Plasma membrane affiliated AMPA GluA1 in estrogen receptor β-containing paraventricular hypothalamic neurons increases following hypertension in a mouse model of post-menopause

Abstract

Sex and ovarian function contribute to hypertension susceptibility, however, the mechanisms are not well understood. Prior studies show that estrogens and neurogenic factors, including hypothalamic glutamatergic NMDA receptor plasticity, play significant roles in rodent hypertension. Here, we investigated the role of sex and ovarian failure on AMPA receptor plasticity in estrogen-sensitive paraventricular nucleus (PVN) neurons in naïve and angiotensin II (AngII) infused male and female mice and female mice at early and late stages of accelerated ovarian failure (AOF). High-resolution electron microscopy was used to assess the subcellular distribution of AMPA GluA1 in age-matched male and female estrogen receptor beta (ERβ) enhanced green fluorescent protein (EGFP) reporter mice as well as female EGFP-ERβ mice treated with 4-vinylcyclohexene diepoxide. In the absence of AngII, female mice at a late stage of AOF displayed higher levels of GluA1 on the plasma membrane, indicative of functional protein, in ERβ-expressing PVN dendrites when compared to male, naïve female and early stage AOF mice. Following slow-pressor AngII infusion, males, as well as early and late stage AOF females had elevated blood pressure. Significantly, only late stage-AOF female mice infused with AngII had an increase in GluA1 near the plasma membrane in dendrites of ERβ-expressing PVN neurons. In contrast, prior studies reported that plasmalemmal NMDA GluN1 increased in ERβ-expressing PVN dendrites in males and early, but not late stage AOF females. Together, these findings reveal that early and late stage AOF female mice display unique molecular signatures of long-lasting synaptic strength prior to, and following hypertension.

Graphical Abstract

INTRODUCTION

Up to middle age, hypertension is more prevalent in men compared to women, however, as women enter menopause, they become increasingly susceptible to hypertension (Martins et al., 2001; Maranon and Reckelhoff, 2013; Van Kempen et al., 2016). The mechanisms subserving gender differences in hypertension susceptibility are uncertain, although neurogenic processes are strongly implicated (reviewed in (Hay et al., 2014)).

Sex differences in hypertension susceptibility are also found in small rodents (Xue et al., 2007; Xue et al., 2014). For example, in a common model of hypertension induced by systemic administration of a subpressor dose of angiotensin II (AngII) young male mice, but not young female mice, show a slowly developing (i.e., slow-pressor) increase in blood pressure and sympathetic tone (Xue et al., 2005; Girouard et al., 2009; Tiwari et al., 2009; Xue et al., 2013b; Marques-Lopes et al., 2015). Differences in estrogen signaling in males and females may contribute to the sex-dependent divergence in the hypertensive response to AngII. This is suggested by evidence that the estrous acyclicity and/or low ovarian hormone levels seen in aged (Fortepiani et al., 2003; Tiwari et al., 2009; Marques-Lopes et al., 2015) or ovariectomized rodents (Xue et al., 2013a; Hay et al., 2014), respectively, increases hypertension susceptibility to slow-pressor AngII-infusion. Additionally, our recent studies (Van Kempen et al., 2016; Marques-Lopes et al., 2017) utilizing the mouse model of accelerated ovarian failure (AOF) have begun to elucidate the mechanisms of hypertension at different stages of ovarian decline. Induction of AOF by systemic injection of 4-vinylcyclohexane diepoxide (VCD), produces an early stage of ovarian failure (i.e. peri-AOF) marked by erratic and extended hormone cycles as well as a later stage (post-AOF) characterized by undetectable serum levels of estrogen and acyclicity (Van Kempen et al., 2014). This feature of the VCD AOF model provides experimental models of peri- and post-menopausal stages (Mayer et al., 2004; Lohff et al., 2005; Van Kempen et al., 2011; Van Kempen et al., 2014; Marques-Lopes, 2018). We have shown that susceptibility to slow-pressor AngII begins at peri-AOF and extends into post-AOF and that the magnitude of elevated blood pressure in both is comparable to that seen in male mice (Marques-Lopes et al., 2017).

The hypothalamic paraventricular nucleus (PVN) may significantly contribute to the emergence of hypertension in AOF females. The PVN is a critical neural coordinator of brain cardiovascular circuits involved in sympathetic and neuroendocrine systems pivotal for blood pressure regulation (Ferguson et al., 2008; Braga et al., 2011; Li and Pan, 2017; Basting et al., 2018; Dampney et al., 2018). Signaling via estrogen receptors (ER) α and β, estrogens have been shown to contribute to blood pressure regulation by acting in the PVN and other neural components of cardiovascular regulatory circuitry (Van Kempen, 2016). Studies from our lab and others implicate neurons containing ERβ, but not ERα, within the PVN as a significant target of estrogen’s actions on blood pressure regulation (Ogawa et al., 2000; Xue et al., 2013a; Marques-Lopes et al., 2014a; Marques-Lopes et al., 2017). In the rodent PVN, ERβ is contained in pre-sympathetic neurons that project to the rostral ventrolateral medulla and intermediolateral spinal cord (Stern and Zhang, 2003; Bingham et al., 2006; Marques-Lopes et al., 2014a). Therefore, PVN ERβ-containing neurons are positioned to regulate sympathoexcitation both indirectly and directly (Benarroch, 2005).

In several models of hypertension, increases in glutamatergic excitatory input to PVN pre-sympathetic neurons result in elevations of sympathetic outflow (reviewed in (Li and Pan, 2017; Zhou et al., 2019)). Many of glutamate’s key actions on neural signaling and plasticity are mediated by the NMDA receptor (reviewed in (Hunt and Castillo, 2012)), a heteromer formed by the obligatory GluN1 subunit along with GluN2 subunits (Dingledine et al., 1999). The PVN is populated by neurons that express NMDA receptor subunit genes (Herman et al., 2000; Eyigor et al., 2001) and proteins (Petralia et al., 1994). Neurophysiological, immunocytochemical, and spatiotemporal gene silencing studies demonstrate that NMDA receptors within the PVN contribute to AngII-mediated hypertension (Coleman et al., 2013; Wang et al., 2013; Glass et al., 2015; Ma et al., 2018).

