Reciprocality Between Estrogen Biology and Calcium Signaling in the Cardiovascular System

Abstract

17β-Estradiol (E₂) is the main estrogenic hormone in the body and exerts many cardiovascular protective effects. Via three receptors known to date, including estrogen receptors α (ERα) and β (ERβ) and the G protein-coupled estrogen receptor 1 (GPER, aka GPR30), E₂ regulates numerous calcium-dependent activities in cardiovascular tissues. Nevertheless, effects of E₂ and its receptors on components of the calcium signaling machinery (CSM), the underlying mechanisms, and the linked functional impact are only beginning to be elucidated. A picture is emerging of the reciprocality between estrogen biology and Ca²⁺ signaling. Therein, E₂ and GPER, via both E₂-dependent and E₂-independent actions, moderate Ca²⁺-dependent activities; in turn, ERα and GPER are regulated by Ca²⁺ at the receptor level and downstream signaling via a feedforward loop. This article reviews current understanding of the effects of E₂ and its receptors on the cardiovascular CSM and vice versa with a focus on mechanisms and combined functional impact. An overview of the main CSM components in cardiovascular tissues will be first provided, followed by a brief review of estrogen receptors and their Ca²⁺-dependent regulation. The effects of estrogenic agonists to stimulate acute Ca²⁺ signals will then be reviewed. Subsequently, E₂-dependent and E₂-independent effects of GPER on components of the Ca²⁺ signals triggered by other stimuli will be discussed. Finally, a case study will illustrate how the many mechanisms are coordinated to moderate Ca²⁺-dependent activities in the cardiovascular system.

Main Components of the Calcium Signaling Machinery (CSM) in Cardiovascular Tissues

The CSM herein refers to proteins responsible for the generation or sequestration of intracellular Ca²⁺ signals and their transduction to target activities. In this section, key CSM components in cardiovascular tissues will be briefly described to facilitate review of the relevant effects and mechanisms of estrogenic agonists and receptors.

Intracellular Ca²⁺ Stores, Release, and Uptake Mechanisms

Organelles Functioning as Intracellular Ca²⁺ Stores

The sarcoplasmic/endoplasmic reticulum (SR/ER) is the main Ca²⁺ store in cardiomyocytes, vascular smooth muscle cells (VSMCs) (1, 2), and endothelial cells (ECs), where the ER stores ~75% Ca²⁺ and mitochondria house ~25% (3). The Golgi (4, 5) and lysosomes have more recently been recognized as Ca²⁺ reservoirs (6, 7). Ca²⁺ reaches 5 × 10⁻⁴ M in the ER/SR and lysosomes and 1.3–2.5 × 10⁻⁴ M between the trans-Golgi and cis-Golgi (5, 8). The medial Golgi also releases Ca²⁺ in response to inositol-triphosphate receptor (IP₃R) and ryanodine receptor (RyR) stimulation (9). Crosstalk between the ER/SR and other organelles affects their Ca²⁺ fluxes (10–14). In neonatal cardiomyocytes, beat-to-beat oscillations in mitochondrial and cytosolic Ca²⁺ occur in parallel (15), and mitochondrial uptake reduces cytosolic Ca²⁺ (16).

Mechanisms of Ca²⁺ Uptake Into Ca²⁺ Stores

SR/ER Ca²⁺-ATPases (SERCAs) are the key Ca²⁺ uptake mechanisms. For each ATP hydrolyzed, they pump 2 Ca²⁺ ions into the ER/SR in exchange for less than four H⁺ ions (17). SERCA2b is ubiquitously expressed. SERCA2a predominates in cardiomyocytes and is essential for cardiac development (18). SERCA3 is the predominant vascular isoform; its deletion causes smooth muscle relaxation abnormality (19, 20). SERCA3 has lower affinity for Ca²⁺ and is only active at high Ca²⁺ levels. Non-phosphorylated phospholamban interacts with SERCA1a, SERCA2a, and SERCA2b and reduces their Ca²⁺ affinity. Phosphorylation at Ser16 and Thr17 removes phospholamban–SERCA interaction, promoting SERCA activity (21, 22). Sarcolipin also binds SERCAs and reduces their Ca²⁺ affinity. Its deletion increases SR Ca²⁺ uptake (23).

The secretory pathway Ca²⁺ pump (SPCA) mediates Ca²⁺ uptake into the Golgi with nanomolar affinity for Ca²⁺. Unlike the SERCA, Ca²⁺ transport by SPCA is not associated with counter transport of H⁺. In the medial Golgi, both SERCA and SPCA participate in Ca²⁺ uptake (9).

Mitochondrial Ca²⁺ uptake is mediated by the voltage-dependent anion channel (VDAC) and the mitochondrial Ca²⁺ uniporter (MCU). VDACs are non-selective anion channels in the open state yet in the “closed” state permit influxes of cations such as K⁺, Na⁺, and Ca²⁺ into the mitochondria (24). VDAC isoforms participate equally in transporting Ca²⁺ triggered by IP₃-producing agonists; however, VDAC1 selectively transports apoptotic Ca²⁺ signals (25). Myocardial VDAC2 regulates rhythmicity by influencing the spatial and temporal properties of cytoplasmic Ca²⁺ signals (26). The MCU constitutes a low-affinity yet selective Ca²⁺ channel pore as part of a mitochondrial Ca²⁺ uptake protein complex (MICU) and the essential MCU regulator (27, 28).

Mechanisms of Ca²⁺ Release From Ca²⁺ Stores

In IP₃Rs, IP₃ binds with IP₃R2 > IP₃R1 > IP₃R3 affinity order (29) and cooperatively switches IP₃R tetramers to an open conformation to form clusters and release Ca²⁺ (30, 31). IP₃Rs regulate Ca²⁺ release from the ER/SR, Golgi apparatus, and nucleus (32). ER/SR Ca²⁺ release depletes ER Ca²⁺ and triggers store-operated Ca²⁺ entry (SOCE). IP₃R2 predominates in the cardiomyocytes (33). In failing hearts, IP₃R-mediated Ca²⁺ transients are enhanced, and mitochondrial Ca²⁺ uptake is reduced, which facilitates contraction and spontaneous action potentials that increase arrhythmogenicity (34). In VSMCs, all IP₃Rs are expressed and are important for agonist-induced contraction (35). Endothelial IP₃R1 is predominant in the brain (36), whereas IP₃R2 and IP₃R3 are abundant in the aorta and pulmonary arteries (37, 38).

