GPER activation ameliorates aortic remodeling induced by salt-sensitive hypertension

Activation of GPER attenuated aortic remodeling in a model of salt-sensitive hypertension without altering blood pressure and was associated with decreased glycosaminoglycans and oxidative stress in the medial layer. Targeting this novel estrogen receptor in women with cardiovascular disease may convey the protective actions associated with endogenous estrogen.

Abstract

The mRen2 female rat is an estrogen- and salt-sensitive model of hypertension that reflects the higher pressure and salt sensitivity associated with menopause. We previously showed that the G protein-coupled estrogen receptor (GPER) mediates estrogenic effects in this model. The current study hypothesized that GPER protects against vascular injury during salt loading. Intact mRen2 female rats were fed a normal (NS; 0.5% Na⁺) or high-salt diet (HS; 4% Na⁺) for 10 wk, which significantly increased systolic blood pressure (149 ± 5 vs. 224 ± 8 mmHg; P < 0.001). Treatment with the selective GPER agonist G-1 for 2 wk did not alter salt-sensitive hypertension (216 ± 4 mmHg; P > 0.05) or ex vivo vascular responses to angiotensin II or phenylephrine (P > 0.05). However, G-1 significantly attenuated salt-induced aortic remodeling assessed by media-to-lumen ratio (NS: 0.43; HS+veh: 0.89; HS+G-1: 0.61; P < 0.05). Aortic thickening was not accompanied by changes in collagen, elastin, or medial proliferation. However, HS induced increases in medial layer glycosaminoglycans (0.07 vs. 0.42 mm²; P < 0.001) and lipid peroxidation (0.11 vs. 0.51 mm²; P < 0.01), both of which were reduced by G-1 (0.20 mm² and 0.23 mm²; both P < 0.05). We conclude that GPER's beneficial actions in the aorta of salt-loaded mRen2 females occur independently of changes in blood pressure and vasoreactivity. GPER-induced attenuation of aortic remodeling was associated with a reduction in oxidative stress and decreased accumulation of glycosaminoglycans. Endogenous activation of GPER may protect females from salt- and pressure-induced vascular damage.

NEW & NOTEWORTHY

Activation of GPER attenuated aortic remodeling in a model of salt-sensitive hypertension without altering blood pressure and was associated with decreased glycosaminoglycans and oxidative stress in the medial layer. Targeting this novel estrogen receptor in women with cardiovascular disease may convey the protective actions associated with endogenous estrogen.

arterial stiffness predicts cardiovascular mortality in hypertensive patients (16). A high-salt diet (HS) exacerbates target organ damage and increases cardiovascular mortality independently of hypertension (15, 33). Excessive salt intake is associated with remodeling of large arteries due to reorganization of the extracellular matrix and hypertrophy of vascular smooth muscle cells (30). Although aortic stiffening is considered a protective adaptation to higher wall stress, decreased elasticity of this conduit vessel results in increased afterload on the left ventricle and higher pulsatile pressure transmitted downstream to smaller arteries and eventually organs (17). Importantly, an increase in aortic stiffness may precede or exacerbate hypertension, particularly during aging (22). Although women exhibit some protection from cardiovascular disease, menopause increases pulse wave velocity and intimal media thickness (9, 34). Arterial stiffening increases cardiac afterload and promotes heart failure with preserved ejection fraction, a disease twice as common in women vs. men (1, 31).

The protective effects of estrogen on vascular injury are evident in both estrogen receptor (ER) α and ERβ knockout mice, indicating that another receptor may mediate this hormone's effects in the vasculature (12, 14). The novel G protein-coupled estrogen receptor (GPER) is a membrane-bound receptor linked to acute signaling pathways (27, 36). Our previous studies in mRen2 hypertensive female rats showed that GPER activation lowers ovariectomy-induced increases in blood pressure and attenuates salt-induced renal and cardiac damage (18, 19). This novel receptor may play an important role in mediating estrogenic effects on vascular remodeling as well.

In the current study, we hypothesized that chronic GPER activation is protective against aortic remodeling in response to salt-sensitive hypertension. To test this hypothesis, we utilized mRen2 female rats, a unique congenic model of hypertension in which an HS diet profoundly elevates blood pressure (7, 8). We assessed aortic remodeling in response to high salt using quantitative histomorphometric analysis and examined the effects of chronic administration of the selective GPER agonist G-1 on blood pressure and the vascular extracellular matrix and oxidative stress.

