Beyond weight loss: Disentangling the direct effects of semaglutide vs equivalent calorie restriction on reproductive systems of male and female rats

Abstract

Background

Semaglutide (SEMA), a glucagon-like peptide-1 (GLP-1) analogue approved for treatment of obesity, is now one of the most used anti-obesity pharmacotherapeutic agents worldwide. The effects of SEMA on body weight, appetite, and adiposity are well established. Moreover, sex differences in these effects are clearly emerging. Metabolism and reproduction are intimately intertwined; yet, whether and how these blockbuster drugs affect the reproductive system of males and females is poorly understood. To address this gap, we investigated the impact of chronic SEMA treatment on gonads and circulating reproductive hormones in diet-induced obese male and female rats.

Methods and Results

GLP-1 receptor is expressed in ovaries and testes, potentially allowing a direct effect of SEMA. Testicular expression of this receptor was 4.7-fold higher than in ovaries, supporting a potential male bias for potency of the drug effect. To disambiguate the direct effect of the drug from potential downstream effects of weight loss, and ensuing improvements in metabolism produced by the drug, we also evaluated pair-fed (PF) controls. After 4 weeks of treatment, PF and SEMA males displayed improvements in reproductive measures like sperm motility and mucus penetration parameters. Morphological analysis of male and female gonads largely suggested increased fertility. In males, changes in germinal epithelium and seminiferous tubules were detected. In females, increased folliculogenesis was identified in both SEMA and PF rats. Structural changes in the gonads were accompanied by changes in circulating gonadal hormones in a sex-specific manner. In males, SEMA attenuated the weight-loss-induced reduction in all androgens measured. In females, levels of progesterone, pregnenolone, and estradiol were reduced in a treatment specific manner. Pituitary hormones were also affected in both sexes.

Conclusions

Altogether, this study highlights broad SEMA-specific, as well as weight loss induced but rescued by SEMA, effects at transcriptional, functional, and systemic levels on the reproductive systems of males and females.

Highlights

• •Morphological analysis of gonads suggests increased female fertility but also increased ovarian cysts.
• •In females, levels of progesterone and estradiol were potently reduced by SEMA.
• •SEMA males show improvements in reproductive measures like sperm motility and mucus penetration.
• •In males SEMA attenuated the weight-loss-induced reduction in all androgens measured.
• •Surprisingly, testes are more sensitive to GLP-1 than ovaries given much higher GLP-1R expression.

1. Introduction

Obesity affects a significant fraction of the world population, with some states or countries reporting overwhelming majority of the population as obese or overweight. Effective treatments were not readily available until the recent approval of glucagon-like peptide-1 (GLP-1) -based pharmacotherapeutics for weight loss [1]. Since then, the use of GLP-1-based pharmacotherapies for the treatment of obesity has skyrocketed. In addition to or because of the effective weight loss, these medications affect many different organs in the body [2,3]. Many of these effects, including the degree of weight loss achieved, show significant sex differences [[4], [5], [6]]. Moreover, recent analysis of online patient communities identified a range of self-reported reproductive and endocrine side effects of GLP-1 based therapies including menstrual irregularities, reduced libido, and hot-flashes that have not been systematically captured in clinical trials [7]. Some of these reports contrast with recent data indicating that in Polyendocrine Metabolic Ovarian Syndrome (PMOS; formerly Polycystic Ovary Syndrome) patients semaglutide-associated weight loss normalizes menstrual cycles in most patients [8], leaving this issue unresolved. While it has been established that sex and sex hormones affect the action of GLP-1 analogues, whether these medications affect the levels of sex hormones and reproductive organs is less well understood, especially in a chronic and clinically relevant drug application setting.

GLP-1 is an endogenous incretin hormone secreted by L-cells of the distal small intestine in response to nutrient ingestion or made in the brain, primarily by nucleus of the solitary tract neurons [9]. GLP-1 affects weight loss by reducing food intake largely via central satiety pathways [10]. GLP-1 receptors (GLP-1R) are present in reproduction-relevant brain areas including the paraventricular nucleus of the hypothalamus [11] but may also be found in the pituitary and gonads, allowing the drug potential direct influence at every level of the reproductive axes. While the primary actions of this endogenous hormone as well as the synthetic analogues are on metabolism, metabolism and reproduction are intimately intertwined both neuroanatomically and functionally [12]. Metabolic disturbances including diabetes and obesity have detrimental effects on reproductive health and fertility. Growing evidence implicates excess adiposity as a disruptor of gonadal physiology, steroidogenesis, and fertility outcomes in both males and females [[13], [14], [15]].

In females, the consequences of obesity on reproductive function are particularly well characterized at the level of the ovary and its cellular components. Excess adiposity, a hallmark of obesity, promotes a state of chronic hyperinsulinemia and peripheral hyperandrogenism that directly compromises the follicular environment, impairing granulosa cell function, folliculogenesis, and oocyte development [15,16]. Obesity in women with PMOS exacerbates metabolic disturbances, including insulin resistance, hyperinsulinemia, hyperandrogenemia, and abnormal luteinizing hormone (LH) to follicle-stimulating hormone (FSH) ratios, resulting in theca cell hypersecretion of androgens, inhibition of dominant follicle maturation, and chronic anovulation [17]. At the level of the ovary, insulin resistance impairs the normal responsiveness of granulosa cells to FSH signaling, reducing aromatase activity and estradiol production, and increasing the proportion of atretic follicles relative to healthy, developing ones [15]. Whether and which of these changes are corrected by chronic use of GLP-1R analogues, and to what extent these conceivable changes are mediated by weight loss, remains poorly understood.

In male rodents and humans, obesity has been directly linked to impaired testicular steroidogenesis and spermatogenesis, mediated in part through elevated scrotal temperature, increased intratesticular estrogen from adipose aromatase activity, lipid accumulation within Sertoli cells, and oxidative damage to spermatogenic cell lines [18,19]. The negative impact of obesity and diabetes on testicular steroidogenesis and spermatogenesis is well established, with impairment of glucose homeostasis leading to depletion of glucose transporter proteins in spermatozoa, decreased intracellular glucose availability, impaired sperm metabolism and ATP production, and consequently reduced sperm motility [18,20,21]. These gonadal-level perturbations establish obesity as a primary driver of reproductive dysfunction through mechanisms that are both systemic and at the organ level. The potential of weight loss induced by GLP-1 analogues interacting with or correcting these pathologies is not well characterized.

Considering the rapidly expanding clinical use of the GLP-1 analogue, semaglutide, among reproductive-age populations, we sought to determine the effect of this pharmacological agent on ovarian and testicular structure, pituitary and gonadal hormones, steroid-synthesizing enzymes, and related endpoints with a clinically relevant drug application route (subcutaneous) and protocol (chronic administration with ramp-up in dosing) in male and female obese rats. Since an important interpretive challenge is determining whether reproductive changes observed with GLP-1R analogues reflect direct effects of the drug or secondary benefits from weight loss, here we utilized a pair-feeding design to disambiguate the direct influence of semaglutide compared to metabolic benefits derived from the reduced food intake and weight loss produced by the drug.

2. Methods

2.1. Animals

7-week-old male and female rats (Charles River, Wilmington MA), following 1-week of habituation, were individually housed and were offered ad libitum access to a high-fat diet (HFD; 60% kcal from fat diet, Research Diets D12462). Rats were provided HFD for 6 weeks prior to treatment to induce obesity and during the treatment phase were provided standard chow (PicoLab Rodent Diet 5053) in addition to HFD. This protocol leads to the development of obesity, excess caloric intake, and body weight gain compared to age-matched controls fed only the standard chow diet [22]. Furthermore, unpublished laboratory data indicate that the 6-week HFD exposure increases visceral (gonadal) fat in males and females by ∼100%. It is commonly accepted that the fat accumulation that occurs as a consequence of feeding the animals a diet high in fat is sufficient to model obesity in humans. All procedures were approved and adhered to ethical guidelines of The Pennsylvania State University Institutional Animal Care and Use Committee.

