We tested the hypothesis that endothelial nitric oxide (NO) synthase (eNOS)-derived NO modulates rho-kinase-mediated vascular contraction. Because 3-hydroxy-3-methylglutaryl (HMG)-CoA-reductase inhibition can both upregulate eNOS expression and inhibit rhoA/rho-kinase function, a second hypothesis tested was that statin treatment modulates rho-kinase-mediated contraction and that this can occur independently of eNOS. Contractile responses to the receptor-dependent agonists serotonin and phenylephrine but not to the receptor-independent agent KCl were greater in aortic rings from eNOS-null (eNOS−/−) vs. wild-type (eNOS+/+) mice. Similarly enhanced responses were seen in eNOS+/+ rings after acute NOS inhibition. The rho-kinase inhibitor Y-27632 abolished or profoundly attenuated responses to receptor agonists in both eNOS+/+ and eNOS−/− rings, but responses in eNOS+/+ were more sensitive to Y-27632. Mevastatin treatment (20 mg/kg sc per day, 14 days) reduced responses to serotonin and phenylephrine in female mice of both strains. KCl-induced contractions were slightly smaller in eNOS+/+-derived aortic rings only. Levels of plasma cholesterol, and aortic expression of rhoA and rho-kinase, did not differ between groups. Thus eNOS-derived NO suppresses rhoA/rho-kinase-mediated vascular contraction. Moreover, a similar suppressive effect on rho-kinase-mediated vasoconstriction by statin therapy occurs independently of effects on eNOS or plasma cholesterol.
the ca2+-independent increase in vascular smooth muscle tone that can occur through inhibition of myosin light-chain phosphatase (MLCP) is known as Ca2+ sensitization and appears to be largely mediated by activation of the small GTPase rhoA and its downstream effector rho-kinase (35, 37). RhoA may be activated by several signaling pathways, including the binding of G protein-coupled receptor (GPCR) agonists (35). RhoA-activated rho-kinase then phosphorylates and inhibits MLCP (16), leading to vasoconstriction. By contrast, nitric oxide (NO) contributes to regulation of vascular tone by relaxing smooth muscle. NO is formed within vascular endothelium by the endothelial isoform of NO synthase (eNOS) and is the predominant mediator of endothelium-dependent vascular relaxation (29). Thus the rhoA/rho-kinase and eNOS/NO signaling pathways essentially have major opposing roles in the vasculature, although it is poorly understood whether these mechanisms normally interact.
It is well established that bioactivity of endothelium-derived NO is commonly diminished in many cardiovascular disease states (10). Conversely, several studies have now reported that vascular rhoA/rho-kinase activity is elevated in cardiovascular disease states such as hypertension, arteriosclerosis, and cerebral and coronary vasospasm (34). To our knowledge, however, the possibility that vascular rho-kinase function is augmented as a consequence of chronically reduced eNOS-derived NO activity has not yet been directly tested.
3-Hydroxy-3-methylglutaryl (HMG)-CoA-reductase inhibitors, or “statins,” are widely used clinically to lower elevated blood cholesterol levels. Indeed, the beneficial effects of statin therapy appear to extend far beyond their cholesterol-lowering abilities (36). Experimental use of statins has also been reported to confer cholesterol-independent protection against stroke, improve endothelial function, and decrease platelet aggregation (9, 22). These effects have been attributed to upregulation of eNOS expression and consequently increased NO production, due to inhibition of rhoA isoprenylation and subsequently increased eNOS mRNA stability via effects on the endothelial actin cytoskeleton (21, 24). Although, in theory, statins could also modulate vascular contractility through interference with isoprene formation and subsequent rhoA/rho-kinase signaling in vascular smooth muscle, independently of any effect on eNOS/NO, no study has so far tested this hypothesis. First, we therefore tested whether rho-kinase-mediated vascular contraction is enhanced in the chronic and/or acute absence of NO. Second, we tested whether statin treatment can suppress rho-kinase-mediated vascular contraction, and if so, whether this can occur independently of effects on NO activity and plasma cholesterol levels.
