Short-Term Energy Deprivation Alters Activin A and Follistatin But Not Inhibin B Levels of Lean Healthy Women in a Leptin-Independent Manner

Abstract

Context:

Leptin is a potent modulator of the hypothalamic-pituitary-gonadal axis mediating the effect of energy deprivation on several hypothalamic-pituitary-peripheral axes. Activin A, inhibin B, and follistatin (FST) also regulate the hypothalamic-pituitary-gonadal axis in humans. It remains unknown whether energy deprivation affects these hormone levels in a leptin-dependent or -independent manner.

Objective:

We investigated 1) day-night variability patterns of activin, inhibin, and FST in the fed state, 2) whether their levels are affected by fasting, and 3) whether such an effect is mediated by leptin in physiological replacement or pharmacological doses.

Design:

We conducted two studies in healthy, eumenorrheic females, each comprising three separate admissions. In study 1, six women were maintained for 72 h 1) on isocaloric diet, 2) fasting while receiving placebo, or 3) fasting while receiving metreleptin in physiological replacement doses. In study 2, five women were administered physiological or pharmacological metreleptin doses (0.01, 0.1, or 0.3 mg/kg iv four times daily).

Results:

Neither activin A nor FST had a pulsatile or day-night variability pattern. Inhibin B levels were also nonpulsatile, but a trend toward a day-night pattern was noted. When compared with the fed state, inhibin B levels remained unchanged, whereas FST levels increased (P = 0.01) and activin A decreased (P = 0.01) in the fasting state. These changes were not corrected with metreleptin administered in replacement or pharmacological doses.

Conclusions:

Short-term energy deprivation alters levels of activin A and FST, but these effects are not mediated by leptin.

Activins, inhibins, and follistatin (FST) were discovered in the 1980s and found to regulate gonadotropin secretion from the anterior pituitary (1, 2). Activins and inhibins belong to the TGF-β family, whose members exert their effects on a variety of systems, including ovarian regulation, adult tissue homeostasis, bone formation, tissue remodeling, wound healing and repair, muscle development, and angiogenesis (3, 4). Activin primarily stimulates pituitary gonadotropin release, whereas inhibin binds and neutralizes activin by inhibiting ligand-receptor interactions and promoting activin internalization and degradation (5, 6). FST is a ubiquitous, monomeric, secreted glycoprotein that, despite its lack of structural homology with other TGF-β family members, interacts with several of them and is the main activin-binding protein. FST and inhibin down-regulate activin's stimulatory effect and act as potent modulators of the hypothalamic-pituitary-gonadal (HPG) axis (6).

Another powerful modulator of the HPG axis is leptin, the major signal of energy stores in the body. Leptin is a 167-amino-acid adipocyte-secreted protein coded by the leptin gene that was originally described in an obesity mouse model (7). Its role in human physiology has since been extensively studied, and leptin has been found to play a major role in the neuroendocrine regulation of energy homeostasis (8–10). During states of energy deprivation, leptin levels decrease rapidly and trigger the neuroendocrine response to starvation, including down-regulation of the HPG axis in humans with fasting-induced hypoleptinemia (8, 11–13). The latter translates into suppression of the overnight LH peak frequency in females (13) and males (12) and up to 40% decrease in testosterone levels in males (12). Conversely, leptin replacement results in resumption of LH pulsatility and ovulatory menstruation in leptin-deficient patients (14). Thus, in energy-insufficiency states, leptin acts as a signal of energy insufficiency to inhibit energy-requiring processes, such as reproduction, that are not crucial for survival (15, 16).

Despite the major roles that activin, inhibin, and FST play in ovarian function and reproductive physiology in humans, only limited data exist in terms of the secretion patterns of these hormones, and they apply mainly to pubertal girls and males (17, 18). The only study in young, healthy women measured circulating inhibin B over 6 h, which severely limits the interpretation of secretion pattern data (19). Moreover, no studies have been published on the effect of food deprivation on these hormone levels or on the potential role of leptin if such an effect does exist.

We designed two interventional studies to determine whether 1) the 24-h levels of these hormones display any pulsatile and/or day-night pattern in young, healthy females of reproductive age, 2) their levels are affected by short-term energy deprivation, 3) if such an effect exists, it is mediated by food deprivation-induced changes in circulating leptin, and 4) short-term energy deprivation and changes in these hormones alter reproductive hormone levels. Moreover, to fully assess the potential effect of leptin on these hormone levels in the fasting state, we further investigated whether meterleptin administration not only in physiological replacement but also in pharmacological doses would alter levels of these hormones in young, healthy, eumenorrheic women.

