0021-972X/97/$03.00/0 Vol. 82, No. 2 Journal of Clinical Endocrinology and Printed in U.S.A. Copyright © 1997 by The Endocrine Society

Differential Effects of Hormone-Replacement Therapy on Endogenous Nitric (/) Levels in Postmenopausal Women Substituted with 17␤-Estradiol Valerate and Cyproterone Acetate or

Medroxyprogesterone Acetate* Downloaded from https://academic.oup.com/jcem/article/82/2/388/2823193 by guest on 30 September 2021

BRUNO IMTHURN, MARINELLA ROSSELLI, ADRIAN W. JAEGER, PAUL J. KELLER, AND RAGHVENDRA K. DUBEY† Clinic of Endocrinology (B.I., M.R., P.J.K.), Department of Gynecology and Obstetrics, University Hospital Zurich; and Schering (Schweiz) AG (A.E.J.), Zurich, Switzerland; and Department of Medicine, Center for Clinical Pharmacology (R.K.D.), University of Pittsburgh Medical Center, Pittsburgh, Pennsylvania 15213

Ͼ ABSTRACT included in the analysis. Moreover, 30% increase in serum NO2-/ Increased incidence of cardiovascular disease in postmenopausal NO3- levels were observed only in 13 (responders) out of 26 PMW women (PMW) is accompanied by ovarian dysfunction; hormone re- substituted with estradiol valerate, suggesting that estradiol may placement therapy (HRT) can have cardioprotective effects. Because improve endogenous NO synthesis in a differential fashion. Com- hypertension and are associated with impaired re- pared with baseline, no significant increases in serum NO2-/NO3- lease of -derived (NO) and increased levels of were observed in samples collected while the estradiol-treated re- low- lipoproteins (LDL), we investigated whether HRT aug- sponders were taking either CPA or MPA. In contrast to NO2-/NO3-, ments NO release, and whether these increases are accompanied by serum LDL levels were significantly reduced in samples collected a decrease in LDL levels in PMW. We determined serum nitrite/ after 12 months of HRT (P Ͻ 0.05 . baseline). Furthermore, levels nitrate (NO2-/NO3-) and LDL levels at baseline (before initiation of of NO2-/NO3 showed a significant negative correlation with the levels HRT) and during the 6th and 12th months of the study. The PMW (n ϭ of LDL (r2 ϭ 0.17; P Ͻ 0.05) in the responders but not in nonre- 26) received continuous oral administration of estradiol valerate sponders. These results indicate that oral administration of estradiol (Progynova, 2 mg daily) for 21 days supplemented with either oral valerate in PMW for HRT increases circulating NO levels, an effect cyproterone acetate (CPA; 1 mg; n ϭ 11) or medroxyprogesterone that may contribute to the cardioprotective effects of HRT in PMW. acetate (MPA; 5 mg; n ϭ 15) on days 12–21 of each treatment cycle. In addition, our data suggests but does not prove that concomitant Blood samples in the PMW receiving HRT were collected at times administration of a progestin may attenuate the beneficial effects of while the subjects were taking estradiol valerate alone and estradiol estrogen replacement therapy with regard to NO release. Finally, our valerate plus CPA or MPA. Compared with the samples collected at data provides evidence for the existence of responders and nonre- Ϯ baseline, serum NO2-/NO3- levels increased significantly from 20.1 sponders to postmenopausal estrogen treatment with respect to im- 1.58 ␮mol/L at baseline to 30 Ϯ 3.7 ␮mol/L (P Ͻ 0.01) in samples provement of endogenous NO levels, suggesting that a significant collected after 12 months of HRT while the PMW were not taking number, but not all, of the hormonally substituted PMW profit fully progestins (CPA or MPA), and to 25.4 Ϯ 2 ␮mol/L (P Ͻ 0.05) when all from the beneficial properties of a HRT. (J. Clin Endocrinol Metab 82: the samples, regardless of the treatment with CPA or MPA, were 388–394, 1997)

PIDEMIOLOGICAL studies have shown that women ease in postmenopausal women (PMW) (3) and attenuates E within the reproductive age group are protected the development of dietary atherosclerosis in ovariecto- against cardiovascular disease when compared with men mized monkeys (4). (1–3). The observation that this difference decreases with the Although it is now well accepted that in women estrogen onset of menopause (1–3) has led to the hypothesis that induces cardioprotective effects, the mechanism(s) of estro- estrogen may be protective against cardiovascular diseases gen-induced cardioprotective effects remains unclear. There- (3). In support of this hypothesis, estrogen replacement ther- fore, we hypothesize that estrogen-induced cardioprotection apy markedly reduces the incidence of cardiovascular dis- is mediated in part by increased synthesis of nitric oxide (NO). Several observations provide a strong rationale for this Received June 7, 1996. Revision received October 8, 1996. Accepted hypothesis including: 1) NO exerts a number of cardiopro- November 7, 1997. tective actions, e.g. , inhibition of aggre- Address all correspondence and requests for reprints to: Raghvendra K. Dubey, Center for Clinical Pharmacology, Department of Medicine, gation, and vascular cell growth (5–7). 2) 200 Lothrop Street, 623 Scaife Hall, University of Pittsburgh Medical Treatment of cultured human endothelial cells with 17␤- Center, Pittsburgh, Pennsylvania 15213-2582. estradiol stimulates constitutive NO synthase activity (8). 3) * This work was supported in part by NIH Grants HL-40319 and Aortic rings from female vs. male rabbits release more NO, HL-35909. † Visiting scientist at the Department of Gynecology and Obstetrics, and this enhanced release correlates with circulating estra- University Hospital, Zurich, Switzerland. diol levels (9). 4) Pharmacological blockade of NO synthase

