Anti-Müllerian hormone as a predictor of oocyte yield following controlled ovarian stimulation in the rhesus macaque

Abstract

Anti-Müllerian hormone (AMH) is widely used in the clinic as a biomarker for ovarian reserve and to predict ovarian response to gonadotropin stimulation. Patients with higher AMH levels tend to yield more oocytes and have better outcomes from assisted reproductive technology procedures. The goal of this study is to determine if AMH can be used to predict the outcome of controlled ovarian stimulation in rhesus macaques, which are commonly used in biomedical research, to refine animal use while maximizing oocyte yield. We hypothesized that pre-stimulation AMH values can be used to predict oocyte yield and quality. Regularly cycling adult macaques underwent controlled ovarian stimulation and baseline (pre-stimulation) plasma AMH levels were determined using an AMH-specific enzyme-linked immunoassay. Oocytes were collected by laparoscopic or ultrasound-guided aspiration, then counted and evaluated for quality and stage of meiosis. Sperm from established fertile males were used to inseminate the oocytes in vitro with fertilization success checked 14–16 h later. Females were grouped by oocyte yield: low ≤17; mid = 18–41; high ≥42. We found that high and mid yielders had significantly higher AMH than low yielders (p < 0.0001) and the percent of mature oocytes was greater in the high and mid yielders. There were no significant differences in oocyte quality or ova fertilization rate. These data suggest that AMH is a useful measure for controlled ovarian stimulation success in rhesus macaques and can be used to identify suitable animals for oocyte donation before entering them into a stimulation protocol.

Summary Sentence 

Measurement of serum Anti-Müllerian hormone prior to a controlled ovarian stimulation can be used to predict oocyte yield and the recovered proportion of mature ova in rhesus macaques.

Graphical Abstract

Graphical Abstract.

Anti-Mullerian hormone is a useful measure for controlled ovarian stimulation success in rhesus macaques and can be used to refine animal use while maximizing oocyte yield. Created in https://BioRender.com.

Introduction

Anti-Müllerian hormone (AMH) or Müllerian inhibiting substance, is a dimeric glycoprotein that was initially recognized for its role in male sex differentiation [1, 2]. Specifically, its expression by Sertoli cells of the fetal testes initiates regression of the Müllerian ducts, a fetal structure that gives rise to the uterus, fallopian tubes, and upper vagina [3]. While AMH does not play a role in female sex determination in utero, it is expressed in the ovarian granulosa cells of fetal pre-antral follicles as early as 36 weeks of human gestation. In humans, expression continues to increase after birth until 24.5 years of age when AMH levels begin to decline, completing expression just prior to menopause [4, 5]. In reproductive-aged women (15–49 years of age), AMH is expressed by small growing follicles recruited each cycle from the primordial follicle pool; its expression declines when the antrum, the portion of an ovarian follicle filled with follicular fluid, forms and the follicle reaches a diameter of 4–6 mm [6]. In addition to supporting follicle growth, AMH may inhibit premature follicular development by impeding the selection of primordial follicles into the developing cohort, and diminish responsiveness of the growing follicles to follicle stimulating hormone as a mechanism to regulate follicle pool depletion and support dominant follicle selection [2, 7, 8].

In a clinical setting, AMH serum concentration has been utilized as a common biomarker for ovarian reserve and as a predictor of the number of primordial follicles available for recruitment. An additional characteristic that distinguishes AMH from other hormones is that it does not vary considerably over the menstrual cycle [9, 10]. Multiple findings in human studies have observed a strong association between serum AMH levels and ovarian response [2, 11–15]. Specifically, day 3 serum AMH levels in patients that produced eleven or more oocytes were 2.5 times greater than patients that produced six or fewer oocytes after ovarian stimulation [2]. These findings have supported the use of AMH as a common and effective clinical predictor of success for patients requiring assisted reproductive technology (ART) to establish a pregnancy and is a valuable tool for identifying patients that have a higher chance for successful oocyte recovery following an ovarian stimulation.