RyRs (RyR1–RyR3) are the main SR Ca²⁺ release channels (39). Regulation by cytosolic Ca²⁺: In cardiomyocytes, RyR2 predominates (40) and is closed, activated, and inhibited, respectively, at Ca²⁺ <10⁻⁷ M, ~10⁻⁷-10⁻⁵ M, and >10⁻³ M (41). Entry via voltage-dependent Ca²⁺ channels (VDCCs) stimulates Ca²⁺-induced Ca²⁺ release (CICR) via RyR2, contributing to myocardial contraction. In VSMCs, RyR2 predominates in the aorta and pulmonary and cerebral arteries, while RyR3 is the only isoform in basilar arteries (42–44). CICR also contributes to VSMC contraction, but not as critically as in cardiomyocytes; indeed, skinned smooth muscle fiber bundles can contract at Ca²⁺ levels that do not activate RyRs (45). In ECs, RyR2 is on the ER and mitochondria (46); however, RyR agonists only cause a slow Ca²⁺ release that corresponds to a reduction in the IP₃-sensitive Ca²⁺ pool (47, 48). Regulation by SR Ca²⁺ is important in cardiomyocytes. SR Ca²⁺ overload triggers spontaneous RyR2-mediated Ca²⁺ release, a phenomenon called store overload-induced Ca²⁺ release (SOICR) (49, 50). SOICR can cause delayed afterdepolarizations leading to tachycardias and is abolished by an E4872A mutation in the RyR2 gate (51).

Ca²⁺ Entry

Store-Operated Ca²⁺ Entry (SOCE)

SOCE is a ubiquitous mechanism where Ca²⁺ store depletion triggers Ca²⁺ influx (52, 53). Proposed in the 1980s, SOCE was confirmed in the mid-2000s with the discoveries of the stromal interaction molecule 1 (STIM1) (54–56) and Orai Ca²⁺ channels (57–59). STIM1 resides mainly on the ER/SR membrane and has a luminal EF hand that houses a Ca²⁺-binding loop (60). In Ca²⁺-full ER/SR, the loop is in a closed conformation. Upon ER/SR Ca²⁺ depletion, Ca²⁺ leaving the loop promotes STIM1 oligomerization to interact with Orai1 channels and trigger Ca²⁺ entry (61–63). STIM1 also interacts with L-type Ca²⁺ channels (LTCCs) (64), maintains ER/SR structure (65–67), and is upregulated in atherosclerosis and hypertension (68–71). Myocardial SOCE is normally not prominent; however, STIM1 and SOCE are increased in heart failure (67, 72–76). In VSMCs, SOCE contributes significantly to contraction; α₁AR-mediated contraction is reduced ~30% in SM-specific STIM1^−/− animals (77). In ECs, SOCE is the major Ca²⁺ entry and is required for many critical functions such as endothelial nitric oxide synthase (eNOS) activity and proliferation (78–82).

Voltage-Dependent Ca²⁺ Entry (VDCE)

Functional voltage-dependent Ca²⁺ channels (VDCCs) are the hallmark of tissue excitability and are present in cardiomyocytes and VSMCs, but not ECs. In cardiomyocytes, LTCCs are located mostly in transverse T tubules in apposition to RyR2s (83). Ca²⁺ entry via LTCCs triggers CICR via RyR2. In VSMCs, LTCCs also play a critical role in Ca²⁺ entry and contraction (84). The LTCC complex (85) consists of α₁, α₂, β, δ, and γ subunits. Four LTCC members are named according to their α1 pore-forming subunits: Ca_v1.1, Ca_v1.2, Ca_v1.3, and Ca_v1.4 (86). Ca_v1.2 is predominant in cardiac and smooth muscles.

Ca²⁺ Extrusion via the Plasma Membrane/Sarcolemma

The plasma membrane Ca²⁺-ATPases (PMCAs) prevail for Ca²⁺ extrusion in non-excitable tissues while the Na⁺-Ca²⁺ exchanger (NCX) is more important in excitable cells. SERCA2a, NCX, and PMCA sequester, respectively, ~70, 28, and 2% of cytosolic Ca²⁺ in cardiomyocytes (83) and 25, 25, and 50% in ECs (87).

Plasma Membrane Ca²⁺-ATPase

PMCAs extrude one Ca²⁺ ion for each ATP used and function as Ca²⁺-H⁺ exchangers (88–90). PMCAs are regulated by a Ca²⁺-dependent interaction with calmodulin (CaM). At low Ca²⁺, a C-terminal autoinhibitory domain binds to two cytosolic loops and inhibits pump activity. Increased Ca²⁺ promotes CaM–PMCA interaction, which removes inhibition and activates Ca²⁺ efflux (91, 92). PSD-95 promotes expression and distribution of PMCA4b via PDZ binding (93). PMCAs are inhibited by C-terminal tyrosine phosphorylation (94). Myocardial PMCAs play a little role under physiological conditions. However, expressions of PMCA1 and PMCA4 are reduced by up to 70 and 50%, respectively, in end-stage heart failure (95), and cardiac-specific overexpression of PMCA4b improved myocardial functions in ischemia–reperfusion injury and heart failure (96). PMCAs concentrate in the caveolae of VSMCs and ECs (97, 98). PMCA1 suppresses VSMC proliferation (99, 100), while PMCA4 mediates cell cycle (101, 102). In ECs, PMCA1b, and PMCA4b are predominant (87, 103, 104).

