Inadequate Luteal Function Is the Initial Clinical Cyclic Defect in a 12-Day Stress Model that Includes a Psychogenic Component in the Rhesus Monkey

The Journal of Clinical Endocrinology & Metabolism, May 2002

Xiao, Ennian, Xia-Zhang, Linna, Ferin, Michel

A PDF file should load here. If you do not see its contents the file may be temporarily unavailable at the journal website or you do not have a PDF plug-in installed and enabled in your browser.

Alternatively, you can download the file locally and open with any standalone PDF reader:

https://academic.oup.com/jcem/article-pdf/87/5/2232/9158457/jcem2232.pdf

Inadequate Luteal Function Is the Initial Clinical Cyclic Defect in a 12-Day Stress Model that Includes a Psychogenic Component in the Rhesus Monkey

The Journal of Clinical Endocrinology & Metabolism Inadequate Luteal Function Is the Initial Clinical Cyclic Defect in a 12-Day Stress Model that Includes a Psychogenic Component in the Rhesus Monkey ENNIAN XIAO 0 1 LINNA XIA-ZHANG 0 1 MICHEL FERIN 0 1 0 Abbreviations: CV, Coefficient of variation; HPA , hypothalamicpituitary-adrenal 1 Department of Obstetrics and Gynecology and Center for Reproductive Sciences, College of Physicians and Surgeons, Columbia University , New York, New York 10032 , USA As part of our goal to develop nonhuman primate models to prospectively study how different types of stress may affect the menstrual cycle, we have investigated whether a shortterm stress challenge that includes a significant psychogenic component can induce cyclic dysfunction. The study was performed in rhesus monkeys. The stress challenge had several components that included the psychological response to both a tethering system and to a simultaneous move to an unfamiliar environment and the response to the short surgical procedures required to install and disconnect the tethering system. The stress challenge lasted for 12 d and was initiated in the follicular (n 5) or luteal (n 6) phase of the menstrual cycle. At the end of the stress period, the tethering system was removed, and the animal was returned to its regular housing. To monitor cyclicity, FSH, LH, E2, and progesterone were measured daily throughout the two preceding control cycles, the experimental cycle, and the two poststress cycles, whereas the adrenal endocrine axis response was monitored by measuring cortisol. Animals remained ovulatory after the shortterm stress; however, integrated progesterone secretion in the luteal phase (from the day of LH surge 1 to the day of menstruation 1) of the stress cycle was significantly decreased by 51.6% when the stress was initiated in the follicular - T HERE IS AGREEMENT that the functional hypothalamic chronic anovulation syndrome, a common form of ovulatory impairment, is linked to lifestyle variables such as psychogenic stress and to the resultant suppression of the GnRH pulse generator (1). Yet, the independent association between stress and cyclic dysfunction remains to be demonstrated in the human (2). Research into the mechanisms of stress is complicated by the fact that in the established clinical syndrome it may be impossible to trace the original stress and hence to analyze the central neuroregulatory networks that produce the original cyclic disturbance. Furthermore, in the chronic situation, the response to stress may evolve, and different neuroendocrine elements may become involved with time. The research challenge is to identify relevant stress paradigms so that prospective investigations of the putative hypothalamic-pituitary-adrenal (HPA)-hypothalamic-pituitary-gonadal link can then be initiated. Our goal is to develop models in the nonhuman primate so that effects of stress on the menstrual cycle can be studied in a prospective approach and the activated neuroendocrine pathways can be investigated. We have demonstrated prephase and by 30.9% when it started in the luteal phase. Lower integrated LH levels (luteal d 5–13) accompanied the decreased progesterone. Cyclic parameters were still abnormal in the first poststress cycle, such as a prolonged follicular phase after a stress in the preceding follicular phase or inadequate luteal function after a stress in the preceding luteal phase. Within 4 h of the stress, there was a rapid 3-fold increase in cortisol levels over controls. Levels decreased progressively thereafter but remained significantly higher than controls during the entire short-term stress period. They were still significantly higher in the first 2 wk after stress. Overall, the data suggest that secretory inadequacy of the corpus luteum represents a first clinical stage in the damage that stress can inflict on the normal menstrual cycle. Of interest is the observation that this limited 12-d stress, which includes a significant psychogenic component, continues to produce detrimental effects on the menstrual cycle past the period during which it is exerted. Significant decreases in integrated luteal LH values in the poststress cycle suggest that these effects may be related to continuing disturbances in the neuroendocrine component of the reproductive axis. (J Clin Endocrinol Metab 87: 2232–2237, 2002) viously that a short-term inflammatory stress in the monkey disrupts the normal menstrual cycle by significantly altering follicular and luteal function (3, 