Effects of estrogen on cerebrovascular function: age-dependent shifts from beneficial to detrimental in small cerebral arteries of the rat

The results of the present study reveal that the cerebrovascular effects of estrogen are distinctly age-dependent in female rats, exerting protective effects in younger animals, but detrimental effects in older animals. These findings may lead to age-specific estrogen pharmacotherapies in women that maximize beneficial and minimize detrimental effects on the cerebrovasculature.

Abstract

In the present study, interactions of age and estrogen in the modulation of cerebrovascular function were examined in small arteries <150 μM. The hypothesis tested was that age enhances deleterious effects of exogenous estrogen by augmenting constrictor prostanoid (CP)-potentiated reactivity of the female (F) cerebrovasculature. F Sprague-Dawley rats approximating key stages of “hormonal aging” in humans were studied: perimenopausal (mature multi-gravid, MA, cyclic, 5–6 mo of age) and postmenopausal (reproductively senescent, RS, acyclic 10–12 mo of age). Rats underwent bilateral ovariectomy and were given estrogen replacement therapy (E) or placebo (O) for 14–21 days. Vasopressin reactivity (VP, 10⁻¹²–10⁻⁷ M) was measured in pressurized middle cerebral artery segments, alone or in the presence of COX-1- (SC560, 1 μM) or COX-2- (NS398, 10 μM) selective inhibitors. VP-stimulated release of prostacyclin (PGI₂) and thromboxane (TXA₂) were assessed by radioimmunoassay of 6-keto-PGF_1α and TXB₂ (stable metabolites). VP-induced vasoconstriction was attenuated in ovariectomized + estrogen-replaced, multigravid adult rats (5–6 mo; MAE) but potentiated in older ovariectomized + estrogen-replaced, reproductively senescent rats (12–14 mo; RSE). SC560 and NS398 reduced reactivity similarly in ovariectomized multigravid adult rats (5–6 mo; MAO) and ovariectomized reproductively senescent rat (12–14 mo; RSO). In MAE, reactivity to VP was reduced to a greater extent by SC560 than by NS398; however, in RSE, this effect was reversed. VP-stimulated PGI₂ was increased by estrogen, yet reduced by age. VP-stimulated TXA₂ was increased by estrogen and age in RSE but did not differ in MAO and RSO. Taken together, these data reveal that the vascular effects of estrogen are distinctly age-dependent in F rats. In younger MA, beneficial and protective effects of estrogen are evident (decreased vasoconstriction, increased dilator prostanoid function). Conversely, in older RS, detrimental effects of estrogen begin to be manifested (enhanced vasoconstriction and CP function). These findings may lead to age-specific estrogen replacement therapies that maximize beneficial and minimize detrimental effects of this hormone on small cerebral arteries that regulate blood flow.

NEW & NOTEWORTHY

The results of the present study reveal that the cerebrovascular effects of estrogen are distinctly age-dependent in female rats, exerting protective effects in younger animals, but detrimental effects in older animals. These findings may lead to age-specific estrogen pharmacotherapies in women that maximize beneficial and minimize detrimental effects on the cerebrovasculature.

it is well known that the risk of developing cardiovascular disease (CVD) or stroke increases with advancing age (5, 36), yet most research models utilize young animals in their studies. Similarly, the beneficial effects of estrogen on systemic and cerebral vascular function are well documented in numerous experimental animal studies (12, 13, 15, 43, 50, 58); however, nearly all of these studies have used younger animals. This raises the question of the relevance of these findings to understanding the pathogenesis of a variety of CVDs in aging human populations.

The beneficial effects of estrogen include a number of direct effects on the vascular wall as well as indirect systemic effects, all of which are believed to reduce the risk for CVDs and improve overall cardiovascular health (for reviews, see Refs. 17, 42, 43, 46). Numerous experimental animal and human clinical studies have demonstrated that in the vascular wall, estrogen promotes endothelial cell growth, inhibits vascular smooth muscle cell growth and migration, and enhances the production of local vasodilators, such as nitric oxide (NO) and prostacyclin (PGI₂), by increasing expression and/or activity of key enzymes: cyclooxygenase (COX), endothelial NO synthase (eNOS), and PGI₂ synthase (PGIS) (17, 42, 43, 46). Beneficial indirect systemic effects of estrogen include formation of a favorable serum lipid profile, inhibition of atherosclerosis, antioxidant and anti-inflammatory activities, and alterations in coagulation and fibrinolytic system functions (17, 42, 43, 46).

In contrast to these beneficial effects of estrogen, more recent human clinical trials (HERS, HERS II, and WHI) reported increased incidences and severity of neurological (dementia and stroke) and vascular [coronary artery disease (CAD), hypertension, venous thrombosis] diseases in women undergoing estrogen replacement therapy (18–20, 22, 23, 69). These discrepancies between experimental animal and more recent human clinical studies further raise the question of relevance of animal studies to this issue. Historically, estrogen has been implicated as a protective hormone because the risks of developing CVDs and stroke increase with the onset of menopause and the concomitant decline in circulating levels of endogenous estrogen (5, 36). However, in recent animal studies, both endogenous and exogenous estrogen enhanced the production of and reactivity to thromboxane (TXA₂) and other deleterious constrictor prostanoids in the systemic vasculature of the female (F) rat, exacerbating vascular tone and blood pressure (32, 33, 55). These recent studies suggest that the effects of estrogen on vascular function are far more complex and question the dogmatic view that estrogen is protective against CVDs and stroke.

The failure to account for age-related changes in the hormonal mechanisms that modulate vascular tone may help explain the paradox between the beneficial effects of estrogen observed in animal studies and the deleterious effects reported in recent human clinical trials. Indeed, recent experimental studies in the F rat revealed that estrogen replacement therapy reduced brain injury and improved neurological recovery following ischemic stroke in younger animals, but enhanced brain injury and reduced neurological recovery in aged animals (56). These data suggest that estrogen may exert age-dependent effects on cerebrovascular function, which are beneficial in younger F animals, but deleterious in older F. Therefore, it is important to determine whether age exacerbates the deleterious effects of estrogen on cerebrovascular function, as suggested by the discrepancies between animal studies and human clinical trials. Notably, increasing age is strongly associated with increases in endothelial dysfunction (5, 29, 61, 64) and with reductions in human cerebral blood flow (71). Thus, there is a strong potential for interactions between age and estrogen in the regulation of cerebrovascular function.