2.2. Drugs

GLP-1 agonist, semaglutide (TLC Standards, PS-189002), was dissolved in Milli-Q water (Millipore Sigma) to the concentrations of 7 μg/mL, 15 μg/mL, 30 μg/mL, 50 μg/mL, and 70 μg/mL and stored at 4 °C. To mimic clinical protocol, a dose-escalation phase was utilized in the first week, in which every 2 days the concentration of daily injections of semaglutide was increased, beginning at 7 μg until 70 μg; then 70 μg was maintained for the remaining 3 weeks of the experiment. This dose-escalation period allows for the gradual introduction of the drug to minimize negative side effects such as nausea and gastrointestinal distress and improve medication adherence in humans [23,24]. Control and pair-fed animals were injected with vehicle solutions. All injection volumes were calculated based on daily body weight (1 mL/kg) and delivered at the same time each day subcutaneously.

2.3. Pair-feeding paradigm

Rats were assigned to one of three groups balanced for food intake and body weight as follows: control, SEMA-treated (SEMA), or pair-fed (PF) as described in Byun et al., 2025 [24]. Food and body weights were measured daily at the same time. Control and SEMA rats were allowed ad libitum access to chow and HFD. Each SEMA rat was weight-matched to a PF rat of the same sex. On the first day of injections (Day 0), SEMA-assigned animals were given their first dose as previously described; the following day (Day 1), PF rats were provided with the amounts of HFD and standard chow their SEMA-matched counterpart had consumed within the 24 h following injections, provided in one serving. While PF rats initially consumed all food within a few hours, starting the second week, they distributed their intake more evenly throughout the 24 h. The PF groups allow for the separation of semaglutide-specific effects on reproductive systems beyond improvement in metabolic functions secondary to weight loss.

2.4. Estrous cycle sampling

Samples were collected from female rats using methods previously described, at the same time each day, for 12 consecutive days to capture 2–3 full cycles [25,26]. In short, vaginal lavage was performed with a pipette and 30 μL of sterile saline to wash cells from the vaginal walls. The saline wash was then dispensed onto slides, stained, and visually analyzed under a microscope. Cycle phase was determined by a trained blind observer. For tissue collection, rats were equally distributed across estrous cycle phases in each treatment group, to not bias the results of the treatments by the estrous phase.

2.5. Tissue collection

After 4 weeks of treatment, rats were euthanized via decapitation after light anesthesia with isoflurane. Immediately following decapitation, truncal blood was collected into heparinized tubes. Blood was stored on ice and centrifuged at 3,000×g for 10 min. The plasma was subsequently used for measuring sex steroids and pituitary hormones. One of the gonads from each animal was collected and flash frozen to preserve RNA, while the other gonad was preserved in media for sperm analysis and histology. All tissues were stored at −80 °C until further processing. Females were freely cycling throughout the experiment and at the time of sacrifice. Distribution of estrous phases during collection did not differ between treatment groups (5–7 females in low estrogen signaling phases, metestrus-diestrus, and in high estrogen signaling phases, proestrus-estrus, for each treatment group at tissue collection).

2.6. Collection of sperm

Cauda epididymal rat sperm were collected from male rats after euthanasia. Each cauda epididymis was placed in 1000 μL of capacitation media containing 1X Earle's balanced salt solution, sodium pyruvate (0.23 mM), sodium bicarbonate (26.2 mM), 1X essential amino acids, 1X vitamins, EDTA (10 μM), and 3 mg/mL BSA. The cauda epididymis was cut, and sperm were allowed to swim into the capacitation media.

2.7. Computer-assisted sperm analysis (CASA)

A computer-assisted sperm analysis (Sperm Class Analyzer, Microptic, Hamilton Thorne, Spain) was employed to evaluate sperm motility and motion kinematics, following methods previously described by Palacin et al., 2013 [27]. One in ten dilutions of epididymal sperm were made in human tubal fluid media. In brief, a 10 μL sperm sample was placed on a pre-warmed counting slide. Using an Olympus BX41 microscope (Olympus Europe GmbH, Hamburg, Germany) with 10 × magnification in phase-contrast mode, sperm were observed. A minimum of 10 frames per treatment group, each containing at least 100 sperm, was captured. The Sperm Class Analyzer system was used to analyze digitized images. Sperm moving faster than 5 μm/s were classified as motile. Progressive sperm is determined by dividing the number of sperm with time-average velocity ≥25 μm/s and straightness of trajectory ≥80% by the total analyzed sperm count and multiplying by 100. Curvilinear velocity is the average velocity measured along the actual point-to-point path of the cell. Linearity of the curvilinear trajectory is calculated as straight line velocity/curvilinear velocity × 100. Straightness of trajectory is the linearity of the sperm's average path, calculated as straight line velocity/time-average velocity × 100. Amplitude of lateral head displacement represents the maximum lateral displacement of a sperm head from its spatial average trajectory (track width). Sperm mucus penetration was assessed by tracking sperm trajectories within a capillary tube filled with a viscous medium.

2.8. Hematoxylin and eosin (H&E) staining

Following tissue sectioning of testes and ovaries, the H&E staining was conducted according to previously established procedures and optimized for these tissues [28]. Slides were dried and baked in a 50 °C oven overnight. Slides were then deparaffinized and rehydrated through a graded series of solutions: two changes of xylene (10 min each), two changes of 100% ethanol (3 min each), followed by sequential 3-minute incubations in two changes each of 90%, 85%, and 70% ethanol, and finally distilled water for 30 s. Slides were stained in hematoxylin for 10 min, rinsed twice in distilled water (30 s each), and briefly differentiated in differentiation buffer for 10 s. After a 30-second rinse in distilled water, slides were stained in bluing solution for 2 min, followed by a 30-second distilled water rinse. Counterstaining was performed in eosin for 1 min, followed by two 30-second distilled water rinses and a 30-second PBS rinse. Slides were then dehydrated through an ascending ethanol series (70%, 85%, 90%, and 100%; 1–2 min each), cleared in xylene for 10 min, and allowed to air dry. Coverslips were applied using mounting media. Seminiferous tubule diameter was measured from acquired images using ImageJ software, with the minimum Feret's diameter (minFeret) used as the measure of tubule diameter.

2.9. Quantification of ovarian follicles

Follicle counts started from the first section, and ovarian tissue was examined every 10th section of serially sectioned tissue throughout the entire ovary. Classification of follicles is determined by size and cellular components identified [[29], [30], [31]]. Primordial follicles were classified as oocytes with a single layer of squamous granulosa cells. Primary follicles contained an oocyte surrounded by cuboidal granulosa cells. Oocytes partially surrounded by both squamous and cuboidal granulosa cells were also counted as primary follicles. Secondary follicles contain an oocyte with more than one layer of granulosa cells. Antral follicles feature oocytes surrounded by multiple layers of granulosa and theca cells, along with a fluid-filled antrum. The corpus luteum contains granulosa-lutein cells with the characteristic appearance of steroid-producing cells, characterized by pale cytoplasm, as well as smaller, more deeply stained theca-lutein cells. Atretic follicles were identified by a deeply stained zona pellucida that was detached from the granulosa cells, with or without fragmented oocyte and somatic cells, which may have pyknotic nuclei. Cysts were identified as fluid-filled, thin-walled structures lined with a single layer of granulosa cells and lack a cumulus–oocyte complex. To prevent double-counting, only follicles containing an oocyte nucleus were counted. All histology was analyzed by an observer blnded to the treatment.