Materials and Methods

Study designs

Seventy-two-hour fed/fasting study

Following a previously described protocol (13), six healthy lean women [age 22.8 ± 3.4 yr; body mass index (BMI) = 21.7 ± 2.2 kg/m²] with regular menstrual cycles (length 26–32 d) and no history of oral contraceptive use for at least 6 months participated in a clinical study involving three separate 4-d-long admissions, separated by at least 8 wk, at the General Clinical Research Center (GCRC) of the Beth Israel Deaconess Medical Center. Starting at 0800 h on d 3, blood samples were drawn every 15 min for 24 h through an indwelling catheter. The timing of these frequent blood sampling admissions was scheduled within menstrual cycle d 6–11 and within 2 d of the menstrual cycle day on which subjects had participated in the other two studies.

For the baseline pulsatility study, subjects were admitted to the GCRC for 72 h and participated in a baseline study in the isocaloric fed state [20% of calories from breakfast (0800 h), 35% from lunch (1300 h), 35% from dinner (1800 h), and 10% from snack (2200 h)]. Once the initial analysis revealed a trend toward day-night secretion (see below), we enrolled an additional three healthy, lean, eumenorrheic subjects (total n = 9) and measured their inhibin B levels in the fed state.

The 72-h fasting studies were performed in a double-blinded, randomized fashion. Subjects were admitted to the GCRC for 72 h and participated in two fasting studies: 1) a 72-h fasting study with administration of meterleptin (provided by Amylin, Inc., San Diego, CA; previously known as r-metHuLeptin, provided by Amgen, Inc., Thousand Oaks, CA) (replacement dose of 0.08 mg/kg · d on d 1, followed by 0.2 mg/kg · d on d 2–3, in four equal doses given sc every 6 h starting at 0800 h on d 1), and 2) a 72-h fasting study with administration of placebo (same schedule, volume, and mode of administration as the meterleptin dose).

Dose-dependent leptin replacement

Five healthy lean women (age 20.4 ± 0.7 yr; BMI = 21.9 ± 0.7 kg/m²) with regular menstrual cycles (length 26–32 d) and no history of oral contraceptive use for at least 6 months participated in one clinical study, involving three admissions separated by at least 3–10 wk (as previously described) (20): 1) a 72-h fasting study with administration of meterleptin (dose of 0.01 mg/kg once daily at 0800 h), 2) a 72-h fasting study with administration of meterleptin (dose of 0.1 mg/kg once daily at 0800 h), and 3) a 72-h fasting study with administration of meterleptin (dose of 0.3 mg/kg once daily at 0800 h). In all three studies, starting at 0800 h on d 3, blood samples were drawn and available at time 0, +10 min, +30 min, and +1, +2, +3, +4, +5, +6, +8, +10, +12, and +18 h after meterleptin administration.

All studies were approved by the Institutional Review Board of the Beth Israel Deaconess Medical Center and investigator-initiated Investigational New Drug approval was obtained by CSM from the U.S. Food and Drug Administration for use of meterleptin.

Hormone assays

FST levels were measured using a commercially available ELISA (R&D Systems, Minneapolis, MN), with a sensitivity of 29 pg/ml, intraassay coefficients of variation (CV) of 2.0–2.7% and interassay CV of 7.1–9.2%. Inhibin B levels were measured using a commercially available Gen II ELISA (Beckman Coulter, Inc., Brea, CA), with a sensitivity of 2.6 pg/ml, intraassay CV of 2.2–3.8%, and interassay CV of 3.7–5.6%. Activin A levels were measured using a commercially available ELISA (R&D Systems, Minneapolis, MN), with a sensitivity of 3.67 pg/ml, intraassay CV of 4.2–4.4%, and interassay CV of 4.7–7.9%. FSH levels were measured with the automated Immulite 1000 chemiluminescence assay system, with a sensitivity of 0.1 mIU/ml, intraassay CV of 2.6%, and interassay CV of 5.8%. Free fatty acids (FFA), estradiol, insulin, total testosterone, and SHBG levels were measured as previously described (12, 13). Total leptin levels (nanograms per milliliter) were measured using RIA [Linco Research (St. Louis, MO), now Millipore (Billerica, MA)], with a sensitivity of 0.5 ng/ml, intraassay CV of 7%, and interassay CV of 18–20%. This assay detected both endogenous leptin and exogenously administered meterleptin. LH levels were measured using an automated immunoassay system (Immulite 1000; Siemens, Deerfield, IL), with a sensitivity of 0.1 mIU/ml, intraassay CV of 4.8–6.5%, and interassay CV of 7.2–26.0%. All samples were run in dilutions as necessary for values to fall in the linear part of the standard curve and in duplicate within the same assay to decrease interassay variability.

