|Home | About | Journals | Submit | Contact Us | Français|
Psychosocial factors, particularly social stress, may compromise reproduction. However, some individuals may be more susceptible to socially induced infertility. The present study used group-housed, adult, ovariectomized rhesus monkeys to test the hypothesis that exposure to psychosocial stress, imposed by social subordination, would enhance estradiol (E2)-negative feedback inhibition of LH. Because polymorphisms in the gene encoding the serotonin transporter (SLC6A4) may contribute to individual differences in response to adverse environments, we determined whether subordinate females with the short-promoter-length allele (s-variant) would show greater suppression of LH. Subordinate females, particularly those with the s-variant SLC6A4 genotype, received significantly higher rates of noncontact aggression from more dominant cage mates and had consistently lower body weights. Serum LH was not influenced by social status in the absence of E2. In contrast, subordinate females were hypersensitive to E2-negative feedback inhibition of LH. Furthermore, serum LH in subordinate females with s-variant SLC6A4 genotype was maximally suppressed by Day 4 of treatment, whereas nadir concentrations were not reached until later in treatment in other females. Finally, pharmacological elevation of serum cortisol potentiated E2-negative feedback inhibition in all females. The current data suggest that infertility induced by psychosocial stressors may be mediated by hypersensitivity to E2-negative feedback and that polymorphisms in the SLC6A4 gene may contribute to differences in reproductive compromise in response to chronic stress.
The analysis of social status in macaque monkeys provides a model to understand how psychogenic stress compromises reproduction. When housed socially, rhesus and cynomolgus macaque groups are organized by a linear dominance hierarchy that functions to maintain group stability . Continuous harassment characteristic of subordination is associated with enlarged adrenal glands in cynomolgus macaque females , hypercortisolemia due to diminished glucocorticoid-negative feedback [3–5], and increased sensitivity to ACTH  in both rhesus and cynomolgus macaque females. Importantly, puberty is delayed in subordinate rhesus females [7–9], and the incidence of short luteal phase cycles and periods of anovulation increase in both species of macaques [2, 10, 11] in subordinate compared with dominant females. Consequently, the overall reproductive fitness of subordinate females is reduced relative to more dominant animals . Despite this association, the mechanism responsible for this reproductive dysfunction in subordinate rhesus macaques has not been established.
Although social subordination is associated with reproductive deficits ranging from the timing of puberty through adult fertility, some females are more compromised than others [2, 10, 11], suggesting that individual differences in stress responsivity determine the ultimate reproductive phenotype. Previous studies show that female rhesus monkeys that are most reactive to stress are characterized by diminished serotonergic (5HT) activity, evidenced by reduced response to the 5HT transporter (5HTT) inhibitor-5HT releaser fenfluramine , reduced expression for the genes encoding the 5HTT (SLC6A4) and monoamine oxidase A (MAOA) in the dorsal raphe , and reduced 5HT receptor expression in the paraventricular nucleus of hypothalamus and infundibulum , and greater corticotropin-releasing hormone (CRH) expression in limbic-hypothalamic regions . Diminished 5HT function is associated with traitlike disinhibition of behavior in response to social adversity, including increased impulsivity and aggression as well as poor regulation of affect [17, 18]. Thus, in addition to these differences in behavior that emerge in response to social challenges, reduced 5HT activity may predispose female reproductive compromise. Polymorphisms in the SLC6A4 gene that encodes the 5HTT produce differences in 5HT function [19, 20]. The short-promoter-length allele (s-variant) interacts with social adversity to increase the incidence of affective disorders in people [21–24] and increases emotionality [25, 26], stress hormone responsivity , and behavioral reactivity in rhesus monkeys . Although studies of the contribution of SLC6A4 polymorphisms to a reproductive phenotype have produced equivocal results [14, 29], it is possible that the SLC6A4 short-promoter-length genotype may predispose females to stress-induced reproductive compromise. Indeed, we have shown previously that subordinate female rhesus monkeys with the s-variant SLC6A4 genotype are more likely to have delayed puberty . Whether polymorphisms in the gene encoding SLC6A4 influence effects of social subordination on the regulation of reproduction in adult females is not known.