MATERIALS AND METHODS
Adult male and female mice of eNOS+/+ (wild type, C57BL/6J) and eNOS mutant (eNOS−/− on a C57BL/6J background; Jackson Laboratory, stock no. 002684, colony maintained by homozygous sibling matings) genotype were studied (body wt 25 ± 1 g, mean ± SE; n = 121). Thoracic aortic rings (3–4 mm) were mounted at 5 mN in a 5-ml chamber of a four-channel myograph (model 610M, MultiMyograph) containing Krebs-bicarbonate solution bubbled with 5% CO2 in O2 at 37°C. Tension was continuously recorded on a chart recorder (model 3721, Yokogawa, Japan). Each ring was first exposed to an isotonic high-K+-containing physiological saline solution (KPSS), in which Na+ in Krebs solution was replaced by K+ ([K+]KPSS = 124 mM). The KPSS-induced contraction was allowed to achieve a stable level over 30–40 min. After several washouts and return to stable baseline (∼5 mN), each ring was precontracted to ∼50% of its KPSS response with either serotonin (0.03–0.1 μM) or phenylephrine (0.1–0.3 μM). Sustained relaxation (>70% of precontracted tone) of eNOS+/+ rings in response to ACh (10 μM) confirmed the endothelium to be functionally intact. By contrast, the lack of such relaxation to ACh of eNOS−/−-derived rings phenotypically confirmed the absence of eNOS expression. After several washouts and return to stable baseline, concentration-response curves were established for the vascoconstrictor agents serotonin, phenylephrine, or KCl. One or two such curves were typically performed per ring.
Effects of rho-kinase and NOS inhibition on vascular contraction.
The effect of the selective rho-kinase inhibitor Y-27632 (1–10 μM) (7, 13) on vasoconstrictor responses was assessed by treating aortic rings for 30 min before commencing cumulative additions of a contractile agent. Similarly, the effect of acute NOS inhibition by the isoform nonselective NOS inhibitor NG-nitro-l-arginine methyl ester (l-NAME, 100 μM) was assessed after 30-min treatment before addition of constrictors. During both sets of experiments, vehicle (0.9% saline)-treated rings served as controls. In other experiments, the contractile response to 100 μM l-NAME after 2 h was assessed as an estimate of basal NO activity in eNOS+/+-derived rings precontracted to ∼20% of the KPSS response with one of the three contractile agonists. The lack of response of eNOS−/− rings to l-NAME provided further confirmation of the eNOS−/− genotype.
Chronic mevastatin treatment.
Mevastatin (Calbiochem, La Jolla, CA) was activated by alkaline hydrolysis and then neutralized under sterile conditions. This solution was then dried down by vacuum centrifugation and reconstituted in saline at 33 mg/ml. Aliquots of this preparation were stored at −20°C for up to 2 mo before use. Female mice of either strain were treated with vehicle (0.9% saline, n = 3) or mevastatin (20 mg/kg sc per day, n = 11) for 14 days via a surgically implanted osmotic minipump (Alzet model 2002, Alza). Minipump implantation was performed aseptically under brief anesthesia (methohexitone 65 mg/kg ip).
Total rhoA and rho-kinase in mouse aorta were measured by Western blotting using antisera to rhoA (Santa Cruz Biotechnology sc-179, San Diego) or rho-kinase [raised against a GST fusion protein of the rho-kinase (ROKα) coiled coil domain, residues 429–968, expressed in E. coli from a pGEX plasmid provided by Dr. T. Leung (3)].
Analyses and statistics.
All responses to contractile agents are presented as percentage of the KPSS response of that aortic ring. Concentration-response data for serotonin and phenylephrine were fitted to a Sigmoid plot using GraphPad Prism version 3.0a, which estimated the pEC50 value. KCl curves were generally found to be nonsigmoidal, and so pEC50 values were not calculated. Relaxation responses are presented as %inhibition of the precontracted level of tone. Each n represents the number of animals used. Statistical analysis was carried out using Student's paired or unpaired t-test or a one-way ANOVA followed by Newman-Keuls multiple comparisons test (as appropriate). P < 0.05 was considered statistically significant.
KPSS-induced vascular contractions were slightly greater in eNOS+/+ vs. eNOS−/− aortic rings (11.4 ± 0.05 vs. 10.3 ± 0.4 mN, n = 57 and 53, respectively; P < 0.05). In rings from eNOS+/+ mice, ACh elicited a large, sustained relaxant response (79 ± 1%, n = 57), whereas in rings from eNOS−/− mice ACh elicited a large, transient contraction (82 ± 6%, n = 53; P < 0.05 vs. eNOS+/+).