In PVN pre-sympathetic neurons, NMDA receptor plasma membrane trafficking, a process critical for the development of neural plasticity (Hunt and Castillo, 2012), is important for the development of hypertension (Girouard et al., 2009; Glass et al., 2015). Our electron microscopic studies demonstrate that peri-AOF hypertension is associated with a profile of NMDA receptor redistribution in ERβ-containing neurons in the PVN that differs from males and post-AOF females (Marques-Lopes et al., 2014a; Marques-Lopes et al., 2017). In particular, following slow-pressor AngII-induced hypertension, the essential NMDA receptor GluN1 subunit is elevated on and near the plasma membrane of ERβ-containing PVN dendrites in peri-AOF in a manner that is similar to males. However, unlike males, total GluN1 in ERβ-containing PVN dendrites is not elevated in peri-AOF females (Marques-Lopes et al., 2014a; Marques-Lopes et al., 2017). Moreover, the increase in total GluN1 in ERβ-containing PVN dendrites in post-AOF females more closely resembles the pattern of hypertension-induced redistribution of GluN1 seen in males than peri-AOF females (Marques-Lopes et al., 2017). Importantly, GluN1 is decreased on the plasma membrane as well as in the cytoplasm of ERβ-containing PVN dendrites in non-hypertensive young females. Together these findings demonstrate that the increase in blood pressure seen in hypertensive males and peri-AOF females, but not post-AOF females, directly correlates with the upregulation of GluN1 affiliated with the plasma membrane on ERβ-containing PVN neurons. Significantly, these results point to distinct molecular substrates underlying hypertension in peri- and post-AOF mice.

In addition to post-synaptic NMDA receptors, post-junctional AMPA receptors in PVN pre-sympathetic neurons are also altered in hypertension (reviewed in (Li and Pan, 2017)). The AMPA-type glutamate receptors mediate much of the excitatory postsynaptic currents in glutamatergic synapses [reviewed in (Sheng and Kim, 2002)]. AMPA receptors are multiunit proteins consisting of various combinations of GluA1-4 subunits that express distinct biophysical properties. Receptors expressing GluA1 are among the most abundant, play an important role in constitutive excitatory neurotransmission, and traffic to the plasma membrane in an experience-dependent manner. The AMPA GluA1 subunit is found throughout the PVN (Eyigor et al., 2001). Significantly, GluA1 expressing AMPA receptor activity contributes to hyperactivity of PVN-spinal projecting neurons in spontaneously hypertensive rats (SHR) (Li et al., 2012). In other cardiovascular brain regions, both the SHR and DOCA-salt hypertension models are associated with an increase of GluA1 in dendritic spines suggesting that this AMPA receptor subunit adapts to hypertension (Aicher et al., 2003; Hermes et al., 2008). However, the relationship between GluA1 localization in ERβ-containing PVN neurons during hypertension in gonadally intact male and female mice, as well as peri- versus post-AOF mice is unknown. Thus, this study utilized the same mice used in our prior studies (Marques-Lopes et al., 2014a; Marques-Lopes et al., 2017) to examine the effect of slow-pressor AngII on the subcellular distribution of GluA1 between ERβ-containing PVN neurons in peri- and post-AOF female mice compared to young (non-AOF) female and male mice.

EXPERIMENTAL PROCEDURES

Animals:

Experimental procedures were approved by the Institutional Animal Care and Use Committee of Weill Cornell Medicine and were in accordance with the 2011 Eighth Edition of the National Institute of Health Guide for the Care and Use of Laboratory Animals. Female and male bacterial artificial chromosome (BAC) ERβ-enhanced green fluorescent protein (ERβ–EGFP) mice on a C57BL/6 background (N = 30) were used for these studies (Milner et al., 2010). ERβ-EGFP mice were initially developed by the GENSAT project (www.gensat.org) at the Rockefeller University (Gong et al., 2003). The ERβ reporter mice identify neurons with ERβ in either nuclear or extranuclear locations (Milner et al., 2010). Mice were housed 3-4 per cage, kept on a 12:12 light/dark cycle and had ad libitum access to food and water. Four groups of ERβ-GFP mice (6 mice per group) were used for the anatomical experiments: 1) 2-month old male mice, 2) 2-month old female mice, 3) ~4-month old peri-AOF female mice and 4) ~6.5-month old post-AOF female mice. The mice from groups 1 and 2 were the same as those used in our study by (Marques-Lopes et al., 2014a), whereas the mice from groups 3 and 4 were the same as those used in our study by (Marques-Lopes et al., 2017).

AOF procedure:

To induce AOF, gonadally-intact 50-55 postnatal day old female mice were injected with VCD (130 mg/kg i.p.) in vehicle (sesame oil) for 5 sequential days per week for three weeks as described in our prior study (Van Kempen et al., 2014). Control mice were injected with oil (vehicle) only. This VCD injection regimen selectively eliminates primary follicles in the ovary and results in ovarian failure and undetectable levels of estradiol (Mayer et al., 2004; Lohff et al., 2005). Prior studies showed that this VCD regimen does not negatively affect peripheral tissues, including organ weights, liver, and kidney function (Mayer et al., 2005; Haas et al., 2007; Sahambi et al., 2008; Wright et al., 2008). Moreover, VCD at this dose does not have any direct inflammatory effects (as measured by immunolabeling for the microglial marker Iba1) on the brain regions outside (e.g., area postrema and subfornical organ) and inside (e.g., hippocampus and PVN) the blood brain barrier (Van Kempen et al., 2014).

Prior studies have previously determined time-periods for peri-AOF and post-AOF mice that have endpoints corresponding hormonal profiles to those seen in peri- and post-menopausal stages in humans (Mayer et al., 2004; Lohff et al., 2005; Van Kempen et al., 2014). At the peri-AOF timepoint (58 days after initiating the VCD injections), mice have irregular and extended estrous cycles and increased plasma follicle stimulating hormone (FSH) (Mayer et al., 2004; Lohff et al., 2005; Harsh et al., 2007). At the post-AOF timepoint (129 days after initiating the VCD injections), mice show an acyclic, anestrous status, in which persistent estrus is observed and ovulation has ceased (Van Kempen et al., 2014). Post-AOF mice have undetectable serum levels of estrogen, decreased levels of progesterone, and elevated serum levels of FSH, lutenizing hormone and androstenedione (reviewed in (Van Kempen et al., 2011)).

