Evidence that estrogen directly and indirectly modulates C1 adrenergic bulbospinal neurons in the rostral ventrolateral medulla

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –1 78

a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m

w w w. e l s e v i e r. c o m / l o c a t e / b r a i n r e s

Research Report

Evidence that estrogen directly and indirectly modulates C1 adrenergic bulbospinal neurons in the rostral ventrolateral medulla Gang Wang a,⁎, Carrie T. Drake a , Mariya Rozenblit a , Ping Zhou a , Stephen E. Alves b,c , Scott P. Herrick a,b , Shinji Hayashi d , Sudha Warrier c , Costantino Iadecola a , Teresa A. Milner a,b,⁎ a

Division of Neurobiology, Department of Neurology and Neuroscience, Weill Medical College of Cornell University, 411 East 69th Street, New York, NY 10021, USA b Harold and Margaret Milliken Hatch Laboratory of Neuroendocrinology, The Rockefeller University, 1230 York Avenue, New York, NY 10021, USA c Department of Molecular Endocrinology, Merck Research Labs, 770 Sumneytown Pike, West Point, PA 19486, USA d Department of Endocrinology, Graduate School of Integrated Science, Yokohama City University, Yokohama 236-0027, Japan

A R T I C LE I N FO

AB S T R A C T

Article history:

Blood pressure in women increases after menopause, and sympathetic tone in female rats

Accepted 28 March 2006

decreases with estrogen injections in the rostral ventrolateral medulla (RVLM) region that

Available online 11 May 2006

contains bulbospinal C1 adrenergic neurons and is involved in blood pressure control. We investigated the anatomical and physiological basis for estrogen effects in the RVLM.

Keywords:

Neurons with α- or β-subtypes of estrogen receptor (ER) immunoreactivity (-ir) overlapped

Estrogen receptor alpha

in distribution with tyrosine hydroxylase (TH)-containing C1 neurons. Immunoelectron

Estrogen receptor beta

microscopy revealed that ERα- and ERβ-ir had distinct cellular and subcellular

Bulbospinal neurons

distributions. ERα-ir was most commonly in TH-lacking profiles, many of which were

Sex steroids

axons and peptide-containing afferents that contacted TH-containing dendrites. ERα-ir

Electron microscopy

was also in some TH-containing dendrites. ERβ-ir was most frequently in TH-containing

Whole-cell patch clamp

somata and dendrites, particularly on endoplasmic reticula, mitochondria, and plasma membranes. In whole-cell patch clamp recordings from isolated bulbospinal RVLM neurons, 17β-estradiol dose-dependently reduced voltage-gated Ca++ currents, especially the long-lasting (L-type) component. This inhibition was reversed by washing or prevented by adding the non-subtype-selective ER antagonist ICI182780. An ERβ-selective agonist, but not an ERα-selective agonist, reproduced the Ca++ current inhibition. The data indicate that estrogens can modulate the function of RVLM C1 bulbospinal neurons either directly, through extranuclear ERβ, or indirectly through extranuclear ERα in selected afferents. Moreover, Ca++ current inhibition may underlie the decrease in sympathetic tone evoked by local 17β-estradiol application. These findings provide a structural and functional basis for the effects of estrogens on blood pressure control and suggest a mechanism for the modulation of cardiovascular function by estrogen in women. © 2006 Elsevier B.V. All rights reserved.

⁎ Corresponding authors. Division of Neurobiology, Weill Medical College of Cornell University, 411 East 69th Street, New York, NY 10021, USA. Fax: +1 212 988 3672. E-mail addresses: [email protected] (G. Wang), [email protected] (T.A. Milner). 0006-8993/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.brainres.2006.03.089

164

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –17 8

Abbreviations: DPN, 2,3-bis(4-hydroxyphenyl)propionitrile BSA, bovine serum albumin EM, electron microscopy ER, estrogen receptor GAPDH, glyceraldehyde-3phosphate dehydrogenase l-aCSF, lactate-artificial cerebrospinal fluid PNMT, phenylethanolamine N-methyltransferase PB, phosphate buffer PPT, 4,4′,4″-(propyl-[1H]-pyrazole1,3,5-triyl)trisphenol RVLM, rostral ventrolateral medulla TS, Tris–saline TH, tyrosine hydroxylase

1.

Introduction

Cardiovascular disease is a leading cause of death in women, claiming 500,000 lives annually (American Heart Association, 2005). Arterial blood pressure levels tend to be lower in premenopausal women than in age-matched men, but levels increase in postmenopausal women and even surpass those in men (Burl et al., 1995). Mice deficient in estrogen receptors (ERs) are hypertensive, and gonadal steroids are linked to the degree of hypertension in rats (Ashton and Balment, 1991; Rowland and Fregly, 1992; Zhu et al., 2002). These observations have suggested that sex hormones, particularly estrogens, contribute to sexlinked differences in arterial blood pressure. In addition to their well-known peripheral effects, estrogens can directly influence brain networks involved in the regulation of cardiovascular functions (He et al., 1998; Mohamed et al., 1999; Saleh et al., 2000a,b;Saleh and Connell, 1999, 2000). In particular, estrogens can modulate arterial baroreceptor reflex pathways by affecting the rostral ventrolateral medulla (RVLM), a brainstem region that projects to spinal presympathetic neurons and is critical for the tonic (resting) and reflex control of arterial pressure (Aicher et al., 2000; Dampney et al., 2000; Guyenet et al., 1989; He et al., 1999). Local injections of 17β-estradiol into the RVLM region that contains bulbospinal neurons decrease sympathetic tone within minutes (Saleh and Connell, 1999, 2000). Approximately 70% of RVLM bulbospinal neurons belong to the C1 adrenergic cell group, as identified by tyrosine hydroxylase (TH) and phenylethanolamine-N-methyltransferase (PNMT) (Jeske and McKenna, 1992; Phillips et al., 2001). The observation that estrogens can regulate the expression of c-Fos in PNMT-containing neurons (Lee et al., 2000) suggests that the C1 adrenergic neuron subpopulation may be particularly sensitive to estrogens. A growing body of evidence suggests that estrogens can modulate C1 bulbospinal neurons through both genomic and nongenomic mechanisms. There are two known forms of ERs, ERα and ERβ; both ERs have a high and nearly equal affinity for estrogens (Levin, 2001). Light microscopic examination of the rat RVLM shows nuclear ERα immunoreactivity (-ir) in about 20% of PNMT-containing neurons (Lee et al., 2000), supporting the notion that some RVLM bulbospinal neurons are genomically regulated

by estrogens. ERβ-ir is also detected in neurons within the RVLM (Merchenthaler et al., 2004; Mitra et al., 2003; Shughrue and Merchenthaler, 2001); however, it appears to be primarily nonnuclear, raising the possibility that these receptors are involved in rapid, nongenomic estrogen actions. The ER subtype and the subcellular location (i.e., nuclear vs. membrane associated) could influence the type and degree of neuronal response (McEwen and Alves, 1999). In particular, in other brain regions, estrogens can affect voltage-gated Ca++ currents (Joëls and Karst, 1995; Mermelstein et al., 1996). We hypothesize that estrogen exerts rapid effects on RVLM function via ERs present in C1 neurons and/or their afferents. To test this hypothesis, we used (1) electron microscopy to localize ERα- and ERβ-ir in C1 neurons and their afferents; and (2) electrophysiology to determine whether estrogens modulate Ca++ currents of RVLM C1 bulbospinal neurons.

2.

Results

2.1. TH-, ERα-, and ERβ-immunoreactive neurons have overlapping distributions in RVLM Because sympathoexcitatory bulbospinal neurons are located in the rostral portion of the C1 adrenergic cell group (Morrison et al., 1988; Schreihofer and Guyenet, 1997; Verberne et al., 1999), light and electron microscopic analyses were focused on this region. At the light microscopic level, ERα-labeled nuclei were found in the same RVLM region as TH-immunoreactive neurons (Figs. 1A and B). ERβ-ir was found in perikarya scattered throughout the C1 area of the RVLM (Fig. 1C). When the primary antibodies for ERα and ERβ were omitted, no immunoreactivity was detected in the RVLM (Fig. 1D).

