Essential Role of Estrogen for Improvements in Vascular Endothelial Function With Endurance Exercise in Postmenopausal Women

Participants

Postmenopausal women (45–65 y) were recruited from the Boulder and Denver (Colorado) metropolitan areas. Forty-eight women were enrolled into the intervention study and randomized to one of three treatment groups described below. Seven estrogen-deficient habitually endurance-trained postmenopausal women were also enrolled for comparison purposes for the oxidative stress mechanistic substudy. Inclusion criteria were amenorrheic ≥1 year and a FSH concentration ≥30 IU/L, nonsmokers, resting blood pressure (BP) <140/90 mmHg, fasted plasma glucose concentrations <7.0 mmol/L (126 mg/dL), and free of overt chronic diseases as assessed by medical history, physical examination, standard blood chemistries, and hematological evaluation (ie, normal kidney, liver, and thyroid function), and electrocardiogram at rest and during incremental treadmill exercise. Women had not taken any HT, cardiovascular or lipid-lowering medications for ≥6 months, or aspirin, nonsteroidal anti-inflammatory medications, or vitamin supplements for ≥4 weeks. All women enrolled in the estrogen/placebo interventions were required to provide documentation of a normal mammogram in the previous 12 months and Pap tests and gynecological examination within 3 years. Exclusion criteria were history of estrogen-dependent neoplasms, acute liver or gallbladder disease, abnormal vaginal bleeding, venous thromboembolism, triglyceride levels >400 mg/dL; known allergy to transdermal patch, or in women with a uterus, peanut oil; other contraindications to HT. Women enrolled in the exercise intervention study were sedentary or not exercising regularly >2 days per week, whereas the endurance-trained women had been performing regular vigorous endurance exercise (primarily running) >3 days per week for ∼25 years (range 6–54 y). All subjects provided written informed consent to participate. All procedures were reviewed and approved by the University of Colorado Boulder and University of Colorado Anschutz Medical Campus Institutional Review Boards.

Measurements

Participants were studied in the supine position following an overnight fast (water only) and caffeine abstinence. Normal dietary patterns and sodium intake were maintained for the 2-day period immediately prior to any measurements. The study took place at the Colorado Clinical and Translational Science Institute Clinical and Translational Research Center (CTRC) and the University of Colorado Boulder CTRC.

Participant characteristics

Seated and supine brachial artery BP were measured in triplicate at rest with a semiautomated device (Dinamap; Johnson & Johnson), as previously described (10). Total body fat was determined using dual-energy x-ray absorptiometry. Peak oxygen consumption (VO_2peak) was assessed during incremental treadmill exercise performed to exhaustion (8). Leisure time physical activity was determined by the Modifiable Activity Questionnaire (15). Fasting concentrations of glucose, insulin, blood lipids and lipoproteins, and serum and plasma concentrations of estradiol, progesterone, estrone, endothelin-1, interleukin-6, and C-reactive protein were measured by the Colorado Clinical and Translational Science Institute CTRC core laboratory as previously described (8, 13).

Interventions

Following baseline testing, participants were randomly assigned to one of three treatment arms: 1) oral estradiol (1 mg/d) and placebo patch; 2) transdermal estradiol patch (0.05 mg/d) and oral placebo tablet daily; and 3) placebo tablet daily and placebo patch; patches were applied weekly (see experimental study design supplemental Figure 1, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org). Because progestins can counteract the effects of estrogens (16), progestin exposure for women with a uterus was minimized by administration of micronized progesterone (200 mg/d) (or placebo tablet for placebo-treated women) for 13 days every 2 months; testing occurred when women were taking estrogen only and at least 1 month after progesterone administration. All investigators and participants were blinded to treatment status. Because cyclic progesterone induces withdrawal bleeding in women with a uterus, it is likely some participants were aware of treatment status. Of the 48 women randomized, 3 withdrew (1 transdermal estradiol and 2 placebo for personal reasons), and 2 were withdrawn (1 oral estradiol began a statin and 1 placebo for cardiac arrhythmia) during the first 12 weeks, and 2 withdrew after completing the initial 12 weeks of estrogen/placebo intervention because of lack of time (1 transdermal estradiol and 1 placebo).