4). Psychogenic stress is another potentially important and relevant model because psychophysiological and behavioral responses to life events have been speculated to activate central neuroregulatory mechanisms that will interfere with normal cyclic function and produce the functional hypothalamic chronic anovulation syndrome (1). Importantly, psychogenic stress has been reported to result in a long-term activation of the HPA axis, as indicated by an increase in cortisol, in a majority of women with the established syndrome (5–7). In a first step, we investigated in the rhesus monkey the effects of a 12-d stress challenge that includes a significant psychogenic component and is initiated during either the follicular or luteal phase of the menstrual cycle. We report that this stress model is sturdy enough to induce rapid changes in the hypothalamic-pituitary-gonadal axis and that the initial cyclic dysfunction results in inadequate luteal function. Materials and Methods Eleven adult female rhesus monkeys (Macaca mulatta) (body weight, 5.1–11.5 kg), with ovulatory cycles, were used in this study. All animals Animals had been housed in individual cages located within the same areas of a large housing room for several years. The housing facility was temperature- and light-controlled (23–25 C; lights on 0730 –1930 h). The animals were fed a high-protein diet (PNI Feeds Inc., St. Louis, MO), supplemented by fruits or vegetables, and had unlimited access to water. Daily blood samples were obtained at 1100 h by venipuncture without sedation, a procedure to which the animals had previously been habituated. Menstruation was checked by daily vaginal swabbing. Protocols The experimental protocols were conducted in accordance with the NIH Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of Columbia University. The overall experimental objective was to investigate the endocrine and menstrual cycle responses to a 12-d stress challenge with a significant psychogenic component. Experimental protocols were performed between the months of September and April. Only monkeys showing two ovulatory cycles and an integrated value for luteal progesterone equal to or greater than 48 ng/ml, a value previously defined as a normal representative luteal phase in our colony (see group 2 in Ref. 3), were used in this study. The stress challenge had several components that included the psychological response to a headcap and tethering system and to a simultaneous move to an unfamiliar environment, and the response to the short surgical procedures required to install and remove the headcap. To install the headcap, animals were sedated with ketamine (Ketaset, 5–7 mg/kg; Fort Dodge Animal Health, Fort Dodge, IA) and intubated. Gas anesthesia (Isoflurane) was then started. The monkey’s head was mounted on a stereotaxic head holder while the upper part of the body was supported with a soft pillow. The skin on top of the head was incised, and a stainless steel headcap (weight, 20 g) was anchored to the skull with four screws (size 4). The surgery procedure lasted 30 min or less. Within the cage, a tether was then attached to the headcap and connected to a swivel anchored to the top of the cage. This device did not restrict animal movement within its cage. The cage was then installed in an unfamiliar room separate from the housing area. The animals shared this room with a few other animals, but the original close long-term relationship with familiar animals in the large housing area was lost. The stress challenge lasted for 12 d and was initiated either in the follicular phase (d 1–2; n 5) or in the luteal phase (d 1– 4; n 6). At the end of the stress period, the animals were briefly sedated (about 10 min) with ketamine for the removal of the headcap and skin suture and returned to the large housing room within their original cage. Anti-pain therapy (aspirin suppositories, 300 mg) was administered at the completion of each surgical procedure. Both the experimental cycle during which the stress was applied and the two poststress cycles were monitored. FSH, LH, E2, and progesterone were measured daily throughout the entire observation period. Inhibin A levels were measured during the luteal phase (d 5–13) in control cycle 2 and in the stress cycle. The response of the HPA endocrine axis was monitored by measuring cortisol daily during the short-term stress challenge. An additional sample was taken at 1500 h during control cycle 2 and on stress d 1, 3 h after the challenge started, as well as 2 and 4 wk after the stress. Assays and statistical analysis Blood samples were centrifuged, and sera were kept frozen at 20C until assay. LH was measured with a recombinant homologous RIA, as described previously (8). Assay sensitivity (at 95% binding) was 0.06 ng/ml. Intra- and interassay coefficients of variation (CV) were 7.9 and 13.1%, respectively. FSH was measured with a recombinant cynomolgus monkey FSH RIA kit (provided by Dr. A. F. Parlow, Pituitary Hormones and Antisera Center, Harbor-University of California Los Angeles Medical Center, Torrance, CA). Synthetic cynomolgus monkey FSH (AFP6853A; Genzyme, Cambridge, MA) was used for reference and iodination, whereas rabbit antirecombinant cynomolgus monkey FSH antibody (AFP 782594) was used as the first antibody, at a dilution of 