|Home | About | Journals | Submit | Contact Us | Français|
Estrogen has important effects on cardiovascular function including regulation of vascular function, blood pressure, endothelial relaxation, the development of hypertrophy and cardioprotection. However, the mechanisms by which estrogen mediates these effects are still poorly understood. As detailed in this review, estrogen can regulate transcription by binding to two nuclear receptors, ERα and ERβ, which differentially regulate gene transcription. ERα and ERβ regulation of gene transcription is further modulated by tissue specific co-activators and co-repressors. Estrogen can bind to ERα and ERβ localized at the plasma membrane as well as GPER to initiate membrane delimited signaling, which enhances kinase signaling pathways that can have acute and long term effects. The kinase signaling pathways can also mediate transcriptional changes, and can synergize with the estrogen receptor to regulate cell function. This review will summarize the beneficial effects of estrogen in protecting the cardiovascular system through ER-dependent mechanisms with an emphasis on the role of the recently described ER-membrane signaling mechanisms.
Sex differences in cardiovascular disease are well documented. Premenopausal women have reduced cardiovascular disease and the occurrence increases after menopause1, 2. Furthermore, many studies in animal models have reported less hypertrophy and less ischemia-reperfusion (I/R) injury in females compared to males3, 4. Additional studies with ovariectomized females have also shown that addition of estrogen reduces hypertrophy and I/R injury5 (Figure 1 summarize the cardiovascular effects associated with estrogen). Large randomized clinical trials did not show a benefit with hormone replacement therapy (HRT)6, 7. However, a recently update of the woman health initiative (WHI) examined different age groups and concluded that HRT “has a harmful effect on coronary heart disease among older women, whereas the results in younger women remain inconclusive”8. The data in this study (Manson et al.8, Appendix eFigure 3) report that in early menopausal women (age 50–59), conjugated equine estrogen (CEE) treatment alone was protective for myocardial infarction (HR 0.55, 95% CI 0.31–1.00), coronary artery bypass grafting/percutaneous coronary intervention (CABG/PCI) (HR 0.56, 95% CI 0.35–0.88), and all cardiovascular events (HR 0.84, 95% CI 0.66–0.10.6). There was a detrimental effect on deep vein thrombosis (HR 1.66, 95% CI 0.75–3.67). Interestingly the slight beneficial effects on myocardial infarction and CABG/PCI were reversed in older menopausal women, consistent with an age dependence. Although these data do not support the use of longer term HRT for reducing coronary heart disease, the data are consistent with the hypothesis that the beneficial effects of estrogen are reduced and ultimately reversed with age. The mechanisms responsible for the loss of protection with age are unclear, but it demonstrates that we do not fully understand the protective mechanisms in females and the role of estrogen. Although it is recognized that sex differences are due to many factors independent of estrogen, it is clear that estrogen has major effects in mediating sex differences in the cardiovascular system. This review will focus on estrogen signaling mechanisms in the cardiovascular system. In addition to the well-established action of the nuclear estrogen receptor-estrogen complex to modulate gene transcription, estrogen can also bind to several different estrogen receptors (ERs) located at the plasma membrane and activate membrane delimited signaling. The cross-talk between these signaling mechanisms leads to complex downstream signaling. In this review we will focus on estrogen signaling with an emphasis on the role of the more recently described membrane signaling mechanisms.
ER signaling has increased in complexity over the last 2 decades. Prior to 1996, only one ER was identified, which is now referred to as ERα, but at the time was simply called the ER. All the effects of estrogen were attributed to its binding to this one receptor which acted as a ligand regulated transcription factor. The complexity of estrogen signaling has grown over the last 20 years. There are now 3 ERs (ERα, ERβ and an orphan G-protein coupled receptor, GPR30 now known as GPER) as well as several splice variants, and ER signaling occurs via the traditional regulation of transcription as well as by activating membrane signaling cascades.
The need for a better understanding of the mechanisms of estrogen signaling is clearly emphasized by the recent findings suggesting that there are sex differences in response to cardiovascular drug treatment and the suggestion that the failure of some small clinical trials could be related, at least in part, to sex differences in response to the drug9, 10. In fact, the NIH has recently called for the inclusion of both sexes in preclinical trials11. Without a better understanding of the complexity of estrogen signaling in the heart, it will be difficult to unravel the mechanisms by which estrogen might be involved in protection in pre-menopausal females and why it fails to protect in late post-menopausal women. This lack of understanding can also complicate the design and interpretation of clinical trials.
