A selective membrane estrogen receptor agonist maintains autonomic functions in hypoestrogenic states

Abstract

It is well known that many of the actions of estrogens in the central nervous system are mediated via intracellular receptor/transcription factors that interact with steroid response elements on target genes. But there is also a compelling evidence for the involvement of membrane estrogen receptors in hypothalamic and other CNS functions. However, it is not well understood how estrogens signal via membrane receptors, and how these signals impact not only membrane excitability but also gene transcription in neurons. Indeed, it has been known for sometime that estrogens can rapidly alter neuronal activity within seconds, indicating that some cellular effects can occur via membrane delimited events. In addition, estrogens can affect second messenger systems including calcium mobilization and a plethora of kinases within neurons to alter cellular functions. Therefore, this brief review will summarize our current understanding of rapid membrane-initiated and intracellular signaling by estrogens in the hypothalamus, the nature of receptors involved and how these receptors contribute to maintenance of homeostatic functions, many of which go awry in menopausal states.

1. Membrane-initiated signaling of estrogens

It has been known for a number of years that 17β-estradiol (E2) has acute, membrane-initiated signaling actions in the brain (Kelly and Rønnekleiv, 2002; Rønnekleiv and Kelly, 2005; Micevych and Dominguez, 2009). A decade ago the nature and physiological significance of these actions were a matter of debate, but it is now widely accepted that some of the actions of E2 are quite rapid and cannot be attributed to the classical nuclear-initiated steroid signaling of ERα or ERβ. Importantly, ERα and ERβ can associate with signaling complexes in the plasma membrane (Razandi et al., 1999; Boulware et al., 2005; Pedram et al., 2006; Szegõ et al., 2006; Dewing et al., 2007; Bondar et al., 2009). In addition, many of the rapid effects of E2 can be induced by selective ERα or ERβ ligands, antagonized by the ER antagonist ICI 182,780 or are absent in animals bearing mutations in ERα and/or ERβ genes (Couse and Korach, 1999; Singer et al., 1999; Dubal et al., 2001; Wade et al., 2001; Abraham et al., 2003; Boulware et al., 2005, 2007). However, it is also evident that E2 can activate bona fide G-protein coupled receptors (GPCRs), the most notable GPR30 and a Gαq-coupled membrane ER (Gu et al., 1999; Toran-Allerand, 2004, 2005; Qiu et al., 2003, 2006; Noel et al., 2009; Zhang et al., 2010; Kenealy et al., 2011).

Substantial evidence has been generated in the support of a novel Gαq-coupled membrane ER (Gαq-mER). Intracellular sharp electrode and whole cell patch recording from guinea pig and mouse hypothalamic slices have been used to characterize this Gαq-mER (Lagrange et al., 1997; Qiu et al., 2003, 2006). These hypothalamic slice studies established that E2 acts rapidly and stereospecifically within physiologically-relevant concentrations (EC₅₀=7.5 nM) to significantly reduce the potency of μ-opioid and GABA_B agonists (i.e., desensitize) to activate an inwardly rectifying K⁺ conductance (Lagrange et al., 1997; Qiu et al., 2003). Estrogenic desensitization of μ-opioid and GABA_B receptors is mimicked either by stimulation of adenylyl cyclase with forskolin or by direct protein kinase A (PKA) activation with the non-hydrolyzable cAMP analog Sp-cAMP, in a concentration-dependent manner (Lagrange et al., 1997; Qiu et al., 2003). Furthermore, the selective PKA antagonists KT5720 and Rp-cAMP block the effects of E2. As one would predict from the extensive literature on desensitization of GPCRs (Gainetdinov et al., 2004), PKA is downstream in a signaling cascade that is initiated by a Gαq-coupled membrane ER that is linked to activation of phospholipase C (PLC)–protein kinase C (PKC)–PKA (Qiu et al., 2003, 2006). Importantly, E2 does not alter the affinity of the μ-opioid and GABA_B receptors for their respective receptors (Cunningham et al., 1998). Furthermore, the ER antagonists ICI 164,384 and ICI 182,780 block the actions of E2 with subnanomolar affinity (K_i=0.5 nM) that is similar to K_i for antagonism of ERα (Weatherill et al., 1988; Lagrange et al., 1997). These pharmacological findings clearly argue for a novel G-protein-coupled membrane receptor with high selectivity for E2.

