6 weeks. HOE140 was injected subcutaneously twice daily according to the method of a previous study10 ; at this dose HOE140 completely blocks the hypotensive action of bradykinin. Moreover, at doses 10 times lower this drug already exhibits one half-life of protection against bradykinin-induced hypotension of more than 5 hours after one subcutaneous injection.11 The third group received hydralazine 100 mg/kg per day in drinking water for 6 weeks; from a previous study1 and our preliminary experiments this treatment is as potent as ramipril 10 mg/kg per day for reduction of blood pressure. The fourth group served as a control and received subcutaneous injections of saline solution (HOE140 solvent) during the last 2 of the 6 weeks. In a second set of experiments six rats were treated with HOE140 (500 μg/kg per day SC injection BID) for 2 weeks while six others served as controls and received saline injections.
Rat weight was assessed weekly; systolic blood pressure was measured by tail plethysmography (Apollo 179, ITTC, Life Science) under light anesthesia with diazepam 0.1 mg/kg IM. At the end of these treatment periods the rats were killed under ether anesthesia; before dissection blood samples were withdrawn to assess plasma renin activity by angiotensin I radioimmunoassay.12
Vasomotor Response Studies
The thoracic aorta was removed, cleaned of adhering fat and connective tissue, and cut into 3- to 4-mm-long rings. During the dissection utmost care was taken to protect the endothelial lining. In some preparations the endothelium was removed by gently rubbing the intimal surface with forceps. All rings were mounted under 1.5 g resting tension on stainless steel hooks in 20-mL organ baths. These organ chambers were filled with Krebs-Henseleit solution of the following composition (in mmol/L): NaCl 118.1, KCl 4.7, MgSO4 1.2, KH2PO4 1.2, CaCl2 2.5, NaHCO3 25, and glucose 5, aerated with a mixture of 95% O2/5% CO2 and kept at 37°C. Tension was measured isometrically with a transducer (Grass FT O3C) and recorded continuously with a transducer amplifier (Janssen Scientific Instruments) and a pen recorder. After 60 minutes of equilibration the rings were stretched progressively and exposed to 40 mmol/L KCl at each level of stretch until the optimal point of the length-tension relation was reached.
For the endothelium-dependent relaxations the concentration-response curves were constructed on rings contracted with phenylephrine, and the concentration of this amine was adjusted to obtain equivalent plateaus in all preparations (1.3 to 1.8 g tension). After the plateau was reached, the relaxing agent was added cumulatively so that each concentration was allowed to develop its maximal effect. Each cycle of contraction-relaxation was separated by an interval of 90 minutes. For each ring no more than three concentration-response curves were constructed and the maximal contractile response to depolarization was assessed; this response was induced by a modified Krebs’ solution in which NaCl was replaced by KCl, leading to a final isotonic concentration of 122.8 mmol/L KCl. To study contractions evoked by L-NMMA (10 μmol/L to 1 mmol/L) we added increasing concentrations of the drug in a cumulative fashion to quiescent preparations, and each concentration was allowed to develop its maximal effect. These L-NMMA–induced contractions were used as an indirect measurement of the basal nitric oxide formation.13
To test the effect of cyclooxygenase inhibition, two adjacent preparations were studied in parallel; one served as a control, while the other was incubated for 30 minutes with indomethacin (3 μmol/L) before the construction of the concentration-response curves.
Measurement of Tissue Levels of Guanosine 3′5′-Monophosphate
Unrubbed rings were prepared as for studies of vasomotor responses and incubated without tension in the Krebs’ solution aerated with a mixture of 95% O2/5% CO2 and kept at 37°C for 3 hours.
