Increased Microparticle Production and Impaired Microvascular Endothelial Function in Aldosterone-Salt-Treated Rats: Protective Effects of Polyphenols

We aimed to characterize circulating microparticles in association with arterial stiffness, inflammation and endothelial dysfunction in aldosterone-salt-induced hypertension in rats and to investigate the preventive effects of red wine polyphenols. Uninephrectomized male Sprague-Dawley rats were treated with aldosterone-salt (1 µg.h−1), with or without administration of either red wine polyphenols, Provinols™ (20 mg.kg−1.day−1), or spironolactone (30 mg.kg−1.day−1) for 4 weeks. Microparticles, arterial stiffness, nitric oxide (NO) spin trapping, and mesenteric arterial function were measured. Aldosterone-salt rats showed increased microparticle levels, including those originating from platelets, endothelium and erythrocytes. Hypertension resulted in enhanced aortic stiffness accompanied by increased circulating and aortic NO levels and an upregulation of aortic inducible NO-synthase, NFκB, superoxide anions and nitrotyrosine. Flow-induced dilatation was reduced in mesenteric arteries. These effects were prevented by spironolactone. Provinols™ did not reduce arterial stiffness or systolic hypertension but had effects similar to those of spironolactone on endothelial function assessed by flow-mediated vasodilatation, microparticle generation, aortic NO levels and oxidative stress and apoptosis in the vessel wall. Neither the contractile response nor endothelium-dependent relaxation in mesenteric arteries differed between groups. The in vivo effects of Provinols™ were not mediated by mineralocorticoid receptors or changes in shear stress. In conclusion, vascular remodelling and endothelial dysfunction in aldosterone-salt-mediated hypertension are associated with increased circulating microparticles. Polyphenols prevent the enhanced release of microparticles, macrovascular inflammation and oxidative stress, and microvascular endothelial dysfunction independently of blood pressure, shear stress and mineralocorticoid receptor activation in a model of hyperaldosteronism.


Introduction
Mineralocorticoid administration to uninephrectomized rats drinking a concentrated salt solution is a well known model of experimental hypertension and the degree of hypertension is dependent on the duration of the aldosterone-salt regimen. An increase in systolic arterial pressure (SAP) was observed after administration of aldosterone and salt for 4 weeks in rats. Key aspects of this model are the maintenance of the deleterious effect of salt in presence of relatively high levels of aldosterone and the presence of proteinuria and increased oxidative stress compared with uninephrectomized rats on a normal salt diet [1]. Furthermore, an increase in arterial stiffness, assessed by the incremental elastic modulus (Einc), was also observed within this time-frame [2]. The mineralocorticoid receptor antagonists, eplerenone and spironolactone, have been shown to prevent these arterial changes, together with the associated cardiac fibrosis, independently of left ventricular hypertrophy, via blood pressure reduction [3]. In hypertensive patients, eplerenone treatment has been recently shown to reduce stiffness in resistance arteries [4]. For endothelial function, a number of studies have shown that chronic treatment with aldosterone resulted in impaired acetylcholine-induced relaxation of the aorta in both normotensive and hypertensive rats resulting from inflammatory and oxidative processes [5].
Microparticles may also play a key role in endothelial dysfunction and vascular oxidative stress. Indeed, microparticles are small membrane vesicles that are shed from cells in response to activation and apoptosis in various disease states [6]. Because of their presence in the bloodstream, circulating microparticles may be considered to play a major role in interactions with circulating cells or the components of the vessel wall. Recently, it has been reported that hypertensive patients displayed both elevated levels of circulating endothelial microparticles and increased arterial stiffness assessed by brachial-ankle pulse wave velocity [7]. Furthermore, microparticles promote endothelial dysfunction in metabolic syndrome patients with moderate hypertension [8]. To the best of our knowledge, the circulating levels of microparticles according to their cellular origin have never been characterized in aldosterone-salt hypertension.
Red wine polyphenols, including Provinols TM , possess beneficial properties for preventing cardiovascular disorders by their influence on nitric oxide (NO) balance and prevention of oxidative stress. These compounds can decrease blood pressure in normotensive rats [9], an effect which involves the NO pathway because polyphenols are able to induce ex vivo endothelium-dependent relaxation as a result of enhanced NO synthesis via endothelial NO synthase (eNOS) [10]. We [11,12] and other groups [13] have provided evidence that Provinols TM reduce blood pressure in several experimental models of hypertension, and can correct cardiovascular remodelling and vascular dysfunction. Furthermore, our recent manuscript shows that delphinidin in Provinols TM activated the NO pathway (Src, ERK 1/2, eNOS, caveolin-1), leading to NO production in endothelial cells via activation of the a isoform of the oestrogen receptor [14]. An increase in NOS activity, the prevention of oxidative stress, and a reduction in the inflammatory process may explain the effects of Provinols TM .
The uninephrectomized aldosterone-salt rat model is widely used for measuring the cardiovascular effects of aldosterone. We chose to use the classical reference group of uninephrectomized rats as controls and not uninephrectomized animals receiving a high salt diet, since the renin level is markedly decreased in these latter animals [15]. We consider also the model of aldosterone infusion and salt without uninephrectomy as an unsuitable control, because the specific involvement of mineralocorticoid receptor has not been definitively established. Indeed, the effects of mineralocorticoid receptor antagonism have never been described in this model [16].
In the present study, our aim was first to characterize circulating microparticles according to their cellular origin and to study their role in association with arterial stiffness, vascular inflammation and endothelial dysfunction in aldosterone-salt-induced hypertension from uninephrectomized rats. Secondly, we investigated the cardio-protective effects of Provinols TM on the consequences of aldosterone-salt-induced hypertension on the above parameters, i.e. microparticles, large artery stiffness, inflammation and oxidative stress, and flow-induced dilatation in uninephrectomized rats. The effects of Provinols TM alone were compared with those of spironolactone alone in another series of uninephrectomized control rats.

