Functional evidence for alternative ANG II-forming pathways in hamster cardiovascular system

Hikaru Nishimura, Hendrik Buikema, Ovidiu Baltatu, Detlev Ganten, Hidenori Urata


Like human chymase, hamster chymase is an ANG II-forming enzyme, but pathophysiological roles of chymase are still unknown. We determined the functional conversion of ANG I and [Pro11, d-Ala12]ANG I, a chymase-selective substrate, to ANG II in the hamster cardiovascular system. ANG I and [Pro11, d-Ala12]ANG I produced similar dose-dependent pressor responses in conscious hamsters. Captopril and CV-11974, an ANG II type 1 (AT1)-receptor antagonist, inhibited the responses to ANG I; in contrast, the pressor responses to [Pro11,d-Ala12]ANG I were suppressed only by CV-11974. In the isolated aorta, captopril suppressed ANG I-induced contraction by 84%; administration of captopril with either chymostatin or aprotinin eliminated the contraction. [Pro11, d-Ala12]ANG I-induced contraction was not affected by captopril but was attenuated by chymostatin (71%) and aprotinin (57%). CV-11974 abolished the responses to both substrates, whereas PD-123319, an AT2-receptor antagonist, had no effect. In homogenates of the aorta and heart, soybean trypsin inhibitor-inhibitable ANG II formation predominated over captopril- or aprotinin-inhibitable ANG II formation. These data suggest that [Pro11,d-Ala12]ANG I and part of ANG I were functionally converted to ANG II by chymase and other serine protease(s) in hamster vessels, inducing AT1-receptor-mediated vasoconstriction. Biochemical data supported a role for chymase in the alternative pathway.

  • chymase
  • serine proteases
  • angiotensin-converting enzyme
  • aorta

accumulating evidence supports the presence of the local renin-angiotensin system in the mammalian heart and blood vessels (1, 7). The conversion of ANG I to ANG II has been thought to be catalyzed mainly by angiotensin-converting enzyme (ACE). However, Urata et al. (30, 31) showed that the major ANG II-forming enzyme in the human left ventricular particulate fraction was chymase, a finding confirmed by others (2). Several studies suggest the presence of ANG II-generating pathways alternate to ACE in the human cardiovascular system. Despite long and adequate ACE inhibitor treatment, increased plasma ANG II levels were found in patients with progressive heart failure (25). Increased plasma ANG II formation during exercise and in ischemia was inhibited by nafamostat, a serine protease inhibitor, but not by captopril (19, 29). Because ACE is not inhibited by nafamostat, the observed effects of nafamostat are probably due to the inhibition of ACE-independent ANG II formation. In isolated human arteries the ANG I-induced vasoconstriction was more effectively inhibited by chymostatin than by captopril (24).

ACE-independent ANG II formation has also been reported in various hamster tissues such as the cheek pouch arterioles (5), the coronary artery (6), and the heart (9, 15). However, these functional studies failed to specify the responsible enzyme(s) in the ACE-independent pathways, because the evidence for the alternative pathways was based solely on the inhibitory effects of ACE inhibitors and ANG II subtype 1 (AT1)-receptor antagonists. Recently, a chymostatin-sensitive ANG II-generating enzyme in the hamster cheek pouch vascular tissues was purified and identified as chymase (27). Hamster chymase is mostly ANG II-forming enzyme (27), whereas rat chymase (rat mast cell proteinase I) mainly hydrolyzes the Tyr4—Ile5bond of ANG I to yield inactive fragments (16). Thus biochemical characteristics of hamster chymase resemble those of human chymase (31). However, functional evidence for chymase-dependent ANG II formation in hamsters is still lacking.

[Pro11,d-Ala12]ANG I, a selective substrate for chymase, was developed on the basis of the different substrate binding site of chymase and ACE (11, 14). This substrate is an inactive precursor that yields ANG II when incubated with chymase, but not with ACE (11, 14). Functional evidence for chymase-dependent ANG II formation was reported using this substrate. [Pro11, d-Ala12]ANG I produced pressor responses in conscious marmosets (18) and baboons (10). These responses were inhibited by AT1-receptor antagonists, but not by ACE inhibitors. [Pro11,d-Ala12]ANG I also induced a positive inotropic effect in isolated human atrial trabeculae (11). These findings suggest a potential role of chymase in the cardiovascular system.