Although AMH is a commonly used measure in human fertility clinics, such application of AMH values are yet to be fully defined and implemented in rhesus macaques (Macaca mulatta). Rhesus macaques are one of the most commonly utilized nonhuman primates (NHP) in biomedical research, primarily due to their close evolutionary link to humans. This makes them a translatable model for the study of a variety of high-priority human health problems, such as human immunodeficiency virus, obesity, and cognitive aging [16]. ART in NHPs is vital for some aspects of biomedical research as it can provide a source of oocytes and embryos, which are critical for the generation of pluripotent stem cells for cell-based therapies, monozygotic twinning, somatic cell nucleus transfer, and with recent advancements, NHP transgenic models of human disease from gene-modified zygotes [17–19]. Controlled ovarian stimulation (COS) uses exogenous gonadotropin hormones to stimulate the production of multiple mature follicles for oocyte collection [20]. Establishing appropriate AMH concentration values that are accurate in predicting COS outcomes in rhesus macaques is integral in reducing animal use while ensuring a maximal oocyte yield.

Similar to humans, AMH in rhesus macaques has been found to play an important role in follicular development and its serum levels have been indicated as a predictor of ovarian reserve [21, 22]. However, the ability to use AMH as a predictor of COS outcomes, such as the response to the COS protocol, oocyte yield, quality, maturation, and fertilization rate, remains limited. Reports in the cynomolgus macaque (Macaca fascicularis) found serum AMH concentration was positively correlated with number of oocytes retrieved following ovarian stimulation [23]. Due to evolutionary similarities among NHP species, we hypothesized that AMH values could be an effective predictor of COS outcomes in rhesus macaques, and could become a major determinant in whether an animal should be used for ovarian stimulation procedures.

Materials & methods

Animal husbandry and ethics

All animals were socially housed indoors and cared for by Animal Resources & Research Support at the Oregon National Primate Research Center (ONPRC). The temperature was maintained at 22°C and a light-regulated cycle that was 12 h light/12 h dark. Water was available ad libitum; diet consisted of chow (TestDiet, St. Louis, MO, USA), fresh vegetables, and fruit. All protocols were followed in strict accordance as approved by the ONPRC Institutional Animal Care and Use Committee and followed regulations set forth by the Office of Laboratory Animal Welfare.

Ovarian stimulation and oocyte collection

A group (n = 38) of regular cycling adult rhesus macaques (M. mulatta), ages 7–18 years old and weighing 5–10 kg, were monitored daily for the onset of menses. Once detected, females were started on a controlled ovarian stimulation (COS) cycle within the first 4 days of the ovarian cycle, as previously described [20]. Briefly, animals were administered 30 IU BID of recombinant human follicle stimulating hormone intramuscularly (IM) on days 1–8 with at least 8 h between daily doses. A single dose of a gonadotropin releasing hormone antagonist (acyline, 0.75 mg/kg) was given subcutaneously between days 4 and 7 when serum estrogen (E₂) exceeded 200 ng/mL. On days 7–8, 30 IU BID of recombinant human luteinizing hormone was administered IM and a single dose of 1100 IU human chorionic gonadotropin (hCG) was administered IM on day 8. A baseline peripheral blood sample was collected on the morning of the first stimulation day (COS day 1), prior to hormone administration, as well as day 4–6 and 8–9 to ensure response. Blood was centrifuged at 1400 rpm for 10 min, plasma was collected and measured for estradiol (E₂) and progesterone (P₄) concentration by the ONPRC Endocrine Technology Core using radioimmunoassay with a Roche Cobas e411 system (Roche Diagnostics, Indianapolis, IN, USA). Plasma from the pre-treatment on COS day 1 was selected for AMH measurement and stored at −20°C until analysis.