2.10. Gene expression

RNA was extracted from the flash-frozen gonads using TRIZOL (Invitrogen) according to the manufacturer's protocol. The resulting RNA was measured in duplicate using the NanoDrop2000 (Thermo Scientific). 2000 ng RNA was used to create complementary DNA using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Quantitative PCR (qPCR) was performed with TaqMan Fast Advanced Master Mix (Applied Biosystems) on a QuantStudio 3 Real-Time PCR System (Applied Biosystems). TaqMan Gene Expression Probes for GLP-1R, steroidogenesis enzymes, and pituitary hormone receptors were used [probe ID's: GLP1-R (Rn00562406_m1), Star (Rn00580695_m1), Cyp11a1 (Rn00568733_m1), Cyp17a (Rn00562601_m1), Cyp19a (Rn00567222_m1), anti-mullerian hormone (Rn00563731_g1), follicle-stimulating hormone receptor (Rn01648507_m1), luteinizing hormone receptor (Rn00564309_m1)]. Relative fold expression was calculated using the 2^-ΔΔCt method developed by Livak and Schmittgen [32]. Beta- Actin (Rn00667869_m1) was used as a housekeeping gene. All samples were run in duplicate.

2.11. Plasma pituitary hormones

Circulating pituitary hormones LH and FSH were measured using solid-phase sandwich enzyme-linked immunosorbent assay (ELISA) purchased from ThermoFisher Scientific (LH assay: #EEL122; FSH assay: #EEL125). All standards and samples were run in duplicate. Final concentrations were calculated using standard curves according to manufacturer's protocol.

2.12. Plasma sex steroids

20 μL of a 100 μM mixture containing labeled 13C3-17β-estradiol, 13C3-progesterone, 13C3-testosterone, 13C2-pregnenolone, 13C3-DHEA, and 2.0 mL of ethanol were added to a tube containing Waters MassTrak Steroid Internal Standard Mix. The mixture was vortexed vigorously and then diluted 100-fold with Liquid Chromatography-Mass Spectrometry (LC-MS) grade water. This solution was used as the internal standard working solution in subsequent steps. Plasma samples, which had been thawed, were allowed to equilibrate to room temperature and then homogenized by vortexing. An aliquot of 100 μL of the internal standard working solution was added to each individually labeled tube, followed by the addition of 200 μL of plasma. The tubes were vortexed and incubated at room temperature for 30 min 100 μL of 0.5 M ammonium acetate (pH 5.5) was added to each tube, and the mixtures were incubated at room temperature for an additional 30 min, with occasional vortexing. For extraction, 500 μL of ethyl acetate was added to each tube, and the tubes were vortexed for 15 min before being centrifuged at 3700 rpm for 15 min at 5 °C. The organic (upper) layers were transferred to new tubes, which were then re-extracted with 0.2 M ammonium bicarbonate (pH 8.0). After the re-extraction, the organic layers were collected and transferred to new tubes containing 200 μL of ethanol. The aqueous (lower) layers from the ethyl acetate extraction were subjected to sequential extractions with 500 μL of a 70:30 (v/v) mixture of ethyl acetate and hexane, followed by 500 μL of butanol-1. The organic layers from these extractions were then re-extracted with 0.2 M ammonium bicarbonate solution, which had been used during the ethyl acetate extraction step. After the re-extraction, all organic layers were combined with the respective ethyl acetate fractions into the tubes containing ethanol. The extracted samples were dried under vacuum, then reconstituted in 25 μL of a water–ethanol mixture (80:20, v/v). The reconstituted samples were sonicated for 2 min and centrifuged at 12,000 rpm for 15 min at 5 °C. The clear supernatants were carefully transferred into separate labeled LC-MS vials and subjected to LC-MS analysis.

LC-MS analysis was performed using a Nexera 40 HPLC system (Shimadzu) coupled to a ZenoTOF 7600 mass spectrometer (Sciex) equipped with an IonDrive Turbo V ion source. Electrospray ionization was conducted in both positive and negative ion modes in a sequential manner. Chromatographic separation was achieved on a Waters Acquity Premier HSS T3 column (2.1 × 100 mm, 1.8 μm particle size) maintained at 40 °C. The mobile phase flow rate was set at 0.25 mL/min. The mobile phases consisted of LC-MS-grade water containing 0.2 mM ammonium fluoride (solvent A) and LC-MS-grade methanol (solvent B). The gradient elution program was as follows: 0.00–1.20 min, 10% B; 2.00 min, 65% B; 10.00 min, 75% B; 10.50–15.80 min, 95% B; and 16.00–20.00 min, 10% B. A 10 μL injection volume was used. Mass spectrometric data were acquired with a declustering potential of 70 V. The curtain gas pressure was set to 35 psi, nebulizer gas to 50 psi, heater gas to 70 psi, and the ion source temperature to 650 °C. The ion spray voltage was 5500 V in positive ion mode and −4500 V in negative ion mode. Data acquisition was performed in both MS¹ and MS² ] modes. The MS¹ scan range was m/z 100–1000, while the MS [2] scan range was m/z 50–400. Quantification of steroids was based on MS² data acquired in high-resolution multiple reaction monitoring (HRMRM) mode within the mass range of m/z 80–400, using an accumulation time of 10 ms.

2.13. Statistical analysis

Data in bar graphs are visualized as mean ± standard error of the mean (SEM). Statistical significance for single timepoint data was calculated using Student's t-test or one-way ANOVA followed by Holm–Šídák for post-hoc pairwise comparison where appropriate using GraphPad Prism 10. Longitudinal data were analyzed using a two-way ANOVA, with time and treatment as factors, followed by Holm–Šídák for post-hoc analysis where interactions were significant. The presence of cysts was assessed as a binary outcome (0 vs. 1+ cyst per rat), and the proportion of rats developing cysts was compared across treatment groups using a Fisher's exact test of independence. P-values <.05 are considered statistically significant.

3. Results

3.1. SEMA or food restriction reduce body weight in a sex-specific manner

There was a significant effect of treatment on cumulative calorie intake across time (2-way ANOVA: F (8, 128) = 11.48, p < 0.05; Figure 1B). Starting from the first week, SEMA treated females ate significantly fewer calories than controls (p < 0.05). The corresponding PF females ate the same number of calories as the SEMA group. Body weight was also significantly affected in females over time (2-way ANOVA: F (8, 128) = 18.49, p < 0.05; Figure 1C). Divergence between control and SEMA-treated females reached significance at week 3 (p < 0.05). Following 4 weeks of treatment, body weight was significantly affected in females (F (2, 32) = 3.442, p < 0.05; Figure 1D), and post-hoc analysis determined significantly reduced body weight in SEMA-treated females compared to controls (p < 0.05). No significant differences in final body weight were detected between SEMA and PF, nor controls and PF females. Calorie intake in males was also influenced by the treatment (2-way ANOVA: F (8, 116) = 16.82, p < 0.05, Figure 1E). By the second week of treatment, significant differences were detected between controls and SEMA males (p < 0.05). PF males consumed the same number of calories as their SEMA-treated counterparts. Body weight was also significantly altered by the treatment (2-way ANOVA: F (8, 116) = 33.32, p < 0.05; Figure 1F). In contrast to females, both SEMA and PF males weighed significantly less than controls starting at two weeks of treatment (p < 0.05). Final body weight in males differed across groups (F (2, 29) = 7.626, p < 0.05; Figure 1G). Post-hoc Holm–Šídák test highlighted significant decreases in body weight in SEMA and PF males compared to controls (p < 0.05 each). However, there were no differences in body weights between the SEMA-treated and pair-fed males. Gonadal weights in both males and females did not differ between treatments, indicating gonadal tissue was spared during weight loss (Figure 1H,I). This is likely an expected outcome given that moderate weight loss or obesity reversal is thought to preserve or improve organ function and does not shrink reproductive tissues or reduce their mass. Severe starvation or malnutrition, on the other hand, can cause a proportional drop in gonad size and weight due to extreme energy deficits. Lack of change in gonadal weights suggest that the weight loss associated with SEMA or equivalent food restriction is not severe enough to affect gonadal weights.