MATERIALS AND METHODS

Animals.

All procedures were carried out in accordance with the Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee. Hemizygous mRen2 congenic female rats were obtained from the Wake Forest Hypertension Center breeding colony. Rats had free access to food and water in a temperature-controlled room (22 ± 2°C) with a 12-h light to dark cycle. At 5 wk of age, the normal salt diet (NS, 0.5% Na⁺) was switched to HS (4% Na⁺) as previously described (7, 19). The selective GPER agonist G-1 (400 μg·kg⁻¹·day⁻¹; EMD Chemicals, Gibbstown, NJ) or vehicle (veh) was administered for 2 wk beginning at 13 wk of age via subcutaneous osmotic minipump (Model 2ML2; Alzet, Palo Alto, CA). We previously published that this dose of G-1 does not activate ERα and ERβ because there are no alterations in uterine wet weight or body weight after treatment (18). Blood pressure was measured via tail cuff plethysmography (Narco Bio Systems, Houston, TX). Animals were randomly assigned to three experimental groups: NS (n = 7), HS+veh (50% DMSO/saline, n = 9), and HS+G-1 (n = 9).

Vascular reactivity.

After death, the upper thoracic aorta was submerged in formalin for histology and the lower portion used for vascular reactivity as previously described (18). From each animal, four adjacent 2-mm segments from the lower thoracic aorta were cut, and two segments were denuded by rubbing with a stainless steel wire. Rings were mounted on two wires connected to an isometric force transducer (Grass Technologies, West Warwick, RI) and lowered into organ baths filled with Krebs (in mM: 118 NaCl, 25 NaHCO₃, 4.8 KCl, 2.5 CaCl₂, 1.2 MgSO₄, 1.2 KH₂PO₄, and 11 glucose; pH 7.4) and bubbled with 95% O₂ and 5% CO₂ at 37°C. A passive tension of 2 g was previously determined to be optimal for measuring active tension. Rings were contracted with 10⁻⁶ mol/l phenylephrine (PE), and if relaxation to 10⁻⁶ mol/l acetylcholine (Ach) was >50% of the contraction, the ring was considered endothelium-intact. After repeated washings, rings were exposed to 10⁻⁶ mol/l angiotensin II (Ang II). Vessels were again washed and preconstricted with 10⁻⁶ mol/l PE for 10 min, and the relaxation response to 10⁻⁹ to 10^−5.5 mol/l G-1 was measured.

Histology.

Formalin-fixed aortas were embedded in paraffin, cut into 5-μm sections, and mounted on slides. To evaluate aortic wall thickness, slides were stained with hematoxylin, and images analyzed with ImageJ software (National Institutes of Health). The inner and outer circumference of the tunica media was manually traced, and the internal and external cross-sectional areas were recorded. Internal area was subtracted from external area to obtain medial area, and medial area was divided by lumen area to obtain the media-to-lumen ratio. These area measurements were used to back calculate lumen and external diameter by using the equation diameter = 2 × √(area/π). The difference between external and internal diameter is reported as wall thickness. Slides were stained with picrosirius red to determine collagen content. After heating at 55°C for 30 min, deparaffinization in two changes of xylene, and hydration, slides were incubated in a filtered solution containing 0.1% sirius red F3B dissolved in a saturated aqueous solution of picric acid. After 1 h, slides were washed in two changes of 0.5% glacial acetic acid and then dehydrated. Slides were imaged with a polarized light microscope and analyzed with a custom Matlab code to calculate the relative distribution of type I large diameter (red and orange) and type III small diameter (yellow and green) collagen fibers as previously described (5, 37). The NovaUltra Orcein Elastin Stain Kit (IHC World, Woodstock, MD) was used for elastin staining, and glycosaminoglycans were evaluated with the NovaUltra Alcian Blue Stain Kit (IHC World) according to the manufacturer's protocols. For immunohistochemistry, slides were heated at 55°C for 30 min, deparaffinized in two changes of xylene, and hydrated before blocking tissues for 30 min with 5% normal goat serum, 1% bovine serum albumin, and 0.1% Triton dissolved in phosphate-buffered saline. Primary antibodies were diluted in the blocking buffer and incubated at 4°C overnight, followed by secondary antibodies at room temperature for 1 h. 4-Hydroxy-2-noneal (4-HNE, 1:500, EMD Millipore 393207, lot no. 01HR-0620) was used for analysis of oxidative stress, and proliferating cell nuclear antigen (PCNA) (1:1,000, Abcam ab2426, lot no. GR41581-11) and Ki-67 (1:100, Abcam ab66155, lot no. GR134821-1) were used to detect cell proliferation by using standard immunohistochemistry methods and the Metal Enhanced DAB Substrate Kit (Thermo, Fremont, CA) as previously described (19). Alexafluor secondary antibodies (1:200, Molecular Probes) were used for detection of smooth muscle α-actin (1:600, Thermo MS113P0, lot no. 113P1205E), SOD1 (1:500, Abcam ab13498), SOD2 (1:50, Abcam ab13533), nNOS (1:200, BD Biosciences 610310), phospho-nNOS (1:50, Abcam ab5583), Rac1 (1:50, Abcam ab33186), and phospho-Rac1 (1:100, Abcam ab203884). For each tissue section, the entire cross section was imaged. Data are expressed as the area fraction (percent of pixels with positive staining in the medial area) or as the total square area of medial staining (mm²).