The present study was designed to test the hypothesis that social subordination would significantly attenuate luteinizing hormone (LH) secretion in adult female rhesus monkeys and that any such effect would be exacerbated by the presence of the s-variant in the SLC6A4 gene. Furthermore, because stress-induced changes in reproductive hormones may be dependent on the presence of estradiol (E2) [30–32], the study also tested the hypothesis that subordinate rhesus macaque females would be hypersensitive to E2-negative feedback inhibition of LH, a finding that would be consistent with observations in marmoset monkeys . Finally, an unexpected outcome of a primary pharmacologic manipulation (exposure to a CRH receptor antagonist [CRHA]) resulted in a paradoxical increase in cortisol release, allowing us to assess also the differential effect of relative hypercortisolemia on LH suppression in relation to both social status and SLC6A4 genotype.
Subjects were 37 ovariectomized adult female rhesus monkeys (Macaca mulatta) ages 10–13 yr that were housed in one of eight small social groups at the Yerkes National Primate Research Center (YNPRC) Field Station, with five groups having five members and three groups having four members. Groups were maintained in run-type enclosures that had an indoor and an outdoor area that measured 20 × 15 × 8 feet each. The artificial lighting of the indoor living area was maintained on a 12L:12D cycle, with lights on at 0600 h. Animals were fed a commercial low-fat, high-fiber monkey diet (Purina Mills, St. Louis, MO) ad libitum twice daily, supplemented daily with seasonal fruit and vegetables. The Emory University Institutional Animal Care and Use Committee, in accordance with the Animal Welfare Act and the U.S. Department of Health and Human Services “Guide for Care and Use of Laboratory Animals,” approved all procedures.
Animals were assigned to new social groups based on SLC6A4 gene polymorphisms and their social status within their large, natal breeding groups . Initially, 40 females of intermediate social status—20 with l/l and 20 with l/s or s/s (s-variant)—were identified. Females had been removed previously from their natal groups and recombined into groups of five with unfamiliar females having the same SLC6A4 genotype, yielding four social groups containing five l/l animals and four groups containing five s-variant individuals . However, three monkeys developed chronic diarrhea and were subsequently removed prior to the start of this study.
The imposition of social subordination in rhesus macaques is not accomplished through contact aggression, but rather through continual harassment and the threat of aggression [34–36], and thus is determined empirically based on the outcome of dyadic interactions in which a female clearly emits a submissive response to another animal . Subordinate animals reduce the probability of receiving contact aggression by emitting appropriate submission responses toward dominant animals [34, 35]. Using previously established conventions , females ranked 1 or 2 in their group were classified as dominant, and those ranked 3 or higher were considered subordinate. Because groups were homogenous for genotype, each social status position had representation of both SLC6A4 genotypes. The social status by genotype groupings were: 8 dominant l/l, 8 dominant s-variant, 9 subordinate l/l, and 12 subordinate s-variant. Groups had been formed for 18 mo prior to the start of the present study, and females were ovariectomized 6 mo prior to the formation of these groups.
All animals were studied in each of four treatment conditions during the breeding season, between the months of January and April. Gonadally intact female rhesus monkeys housed in this environment are still able to ovulate through April . Each condition lasted 5 days and was separated by a 2-wk, no-treatment washout period. Females living together within a group (n = 4 or 5) received the same treatment as their group mates at the same time. The order of the treatments was counterbalanced across the eight different groups of females. The four treatments were: 1) control, 2) E2 replacement, 3) CRHA, and 4) E2 plus CRHA. Saline was administered as a 0.25-ml s.c. injection at 0830 h every morning for 5 consecutive days as the control condition. Estradiol replacement was achieved by implanting E2-filled Silastic capsules (Dow Corning, Midland, MI) s.c. as described previously . Analysis of selected samples (Day 4 and Day 7) during E2-replacement vs. no-E2-replacement conditions indicated hormone replacement achieved mid-follicular phase concentrations (66.7 ± 2.1 pg/ml vs. <5.0 pg/ml). Capsules were implanted 3 days prior to the initiation of data collection and were removed immediately after the end of the phase. Empty capsules were not implanted as a control for the E2-replacement condition because historical data indicate empty capsules (and associated minor surgery) produce similar results with respect to LH in rhesus monkeys, as does the absence of capsules .