Figure 1.

SEMA induces body weight loss while preserving gonad weight. Timeline and design of the experiment (A). Weekly cumulative calorie intake (B) and body weights (C) of females during treatment. Final body weight of females (D). Weekly cumulative calorie intake (E) and body weights (F) of males during treatment. Body weight of males at the conclusion of experiment (G). Ovary weights (H). Testis weights (I). GLP-1R is expressed on ovaries and testis permitting direct effect of the drug on these tissues. Sex differences in GLP-1R expression in gonads (J). Relative expression of GLP-1R in ovaries (K) and testes (L). n=10-12 per sex and per treatment. B, C, E, and F: ∗∗ p ≤ .01, ∗∗∗p ≤.001, ∗∗∗∗p ≤.0001 indicate significant differences between Control and SEMA. ^$ p ≤ .05, ^$$ p ≤ .01 indicate significant differences between Control and PF. ∗ p ≤ .05, ∗∗ p ≤ .01, ∗∗∗∗p ≤ .0001. HFD: High-Fat Diet, SEMA: Semaglutide, PF: Pairfed, Glp-1r: Glucagon-like Peptide-1 receptor.

3.2. GLP-1R is expressed at higher levels in male than female gonadal tissue and is resilient to treatment

To determine whether GLP-1 analogues can have direct effects on the ovaries and testes, real-time quantitative PCR was used to assess the expression of GLP-1R in the ovaries and testes of rats. Our results demonstrate that the GLP-1R is expressed in the gonads of male and female rats (Figure 1J). Thus, this allows for the direct action of semaglutide on these reproductive organs. Interestingly, testes express 4-5-fold more GLP-1R than ovaries (Figure 1J), potentially indicating a route for more potent influence of GLP-1 and its analogues on male gonads. Additionally, the relative expression of GLP-1R in gonads remains constant even when a chronic GLP-1R agonist is present (Figure 1K,L).

3.3. SEMA or caloric restriction reduces select pituitary and steroid hormones in a sex-specific manner

To determine how circulating hormones released from the anterior pituitary gland and the gonads are influenced by chronic SEMA treatment and weight loss, ELISAs and LC-MS were employed to measure plasma hormone concentrations. Chronic SEMA injections did not alter FSH levels in either sex (Figure 2B,D). However, LH, another hormone secreted from the gonadotrophs of the anterior pituitary, which is important for stimulating steroidogenesis in theca cells and dominant follicles, tended to be reduced by SEMA-treatment in females compared to controls (F (2, 27) = 2.743, p < .09, Holm–Šídák: p < 0.06; Figure 2C). Even though this effect was not statistically significant in PF females, the levels were also not different from the SEMA group, suggesting caloric restriction alone has an intermediate influence on LH levels in females (Figure 2C). In males LH, which plays a key role in stimulating testosterone production, was robustly impacted (F (2, 28) = 6.539, p < 0.05). Concentration of LH was reduced by over 50% in PF rats compared to controls (p < 0.05; Figure 2E). SEMA treatment also decreased LH levels compared to controls (p < 0.05; Figure 2E). In addition to the pituitary hormone changes detected, gonadal hormone levels were also altered by the treatment. The first intermediate hormone in the steroid synthesis pathway after cholesterol, cleaved by the cholesterol side chain cleavage enzyme, is pregnenolone. Levels of pregnenolone were reduced, but exclusively in the PF females (F (2, 22) = 5.229, p < 0.05, Holm–Šídák: p < 0.05; Figure 2F). Interestingly, SEMA-treated females had comparable pregnenolone levels to controls (Figure 2F), indicating that SEMA is ameliorating the pregnenolone decline stemming from caloric restriction. In contrast, SEMA tended to decrease circulating pregnenolone in males (F (2, 25) = 2.526, p < 0.1, Holm–Šídák: p < 0.09; Figure 2L). Progesterone, a hormone synthesized from pregnenolone, was also robustly affected by treatment in females (F (2, 27) = 4.304, p < 0.05). But unlike pregnenolone results, both SEMA and caloric restriction decreased progesterone levels compared to controls (Figure 2G; p < 0.05 each). Progesterone levels were not influenced by either of the treatments in males (Figure 2M). Dehydroepiandrosterone (DHEA), also synthesized from pregnenolone, was unaffected in females (Figure 2H). In contrast, DHEA tended to be decreased by caloric restriction alone in males (Figure 2N; F (2, 25) = 3.250, p < 0.06, Holm–Šídák: p < 0.05). This reduction was rescued by SEMA. In line with the DHEA changes, testosterone was also decreased exclusively by caloric restriction in males (p < 0.05). This potentially indicates that SEMA ameliorates the testosterone changes associated with caloric restriction in males (Figure 2P). Androstenedione, an intermediate hormone produced from either progesterone or DHEA and a substrate for testosterone was affected by treatment in males (F (2, 26) = 4.322, p < 0.05; Figure 2O). PF males had significantly less circulating androstenedione compared to controls (p < 0.05), whereas levels were partially rescued by SEMA (p < 0.07; Figure 2O). Reduced testosterone is consistent with both reduced DHEA and androstenedione. These effects on DHEA, androstenedione, and testosterone did not extend to females (Figure 2H–J). Estradiol, synthesized from testosterone, remained consistent between treatments in males (Figure 2Q). However, caloric restriction decreased plasma estradiol in both SEMA and PF females (Figure 2K). Overall, we found sex- and treatment-specific effects of SEMA and caloric restriction on steroid hormones.

Figure 2.

SEMA and food restriction alter circulating pituitary and steroid hormones in a sex-specific manner. Schematic of hormone analysis along hypothalamic-pituitary-gonadal axis (A). Neither SEMA nor PF altered plasma concentrations of FSH in females (B). Levels of LH tended to be decreased in SEMA-treated females (C). Levels of FSH in males were not altered by SEMA or caloric restriction (D). LH was reduced in both SEMA and PF males (E). Pregnenolone levels in females were reduced by caloric restriction but not SEMA (F). Both SEMA and PF females had decreased levels of progesterone (G). Plasma DHEA, androstenedione, and testosterone concentrations were not altered by the treatment in females (H, I, & J). Caloric restriction reduced circulating estradiol levels in SEMA and PF females (K). Circulating pregnenolone tended to be decreased in SEMA males only (L). Plasma progesterone was not influenced by treatment in males (M). DHEA, androstenedione, and testosterone levels were decreased in the calorie restricted PF group, but this reduction was rescued by SEMA for these hormones (N, O, & P). Treatment did not affect estradiol levels in males (Q). n=10-12 per sex and per treatment. # p ≤ .09, ## p ≤ .07, ### p ≤ .06,∗ p ≤ .05, ∗∗ p ≤ .01. GnRH: Gonadotropin-releasing hormone, FSH: Follicle-stimulating hormone, LH: Luteinizing hormone, ELISA: Enzyme-Linked Immunosorbent Assay, LC-MS: Liquid Chromatography-Mass Spectrometry, SEMA: Semaglutide, PF: Pair-fed, DHEA: Dehydroepiandrosterone.