Na⁺-Ca²⁺ Exchanger

The NCX may function in two modes. In the forward mode, myocardial NCX1 balances LTCC-mediated Ca²⁺ entry and RyR-mediated Ca²⁺ release during cardiac excitation, extruding ~25% of the Ca²⁺ needed to activate myofilaments (105). NCX1 also predominates in VSMCs (106, 107). In ECs, NCX accounts for ~25% of Ca²⁺ removal (87). Endothelial NCX and PMCA dynamically adjust their Ca²⁺ extrusion rates to maintain sufficient efflux (104). In the reverse mode, upon myocardial depolarization, Na⁺ entry causes the NCX to transiently operate in this mode, promoting Ca²⁺ entry. This is much less efficient in triggering SR Ca²⁺ release compared to LTCC-mediated Ca²⁺ entry (108, 109). However, it primes the dyad to increase LTCC-mediated CICR (110). In VSMCs, reverse-mode NCX1 facilitates Ca²⁺ entry and mediates contraction, vascular tone, and blood pressure (111, 112). The reverse mode is not significant in ECs.

Sex Differences in Ca²⁺ Signaling Proteins

Higher mRNA levels of Ca_v1.2, RyR, and NCX, but not of phospholamban and SERCA2, have been observed in female than in male rat hearts (113). However, caffeine-induced Ca²⁺ release is lower in cardiomyocytes from female hearts (114). Ca_v1.2 mRNA is higher in coronary smooth muscle from male than from female pigs (115). In smooth muscle cells (SMCs), expressions of ERα and ERβ, but not G protein-coupled estrogen receptor 1 (GPER), are higher in female than in male rats (116). These differences and the lower Ca_v1.2 expression (115) may be responsible for less contraction of VSMCs from females (116). No studies have examined sex differences in Ca²⁺ handling proteins in ECs.

Transduction of Ca²⁺ Signals—The Essential Role of Calmodulin (CaM)

While some Ca²⁺-dependent proteins are activated directly by Ca²⁺, many are activated by a complex between Ca²⁺ and CaM. CaM has two lobes linked by a flexible helix and can interact with ~300 target proteins (117, 118). Ca²⁺-free CaM binds or serves as structural subunits of ~15 proteins (119). However, each CaM lobe has two Ca²⁺-binding sites, and cooperative Ca²⁺ binding induces conformations that allow CaM to interact with many proteins, aided by the flexibility of the central helix (120, 121). Thus, CaM is the ubiquitous Ca²⁺ signal transducer. Activities of Ca²⁺/CaM-binding proteins depend on the Ca²⁺ signals, CaM availability, and properties of the interaction between Ca²⁺-CaM and the target proteins. Many of these factors are subject to estrogenic moderation.

Despite being required for activation of many Ca²⁺-dependent proteins, up to 50% of cellular CaM is engaged in inseparable interactions, leaving much less available for dynamic target binding (122). This generates an environment of limited CaM (123), as has been demonstrated in ECs (124), VSMCs (125), and cardiomyocytes (126). Consequently, competition for CaM generates a unique crosstalk among CaM-dependent proteins (124, 127), and factors that alter CaM level are predicted to have pervasive functional impact. It is noteworthy that virtually all CSM components interact with CaM and, in the context of reciprocality between estrogenic and Ca²⁺ signaling pathways, that ERα and GPER are both regulated by direct interactions with Ca²⁺-CaM.

Estrogen Receptors and Their Calcium-Dependent Regulation

Estrogen Receptor α (ERα)

ERα (128–130) is a nuclear receptor that, upon E2 binding (K_d ~ 10⁻¹⁰ M), assumes an active conformation to bind estrogen-responsive elements (EREs) in the promoters of target genes, modulating their transcription (131). Its N-terminus has a transcriptional activation function (AF-1) domain, a DNA-binding domain, and a hinge region; the C-terminus houses the ligand-binding domain and a second AF-2 domain. ERα is robustly expressed in the heart (132), VSMCs, and ECs (133–136).

ERα activities are strongly regulated by the Ca²⁺-dependent interaction with CaM. ERα binds CaM in a Ca²⁺-dependent fashion with a K_d of 1.6 ×10⁻¹⁰ M and an EC₅₀(Ca²⁺) value of ~3 × 10⁻⁷ M (137). When ERα from Wistar rats' uteri is used, CaM decreases ERα-E₂ binding but increases liganded ERα-ERE interaction (138, 139). A comparison of the CaM-bound/CaM-unbound ERα ratio in the cytosolic (unliganded) and nuclear (liganded) ERα pools isolated from MCF-7 cells suggests that E₂ binding induces a conformation that favors ERα-CaM interaction (138). The CaM-binding domain was initially predicted to be a.a. 298–310 (137) but was later determined to be a.a. 298–317, with a.a. 248–317 required for maximal interaction (140). Further studies revealed that a.a. 287–311 is required to interact with both CaM lobes (141). CaM binding promotes ERα homodimerization that is critical for transcription activity (140, 142). With two lobes, each CaM binds two ERα molecules and thus stabilizes ERα dimerization (143). Notably, analogs of ERα17p (a.a. 295–311) that are unable to bind CaM downregulates ERα, stimulates ERα-dependent transcription, and enhances proliferation of MCF-7 cells, as does the wild-type ERα17p, indicating that this domain may also be involved in CaM-independent posttranslational regulation of ERα (144).

Estrogen Receptor β (ERβ)

ERβ has ~96% and 55–58% sequence homology with ERα in the DNA- and ligand-binding domains, respectively (145, 146). ERβ binds E₂ with a K_d of ~4–6 × 10⁻¹⁰ M. ERβ forms homodimers but more preferentially forms heterodimers with ERα, which bind E₂ with a K_d of ~2 × 10⁻⁹ M and are transcriptionally active (147). ERβ is abundantly expressed in the vasculature (133–136). However, its expression and direct actions in the heart are controversial; cardiac manifestations in ERβ^−/− animals have been attributed to indirect effects from vascular changes (148). ERβ is not regulated by Ca²⁺ or CaM (149).