The CRHA used was the CRH type 1 receptor antagonist CP154526 (Pfizer, Groton, CT). This CRHA and its close analogue, antalarmin, were developed to penetrate the blood-brain barrier and antagonize peripheral and central CRH type 1 receptors as a possible treatment for stress-induced disorders [39, 40]. Based on the existing literature in monkeys using antalarmin [41–44], we chose to administer a dose of 10 mg/kg daily (s.c.; volumes ranging from 0.20 to 0.55 ml depending on the weight of each animal) for 5 consecutive days at 0830 h with the expectation that it would attenuate cortisol secretion. However, as described in Results and reported previously , this dose paradoxically stimulated cortisol secretion, allowing us to evaluate the impact of an increase in cortisol on LH secretion in female monkeys.
The social housing design precluded the use of indwelling catheters for the frequent sampling of cortisol and LH in peripheral blood. Consequently, we measured both cortisol and LH in single morning serum samples obtained 30 min after the saline or CRHA injection (i.e., at 0900 h) on Days 2–5 of each phase. A similar approach has been used to quantify the seasonal [37, 45] and developmental [8, 46] regulation of LH in socially housed rhesus monkeys. Because a single morning cortisol sample does not reliably discriminate dominant vs. subordinate females [28, 47, 48], the intent of the cortisol measurements was to assess treatment effects of E2 and the CRHA rather than consequences of social subordination. All subjects were habituated to being removed from their group for venipuncture. Samples were generally obtained within 10–15 min from entering the animal area to minimize arousal. This approach reliably results in lower baseline cortisol measures in capture-acclimated monkeys  and does not compromise parameters of reproduction . Body weights were obtained weekly thoughout the duration of the study. Finally, using an established ethogram, aggressive and submissive behavior for each female was recorded for 30 min at 5 h after the saline or CRHA injection on the 5 consecutive days of each phase to assess social interactions and verify psychogenic stress . Data were recorded using a Palm PDA and the “Hands Obs” program developed by the Center for Behavioral Neuroscience . Data were collected in the format of actor-behavior-recipient. Dominance status was determined by the outcome of unequivocal dyadic agonistic interactions . Interobserver reliability exceeded 90%.
All assays were done in the Biomarkers Core Laboratory at the YNPRC. Selected samples were assayed for E2 to verify Silastic capsule efficacy using a modification of a previously validated commercial assay (Siemens/DPC, Los Angeles, CA) . Using 200 μl of serum, the assay has a sensitivity of 5 pg/ml and intraassay and interassay coefficients of variation (CVs) of 5.2% and 11.1%, respectively. Serum levels of cortisol were determined by radioimmunoassay with a commercially available kit (Beckman-Coulter/DSL, Webster, TX) described previously for rhesus monkeys [5, 48, 53]. Using 25 μl, the assay has a range from 0.5 to 60 μg/dl, with interassay and intraassay CVs of 4.9% and 8.7%, respectively. Assaying monkey serum at 10 and 25 μl produces similar estimates of cortisol concentrations after dilution correction (20.85 μg/dl vs. 21.30 μg/dl), and adding exogenous cortisol to known monkey sample with an established cortisol level yields estimates of cortisol similar to what would be expected (13.83 μg/dl vs. 13.68 μg/dl). Serum concentrations of LH were measured by radioimmunoassay using reagents provided by the National Institute of Diabetes and Digestive and Kidney Diseases-National Hormone and Peptide Program (Harbor-UCLA Medical Center, Torrance, CA) that has been validated previously . The assay uses macaque LH as the standard and a polyclonal antibody directed against macaque LH. Using 100 μl of serum, the assay has a range from 0.2 to 10 ng/ml and interassay and intraassay CVs of 4.9% and 8.7%, respectively.