Pulse analysis

Pulse_XP software (UVA Pulse Analysis Software, Charlottesville, VA) was used to analyze pulsatile (episodic) and day-night (periodic) secretion. Episodic secretion was examined using the previously evaluated Cluster8 computerized pulse-analysis algorithm (21). This algorithm uses the assay's measurement error, as expressed in sd between duplicate measurements, to detect clusters of statistically significant peaks or nadirs compared with preceding and following values, using a pooled t statistic. For the FST data that had a temporal resolution of 15 min, we used an MxN cluster of 1 × 2 and a T-score of 1.5. For inhibin B and activin A data that had a temporal resolution of 30 min a MxN cluster of 1 × 1 and a T-score of 1.5 was used. The periodic component of the time series was evaluated by the cosine algorithm that fits a four-parameter cosine function with a fixed period of 24 h at the time series and allows for evaluation of the amplitude, the residual variability, and the phase agreement between subjects. We also used a nonlinear ordinal least-squares trigonometric regression model, with a fixed period at 1 d, applied to all subjects to evaluate potential periodicity, and the adjusted nonlinear coefficient of determination (R²) was calculated.

Statistical analysis

Data are presented as mean ± sem and statistical analysis used Stata version 11.1 (Stata Corp., College Station, TX). Data were examined for normality using probability-probability plots and the Shapiro-Wilkes test. For study 1, we analyzed our data using hierarchical, mixed-effects linear regression models. Hormone levels were modeled as a linear function of time and condition (fed, fasting with placebo, or fasting with meterleptin) using dummy encoding. Time and condition were the fixed effects of the model, and we included random intercepts at both the subject and condition level. For study 2, we compared hormone levels in the three conditions (0.01, 0.1, and 0.3 mg/kg meterleptin) by repeated-measurements analysis of covariance across the evaluated time points. Both time and condition were used as within-subject factors in the repeated-measurements analysis of covariance model, because the subjects participating in the three studies were the same, whereas the baseline hormone levels were used as a covariate to adjust for baseline fluctuations between the subjects. Post hoc tests were performed, with Tukey's honestly significant difference test for multiple comparisons.

Results

Day-night variation and pulsatility of FST, inhibin B, and activin A (Fig. 1)

Fig. 1.

Average 24-h (0800–0800 h) leptin (n = 6) (A), FST (n = 6) (B), activin A (n = 6) (C), and inhibin B (n = 9) (D) levels on d 3 of baseline fed state (nonlinear adjusted R² is displayed on top center of each panel; dashed lines represent 95% prediction band; solid lines represent 95% confidence bands).

The Cluster8 pulse detection algorithm failed to detect any statistically significant pulses in the FST, inhibin B, or activin A secretion patterns. Nonlinear ordinal least-squares regression revealed no day-night pattern for FST (adjusted R² = 4.80%) or activin A (adjusted R² = 0.11%). Analysis of inhibin B secretion revealed a trend toward a day-night pattern, which was confirmed by the analysis of three additional subjects (n = 9; adjusted R² = 12.64%), with higher levels occurring at midday (peak at 1545 h) and lower levels occurring in the early morning (nadir at 0630 h; peak-to-nadir level difference of 48.74%). However, when we tried to replicate these results in the fasting plus placebo and fasting plus meterleptin states, we failed to demonstrate a day-night variability pattern.

Effects of 72-h fasting on FFA and hormone levels (Figs. 2 and 3)

Fig. 2.

Average (n = 6) 24-h profile (0800–0800 h) of FST (A), activin A (B), and inhibin B (C) levels on d 3 of baseline fed state (blue), 72 h fasting with placebo (red), and 72 h fasting with replacement-dose meterleptin (green) (P values are shown next to brackets, lower P values shown in italics are adjusted for age and BMI).

Fig. 3.