The physiological effects of estrogen are mediated by estrogen binding to ERs. With classical ER signaling, estrogen binding to ERα or ERβ initiates translocation to the nucleus where the ER-estrogen complex acts as a ligand gated transcription factor initiating changes in gene expression. However estrogen can also bind receptors localized at the plasma membrane, leading to activation of acute signaling pathways often referred to as “non-genomic” or membrane delimited signaling. However, the term non-genomic is somewhat misleading because activation of acute signaling pathways at the membrane can also indirectly lead to changes in gene expression. Recent studies have demonstrated that post-translational modifications of ERα and ERβ can influence their cell localization and activity12–14. In addition to ERα and ERβ, GPER has been shown to localize to membranes and activate acute signaling pathways following estrogen binding.
ERα and ERβ are localized in both the cytosol and the nucleus and upon binding estrogen they can translocate to the nucleus, bind to DNA and modulate gene expression. ERα and ERβ are highly homologous; they share about 95% homology in their DNA binding domains, but their ligand-binding domains share only 56% of amino acid sequence identity15. Estrogen-activated ERs alter gene expression by (i) directly binding to DNA, (ii) indirectly binding DNA through other transcription factors or (iii) ligand-independent binding (Fig 2). Estrogen binding to ERα promotes receptor dimerization and facilitates receptor binding to the consensus estrogen response element (ERE) in the DNA sequence, thus mediating its genomic effects. Co-activators and co-repressors are recruited to activate or inhibit gene expression16 (Fig 2, Panel A). Estrogen can also alter gene expression via non-ERE regulatory mechanisms. ERs can tether to activator protein-1 (AP1) and specificity protein 1 (Sp1) regulating gene expressions. Burns et al17 showed that mutation of a region of ER inhibits binding between ERs and transcription factors and blocks this signaling. Also consistent with the concept that ER binding via transcription factors is independent of EREs, Jakacka and colleagues reported that an ERE is not necessary for activation and inactivation of API-dependent transcription18 (Fig 2, Panel B). Interestingly, activation of kinase signaling pathways by growth factors or estrogen activation of membrane ERs can lead to phosphorylation of ER and trigger its binding to DNA, thereby modulating of transcription. A specific serine phosphorylation site on ER has been shown to be important for ligand-independent activation of transcription19 or for binding to ERE (Fig 2, Panel C and D). This mechanism was elucidated when it was observed that growth factor activation of uterine proliferation was lost in mice that lacked ERα20. Prior to these studies it was not appreciated that ER was required for growth factor dependent proliferation of the uterus20. These results led to the concept that phosphorylation of ER can lead to its ligand independent effects on transcription.
Estrogen has been shown to regulate the expression of a large number of cardiac genes21–34. Overall, different genes are found to be regulated in different studies and it is difficult to draw general conclusions. Gene regulation depends on the context of the cell, thus the estrogen dependent changes in transcription might vary with age, sex, disease and tissue. To address some of these issues, Devanathan et al developed a mouse model of cardiomyocyte-specific deletion of ERα to characterize the role of ERα in the heart independent of systemic effects. Microarray data revealed differential variations in the expression of 208 genes compared to wild-type heart35. A genome-wide expression profiling of estrogen-treated cardiomyocyte from men and women showed that 36 estrogen-dependent genes are regulated in a sex-specific manner confirming that estrogen effects are tissue- and sex-specific36.
To further complicate the analysis of ER regulated genes, ERα and ERβ differentially regulate gene expression, and there can be tissue and temporal variations in the protein levels of ERα and ERβ37, 38. Also, expression of co-activators and co-repressors can differ37, 39–41 and the recruitment of co-regulators depends in part on the ligand bound. For example, tamoxifen recruits co-activators in endometrium, but it is an antagonist in breast recruiting co-repressors42. See Hall et al for a more detailed review on genomic estrogen receptor signaling16 and Rosenfeld et al for a discussion of the roles of coactivators and corepressors in ER signaling43. In mouse aorta, O’Lone et al reported that ERα and ERβ regulate a distinct set of genes40. They reported that estrogen activation of ERα primarily results in up-regulated gene expression, whereas ERβ primarily enhances down-regulation of gene expression. In contrast to the study of O’Lone, a gene array study in ovariectomized female mice showed that perfusion of the heart for 2h with an ERβ selective agonist primarily led to up-regulation of gene expression with only a few genes down-regulated41. These conflicting results might be due to differences in aorta versus ventricular myocytes (e.g. differences in levels of co-repressors and co-activators) or because of different durations of estrogen treatment; as mentioned estrogen can act in a time- and tissue-dependent manner44. For example, in endothelial cells ERα upregulates endothelial nitric oxide synthase (eNOS), while in cardiac muscle ERβ mediates up-regulation of eNOS45, 46. Also, studies in vascular smooth muscle cells showed that inducible NOS (iNOS) expression is enhanced by ERβ and repressed by ERα47. Furthermore, ERβ blocks iNOS gene up-regulation in endothelial cells48. ERα and ERβ can also synergize to regulate protein and enzyme activity. For example, in vascular endothelial cells, ERα regulates activation of superoxide dismutase 2 (SOD2), whereas ERβ regulates SOD2 basal expression. SOD2 up-regulation results in a decrease in reactive oxygen species (ROS) generation with lower mitochondrial dysfunction and vascular damage49. Table 1 summarizes the known functions associated with ERα and ERβ gene regulations.