Data were summarized as mean ± SEM. The main effects of status (dominant vs. subordinate), SLC6A4 genotype (l/l vs. s-variant), E2 treatment (placebo vs. hormone), CRHA treatment (saline vs. drug), and time on treatment (Days 1 or 2 through 5), as well as their interactions on hormonal data were analyzed with ANOVA for repeated measures. Because the behavioral data were not normally distributed, these data were analyzed with nonparametric statistics using Mann-Whitney tests for two independent samples (e.g., dominant vs. subordinate, l/l vs. s-variant, or subordinate l/l vs. subordinate s-variant), the Kruskal-Wallis test for more than two independent samples (e.g., ranks 1–5 or each social status category at each genotype), the Wilcoxon test for two related samples (e.g., subordinates at two treatment conditions), and the Friedman test for more than two related samples (e.g., comparison of four treatment conditions). To support the categorization of monkeys ranked 1 and 2 in their groups as dominant and those ranked 3–5 as subordinate, analysis of agonistic behavior was also performed on a female's individual rank. Two-tailed tests with a P ≤ 0.05 were considered significant.
Table 1 shows hourly rates, collapsed across the four treatment conditions, for aggressive and submissive behaviors directed toward others as a function of individual dominance ranks and social status categories. Rates of aggression directed toward others varied significantly across the five dominance ranks (P < 0.01). When animals were categorized as dominant or subordinate, rates of aggression were significantly higher in dominant females (P < 0.01). Furthermore, females with an s-variant SLC6A4 genotype had significantly higher rates of aggressive behavior compared with females with an l/l genotype (P = 0.05). Indeed, rates of aggression directed toward others were significantly higher in dominant s-variant compared with dominant l/l females (P = 0.03). Rates of submissive behavior emitted by animals increased significantly with lower dominance ranks (P < 0.01). Thus, higher rates of submissive behavior were also evident in females considered subordinate vs. those categorized as dominant (P < 0.01). In addition, females with an s-variant genotype emitted significantly higher rates of submissive behavior (P = 0.01) and subordinates with the s-variant genotype elicited higher rates of submissive behavior compared with subordinate l/l females (P = 0.01). Rates of aggression received, reflecting the amount of harassment subordinate monkeys received from more dominant animals (Fig. 1), increased significantly with lower dominance ranks (P < 0.01), which were again reflected in higher rates for females classified as subordinate as opposed to dominant females (P < 0.01). In addition to significantly higher rates of harassment received by s-variant females (P = 0.02), subordinate s-variant females received more aggression and dominant females than l/l subordinates (P = 0.02; Fig. 1).
Hourly rates of aggression toward others (and its reciprocal, aggression received from others) were significantly increased by E2 (1.30 ± 0.28 vs. 2.96 ± 0.76; P < 0.01) but were unaffected by CRHA treatment (2.67 ± 0.74 vs. 1.59 ± 0.31; P = 0.25). Hourly rates of submissive behavior toward others were not affected by E2 (1.32 ± 0.35 vs. 1.95 ± 0.43; P = 0.39) but were significantly attenuated by CRHA (2.20 ± 0.52 vs. 1.07 ± 0.17; P < 0.01). However, the interaction of the CRHA did not significantly attenuate submissive behavior when combined with E2 (2.41 ± 0.96 vs. 1.19 ± 0.39; P = 0.39). The effects of status on agonistic behavior described above were not significantly influenced by the specific treatments (P > 0.05). Taken together, these data show that subordinate females are most frequently the recipients of increased aggression from more dominant animals, and this pattern of harassment is greatest in females with s-variant genotype. Furthermore, rates of aggression were increased by E2 but not CRHA treatment.