GPER

GPER (150), aka GPR30, was cloned from various tissues in the 1990s (151–156). GPR30 is required for estrogenic activation of extracellular signal-related kinase (ERK)1/2 via transactivation of the epidermal growth factor receptor (EGFR) and release of the heparan-bound epidermal growth factor (EGF) (157, 158). It was shown to bind E₂ in 2005 (159, 160), and the designation GPER was adopted by the International Union of Basic and Clinical Pharmacology in 2007 (161). A host of steroidal and non-steroidal agents and specific GPER agonists can activate GPER (150). GPER couples with Gα_s or Gα_i/o. Supporting Gα_s coupling are data that (1) most membrane-bound [³⁵S]GTPγ-S from cells overexpressing GPER and treated with E₂ coimmunoprecipitate with Gα_s (159), (2) GPER is present in Gα_s-pull-down fraction from GPER-expressing cells, and (3) E₂ promotes GPER-dependent cyclic adenosine monophosphate (cAMP) production (162). Supporting GPER-Gα_i/o association are results that pertussis toxin prevents (1) E₂-induced, GPER-mediated ERK1/2 phosphorylation in cells transfected with GPER (134, 157); (2) upregulation of c-fos in ERα/ERβ-negative, GPER-positive SKBr3 cells (163); and (3) E₂-induced Ca²⁺ signals in ECs (164).

GPER is robustly expressed in cardiovascular tissues (133–136). In ECs, GPER mRNA is increased 8-fold by shear stress (154). GPER is localized on the ER/SR membrane (160) and responds to cell-permeable ligands (165). However, it also resides on the plasma membrane (166) and requires its C-terminal PDZ-binding motif to do so (167). The plasmalemmal GPER pool seems to constitutively undergo clathrin-dependent endocytosis and accumulate in the trans-Golgi network for ubiquitination in the proteasome without recycling to the plasma membrane, a process unaffected by agonist stimulation (168). Despite its predominant expression in the ER/SR, the sequence that drives GPER localization here has not been identified.

GPER is directly regulated by Ca²⁺-CaM complexes. In VSMCs and ECs, GPER coimmunoprecipitates with CaM in a constitutive association that is promoted by treatment with E₂, G-1, or receptor-independent stimulation of Ca²⁺ entry (169, 170). GPER is the first G protein-coupled receptor (GPCR) shown to possess four CaM-binding sites on its respective four submembrane domains (SMDs) (169). Fluorescence resonance energy transfer (FRET) biosensors based on SMDs of GPER bind CaM with K_d from 0.4 to 136 × 10⁻⁶ M and affinity ranking SMD2 > SMD4 > SMD3 > SMD1. These interactions are Ca²⁺ dependent, with an EC₅₀ (Ca²⁺) of 1.3 × 10⁻⁷-5 × 10⁻⁶ M, values within the physiological Ca²⁺ range (169). Due to technical challenges with purifying full-length GPCRs, the K_CaM for GPER as a holoreceptor is not available. The presence of four CaM-binding sites makes this task even more challenging and, in some way, not useful functionally. Functionally, mutations that reduce CaM binding but that do not perturb GPER-G_βγ preassociation drastically prevent GPER-mediated ERK1/2 phosphorylation (170).

Stimulation of Calcium Signals by Estrogen and GPER Agonists

Observations

In rat hearts, E₂ (10⁻¹²-10⁻⁸ M) triggers ⁴⁵Ca²⁺ uptake that is inhibited by LTCC antagonists (171). In VSMCs, GPER agonist G-1 triggers a slow-rising Ca²⁺ signal that is <2 × 10⁻⁷ M (172). In MCF-7 cells, E₂ (10⁻⁷ M) induces Ca²⁺ store release and entry, yet only the former is required to activate mitogen-activated protein kinase (MAPK) (173). Interestingly, the ERα/ERβ antagonist ICI182,780 (10⁻⁶ M) also triggered Ca²⁺ signals in these cells. In ECs, E₂ (10⁻¹⁰-10⁻⁹ M) triggers Ca²⁺ store release and entry, effects not affected by ERα/ERβ inhibitor tamoxifen (164, 174). The data with ICI182,780 and tamoxifen implicate a receptor other than ERα or ERβ in mediating the Ca²⁺ signal. Both reagents were later shown to be GPER agonists, triggering ERK1/2 phosphorylation only in cells expressing GPER (157, 159). Later studies confirmed Ca²⁺ signals stimulated by E₂, GPER agonist G-1, and ICI182,780 in cells expressing GPER endogenously and absence of this effect in GPER^−/− cells (160, 175, 176).

Mechanisms (Figure 1)

Figure 1.

Components of the CSM that are affected by E₂ and/or GPER in excitable and non-excitable cardiovascular tissues. See text for details.

Direct E₂-Ca_v1.2 Interaction

E₂ (10⁻¹¹-10⁻⁹ M) potentiates I_Ca,L in neurons and HEK293 cells overexpressing the α1C subunit; nifedipine displaces membrane E₂ binding; and E₂'s effect is reduced by a dihydropyridine-insensitive LTCC mutant, indicating that E₂ binds to the dihydropyridine-binding site (177). Intriguingly, E₂ and the dihydropyridines exert opposite effects on I_Ca,L.

Direct, Membrane-Delimited Activation of Ca²⁺ Channels by Gα Subunits

GPCR stimulation can trigger Ca²⁺ signals independently of the second messenger (178–180). GPER couples with Gα_s and Gα_i/o, which can interact with LTCC (178, 181, 182) and trigger Ca²⁺ entry.

Release of G_βγ Subunit Upon GPER-Associated Gα_i Stimulation

G_βγ stimulates PLCβ (183–185) and activates IP₃R1 (186), both of which trigger Ca²⁺ store depletion and SOCE. Consistently, E₂-induced Ca²⁺ store release and entry in ECs are completely inhibited by pertussis toxin and PLCβ inhibitor U73122 (164). Also, HEK293 cells only produce a Ca²⁺ response to E₂ when expressing HA-tagged GPER (162). Since (1) Ca²⁺ entry channels are located on the membrane and (2) G_βγ activates IP₃Rs by interacting with the IP₃-binding sites (186) on IP₃Rs' cytosolic domains, both the membrane-delimited/Gα-mediated and G_βγ-mediated mechanisms should only be operable by the plasmalemmal GPER pool. A distinguishing feature is that the former mechanism would not trigger SR/ER Ca²⁺ release in the absence of extracellular Ca²⁺, whereas the latter would. Based on this feature, data fitting the former are available from renal tubular cells (176); and data fitting the latter, from vascular ECs (164).