388 ESTROGEN INCREASES NO SYNTHESIS IN PMW 389 attenuates estradiol-induced vasodilation of the uterine cir- on the last 3 days of the combined estradiol-progestin treatment. All culation (10). 5) In humans, circulating levels of NO increase blood samples were taken in the morning after a 12-h fast period, and with follicular development and correlate with 17␤-estradiol immediately after the patient’s arrival. Serum was separated by centri- fuging the samples at 800 ϫ g for 10 min. The samples were stored at levels (11). To test our hypothesis, we recently examined in Ϫ20 C until analysis. PMW the effects of estrogen replacement therapy using transdermal patches on circulating levels of nitrite/nitrate Nitrite/nitrate analysis (metabolites of NO; 12). These studies provided the first Serum NO2-/NO3- levels were measured by reacting the samples clinical evidence that estrogen replacement therapy increases with the Griess reagent, by our previously described method (12) and NO production in PMW. a commercially available kit (Alexis Corp., Laeufelingen, Switzerland). Our previous study, however, raises several important Briefly, aliquots (40 ␮L) of serum in duplicate were incubated at room questions: 1) Is the effect of estrogen on NO production temperature with cofactors and for1hto convert NO3-toNO2-. Total NO2- was then analyzed by reacting the Downloaded from https://academic.oup.com/jcem/article/82/2/388/2823193 by guest on 30 September 2021 caused by a direct action of estrogen on NO synthase or ␮ samples with 50 L of Griess reagent component. Amounts of NO2-in caused by an indirect effect? In this regard, low-density li- serum were estimated from a standard curve measuring the absorbance poprotein (LDL) is known to inhibit NO synthase (13, 14) and at 540 nm. is increased in PMW (2, 15). Because estrogen replacement therapy reduces LDL levels in PMW (2, 3, 16), it is possible Hormone analysis that estrogen increases NO production via reducing LDL Serum 17␤-estradiol levels were analyzed by RIA (Diagnostics Prod- levels. 2) Is the effect of estrogen on NO production atten- ucts Corp., Los Angeles, CA). Serum FSH levels were estimated by uated by concomitant treatment with a progestin? In our means of microbead enzyme immunoassay (IMX, Abbott Park, IL). previous study (12), increases in circulating NO levels in response to 17␤-estradiol were substantially diminished LDL and high-density lipoprotein (HDL) analysis while PMW were taking the progestin norethisterone acetate. Serum levels of LDL were measured by the methods of 3) Does only transdermal delivery of estrogen increase NO Burstein et al. (18) and Friedwald et al. (19) using a commercially avail- synthesis or is orally administered estrogen also effective? able kit (bioMerieux, MarcyI’Etolle, France). Briefly, the LDL fraction in Previous studies (17) suggest that the therapeutic effects of 50-␮l aliquots of serum was precipitated by addition of 1 mL of am- estrogen may depend on the route of administration. The phophillic polymers [anionic-polycyclic activator (0.4 mg/mL), anionic- polycondensated activator (0.6 mg/mL), and polysubstituted dioxan goal of the current study was to further investigate the role (12.4 mmol/L) in an -buffer (25 mmol/L), pH 6.1]. After in- of NO synthase in estrogen-induced cardioprotection by ad- cubation for 30 min at 2–8 C, the samples were centrifuged at 4 ϫ 103 dressing each of the three aforementioned question. To rpm. The supernatant was discarded, and the pellet resuspended in 0.5 achieve this objective, we examined the effects of oral ad- mL trisodium citrate (0.15 mol/L) containing NaCl (0.11 mol/L). Ali- quots (50 ␮L) of this suspension were reacted with an enzyme solution ministration of estradiol valerate on serum levels of nitrite/ (provided in the kit), and the levels of LDL measured spectrophoto- nitrate and LDL in PMW with and without concomitant metrically at 546 nm. The amounts of LDL were estimated from a administration of two chemically distinct progestins. standard curve of LDL run in parallel. Serum HDL levels were measured spectrophotometrically at 500 nm by the methods of Burstein et al. (18) and Lopes-Virella et al. (20) and by Subjects and Methods using a commercially available kit (Boehringer-Mannheim, Mannheim, Subjects Germany).