Animals
Male Sprague-Dawley (SD) rats (n = 149), uninephrectomized at 7 weeks of age, were purchased when 8 weeks old from Charles River (France). All rats were housed in individual cages and were fed a standard rat chow and tap water ad libitum. They were maintained in a quiet room at constant temperature (20 to 22uC) and humidity (50 to 60%). The rats were randomly divided into 6 groups. In the control group (n = 41), the uninephrectomized SD rats received normal food for 4 weeks without aldosterone (Sigma-Aldrich, St Louis, MO) administration. In the spironolactone (Sigma-Aldrich) group (n = 12), rats received spironolactone (30 mg.kg 21 .day 21 ) in the food (based on an average consumption of 25 g.day 21 ) for 4 weeks without aldosterone administration. In the Provinols TM group (n = 12), rats received Provinols TM 20 mg.kg 21 .day 21 in their food [17] without aldosterone administration. In the aldosterone-salt group (n = 28), osmotic mini-pumps were implanted subcutaneously between the shoulder blades (Alzet model 2004, 1 mg.h 21 ) for continuous subcutaneous infusion for 4 weeks. Osmotic mini-pumps were filled with aldosterone and incubated for several hours at 37uC prior to implantation. Surgery was performed using an aseptic technique and the rats were anaesthetized with isofluorane 1.5% in oxygen. The aldosterone-salt group rats received 1% NaCl solution (plus 0.1% KCl) to drink, provided in a single drinking burette. In the aldosterone-salt spironolactone group (n = 26), aldosteronesalt rats received spironolactone (30 mg.kg 21 .day 21 ) in the food and in the aldosterone-salt Provinols TM group (n = 30), aldosterone-salt rats received Provinols TM (20 mg.kg 21 .day 21 ) in their food. Provinols TM were obtained from the Société Française des Distilleries Union de Coopératives Agricoles (Vallon Pont d'Arc, France). The composition, in mg.g 21 of dry powder, is: 480 proanthocyanidins, 61 total anthocyanins, 19 free anthocyanins, 38 catechin, 18 hydroxycinnamic acids, 14 flavonols and 370 polymeric tannins. All procedures were conducted in accordance with the Guide for the Care and Use of Laboratory Animals published by the US National Institute of Health and were approved by the local ethical committee ''Comité regional Nancy-Lorraine/Nord-Est'' (nu B54-547-20).

Microparticle Isolation and Characterization
Blood obtained by intracardiac puncture was collected in citrated tubes and processed for assay within 2 hours. Samples were centrifuged 3 minutes at 1900 g and then 4 minutes at 5000 g to obtain platelet-free plasma (PFP) and stored at 280uC until subsequent use. Discrimination between the various membrane microparticle subpopulations present in PFP was achieved by analyzing the expression of membrane-specific antigens, as previously described [18]. Quantification and phenotype of endothelial, platelet, and erythrocyte microparticles were performed using anti-CD54-biotin (1.5 mg.ml 21 ), anti-CD61-FITC (2.5 mg.ml 21 ) (BioLegend, San Diego, CA) and anti-erythroid cellbiotin (2.5 mg.ml 21 ) (BD Biosciences, San Jose, CA) labeling. For biotin-conjugated antibodies, streptavidin-FITC (2.5 mg.ml 21 ) (Sigma-Aldrich) was added for 15 minutes. Eight ml of plasma were incubated with 5 ml of specific antibody (Beckman Coulter, Villepinte, France). Irrelevant rat IgG was used as an isotypematched negative control for each sample. After 45 minutes of incubation at room temperature, samples were diluted in 300 ml of 0.9% saline salt solution. Then 8 ml of Flowcount beads were added and samples were analyzed in a flow cytometer 500 MPL system (Beckman Coulter).

