Vasodilation by GPER in mesenteric arteries involves both endothelial nitric oxide and smooth muscle cAMP signaling

Abstract

Our previous work showed that chronic activation of the membrane-bound estrogen receptor GPR30/GPER significantly lowers blood pressure in ovariectomized hypertensive mRen2.Lewis female rats which may, in part, reflect direct vasodilatory actions. The current study assessed the hypothesis that cyclic adenosine monophosphate (cAMP) signaling contributes to GPER-mediated vasorelaxation. In mesenteric resistance arteries from intact Lewis females, relaxation to 17-β-estradiol (E2; 47±3% of phenylephrine contraction vs. vehicle 89±2%, P<0.001) or G-1 (44±8%, P<0.001) was blunted to a similar extent by denuding (P<0.001) or the nitric oxide synthase inhibitor L-NAME (P<0.001). In contrast, the cyclooxygenase inhibitor indomethacin did not alter vasodilation (P>0.05). The cAMP analogue Rp-cAMPS partially attenuated vasodilation (65±7%, P<0.001), while the combination of L-NAME and Rp-cAMPS exhibited additive effects to effectively abolish vasorelaxation (P>0.05 vs. vehicle). Pretreatment of endothelium-intact vessels with the adenylyl cyclase inhibitor SQ (63±6%) or the guanylyl cyclase inhibitor ODQ (62±9%) both partially inhibited the response to G-1 (P<0.01), while pretreatment with the both inhibitors completely abolished vasorelaxation (P>0.05 vs. vehicle). In denuded vessels only SQ reduced the response (88±3%, P<0.001). Moreover, G-1 significantly increased intracellular cAMP levels in cultured mesenteric smooth muscle cells (P<0.05). We conclude that GPER-dependent vasorelaxation apparently involves both endothelial release of nitric oxide which activates guanylyl cyclase and smooth muscle cell activation of adenylyl cyclase. Downstream production of cyclic nucleotides and stimulation of protein kinases may phosphorylate proteins to promote vascular smooth muscle cell relaxation. The ability of GPER to initiate these signaling pathways may contribute to the beneficial vascular effects of estrogen.

Introduction

In ovariectomized hypertensive mRen2.Lewis rats, chronic administration of the selective GPER agonist G-1 significantly reduces blood pressure [1]. While this antihypertensive response may be partially due to local alterations in the renin-angiotensin system, GPER activation may also have direct effects on vascular tone. Indeed, we previously showed that acute administration of the GPER agonist G-1 induces vasodilation of mesenteric resistance arteries that is not different from the nonselective agonist 17-β-estradiol in hypertensive mRen2.Lewis females and normotensive Lewis controls [2]. The vasorelaxation induced by G-1 and 17-β-estradiol is partially dependent on endothelial nitric oxide production but also comprises an endothelium-independent component. Moreover, we demonstrated GPER immunoreactive staining on both intimal and medial aspects of the mesenteric vasculature.

In the initial characterization of its expression and function in breast cancer cell lines, GPER was found to activate acute signaling pathways including cAMP accumulation [3]. Despite the numerous studies verifying GPER expression in vascular smooth muscle [1, 4–8] and demonstrating endothelium-independent relaxation to G-1 [2, 7, 9], no studies as of yet have identified the intracellular signaling pathways that mediate GPER’s vasodilatory effects in smooth muscle. Therefore, the current study assessed the hypothesis that stimulation of cyclic AMP is an integral pathway in the vasorelaxant actions of GPER in vascular smooth muscle cells.

Experimental Methods

Animals

Lewis female rats (Harlan, Indianapolis, IN) were used for both myography and cell isolation at 15 weeks of age. Adequate measures were taken to minimize pain or discomfort, and all experiments were approved by the Tulane University Animal Care and Use Committee.

