| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regulates Follicle-Stimulating Hormone Secretion, Whereas Central Estrogenic Pathways Direct Growth Hormone and Prolactin Secretion in Postmenopausal WomenEndocrine Research Unit (M.C., J.B., J.M.M., J.D.V.), Department of Internal Medicine, Clinical Translational Science Unit, Mayo Medical and Graduate Schools of Medicine, Mayo Clinic, Rochester, Minnesota 55905; and Division of Endocrinology (C.Y.B.), Department of Internal Medicine, Tulane University Health Sciences Center, New Orleans, Louisiana 70112
Address all correspondence and requests for reprints to: Johannes D. Veldhuis, Endocrine Research Unit, Department of Internal Medicine, Clinical Translational Science Unit, Mayo Medical and Graduate Schools of Medicine, Mayo Clinic, 200 First Street S.W., Rochester, Minnesota 55905. E-mail: veldhuis.johannes{at}mayo.edu.
| Abstract |
|---|
|
|
|---|
Objective: Our objective was to distinguish between hypothalamic and pituitary or peripheral (hepatic) actions of E2.
Setting: This study was performed in an academic medical center.
Design: This was a double-blind, prospectively randomized, placebo (Pl)-controlled study.
Methods: The capability of a selective, noncompetitive, non-CNS permeant estrogen receptor (ER)-
antagonist, fulvestrant (FUL) to antagonize the effects of transdermal E2 and Pl on GH, prolactin, and FSH secretion was assessed in 43 women (ages 50–80 yr) in a four parallel-cohort study. Each woman received four secretagogue infusions to stimulate GH secretion. IGF-I and its binding proteins were measured secondarily.
Results: Administration of Pl/E2 increased GH and prolactin concentrations by 100%, and suppressed FSH concentrations by more than 50% (each P
0.004 compared with Pl/Pl). Treatment with FUL/E2 compared with Pl/E2 partially relieved estrogens inhibition of FSH secretion (P = 0.041), without altering E2s stimulation of prolactin secretion. ANOVA further revealed that: 1) estrogen milieu (P = 0.014) and secretagogue type (P < 0.001) each determined GH concentrations; 2) FUL/Pl suppressed IGF-I concentrations (P < 0.001); 3) FUL abrogated estrogens elevation of IGF binding protein-1 concentrations (P < 0.001); and 4) FUL did not oppose estrogens suppression of IGF binding protein-3 concentrations (P < 0.001).
Summary and Conclusions: Responses to a non-CNS permeant ER
antagonist indicate that E2 inhibits FSH secretion in part via pituitary/peripheral ER
, drives prolactin output via nonpituitary/nonperipheral-ER
effects, and directs GH secretion and IGF-I-binding proteins by complex mechanisms.
| Introduction |
|---|
|
|
|---|
GH secretion is under the control of multiple peptidyl signals, including GHRH and somatostatin released by hypothalamic neurons, the GH-releasing peptide (GHRP), ghrelin, produced in the stomach, hypothalamus, and pituitary gland, and circulating GH and IGF-I, which feed back negatively on the hypothalamo-pituitary unit (10, 11, 12, 13, 14, 15). Clinical studies indicate that E2 regulates hypothalamo-pituitary responses to each of these peptides (16, 17, 18, 19, 20). However, whether estrogen drives GH secretion predominantly via hypothalamic or pituitary actions remains unknown (2, 10, 21, 22). Indeed, whether E2 influences the production of other human hypophyseal hormones principally by effects on the brain or pituitary has not been elucidated.
The present study attempts to distinguish central nervous system (CNS) (hypothalamic) and pituitary effects of E2 in postmenopausal women by administering E2 in the presence or absence of fulvestrant (FUL) (ICI 182,780), a highly specific nonbrain-permeant estrogen receptor (ER) antagonist (23, 24). The drug is a systemically active 7
-alkylsulfinyl analog of E2, which binds ER
, thereby preventing receptor dimerization, forcing receptor degradation, and inhibiting gene transcription (25, 26). In breast cancer, FUL decreases ER
protein and inhibits cellular replication without exerting intrinsic estrogen-agonist activity (23). An important mechanistic property of this antiestrogen is that it does not cross the blood-brain barrier. In particular, peripheral administration of FUL, unlike tamoxifen, does not antagonize the E2-induced prolactin surge, brain progesterone-receptor expression, or radiolabeled E2 binding in the hypothalamus of the rat (23, 24), and does not induce hot flushes in women.