Von Willebrand Factor ELISA
Circulating von Willebrand factor (vWF) antigen was measured by ELISA (AsserachromH, Diagnostica Stago, Asnières, France). Each plasma sample was assayed at two different dilutions chosen in order to interpolate results using the calibration curve obtained with the human calibrator supplied within the kit.

Carotid Artery Stiffness
We performed simultaneous recording of arterial diameter (left carotid artery) and blood pressure (right carotid artery) in pentobarbital-anesthetized rats as has been described previously [2]. Arterial diameter measurement was obtained by using an ultrasonic echo-tracking device (Diarad-00, Asulab SA). The relationship between the pressure (P) and the lumen cross-sectional area (LCSA) was fitted with the model of Tardy et al. [19] by using an arctangent function and three optimal fit parameters (a, b, c) as follows: Carotid cross-sectional distensibility (Dist), a derivative of this function, was used to assess the global elastic behaviour of the artery. Circumferential wall stress (s) and Einc, which characterises the intrinsic mechanical properties of the wall material, were calculated with the above-mentioned parameters. Dist, s and Einc are given by the following equations: where the media cross sectional area (MCSA) was determined by histomorphometry. The reproducibility was 361% (inter-observer coefficient of variation of carotid diameter).

Aortic Composition
Insoluble elastin, total collagen and cell protein contents were measured on descending thoracic aortae without homogenization, as described previously [20]. Briefly, aortic segments were opened longitudinally, the media separated from the adventitia and the medial length measured under a microscope. Media were then defatted, dried and their dry weight recorded. Medial cell proteins were extracted by 0.3% sodium dodecyl sulfate (SDS) and subsequently assayed and insoluble elastin was purified by the hot alkali method and quantified by weighing. Proteins in the NaOH extract were then hydrolysed, and total medial collagen was quantified by assaying hydroxyproline in the hydrolysate, using a colorimetric assay.

NO Spin Trapping and Electronic Paramagnetic Resonance Technique (EPR) Studies
The detection of NO production was performed using the technique with Fe 2+ diethyldithio-carbamate (DETC, Sigma-Aldrich) as spin trap [21]. To measure NO production in blood, we injected 400 mg.kg 21 of DETC (1 ml.g 21 weight) intraperitoneally and 40 mg.kg 21 of FeSO 4 .7H 2 O (1 ml.g 21 weight), in a solution containing sodium citrate dehydrate 200 mg.ml 21 , subcutaneously. After 30 minutes, rats were anesthetized with isoflurane and killed, and venous blood was obtained for NO measurements. These studies were performed on a table-top xband spectrometer Miniscope (Magnettech, Berlin, Germany). Recordings were made at 77uK, using a Dewar flask. Instrument settings were 10 mW of microwave power, 1 mT of amplitude modulation, 100 kHz of modulation frequency, 60 seconds of sweep time and 3 scans. Blood samples exhibited an EPR characteristic of signals derived from NO-Fe(DETC) 2 [22]. The quantitative measurement of the NO-Fe(DETC) 2 signal amplitude was expressed in relative units.ml 21 of blood (amplitude per ml).
In another set of experiments, animals were killed and aorta dissected and incubated for NO production in 250 ml of Krebs-Hepes buffer solution, then treated with 250 ml of Fe(DETC) 2 and incubated for 45 minutes at 37uC. NO detection was measured in situ by EPR as described above, and values were corrected for the dry weight of the sample (dehydrated sample) in mg and expressed as amplitude signal per mg (W ds ).

Aortic Nitrotyrosine Staining
Fixed sections from the aortae were incubated in blocking buffer (5% non-fat dry milk) for nitrotyrosine. Then, tissue sections were incubated overnight (4uC) with a mouse monoclonal antinitrotyrosine (clone 1A6) antibody (1:100, Millipore, Billerica, MA, ref: 05-233) in blocking buffer. Three washes were followed by incubation with secondary murine fluorescence-labeled antibody Alexa fluor-488 (1:100, Molecular Probes, Cergy Pontoise, France). Controls were performed by omission of the first or second antibody. After washing, vessel sections were mounted on glass slides. Confocal equipment mounted on a Nikon Eclipse, TE 200-S, inverted microscope was used for optical sectioning of the tissue. Digital image recording was performed using the QED In Vivo Software.