Recent studies reported that age and sex alter cerebrovascular reactivity in intact Sprague-Dawley rats (7, 8). Vasopressin (VP)-induced vasoconstriction of the middle cerebral artery (MCA) was reduced with age in intact F rats but was unchanged in intact male rats. The roles of COX-1- and COX-2-derived prostanoids in modulating VP-induced vasoconstriction of the MCA were age-dependent in F but not in male rats. Additionally, the production of PGI₂ and TXA₂ by the MCA was reduced with age in F rats (8). Thus, in the present study, the interactions between age and estrogen in the modulation of cerebrovascular function were examined in sexually mature (perimenopausal) and older reproductively senescent (postmenopausal) F rats. The MCA was employed as a model of a small muscular artery <150 μM that regulates cerebral blood flow. This study tested the central hypothesis that age enhances the deleterious effects of exogenous estrogen by enhancing constrictor prostanoid-potentiated reactivity of the F cerebrovasculature. These studies are believed to be the first to examine the combined effects of age and estrogen on the vasoconstrictor reactivity to VP in the F cerebrovascular circulation. The aims of the present study were to determine the effects of age and estrogen on 1) cerebrovascular reactivity of the F MCA to VP, a systemic vasoconstrictor hormone important in the regulation in systemic and cerebrovascular function, 2) basal and VP-stimulated production of the primary dilator (PGI₂) and constrictor prostanoids (TXA₂) that modulate cerebrovascular reactivity to VP in the F MCA, and 3) the roles of COX-1- and COX-2-derived prostanoids in modulating cerebrovascular reactivity to VP. The results of the present study reveal that the cerebrovascular effects of estrogen are distinctly age-dependent in F rats, exerting protective effects in younger animals, but detrimental effects in older animals. These findings may lead to age-specific estrogen pharmacotherapies that maximize beneficial and minimize detrimental effects of this hormone on the cerebrovasculature.

MATERIALS AND METHODS

Ethical approval.

All animal protocols were in accordance with the “U.S. Government Principles for the Utilization and Care of Vertebrate Animals Used in Testing, Research and Training”, as detailed in the National Institutes of Health “Guide for the Care and Use of Laboratory Animals”, and were approved by the Texas A&M University Institutional Animal Care and Use Committee.

Animals and maintenance.

F Sprague-Dawley rats of distinctly different age and function were studied. These different groups of F rats are intended to model two key stages of “hormonal age” in women, namely, mature (perimenopausal) and aged (postmenopausal). Mature, multigravid adult F rats (MA, 5–6 mo) are older than the virgin F rats used in most studies, are multiparous (2–3 previous pregnancies), and exhibit a longer and somewhat irregular 5–9 day estrous cycle. Reproductively senescent F rats (RS, 12–14 mo) are older, have carried 5–6 previous pregnancies, and are acyclic. This model has been used successfully to study the effects of age and estrogen on neurological function and excitotoxic injury (2, 25, 26, 47, 56). All rats were purchased from Harlan (Houston, TX) and were housed at the main animal facility at Texas A&M University. Rats were housed in pairs in standard plastic laboratory rat cages in a well-ventilated room and were maintained at constant temperature (21°C–26°C), with controlled photoperiod (12 h light: 12 h dark). Sixteen percent protein global diet (soy and alfalfa-free to minimize dietary phytoestrogens; Harlan, Houston, TX) and water were provided ad libitum.

Ovariectomy and estrogen replacement therapy.

All F rats underwent bilateral ovariectomy (O) using standard surgical methods. Briefly, a dorsomedial incision through the skin was made, subcutaneous fascia was cleared, and 1 cm bilateral dorsolateral incisions were made in the abdominal wall to expose the ovarian fat pads. The ovaries and uterine horns were carefully externalized, ligated and then transected. The ligated uterine horns were swabbed with Betadyne and allowed to retract into the abdominal cavity. The abdominal wall was closed with 4-0 vicryl absorbable suture, and the skin incision was approximated using wound clips. Prior to anesthetic recovery, estrogen (E) replacement therapy was initiated by subcutaneous implantation of three 0.05 mg 17β-estradiol 60-day time release pellets (Innovative Research, Sarasota, FL). This dose produces plasma levels of 17β-estradiol quite similar to the nonsurge levels measured in normally cycling F rats (32, 33, 68). Rats not receiving estrogen replacement were implanted with vehicle-control placebo pellets. All rats were killed for in vitro experiments 14–21 days postsurgery.

Pressurized cannulated MCA vessel preparation.

Rats were humanely euthanized by rapid decapitation to avoid artifactual effects of anesthetics and to minimize activation of neural and humoral pathways. The MCAs were isolated immediately and placed in chilled, Krebs-Henseleit-bicarbonate solution (KHB). The KHB was composed of (in mM) 118.0 NaCl, 25.0 NaHCO₃, 10.0 glucose, 4.74 KCl, 2.5 CaCl₂, 1.18 MgSO₄, and 1.18 KH₂PO₄. MCAs from each animal were cleaned of connective and brain tissue, and arterial segments were prepared in triplicate. MCA segments were cannulated and tied securely to glass micropipettes (75–125-μm diameter) using 11-0 ophthalmic suture. The micropipettes were filled with physiological salt solution (PSS) with albumin, which contained the following (in mM): 145 NaCl, 4.7 KCl, 2.0 CaCl₂, 1.17 MgSO₄, 3.0 MOPS, 1.2 NaH₂PO₄, 5.0 glucose, 2.0 pyruvate, 0.02 ethylenediaminetetraacetic acid, and 1% BSA. The cannulated vessel was then transferred to the stage of an inverted microscope (Olympus CKX41, Olympus, Shinjuku, Japan) equipped with a 4× objective (numerical aperture of 0.13) and coupled with a video camera (Hitachi KP-M3AN, Hitachi, Chiyoda-ku, Japan), video monitor (Pelco PMM12A, Pelco, Clovis, CA), DVD recorder (Phillips DVDR3475, Phillips, Amsterdam, The Netherlands), and video micrometer (Colorado Video 307A, Boulder, CO). Both micropipettes were connected to a single reservoir system and were gradually adjusted to set the intraluminal pressure of the vessel at 85 mmHg without allowing flow through the vessel lumen. Leaks were detected by verifying that intraluminal diameter of the pressurized arteriole remained constant when the valve to the reservoir system was closed. Only arterioles that were free of leaks were studied. The vessel chamber bath (Living Systems TC-09S, Living Systems, Miami, FL) containing PSS + albumin was gradually warmed and maintained at 37°C for the duration of the experiment. Luminal diameter was monitored continuously throughout the experiment. The vessels were allowed to equilibrate for 1 h before being pretreated with pharmacological agents indicated below for 20 min. Cumulative concentration-response curves to arginine vasopressin (VP; 10⁻¹² to 10⁻⁷ mol/l) were obtained by direct, cumulative additions of VP into the tissue baths, in the absence or presence of inhibitors, including 1) selective COX-1 inhibitor (SC560, 1 μM) or 2) selective COX-2 inhibitor (NS398, 10 μM). Diameter measurements were determined in response to cumulative concentrations of VP. Percent constriction was determined by the following equation: % constriction = (B_D − B_X)/B_D·100, where B_D is the steady-state baseline diameter after inhibitor incubation and B_X is the diameter after each VP concentration. The concentration of VP that produced 50% of the maximal response (EC₅₀) was calculated individually from the log concentration-response curve of each MCA segment.