3.4. SEMA did not alter gonadal expression of gonadotropin receptors

Testes and ovaries of rats were analyzed for the expression of the gonadotropin receptors, as these could buffer or exacerbate the influence of changes in pituitary hormone levels on the reproductive axis. The gonadal expression of follicle-stimulating hormone receptor (FSHR) was not significantly influenced by treatments in females or males (Figure 3B,D). Although SEMA and PF reduced circulating LH, luteinizing hormone receptor (LHR) expression was not affected in the ovaries or testes (Figure 3C,E).

Figure 3.

SEMA and caloric restriction influence transcription of genes involved in steroidogenesis. Image of workflow for analysis of pituitary hormone receptors and enzymes involved in steroidogenesis (A). Expression of FSH and LH, pituitary hormone receptors is not significantly affected by treatment in ovarian tissue (B & C). Similarly, testicular Fshr and Lhr expression in males remained consistent between treatments (D & E). Ovarian Star expression is decreased in SEMA and PF groups (F). Other enzymes involved in steroidogenesis such as Cyp11a, Cyp17a, and Cyp19a are not affected by treatment (G, H, & I). Ovarian Amh expression is not influenced by SEMA or caloric restriction (J). In contrast to female results, testicular Star expression is not impacted by chronic SEMA or food restriction (K). Testicular expression of Cyp11a and Cyp17a remained unchanged (L & M). Cyp19a was increased in PF males compared to SEMA males (N). Testicular Amh expression did not significantly differ between treatment groups (O). n=10-12 per treatment, per sex. ∗ p ≤ .05. Fshr: Follicle-stimulating hormone receptor, Lhr: Luteinizing hormone receptor, SEMA: Semaglutide, PF: Pair-fed, DHEA: Dehydroepiandrosterone, Star: Steroidogenic Acute Regulatory protein, Cyp11a: Cytochrome P450 Family 11 Subfamily A, Cyp17a: Cytochrome P450 Family 17 Subfamily A, Cyp19a: Cytochrome P450 Family 19 Subfamily A, Amh: Anti-Müllerian Hormone, RT-PCR: Reverse Transcription Polymerase Chain Reaction.

3.5. Effect of SEMA on gonadal expression of hormone receptors and steroidogenesis enzymes

Since specific gonadal steroid levels were affected by the treatments, testes and ovaries of rats were analyzed for the expression of enzymes involved with steroidogenesis to further understand the potential effects of SEMA on reproductive function at the level of the gonad. We found that the expression of Steroidogenic Acute Regulatory (STAR), which regulates the transport of cholesterol from the outer to inner mitochondrial membranes, was uniquely affected in females (F (2, 32) = 4.734, p < 0.05). Further post-hoc analysis determined that both SEMA and PF females had decreased expression of STAR in the ovaries compared to controls (p < 0.05 each: Figure 3F). STAR mediates a critical rate-limiting step in the steroidogenesis pathway and can affect the downstream production of steroid hormones. Enzymes further downstream in steroidogenesis were not impacted by SEMA or food restriction in females (Figure 3 G, H, & I). In males, expression of early enzymes involved in steroidogenesis was not altered by any of the treatments (Figure 3K, L, & M). Males, however, had distinctive effects in expression of aromatase (CYP19a), the enzyme responsible for converting testosterone into estradiol: PF males had increased expression of CYP19a compared to SEMA-treated males (p < 0.05; Figure 3N), a change supporting the reduced testosterone in males.

The gene expression of Anti-Mullerian hormone (AMH), important for sexual differentiation in utero as well as an indicator of ovarian reserve in adult females, was not altered by the treatment in females (Figure 3J). Additionally, testicular AMH expression was also not affected by the treatment (Figure 3O).

3.6. GLP-1 analogue treatment and caloric restriction affect sperm and testes morphology

To determine whether GLP-1 analogue treatment influences the male reproductive function, computer-assisted sperm analysis (CASA) was utilized. Overall, sperm concentration was not significantly altered following treatment (Figure 4B). However, SEMA and caloric restriction increased total motile sperm compared to controls (F (2, 30) = 4.603, p < 0.05, post-hoc: p < 0.05 each; Figure 4C). Progressive motility, the ability of sperm to actively move, was also different between groups (F (2, 30) = 3.683, p < 0.05). Post-hoc analysis detected trends for increased motility in SEMA and PF males compared to controls (p < 0.09 for both; Figure 4D). Other characteristics of sperm movement, including curvilinear velocity, speed of the sperm heads, and linearity index, path of sperm progression, did not significantly differ between treatment groups (Figure 4E,F). Mucus penetration, i.e. the ability of sperm to move through mucus, a critical step in fertilization was affected by the treatment (F (2, 30) = 3.424, p < 0.05). Post-hoc analysis determined strong trends for increased mucus penetration in both SEMA and PF males compared to controls (p < 0.06 each; Figure 4G).

Figure 4.

SEMA and caloric restriction alter sperm characteristics and tubule structure. Schematic of testicular tissue processing and sperm analysis (A). Concentration of sperm harvested from the caudal epididymis was not affected by SEMA or food restriction (B). The number of total motile sperm cells was increased by both SEMA and equivalent caloric restriction (C). SEMA and caloric restriction tended to increase sperm progressive motility (D). Other characteristics including curvilinear velocity and linearity index were not significantly altered by the treatments (E & F). Sperm of SEMA and PF tended to have an increased ability to penetrate mucus compared to controls (G). Both SEMA and caloric restriction increased germinal epithelial thickness, with SEMA resulting in increased thickness also compared to PF rats (H). Seminiferous tubule diameter was uniquely decreased in SEMA treated males (I). Similarly, seminiferous tubule area was decreased specifically in the SEMA group (J). n=10-12 per treatment (B-G). n=50-55 samples per treatment (H-J). # p ≤ .09, ###p ≤ .06, ∗ p ≤ .05, ∗∗ p ≤ .01, ∗∗∗p ≤ .001, ∗∗∗∗p ≤ .0001. CASA: Computer-assisted sperm analysis, SEMA: Semaglutide, PF: Pair-fed.

Structural changes to testes were assessed using hematoxylin and eosin staining. Thickness of the epithelial wall surrounding the tubules was measured. This was significantly affected by SEMA treatment and caloric restriction (F (2, 157) = 33.90, p < 0.05). Thickness of the germinal epithelium was robustly increased in SEMA-treated males compared to controls and PF (p < 0.05 for both; Figure 4H). Calorie restriction alone, increased germinal epithelial thickness compared to ad libitum fed controls, but not to the level in the SEMA-treated males (p < 0.05; Figure 4H). Seminiferous tubule diameter was also affected by the treatment (F (2, 157) = 5.572 p < 0.05). There was a SEMA-specific reduction in tubule diameter (vs. controls p < 0.05, vs. PF p < 0.05; Figure 4I). Similar effects were found in seminiferous tubule area (F (2, 157) = 5.500, p < 0.05), where the tubule area of SEMA-treated rats decreased compared to control and PF males (p < 0.05 each; Figure 4J). These represent one of the few parameters where a change was unique to the SEMA group and not present in either of the control groups.