Thereafter, tissues were quickly taken out of the bath and snap-frozen in liquid nitrogen. Frozen tissue was homogenized for 5 minutes with a Polytron homogenizer in 6% ice-cold trichloroacetic acid, and the homogenate was centrifuged for 10 minutes at 2000g at 4°C. Hydrochloric acid 20 μL (1 mol/L) was added to each sample, and the trichloroacetic acid was extracted fourfold with 5 vol ether saturated with water. The samples were evaporated to dryness at 50°C under a stream of nitrogen and resuspended in 0.150 mL of 50 mmol/L sodium acetate (pH 5.8) containing bovine serum albumin (0.02%) and thimerosal (0.005%). Fifty milliliters of trichloroacetic acid was extracted in parallel to the samples and used to dilute the standards (2 to 512 fmol per assay) to establish a standard curve.
Aliquots (50 μL) of the extracted samples and of the standards were assayed for cGMP by use of an enzyme immunoassay14 (Amersham). The acetylation technique was used. Each assay was performed in duplicate.
The following drugs were used: acetylcholine chloride (Sterop), histamine, bradykinin, phenylephrine hydrochloride, L-NMMA (all from Sigma Chemical Co), indomethacin (Merck, Sharp & Dohme), serotonin (5-hydroxytryptamine maleate, Fluka), nitroglycerin (Nysconitrine, Bio-Therabel), and nitroprusside (Roche). Ramipril and HOE140 were generously provided by Hoechst AG. Stock solutions of the drugs were prepared in distilled water except phenylephrine and serotonin, which were dissolved in distilled water containing ascorbic acid 1 mmol/L, and indomethacin, which was dissolved in 70% ethanol.
Results are mean±SEM. The number of experiments is also the number of rats used. Contractions are expressed as percentages of maximal response to KCl 122.8 mmol/L (isotonic solution), whereas relaxations are expressed as percentages of inhibition of tension developed to phenylephrine. The AUC was calculated for each concentration-response curve and expressed in arbitrary units. Intergroup differences were tested by two-tailed unpaired t test or by one-way ANOVA followed by Fisher’s least significant difference test for multiple comparisons. Student’s t test for paired observations was applied to detect differences in the means within the same group. Significance was accepted at P<.05.
Comparison of Treatment With Ramipril Versus Hydralazine
Blood Pressure, Weight, and Laboratory Tests
After 6 weeks of treatment systolic blood pressure decreased from 125±4 to 105±3 mm Hg (n=9, P<.01) in the ramipril-treated group, from 129±5 to 115±3 mm Hg (n=9, P<.05) in the group cotreated with ramipril and HOE140, and from 126±3 to 100±3 mm Hg (n=8, P<.01) in the hydralazine-treated group. In the control group systolic blood pressure did not change significantly (from 125±5 to 132±4 mm Hg).
For the control group, ramipril-treated group, ramipril and HOE140–cotreated group, and hydralazine-treated group, weights before treatment were 112±6, 114±5, 114±6, and 113±7 g, respectively; at the end of the treatment period, they were 234±9, 250±9, 220±6, and 255±8 g (P=NS), respectively. Plasma renin activity was significantly increased in the ramipril-treated group and in the ramipril and HOE140–cotreated group: 55±5 ng/mL per hour (n=9) and 39±4 ng/mL per hour (n=9) versus 6±2 ng/mL per hour (n=10, P<.01) in the control group. In the hydralazine-treated group, plasma renin activity was 7±3 ng/mL per hour (n=8, P=NS versus control group).
The maximal relaxation to bradykinin was only 17.2±3.1% (n=10, percent of inhibition of phenylephrine-induced contraction) and 15.1±5% (n=8) in the control group and the hydralazine-treated group, respectively (Fig 1⇓). In the ramipril-treated group the bradykinin-induced relaxation was potentiated: the maximal response was 39±4.1% (n=9, P<.01 versus the control group) (Figs 1⇓ and 2⇓); the AUC was significantly decreased compared with the control and the hydralazine-treated groups (P<.01, Table 1⇓). In the ramipril and HOE140–cotreated group the bradykinin-induced relaxations were completely inhibited (Fig 1⇓).