Following 12-week testing, all participants (36 of 41 with analyzable images; 15 oral estradiol, 11 transdermal estradiol, and 10 placebo) began a 12-week endurance exercise training program as previously described (11) while maintaining their HT or placebo regimen. Briefly, after orientation to the exercise training program, participants were asked to walk 40 to 45 minutes per day, 5 to 7 days per week, at 65% to 80% of maximal heart rate on their own at home. Exercise program adherence was documented through physical activity logs and by the use of heart rate monitors (Polar Wireless Heart Rate Monitor). Four women (one placebo, one transdermal, and two oral estradiol) did not complete the exercise intervention but continued on estradiol/placebo treatment and completed 24-week testing and are included in the analysis. Key outcome variables were repeated at the end of the exercise training program, and ≥20 hours after the last exercise bout.

Statistical analysis

Outcomes were examined using descriptive statistics. Parameters with skewed distribution were log transformed and presented as median and interquartile range. ANOVA was used to assess group differences at baseline in clinical characteristics, circulating factors, and FMD. Intention-to-treat using two-way repeated measures ANOVA (group × time) was conducted to determine the effects of each estradiol group alone and combined with exercise vs placebo on FMD, and on clinical characteristics and humoral factors. Paired t tests were used to determine the effects of ascorbic acid on FMD at baseline and following the exercise program. As an exploratory analysis, we evaluated whether responses to exercise were different in women with previous HT use vs nonusers. Bivariate correlations were determined using the Pearson correlation coefficient. The study was powered to detect within-group changes and we evaluated between-group differences. Based on our previous investigations involving exercise and estrogen interventions, there was >80% power at an α of .05 to detect significant within-group differences in FMD with sample sizes of 10 to 15 per group (11, 13). Data analysis was performed with IBM SPSS Statistics 21.

Participant characteristics

Women were 7.9 ± 6.2 years past menopause and 69% (n = 25) were prior HT users. There were no group differences in time since menopause, prior HT use, or any other clinical characteristic or circulating factor (Tables 1 and 2).

Exercise intervention

Estrogen modulation of endothelial function adaptations to exercise training

Regular endurance exercise has been shown to prevent or ameliorate age-related endothelial dysfunction in older men (11, 17, 18). However, this is not consistently observed in older women (11, 12), suggesting that the adaptive response to exercise in older adults may be sex-specific. Our laboratory recently demonstrated that brachial artery FMD increased ∼50% in response to 8 weeks of moderate intensity walking in previously sedentary middle-aged and older men, but did not change in postmenopausal women (11). The findings of the intervention study were corroborated by the results of a separate larger (N = 167) cross-sectional analysis of endurance-trained and sedentary older men and women (11), previous cross-sectional findings in sedentary vs endurance-trained older men (10, 18), and the results of a previous endurance exercise training intervention study in postmenopausal women (12). Because ovarian function has a modulatory influence on vascular aging in women (8, 13), we hypothesized that the sex-specific endothelial adaptations to exercise may be related to differential exposure to sex hormones, as suggested previously (5, 11). Unlike women, whose endogenous estrogen concentrations undergo a relatively abrupt decrease with menopause, a parallel change is not observed in aged-matched men (19). Thus, we postulated that the lack of exercise training benefit on endothelial function may occur only in estrogen-deficient postmenopausal women (11).

In support of this, in the present study brachial artery FMD increased following 12 weeks of moderate intensity endurance exercise training in previously sedentary postmenopausal women who were treated with either oral or transdermal estradiol, but not in placebo-treated women (Figure 1). Moreover, in a control group of estrogen-deficient postmenopausal women who habitually performed vigorous endurance exercise, FMD was similar to the average values for all the women at baseline (ie, prior to beginning the interventions), consistent with our previous observations (11). These findings provide the first direct evidence that estrogen is necessary to induce the beneficial effects of endurance exercise training on endothelial function in postmenopausal women (Figure 3). Previous studies reporting that FMD was reduced in highly trained amenorrheic premenopausal athletes compared with eumenorrheic athletes and sedentary controls, and was restored after recovery of the menstrual cycle or with oral contraceptives (20, 21), are consistent with the present observations in postmenopausal women.

