| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Original Articles: Hormones and Reproductive Health |
Department of Psychiatry, Weill Medical College, Cornell University (M.A., C.R.), New York, New York 10021; Behavioral Endocrinology Branch, National Institute of Mental Health, National Institutes of Health (C.R.), Bethesda, Maryland 20892; and Department of Military and Emergency Medicine, Uniformed Services University of the Health Sciences (E.G., P.D.), Bethesda, Maryland 20814
Address all correspondence and requests for reprints to: Dr. Margaret Altemus, Box 244, Weill Medical College, Cornell University, 1300 York Avenue, New York, New York 10021. E-mail: maltemus{at}mail.med
Abstract
Accumulating evidence indicates that gonadal steroids modulate functioning of the hypothalamic-pituitary-adrenal (HPA) axis, which has been closely linked to the pathophysiology of anxiety and depression. However, the effect of the natural menstrual cycle on HPA axis responsivity to stress has not been clearly described. In nine healthy women, metabolic and hormonal responses to treadmill exercise stress during the early follicular phase of the menstrual cycle, when gonadal steroid levels are low, were compared with responses in the midluteal phase of the cycle, when both progesterone and estrogen levels are relatively high. Exercise intensity was gradually increased over 20 min to reach 90% of each subjects maximal oxygen consumption during the final 5 min of exercise. Basal plasma lactate, glucose, ACTH, vasopressin, oxytocin, and cortisol levels were similar in the two cycle phases. However, in response to exercise stress, women in the midluteal phase had enhanced ACTH (P < 0.0001), vasopressin (P < 0.01), and glucose (P < 0.001) secretion. These findings suggest that relatively low levels of gonadal steroids during the early follicular phase of the menstrual cycle provide protection from the impact of stress on the HPA axis.
CLINICAL OBSERVATIONS suggest that fluctuations in reproductive hormones influence the course of depression and anxiety disorders, but the biological mechanisms that may mediate the effects of reproductive hormones on emotional regulation have not been identified. Accumulating evidence suggests that gonadal steroids modulate functioning of the hypothalamic-pituitary-adrenal (HPA) axis, which has been closely linked to the pathophysiology of anxiety and depression. On the one hand, HPA axis responses to acute stress can reflect the sensitivity of the individual to a physical or emotional stress. On the other hand, glucocorticoids secreted during stress may, in turn, act in the central nervous system to regulate mood and other mental processes (1). Glucocorticoids are also the primary endocrine feedback signal for suppression of the HPA axis and noradrenergic responses to stress (2, 3, 4). Alterations in glucocorticoid secretion and glucocorticoid receptor sensitivity are postulated to play a role in the development of psychiatric disorders (4) and the mechanism of action of antidepressant medication (5, 6).
Our previous clinical studies have shown enhanced sensitivity of the HPA axis to dexamethasone suppression and increases in type II glucocorticoid receptor messenger ribonucleic acid (mRNA) expression in circulating mononuclear cells during the early follicular phase of the menstrual cycle relative to the midluteal phase (7). The study described here was designed to test whether HPA axis responsivity to stress is restrained in the early follicular, compared with the midluteal phase of the menstrual cycle. Restrained HPA axis reactivity in the early follicular phase would be expected as a consequence of increased glucocorticoid receptor sensitivity.
We chose to deliver exercise as an HPA axis stressor across the menstrual cycle because it is a reproducible, quantifiable stressor that elicits neuroendocrine and metabolic responses proportional to the intensity and duration of the exercise (8, 9). The relative intensity of the exercise can be standardized across subjects who have different levels of physical conditioning by setting each subjects exercise intensity to elicit a specific percentage of that individuals maximal aerobic capacity (VO2max). Exercise stress stimulates release of cortisol, ACTH, and vasopressin, which, in turn, promotes mobilization of the energy stores needed to meet the exercise stress.
Subjects and Methods
Subjects
Nine healthy women participated in the study after giving written informed consent. The protocol had been approved by the intramural institutional review boards of the NIMH and the Uniformed Services University of the Health Sciences. Subjects were recruited by local advertisements for paid participation in an exercise study. Subjects were 34.2 ± 1.8 (range, 2542) yr of age. All subjects were Caucasian, were in good physical health, had regular menstrual cycles, were nonsmokers, and were not taking any medications other than vitamins. All subjects completed a medical history and underwent a physical examination and screening laboratory tests, including complete blood count, chemistry panel, thyroid function tests, toxicology screen, urinalysis, and pregnancy test. No subject was suffering from anxiety disorders, depression, or other mental illnesses, as assessed by a structured diagnostic interview (10). No subject suffered from premenstrual dysphoric disorder, as assessed by clinical interview and 3 months of daily mood ratings (10, 11).