The acetylcholine-induced and histamine-induced relaxations were also significantly enhanced in the ramipril group compared with the control group (Figs 3⇓ and 4⇓, Table 1⇑). In contrast, the AUC values in the ramipril and HOE140–cotreated group and the hydralazine-treated group were not significantly different from the control group (Table 1⇑). With indomethacin added (3 μmol/L, 30 minutes) the potentiation of relaxations evoked by bradykinin, acetylcholine, and histamine persisted in the ramipril-treated group (Table 1⇑).
Contractions Induced by Depolarization and Serotonin
The maximal response to KCl (122.8 mmol/L, isotonic solution) was similar in the four groups: 1.5±0.4 g (n=10) in the control group, 1.8±0.5 g (n=9) in the ramipril-treated group, 1.8±0.3 g (n=9) in the ramipril and HOE140–cotreated group, and 1.5±0.3 g (n=8) in the hydralazine-treated group.
The serotonin-induced contractions were significantly diminished in the ramipril-treated group compared with the control group (Fig 5A⇓): the AUC was 109±10 arbitrary units (n=9) versus 147±20 arbitrary units (n=10) in the control group (P<.05). In the ramipril and HOE140–cotreated group and the hydralazine-treated group the AUC values were 140±12 arbitrary units (n=9) and 152±9 arbitrary units (n=8), respectively (P=NS versus the control group).
When indomethacin 3 μmol/L (30 minutes) was added, these concentration-response curves were not significantly modified in the four groups (data not shown). Removal of the endothelium attenuated the difference between the ramipril-treated group and the control group: the AUC was 243±11 arbitrary units (n=9) versus 278±18 arbitrary units (n=10) in the control group (P=NS). In the ramipril and HOE140–cotreated group and the hydralazine-treated group the AUC values were 245±15 arbitrary units (n=9) and 271±16 arbitrary units (n=8), respectively (P=NS versus the control group). At 10 μmol/L of serotonin (Fig 5B⇑) however the response in both ramipril-treated group and the ramipril and HOE140–cotreated group was slightly decreased (P=.055) compared with the control group: 88.1±5.3% (n=9, percent KCl) and 93±4.5% (n=9) versus 107.1±6.8% (n=10).
Ramipril and the Nitric Oxide–Guanylate Cyclase Pathway
Endothelium-Dependent Contractions to L-NMMA
The inhibitor of nitric oxide formation L-NMMA evoked concentration-dependent contractions (Fig 6⇓); in the ramipril group (n=6) the maximal contraction averaged 21±4% of the maximal response to KCl (1.8±0.3 g) compared with 11±1% (percent KCl, 1.8±0.2 g) in the control group (n=6) and 9±2% (percent KCl, 1.7±0.3 g) in the group cotreated with ramipril and HOE140 (n=6) (P<.05 for both groups versus the ramipril-treated group). The AUC also significantly increased in the ramipril-treated group: 39±11 versus 20±2 arbitrary units in the control group (P<.05) and 15±5 arbitrary units in the group cotreated with ramipril and HOE140 (P<.05). In rubbed preparations the increase in tone induced by the l-arginine analogue L-NMMA was abolished.
Relaxations in Response to Nitrovasodilators
The concentration-response curves were constructed in rubbed rings to create optimal conditions.15 The relaxations to nitroglycerin (0.1 nmol/L to 0.3 μmol/L) (Fig 7⇓) did not differ statistically from each other in all three groups. The AUC was 480±14 arbitrary units in the ramipril-treated group (n=9), 489±21 arbitrary units in the control group (n=9), and 467±20 arbitrary units in the group cotreated with ramipril and HOE140 (n=9). Similar results were observed for nitroprusside-induced relaxations (0.1 nmol/L to 0.1 μmol/L); the AUC was 304±15 arbitrary units in the ramipril-treated group (n=9), 314±16 arbitrary units in the control group (n=9), and 296±15 arbitrary units in the group cotreated with ramipril and HOE140 (n=9) (P=NS, graph not shown).