ER
is expressed in normal human pituitary cells producing GH (2.3%), FSH (70%), and prolactin (50%) (27, 28). ERβ is present in normal pituitary cells, the majority of pituitary tumors, and coexpressed with ER
in prolactinomas and gonadotrope tumors (29, 30). We could find no direct experimental evidence for ERβ effects on human pituitary hormone secretion in vivo, but ERβ induces progesterone-receptor immunoreactivity in rat gonadotrope cells in vitro and antagonizes ER
effects on stress-induced ACTH secretion in the rat in vivo (31, 32). Both ER subtypes also exist in the hypothalamus (33). In the arcuate nucleus, the majority of GHRH neurons express ER
and GnRH neurons ERβ, whereas few (<5%) somatostatin neurons have either ER
or ERβ (34, 35). Accordingly, estrogens could act on both the pituitary gland and the brain. Which action predominates in the human is not known.
To assess the importance of pituitary vis-à-vis CNS mechanisms in mediating estrogen actions, we developed a novel investigative paradigm comprising: 1) transdermal delivery of placebo (Pl) or E2 in postmenopausal women to mimic the young-adult estrogen milieu, and 2) concomitant administration of Pl or FUL to distinguish between central-neural (FUL resistant) and pituitary or hepatic (FUL accessible) loci of E2 action. GH secretion was the primary endpoint, and FSH and prolactin responses were secondary outcomes.
| Subjects and Methods |
|---|
|
|
|---|
Healthy postmenopausal women provided written informed consent approved by the Mayo Institutional Review Board. Inclusion criteria were stable body mass index (BMI) less than or equal to 30 kg/m2, no night-shift work, unremarkable medical history, physical examination, and renal screening, and hepatic, endocrine, metabolic, and hematological function. Exclusion criteria included any sex hormone replacement 4 wk before participation, vascular or gallbladder disease, breast or endometrial cancer, anemia, psychiatric illness, alcohol abuse, and use of neuroactive medications. Menopausal status was confirmed by concentrations of FSH more than 50 IU/liter, LH more than 20 IU/liter, and E2 less than 35 pg/ml (<120 pmol/liter).
Protocol design
This was a prospectively randomized, Pl-controlled, double-blind, parallel cohort design (Fig. 1
). Randomization was by a randomized-number table to Pl/Pl, Pl/E2, FUL/Pl, and FUL/E2 administration. A total of 43 women were randomly assigned to receive double Pl [Pl/Pl (n = 10)], Pl/E2 (n = 10), FUL/Pl (n = 12), and FUL/E2 (n = 11). Median (range) age and BMI were 61 yr (50–80) and 25 kg/m2 (18–30), respectively. Before E2 and Pl replacement, a baseline outpatient screening blood sample was obtained (Table 1
data). Thereafter, eligible subjects received three consecutive weekly im injections of Pl or FUL (250 mg) under an investigator-initiated Food and Drug Administration new drug number (Fig. 1
). The antiestrogen is approved for single monthly injections of 250 mg for the therapy of breast cancer (26). The higher schedule and dose were selected to allow the maximal clinically safe blockade of ERs during E2 administration. Beginning 1 wk after the second FUL injection, transdermal Pl or E2 supplementation was begun daily for 18 d (Fig. 1
). Doses were 0.05 mg daily for 4 d, 0.10 mg for 4 d, 0.15 mg for 4 d, and then 0.20 mg for 7 d to mimic the normal menstrual cycle pattern in young women (36). Oral micronized progesterone of 100 mg was prescribed for 12 d beginning on the last day of the study.
|
|
Serum was collected at 0800 h every fourth outpatient day and on the last day of the entire study to assay E2, testosterone, LH, FSH, SHBG, prolactin, IGF-I, and IGF binding protein (IGFBP)-1 and IGFBP-3.