Effect of eNOS expression on responses to contractile agents.
Serotonin, phenylephrine, and KCl each caused concentration-dependent contraction of aortic rings from both eNOS+/+ and eNOS−/− mice (Fig. 1, A, C, and E). Maximum responses to the receptor agonists serotonin and phenylephrine were greater in aortic rings from eNOS−/− mice than in rings from eNOS+/+ mice (P < 0.05, Fig. 1, A and C, respectively). However, maximum responses to the nonreceptor contractile agent KCl did not differ between strains (Fig. 1E). pEC50 values were significantly lower (∼2-fold) in eNOS−/− vs. eNOS+/+ rings for serotonin and phenylephrine (Table 1). Interestingly, maximum contractile responses to serotonin were greater in aortic rings from male vs. female eNOS+/+ (129 ± 5 vs. 102 ± 7%, n = 10–12; P < 0.05), but not eNOS−/− (161 ± 6 vs. 150 ± 6%, n = 10–12; P > 0.05) mice. There was no such gender difference in responses to phenylephrine or KCl in either eNOS+/+ or eNOS−/− mice (data not shown).
Effect of acute NOS inhibition.
In eNOS+/+aortic rings treated with l-NAME, responses to serotonin, phenylephrine, and KCl resembled those of control rings from eNOS−/− mice (Fig. 1, B, D, and F; compare with Fig. 1, A, C, and E). l-NAME had no effect on contractile responses to any agent in eNOS−/− aortic rings (data not shown). After precontraction with serotonin, vascular contractions in response to l-NAME were smaller in male vs. female eNOS+/+ aortic rings (45 ± 8 vs. 75 ± 6%, n = 7–9; P < 0.05). However, no such gender difference was observed in response to l-NAME when eNOS+/+ aortic rings were precontracted with KCl (61 ± 17 vs. 59 ± 12%, n = 6–8; P > 0.05) or phenylephrine (154 ± 33 vs. 99 ± 5%, n = 2–3).
Effect of rho-kinase inhibition on contractile responses.
Maximum responses to serotonin and phenylephrine were profoundly inhibited in a concentration-dependent manner by the selective rho-kinase inhibitor Y-27632 (1–10 μM) in aortic rings from both eNOS+/+ and eNOS−/− mice (Fig. 2, A–D). Inhibitory effects of Y-27632 were more pronounced in aortic rings from eNOS+/+ than eNOS−/− mice. By contrast, although slightly reduced by Y-27632, maximum responses to KCl were relatively resistant to effects of rho-kinase inhibition in eNOS+/+ and eNOS−/− aortic rings (Fig. 2, E and F, respectively). Effects of Y-27632 on pEC50 values, if any, were modest (Table 1).
Effects of chronic mevastatin treatment.
Plasma cholesterol concentration was not altered by mevastatin treatment (control eNOS+/+ = 1.28 ± 0.08 mM, n = 8; mevastatin-treated eNOS+/+ = 1.31 ± 0.19 mM, n = 6; control eNOS−/− = 1.53 ± 0.17 mM, n = 6; mevastatin-treated eNOS−/− = 1.57 ± 0.24 mM, n = 5). Mevastatin treatment resulted in 20–30% smaller responses to KPSS in both eNOS+/+ (control = 11.0 ± 0.6 mN, n = 29; mevastatin-treated = 8.7 ± 0.8 mN, n = 6; P > 0.05) and eNOS−/− (control = 10.3 ± 0.5 mN, n = 20; mevastatin-treated = 6.8 ± 0.2 mN, n = 5; P < 0.05) mice. There were no differences in responses to ACh between aortic rings from control and mevastatin-treated mice (control eNOS+/+ = 80 ± 2%, mevastatin-treated eNOS+/+ = 81 ± 4%; control eNOS−/− = −88 ± 9%, mevastatin-treated eNOS−/− = −80 ± 2%).