Functional Impact

Do Ca²⁺ signals stimulated by estrogenic agonists activate Ca²⁺-dependent activities? When reported, the concentration of a Ca²⁺ signal allows for prediction of proteins that may or may not be affected by it. For example, E₂ induces ER Ca²⁺ release signals of ~2 × 10⁻⁷ M and activates MAPK (173), because this Ca²⁺ level is sufficient for MAPK activity (187); indeed, Ca²⁺ chelation abolishes E₂'s effect (173). Considering that GPER mediates the effect of E₂ to trigger Ca²⁺ signals that activate MAPK, GPER activity can promote many downstream effects (163, 170, 188). In ECs, E₂ (10⁻⁹-10⁻⁶ M) stimulates very small Ca²⁺ signals (<10⁻⁷ M) (174). One can predict that only proteins with very high Ca²⁺ sensitivity, for example, phosphorylated eNOS (170, 189, 190), would be activated by these signals. Whether a Ca²⁺ signal can produce a predicted effect also depends on other factors. For example, the Ca²⁺ signal of ~2 × 10⁻⁷ M triggered by G-1 in VSMCs (172) would be sufficient to activate myosin light-chain kinase (MLCK) and cause vasoconstriction, based on MLCK's properties (191). However, G-1 causes vasodilation (172, 192–194), likely by activating eNOS (170, 193, 195–198), inhibiting VSMC Ca²⁺ (199), and stimulating SMC K⁺ efflux (200).

Calcium Entry Inhibition by Estrogenic Agonists and Estrogen Receptors

To a large extent, estrogenic regulation of Ca²⁺ signaling involves effects of estrogenic agonists and receptors on the Ca²⁺ signals triggered by other stimuli, via both E₂-dependent and E₂-independent mechanisms.

Store-Operated Ca²⁺ Entry (Figure 1)

In VSMCs, E₂ (10⁻⁸-10⁻⁵ M) inhibits norepinephrine- and phenylephrine-induced arterial constriction in the presence of extracellular Ca²⁺ but not that induced in Ca²⁺-free medium (201, 202). These effects may be attributed to inhibition of both VDCE and SOCE, as α₁ adrenoceptor agonists can activate both (77). GPER-mediated inhibition of SOCE has been shown in ECs, where G-1 (10⁻⁸-10⁻⁶ M) suppresses SOCE induced by thapsigargin or bradykinin (203). Interestingly, the observations that in the absence of any treatment with agonists, thapsigargin-induced SOCE is increased by 80% in GPER-knockdown ECs and is reduced by 40% in GPER-overexpressing HEK293 cells implicate E₂-independent mechanisms (203).

How E₂/GPER suppresses SOCE seems to involve STIM1. G-1 treatment prevents thapsigargin-induced STIM1 puncta, indicating inhibition of STIM1's association with the Ca²⁺ channel; and Ser575/608/621Ala mutations of STIM1 reduce the inhibitory effect of G-1 (203). Consistently, E₂ inhibits Ser575 STIM1 phosphorylation in bronchial epithelial cells, thus suppressing STIM1 mobility and SOCE (204). Our initial data also indicate that dynamic physical interaction between them contributes importantly to GPER's inhibition of SOCE (205).

Voltage-Dependent Ca²⁺ Entry (Figure 1)

Electrically induced Ca²⁺ signals are increased in cardiomyocytes from ovariectomized (OVX) animals (206–208). Many lines of evidence indicate that GPER mediates the inhibitory effect of E₂ on I_Ca,L. These include inhibitory effects of E₂ (1–3 × 10⁻⁵ M) and combined ERα/ERβ antagonists/GPER agonists (ICI182,780, tamoxifen, or raloxifene) on I_Ca,L in cardiomyocytes from both WT and ERα^−/−/ERβ^−/− animals, as reviewed in (132). Similarly, in VSMCs, E₂ inhibits electrically induced I_Ca,L (209, 210), and ERα/ERβ antagonists/GPER agonists tamoxifen and ICI164,384 inhibit high-K⁺-induced contraction (202). GPER agonist G-1 (10⁻⁶ M) inhibits nifedipine-sensitive Ca²⁺ spikes in LTCC-expressing A7R5 SMCs, an effect prevented by GPER antagonist G-15 (10⁻⁶ M) (199); these concentrations are specific for GPER (175, 211). Consistently, ERα knockout does not affect E₂'s inhibition of KCl-induced ⁴⁵Ca²⁺ uptake in VSMCs and vasorelaxation (212).

How E₂ inhibits electrically induced VDCE is still unknown. Hypothetically, at high levels, E₂ binding to the dihydropyridine-binding site on LTCC (177) may instead inhibit I_Ca,L. As for prevention of β adrenoceptor (βAR)-mediated potentiation of VDCE, recent evidence suggests that GPER may be an intrinsic component of β₁AR activation. Thus, G-1 inhibits isoproterenol-induced increases in left ventricle (LV) pressure, heart rate, ectopic contractions, I_Ca,L, LTCC phosphorylation, and total myocardial Ca²⁺ signal, while the GPER inhibitor G-36 promotes ISO-induced Ca²⁺ signal and LTCC phosphorylation (213). Speculatively, GPER may do so in part by interacting with β₁AR or with A kinase-anchoring protein 5, thus inhibiting cAMP production (167). These may represent some E₂-independent effects of GPER. Studies in GPER-knockout tissues are needed to further clarify the mechanisms.

Estrogenic Regulation of Cytoplasmic Calcium Removal Mechanisms

SERCA Activity

Few studies, mostly in cardiac tissues, have examined the effects of E₂ on SERCA activity, with somewhat conflicting results. E₂ (1–30 × 10⁻⁶ M) does not affect the V_max of SR vesicle Ca²⁺ uptake in canine LV tissue (214). However, ovariectomy reduces the V_max but increases the Ca²⁺ sensitivity for SR Ca²⁺ uptake of rat LV homogenates or SR-enriched membrane fractions; mechanistically, these effects appear to be associated with reduced Thr17 phosphorylation of phospholamban and are restored by treatment with either E₂ or progesterone (215) (Figure 1). How E₂ and progesterone promote Thr17 phosphorylation of phospholamban is unknown, perhaps by inhibiting CaM kinase II (216), the enzyme that phosphorylates phospholamban (21). The effect of E₂ on SERCA activity in VSMCs has not been examined.