Statistics.

Data were analyzed with GraphPad Prism version 5.01 (GraphPad Software, San Diego, CA) and expressed as the means ± SE. Unpaired t-test was used to analyze data comparing two groups, with P < 0.05 considered significant. One-way ANOVA with Tukey's post hoc test was used for comparisons of three groups, with a confidence limit of 95% considered significant. Repeated measures two-way ANOVA and Holm-Sidak's multiple comparisons test with α = 0.05 was used when comparing three groups +/− endothelium or when analyzing concentration response data. An additional comparison of the concentration response data from only the two HS groups was made by using the same parameters.

RESULTS

Neither the HS diet (administered from 5 to 15 wk of age) nor G-1 treatment (administered from 13 to 15 wk of age) altered body weight in the mRen2 female rat (Fig. 1A). As previously reported in the mRen2 female rat (8), HS significantly increased systolic blood pressure from 149 ± 5 to 224 ± 8 mmHg (Fig. 1B). Chronic treatment with the selective GPER agonist G-1 for 2 wk did not influence blood pressure (216 ± 4 mmHg). In aortic rings preconstricted with phenylephrine, the vasodilatory response to acetylcholine was not different between groups (Fig. 1C). In endothelium-intact vessels, the response to PE (Fig. 1D) and Ang II (Fig. 1E) was significantly higher in response to an HS diet. Endothelial denuding increased the response to both PE and Ang II only in NS vessels. Chronic treatment of HS animals with G-1 did not influence these vasoconstrictor responses in either endothelium-intact or denuded aortic rings. Ex vivo application of G-1 induced a significant vasodilatory effect in NS vessels that was similarly attenuated in denuded NS rings or intact aortic rings from HS+veh and HS+G-1 animals (Fig. 1F). Further comparison of intact and denuded HS+veh vs. HS+G-1 rings showed that in vivo G-1 treatment slightly but significantly enhanced the ex vivo vasodilatory response both in the presence and absence of the endothelium.

Fig. 1.

A: body weight was not significantly different between groups (P = 0.12). B: blood pressure as measured by tail cuff was significantly increased in high-salt (HS) vs. normal salt (NS) animals and was not altered by 2-wk treatment with G-1 (*P < 0.0001 vs. NS). C: intact aortic rings (solid bars) isolated from NS, HS+veh, and HS+G-1 animals and preconstricted with 10⁻⁶ mol/l phenylephrine (PE) showed no significant differences in the relaxation response to 10⁻⁶ mol/l acetylcholine (Ach; P = 0.54). Endothelium-denuded rings (hatched bars) show a lack of vasodilation to Ach. D: in endothelium-intact but not denuded vessels, the response to PE was greater in animals treated in vivo with HS+veh and HS+G-1 (*P < 0.01). Denuding significantly increased the contraction only in NS vessels (†P < 0.05). E: in both endothelium-intact and denuded vessels, contraction to Ang II was greater in HS+veh and HS+G-1 vs. NS vessels (*P < 0.0001). Denuding significantly increased the response only in NS vessels (†P < 0.0001). F: relaxation to increasing concentrations of ex vivo G-1 (10⁻⁹ to 10^−5.5 mol/l) was significantly greater in NS rings than HS+veh and HS+G-1 rings (*P < 0.0001, **P < 0.0001). G: additional analysis of the ex vivo G-1 response in intact and denuded HS+veh and HS+G-1 rings showed that chronic G-1 treatment enhanced the vasodilatory response in both the presence and absence of the endothelium (**P < 0.01).