Overall, estrogen regulation of gene expression is complex; ERα up-regulates some genes, while others are unaffected or down-regulated by ERβ and vice versa. Also, the relative levels of ERα and ERβ and the co-repressors and co-activators can change with sex, age, disease, and tissue thereby altering the response to estrogen. Additional studies are needed to better characterize the factors that contribute to tissue specific ER regulation of gene expression in the cardiovascular system. Furthermore, although there are a number of studies examining ER regulation of gene expression, very few studies have examined sex differences or estrogen mediated differences in protein expression50–52. There are a number of studies showing that changes in gene expression do not always translate to changes in protein expression, and therefore it is important to characterize the effects of sex and estrogen at the protein level. In addition to regulating mRNA transcription, estrogen is likely to regulate RNA processing, splicing, mRNA and protein stability, and protein post-translational modifications53–57. We have just begun to understand these sex differences in protein levels and additional studies are needed.
A rapid response to estrogen was first reported in the 1960s. Szego et al showed that estrogen increased the cAMP concentration within minutes, paving the way for recent studies on the non-genomic response of estrogen58. It is established that physiological concentrations of estrogen can have rapid effects, which occur independent of protein synthesis or gene activation, and it is generally accepted that these acute effects of estrogen are transmitted by signaling pathways through activation of ERs localized at the plasma membrane59, 60. In addition to the plasma membrane, ERs have also been localized to organelles including mitochondria61 and endoplasmic reticulum30, 62, 63; however the effects of ERs in mitochondria and endoplasmic reticulum are poorly understood and will not be discussed here.
Investigations of non-genomic estrogen action and its contribution to cellular function have been facilitated by the development of an estrogen conjugated to bovine serum albumin (BSA). However, the cross-link between estrogen and BSA was through a key binding site for ERs making the experiments problematic as the conjugation inhibited the physiological effects of estrogen binding64–66. Recently, Katzenellenbogen and coworkers developed an estrogen-dendrimer conjugate (EDC), an estrogen-macromolecule conjugate in which estrogen is attached to a large and positively charged polyamidoamine dendrimer67. EDC provides a new tool to investigate the non-nuclear signaling mechanisms of estrogen.
It is estimated that approximately 5–10% of total ERs are localized to the plasma membrane68. The relative amount of ERα and ERβ at the plasma membrane can differ among cell types; for example in breast cancer cells ERα is found to be less than ERβ69, meanwhile in reproductive tissue ERα is more abundant than ERβ70. ERs distribution is poorly studied in cardiovascular tissue, however both the ERα and ERβ are found to be present at the plasma membrane of vascular endothelial cells (EC)71. ERα and ERβ are tethered to the plasma membrane by a mechanism involving post translational modifications such as palmitoylation, which has been shown to increase association of the receptor with the plasma membrane30, 62.
Several laboratories have conducted investigations to confirm the localization of ERα to the plasma membrane. Using three different ERα antibodies, Chambliss and colleagues found ERα in EC plasma membranes, and they further demonstrated the presence of the ERα protein in a plasma membrane fraction isolated from the cell. They were also able to detect ERα in the caveolae using subfractions of caveolae and non caveolae62. The same group later reported that ERα and eNOS co-localized in the caveolae. This group also found that ERβ is present in caveolae from EC and has a non-genomic signaling response, through eNOS stimulation72. Chambliss et al. further confirmed that the same isoform of ERα (or ERβ) was present in purified plasma membranes and in the nucleus30. Finally, Pedram et al confirmed in human breast cancer cells, using a mass spectrometry approach, that membrane and nuclear ERs are the same protein73.