The aim of this study was to investigate the functional contribution of ACE-independent ANG II-forming pathways in the hamster cardiovascular system, with ANG I and [Pro11,d-Ala12]ANG I as substrates. The effects of inhibitors of the renin-angiotensin system and serine proteases on the responses to these substrates were evaluated in conscious animals and in isolated aorta. Biochemical conversion of ANG I to ANG II was additionally assessed in the particulate fractions of the hamster heart and aorta in the absence or presence of several inhibitors.


Male Syrian hamsters (average body weight 127 ± 3 g) were housed in plastic cages and given standard hamster chow and water ad libitum. They were maintained in a quiet room at constant temperature (20–22°C) and under a 12:12-h light-dark cycle. After a 7-day acclimatization period, the following functional and biochemical studies were done. All experiments were performed in accordance with institutional guidelines of the Max Delbrück Center for Molecular Medicine for use of experimental animals.

Blood pressure measurements in conscious hamsters.

After anesthesia with 4% chloral hydrate (350 mg/kg ip), polyethylene catheters (PE-50; 0.58-mm ID, 7 cm long) were inserted into the right carotid artery and the right jugular vein for measurement of arterial pressure and injection of agents, respectively. The catheters, filled with heparinized saline (10 IU/ml), were exteriorized in the interscapular area. Hamsters were allowed to recover for 48 h with free access to food and water. The carotid catheter was connected to a standard water-filled blood pressure transducer (model 101021-2, TSE, Bad Homburg, Germany) positioned at the level of the heart. The transducer signal was preamplified before analog-to-digital conversion for data recording and analysis by the TSE Data Acquisition Software Package. After a 30-min stabilization period, blood pressure and heart rate were continuously monitored and stored. Data sets with evidence of arrhythmias or cycle-to-cycle instability were discarded.

[Pro11,d-Ala12]ANG I was administered in doses of 7, 21, 70, 210, and 700 μg/kg iv (n = 14). The hamsters were randomly assigned to one of two groups (n = 7 each): pretreatment with CV-11974 (1 mg/kg iv), an AT1-receptor antagonist (12), or captopril (4 mg/kg iv); then dose responses to [Pro11, d-Ala12]ANG I were repeated. In a separate group of hamsters (n = 14) the same protocol was performed with ANG I in doses of 0.1, 1, 3, 10, and 30 μg/kg iv. The concentrations of [Pro11,d-Ala12]ANG I and ANG I were chosen to obtain comparable pressor responses (18). Captopril and CV-11974 were administered intravenously in a volume of 0.25 ml/kg.

Isolated aorta.

Another set of hamsters was anesthetized with 4% chloral hydrate (350 mg/kg ip) and killed by exsanguination. The thoracic aorta was rapidly excised and cut into 2-mm-long rings in modified ice-cold Krebs buffer with the following composition (mM): 119 NaCl, 4.7 KCl, 1.2 MgSO4 ⋅ 7H2O, 2.5 CaCl2 ⋅ 2H2O, 1.2 KH2PO4, 25.0 NaHCO3, and 5.5 glucose. The aortic rings were mounted in 20-ml baths containing the buffer (pH 7.4), maintained at 37°C, and oxygenated continuously with 95% O2-5% CO2. Contractile responses were recorded with an isometric transducer (model 837004, TSE) and stored (TSE Data Acquisition Software Package). After 45 min of equilibration at a resting tension of 2.1 g, the rings were primed and checked for viability by repeated exposures to 50 mM KCl (2–3 times), with intervening washing and stabilization periods. Then phenylephrine (10 μM) was administered to induce a contraction, which was used as a reference response. After phenylephrine was washed from the system, the rings were incubated with vehicle, captopril (100 μM), chymostatin (100 μM), aprotinin (100 μM), CV-11974 (10 μM), or the type 2 ANG II (AT2)-receptor antagonist PD-123319 (10 μM) for 30 min before stimulation with ANG I (20 nM) or [Pro11,d-Ala12]ANG I (200 nM). The concentrations of these substrates were selected to achieve comparable vasoconstrictor responses on the basis of dose-response curves for each peptide (∼35% effective dose). To test the contribution of kinin-mediated formation of nitric oxide (NO) to the effect of captopril, the response to ANG I was also assessed in the presence of an inhibitor of NO synthesis, 100 μMN G-nitro-l-arginine methyl ester (l-NAME), in an additional set of rings (n = 5). Finally, to exclude the possibility of any nonspecific decrease in contractility during the experiment, we confirmed the preserved response to 10 μM phenylephrine at the end of the protocol in all rings (4).