Table 2 lists body weights for subjects throughout the four treatment phases. Overall, dominant females were significantly heavier than subordinates (F1,33 = 9.71; P < 0.01). Although females with an l/l SLC6A4 genotype had higher baseline body weights than those with an s-variant genotype, differences were not significant (F1,33 = 3.45; P = 0.07). The social status difference in body weight was not significantly modified by genotype (F1,33 = 1.94; P = 0.18). Treatment with E2 decreased body weights in all females (F1,33 = 10.26; P < 0.01). Although this effect of E2 was not modified by status (F1,33 = 0.60; P = 0.44), females with the s-variant genotype lost significantly less weight during the E2 phases than did females with an l/l genotype (F1,33 = 21.74; P < 0.01). However, this interaction of E2 treatment with genotype was not further modified by status (F1,33 = 0.27; P > 0.61). Treatment with the CRHA also significantly decreased body weight (F1,33 = 6.98; P = 0.01), but this effect was not modified by status, genotype, or their interaction (P > 0.05). Finally, the reduction in body weight during the combined combination E2 plus CRHA treatment was not significantly different than that observed with E2 or CRHA alone (F1,33 = 0.01; P = 0.95).
As illustrated in Figure 2, E2 decreased (F1,33 = 14.90; P < 0.001) and CRHA increased (F1,33 = 12.55; P = 0.001) morning cortisol values in all subjects. Daily cortisol concentrations were consistently lower throughout the week of E2 treatment compared with placebo (F3,99 = 0.24; P = 0.87) but did increase progressively throughout the week during the CRHA treatment (F3,99 = 6.55; P < 0.01). The effect of E2 on cortisol concentrations was not significantly affected by social status (F1,33 = 2.08; P = 0.16), SLC6A4 genotype (F1,33 = 0.01; P = 0.96), or the interaction of status and genotype (F1,33 < 0.01; P = 0.99). Similarly, the CRHA-induced elevation in cortisol was unaffected by status (F3,99 = 0.96; P = 0.41), genotype (F3,99 = 2.22; P = 0.90), or their interaction (F3,99 = 1.23; P = 0.30).
Serum LH was significantly lower during E2 treatments compared with the non-E2 treatment conditions (3.94 ± 0.17 vs. 6.31 ± 0.28 ng/ml; F1,33 = 85.16; P < 0.01). Consequently, the effects of status, genotype, and the CRHA treatment were evaluated separately for the non-E2 and the E2 conditions. In the absence of E2, serum LH did not vary significantly by time between dominant and subordinate females of either genotype (F3,99 = 0.99; P = 0.73), so data are collapsed across treatment days (Fig. 3). As can be seen, there were no differences in morning LH between dominant and subordinate females (F1,33 = 0.01; P = 0.99), but serum LH was significantly higher in females with an s-variant compared with the l/l SLC6A4 genotype (F1,33 = 5.12; P = 0.03). Moreover, in the absence of E2, LH was significantly lower during CRHA treatment (F1,33 = 3.95; P = 0.057), but this was unaffected by days on treatment (F3,99 = 1.25; P = 0.30), social status (F1,33 = 2.14; P = 0.16), or genotype (F1,33 = 2.33; P = 0.15).
A different pattern emerged during E2 replacement, because morning LH was significantly lower in subordinate (3.51 ± 0.22 ng/ml) compared with dominant (4.38 ± 0.25 ng/ml; F1,33 = 6.62; P = 0.02) females. However, the response in LH to E2 treatment did not vary significantly by SLC6A4 genotype (l/l, 3.83 ± 0.24 ng/ml; s-variant, 4.05 ± 0.23 ng/ml; F1,33 = 0.47; P = 0.53) or a status-by-genotype interaction (F1,33 = 0.04; P = 0.84). Importantly, this main effect of status varied significantly by day of E2 treatment (F3,99 = 5.53; P < 0.01), because serum LH was significantly lower in subordinates during the early portion of E2 treatment (Days 4–5 after implantation of E2 capsule) but not the later portion (Days 6–7; Fig. 4).