In view of the differences between circulating levels of E2 and working concentrations of E2 used for in vitro analysis, it is important to clarify the pharmacological analysis of in vitro E2 physiological responses: When applying E2 in a bath superfusing hypothalamic slices, the physiological actions depend on the pharmacokinetics of E2 in the slice as it does for all other tissues. Therefore, it is important to do a dose–response to establish the potency and efficacy of E2. The potency (EC₅₀) of an agonist is the concentration required to produce 50% of the maximum effect in an experimental preparation. The value is obtained from a mathematical curve fitted to experimental data points. The potency (EC₅₀) is dependent on the binding affinity, the efficacy of agonist, the receptor reserve in the tissue or cell, and the ability of the agonist (i.e., E2) to penetrate to the site of action (Furchgott, 1978). As discussed above, the EC₅₀ value for E2 to rapidly attenuate (desensitize) the μ-opioid response is 7.5 nM (Lagrange et al., 1997), whereas the EC₅₀ for E2 to augment the K_ATP channel activity in GnRH neurons is an order of magnitude lower (0.6 nM), which is probably reflective of the receptor reserve in GnRH neurons versus POMC neurons (Zhang et al., 2010). Most importantly, critical requirement for establishing a specific receptor-mediated response is the blockade by a selective antagonist. Indeed, selective ER antagonists block the actions of E2 in POMC neurons and GnRH neurons with subnanomolar affinity (K_i=0.5 nM) similar to K_i for inhibition of E2 binding to ERα (Weatherill et al., 1988; Lagrange et al., 1997). Therefore, based on all of the criteria for establishing a physiological response, these rapid actions of E2 are physiological. The importance of this rapid response in physiological systems is discussed below.

About a decade ago a diphenylacrylamide compound, STX, that does not bind ERα or ERβ (Qiu et al., 2003, 2006) was developed to selectively target the Gαq-mER and its down-stream signaling cascade – phospholipase Cβ-protein kinase Cδ-protein kinase A pathway – that mediates μ-opioid and GABA_B desensitization in hypothalamic neurons. The design arose out of studies in which we found that E2 stereospecifically (17α-estradiol is not active) activates the Gαq-mER signaling pathway (see above), and these actions were blocked by the ER antagonist ICI 182,780 (Lagrange et al., 1997; Qiu et al., 2003, 2006). The results from these physiological and pharmacological experiments led to the design of STX, which is structurally similar to 4-OH tamoxifen, to target the Gαq-mER signaling pathway (Qiu et al., 2003). As predicted, STX had greater affinity (~20-fold) for the Gαq-mER than E2 and does not bind to ERα or ERβ (Qiu et al., 2006; Tobias et al., 2006). Most importantly, both STX and E2 activated this Gαq signaling pathway in mice lacking both ERα and ERβ and in GPR30-knockout mice (Qiu et al., 2006, 2008). Definitive characterization (i.e., cloning) of this novel Gαq-mER is currently a work in progress.