Average (n = 6) 24-h profile (0800–0800 h) of leptin, LH (12, 13), FSH, FST, inhibin B, and activin A levels on d 3 of baseline fed state (left), 72 h fasting with placebo (center), and 72 h fasting with replacement-dose meterleptin (right) (blue arrows indicate meals; green arrows indicate iv placebo/leptin).

As previously described (12, 13), 72-h fasting led to a significant decrease of serum leptin values (2.8 ± 0.3 ng/ml) from the baseline fed state (16.7 ± 1.2 ng/ml), but these levels fully recovered to the physiological range upon meterleptin administration (28.8 ± 2.0 ng/ml, overall P = 0.002). Similarly, overnight LH peak frequency had a statistically significant decrease during 72-h fasting (7.3 ± 0.4 vs. 4.5 ± 1.1 peaks per 12 h, P < 0.05), with complete recovery after meterleptin replacement (6.4 ± 0.5 peaks per 12 h, P = 0.04). Furthermore, 72-h fasting significantly increased FFA levels (0.04 ± 0.0 vs. 0.9 ± 0.2 mEq/liter, P = 0.03), significantly decreased insulin levels (8.1 ± 1.1 vs. 1.2 ± 0.2 μIU/ml, P < 0.01), but did not alter estradiol, total testosterone, or SHBG levels (all P > 0.05). Meterleptin administration in the fasting state did not change significantly any of these hormone levels (all P > 0.05).

On d 3 of the 72-h fasting, FST levels increased significantly compared with levels in the fed state (3499.7 ± 706.8 vs. 1522.4 ± 316.7 pg/ml, P = 0.01). Leptin replacement during fasting resulted in FST levels (2785.1 ± 537.6 pg/ml) that were not statistically different from the fasting plus placebo (P = 0.39) or fed states (P = 0.10). Conversely, fasting significantly decreased activin A levels compared to the fed state (154.5 ± 36.9 vs. 206.5 ± 34.3 ng/ml, P = 0.01). However, in this case, treatment of fasting subjects with replacement-dose meterleptin did not cause a statistically significant change in activin A levels (154.0 ± 38.1 ng/ml) compared with fasting plus placebo (P = 0.98), and levels remained statistically higher compared with the fed state (P = 0.01). After short-term energy deprivation, inhibin B levels did not change compared with the fed state, with and without leptin replacement (69.5 ± 10.5 vs. 73.0 ± 7.7 vs. 62.0 ± 6.4 ng/ml, all P > 0.05), and neither did FSH levels (3.9 ± 0.6 vs. 4.2 ± 0.8 vs. 4.0 ± 0.2 mIU/ml, all P > 0.05). These findings remained unchanged after controlling for age and BMI (italicized P values in Fig. 2).

Effects of physiological and pharmacological leptin replacement on activin A, inhibin B, and FST levels

As previously described (20), on the third day of a 72-h fast, circulating levels of leptin were decreased from the baseline fed state. However, these levels were progressively restored with increasing doses of leptin replacement and eventually eliminated at the 0.3 mg/kg meterleptin dose.

After meterleptin administration, there was no effect of dose on any of the levels of FST (P = 0.84), activin A (P = 0.30), or inhibin B (P = 0.09), after adjusting for baseline values.

Discussion

We demonstrate for the first time that FST, inhibin B, and activin A are not secreted in a pulsatile manner in young, healthy females, nor do they display a day-night pattern of secretion. Only inhibin B exhibited a clinically nonsignificant trend toward day-night secretion with peak levels in midday in the first six subjects studied, and despite increasing the sample size by 50%, we were unable to demonstrate a significant day-night pattern. Another novel finding of this study is that during periods of short-term energy deprivation, activin A levels decrease significantly, whereas FST levels increase significantly compared with the fed state, but inhibin B and FSH levels are not altered. Our interventional studies involving physiological and pharmacological leptin replacement demonstrated that the underlying mechanism for these changes is not leptin dependent and thus remains to be elucidated.

These findings are consistent with the teleological theory of decreased reproductive capacity during hypocaloric states, when the limited energy stores are shifted toward functions that are essential for survival. They also suggest that, in addition to leptin, hormones such as activin A and FST may also play a role in down-regulating the HPG axis during states of energy deprivation, but the mechanisms through which they exert their effects warrant further investigation. It remains to be conclusively demonstrated by interventional studies involving administration of activin A or inhibitors of FST, when these reach a clinical stage of development, whether these changes may act in parallel and/or in concert with leptin to regulate the HPG axis in healthy, lean females.