Estrous cycle determination:

Estrous cycle assessment using vaginal smear cytology (Turner and Bagnara, 1971) was performed 8-10 days prior to osmotic minipump insertion (below). Estrous cycles in young mice are 4-5 days long and have three primary phases: proestrus (high estrogen levels; 0.5 - 1 day long), estrus (declining estrogen levels and rising progesterone levels; 2 - 2.5 days long) and diestrus (low estrogen and progesterone levels; 2 - 2.5 days long). Male mice were removed from their cage and handled daily. Prior to osmotic mini-pump implantation, young female mice and control female mice for VCD experiments had normal estrous cycles. In contrast, peri-AOF mice had irregular estrous cycles and post-AOF mice were acyclic before the minipump implantation (Supplementary Tables 1 and 2 in (Marques-Lopes et al., 2017)).

Slow-pressor AngII administration:

To minimize stress, the same experimenter (JM-L) handled the mice at the same time of day throughout the study. Osmotic mini-pumps (Alzet, Durect Corporation, Cupertino, CA) containing AngII (600 ng · kg–1 · min–1) in vehicle (saline+0.1% bovine serum albumin (BSA)) or vehicle alone were implanted subcutaneously under isofluorane anesthesia. Tail-cuff plethysmography (Model MC4000; Hatteras Instruments, Cary, NC) was used to measure systolic blood pressure (SBP) the day before mini-pump implantation (baseline), and 2, 5, 9 and 13 days after mini-pump implantation in awake mice, as previously described (Marques-Lopes et al., 2014a; Marques-Lopes et al., 2017). Mice were euthanized 14 days after mini-pump implantation (i.e., one day after the final SBP measurements). A timeline of the experiments in the AOF mice is shown in Supplementary Figure 1 in (Marques-Lopes et al., 2017).

The young female mice used in this study were in estrus or diestrus on the day of euthanasia. Peri-AOF VCD mice had equal proportions of estrus and metestrus / diestrus phases in the saline and AngII infused groups. Post-AOF VCD mice were in metestrus / diestrus on the day of euthanasia (Marques-Lopes et al., 2017).

Immunocytochemical procedures

Antibodies:

A chicken polyclonal anti-GFP antibody (GFP-1020; RRID: AB_10000240; Aves Lab Inc., San Diego, CA) generated against recombinant GFP (Encinas et al., 2006) was used. On Western blots this antibody recognizes one major band at ~27 kD and immunolabels cells in brain sections from transgenic mice expressing EGFP (manufacturer’s data sheet, www.aveslab.com). Also, immunolabeling for this antibody is absent in brain sections from mice not expressing EGFP (Volkmann et al., 2010; Milner et al., 2011). Immunolabeling of ERβ protein as well as mRNA for ERβ is co-localized in ERβ-EGFP neurons (Milner et al., 2010; Oyola et al., 2017).

A polyclonal rabbit antibody raised against the cytoplasmic domain of GluA1 (AB_1504, Millipore Sigma, Billerica MA) was used in this study. On Western blots this antibody recognizes one major band at about 106 kD on 10ug of mouse brain lysate [manufacturer’s data sheet, www.emdmillipore.com and (Yokoi et al., 2016)]. Moreover, immunoprecipitation, Western blotting and electron microscopy of anti-GluA1 in rat hippocampus demonstrate the expected locations in pre- and post-synaptic profiles (Hussain et al., 2015).

Tissue preparation:

Mouse brains were prepared and processed for pre-embedding dual immunolabeling electron microscopic studies according to previously described methods (Milner, 2011). For this, mice were deeply anesthetized with sodium pentobarbital (150 mg/kg, i.p.) and their brains fixed by aortic arch perfusion consecutively with ~5 ml 2% heparin in normal (0.9%) saline followed by 30 ml of 3.75% acrolein and 2% paraformaldehyde (PFA) in 0.1M phosphate buffer (PB; pH 7.4). After perfusion, the brains were removed from the skull and post-fixed in 2% acrolein and 2% PFA in PB at room temperature for 30 min. Brains were cut into 5-mm coronal blocks using a brain mold (Activational Systems, Inc., Warren, MI) and sectioned (40 μm thick) on a VT1000X Vibratome (Leica Microsystems, Buffalo Grove, IL). Brain sections were placed in cryoprotectant (30% sucrose, 30% ethylene glycol in PB) and then kept at −20°C until the immunocytochemistry experiment.

To ensure that sections from each experimental group were processed with identical immunolabeling conditions (Pierce et al., 1999), sections were marked in the cortex with punch codes and pooled into single containers. The region of the PVN (0.70-0.94 mm caudal to bregma (Hof et al., 2000) were chosen from two sections per animal (3 animals per group).

Dual label electron microscopic immunocytochemistry:

Free-floating tissue sections were treated with 1% sodium borohydride in PB for 30 min to neutralize reactive aldehydes followed by 8-10 rinses in PB. The sections then were placed in 0.5% bovine serum albumin (BSA) in 0.1 M Tris-buffered saline (TS) for 30 min. The tissue then was incubated in a solution of anti-GluA1 (1:150) antiserum in 0.1% BSA in 0.1 M TS for 1 day at room temperature. The antiserum to GFP (1:2500) was added to the primary antibody diluent and the tissue was moved to 4°C and incubated for an additional day.

For GFP immunoperoxidase labeling, sections were incubated for 30 min in goat anti-chicken IgG (1:400; Jackson ImmunoResearch Inc., West Grove, PA), followed by a 30 min incubation in avidin–biotin complex at half the manufacturer’s recommended dilution (Vector Laboratories, Burlingame, CA). The bound peroxidase in the sections was visualized by reaction for 6 min in 3,3′-diaminobenzidine (DAB; Sigma-Aldrich Chemical Co., Milwaukee, MI) and hydrogen peroxide. The tissues were rinsed in TS between all steps.

For GluA1 silver-intensified immunogold (SIG) labeling, the sections were rinsed in PB, and incubated overnight in a 1:50 dilution of donkey anti-mouse IgG with bound 1 nm colloidal gold (Electron Microscopy Sciences (EMS), Fort Washington, PA) in 0.01% gelatin and 0.08% BSA in 0.01M phosphate-buffered saline (PBS). After rinsing in PBS, the tissue was placed in 2% glutaraldehyde in PBS for 10 minutes, rinsed in PBS and then placed in 0.2 M sodium citrate buffer (pH 7.4). Bound gold particles were enhanced using a Silver IntenSEM kit (RPN491; GE Healthcare, Waukeska, WI) for 7 min.

Dual labeled sections were post-fixed in 2% osmium tetroxide for 1 hr, dehydrated through a graded series of ethanols and propylene oxide, and then embedded in EMbed-812 (EMS) between two sheets of Aclar plastic. Ultrathin sections (70 nm thick) from the PVN were cut with a Diatome diamond knife (EMS) using a Leica EM UC6 ultratome. The sections were collected on 400-mesh thin-bar copper grids (EMS) and counterstained with uranyl acetate and Reynold’s lead citrate.