Vascular Reactivity
Second generation branches of small mesenteric arteries (SMA) were carefully removed and arterial segments were mounted on a wire myograph filled with physiological salt solution (PSS) of the following composition in mM: NaCl 130, NaHCO 3 [10]. Briefly, SMA were stretched to 200 mg, lengths that yielded circumferences equivalent to 90% of those which the vessels would have had with an intramural pressure of 100 mmHg (i.e., the pressure corresponds to the physiological value). PSS was kept at 37uC and gassed with 95% O 2 and 5% CO 2 at pH 7.4. Mechanical activity was recorded isometrically by a force transducer (Danish Myo Technology, Aarhus, Denmark). After setting the vessels to their working length, the contractile response to KCl (80 mM) was assessed. After washing of arteries, challenges with phenylephrine (Sigma Aldrich) were performed to elicit a contractile response and to test their maximal contractile capacity. Arteries were pre-contracted at 80% of their maximal contraction with phenylephrine. When the contraction reached a plateau, cumulative application of acetylcholine (1 nM to 10 mM, Sigma Aldrich) was performed. The presence of functional endothelium was assessed by the ability to induce more than 50% relaxation of vessels pre-contracted with phenylephrine. Parameters of vascular sensitivity were analyzed by means of pD 2 = -log EC 50 , EC 50 being the molar concentration of the agonist that produces 50% of the maximal effect; EC 50 values were calculated by logit-log regression.
The presence of endothelium was tested by applying acetylcholine after pre-contraction of the arteries with KCl-depolarization (80 mM). Diameter changes were measured when intra-luminal pressure was increased from 10 to 125 mmHg. Pressure was then set at 75 mmHg and dilatation in response to flow was evaluated. Dilatation response-to-flow experiments were repeated in series in presence of the NOS inhibitor, N $ nitro-L-arginine (L-NA, 100 mM, Sigma-Aldrich). The NO-dependent dilatation was calculated as the difference between the dilatation without inhibitor and that in the presence of L-NA.

Shear Stress-induced Microparticle Formation
Shear-induced blood cell activation was evaluated in vitro using an erythroaggregometer (Couette flow, SD Medical), subjecting citrated anticoagulated whole blood (from Sprague Dawley rats via a carotid catheter under isoflourane anaesthesia) (129 mM; v/v 9:1) to shear rates of 300 s 21 for 2 minutes at 37uC. Samples were incubated with aldosterone at 10 nM for 10 min. The following antagonists were added 10 min prior to the addition of aldosterone: the mineralcorticoid receptor antagonist, spironolactone, at 1 mM, Provinols TM at 0.01 g.l 21 or the oestrogen receptor antagonist, fulvestrant (Fulv; Sigma-Aldrich), at 1 mM. Controls were performed in which citrated whole blood was in contact with the cup and bob for 2 minutes at 37uC but not exposed to shear stress. Samples were then centrifuged for 10 minutes at 190 g to obtain platelet-rich plasma (PRP) and then for 3 minutes at 1900 g. The supernatant was centrifuged for 4 min at 5000 g to obtain the platelet-poor plasma (PPP). Platelet activation and the generation of cell-derived microparticles in sheared blood samples were evaluated using the chromogenic assay measuring the phospholipid-related procoagulant activity (PPA) as previously described [24].

Vascular Smooth Muscle Cell Isolation and Culture
Primary rat aortic vascular smooth muscle cells (VSMCs) were isolated by using standard enzymatic dissociation techniques, as previously reported [25]. Cells were plated and grown in Dulbecco's modified Eagle's medium (DMEM) supplemented with 5% fetal bovine serum, and cultured at 37C in air containing 5% CO 2 . Cells were used at passage 2. Before treatment, cells were serum-starved for 12 hours.

Statistical Analysis
All values are expressed as means 6 SEM. One-way analysis of variance was performed to compare the different groups of rats. The Fisher test was used for inter-group comparisons. Einc was log transformed to generate linear relationships [28]. After this transformation, we calculated the mean slopes of the curves. If the slopes were not significantly different, we compared the curves within the common range of Einc for WS by calculating WS at 1500 kPa of Einc. Differences were considered significant at values of P,0.05. Statistical analyses of PD 2 values and maximal effects of acetylcholine were performed by two-way analysis of variance, and nonparametric Mann-Whitney U tests or analysis of variance for repeated measures and subsequent Bonferroni post hoc test.