Myography

Second-order mesenteric vessels (~200 μm I.D.) were isolated and sectioned into 2 mm rings. Vessels were mounted in a DMT wire myograph containing Krebs-Henseleit solution (in mM): NaCl (118), KCl (4.8), CaCl₂ (2.5), MgSO₄ (1.2), NaHCO₃ (25), KH₂PO₄ (1.2) and glucose (11) at 37°C and oxygenated with 95% O2 - 5% CO2. Internal circumference was normalized to 0.9·IC100, where IC100 is the internal circumference at a transmural pressure of 100 mmHg. Viability and endothelial integrity were tested with phenylephrine (PE, 10⁻⁵ M; MP Biomedicals, Solon, OH) followed by acetylcholine (Ach, 10⁻⁶ M; Enzo, Farmingdale, NY) and rings with > 50% relaxation to Ach were considered endothelium intact. Rings were washed and again constricted with PE before cumulative additions of vehicle (DMSO), 17-β-estradiol (E2; EMD Millipore, Billerica, MA), or G-1 (EMD Millipore). The following inhibitors were added to the bath 10 minutes prior to PE preconstriction for the concentration response curve: N^G-Nitro-L-arginine-methyl ester (L-NAME, 10⁻⁴ M; Enzo), indomethacin (10⁻⁵ M; Sigma, St. Louis, MO), Rp-Diastereomer of adenosine-3′,5′-cyclic monophosphothioate (Rp-cAMPS, 10⁻⁵ M; Santa Cruz Biotechnology, Santa Cruz, CA), SQ 22536 (SQ, 10⁻⁴ M; Enzo), and ODQ (10⁻⁴ M; Enzo). All data are expressed as percent of the PE preconstriction, and two-way ANOVA was used to compare treatment groups and agonist concentrations.

cAMP measurements

Smooth muscle cells were isolated from the mesenteric arterial branch as previously described [10]. Cells were passaged a maximum of four times, as previous studies show that GPER expression is only stable for this time in culture [7]. Cells were cultured in Medium 199 containing 10% fetal bovine serum and penicillin/streptomycin, then switched to phenol red-free medium containing 5% charcoal-stripped serum two days before experiments. Cells were pretreated with the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (IBMX, 10⁻⁹ M; Sigma) for 20 minutes before the addition of vehicle (DMSO) or G-1. In some experiments, cells were also pretreated with the GPER antagonist G15 (10⁻⁸ M; EMD Millipore). Cells were incubated at 37 °C for 15 minutes, and cAMP was measured by EIA kit (Cayman Chemical, Ann Arbor, MI). One-way ANOVA was used to determine differences between groups.

Results

In isolated mesenteric resistance arteries from 15 week-old Lewis females, vasorelaxation to 17-β-estradiol (E2) and the GPER agonist G-1 was dose-dependent and became significant at 10⁻⁶ M. The maximum dilation reached was 47 ± 3% for E2 and 44 ± 8% for G-1 (P<0.001 vs. vehicle, Figure 1A–B). The cyclooxygenase inhibitor indomethacin had no effect on vasodilation to either E2 or G-1 (P>0.05 vs. agonist). Denuding or pretreatment with the nitric oxide synthase inhibitor L-NAME (10⁻⁴ M) similarly attenuated vasodilation to E2 and G-1 (P<0.001) as previously reported in mRen2.Lewis female vessels [7].

Figure 1.

(A) Estradiol (E2) and (B) G-1 significantly relaxed mesenteric arteries (*P<0.01). While the response to E2 was not altered in the presence of indomethacin (10⁻⁵ M, P>0.05 vs E2 or G-1), both L-NAME (10⁻⁴ μM) and endothelial denuding inhibited vasorelaxation to a similar extent (*P<0.001).

We next determined the signaling mechanisms involved in the endothelium-independent portion of the vasodilatory response. To ascertain whether the signaling molecule cyclic adenosine monophosphate (cAMP) was involved in the smooth muscle cell mechanism, endothelium-intact vessels were pretreated with the cell permeable cAMP analogue Rp-cAMPS, and the response to G-1 was partially inhibited (P<0.001; Figure 2). Pretreatment of intact vessels with a combination of L-NAME, to inhibit the endothelial component, and Rp-cAMPS completely abolished the response (P>0.05 vs. vehicle).