2.2. ERα- and ERβ-ir have distinct neuronal and subcellular distributions Dual labeling immunoelectron microscopy was used to localize ERα- and ERβ-ir in C1 neurons (identified by TH immunoreactivity) and their afferents. Immunoreactivities for both ERs were found at extranuclear sites in TH-labeled

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –1 78

165

Fig. 1 – By light microscopy, neurons with TH, ERα, and ERβ immunoreactivities have overlapping distributions in the RVLM. (A) TH-ir is found in perikarya and dendrites (arrowheads indicate examples). (B) ERα-ir is found in scattered cell nuclei (arrowheads indicate example). (C) ERβ-ir is found in a few perikarya (arrowheads). (D) In the absence of primary antibody, no immunoreactivity is detected in the RVLM. All photographs taken from coronal sections at approximately level 65 of Swanson (1992). D, dorsal; M, medial. Scale bar, 500 μm.

perikarya, dendrites, and terminals and in neuronal profiles associated with TH-labeled profiles. However, the subcellular distribution and proportion of cellular profiles that contained ERα-ir differed from ERβ-ir (Tables 1 and 2). The distribution of TH-ir in the RVLM was consistent with our earlier studies (Milner et al., 1988, 1989). Neither ER was localized to nuclei at the EM level, likely because detergents to enhance penetration were not added to the primary antibody incubation. ERα-ir was found in four major extranuclear locations (Fig. 2; Tables 1 and 2). First, ERα-ir was contained within TH-labeled dendritic shafts, axons, and terminals. Within dendrites, ERα-ir

Table 1 – Relative distributions of ERα and ERβ immunoreactivities in RVLM cellular profiles ERα (N = 156) Soma Dendrite Axon Terminal Glia

2% 37% 34% 18% 9%

(3) (58) (53) (28) (14)

ERβ (N = 213) 8% 57% 21% 5% 9%

(17) (121) (45) (11) (19)

For each antibody, analysis was conducted on four animals, one block per animal. The total area examined for each antibody was 2.4 × 106 μm2. The numbers in parentheses indicate the number of profiles containing the ER labeling.

usually was diffuse and was not affiliated with any organelles or the plasma membrane (Fig. 2A), although it was sometimes associated with large neurosecretory-like granules (Fig. 2B). Second, ERα-ir was found in non-TH-containing axon terminals (Figs. 2A, C, and D), and a few dendrites and glia, that apposed THlabeled dendrites. Axon terminals with ERα-ir were usually large (0.9 ± 0.09 μm in diameter; n = 18) and contained numerous densecore vesicles (DCVs). Peroxidase reaction product for ERα often was associated with the large DCVs (Figs. 2A, C, and D). Rarely, smaller terminals (< 0.5 μm) with diffuse ERα-ir were observed (not shown). A few ERα-containing terminals were immunoreactive for TH (Table 2). Third, ERα-ir was found in dendrite and terminal profiles that were within 0.5 μm of TH-labeled dendrites (Fig. 2A; Table 2). Fourth, ERα-ir was in TH-lacking dendrites that were contacted by TH-containing axon terminals (Table 2). Unlike ERα-ir, ERβ-ir was found in TH-immunoreactive perikarya (Figs. 3A and B). In the single planes of section analyzed, ERβ-ir was detected in 8 out of 15 (53%) TH-immunoreactive perikarya profiles. Within these perikarya, ERβ–peroxidase reaction product often was in clusters affiliated with endoplasmic reticula (Fig. 3A), mitochondria (Fig. 3B), or the plasma membrane (Fig. 3A). Like ERα-ir, ERβ-ir was in dendritic profiles (Figs. 3C and D); however, nearly twice as many ERβ-ir dendrites contained TH labeling (Table 2). Within dendritic profiles, peroxidase reaction product for ERβ-ir was in discrete aggregates (Figs. 3C and D) that often were affiliated with the plasma membrane and/or

166

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –17 8

Table 2 – Interactions of ERαERα- or or ERβ-labeled ERβ-labeled profiles profiles with with TH-labeled TH-labeledprofiles profiles

Relationship type

% of ERα (n = 156)

% of ERβ (n = 213)

COLOCALIZATION ER within TH-containing perikarya

0 (0)

8 (17)

ER within TH-containing dendrites

24 (37)

51 (109)

ER within TH containing axons/terminals

4.5 (7)

ER-axons/terminals apposed to TH-dendrites

16 (25)

ER-dendrites apposed to TH-dendrites

4.5 (7)

4 (9)

2 (3)

3 (6)

49 (77)

16 (34)

1 (2)

APPOSITION

ER-glia apposed to TH-dendrites

17 (36)

NOT DIRECT ER-profiles near TH-containing profiles

For each antibody, analysis was conducted on four animals, 1 block per animal. The total area examined for each antibody was 2.4 × 106 μm2. Numbers of profiles are indicated in parentheses.

endomembranes (Fig. 3C). Additional analysis of the subcellular distribution of ERβ-ir using immunogold-silver confirmed discrete localization on mitochondria, endomembranes, and plasma membrane (Fig. 3E; Table 3). ERβ-ir was found frequently in axon terminals, and occasionally in dendrites and glia, that directly apposed THlabeled profiles (Fig. 3F; Table 2). Compared to ERα-containing terminals, ERβ-labeled axon terminals tended to be smaller (0.7 ± 0.08 μm in diameter; n = 13), with small synaptic vesicles distributed throughout the terminals and usually no large DCVs (Fig. 3F). Peroxidase reaction product for ERβ was mostly found in aggregates affiliated with small synaptic vesicles; however, it was sometimes diffusely distributed throughout labeled terminals. Compared to ERα-labeled profiles, fewer ERβ single-labeled profiles were found in the vicinity of TH-labeled profiles (Table 2). Some differences in the proportion of profiles containing ERα-ir or ERβ-ir (regardless of TH labeling) were found (Table 1). In particular, ERα-ir was more common in axonal and terminal profiles than was ERβ, and ERβ-ir was more often found in somata and dendrites than was ERα. Like ERα, ERβ-ir also was found in TH-labeled somata and dendrites. However, in contrast to ERα-ir, ERβ-ir was prominently detected on plasma membranes, endomembranes and mitochondria.

2.3. 17β-estradiol inhibited Ca++ currents in RVLM bulbospinal neurons Dissociated RVLM bulbospinal neurons containing rhodamine microbeads were identified in vitro. Each brainstem slice contained approximately 40 retrogradely labeled cells (N = 4; range 29–56). As seen in younger (P6–P16) rats (Kawai et al.,

1999; Lipski et al., 1998; Wang et al., 2001), many isolated RVLM bulbospinal neurons from older rats appeared healthy (i.e., not swollen, dark, or with grainy membranes) and were chosen for patch clamp recording (Fig. 4A). The conventional whole-cell configuration of the patch clamp technique was used to record voltage-activated Ca++ channel currents (ICa) that were evoked by depolarization from the holding potential of −60 mV to the stepping potentials (between −50 and +20 mV). At stepping potentials of −10 or −20 mV, the ICa displayed two components: transient and long lasting (Bean, 1989;Catterall, 1998;Li et al., 1998), both of which were inhibited by application of Cd++ (100 μM), a nonspecific Ca++ channel blocker (Fig. 4B). The long-lasting component of ICa was specifically inhibited by the L-type Ca++ channel blocker nicardipine (2 μM) (Fig. 4B). 17β-Estradiol at concentrations ranging from 10 nM to 10 μM suppressed the amplitude of both transient and longlasting components of Ca++ currents in a dose-dependent manner (Figs. 4C and 5). From 10 nM, 17β-estradiol significantly inhibited both the transient and the L-type Ca++ currents (P < 0.01; n = 3 neurons). At 10 μM, 17β-estradiol-induced suppression of the transient and long-lasting Ca++ currents seemed to reach a maximal effect, reducing them by 48 ± 6.7% (n = 6 neurons) and 62 ± 9.8% (n = 5 neurons), respectively (Figs. 4C and 5). The IC50 values were similar for transient and L-type current inhibition by 17β-estradiol (Fig. 5A). The effect of 17βestradiol was largely reversed by washing with the control buffer (Figs. 4C and 5B) and was prevented by co-application of ICI 182780 (3 μM), a nonspecific ER antagonist (n = 5 neurons; Figs. 4C and 5B). In contrast, a tetrodotoxin (TTX)-sensitive voltage-gated Na+ current, elicited by stepping from −60 to −20 mV, was not significantly altered by 17β-estradiol (vehicle: 100 ± 6.4%; 10 μM 17

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –1 78

167

Fig. 2 – ERα-ir is present in some TH-labeled dendrites and in terminals that contact TH-labeled dendrites. (A) A dual-labeled dendrite (ERa + TH-d) contains diffuse ERα peroxidase labeling (arrowhead) and TH-immunogold particles (arrows) throughout its cytoplasm. The dendrite is near an axon terminal with light ERα labeling (ERa-t) associated with large dense-core vesicles (arrowheads). (B) A dendrite contains TH-immunogold (examples, black arrows) and an ERα-labeled multivesicular body (arrowhead). (C) An ERα-labeled terminal (ERa-t) with an aggregate of immunolabeling (arrowhead) forms an asymmetric synapse (curved arrow) with a TH immunogold-labeled dendrite (TH-d). (D) ERα-ir in a large terminal (ERa-t) is affiliated with many dense-core vesicles (arrowheads). This terminal is adjacent to a TH-labeled dendrite (TH-d). Scale bar, 500 nm.

β-estradiol: 97.9 ± 8.5%; P > 0.05, n = 7). Also, a tetraethylammonium (TEA)- and charybdotoxin-sensitive sustained outward K+ current (Wang et al., 2001) elicited by stepping from −60 mV to +60 mV was not significantly inhibited by 17β-estradiol (vehicle: 100 ± 16.3% of control; 3 μM 17β-estradiol: 92 ± 3.2%; P > 0.05, n = 4).