1:750,000. In the 10 most recent assays, mean total binding was 27.8% whereas the slope of the dose-response curve was 2.37. Assay sensitivity (at 95% binding) was 0.045 ng/ml, and mean intra- and interassay CV were 5.0 and 6.1%. Dilution curves with pools of sera from ovariectomized rhesus monkeys were parallel to standard curves. E2, progesterone, and cortisol were measured by chemiluminescent immunoassays using the Immulite system (Diagnostic Products Corporation Inc., Los Angeles, CA). Assay sensitivity was 20 pg/ml for E2, 0.2 ng/ml for progesterone, and 1 g/dl for cortisol. Interassay CV were 11.9% for E2, 11.1% for progesterone, and 11.6% for cortisol. Inhibin A was measured by ELISA (Diagnostics Systems Laboratories, Inc., Webster, TX). Assay sensitivity was 10 pg/ml, and intra- and interassay CV were 10.8 and 14%. Cycle parameters, such as the length of the follicular and luteal phase, hormone concentrations, as well as luteal function as reflected by integrated progesterone and inhibin A values in the luteal phase, were compared in the control, stress, and poststress periods. For luteal progesterone evaluation, the areas under the luteal phase progesterone curves (from the day of LH surge 1 to the day of menstruation 1) were calculated by trapezoidal analysis. For luteal inhibin A evaluation, integrated values in control cycle 2 and in the stress cycle were calculated from luteal d 5–13 to reflect the active days of inhibin A secretion. LH and FSH luteal-follicular phase transition levels (d 3 to d 2 of menses) and integrated values for luteal d 5–13 were also compared in control and experimental cycles. Comparisons were made by multiple ANOVA followed by the Tukey test. The level of significance was established at P value less than 0.05. A t test was used for the comparison of inhibin A values between control cycle 2 and the stress cycle. Results Stress challenge in the follicular phase When the stress challenge was initiated in the follicular phase, all monkeys but one continued to show regular ovulatory menstrual cycles (Fig. 1). Follicular phase length and E2 peaks (overall control cycles, 195.8 10.8 pg/ml; stress cycle, 187.0 26.2; mean se) were not significantly changed in the stress cycle. However, progesterone secretion in the luteal phase after the stress challenge was significantly decreased by 51.6%. Integrated LH secretion during the luteal phase of the stress cycle (luteal d 5–13) was also significantly decreased. Integrated luteal inhibin A levels (d 5–13) were 292.6 75.9 (se) pg/ml in control cycle 2 and lower but not significantly in the stress cycle (128.7 24.3; P 0.0555). The length of the follicular phase of the first poststress cycle was significantly prolonged, and the E2 peak (211.6 41.1) was delayed. FSH levels during the luteal-follicular transition period were not altered (control cycles, 1.80 0.10, 1.74 0.06; stress cycle, 1.68 0.08; poststress cycles, 1.72 0.08, 1.69 0.01 ng/ml; mean se). No differences in follicular basal LH concentrations were observed between cycles. Luteal progesterone levels in the first poststress cycle were still low in three monkeys but not statistically different as a group, whereas there was no difference with control in poststress cycle 2. Length of the luteal phase remained unchanged at all times. Surge LH and FSH levels were unchanged (data not shown). [In one monkey (J47), the follicular phase was immediately prolonged to 63 d by the stress challenge, followed by ovulation; E2 levels remained low ( 20 –38 pg/ml) for a period of 50 d, and an E2 peak occurred 12 d later]. Effects of this 12-d stress in the follicular phase in an individual monkey are illustrated on Fig. 2. Stress challenge in the luteal phase When the stress was started in the luteal phase, all monkeys continued to show regular ovulatory menstrual cycles (Fig. 3). Integrated luteal progesterone during the stress period was significantly decreased by 30.9%. Integrated luteal LH values were also significantly decreased, whereas inhibin A values were 238.5 36.9 pg/ml in control cycle 2 and 173.7 42.3 in the stress cycle (NS; P 0.2908). Integrated luteal progesterone and LH levels during the first poststress cycle remained significantly lower. Luteal phase length remained unchanged at all times. Follicular phase length and E2 peaks (poststress cycle 1, 186.0 28.3; poststress cycle 2, 233.3 30.9), as well as baseline and peak gonadotropin values, were not significantly different from controls. Effects of a 12-d stress challenge in the luteal phase in an individual monkey are illustrated on Fig. 4. Cortisol response Because data from stress initiated in the follicular and luteal phase were similar, they were pooled. Baseline cortisol concentrations in the control cycles were 20.3 1.8 (morning) and 16.4 1.8 (afternoon) g/dl. Within 3 h of the initiation of the stress challenge, there was a rapid increase in cortisol levels to 56.3 4.7 (P 0.05 vs. afternoon control) (Fig. 5). Cortisol levels decreased by d 2 of the stress to 37.5 3.1 and continued to decrease thereafter. However, they remained FIG. 2. E2 (line with triangles) and progesterone (shaded area) profiles in monkey J23 throughout two control cycles, the stress cycle during which a 12-d stress challenge (white box) was applied in the follicular phase, and