Salt-sensitive hypertension stimulated aortic remodeling, measured as a significant increase in the media-to-lumen ratio from 0.43 ± 0.02 to 0.89 ± 0.09 (P < 0.001; Fig. 2, A–B). Analysis of wall thickness and lumen size showed that the increased ratio was due to a thicker wall (1.6 ± 0.07 to 2.7 ± 0.25 mm/kg, P < 0.01; Fig. 2C) but no change in lumen diameter (8.0 ± 0.34 to 7.2 ± 0.33 mm/kg, P = 0.19; Fig. 2D). HS also increased total medial cross-sectional area from 4.6 ± 0.28 to 7.3 ± 0.97 mm²/kg (P < 0.05; Fig. 2E). Chronic G-1 administration attenuated aortic remodeling, as evidenced by a decrease in the media-to-lumen ratio to 0.61 ± 0.02 (P < 0.05). The G-1 effect was attributed to a decreased wall thickness (2.0 ± 0.08 mm/kg; P < 0.05; Fig. 2C) with no significant change in lumen diameter (7.5 mm/kg; P = 0.81; Fig. 2D). Medial cross-sectional area was not different in HS+G-1 aortas compared with NS aortas (6.1 ± 0.33 mm²/kg; P = 0.30 vs. NS; Fig. 2E).

Fig. 2.

A: sample hematoxylin stained aortic cross sections from animals treated in vivo with NS, HS+veh, and HS+G-1. B: the media-to-lumen ratio was significantly increased by HS and ameliorated by G-1 (***P < 0.001, *P < 0.05, n = 4–5). C: HS significantly increased (**P < 0.01), whereas G-1 decreased (*P < 0.05) wall thickness normalized to body weight (BW). D: there were no changes in lumen diameter normalized to body weight (P = 0.21). E: medial cross-sectional area (CSA) normalized to body weight was significantly increased in the HS+veh group (*P < 0.05) but was not different between HS+G-1 and NS groups (P = 0.24).

Aortic sections were assessed for hyperplasia by using antibodies against PCNA and Ki-67. No evidence of immunostaining was found in the medial sections of aorta for these two nuclear proteins necessary for cellular proliferation (data not shown). To determine whether there was hypertrophy of smooth muscle cells in response to salt, we used immunofluorescence to detect expression of the contractile protein α-actin. Actin staining was not different between groups, when either analyzing data as the percent of stained pixels (area fraction; Fig. 3B) or when taking the medial cross-sectional size into consideration and expressing the total stained square area (Fig. 3C).

Fig. 3.

A: sample aortic cross sections stained with α-smooth muscle actin and fluorescent secondary antibody. Area fraction (B) and square area (C) for actin staining was not changed with salt or G-1 (P > 0.05, n = 4–5).

To determine whether the extracellular matrix was altered by HS and G-1, aortic sections were analyzed for collagen by picrosirius red staining by using quantitative polarized light microscopy. Staining was quantified according to the relative distribution of type I large diameter (red and orange) and type III small diameter (yellow and green) collagen fibers as previously described (4, 34). Collagen was not affected by salt or G-1 even when assessing individual fiber types in both the media (Fig. 4) and adventitia (data not shown). The area fraction of elastin staining was significantly decreased by HS, indicating decreased density of elastin fibers (Fig. 5, A--B), but the total amount of elastin per cross section was not altered, suggesting no degradation of elastin due to salt (Fig. 5C). Furthermore, G-1 altered neither of these elastin measurements. To determine whether changes in other extracellular matrix components were altered by salt or G-1 in the mRen2 model, we assessed glycosaminoglycans by alcian blue staining. HS significantly increased alcian blue staining in the medial layer of the aorta and was characterized by large pools of glycosaminoglycans between elastin fibers (Fig. 6, arrows). Treatment with the GPER agonist for 2 wk attenuated both the area fraction and square area of alcian blue staining, and glycosaminoglycan staining significantly and positively correlated with media-to-lumen ratio (Fig. 6D).