Thirty-two healthy, caucasian PMW participated in this study. In- Statistical analyses clusion criteria were as follows. Participants 1) were 45–55 yr of age; 2) were more than 1 yr after menopause; 3) exhibited symptoms requiring The person performing the LDL, HDL, and NO2-/NO3- assay was hormone replacement therapy (HRT) for therapeutic or preventive rea- unaware of whether the sample was from a subject in the CPA or MPA sons; 4) were never treated with HRT or, if ever on HRT, had a wash-out group. Baseline (before HRT) measurements of serum LDL, HDL, and period of at least 4 weeks (on conjugated estrogen at least 12 weeks) NO2-/NO3- levels were obtained from all 32 individuals. Five of the 32 before to the beginning of the trial; 5) had FSH serum levels more than subjects (1 in the MPA, 4 in the CPA group) retrospectively showed 40 IU/L; 6) had estradiol serum levels less than 100 pmol/L; and 7) had serum estradiol levels higher than 100 pmol/L, and therefore these Ϯ a body weight within the range of standard weight 15%. The women individuals were excluded from the statistical analyses. Two subjects (1 enrolled in the study had similar lifestyles and dietary habits and were in the MPA, 1 in the CPA group) had incomplete data, missing the instructed to avoid any changes in dietary habits during the investiga- sample of the combined estradiol-progestin application at treatment tion. All subjects gave written informed consent to the study, which was month 6. Out of the 27 subjects with complete sets of data, one had approved by the Institutional Medical Research Committee on Clinical significantly high levels of nitrite/nitrate (60 ␮mol/L vs. 20.1 Ϯ 1.58 Investigation. averaged from 26 subjects) and was an outlier according to the box plot statistical analysis and was excluded. Statistical analysis was performed Protocol using ANOVA and paired Student’s t test, as appropriate. The criterion of significance was a value of P Ͻ 0.05. Data are presented as mean Ϯ HRT was applied in a sequentially combined fashion. The estrogen sem. used was estradiol valerate, which is immediately metabolized to es- To study whether estradiol valerate administration altered serum tradiol, the natural human estrogen. Estradiol valerate (2 mg daily; NO2-/NO3-, LDL, and HDL levels with the time of treatment, the 26 Progynova, Schering, Switzerland) was administered orally for 21 days. subjects with complete data sets were analyzed by making between Subjects were randomly assigned to receive with the last 10 tablets of group comparisons by ANOVA, and Fisher’s least significant test was Progynova in each cycle either 1 mg of cyproterone acetate (CPA) or 5 used to detect differences between specific groups. The effects of the two mg of medroxyprogesterone acetate (MPA). At the end of each 21-day different progestins (CPA and MPA) on the NO2-/NO3- levels were treatment cycle, there was a 7-day treatment-free period. Blood samples compared using ANOVA. To evaluate whether sequential treatment for serum NO2-/NO3- levels were drawn at baseline (i.e. before HRT was with the progestins (CPA and MPA) modulate the effects of estradiol initiated) and then again at 6 and 12 months into HRT. The samples were valerate on NO synthesis, only the data from the PMW who responded withdrawn on one of the last 3 days of unopposed estradiol intake and to estradiol valerate with an increase in NO2-/NO3- levels were ana- 390 IMTHURN ET AL. JCE&M• 1997 Vol 82 • No 2

lyzed. This analysis was conducted using two-factor ANOVA with repeated measures (factor A: treatment group; factor B: time period) and Friedmans repeated measure ANOVA on ranks; all pairwise multiple comparisons were by Newman Keuls method. A response to estradiol of greater than a 30% increase in NO2-/NO3- levels compared with baseline was used to define subjects as responders and a scatter plot was constructed to show a bimodal distribution. Linear regression analysis was performed to analyze whether there was a correlation between the levels of NO2-/NO3- and LDL in responders and nonresponders.

Results Serum estradiol levels (average) in all 26 PMW before HRT

Ϯ Downloaded from https://academic.oup.com/jcem/article/82/2/388/2823193 by guest on 30 September 2021 was 38 6 pmol/L, and the average baseline serum NO2- /NO3- levels in all 26 PMW before initiation of HRT was 20.1 Ϯ 1.58 ␮mol/L. Pharmacokinetic studies with estradiol FIG. 2. A, Scatterplot showing serum NO2-/NO3- levels in samples collected from PMW (n ϭ 26) after 12 months of HRT and during valerate given orally shows that a dose of 2 mg attained estradiol administration alone. Levels of nitrite/nitrate in these PMW serum estradiol levels of 230–930 pmol/L (21). In samples demonstrate presence of two distinct populations and suggest that obtained after 6 months of HRT during which individuals there is a differential increase in nitrite/nitrate levels in PMW re- were taking estradiol valerate without progestin (CPA or ceiving estradiol valerate. Subjects with increases in nitrite/nitrate levels of more than 30% (subjects above dashed line) were classified MPA), serum NO2-/NO3- levels increased marginally to Ϯ ␮ Ͼ as responders, whereas, subjects with increases of less than 30% were 23.2 2.3 mol/L (P 0.05; Fig. 1). However, in samples classified as nonresponders. Out of 26 subjects, 13 were nonre- collected after 12 months of treatment, serum NO2-/NO3- sponders, and 13 subjects were responders. B, Changes in nitrite/ concentrations increased significantly and were 30.0 Ϯ 3.7 nitrate levels in PMW (responders, Ͼ30% increase vs. baseline, n ϭ ␮mol/L (P Ͻ 0.01 vs. baseline and P Ͼ 0.05 vs. 6 month). 13; and nonresponders Ͻ30% increase vs. baseline, n ϭ 13) after 12 months of treatment with estradiol valerate. Samples were collected Serum NO2-/NO3- levels did not increase in all the PMW from PMW receiving HRT while subjects were receiving estradiol subjects receiving HRT, but rather increased differentially. A valerate alone. Circulating nitrite/nitrate levels increased signifi- Ͻ scatterplot of increases in NO2-/NO3- levels in PMW in re- cantly in responders but not in nonresponders. *, P 0.01 vs. pre- sponse to HRT showed a bimodal distribution, with two treatment. One subject from the responders group with nitrite/nitrate ␮ ␮ distinct PMW populations (Fig. 2A). A detailed analysis levels of 60 mol/L at baseline and 91 mol/L after 12 months of HRT while taking estradiol valerate (an increase of 52%) was excluded from demonstrated that out of 26 subjects, only 13 individuals had analysis because her nitrite/nitrate levels were beyond levels ob- an increase of greater than 30% in serum NO2-/NO3- levels, served in 26 PMW. i.e. from 20.5 Ϯ 3.4 ␮mol/L at baseline to 39.3 Ϯ 5.5 ␮mol/L in samples collected after 12 months of HRT (P ϭ 0.007 vs, ential increases in serum NO2-/NO3- levels were observed in baseline; Fig. 2, A and B), and these subjects were classified PMW treated for 12 months with estradiol valerate, there as responders. In the remaining 13 subjects serum NO2-/ were no differences in serum NO2-/NO3- levels in samples NO3- serum levels either increased less than 30% or de- taken after 6 months of substitution, which were 22 Ϯ 2 creased or remained unchanged after 12 months of HRT ␮mol/L in responders and 23.7 Ϯ 4.2 ␮mol/L in nonre- Ϯ ␮ (baseline: 22.8 2.5 mol/L vs. after 12 months of HRT: sponders (P Ͼ 0.05). The subjects in these two sets, classified Ϯ ␮ Ͼ 19.8 2.8 mol/L; P 0.05) and these subjects were clas- as responders (increase in NO -/NO - levels) and nonre- sified as nonresponders (Fig. 2, A and B). Although differ- 2 3 sponders (no change or decrease in NO2-/NO3- levels) did not have any significant difference with respect to age, base-