Enzymatic assay for ANG II formation from ANG I.

After the animals were killed by exsanguination (n = 4), the left ventricle and thoracic aorta were removed and immediately snap frozen in liquid nitrogen. Biochemical assay was done as previously described (3). Briefly, frozen tissues (0.05–0.1 g) were homogenized in 500 μl of chilled 20 mM phosphate buffer (pH 7.4) and centrifuged at 14,000 rpm for 30 min at 4°C. This procedure was repeated one more time. The pellet was resuspended in 200 μl of 20 mM phosphate buffer (pH 7.4). To determine total ANG II formation from ANG I, particulate fractions of the tissues were incubated with ANG I (final concentration 500 μM) in phosphate-buffered saline (pH 7.4, total incubation volume 50 μl). The following inhibitors were used to determine each inhibitor-blockable enzymatic activity: 100 μM captopril, 0.1 mg/ml soybean trypsin inhibitor (SBTI), and 10 μM aprotinin. The incubation was carried out for 30 min at 37°C, and the reaction was terminated by addition of 300 μl of ice-cold ethanol. The precipitated proteins were removed by centrifugation at 12,900 g for 10 min, and the supernatant containing ANG I, ANG II, and their metabolites was evaporated to dryness. The residues were resuspended in 100 μl of water, 40 μl of which were applied to a C18 reverse-phase HPLC column (Vydac, Hesperia, CA) using an 8-min linear acetonitrile gradient (5–20%) in 25 mM triethylammonium phosphate buffer (pH 3.0, flow rate 2 ml/min). The assays were done in duplicate. The peak area corresponding to a synthetic ANG II standard was integrated to calculate ANG II formation, expressed as nanomoles of ANG II formed per minute per gram of tissue wet weight. An inhibitable ratio of each inhibitor was calculated as follows: (total ANG II formation − ANG II formation in the presence of each inhibitor) / total ANG II formation.

Pharmacological agents.

Captopril, phenylephrine, andl-NAME were purchased from Sigma Chemical (St. Louis, MO). ANG I, ANG II, chymostatin, aprotinin, and SBTI were obtained from Bachem Biochemica (Bubendorf, Switzerland). Chymostatin was dissolved in DMSO, and aprotinin and SBTI were dissolved in water. We confirmed that DMSO in the concentration used in this study did not affect the responses to ANG I or phenylephrine. Purchased ANG I was further purified to 99.9% homogeneity by BioTez (Berlin-Buch, Germany). CV-11974 was supplied by Takeda Chemical (Osaka, Japan). PD-123319 was a kind gift from Parke-Davis (Ann Arbor, MI). [Pro11,d-Ala12]ANG I was synthesized and purified by BioTez. Concentrations of the chemicals are expressed as final concentration in the organ bath.


Values are means ± SE. Contractile responses of aortic rings to ANG I and [Pro11,d-Ala12]ANG I were expressed as a percent response to 10 μM phenylephrine. Differences in inhibitory effects of various inhibitors in aortic rings were analyzed by Scheffé’s test after ANOVA.P < 0.05 (2-tailed) was considered statistically significant.


Blood pressure measurement in vivo.