2.4. The ERβ agonist DPN, but not the ERα agonist PPT, inhibited Ca++ currents in RVLM bulbospinal neurons To determine whether the 17β-estradiol-induced inhibition of voltage-gated Ca++ currents in RVLM bulbospinal neurons

is mediated by the ERα and/or ERβ subtypes, we examined the effects of the ERβ agonist, DPN, and the ERα agonist, PPT. Like 17β-estradiol, DPN (1–300 nM) dose-dependently suppressed the transient and L-type Ca++ currents (Fig. 6C). At 30 nM, DPN preferentially inhibited the L-type Ca++ current (by 42.3 ± 14.6%; P < 0.05). At 100 and 300 nM, DPNinduced inhibition of both the transient and L-type Ca++ currents reached a maximum (Fig. 6C). This inhibition could be reversed by washing with the control buffer (data not shown). In contrast, PPT (300 nM) exerted little influence on the transient Ca++ current (100.5 ± 8.6% of controls, P > 0.05,

168

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –17 8

Fig. 3 – ERβ is present in discrete aggregates in many TH-labeled somata and dendrites and in some terminals. (A) A TH-labeled soma (ERb + TH-s) contains a cluster of ERβ immunoperoxidase reaction product (arrowhead) near the endoplasmic reticulum (er). A nearby dendrite (ERb + TH-d) contains TH labeling, and ERβ-ir associated with a synapse (curved arrow). (B) In a TH-labeled soma (ERb + TH-s), a patch of ERβ immunoperoxidase labeling (arrowhead) is near a mitochondrion (m). (C) An aggregate of ERβ labeling (arrowhead) is associated with an endomembrane (em) in a TH-immunogold-labeled dendrite (ERb + TH-d). The aggregate of ERβ-ir is beneath a synapse (curved arrow) formed by an unlabeled terminal (uT). (D) A patch of ERβ–peroxidase (arrowhead) is near the plasma membrane (pm) of a TH-labeled dendrite. (E) In a dendrite (ERb + TH-d) labeled for TH using immunoperoxidase, ERβ immunogold-particles (arrowheads) are found in the cytoplasm, associated with the plasma membrane (pm) or a mitochondrion (m). (F) A terminal with diffuse ERβ labeling (ERb-t) forms a symmetric synapse (curved arrow) with a TH-labeled soma (TH-s). Scale bar, 500 nm. n = 6 neurons) or the L-type Ca++ current (91.4 ± 6.7% of controls, P > 0.05, n = 6 neurons) (Figs. 6A and B). The phenotype of neurons from which recordings were made was assessed by single-cell RT-PCR. mRNA for the C1 neuron marker TH and the GABAergic marker GAD67 were assessed from aspirates of some bulbospinal cells from which recordings were obtained. All RVLM neurons examined (n = 6) contained TH but not GAD67 (Fig. 7), indicating that they were C1 neurons.

3.

Discussion

We have demonstrated that extranuclear ERs are present in C1 bulbospinal neurons and their afferents. Whereas ERα-ir is located primarily in afferents, ERβ-ir is commonly observed in somatodendritic regions of C1 neurons. Furthermore, estrogen modulates Ca++ currents in C1 neurons, an effect mediated

169

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –1 78

Table 3 – Subcellular distribution of ERβ immunogold particles in TH-labeled dendrites Cytoplasm 159 49.8%

Mitochondria 29 9%

Mitochondrial-associated endomembrane

Endoplasmic reticula

Microtubules

Plasma membrane

67 21%

21 6.6%

2 0.6%

41 12.9%

Three animals, 50 dual-labeled dendritic profiles per animal, were analyzed.

predominantly by ERβ. These findings provide a morphological and functional basis for direct effects of estrogen on C1 neurons and indirect effects mediated via modulation of selected afferents.

3.1.

Evidence for direct estrogen effects on C1 neurons

Our anatomical and physiological findings strongly suggest that RVLM C1 bulbospinal neurons contain functional ERs. 17β-Estradiol diminished Ca++ currents in isolated bulbospinal neurons, and the effect was blocked by an ER antagonist, ICI 182780. This shows that functional, rapidly activated ERs are present in at least the somata and proximal dendrites of C1

bulbospinal neurons. Our localization of extranuclear ER labeling in C1 neurons supports the presence of particular ER subtypes in these somatodendritic compartments and demonstrates ERs in small dendrites, extending previous light microscopic studies in rat and mouse showing nuclear ERα labeling in catecholaminergic neurons (Lee et al., 2000; Vanderhorst et al., 2005). The present results suggest a major role for ERβ in directly influencing C1 neurons. The effect of 17β-estradiol on isolated spinally projecting RVLM neurons was mimicked by the ERβselective agonist but not by the ERα agonist. Additionally, ERβ was frequently in perikarya and dendrites of TH-containing neurons, whereas ERα (although present in nuclei) was absent

Fig. 4 – 17β-estradiol inhibits voltage-gated Ca++ currents in bulbospinal RVLM neurons. (A) Photograph of a dissociated RVLM neuron retrogradely labeled with rhodamine microbeads that were injected into the thoracic spinal cord. (B) Representative voltage-gated Ca++ currents recorded from an RVLM neuron in the absence and presence of nicardipine (an L-type Ca++ channel blocker) and Cd++ (a nonspecific Ca++ channel blocker). (C and D) Voltage-gated Ca++ currents were elicited by depolarization to −20 mV from a holding potential of −60 mV. Application of 17β-estradiol inhibited transient and long-lasting Ca++ currents. This effect was prevented by with co-application of 3 μM ICI 182780, an ER antagonist.

170

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –17 8

Fig. 5 – 17β-estradiol inhibits Ca++ currents via estrogen receptors and in a dose-dependent manner. (A and B) Dose–response relationship of 17β-estradiol-induced inhibition of Ca++ currents in rhodamine labeled and isolated RVLM bulbospinal neurons. 17β-estradiol began to block the Ca++ currents at 10 nM. This inhibition became significant with higher 17β-estradiol concentrations (**P < 0.01; paired Student's t test). The inhibited Ca++ currents were largely recovered by wash with the control buffer. Subsequent co-administration of ICI 182780 with 17β-estradiol (100 nM) completely prevented 17β-estradiol-mediated inhibition. Panel A shows fitted dose–response curves with IC50 for the transient (left) and L-type (right) Ca++ currents. IC50 was obtained using the equation: I = Imax[1 − (x / IC50 + x)], where I was the Ca++ channel current amplitude in the presence of a given dose of 17β-estradiol, Imax was the Ca++ current amplitude in the control, and x was the concentration of 17β-estradiol. Data shown as mean ± SEM; number of neurons sampled indicated in bars (B).

from TH-labeled perikarya and considerably less frequent than ERβ in TH-labeled dendrites. Finally, ERβ-ir was frequently on the plasma membrane, consistent with the possibility of rapid activation by extracellular ligands (Cahill et al., 2001). In contrast, ERα-ir was usually associated with intracellular organelles. Taken together, our findings suggest that ERβ is more likely to be responsible for the rapid effects of 17β-estradiol on RVLM bulbospinal neurons. Consistent with observations in other brain regions (Hrabovszky et al., 2004; Milner et al., 2005), ERβ-ir in the RVLM was associated with cytoplasmic organelles, particularly endoplasmic reticula, mitochondria, and endomembranes near mitochondria. Light and confocal microscopic studies on cultured primary neurons and cardiomyocytes have suggested that ERβ-ir is associated with mitochondria (Cammarata et al., 2004; Yang et al., 2004). Whereas our study supports this, it suggests that ERβ is primarily associated with the endomembranes adjacent to mitochondria. Endoplasmic reticula interconnected with the mitochondria are thought to generate Ca++ signals that control processes such as synaptic plasticity and neuronal excitability (Berridge, 2002; Mironov et al., 2005). The association of ERβ-ir with mitochondria or nearby endomembranes may be

important in estrogen-induced production of NAD(P)H-derived reactive oxygen species (Wagner et al., 2001). Consistent with this, NAD(P)H oxidase immunolabeling is associated with mitochondria in brain (Wang et al., 2004).

3.2.

Evidence for indirect estrogen effects on C1 neurons

ERα-ir, and less frequently ERβ-ir, was in terminals contacting TH-labeled dendrites and somata, suggesting that estrogens presynaptically influence neurotransmitter release onto C1 neurons. Consistent with this possibility, estrogens can modulate the release of several neurotransmitters, including norepinephrine and hypothalamic peptides in other brain regions (Crowley, 1988; Karkanias and Etgen, 1993; MartínezMorales et al., 2001). Within terminals, ERα-ir was frequently located on large DCVs, which are peptide-containing neurosecretory vesicles that require relatively intense stimulation to fuse with the plasma membrane and release their contents (Thureson-Klein et al., 1986). This location suggests that ERα may belong to a reserve pool of ERs that could be presented to the plasma membrane under conditions of altered hormone levels and/or high frequency stimulation, such as baroreceptor activation. The newly incorporated ERα would then be

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –1 78

171

Fig. 6 – The ERβ agonist, DPN, but not the ERα agonist, PPT, inhibits L-type Ca++ currents. (A) Representative traces of voltage-gated Ca++ currents in a rhodamine-labeled bulbospinal RVLM neuron are shown in control buffer, in the presence of the ERα agonist PPT (300 nM) or ERβ agonist DPN (300 nM), and after wash with the control buffer. (B) Cumulative data showing the different effects of the ERβ agonist DPN (300 nM) and ERα agonist PPT (300 nM) on the L-type Ca++ currents in bulbospinal RVLM neurons. (C) Dose–response relationship of DPN-induced inhibition of transient and L-type Ca++ currents in isolated RVLM bulbospinal neurons. Inhibition of L-type Ca++ currents became significant at 30 nM. At 100 nM, inhibition of transient Ca++ currents became significant (*P < 0.05; **P < 0.01; paired Student's t test). Data shown as mean ± SEM; number of neurons sampled indicated in bars (B) or above data points (C).

positioned to influence subsequent transmitter release. Our observation that ERα-containing terminals were found within 0.5 μm of TH-containing profiles suggests that peptides released from ERα-labeled terminals could activate C1 neurons through volume transmission (see discussion in Drake et al., 2002). Terminals with ERα-ir and ERβ-ir had different morphological features, suggesting that subtype-selective ER agonists would activate different sets of afferents. The large size and numerous DCVs of many ERα-immunoreactive terminals are suggestive of peptide-containing afferents that arise from magnocellular hypothalamic neurons and can directly alter arterial pressure (Gomez et al., 1993; Milner et al., 1993). The

rarer ERα-labeled terminals that were smaller and contained DCVs resemble the afferents from the nuclei of the solitary tracts, which conduct pressor responses to chemoreceptor stimulation (Aicher et al., 1995, 2000). ERβ-labeled terminals, which were smaller than ERα-labeled terminals, lacked DCVs and formed synapses with TH-labeled somata, resemble the GABAergic afferents from the caudal ventrolateral medulla (Jeske et al., 1995; Milner et al., 2001a) that regulate sympathoinhibition (Jeske et al., 1993, 1995). However, our finding that few terminals contained ERβ-ir suggests that this receptor subtype may not play a significant role, at least under resting conditions, in nongenomically regulating afferents to RVLM bulbospinal neurons.