Fig. 4.

A: sample aortic cross sections stained with picrosirius red and imaged with polarized light microscopy. B: area fractions of individual fiber colors were not changed by salt or G-1.

Fig. 5.

A: sample aortic cross sections stained for elastin. B: the area fraction was significantly decreased by HS and not altered by G-1 (*P < 0.05 vs. NS). C: the total square area of elastin staining was not different between groups (P > 0.05, n = 4–5).

Fig. 6.

A: sample aortic cross sections stained with alcian blue for detection of proteoglycans. Arrows denote areas of pooling in HS+veh sections. The area fraction (B) and square area (C) of alcian blue staining was increased with HS (**P < 0.01) and attenuated by G-1 (*P < 0.05, n = 4–5). D: alcian blue staining significantly and positively correlated with media-to-lumen ratio.

We previously showed that urinary 8-isoprostane and proximal tubule oxidative stress was reduced in response to G-1 in this model (19). To determine whether changes in oxidative stress also contributed to vascular remodeling, we stained for the lipid peroxidation product 4-hydroxynonenal (4-HNE) in aortic sections. Compared with aortas from NS rats, HS significantly increased oxidative stress (Fig. 7, A–C), whereas chronic G-1 treatment attenuated the effect of HS. Furthermore, 4-HNE staining significantly correlated with media-to-lumen ratio (Fig. 7D) and alcian blue staining (Fig. 7E).

Fig. 7.

A: sample aortic cross sections immunostained for 4-HNE as a marker of oxidative stress. The area fraction (B) and square area (C) of 4-HNE staining was increased with HS (**P < 0.01, ***P < 0.001) and attenuated by G-1 (*P < 0.05, n = 4–5). D: 4-HNE staining significantly and positively correlated with media-to-lumen ratio. E: 4-HNE staining significantly and positively correlated with staining for glycosaminoglycans.

To identify a mechanism for G-1's antioxidant effects, we used immunofluorescence to stain for various players in the production and scavenging of reactive oxygen species. We first assessed expression of superoxide dismutase enzymes 1 and 2 (SOD1 and SOD2) but did not find a significant difference between groups (Fig. 8, A–B). Next, we concentrated on proteins that could be acutely activated in response to the GPER agonist and therefore fit the time course for nongenomic G protein-coupled signaling events. Our previous work in the mRen2 female rat demonstrate a salt-induced increase in neuronal nitric oxide synthase (nNOS) and oxidative stress that is ameliorated by nNOS inhibition (40). However, we saw no differences in aortic nNOS, phosphorylated nNOS, or the activation of nNOS (Fig. 8, C–E). Next, we assessed staining for Rac1, a small GTP-ase that is integral to the activation of NADPH oxidases. Both estradiol and a membrane-impermeable estrogen dendrimer conjugate acutely phosphorylate and deactivate Rac1 (41). However, we saw no changes in either total or phosphorylated Rac1 in response to HS or chronic G-1 treatment (Fig. 8, D–F).

Fig. 8.

Immunofluorescence of aortic sections was performed for the following proteins: superoxide dismutase (SOD) 1 (A); SOD2 (B); phosphorylated nNOS (pnNOS) (C); total nNOS (D); pnNOS/nNOS (E); phosphorylated Rac1 (F); total Rac1 (G); and pRac1/Rac1 (H). No significant differences were found between groups (P > 0.05).