Each volunteer was admitted to the Clinical Research Unit on four separate evenings once before each infusion study to allow overnight adaptation. Sleep was deferred until 2200 h. To obviate food-related confounds, subjects ate a standard meal at 2000 h, and then fasted until 1400 h the next day. At 0700 h two iv catheters were inserted into contralateral forearm veins. Between 0800 and 1400 h, plasma was collected every 10 min to assay GH.
Inpatient secretagogue infusions
Infusion studies were performed in randomized order during any four of the last 5 d of the Pl and E2 (highest dose) supplementation schedule. The four iv infusion paradigms were: 1) saline continuously at 50 ml/h, from 0800–1400 h; 2) saline (0800–1000 h), followed by GHRH and GHRP-2 infused constantly from 1000–1400 h (both at a rate of 1 µg/kg·h); 3) saline (0800–1000 h), L-arginine (30 g iv over 30 min, from 0930–1000 h), and GHRH (1 µg/kg bolus iv) at 1000 h; and 4) saline/L-arginine infusion as in the third paradigm, followed by GHRP-2 (3 µg/kg bolus iv) at 1000 h. The combined-peptide infusion permits indirect appraisal of potential estrogenic modulation of somatostatin outflow (37). The two interventions, which are preceded by L-arginine and followed by a maximally stimulatory peptide dose, are designed a priori to identify possible evidence of estrogen-dependent effects on endogenous ghrelin and GHRH-sensitive pathways, respectively (36, 38, 39, 40). Constant infusion of GHRH/GHRP-2 permits evaluation of somatostatin restraint, as verified in model simulations (38, 39). Thus, this intervention allows one to compare how the four estrogen milieus influence somatostatin outflow. The comparison is not among secretagogues, but among estrogen milieus.
Hormone assays
Plasma GH concentrations were measured in duplicate by automated ultrasensitive double-monoclonal immunoenzymatic, magnetic particle-capture chemiluminescence assay using 22-kDa recombinant human GH as standard (Sanofi Diagnostics Pasteur Access, Chaska, MN), as described recently (36, 40). Details of the LH, FSH, E2, and testosterone assays were reported earlier (36, 40).
Total IGF-I concentrations were measured by immunoradiometric assay after extraction (Diagnostic Systems Laboratories, Webster, TX). Interassay coefficients of variation (CVs) were 9% at 64 µg/liter and 6.2% at 157 µg/liter. Intraassay CVs were 3.4, 55.4, and 1.5% at 9.4, 55.4, and 264 µg/liter, respectively. IGFBP-1 and IGFBP-3 concentrations were measured by two-site immunoradiometric assay (Diagnostic Systems Laboratories), with interassay CVs of 3.5% at 5.16 µg/liter and 3.6% at 142 µg/liter, and intraassay CVs of 5.2% at 5.23 µg/liter and 2.7% at 144.6 µg/liter (IGFBP-1), and interassay CVs 1.9% at 7690 µg/liter and 0.6% at 803 µg/liter, and intraassay CVs 1.8% at 8272 µg/liter and 3.9% at 735 µg/liter (IGFBP-3).
Statistical comparisons
Baseline (preintervention) characteristics in the four study groups were evaluated by one-way ANOVA. Post hoc comparisons of means were made using Tukeys honestly significantly different (HSD) criterion (41). Significance was construed at experiment-wise P < 0.05. Time-dependent effects of E2 and/or FUL on outpatient fasting hormone measurements were analyzed using two-way ANOVA in a repeated-measures design (42). The model examined the individual and interactive effects of time (four factors) and drug (four other factors, Pl, FUL, and/or E2). Logarithmic transformation was first applied to limit the dispersion of residual variance. Data obtained in the Clinical Research Unit after the highest dose of E2 or Pl, were also analyzed by two-way ANOVA to compare peak and incremental GH-concentration responses to peptide infusions, wherein four types of secretagogues replaced the four time factors. The same statistical structure was applied to evaluate the incremental effects of Pl or FUL on GH, IGF-I, IGFBP-1, and IGFBP-3 concentrations. Incremental responses were assessed to minimize variability due to larger interindividual variations in starting measurements.