Ultrastructural data analyses:

A person blind to experimental condition conducted all data collection and analyses. In each PVN block, 50 dual-labeled dendritic profiles were randomly selected and photographed using a Tecnai transmission electron microscope (Tecnai 12 BioTwin, FEI, Hillsboro, OR). Prior studies have determined that 50 dendrites per animal were sufficient to make quantitative comparisons on the subcellular distribution of proteins (Znamensky et al., 2003). Dendrites were identified using standard criteria (Peters et al., 1991). Fields selected for analysis were adjacent to the tissue-plastic interface (i.e., the surface of the tissue) to control for differences in antibody penetration (Milner, 2011). Dual labeled dendrites were usually post-synaptic to axon terminal profiles and contained regular microtubular arrays and mitochondria. Immunoperoxidase labeling for EGFP was distinguished as a diffuse precipitous electron-dense reaction product, whereas SIG labeling of GluA1 appeared as punctate black electron-dense particles.

The subcellular distribution of GluA1-SIG particles in ERβ-EGFP-labeled dendrites was determined according to previously described quantitative methods (Coleman et al., 2013; Marques-Lopes et al., 2017). For this, Microcomputer Imaging Device software (MCID, Imaging Research Inc., Ontario, Canada) was used to assess perimeter, cross-sectional diameter and major and minor axis lengths of each dual labeled dendrite. Dendrites that lacked a distinguishable plasma membrane were excluded from the analysis. The distribution of GluA1-SIG was analyzed with the following parameters: (1) the density of GluA1-SIG particles on or near (GluA1-SIG particles located within 70nm of the plasmalemma) the plasma membrane (PM / μm; referred to as plasmalemmal), (2) the density of GluA1-SIG particles localized to the cytoplasm per cross-sectional area of the dendrite (CY / μm²), and (3) the total number of GluA1-SIG particles (sum of plasmalemmal and cytoplasmic) per dendritic cross sectional area (Total / μm²). Based on average diameter, dendrites were further divided in large (> 1.0 μm) and small (≤ 1.0 μm), which correspond with proximal and distal to the cell body (Peters et al., 1991).

Figure preparation:

Image adjustments were made to the entire image. For light microscope photomicrographs, adjustments to brightness, sharpness and contrast were made in Microsoft Powerpoint 2010. For electron micrographs, resolution was increased to 400 dpi and then images were adjusted for levels and sharpness using unsharp mask in Adobe Photoshop 9.0 (RRID:SCR_014199). Electron micrographs then were imported into Microsoft Powerpoint 2010, for additional changes to brightness, contrast, and sharpness. These latter adjustments were made to achieve uniformity in the appearance between electron micrographs. Graphs were made using Prism 7 software (Graphpad Prism, RRID:SCR_002798).

Statistical analysis:

Data are expressed as means ± SEM. Significance was set to an alpha < 0.05. Statistical analyses for electron microscopic data was conducted on JMP 12 Pro software (JMP, RRID:SCR_014242). Two-way analysis of variance (ANOVA), followed by a Tukey's HSD post-hoc analyses, was used for comparisons between groups (young males, non-AOF females, peri-AOF females, and post-AOF females) and treatment (Sal vs. AngII) in the EM dual-labeling studies. A Student’s t-test was used to analyze the electrophysiology data.

RESULTS

In post-AOF vehicle-infused female mice, GluA1-SIG particles are differentially distributed in ERβ-EGFP dendrites compared to the other three groups

There were no differences in systolic blood pressure between any of the four groups (males, and non-AOF, peri-AOF, and post-AOF females) when administered vehicle at any time point (Fig. 1). To assess the pattern of GluA1 localization in the absence of AngII administration, a quantitative analysis was performed to examine the density of GluA1 within subcellular compartments (i.e., SIG particle density) of ERβ-EGFP dendrites in vehicle-infused mice only. ERβ-EGFP dendrites were primarily sampled from the dorsal parvocellular region of the caudal PVN (Fig. 2), which harbors the majority of autonomic neurons (Biag et al., 2012; Oyola et al., 2017). Representative micrographs showing GluA1-SIG labeling in ERβ-EGFP dendrites for all eight groups are shown in Figure 3A-H. The ERβ-EGFP containing dendrites lacked spines and were usually contacted by one or more unlabeled terminals (examples Fig. 3A, E).

Fig. 1. Slow-pressor AngII increases blood pressure in males, and peri- and post-AOF mice.

Systolic Blood Pressure (SBP) was not different in vehicle-infused mice from any of the four groups. SBP significantly increases in young males and peri- and post-AOF females following slow-pressor AngII infusion. SBP measurements did not significantly differ between saline and AngII infusion in non-AOF females. **p < 0.01; ***p < 0.001. (Data from (Marques-Lopes et al., 2014b; Marques-Lopes et al., 2017))

Fig 2. Representative light micrograph showing ERβ-EGFP labeled neurons in the PVN.

The mediocaudal region of the PVN, which contains the majority of autonomic neurons (Biag et al., 2012; Oyola et al., 2017) was sampled for electron microscopy. dp, dorsal parvicellular; mpv, mediocentral parvicellular; pv, periventricular. Bar = 50 μm.

Fig. 3. Representative electron micrographs of GluA1-SIG particles in ERβ-EGFP dendrites in the PVN from saline- and AngII-infused mice.

A-F. Electron micrographs of GluA1-SIG particles in ERβ-EGFP dendrites of a saline-infused male (A), an AngII-infused male (B), a saline-infused non-AOF female (C), an AngII-infused non-AOF female (D), a saline-infused peri-AOF female (E), an AngII-infused peri-AOF female (F), a saline-infused post-AOF female (G) and an AngII-infused post-AOF female (H). Examples of GluA1-SIG particles on the plasma membrane (chevron), near plasma membrane (arrowhead) and cytoplasm (arrow) are shown. Bars = 500nm.