line serum FSH, estradiol, and NO2-/NO3- levels. In the two groups of PMW receiving either CPA (n ϭ 11) or MPA (n ϭ 15), no significant difference in the baseline

serum FSH, estrogen, and NO2-/NO3- levels were observed. Serum NO2-/NO3- levels increased significantly in samples collected at 12 months while the subjects were taking estra-

diol valerate alone (Fig. 1). In contrast, the serum NO2-/NO3- levels in samples collected while subjects were taking estra- diol plus CPA or MPA (after 12 months of HRT) were not

significantly different from the serum NO2-/NO3- levels ob- served at baseline and were 17.6 Ϯ 3.6 at baseline vs.19Ϯ 3 ␮mol/L (P Ͼ 0.05) in the CPA-treated group and 17.4 Ϯ 1.5 at baseline vs. 22.5 Ϯ 4.0 ␮mol/L (P Ͼ 0.05) in MPA-treated group. However, when 12-month samples were included in the statistical analysis regardless of collection with respect to FIG. 1. Bar graph showing time-dependent changes in circulating estradiol valerate and the type of progestin used (CPA and NO (NO -/NO -) levels in response to estradiol valerate in all post- 2 3 MPA combined), the increase in serum NO -/NO - levels menopausal women (n ϭ 26) receiving HRT. Samples were collected 2 3 was significant compared with the levels before the initiation while subjects were receiving estradiol valerate alone. Compared with Ϯ ␮ baseline, serum NO2-/NO3- were significantly increased in samples of the HRT (25.4 1.58 mol/L after 12 months of HRT vs. collected after 12 months of treatment. 20.1 Ϯ 1.6 ␮mol/L at baseline; P Ͻ 0.05). ESTROGEN INCREASES NO SYNTHESIS IN PMW 391

To investigate the effects of CPA and MPA on estradiol- induced NO synthesis, only the samples collected at 12 months from the responders group taking either MPA or