Treatment with CRHA, which elevated cortisol levels in all females (Fig. 2), resulted in significantly lower morning LH when combined with E2 (3.72 ± 0.17 ng/ml) compared with the E2-only condition in all females (4.17 ± 0.22 ng/ml; F1,33 = 5.18; P = 0.03; Fig. 4). This effect of CRHA was consistent across the treatment period because there was not a significant CRHA-by-day interaction (F3,99 = 0.82; P = 0.49). Furthermore, the interaction of status, genotype, and treatment day was not significant (F3,99 = 0.32; P = 0.81). However, there was a significant status-by-genotype-by-CRHA treatment (E2 vs. E2 plus CRHA)-by-day interaction (F3,99 = 2.89; P = 0.04). Posthoc analyses were focused on how serum LH differed during days from E2 implantation in each of the four groups. Analyses indicated subordinate, s-variant females were maximally suppressed by Day 5 of treatment during the E2-only phase, whereas dominant, s-variant females were maximally suppressed by Day 6. The dominant and subordinate females with the l/l SLC6A4 genotype both showed a progressive and significant decline in serum LH from Days 4 through 7 of E2 only (Fig. 4, left; posthoc tests P < 0.05). In contrast, during the E2 plus CRHA phase, all females reached nadir concentrations of serum LH by Day 6 of treatment (Fig. 4, right; posthoc tests P < 0.05).
Consistent with observations in marmoset monkeys , the current study demonstrates that socially subordinate female rhesus monkeys are hypersensitive to the negative feedback inhibition of LH secretion by follicular phase levels of E2 compared with dominant animals. This suggests that the hypersensitivity to E2-negative feedback inhibition of LH likely accounts for the increased frequency of anovulation or short luteal phase ovulations known to be a characteristic of female macaques exposed to psychogenic stressors [2, 10, 11, 55, 56]. Furthermore, the present results emphasize that genetic polymorphisms should be evaluated in the analysis of stress-induced reproductive deficits, because this hypersensitivity to E2 inhibition was accelerated in subordinate females with the s-variant SLC6A4 genotype. Importantly, by the end of treatment, social status differences in serum LH were no longer evident, suggesting the system was maximally suppressed at that time by this dose of E2. Nevertheless, these observations add support to pervious reports that individuals with this genotype are more vulnerable to the adverse consequences of social subordination  or other types of psychosocial stress [20, 25, 57–59]. This finding should be expanded to a larger population analysis to confirm whether SLC6A4 polymorphisms potentiate stress-induced reproductive compromise via enhanced E2-negative feedback inhibition.
The importance of E2 in differentiating LH secretion in dominant and subordinate females was underscored by the lack of social status differences in serum LH during the placebo condition. These observations contrast with those from sheep, in which psychosocial stress can inhibit LH release in ovariectomized animals not replaced with E2 . Indeed, under the placebo conditions of the present study, females with the s-variant SLC6A4 genotype, regardless of dominance rank, had higher concentrations of serum LH. Assuming individuals with the short-promoter-length allele of the SLC6A4 gene may have reduced 5HT activity , these higher levels of serum LH during the nonestradiol treatment periods in s-variant females could be explained by reduced serotonergic inhibition, because 5HT inhibits LH release during hypoestrogenic conditions [61, 62]. The biological significance of these genotype differences in the open-loop secretion of LH is unclear at this time.