172

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –17 8

pal and neostriatal neurons (Joëls and Karst, 1995; Kurata et al., 2001; Mermelstein et al., 1996). Our finding that ER activation also affects Ca++ currents in RVLM bulbospinal neurons, which are known to contain glutamate (Aicher et al., 2000), is consistent with these studies.

3.4.

Fig. 7 – RT-PCR revealed that many bulbospinal neurons used in the physiological studies are catecholaminergic. Three retrogradely labeled RVLM neurons expressed TH but not GAD67 mRNA, as assessed by single-cell RT-PCR. The housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was a control for amplification and gel loading.

3.3. Estrogen diminishes C1 bulbospinal neuron Ca++ conductance The present study revealed that concentrations of 17βestradiol in the physiological range produce a dramatic blockade of voltage-activated Ca++ channels of rat RVLM bulbospinal neurons without significant effects on voltagegated K+ and Na+ channels. In agreement with previous studies (Stornetta and Guyenet, 1999), RVLM bulbospinal neurons expressed TH but not GAD67. 17β-estradiol inhibited both the transient and L-type Ca++ currents, but the greater magnitude of L-type Ca++ current inhibition suggests that estrogens may play an important role in modulating exocytosis of the “slow” neurotransmitters (Wang et al., 1997, 1999) in presympathetic RVLM bulbospinal neurons. Further supporting this speculation, the ERβ agonist DPN also exerted a preferential inhibitory effect on the L-type Ca++ current (Catterall, 1998; Li et al., 1998). Finally, the inhibitory actions of 17β-estradiol and DPN on Ca++ currents suggest an ERβdependent mechanism that might underlie the previous finding that local microinjection of 17β-estradiol into the RVLM decreases sympathetic tone (Saleh and Connell, 2000). Evidence from other brain regions suggests that ER activation can influence neuronal excitability, and that the degree of excitability depends upon the subtype and the subcellular location (i.e., nuclear vs. membrane) of the receptor. Short-term effects of acute estradiol exposure include increases in glutamate receptor-mediated currents and long-term potentiation in hippocampal pyramidal cells (Foy et al., 1999; Wilson, 1996; Woolley and Schwartzkroin, 1998). Estrogens also can inhibit NMDA-induced intracellular Ca++ increases and voltage-gated Ca++ channel currents, primarily by binding to plasma membrane ERs, in hippocam-

Implications for blood pressure control in females

Estrogens can have rapid effects by binding to ERs on cell membranes (Kelly and Levin, 2001; Razandi et al., 1999; Watson et al., 1999). The possibility that this also occurs in the RVLM was first suggested by the relatively rapid (within minutes) changes in sympathetic tone following local injections of 17βestradiol into the bulbospinal neuron-containing region of the RVLM (Saleh et al., 2000b; Saleh and Connell, 2000). The present findings that estrogen effects on bulbospinal neurons occur within 10 s and that C1 neurons contain plasmalemmal ERs provide strong evidence that the RVLM is poised to respond rapidly to changes in ovarian hormone levels. Previous data in rats have suggested that estrogens regulate central cardiovascular functions by producing a shift in sympathovagal balance toward the parasympathetic limb (Saleh et al., 2000a), resulting in an enhanced baroreflex function that provides protection against cardiac arrhythmias (Esler, 1992). In women, parasympathetic and sympathetic functions are altered in a complementary manner during the menstrual cycle (Minson et al., 2000; Saeki et al., 1997). The present study provides new evidence that estrogens can play a role in centrally mediated cardiovascular functions by directly and indirectly affecting C1 neurons in the RVLM. The cardioprotective effects of hormone replacement therapies in postmenopausal women are the subject of active debate. In several recently completed clinical trials, prolonged estrogen + progestin, or estrogen-only replacement did not reduce the risk of cardiovascular events in postmenopausal women with coronary heart disease; however, the steroid preparations and dosing schedules typically used in humans vary widely and may not be optimal for producing potential cardioprotective effects (Grady et al., 2002; Hodis et al., 2003; Wassertheil-Smoller et al., 2003). Importantly, an initiation of hormone therapy shortly after the onset of perimenopause may promote protection or benefit, as opposed to an increased risk from treatment if therapy is delayed 10–15 years postmenopause. This issue is being actively investigated in the ongoing Kronos Early Estrogen Prevention Study (KEEPS), a 5year study of women ages 42–58 (Harman et al., 2005). Also, continuous dosing with constant levels of hormones that normally cycle may be ineffective or even reverse positive actions (Gibbs et al., 2002). A better understanding of the mechanisms underlying ER actions in central cardiovascular circuits is needed to develop more effective therapeutic approaches. Our findings contribute to understanding the fundamental processes underlying the role of estrogens in cardiovascular control.

4.

Experimental procedures

All experiments were conducted in accordance with the NIH Guide for the Care and Use of Laboratory Animals and were

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –1 78

approved by the Weill Medical College of Cornell University Institutional Animal Care and Use Committee.

4.1.

Immunolocalization

4.1.1.

Animals

Female adult Sprague–Dawley rats (N = 8 intact; 3–6 months old; 300–400 g), from Charles River Laboratories (Wilmington, MA) or Taconic Laboratories (Chatham, NY) were used. All rats were housed with 12:12-h light/dark cycles (lights on 0600– 1800). Estrous cycle stage was determined using vaginal smear cytology (Turner and Bagnara, 1971), and females were assessed for at least two full cycles. The diestrus stage was chosen in this study because it is characterized by low to moderate circulating estrogen levels and produces maximal expression of nuclear ERs (Weiland et al., 1997).

4.1.2.

Antibodies

4.1.2.1. ERα.

A rabbit polyclonal antiserum (AS409; diluted 1:10,000) against the near full-length peptide of the native rat ERα (amino acids 61 through the carboxyl terminus) was produced and characterized previously in S. Hayashi's laboratory (Okamura et al., 1992). This antibody has been tested previously for specificity using Western blots and has been shown to recognize both the ligand-bound and unbound receptor migrating at ∼110 kDa (likely an ERα/fusion protein complex), migrating at ∼67 kDa, and migrating at ∼41–45 kDa (the degradation products of ERα from the fusion protein) (Alves et al., 1998; Milner et al., 2001b; Okamura et al., 1992). Following preadsorption of the AS409 antiserum with purified ERα protein, no bands were detected in any of these locations (Milner et al., 2001a). Light microscopic analysis of tissue sections fixed with 3.75% acrolein and 2% paraformaldehyde and labeled for the ERα (AS409) antiserum preadsorbed with purified ERα protein in the presence of 0.25% Triton X-100 resulted in no detectable nuclear labeling in the hypothalamus and hippocampus (Milner et al., 2001b). Light microscopic analysis of acrolein/paraformaldehyde tissue from α estrogen receptor knockout mice labeled for ERα in the presence of 0.1% Triton resulted in no detectable nuclear labeling in the midbrain and hippocampus (Alves et al., 2000). Electron microscopic analysis of hippocampal tissue sections fixed with 3.75% acrolein and 2% paraformaldehyde and labeled for the ERα (AS409) antiserum preadsorbed with purified ERα protein or preimmune serum for AS409 (AS401) using conditions identical to those used in the present study resulted in no detectable extranuclear labeling in dendrites or presynaptic profiles; in both instances, only one or two immunoperoxidase-labeled glial process per 3025 μm2 field were found.

4.1.2.2. ERβ. A polyclonal antibody (485; Merck Research Laboratories 80424; diluted 1:5000) generated in rabbit against a conserved sequence (rat aa 64–82) of the mouse, human, and rat ERβ that is located within the A/B domain of ERβ (exons 2– 3) and is not found in ERα (Mitra et al., 2003) was used. Previous Western blot analyses showed that the ERβ 485 antibody recognizes a protein that migrates at about 60 kDa on SF9 cell blots, 55 kDa on human ovary and testes blots, and about 70 kDa on whole tissue extracts of rat and mouse brain (Mitra

173

et al., 2003). The antibody labels nuclei in cos-7 cells transfected with ERβ and is eliminated in these cells and in mouse and rat brain hippocampal tissue by preadsorption with the antigenic peptide (Milner et al., 2005; Mitra et al., 2003). Light microscopic analysis of mouse brain sections fixed with 3.75% acrolein and 2% paraformaldehyde and labeled for the ERβ 485 antiserum preadsorbed with the antigenic peptide in the presence of 0.1% Triton X-100 resulted in no detectable nuclear or extranuclear labeling in any brain region (Mitra et al., 2003). Electron microscopic analysis of hippocampal tissue sections fixed with 3.75% acrolein and 2% paraformaldehyde and processed through the same labeling conditions as those used in the present study in the absence of primary antiserum for ERβ revealed no labeling in neuronal profiles; only one or two immunoperoxidase-labeled glial process per 3025 μm2 field were found. At least 5 splice variants of ERβ have been identified (Price et al., 2000). The selectivity of this antiserum for individual splice variants has not been determined. Although all precautions were taken to assure that the antibodies specifically recognize ERα or ERβ, it is possible that they could recognize peptide sequences contained in other proteins. Thus, the labeling described in the present study should be interpreted as ER-“like” immunoreactivity.