Several sites have been identified on ERα, which are involved in its localization to the plasma membrane. The interaction between ERα and caveolin-1 requires Ser-522, located in the ligand-binding domain, which has been reported to facilitate the binding of ERα to caveolae12, 62. Also, palmitoylation of Cys447 on the ligand-binding domain of ERα has been demonstrated to be crucial for ERα localization to the membrane12, 74. Finally immunoprecipitation studies demonstrated that amino acids 251–260 and 271–595 of ERα bind directly to Gαi and Gβγ75. These modifications and their role in specific cell types will be discussed later.
The full-length ERα seems to be the predominant form associated with the plasma membrane and has been shown to initiate rapid actions of estrogen. However, an N-terminus-deleted splice variant, ERα 46, has been shown to co-localize with caveolin-1 in caveolae and it seems to have a reduced transcriptional activity compared to the full-length ERα76. ERα 46 has been found co-immunoprecipitated with eNOS77 leading to eNOS activation and nitric oxide release78.
As mentioned above, Ser522 is important for physical association of ERα with caveolin-1, and subsequent translocation to the plasma membrane12. Indeed, one study showed that ERα localization at the plasma membrane is lost in caveolin-1 null cells, and ERα is only found in the nucleus12. However, another study reported that the mutation of Ser522 in Chinese hamster ovary cells is not sufficient to block the ERα translocation to the plasma membrane, suggesting that Ser522 is not the only residue responsible for membrane association12. Acconcia et al found that Cys447 of human ERα is a site for palmitoylation and this modification promotes the association of ERα with caveolin-113, 74. Mutation of this site blocked ERα translocation to the plasma membrane in cancer cell lines53. In 2007, Levin’s group identified a highly conserved amino acid motif in the estrogen binding domain of both human and mouse ERα and ERβ that mediates membrane translocation via palmitoylation, and Cys447 is part of this motif14. Therefore, palmitoylation of Cys447 is essential for caveolin-1 interaction and membrane translocation. However, additional sequences flanking this cysteine residue are needed for optimal palmitoylation. Mutation of key residues of ER for palmitoylation, such as phenylalanine or tyrosine at position −2, significantly reduces membrane localization and non-genomic actions of ERα14. Palmitoylation occurs only on ER monomers, therefore enhancing ER dimerization, such as with estrogen binding, limits the available receptors that can undergo palmitoylation and translocate to the membrane. Since the percentage ERs localized at the plasma membranes are only 5–10% of the total amount present in the cell, it is important to define the proteins and mechanisms that regulate the ERs trafficking to plasma membrane. Heat shock protein 27 (HSP27) was identified by Razandi et al as a promoter of ER palmitoylation79. Using a proteomic approach, they found that HSP27 binds to the palmitoylation site of ERα and promotes palmitoylation of ERα monomer. Another group found that ERα can localize to caveolae indirectly via the scaffold protein striatin, forming complexes with Gαi and Gα proteins80.
ERα was first shown to localize to the plasma membrane and to activate phosphatidylinositol-3-OH kinase (PI3K) signaling81. ERα and ERβ have been shown to play important roles in modulating cardiovascular disease. As discussed above, ERs localized at the plasma membrane initiate rapid signal transduction through kinase cascades involving PI3K82, 83, mitogen-activated protein kinase (MAPK)84, 85, and endothelial nitric oxide synthase (eNOS) activation86, 87. Rapid signaling by estrogen action was shown to lead to increased nitric oxide within minutes after addition of estrogen88. Nitric oxide generation is involved in many beneficial effects on the cardiovascular system. Nitric oxide induces EC growth and migration, vasodilatation, and platelet inhibition, resulting in prevention of thrombus formation and leukocyte-EC adhesion71, 89.