Although overall levels of LH were reduced significantly in subordinate compared with dominant females during E2 treatments, the differences were most pronounced during the initial days of E2 treatment, because serum LH was maximally suppressed in all females by Day 7 of E2 treatment. This hypersensitivity to E2-negative feedback inhibition is similar to that observed during the initial stages of puberty [63–65], during lactational infertility , and in the regulation of LH secretion in seasonally breeding animals [45, 67]. The inhibitory action of E2 is likely mediated through a number of neurochemical changes acting on hypothalamic gonadotropin-releasing hormone (GnRH) release as well as directly affecting pituitary gonadotropin secretion [68–70]. However, the broader question is how this inhibitory action of E2 on LH is enhanced or exacerbated in socially subordinate females. Because social subordination in macaques is considered a potent psychosocial stressor [71–73], typically characterized by a dysregulated limbic-hypothalamic-pituitary-adrenal (LHPA) axis [4, 74, 75], the most parsimonious explanation is that the increased suppression of LH in subordinate females is due to stress hormones acting synergistically with E2. However, because body weights were significantly lower in subordinate compared with dominant females, one cannot rule out the possibility that metabolic signals also synergize with stress hormones  to enhance E2-negative feedback inhibition of LH.
A large body of data supports the hypothesis that CRH and glucocorticoids inhibit LH release. Central administration of CRH inhibits the secretion of LH in female rats and monkeys [76–78], and the continuous expression of CRH from the central nucleus of the amygdala disrupts estrous cycles in female rats , because CRH may act directly on GnRH cell bodies that project to the median eminence [80, 81]. Furthermore, activation of pituitary glucocorticoid receptors inhibits LH pulsatility [78, 82, 83]. There are few studies showing the importance of E2 on the stress-induced inhibition of LH. Estradiol potentiates the inhibition of LH by CRH in rats  as well as the hypoglycemic stress-induced inhibition of LH [85–87] at doses that have little effect on LH in nonstressed animals . Furthermore, the reduction of LH pulse amplitude  and inhibition of GnRH release in ewes [32, 89] by the pharmacological elevation of serum cortisol is dependent on E2. Despite these observations that E2 enables or facilitates stress-induced inhibition of LH secretion, the mechanisms responsible for this synergistic action are not understood.
We found that midfollicular levels of E2 decreased morning cortisol in all animals, even during treatment with CRHA. The effect of E2 on the LHPA axis is complex. Estradiol increases activation of the LHPA axis  by decreasing glucocorticoid-negative feedback [91–93], increasing adrenal sensitivity to ACTH  as well as hypothalamic CRH content [95–97], and enhancing diurnal [98, 99] or morning cortisol secretion [100, 101]. In contrast, other studies show that E2 decreases basal or stress-induced activation of the LHPA axis [91, 102–105]. The aim of this study was not to understand how E2 regulates cortisol secretion and LHPA activity, because the design was chosen to balance both E2 and the CRH receptor analogue treatments. Nevertheless, the finding that E2 lowered serum cortisol underscores the need to better define under what circumstances E2 may stimulate or inhibit stress hormone secretion.
The hypersensitivity to E2-negative feedback inhibition of LH in subordinate females was associated with significantly higher rates of harassment from more dominant animals, a characteristic of macaque social status relations [35, 106]. Importantly, the current data indicate that subordinate, s-variant females received significantly more harassment from their more dominant s-variant cage mates, and these behaviors were associated with an earlier maximum suppression of LH during the E2 treatment phase. The higher rates of aggression and affiliation exhibited by dominant s-variant females suggest these animals initiate more social interactions. Dysfunction of the 5HT system is linked to increased incidences of aggressive behavior [107–109] because 5HT usually acts to inhibit aggression  and limit impulsivity . Previous studies indicate 5HT tone is lower in individuals with an s-variant genotype [29, 112, 113], and reduced central 5HT activity is associated with increased impulsivity and aggression [17, 111, 114–116] as well as hostility in humans [117, 118]. Although s-variant females may be more predisposed to be aggressive given the correct social circumstances (i.e., high dominance status), simply engaging in more social interactions could increase the likelihood that an agonistic episode will occur.