4.1.2.3. TH. A mouse antibody to tyrosine hydroxylase (TH; diluted 1:2000) was obtained from Incstar (Stillwater, MN). This antibody has been extensively characterized and localized in fixed rat brain (Milner et al., 1999). 4.1.3.

Immunocytochemistry

Eight adult female rats in diestrus were deeply anesthetized with sodium pentobarbital (150 mg/kg, i.p.) and perfused through the ascending aorta sequentially with solutions of (1) 10–15 ml saline (0.9%) containing 1000 U of heparin; (2) 50 ml of 3.75% acrolein and 2% paraformaldehyde in 0.1 M phosphate buffer (PB; pH 7.4); and (3) 200 ml of 2% paraformaldehyde in PB. Each brain was removed from the skull, cut into 5-mm coronal blocks, and postfixed with 2% paraformaldehyde in PB for 30 min. The block containing the RVLM was cut into 40-μmthick sections on a vibrating microtome (Vibratome; Leica, Wien, Austria). The sections were collected in PB and treated with 1% sodium borohydride in PB for 30 min prior to immunocytochemical labeling. Free-floating sections were incubated in 0.5% bovine serum albumin (BSA) in 0.1 M Tris–saline (pH 7.6; TS) for 30 min then anti-ER antiserum in 0.1% BSA in TS for 1 day at room temperature followed by an additional 4–5 days at 4°C with TH antiserum included for the last 24 h. ER labeling was visualized using the avidin–biotin complex (ABC) peroxidase technique (Hsu et al., 1981). Briefly, tissue was incubated in (1) biotinylated goat anti-rabbit IgG in 0.1% BSA (1:400, Vector Labs, Burlingame, CA) for 30 min; (2) peroxidase–avidin complex for 30 min; and (3) diaminobenzidine (Aldrich, Milwaukee, WI) and H2O2 for 6 min. Following ER labeling, sections were processed for TH labeling using the silverenhanced immunogold technique (Chan et al., 1990). For this, sections were rinsed in TS and incubated in donkey antimouse IgG conjugated to 1 nm gold particles (1:50, Electron Microscopy Sciences, Fort Washington, PA) in 0.01% gelatin and 0.08% BSA in phosphate-buffered saline (PBS; pH 7.4) for

174

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –17 8

2 h at room temperature. Sections then were rinsed in PBS, postfixed in 2% glutaraldehyde in PBS for 10 min, and rinsed in PBS followed by 0.2 M sodium citrate buffer (pH 7.4). The conjugated gold particles were enhanced by treatment with silver solution (EMS) for 5–7 min. Some sections were processed for EM localization of ERβ labeling using the immunogold-silver method and for TH using the immunoperoxidase method. All procedures were the same as described above except that the ERβ antiserum was diluted 1:2000 and the TH antiserum was diluted 1:10,000, and the secondary antibodies were donkey anti-rabbit IgG conjugated to 1 nm gold particles (EMS) and biotinylated horse-antimouse IgG (Vector). Sections for EM were postfixed for 1 h in 2% osmium tetroxide, dehydrated through alcohols and propylene oxide, and embedded between two sheets of plastic in EMbed 812 (EMS). Ultrathin sections (70 nm thick) were collected from the RVLM and cut on an ultratome (Leica) as previously described (Milner et al., 1999). Sections were counterstained with uranyl acetate and Reynold's lead citrate, and final preparations were analyzed on a FEI Tecnai Biotwin transmission electron microscope. Images were acquired with a digital camera system (Advanced Microscopy Techniques, software version 3.2) and adjusted for levels, brightness, and contrast in Adobe Photoshop 7.0 on an Apple Power Macintosh G5 computer. Final figures were assembled in Quark-Xpress 6.1.

4.1.4.

Immunocytochemical analysis

The subcellular location of ER–peroxidase labeling relative to TH-labeled profiles in the RVLM was examined. Quantitative analysis was performed on four blocks for sections labeled with ERα and TH (N = 4 rats) and four blocks for sections labeled with ERβ and TH (N = 4 rats). The total area examined for each antibody combination was identical (2.4 × 106 μm2). Blocks for EM analysis were selected from medulla sections that contained TH-immunolabeled perikarya in the RVLM (between levels 63 and 66 of Swanson, 1992) by light microscopy. Only grid squares within a depth of 1.5 μm from the tissue-plastic interface were used. To insure that differences in profile number were not due to differences in antibody penetration and sampling size (Leranth and Pickel, 1989), profiles were sampled from fields near the plastic-tissue interface and all fields sampled in each block were identical in area. As we and others have found (Blaustein, 1992; Milner et al., 2001b; Tabori et al., 2004; Wagner et al., 1998), patchy immunoreactivity and immunoreactivity in small, sparsely distributed profiles (e.g., axons) is more easily resolved by EM than light microscopy. From each block, all profiles containing ER labeling were photographed. ER-immunolabeled profiles were counted and classified according to the nomenclature of Peters et al. (1991). Soma profiles were identified by the presence of a nucleus. Dendrite profiles contained regular microtubule arrays and were usually postsynaptic to axon terminal profiles. Unmyelinated axon profiles had a diameter less than 0.2 μm, had a few small synaptic vesicles, and lacked synaptic junctions in the plane of section. Terminal profiles had minimal diameters greater than 0.2 μm, contained numerous small synaptic vesicles, and often contacted other neuronal profiles. Astrocyte profiles were identified by the presence of glial filaments, the tendency to conform to the

boundaries of surrounding profiles, and/or the absence of microtubules. The term “contact” is used here to include asymmetric and symmetric synapses and appositions. Asymmetric synapses had thickened postsynaptic densities, whereas symmetric synapses exhibited narrow synaptic densities. Appositions were defined as contacts not separated by glial profiles but lacking intercleft material or conventional synapses in the plane of section analyzed. Profiles that could not be definitively identified were categorized as “unknown”.

4.2.

Electrophysiology

4.2.1.

Preparation of dissociated RVLM bulbospinal neurons

Timed-pregnant Sprague–Dawley rats (N = 12) were purchased from Charles River Laboratories. Female offspring (ages 19– 26 days) were anesthetized with a mixture of ketamine (80 mg/ kg) and xylazine (5 mg/kg) (i.p.) and stereotaxically injected with rhodamine-conjugated microbeads bilaterally (2–3 μl each side) into the interomediolateral region of the thoracic (T2–T4) spinal cord (Kawai et al., 1999; Li and Bayliss, 1995; Li et al., 1998; Lipski et al., 1998; Wang et al., 2001). Two to four days later, rats were sacrificed in a CO2 chamber and decapitated, and a pie-shaped piece of the ventral medulla (with the vertex at the nucleus ambiguus extending down to the ventral surface) was cut from a 350-μm-thick coronal brainstem slice and placed in a chamber containing lactate-artificial cerebrospinal fluid (l-aCSF) composed of (in mM) 124 NaCl, 26 NaHCO3, 1 Na2PO4, 5 KCl, 2 MgSO4, 2 CaCl2, 10 glucose, 4.5 lactic acid, with pH 7.35, when oxygenated with 95% O2 and 5% CO2. This slice contains most retrogradely labeled RVLM neurons and excludes most noncardiovascular bulbospinal neurons that might have been labeled with the injection (Randich et al., 1991). Mild enzymatic digestion was performed in 7.0 ml of laCSF containing 140 U papain, 1.6 mM L-cystein, 0.2 mM EDTA, and 13.4 μM β-mercaptoethanol. Isolated RVLM neurons were placed in a Petri dish and perfused with l-ACSF.

4.2.2.

Recording Ca++, Na+, and K+ currents

Whole-cell voltage-clamp recordings (Hamill et al., 1981) from retrogradely labeled RVLM neurons (n = 45 neurons) were performed at room temperature. Cells that were swollen, dark, or had grainy membranes were discarded (Kawai et al., 1999). For Ca++ and Na+ current recordings, patch pipettes were filled with (in mM) 145 Cs-glutamate, 1 MgCl2, 10 HEPES, 10 Cs-EGTA, 3 Mg2-ATP, 0.2 cAMP, pH 7.3. For K+ current recording, patch pipettes were filled with (in mM) 114 K-gluconate, 17.5 KCl, 4 NaCl, 4 MgCl2, 10 HEPES, 0.2 K-EGTA, 3 Mg2-ATP, 0.3 NaGTP, pH 7.3. An Axonpatch 200A amplifier was used with pClamp6.01 or Windows 8.0 data acquisition programs. Junction potential was zeroed prior to gigaΩ-seal formation. Recorded currents were low-passed at 5 kHz, digitized at 10 kHz, and stored on disk for later analysis. Current–voltage curves of Ca++ currents were established using 500 ms voltage steps from a holding potential of −60 mV. The effect of drugs or hormones was examined using a local double-barrel perfusion system. Because of the large size (>20 μm in diameter) and extended processes of dissociated RVLM neurons, voltageclamp errors due to unsatisfied space-clamp will occur. Thus, any currents that were not smooth or did not have continuous voltage–time functions were excluded from further analysis

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –1 78

175

(Li et al., 1998). The amplitude of transient Ca++ currents was measured 65 ms following the start of the depolarizing pulse, and the amplitude of L-type Ca++ currents was measured 500 ms after initiating the depolarizing pulse.

ferroni's correction following one-way ANOVA was used for multiple comparisons between the control and treated groups (e.g., Figs. 5B and 6C).