Prostanoid release assay (TXA₂ and PGI₂).

Vascular prostanoid production by the MCA was measured using incubation and radioimmunoassay methods adapted for microvessels, as described previously (32). Briefly, isolated MCA (3–4-mm axial length) were cleaned of all connective tissue and fat and placed into chilled PSS without BSA (PSS-BSA) to rest for 60 min. The arteries were then transferred into 0.5-ml microcentrifuge tubes with 450-μl chilled solution and gradually warmed in a water bath to 37°C for a 45-min preincubation. The preincubation medium was carefully aspirated, and 300 μl PSS-BSA alone (basal) or PSS-BSA with VP 10⁻⁹ M (low) or PSS-BSA with VP 10⁻⁷ M (high) was added and incubated at 37°C for 45 min. After incubation, the incubation media were collected and stored at −80°C until RIA of stable metabolites of PGI₂ (6-keto-PGF_1α) and TXA₂ (TXB₂). MCA segments were saved and stored at −80°C for dry weight analysis.

Chemical reagents and drugs.

The following reagents and drugs were used: 17β-estradiol (Innovative Research of America, Sarasota, FL), SC560 and NS398 (Cayman Chemical, Ann Arbor, MI), arginine VP (Bachem, Torrance, CA). All other chemicals were purchased from Sigma Chemical (St. Louis, MO). Stock solutions of all drugs were prepared fresh daily except for VP (which was diluted daily from aliquots of 1 × 10⁻³ M stock solution stored at −80°C) and NS (which was diluted daily from aliquots of 1 mg/ml stock solution stored at −20°C).

Statistics.

All data are expressed as means ± SE; n indicates the number of animals studied. One- or two-way ANOVAs were used to detect significant differences among means of all experimental groups. If a main effect was identified, pairwise Student's t-tests were performed to detect significant differences between any two means of the data groups, with a Bonferroni correction for multiple comparisons. Vascular function and prostanoid release data were analyzed using a two-way ANOVA for estrogen (O vs. E) and age (MA vs. RS). The effects of treatment [Control (CTL), COX-1 inhibition, COX-2 inhibition] were analyzed in each experimental group using a one-way ANOVA with Bonferroni multiple-comparison correction. Plasma estradiol levels, body weight, and uterine weight were analyzed by estrogen and age using a two-way ANOVA and Student's t-tests. A P value < 0.05 was considered significant.

RESULTS

Effects of age and estrogen levels on body weight and uterine weight.

Plasma 17β-estradiol concentrations, body weights, and uterine weights are summarized in Table 1. Both younger MA and older RS F that were ovariectomized and given estrogen replacement (MAE, RSE) had significantly lower body weights and significantly greater uterine weights compared with ovariectomized F of the same age (MAO, RSO). Plasma estradiol levels followed the same trends; in MAO and RSO, ovariectomy dramatically reduced estradiol levels (≥95% compared with typical values for intact F), while estrogen replacement restored plasma estradiol to physiological, nonsurge levels in both MAE and RSE.

Table 1.

Plasma 17β-estradiol concentrations, body weights, and uterine weights of MAO, MAE, RSO, and RSE female rats

Group	Estradiol, pg/ml n = 13–15	Body Weight, g n = 13–15	Uterine Weight, g/100 g body wt n = 12–14
MAO	1.9 ± 1.3a	308.9 ± 12.9b	0.14 ± 0.03a
MAE	33.6 ± 7.3b	259.9 ± 6.9a	0.24 ± 0.02b
RSO	0.2 ± 0.1a	339.6 ± 4.1c	0.10 ± 0.02a
RSE	37.4 ± 8.4b	302.8 ± 7.0b	0.29 ± 0.03b

Values are expressed as means ± SE; n = number of animals studied. The rats were divided into four groups: mature multigravid adult female rats (MA; 4–6 mo), either ovariectomized (MAO) or ovariectomized and estrogen-replaced (MAE), and reproductively senescent female rats (RS; 10–12 mo), either ovariectomized (RSO) or ovariectomized and estrogen-replaced (RSE). ^a–c0.0001 ≤ P ≤ 0.04, values within each column (estradiol, body weight, uterine weight) with different superscripts are significantly different (MAO vs. MAE vs. RSO vs. RSE).

Effects of age and estrogen on vascular reactivity to VP.

The effects of age and estrogen on VP-induced vasoconstriction are shown in Fig. 1, Fig. 2, and Table 2. Comparison of the control concentration-response curves to VP among the four experimental groups (MAO, MAE, RSO, RSE) revealed clear age- and estrogen-dependent differences, which differed significantly at both middle- and maximal-VP concentrations (Fig. 1). In MA rats, estrogen replacement reduced reactivity of MCA to VP throughout the concentration-response curve (21% at maximal VP); in sharp contrast, in RS rats, estrogen replacement increased reactivity throughout the concentration-response curve (27% at maximal VP).

Fig. 1.