Currently, very little is known about the secretion pattern of the hormones studied herein. In animal studies, rams exhibited a nonpulsatile secretion pattern for FST (22), and male rhesus monkeys were found to have no diurnal rhythm of inhibin secretion (23). In humans, the majority of data are derived from studies of healthy male volunteers in the fed state. The earliest study, by Carlsen et al. (18), measured inhibin B levels every 30 min for 24 h in 13 healthy male volunteers (21–36 yr old) and identified a diurnal pattern with peak values in the early morning and nadirs in the late afternoon and evening hours. Similar patterns of secretion were noted when measuring immunoreactive inhibin concentrations in 24- to 27-yr-old healthy males every 15 min (24) and every hour for 24 h (25) as well as after FSH or testosterone injections to 18- to 50-yr-old male volunteers (26, 27). In a study of seven pubertal girls that underwent serum sampling every 15 min for 24 h, FST levels were highest between 0500 and 1100 h, activin A decreased during the night, and inhibin B decreased during the day, reaching a nadir between 1700 and 2245 h (17). Finally, in the only study of young women, where midfollicular inhibin B levels are measured every 10 min for 6 h in 10 women with polycystic ovary syndrome vs. five controls, the pattern of peaks every 60–70 min seen in controls was abolished in polycystic ovary syndrome patients (19).

Our results thus represent the first study to examine the secretion pattern of all three hormones in healthy, regularly menstruating females, using samples drawn every 15 (for FST) to 30 (for inhibin B and activin A) minutes for 24 h. Interestingly, inhibin B was the only hormone where a nonsignificant tendency toward a day-night pattern was noted. In contrast to the previously published data in healthy men or pubertal girls described above, we demonstrate a statistically nonsignificant pattern of inhibin B peaks in midday and nadirs in the early hours of the morning. This variation in secretion patterns could be explained by a gender-specific mechanism in addition to the HPG axis maturation process that occurs during and after puberty, but we feel confident about our findings, especially because they were replicated with the additional three subjects.

Another important component of our interventional studies was examining the effect of energy deprivation on these hormones, a question that has also not been studied in the past. No studies exist comparing activin, inhibin, and FST levels in fed vs. fasting men or women, although such data would have clinical and therapeutic implications. Clinically, energy deprivation is manifested mainly in the syndromes of anorexia nervosa (AN) and/or hypothalamic amenorrhea (HA). In a study of 23 women with HA, inhibin levels did not differ compared with controls, but their activin A levels were significantly higher (28). Because these patients start to regain their body weight and leptin levels increase, so do inhibin B levels, ultimately leading to restoration of their menstrual function (29). These observational studies cannot prove causality and thus cannot prove with certainty whether the observed restoration of inhibin B levels was due to the increase in leptin levels with refeeding and/or to a third confounding factor that also changed during refeeding, thus driving hormonal and reproductive changes. Our interventional data built upon and expanded on these observations.

We examined directly the effects of short-term (72-h) energy deprivation on the hormones activin A, inhibin B, and FST to test our underlying hypothesis that inhibin B and FST levels would increase, whereas activin A levels would decrease in response to caloric restriction. We also examined whether such effects could be mediated by leptin by physiologically and pharmacologically replacing it. We demonstrated that only activin A and FST levels changed in response to fasting, whereas inhibin B levels remained unchanged. We speculate that the effects of caloric restriction on inhibin B, if any, occur after at least 4 d and were therefore not detected during our study. Whether this is indeed the case would remain to be investigated by long-term energy-deprivation studies.

In recent years, leptin has emerged as one of the primary signals of energy homeostasis with a key role in the neuroendocrine adaptation to both over- and undereating, acting as a metabolic gatekeeper for reproductive function (16). The primary site of leptin's action is on the GnRH neurons of the hypothalamus where it stimulates orexigenic and anorexigenic pathways that in turn adjust food intake and body weight (30–32). In addition to this central mechanism, evidence exists that peripheral tissues, namely the ovary, express not only leptin but also its receptor, therefore providing an additional mechanism for leptin's effect on folliculogenesis and ovulation (31, 33–35). During states of energy deprivation, leptin concentrations decrease much more dramatically and out of proportion to fat mass changes, compared with states of energy excess (12, 36). During caloric restriction, leptin levels fall and down-regulate not only the HPG axis but also the hypothalamic-pituitary-thyroid and -GH axes, ultimately leading to energy conservation for immediate survival (37). Similarly, in patients with AN, falling leptin levels correlate closely with decreasing LH levels and menstrual aberrations, but both levels normalize in response to refeeding and weight gain (15). Several studies of HA subjects who were treated with meterleptin have demonstrated improvement in LH pulse frequency as well as resumption of follicle growth, ovulation, menstrual cyclicity, and bone mineral density (14, 38, 39).