4.2.3.

Acknowledgments

Drugs

The following drugs were used: 17β-estradiol (10 nM–10 μM; Sigma), the most potent naturally occurring estrogen in humans (Williams and Stancel, 1996); ICI182,780 (3 μM; Sigma), a specific ER antagonist, although under certain conditions a partial agonist at ERs (Couse et al., 2000); 4,4′,4″(propyl-[1H]-pyrazole-1,3,5-triyl)trisphenol (PPT), an ERα-selective agonist (300 nM; Tocris Cookston Ltd., Avon, United Kingdom) (Lund et al., 2005); 2,3-bis(4-hydroxyphenyl)-propionitrile (DPN), an ERβ-selective agonist (1–300 nM; Tocris Cookston Ltd.) (Lund et al., 2004); nicardipine (2 μM; Sigma), an L-type Ca++ channel blocker (Wang et al., 1997, 1999); and cadmium chloride (100 μM; Sigma), a nonspecific Ca++ channel blocker (Wang et al., 1997, 1999).

4.2.4.

Single-cell RT-PCR

During whole-cell patch clamp recordings, the pipette buffer contained carrier tRNA at 20 μg/ml. At the end of the experiments, the contents of recorded neurons were aspirated and analyzed for the expression of TH, glutamic acid decarboxylase (GAD67), and the housekeeping gene glyceraldehyde-3-phosphate dehydrogenase (GAPDH), using the reverse transcription polymerase chain reaction (RT-PCR) procedures of Kacharmina et al. (1999). Synthetic oligonucleotides were designed based on amino acid sequence information and/or primary DNA sequence information, using database accession available at the National Center for Biotechnology Information; primers were validated with the public domain software Primer3 (Rozen and Skaletsky, 2000). PC12 cells served as a control. Cytosol contents were expelled into silanized and RNAse-free tubes and the cDNA synthesis reaction was initiated by adding 10 μl reverse transcription buffer [50 mM Tris–Cl (pH 8.3), 10 mM dithiothreitol, 3 mM MgCl2, 75 mM KCl, 5 mM random hexamer primer, 0.5 mM of each deoxynucleotide (dNTPs), 2U/μl RNAse inhibitor (Rnasin, Promega), and 10 U of reverse transcriptase (Superscript II, Invitrogen)]. After 15 min incubation at room temperature, cDNA synthesis was carried out at 37°C for 60 min. RT-PCR amplification of gene fragments was accomplished with genespecific primers. Forty cycles of PCR reaction were carried out to ensure detection of low levels of expression. RNA from neurons of nearby regions that were not expected to contain the gene(s) of interest were used as a negative control. Primers: Rat TH—forward: 5′-AGGGCTGCTGTCTTCCTACGGAG; reverse: 5′-GAAAGGCCCTGGCACCTGTGG. GAD67—forward: 5′-CCCTTTGCAGAACCGTAATC; reverse: 5′-ACTTCCTGCCATCCATCATC. Rat GAPDH—forward: 5′-ACCACAGTCCATGCCATCAC; reverse: 5′-TCCACCACCCTGTT-GCTGTA.

4.2.5.

Statistical analysis

All data are expressed as the means ± standard error of the means (SEM), which were determined for each group of recorded neurons before and after application of drugs. Student's t test (two tailed) was used to compare the data for the two groups (e.g., Fig. 6B). Two-tailed t tests with Bon-

We thank Ms. Nora Tabori for technical assistance. Grant support: NIH grants HL18974; MH 59251 (S.E.A.).

REFERENCES

Aicher, S.A., Kurucz, O.S., Reis, D.J., Milner, T.A., 1995. Nucleus tractus solitarius efferent terminals synapse on neurons in the caudal ventrolateral medulla that project to the rostral ventrolateral medulla. Brain Res. 693, 51–63. Aicher, S.A., Milner, T.A., Pickel, V.M., Reis, D.J., 2000. Anatomical substrates for baroreflex sympathoinhibition in the rat. Brain Res. Bull. 51, 107–110. Alves, S.E., Weiland, N.G., Hayashi, S., McEwen, B.S., 1998. Immunocytochemical localization of nuclear estrogen receptors and progestin receptors within the rat dorsal raphe nucleus. J. Comp. Neurol. 391, 322–334. Alves, S.E., McEwen, B.S., Hayashi, S., Korach, K.S., Pfaff, D.W., Ogawa, S., 2000. Estrogen-regulated progestin receptors are found in the midbrain raphe but not hippocampus of estrogen receptor alpha (ERα) gene-disrupted mice. J. Comp. Neurol. 427, 185–195. American Heart Association, 2005. Women and Cardiovascular Disease. Ashton, N., Balment, R.J., 1991. Sexual dimorphism in renal function and hormonal status of New Zealand genetically hypertensive rats. Acta Endocrinol. (Copenh) 124, 91–97. Bean, B.P., 1989. Classes of calcium channels in vertebrate cells. Annu. Rev. Physiol. 51, 367–384. Berridge, M.J., 2002. The endoplasmic reticulum: a multifunctional signaling organelle. Cell Calcium 32, 235–249. Blaustein, J.D., 1992. Cytoplasmic estrogen receptors in rat brain: immunocytochemical evidence using three antibodies with distinct epitopes. Endocrinology 131, 1336–1342. Burl, V.L., Whelton, P., Roccella, E.J., Brown, C., Cutler, J.A., Higgins, M., Horan, M.J., Labarthe, D., 1995. Prevalence of hypertension in the US adult population: results from the Third national Health and Nutrition examination Survey, 1988–1991. Hypertension 25, 305–313. Cahill, C.M., Morinville, A., Lee, M.C., Vincent, J.P., Collier, B., Beaudet, A., 2001. Prolonged morphine treatment targets delta opioid receptors to neuronal plasma membranes and enhances delta-mediated antinociception. J. Neurosci. 21, 7598–7607. Cammarata, P.R., Chu, S., Moor, A., Wang, Z., Yang, S.H., Simpkins, J.W., 2004. Subcellular distribution of native estrogen receptor alpha and beta subtypes in cultured human lens epithelial cells. Exp. Eye Res. 78, 861–871. Catterall, W.A., 1998. Structure and function of neuronal Ca2+ channels and their role in neurotransmitter release. Cell Calcium 24, 307–323. Chan, J., Aoki, C., Pickel, V.M., 1990. Optimization of differential immunogold-silver and peroxidase labeling with maintenance of ultrastructure in brain sections before plastic embedding. J. Neurosci. Methods 33, 113–127. Couse, J.F., Curtis, H.S., Korach, K.S., 2000. Receptor null mice reveal contrasting roles for estrogen receptor alpha and beta in reproductive tissues. J. Steroid Biochem. Mol. Biol. 74, 287–296. Crowley, W.R., 1988. Sex differences in the responses of