3.7. SEMA treatment or caloric restriction affect specific stages of folliculogenesis

To assess whether chronic SEMA treatment affects female reproductive function, one ovary from each female rat was collected for H&E stain and quantification of ovarian follicles. Early follicle development was impacted by treatment. Primordial follicle counts tended to differ due to treatment (F (2, 31) = 2.691, p < 0.09; Figure 5B). Further analysis determined SEMA-treated females tended to have more primordial follicles detected than controls (p < 0.07; Figure 5B). PF females had an intermediate phenotype between controls and SEMA groups. The number of primary follicles was influenced by SEMA as SEMA-treated females had significantly more primary follicles in their ovaries compared to controls (F (2, 31) = 5.993, p < 0.05: Holm–Šídák: p < 0.05; Figure 5C). Equivalent calorie-restricted PF females again were an intermediate between controls and SEMA females, however there were no statistically significant differences between either group. Secondary follicle count was affected (F (2, 33) = 4.054, p < 0.05; Figure 5D). Specifically, SEMA-treated females and PF females had significantly increased secondary follicle count compared to controls (p < 0.05 each). Antral follicle numbers were also increased, though this effect reached significance in PF females only (p < 0.05; Figure 5E). There were no changes in corpus luteum counts (Figure 5F). Lastly, indicators of reproductive dysfunction: atretic follicles and ovarian cysts were measured. Atretic follicles were not influenced by SEMA treatment or caloric restriction (Figure 5G). However, treatments significantly affected cyst development (Fisher's exact test, p < 0.05; Figure 5H). No cysts were detected in any of the control rats; 66% of SEMA-treated rats and 42% of calorie-restricted rats had 1 or more cysts detected. In contrast, total cyst counts per group did not differ significantly, suggesting treatment influenced whether cysts developed but not the overall burden in affected animals (Figure 5I).

Figure 5.

Late-stage folliculogenesis is increased by SEMA-treatment and equivalent calorie restriction. Workflow of processing ovaries and analysis of histological appearance of folliculogenesis (A). Primordial follicle counts tended to be higher in SEMA-treated females (B). The number of primary follicles was robustly increased due to SEMA treatment (C). Later stage follicles, secondary follicles, were increased in both SEMA and PF females (D). Similarly, preovulatory antral follicles were also increased, although this change was only significant for PF females (E). Corpus lutea counts were not different between groups (F). The number of atretic follicles was unaffected by treatment (G). Treatments significantly affected cyst development with cysts detected in only SEMA and PF groups as determined by Fisher’s exact test (H). This change did not reach statistical significance when the total numbers of cysts were compared (I). n=10-12 per treatment. ## p ≤ .07, ∗ p ≤ .05, ∗∗ p ≤ .01. SEMA: Semaglutide, PF: Pair-fed.

3.8. Effects of SEMA treatment on the estrous cycle

To investigate how daily SEMA injections influence the estrous cycle of a rat, vaginal cytology analysis was conducted on samples taken at the same time each day for 12 days. Overall, estrous cycle length did not change between groups (Figure 6B). However, when time spent in specific phases of the cycle was assessed, significant differences in specific phases were found. Proestrus was affected (F (2, 32) = 4.528, p < 0.05; Figure 6C), where SEMA-treated females spent less time in proestrus (p < 0.05). PF females also spent significantly fewer days in proestrus compared to controls (p < 0.05; Figure 6C). This indicates that this phase may be sensitive to weight loss. The reduction in proestrus was compensated for by increase in time spent in estrus phase: estrus phase was influenced by calorie restriction (F (2, 32) = 3.836, p < 0.05) and PF females spent more time in estrus (p < 0.05; Figure 6D), with SEMA rats showing an intermediate phenotype. Amount of time spent in metestrus and diestrus was not affected (Figure 6E,F).

Figure 6.

SEMA and calorie restriction influence estrous cycle. A schematic of collecting and analyzing estrous smears (A). The overall average length of the estrous cycle was not altered by the treatments (B). Days captured in proestrus were significantly decreased in PF females and were reduced in SEMA treated females (C). The percentage of cycle captured in estrus was increased in PF females compared to controls (D). Metestrus remained unchanged due to treatment (E). Detected time spent in diestrus was also resistant to treatment (F). n=10-12 per treatment. ∗ p ≤ .05. SEMA: Semaglutide, PF: Pair-fed.

4. Discussion

Reproduction and metabolism interact; they influence each other in their normal function but also in dysfunction. This is supported by the reproductive issues found in patients with metabolic dysfunction, like diabetes and obesity. Weight loss is generally thought to improve reproductive outcomes in obesity, but whether weight loss induced by the most used anti-obesity pharmacotherapy, SEMA, has the same outcome is less well understood. Self-reports of reproductive and hormonal side effects of GLP-1 based therapies are emerging and suggest some undesired or unexpected effects in this realm were not captured in clinical trials [7]. Our results demonstrate both male and female SEMA-treated rats consumed fewer calories and weighed less than controls at the conclusion of the study, and that indeed SEMA affects the reproductive axis at all levels and in a sex-specific manner. Importantly, while many of the parameters were influenced by the GLP-1 analogue treatment and caloric restriction alike, some parameters were altered exclusively by caloric restriction, while a few others were altered solely by SEMA. This suggests that much of the reproductive impact is mediated by weight loss itself, but that SEMA is also capable of altering some of these changes by either attenuating the changes brought by caloric restriction or inducing a direct effect on the tissues of interest independently of weight loss.

Consistent with previous reports, GLP-1R expression was detected in both the testes and ovaries, supporting the possibility that SEMA can directly influence gonadal tissue [[33], [34], [35], [36], [37], [38]]. While previous work detected GLP-1R in both tissues, direct comparison of expression between the sexes was not conducted. Interestingly, comparison between sexes revealed that testicular GLP-1R expression was nearly five-fold higher than ovarian expression, which may suggest greater sensitivity of male gonadal tissue to GLP-1 signaling. This is supported by more widespread and robust influence of the drug on sperm and testicular morphology demonstrated here. Prior work suggests that testicular GLP-1R are primarily found in Leydig and Sertoli cells, while ovarian expression was reported predominantly in granulosa cells [33,37,39]. Cell culture studies have identified GLP-1R on Sertoli and Leydig cells to play a mechanistic role in preserving testosterone production and thickening of germinal epithelium, similar to what was seen in SEMA males [21,40]. Activation of GLP-1R in primary cultured granulosa cells directly reduced FSH-induced steroidogenesis and progesterone output [35]. In contrast to findings previously reported in mice, relative GLP-1R expression in gonads in the present study was not significantly altered by chronic exposure to SEMA or food restriction, whereas one previous report indicates that liraglutide increases GLP-1R expression in mouse testes while another suggests that exendin-4 does not alter ovarian GLP-1R expression in obese mice [34,41].

We demonstrated that chronic SEMA reduces circulating LH, but not FSH, in both sexes. This may suggest that not only direct gonadal action on GLP1-R affects the reproductive axis, but also upstream effects potentially on the hypothalamus and pituitary are possible. This is perhaps not surprising, as obesity was previously shown to affect the secretion of gonadotropin-releasing hormone (GnRH) and consequently LH and FSH. In preclinical models, circulating LH is increased in obese females [42,43] but decreased in males, whereas FSH remains unchanged [[42], [43], [44], [45]]. Conversely, weight loss via subcutaneously injected liraglutide, in metabolically challenged rodents, increased LH levels whereas controlled caloric restriction decreased LH in both males and females [34,[46], [47], [48]]. Our results extend previous work in female rats, which found that only acute brain application of GLP-1 increases, while acute or week-long exendin-4 reduces LH [38]. The latter is more in line with the current study design. Since that work used brain infusions of the drug, this may suggest that LH changes are driven by a direct effect on the hypothalamus, likely via the paraventricular nucleus GLP-1R activation. However, since female rats in that study also lost weight, there is still a possibility that weight loss was the driver of the LH suppression. The reduction in LH identified here in both sexes with chronic peripheral SEMA is therefore consistent with suppression of the hypothalamic-pituitary-gonadal (HPG) axis by caloric restriction. Receptivity of the FSH and LH signals, on the other hand, was not different, as the gonadal expression of gonadotropin receptors, FSHR and LHR, was not altered by weight loss alone or SEMA, in either sex.