Approximately 36 h after hCG administration (COS day 10), follicles on both ovaries were aspirated by a trained surgeon by either a laparoscopic or an ultrasound-guided approach. Aspirates were kept in pre-warmed (37°C) Hepes-buffered Tyrode’s Albumin Lactate Pyruvate (TALP-HEPES) supplemented with 3 mg/mL bovine serum albumin (BSA) and 10 IU/mL heparin for transportation (<5 min) back to the lab. Aspirates were diluted to a final volume of 10 mL with TALP-HEPES, and 300 μL of hyaluronidase (10 mg/ml) was added to facilitate removal of granulosa cells. The aspirates were then filtered through a 70 μM cell strainer to remove cellular debris and red blood cells. The cell strainer was backwashed into a 60 mm petri dish with 10 mL TALP-HEPES + BSA to retrieve oocytes. Samples were searched for oocytes under a dissecting microscope, cumulus was removed by passing repeatedly through a STRIPPER pipette (MXL3-STR, CooperSurgical Fertility Solutions, Denmark) and transferred to a 30 mm petri dish containing fresh TALP-HEPES for grading. Oocytes were then sorted into a 4-well culture plate containing 500 μl of pre-equilibrated TALP medium (37°C in 5% CO₂ humidified air) according to their stage of meiosis at the time of collection—germinal vesicle (GV), metaphase-I (MI), and metaphase-II (MII). Total oocyte yield was sorted into three groups of responders: low (≤17 oocytes), medium (18–41 oocytes), and high (≥42 oocytes). Failed responders were defined as animals that were not taken to follicle aspiration due to either an E₂ value that does not exceed 150 pg/ml by day 5 of stimulation, an E₂ value that drops during stimulation, a P₄ that increases above 0.5 ng/mL during stimulation (indicating premature luteinization), or a P₄ value of <1.0 ng/ml 12 h following hCG administration.

Semen collection and in vitro fertilization

One to eleven semen samples were collected from male rhesus macaques (n = 13) that are proven breeders and trained for collaborative non-sedated donation, as previously described [24]. Samples were collected into a sterile tube and kept at 37°C for 30 min to allow semen liquefaction. The liquid component of the ejaculate was aspirated and transferred to a sterile 15 mL conical vial while the coagulated portion was rinsed with 5 ml 37°C TALP-HEPES and added to the liquid portion. The final volume was quantum satis (Q.S.) to 12 mL with TALP-HEPES and centrifuged at 300 × g for 7 min. After the first wash, 10 mL of the supernatant was aspirated and sperm pellet was re-suspended in TALP-HEPES and Q.S to 12 mL for a second centrifugal wash. Following the second wash, 11 mL of supernatant was discarded and the sperm pellet was re-suspended in the remaining 1 mL of medium. Sperm were assessed for concentration and motility using a computer assisted sperm analysis (CASA) system (IVOS II-Animal Motility software, version 1.11, Hamilton Thorne, Beverly, MA, USA), programmed for rhesus macaque spermatozoa. To do this, 10 μL of the sperm solution was added to 190 μL of TALP-HEPES + BSA to dilute sperm 1:20 before adding 6.5 μL to a pre-warmed slide for analysis. Using calculated semen parameters from the CASA, a 2 mL sperm solution was prepared containing pre-equilibrated TALP medium to create a motile sperm concentration of 20 × 10⁶ cells/mL, and maintained in the incubator until in vitro fertilization (IVF). Approximately 6 h after collection isolation, in vitro maturation changes in meiosis were recorded and only MII oocytes were sorted into wells containing 100 μL pre-equilibrated IVF medium (BO-IVF, IVF Bioscience, Falmouth, England, UK) to be fertilized. Sperm were activated by combining 900 μL of the incubated sperm solution with 100 μL of IVF activator (5 mg/mL cAMP and 2 mg/mL caffeine in saline) and incubated for 15 min prior to IVF. Following incubation, 10 μL of IVF activator were added to each well prior to insemination with 1 μL of activated sperm. Oocytes and sperm were co-cultured for 14–16 h before fertilization rate was determined by the presence of two polar bodies and/or two pronuclei.