Mevastatin treatment reduced maximum responses to serotonin and phenylephrine in aortic rings from both eNOS+/+ and eNOS−/− mice (Fig. 3, A–D). pEC50 values for serotonin were decreased slightly in aortic rings from eNOS+/+ mice and increased slightly in rings from eNOS−/− mice (Table 1). pEC50 values for phenylephrine were higher in aortic rings from eNOS−/− mice after mevastatin treatment (Table 1).
Maximum responses of eNOS+/+ aortic rings to KCl were slightly but significantly reduced by mevastatin (P < 0.05, Fig. 3E). However, mevastatin treatment had no effect on responses to KCl in aortic rings from eNOS−/− mice (Fig. 3F).
Contractions in response to l-NAME (100 μM) of aortic rings precontracted with serotonin tended to be smaller in rings from mevastatin-treated eNOS+/+ mice (48 ± 11%, n = 6) than in those from untreated eNOS+/+ mice (75 ± 6%, n = 9; P = 0.08).
Effect of chronic mevastatin treatment on rho-kinase expression in mouse thoracic aorta.
Western blot analysis revealed no effect of chronic mevastatin treatment on rhoA (data not shown) or rho-kinase expression in thoracic aorta of both eNOS+/+ and eNOS−/− mice (Fig. 4).
The present study is the first to directly investigate the importance of eNOS expression in modulation of rho-kinase-mediated vascular contraction after chronic statin treatment. There are several important findings of this study. First, vascular contraction in response to the GPCR agonists serotonin and phenylephrine was selectively greater in aortic rings from eNOS−/− vs. eNOS+/+ mice. Second, acute inhibition of NO synthesis markedly augmented responses of eNOS+/+ aortic rings to these agonists, to become functionally identical to those of eNOS−/− rings. Thus both acute and chronic NO deficiency lead to augmented vascular contractility to serotonin and phenylephrine. Third, vascular responses to serotonin and phenylephrine were profoundly attenuated by the rho-kinase inhibitor Y-27632 in both eNOS+/+- and eNOS−/−-derived aortic rings. However, responses in rings from eNOS+/+ mice were considerably more sensitive to inhibition by Y-27632. Decreased sensitivity of GPCR agonist responses to rho-kinase inhibition during NO deficiency is consistent with the concept that NO normally inhibits vascular rhoA/rho-kinase function. Fourth, chronic mevastatin treatment impaired vascular responses of female eNOS+/+- and eNOS−/−-derived aortic rings to serotonin and phenylephrine without affecting plasma cholesterol concentration. Thus statins modulated vascular function independently of increased endothelial NO synthesis or reduced plasma cholesterol and conceivably acted by interfering with the activity of the rhoA/rho-kinase signaling pathway, which was found to be essential for contraction by these two agonists in mouse aorta.
Physiological roles of vascular NO.
NO is formed within the vascular endothelium by eNOS and modulates vascular tone by mediating endothelium-dependent relaxation in response to certain vasodilators such as ACh (29). It has been known for some time that endothelium-derived NO can also modulate responses to numerous vasoconstrictor stimuli, including serotonin and α-adrenoceptor agonists (6). Thus NO removal either by endothelial denudation (6, 20) or pharmacological block of NO production (2, 29) leads to enhanced vasoconstrictor responses. However, because highly selective eNOS inhibitors are not yet available, pharmacological investigation of specific role(s) of eNOS have been restricted. The recent availability of genetically manipulated animals lacking the eNOS gene has enabled a more definitive means for assessing the importance of endothelial NO in vascular function (12, 19).
In the present study, serotonin- and phenylephrine-induced contractile responses of aortic rings from eNOS−/− mice were greater than those from eNOS+/+ mice. This is consistent with findings in aorta (17) and femoral (39) and carotid arteries (19) of eNOS−/− mice and suggests that genetically based chronic deficiency of eNOS-derived NO leads to augmented vascular contractility, particularly to GPCR agonists. Despite its lack of isoform selectivity, we used the NOS inhibitor l-NAME to investigate whether acute NO deficiency in normal (eNOS+/+) mice could lead to similarly augmented vascular contractions as those observed in aortas from eNOS−/− mice. Indeed, responses were greater in aortas from eNOS+/+ mice after l-NAME treatment, becoming phenotypically identical to those from eNOS−/− mice and indicating that both chronic and acute NO deficiency equally lead to augmented contractility. Importantly also, l-NAME treatment had no effect on responses of aortic rings from eNOS−/− mice, showing that other sources of NO (e.g., inducible NOS and neuronal NOS) do not typically affect vascular contractility to these agents in mouse aorta.