Arterial Pressure and Organ Weights
Aldosterone-salt treatment increased significantly SAP and pulse pressure (PP) compared to controls. There were no significant changes observed in body weight, diastolic or mean blood pressures or in carotid artery diameter. The significant increase in SAP and PP due to aldosterone-salt treatment was prevented by spironolactone treatment but not by Provinols TM (Table 1). When administered alone, spironolactone but not Provinols TM exerted an inhibitory effect on blood pressure (Table  S1). Spironolactone and Provinols TM treatment increased body weight compared to the aldosterone-salt group ( Table 1). The increases in heart and kidney weights induced by the aldosteronesalt treatment were not influenced by Provinols TM whereas spironolactone partially prevented the increase in heart weight.

Microparticle Counts and their Cellular Origin
Aldosterone-salt treatment increased significantly the total number of circulating microparticles when compared to control ( Figure 1A). This increase in the total number of microparticles was completely prevented by both spironolactone and Provinols TM . Neither spironolactone nor Provinols TM alone had any effect on the total number of microparticles ( Figure 1A). Phenotypical characterization of the cellular origin of microparticles showed significant increases in platelet-, endothelial-and erythrocyte-derived microparticles in aldosterone-salt-treated rats ( Figures 1B-D) which were all inhibited by both spironolactone and Provinols TM . In addition, both spironolactone and Provinols TM alone significantly reduced platelet-and endothelialderived microparticles, without modifying erythrocyte-derived microparticle number ( Figure 1B-D).

Von Willebrand Factor
There was an increase in circulating vWF in aldosterone-salt rats (3.460.4 versus 2.260.3 mg.ml 21 in controls; n = 7 and 11 respectively; p = 0.02) which was inhibited by both spironolactone (2.260.3 mg.ml 21 ; n = 9) and Provinols TM (2.160.3 mg.ml 21 ; n = 10). Figure 2A shows Einc-WS curves in four groups of rats. To compare these curves, we calculated the mean WS at 1500 kPa of Einc ( Figure 2B). Comparison of Einc-WS curves from aldosterone-salt and control rats showed a significant reduction in WS at 1500 kPa of Einc in aldosterone-salt rats indicating increased arterial stiffness. In the spironolactone group, this increase was not observed. However, Provinols TM did not significantly prevent aldosterone-salt-induced arterial stiffness. The Einc-WS curves from rats treated with spironolactone or Provinols TM alone were not different from the control group ( Figure S1).

Aortic Composition
Thoracic aortic medial dry weight per cm length, and cell protein content were increased significantly by the aldosteronesalt treatment compared with control rats (Table 2), indicating that hypertrophy of the media had occurred. The increased dry weight reflects not only the increase in smooth muscle cell content but also an increase in some extracellular components not quantified here (principally proteoglycans and glycoproteins) and deduced by substraction. The significant increase in aortic dry weight due to aldosterone-salt was not observed under co- treatment with spironolactone or Provinols TM . However, only spironolactone, but not Provinols TM , was able to significantly attenuate the aldosterone-salt-induced increase in cell proteins.
A significant decrease in the percentage of elastin in the thoracic aorta with aldosterone-salt treatment was observed, which was only partially corrected by both spironolactone and Provinols TM treatments. There were no changes in collagen with aldosteronesalt treatment. However, as the proportion of elastin was decreased, there was a significant reduction in the elastin to collagen ratio after aldosterone-salt treatment. Both spironolactone and Provinols TM partially prevented this reduction in elastin/ collagen providing values approaching those of controls.

NO production, iNOS, NFkB p65 and phosphorylated-IkB alpha activation
The NO-Fe(DETC) 2 EPR signal was greater in blood and aorta from aldosterone-salt treated rats than in control rats ( Figures 3A  and 3B, respectively). This increase was completely prevented by spironolactone or Provinols TM . Neither spironolactone nor Provinols TM alone had any effect on aortic NO production ( Figure S2).
Western blot analysis showed an increase in iNOS expression in aortae from aldosterone-salt rats compared to those from control rats (Figures 3C and 3D). Much less staining was observed following treatment with spironolactone or Provinols TM . The iNOS expression was associated with an increase in NFkB p65 expression in aortae from aldosterone-salt rats compared to control rats (Figures 3C and 3E). Activation of the transcriptional factor, NFkB p65, in aortae from aldosterone-salt treated rats was confirmed by the increased levels of phosphorylated-IkB alpha compared to control vessels ( Figures 3C and 3F). All these effects of aldosterone-salt treatment on iNOS and NFkB p65 expression and phosphorylated-IkB alpha were completely prevented both by spironolactone and by Provinols TM .