The baseline heart rate of the conscious hamsters (n = 28) was 317 ± 13 beats/min. The mean systolic and diastolic blood pressures were 122 ± 5 and 87 ± 3 mmHg, respectively. There were no significant differences in these baseline parameters between the ANG I- and [Pro11,d-Ala12]ANG I-injected groups; pretreatment with captopril or CV-11974 did not affect basal blood pressure or heart rate (data not shown). Intravenous injections of ANG I (Fig.1 A) and [Pro11,d-Ala12]ANG I (Fig. 1 B) caused similar dose-dependent pressor responses. CV-11974 significantly inhibited both of these responses, whereas captopril inhibited only the ANG I-induced responses, indicating that the pressor response to [Pro11, d-Ala12]ANG I was ACE independent.

Fig. 1.

Effects of captopril (4 mg/kg iv) and CV-11974 (1 mg/kg iv) on increases in mean arterial pressure (ΔMAP) in conscious hamsters in response to ANG I (A) and [Pro11,d-Ala12]ANG I (B).n, Number of animals. * P < 0.01 vs. respective control values.

Isolated aorta.

Figure 2 Ashows the vasoconstrictor response to ANG I (20 nM) with or without various inhibitors. Captopril (100 μM) significantly attenuated the ANG I-induced contraction (84% inhibition). To exclude the possible role of NO in the effect of captopril, we pretreated five aortic rings with l-NAME (100 μM). Although the response to ANG I was increased (from 38.2 to 56.9% of phenylephrine), captopril similarly suppressed it (81%), as in the absence of l-NAME (84%). Although chymostatin (100 μM) or aprotinin (100 μM) alone did not affect the response to ANG I (aprotinin data not shown), administration of captopril with either chymostatin or aprotinin virtually abolished it (Fig. 2 A). CV-11974 (10 μM) eliminated the ANG I-induced contraction (Fig.2 A), but PD-123319 (10 μM) had no effect (data not shown).

Fig. 2.

Effects of captopril (100 μM), chymostatin (100 μM), aprotinin (100 μM), and CV-11974 (10 μM) on vasoconstrictor responses to ANG I (A) and [Pro11,d-Ala12]ANG I (B) in isolated aorta.n, Number of experiments. * P < 0.05 vs. respective control values; ** P < 0.05 vs. captopril (A) and chymostatin (B) treatment.

[Pro11,d-Ala12]ANG I (200 nM) produced contraction similar to that produced by ANG I (Fig.2 B). Although captopril had no effect on the response, chymostatin (100 μM) or aprotinin (100 μM) significantly attenuated it (71 and 57% inhibition, respectively). Lower doses of chymostatin or aprotinin had no inhibitory effect (data not shown). The combination of either captopril and chymostatin (data not shown) or chymostatin and aprotinin showed no additive inhibitory effect. CV-11974 completely suppressed the response to [Pro11,d-Ala12]ANG I (Fig. 2 B), whereas PD-123319 did not affect it (data not shown), indicating that vasoconstriction by [Pro11,d-Ala12]ANG I was mediated by the AT1receptor, as was vasoconstriction by ANG I.

Enzymatic assay for ANG II formation from ANG I.

Table 1 summarizes the total ANG II formation from ANG I in hamster aorta and left ventricular particulate fractions and the effects of various inhibitors on the conversion. In both tissues, ANG II formation from ANG I was biochemically confirmed. We analyzed a relative contribution of ACE-dependent and ACE-independent pathways to the ANG II generation on the basis of the inhibitory ratios of captopril, SBTI, and aprotinin. The SBTI-inhibitable ANG II formation predominated over captopril- or aprotinin-inhibitable ANG II formation in the left ventricle and the aorta.

View this table:
Table 1.

Effect of various inhibitors on ANG II formation from ANG I


The present study provides evidence for the presence of ACE-independent ANG II-forming pathways from ANG I or [Pro11,d-Ala12]ANG I in hamster heart and aorta. Our functional and biochemical results further suggest that multiple enzymes, including chymase, play a role in the ACE-independent ANG II formation.