Concentration-response curves for vasopressin (VP) in endothelium-intact pressurized middle cerebral artery segments prepared from MAO, MAE, RSO, and RSE Sprague-Dawley female rats. Mature multigravid adult female rats (MA, 4–6 mo), either ovariectomized (MAO) or ovariectomized and estrogen-replaced (MAE), and reproductively senescent female rats (RS; 10–12 mo), either ovariectomized (RSO) or ovariectomized and estrogen-replaced (RSE). Data points represent means ± SE (n = 6 or 7 rats/group). MAO, MAE, RSO, and RSE were compared statistically; ^a–f0.0001 ≤ P ≤ 0.009, mean values without common superscript differ significantly at middle and maximal concentrations of VP. At middle VP, MAE, and RSO differ significantly from MAO and RSE, and MAO differs significantly from RSE. At maximal VP, MAE differs significantly from MAO, RSO, and RSE. MAO and RSO do not differ.

Fig. 2.

Concentration-response curves for VP in endothelium-intact pressurized middle cerebral artery segments prepared from MAO, MAE, RSO, and RSE female Sprague-Dawley rats in the presence of selective COX inhibitors SC560 (COX-1; 1 μM), NS398 (COX-2; 10 μM), or vehicle-control (CTL). Vessels were prepared in triplicate from each experimental group: mature multigravid adult female rats (MA, 4–6 mo.), either ovariectomized (MAO) (A) or ovariectomized and estrogen-replaced (MAE) (B), and reproductively senescent female rats (RS, 10–12 mo), either ovariectomized (RSO) (C) or ovariectomized and estrogen-replaced (RSE) (D). Data points represent means ± SE (n = 6 or 7 rats/group). CTL, SC560 (COX-1), and NS398 (COX-2) treatments were compared statistically for each experimental group (i.e., MAO, MAE, RSO, or RSE). ^a–fP ≤ 0.0001, mean values without common superscript differ significantly at middle and maximal concentrations of VP. For MAO, CTL differs significantly from SC560 (COX-1) and NS398 (COX-2) at both middle and maximal VP. SC560 (COX-1) and NS398 (COX-2) differ significantly at middle VP only. For MAE, CTL, SC560 (COX-1), and NS398 (COX-2), all differ significantly from one another at both middle and maximal VP. For RSO, NS398 (COX-2) differs significantly from CTL and SC560 (COX-1) at middle VP; CTL differs significantly from SC560 (COX-1) and NS398 (COX-2) at maximal VP. For RSE, CTL, SC560 (COX-1), and NS398 (COX-2) all differ significantly from one another at both middle and maximal VP.

Table 2.

Vasoconstrictor reactivity to vasopressin in rat middle cerebral arteries at middle- and maximal-VP concentrations and potency (EC₅₀) in MAO, MAE, RSO, and RSE female rats

Group n = 6 or 7	Middle VP Concentration, 10−9.5 M, % Constriction	Maximal VP Concentration, 10−7 M, % Constriction	EC50, nM
Control
MAO	40.2 ± 1.8b	63.6 ± 0.7b	0.14 ± 0.04b
MAE	23.6 ± 1.3a	50.1 ± 0.9a	0.40 ± 0.15b,c
RSO	26.2 ± 1.4a	59.3 ± 1.1b	0.40 ± 0.04c
RSE	52.0 ± 1.9c	75.4 ± 1.1c	0.07 ± 0.01a
SC560
MAO	29.0 ± 1.3b	54.5 ± 1.1c	0.37 ± 0.07c
MAE	4.7 ± 0.5a	20.8 ± 0.7a	2.72 ± 0.41d
RSO	25.7 ± 0.8b	45.6 ± 1.0b	0.21 ± 0.02b
RSE	30.7 ± 1.4b	47.9 ± 1.0b	0.13 ± 0.04a
NS398
MAO	23.6 ± 0.9c	53.0 ± 0.8d	0.46 ± 0.10a
MAE	13.4 ± 1.2a	32.8 ± 0.2b	0.98 ± 0.40a
RSO	17.4 ± 1.1b	43.9 ± 1.3c	1.00 ± 0.26a
RSE	10.7 ± 0.9a	26.8 ± 1.7a	1.02 ± 0.42a

Values are means ± SE and are derived from Fig. 2; n = no of animals studied. Middle cerebral arteries were prepared in triplicate and pretreated with selective inhibitors of either COX-1 (SC560, 1 μM) or COX-2 (NS398, 10 μM) or vehicle control. Experimental groups (MAO, MAE, RSO, RSE) were compared statistically within each treatment [Control, SC560 (COX-1) and NS398 (COX-2)]. ^a–d 0.0001 ≤ P ≤ 0.015, values within each column [middle vasopressin (VP) concentration, maximal VP concentration, EC₅₀] with different superscripts are significantly different (e.g., Control: MAO vs. MAE vs. RSO vs. RSE; SC560: MAO vs. MAE vs. RSO vs RSE; NS398: MAO vs. MAE vs. RSO vs. RSE).

In ovariectomized F rats, age reduced reactivity of the MCA to lower concentrations of VP. Thus, in older RSO rats, VP-induced vasoconstriction was significantly lower at mid-VP concentration compared with younger MAO; however, MAO and RSO did not differ at max-VP concentration (Fig. 1). Sensitivity to VP (EC₅₀) was also significantly lower in RSO compared with MAO (Table 2). Estrogen replacement reversed the effects of age on reactivity of the MCA to VP. Thus, maximal vasoconstriction to VP was augmented dramatically in older RSE F (50.6%), compared with younger MAE F (Fig. 1). In contrast, estrogen replacement attenuated VP-induced vasoconstriction by 21% at maximal VP concentration in younger MAE F, compared with ovariectomized MAO F, while sensitivity to VP (EC₅₀) was reduced even more dramatically (nearly threefold) in MAE compared with MAO (0.14 ± 0.04 vs. 0.40 ± 0.15 nM). Interestingly, in older RSE rats, estrogen replacement augmented VP-induced vasoconstriction of the MCA by 27% compared with RSO F, and increased sensitivity to VP (EC₅₀) by nearly sixfold in the older RSE F rats (0.40 ± 0.04 vs. 0.07 ± 0.01 nM).

In the ovariectomized mature adult rats (MAO), VP produced concentration-dependent vasoconstriction with a maximal response of 63.6 ± 0.7% and an EC₅₀ of 0.14 ± 0.04 nM. Both COX-1- (SC560) and COX-2- (NS398) selective inhibitors significantly attenuated vasoconstriction at both middle and maximal concentrations of VP (Fig. 2A). Compared with the control MAO group, maximal vasoconstriction was reduced by 14% and 17% by SC560 and NS398, respectively.