Interactions between rhoA/rho-kinase and NO.
The small G protein rhoA and its downstream effector rho-kinase contribute to agonist-induced vascular smooth muscle contraction via Ca2+ sensitization (11, 37). Both the GPCR agonists serotonin and phenylephrine have been reported to elicit vascular contraction via this pathway (30, 37). Recent data suggest that there may be several levels of interaction between the rhoA/rho-kinase signaling pathway and eNOS/NO within the vasculature. First, rhoA is reported to downregulate eNOS expression by decreasing its mRNA stability in cultured endothelial cells, possibly through effects on the cellular actin cytoskeleton (21, 25). Several other studies (26, 32, 40) have shown that both cGMP (which is stimulated by NO) and its downstream effector, cGMP-dependent protein kinase, can functionally antagonize Ca2+ sensitization by activating MLCP or inhibiting rhoA, respectively. Thus there is evidence that both the NO/cGMP and rhoA/rho-kinase signaling pathways can each suppress the function of the other. We therefore speculated that the augmented contractility to serotonin and phenylephrine in aorta from eNOS−/− mice is due to rho-kinase function being augmented in the absence of opposition from endothelium-derived NO. Our data, which showed that Y-27632 abolished or profoundly inhibited contractile responses to serotonin and phenylephrine in eNOS−/− mice, is consistent with this proposal.
Rho-kinase and NO function in vascular disease.
Several recent reports have provided strong evidence that vascular rhoA/rho-kinase activity is elevated in disease states such as hypertension (5, 27, 33, 37), cerebral (31) and coronary (15) vasospasm, arterial lesion formation (28), and vascular hypertrophy (41). This phenomenon therefore appears to occur in parallel with the widely accepted characteristic of diminished levels of endothelium-derived NO in such vascular diseases (10). Thus, given that activated rhoA can reportedly inhibit eNOS expression (24), excessive stimulation of the procontractile rhoA/rho-kinase pathway could exacerbate vascular dysfunction, in part by reduction of eNOS expression. On the other hand, and consistent with the present findings, is the concept that NO deficiency may result in increased rho-kinase-mediated Ca2+ sensitization in vascular smooth muscle due to lack of rhoA/rho-kinase suppression by endothelial NO. Thus elevated rho-kinase activity in the absence of NO would be expected to require higher concentrations of a rho-kinase inhibitor to normalize or attenuate exaggerated contractile responses to certain agonists, as was observed here. Indeed, although we found responses of aortic rings from both mouse strains to serotonin and phenylephrine to be dependent on rho-kinase activity, GPCR responses of eNOS+/+-derived rings were much more sensitive to rho-kinase inhibition than were eNOS−/− vessels. By contrast, however, responses of both strains to KCl were relatively unaffected by Y-27632, with maximum KCl contractions remaining ≥85% of the KPSS response in the presence of 10 μM Y-27632. This may indicate that in both strains, rho-kinase substantially (and probably completely) mediates aortic contractile response to these receptor agonists, but not to KCl, as predicted. It is unlikely that the modest effects of Y-27632 on KCl responses were actually due to inhibition of a component of contraction caused by norepinephrine released during depolarization of adventitial nerves, and thus mediated by activation of α-adrenoceptors and rhoA/rho-kinase, because similar effects were observed in the presence of the α1-adrenoceptors antagonist prazosin (data not shown). As mentioned, our findings support the concept that responses to serotonin and phenylephrine in eNOS−/− vessels are indeed augmented due to increased activity of rhoA/rho-kinase in the absence of NO. Another recent study in rat aortic segments either endothelium denuded or treated with a NOS inhibitor also concluded that NO can oppose rho-kinase-mediated contractions and that there is greater sensitivity to rho-kinase inhibition by Y-27632 (and therefore perhaps less rho-kinase activity) when NO is present (4).