In conscious hamsters, [Pro11,d-Ala12]ANG I produced dose-dependent pressor responses, which were inhibited by an AT1-receptor antagonist, but not by an ACE inhibitor, in agreement with the findings in conscious marmosets (18) and baboons (10). Because [Pro11,d-Ala12]ANG I does not directly interact with ANG II receptors and ACE does not cleave this substrate (14), the pressor responses to [Pro11,d-Ala12]ANG I are due to ACE-independent ANG II formation (10, 18). Using AT1- and AT2-receptor antagonists, we further investigated the angiotensin receptor subtypes mediating the vasoconstrictor responses to [Pro11, d-Ala12]ANG I in isolated hamster aorta. Our data clearly showed that the [Pro11,d-Ala12]ANG I-induced vasoconstriction was mediated by AT1 receptors.

The observed conversion of [Pro11,d-Ala12]ANG I to ANG II is probably mediated by chymase, as previously reported (10, 11, 18). However, we cannot rule out the possibility that [Pro11,d-Ala12]ANG I is also cleaved by enzyme(s) other than chymase for two reasons:1) the inhibition by chymostatin, even in a large dose, was only partial (71%), and chymostatin inhibits cathepsin G in addition to chymase; and2) aprotinin, which is not a chymase inhibitor, partially inhibited [Pro11,d-Ala12]ANG I-induced contraction. Thus aprotinin-sensitive enzyme(s), in addition to chymase, may be responsible, at least partially, for the vasoconstriction induced by [Pro11,d-Ala12]ANG I.

Captopril (4 mg/kg) and CV-11974 (1 mg/kg) similarly suppressed the pressor responses to ANG I in conscious hamsters, although the suppression was not complete. The reason for the incomplete inhibition is not clear. Insufficient doses are unlikely, because we used hypotensive doses for both inhibitors (12, 18). The effect of an even higher dose of CV-11974 (3 mg/kg) was comparable to that of trandolapril (3 mg/kg) in a renovascular hypertension model of the hamster (13). These findings suggest that circulating ANG I conversion is mediated predominantly by ACE in hamsters. This may be related to the fact that ACE is located in the vascular endothelium, whereas chymase is mainly found in the adventitia (22). Interestingly, in conscious marmosets a lower dose of captopril (1 mg/kg) inhibited the pressor responses to ANG I completely, whereas losartan (5 mg/kg) did not completely suppress them (18). In marmosets, captopril partially suppressed even the responses to [Pro11,d-Ala12]ANG I. Although the reason for the discrepancy between the present and previous studies (18) is not known, species difference is a most plausible explanation. Bradykinin and NO may play a greater role in marmosets than in hamsters. This notion was substantiated by our results in the isolated aorta usingl-NAME. However, further studies are needed to clarify this issue using several ACE inhibitors and AT1-receptor antagonists.

Captopril suppressed the ANG I-induced contraction of the hamster aorta by 84%. A similar degree of suppression was also found in the endothelium-denuded dog mesenteric artery (18). Because we used the endothelium-intact aorta, the inhibitory effect of captopril could be due not only to the blockade of the renin-angiotensin system but also to enhanced kinin-mediated formation of NO. To exclude this possibility, we compared the effect of captopril in the presence and absence of l-NAME. In this setting, captopril similarly suppressed the ANG I-induced contraction, suggesting that the inhibition by captopril in hamster aorta was mainly due to the blockade of ANG II formation. Our finding, however, is in disagreement with the results by Cornish et al. (5). They demonstrated that the ANG I-mediated vasoconstriction in hamsters was not at all affected by an ACE inhibitor but was prevented by an ANG II receptor antagonist. The discrepancy may be explained by the difference in the method and/or vessels used. Cornish et al. measured the changes in the diameter of cheek pouch arterioles, whereas we determined the contractile force of the aorta. It is well known that vascular responses to ANG II differ in various blood vessels (28). Indeed, the same authors reported that an ACE inhibitor suppressed the ANG I-induced contraction of hamster coronary arteries by 82% (6).