In estrogen-replaced mature adult rats (MAE), concentration-dependent vasoconstriction to VP was reduced 21% at maximal VP compared with MAO, with a maximal response of 51 ± 0.9% and an EC₅₀ of 0.40 ± 0.15 nM (Fig. 2B). Both COX-1 (SC560) and COX-2 (NS398) selective inhibitors attenuated vasoconstriction substantially more in MAE than in MAO at both middle and maximal VP concentrations. Compared with the control MAE group, maximal vasoconstriction was reduced to a noticeably greater extent by SC560 (59%) than by NS398 (34%), suggesting a greater role for COX-1-derived constrictor prostanoids in MAE.

In the older ovariectomized rats (RSO), maximal response to VP was 59.3 ± 1.1%, with an EC₅₀ of 0.40 ± 0.04 nM (Fig. 2C). Similar to MAO, in RSO, SC560 and NS398 both reduced VP-induced vasoconstriction in RSO to a similar extent, by 23% and 26%, respectively, at maximal VP, although the effects of COX inhibition were greater in older RSO (23–26%) than in younger MAO (14–17%).

In the older estrogen-replaced rats (RSE), both maximal response (75.4 ± 1.1%) and sensitivity to VP (EC₅₀, 0.07 ± 0.01 nM) were noticeably higher than the other three experiment groups (Figs. 1 and 2D). Both COX-1- (SC560) and COX-2- (NS398)-selective inhibitors dramatically attenuated vasoconstriction at both middle and maximal VP concentrations. However, compared with the control RSE group, maximal vasoconstriction was reduced to a much greater extent by NS398 (65%) than by SC560 (36%), suggesting a substantially greater role for COX-2-derived constrictor prostanoids in the older RSE group.

Effects of age and estrogen on basal and VP-stimulated PGI₂ production.

Basal and VP-stimulated (low concentration 10⁻⁹ M; or high concentration 10⁻⁷ M) release of 6-keto-PGF_1α (stable metabolite of PGI₂) are shown in Fig. 3. Basal release of 6-keto-PGF_1α did not differ significantly among the four experimental groups (MAO, MAE, RSO, or RSE). Within each group, VP increased 6-keto-PGF_1α production in a concentration-dependent manner. Low-concentration VP increased 6-keto-PGF_1α production approximately threefold in all groups. Age had no significant effects on a low concentration of VP-stimulated 6-keto-PGF_1α production in ovariectomized or estrogen-replaced F (MAO vs. RSO and MAE vs. RSE); however, estrogen replacement increased 6-keto-PGF_1α production significantly in both ages (52% in MAE vs. MAO and 45% in RSE vs. RSO). With advancing age, high-concentration VP-stimulated production of 6-keto-PGF_1α decreased significantly with estrogen replacement (20% in RSE vs. MAE); in the absence of estrogen, there were no significant differences in RSO vs. MAO. Estrogen replacement increased high concentration VP-stimulated 6-keto-PGF_1α release markedly in both mature (MA) and older (RS) age groups (58% in MAE vs. MAO and 56% in RSE vs. RSO). High-concentration VP increased 6-keto-PGF_1α production more than sixfold in MAE and nearly fivefold in MAO, RSO, and RSE from their respective basal levels. At high-concentration VP, MAE produced significantly more 6-keto-PGF_1α than any of the other groups (by 25% in RSE, 58% in MAO, and 94% in RSO), suggesting that estrogen exerts a much greater stimulatory effect on 6-keto-PGF_1α production by the F rat MCA than the minor inhibitory effect of age.

Fig. 3.

Basal and VP-stimulated (low concentration, 10⁻⁹ M; or high concentration 10⁻⁷ M) release of 6-keto-PGF_1α by middle cerebral artery segments from MAO, MAE, RSO, and RSE female Sprague-Dawley rats. Mature multigravid adult female rats (MA; 4–6 mo.), either ovariectomized (MAO) or ovariectomized and estrogen-replaced (MAE), and reproductively senescent female rats (RS; 10–12 mo.), either ovariectomized (RSO) or ovariectomized and estrogen-replaced (RSE). Values are means ± SE; n = 6 rats/group. Basal, low-VP, and high-VP prostanoid release were compared statistically within each experimental group (i.e., MAO, MAE, RSO, or RSE). ^a–cP ≤ 0.0001, mean values within groups (MAO, MAE, RSO, RSE) without common superscript are significantly different. Basal, low-VP, and high-VP all differ significantly from one another in each experimental group. Basal, low-VP, or high-VP treatments were compared statistically among the four experimental groups (i.e., MAO vs. MAE vs. RSO vs. RSE). *^,#,+0.0001 ≤ P ≤ 0.02 mean values between groups (MAO vs. MAE vs. RSO vs. RSE) with different superscripts are significantly different. Basal prostanoid release did not differ among the four experimental groups. In low-VP prostanoid release, MAO and RSO differ significantly from MAE and RSE. In high-VP prostanoid release, MAO and RSO differ significantly from MAE and RSE, and MAE and RSE differ from each other.

Effects of age and estrogen on basal and VP-stimulated TXA₂ production.

Basal and VP-stimulated (low concentration: 10⁻⁹ M; high concentration: 10⁻⁷ M) release of TXB₂ (stable metabolite of TXA₂) are shown in Fig. 4. Basal release of TXB₂ did not differ significantly among the four experimental groups (MAO, MAE, RSO, or RSE). VP-stimulated TXB₂ production increased in a concentration-dependent manner in both MAE and RSE, but not in MAO or RSO. Thus, in MAO and RSO, there were no differences between basal and low-concentration VP-stimulated TXB₂ production; however, estrogen replacement increased low concentration VP-stimulated TXB₂ production twofold compared with basal, in both MAE and RSE, indicating increased sensitivity to VP with estrogen replacement. Age had no significant effect on low-concentration VP-stimulated TXB₂ release in ovariectomized or estrogen-treated groups (MAO vs. RSO and MAE vs. RSE); however, estrogen replacement increased TXB₂ release markedly in both younger (90% in MAO vs. MAE) and older (78% in RSO vs. RSE) F rats. Low-concentration VP-stimulated TXB₂ production did not differ between MAO and RSO; however, age and estrogen replacement both increased TXB₂ production to a greater extent in older RSE (150%) compared with MAE (104%). High-concentration VP-stimulated TXB₂ release increased significantly with estrogen in both mature (MA) and older (RS) F rats (65% in MAE vs. MAO and 165% in RSE vs. RSO). High-VP concentration increased TXB₂ production threefold in MAO, fourfold in MAE, threefold in RSO, and sevenfold in RSE from their respective basal levels. At high-concentration VP, RSE produced significantly more TXB₂ than all other groups (by 29% in MAE, 108% in MAO, and 165% in RSO). These findings suggest that age and estrogen interact synergistically to enhance TXB₂ production by the F rat MCA in response to VP.