Quantification of AMH

Collected baseline blood plasma from 72 cycles (n = 38 females, 1–4 cycles per female) from different projects was retrospectively utilized for the quantification of AMH in aims to reduce research animal use. AMH was quantified using an enzyme-linked immunoassay, specific for human and NHP AMH (Ansh Labs LLC, Webster, TX, USA) and the protocol was carried out in accordance with the manufacturer’s instructions with samples run in duplicate. The assay has an analytical measurement range of 0.084–14.2 ng/mL. Samples were diluted 1:1–1:2 for analysis with the final concentration being determined by matching the mean optical density, measured using a tunable microplate reader (Molecular Devices LLC, San Jose, CA, USA) with the respective AMH concentration and applying dilution factors. The intra-assay coefficients of variation were ˂14% and the inter-assay coefficient was ˂5%.

Statistical analysis

Statistical interrogation of the data was blindly conducted using Graphpad Prism Version 9.3.1. One-way ANOVA with Tukey’s multiple comparisons test was used to compare plasma AMH levels between groups. Differences were considered significant at p < 0.05 and values are presented as mean ± standard deviation. To examine relationships between AMH levels and other variables, a correlation test was used to compute Pearson correlation coefficients, with p < 0.05 considered significant. To identify and remove outliers, we performed the ROUT method with Q = 1%.

Results

Baseline plasma AMH concentration was determined to be positively correlated to oocyte yield following a COS (R² = 0.42, Figure 1A). When grouped into different yields (low: ≤17 oocytes, mid: 18–41, high: ≥42, and failed responders), the baseline AMH concentration of the high responding group (38.3 ± 12.9 ng/mL) and the mid responding group (31.9 ± 12.8 ng/mL) were significantly different (p < 0.0001) from the low responding group (14.9 ± 6.3 ng/mL, Figure 1B). It should be noted that the failed responders group includes animals whose AMH levels may have initially been sufficient to predict oocyte yield but experienced an overall failed cycle due to reasons other than inadequate E2 and/or P4, e.g. stress. Therefore, we did not perform a statistical comparison between the failed responders and the other groups; however, we have included to oocyte yield data in Figure 1 for transparency. The average percent of MII oocyte composition among high (55.6%) and mid (56.12%) oocyte yielding groups was similar; both were significantly higher (p ≤ 0.05) than the low yielding group (38.8%, Figure 1B and D). No difference was observed among groups in MII fertilization rate (Figure 1C and F).

Figure 1.

Baseline AMH levels are positively correlated to oocyte yield following controlled ovarian stimulation. A, D) Higher baseline AMH levels were correlated with higher oocyte yield following COS; High and mid responding groups had higher AMH levels than low responders. B, E) Higher baseline AMH levels were correlated with higher % MII; Both high and mid groups had higher % MII oocytes than the low responders. C, F) Fertilization rate of MII oocytes among all responding groups were similar and there was no correlation between fertilization rate and AMH levels. High (≥42 oocytes, n = 18), mid (18–41 oocytes, n = 20), low (≤17 oocytes, n = 23), and failed (n = 11). ^*p < 0.05, ^****p < 0.0001.

When investigating individual animal ages, no significant difference was observed in age (years) or weight (kg) between oocyte yielding groups (Figure 2A and B). It was observed that there were no differences in E₂ and P₄ levels prior to the initiation of a COS cycle (Figure 3B and D), however, there were significant variations in post-COS E₂ and P₄ across oocyte yielding groups (Figure 3F and H). Post-COS E₂ was significantly higher in both the high (3905 ± 1156 pg/mL; p < 0.01) and mid (4262 ± 2450 pg/mL; p < 0.001) yielding groups when compared to the low yielding group (2164 ± 844 pg/mL). There was no observed variation in P₄ levels among yielding groups. No difference in oocyte quality was observed between groups. To address the concern that multiple COS cycles may affect how individual animals respond to the stimulation, we looked at AMH levels and oocyte yield after each successive COS. Both AMH and the number of oocytes retrieved remained similar across cycles (Figure 4).

Figure 2.

No difference in age or weight among responding groups. A) No difference was observed among any of the groups in age. B) Responding groups had similar weights. High (≥42 oocytes, n = 18); mid (18–41 oocytes, n = 20), low (≤17 oocytes, n = 23).

Figure 3.