As in dog mesenteric arteries (18), chymostatin (100 μM) alone had no effect on the ANG I-induced vasoconstriction in hamster vessels. This is in contrast to the finding in human visceral arteries, in which an even lower concentration of chymostatin (50 μM) reduced the response to ANG I by 60–70%, whereas captopril attenuated it by only 30–40% (24). Similar results were also reported using human internal mammary arteries (32). These findings indicate that vascular chymase plays a greater role in human arteries than in dog or hamster arteries. Interestingly, the combined pretreatment of hamster vessels with chymostatin and captopril completely inhibited the response to ANG I, as found in dog arteries (18, 23). A similar synergistic inhibitory effect on the response to ANG I was also observed in human arteries (24, 32) and detrusor muscle (17), but not in rat or rabbit arteries (24). Because the combination of ACE inhibitors and serine proteinase inhibitors did not affect the ANG II-induced contraction (17, 23), the synergistic inhibitory effect is specific for the conversion of ANG I to ANG II. We further found that the coadministration of captopril and aprotinin also completely eliminated the ANG I-induced asoconstriction, suggesting again the presence of an aprotinin-inhibitable ANG II-forming pathway in hamster aorta. This is supported by our biochemical data showing the existence of aprotinin-inhibitable ANG II formation.

Taken together, in vivo and in vitro studies by us and others (18) suggest that ACE is the main enzyme to convert exogenously administered ANG I to ANG II in dog and hamster vasculature. However, other enzyme system(s) may become physiologically competitive with ACE in the presence of an ACE inhibitor. This “escape phenomenon” could explain, at least in part, why plasma ANG II levels did not sufficiently decrease, despite the administration of ACE inhibitors in normal volunteers (21) and in patients with heart failure (25). Furthermore, a serine protease inhibitor, nafamostat, but not captopril, inhibited the increase in plasma ANG II after exercise and improved the leg circulation in patients with arteriosclerosis (19, 29). The increase in ANG II concentration in the dog coronary sinus after coronary ligation was not inhibited by captopril but was significantly decreased by aprotinin (8, 20). Because aprotinin and nafamostat do not inhibit ACE or chymase, these findings and the present results suggest the presence of the ACE- and chymase-independent ANG II-forming pathway in the cardiovascular system of dogs, hamsters, and humans. Although our study could not specify the enzyme(s) responsible for the aprotinin-sensitive ANG II formation, tissue kallikrein or kallikrein-like protease has been implicated (8,19, 20, 29). Recently, kallikrein-like enzyme, which is capable of generating kinin and ANG II, has been purified and characterized in the dog heart (26).

Biochemical data demonstrated 1) the in vitro conversion of ANG I to ANG II in the hamster heart and aorta and 2) the predominance of ACE-independent ANG IIforming pathways in these tissues. As in human left ventricular tissues (2, 30, 31), chymase may also play an important role in the alternative pathway in hamster tissues. The predominance of the SBTI-inhibitable ratio over the aprotinin-inhibitable ratio may support this notion, because SBTI inhibits chymase and several other ANG II-forming enzymes such as kallikrein and cathepsin G (31), whereas aprotinin inhibits the same enzymes except for chymase. However, care should be exercised in the interpretation of these data, because marked species differences exist in the ACE- vs. chymase-mediated ANG II formation (2, 3, 24). To clarify this issue, a chymase-specific inhibitor is needed. Unlike human heart (31) or brain (3), we demonstrated the aprotinin-inhibitable ANG II formation in hamster heart and aorta, suggesting the existence of an ACE- and chymase-independent ANG II-generating system. This is substantiated by our functional results in the isolated aorta.

In summary, this study showed the presence of functional ACE-independent ANG II-forming pathways in the hamster cardiovascular system, although their contributory ratio is small. Chymase, as well as aprotinin-sensitive serine protease(s), appears to be responsible for the alternative ANG II formation. Further studies are needed to clarify the pathophysiological roles of the ACE-independent ANG II formation in cardiovascular disease.


This study was partly supported by European Commission Grant ERBCHBGCT 940725 to H. Buikema.


  • Address for reprint requests: H. Nishimura, Third Dept. of Internal Medicine, Osaka Medical College, 2-7 Daigakumachi, Takatsuki, Osaka 569, Japan.


View Abstract