Fig. 4.

Basal and VP-stimulated (low concentration, 10⁻⁹ M; or high concentration 10⁻⁷ M) release of TXB₂ by middle cerebral artery segments from MAO, MAE, RSO, and RSE female Sprague-Dawley rats. Mature multigravid adult female rats (MA; 4–6 mo.), either ovariectomized (MAO) or ovariectomized and estrogen-replaced (MAE), and reproductively senescent female rats (RS; 10–12 mo.), either ovariectomized (RSO) or ovariectomized and estrogen-replaced (RSE). Values are means ± SE; n = 6 rats/group. Basal, low-VP, and high-VP prostanoid release were compared statistically within each experimental group (i.e., MAO, MAE, RSO, or RSE). ^a–c0.0001 ≤ P ≤ 0.003, mean values within groups (MAO, MAE, RSO, RSE) without common superscript are significantly different. Basal, low-VP, and high-VP all differ significantly from one another in MAE and RSE; basal and low-VP differ significantly from high-VP in MAO and RSO. Basal, low-VP, or high-VP treatments were compared statistically among the four experimental groups (i.e., MAO vs. MAE vs. RSO vs. RSE). *^,#,+0.0001 ≤ P ≤ 0.003 mean values between groups (MAO vs. MAE vs. RSO vs. RSE) with different superscripts are significantly different. Basal prostanoid release did not differ among the four experimental groups. In low-VP prostanoid release, MAO and RSO differ significantly from MAE and RSE. In high-VP prostanoid release, MAO and RSO differ significantly from MAE and RSE, and MAE and RSE differ from each other.

DISCUSSION

In the present investigation, the effects of age and exogenous estrogen on cerebrovascular reactivity and prostanoid production were examined in the MCA of ovariectomized and estrogen-replaced F Sprague-Dawley rats. The results reveal that estrogen is an important regulator of the vasoconstrictor responses of the MCA to VP in the F rat and that it exerts opposing effects on cerebrovascular reactivity with advancing age. The major findings of this study are that 1) estrogen treatment had age-dependent divergent effects on VP-induced cerebrovascular vasoconstriction; 2) age altered the specific COX isoform and resultant constrictor/dilator prostanoid production enhanced by estrogen; and 3) estrogen enhanced VP-stimulated PGI₂ and TXA₂ production in both MA and RS, while age decreased VP-stimulated PGI₂ and increased in TXA₂ production in older RS F with estrogen replacement. Taken together, these data reveal that the vascular effects of estrogen are distinctly age-dependent. In younger ovariectomized MA (perimenopausal) F with estrogen replacement, the beneficial and protective effects of estrogen are clearly expressed (decreased vasoconstriction, increased dilator prostanoid function). Conversely, in older ovariectomized RS (postmenopausal) F with estrogen replacement, the detrimental effects of estrogen begin to be manifested (enhanced vasoconstriction, increased constrictor prostanoid function). Plasma concentrations of 17β-estradiol were attenuated substantially by ovariectomy in both age groups. Estrogen replacement therapy restored plasma estradiol levels to concentrations that did not differ between MA or RS groups and are consistent with endogenous, nonsurge estradiol levels in intact F rats reported previously (32, 33, 68). The present findings contribute to the understanding of the interactive effects of age and estrogen in the modulation of cerebrovascular function. Figure 5 is a schematic diagram of a proposed model of cerebrovascular endothelial and VSM interactions, based upon the present findings, which aids in summarizing the interactions between age and estrogen in the regulation of cerebrovascular tone in the F rat.

Fig. 5.

Schematic diagram of a proposed model of the interactions between age and estrogen in the regulation of cerebrovascular function, based upon the present studies. The distinctly age-dependent effects of estrogen in the cerebrovasculature of the female rat are depicted in this model. In younger MA, beneficial and protective effects of estrogen are evident (decreased vasoconstriction, increased dilator prostanoid function). In older RS, detrimental effects of estrogen begin to be manifested (enhanced vasoconstriction and constrictor prostanoid function). AA, arachidonic acid; COX-1, cyclooxygenase-1; COX-2, cyclooxygenase-2; MA, mature multigravid adult; MCA, middle cerebral artery; PGI₂, prostacyclin; RS, reproductively senescent; TXA_2, thromboxane; V₁, vasopressin-1 receptor; V₂, vasopressin-2 receptor; VP, vasopressin.

Effects of age on cerebrovascular reactivity.

Interestingly, in the current study, advancing age had little or no effect on VP-induced vasoconstriction in ovariectomized rats. However, VP-stimulated PGI₂ production declined, while COX-2-derived TXA₂ production increased dramatically in older RS F with estrogen replacement, suggesting age- and estrogen-dependent changes in endothelial function, particularly prostanoid production.

Endothelial dysfunction occurs earlier in men than in women (age 40 vs. 55) (5). Aging is associated with a reduction in endothelium-dependent vasodilation in both experimental animal and human models via numerous mechanisms related to NO-mediated dilation (29, 61, 64). Additionally, consistent with the present findings, numerous studies have demonstrated enhancement of constrictor prostanoids with advancing age, via upregulation of both COX-1 and COX-2 expression (21, 27, 59, 62). In numerous disease states and with advancing age, PGI₂ receptor (IP) expression is also reduced, which reduces the vasodilatory efficacy of endogenously produced PGI₂ (48, 62). Furthermore, there is also indirect evidence suggesting that untransformed PGH₂ [which also causes vasoconstriction via the TXA₂ receptor (TP)] is augmented due to COX-1/COX-2 upregulation (6, 34, 51), and several studies have reported increases in TXA₂ and TXS mRNA in aorta and mesenteric arteries with age (38, 62). Thus, consistent with previous studies, the present findings demonstrate that the balance of dilator to constrictor prostanoids is altered with age, including reductions in the release of and/or reactivity to dilator PGI₂ and augmentation in the release of and/or reactivity to constrictor prostanoids PGH₂ and TXA₂. The mechanism(s) underlying these age-dependent changes in the roles of COX-1 vs. COX-2 isoforms are unclear; however, cellular senescence-associated changes such as telomeric shortening and/or upregulation of cyclin-dependent kinase inhibitor p16 have been associated with age-related pathologies, including cardiovascular disease, diabetes, and Alzheimer's disease (for review, see Ref. 24). Thus, it is possible that senescence-associated changes in the expression of COX isoforms in the F MCA are involved.

