• The mechanisms involved in the vasodilator actions of angiotensin II (Ang II) have not yet been completely elucidated. We investigated the potential mechanisms that seem to be involved in the Ang II vasodilator effect using rat isolated mesenteric vascular bed (MVB).
• Under basal conditions, Ang II does not affect the perfusion pressure of MVB. However, in vessels precontracted with norepinephrine, Ang II induces vasodilation followed by vasoconstriction. Vasoconstrictor, but not the vasodilation of Ang II, is inhibited by AT₁ antagonist (losartan). The vasodilator effect of Ang II was not inhibited by AT₂, angiotensin IV and angiotensin 1–7 receptor antagonists alone (PD 123319, divalinal, A 779, respectively).
• The vasodilator effect of Ang II is significantly reduced by endothelial removal (deoxycholic acid), but not by indomethacin. Inhibition of NO-synthase by N^G-nitro-L-arginine methyl ester (L-NAME) and guanylyl cyclase by 1H-[1,2,3] oxadiazolo [4,4-a] quinoxalin-1-one (ODQ) reduces the vasodilator effect of Ang II. This effect is also reduced by tetraethylammonium (TEA) or L-NAME, and a combination of L-NAME plus TEA increases the inhibitory effect of the antagonists alone. However, indomethacin does not change the residual vasodilator effect observed in vessels pretreated with L-NAME plus TEA.
• In vessels precontracted with norepinephrine and depolarized with KCl 25 mM or treated with Ca²⁺-dependent K⁺ channel blockers (charybdotoxin plus apamin), the effect of Ang II was significantly reduced. However, this effect is not affected by ATP and voltage-dependent K⁺ channel blockers (glybenclamide and 4-aminopyridine).
• Inhibition of kininase II with captopril significantly potentiates the vasodilator effect of bradykinin (BK) and Ang II in the rat MVB. The inhibitory effect of the B₂ receptor antagonist HOE 140 on the vasodilator effect of Ang II is further enhanced by PD 123319 and/or A 779.
• The present findings suggest that BK plays an important role in the endothelium-dependent vasodilator effect of Ang II. Probably, the link between Ang II and BK release is modulated by receptors that bind PD 123319 and A 779.

The renin–angiotensin system modulates cardiovascular function, and angiotensin II (Ang II), the most important autacoid of this control system, induces vasoconstriction, facilitates sympathetic transmission and promotes renal salt and water retention (Chung & Unger, 1999). The effects of Ang II on the cardiovascular system are clearly mediated by the activation of the AT₁ receptor. In contrast, the exact significance of AT₂ receptors on the action of Ang II is not yet clear. The vasodilator effect of Ang II has been shown to occur not only ‘in vitro' (Arima et al., 1997; Endo et al., 1998; Kimura et al., 2001; Soares de Moura et al., 2001; Katada & Majima, 2002), but also ‘in vivo' (Scheuer & Perrone, 1993; Tsutsumi et al., 1999). We have recently demonstrated (Soares de Moura et al., 2001) that Ang II has no vasoconstrictor effect in the rat mesenteric vascular bed (MVB), but, when these vessels are precontracted with norepinephrine (NE), Ang II induces an endothelium-dependent vasodilation.

In the present study, we investigated the mechanism of the vasodilator effect of Ang II on the MVB of the rat. Ang II induces an endothelium-derived relaxing factor (EDRF)-dependent effect that seems to be modulated by an interaction among Ang II, bradykinin (BK) and angiotensin 1–7 (Ang 1–7). Preliminary results have been presented to the British Pharmacological Society (Soares de Moura et al., 2001).

All experiments were reviewed and approved by the Ethics Committee of Animal Experiments of the State University of Rio de Janeiro. The rat superior MVB was isolated according to McGregor (1965). Briefly, male Wistar rats were killed with inhaled CO₂ and the MVB was cannulated and perfused at a flow rate of 4 ml min⁻¹ with a physiological salt solution (PSS) by a peristaltic pump (Lifecare Model 4, Abbott'Shaw). The PSS had the following composition (mM): NaCl 118, KCl 4.7, CaCl₂ 2.5, MgSO₄ 1.2, KH₂PO₄ 1.2, NaHCO₃ 25, EDTA 0.02 and glucose 11. All experimental groups were performed in the presence of losartan (30 μmol), with the exception of the losartan group, where the effect of Ang II was obtained before and after losartan. The PSS (37°C) was bubbled with 95% O₂/5% CO₂. Perfusion pressure (PP) was measured with a transducer connected to a preamplifier and chart recorded. Drugs were either dissolved in PSS and perfused at the desired concentration, or were administered as bolus injections directly into the perfusion stream (volume <300 μl).