Quantitative analysis revealed a significant main effect of condition (male, female non-AOF, female peri-AOF and female post-AOF) on plasma membrane [F(3,604) = 3.85, p = 0.0095], cytoplasmic [F(3,604) = 10.51, p < 0.0001] and total [F(3,604) = 7.1867, p < 0.0001] GluA1-SIG particles in ERβ-EGFP dendrites. Post-hoc comparisons showed that the density of plasma membrane GluA1-SIG particles in ERβ-GFP dendrites in the post-AOF vehicle-treated females was significantly greater than seen in vehicle-infused males (p = 0.0038) and non-AOF vehicle-treated females (0.0058; Fig. 4). Further, post-AOF females had significantly fewer cytoplasmic GluA1-SIG particles in ERβ-GFP dendrites compared to males (p < 0.0001), non-AOF (p < 0.0001) and peri-AOF (p = 0.0061) female mice. In addition, vehicle-infused males had significantly more cytoplasmic GluA1-SIG particles in ERβ-EGFP dendrites compared to peri-AOF females (p = 0.0126). The post-AOF females had significantly less total GluA1-SIG particles in ERβ-EGFP dendrites compared to the male (p < 0.0001), non-AOF (p = 0.0003) and peri-AOF (p = 0.0008) females.

Fig. 4. Both sex and hormonal status alter the distribution of GluA1-SIG particles in ERβ-EGFP dendrites in saline-infused mice.

Following saline-infusion, males, non-AOF females and peri-AOF females have significantly more (*** p < 0.001) GluA1-SIG particles in the cytoplasm and in total in ERβ-EGFP dendrites than post-AOF female mice. Males also have significantly greater (* p < 0.05) cytoplasmic GluA1-SIG particles than peri-AOF female mice. N = 3 mice per group; n = 50 dendrites per mouse.

Thus, in the absence of AngII, post-AOF mice had a higher level of plasma membrane GluA1 compared to males, non-AOF females and peri-AOF females. In addition, post-AOF females also exhibited lower levels of cytoplasmic and total GluA1 in ERß-expressing dendrites of PVN neurons compared the other three groups.

GluA1 density in ERβ-EGFP dendrites is differentially altered in females following AngII depending on hormonal status

We further examined the effect of condition and treatment on blood pressure and the subcellular distribution of GluA1 in PVN ERβ-EGFP dendrites. There were significant differences in systolic blood pressure between the different treatment groups. Whereas non-AOF mice infused with AngII did not differ from those administered vehicle, peri-AOF, post-AOF, and male mice all showed comparable increases in systolic blood pressure following AngII with respect to vehicle (Fig. 1).

There were a main effects of condition [F(3,1205) = 2.74; p = 0.042] and treatment [F(1,1205) = 4.87; p = 0.028], in addition to a condition by treatment interaction [F(3,1205) = 4.08; p = 0.007] on GluA1-SIG particle localization in ERβ-labeled dendrites when examined without consideration of size (Fig. 5A). Tukey’s post-hoc analysis revealed significantly greater GluA1-SIG density on and near the plasma membrane in post-AOF mice infused with AngII compared to their vehicle controls (p = 0.019). When the ERβ-EGFP dendrites were separated into large (i.e., proximal) and small (i.e. distal) categories based on cross-sectional area, a difference in GluA1 density emerged in large (Fig. 5B), but not small (Fig. 5C) dendrites. There were main effects of treatment [F(1,350) = 12.36; p = 0.0005] and condition [F(3,350) = 5.05; p = 0.002] with respect to the plasmalemmal density of GluA1-SIG particles in large ERβ-EGFP dendrites. Post-hoc analysis showed that among all groups only AngII-infused post-AOF females had a significantly greater GluA1-SIG density on and near the plasma membrane in large ERβ-EGFP dendrites compared to their vehicle-infused counterparts (p = 0.002; Fig. 5B). No differences in GluA1 density were seen in small ERβ-EGFP dendrites across any of the treatment groups (Fig. 5C).

Fig. 5. Redistribution of GluA1 on and near the plasma membrane of ERβ-EGFP dendrites following AngII infusion.

A. Following AngII infusion, on and near plasma membrane GluA1-SIG particle density in small ERβ-EGFP dendrites is significantly increased in post-AOF females (p < 0.05). AngII-infused post-AOF has a greater density of GluA1-SIG on and near the PM compared to AngII-infused peri-AOF mice (^ap = 0.002). B. In large dendrites, a similar pattern is observed. AngII-infused post-AOF mice have a greater plasma membrane associated GluA1 density compared to controls (p < 0.01). Densities are greater in post-AOF AngII-infused mice compared to AngII-infused males (^ap = 0.01) and peri-AOF females (^bp = 0.006). C. No differences in on and near plasmalemma density were observed in small dendrites. N = 3 mice per group; n = 50 dendrites per mouse.

There were also significant main effects of condition [F(3, 1205) = 18.73; p <0.0001] and treatment [F(1, 1205) = 6.678; p = 0.009], as well as a condition by treatment interaction [F(3, 1205) = 5.3869; p = 0.0011] with regard to cytoplasmic GluA1-SIG particle density in ERβ-GFP dendrites. There was a significantly greater density of GluA1-SIG particles in the cytoplasm of ERβ-EGFP dendrites of AngII-infused non-AOF females compared to those infused with vehicle (p = 0.0068; Fig. 6A). Additionally, AngII-infused non-AOF females had a greater cytoplasmic GluA1 particle density in small ERβ-EGFP dendrites compared to AngII-infused peri-AOF (p = 0.0025) and post-AOF females (p < 0.0001). Finally, AngII-infused males showed a greater cytoplasmic GluA1-SIG particle density in ERβ-EGFP dendrites compared to AngII-infused peri-AOF females (p = 0.0001) and post-AOF females (0.0007; Fig 6A). When separated by size, small dendrites showed a greater density of cytoplasmic GluA1-SIG particles in ERβ-EGFP dendrites from AngII-infused non-AOF females compared to peri- (p = 0.0001) and post-AOF (p = 0.0005; Fig 6B) females infused with AngII.

Fig. 6. Redistribution of GluA1 in the cytoplasm in ERβ-EGFP dendrites following AngII infusion.

A. Following AngII infusion, cytoplasmic GluA1-SIG particles significantly increase in ERβ-EGFP dendrites (all sizes) in non-AOF (young) females relative to saline-infused non-AOF females (**p < 0.01). AngII-infused males show significantly greater GluA1-SIG density in the cytoplasm compared to AngII-infused peri-AOF females (^ap = 0.0007). B. In small ERβ-EGFP dendrites, cytoplasmic GluA1-SIG particles were significantly elevated in AngII-infused non-AOF females compared to AngII-infused peri-AOF (^ap = 0.0001) and post-AOF females (^bp = 0.0005). N = 3 mice per group; n = 50 dendrites per mouse.