Effects of estrogen on mechanisms of cerebrovascular reactivity.

The present findings reveal that estrogen is an important regulator of VP-induced vasoconstriction in the F rat MCA and that it clearly exerts divergent effects on the modulation of cerebrovascular function with advancing age, enhancing beneficial vasodilator effects in younger F, while augmenting deleterious vasoconstrictor effects in older F. The second major finding is that estrogen enhanced the roles of COX-1- and COX-2-dependent prostanoid pathways in VP-induced cerebrovascular vasoconstriction in an age-dependent manner. Thus, while COX-1-dependent vasodilatory prostanoids dominated in younger MA F, this effect of estrogen was reversed with advancing age, revealing a dominant role for COX-2-derived constrictor prostanoids in older RS F. In support of the present findings, a variety of previous studies reported that estrogen enhances the production and/or the sensitivity of cerebral arteries to vasodilatory factors, such as NO and PGI₂ and shifts the balance of prostanoid production toward greater production of vasodilator forms (15, 16, 49, 50, 52, 58). More specifically, in cerebral vessels of young rats, estrogen elevates both COX-1 and PGIS, enhancing PGI₂ production (15, 49, 50). Interestingly TXA₂ production is also elevated slightly but significantly in young estrogen-treated animals, perhaps due to increased expression or activity of COX-1 (50). However, the present studies reveal important age-dependent changes in the effects of estrogen on the patterns of constrictor/dilator prostanoid production and COX isoform activity in the MCA of F rats. The combined effects of age and estrogen replacement enhance both COX-2-derived constrictor prostanoid production and VP-induced vasoconstriction of the F rat MCA.

Taken together with the functional data, the attenuation of VP-induced vasoconstriction in younger MA F observed in the present study appears to be due to estrogen-enhanced production of PGI₂, while the age-dependent augmentation of vasoconstriction in older RS F is due to enhanced production of TXA₂ and/or PGH₂. These data provide clear evidence that estrogen exerts different effects depending upon age. In younger F, estrogen replacement therapy is beneficial; while, in older F, it is detrimental. These findings may help to explain why estrogen is beneficial and neuroprotective in younger animals and women but deleterious in older animals and postmenopausal women. The age- and estrogen-dependent shifts in dilator and constrictor prostanoids observed in the present study reveal important new and novel roles for estrogen and prostanoids in cerebrovascular function during aging that have not been previously reported.

Beneficial vs. deleterious effects of age and estrogen on cerebrovascular reactivity.

Earlier findings from human epidemiological (31, 44) and experimental animal studies (13, 28, 57) led to the dogma that estrogen replacement therapy exerts beneficial effects on neurological and cardiovascular health, and that estrogen is protective against diseases such as dementia, CAD, hypertension, and stroke. An abundance of evidence from experimental animal studies has established that estrogen does exert beneficial or protective effects on the cerebrovasculature by reducing vascular reactivity and thereby increasing blood flow through NO- and vasodilator prostanoid-dependent mechanisms (15, 16, 40, 41, 49, 50). In contrast, more recent human epidemiological findings such as the HERS (19, 23), HERS II (18, 22), and WHI studies (20, 69) all suggest that in older women, estrogen replacement therapy increases the incidences of neurological (dementia and stroke) and vascular diseases (CAD, hypertension, and venous thrombosis). Thus, the role of estrogen in cardiovascular health and disease has become controversial.

Numerous studies have revealed that age-related changes in vascular function are due, at least in part, to increases in oxidative stress. For example, advancing age impairs eNOS-dependent vasodilation by increases in superoxide formation via activation of NADPH oxidase in pial arterioles from male rats (39). Similar studies in the cerebral vasculature of young and aged F rats revealed that estrogen replacement therapy reduces bacterial LPS-induced increases in inducible nitric oxide synthase (iNOS) and COX-2 formation significantly in younger but not in older animals (60). Since both iNOS and COX-2 can lead to the formation of reactive oxygen species (ROS), vasoconstrictor tone of cerebral vessels would be enhanced by estrogen in aged compared with younger F rats. Age-dependent decreases in the coupled state of NOS may also occur. In the uncoupled state, NOS catalyzes NADPH to superoxide ion, rather than producing NO. Additionally, there is evidence that aging reduces the levels of l-arginine (4) and tetrahydrobiopterin (10), providing a greater potential for uncoupling of NOS and increased production of superoxide (70). These findings provide a clear mechanistic explanation for an age-dependent change in the effects of estrogen on the vascular wall, from the beneficial expression of eNOS and formation of NO in younger animals to the uncoupling of eNOS and the deleterious formation of superoxide in older animals. Indeed, preliminary experiments suggest that estrogen exerts beneficial effects on the F rat MCA by suppressing formation of superoxide and reducing vasoconstrictor responses to VP in younger MA females, while in older RS F, estrogen exerts deleterious effects by exacerbating the formation of superoxide and vasoconstrictor responses to VP (7). In summary, in younger F, estrogen leads to the formation of beneficial or protective substances, such as NO and vasodilatory prostanoids; however, with advancing age, estrogen therapy may result in the increased formation of detrimental substances, including vasoconstrictor prostanoids, superoxide, and other ROS.

Relevance to cerebrovascular health and disease.

The enhanced release of NO and other vascular mediators (PGI₂, TXA₂, etc.) and upregulation of their associated enzymes occur via increased shear stress, hormones, exercise, and diet, while decreased secretion and/or downregulation occurs via oxidative stress, smoking, obesity, and vascular diseases (67). Endothelial dysfunction is exhibited in many disease states, including CAD, diabetes, and hypertension, and is also associated with advancing age. Mechanistically, endothelial dysfunction is associated with decreased NO bioactivity, increased superoxide and TXA₂ production, attenuation of endothelium-dependent dilation, increased vascular tone, and atherogenesis (14, 37, 65–67). A decrease in NO bioavailability (decreased synthesis or increased degradation) and alteration in the balance of constrictor vs. dilator prostanoids clearly alters both tone and overall vascular health.