Parallel studies initiated in hippocampal slices some 20 years ago showed that E2 enhanced N-methyl-D-aspartate (NMDA)-mediated excitatory postsynaptic potentials (EPSPs) and long-term potentiation (LTP) following Schaffer (collateral) fiber stimulation (Wong and Moss, 1992; Foy et al., 1999; Rudick and Woolley, 2003). Also, E2 potentiated non-NMDA (kainate)-mediated excitation of hippocampal CA1 pyramidal neurons via activation of a cAMP/PKA pathway (Gu and Moss, 1996, 1998). Importantly, these rapid actions of E2 on glutamate excitation of hippocampal neurons were still present in animals deficient in ERα (from Dr. Ken Korach, NIH), suggesting a novel mechanism (receptor) for the rapid actions of E2 in the hippocampus (Gu et al., 1999). In addition E2 and E2-BSA (E2 conjugated to bovine serum albumin to limit membrane penetration) applied acutely to the hippocampus in ovariectomized animals produced a sustained reduction of the slow after hyperpolarization current (I_AHP) in CA1 pyramidal neurons (Carrer et al., 2003). This provided further evidence for the involvement of a membrane ER, although, the mechanism by which E2 regulates Ca²⁺ influx into CA1 neurons is currently unknown. More recent studies from Catherine Woolley’s laboratory have shown that E2 via ERβ also has presynaptic effects to enhance glutamate release (increased vesicle release probability) and hence, excitation of CA1 neurons (Smejkalova and Woolley, 2010). In addition, E2 through the association of ERα with the metabotropic glutamate receptor 1 (mGluR1) attenuates GABA-A mediated inhibitory postsynaptic currents (IPSCs) (Huang and Woolley, 2012). The ERα-mediated effects are via mGluR1-dependent mobilization of the endocannabinoid anandamide that retrogradely suppresses GABA from CB-1 receptor-containing presynaptic GABAergic boutons. Interestingly, this signaling pathway exists in females but not males, which is congruent with the findings from Boulware et al. (2005, 2007)) who showed in female and not male hippocampal CA3–CA1 neuronal cultures that E2 rapidly stimulates MAPK-dependent cAMP-responsive element binding protein (CREB) phosphorylation. This modulation of CREB activity also occurs via ERα interactions with mGluR 1. The protein–protein interaction between the “classical” ERs and mGluRs has not been fully elucidated, but ERα and GluR1a have been co-immunoprecipitated in membrane fractions from both astrocytes and neurons (Dewing et al., 2007; Kuo et al., 2009; Dominguez and Micevych, 2010).

Recently, the orphan GPCR GPR 30 has received notoriety because of its role in mediating E2’s effects in the CNS (Noel et al., 2009; Lebesgue et al., 2010; Kenealy et al., 2011). GPR30 exhibits the signaling characteristics of a bonafide ER (Thomas et al., 2005; Prossnitz et al., 2008). In breast cancer cells that are transfected with GPR30, E2 activates the mitogen-activated protein kinases (MAPK), ERK1 and ERK2, and these actions are independent of ERα or ERβ (Filardo, 2002; Filardo and Thomas, 2005). Accordingly, E2 activates Gβγ-subunits that promote the release and activation of an epidermal growth factor precursor (proHB-EGF). The active HB-EGF binds to the EGF receptor (ErbB) to facilitate receptor dimerization and downstream activation of ERK (Filardo et al., 2000, 2002; Filardo, 2002). Interestingly, the selective estrogen receptor modulator (SERM) tamoxifen and ER antagonist ICI 182,780 both promote GPR30-dependent transactivation of the EGF receptor and subsequent MAPK activation. It is important to note that the Gαq-mER-mediated response to E2 in arcuate neurons is still present in GPR30 KO mice (Qiu et al., 2008). GPR30 has been localized in the brain and E2 binds to this receptor (Funakoshi et al., 2006; Bologa et al., 2006; Brailoiu et al., 2007; Prossnitz et al., 2007), and recent studies from the Terasawa lab have shown that GPR30 is involved in mediating the rapid actions of E2 in monkey olfactory placode GnRH neurons (see below). Therefore, it is evident that estrogen can rapidly alter cell function through ERα, ERβ and/or novel GPCRs that bind E2.