There were also between-group differences in the total density of GluA1-SIG in ERβ-EGFP dendrites. There were significant main effects of both condition [F(3,1205) = 12.14; p <0.0001] and treatment [F(1,1205) = 12.91; p = 0.0003], as well as a condition by treatment interaction [F(3,1205) = 10.831; p < 0.0001] with regard to total GluA1-SIG labeling. The total GluA1-SIG particle density in ERβ-EGFP dendrites in AngII-infused non-AOF females was significantly greater compared to those infused with vehicle (p = 0.0006; Fig. 6C). AngII-infused post-AOF females also showed a significantly greater total GluA1-SIG density in ERβ-EGFP dendrites compared to vehicle-treated post-AOF females (p = 0.0002). Finally, AngII-administered peri-AOF females had a significantly lower total GluA1-SIG density in ERβ-GFP dendrites compared to AngII-treated non-AOF (p < 0.0001) and post-AOF (p = 0.0021) females (Fig. 7).

Fig. 7. Changes in total GluA1 levels in ERβ-EGFP dendrites following AngII infusion.

Total GluA1-SIG particles significantly increased in ERβ-EGFP dendrites of all sizes in AngII-infused non-AOF female mice (***p < 0.001) and post-AOF female mice (***p < 0.001) compared to their saline-infused counterparts. Total GluA1-SIG particle density was significantly lower in ERβ-EGFP dendrites in AngII-infused peri-AOF mice compared to AngII-infused non-AOF mice (^ap < 0.0001) and post-AOF mice (^bp = 0.0021). N = 3 mice per group; n = 50 dendrites per mouse.

The diverse changes in GluA1-SIG density in ERβ-EGFP dendrites following AngII administration demonstrate a significant effect of sex, hormonal status and hypertensive state on GluA1 receptor trafficking. Specifically, the heightened GluA1 density on and near the plasma membrane in hypertensive AngII post-AOF females is suggestive of a potentiation in AMPA GluA1 receptor mediated excitatory signaling in ERβ-containing dendrites following AngII hypertension.

Distal ERβ-GFP dendrites are larger in male and post-AOF mice compared to non-AOF and peri-AOF females

Morphological measurements (i.e., cross-sectional area and maximum diameter) were compared in ERβ-EGFP dendrites dually labeled for GluA1 in all the groups of mice. There was a main effect of condition [F(3, 1205) = 7.195; p <0.0001] and a treatment by condition interaction [F(3, 1205) = 3.221; p =0.022] with respect to dendritic area. Dual-labeled dendrites in AngII-treated in males had a significantly larger area when compared to those of AngII-infused non-AOF females (p = 0.0004; Fig 8A).

Fig. 8. The size of dendrites dually labeled for GluA1 and ERβ-EGFP differs depending on sex, hormonal status and treatment.

A. Following AngII infusion, the area of GluA1/ERβ-EGFP dendrites is significantly greater in males compared to non-AOF females (***p < 0.001). B. The area of small GluA1/ERβ-EGFP dendrites in males is greater compared to both non-AOF and peri-AOF females (***p < 0.001). Likewise, the area of GluA1/ERβ-EGFP small dendrites in post-AOF females is greater than non-AOF females (^a,bp < 0.05). Additionally, the area of small GluA1/ERβ-EGFP dendrites in AngII-infused post-AOF females is greater compared to AngII-infused peri-AOF females (**p < 0.01). N = 3 mice per group; n = 50 dendrites per mouse.

When dual labeled dendrites were separated by size additional differences in area were seen. There was a significant main effect of condition [F(3,847) = 28.7864; p < 0.0001] on dendritic area in small dual labeled dendrites. Regardless of treatment (i.e., vehicle vs AngII) males had significantly larger dendrites in the small category than vehicle non-AOF (p < 0.0001), AngII-infused non-AOF (p < 0.0001), vehicle peri-AOF (p = 0.0001), and AngII-infused peri-AOF females (p < 0.0001). Interestingly, the size of vehicle post-AOF small dendrites was no different from that of AngII-infused males (vehicle: p = 0.666; AngII: p = 0.984). Further, AngII-infused post-AOF females had significantly larger dendrites in the small category compared to AngII-infused peri-AOF females (p = 0.0015; Fig 8B).

These findings suggest that the size of distal ERβ-EGFP dendrites containing GluA1 is reduced in mice with low estrogen states (males and post-AOF) compared to higher estrogen states (non-AOF and peri-AOF).

DISCUSSION

This study provides the first evidence that sex, hormonal status and AngII-mediated slow-onset hypertension impact the subcellular location of the key neuroplasticity-associated AMPA GluA1 protein in estrogen-sensitive neurons in the PVN. Specifically, we demonstrate that in the absence of AngII, post-AOF females compared to young males, non-AOF females and peri-AOF females have a higher density of plasma membrane associated GluA1 in PVN ERβ-containing dendrites. These findings indicate that GluA1 expressing AMPA receptors are strategically positioned to enhance glutamate signaling and glutamate-dependent synaptic plasticity even in the absence of a hypertensive stimulus in post-AOF mice (Fig. 9). Following slow-pressor AngII infusion, males and both peri- and post- AOF females have elevated blood pressure. Significantly, we find that following slow-pressor AngII administration GluA1 traffics towards the plasma membrane of ERβ-expressing PVN neurons in hypertensive post-AOF females only (Fig. 9). Interestingly, non-AOF females do not develop hypertension following AngII, and also show an increase in cytoplasmic, not plasma membrane GluA1 in ERβ-expressing dendrites of PVN neurons. These findings contrast with our previous studies showing that plasmalemmal associated NMDA GluN1 increases in ERβ-expressing PVN dendrites in males and peri- AOF females but not post-AOF females (Marques-Lopes et al., 2014b; Marques-Lopes et al., 2017). Thus, it appears that prior to and following hypertension, distinct patterns of glutamate receptor localization and plasticity in estrogen-sensitive PVN neurons distinguish male and female mice, as well as female mice at post-AOF, peri-AOF or with normal ovarian function.

Fig. 9. Schematic diagram summarizing the effects of AngII-infusion on the subcellular density distribution of GluA1 in ERβ-EGFP dendrites in the PVN.