Not only pituitary hormones, but also gonadal hormone levels were altered by SEMA in both sexes, though in a sex-specific manner. Gonadal hormone changes may be driven by changes in pituitary hormone levels, but also by direct effects on gonadal steroid synthesis enzymes. It appears that both processes might contribute to the alterations in circulating gonadal hormones identified here. STAR expression was robustly reduced by both SEMA and caloric restriction in females, yet unchanged in males. This female-specific response is consistent with previous in vitro findings indicating that in cultured rat granulosa cells treatment with GLP-1 or glucose-dependent insulinotropic polypeptide (GIP) inhibits the expression of STAR [35]. Furthermore, STAR expression is also decreased in PMOS models by GLP-1R agonists [49,50]. Our results also pinpoint caloric restriction and weight loss as sufficient to reduce STAR expression in ovaries. Nevertheless, these results, together with the prior in vitro work, indicate that the impact of GLP-1 analogues on STAR might be two-pronged: direct effect on ovarian cells as well as secondary impact of reduced food intake. Our data also indicate that the effect on STAR was sex-specific – no effect of SEMA or caloric restriction on this enzyme was found in the testes. Our in vivo male results are consistent with prior in vitro work showing that the resilience of STAR expression to GLP-1 treatment was also demonstrated in isolated Leydig cells [40].

Expression of other enzymes in the steroidogenesis pathway was not altered in females. This contrasts with results obtained from in vitro work, which indicates transient changes in expression of other enzymes in cell cultures [35], potentially hinting that for some changes, the direct effect of the drug on the ovary is counteracted by the whole-body changes induced by the drug or weight loss. Yet, despite the lack of enzyme expression changes beyond STAR in females, there were quite drastic changes in pregnenolone, progesterone, and estradiol. It may be that STAR reduction creates a bottleneck that is carried to the synthesis of some of the hormones downstream of this enzyme. Moreover, gene expression is only one facet of enzymatic function; enzyme activity can be regulated in post-transcription or translation, thus there might be changes which escaped our current measurements. Yet, the reduction in progesterone identified here is not surprising. Varied effects of other GLP-1 agonists delivered directly to the brain on progesterone have been previously reported, measured by ELISA; they show reduced or increased progesterone depending on the agonist used and estrous phase of the rat [38]. However, our results are in line with another study showing that chronic systemic injections of liraglutide or calorie restriction decrease circulating progesterone in females [48,51].

Progesterone changes were clearly in line with STAR changes, and this hormone was altered by both drug and weight loss in females. Yet, the fact that pregnenolone levels are reduced by weight loss but when that weight loss takes place with GLP-1 analogue on board, pregnenolone levels remain constant, at first glance might be puzzling, at least when ovarian production is concerned. However, as adrenal glands also synthesize pregnenolone, and GLP-1 and its analogues have long been known to increase hypothalamic-pituitary-adrenal (HPA) axis activity, it is possible that the “rescue” of the drastic reduction in pregnenolone levels by SEMA is driven by increased adrenal production. While the effect of the drug on the HPA axis is well established, less is known about direct adrenal effects. A question ripe for investigation. Pregnenolone levels in males reveal a sex-divergent pattern. SEMA males tended to have reduced pregnenolone while PF males did not, implicating direct GLP-1R action on CYP11a rather than an effect of energy deficit consistent with prior reports that GLP-1-treated Leydig cells modulate CYP11a [40]. This reduction in pregnenolone did not translate to reductions in downstream hormones, suggesting SEMA redirects steroidogenic flux by reducing precursor availability but rescuing end products. Irrespective of the source of pregnenolone rescue, the fact that levels are preserved in females by SEMA compared to caloric restriction alone might be an advantage of the drug, as low levels of pregnenolone in females are associated with a wide array of negative symptoms from fatigue and decreased libido to dry skin; symptoms not unfamiliar in a fasting condition.

Testosterone and androstenedione were not significantly impacted in females. Estradiol, however, was reduced in both SEMA and PF females, indicating that the effect is a consequence of the reduced food intake rather than a SEMA-specific action. Estradiol is produced following gonadotropin stimulation in the granulosa cells, and the reductions we observed in our SEMA females parallel those reported in lean female rats given chronic systemic liraglutide treatment, which reduced both LH and progesterone and subsequently estrogen measured by electrochemiluminescence immunoassay [51]. Furthermore, other findings align with phase-specific reductions in circulating estradiol following acute intraventricular injection of GLP-1 and exendin-4 [38]. Moderate to severe calorie restriction in female rats also reduced gonadotropins and circulating estradiol [52]. In contrast, in isolated rat granulosa cells, acute GLP-1R activation suppresses FSH-induced progesterone synthesis without affecting estradiol production [35]. This suggests that the in vivo estradiol reduction we observed is driven by central HPG-axis suppression. This convergence across administration routes supports that the effects are mediated upstream of the granulosa cells.

In males, SEMA fully restored DHEA and testosterone to control levels, with partial restoration of androstenedione, while PF males showed reductions across all three androgens. The dissociation between expression of CYP17 and DHEA may be due to differences in post-translational regulation and activity of the enzyme in the gonad or contribution of DHEA from the adrenal glands. In line with reduced DHEA, but only in PF rats, testosterone levels were also reduced. The reduction in testosterone detected here is in line with previous literature showing similar changes after both short -term fasting and chronic calorie restriction [43,45,47]. This reduction might be affected by one of two factors identified here: the reduced availability of substrate (DHEA and androstenedione) or increased conversion to a product (estrogen). CYP19a1 expression was elevated in PF male testes, supporting the possibility that aromatase upregulation contributes to testosterone loss; however, circulating estradiol was not significantly elevated in PF males, suggesting that any increased aromatization may be locally compartmentalized within the testes. Overall, SEMA restores the balance for DHEA, testosterone, and androstenedione disrupted by caloric restriction, preserving androgens as the end products of steroidogenesis in males.

Not only were circulating hormones and gonadal enzymes affected by the treatment, but also morphological changes in gonads and reproductive cells were identified. In males, while the total concentration of sperm did not vary with treatment, there was an increase in total motile sperm and progressive motility in both SEMA and PF males. These changes clearly indicate greater functional capacity. This capacity might be further potentiated by increased mucus penetration ability of the sperm, indicating that SEMA and PF sperm can permeate cervical mucus to reach an egg. This is perhaps not surprising when considered as a correction of reduced capacity of sperm to perform in obesity [20,34]. As these effects were found in both SEMA and PF rats, they are likely a consequence of improved metabolism rather than a direct effect of SEMA treatment. Our results add multiple novel aspects of SEMA effect on sperm. Previous literature on this topic was quite inconsistent, as studies have used heterogeneous designs, including liraglutide and exenatide delivered by subcutaneous injection in obese mice, continuous GLP-1 infusion in lean rats, and liraglutide in diabetic, aged, and GLP-1R knockout mouse models, with reports of improved or unchanged sperm motility [20,34,53,54]. Chronic caloric restriction is associated with increased sperm defects [53]. Differences in baseline metabolic state, GLP-1 analogue, and dosing likely contribute to these inconsistencies. Our pair-fed design allowed for the contributions of weight loss and drug action to be dissociated. One important caveat to the PF design is that PF animals receive their food in one serving. Interestingly, while PF rats initially consume all their food within a few hours of receiving it, after the first week they distributed the food consumption more evenly throughout the 24 h. This potentially ameliorates the differences in timing of the food consumption between SEMA and PF rats.