CPA were considered for statistical analysis. Serum NO2-/ Ͻ NO3- levels were significantly increased (P 0.05 vs. baseline in MPA group and P Ͻ 0.05 vs. baseline in CPA) in samples collected while the subjects were taking estradiol valerate alone (Fig. 3); in contrast, the serum NO2-/NO3- levels were decreased to basal levels in samples collected while the sub- jects were taking estradiol valerate with CPA or MPA (Fig. 3). However, when all of the 12-month samples were in- cluded in the statistical analysis in the responders group, Downloaded from https://academic.oup.com/jcem/article/82/2/388/2823193 by guest on 30 September 2021 regardless of collection with respect to estradiol valerate with FIG. 4. Bar graph showing changes in serum nitrite/nitrate levels in or without CPA or MPA treatment, the change in NO2-/NO3- PMW from the responders group after 12 months of HRT with es- levels in HRT-PMW receiving CPA or MPA was significantly tradiol valerate and sequential administration of the progestins MPA increased and was 26.5 Ϯ 5.4 ␮mol/L (P Ͻ 0.05 vs. baseline alone (n ϭ 9) (left panel) or CPA alone (n ϭ 4) (middle panel). Right Ϯ ␮ Ϯ ␮ Ͻ panel shows effects of overall HRT irrespective of type of progestin 17.6 35.4 mol/L) and 37.4 5 mol/L (P 0.002 vs. ϭ ϭ ϭ Ϯ ␮ used, i.e. MPA (n 9) and CPA (n 4) data combined (n 13). Data baseline 17.4 1.9 mol/L), respectively (Fig. 4). Finally, are mean Ϯ SEM of all samples collected after 12 months, while sub- when all the samples in the responders group, regardless of jects were receiving estradiol valerate alone and estradiol valerate the type of progestin used (CPA and MPA combined), were plus progestins (CPA or MPA). HRT significantly increased NO syn- included in the statistical analysis, the increase in serum thesis in PMW irrespective of type of progestins used along with estradiol valerate. P Ͻ 0.05 value compared with respective baseline NO2-/NO3- levels at 12 months was significant compared nitrite/nitrate levels in each group. with the levels before the initiation of the HRT (33.2 Ϯ 4.2 ␮mol/L after 12 months of HRT vs. 17.5 Ϯ 1.63 ␮mol/L at baseline; P ϭ 0.002; Fig. 4). The average circulating LDL levels in PMW before HRT were 3.9 Ϯ 0.15 mmol/L and decreased significantly to 3.5 Ϯ 0.13 mmol/L (8% decrease; P Ͻ 0.05) after 12 months of HRT (Fig. 5A). Compared with the group of PMW in which no increases in NO synthesis were observed after HRT (nonre- sponders), the basal levels of LDL were substantially higher in the group of PMW in which increases in NO synthesis were observed after HRT (responders). Furthermore, the lev- els of LDL decreased significantly from baseline 4.1 Ϯ 0.2 mmol/L to 3.7 Ϯ 0.2 mmol/L (64 10% decrease; P Ͻ 0.05) after HRT in the responders but not in the nonresponders FIG. 5. A, Bar graph showing time-dependent changes in circulating (P Ͼ 0.05). Nonparameteric regression analysis between the LDL levels in postmenopausal women (n ϭ 23) receiving HRT without regard to groups receiving CPA or MPA. Compared with baseline, NO2-/NO3- levels and LDL levels showed a significant neg- ative correlation in the responders’ group (r2 ϭ 0.17; P Ͻ 0.05; serum LDL were significantly decreased in samples collected after 12 months of treatment. B and C, Linear regression analysis showing Fig. 5B) but not in the nonresponders’ group (Fig. 5C). In correlation between serum nitrite/nitrate and LDL levels in samples collected from PMW before and after substitution with 17␤-estradiol valerate. B, Correlation in responders group (n ϭ 11); C, correlation in nonresponders group (n ϭ 12). A significant negative correlation was observed between LDL and nitrite/nitrate levels in responders but not in nonresponders.

contrast to LDL, HRT increased the circulating HDL levels in PMW receiving HRT for 12 months by 6% (P ϭ 0.051 com- pared with levels of HDL before initiation of HRT, which were 1.7 Ϯ 0.1 mmol/L. In contrast to LDL, the increases in HDL levels after HRT were not differential in responders (P ϭ 0.25) and nonresponders (P ϭ 0.13). FIG. 3. Comparison of changes in serum nitrite/nitrate levels after 12 months of HRT in the responders group of PMW receiving estradiol Discussion valerate sequentially with progestins MPA (n ϭ 9) or CPA (n ϭ 4). Nitrite/nitrate levels were significantly increased in samples collected In a previous study (12), we provided the first evidence while subjects were receiving estradiol valerate alone. In contrast, that 17␤-estradiol, administered via transdermal patches, in- levels of nitrite/nitrate were not significantly increased in samples creased endogenous NO levels in postmenopausal women. collected when subjects were taking MPA (E ϩ MPA) or CPA (E ϩ CPA) along with estradiol valerate. *,PϽ 0.05 compared with base- Furthermore, the increases in endogenous NO levels were line; †, P Ͻ 0.05 compared with levels of nitrite/nitrate in samples decreased when an artificial progestin, norethisterone ace- collected when subjects were taking estradiol valerate (E) alone. tate, was coadministered with 17␤-estradiol. The aim of the 392 IMTHURN ET AL. JCE&M• 1997 Vol 82 • No 2 present study was to investigate whether other chemically LDL (2, 3, 27, 28) and increases in HDL (3) should improve different forms of estradiol and progestins that are used overall endothelial/vascular function and diminish vascular clinically for HRT and administered orally also increase NO disease and thereby augment vascular NO production. In- levels. The results of the present study demonstrate the fol- deed, this latter hypothesis fits well with the slow time course lowing: 1) Oral administration of estradiol valerate to PMW of enhanced NO levels (i.e. Ͼ6 months) observed in the for 12 months results in a significant increase in circulating present study. levels of nitric oxide (nitrite/nitrate) in some women. 2) The Our finding that serum NO2-/NO3- levels were increased increase in circulating NO2-/NO3- levels are accompanied by after 12 but not 6 months of treatment indicate that effects of a decrease in LDL levels and an increase in HDL levels. 3) No estrogen on NO levels are not rapid. In contrast, short-term increases in circulating NO2-/NO3- levels in response to increases in estrogen levels in premenopausal women during estradiol valerate occur when subjects are taking the pro- the menstrual cycle increases NO levels dynamically (11). gestins CPA or MPA. 4) HRT increases endogenous NO Estrogen increases transcription of NO synthase in male Downloaded from https://academic.oup.com/jcem/article/82/2/388/2823193 by guest on 30 September 2021 levels in only approximately half of PMW. 5) Increases in NO guinea pigs, but with a marked delay compared with females levels in responders are accompanied by a significant de- (29). It has been suggested that these differential vasodilatory crease in circulating LDL levels, whereas no significant de- effects in males and females are receptor dependent, and that crease in LDL levels occur in nonresponders. However low- the lag phase in NO synthase transcription in male guinea ering of LDL is not solely responsible for the increase in NO pigs could be caused by the presence of fewer estrogen re- production. ceptors. Estrogen receptors within the blood vessels of PMW