Statistical power analysis performed a priori assumed a 2.1-fold stimulatory effect of E2 over Pl on mean fasting baseline GH concentrations [obtained from earlier aggregate data (n = 73 subjects)], and postulated 33% inhibition of the E2 effect by coadministration of FUL. Assuming nominal mean fasting GH concentrations of 0.24 ± 0.08 (SD) µg/liter in postmenopausal women (n = 52), E2 would be expected to elevate GH concentrations to 0.48 ± 0.17 µg/liter. A 33% decline would then represent a decrement of 0.16 µg/liter. By an unpaired one-tailed t test, the latter decrease could be detected with 90% power at P < 0.05 if a total of 20 subjects were studied.
| Results |
|---|
|
|
|---|
Preinterventional (baseline) data
Table 1
summarizes mean age, BMI, and baseline hormone concentrations obtained before Pl, E2, or FUL administration in the four study groups. There were no differences in these variables at study outset. E2 concentrations were 11 ± 1.6 pg/ml at baseline in the double-Pl group. Values increased significantly and comparably in response to successive 4 d outpatient increments in the E2 dose (P < 0.001) (Fig. 2
). Maximal E2 concentrations (pg/ml) approximated those of the normal late-follicular phase whether or not FUL was administered concomitantly (152 ± 15 and 137 ± 15 in Pl/E2 and FUL/E2, respectively; P = 0.16).
|
One-way ANOVA revealed a significant overall drug (E2 or FUL) treatment effect on mean fasting GH concentrations obtained on the day of saline infusion (d 14–18 of Pl or E2 exposure; P = 0.004; Fig. 3
). Post hoc analysis using the Tukey HSD criterion showed that GH concentrations were 2- and 2.6-fold higher in the Pl/E2 (P = 0.047) and FUL/E2 (P = 0.017) groups, respectively, than in the Pl/Pl cohort. GH concentrations did not differ between Pl/E2 and FUL/E2 or between FUL/Pl and Pl/Pl (both P > 0.50).
|
|
LH, FSH, and prolactin
Fasting gonadotropin and prolactin concentrations measured at the end of study (last day of peptide infusions) are given in Table 3
. FSH concentrations were significantly lower and prolactin concentrations significantly higher in the Pl/E2 and FUL/E2 groups than in the Pl/Pl or FUL/Pl groups (both P < 0.001). LH concentrations (P = 0.08) did not differ among the four estrogen treatment groups. Because of prominent intersubject variability, ANOVA was also applied to intraindividual changes (increments) in FSH (P < 0.001), LH (P = 0.042), prolactin (P < 0.001), and SHBG (P = 0.002) concentrations from baseline (Fig. 4
). Post hoc analysis disclosed that FUL partially muted E2s inhibition of FSH secretion (P = 0.024) without altering E2s stimulation of prolactin secretion. The effects of FUL/E2 on (incremental) SHBG and (decremental) LH concentrations did not differ from those due to E2 or FUL individually.
|
|
There was no effect of Pl vs. FUL before E2 or Pl was initiated (P = 0.81). Successive comparisons of IGF-I, IGFBP-1, and IGFBP-3 concentrations measured every 4 d with individual baseline values are given in Fig. 5
. Two-way ANOVA of these differences revealed significant effects of treatment, but not a treatment-by-time interaction (P = 0.37). Primary treatment effects were P = 0.009 (IGF-I), P < 0.001 (IGFBP-1), and P < 0.001 (IGFBP-3). Post hoc analysis showed that FUL/Pl lowered IGF-I concentrations compared with Pl/Pl (P = 0.001) and Pl/E2 (P = 0.015), but not FUL/E2 (P = 0.096). Pl/E2 elevated IGFBP-1 concentrations compared with the other three interventions (each P
0.035). Pl/E2 decreased IGFBP-3 concentrations compared with Pl/Pl (P < 0.001) and FUL/Pl (P = 0.014). Finally, FUL alone increased IGFBP-1 (P = 0.002) and decreased IGFBP-3 (P = 0.009), and with E2 opposed the effect of E2 on IGFBP-1 (P = 0.005), but not on IGFBP-3 (P = 0.094).