Figure 2.

The cAMP inhibitor Rp-cAMPS (10⁻⁵ M) partially reduced the response to G-1 (P<0.001), while pretreatment with Rp-cAMPS and L-NAME abolished the response (P>0.05 versus vehicle).

Nitric oxide activates guanylyl cyclase in vascular smooth muscle to increase the production of cyclic GMP, while adenylyl cyclase is responsible for the production of cAMP. To ascertain the role of these two enzymes in the GPER vasodilatory response, we utilized the soluble guanylyl cyclase inhibitor ODQ and the adenylyl cyclase inhibitor SQ. As shown in Figure 3A, both ODQ and SQ partially attenuated GPER-mediated vasodilation in endothelium-intact vessels (P<0.01), and the combination of the two inhibitors completely blocked relaxation (P>0.05 versus vehicle). Additional studies showed that ODQ was ineffective in the absence of the endothelium, while the adenylyl cyclase inhibitor SQ completely abolished relaxation in denuded vessels (P>0.05 versus vehicle; Figure 3B).

Figure 3.

(A) The adenylyl cyclase inhibitor SQ (10⁻⁴ M) and the guanylyl cyclase inhibitor ODQ (10⁻⁴ M) both partially inhibited the response to G-1 in intact vessels (P<0.01). The combination of the two inhibitors abolished the response (P>0.05 versus vehicle). (B) In denuded vessels, ODQ did not alter the response to G-1 (P>0.05 vs. G-1) while SQ abolished endothelium-independent vasorelaxation (P>0.05 versus vehicle).

In order to confirm that GPER activation by G-1 increases cAMP in vascular smooth muscle, we isolated cells from the same arteries used in the myograph experiments. We previously showed that GPER expression in these cells that was stable in culture through passage four [7]. As shown in Figure 4, G-1 elicited a significant increase in cAMP production in mesenteric artery smooth muscle cells that was significant at nanomolar concentrations. Pretreatment with the GPER antagonist G15 abolished the cAMP response to G-1.

Figure 4.

In cultured cells, 15 minute treatment with G-1 (10⁻⁹ M to 10⁻⁵ M) increased the production of cAMP in comparison to DMSO (*P<0.05, n=4–13). The GPER antagonist G15 (10⁻⁸ M) abolished the response to 10⁻⁹ M G-1 (P>0.05 vs. DMSO, n=5).

Discussion

The present study demonstrates that GPER-dependent relaxation of female Lewis mesenteric arteries involves both endothelial and smooth muscle cell signaling pathways. In endothelial cells, GPER stimulates the release of nitric oxide which diffuses to smooth muscle and activates guanylyl cyclase. In smooth muscle cells, GPER activates adenylyl cyclase resulting in an increase in cAMP. The ability of G-1 to increase cAMP was confirmed in cultured vascular smooth muscle cells isolated from the mesenteric vessels. Because the G protein subunit Gαs is known to activate adenylyl cyclase and stimulate the production of cAMP, we propose that in vascular smooth muscle cells GPER is coupled to this alpha subunit and upon ligand binding utilizes this signaling mechanism to induce vasodilation (Figure 5).

Figure 5.

GPER activation causes increases in nitric oxide (NO) through an endothelium-dependent mechanism. In vascular smooth muscle cells, GPER may couple to the αs subunit to activate adenylyl cyclase (AC), increase cAMP, and trigger protein kinases (PKA, PKG) to phosphorylate proteins involved in vascular smooth muscle cell contractility. IP₃R: inositol trisphosphate receptor, SERCA: sarcoplasmic reticulum Ca²⁺-ATPase, PLB: phospholamban.