Besides lowering elevated plasma cholesterol levels, statins are now known to confer several other benefits for cardiovascular function, including stabilization of atherosclerotic plaques (36), limiting cerebral infarct size (1, 9, 22), and reducing oxidative stress (8, 38). Thus far, such beneficial effects of statins have been thought to be solely due to their ability to increase eNOS expression and activity and hence increase NO bioavailability in the vasculature (14, 18, 23). The current concept of how this occurs is that by blocking the isoprenylation of rhoA in endothelium, statins prevent its ability to translocate from the cytosol and embed in the membrane, thereby preventing its activation by binding GTP in exchange for GDP (24, 36). In this way, prevention of the isoprenylation of rhoA results in disinhibition of eNOS expression by increasing eNOS mRNA stability (21, 24). Theoretically, however, statins may be equally able to modulate vascular contraction through direct interference with rhoA/rho-kinase signaling in vascular smooth muscle, independently of any additional indirect effects on eNOS function.
Using eNOS−/− mice, we directly investigated the importance of eNOS expression in modulation of rho-kinase-mediated vascular contraction after chronic statin treatment. Indeed, chronic mevastatin treatment suppressed contractile responses of aortic rings from both eNOS+/+ and eNOS−/− female mice to serotonin and phenylephrine, indicating that such an effect of statins can occur independently of NO. The present dose of mevastatin did not affect plasma cholesterol levels, as reported previously (1, 9). Also noteworthy was the relative lack of effect of mevastatin treatment on KCl-induced contractile responses, analogous to the selectivity of rho-kinase inhibition in attenuating contractions to serotonin and phenylephrine. Thus the data are consistent with the effect of mevastatin occurring via interference with rhoA/rho-kinase signaling in vascular smooth muscle. In any case, it is apparent that even in eNOS+/+ mice, mevastatin treatment probably did not substantially increase levels of endothelial NO release because neither ACh-induced relaxation nor l-NAME-induced contraction were enhanced. These effects of mevastatin were found to occur without any change in total rhoA or rho-kinase expression, although it is possible that the level of activation or membrane association of these proteins was modulated by the treatment.
Gender difference in NO activity.
We noted a gender difference in contractile responses to serotonin (but not phenylephrine or KCl) in aortic rings from eNOS+/+ but not eNOS−/− mice, with responses being greater in males than in females. As mentioned, vascular contractions induced by serotonin can be modulated in some vessels by the simultaneous release of NO via activation of endothelial serotonergic receptors (6). A similar gender difference in response to serotonin was recently reported in mouse carotid artery (19). This effect could be due to estrogen-dependent upregulation of endothelial serotonergic receptors or their subsequent signaling to eNOS, to increase NO release in response to serotonin in females, as opposed to any general effect of estrogen on endothelial NO synthesis. This notion that serotonin-induced NO release from eNOS is greater in females than males is further supported by our observations that 1) contractions to l-NAME after serotonin (but not phenylephrine or KCl) precontraction were greater in females than in males; 2) there was no gender difference in responses to serotonin in eNOS−/− mice; 3) there was no gender difference in contractile responses to phenylephrine or KCl in either strain; and 4) there was no gender difference in the relaxant response to ACh in eNOS+/+ mice. Thus this gender-related difference in response to serotonin is clearly both agonist and eNOS related.
The results of this study demonstrate that both chronic and acute NO deficiency lead to augmented contractility of the mouse aorta, particularly in response to two GPCR agonists. Furthermore, sensitivity to rho-kinase inhibition is decreased in the absence of endothelial NO production. Taken together, these observations strongly suggest a novel physiological role for NO in specifically opposing vascular contraction occurring through GPCR-related rho-kinase-mediated Ca2+ sensitization. This study has also shown that chronic mevastatin treatment can suppress vascular contractile responses dependent on rho-kinase. This effect of statin treatment can occur independently of eNOS/NO, suggesting a more direct effect perhaps by disrupting rhoA/rho-kinase signaling in vascular smooth muscle. Given that excessive activity of the rhoA/rho-kinase pathway has been implicated in the pathology of a range of vascular disorders, our data support the notion that this could be the direct result of decreased NO bioactivity and are consistent with the observed cholesterol-independent clinically beneficial effects of statins in such conditions.
These studies were supported by a Project Grant from the National Health and Medical Research Council of Australia (208969).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2004 the American Physiological Society