Aortic cGMP content was 439.6±70.4 fmol/mg wet wt in the control group (n=6); it increased by 76% in the ramipril-treated group (774.3±99.8 fmol/mg wet wt, n=6, P=.05 versus control), whereas in the group cotreated with ramipril and HOE140 (n=6) it was 422.7±76.5 fmol/mg wet wt (P=NS versus control).
Effects of HOE140 Treatment
In the group treated with only HOE140 the concentration-response curves to acetylcholine, histamine, and serotonin were similar to the control group (Table 2⇓), whereas the bradykinin-induced relaxations were completely inhibited. Systolic blood pressure was not altered by the treatment (129±7 versus 124±6 mm Hg before treatment).
Our results show that long-term ACE inhibition potentiates the responses to all endothelium-dependent vasodilators tested. Blood pressure reduction does not play a major role in this potentiation of EDRF-mediated effects; indeed, there was no alteration in the endothelial function in the hydralazine-treated group despite a hypotensive action similar to that of ramipril. Therefore, the potentiation of EDRF-mediated effects by long-term ACE inhibition does not seem to be related to an increase in flow induced by a decrease in peripheral vascular resistance. In agreement with this hypothesis, the increase in endothelium-dependent relaxations was abolished in the group cotreated with ramipril and HOE140 despite a significant reduction of the blood pressure. Moreover, the absence of effect of treatment with only HOE140 on blood pressure and on endothelium-dependent responses (except those induced by bradykinin) (Table 2⇑) demonstrates that at the doses used the drug has selective antagonistic property on β2-receptors.
It is tempting to ascribe the ramipril-induced increase in endothelium-dependent relaxations to an upregulation of guanylate cyclase because reduced generation of angiotensin II may alter the activity of this nitric oxide target enzyme.16 17 However, this mechanism is quite unlikely because long-term ACE inhibition did not affect the responses to endothelium-independent activators of guanylate cyclase such as nitroglycerin and nitroprusside.
Under physiological conditions the effect of endogenously formed bradykinin is limited by the converting enzyme. If the breakdown of bradykinin is inhibited by converting enzyme inhibition, however, endothelium-derived bradykinin may accumulate extracellularly near the β2-kinin receptor.18 In the coronary circulation this enhanced bradykinin availability seems to facilitate the release of nitric oxide and prostacyclin and thereby to account for the potentiation of endothelium-dependent responses by ACE inhibition.6 In cultured endothelial cells such increased formation of nitric oxide has been demonstrated with ramiprilat19 ; via a bradykinin-dependent mechanism ACE inhibition19 increases intracellular calcium, which upregulates the constitutive nitric oxide synthase.20 Rat aorta is able to synthesize kinins from stored kininogen.21 Interestingly, our results demonstrate bradykinin-evoked relaxations and therefore the presence of β2-kinin endothelial receptors in the rat aorta. Because ramipril improves all endothelium-dependent relaxations a similar facilitation of nitric oxide release may occur in the present study. This mechanism also seems bradykinin dependent because this improvement of endothelial reactivity to muscarinic and histaminergic stimuli was abolished by the kinin β2-antagonist. In previous studies5 6 7 on coronary vessels, however, the effect of ACE inhibition was assessed only under short-term conditions, whereas long-term ACE inhibition may increase endothelial-converting enzyme mRNA and activity,22 which may attenuate the short-term effect on endogenous bradykinin during a longer treatment period. It is likely that the high doses of ramipril used cause an excess of inhibitor that can counteract this ACE induction.
This hypothesis of an enhanced nitric oxide availability also may explain the higher contraction evoked by L-NMMA. Indeed, the basal release of EDRF counteracts the inherent propensity of the vascular smooth muscle to develop intrinsic tone,23 and therefore the increase in tension of quiescent preparations induced by this nitric oxide synthase inhibitor must reflect the amount of nitric oxide released under basal conditions.13 24
In line with this interpretation, ramipril treatment resulted in enhanced aortic cGMP content, suggesting an increased production of the nucleotide in smooth muscle cells with the stimulus derived from the endothelium. Indeed, as mentioned above the similar responses to the nitrovasodilators allow us to exclude in smooth muscle an alteration of the guanylate cyclase activity by long-term ACE inhibition.