Our results showing a contribution of both cell types to vasorelaxation are similar to those recently reported by Li and colleagues in the thoracic aorta of ovariectomized Sprague-Dawley rats [11]. However, others have shown that the GPER response is completely endothelium-dependent in epicardial porcine coronary arteries [12] and Sprague-Dawley carotid arteries [6]. Still others have shown a response that is completely endothelium-independent in female Sprague-Dawley inferior vena cava [13] and porcine coronary arteries [9]. The reason for these disparate results is unclear and may suggest differing levels of GPER expression and/or signaling in endothelial versus smooth muscle cells depending on the model and arterial bed investigated. For example, Reslan et al. showed greater vasodilation to G-1 in mesenteric and renal arteries versus carotid and pulmonary arteries [14]. Furthermore, we and others have demonstrated sex and age differences in the contribution of endothelium versus smooth muscle in GPER-mediated vasodilation [2, 15] and have also shown that disease conditions such as salt-loading influence this response [16].

The cAMP analogue Rp-cAMPs and the adenylyl cyclase inhibitor SQ partially inhibited the vasodilatory response in intact vessels. However, in denuded or L-NAME pretreated vessels these inhibitors were completely abolished vasorelaxation. Therefore, the signaling mechanism in endothelial cells is distinct from the activation of adenylyl cyclase. In endothelial cells, GPER may couple to a different alpha subunit or utilize a signaling pathway which does not include adenylyl cyclase. Additional studies using isolated endothelial cells may elucidate the mechanisms by which GPER influences nitric oxide synthase and release of nitric oxide.

In 1967, Szego and Davis showed that 17-β-estradiol stimulates uterine cAMP levels in less than one minute [17]. This acute cAMP response results from stimulation of adenylyl cyclase and is not dependent on RNA or protein synthesis, suggesting a non-genomic mechanism [18]. The estrogen-induced increase in cAMP is one possible mechanism for the acute vasodilatory effects of the hormone. In rat aorta, Rp-cAMPs inhibits the vasodilatory response to estrogen and diethylstilbestrol [19, 20]. In addition, cAMP and adenylyl cyclase are implicated in estradiol-induced relaxation of porcine coronary arteries [21]. We show here that the vasodilatory response to GPER activation also involves the adenylyl cyclase/cAMP signaling pathway in vascular smooth muscle.

Increases in cAMP can activate protein kinase A and can also cross-activate other protein kinases such as protein kinase G [22]. Subsequently, these protein kinases phosphorylate effector proteins that decrease the contractile state of smooth muscle cell. Possible targets include plasma membrane ion channels such as L-type calcium channels [23] and large-conductance potassium channels [24]. Alterations in sarcoplasmic reticulum proteins, such as the inositol trisphosphate receptor [25] and phospholamban [26], may increase intracellular calcium uptake. Alternative targets may include proteins which directly regulate the contractile apparatus, including myosin light chain kinase [27] and myosin phosphatase [28]. While the ability of GPER to alter many of these targets remains to be studied, Yu and colleagues recently found that G-1 alters large-conductance potassium channel activity in porcine coronary arteries [9]. In addition, Gros et al. showed that GPER increases myosin light chain phosphorylation in rat aortic smooth muscle cells [8]. Studies are in progress to identify the protein kinases and phosphorylation targets of GPER in mesenteric artery smooth muscle cells of the Lewis female rat.

Highlights.

• Activation of the estrogen receptor GPER relaxes Lewis female mesenteric arteries.

• Vasorelaxation is partially dependent on endothelial production of nitric oxide.

• In vascular smooth muscle, vasorelaxation is dependent on cAMP production.

• In isolated vascular smooth muscle cells, GPER activation increases cAMP.

• These signaling pathways may promote vasorelaxation and oppose hypertension.

Abbreviations

GPER

G protein-coupled estrogen receptor

cAMP

cyclic adenosine monophosphate

L-NAME

Nitro-L-arginine-methyl ester

PE

phenylephrine

Ach

acetylcholine

E2

estrogen

G-1

GPER agonist