Since the late 1990s, findings from both experimental animal studies and human clinical trials have repeatedly revealed that estrogen exerts neuroprotective effects on cerebral ischemia or stroke (11, 35, 53). Younger premenopausal women are protected from ischemic stroke compared with males, yet this protective effect is lost after menopause (31, 44). Further, premenopausal women experience less damage and greater functional and cognitive recovery from neurologic insult than do males (3). Because of these findings, as well as supporting evidence from experimental animal studies, exposure to estrogen was widely considered to be neuroprotective. Indeed, by the year 2000, an estimated 10 million women were receiving HRT for the alleviation of menopausal symptoms with the belief that estrogen would also exert beneficial cardioprotective and neuroprotective effects, reducing the incidences of CAD, dementia, hypertension, and stroke. However, after only a few years, reports from the WHI clinical trials revealed that estrogen therapy significantly increased the incidence and severity of cardiovascular and neurological diseases (20, 54, 69). These findings led to a major controversy on the effects of estrogen on women's health, resulting in substantial reductions in the use of estrogen replacement therapy in older postmenopausal women. Thus, it is important to elucidate how and why estrogen exerts beneficial cardioprotective and neuroprotective effects in animal models and younger women, but deleterious effects in older postmenopausal women.

More recent experimental animal studies reveal that estrogen exerts divergent effects on neurological and neurovascular functions, which are clearly age-dependent in nature. Using the same aging model as the present study [i.e., younger reproductively mature (MA) and older (RS) F rats], MCA occlusion was used to determine the effects of age and estrogen on stroke. These studies revealed that estrogen reduced the severity of cortical-striatal stroke damage and improved recovery in younger MA F rats, while it enhanced stroke damage and reduced recovery in older RS F rats (56). Similar age-dependent differences in the effects of estrogen were also observed in N-methyl-d-aspartate (NMDA)-induced local inflammation in the forebrain of the F rat; thus, estrogen exacerbated NMDA-induced damage in older RS F rats compared with their younger MA counterparts (47). These findings are entirely consistent with the age-dependent differences in the effects of estrogen on cerebrovascular reactivity to VP observed in the present study, as well as the effects of age and sex on cerebrovascular function reported in another recent study (8). Taken together, these findings reveal that estrogen exerts clear age-dependent divergent effects on neurological and cerebrovascular function in the F rat brain, which are beneficial in younger animals, but deleterious in older animals.

The question arises as to the relevance of VP-induced vasoconstriction of the MCA observed in the present study to the regulation of cerebrovascular function. Magnetic resonance imaging studies reveal that the middle cerebral and basilar arteries carry the largest single fractions of total cerebral blood flow (CBF) in humans (21% and 20%, respectively); thus, changes in their tone should impact CBF significantly (71). These studies also revealed that CBF declines with age in all of the major cerebral arteries, perhaps due to changes in cerebrovascular tone and autoregulation. The vascular (V_1a) VP receptor is widely distributed throughout the brain in neurons, astrocytes, vascular endothelial and smooth muscle cells, and choroid plexus (30), and VP causes vasoconstriction of human as well as cat, goat, and rat cerebrovasculatures (9, 30). Interestingly, plasma and cerebrospinal VP levels are elevated following cerebral ischemia, subarachnoid hemorrhage, and traumatic brain injury, both in human patients and experimental animal models (30, 63). The enhanced VP-induced vasoconstriction that likely follows brain injuries may, therefore, lead to enhanced tissue damage and brain edema due to the loss of CBF autoregulation (63). Thus, VP may influence CBF and autoregulatory function both under normal conditions and in pathophysiological states following brain injury. Indeed, administration of VP antagonists or antiserum alleviates the acute cerebral vasospasm and edema that follow experimental subarachnoid hemorrhage (9, 30).

On the basis of the forgoing discussion, the enhanced VP-induced vasoconstriction and constrictor prostanoid function observed in older estrogen-treated F rats in the present study should be viewed as deleterious and may be relevant to the negative prognosis for ischemic stroke observed in human patients with impaired collateral cerebral circulation (45). Similarly, it has been suggested that increased COX-2 activity may account for the observed role of oxygen free radicals in ischemia-associated cerebrovascular derangement (1). This suggestion is consistent with the enhanced role of COX-2-derived constrictor prostanoids observed in the MCA of older estrogen-treated F rats in the present study and the preliminary findings that oxygen free radicals enhance VP-induced vasoconstriction in this group, but not in the MCA of younger estrogen-treated F rats (7). Taken together, these past and present findings provide compelling evidence that increasing age and estrogen produce synergistic effects that are deleterious to the F cerebrovasculature, which could exacerbate tissue damage following brain injuries.

Perspectives and Significance

Although multiple studies have demonstrated beneficial effects of estrogen replacement on the cerebral vasculature of young animals (13, 15, 42, 43, 49, 50), there is a clear conflict in the beneficial vs. deleterious effects of estrogen observed in human clinical trials vs. experimental animal studies. The present study is the first designed to compare the effects of estrogen replacement on cerebrovascular function in younger vs. older F rats. The results of this study reveal important new and novel information on the effects of estrogen and advancing age on cerebrovascular function and help to explain the apparent conundrum of beneficial vs. deleterious effects of estrogen on cerebrovascular function. Further understanding of the mechanisms underlying the effects of age and estrogen on the regulation of cerebrovascular function will aid in our understanding of pathophysiological states such as stroke and traumatic brain injury, and in the development of more effective age- and sex-dependent pharmacotherapies for stroke and other cerebrovascular diseases.

GRANTS

This work was supported by a grant from the National Institutes of Health, Heart Lung Institute: HL-080402 to J. N. Stallone.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

Author contributions: R.R.D. and J.N.S. conception and design of research; R.R.D. performed experiments; R.R.D. analyzed data; R.R.D. and J.N.S. interpreted results of experiments; R.R.D. and J.N.S. prepared figures; R.R.D. and J.N.S. drafted manuscript; R.R.D. and J.N.S. edited and revised manuscript; R.R.D. and J.N.S. approved final version of manuscript.