In the current study, changes in circulating levels of NO2- are lost or decreased (30), and estrogen up-regulates its own /NO3- were used as an index for changes in NO synthesis. receptors (31). Therefore, it is possible that the effects of An important issue therefore is whether this methodological estrogen are receptor operated, and that the delayed effects approach is valid. In this regard, NO is a labile (5) of estrogen on NO levels are caused by the initial lack of and that decomposes into nitrite and nitrate within seconds of its subsequent up-regulation of estrogen receptors in PMW; release. Several in vivo studies have shown that endogenous whereas, in premenopausal women the receptors are con- changes in NO synthesis induced by either l- tinuously expressed and primed because of the constant methyl ester or endotoxin can be measured by analyzing presence/generation of estrogen, thus leading to a rapid changes in circulating or urinary levels of NO2-/NO3- (11, 12, receptor mediated increase in NO levels. Alternatively, be- 22, 23). More importantly, stoichometric metabolic tracer cause increases in NO2-/NO3- levels are accompanied with 15 studies by Hibbs et al. (24) using l-[guanidino- N2] a decrease in LDL levels, it is feasible that the delayed effects as a substrate for NO synthesis demonstrate that in humans of estrogen replacement on NO levels/production also are increased nitrate production in serum is derived from NO indirect and may in part be caused by the effects of estrogen generated from the terminal guanidino of the on LDL, free generation, and -induced labeled l-arginine. endothelial injury. Although the above studies suggest that NO is the primary An important observation of the present study is that only source for circulating NO2-/NO3-, dietary must also half of the investigated subjects showed an increase in serum be considered. However, it is extremely unlikely that the NO2-/NO3- levels during the treatment period. This obser- changes in circulating levels of NO2-/NO3- measured in the vation suggests the existence of responders and nonre- present study were caused by dietary factors. This conclusion sponders to postmenopausal estrogen treatment with respect is based on two lines of reasoning. First, dietary nitrates are to improvement in endogenous NO levels/production; how- excreted in the urine within 18 h of ingestion (25, 26). Because ever, the explanation for the dichotomonus response remains in the present study serum NO2-/NO3- levels were measured obscure. Similar to this finding, is the finding that a heter- approximately 14 h after the last meal (overnight fast), most ogeneous response of HRT on LDL previously has been of the dietary nitrates would have been eliminated at the time observed (28). One reason for the different reaction of these of blood sampling. Second, because each subject served as two groups on estrogen treatment could be the inability of her own control, dietary habits would be the same before and nonresponders to reestablish the estrogen receptors nor- during HRT, thus nullifying any effect of dietary nitrates. mally lost in postmenopause (30). It has been shown that The mechanism(s) by which estradiol induces NO syn- PMW lose estrogen receptors in the vascular endothelium thesis are unknown. Presumably the effects are receptor me- (30), hence subjects with a widely damaged endothelium, as diated because estrogens are well known to interact with in atherosclerosis, may be unable to reconstitute estrogen high-affinity estrogen receptors in the nucleus and thereby receptors, and therefore there is no increase in NO levels in direct the expression of numerous target genes. The findings response to estrogen. Another possibility is that reductions that estrogen stimulates constitutive in LDL mediate the increases in NO levels/production. In activity in cultured human vascular smooth muscle cells (8), this regard in our study, serum LDL levels decreased sig- and that estrogen-induced increases in blood flow are nificantly in responders (P ϭ 0.01) but not in nonresponder blocked by l-nitroarginine methyl ester (10) suggest that (P Ͼ 0.05) after 12 months HRT compared with baseline estrogen receptors might directly and/or indirectly regulate values. Linear regression analysis showed a significant neg- constitutive nitric oxide synthase gene expression. Another ative correlation between LDL and NO2-/NO3- levels in the possibility is that estrogen-induced changes in lipoprotein responders but not in the nonresponders. Although, there is levels indirectly result in increased production of NO. For a significant correlation between LDL and NO2-/NO3-in instance, estrogen-induced reductions in LDL and oxidized responders, it is not strong enough to establish a cause-and- ESTROGEN INCREASES NO SYNTHESIS IN PMW 393 effect relationship (correlation coefficient value of r2 ϭ 0.185; ens our contention that the cardioprotective effects of estra- P ϭ 0.03). These findings suggest that although lowering of diol are mediated at least in part through estradiol-induced LDL does increase NO levels, it is not the sole factor influ- NO synthesis. Additionally, these data suggest that irrespec- encing NO levels/production. Hence, based on the above tive of the route of administration and chemical form of information it could be argued that apart from the largely estradiol used, HRT increases endogenous NO activity. Fi- receptor-mediated increases in NO levels, the increases in nally, our data also provides evidence for the existence of