NCX Activity

As with SERCA activity, few studies have measured the effects of E₂ on NCX activity. Na⁺-dependent ⁴⁵Ca²⁺ uptake in rat LV myocytes is increased by ~3-fold after 60 days of ovariectomy, which is restored by replenishment with E₂ (1.5 mg/60 days) (208). During myocardial ischemia, intracellular Na⁺ concentration is higher in male than in female cardiomyocytes and is associated with increased Ca²⁺ concentration as a result of increased NCX activity (217). These studies are consistent with an inhibitory effect of E₂ on NCX activity in both the forward and reverse modes (Figure 1). However, the mechanisms of this inhibition are unclear.

Mitochondrial Ca²⁺ Uptake

In the heart, diethylstilbestrol (0.9–1.8 × 10⁻³ M) inhibits mitochondrial ⁴⁵Ca²⁺ uptake (218). Mitochondrial Ca²⁺ retention capacity (mCRC), a combination of mitochondrial Ca²⁺ uptake, total mitochondrial Ca²⁺-binding sites, and mitochondrial Ca²⁺ release mechanisms, is a determinant of the protective role of the mitochondria during cytoplasmic Ca²⁺ overload. E₂ (4 × 10⁻⁸ M) increases myocardial mCRC following ischemia–reperfusion, an effect abolished by genetic deletion of GPER but not of ERα or ERβ; mechanistically, this effect seems to involve PKC-dependent, MAPK-dependent phosphorylation of glycogen synthase kinase (GSK)-3β, leading to inhibition of the mitochondrial permeability transition pore (219). Consistently, E₂ (10⁻⁸ M) inhibits high Ca²⁺-induced cytochrome c release from myocardial mitochondria (220). In ECs, 48-h E₂ (10⁻⁸ M) treatment inhibits mitochondrial Ca²⁺ uptake, an effect abolished by the ERα/ERβ antagonist ICI182,780 (10⁻⁸ M) (221). The mechanisms whereby E₂ inhibits mitochondrial Ca²⁺ uptake are still unknown (Figure 1).

PMCA Activity

Recent data show that GPER inhibits PMCA activity via both E₂-dependent and E₂-independent mechanisms (Figure 1). E₂-dependent mechanisms are evidenced by the effects of G-1 (10⁻⁸-10⁻⁶ M) and E₂ (1–5 × 10⁻⁹ M) to inhibit PMCA-mediated efflux in primary ECs without affecting PMCA expression levels and to promote PMCA phosphorylation at Tyr1176 (135, 170), which is known to inhibit pump activity (94). Notably, this phosphorylation masks the stimulatory effect of enhancing the PMCA–CaM interaction produced by 48-h E₂ treatment (170). E₂-independent mechanisms are indicated by the findings that (1) GPER constitutively interacts with PMCA4b via the anchoring action of PSD-95 at their C-terminal PDZ-binding motifs; (2) overexpression of GPER decreases PMCA activity; (3) GPER knockdown promotes PMCA activity; and (4) PSD-95 knockdown or truncation of the PDZ-binding motif on GPER releases GPER–PMCA association and promotes PMCA activity (135). Functionally, these mechanisms collectively prolong agonist-induced Ca²⁺ signal and enhance eNOS activity in ECs (135, 170, 203). Consistent with suppressed Ca²⁺ efflux, the Ca²⁺ signals stimulated by E₂ and the GPER agonist G-1 in cells overexpressing GPER reported by various laboratories display much more prolonged plateau phases compared to Ca²⁺ signals in cells not overexpressing GPER or those stimulated by other agonists such as ATP or bradykinin (160, 162, 164, 175). GPER–PMCA4b interaction seems to be mutually influential, such that knockdown of PMCA decreases GPER-mediated ERK1/2 phosphorylation, while GPER knockdown does the opposite on PMCA activity (135).

Estrogenic Regulation of Calcium Signal Transduction—The Calmodulin Network

Since CaM is the universal Ca²⁺ signal transducer for numerous proteins (117, 118), is insufficiently expressed for its targets (122, 125, 126), and is a source of competition among target proteins (124, 127), factors that regulate its expression and target interactions are predicted to have a pervasive impact. The effects of E₂ on the CaM network have been examined in some detail in vascular ECs in recent studies (135, 169, 170). E₂ treatment (1–5 × 10⁻⁹ M, 48 h) upregulates total CaM by around 7-fold and free Ca²⁺-CaM by ~15-fold in primary ECs. Data obtained using specific estrogen receptor agonists, gene silencing, and receptor overexpression indicate that GPER, but not ERα or ERβ, mediates this effect. Thus, the GPER agonist G-1 (10⁻⁹-10⁻⁷ M), but not the ERα agonist propyl pyrazole triol (PPT) (3 × 10⁻¹⁰-2 × 10⁻⁷ M) or the ERβ agonist diarylpropionitrile (DPN) (10⁻¹⁰-5 × 10⁻⁸ M), increases CaM expression; GPER knockdown reduces the effect of E₂ to upregulate CaM; and E₂ upregulates CaM in SKBR3 cells that express only GPER and not ERα or ERβ (170). Consistently, the ERα/ERβ antagonist/GPER agonist ICI182,780 dose-dependently upregulates CaM. Mechanistically, GPER exerts this action via the activities of EGFR and MAPK/ERK kinase 1 (MEK1). Functionally, E₂ upregulates CaM and promotes the PMCA–CaM interaction; however, the predicted stimulatory effect on Ca²⁺ extrusion is masked by E₂-induced inhibitory phosphorylation at Tyr1176 of PMCA (170); additionally, GPER exerts E₂-dependent and E₂-independent effects to inhibit PMCA (135). These collective actions prolong Ca²⁺ signals, promote Ca²⁺-CaM complex formation, and increase Ca²⁺-CaM associations with low- to high-affinity CaM network members, represented by GPER itself, ERα, and eNOS (170). Considering that CaM binding stabilizes ERα homodimers, these effects are expected to promote other genomic actions of E₂ as well. Thus, a feedforward mechanism exists in which GPER mediates E₂'s effects to increase CaM and inhibits Ca²⁺ efflux, prolonging cytoplasmic Ca²⁺ signals, and the resultant increases in Ca²⁺-CaM complexes in turn promote the activities of GPER itself and other CaM network members (170) (Figure 1).