176

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –17 8

hypothalamic luteinizing hormone-releasing hormone and catecholamine systems to ovarian hormones and naloxone: implications for sexual differentiation of luteinizing hormone secretion in rats. Brain Res. 461, 314–321. Dampney, R.A.L., Tagawa, T., Horiuchi, J., Potts, P.D., Fontes, M., Polson, J.W., 2000. What drives the tonic activity of presympathetic neurons in the rostral ventrolateral medulla? Clin. Exp. Pharmacol. Physiol. 27, 1049–1053. Drake, C.T., Chang, P.C., Harris, J.A., Milner, T.A., 2002. Neurons with mu opioid receptors interact indirectly with enkephalin-containing neurons in the rat dentate gyrus. Exp. Neurol. 176, 254–261. Esler, M., 1992. The autonomic nervous system and cardiac arrhythmias. Clin. Auton. Res. 2, 133–135. Foy, M.R., Xu, J., Xie, X., Brinton, R.D., Thompson, R.F., Berger, T.W., 1999. 17beta-estradiol enhances NMDA receptor-mediated EPSPs and long-term potentiation. J. Neurophysiol. 81, 925–929. Gibbs, R.B., Nelson, D., Anthony, M.S., Clarkson, T.B., 2002. Effects of long-term hormone replacement and of tibolone on choline acetyltransferase and acetylcholinesterase activities in the brains of ovariectomized, cynomologus monkeys. Neuroscience 113, 907–914. Gomez, R.E., Cannata, M.A., Milner, T.A., Anwar, M., Reis, D.J., Ruggiero, D.A., 1993. Vasopressinergic mechanisms in the nucleus reticularis lateralis in blood pressure control. Brain Res. 604, 90–105. Grady, D., Herrington, D., Bittner, V., Blumenthal, R., Davidson, M., Hlatky, M., Hsia, J., Hulley, S., Herd, A., Khan, S., Newby, L.K., Waters, D., Vittinghoff, E., Wenger, N., 2002. Cardiovascular disease outcomes during 6.8 years of hormone therapy, Heart and Estrogen/progestin Replacement Study follow-up (HERS II). JAMA 288, 49–57. Guyenet, P.G., Haselton, J.R., Sun, M.-K., 1989. Sympathoexcitatory neurons of the rostroventrolateral medulla and the origin of the sympathetic vasomotor tone. Prog. Brain Res. 81, 105–116. Hamill, O.P., Marty, A., Neher, E., Sakmann, B., Sigworth, F.J., 1981. Improved patch-clamp techniques for high-resolution current recording from cells and cell-free membrane patches. Pflugers Arch. 391, 85–100. Harman, S.M., Brinton, E.A., Cedars, M., Lobo, R., Manson, J.E., Merriam, G.R., Miller, V.M., Naftolin, F., Santoro, N., 2005. KEEPS, The Kronos Early Estrogen Prevention Study. Climacteric 8, 3–12. He, X.R., Wang, W., Crofton, J.T., Share, L., 1998. Effects of 17beta-estradiol on sympathetic activity and pressor response to phenylephrine in ovariectomized rats. Am. J. Physiol. 275, R1202–R1208. He, X.R., Wang, W., Crofton, J.T., Share, L., 1999. Effects of 17beta-estradiol on the baroreflex control of sympathetic activity in conscious ovariectomized rats. Am. J. Physiol. 277, R493–R498. Hodis, H.N., Mack, W.J., Lobo, R.A., 2003. Randomized controlled trial evidence that estrogen replacement therapy reduces the progression of subclinical atherosclerosis in healthy postmenopausal women without preexisting cardiovascular disease. Circulation 108, e5. Hrabovszky, E., Kallo, I., Steinhauser, A., Merchenthaler, I., Coen, C. W., Petersen, S.L., Liposits, Z., 2004. Estrogen receptor-beta in oxytocin and vasopressin neurons of the rat and human hypothalamus, Immunocytochemical and in situ hybridization studies. J. Comp. Neurol. 473, 315–333. Hsu, S.M., Raine, L., Fanger, H., 1981. Use of avidin-biotinperoxidase complex (ABC) in immunoperoxidase techniques: a comparison between ABC and unlabeled antibody (PAP) procedures. J. Histochem. Cytochem. 29, 557–580. Jeske, I., McKenna, K.E., 1992. Quantitative analysis of bulbospinal projections from the rostral ventrolateral medulla: contribution of C1-adrenergic and nonadrenergic neurons. J. Comp. Neurol. 324, 1–13.

Jeske, I., Morrison, S.F., Cravo, S.L., Reis, D.J., 1993. Identification of baroreceptor reflex interneurons in the caudal ventrolateral medulla. Am. J. Physiol.: Regul., Integr. Comp. Physiol. 264, R169–R178. Jeske, I., Reis, D.J., Milner, T.A., 1995. Neurons in the barosensory area of the caudal ventrolateral medulla project monosynaptically on to sympathoexcitatory bulbospinal neurons in the rostral ventrolateral medulla. Neuroscience 65, 343–353. Joëls, M., Karst, H., 1995. Effects of estradiol and progesterone on voltage-gated calcium and potassium conductances in rat CA1 hippocampal neurons. J. Neurosci. 15, 4289–4297. Kacharmina, J.E., Crino, P.B., Eberwine, J., 1999. Preparation of cDNA from single cells and subcellular regions. Methods Enzymol. 303, 3–18. Karkanias, G.B., Etgen, A.M., 1993. Estradiol attenuates alpha 2-adrenoceptor-mediated inhibition of hypothalamic norepinephrine release. J. Neurosci. 13, 3448–3455. Kawai, Y., Qi, J., Comer, A.M., Gibbons, H., Win, J., Lipski, J., 1999. Effects of cyanide and hypoxia on membrane currents in neurones acutely dissociated from the rostral ventrolateral medulla of the rat. Brain Res. 830, 246–257. Kelly, M.J., Levin, E.R., 2001. Rapid actions of plasma membrane estrogen receptors. Trends Endocrinol. Metab. 12, 152–156. Kurata, K., Takebayashi, M., Kagaya, A., Morinobu, S., Yamawaki, S., 2001. Effect of beta-estradiol on voltage-gated Ca(2+) channels in rat hippocampal neurons: a comparison with dehydroepiandrosterone. Eur. J. Pharmacol. 416, 203–212. Lee, E.J., Moore, C.T., Hosny, S., Centers, A., Jennes, L., 2000. Expression of estrogen receptor-α and c-Fos in adrenergic neurons of the female rat during the steroid-induced LH surge. Brain Res. 875, 56–65. Leranth, C., Pickel, V.M., 1989. Electron microscopic preembedding double-labeling methods. In: Heimer, L., Zaborszky, L. (Eds.), Neuroanatomical Tract-Tracing Methods 2. Plenum Press, New York, pp. 129–172. Levin, E.R., 2001. Cell localization, physiology, and nongenomic actions of estrogen receptors. J. Appl. Physiol. 91, 1860–1867. Li, Y.W., Bayliss, D.A., Guyenet, P.G., 1995. C1 neurons of neonatal rats: intrinsic beating properties and α2-adrenergic receptors. Am. J. Physiol.: Regul., Integr. Comp. Physiol. 269, R1356–R1369. Li, Y.W., Guyenet, P.G., Bayliss, D.A., 1998. Voltage-dependent calcium currents in bulbospinal neurons of neonatal rat rostral ventrolateral medulla: modulation by α2-adrenergic receptors. J. Neurophysiol. 79, 583–594. Lipski, J., Kawai, Y., Qi, J.G., Comer, A., Win, J., 1998. Whole cell patch-clamp study of putative vasomotor neurons isolated from the rostral ventrolateral medulla. Am. J. Physiol.: Regul., Integr. Comp. Physiol. 274, R1099–R1110. Lund, T.D., Rovis, T., Chung, W.C., Handa, R.J., 2005. Novel actions of estrogen receptor beta on anxiety-related behaviors. Endocrinology 146, 797–807. Martínez-Morales, J.R., López-Coviella, I., Hernández-Jiménez, J.G., Reyes, R., Bello, A.R., Hernández, G., Blusztajn, J.K., Alonso, R., 2001. Sex steroids modulate luteinizing hormone-releasing hormone secretion in a cholinergic cell line from the basal forebrain. Neuroscience 103, 1025–1031. McEwen, B.S., Alves, S.E., 1999. Estrogen actions in the central nervous system. Endocr. Rev. 20, 279–307. Merchenthaler, I., Lane, M.V., Numan, S., Dellovade, T.L., 2004. Distribution of estrogen receptor alpha and beta in the mouse central nervous system: in vivo autoradiographic and immunocytochemical analyses. J. Comp. Neurol. 473, 270–291. Mermelstein, P.G., Becker, J.B., Surmeier, D.J., 1996. Estradiol reduces calcium currents in rat neostriatal neurons via a membrane receptor. J. Neurosci. 16, 595–604. Milner, T.A., Pickel, V.M., Abate, C., Joh, T.H., Reis, D.J., 1988. Ultrastructural characterization of substance P-containing neurons in the rostral ventrolateral medulla in relation to

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –1 78

neurons containing catecholamine synthesizing enzymes. J. Comp. Neurol. 270, 427–445. Milner, T.A., Pickel, V.M., Reis, D.J., 1989. Ultrastructural basis for interactions between central opioids and catecholamines: I. Rostral ventrolateral medulla. J. Neurosci. 9, 2114–2130. Milner, T.A., Reis, D.J., Pickel, V.M., Aicher, S.A., Giuliano, R., 1993. Ultrastructural localization and afferent sources of corticotropin-releasing factor in the rat rostral ventrolateral medulla: implications for central cardiovascular regulation. J. Comp. Neurol. 333, 151–167. Milner, T.A., Rosin, D.L., Lee, A., Aicher, S.A., 1999. Alpha2Aadrenergic receptors are primarily presynaptic heteroreceptors in the C1 area of the rat rostral ventrolateral medulla. Brain Res. 821, 200–211. Milner, T.A., Drake, C.T., Aicher, S.A., 2001a. Cellular relations between μ-opioid receptive, GABAergic and reticulospinal neurons in the rostral ventrolateral medulla. Brain Res. 917, 1–14. Milner, T.A., McEwen, B.S., Hayashi, S., Li, C.J., Reagan, L.P., Alves, S.E., 2001b. Ultrastructural evidence that hippocampal alpha estrogen receptors are located at extranuclear sites. J. Comp. Neurol. 429, 355–371. Milner, T.A., Ayoola, K., Drake, C.T., Herrick, S.P., Tabori, N.T., McEwen, B.S., Warrier, S., Alves, S.E., 2005. Ultrastructural localization of estrogen receptor beta immunoreactivity in the rat hippocampal formation. J. Comp. Neurol. 491, 81–95. Minson, C.T., Halliwill, J.R., Young, T.M., Joyner, M.J., 2000. Influence of the menstrual cycle on sympathetic activity, baroreflex sensitivity, and vascular transduction in young women. Circulation 101, 862–868. Mironov, S.L., Ivannikov, M.V., Johansson, M., 2005. [Ca2+]i Signaling between mitochondria and endoplasmic reticulum in neurons is regulated by microtubules: from mitochondrial permeability transition pore to Ca2+-induced Ca2+ release. J. Biol. Chem. 280, 715–721. Mitra, S.W., Hoskin, E., Yudkovitz, J., Pear, L., Wilkinson, H.A., Hayashi, S., Pfaff, D.W., Ogawa, S., Rohrer, S.P., Schaeffer, J.M., McEwen, B.S., Alves, S.E., 2003. Immunolocalization of estrogen receptor beta in the mouse brain: comparison with estrogen receptor alpha. Endocrinology 144, 2055–2067. Mohamed, M.K., El Mas, M.M., Abdel-Rahman, A.A., 1999. Estrogen enhancement of baroreflex sensitivity is centrally mediated. Am. J. Physiol. 276, R1030–R1037. Morrison, S., 1988. Reticulospinal vasomotor neurons of the rat rostral ventrolateral medulla: relationship to sympathetic nerve activity and the C1 adrenergic cell group. J. Neurosci. 8, 1286–1302. Okamura, H., Yamamoto, K., Hayashi, S., Kuroiwa, A., Muramatsu, M., 1992. A polyclonal antibody to the rat oestrogen receptor expressed in Escherichia coli: characterization and application to immunohistochemistry. J. Endocrinol. 135, 333–341. Peters, A., Palay, S.L., Webster, Hd., 1991. The Fine Structure of the Nervous System, 3rd ed. Oxford Univ. Press, New York. Phillips, J.K., Goodchild, A.K., Dubey, R., Sesiashvili, E., Takeda, M., Chalmers, J., Pilowsky, P.M., Lipski, J., 2001. Differential expression of catecholamine biosynthetic enzymes in the rat ventrolateral medulla. J. Comp. Neurol. 432, 20–34. Price Jr., R.H., Lorenzon, N., Handa, R.J., 2000. Differential expression of estrogen receptor beta splice variants in rat brain: identification and characterization of a novel variant missing exon 4. Mol. Brain Res. 80, 260–268. Randich, A., Thurston, C.L., Ludwig, P.S., Timmerman, M.R., Gebhart, G.F., 1991. Antinociception and cardiovascular responses produced by intravenous morphine: the role of vagal afferents. Brain Res. 543, 256–270. Razandi, M., Pedram, A., Greene, G.L., Levin, E.R., 1999. Cell membrane and nuclear estrogen receptors (ERs) originate from a single transcript: studies of ERα and ERβ expressed in Chinese Hamster ovary cells. Mol. Endocrinol. 13, 307–319.