Oxidative and Nitrosative Stresses
EPR measurement of oxidative stress demonstrated that aortae from aldosterone-salt-treated rats displayed an increased O 2 production compared with controls that was prevented by both spironolactone and Provinols TM ( Figure 4A). Neither spironolactone alone nor Provinols TM alone affected O 2 production ( Figure 4).
Aortae from aldosterone-salt rats also showed a marked increase in nitrotyrosine staining, reflecting an increased degree of protein nitration in the vascular wall ( Figures 4B and C). Weak staining, comparable to that obtained in control aorta, was found in vessels from aldosterone-salt rats treated either with spironolactone or Provinols TM (Figures 4D and E).

Vascular Reactivity in Small Mesenteric Arteries
The contractile response to KCl-depolarization (80 mM) did not differ between control rats and the other groups. However, contraction to KCl was increased in aldosterone-salt rats when compared with the treatment spironolactone alone. Endotheliumdependent relaxation in response to acetylcholine was not significantly different in vessels harvested from the different groups of rats (Table 3). Flow increases induced a significant vasodilator response in all the 4 groups of rats studied ( Figure 5A). This flow-induced dilatation was significantly reduced in small mesenteric arteries from the aldosterone-salt-treated rats compared to those from control. Treatment by spironolactone and Provinols TM normalized the effect of aldosterone-salt on flow-induced dilatation.
Moreover, the NO-dependent dilatation ( Figure 5B) was calculated as the difference between the dilatation without L-NA, represented in Figure 5A, and the dilatation in presence of L-NA (not shown). The NO-component of flow-induced dilatation was reduced in arteries from aldosterone-salt-treated rats compared to vessels from control rats. Treatment either with spironolactone or Provinols TM largely corrected this aldosterone-salt-induced effect. Figure 6A shows the PPA ratio values in PPP obtained from sheared and unsheared samples of total rat blood. Shear stress induced an increase in microparticle production in control samples. Treatment with aldosterone potentiated this shearinduced production. This effect was completely inhibited by spironolactone but not by Provinols TM . This absence of mineralocorticoid receptor antagonist activity of Provinols TM was further supported by the inability of Provinols TM to block the aldosterone-induced increase in the expression of MDM2, gp91 and cardiotrophin-1 (CT-1) by cultured primary rat vascular smooth muscle cells (Figure 6 B-E). Estrogen receptor blockade with fulvestrant inhibited the effect of shear stress. However, in the absence of shear stress, aldosterone did not induce microparticle production by cultured human aortic endothelial cells (12266 versus 13064 nM equivalent phosphatidylserine in control cells).

Spironolactone and Provinols Prevent Aldosteroneinduced Apoptosis in the Vessel Wall
As shown in Figure 7, vessels from aldosterone-salt group display expression of cleaved caspase-8 and caspase-3, indicating that aldosterone-salt treatment induces apoptosis of vessel wall. In addition, both Spirolactone and Provinols prevented aldosteronesalt-induced apoptosis as reflected by the reduction of caspase cleavage.