Experimental protocol
Exercise testing was performed in both the early follicular (36 days after the onset of menses) and midluteal (710 days after the LH surge) phases of the menstrual cycle. Each woman determined the time of her LH surge by using a home urine test kit (Clearplan Easy, Unipath Ltd., Bedford, UK). Four women performed the follicular phase exercise test first, and five women started with the luteal phase exercise test. Five of the nine subjects completed exercise testing within one menstrual cycle. The other subjects required up to three cycles to complete testing due to scheduling difficulties and failure to detect an LH surge needed for scheduling of the luteal test.
Details of the exercise protocol have been reported previously (12). Subjects were instructed not to eat on the morning before the exercise tests and to abstain from caffeine and alcohol consumption and from running or other strenuous activity during the 24 h before testing.
VO2max was determined once by a progressive treadmill exercise protocol conducted to volitional exhaustion. Oxygen consumption and carbon dioxide production during the maximal exercise test and the two subsequent follicular and luteal exercise test sessions were measured with a Metabolic Measurement Cart 2900c (SensorMedics, Inc., Yorba Linda, CA). Electrocardiogram and heart rate were also monitored continuously during all exercise tests.
For the follicular and luteal exercise test sessions, subjects reported to the laboratory between 07000800 h. On arrival, all subjects had an iv catheter inserted into an antecubital vein and then consumed 5 mL/kg BW water to ensure uniform hydration. The exercise test began 60 min after the subject finished drinking. The treadmill exercise test consisted of 5 min of exercise at 50% of VO2max, 5 min at 70%, a 2-min break for blood drawing, another 5 min at 70%, and 5 min at 90% of VO2max. Exercise intensities of 50%, 70%, and 90% of VO2max were enforced by adjusting the slope and speed of the treadmill. The exercise was followed by a cool-down period of 5 min at a minimal workload, then 35 min of rest in a semirecumbent position. Expired oxygen and carbon dioxide, electrocardiogram, and heart rate were monitored throughout the exercise test and for the first 15 min of the recovery period.
Subjects stood for at least 15 min before the first baseline blood sample was drawn. Samples were collected before exercise (-10 and 0 min), during exercise (+10 min), and immediately after exercise (+20 min) in a standing position and subsequently (+30, +40, +50, and +60 min) in a semirecumbent position. Blood was collected into a syringe and dispensed into chilled tubes containing ethylenediamine tetraacetate for ACTH, cortisol, vasopressin, oxytocin, estradiol, and progesterone assays and into tubes containing heparin and fluoride for lactate and glucose measurements. Chilled tubes were stored on ice until plasma was separated by centrifugation at 4 C (within 30 min). Plasma for ACTH, vasopressin, and cortisol determinations was stored frozen at -80 C until assayed, whereas plasma for lactate, glucose, estrogen, and progesterone determinations was refrigerated and assayed on the day of the exercise test.
Biochemical assays
Plasma lactate and glucose concentrations were determined in duplicate (analyzer model 27, YSI, Inc., Yellow Springs, OH). The intra- and interassay coefficients of variation (CVs) for lactate and glucose were less than 5%. ACTH, vasopressin, and cortisol were assayed in three batches, with follicular and luteal samples from individual subjects run in the same batch to minimize effects due to interassay variation. All oxytocin samples were run in one assay. Commercial RIA kits were used to measure ACTH (immunoradiometric assay, Nichols Institute Diagnostics, San Juan Capistrano, CA) and cortisol (Coat-a-Count, Diagnostic Products, Los Angeles, CA). The limits of detection for these kits in our laboratory are 0.55 pmol/L for ACTH and 27 nmd/L for cortisol. The intraassay CV averaged 5% for cortisol and ACTH, and the interassay CV was 12% for ACTH and 8% for cortisol. Arginine vasopressin and oxytocin RIAs were performed as previously described (13, 14) with a 3-fold concentration during the extraction procedure. The intraassay CVs were less than 10% for vasopressin and oxytocin. The interassay CV for vasopressin was 13%. The detection limit was 0.60 pmol/L for vasopressin and 1.0 pmol/L for oxytocin. Oxytocin levels were not measured in one subject. Basal estradiol and progesterone were measured by RIA at the NIH Clinical Center Laboratory at the time of exercise testing. The intraassay CVs were less than 5% for the estradiol and progesterone assays. Detection limits were 37 pmol/L for estradiol and 1.27 nmol/L for progesterone.