Endothelial homeostasis is important in cardiovascular health and endothelial dysfunction is associated with an increased cardiovascular risk90. The rapid activation of eNOS is an important mechanism in vascular relaxation, EC migration and proliferation. Initial studies showed that an estrogen-BSA conjugate stimulates eNOS phosphorylation in cultured EC66. Estrogen can activate eNOS by multiple signaling pathways as illustrated in Figure 3A. Upon estrogen binding, membrane localized ERα forms a complex with p85α, the regulatory subunit of PI3K83 and with SH2 domain of c-Src mediated by the phosphorylated Tyr-537 of ERα78, 91 and/or by methylation of Arg260 of ERα92. The formation of the ERα/PI3K complex leads to activation of protein kinase B (Akt), extracellular-signal-regulated kinases 1/2 (ERK1/2), and phosphorylation and activation of eNOS. ERα can also activate MAPK signaling and stimulate nitric oxide production by eNOS87. Indeed, mice treated with estrogen showed increased eNOS activity and decreased vascular leukocyte accumulation after I/R injury and this vascular protective effect of estrogen was abolished in the presence of inhibitors of PI3K or eNOS83. ERα can also interact with the G proteins Gαi and Gβ through amino acids 251–260, and this interaction activates eNOS and monocyte adhesion, as well as stimulating cell migration75. A single alanine substitution generated within the Gαi binding domain of ERα blocked the ability of estrogen to induce ERK1/2 and eNOS activation. Also, blocking formation of the ERα-Gαi complex prevented EC cell migration93. Recently, Chambliss et al used EDC to demonstrate that EC proliferation and migration occurs via ERα-Gαi-Src kinase complex formation and resultant eNOS activation94. They also demonstrated that EDC can attenuate the development of neointimal hyperplasia following vascular injury. Interestingly, EDC did not cause endometrial carcinoma cell growth in vivo or increase cell proliferation in breast cancer cells. The activation of PI3K upon ERα-Gαi complex formation increases the activity and expression of the metalloproteinase (MMP) MMP2 and MMP9 leading to EGF receptor stimulation, which activates the MAPK kinase signaling pathway84.
Immunoprecipitation experiments demonstrated that ERα can bind the adaptor striatin through amino acids 183–253 of ERα. Estrogen promotes the formation of a complex containing ERα, striatin and Gαi leading to activation of ERK1/2, Akt and eNOS pathways stimulating EC migration80, 95. Beside nitric oxide generation, the non-genomic ER signaling pathway also modulates intracellular calcium homeostasis at physiological estrogen concentration (10−9M) in human EC, causing a rapid increase in intracellular calcium96. Treatment with estrogen conjugated to BSA caused a rise in calcium concentration in rat EC whereas treatment with the ER antagonist ICI182,780 blocked this effect97. Taken together, estrogen activation of membrane ERs leads to rapid signaling mechanisms that target nitric oxide signaling, thus modulating EC proliferation, migration and vascular relaxation.
The proliferation of vascular smooth muscle cells (VSMC) plays a crucial role in vascular diseases, especially in atherosclerosis. Similar to EC, ERα and ERβ are associated with caveolin-1 at the plasma membrane of VSMCs. Several studies demonstrate that estrogen rapidly (within 30 minutes) inhibits VSMC proliferation98, 99. Upon estrogen treatment, ERα stimulates the activity and expression of several phosphatases, such as MAP kinase phosphatase-1 (MKP-1), Src homology region 2 domain-containing phosphatase-1 (SHP-1), phosphatase and tensin homolog (PTEN), and protein phosphatase 2A (PP2A). Activation of these phosphatases, which reverses phosphorylation, result in a decrease of cell growth and migration99–101 (Fig 3, Panel B). Consistent with these findings, a recent study showed that estrogen significantly blocks VSMC proliferation, and decreases Akt phosphorylation due to enhanced PP2A activation mediated by a complex formed between ERα and PP2A102. They also used VSMC derived from a transgenic mouse line overexpressing a peptide that prevents ER trafficking to the plasma membrane and demonstrated that in this transgenic mouse, cell proliferation was not inhibited upon addition of estrogen. Similar results were obtained by other groups using transgenic mice overexpressing a peptide that prevents ERα from interacting with striatin, a scaffold protein essential for ER binding to the caveolae. These studies demonstrated that the ability of estrogen to stimulate VSMC growth after carotid artery injury was lost in mice overexpressing the disrupting peptide95. Microarray analysis of ex vivo aorta from these mice showed that the transcriptional response was greatly altered by blocking the non-genomic signaling. The genes showing different responses in the transgenic versus wild type mice were associated with vascular function, indicating that acute estrogen signaling also plays a key role in physiological vascular gene regulatory responses. Further studies are needed to better understand the molecular mechanisms of non-genomic signaling as well as the effects on gene transcription. A better understanding of these mechanisms could contribute to the development of new treatments to improve vascular health.
There are very few studies examining the non-genomic role of ERα and ERβ in cardiomyocytes. No studies have been done to distinguish genomic and non-genomic actions using transgenic models or extranuclear ER agonists. Estrogen was shown to have a role in cardioprotection against I/R injury. Short-term estrogen treatment (15 minutes before ischemia and 5 minutes at the beginning of reperfusion) improved heart function similar to that seen with pacing postconditioning3, 103. In addition, these beneficial effects of estrogen were not observed in the presence of a NOS inhibitor, suggesting that nitric oxide production via ER activation plays a key role in this process103. Also, there are a number of studies showing that estrogen prevents cardiac hypertrophy; in particular through ERβ4, 104. Levin’s group demonstrated that estrogen is not able to prevent angiotensin-induced hypertrophy and fibrosis in ERβ knock out mice105. They provide data showing that rapid estrogen signaling acts on cardiac hypertrophy sequestering the transcription factors in the cytoplasm of cardiomyocytes to prevent target gene activation. Indeed, estrogen through ERβ can activate the modulatory calcineurin-interacting protein 1 gene blocking calcineurin activation in myocytes22. Histone deacetylase (HDACs) are important in regulation of cardiac hypertrophy: class II is involved in signaling that inhibit this disorder, whereas class I promotes hypertrophy106. Recently, Pedram at al demonstrated that angiotensin regulates gene and protein level of both class I and II HDACs and rapid estrogen action through ERβ blocks these angiotensin actions107. Estrogen causes inhibition of calcium flux trough a Gαi-dependent mechanism, inhibiting Ca2+/calmodulin-dependent protein kinase II-induced class II HDAC4 phosphorylation and nuclear translocation. Also, angiotensin stimulates protein kinase C and D with consequent phosphorylation of class II HDC5 and estrogen is able to prevent these effects107. In summary, these findings suggest an important role for estrogen in cardioprotection due to rapid estrogen action. Additional studies should be conducted using the novel extranuclear ER agonists together with ER null transgenic mice to better understand the mechanisms involved in the protection against heart disease.
It has recently been shown that estrogen can signal through an orphan G-protein-coupled receptor, GPR3063, 108–110, also known as G protein-coupled estrogen receptor 1 (GPER). Two independent groups found that estrogen can directly bind to GPER at the membrane in monkey kidney fibroblast cells and breast cancer cells63, 109. Haas and colleagues later discovered GPER expression in VSMC of human arteries and veins and confirmed that this novel receptor is regulated by estrogen111. There is some debate regarding the cellular location of GPER. Prossnitz’s group expressed GFP-GPER in monkey kidney fibroblasts and images obtained using confocal fluorescence microscopy revealed that GPER is localized to the endoplasmic reticulum and the Golgi apparatus, but not to the plasma membrane63. They also found that estrogen binding to GPER increases calcium mobilization by a signaling pathway involving epidermal growth factor receptor transactivation108 rather than the more common mechanism involving ER activation of phospholipase C (PLC)–dependent inositol 1,4,5-trisphosphate production. In contrast, other groups reported that GRER is localized to the plasma membrane112, 113, 114. A possible explanation for these discrepant data has been proposed by Filardo’s group, which showed that GPER undergoes constitutive endocytosis resulting in a rapid (half live of 30 min) down-regulation. The recycled receptor that is present in the endosomes and endoplasmic reticulum levels might result from high levels of synthesis of the protein115. In summary, GPER localization appears to be heterogeneous indicating that GPER can localize both at the plasma membrane and in specific intracellular sites. However, whether GRER is functionally active in the endoplasmic reticulum membrane is debated.
Several studies investigated the role of GPER in vascular cells. Interestingly, the classical ERα and ERβ antagonist, ICI182,780, was not able to attenuate estrogen-induced vasorelaxation in either canine coronary or small rat arteries116–118. Although ICI 182,780 is an antagonist of ERα and ERβ, it is an agonist of GPER. Furthermore, removal of ERα and ERβ did not abolish estrogen-dependent responses in the cardiovascular system119. These results indicated that estrogen was not acting solely through ERα or ERβ, but that another receptor was involved. This led to the suggestion that activation of GPER might be responsible for the vasodilatory effects of estrogen. Several studies using the selective GPER agonist (G-1) were conducted confirming its ability to initiate vascular relaxation120. Acute G-1 treatment induces relaxation on porcine coronary arteries121, on rat aorta122, and human internal mammary arteries123. Infusion of G-1 to normotensive rats, results in an acute reduction in blood pressure. Furthermore, GPER knock out mice did not exhibit a vascular response on addition of estrogen123. Moreover, in hypertensive mRen2.Lewis rats, G-1 treatment leads to improved cardiac relaxation and reduced myocyte hypertrophy124. Furthermore, GPER agonist treatment induces vasodilation in female rats125. It has been suggested that G-1 results in 30–40% vasorelaxation in vivo126, however the mechanism mediating this effect is poorly understood. There are data reporting actions of GRER via endothelium-dependent and endothelium-independent mechanisms. Estrogen binding to GPER leads to activation of eNOS, thus increasing the production of nitric oxide in coronary ECs to relax these arteries116. Endothelium denudation or treatment with L-NAME, a NOS inhibitor, completely abolished the relaxation of porcine coronary arteries induced by G-1 action, consistent with a role for eNOS activation116. In contrast, several other studies demonstrated that G-1 treatment can still have an effect on vasorelaxation in endothelium-denuded aorta122 and porcine coronary arteries127. It was also shown that inhibition of eNOS does not affect this vasodilation response. Yu et al suggested that the endothelium-independent effect of coronary smooth muscle was mediated by a large conductance calcium-activated potassium channel in porcine and human arteries127. Moreover, treatment with an eNOS inhibitor did not block the G-1-induced relaxation, but relaxation was inhibited by blocking the potassium channel. A scheme of these mechanisms is shown in Figure 4.
The studies, suggesting that activation of GPER alone mediates vasodilation, contrast with studies suggesting that ERα and/or ERβ active NOS and mediate vasodilation. The reason for this discrepancy is unclear, but it may be due to differences in models, specificity of agonists and antagonists128 or cross-talk between different receptors.
Normally, VSMC proliferation is slow, however during stress, such as atherosclerosis or hypertension, these cells increase the rate of proliferation and migration129. Estrogen decreases VSMC proliferation after injury.130–132 However, the anti-proliferative actions of estrogen do not require ERα or ERβ as estrogen retains its anti-proliferative effect in ERα and ERβ knock out mice, suggesting that other receptors such as GPER might play a role133, 134.131 Li and colleagues recently demonstrated in coronary VSMCs that estrogen activation of GPER can decrease cell proliferation and migration135. Estrogen also has a role in regulating proliferation of vascular ECs, thus preventing excessive EC proliferation that can occur during atherosclerosis94. Treatment with G-1 reduced DNA synthesis and endothelial proliferation by blocking the transition between G2 and M phases of the cell cycle in mouse microvascular ECs and also in human umbilical vein ECs136, again showing a role for GPER.
Ullrich et al have shown in cardiomyocytes that estrogen-dependent intracellular calcium signaling is independent of classic ERs, ERα and ERβ119. They treated cardiomyocytes from ERα and ERβ knock out mice with serotonin and showed that the estrogen-mediated inhibition of Ca2+ influx and muscle contraction was not altered. Haas et al developed a GPER knock out mouse and confirmed that GPER stimulation is important for changing the intracellular Ca2+ concentration123. Thus, estrogen may play a role through GPER by regulating the myofilament sensitivity to Ca2+ influx.
It has also been shown that estrogen-GPER activation protects against cardiac I/R injury137. Deschamps et al demonstrated that acute treatment of perfused rat heart with G-1 is cardioprotective, as it improved the contractile dysfunction after ischemia. The protection afforded by G-1 was shown to be dependent on PI3K signaling as a PI3K inhibitor blocked G1-mediated cardiac protection138. Also, GPER activation has been shown to inhibit mitochondria permeability transition pore opening after I/R injury, thereby protecting hearts from cell death139. It was also reported that G-1 can reduce the inflammatory response in global ischemic injury140. The discovery that GPER activation of acute signaling mechanisms results in myocardial protection provides the exciting possibility that GPER agonists can be used to initiate cardioprotection while avoiding the side effects of estrogen due to nuclear ER stimulation in the uterus and breast.
GPER studies in the cardiovascular field have been growing rapidly and overall these findings indicate a protective role of GPER in cardiovascular disease. Agonists of GPER are potential selective ER modulators (SERMs) and are likely to be good candidates for intervention in many cardiac dysfunctions. However, the molecular mechanisms downstream of GPER stimulation are still unclear and future research needs to be done to clarify the signaling pathways involved and to understand the potentially beneficial effects of this receptor.