NO2-/NO3- in PMW may in part be caused by the lowering responders and nonresponders to postmenopausal estrogen of LDL. Furthermore, the above finding implies that the treatment with respect to improvement of endogenous NO cardiovascular benefit of HRT is higher for a responder than levels, suggesting that a significant number, but not all, of the for a non-responder. hormonally substituted PMW profit fully from the beneficial Progestins have been reported to impair the cardiopro- properties of a HRT. tective effects of estrogen (32), as well as enhance the vascular Downloaded from https://academic.oup.com/jcem/article/82/2/388/2823193 by guest on 30 September 2021 damage associated with hypertension (33). We have shown Acknowledgment previously that luteal, but not follicular, increases in pro- gesterone levels in hormonally stimulated and in spontane- We thank Dr. Edwin K. Jackson (Center for Clinical Pharmacology, ous menstrual cycles decrease serum NO -/NO - levels, University of Pittsburgh Medical Center, Pittsburgh, PA) for critical 2 3 suggestions in writing the manuscript. even in the presence of high concentrations of 17␤-estradiol (11). Furthermore, we have shown recently that norethis- terone acetate (NETA), a commonly used progestin for HRT References in Europe, prevented 17␤-estradiol-induced increases in cir- 1. Wenger NK, Speroff L, Packard B. 1993 Cardiovascular health and disease in culating NO -/NO - levels during HRT (12). Combined ther- women. N Engl J Med. 329:247–256. 2 3 2. Matthews KA, Meilahn E, Kuller LH, Kelsey SF, Caggiula AW, Wing RR. apy of estrogen with MPA has been shown recently to have 1989 Menopause and risk factors for coronary heart disease. N Engl J Med. better cardioprotective effects even compared with estrogen 321:641–646. alone (3). NETA is a -derived progestin with 3. Nabulsi AA, Folsom AR, White A, et al. 1993 Association of hormone-re- placement therapy with various cardiovascular risk factors in postmenopausal androgenic effects, whereas MPA and CPA are derivatives of women. N Engl J Med. 328:1069–1075. 17␣-OH-. Therefore, we speculated that com- 4. Adams MR, Kaplan JR, Manuck SB, et al. 1990 Inhibition of coronary artery atherosclerosis by 17-beta estradiol in ovariectomized monkeys: lack of an pared with NETA, MPA and CPA may have a different effect of added progesterone. Arteriosclerosis. 10:1051–1057. influence on circulating NO2-/NO3- levels. Similar to our 5. Moncada S, Palmer RM, Higgs EA. 1991 Nitric oxide: physiology, patho- observation with NETA, in contrast to unopposed estrogen physiology, and pharmacology. Pharmacol Rev. 43:109–142. 6. Dubey RK. 1994 Vasodilator-derived nitric oxide inhibits angiotensin-II and treatment, no significant increase in serum NO2-/NO3- levels fetal calf serum-induced growth of arteriolar smooth muscle cells. J Pharmacol were observed when estradiol valerate was coadministered Exp Therap. 269:402–408. 7 Dubey RK, Jackson EK, Lu¨ scher TF. 1995 Nitric with MPA or CPA. Our data suggests that progestins, oxide inhibits angiotensin II-induced migration of rat aortic smooth muscle cell: role of cyclic-nucleotides and angiotensin1 receptors. J Clin Invest. whether natural or artificial and whether 17␣-OH-proges- 96:141–149. terone or testosterone derivatives, attenuate estrogen-in- 7. Dubey RK, Jackson EK, Lu¨ scher TF. 1995 Nitric oxide inhibits angiotensin duced NO levels/production. In this regard, progesterone II-induced migration of rat aortic smooth muscle cell: role of cyclic-nucleotides and angiotensin1 receptors. J Clin Invest. 96:141–149. inhibits estrogen-induced endothelium-dependent re- 8. Hishikawa K, Nakaki T, Marumo T, Suzuki H, Kato R, Saruta T. 1994 sponses associated with the production of NO (34). However, Induction of constitutive nitric oxide synthase by estradiol in human endo- thelial cells. Hypertension. 24:386 (Abstract 83). the exact mechanism(s) by which progesterone decreases 9. Hayashi T, Fukuto JM, Ignarro LJ, Chaudhuri G. 1992 Basal release of nitric estradiol-induced NO levels is unclear. oxide from aortic rings is greater in female rabbits than in male rabbits: implications for atherosclerosis. Proc Natl Acad Sci USA. 89:11259–11263. Because of the rapid clearance of orally administered es- 10. Van-Buren GA, Yang DS, Clark KE. 1992 Estrogen-induced uterine vasodi- tradiol valerate, the levels of estradiol in the serum were not lation is antagonized by L-nitroarginine methyl ester, an inhibitor of nitric measured, and this raises the issue as to whether the differ- oxide synthesis. Am J Obstet Gynecol. 167:828–833. 11. Rosselli M, Imthurn B, Macas E, Keller PJ, Dubey RK. 1994 Circulating ential responses observed in the PMW were real or caused nitrite/nitrate levels increase with follicular development: indirect evidence by noncompliance. To address this issue, we reevaluated the for estradiol mediated NO-release. Biochem Biophys Res Commun. data from our previous study (12) in which PMW were 202:1543–1552. 12. Rosselli M, Imthurn B, Keller PJ, Jackson EK, Dubey RK. 1995 Circulating substituted with transdermal patches, and both estradiol and nitric oxide (nitrite/nitrate) levels in postmenopausal women substituted with ␤ NO2-/NO3- levels were measured. Similar to the present 17 -estradiol and norethisterone acetate. A two year follow-up study. Hy- pertension. 25:848–853. study, estrogen replacement differentially increased NO2-/ 13. Andrews HE, Bruckdorfer KR, Dunn RC, Jacobs M. 1987 Low-density li- NO3- levels in PMW (66% responders and 34% nonre- poproteins inhibit endothelium-dependent relaxation in rabbit aorta. Nature. sponders), yet circulating estradiol levels in both responders 327:237–239. 14. Chin JH, Azhar S, Hoffman BB. 1992 Inactivation of endothelial derived and nonresponders were similar and not statistically differ- relaxing factor by oxidized lipoproteins. J Clin Invest. 89:10–18. ent. These findings strongly suggest that the phenomena of 15. Stevenson JC, Crook D, Gogsland IF. 1993 Influence of age and menopause differential increases in NO levels in response to estradiol on serum and lipoproteins in healthy women. Atherosclerosis. 98:83–90. 16. Kushwaha RS, Lewis DS, Carey KD, McGill, Jr, HC. 1991 Effects of estrogen replacement observed in the present study is real. and progesterone on plasma lipoproteins and experimental atherosclerosis in In conclusion, we provide the first clinical evidence that the baboon (Papio sp). Arterioscler Thromb. 11:23–31. 17. Walsh BW, Schiff I, Rosner B, Greenberg L, Ravnikar V, Sacks FM. 1991 oral administration of estradiol valerate in PMW for HRT Effects of postmenopausal estrogen replacement on the concentrations and increases circulating NO2-/NO3- levels and lowers LDL lev- metabolism of plasma lipoproteins. N Engl J Med. 325:1196–1204. els. Furthermore, the progestins MPA and CPA attenuate 18. Burstein M, Scholnick HR, Morfin R. 1970 Rapid method for the isolation of lipoproteins from human serum by precipitation with polyanions. J Res. estrogen-induced increases in circulating NO2-/NO3- levels. 11:583–595. Finally, our data confirm our previous finding and strength- 19. Friedewald WT, Levy RI, Fredrickson DS. 1972 Estimation of the concentra- 394 IMTHURN ET AL. JCE&M• 1997 Vol 82 • No 2