Estrogenic Moderation of Calcium-Dependent Activities

How do the various mechanisms discussed so far come together in regulating cardiovascular functions? An immediate challenge is how to reconcile the effects of estrogenic agonists to both trigger acute Ca²⁺ signals by themselves and inhibit otherwise stimulated Ca²⁺ signals. The Ca²⁺ signals triggered by estrogenic agonists in primary cardiovascular cells are generally of very low amplitude. Furthermore, as in experiments testing their effects on Ca²⁺ signals otherwise triggered, estrogenic agonists are present in situ with other stimuli whose Ca²⁺ signals they inhibit. Thus, for mechanisms that generate cytoplasmic Ca²⁺ signals, E₂ and GPER exert ultimate inhibitory effects. For cytoplasmic Ca²⁺ removal mechanisms, estrogenic agonists and GPER also are inhibitory. For Ca²⁺ signal transduction, E₂, via a feedforward at GPER, increases CaM expression and enhances linkage in the CaM-binding proteome.

All things considered, E₂ and GPER, via both E₂-dependent and E₂-independent mechanisms, act to moderate Ca²⁺-dependent activities in the cardiovascular system. They “clamp” cytoplasmic Ca²⁺ signals by lowering peaks (inhibition of signal generation) and raising troughs (inhibition of signal removal), collectively confining tissues in a narrower yet more sustained operating range of Ca²⁺. Also, GPER-mediated increases in CaM expression and CaM network linkage improve Ca²⁺ signal transduction efficiency. Considering the Ca²⁺ sensitivity of Ca²⁺-dependent proteins in this context, one can predict that those with low Ca²⁺ sensitivity (requiring high Ca²⁺ for activation) are more likely to be affected by the inhibition of Ca²⁺ signal generation. On the other hand, proteins with high Ca²⁺ sensitivity (requiring low Ca²⁺ for activation) are more likely to be promoted by the inhibition of Ca²⁺ removal and less affected by the suppression of Ca²⁺ signal generation (Figure 2).

Figure 2.

Moderation of cardiovascular functions by E₂ and GPER via effects on Ca²⁺ signal generation, Ca²⁺ signal removal, and Ca²⁺ signal transduction. See text for details. Modified with permission from the author's previous publication (170).

This notion has been demonstrated experimentally via the case of eNOS, a Ca²⁺-dependent CaM-binding protein (222) with sub-nanomolar affinity for CaM (127). CaM interaction and subsequent activation of wild-type eNOS have high Ca²⁺ sensitivities, with respective EC₅₀(Ca²⁺) values ~1.8 × 10⁻⁷ and 4 × 10⁻⁷ M (190). eNOS is also regulated by multisite phosphorylation (223). Notably, its bi-phosphorylation at Ser617 and Ser1179 promotes NO production by increasing the Ca²⁺ sensitivity for both CaM binding and enzyme activation, reducing their respective EC₅₀ (Ca²⁺) values to ~0.7 × 10⁻⁷ and 1.3 × 10⁻⁷ M, thus rendering the synthase active at resting cytoplasmic Ca²⁺ (189). E₂ and GPER (1) prolong endothelial cytoplasmic Ca²⁺ signal by inhibiting Ca²⁺ efflux (135, 170), (2) promote eNOS phosphorylation at Ser617 and Ser1179 (170, 198), (3) increase CaM expression and eNOS–CaM interaction (170), and (4) suppress endothelial SOCE (203). When we incorporate these effects into a verified sequential “CaM binding eNOS activation” model (189, 190), eNOS activity and NO accumulation are shown to substantially increase across the time course of bradykinin-induced Ca²⁺ signal in ECs by treatment with G-1 (203). Importantly, major contributions to this outcome include the increases in CaM binding, phosphorylation, Ca²⁺ sensitivity, and duration of Ca²⁺ signals due to Ca²⁺ efflux inhibition, but little or no effect of the inhibition of SOCE (203), due obviously to the synthase's high Ca²⁺ sensitivity (Figure 3). Thus, via multifaceted actions on components of the CSM, E₂ and GPER moderate Ca²⁺-dependent activities by differentially affecting the continuum of Ca²⁺-dependent proteins based on their Ca²⁺ sensitivities for Ca²⁺ or Ca²⁺-CaM complexes.

Figure 3.

Moderation of Ca²⁺-dependent eNOS activity by GPER activation. (A) Cytoplasmic Ca²⁺ clamping by GPER activation in ECs (203). The solid line represents Ca²⁺ signals produced in response to agonist stimulation in the absence of GPER activation. The sparsely dotted area represents the range of cytoplasmic Ca²⁺ signals, in which peak and trough are seen due to maximal effects of Ca²⁺ entry and Ca²⁺ efflux. The stippled blue line represents Ca²⁺ signals produced in the presence of GPER and its activation. These signals are clamped in a narrower range (the blue area) due to inhibitory effects on both SOCE [green stripes (203)] and PMCA4b-mediated Ca²⁺ efflux [red stripes (135, 170)]. (B) Average time courses of cytoplasmic Ca²⁺ signals measured in primary ECs treated with bradykinin in the absence of extracellular Ca²⁺ followed by treatment with vehicle or G-1; total Ca²⁺ signals were triggered by re-addition of extracellular Ca²⁺ [arrow (203)]. (C) Calculated eNOS point activity corresponding to each Ca²⁺ value in (B) considering only changes in Ca²⁺ due to GPER activation using a verified sequential eNOS–CaM binding eNOS activation model [equation, where (K₁, K₂) and (K₃, K₄) are derived products of the binding constants of Ca²⁺ at the Ca²⁺-binding sites on the N and C lobes of CaM in binding to CaM and interaction of Ca²⁺-CaM and eNOS (189, 190). (D) Calculated eNOS point activity corresponding to each Ca²⁺ value measured in (B), factoring in changes in Ca²⁺, CaM binding, and eNOS phosphorylation (170, 203). See details in text and (170, 203). Reproduced with permission from the author's previous publication (203).