Discussion
We report that aldosterone-salt-induced hypertension produced a marked increase in both circulating microparticles, especially those from platelets, endothelium and erythrocytes, and circulating vWF. The microparticle increase was accompanied by increases in aortic stiffness, oxidative and nitrosative stresses and microvascular endothelial dysfunction as well as induction of apoptosis in the vessel wall. Of particular interest is that Provinols TM prevented changes in circulating microparticles, microvascular endothelial dysfunction, oxidative and nitrosative stress and apoptosis. The effects of these polyphenols were not mediated by mineralocorticoid receptors or changes in shear stress. However, Provinols TM did not block the aldosterone-salt-induced increases in aortic hypertrophy, stiffness or SAP.
Microparticle levels reflect the severity of endothelial alterations and of platelet activation. Although the precise mechanisms of microparticle production remain unclear, it has been emphasized that shear stress contributes to microparticle release whatever their cellular origin, as we have demonstrated in our in vitro studies. An effect of hypertension on microparticle production is supported by the correlation between plasma levels of endothelial and platelet microparticles and blood pressure in patients with severe untreated hypertension [29]. Besides reflecting the state of the parent cells, there is growing consensus on the many facets of microparticles which can act as diffusible messengers with either beneficial or deleterious missions [6,30]. In particular, a functional or causative role of microparticles on endothelial dysfunction has been reported in several studies involving in vitro incubation of microparticles with endothelial cells or rings from thoracic aortae or their injection into mice [8,[31][32][33][34][35][36]. Although a possible direct effect of microparticles was not assessed on different types of arteries, they might contribute to the blood vessel inflammation observed in the present study (see below). For example, microparticles harvested from patients with pre-eclampsia induce activation of the transcription factor RelA/NFkB associated with an up-regulation of proinflammatory protein expression, namely iNOS and COX-2, in the vessel wall. The proinflammatory properties of these microparticles lead to the observation of surrogate signs of oxidative and nitrosative stress in the vascular wall [26]. The same type of inflammation was seen in the aorta from aldosteronesalt-treated rats. Finally, spironolactone or Provinols TM alone decreased endothelial-derived microparticle levels. Spironolactone [37] as well as red wine polyphenols [38] protect endothelial cells against apoptosis, a cellular process releasing microparticles. Similar mechanisms may explain the effects of spironolactone and Provinols TM on endothelial-derived microparticle release. This hypothesis is reinforced by the fact that spironolactone or Table 3. Contractile effects of KCl (80 mM) and endothelium-dependent relaxation to acetylcholine in rat mesenteric resistance arteries.