Hormonal differences after COS. A, C) No significant correlation was found between baseline AMH levels versus pre-COS E2 levels (n = 72, p = 0.25) or pre-COS P4 levels (n = 72, p = 0.74). B, D) We did not observe any significant differences in pre-COS E2 or pre-COS P4 between high (≥42 oocytes, n = 18), mid (18–41 oocytes, n = 20) and low (≤17 oocytes, n = 23) responding groups. E) There was a significant positive correlation between baseline AMH and post-COS E2 (n = 72, p = 0.0004). F) Significant differences were observed in post-COS E2 among the responding groups. G, H) There was no correlation between baseline AMH and post-COS P4 (n = 72) and no post-COS P4 differences between the responding groups. ^***p < 0.001, ^**p < 0.01.

Figure 4.

AMH levels and oocyte yield are not changed by successive COS cycles. A) Changes in individual macaque AMH plasma concentration after multiple COS cycles were not observed. B) No changes were observed in oocyte yield after multiple COS cycles.

Discussion

Assessment of AMH levels is used clinically as an indicator of ovarian reserve and as a predictor of reproductive success in women. Current knowledge on how plasma AMH levels can predict ovary stimulation outcomes in NHPs remains limited, which hinders biomedical research due to a high degree of variability in ovarian response and ART outcomes among animals. Understanding the reproductive value of AMH in rhesus macaques is a necessary refinement to optimize selection of suitable animals for oocyte collection.

Our study found that in rhesus macaques, baseline AMH concentration and oocyte yield following a COS are positively correlated. Animals that showed higher AMH values produced a higher oocyte yield, which is consistent with what is described in humans and cynomolgus macaques [2, 11, 12, 23]. As expected, high and mid responders had higher post-COS E₂ values than low responders due to increased follicular development and oocyte output.

The similar AMH levels between failed, mid, and high responders suggest that animals in the failed response group were ultimately unable to respond to COS due to reasons other than insufficient AMH levels and ovarian reserve. Poor ovarian response, in humans, is categorized similarly by low E₂ levels from only a small number of follicles developing after standardized stimulation protocol [25], which is consistent with the low E₂ and P₄ measured among failed responders in this study. These low E₂ and P₄ values during stimulation were already used to identify and prevent animals from undergoing follicle aspiration. This indicates that pre- stimulation AMH can be used to reflect oocyte yield, however, COS and E2/P4 values are needed to identify individuals with impaired ovarian response to stimulation. Factors for poor response include not only inadequate AMH levels, but also the development of gonadotropin antibodies, advanced maternal age, and higher weight, both in humans and rhesus macaques [25–27]. No difference in egg donor age was observed between yield groups, which was expected, primarily due to the limited sample pool of all rhesus macaques being of ideal reproductive age for COS. However, the significantly higher mean weight among animals that failed to respond compared to the mid yielding group suggests that weight could be an underlying cause for this group of animals.

High and mid responders also yielded more MII oocytes compared to low responders (High = 28.3 ± 10.0; Mid = 17.7 ± 7.0; Low = 3.7 ± 2.7; Table 1). This suggests that animals with higher AMH values developed more follicles after stimulation but also had a higher ratio of follicles respond to exogenous hCG, which mimicked an LH peak to initiate resumption of meiosis and oocyte maturation [28]. Metaphase II is the only oocyte stage capable of undergoing fertilization, so rhesus with higher AMH values produce more oocytes capable of being fertilized in total. Moreover, there was no difference in MII fertilization rates across groups, suggesting there is not a tradeoff between oocyte number and fertilization capability of the oocytes themselves. This has also been observed in humans but remains contradictory [15, 29].

Table 1.

Distribution of oocytes at time of retrieval by stage of meiosis and responder category. The number of GV, MI, and MII oocytes were compared based on the category of response (total number of oocytes) from the controlled ovarian stimulations. ^**p < 0.01, ^***p < 0.001 versus High Responders; ^###p < 0.001 versus mid responders.