Slow pressor AngII elevates SBP in males as well as peri- and post-AOF females but not non-AOF (young) females (Marques-Lopes et al., 2014a; Marques-Lopes et al., 2017). Non-AOF female mice have regular 4-5 day estrous cycles whereas peri-AF females have erratic estrous cycles. Post-AOF female mice are acyclic and have negligible estrogen levels [reviewed in (Marques-Lopes, 2018)]. Baseline (saline): Males and non-AOF females have identical subcellular density distributions of GluA1-SIG particles in ERβ-EGFP dendrites. Compared to males and non-AOF females, peri-AOF females and post-AOF females have an elevated plasma membrane GluA1-SIG particle density and a lesser cytoplasmic density in ERβ-EGFP dendrites. Slow-pressor AngII: In hypertensive males and peri-AOF females, the subcellular density distribution of GluA1-SIG particles in ERβ-EGFP dendrites is similar to their saline counterparts. Moreover, the size of GluA1/ERβ-EGFP dendrites is reduced in hypertensive AngII peri-AOF females. In non-hypertensive non-AOF females, cytoplasmic GluA1-SIG particles in ERβ-EGFP dendrites is elevated compared to saline-infused non-AOF female mice. In hypertensive post-AOF females near plasma membrane GluA1-SIG particles in ERβ-EGFP dendrites are elevated compared to saline-infused post-AOF females. Moreover, AMPA currents are elevated in hypertensive AngII post-AOF females compared to saline post-AOF females and this elevation is blocked by an ERβ agonist.

Heightened plasma membrane GluA1 in dendrites of ERβ-expressing PVN neurons selectively in post-AOF females given vehicle

Even in the absence of AngII, post-AOF mice show a divergence in the subcellular distribution of GluA1 in dendrites of ERβ-expressing PVN neurons when compared to male mice or female mice differing in hormonal status. We report that the subcellular distribution of GluA1-SIG particles in ERβ-EGFP dendrites is similar in saline-infused young male mice, young non-AOF females mice and peri-AOF female mice. However, at the post-AOF time-point GluA1-SIG particles are elevated on the plasma membrane of ERβ-EGFP dendrites of post-AOF mice (Fig. 9). These findings indicate that more GluA1-containing AMPA receptors in ERβ-containing dendrites are available for ligand (glutamate) binding in post-AOF mice compared to the other conditions (Boudin et al., 1998; Ladepeche et al., 2014).

The higher levels of plasma membrane GluA1 in PVN ERβ-EGFP dendrites in post-AOF mice may have important functional consequences. The elevation or reduction of excitatory activity results in increases or decreases, respectively, of AMPA receptors at synapses (Lissin et al., 1998; O'Brien et al., 1998). About 10-20% of AMPA receptors are recycled from the surface within 10 min of synaptic activity (Ehlers, 2000; Lin et al., 2000). Rapid insertion of AMPA receptors, thought to be composed primarily of GluA1/2 subunits, in the plasma membrane increases synaptic transmission in long-term potentiation (LTP) an established model of synaptic plasticity [reviewed in (Turrigiano and Nelson, 2004)]. Thus, the presence of more GluA1 on and near the plasma membrane of ERβ-EGFP dendrites in post-AOF mice may predispose estrogen-sensitive neurons to enhanced excitatory signaling and facilitate neural plasticity during AngII-infusion.

Select increase in GluA1 levels and cytoplasmic accumulation in dendrites of PVN ERβ-expressing neurons in non-AOF females following AngII

Non-AOF female mice infused with AngII did not develop hypertension, a finding consistent with previous reports. However, these mice did show an increase in cytoplasmic GluA1 in ERβ-EGFP dendrites compared to vehicle-treated animals (Fig. 9). In our previous study (Marques-Lopes et al., 2014b) it was found that both cytoplasmic and plasma membrane-associated GluN1 decreased in ERβ-EGFP dendrites in non-hypertensive AngII-infused non-AOF females. Together, these findings would be consistent with diminished synaptic strength for promoting plasticity processes (Fernandez-Monreal et al., 2012) in the PVN of non-hypertensive non-AOF females following AngII.

In addition to the increased cytoplasmic GluA1 accumulation, the size of GluA1 and ERß-EGFP co-expressing dendrites as measured by cross-sectional area decreased in non-AOF females following AngII infusion. This decrease in area was more robust in small dendrites, which generally are contacted by greater numbers of excitatory-type axon terminals (Froemke et al., 2005; Lovett-Barron et al., 2014). The decrease in dendritic size could indicate a decrease in receptor recycling or a disruption in membrane endocytosis leading to loss of the plasma membrane (Hanley, 2008). However, it is also possible that AngII infusion may alter the subcellular distribution of GluA1 in other phenotypes (i.e. non-ERβ expressing) of PVN neurons. Regardless, the decreased dendritic size together with the increased localization of GluA1 to the cytoplasm in ERβ PVN dendrites is consistent with decreased sensitivity of ERβ-containing neurons to glutamate in young non-AOF females who do not develop hypertension in response to administration of AngII.

In contrast to intact female mice, male mice experienced elevated blood pressure, but no change in the distribution of GluA1. Prior studies have shown that slow-pressor AngII is associated with increased plasma membrane-affiliated GluN1-SIG particles in ERβ-EGFP dendrites in male mice (Marques-Lopes et al., 2014b). The latter finding is consistent with studies demonstrating that NMDA receptors in the PVN are important for the slow-pressor AngII model of hypertension [reviewed in (Zhou et al., 2019)]. In contrast, the present study demonstrates that the subcellular distribution of GluA1-SIG particles in ERβ-EGFP dendrites is not altered in males following slow-pressor AngII (Fig. 9). This finding is in agreement with previous studies showing that the subcellular protein levels of GluA2, but not GluA1, are altered in the PVN of male SHRs (Li et al., 2012).

AngII hypertension in post-AOF mice is accompanied by an increase in near plasma membrane GluA1 in ERβ-expressing PVN dendrites

Although AngII treatment was associated with hypertension in both peri- and post-AOF females, the density of GluA1-SIG particles in ERβ-containing dendrites increased near the plasma membrane in post-AOF females only (Fig. 9). As many PVN ERβ-containing neurons project to the spinal cord (Bingham et al., 2006; Milner et al., 2010), this finding would be consistent with an increase in the excitatory state of sympathoexcitatory neurons.

The redistribution of GluA1-SIG particles toward the plasma membrane of ERβ-containing dendrites of AngII-infused post-AOF mice contrasts with what has previously been reported with the NMDA GluN1 receptor subunit (Marques-Lopes et al., 2017). In the case of GluN1, slow-pressor AngII resulted in an increase in cytoplasmic, but not plasma membrane labeling in ERβ-containing dendrites in AngII-infused post-AOF mice (Marques-Lopes et al., 2017). Together, these findings may indicate that AMPA receptors, rather than NMDA receptors, have a more prominent role in the glutamate mediation of AngII hypertension in post-AOF female mice compared to peri-AOF mice.