2. Membrane-initiated signaling of E2 and hypothalamic control of autonomic functions

Besides its quintessential role in the feedback control of the reproductive axis, E2 modulates a number of hypothalamic-regulated autonomic functions, most notably energy homeostasis and temperature. E2 signaling via ERα is a critical component in the hypothalamic regulation of energy balance (Geary et al., 2001). In rodents, hypo-estrogenic states are clearly associated with decreased activity and an increase in body weight (Czaja and Goy, 1975; Butera and Czaja, 1984; Czaja, 1984; McCaffrey and Czaja, 1989; Jones et al., 2000; Asarian and Geary, 2002; Qiu et al., 2006; Clegg et al., 2006, 2007). In humans, a loss-of-function mutation in ERα has a clear metabolic phenotype in man with expression of type 2 diabetes, hyperinsulinemia and obesity (Smith et al., 1994). However, global reinstatement of an ERα that is lacking the ERE targeting domain is sufficient for “rescuing” the metabolic deficits in mice (Park et al., 2011). These findings suggest an important role for non-ERE mediated E2 signaling. Moreover, brain-specific knockout of ERα causes hyperphagia and hypometabolism (Musatov et al., 2007; Xu et al., 2011), and selective knockdown of ERα in POMC neurons appears to recapitulate the hyperphagic phenotype in female mice (Xu et al., 2011). However, there are at least two caveats that impact the interpretation of gene deletion experiments. Firstly, ERα is a transcription factor affecting the expression of hundreds of genes important for cell signaling, and many of these genes are essential for mER initiated responses that contribute to POMC excitability and hence control of energy homeostasis (Hewitt et al., 2003; Malyala et al., 2004). Secondly, POMC-Cre (mice utilized in Xu et al. (2011)) is also expressed in progenitor neurons that are destined to become NPY neurons and perhaps other hypothalamic and extrahypothalamic neurons (Padilla et al., 2010, 2012). Therefore, ERα knockout spreads well beyond the single neuron phenotype as originally proposed. Hence, one must be cautious in interpreting ERα knockout (global or targeted) experiments.

Experiments dating back three decades determined that the anorectic effects of E2 in rodents are mediated through CNS sites of action since direct injections of E2 into the paraventricular nucleus of the hypothalamus (PVH) or arcuate/ventromedial nucleus are effective to reduce food intake, body weight and increase wheel running activity in females (Colvin and Sawyer, 1969; Ahdieh and Wade, 1982; Butera and Czaja, 1984). This is due, in part, to the actions of E2 on arcuate POMC and neuropeptide Y (NPY) neurons, the former being anorexic and the latter being orexigenic (Roepke, 2009). Thus, a number of experiments have shown that E2 upregulates the expression of the peptide β-endorphin in POMC neurons in ovariectomized female guinea pigs and increases the mRNA expression of POMC in both mice and guinea pigs (Thornton et al., 1994; Bethea et al., 1995; Pelletier et al., 2007; Roepke et al., 2008). Furthermore, there is a decrease in hypothalamic POMC mRNA levels in postmenopausal women (Abel and Rance, 1999). In addition, E2 reverses the ovariectomy-induced increase in arcuate NPY protein and mRNA expression in rodents (Crowley et al., 1985; Shimizu et al., 1996; Pelletier et al., 2007). Therefore, it appears that the arcuate nucleus and particularly POMC and NPY neurons are major targets for the anorectic actions of estrogen, which underscores their importance in the control of energy homeostasis. In addition, POMC neurons are critical for the regulation of feeding behavior and are also involved in the rewarding aspects of food intake (Hayward et al., 2002; Appleyard et al., 2003).