the two poststress cycles. Numbers in the shaded area are integrated luteal progesterone values. Note that luteal progesterone secretion was lower in the stress cycle and that follicular phase length was prolonged in the first poststress cycle. (Although peak E2 appears decreased after the stress, overall mean E2 peaks were not statistically different as a group.) significantly higher than morning controls over the entire challenge period (25.9 1.5 on d 11). Cortisol was still significantly elevated 2 wk poststress (25.6 1.1) but was similar to controls at 4 wk poststress 21.0 1.4. Mean body weight in 9 of 11 animals (including J47) was not significantly affected by the short-term stress: 7.2 0.6 kg (control), 7.0 0.6 kg (end of stress period), and 7.4 0.6 kg (1 month later). Two monkeys, however, had a 10 –13% decrease in body weight at the end of the 12-d stress period, but one of these had recovered 1 month later. Discussion This study was designed to develop a model of stress challenge that includes a significant psychogenic aspect in the nonhuman primate to allow a prospective evaluation of the initial effects of this stress on the menstrual cycle. The stress challenge had several components that included the psychological response to both a tethering system and a simultaneous move to an unfamiliar environment. It also included the response to the short surgical procedures required to install and disconnect the tethering system. The data demonstrate that this 12-d stress challenge, whether initiated in the follicular or luteal phase of the menstrual cycle, can rapidly induce menstrual cycle dysfunction, mainly in the form of a significant decrease in luteal function reminiscent of the inadequate luteal phase syndrome. Significantly, the period of disruption of the cycle is longer than that of the stress itself. An important result of the model is to highlight the first clinical sign of cyclic dysfunction in an otherwise normal ovulatory primate subjected to this stress. This occurs in the form of a significant reduction in integrated progesterone concentrations during the luteal phase, compared with control luteal phases in the same individual monkeys. Of clinical relevance is the observation that the calculated percentage of progesterone decrease from controls in these animals (from 30.9 to 51.6%) is similar or greater than that calculated in a clinical study of patients with the diagnosis of luteal phase deficiency (9). These patients presented with a complaint of infertility or recurrent abortion and had an endometrial biopsy that was more than 2 d out of phase. Whether the observed decrease in progesterone in our animals would similarly affect fertility remains to be determined. An equivalent decrease in luteal function was also reported in monkeys during strenuous exercise training (10). Other investigators have also noted a prevalence of luteal defect in cycling women as a result of exercise or eating disorders (11–15). The corpus luteum also secretes inhibin A (16, 17), and our data suggest that luteal inhibin A levels during the stress cycle are probably also decreased. That they are not significantly so probably is due to large individual variations. Overall, these data suggest that secretory inadequacy of the corpus luteum may represent the first clinical stage in the damage that stress and other conditions can inflict upon the normal menstrual cycle. The mechanism whereby this stress challenge induces a luteal defect in the monkey remains to be completely investigated. So far, data show that this reduction in overall luteal progesterone secretion does not reflect a shorter luteal phase, indicating that premature luteolysis does not occur. A similar observation was made in monkeys during exercise (10). Rather, the progesterone and inhibin A decrease can most probably be attributed to the observed lower luteal LH concentrations, because it is known that normal luteal function requires the presence of adequate gonadotropin levels (18 –22). Of particular interest is the observation that this short-term stress challenge continues to produce detrimental influences on the menstrual cycle and HPA axis past the period during which it is exerted and in some instances into the subsequent cycle. For example, integrated progesterone concentrations are lower in the luteal phase that follows a 12-d psychogenic stress in the follicular phase and remain lower in the first cycle after a psychogenic stress initiated in the luteal phase. Data showing that integrated LH values during the poststress deficient luteal phases are still significantly decreased suggest that these poststress effects may be related to continuing disturbances in the neuroendocrine component of the reproductive axis. Measurements of cortisol indicate that HPA axis activity remains increased for up to 2 wk after termination of the stress. In the present stress model, it is unfortunately not possible to determine whether the reintroduction into the familiar environment or the short physical stress associated with the removal of the tethering system may have been construed as a new stress and played a role. In our view, it is unlikely that these could have been responsible for the continued HPA state of activation over a 2-wk period. In fact, our data are in line with