Oocyte Yield	GV (#)	MI (#)	MII(#)
High (>42)	11.1 ± 6.0	11.7 ± 7.1	28.3 ± 10.0
Mid (18–41)	7.8 ± 4.3	6.0 ± 4.6**	17.7 ± 7.0***
Low (<17)	3.8 ± 3.1***	2.6 ± 2.2***	3.7 ± 2.7***###

As previously stated, this study’s purpose was to show the relationship between serum baseline AMH and ovarian stimulation outcomes in reproductive age females. All animals were screened for ideal weight, maternal age, parity, and temperament, to warrant undergoing COS and follicle aspiration. It is possible that a greater difference in AMH levels and COS outcomes would be observed if these animals had a greater distribution of age and weight, however only regularly cycling animals of reproductive age (7–18 years old) and weight (5–10 kg) were evaluated for ovarian stimulation. Previous studies have shown that individuals with advanced maternal age and obesity have lower AMH levels and therefore limited outcomes from ovarian stimulation [11, 30, 31]. Through limiting these variables, this study aimed to utilize the AMH assay as another screening method for rhesus macaques prior to enrollment in a COS cycle in an effort to maximize oocyte yield and limit animal use, and therefore represents a significant improvement in animal use and welfare. AMH values were higher than previously observed in cynomolgus macaques, where the cutoff for high responders (≥51 oocytes) was 15.6 ng/mL and low responders (≤14 oocytes) was 9.7 ng/mL [23, 29, 32, 33]. These differences could be credited to differences in species and resources, although it should be noted that our results were independently validated by the Endocrine Technology Services Core at the Oregon National Primate Research Center. Another limitation is that this study only looked at the influence of AMH on COS outcomes until fertilization. Resultant zygotes were used for other studies that prevented the collection of any developmental data including blastocyst formation or pregnancy rates following embryo transfer. Although higher AMH values in patients may increase the likelihood of obtaining a pregnancy due to the availability of a higher number of blastocysts for transfer, the existing evidence regarding the influence of AMH beyond folliculogenesis remains under debate [34]. Although there were no differences in oocyte yield or AMH levels after subsequent stimulation cycles in each animal, it is possible that those in the failed response group developed antibodies to the gonadotropin treatments, thus limiting their response [27]. Properly quantifying gonadotropin antibodies to stimulation material or developing rhesus-specific gonadotropins is a method that could be later employed to distinguish factors contributing to these observed effects.

As previously suggested, further steps to advance this work include assessing the applicability of AMH to predict pregnancy outcomes based on blastocyst rates, oocyte and blastocyst survival rates after cryopreservation, and implantation rates, which are vital in increasing the success of embryo transfers in rhesus macaques. Beyond in vitro ART techniques, the value of AMH as a predictor can be further investigated in the time-mated breeding of rhesus macaques to not only increase access to the limited number of animals available, but also to gain further knowledge of NHP reproductive physiology. Further studies will aim to gain insight on why and how animals in the failed response group continually fail to respond to COS, and may lead to the development of methods that would allow poor responders to become viable oocyte donors.

In conclusion, AMH can be an effective predictor of both oocyte number and in vivo maturation rates following a COS in rhesus macaques, but further evidence is needed to determine the full extent of AMH influence on NHP reproductive physiology. These investigations are integral for the identification of individuals with a higher ovarian reserve, and thus more suitable for successful COS-related studies, ultimately reducing the number of animals needed. Additionally, advancing knowledge of NHP reproduction and ART techniques is crucial in aiding the increasing demand for implementation of conservation efforts, given that the International Union for the Conservation of Nature states that ⁓75% of the world’s primates have declining populations and 60% are currently threatened with extinction [35].

 

Conflict of interest: No potential or actual conflicts of interest to disclose for any of the authors.

Author contributions

J.V.J.: Project administration, methodology, investigation, formal analysis, writing—original draft, writing—review and editing. P.Y.L.: Methodology, formal analysis, visualization, writing—review and editing. E.C.M.: Investigation, writing—review and editing. F.C.B.: Investigation, writing—review and editing. N.P.: Writing—review and editing. C.V.B.: Conceptualization, methodology, writing—review and editing. C.B.H.: Conceptualization, methodology, resources, funding acquisition, writing—review and editing.

Data availability

The data underlying this article will be shared on reasonable request to the corresponding author.