DISCUSSION

The present study demonstrates that GPER induced beneficial vascular effects during salt-sensitive hypertension in the absence of changes in systolic blood pressure and vascular reactivity. The mechanism for the G-1 effect most likely involves a reduction in glycosaminoglycans in the medial layer and a decrease in oxidative stress. These results nicely complement our previous work showing beneficial effects of GPER activation in the kidney and heart (13, 19). GPER's renoprotective effects were similarly associated with a reduction in oxidative stress, while G-1's beneficial cardiac effects were also independent of collagen alterations. Taken together, our results show that in the mRen2 female rat, activation of the novel estrogen receptor GPER elicited protective effects throughout the cardiovascular system. Furthermore, because G-1 exerted beneficial actions in ovary-intact animals, this study indicates that continuous GPER activation provides added benefits during fluctuating levels of endogenous estrogen that occur with the normal estrous cycle.

The lack of blood pressure effects in the current study may be due to the presence of the endogenous ligand, as we previously showed that G-1 markedly reduces blood pressure in ovariectomized female mRen2 rats (18). The GPER agonist also did not lower pressure in intact female mRen2 rats on a normal salt diet, although the blood pressure in these animals is much lower (∼140 mmHg systolic) compared with the salt-loaded animals in the current study (>220 mmHg) (18). Similar to our work in mesenteric vessels, we found a significant attenuation of aortic relaxation to ex vivo G-1 in HS-fed rats (19). Chronic G-1 treatment enhanced the ex vivo response to the agonist in both endothelium-intact and denuded vessels, indicating a direct effect on vascular smooth muscle. The beneficial effect of G-1 treatment on GPER-mediated vasodilation, however, did not restore vasodilation to levels seen in NS vessels and therefore could underlie the lack of a pressure effect.

We found that vasoconstrictor responses to both PE and Ang II were higher in vessels from HS animals and were not affected by denuding, indicating salt- and pressure-induced endothelial dysfunction. A similar effect of HS on the vasoconstrictor response to PE is observed in aortic strips from pregnant Sprague-Dawley rats, where endothelial denuding significantly enhances the response in NS- but not HS-fed animals (4). HS induces oxidative stress which in turn impairs endothelial cell calcium signaling and attenuates nitric oxide production in aortic rings from male Sprague-Dawley rats (42). Interestingly, the enhanced vasoconstriction in the current study occurred in the absence of changes in the acetylcholine response despite the fact that a multiple of other studies demonstrate endothelial dysfunction in hypertensive vessels (35). The effect of an HS diet on endothelial function is mixed, with acetylcholine-induced aortic relaxation not different in Sprague-Dawley rats fed an HS diet (24) but reduced in Dahl salt-sensitive rats (20). Compensatory mechanisms may maintain the acetylcholine response in HS vessels even though endothelial dysfunction exists. For example, the response to acetylcholine is not reduced in mesenteric arteries from salt-loaded Sprague-Dawley rats unless vessels are pretreated with L-NAME, indicating a reduced contribution of nitric oxide but an enhanced role for endothelium-derived hyperpolarizing factor (32). Furthermore, the enhanced vasoconstrictor responses induced by HS were not improved by chronic G-1, suggesting that other estrogen receptors may have greater impact on large vessel reactivity in HS conditions. Nevertheless, the lack of a blood pressure effect in the current study allowed us to assess the beneficial effects of GPER activation that are pressure independent.

In the mRen2 female rat, an HS diet for 10 wk significantly increased the media-to-lumen ratio. This remodeling encompassed a thicker wall and cross-sectional area with no change in lumen diameter. One caveat of this study is that measurements were not obtained in perfusion-fixed vessels. Instead, the upper thoracic aorta was immersed in formalin and the lower portion used for vascular reactivity studies. Future studies will characterize the effects of HS and G-1 using a pressure myograph to allow structural measurements in fully relaxed vessels. However, Lee et al. (16a) showed that cross-sectional area of the medial layer is independent of contractile state and perfusion pressure, and this variable was increased by HS but was not different between NS and HS+G-1 groups. While HS induced a thicker wall coupled with a decreased lumen diameter, G-1 normalized the media-to-lumen ratio by decreasing wall thickness. This remodeling occurred in the absence of changes in systolic blood pressure or vascular reactivity, suggesting direct effects on aortic structure. Evidence from animal models and clinical studies indicates that HS-induced vascular damage occurs independently of hypertension, and drug treatments induce disparate effects on pressure and stiffness (30). Moreover, in hypertensive postmenopausal women taking hydrochlorothiazide, estrogen plus progestin therapy significantly attenuates pulse wave velocity without lowering mean arterial pressure (26). These results in concert with the current study indicate that estrogen is protective in female vasculature exposed to high salt or pressure.