reports by other investigators who have also reported persistence of elevated adrenal activity after stress termination (23–25). This inability to shut off the allostatic response, i.e. the activation of the complex adaptive pathways in response to stress, after the psychogenic stress is terminated, may be representative of a type of allostatic load and of its pathophysiological consequences (26). Such a failure to turn off the response has been reported to be a feature of age-related functional decline (27) and also in depressive illness (28). Delayed effects of the stress on the luteal phase in our animals cannot be explained by deficient FSH/LH ratios during the preceding lutealfollicular transition period, which have been reported to accompany the inadequate luteal phase syndrome (29, 30), because these were normal in our data. Results obtained with this stress model, including delayed effects, parallel observations in a previously described inflammatory stress model in the monkey (3, 4). Yet, there are important differences. First, the neuroendocrine response to the inflammatory stress in the follicular phase differs in that there is an increased tonic LH secretion, a phenomenon observed both in monkeys and women (3, 31, 32). This inappropriate release of LH in the follicular phase following endotoxin does not appear to be directly related to an estrogen feedback action, because this is also observed in ovariectomized or menopausal individuals under constant estrogen replacement (8, 32). This phenomenon is not observed in the present stress model. Second, the acute inflammatory stress appears to produce more prolonged immediate delays in the follicular phase. Differences between the two models may relate to the large amount of cytokines released in the inflammatory stress (33, 34). The present stress challenge (like most stress modalities) is of course a complex entity that may impact on several systems. At this early stage of our investigation, it is not possible to speculate on which behavioral or cognitive variables activate the neuroendocrine or other central pathways that disrupt the cycle. One behavioral variable frequently associated with stress is a disruption in eating behavior; weight loss can activate the processes that disrupt the function of the GnRH pulse generator (14, 15). We do not believe, however, that food deprivation had a major impact on our data because the 12-d stress did not result in a change in body weight in 9 of the 11 animals. In the other 2 monkeys, there was no correlation between a 10% loss in body weight and the severity of cyclic damage. Similarly, other investigators have reported that a transition from normal cyclicity to an amenorrheic state during strenuous exercise training in monkeys is not necessarily associated with weight loss, although low energy availability may play a causal role in this transition (10, 35). In conclusion, we have demonstrated significant disturbances in the menstrual cycle after a 12-d stress challenge with a significant psychogenic component in a nonhuman primate model. First, the data clearly indicate that the initial clinical response to this stress is inadequate luteal function, emphasizing the concept that, even though a short-term stress may be insufficient to arrest the menstrual cycle, it potentially may still have a significant impact on fertility. Second, they demonstrate that the impact of a stress confined to one stage of the menstrual cycle exceeds the duration of the stress stimulus significantly, because effects are still observed in the first poststress cycle. It should be pointed out that in the present study only animals with a normal luteal function (as judged by integrated progesterone levels) (3) in the control cycles were used, thereby probably minimizing the effects of the stress. Indeed, our previous data with the inflammatory stress model have suggested that greater cyclic damage is seen in individuals in which subnormal luteal function is already present in the control cycles (3). Overall, we expect that the short-term psychogenic and inflammatory stress nonhuman primate models developed in our laboratory, and the planned longer term models, will be useful in unraveling some of the neuroendocrine pathways activated by stress. Acknowledgments We thank the National Hormone and Pituitary Program (NIDDK) and Dr. A. F. Parlow (Harbor-University of California Los Angeles Medical Center, Torrance, CA) for providing the reagents for monkey LH and FSH assays. We also acknowledge the help of Alinda Barth and Nancy Cotui in performing the gonadotropin and steroid assays. Endocrinology 136 : 897 - 902 35. Williams NI , Helmreich DL , Parfitt DB , Caston-Balderrama A , Cameron JL 2001 Evidence for a causal role of low energy availability in the induction of menstrual cycle disturbances during strenuous exercise training . J Clin Endocrinol Metab 86 : 5184 - 5193


This is a preview of a remote PDF: https://academic.oup.com/jcem/article-pdf/87/5/2232/9158457/jcem2232.pdf

Xiao, Ennian, Xia-Zhang, Linna, Ferin, Michel. Inadequate Luteal Function Is the Initial Clinical Cyclic Defect in a 12-Day Stress Model that Includes a Psychogenic Component in the Rhesus Monkey, The Journal of Clinical Endocrinology & Metabolism, 2002, 2232-2237, DOI: 10.1210/jcem.87.5.8500