The preparations were left to equilibrate for 30 min, and then injections of 120 μmol KCl were administered every 10 min until consistent responses were obtained. The basal pressure after the equilibration period was 23.2±1 mmHg (n=198). Ang II (3–300 nmol) was injected before and after the PP had been elevated (80–110 mm Hg) with NE (6–30 μM) added to the perfusion fluid. The vasodilator effect of all the drugs was expressed as a percent decrease in relation to the pressor effect of NE. When the pressor effect of NE reached a plateau, dose–response curves to bolus injections of Ang II (3–300 nmol), Ang 1–7 (10–300 nmol), angiotensin IV (Ang IV; 10–300 nmol) and CGP 42112 A (10–300 nmol), a putative AT₂ agonist, were obtained. Acetylcholine (ACh; 10 pmol), BK (10 pmol) and nitroglycerin (NG; 1 nmol) were injected as single doses.

The vasodilator effects of Ang II were studied before and after perfusion with losartan (30 μM), an inhibitor of the AT₁ receptor, PD 123319 (10 μM) a putative inhibitor of the AT₂ receptor, A 779 (1 μM), a putative inhibitor of the Ang 1–7 receptor, glybenclamide (1 μM), an inhibitor of the ATP-dependent K⁺ (K_ATP) channel, aminopyridine (4-AP; 1 mM), an inhibitor of the voltage-dependent K⁺ (K_V) channel, indomethacin (3 and 10 μM), an inhibitor of cyclooxygenase (COX), divalinal-Ang IV, an inhibitor of the Ang IV receptor PD 123319 (10 μM) plus A 779 (1 μM); HOE 140 (0.01 μM), an antagonist of the BK B₂ receptor, HOE 140 (0.01 μM) plus PD 123319 (10 μM), HOE 140 (0.01 μM) plus A 779 (1 μM) and HOE 140 (0.01 μM) plus PD 123319 (10 μM) plus A 779 (1 μM).

The vasodilator effects of Ang II, ACh and NG were studied before and after perfusion with deoxycholic acid (2.5 mM) dissolved in PSS for 3 min to chemically remove the endothelium, N^G-nitro-L-arginine methyl ester (L-NAME; 0.3 mM), an inhibitor of nitric oxide (NO) synthase or 1H-[1,2,3] oxadiazolo [4,4-a] quinoxalin-1-one (ODQ; 10 μM), an inhibitor of guanylyl cyclase (GC), tetraethylammonium (TEA), a nonspecific inhibitor of K⁺ channels, TEA plus L-NAME, TEA plus L-NAME plus indomethacin, charybdotoxin (ChTx) plus apamin inhibitors of Ca²⁺-dependent K⁺ (K_Ca²⁺) channels or in vessels perfused with high K⁺ solution (25 mM). In vessels pretreated with deoxycholic acid, L-NAME and indomethacin, the concentration of NE was adjusted in order to maintain the same increase in PP. The vasodilator effects of Ang II, ACh, NG and BK were studied before and after perfusion with HOE 140 (0.2 μM). Dose–response curves were constructed to CGP 42112 A and Ang IV in vessels perfused with PSS, and the vasodilator effects of Ang IV (300 nmol) and Ang II (30 nmol) were studied before and shortly after (1–5 min) injections of Divanilal-Ang IV (300 nmol). The vasodilator effects of Ang II, ACh, NG and BK were studied before and after perfusion with captopril (10 μM), an inhibitor of kininase II. The effects of Ang 1–7 were studied before and after perfusion with A 779 (1 μM).