In contrast to the present and previous (11, 12) findings, endothelial function has been reported to be higher in estrogen-deficient former elite endurance-trained postmenopausal women compared with sedentary women, although not different between endurance-trained and sedentary women habitually taking combined estrogen plus progestin HT (22). Brachial artery FMD also has been observed to increase in response to an acute exercise bout alone and following 4 weeks of estrogen alone, without any apparent additive effects of combining the two treatments (23). The disparate findings compared with the present and previous (11, 12) studies may be attributed in part to differences in the FMD technique used including the use of upper-arm occlusion to induce reactive hyperemia (23), and calculating FMD from diameters measured 50 to 60 seconds after cuff deflation rather than continuous (22). Finally, because some progestins can antagonize the favorable effects of estrogen (16), it is possible that the lack of HT effect in the earlier study (22) was related to antagonistic effects of progestin. For this reason, in the present study, progesterone (or placebo) was administered cyclically to women with a uterus so that testing would not occur when women were taking combined estrogen and progesterone.

Potential mechanisms

The reasons the vascular endothelium is unresponsive to exercise training in hypoestrogenic women are unknown. The endothelial dysfunction observed in estrogen-deficient postmenopausal women and amenorrheic female athletes is related, in part, to reduced nitric oxide (NO) bioavailability (14, 24). Oxidative stress is a key mechanism contributing to the reduced NO and endothelial function in estrogen-deficient postmenopausal women (14) and is suspected to play a role in the endothelial dysfunction observed in amenorrheic athletes who have elevated reactive oxygen species levels (24). Reactive oxygen species impairs endothelial function by scavenging NO (25), and estrogen can increase NO bioavailability through antioxidant effects (14, 26). Consistent with this, in the present study FMD increased during a systemic ascorbic acid infusion, a well-described model to assess oxidative stress-related suppression of endothelial function (10), in all postmenopausal women at baseline (ie, before treatment) and in the habitually endurance-trained estrogen-deficient postmenopausal women (Figure 2). Moreover, ascorbic acid increased FMD after the exercise intervention in the placebo-treated women only. Importantly, however, no effect was observed after the exercise intervention in estrogen-treated women. Collectively, these results suggest that estrogen status plays a permissive role in the adaptive response of the endothelium to endurance exercise training in women by mitigating the effects of oxidative stress. Consistent with this broad concept, the enhanced endothelial function observed in endurance-trained older men is associated with an absence of vascular oxidative stress (10, 27).

Estrogen treatment may also modulate exercise-induced increases in shear stress-associated signal transduction to increase NO production and vasodilation. Estrogen and exercise share common intracellular signaling pathways to mediate NO release (28–32). Exercising blood flow increases frictional forces (ie, shear stress) along the surface of the endothelium, stimulating mechanosensors (eg, integrins, G-protein coupled receptors) that transduce mechanical forces into biochemical signals to phosphorylate and activate endothelial nitric oxide synthase (eNOS) and increase NO-dependent vasodilation (28, 29). Estrogen causes NO to be released through eNOS activation via estrogen receptor (ER) α-mediated nongenomic signaling of these same pathways (30, 32). In addition, both exercise and estrogen increase eNOS protein via transcriptional regulation of eNOS (31, 32). In vitro studies demonstrate that arterioles exposed to estrogen exhibit up-regulated eNOS and augmented NO-mediated vasodilation in response to increased flow and shear stress (33). The increased NO-mediated vasodilation was prevented with ER antagonism, indicating that the increased dilation was mediated via ER up-regulation of eNOS (33). Because prolonged estrogen deficiency decreases ERα expression, resulting in impaired ERα/eNOS signaling (34), the lack of exercise benefit in estrogen-deficient women may be related to reduced ERα and eNOS activation. We recently demonstrated that in vivo endothelial cell ERα is reduced in estrogen-deficient postmenopausal compared with premenopausal women, and that ERα was strongly correlated with eNOS protein and FMD (35). Whether ERα/eNOS signaling is important for vascular adaptations to exercise training in postmenopausal women warrants further study.