Although nuclear and acute effects of estrogen can individually lead to changes in cell signaling, gene expression and cell function, there is also cross-talk between the non-genomic and genomic pathways which work together to modulate cell and organelle function. Estrogen acts synergistically, coordinating the activation of kinase signaling and the genomic signaling to regulate gene transcription. Mice lacking the ERα palmitoylation site, in which signaling is only via nuclear DNA binding and not via membrane signaling, were generated and offer insight into the collaboration between non-nuclear and nuclear ERα and ERβ signaling141,142. Pedram et al found that nuclear ERα and ERβ alone are not able to maintain normal reproductive organ development; these mice showed infertility, abnormal ovaries and pituitary hormone regulation141. Overall, they also demonstrated that loss of membrane-ERα compromises estrogen-induced expression of some important mRNAs in the uterus, mammary glands and ovaries. Moreover, Pedram et al developed a mouse that expresses the ligand binding domain of ERα that is targeted exclusively to the plasma membrane. After ovariectomy and treatment with an ERα agonist, they carried out DNA microarray from liver mRNA. The majority of genes were regulated only by nuclear ERα, however mRNAs for cholesterols, triglyceride and fatty acid synthesis were comparably suppressed in membrane-only-ERα and wild type mice and completely absent in ERα knock out mice. Several transcription factors are regulated by protein kinase-mediated phosphorylation. These transcription factors might regulate gene expression by ERs directly binding to DNA (Fig. 2, Panel B). For example, AP-1 activity is regulated via MAPK-mediated phosphorylation143. Estrogen can activate MAPK signaling that results in enhanced AP-1 DNA binding and lead to gene expression activation144, 145. Also, activation of the PI3K-Akt signaling pathway by estrogen leads to nuclear factor kappa-light-chain-enhancer of activated B cells (NF-kB) phosphorylation and activation resulting in enhanced expression of genes that contain an NF-kB binding site146. Katzenellenbogen’s lab recently reported that ~25% of estrogen-regulated genes are responsive with EDC treatment147. Also, knockdown of ERα abolished gene expression due to EDC treatment. Conversely, the knockdown of GPER showed no effect147, indicating that in this study the cross-talk between the non-genomic and genomic effects is mediated through ERs.
Premenopausal women have a lower incidence of cardiovascular disease compared with age-matched men. However, this advantage for women gradually disappears after menopause with the cessation of ovarian function and reductions in estrogen levels. Both the ERα and ERβ have been localized in the cardiovascular system148. The expression level of ERs appears to vary with gender38, 149, 150. Whether ER isoforms are distributed differently in cardiovascular system and whether they change with aging or gender is an important area for further investigation and would partially explain the clinical conflicting evidence of estrogen treatment in cardiovascular disease. However, changes in ER expression and signaling pathway are poorly investigated. Estrogen has been shown to positively regulate ER levels151, and also other hormones such as progesterone and vitamin D negatively regulate ER levels152. Measured by western blotting ERα and ERβ did not differ with age in aorta from spontaneously hypertensive rats153. A study using immunofluorescence reported that vascular EC ERα expression in post-menopausal women is 33% lower than in the late follicular phase in premenopausal women153, 154. In astrocytes, using immunohistochemistry and western blot analysis Arimoto et al reported an increase in ERα levels with aging155. Using immunofluorescence Wu et al also found an increase in ERα with age in the hypothalamus156. Alterations in methylation of CpG islands, a cytosine and guanine rich area in gene promoter regions, in the ER promoter have been reported to occur with age and disease and these changes could be involved in altering ER levels. Methylation of the ER promoter has been reported to increase with athrosclerosis157, but to decrease in brain following stroke158. In future studies it will be important to better define changes in expression and localization of ER with age, sex and disease in the cardiovascular system.
Although estrogen signaling has been studied for years in reproductive tissues, it is only in the past decade or so that its role in the cardiovascular system has been appreciated. It has become apparent that estrogen has effects on cardiovascular tissues. Estrogen has important effects on lipid profile (such as lowering LDL cholesterol), vascular remodeling, blood pressure, endothelial relaxation, development of hypertrophy and cardioprotection. However, the mechanisms by which estrogen mediates these effects are still poorly understood. As detailed in this review, estrogen can regulate transcription by binding to two nuclear receptors, ERα and ERβ, which differentially regulate gene transcription. These receptors are present at different ratios on different cells and their levels can change with time. ERα and ERβ regulation of gene transcription is also modulated by tissue specific co-activators and co-repressors. Taken together estrogen can result in tissue- and temporal-specific regulation of gene transcription. Layered on top of tissue specific estrogen regulation of transcription, and adding to the complexity, is membrane delimited signaling. Estrogen can bind to ERα and ERβ localized at the plasma membrane as well as GPER; estrogen binding enhances acute kinase signaling pathways that synergize with the estrogen receptor to mediated transcriptional changes. In addition, acute signaling cascades can lead to phosphorylation of estrogen receptors that can lead to ligand independent ER regulation of transcription (see Figure 2). Understanding these complexities of estrogen-ER signaling is necessary to better understand why premenopausal women have reduced cardiovascular disease but yet HRT was not beneficial, at least in older menopausal women. Furthermore a better understanding of what regulates ERα versus ERβ and how the nuclear ER transcriptional activation synergizes with acute membrane ER signaling will allow better design of selective estrogen modulator which may provide beneficial effects of estrogen without unwanted side effects.
SOURCES OF FUNDING
This work was supported by NHLBI Intramural funds.