|
| Discussion |
|---|
|
|
|---|
antagonist that does not traverse the blood-brain barrier: 1) does not decrease GH secretion whether or not it is combined with E2, 2) partially opposes E2s suppression of FSH concentrations, and 3) does not attenuate E2s stimulation of prolactin secretion (Fig. 6
knockout in adult female mice (43). Thus, low endogenous E2 concentrations in postmenopausal women acting via ER
appear to maintain IGF-I availability. The collective data indicate that E2 suppresses FSH secretion in part via a direct ER
-mediated effect on the pituitary gland; augments fasting GH and prolactin secretion principally via CNS actions, which are not inhibitable by a peripheral antiestrogen; regulates IGFBP-1 availability partially by peripheral ER
mechanisms; and directs IGFBP-3 production via unknown non-ER
pathways.
|
pituitary/peripheral and toward central actions of E2 given that a priori estimates of statistical power predicted a greater than 90% probability of detecting at least 33% inhibition by FUL of ER
-dependent stimulation of GH secretion (see Subjects and Methods). In mechanistic clinical studies, E2 supplementation augments GHRH potency, blunts inhibition by somatostatin, potentiates stimulation by synthetic GHRP, and amplifies GH release evoked by somatostatin withdrawal (16, 17, 18, 44). Although the definitive in vivo mechanisms are not known, E2-stimulated CNS sites include GHRH neurons that express ER
and hypothalamic cells expressing IGF-I receptors, GHRP receptors, and the IGF-I gene; conversely, E2-inhibited CNS loci include GH and somatostatin receptors (45, 46, 47, 48, 49). Whereas ERβ has not been established directly as a clinical mediator of any in vivo pituitary effects, somatotrope cells contain both ER
and ERβ (27, 28, 34). In various animal species, E2 can augment pituitary IGFBP-2 concentrations, IGF-I binding sites, and somatostatin receptor subtype 2, and repress somatostatin receptor subtype 5 and GHRH receptors (50, 51, 52, 53). Depending upon dose, estrogens are able to inhibit, not affect, or stimulate GH synthesis and secretion by pituitary cells in vitro (54, 55, 56, 57). In the accompanying analyses, antagonizing the ER
pathway sufficiently to impede the inhibitory effect of E2 on FSH secretion by almost 50% had no detectable effect on GH secretion before or after infusion of three mechanistically distinct types of GH secretagogues (Subjects and Methods). We interpret this outcome to indicate that estrogens 2-fold stimulation of GH secretion when FUL is present is not mediated via pituitary ER
pathways, but rather requires CNS mechanisms or far less well-studied peripheral or pituitary ERβ (or other) pathways.
Exogenous estrogen elevates IGFBP-1 but suppresses IGFBP-3 and acid-labile subunit concentrations in a dose-dependent manner (58, 59, 60). If ER
were required to maintain physiological concentrations of IGFBP-1 or IGFBP-3 as we postulate for IGF-I, then administration of FUL should inhibit E2s induction of IGFBP-1 and suppression of IGFBP-3. Because this prediction was true for IGFBP-1 but not IGFBP-3, we believe that peripheral ER
is not the exclusive regulator of IGFBP-3. Whether ERβ is such a countervailing mediator is not known.
Studies in animal models indicate that estrogens inhibit gonadotropin and stimulate prolactin secretion via ER
(61, 62, 63), which is localized in lactotropes and GnRH-stimulatory kisspeptin neurons (64). Concentrations of FSH and LH increase and those of prolactin decrease in gonadectomized wild-type and ER
knockout mice, and are normalized by E2 administration in wild-type but not ERβ knockout animals (63, 65). Although the endocrine effects of FUL are reversed by selective ER
agonists in animals (66), in the present clinical study, estrogenic suppression of FSH secretion was not totally overcome by the ER
antagonist. This may be due to an inadequate dose of FUL, or involvement of CNS pathways or pituitary/peripheral ERβ. Because depletion of ER
can augment ERβ, which opposes the effects of ER
on certain gene promoters (67, 68, 69), further investigations should appraise whether ERβ controls pituitary-hormone secretion.