tion of low-density lipoprotein cholesterol in plasma, without use of the pre- and noncontraceptive use of estrogen in women: results from the Lipid Re- parative ultracentrifuge. Clin Chem. 18:499–502. search Clinics Program follow-up study. Circulation. 75:1102–1109. 20. Lopes-Virella MF, Stone P, Ellis S, Colwell JA. 1977 Cholesterol determina- 29. Weiner CP, Lizasoain I, Baylis SA, Knowles RG, Charles IG, Moncada S. tion in high-density lipoproteins separated by three different methods. Clin 1994 Induction of -dependent nitric oxide synthases by sex hormones. Chem. 23:882–884. Proc Nat Acad Sci USA. 91:5212–5216. 21. Dusterberg B, Nishino Y. 1982 and pharmacological fea- 30. Losordo DW, Kearney M, Kim EA, Jekanowski J, Isner JM. 1994 Variable tures of oestradiol valerate. Maturitas. 4:315–324. expression of the in normal and atherosclerotic coronary 22. Ochoa JB, Udekwu AO, Billiar TR, et al. 1991 levels in patients arteries of premenopausal women. Circulation. 89:1501–1510. after trauma and during . Ann Surg. 214:621–626. 31. Rosser M, Chorich L, Howard E, Zamorano P, Mahesh VB. 1993 Changes in 23. Shultz PJ, Tolins JP. 1993 Adaptation to increased dietary salt intake in the rat uterine receptor messenger ribonucleic acid levels during estrogen and rat. Role of endogenous nitric oxide. J Clin Invest. 91:642–650. progesterone induced estrogen receptor depletion and subsequent replenish- 24. Hibbs JB, Westenfelder C, Taintor R, et al. 1992 Evidence for cytokine- inducible nitric oxide synthesis from L-arginine in patients receiving inter- ment. Biol Reprod. 48:89–98. leukin-2 therapy. J Clin Invest. 89:867–877. 32. Lobo RA. 1992 The role of progestins in hormone replacement therapy. Am J 25. Wasserman AE. 1978 The nitrite- situation: a review. Food Engi- Obstet Gynecol. 166:1997–2004. neering. 50:110–116. 33. Wolinsky H. 1972 Effects of estrogen and progesterone treatment on the Downloaded from https://academic.oup.com/jcem/article/82/2/388/2823193 by guest on 30 September 2021 26. Evans TG, Ramussen K, Wiebke G, Hibbs, Jr, JB. 1994 Nitric oxide synthesis response of the aorta of male rats to hypertension. Morphological and chemical in patients with advanced HIV infection. Clin Exp Immunol. 97:83–86. studies. Circ Res. 30:341–349. 27. Sack MN, Rader DJ, Cannon RO. 1994 Oestrogen and inhibition of oxidation 34. Miller VM, Vanhoutte PM. 1991 Progesterone and modulation of endothe- of low-density lipoproteins in postmenopausal women. Lancet. 343:269–270. lium-dependent responses in canine coronary arteries. Am J Physiol. 261: 28. Bush TL, Barrett-Connor E, Cowan LD, et al. 1987 Cardiovascular mortality R1022–R1027.