With regard to the inhibitory effect of ramipril treatment on serotonin-induced contractions, the potentiation of EDRF-mediated effects also seems to be the underlying mechanism. In rat aorta contraction to this amine is inhibited by the basal release of EDRF,25 and in the present study removal of the endothelium markedly attenuates this decrease in serotonin-induced contraction in the ramipril group. However, the persistence of a trend for a diminished response to this amine in rubbed preparations of the ramipril-treated group (Fig 5⇑) also suggests changes in smooth muscle cells. According to Auch-Schwelk et al26 the chronically reduced generation of angiotensin II induced by ACE inhibition may alter the serotonergic receptor in the vascular smooth muscle. In the present study this trend was nevertheless present for only one concentration of serotonin, and there were no significant changes in the AUC values.
Because indomethacin did not alter the ramipril-induced potentiation of the endothelium-dependent responses, the release of prostacyclin does not seem to intervene in the ramipril effects. In fact, according to Levy27 prostacyclin has no effect in the isolated rat aorta except that at high concentrations it contracts the preparations. The finding that the potentiation of all endothelium-dependent responses tested and the increase in aortic cGMP were abolished in the group cotreated with ramipril and HOE140 supports the concept of a bradykinin-dependent mechanism facilitating nitric oxide release via endothelial β2-receptors. Of note, this increase in nitric oxide availability is highly sensitive to this β2-kinin antagonist as the vascular effect of long-term administration of ramipril was abolished by the coadministration of HOE140 for only 2 weeks.
In hypertensive rats similar findings have been reported, namely long-term ACE inhibition enhanced endothelium-dependent relaxations and aortic cGMP; the latter was abolished by a coadministration of HOE140.28 However, much lower doses of ramipril were necessary to increase the endothelium-dependent responses.28 These two settings are nevertheless different: On one hand we have some functional defect and a restoration of a normal function; on the other hand we are faced with a normal endothelial function and a potentiation of EDRF-mediated effects. Moreover, it might be that long-term ACE inhibition may increase endothelium-converting enzyme mRNA and activity to a lower extent in the hypertensive strain. In line with this hypothesis, the influence of angiotensin II on converting enzyme gene expression appears lower in glial cells from spontaneously hypertensive rats than in those derived from Wistar-Kyoto rats.22 Therefore, to overcome converting enzyme induction excess of inhibitor should be necessary, mainly in normotensive animals.
In conclusion, our results indicate that in rat aorta long-term ACE inhibition with ramipril potentiates endothelium-dependent responses by enhancing nitric oxide availability. This effect is not directly related to blood pressure reduction and seems to involve an inhibition of bradykinin breakdown facilitating nitric oxide release via endothelial β2-receptors.
Selected Abbreviations and Acronyms
|AUC||=||area under the curve|
|EDRF||=||endothelium-derived relaxing factor|
We thank Astra Pharma and Hoechst AG for the supply of ramipril and HOE140. This work was supported by research grants from “Fonds National de la Recherche Scientifique” (FNRS 9.4553.92), the Bekales Foundation, and “Fondation pour la Chirurgie Cardiaque.”
- Received December 6, 1994.
- Revision received January 20, 1995.
- Accepted August 10, 1995.
Clozel M, Kuhn H, Hefti F. Effects of angiotensin-converting enzyme inhibitors and hydralazine on endothelial function in hypertensive rats. Hypertension.. 1990;16:532-540.
Becker RHA, Wiemer G, Linz W. Preservation of endothelial function by ramipril in rabbits on a long-term atherogenic diet. J Cardiovasc Pharmacol. 1991;18(suppl 2):S110-S115.