177

Rowland, N.E., Fregly, M.J., 1992. Role of gonadal hormones in hypertension in the Dahl salt-sensitive rat. Clin. Exp. Hypertens., A 14, 367–375. Rozen, S., Skaletsky, H., 2000. Primer3 on the WWW for general users and for biologist programmers. Methods Mol. Biol. 132, 365–386. Saeki, Y., Atogami, F., Takahashi, K., Yoshizawa, T., 1997. Reflex control of autonomic function induced by posture change during the menstrual cycle. J. Auton. Nerv. Syst. 66, 69–74. Saleh, T.M., Connell, B.J., 1999. Centrally mediated effect of 17beta-estradiol on parasympathetic tone in male rats. Am. J. Physiol. 276, R474–R481. Saleh, T.M., Connell, B.J., 2000. 17β-estradiol modulates baroreflex sensitivity and autonomic tone of female rats. J. Auton. Nerv. Syst. 80, 148–161. Saleh, M.C., Connell, B.J., Saleh, T.M., 2000a. Autonomic and cardiovascular reflex responses to central estrogen injection in ovariectomized female rats. Brain Res. 879, 105–114. Saleh, M.C., Connell, B.J., Saleh, T.M., 2000b. Medullary and intrathecal injections of 17β-estradiol in male rats. Brain Res. 867, 200–209. Schreihofer, A.M., Guyenet, P.G., 1997. Identification of C1 presympathetic neurons in rat rostral ventrolateral medulla by juxtacellular labeling in vivo. J. Comp. Neurol. 387, 524–536. Shughrue, P.J., Merchenthaler, I., 2001. Distribution of estrogen receptor beta immunoreactivity in the rat central nervous system. J. Comp. Neurol. 436, 64–81. Stornetta, R.L., Guyenet, P.G., 1999. Distribution of glutamic acid decarboxylase mRNA-containing neurons in rat medulla projecting to thoracic spinal cord in relation to monoaminergic brainstem neurons. J. Comp. Neurol. 407, 367–380. Swanson, L.W., 1992. Brain Maps: Structure of the Rat Brain. Elsevier, Amsterdam. Tabori, N.T., Stewart, L.S., Znamensky, V., Romeo, R.D., Alves, S.E., McEwen, B.S., Milner, T.A., 2004. Ultrastructural evidence that androgen receptors are located at extranuclear sites in the rat hippocampal formation. Neuroscience 130, 151–163. Thureson-Klein, A.K., Klein, R.L., Zhu, P.C., 1986. Exocytosis from large dense cored vesicles as a mechanism for neuropeptide release in the peripheral and central nervous system. Scand. Electron. Microsc. 1, 179–187. Turner, C.D., Bagnara, J.T., 1971. General Endocrinology. W.B. Saunders, Philadelphia. Vanderhorst, V.G., Gustafsson, J.A., Ulfhake, B., 2005. Estrogen receptor-alpha and -beta immunoreactive neurons in the brainstem and spinal cord of male and female mice: relationships to monoaminergic, cholinergic, and spinal projection systems. J. Comp. Neurol. 488, 152–179. Verberne, A.J.M., Sartor, D.M., Berke, A., 1999. Midline medullary depressor responses are mediated by inhibition of RVLM sympathoexcitatory neurons in rats. Am. J. Physiol.: Regul., Integr. Comp. Physiol. 276, R1054–R1062. Wagner, C.K., Silverman, A.-J., Morrell, J.I., 1998. Evidence for estrogen receptor in cell nuclei and axon terminals within the lateral habenula of the rat: regulation during pregnancy. J. Comp. Neurol. 392, 330–342. Wagner, A.H., Schroeter, M.R., Hecker, M., 2001. 17beta-estradiol inhibition of NADPH oxidase expression in human endothelial cells. FASEB J. 15, 2121–2130. Wang, G., Dayanithi, G., Kim, S., Hom, D., Nadasdi, L., Kristipati, R., Ramachandran, J., Stuenkel, E.L., Nordmann, J.J., Newcomb, R., Lemos, J.R., 1997. Role of Q-type Ca2+ channels in vasopressin secretion from neurohypophysial terminals of the rat. J. Physiol. 502, 351–363. Wang, G., Dayanithi, G., Newcomb, R., Lemos, J.R., 1999. An R-type Ca(2+) current in neurohypophysial terminals preferentially regulates oxytocin secretion. J. Neurosci. 19, 9235–9241. Wang, G., Zhou, P., Repucci, M.A., Golanov, E.V., Reis, D.J., 2001. Specific actions of cyanide on membrane potential and

178

BR A I N R ES E A RC H 1 0 9 4 ( 2 00 6 ) 1 6 3 –17 8

voltage-gated ion currents in rostral ventrolateral medulla neurons in rat brainstem slices. Neurosci. Lett. 309, 125–129. Wang, G., Anrather, J., Huang, J., Speth, R.C., Pickel, V.M., Iadecola, C., 2004. NADPH oxidase contributes to angiotensin II signaling in the nucleus tractus solitarius. J. Neurosci. 24, 5516–5524. Wassertheil-Smoller, S., Hendrix, S.L., Limacher, M., Heiss, G., Kooperberg, C., Baird, A., Kotchen, T., Curb, J.D., Black, H., Rossouw, J.E., Aragaki, A., Safford, M., Stein, E., Laowattana, S., Mysiw, W.J., 2003. Effect of estrogen plus progestin on stroke in postmenopausal women: the Women's Health Initiative: a randomized trial. JAMA 289, 2673–2684. Watson, C.S., Norfleet, A.M., Pappas, T.C., Gametchu, B., 1999. Rapid actions of estrogens in GH3/B6 pituitary tumor cells via a plasma membrane version of estrogen receptor-alpha. Steroids 64, 5–13. Weiland, N.G., Orikasa, C., Hayashi, S., McEwen, B.S., 1997. Distribution and hormone regulation of estrogen receptor immunoreactive cells in the hippocampus of male and female rats. J. Comp. Neurol. 388, 603–612.

Williams, C.L., Stancel, G.M., 1996. Estrogens and progestins. In: Hardman, J.G., Limbird, L.E., Molinoff, P.B., Ruddon, R.W., Goodman Gilman, A. (Eds.), Goodman & Gilman's The Pharmacological Basis of Therapeutics. McGraw Hill, New York, pp. 1411–1440. Wilson, M.A., 1996. GABA physiology: modulation by benzodiazepines and hormones. Crit. Rev. Neurobiol. 10, 1–37. Woolley, C.S., Schwartzkroin, P.A., 1998. Hormonal effects on the brain. Epilepsia 39 (Suppl. 8), S2–S8. Yang, S.H., Liu, R., Perez, E.J., Wen, Y., Stevens Jr., S.M., Valencia, T., Brun-Zinkernagel, A.M., Prokai, L., Will, Y., Dykens, J., Koulen, P., Simpkins, J.W., 2004. Mitochondrial localization of estrogen receptor beta. Proc. Natl. Acad. Sci. U. S. A. 101, 4130–4135. Zhu, Y., Bian, Z., Lu, P., Karas, R.H., Bao, L., Cox, D., Hodgin, J., Shaul, P.W., Thoren, P., Smithies, O., Gustafsson, J.A., Mendelsohn, M.E., 2002. Abnormal vascular function and hypertension in mice deficient in estrogen receptor beta. Science 295, 505–508.