Statistical analyses
Data are presented as the mean ± SEM. Two-way repeated measures ANOVA was used to investigate the effects of exercise and cycle phase on hormonal variables. ACTH data were log transformed before the ANOVA because the distribution among subjects was not normal. If significant differences were indicated by ANOVA, post-hoc contrasts were used to evaluate differences at each time point. Single time point data were compared using two-tailed Students t-tests. Relationships among variables were evaluated using Pearsons correlation coefficient. The area under the curve was used for correlational analyses of multiple time point hormonal data and was calculated by the trapezoidal method with subtraction of the baseline. Significance was set at the 0.05 level.
Results
Basal hormone levels
As expected, basal plasma estradiol and progesterone levels were
lower in the early follicular compared with the midluteal phase of the
menstrual cycle [estradiol, 242 ± 48 vs. 580 ±
48 pmol/L (P < 0.001); progesterone, 2.9 ± 1.3
vs. 47.7 ± 5.1 nmol/L (P < 0.0001)].
Basal plasma lactate, glucose, ACTH, vasopressin, oxytocin, and
cortisol levels were similar in the early follicular and midluteal
phases of the cycle (Table 1
).
|
Hormonal responses to exercise are shown in Fig. 1
. Induction of anaerobic metabolism
during exercise was indicated by marked increases in plasma lactate
levels over time [F(6, 48) = 29.3; P < 0.0001];
values did not differ between phases of the menstrual cycle
[F(6, 48) = 0.94; P = 0.47].
|
A significant increase in vasopressin during the exercise test was also found [F(7,56) = 5.8; P < 0.0001] along with a significant menstrual cycle phase interaction [F(7, 56) = 3.0; P < 0.01]. As noted for ACTH, the vasopressin response was enhanced in the luteal phase.
As expected, a significant change in plasma glucose occurred during the exercise [F(6, 48) = 6.0; P < 0.0001]. As noted for ACTH and vasopressin, there was a significant interaction of glucose response to exercise and cycle phase [F(6, 48) = 2.5; P < 0.04], with exercise-induced glucose levels being higher in the luteal phase.
There was only a trend toward an effect of exercise on plasma cortisol release during the exercise test [F(7, 56) = 2.2; P < 0.06], and no interaction between the cortisol response to exercise and cycle phase [F(7, 56) = 1.1; P = 0.40]. There also was no main effect of cycle phase on cortisol levels [F(1, 8) = 1.2; P = 0.29].
In contrast to the other metabolic and hormonal parameters, there was no effect of exercise on plasma oxytocin [F(7, 42) = 0.617; P = 0.74]. Moreover, oxytocin secretion was not affected by cycle phase [F(1, 6) = 0.016; P = 0.90], and there was no interaction between the oxytocin response to the exercise test and cycle phase [F(7, 42) = 0.954; P = 0.48].
A significant positive correlation was found between basal plasma progesterone and the AUC for both ACTH (r = 0.54) and cortisol (r = 0.56) in the luteal phase. Basal estradiol levels correlated positively with AUC for ACTH (r = 0.80), cortisol (r = 0.55), and glucose (r = 0.75) in the follicular phase and the AUC for cortisol (r = 0.50) and glucose (r = 0.82) in the luteal phase.
Discussion
We found increased ACTH, vasopressin, and glucose responses to treadmill exercise in the midluteal compared with the early follicular phase of the menstrual cycle. The results of this study are compatible with those of a few human studies of HPA axis responsivity across the menstrual cycle, including two studies that found increased cortisol response to psychological stress in the luteal phase (15, 16), another study that reported increased cortisol response to prolonged (90-min) submaximal exercise in the luteal phase (17), and a fourth study that found a significant rise in cortisol after 30 min of submaximal exercise stress only the in midluteal phase of the cycle (18). However, many stress studies have not found menstrual cycle-related differences in HPA axis responsivity, including studies using intense (19) and moderate physical exercise (20, 21), endotoxin infusion stress (22), or psychological stressors (23, 24). The lack of a menstrual cycle effect on ACTH secretion in an earlier exercise stress study from our laboratory (21) and in an exercise study from another laboratory (20), both using lower intensity exercise, suggests that cycle phase differences in ACTH secretion are evident only during high intensity exercise. Our use of narrow windows for timing of tests, our choice of testing times to maximize differences in gonadal steroid levels, measurement of ACTH and vasopressin secretion in addition to cortisol secretion, and use of a relatively severe stress stimulus may have increased our ability to detect these cycle-related differences. Our finding of increased HPA axis responsivity in the luteal phase parallels our previous report of enhanced HPA axis responses to exercise in nonlactating postpartum women compared with lactating women who are in a state of relative gonadal steroid suppression (12).