KCl-induced contraction mN/mm
Endothelium-dependent relaxation to acetylcholine Provinols TM prevented apoptosis induced by aldosterone-salt treatment at the level of the vessel wall.
Blockade of the aldosterone pathway by spironolactone or treatment by Provinols TM prevented the increase in circulating microparticles and plasma vWF as well as the associated vascular inflammation. Recently, aldosterone has been reported to activate endothelial vWF exocytosis at the cellular level [39]. Although microparticle generation and vWF release represent distinctive phenomena they may both be involved in aldosterone-salt-induced vascular inflammation. Even though it has been shown that selective mineralocorticoid receptor inhibition decreases platelet activation, the mechanism responsible for these effects has not been defined [40]. Lijnen and Petrov [41] described that spironolactone decreased intraerythrocyte and intraplatelet Na + and Ca 2+ by decreasing Na + -pump activity. It is possible that the reduction in platelet microparticle levels observed in the present study is associated with a lower increase in intracellular Ca 2+ in platelets because microparticle release is a Ca 2+ -dependent process [6]. Polyphenols lead to inhibition of whole-blood aggregation suggesting a potential mechanism for the beneficial effects of polyphenols by the suppression of platelet-mediated thrombosis [42].
A significant correlation between endothelial microparticles and the loss of flow-mediated dilatation has been previously reported in obstructive sleep apnea [43] and in patients with end-stage renal failure [44] supporting the idea that these measurements are reliable indices of endothelial dysfunction. Thus, reduction of circulating microparticles may protect against endothelial dysfunction as evidenced by normalization of flow-mediated dilatation and intense vascular wall inflammation in our experimental model of hypertension. It is noteworthy that Provinols TM did not prevent the in vitro aldosterone-induced increase in microparticle production via shear stress. Although the in vitro conditions may not completely mimic the in vivo situation regarding extracellular Ca 2+ and the level of shear stress, this result suggests that the in vivo effect of Provinols TM on microparticle levels was not dependent on shear stress.
The systolic hypertension observed in this model may be the result of the increase in stiffness of the vascular wall, as previously shown [2]. The significant increase in aortic dry weight and in cell proteins in aldosterone-salt rats reflects arterial wall hypertrophy, in accordance with previous studies [2,45]. The finding of an increase in cell protein content but not of collagen is consistent with previous data showing an absence of collagen accumulation within the media in rats receiving aldosterone for 6 weeks [46]. The decreased elastin/collagen ratio and the increased cell protein content observed could partly explain the increased arterial stiffness, which is independent of any change in collagen.
In our study, in contrast to spironolactone, treatment with Provinols TM was not able to correct the hypertension. Previous studies have, however, reported antihypertensive effects of red wine polyphenol compounds, including Provinols TM , in a rat DOCA-salt model of hypertension [47] and in other rat models; salt-sensitive, angiotensin II administration, NO-inhibition or the spontaneously hypertensive rat. However, in none of these models were the effects of Provinols TM compared to that of a mineralocorticoid receptor antagonist. In our model, systolic hypertension is dependent on the action of aldosterone on the mineralocorticoid receptor as demonstrated by the preventive effect of spironolactone. The hypotensive effect in control rats was also consistent with the permanent vascular tone driven by the microparticles. The absence of effect of Provinols TM on systolic hypertension could support the hypothesis that the effect in other models is related to interference with the renin angiotensin system, the NO pathway or the sympathetic nervous system [11,12]. In agreement with this hypothesis, Provinols has no effect in control rats without changes in these pathways. Furthermore, spironolactone has been shown not to significantly lower systolic blood pressure in Ang IIinfused rat [48]. This suggests that the effects of Provinols TM are not mediated via the mineralocorticoid receptor.
Aldosterone-salt-treated rats exhibited increased circulating and aortic NO levels. These effects were associated with an upregulation of iNOS in the vascular wall which was prevented by spironolactone and Provinols TM . In line with this, we have reported in L-NAME hypertensive rats that the increase of iNOS expression in the aorta is reduced in rats treated with Provinols TM [12]. Moreover, we provide evidence that the effect of aldosterone is also associated with an increased oxidative stress, marked by increased aortic O 2 2 production, probably via activation of the NADPH oxidase pathway [49]. An increase in O 2 2 production leads to an enhancement of its interaction with NO and to the formation of the highly toxic ONOO 2 , as illustrated by the protein nitration in the aorta. Peroxynitrite is produced in large amounts here and has harmful effects in various tissues, especially in blood vessels. We demonstrate that spironolactone prevents the increase in O 2 2 production in the aorta as well as attenuating aldosterone-salt-induced ONOO 2 production. Provinols TM prevented the increase in aortic O 2 2 production, probably due to their capacity to reduce aortic p22phox gene overexpression, observed in accordance with the work of Jimenez et al. [47] in DOCA-salt animals. This effect leads to reduced formation of ONOO 2 and is also concomitant with a decrease in NFkB which is upstream to iNOS induction. Taken together, these data highlight both the anti-inflammatory effect and the decrease in oxidative stress induced by Provinols TM in the aorta, and therefore extend this beneficial action of polyphenols to aldosterone-saltinduced hypertension.
As for endothelial function, studies have shown that chronic treatment with aldosterone results in impaired acetylcholineinduced relaxation of the aorta in both normotensive and hypertensive rats via the production of cyclo-oxygenase 2 metabolites [5]. In small mesenteric arteries, we found that the response to acetylcholine was not altered whereas the vasodilator response to flow was reduced. This endothelial dysfunction in aldosterone-salt-treated rats was associated with a decrease in the NO component of shear stress. The mechanism by which aldosterone impairs flow-induced NO vasodilatation may result from the increase in oxidative stress that reduces endothelial NO bioavailability. This finding is in agreement with a previous report suggesting that increased NADPH-derived peroxide may be one possible mechanism for the reduced NO component of flowmediated vasodilatation in SHR [50]. Provinols TM prevented endothelial dysfunction at the level of the microcirculation which is associated with enhanced endothelium-dependent relaxation resulting from augmented NO generation via increased eNOS activity and expression [9,11,12]. Also, we have reported that red wine polyphenols evoke endothelium-dependent relaxation associated with ERa-stimulation, c-Src/ERK1/2-mediated activation of eNOS, with consequent endothelial NO release [14]. Such a mechanism contributes probably to the beneficial effect of red wine polyphenols.
In conclusion, this study demonstrates that aldosterone-salt administration in rats is able to induce an increase in microparticles which is associated with endothelial dysfunction and vascular inflammation, and leads to the appearance of various oxidative stress markers. These arterial modifications could represent an early step in the development of hypertension. These changes were all reversed by spironolactone in all vascular territories, whereas the action of Provinols TM , which is independent of mineralocorticoid receptor activity, preferentially affected the microcirculation and the NO pathway. The impact of red wine polyphenols on the production of microparticles and flow-induced dilatation provides new insights for future development of therapeutic strategies in aldosterone-mediated hypertension. Figure S1 Carotid artery incremental elastic moduluswall stress (Einc-WS) curves. A. Einc-WS curves from Control, Spironolactone (Spiro) and Provinols TM (Prov) rats (n = 5-9 for each group). B. Mean value of WS at 1500 kPa of Einc. Values are means 6 SEM. (TIF) Figure S2 NO production in aorta. Quantification of the amplitude of NO-Fe(DETC)2 signal (amplitude per mg of dried sample Wds) in aorta from Control, Spironolactone (Spiro) and Provinols TM (Prov) rats (n = 6 for each group).

(TIF)
Table S1 Effects of treatment by spironolactone (Spiro) or Provinols TM (Prov) on blood pressure, carotid diameter and organ weights in rats.