The present results confirm the previous findings of our group, showing that, Ang II in the MVB of the rat, unlike in most vascular beds, does not produce vasoconstriction (Soares de Moura et al., 2001). However, when the tonus of the mesenteric vessels is increased with infusions of NE, Ang II produces an initial vasodilation, then followed by vasoconstriction. The vasoconstrictor, but not the vasodilator effect of Ang II, depends on AT₁ receptor activation, because this effect is completely abolished by losartan or irbesartan (Soares de Moura, personal observation). The vasodilator effect of Ang II depends on the endothelium since Ang II- and ACh-induced vasodilation are significantly inhibited in vessels pretreated with deoxycolic acid, while the effect of NG, an endothelium-independent vasodilator, is not affected by endothelium removal. The endothelial-dependent vasodilator effect of Ang II has also been demonstrated in the isolated microperfused rabbit afferent arteriole (Arima et al., 1997) and in perfused rat mesenteric arteries (Katada & Majima, 2002). However, the results obtained by Katada & Majima (2002) differ from our present results, given that the vasodilator effect of Ang II was only observed after the blockade of AT₁ receptors by losartan.

The vasodilator effect of Ang II in the rat MVB might be due to release of vasodilator autacoids by the endothelial cells, including prostacyclin, NO and hyperpolarizing factor(s). In this study, the involvement of prostaglandins on the vasodilator effect of Ang II seems improbable, since indomethacin, even at a high concentration (10 μM), does not affect the vasodilator effect of Ang II. Our results are in accordance with previous findings on the microperfused rabbit afferent arteriole (Arima et al., 1997). The lack of involvement of prostaglandins on the vasodilator effect of Ang II in the MVB of the rat is further supported by the present results, which demonstrate that the residual vasodilator effect induced by Ang II in vessels pretreated with TEA plus L-NAME is not reduced by indomethacin. In contrast, our present results differ from those obtained by Katada & Majima (2002), showing that indomethacin (10 μM) modestly inhibits Ang II vasodilation in perfused rat mesenteric arteries. A possible reduction in the responsiveness of vascular preparation in Katada & Majima's (2002) experiments might have been caused by the continuous infusion of Ang II, rather than by the inhibition of vasodilation mediated by indomethacin, since the Ang II vasodilator response decreased as the number of applications increased.

In this study, NO seems to play an important role in the vasodilator effect of Ang II, since L-NAME substantially reduces the vasodilator effect of Ang II in the rat MVB. The involvement of the NO-guanosine 3′5′ cyclic monophosphate (cGMP) pathway in the vasodilator effect of Ang II is further corroborated by our results with ODQ treatment, showing a significant reduction in the vasodilator effect of Ang II. It is conceivable that NO is not the only autacoid released from endothelium in response to Ang II, given that the vasodilator effect of Ang II is not completely inhibited by L-NAME. Since the vasodilator effect of Ang II and ACh, but not NG, is inhibited by ChTx plus apamin, it is possible that K_Ca²⁺ channel activation might be involved in the vasodilator effect of Ang II. Probably, K_ATP and K_v channels are not involved in the Ang II-dependent vasodilation, because this effect is not inhibited by glybenclamide or 4-AP. The inhibitory effect of L-NAME or TEA on the vasodilator effect of Ang II was further enhanced when those two inhibitors were used together, but the inhibitory effect of ChTx plus apamin was not further enhanced by L-NAME, suggesting that Ang II-induced hyperpolarization is a crucial mechanism involved in the vasodilator effect of Ang II in the rat MVB. Ang II could induce hyperpolarization not only by releasing of endothelium-derived hyperpolarizing factor (EDHF), which activates K⁺ channels inhibited by ChTx plus apamin or TEA, but also by a mechanism modulated by cGMP that induces hyperpolarization by activating a K_Ca²⁺ channel inhibited by ChTx (Robertson et al., 1993; Vanhoutte, 1999). This hypothesis is supported by our findings, which show that, in vessels depolarized with high K⁺ solution and constricted with NE, the vasodilator effect of Ang II is almost completely abolished.

It has been shown that the vasodilator effect of Ang II on the microperfused rabbit afferent arteriole (Arima et al., 1997) and rat mesenteric arterial segments (Katada & Majima, 2002) is inhibited by PD 123319. However, in the current study, the vasodilator effect of Ang II was not inhibited by PD 123319 alone. Furthermore, we failed to demonstrate a vasodilator effect of CGP 42112 A. However, when the B₂ receptors are not completely inhibited by HOE 140 (0.01 μM), a combination of HOE 140 plus PD 123319 further enhances the inhibitory effect of HOE 140, suggesting that AT₂ receptors could participate in the BK-dependent vasodilation induced by Ang II.