The less robust cycle-related difference in cortisol secretion in response to the exercise testing is surprising in light of the enhanced ACTH and vasopressin responses in the luteal phase, both of which stimulate cortisol secretion, and in light of reports that adrenal sensitivity to ACTH is enhanced in the luteal phase (16, 25). This discrepancy most likely is due to the circadian fall in cortisol, which continued through the early morning testing. The circadian fall completely obscured any rise in plasma cortisol in response to the exercise test in half of the subjects. In contrast, cortisol responses were larger and more distinct in an exercise study that was performed in postpartum women in our laboratory using the same protocol but at midday, when ambient cortisol levels are lower (12).
The physiological mechanism that produces the increase in glucose level at baseline and throughout testing during the luteal phase is unclear. An earlier study from our laboratory using moderate intensity exercise during the follicular and luteal phases of the cycle also demonstrated higher glucose levels during the luteal phase (21). Another study noted the typical drop in glucose in response to submaximal exercise only in the luteal phase (17); this finding also suggests that glucose responses to stress are enhanced in the luteal phase. However, several other exercise studies found no difference in glucose levels across the menstrual cycle (19, 26). Impairment of glucose uptake and increases in fasting glucose in the luteal phase have been described in two studies that more directly focused on glucose regulation across the menstrual cycle (27, 28). Increased glucose responsivity noted in the luteal phase in our current study may be a consequence of increased vasopressin and cortisol responses in the luteal phase, as both of these hormones mobilize glucose. The lack of cycle phase differences in glucose response in other studies may result from less robust cycle-related differences in hormones that mobilize glucose.
The failure of exercise to stimulate oxytocin release in healthy women is consistent with the findings of other exercise and psychological stress studies (12, 29, 30). On the other hand, massage (31) and breast stimulation in the luteal phase (32) have been shown to provoke oxytocin release in women. The lack of oxytocin response to exercise in women stands in contrast to the oxytocin release in rats that occurs in response to a wide variety of physiological stressors (33). This species difference parallels observations that oxytocin increases plasma ACTH release in rats (33) and inhibits plasma ACTH release in humans (34, 35).
Menstrual cycle-related changes in several different regulatory elements of the HPA axis may contribute to the increased HPA axis responsivity in the luteal phase. Perhaps most important, gonadal steroids appear to regulate glucocorticoid receptor activity in the limbic-hypothalamic- pituitary-adrenal axis and are likely to thereby modulate feedback sensitivity of the HPA axis. In a prior study using the same sampling windows across the menstrual cycle, we found decreased lymphocyte glucocorticoid receptor mRNA expression in the midluteal compared with the early follicular phase. Reduced glucocorticoid receptor mRNA expression in the luteal phase was accompanied by reduced sensitivity to dexamethasone suppression of plasma cortisol (7). Animal studies are consistent with these observations, showing increases in HPA axis feedback sensitivity after ovariectomy (36). Enhanced plasma vasopressin release in the midluteal phase may also contribute to the enhanced ACTH response found in the midluteal phase by stimulating pituitary vasopressin receptors (37, 38). In addition, catecholamine activity is enhanced in the luteal phase (39, 40) and may contribute to enhanced HPA axis responsivity. Central serotonergic activity also may fluctuate across the menstrual cycle and contribute to changes in HPA axis regulation (41, 42, 43).
Because of the naturalistic design of this study, it was not possible to definitely sort out the differential effects of specific gonadal steroids on HPA axis responsivity. Both estradiol and progesterone levels correlated positively with at least some components of HPA axis responsivity. Although correlations between estradiol and HPA axis responses were seen in both the follicular and luteal phases of the cycle, progesterone levels all were close to the detection limit of the assay in the follicular phase; this prevented any meaningful associations with hormonal responses. Further studies using specific estrogen and progesterone receptor antagonists or single hormone replacement strategies in postmenopausal women should help to clarify the roles of individual gonadal steroid hormones in regulation of the HPA axis in humans.