Considering the two Ca²⁺-dependent estrogen receptors—ERα and GPER—how does the presence of one influence the effects of the other on Ca²⁺ signaling? A complex relationship is predicted to exist in which ERα transcriptional activities affect the expression of certain Ca²⁺ signaling proteins but are themselves influenced by the amplitudes and dynamics of Ca²⁺ signals limited by GPER activation and the availability of CaM that is promoted by GPER action (170). In turn, as CaM is limited in cells (122, 124, 126, 127), the high affinity binding of CaM by ERα and GPER further limits CaM availability and will influence CaM-dependent regulation of each other at the receptor level, a predictable outcome of the functional crosstalk via competition for limited CaM (124, 127). These relationships may represent but a small aspect of the reciprocality between estrogen and Ca²⁺ signaling.

Conclusion and Future Perspectives

Reciprocality between estrogen signaling and Ca²⁺-dependent activities is becoming evident. Considering the impact of estrogen and its receptors on Ca²⁺ signaling, E₂, and in many cases, GPER exert inhibitory effects on many components of the CSM in cardiovascular tissues, from Ca²⁺ store release and uptake (214, 215, 221) and Ca²⁺ entry (199, 201–210, 212, 213) to cytosolic Ca²⁺ removal mechanisms (135, 170, 208, 217–221). Considering the impact of Ca²⁺ signaling on estrogen biology, both ERα and GPER are strongly regulated by direct Ca²⁺-dependent interactions with CaM. These interactions serve to stabilize receptor dimerization and enhance subsequent transcriptional activities [the case of ERα (137, 138, 142, 143)] or promote receptor-mediated downstream signaling [the case of GPER (169, 170)]. Also, E₂-induced MAPK activation has long been known to be dependent on the Ca²⁺ signal produced (173). Reciprocality between estrogen biology and Ca²⁺ signaling is further evidenced by the demonstration of a feedforward mechanism, in which E₂, via GPER activation, upregulates total cellular CaM expression and free intracellular Ca²⁺-CaM concentration, which promotes functions of GPER and ERα and other classes of Ca²⁺-CaM-dependent proteins (170). The combination of these various actions is predicted to affect Ca²⁺-dependent functions depending on the affinity and Ca²⁺ sensitivities of the proteins involved, as exemplified by the case of eNOS (Figures 2, 3) (170, 203).

The moderating effects that estrogenic agonists and receptors exert on the CSM can explain many of their cardiovascular effects, such as preventing excessive cardiac contraction during sympathetic stress, limiting adverse outcomes related to Ca²⁺ overload, and reducing vascular tone. Nevertheless, the effects of E₂ and estrogen receptors on many CSM components have not been examined. Additionally, many questions remain regarding mechanisms of the observed effects that estrogenic agonist and receptors produce on the CSM. For example, how do E₂ and GPER inhibit I_Ca,L? What are the mechanisms that position GPER as an intrinsic component of β₁AR signaling in the myocardium? What are the mechanisms whereby E₂ inhibits the activities of SERCA and NCX? What are the mechanisms whereby E₂ inhibits mitochondrial Ca²⁺ uptake? Further studies are needed to answer these questions. Through many examples, however, it is clear that GPER produces both E₂-dependent and E₂-independent effects on the CSM. While the search is ongoing for approaches to apply specific estrogen receptor agonists to the prevention of cardiovascular disease, the therapeutic potential of E₂-independent effects of GPER and other estrogen receptors is as yet an unexplored territory.

Author Contributions

Q-KT conceived the ideas, generated the figures, and wrote the manuscript.

Conflict of Interest

The author declares that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Glossary

Abbreviations

AF domain

transcriptional activation function domain

CaM

calmodulin

Ca²⁺-CaM

Ca²⁺-bound calmodulin

cAMP

cyclic adenosine monophosphate

CICR

Ca²⁺-induced Ca²⁺ release

CRAC

Ca²⁺ release-activated channels

CSM

Ca²⁺ signaling machinery

E₂

17β-estradiol

ECs

endothelial cells

EGFR

epidermal growth factor receptor

eNOS

endothelial nitric oxide synthase

ERβ

estrogen receptor β

ERα

estrogen receptor α

ERK1/2

extracellular signal-related kinases 1 and 2

FRET

fluorescence resonance energy transfer

GPER

G protein-coupled estrogen receptor 1

GPR30

G protein-coupled estrogen receptor 1

HEK293 cells

human embryonic kidney 293 cells

I_{Ca, L}

L-type Ca²⁺ channel current

IP₃Rs

inositol-trisphosphate receptors

LTCC

L-type Ca²⁺ channels

LV

left ventricle

MAPK

mitogen-activated protein kinases

mCRC

mitochondrial Ca²⁺ retention capacity

MCU

mitochondrial Ca²⁺ uniporter

MEK1

MAP (mitogen-activated protein) kinase/ERK (extracellular signal-regulated kinase) kinase 1

mPTP

mitochondrial permeability transition pore

NCX

Na⁺-Ca²⁺ exchanger

OVX

ovariectomy/ovariectomized

PDZ

PSD-95/Dlg/ZO

PKC

protein kinase C

PLCβ

phospholipase C-β

PMCA

plasma membrane Ca²⁺-ATPase

PSD-95

post-synaptic density protein 95

RMP

resting membrane potential

RyRs

ryanodine receptors

SCPA

secretory pathway Ca²⁺ pump

SERCA

sarcoplasmic/endoplasmic reticulum-ATPase

SMD

submembrane domains of G protein-coupled receptors

SOCE

store-operated Ca²⁺ entry

SOICR

store overload-induced Ca²⁺ release

SR/ER

sarco-endoplasmic reticulum

STIM1

stromal interaction molecule 1

VDAC

voltage-dependent anion channel

VDCC

voltage-dependent Ca²⁺ channels

VDCE

voltage-dependent Ca²⁺ entry

VSMCs

vascular smooth muscle cells.