Proteoglycans are complex macromolecules composed of glycosaminoglycans covalently bound to proteins. These hydrophilic molecules interact with collagen and elastin fibers to influence large artery stiffness and vascular mechanohomeostasis (29). A change in glycosaminoglycan content precedes and promotes a variety of vascular pathologies, including hypertension (38). In fact, adult spontaneously hypertensive rats have 30% more aortic glycosaminoglycans compared with the normotensive control Wister Kyoto rat (28). We found a significant increase in glycosaminoglycans in HS aortas vs. NS controls that was attenuated by chronic treatment with G-1. This increase in glycosaminoglycans is most likely the underlying cause for increased wall thickness in the absence of other extracellular matrix alterations and may increase stiffness and compromise mechanosensing in the aorta (29). This study is the first to show that a GPER-induced decrease in glycosaminoglycans may offer vascular protection from salt and pressure. Estradiol also attenuates vascular damage in a mouse carotid injury model by reducing proteoglycan deposition and reduces the amount of hyaluronic acid (one type of glycosaminoglycan) in the aorta of ovariectomized rabbits (2, 25). The mechanism for these estrogenic effects may be transcriptional, as estrogen reduces gene expression for proteoglycan 4 as well as the enzyme that synthesizes hyaluronic acid (10, 21). Since other studies have implicated ERα, our results showing that GPER activation significantly reduces glycosaminoglycans indicate either functional redundancy in the receptor which mediates this estrogenic effect or an alternative mechanism (10, 25).

One such mechanism may include a reduction in oxidative stress, as proteoglycan synthesis is inhibited by antioxidants in vascular smooth muscle cells (6). Oxidative stress contributes to vascular remodeling, and estrogen protects against oxidative damage (39). Staining for 4-HNE, a product of lipid peroxidation and a marker for oxidative stress, was significantly increased in the medial layer of the aorta in response to HS and was attenuated by treatment with G-1 for 2 wk. Estrogen modulates oxidative stress to attenuate vascular remodeling in injured arteries, and either estradiol or selective activation of ERα or ERβ attenuates aldosterone-induced vascular remodeling and oxidative stress (3, 23, 39). While we did not find significant differences in the expression of SOD1, SOD2, nNOS, or Rac1, our current data indicate that GPER induces antioxidant effects, and elucidation of the underlying mechanisms awaits future studies.

In the mRen2 model of salt-sensitive hypertension, activation of the novel estrogen receptor GPER attenuated aortic remodeling, glycosaminoglycans, and oxidative stress. Although not assessed in this study, we predict that these beneficial effects increased compliance of the aorta, which would reduce cardiac afterload and potentially contribute to the protective effects previously reported in the heart (13). Likewise, if GPER activation similarly reduced remodeling of small arterioles, this could be one mechanism for the renoprotective effects previously reported in the kidney (19). Clinically, menopause is associated with an increase in carotid artery thickness and aortic stiffness while hormone therapy is associated with healthier vasculature (11). These clinical data in addition to our current results suggest that estrogen has beneficial effects to protect end organs from hypertensive damage that occur even when blood pressure is not lowered. Future studies will determine whether targeting GPER decreases aortic pulse wave velocity and whether these beneficial effects are present in other models of hypertension and cardiovascular diseases.

GRANTS

This work was supported by National Heart, Lung, and Blood Institute Grants HL-103974, HL-51952, HL-56973, and HL-116769 and American Heart Association Grants 10BGIA3080005 and 14GRNT20480131.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the author(s).

AUTHOR CONTRIBUTIONS

L.L., S.K., B.M., D.D.H., R.A.B., E.H.T., M.A.Z., and S.H.L. performed experiments; L.L., S.K., B.M., D.D.H., K.S.M., and S.H.L. analyzed data; L.L. and S.H.L. drafted manuscript; D.D.H. and S.H.L. prepared figures; A.J.T., K.S.M., M.C.C., and S.H.L. conception and design of research; A.J.T., K.S.M., M.C.C., and S.H.L. edited and revised manuscript; S.H.L. interpreted results of experiments; S.H.L. approved final version of manuscript.