FUL inhibits E2-stimulated lactotrope proliferation in vitro (66) but does not block suckling- or E2-evoked prolactin secretion in the rat in vivo (24). The topographical restriction of FUL action to outside the CNS is not shared by another antiestrogen (tamoxifen), which is able to antagonize brain actions of E2 (24, 66, 70). Accordingly, earlier laboratory outcomes and current clinical data together argue that estrogen stimulates prolactin secretion primarily by CNS mechanisms that do not involve peripheral ER
.
In conclusion, in healthy postmenopausal women, E2s suppression of FSH secretion in part involves pituitary ER
, whereas estrogenic stimulation of prolactin and GH secretion entails principally nonperipheral ER
(putatively CNS) mechanisms. Estrogenic effects on IGF-I and IGFBPs are more complex with evidence of ER
dependence (IGF-I and IGFBP-1) and independence (IGFBP-3).
| Acknowledgments |
|---|
| Footnotes |
|---|
Disclosure Statement: The authors have nothing to declare.
First Published Online December 18, 2007
Abbreviations: BMI, Body mass index; CNS, central nervous system; CV, coefficient of variation; E2, estradiol; ER, estrogen receptor; FUL, fulvestrant; GHRP, GH-releasing peptide; HSD, honestly significantly different; IGFBP, IGF binding protein; Pl, placebo.
Received June 13, 2007.
Accepted December 12, 2007.
| References |
|---|
|
|
|---|
and β isoforms in human pituitary tumors. J Clin Endocrinol Metab 83:3965–3972
and a tumor-specific splice variant. J Clin Endocrinol Metab 83:3308–3315
-dihydrotestosterone and its metabolite 5β-androstan-3β, 17β-diol inhibit the hypothalamo-pituitary-adrenal response to stress by acting through estrogen receptor β-expressing neurons in the hypothalamus. J Neurosci 26:1448–1456
mRNA. J Clin Endocrinol Metab 85:3840–3846
, but not ER β, gene is expressed in hypothalamic growth hormone-releasing hormone neurons of the adult female rat. Neurosci Res 52:121–125[CrossRef][Medline]
but not ERβ. Mol Endocrinol 17:1039–1053
in the anterior pituitary gland. Mol Endocrinol 11:674–681
. Endocrine 11:137–143[CrossRef][Medline]
. Neurosci Lett 401:225–230[CrossRef][Medline]
in negative feedback regulation of gonadotropin-releasing hormone mRNA expression in the female mouse. Neuroendocrinology 78:204–209[CrossRef][Medline]
and ER β. Mol Interv 3:281–292
and ERβ at AP1 sites. Science 277:1508–1510This article has been cited by other articles:
![]() |
J. D Veldhuis, D. M Keenan, J. N Bailey, J. M Miles, and C. Y Bowers Preservation of GHRH and GH-releasing peptide-2 efficacy in young men with experimentally induced hypogonadism Eur. J. Endocrinol., August 1, 2009; 161(2): 293 - 300. [Abstract] [Full Text] [PDF] |
||||
![]() |
J. D. Veldhuis, K. L. Mielke, M. Cosma, C. Soares-Welch, R. Paulo, J. M. Miles, and C. Y. Bowers Aromatase and 5{alpha}-Reductase Inhibition during an Exogenous Testosterone Clamp Unveils Selective Sex Steroid Modulation of Somatostatin and Growth Hormone Secretagogue Actions in Healthy Older Men J. Clin. Endocrinol. Metab., March 1, 2009; 94(3): 973 - 981. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Shao, M. Nutu, B. Weijdegard, E. Egecioglu, J. Fernandez-Rodriguez, E. Tallet, V. Goffin, C. Ling, and H. Billig Differences in Prolactin Receptor (PRLR) in Mouse and Human Fallopian Tubes: Evidence for Multiple Regulatory Mechanisms Controlling PRLR Isoform Expression in Mice Biol Reprod, October 1, 2008; 79(4): 748 - 757. [Abstract] [Full Text] [PDF] |
||||
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |