Interactions between the Hypothalamic-Pituitary-Adrenal Axis and the Female Reproductive System: Clinical Implications

Dr. George P. Chrousos (Developmental Endocrinology Branch, National Institute of Child Health and Human Development [NICHD], National Institutes of Health [NIH], Bethesda, Maryland): Ancient physicians knew of the adverse effects of stress on the reproductive system [1, 2]. In the 5th century BCE, Hippocrates of Cos explained the impotence and infertility of the Scythians, nomadic tribes living in what is now southern Ukraine, as a result of their rough lives. About the men, he wrote, From the cold and tiredness they forget their sexual drive and their desire to come into union with the other sex; about the women, he stated, nor is their menstrual discharge such as it should be, but scanty and at too long intervals. About 500 years later, Soranos of Ephesus published the following differential diagnosis of amenorrhea in his pioneering treatise on gynecology and perinatology: Of those who do not menstruate, some have no ailment and it is physiological for them not to menstruate, either because of their age, as in those too young or too old, or because they are pregnant, or barren singers and athletes. Others, however, do not menstruate because of a disease of the uterus or of the rest of the body, for example when subjected to under-nourishment, great emaciation and wasting or to the accumulation of fatty flesh, or cachexia, or fevers and long ailment. The hypothalamic-pituitary-adrenal axis, together with the arousal and autonomic nervous systems, constitutes the stress system (Figure 1). This system is activated during stress and produces the clinical phenomenology of what Hans Selye described as the stress syndrome [3]. Indeed, during stress, several changes take place in the central nervous system and periphery of mammals, changes that help preserve the individual and the species. These include the mobilizing of adaptive behaviors and peripheral functions and the inhibiting of biologically costly behaviors and vegetative functions, such as reproduction, feeding, and growth. Figure 1. Interactions of the reproductive system with the hypothalamic-pituitary-adrenal axis and locus ceruleus-norepinephrine system (LC/NE). middle right The principal molecular regulators of the hypothalamic-pituitary-adrenal axis are corticotropin-releasing hormone (CRH), a 41-amino acid peptide, and the nonapeptide arginine-vasopressin, both of which are secreted by parvicellular neurons of the paraventricular nucleus of the hypothalamus into the hypophyseal portal system [3]. There, they synergistically stimulate pituitary adrenocorticotropic hormone (ACTH) secretion and, consequently, cortisol secretion by the adrenal cortex. The noradrenergic brainstem neurons that regulate the central arousal (locus ceruleus) and systemic sympathetic-adrenomedullary systems are innervated and stimulated by and reciprocally innervate and stimulate the parvicellular hypothalamic CRH and arginine-vasopressin neurons of the paraventricular nucleus. The female reproductive system is regulated by the hypothalamic-pituitary-ovarian axis (Figure 1). Neurons that secrete gonadotropin-releasing hormone in the preoptic and arcuate nucleus areas of the hypothalamus secrete into the hypophyseal portal system and stimulate the production of follicle-stimulating and luteinizing hormones, which then activate the ovary to secrete estradiol and progesterone [4]. In addition to acting on their other target tissues (other components of the central nervous system, uterus, genitalia, and skin), both of the gonadal steroids and another ovarian hormone, inhibin, exert negative feedback effects on the secretion of follicle-stimulating and luteinizing hormones. An excellent example of the effect of stress on the female reproductive system is so-called stress-induced or functional hypothalamic amenorrhea [5, 6]. Indeed, the prevalence of sustained secondary amenorrhea in normal young women is about 2%. This rate increases markedly in proportion to chronic stress, all the way up to 100% in prisoners before execution. Thus, if severe enough, stress can completely inhibit the female reproductive system. During her reproductive years, a normal woman is exposed to a monthly fluctuation of circulating estradiol and progesterone that may affect her behavior, mood, and immune and other functions. Indeed, epidemiologic data underscore the effect of gonadal function on nonreproductive female processes [7, 8]. Thus, suicide attempts and allergic bronchial asthma attacks correlate with the phase of the menstrual cycle, with fourfold increases in prevalence seen when the plasma estradiol level is at its lowest (that is, in the late luteal and menstruation phases) [9, 10]. Other studies have suggested that the period of peak estradiol secretion in the state immediately before ovulation is associated with elevations in mood, a phenomenon that might contribute to fecundity. Hypothalamic-Pituitary-Adrenal Axis and the Female Reproductive System Dr. David Torpy (Developmental Endocrinology Branch, NICHD, NIH): The hypothalamic-pituitary-adrenal axis, when activated by stress, has an inhibitory effect on the reproductive system; teleologically, this makes sense (Figure 1; Table 1). Indeed, the hypothalamic CRH neurons innervate and inhibit directly or indirectly, through proopiomelanocortin neurons, the hypothalamic control center of the gonadal axis [11]. In addition, glucocorticoids secreted from the adrenal cortex act at the levels of the hypothalamic, pituitary, gonadal, and end-target tissues to suppress the gonadal axis. On the other hand, estradiol exerts a negative, although indirect, effect on the activity of the gonadotropin-releasing hormone neuron, which has no detectable estrogen receptor [12]. Table 1. Interactions between the Stress System and the Female Reproductive System* The interaction between the hypothalamic-pituitary-adrenal and gonadal axes at the level of the hypothalamus was directly examined in rhesus monkeys [13]. Insulin-induced hypoglycemia caused an increase in cortisol levels and a decrease in plasma luteinizing hormone levels associated with reduced electrical activity measured directly at the gonadotropin-releasing hormone neuron. When a CRH antagonist was given intracerebroventricularly, the effect of insulin hypoglycemia on electrical activity at the gonadotropin-releasing hormone pulse generator was greatly attenuated; this finding suggests that CRH has a direct effect on the hypothalamic neurons that secrete gonadotropin-releasing hormone. Glucocorticoids inhibit gonadal axis function at the hypothalamic, pituitary, and uterine levels [14-16]. Sakakura and colleagues studied women who had received prednisolone for various indications for 1.5 to 5 months in daily doses ranging from 10 mg to 40 mg [15]. All of these women had menstrual disturbances associated with glucocorticoid treatment, and the investigators found that prednisolone reduced the peak luteinizing hormone response to intravenous gonadotropin-releasing hormone by about 60%. This suggests an inhibitory effect of glucocorticoids on the pituitary gonadotroph. Glucocorticoids also inhibit estradiol-stimulated uterine growth [16]. In one placebo-controlled experiment done in rats, dexamethasone and estradiol were administered for 5 days. Estradiol alone produced the expected increase in uterine weight; this increase was significantly attenuated by daily coadministration of dexamethasone, the effect of which may be partly explained by the reduced intracellular estrogen receptor concentrations measured in this experiment. Most likely, however, glucocorticoid receptor-mediated inhibition of the c-fos/c-jun transcription factor by protein-protein interaction is primarily responsible for this inhibition [17]; this factor is used in the signal transduction pathways of many growth factors and is directly or indirectly stimulated by estrogen [18]. Estrogen, which is derived principally from the ovaries, stimulates the hypothalamic-pituitary-adrenal axis (Figure 1). This had been suspected on the basis of sex differences in hypothalamic-pituitary-adrenal axis responses to stimuli in both animals and humans [19]. Compared with controls, pregnant women and women receiving high-dose estrogen therapy had elevated levels of free cortisol in both morning and evening plasma samples [20]. In addition, hypothalamic-pituitary-adrenal axis responsiveness is greater in women than in men. When ACTH and cortisol responses to ovine CRH were compared in 24 men and 19 women [21], the ACTH peak response was significantly greater in women and the cortisol response was characteristically prolonged in response to higher peak ACTH levels. Estrogen can induce hyperresponsiveness of the hypothalamic-pituitary-adrenal axis to stimuli in normal men; thus, this effect seems to be due to estrogen rather than to other factors specific to female physiology [22]. Recently, estradiol patches were given to normal men who were then subjected to a psychosocial stressor-unprepared public speaking-for 15 minutes. Cortisol and ACTH responses were greater in the estradiol recipients than in the placebo recipients. Similarly, the plasma norepinephrine response in these men was augmented by estrogen, possibly because of the stimulation of CRH neurons (which innervate and stimulate central noradrenergic neurons) or because of direct effects on the production or metabolism of norepinephrine [23, 24]. Estrogen stimulation of the hypothalamic-pituitary-adrenal axis may be exerted through interaction of the ligand-activated estrogen receptor with specific DNA sequences, the estrogen-responsive elements, in the promoter of the human CRH gene [25, 26]. Estrogen may exert some of its physiologic negative feedback effect on the reproductive axis through a subpopulation of CRH and proopiomelanocortin neurons that inhibit gonadotropin-releasing hormone and, hence, follicle-stimulating hormone and luteinizing hormone secretion. Evidence from studies in

D r. George P. Chrousos (Developmental Endocrinology Branch, National Institute of Child Health and Human Development [NICHD], National Institutes of Health [NIH], Bethesda, Maryland): Ancient physicians knew of the adverse effects of stress on the reproductive system (1,2). In the 5th century BCE, Hippocrates of Cos explained the impotence and infertility of the Scythians, nomadic tribes living in what is now southern Ukraine, as a result of their rough lives. About the men, he wrote, "From the cold and tiredness they forget their sexual drive and their desire to come into union with the other sex"; about the women, he stated, "... nor is their menstrual discharge such as it should be, but scanty and at too long intervals." About 500 years later, Soranos of Ephesus published the following differential diagnosis of amenorrhea in his pioneering treatise on gynecology and perinatology: Of those who do not menstruate, some have no ailment and it is physiological for them not to menstruate, either because of their age, as in those too young or too old, or because they are pregnant, or barren singers and athletes. Others, however, do not menstruate because of a disease of the uterus or of the rest of the body, for example when subjected to under-nourishment, great emaciation and wasting or to the accumulation of fatty flesh, or cachexia, or fevers and long ailment.
The hypothalamic-pituitary-adrenal axis, together with the arousal and autonomic nervous systems, constitutes the stress system ( Figure 1). This system is activated during stress and produces the clinical phenomenology of what Hans Selye described as the stress syndrome (3). Indeed, during stress, several changes take place in the central nervous system and periphery of mammals, changes that help preserve the individual and the species. These include the mobilizing of adaptive behaviors and peripheral functions and the inhibiting of biologically costly behaviors and vegetative functions, such as reproduction, feeding, and growth.
The principal molecular regulators of the hypothalamic-pituitary-adrenal axis are corticotropinreleasing hormone (CRH), a 41-amino acid peptide, and the nonapeptide arginine-vasopressin, both of which are secreted by parvicellular neurons of the paraventricular nucleus of the hypothalamus into This paper is also available at http://www.acponline.org.
For a glossary of terms used in this paper, see end of text.
An edited summary of a Clinical Staff Conference held on 20 November 1996 at the National Institutes of Health, Bethesda, Maryland.
Authors who wish to cite a section of the conference and specifically indicate its author may use this example for the form of the reference: Torpy D. Hypothalamic-pituitary-adrenal axis and the female reproductive system, pp 230-232. In: Chrousos GP, moderator. Interactions between the hypothalamic-pituitary-adrenal axis and the female reproductive system: clinical implications. Ann Intern Med. 1998;129:229-240. Stress generally inhibits the female reproductive system (middle) primarily through the hypothalamic-pituitary-adrenal axis (left) through 1) suppression of hypothalamic gonadotropin-releasing hormone secretion by corticotropin-releasing hormone (CRH) and CRH-induced /3endorphin; 2) inhibition of hypothalamic gonadotropin-releasing hormone (GnRH), pituitary luteinizing hormone (LH), and ovarian estradiol (E2) secretion by Cortisol; and 3) cortisol-induced target tissue resistance to estradiol. The locus ceruleus-norepinephrine system (right) provides positive input to the reproductive system, which is frequently overcome by the stress-activated hypothalamic-pituitary-adrenal axis. However, sexual stimulation and GnRH neuron activation may render the gonadal axis resistant to suppression by the hypothalamic-pituitary-adrenal axis. Through estradiol, the reproductive system provides positive input to both components of the stress system by stimulating CRH secretion and inhibiting reuptake and catabolism of catecholamines. a-MSH = melanocyte-stimulating hormone; ACTH = adrenocorticotropic hormone; AVP = arginine-vasopressin; FSH = follicle-stimulating hormone; NE (a) = norepinephrine stimulation via a-noradrenergic receptors; POMC = proopiomelanocortin. Solid line = stimulation; dotted line = inhibition. the hypophyseal portal system (3). There, they synergistically stimulate pituitary adrenocorticotropic hormone (ACTH) secretion and, consequently, Cortisol secretion by the adrenal cortex. The noradrenergic brainstem neurons that regulate the central arousal (locus ceruleus) and systemic sympatheticadrenomedullary systems are innervated and stimulated by and reciprocally innervate and stimulate the parvicellular hypothalamic CRH and argininevasopressin neurons of the paraventricular nucleus. The female reproductive system is regulated by the hypothalamic-pituitary-ovarian axis (Figure 1). Neurons that secrete gonadotropin-releasing hormone in the preoptic and arcuate nucleus areas of the hypothalamus secrete into the hypophyseal portal system and stimulate the production of folliclestimulating and luteinizing hormones, which then activate the ovary to secrete estradiol and progesterone (4). In addition to acting on their other target tissues (other components of the central nervous system, uterus, genitalia, and skin), both of the gonadal steroids and another ovarian hormone, inhibin, exert negative feedback effects on the secretion of follicle-stimulating and luteinizing hormones.
An excellent example of the effect of stress on the female reproductive system is so-called stress-induced or functional hypothalamic amenorrhea (5,6). Indeed, the prevalence of sustained secondary amenorrhea in normal young women is about 2%. This rate increases markedly in proportion to chronic stress, all the way up to 100% in prisoners before execution. Thus, if severe enough, stress can completely inhibit the female reproductive system.
During her reproductive years, a normal woman is exposed to a monthly fluctuation of circulating estradiol and progesterone that may affect her behavior, mood, and immune and other functions. Indeed, epidemiologic data underscore the effect of gonadal function on nonreproductive female processes (7,8). Thus, suicide attempts and allergic bronchial asthma attacks correlate with the phase of the menstrual cycle, with fourfold increases in prevalence seen when the plasma estradiol level is at its lowest (that is, in the late luteal and menstruation phases) (9,10). Other studies have suggested that the period of peak estradiol secretion in the state immediately before ovulation is associated with elevations in mood, a phenomenon that might contribute to fecundity.

Hypothalamic Pituitary Adrenal Axis and the Female Reproductive System
Dr. David Torpy (Developmental Endocrinology Branch, NICHD, NIH): The hypothalamic-pituitaryadrenal axis, when activated by stress, has an inhibitory effect on the reproductive system; teleologically, this makes sense ( Figure 1; Table 1). Indeed, the hypothalamic CRH neurons innervate and inhibit directly or indirectly, through proopiomelanocortin neurons, the hypothalamic control center of the gonadal axis (11). In addition, glucocorticoids secreted from the adrenal cortex act at the levels of the hypothalamic, pituitary, gonadal, and end-target tissues to suppress the gonadal axis. On the other hand, estradiol exerts a negative, although indirect, effect on the activity of the gonadotropin-releasing  hormone neuron, which has no detectable estrogen receptor (12). The interaction between the hypothalamic-pituitary-adrenal and gonadal axes at the level of the hypothalamus was directly examined in rhesus monkeys (13). Insulin-induced hypoglycemia caused an increase in Cortisol levels and a decrease in plasma luteinizing hormone levels associated with reduced electrical activity measured directly at the gonadotropin-releasing hormone neuron. When a CRH antagonist was given intracerebroventricularly, the effect of insulin hypoglycemia on electrical activity at the gonadotropin-releasing hormone pulse generator was greatly attenuated; this finding suggests that CRH has a direct effect on the hypothalamic neurons that secrete gonadotropin-releasing hormone.
Glucocorticoids inhibit gonadal axis function at the hypothalamic, pituitary, and uterine levels (14)(15)(16). Sakakura and colleagues studied women who had received prednisolone for various indications for 1.5 to 5 months in daily doses ranging from 10 mg to 40 mg (15). All of these women had menstrual disturbances associated with glucocorticoid treatment, and the investigators found that prednisolone reduced the peak luteinizing hormone response to intravenous gonadotropin-releasing hormone by about 60%. This suggests an inhibitory effect of glucocorticoids on the pituitary gonadotroph.
Glucocorticoids also inhibit estradiol-stimulated uterine growth (16). In one placebo-controlled experiment done in rats, dexamethasone and estradiol were administered for 5 days. Estradiol alone produced the expected increase in uterine weight; this increase was significantly attenuated by daily coadministration of dexamethasone, the effect of which may be partly explained by the reduced intracellular estrogen receptor concentrations measured in this experiment. Most likely, however, glucocorticoid receptor-mediated inhibition of the c-fos/c-jun transcription factor by protein-protein interaction is primarily responsible for this inhibition (17); this factor is used in the signal transduction pathways of many growth factors and is directly or indirectly stimulated by estrogen (18).
Estrogen, which is derived principally from the ovaries, stimulates the hypothalamic-pituitary-adrenal axis (Figure 1). This had been suspected on the basis of sex differences in hypothalamic-pituitaryadrenal axis responses to stimuli in both animals and humans (19). Compared with controls, pregnant women and women receiving high-dose estrogen therapy had elevated levels of free Cortisol in both morning and evening plasma samples (20). In addition, hypothalamic-pituitary-adrenal axis responsiveness is greater in women than in men. When ACTH and Cortisol responses to ovine CRH were compared in 24 men and 19 women (21), the ACTH peak response was significantly greater in women and the Cortisol response was characteristically prolonged in response to higher peak ACTH levels.
Estrogen can induce hyperresponsiveness of the hypothalamic-pituitary-adrenal axis to stimuli in normal men; thus, this effect seems to be due to estrogen rather than to other factors specific to female physiology (22). Recently, estradiol patches were given to normal men who were then subjected to a psychosocial stressor-unprepared public speaking-for 15 minutes. Cortisol and ACTH responses were greater in the estradiol recipients than in the placebo recipients. Similarly, the plasma norepinephrine response in these men was augmented by estrogen, possibly because of the stimulation of CRH neurons (which innervate and stimulate central noradrenergic neurons) or because of direct effects on the production or metabolism of norepinephrine (23,24). Estrogen stimulation of the hypothalamic-pituitary-adrenal axis may be exerted through interaction of the ligand-activated estrogen receptor with specific DNA sequences, the estrogenresponsive elements, in the promoter of the human CRH gene (25,26).
Estrogen may exert some of its physiologic negative feedback effect on the reproductive axis through a subpopulation of CRH and proopiomel- anocortin neurons that inhibit gonadotropin-releasing hormone and, hence, follicle-stimulating hormone and luteinizing hormone secretion. Evidence from studies in nonhuman primates suggests that in the period immediately before ovulation, a decrease in estradiol levels leads to reduced hypothalamic CRH secretion. This effectively disinhibits the gonadotropin-releasing hormone (GnRH) neuron and possibly participates in the generation of the ovulatory luteinizing hormone surge (27). This takes place simultaneously with a delayed estrogen-induced central noradrenergic surge that has an additional positive effect on the gonadotropin-releasing hormone neuron (24).
Estradiol also downregulates glucocorticoid receptor binding in the anterior pituitary, hypothalamus, and hippocampus; this tends to increase hypothalamic-pituitary-adrenal axis activity by interfering with glucocorticoid negative feedback, whereas progesterone opposes these effects (28). It is not known whether the changes induced by estrogen in neuronal CRH and glucocorticoid receptor activities are mechanistically related, but they do alter the system in the same direction. One should also keep in mind the regulatory feedback loops and adaptations that take place over time as new equilibria are established in the relation shown in Figure 1 (29,30).
The newly discovered adipocyte-derived peptide hormone leptin interacts directly and indirectly with both the adrenal and gonadal axes, and its levels are higher in women than in men (31, 32) (Figure 2). By promoting satiety and sympathetic system out-flow, leptin is thought to provide the peripheral signal to a central mechanism regulating the size of body fat stores (33). Leptin suppresses the hypothalamic-pituitary-adrenal axis by inhibiting hypothalamic CRH and adrenocortical Cortisol secretion (34,35) while it stimulates gonadal function by potentiating the activity of the gonadotropin-releasing hormone neuron (36). Thus, increasing leptin may be involved in the control of the onset of puberty, a phenomenon long known to be temporally related to the acquisition of a certain fat mass (36)(37)(38). Low leptin levels may be involved in the adaptive activation of the hypothalamic-pituitary-adrenal axis and the inhibition of gonadal function that takes place in starvation and anorexia nervosa (39)(40)(41). Some of the effects of leptin on the central nervous system are mediated by inhibition of the potent orexogen neuropeptide Y, which normally stimulates the CRH neuron and inhibits the locus ceruleus-norepinephrine system (42)(43)(44).

Central and Peripheral Roles of Corticotropin-Releasing Hormone
Dr. George P. Chrousos: The marked changes that take place in a woman's reproductive system during her life are bound to affect the functioning of the stress system. The first of these changes takes place at puberty, when gonadarche is slowly established with increasing ovarian follicle growth and circulating estradiol levels first and then the estab-lishment of ovulatory menstrual cycles within the next 2 to 3 years (Figure 3). During this time, the stress system receives increasing intermittent positive input from estradiol. Puberty is a period of increasing vulnerability to disorders or states characterized by disturbances or changes in hypothalamic CRH secretion (3,45), such as melancholic and atypical depression, eating disorders, chronic active alcoholism or other addictions, and chronic active athleticism, as well as seasonal affective disorder, the chronic fatigue and fibromyalgia syndromes, and several autoimmune disorders (Table 2).
Once established, the monthly fluctuations of estradiol that accompany menstrual cycles are expected to influence the secretion of central nervous system CRH and catecholamines until menopause (Figure 4). Decreased secretion of CRH in the late luteal and menstruation phases would be expected and might help explain the presence of luteal dysphoric mood disorder (the premenstrual tension syndrome) and the increased incidence of suicides and enhanced vulnerability to autoimmune and allergic inflammatory phenomena seen during these periods (7,8,46,47) (Table 2). Finally, during the perimenopausal period and early menopause, there is a progressive, intermittent decrease in estradiol levels that would be expected to be associated with decreased activity of the CRH and locus ceruleusnorepinephrine systems and might help explain the characteristic "hot flashes" and so-called climacteric depression (Table 2). "Reproductive" CRH has been identified in various reproductive tissues and can, accordingly, be ovarian, testicular, endometrial, or placental. It is a form of "tissue" corticotropin-releasing factor (CRH found in peripheral tissues) and is analogous to the "immune" CRH found in immune organs and inflammatory sites (48). The functions of immune CRH may shed light on those of reproductive CRH and are briefly discussed below.
Inflammatory sites examined by immunohistochemistry and extraction-chromatography contain large amounts of immune CRH, which is identical to hypothalamic CRH (48). Endothelial cells, macrophages, and tissue fibroblasts all have CRH in their cytoplasm. Immune neutralization and CRH antagonist experiments have demonstrated marked inhibition of inflammation indices, such as the volume of the inflammatory exudate and its leukocyte concentration (48)(49)(50)(51). Immune CRH is present at high levels in many sites of experimental inflammation in the rat and mouse and in all natural inflammatory sites examined thus far in humans. The latter include the inflamed joints of patients with rheumatoid arthritis and osteoarthritis and the thyroid glands of patients with Hashimoto thyroiditis (52,53). The exact mechanisms by which immune CRH exerts its proinflammatory actions are not known, but one mechanism is the degranulation of mast cells (54). Indeed, CRH causes vasodilation, increases vascular permeability, and allows extravasation of plasma through the capillary vessel walls (48).
In inflammatory sites, CRH is not only generated by immune system cells but is also secreted from the terminals of sympathetic postganglionic nerves and primary afferent nerves, whose cell bodies in the sympathetic and dorsal root ganglia contain large amounts of CRH (48). Secretion of immune CRH is suppressed by glucocorticoids and somatostatin. Female rats have greater inflammatory responses and produce more immune CRH in inflammatory sites than male rats do, and the presence of estrogen seems to cause the difference. Despite high local production of immune CRH in inflammatory sites, concurrent plasma concentrations are extremely low, probably as a result of rapid clearance mechanisms.
Corticotropin-releasing hormone and its receptors are also present in rat and human ovaries (Table 3). Ovarian CRH is primarily found in the theca and stroma and also in the cytoplasm of the ovum itself (55,56). Corticotropin-releasing hormone receptors, which are type 1 (similar to those of the anterior pituitary), are also found primarily in the stroma and theca and in the cumulus oophorus, whereas the follicular fluid contains CRH as well. The findings suggest that CRH may participate in the communication between the ovum and the cumulus oophorus and may influence ovarian steroid biosynthesis. Incubation of granulosa-lutein cells with CRH suppresses estradiol and progesterone secretion in a dose-dependent, interleukin-1-mediated manner (57,58). In this sense, ovarian CRH has antireproductive actions that might be related to the earlier menopausal failure of ovaries in women exposed to high psychosocial stress (59). We believe that a major physiologic function of ovarian CRH is its participation in the "aseptic" inflammatory phenomena of the ovary, including ovulation and luteolysis.
The human endometrium also contains CRH (Table 3). In fact, the endometrial glands are full of CRH during both the proliferative and the secretory phases of the cycle (60). In the luteal phase, CRH is probably secreted into the lumen of the uterus, where it may participate in the inflammatory phenomena of blastocyst implantation and (later in the cycle) of menstruation. Compared with interimplantation sites, implantation sites in rat endometrium show local extravasation of plasma and contain increased amounts of CRH messenger RNA and CRH (61). We found CRH expression in human decidualized endometrial stroma, and other researchers demonstrated that CRH itself decidualized endometrial stroma cells synergistically with progesterone (60,62). It is interesting that the effect of estradiol on CRH transcription in immortalized human uterine epithelial cells seems to be inhibitory rather than stimulatory; this may explain, to some extent, the contraceptive properties of the "day after" pill, which contains high doses of estrogen (63).
The latter half of human pregnancy is associated with hypercortisolism ( Figure 5). Indeed, the levels of free plasma Cortisol and 24-hour urinary free Cortisol excretion in pregnancy overlap with levels in patients with mild Cushing syndrome (64). In the same vein, dexamethasone cannot properly suppress Cortisol in late pregnancy, just as it cannot in the Cushing syndrome. Placental CRH causes this hypercortisolism of human pregnancy (Table 3). By 28 to 30 weeks' gestation, CRH levels in plasma are similar to those in the portal system, whereas the levels of CRH-binding protein are similar to those in nonpregnant women and normal men (65,66). At 34 to 35 weeks' gestation, CRH-binding protein concentrations decrease by two thirds, whereas total and free CRH levels are markedly increased in plasma during labor and return to undetectable amounts within hours after delivery.
In the early 1980s, several groups demonstrated that placental CRH was produced by the syncytiotrophoblast, chorion, amnion, and decidua and that it was the product of the same gene that produces hypothalamic CRH (65,67). Incubation of human placental tissue with CRH caused secretion of j3-endorphin and a-melanocyte-stimulating hormone in a dose-dependent manner (68). In addition, CRH caused stimulation of prostaglandin-E 2 and prostaglandin-F 2 a, both of which have a role in labor and delivery; in contrast, CRH receptors were shown in the myometrium, where CRH had a constrictive effect in synergy with oxytocin (69)(70)(71). In addition, CRH was found to stimulate nitric oxide production by the endothelium of placental vessels and to cause the dilation of these vessels, thus facilitating fetoplacental circulation (72).
To determine whether placental CRH was secreted in a pulsatile or circadian fashion in the third trimester of pregnancy and whether there were any correlations over time between placental CRH and the hypothalamic-pituitary-adrenal axis hormones, we studied normal pregnant women in the 34th week of pregnancy (Figure 6). We found CRH, ACTH, and Cortisol pulsations but, in contrast to ACTH and Cortisol (which were secreted in a circadian fashion), plasma CRH did not have a circadian rhythm (73). In the time cross-correlation analyses, CRH levels correlated positively with those of ACTH and Cortisol, which means that either CRH causes secretion of ACTH and Cortisol or Cortisol stimulates placental CRH secretion, or both. Indeed, glucocorticoids stimulate placental CRH secretion in cultured human placental cells (74). Thus, in the last trimester of pregnancy, the placenta secretes CRH, which seems to be under the positive influence of Cortisol. Circulating placental CRH then causes ACTH secretion, in synergy with portal parvicellular arginine-vasopressin, which thus seems to be responsible for generating the pulsations and circadian rhythm of ACTH and Cortisol (75). The persistent elevation of plasma ACTH and a-melanocyte-stimulating hormone levels then may cause some hypertrophy of the adrenal cortices in normal pregnant women. Thus, placental CRH seems to be responsible for the maternal hypercortisolism of pregnancy; for maintaining proper blood supply to the fetus (probably by activating the nitric oxide synthase of these vessels); and, later, before labor begins, for causing increased myometrial contractility. Plasma levels of CRH are markedly elevated in preeclamptic or eclamptic mothers, in mothers with intrauterine infections, and in healthy pregnant women during normal labor. Longitudinal studies of many hundreds of pregnant women demonstrated that in the latter half of pregnancy, one could predict the onset of labor by plasma levels of CRH (76). Thus, women who delivered prematurely had higher levels of CRH, equivalent to those of women close to term; the opposite biochemical profile was seen in women with postmature labor. On the basis of these data, CRH was proposed to be the biological clock that times labor and delivery.
Anoxia, the inflammatory cytokines, several prostaglandins, and glucocorticoids themselves cause placental CRH secretion in vitro and in vivo. This means that sustained anoxia caused by preeclampsia or eclampsia, increases of circulating cytokine levels caused by infection or inflammation, and increases of glucocorticoid concentration caused by physical or emotional stress may all initiate premature labor through increases in CRH secretion (Figure 7). Potent CRH receptor antagonists that might be helpful in delaying premature labor and delivery are being developed, and preliminary results are promising (51).
The clinical implications of placental CRH extend beyond pregnancy, labor, and delivery (76, 77) ( Figures 5 and 6). The postpartum period is characterized by an increased incidence of psychiatric and autoimmune manifestations. Indeed, the "post- In the immediate postpartum period, the loss of placental CRH and of estradiol input to the hypothalamic CRH neuron results in a period of low hypothalamic CRH secretion and, hence, increased vulnerability to mood disturbances, such as postpartum blues, depression, or psychosis, or to autoimmune disorders, such as postpartum thyroiditis. In the nonpregnant state, the pulsatility and circadian rhythm of the hypothalamic-pituitary-adrenal axis are maintained by pulsations of hypothalamic CRH and arginine-vasopressin (AVP). In the pregnant state, peripheral placental CRH usurps the role of hypothalamic CRH and causes hypercortisolism while pulsatility and circadian rhythm are maintained by arginine-vasopressin. In the postpartum state, hypothalamic (portal) CRH is initially suppressed and gradually returns to normal. During this period, pulsatility and circadian rhythm are possibly maintained by portal argininevasopressin as input from portal CRH gradually increases over time; the predominance of arginine-vasopressin might explain the decreased Cortisol suppressibility by dexamethasone previously seen in the early postpartum period. ACTH = adrenocorticotropic hormone; E2 = estradiol; P4 = progesterone. Solid line = stimulation; dotted line = inhibition. Adapted from reference 73. In addition, autoimmune diseases, such as "postpartum thyroiditis" and rheumatoid arthritis, frequently develop or are acutely exacerbated during the first few months postpartum.
Although several depressive conditions, such as melancholic depression and anorexia nervosa, are typically associated with high hypothalamic CRH secretion, other states, such as atypical or seasonal depression, the chronic fatigue and fibromyalgia syndromes, and the Cushing syndrome before and during the first year after cure, are all associated with decreased production of hypothalamic CRH (3,(77)(78)(79). We hypothesized that the postpartum period might be associated with low hypothalamic CRH secretion, which would predispose patients to atypical depression and autoimmune phenomena. We prospectively studied 17 pregnant women who were healthy and had no personal or family history of depression (77). Psychometric testing was done serially beginning with the 20th week of pregnancy and continuing up to a year postpartum. We did ovine CRH stimulation tests at 3, 6, and 12 weeks postpartum. Nine women had normal affect throughout, but 7 developed postpartum blues and 1 developed full-fledged postpartum depression. Plasma levels of ACTH before and after ovine CRH showed little response at 3 weeks, a better but still suppressed response at 6 weeks, and an almost normal response at 12 weeks. Cortisol levels remained in the upper normal range throughout, mostly because plasma cortisol-binding globulin levels were about twice the normal level at 3 weeks, and it took about 3 months or longer for them to decrease to within the normal range. When we separated the women at 3, 6, and 12 weeks into those with the blues or depression and those without a mood dis-order, the former had marked suppression of the ACTH response to ovine CRH compared with the latter. Thus, whereas the overall data showed a transient suppression of hypothalamic CRH secretion in the postpartum period, women with the blues or depression had a more severe suppression that lasted longer.
In the postpartum state, the major source of CRH and estrogen, the placenta, is no longer present, whereas the hypothalamic CRH neuron is probably suppressed as a result of previous exposure to high levels of Cortisol for more than 3 months and because of concurrent estrogen deficiency (Figures 5 and 6). Despite the suppressed ACTH responses, total plasma Cortisol levels were not low in our postpartum women, probably because of the elevated cortisol-binding globulin concentrations and because the adrenal glands were probably hypertrophic and, hence, hyperresponsive to ACTH from their increased function during the last months of pregnancy. High-dose estrogen has a marked antidepressant effect during this time, possibly because it reestablishes normal stress system secretion of CRH and norepinephrine (80).
These data may be extended to two other forms of estrogen-related depressive syndromes in women. One is the premenstrual tension syndrome, a condition associated with mild alterations of hypothalamic-pituitary-adrenal axis function throughout the cycle (46). The time of maximal emotional disturbance coincides with the decrease of plasma estradiol levels during the latter part of the luteal phase of the cycle, when plasma progesterone levels are still elevated. The other is climacteric depression, a condition about which we have a suggestion but no conclusive evidence of hypothalamic-pituitary-adrenal axis dysfunction (81). Studies measuring morning plasma Cortisol levels indicated that these levels were decreased by about 40% to 50%. These studies suggest that estrogen withdrawal took place during the perimenopausal and early menopausal period and was probably accompanied by transient hypoactivation of the stress system. Women with both of these estrogen and, hence, CRH withdrawal states show mood improvements with estrogen therapy.

Conclusions
Dr. Philip W. Gold (Clinical Neuroendocrinology Branch, National Institute of Mental Health, Bethesda, Maryland): Corticotropin-releasing hormone and estrogens seem to play crucial roles in the sexual dimorphism of the physiologic and pathophysiologic manifestations of the human stress response. In addition to coordinating the behavioral, neuroendocrine, metabolic, and immune compo-nents of the stress response, CRH seems to have direct reproductive regulatory roles at the hypothalamus, influencing gonadotropin-releasing hormone secretion, and at the periphery, promoting inflammatory phenomena, such as ovulation and implantation. Placental CRH drives the pituitary-adrenal axis to produce high Cortisol secretion during the latter part of pregnancy. Withdrawal of placental CRH after delivery provokes a secondary hypothalamic CRH deficiency with varying consequences for mood disorders and autoimmune phenomena. Corticotropin-releasing hormone may be the placental clock determining the onset of parturition. Thus, disturbances in placental CRH production may account for premature or delayed parturition in some cases. In addition to its profound feminizing effects, estrogen stimulates CRH gene expression and noradrenergic function. Alterations in estrogenic tone during the menstrual cycle, the postpartum period, and the climactery may be involved in the mood disturbances and alterations in immune function seen at these times.
The World Health Organization has determined that unipolar depression confers greater overall morbidity on women than any other cause does. It is conservatively estimated that 9% to 12% of adult women are affected by unipolar depression, with a ratio of affected females to affected males of at least 2:1. The classic form of depression, melancholia, is a state of pathologic hyperarousal characterized by profound anxiety about the adequacy of self, dread for the future prospects of such a deficient self, insomnia, anorexia, loss of libido, and other manifestations compatible with a hyperfunctional stress system (3,82). We and others have demonstrated hypersecretion of central nervous system CRH in melancholia, which results not only in the hypercortisolism and increased sympathetic activity of this condition but also in many of its other clinical and biochemical manifestations, including hypogonadism, inhibition of the growth hormone and thyroid axes, and mild immunosuppression (3,48,83). The constellation of these biochemical changes leads to several serious public health consequences, including osteoporosis and increased risk for bone fracture (84) and shortened life expectancy, mostly from cardiovascular disease (85). Patients with melancholia have a twofold increased risk for dying from ischemic heart disease and a mortality rate of approximately 50% after acute, serious physical illness (this rate is 10% in nondepressed patients) (86,87). This may be due to the adverse effects of chronic hypercortisolism on visceral fat, lipid metabolism, insulin sensitivity, blood coagulation, and arterial pressure (for example, changes constituting the metabolic syndrome X and participating in the development of atherosclerosis) (88). Long-term adminis-tration of antidepressant agents leads to suppression of CRH secretion and, it is hoped, to alleviation of its long-term behavioral, biochemical, and somatic sequelae (89,90).
Although melancholic depression is easier to recognize, another prevalent form of major depression-atypical depression-is also more common in women than in men (3,82). Atypical depression seems to be the antithesis of melancholia. Patients feel lethargic, fatigued, and unmotivated and demonstrate hyperphagia and hypersomnia. Dysphoria in patients with atypical depression more closely reflects feeling less alive than usual rather than intense anxiety about self. Several lines of evidence suggest that the lethargy, fatigue, hypersomnia, and hyperphagia of atypical depression are associated with hyposecretion of CRH (3). This hyposecretion may contribute to a significant increase in the incidence of allergic and autoimmune phenomena in this form of depression and in its two homolog transient states, the postpartum blues or depression and the postcure state of patients with the Cushing syndrome (48,77,79).

Glossary
Arcuate nucleus: Hypothalamic nucleus that contains neurons secreting peptides of importance to reproduction and the stress response, including gonadotropin-releasing hormone; neuropeptide Y; and proopiomelanocortin-derived peptides, such as j3-endorphin and a-melanocytestimulating hormone.
c-fos/c-jun transcription factor: A heterodimeric factor that stimulates the promoters of many genes related to cellular growth and replication.
Locus ceruleus-norepinephrine system: Noradrenergic nuclei of the brain stem that regulate arousal and sympathetic system activity.
Orexogen: Any substance with appetite-stimulating properties, such as neuropeptide Y.
Parvicellular neurons: Small-cell body neurons of the hypothalamus that secrete CRH or arginine-vasopressin in the hypophyseal portal system. Contrast with magnocellular neurons that secrete arginine-vasopressin into the systemic circulation.
Proopiomelanocortin: Precursor molecule for corticotropin, j3-endorphin, and a-melanocyte-stimulating hormone expressed primarily in the arcuate nucleus of the hypothalamus and the anterior pituitary gland.
Stress: State of threatened homeostasis.