Bossaller C, Auch-Schwelk W, Weber F, Götze S, Gräfe M, Graf K, Fleck E. Endothelium-dependent relaxations are augmented in rats chronically treated with the angiotensin-converting enzyme inhibitor enalapril. J Cardiovasc Pharmacol. 1992;20(suppl 9):S91-S95.
Zanzinger J, Zheng X, Bassenge E. Endothelium-dependent vasomotor responses to endogenous agonists are potentiated following ACE inhibition by a bradykinin-dependent mechanism. Cardiovasc Res.. 1994;28:209-214.
Ehring T, Baumgart D, Krajcar M, Hümmelgen M, Kompa S, Heusch G. Attenuation of myocardial stunning by the ACE inhibitor ramiprilat through a signal cascade of bradykinin and prostaglandins but not nitric oxide. Circulation. 1994;90:1368-1385.
Miller VM, Vanhoutte PM. Enhanced release of endothelium-derived relaxing factor(s) by chronic increases in blood flow. Am J Physiol.. 1988;255:H446-H451.
Bao G, Gohlke P, Qadri F, Unger T. Chronic kinin receptor blockade attenuates the antihypertensive effect of ramipril. Hypertension.. 1992;20:74-79.
Tschudi MR, Noll G, Arnet U, Novosel D, Ganten D, Lüscher TF. Alterations in coronary artery vascular reactivity of hypertensive Ren-2 transgenic rats. Circulation. 1994;89:2780-2786.
Steiner AL, Parker CW, Kipnis DM. Radioimmunoassay for cyclic nucleotides. J Biol Chem.. 1972;247:1106-1113.
Pohl U, Busse R. Endothelium-derived relaxant factor inhibits the effect of nitrocompounds in isolated arteries. Am J Physiol.. 1987;252:H307-H313.
Vesely DL. Angiotensin II stimulates guanylate cyclase activity in aorta, heart and kidney. Am J Physiol.. 1981;240:E391-E393.
Sumners C, Myers LM. Angiotensin II decreases cGMP levels in neuronal cultures from rat brain. Am J Physiol.. 1991;260:C79-C87.
Mombouli JV, Nephtali M, Vanhoutte PM. Effects of the non-sulfhydryl angiotensin I converting enzyme inhibitor cilazaprilat on endothelium-dependent responses in isolated canine arteries. Hypertension. 1991;18(suppl II):II-22-II-29.
Hecker M, Dambacher T, Busse R. Role of endothelium-derived bradykinin in the control of vascular tone. J Cardiovasc Pharmacol. 1992;20(suppl 9):S55-S61.
Oza NB, Schwartz JH, Goud HD, Levinsky NG. Rat aortic smooth muscle cells in culture express kallikrein, kininogen, and bradykininase activity. J Clin Invest.. 1990;85:597-600.
King SJ, Oparil S. Converting-enzyme inhibitors increase converting-enzyme mRNA and activity in endothelial cells. Am J Physiol.. 1992;263:C743-C749.
Lüscher TF, Vanhoutte PM. The Endothelium: Modulator of Cardiovascular Function. Boca Raton, Fla: CRC Press; 1990:1-215.
Martin W, Furchgott RF, Villani GM, Jothianandan D. Depression of contractile responses in rat aorta by spontaneously released endothelium-derived relaxing factor. J Pharmacol Exp Ther.. 1986;237:529-538.
Auch-Schwelk W, Duske E, Hink U, Betz M, Unkelbach M, Fleck E. Vasomotor responses in cyclosporin A–treated rats after chronic angiotensin blockade. Hypertension. 1994;23(pt 2):832-837.
Gohlke P, Lamberty V, Kuwer I, Bartenbach S, Schnell A, Linz W, Schölkens BA, Wiemer G, Unger T. Long-term low-dose angiotensin converting enzyme inhibitor treatment increases vascular cyclic guanosine 3′,5′-monophosphate. Hypertension.. 1993;22:682-687.