Numerous studies of ovariectomized rats have found a stimulatory effect of estradiol on HPA axis responsivity (44, 45) and an inhibitory effect of estradiol on HPA axis feedback sensitivity (44, 46), but effects of progesterone on HPA axis regulation are less clear (44, 45, 47, 48, 49). Studies in ovariectomized primates have shown that ACTH and cortisol responses to interleukin-1 challenge were enhanced during replacement of late follicular compared with early follicular levels of estradiol (50, 51). The effects of estrogen on HPA axis responsivity in humans have been mixed, with increased HPA axis responsivity after 1 day of estradiol treatment in men (52), reduced responsivity after 68 weeks of estrogen treatment in postmenopausal women (53, 54), and no effect on HPA responses to stress after longer term treatment (55). Also, a recent study in men showed that 24-h administration of exogenous estrogen significantly enhanced HPA axis responses to interview stress. These effects of estrogen and progesterone on HPA axis regulation may be mediated by the effects of these gonadal steroids on glucocorticoid receptor feedback sensitivity. Preclinical data indicate that progesterone serves as an antagonist at glucocorticoid receptors (56, 57) and that short-term estrogen treatment in ovariectomized rats can down-regulate glucocorticoid receptors in the hippocampus, hypothalamus, and pituitary (47, 58, 59, 60). Preclinical data also indicate that estrogen may enhance the magnocellular release of vasopressin. In rats ovariectomy prevented the release of vasopressin into the systemic circulation in response to salt loading and also attenuated the induction of vasopressin mRNA expression in the magnocellular cells of the supraoptic and paraventricular nuclei of the hypothalamus (61). In addition, arginine vasopressin mRNA in the magnocellular division of the paraventricular nucleus was increased in ovariectomized rats in response to estrogen treatment (59).
Changes in HPA axis and plasma vasopressin responsivity across the menstrual cycle may be linked to fluctuations in mood and somatic symptoms. Multiple studies in nonclinical populations have documented increases in irritability, anxiety, and depression symptoms in the mid- and late luteal phases of the menstrual cycle (23, 62, 63). In addition, 38% of women are particularly sensitive to menstrual cycle hormonal changes and experience severe mood worsening in the luteal phase, with significant impairment of social and occupational functioning (11, 63, 64, 65). Suppression of HPA axis activation during the follicular phase of the menstrual cycle may be a reflection at least in part of suppression of central neuropeptides that drive the axis, particularly CRH and vasopressin. These peptides have arousal-producing, anxiogenic effects and have been argued to play an important role in the generation of depression and anxiety disorder symptoms (66). In addition, increased vasopressin responsivity in the midluteal phase of the menstrual cycle may contribute to subjective sensations of fluid retention premenstrually.
In summary, this study demonstrates suppression of HPA axis and plasma vasopressin reactivity to exercise stress in the early follicular compared with the midluteal phase of the menstrual cycle. These findings in combination with prior findings of reduced HPA axis and vasopressin responses to exercise in lactating women suggest that relatively low levels of gonadal steroids in premenopausal women may reduce the effects of stress on the HPA axis.
Received September 12, 2000.
Revised December 14, 2000.
Accepted December 23, 2000.
References
-OH-progesterone, and androgen response to a standardized
ACTH-stimulation in different stages of the normal menstrual cycle. Acta Endocrinol (Copenh). 100:427433.[Medline]
in the
ovariectomized rhesus monkey. J Neuroendocrinology. 7:387392.[CrossRef][Medline]
This article has been cited by other articles:
![]() |
M. Bloch, D. R. Rubinow, P. J. Schmidt, A. Lotsikas, G. P. Chrousos, and G. Cizza Cortisol Response to Ovine Corticotropin-Releasing Hormone in a Model of Pregnancy and Parturition in Euthymic Women with and without a History of Postpartum Depression J. Clin. Endocrinol. Metab., February 1, 2005; 90(2): 695 - 699. [Abstract] [Full Text] [PDF] |
||||
![]() |
E. A. YOUNG and M. ALTEMUS Puberty, Ovarian Steroids, and Stress Ann. N.Y. Acad. Sci., June 1, 2004; 1021(1): 124 - 133. [Abstract] [Full Text] [PDF] |
||||
![]() |
L. Zhao and R. D. Brinton Suppression of Proinflammatory Cytokines Interleukin-1{beta} and Tumor Necrosis Factor-{alpha} in Astrocytes by a V1 Vasopressin Receptor Agonist: A cAMP Response Element-Binding Protein-Dependent Mechanism J. Neurosci., March 3, 2004; 24(9): 2226 - 2235. [Abstract] [Full Text] [PDF] |
||||
![]() |
C. A. Roca, P. J. Schmidt, M. Altemus, P. Deuster, M. A. Danaceau, K. Putnam, and D. R. Rubinow Differential Menstrual Cycle Regulation of Hypothalamic-Pituitary-Adrenal Axis in Women with Premenstrual Syndrome and Controls J. Clin. Endocrinol. Metab., July 1, 2003; 88(7): 3057 - 3063. [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 |