2.1. Energy homeostasis and obesity

The ability of STX to robustly mimic the effects of E2 on POMC neuronal activity led to the hypothesis that the Gαq-mER may have a role in the control of energy homeostasis. Indeed, in translational animal experiments we have demonstrated that peripheral administration of STX mimics the effects of E2 in controlling energy homeostasis (Qiu et al., 2006; Roepke et al., 2008, 2010). As predicted from the E2- or STX-induced increase in POMC neuronal activity, both E2 and STX reduce food intake and, subsequently, the post-ovariectomy body weight gain. STX and E2 inhibit food intake in ovariectomized guinea pigs by reducing meal frequency, and there is a subsequent reduction in abdominal fat accumulation (Roepke et al., 2010). In support of a hypothalamic site of action, treatment with STX, similar to E2, induces new gene transcription in the arcuate nucleus (Roepke et al., 2008). Many of the regulated genes in the arcuate nucleus are involved in the control of neuronal excitability (e.g., Cav3.1) and intracellular signaling in POMC neurons (Roepke et al., 2008). For example, the PI3 kinase regulatory subunits are regulated by E2 and STX: PI3 kinase p55γ mRNA is increased by E2 treatment (Malyala et al., 2008) and PI3 kinase p85α mRNA is upregulated by STX (Roepke et al., 2008). Therefore, Gαq-mER appears to function in the estrogenic control of energy homeostasis through activation of POMC neurons in the arcuate nucleus, although other hypothalamic neurons may be involved via synaptic input from POMC neurons and/or via direct actions of E2.

2.2. Core body temperature and hot flashes

Another critical homeostatic function modulated by circulating estrogens is the maintenance of core body temperature (Tc). In fact, hot flashes affect 75–85% of perimenopausal and postmenopausal women (Moline et al., 2003). Hot flashes are characterized by periods of sweating and peripheral vasodilation and are often associated with increased environmental temperature (Rapkin, 2007). Therefore, a hot flash can be defined as an exaggerated heat dissipation response initiated by the preoptic temperature sensitive neurons. Although the mechanism behind this response is not known, repeated observations have found that the majority of hot flashes are preceded by elevation in Tc independent of peripheral vasoconstriction or elevated metabolic rate (Freedman et al., 1995; Freedman, 2005). Therefore, it has been postulated that elevated Tc may serve as one trigger of menopausal hot flashes (Freedman and Blacker, 2002; Rapkin, 2007). In general, there is compelling experimental evidence that the incidence of hot flashes in hypo-estrogenic females is decreased by E2 treatment (Tankersley et al., 1992; Brooks et al., 1997; Freedman and Blacker, 2002). The estrogenic reduction in the naloxone-induced rise in tail skin temperature of morphine-dependent ovariectomized rats has been the industry standard for a hot flash model (Simpkins et al., 1983; Merchenthaler et al., 1998; Komm et al., 2005). The morphine-dependence, naloxone-precipitated withdrawal rat model has a number of drawbacks because of the adverse autonomic reactions independent of an elevation in core and skin temperature. But the model does highlight the involvement of CNS opioid effects (e.g., from POMC neurons) directly or indirectly on preoptic temperature sensitive neurons (Yoshida et al., 2009).

We have established a guinea pig “hot flash” model based on the hypothesis that the expression of vasomotor symptoms in menopausal women is due to the reduced thermoneutral zone in core body temperature (Freedman, 2005). In the ovariectomized female guinea pig, both E2 and STX significantly reduce Tc compared to animals receiving vehicle injections (Roepke et al., 2010). The exact cellular target for E2 and specifically the Gαq-mER modulation of Tc has not been identified. However, one potential cellular mechanism is the direct action of E2 via the Gαq-mER on thermosensitive (GABAergic) neurons in the medial preoptic area of the hypothalamus (Griffin et al., 2001; Nakamura et al., 2002; Boulant, 2006). Previous studies in ovariectomized female guinea pigs have shown that medial preoptic GABAergic neurons respond to acute E2 treatment via the Gαq-mER signaling pathway to attenuate GABA_B inhibitory input, leading to increased GABA neuronal activity (Wagner et al., 2001). Moreover, activation of medial preoptic GABA neurons are responsible for evoking the vasomotor responses in rodents (Nakamura and Morrison, 2010) underlying heat dissipation responses (i.e., vasodilatation, sweating) as seen in women experiencing hot flashes.