The vasodilator effect of Ang II could be modulated by Ang IV, a vasodilator component of the renin–angiotensin system (Haberl et al., 1991; Coleman et al., 1998; Patel et al., 1998). Alternatively, Ang II could be metabolized to Ang IV by aminopeptidase A and aminopeptidase M. However, our results demonstrate that Ang IV most likely does not participate in the vasodilator effects of Ang II, since its vasodilator effect is much smaller than the vasodilator effect of Ang II. Moreover, divalinal-Ang IV, an antagonist of the Ang IV receptor (Krebs et al., 1996) which reduces the vasodilation effect of Ang IV in the MVB, does not inhibit the vasodilator effect of Ang II.

BK seems to play an important role in the vascular effects of Ang II (Seyedi et al., 1995; Gohlke et al., 1998; Tsutsumi et al., 1999; Sosa-Canache et al., 2000; Katada & Majima, 2002). Probably, the vasodilator effect of Ang II in the rat MVB is modulated by BK mechanisms, since we showed that captopril enhanced the vasodilator effects of BK and Ang II. In the present results, we also showed that HOE 140, at a concentration (0.01 μM) that significantly reduced the vasodilator effect of BK, induced a small but significant reduction in the vasodilator effect induced by high doses of Ang II. However, HOE 140, at a concentration that almost abolished the vasodilator effect of BK, significantly reduced the vasodilator effect of Ang II in all doses studied, confirming that BK must play an important role in the vasodilator effect of Ang II in the rat MVB.

Ang 1–7, a potent vasodilator compound (Meng & Busija, 1993; Osei et al., 1993; Pörsti et al., 1994; Brosnihan et al., 1996; Ren et al., 2002), could be involved in the mechanism of the action of Ang II in the rat MVB. Ang II could be transformed into Ang 1–7 by carboxypeptidase P, or might also stimulate the Ang 1–7 receptor (Ferrario et al., 1990). Our study demonstrates that Ang 1–7 induces a significant vasodilator effect on the MVB in the same dose range of Ang II. Unlike other vascular beds, where A 779, a putative Ang 1–7 antagonist (Santos et al., 1994), inhibits the vasodilator effect of Ang 1–7 (Bayorh et al., 2002; Ren et al., 2002), our results demonstrate that the vasodilator effect of Ang 1–7 on the MVB could involve other receptors, given that A 779 does not inhibit Ang 1–7 vasodilation. BK appears to modulate Ang 1–7-induced vasodilation, since several reports have shown that the vasodilator effect of BK is enhanced by Ang 1–7 and vice versa (Brosnihan et al., 1996; Gorelik et al., 1998; Almeida et al., 2000; Fernandes et al., 2001; Tom et al., 2001). Also, the involvement of BK in the vasodilator effect of Ang 1–7 is suggested by Brosnihan et al. (1996), who demonstrated that the vasodilator effect of Ang 1–7 was inhibited by HOE 140. BK and Ang 1–7 may share a common mechanism involving NO, since the Ang 1–7 potentiation of BK-induced vasodilation in rat mesenteric vessels is abolished by L-NAME (Oliveira et al., 1999). In the present study, the partial inhibition of Ang II vasodilation induced by HOE 140 (0.01 μM) is significantly enhanced by a combination of HOE 140 (0.01 μM) plus A 779, suggesting that Ang 1–7 might participate in the BK-dependent Ang II vasodilation. Therefore, the BK link to the EDRF-dependent vasodilator effect of Ang II probably involves the participation of receptors inhibited by PD 123319 and A 779, since the inhibitory effect of HOE 140 plus PD 123319 or HOE 140 plus A 779 on the vasodilator effect of Ang II is further enhanced by a combination of HOE 140, PD 123379 and A 779.

In conclusion, in the present study, we demonstrated that Ang II induces an EDRF-dependent vasodilator response in the rat MVB, which depends on NO release. Hyperpolarization due to activation of the K_Ca²⁺ channel could be involved in the vasodilator effect of Ang II. The results also suggest that BK could play a very important role in the mechanism of the vasodilator effect of Ang II. Probably, the link between Ang II and BK release might be modulated by receptors that bind PD 123319 and A 779. Finally, the Ang II-dependent vasodilation does not appear to be modulated by prostaglandin release, activation of K_ATP or K_V channels and AT₁ or Ang IV receptors.

This work was supported in part by CNPq and FAPERJ .