Although we observed a significant effect of social status on serum LH during the E2 treatment condition, the present study failed to show a difference in morning cortisol between dominant and subordinate females. As noted previously, a dysregulation of the LHPA axis, assessed by a dexamethasone suppression test, is a characteristic of socially subordinate macaques [2, 3, 5, 75]. The use of a single sample to show status differences in cortisol may yield variable results [47, 48, 119, 120] and, depending on when the group was formed [28, 47, 121] or the training history of the animals , suggesting the use of a single morning sample for cortisol analyses may provide limited power to detect these group differences. Furthermore, previous studies showing how exposure to a range of stressors induces ovulatory defects do not report serum cortisol [16, 55], but rather report that these reproductive deficits are associated with greater CRH expression in hypothalamic-limbic regions . Other parameters of LHPA status, including provocative tests and more frequent sampling, may provide better biomarkers of psychosocial stress exposure that link stress exposure to reproductive compromise.
As described in Materials and Methods, we chose a 10 mg/kg dose of the CRH type 1 receptor antagonist CP154526, an analogue of the widely used antalarmin, to test the hypothesis that antagonism of CRH type 1 receptors would normalize LHPA dysregulation in subordinates, and thereby improve LH secretion. However, we found the daily administration of this drug for 5 consecutive days increased serum cortisol in all females and potentiated E2-negative feedback suppression of circulating LH in all females but the subordinate s-variant females. Because serum LH in these subordinate females was the lowest during the E2-only treatment condition compared with other females, it is possible that LH could not be further reduced by the effect of the drug. Previous studies show that the effects of CP154526 on corticosteroid release and behavior are inconsistent, because both low and high doses of the antagonist have elicited increases and decreases in stress hormone release and anxietylike behavior in both rodents [122, 123] and nonhuman primates [41, 43, 44]. Specifically, in adult male monkeys, a low dose of antalarmin (<3.2 mg/kg) antagonizes CRH-induced increases in ACTH, whereas a high dose of 10 mg/kg stimulates both cortisol and ACTH in adult male monkeys . Although the mechanism of this effect is unknown, our results corroborate these data. Finally, the administration of CP154526 alone decreased serum LH, but the effect was marginally significant (P = 0.057). Although the effect of the drug in combination with E2 could be attributed to the small but statistically significant elevation in serum cortisol, one cannot rule out the possibility that the drug itself is acting on hypothalamic or pituitary targets to directly reduce LH secretion.
In summary, our data suggest that the reproductive compromise characteristic of subordinate females is due to a hypersensitivity to the negative feedback inhibition of E2 on circulating LH. Furthermore, serum LH was maximally suppressed sooner during E2 treatment in those subordinates with the s-variant SLC6A4 genotype, suggesting they may have increased susceptibility to this hypersensitivity. This reproductive phenotype exhibited by subordinate females was associated with greater harassment by more dominant animals. The importance of cortisol as a signal on the hypothalamic-pituitary axis to inhibit LH must be further evaluated because the paradoxical increase in serum cortisol by CRHA potentiated this E2 inhibitory activity in all females, obscuring social status differences in LH secretion as the treatment progressed. Although our data indicate that the SLC6A4 genotype should be considered when evaluating risk factors for stress-induced infertility in women, these results must be considered preliminary, because much larger cohorts of either animals or women are needed to establish a link between this gene variant and susceptibility to stress-induced anovulation.
The study was conducted with the invaluable assistance of Jennifer Whitley, Marta Checchi, Jeff Fisher, and Dr. Jackie Hoffman. The CP154526 was a provided by Pfizer (Groton CT). We thank the Assay Services Unit at the Wisconsin NPRC for supplying the iodinated LH for the assays. The YNPRC is fully accredited by the Association for Assessment and Accreditation of Laboratory Animal Care, International.
1Supported by National Institutes of Health grants HD46501 and RR00165.