We also discovered that SEMA had unique effects on the structure of the seminiferous tubules, the primary location for sperm production and maturation. The germinal epithelial layer was thicker, particularly in the SEMA-treated males, also resulting in reduced tubule diameter and area. One previous study indicated that in diabetic rats, another GLP-1 analogue, liraglutide, restores germinal epithelial thickness and architecture, indicative of an improved blood-testes-barrier [55]. Together, these data suggest that SEMA directly affects this function, and this effect is consistent between GLP-1 analogues. Moreover, this effect was clearly mediated by the drug alone and not weight loss, a unique combination in this study, but one suggesting that some effects on reproductive tissues result from direct effect of the drug rather than from metabolic improvements or corrections of calorie restriction-induced changes.

Obesity negatively influences follicle development; obese female mice have reduced numbers of early-stage follicles compared to lean controls [56,57]. In the current study, SEMA specifically had positive effects on primordial and primary follicle numbers. But both SEMA and weight loss increased the number of secondary follicles. Furthermore, antral follicles were significantly increased in PF animals, while SEMA rats showed an intermediate phenotype for antral follicles. Thus, folliculogenesis analysis suggests a potential for improved fertility resulting from metabolic improvements with weight loss and additive benefits of SEMA in early-stage follicles. These results are in line with a previous study showing that even an acute treatment with GLP-1 amide increases the number of follicles in lean rats [38]. Following ovulation, the corpus luteum acts like a temporary endocrine organ. Despite having less circulating progesterone due to weight loss, the observed numbers of corpus lutea did not differ between groups, potentially suggesting functional loss of progesterone production. This phenomenon was observed previously in calorie-restricted rats, further suggesting a complex relationship between progesterone production and energy balance [48].

On the other hand, negative indicators of ovarian health produced mixed results. The number of atretic or degraded follicles did not differ between groups. However, ovarian cysts were detected exclusively in SEMA and PF rats, with most of the SEMA rats (66% vs 0% for controls) having at least one cyst. This is an unexpected but potentially problematic finding, as these drugs are increasingly suggested for targeted use in PMOS patients [17]. However, an ameliorating factor is that the number of cysts detected in the SEMA rats was generally on the lower end, with most rats having 1–2 cysts.

Length of the estrous cycle as well as of changes to specific phases are sensitive to energy state, and this interaction is modulated by length and severity of metabolic disturbances. Frequent, prolonged caloric restrictions lead to acyclicity in female mice [58], while excess energy intake and obesity also disrupt normal cycling in rats [59,60]. In PMOS rodent models, GLP-1 receptor agonists, liraglutide and exendin-4, restored regular cycling [61]. In this study, all rats displayed normal estrous cycling and the average length of the complete cycle did not differ between groups. Yet changes in specific phases were identified. Proestrus length was significantly decreased in both PF and SEMA females consistent with greater antral follicle development. Estrus phase duration, however, was increased in PF females, with SEMA females as an intermediate phenotype between control and PF. This indicates a possible partial rescue of the phase in which the female is in heat despite having a higher weight loss compared to calorie restriction alone and is consistent with lower serum estradiol potentially prolonging estrus until estradiol is high enough to trigger a GnRH surge and ovulation. The percent of the cycle observed in metestrus and diestrus phases remained unchanged. This is not the first time that a GLP-1 analogue has been linked with altering estrous phase length, although specific changes detected differed, possibly because of different agonist half-lives and other characteristics, such as species differences and metabolic baselines [41]. Moreover, together these findings illustrate that the reproductive consequences of weight loss are dependent on the baseline metabolic state. In obese animals, energy deficit shifts cycling toward fertility-favorable phases rather than suppressing cyclicity, and SEMA females display a milder version of this shift.

As the use of GLP-1 receptor agonists becomes increasingly common in reproductive-age populations for the treatment of obesity, understanding their off-target effects on hormones and gonadal function is crucial. While some past literature hinted at effects of GLP-1 analogues in this arena, there was a paucity of data with the currently most-used analogue: SEMA. Moreover, potential side-effects of SEMA linked to reproductive and endocrine function that were not captured in clinical trials are beginning to emerge and might depend on whether the patient has dysregulated cycles at baseline [7,8]. Together, our findings demonstrate that SEMA alters pituitary hormone levels, steroidogenesis, sperm quality, as well as ovarian and testicular morphology in a sex-divergent and tissue-specific manner. Importantly, a complex and sex-specific interaction of SEMA with calorie restriction was identified, where for some parameters the impact of SEMA and equivalent weight loss without the drug was comparable, while for others SEMA diverged from drugless weight loss, most often ameliorating what could be construed as deficits induced by caloric restriction. For males, also effects of SEMA independent of weight loss were found. Our results can be interpreted as SEMA leading to mixed effects on female reproductive capacity and largely an improvement in males. Future studies should follow up on our findings to determine whether these endocrine, morphological, and reproductive cell changes can be translated to the functional outcome of reproductive success. This will be quite important, as some studies indicate that the motivation for sex might be reduced by GLP-1 analogues [62,63], and the health of the offspring might also be compromised [64], while our data and recent review of clinical findings [65] suggest that chances of conception might be increased. This future hypothesis is also consistent with some anecdotal reports, indicating unexpected pregnancies in women with fertility issues while on GLP-1 analogues. Our data showing massively reduced levels of estrogen should also be considered in light of recent reports indicating that Tirzepatide, dual GLP-1 and GIP receptor agonist, reduces plasma levels of ethinyl estradiol, necessitating backup methods of contraception [66]. Another important future aspect suggested by present work, is determining how lasting is the impact of the drug on reproductive function, as currently women are asked to stop SEMA two months beforeconception. It's possible that the changes persist, but it's also possible that sudden withdrawal from the drug will dysregulate the reproductive axis, especially in females.

CRediT authorship contribution statement

Morgan R. Sotzen: Writing – review & editing, Writing – original draft, Visualization, Formal analysis, Data curation, Conceptualization. Suyeun Byun: Writing – review & editing, Data curation. Ngozi O. Ibadin: Writing – review & editing, Visualization, Formal analysis, Data curation. Mya A. Knappenberger: Writing – review & editing, Data curation. Madison T. Bento: Writing – review & editing, Data curation. Mohammed Asker: Writing – review & editing, Data curation. Sergei Koshkin: Data curation, Formal analysis, Writing – review & editing. Francisco J. Diaz: Writing – review & editing, Visualization, Formal analysis, Data curation. Karolina P. Skibicka: Writing – review & editing, Writing – original draft, Resources, Project administration, Funding acquisition, Conceptualization.

Ethics declaration

This study was conducted in accordance with the ARRIVE (Animal Research: Reporting of In Vivo Experiments) guidelines. This study was approved by the PSU IACUC.

Funding

This research was funded by the National Institutes of Health R01DK129321 to KPS, Hotchkiss Brain Institute, CIHR (KPS), and NSERC (KPS). T32 fellowship T32GM154124 (NOI).

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Karolina Skibicka reports financial support was provided by Canadian Institutes of Health Research. Karolina Skibicka reports financial support was provided by National Institute of Diabetes and Digestive and Kidney Diseases. Karolina Skibicka reports financial support was provided by The Hotchkiss Brain Institute at the University of Calgary. If there are other authors, they declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

Data will be made available on request.