Limitations

We wish to emphasize certain limitations and experimental considerations. The study was powered to find within-group changes in FMD in response to estradiol and exercise. The lack of significant differences between the placebo- and estradiol-treated groups in the change in FMD with exercise training was likely related to a lack of power. The findings provide initial evidence for an essential role of estradiol in mediating endothelial adaptations to exercise, but this should be confirmed in a larger intervention trial. In addition, we studied healthy postmenopausal women and exercise training may benefit endothelial function only in postmenopausal women with CVD risk factors or established CVD. However, when compared with premenopausal women, our subjects had impaired FMD at baseline (8, 11, 35) and healthy men with similar baseline impairments had a robust increase in FMD with the same exercise intervention (11).

It is possible that the improvement in FMD following exercise in the estrogen-treated women was related to the prolongation of estrogen treatment itself and not to the exercise intervention. However, estradiol's vasodilatory effects are realized immediately and the magnitudes of improvement in FMD in response to acute (ie, 1 h) or chronic (ie, 3 mo) estrogen replacement in postmenopausal women are similar (36). In addition, it is possible a longer exercise intervention period may have elicited improvements in FMD in placebo-treated (estrogen-deficient) postmenopausal women. Black et al reported no change in FMD after 12 weeks of exercise training, but a trend (P = .07) for FMD to improve after 24 weeks of exercise training in estrogen-deficient postmenopausal women (37). However, in the present study and in our previous investigation (11) we found that, among estrogen-deficient postmenopausal women, FMD was similar between sedentary controls and endurance exercise-trained women who had been exercising regularly and vigorously for years.

Although not statistically significantly different, women in the estrogen-treated groups who had used HT previously demonstrated an ∼20% greater mean increase in FMD following the exercise training compared with women who had never used HT. This suggests the possibility that previous HT influences improvements in endothelial function with exercise-training and clearly warrants further study.

Because we administered different doses of oral and transdermal estradiol, the oral estradiol group had higher circulating estradiol concentrations than the transdermal group. However, these commonly prescribed doses have been shown to have similar clinical efficacy despite differences in circulating total estradiol concentrations (38). Oral estradiol undergoes first-pass metabolism resulting in lower free estradiol concentrations compared to transdermal estradiol, which bypasses the liver. Thus, in the present study, the dose of transdermal estradiol likely produced higher concentrations of free estradiol than the oral estradiol dose (39). Importantly, in the present study, the magnitude of improvement in FMD following oral and transdermal estradiol treatment alone was the same, and there was no appreciable difference in the magnitude of improvement with exercise training. Finally, it is important to note that there are risks associated with either estradiol regimen. However, because of the greater increase in venous thromboembolism risk with oral estradiol, transdermal estradiol may be a safer option for vascular protection (40).

Conclusions

The present study provides the first direct evidence for an essential role of estrogen in mediating improvements in endothelial function in response to regular endurance exercise in postmenopausal women. It will be important to confirm this finding in a larger randomized controlled trial and to identify key factors (eg, estrogen status, age, type/dose of estrogen, prior HT use) that may influence vascular adaptations to exercise in postmenopausal women. It also will be important to determine if an optimal level of estradiol is needed to induce improvements in endothelial function with exercise. Because regular exercise is a first-line strategy for primary prevention of CVD in postmenopausal women, a better understanding of the role of ovarian status in the potential beneficial effects of endurance exercise on vascular endothelial function is needed to inform the development of future sex-specific exercise guidelines or other therapies for CVD prevention.