Based on these established data supporting leptin's key role in the neuroendocrine adaptation to starvation, we hypothesized that meterleptin administration to our subjects would lead to restoration of altered activin A and FST levels after short-term energy deprivation. Leptin replacement tended to restore FST values to those seen in the fed state but did not lead to complete return to fed baseline levels. In contrast, activin A levels were not at all affected by meterleptin administration. Furthermore, varying doses of meterleptin (from replacement to supraphysiological) also did not confer a statistical change in any of the hormone levels studied, despite changes in leptin values. The only existing evidence for a potential interplay between leptin (specifically its soluble receptor, sOB-R) and FST is derived from follicular fluid of infertile women where soluble leptin receptor correlated with FST and activin levels (40, 41), but association studies cannot prove causality. Furthermore, the two interventional studies of meterleptin replacement in HA patients also support a lack of change in inhibin B levels over the 3-month treatment period (14, 39). Our interventional study does not provide any support to the hypothesis that leptin accounts for starvation-induced changes in circulating levels of FST, activin A, and inhibin B. It is very possible that mechanisms other than leptin, which mainly regulates the HPG axis centrally, may also be important in regulating peripheral molecules that feed back to regulate gonadotropin secretion by the pituitary (41).

We also measured FSH levels every 15 min during each one of the three conditions and demonstrated that they remain unchanged despite short-term energy deprivation or leptin replacement. Studies of energy deprivation in animals have almost uniformly demonstrated a decrease in FSH levels after fasting (42–47) or altered FSH signaling (48). The data from men are not as conclusive, with some studies demonstrating a decrease in FSH levels after fasting (49–51) and others showing no change at all (52–54). Chan et al. (12) showed that 72-h fasting in lean, healthy males resulted in a significant decrease in leptin levels from 2.24 ± 0.77 to 0.27 ± 0.08 ng/ml, accompanied by a significant decrease of insulin and testosterone levels (55) as well as a significant increase in levels of FFA and SHBG. In contrast, 72-h fasting in lean, healthy females, which decreased circulating leptin levels from 14.7 ± 2.6 to 2.8 ± 0.3 ng/ml, demonstrated no change in any of these hormone levels, including FSH (13). This finding is further supported by other studies of short-term energy deprivation in lean reproductive-age women (56–58). We speculate that the changes in FSH, if any, in lean healthy women might occur after the 72 h of fasting that we investigated here. It is also possible that a more significant decrease in leptin levels, below 2.8 ng/ml, achieved herein could possibly be necessary for activation of leptin-dependent mechanisms. This remains to be studied further. Finally, it is possible that because activin A and FST act primarily in a paracrine/autocrine manner to exert their effects at the local tissue level (e.g. the ovary), these effects may be differentially regulated by fasting and/or leptin, as opposed to their central effect to alter gonadotropin levels.

Our study also has several strengths over some of the already published ones: 1) our day-night pattern analysis had a high temporal resolution of every 15–30 min over 24 h, 2) our study was fully standardized for meal timing, caloric content, and menstrual cycle days between and within subjects, and 3) we used state-of-the art assays, including a well-validated second-generation inhibin B assay (59). Future studies need to evaluate whether more long-term caloric deprivation leads to similar changes in healthy women and whether chronic caloric deprivation in the context of disease states, such as AN and HA, is also associated with similar changes. In addition, important information could be derived from studies elucidating the underlying mechanisms of these changes as well as their effect on menstrual regulation, fertility, and overall health.

In conclusion, our results demonstrate for the first time that levels of FST, activin A, and inhibin B do not display a significant day-night or pulsatile pattern in young healthy women in the fed state. Short-term energy deprivation in female subjects leads to a significant decrease in activin A and increase in FST levels. This effect could not be explained through food deprivation-induced changes in leptin levels, suggesting that alternative mechanisms do exist. Future studies could investigate how these hormones are affected by both short- and long-term energy restriction and evaluate additional outcome measures, such as menstrual patterns and fertility potential.