Importantly, the coupling of Gαq-mER to downstream signaling pathways is similar to the serotonin 5HT2A/2C receptors, which when activated, lower Tc and are implicated in thermoregulation dysfunction caused by ovariectomy (Berendsen et al., 2001; Sipe et al., 2004). Selective serotonin reuptake inhibitors elevate endogenous serotonin levels and are therefore efficacious for treating hot flashes (Stearns et al., 2002) and can significantly attenuate the effects of ovariectomy on thermoregulation in rodents (Deecher et al., 2007). Moreover, both the Gαq-mER and 5HT2C receptors increase POMC neuronal excitability (Qiu et al., 2007). These similarities imply that serotonin via its Gαq-coupled receptor and E2 via Gαq-mER have similar cellular targets in the hypothalamic neurons that regulate Tc and energy homeostasis.

2.3. Bone remodeling and osteoporosis

Recently, STX has been found to mimic the ability of E2 to maintain bone density in the ovariectomized female guinea pigs (Roepke et al., 2010). Although E2 has direct effects on the osteoclast/osteoblast cells involved in bone remodeling (Sylvia et al., 2002; Endoh et al., 1997), E2 can reduce bone loss, a hallmark of hypo-estrogenic states, in part, by controlling the hypothalamic regulation of bone remodeling. It is known that the preautonomic paraventricular (PVN) neurons that drive sympathetic activity are involved in bone remodeling (Yadav et al., 2009; Takeda et al., 2002). These PVN neurons receive a robust μ-opioid receptor inhibitory input, presumably from POMC neurons (Wuarin and Dudek, 1990), and therefore, E2 via Gαq-mER can affect their excitability via inhibitory POMC input. Indeed, we have found that STX, like E2, preserves cancellous bone density, which supports a potential role of this hypothalamic signaling pathway in maintenance of bone density. However, STX, like E2, may also have peripheral effects (e.g., directly on osteoblasts) which cannot be ruled out at this time. Certainly, the decreased PVN output would dampen the sympathetic nervous system activity that controls bone remodeling via the adrenergic β2 receptor activity in the osteoblast cells (Takeda et al., 2002). In fact, there is an increase in bone formation and a decrease in bone reabsorption in adrenergic β2 receptor knockout mice (Elefteriou et al., 2005). Unlike in wild-type mice, gonadectomy of adrenergic β2 receptor knockout mice does not alter bone mass or bone resorption parameters, indicating that increased sympathetic activity may be responsible for the bone loss in hypo-estrogenic states such as menopause (Elefteriou et al., 2005).

3. Summary

It is obvious from the plethora of studies using membrane-delimited E2 ligands and the mER selective ligand STX that “genomic” actions of E2 in the brain do not require the direct nuclear targeting of estrogen receptors (ERα and ERβ). Signals that are initiated by E2 at the plasma membrane can trigger multiple intracellular signaling cascades including activation of MAPK, PI3K, and PKC pathways (Watters et al., 1997; Bi et al., 2001; Cato et al., 2002; Yang et al., 2003; Deisseroth et al., 2003) that result in the phosphorylation of hundreds of proteins that ultimate can affect cell excitability and gene transcription. E2 can bind to the novel Gαq-mER to upregulate cAMP in hypothalamic neurons by increasing adenylyl cyclase activity within minutes (Lagrange et al., 1997). Cyclic-AMP activates PKA, which in turn phosphorylates K⁺ channels (Zhang et al., 2010) and/or calcium channels (Zhang et al., 2009) to alter their activity. In addition, PKA can phosphorylate CREB to elicit new gene transcription (Zhou et al., 1996; Gu et al., 1996; Watters and Dorsa, 1998; Abrahám et al., 2004). Therefore, not only does membrane excitability change within a matter of seconds, but gene transcription can be activated within a relatively short time course (within tens of minutes) in neurons independent of classic estrogen receptors (ERα and ERβ) interacting with EREs. Thus, we need to elucidate the role of membrane estrogen receptors not only in hypothalamic homeostatic functions but also in higher cortical functions. By developing selective non-steroidal agonists for targeting these mERs we would be able to expand the therapeutic window for treating menopausal symptoms associated with aging without concern for the deleterious side effects of the gonadal steroids.