Several lines of evidence support direct or indirect roles for ERβ activation in the modulation of GluA1 subcellular localization. In the hippocampus, ERβ activation regulates GluA1 phosphorylation as well as trafficking to and from the plasma membrane (Liu et al., 2008). The modulation of GluA1 trafficking by ERβ involves phosphorylation at Ser845 (Liu et al., 2008), which plays a key role in the subcellular localization of GluA1. Activation of ERβ may also influence GluA1 localization via modulation of NMDA receptors. In various brain regions, activation of NMDA receptors induces the delivery of GluA1-containing AMPA receptors to synapses (Hayashi et al., 2000). Significantly, ERβ activation, in turn, has been reported to modify NMDA receptor signaling (Cyr et al., 2001; Tan et al., 2012; Sellers et al., 2015). These actions may involve ERβ-dependent modulation of various protein kinases (e.g. ERK1/2, PKC, and PKA) known to impact NMDA receptor signaling and GluA1-expressing AMPA receptor subcellular localization (Norman et al., 2000; Roepke et al., 2011).

Functional considerations

The present results in combination with our previous studies of the NMDA receptor conducted under similar conditions (Marques-Lopes et al., 2014a; Marques-Lopes et al., 2017) suggest that the glutamatergic mechanisms for regulating the excitatory state of PVN neurons may differ between young males and age-matched gonadally-intact females, as well as female mice at early and late stages of ovarian failure. Moreover, it may be hypothesized that dysregulation in estrogen levels over the course of AOF contribute to alterations in glutamate receptor localization in estrogen-sensitive hypothalamic neurons. These actions may have significant implications for the development of hypertension during the process of ovarian decline. In young ovarian-intact female mice, estrogen at normally circulating levels appears to favor insensitivity to the hypertensive actions of AngII and an associated resistance to increased glutamate signaling by accumulating receptors in the cytosolic compartment and away from functional sites on the plasma membrane. In contrast, during irregular estrogen levels and cycling seen during peri-AOF, mice become sensitive to the hypertensive actions of AngII and also show an increase in surface NMDA, but not AMPA receptors. However, during a state of estrogen depletion post-AOF mice, while also remaining sensitive to the hypertensive actions of AngII, also show an increase in plasma membrane AMPA GluA1 receptors but not NMDA receptors.

Alterations in the ratio of AMPA GluA1: NMDA receptors at different stages of AOF could indicate a change in the proportion of synapses that demonstrate an NMDA but no AMPA receptor response. This may be similar to the phenomenon of “silent synapses” (Malinow and Malenka, 2002) characterized during LTP. Expression of LTP occurs when silent synapses are activated after recruiting AMPA receptors to the post-synaptic membrane without altering NMDA receptor excitability (Montgomery et al., 2001). The awakening of silent synapses activates post-synaptic AMPA receptors, (Montgomery et al., 2001). Initially, LTP-induced AMPA receptors inserted into the plasma membrane are primarily composed of GluA1/2 subunits (Turrigiano and Nelson, 2004); however, these receptors subsequently are replaced with AMPA receptors composed of GluA2/3 subunits through constitutive receptor turnover (Malinow and Malenka, 2002). In hippocampus, estrogens are important for the presentation of GluA2 in synapses (Hara et al., 2012). Moreover, GluA2 levels in the hippocampus of ovariectomized rats are elevated following administration of an ERβ agonist (Waters et al., 2009). Thus, it is possible that the diminished availability of synaptic GluA2 due to the absence of estrogen in post-AOF mice impedes the turnover of GluA1 on the plasma membrane of ERβ-dendrites. An increase in synaptic GluA1 incorporation in a silent synapse-like event would be another means by which glutamate could excite PVN ERβ neurons. These processes may parallel the apparent two-step series of adaptations in glutamate receptor localization in PVN ERß-expressing neurons across the AOF transition from early to late stages. The early stage of AOF (i.e. peri-AOF) hypertension is marked by an increase in plasma membrane NMDA receptors, and the later stage of AOF (i.e. post-AOF) characterized by an increase in plasma membrane AMPA GluA1 receptors.

Ultimately, the unique signatures of glutamate receptor plasticity in peri- versus post-AOF mice suggest that separate pathways signaling through different ionotropic glutamate receptor subtypes contribute to early and late stages of AOF, and in turn, these may be further accentuated during hypertension. Distinctions in the mechanisms contributing to neural plasticity as females transition from the estrogen dysregulation seen during peri-AOF to depletion occurring at post-AOF point to the possibility of stage-dependent “windows of opportunity” for intervention. Future studies may investigate more specific mechanisms by which NMDA and AMPA receptors, possibly under the influence of ERβ, act to modulate sympathetic activity, leading to hypertension, during this critical window at the start of ovarian failure.

Highlights.

1. Sex and hormonal status influence GluA1 trafficking in estrogen receptor β (ERβ)-containing paraventricular (PVN) neurons.

2. Saline-infused late stage accelerated ovarian failure (AOF) female mice had higher plasmalemmal GluA1 in ERβ-PVN dendrites.

3. Males and AOF females, but not young females, had elevated blood pressure after slow-pressor angiotensin II (AngII) infusion.

4. Only late stage AOF females infused with slow-pressor AngII showed increased near plasmalemmal GluA1 in ERβ-PVN dendrites.

5. Late stage AOF females display a molecular signature of long-lasting synaptic strength prior to and following hypertension.

ABBREVIATIONS

AngII

angiotensin II

AOF

accelerated ovarian failure

BAC

bacterial artificial chromosome

BSA

bovine serum albumin

DAB

diaminobenzidine

EGFP

enhanced green fluorescent protein

ERα

estrogen receptor α

ERβ

estrogen receptor β

EM

electron microscopic

FSH

follicle stimulating hormone

GFP

green fluorescent protein

ir

immunoreactivity

PVN

paraventricular nucleus of the hypothalamus

PFA

paraformaldehyde

PB

phosphate buffer

PBS

phosphate-buffered saline

SHR

spontaneously hypertensive rat

SIG

silver-intensified immunogold

SBP

systolic blood pressure

TS

tris-buffered saline

VCD

4-vinylcyclohexane diepoxide