Positive effects of amphiregulin on human oocyte maturation and its molecular drivers in patients with polycystic ovary syndrome

Abstract

STUDY QUESTION

Does use of medium containing amphiregulin improve meiotic maturation efficiency in oocytes of women with polycystic ovary syndrome (PCOS) undergoing in vitro maturation (IVM) preceded by a capacitation culture step capacitation IVM (CAPA-IVM)?

SUMMARY ANSWER

Use of medium containing amphiregulin significantly increased the maturation rate from oocytes retrieved from follicles with diameters <6 or ≥6 mm pre-cultured in capacitation medium.

WHAT IS KNOWN ALREADY

Amphiregulin concentration in follicular fluid is correlated with human oocyte developmental competence. Amphiregulin added to the meiotic trigger has been shown to improve outcomes of IVM in a range of mammalian species.

STUDY DESIGN, SIZE, DURATION

This prospective, randomized cohort study included 30 patients and was conducted at an academic infertility centre in Vietnam from April to December 2019. Patients with PCOS were included.

PARTICIPANTS/MATERIALS, SETTING, METHODS

In the first stage, sibling oocytes from each patient (671 in total) were allocated in equal numbers to maturation in medium with (CAPA-AREG) or without (CAPA-Control) amphiregulin 100 ng/ml. After a maturation check and fertilization using intracytoplasmic sperm injection (ICSI), all good quality Day 3 embryos were vitrified. Cumulus cells (CCs) from both groups were collected at the moment of ICSI denudation and underwent a molecular analysis to quantify key transcripts of oocyte maturation and to relate these to early embryo development. On return for frozen embryo transfer (second stage), patients were randomized to have either CAPA-AREG or CAPA-Control embryo(s) implanted. Where no embryo(s) from the randomized group were available, embryo(s) from the other group were transferred. The primary endpoint of the study was meiotic maturation efficiency (proportion of metaphase II [MII] oocytes; maturation rate).

MAIN RESULTS AND THE ROLE OF CHANCE

In the per-patient analysis, the number of MII oocytes was significantly higher in the CAPA-AREG group versus the CAPA-Control group (median [interquartile range] 7.0 [5.3, 8.0] versus 6.0 [4.0, 7.0]; P = 0.01). When each oocyte was evaluated, the maturation rate was also significantly higher in the CAPA-AREG group versus the CAPA-Control group (67.6% versus 55.2%; relative risk [RR] 1.22 [95% confidence interval (CI) 1.08–1.38]; P = 0.001). No other IVM or embryology outcomes differed significantly between the two groups. Rates of clinical pregnancy (66.7% versus 42.9%; RR 1.56 [95% CI 0.77–3.14]), ongoing pregnancy (53.3% versus 28.6%; RR 1.87 [95% CI 0.72–4.85]) and live birth (46.7% versus 28.6%; RR 1.63 [95% CI 0.61–4.39]) were numerically higher in the patients who had CAPA-AREG versus CAPA-Control embryos implanted, but each fertility and obstetric outcome did not differ significantly between the groups. In the CAPA-AREG group, there were significant shifts in CC expression of genes involved in steroidogenesis (STAR, 3BHSD), the ovulatory cascade (DUSP16, EGFR, HAS2, PTGR2, PTGS2, RPS6KA2), redox and glucose metabolism (CAT, GPX1, SOD2, SLC2A1, LDHA) and transcription (NRF2). The expression of three genes (TRPM7, VCAN and JUN) in CCs showed a significant correlation with embryo quality.

LIMITATIONS, REASONS FOR CAUTION

This study included only Vietnamese women with PCOS, limiting the generalizability. Although 100 ng/ml amphiregulin addition to the maturation culture step significantly improved the MII rate, the sample size in this study was small, meaning that these findings should be considered as exploratory. Therefore, a larger patient cohort is needed to confirm whether the positive effects of amphiregulin translate into improved fertility outcomes in patients undergoing IVM.

WIDER IMPLICATIONS OF THE FINDINGS

Data from this study confirm the beneficial effects of amphiregulin during IVM with respect to the trigger of oocyte maturation. The gene expression findings in cumulus indicate that multiple pathways might contribute to these beneficial effects and confirm the key role of the epidermal growth factor system in the stepwise acquisition of human oocyte competence.

STUDY FUNDING/COMPETING INTEREST(S)

This work was funded by the Vietnam National Foundation for Science and Technology Development (NAFOSTED; grant number FWO.106-YS.2017.02) and by the Fund for Research Flanders (FWO; grant number G.OD97.18N). L.N.V. has received speaker and conference fees from Merck, grants, speaker and conference fees from Merck Sharpe and Dohme, and speaker, conference and scientific board fees from Ferring. T.M.H. has received speaker fees from Merck, Merck Sharp and Dohme and Ferring. J.S. reports speaker fees from Ferring Pharmaceuticals and Biomérieux Diagnostics and grants from FWO Flanders, is co-inventor on granted patents on CAPA-IVM methodologies in USA (US10392601B2), Europe (EP3234112B1) and Japan (JP 6806683 registered 08-12-2020) and is a co-shareholder of Lavima Fertility Inc., a spin-off company of the Vrije Universiteit Brussel (VUB, Brussels, Belgium). NA, TDP, AHL, MNHN, SR, FS, EA and UDTH report no financial relationships with any organizations that might have an interest in the submitted work in the previous three years, and no other relationships or activities that could appear to have influenced the submitted work.

TRIAL REGISTRATION NUMBER

NCT03915054.

oocyte, in vitro maturation, embryo, epidermal growth factor, amphiregulin, meiosis, polycystic ovary syndrome

Introduction

Invitro maturation (IVM) is a potential alternative to traditional invitro fertilization (IVF) in specific patient groups. In IVM, immature cumulus–oocyte complexes (COCs) are retrieved from small- and medium-sized follicles before dominance is reached in the ovaries and then matured in vitro. The optimal strategy is still being determined, and this relies on thorough knowledge of the intra-follicular ovarian physiology of small healthy (pre-dominant) follicles.

Human research in this field has been facilitated by basic reproductive biology in rodent models and, especially, livestock breeding (Tsafriri et al., 1996; Conti et al., 1998; Blondin et al., 2002; Nivet et al., 2012; Sugimura et al., 2015; Dieci et al., 2016). In cattle breeding programs, specific factors (e.g. epidermal growth factor [EGF], adenylyl cyclase inhibitors and stimulators, phosphodiesterase inhibitors) have been systematically included in culture media to trigger the variety of pathways known to be involved in oocyte maturation (Lonergan et al., 1996; Le Beux et al., 2003; Sirard et al., 2007; Gilchrist et al., 2016). Another critical aspect of the successful application of IVM in large mammals is that only a very restricted size-class of predominant follicles is considered to provide the COC for IVM culture. This ensures that the oocyte cohort for IVM culture is always well synchronized (Luciano et al., 2011; Sugimura et al., 2015; Dieci et al., 2016). However, in human IVM the clinician is often confronted with a dyssynchronized cohort of follicles, from which as many COCs as possible need to be harvested.

Over the last five years, we have developed a two-step culture system for clinical human application, called capacitation IVM (CAPA-IVM) for COCs from pre-dominant follicles (2–10 mm in diameter) which are retrieved without prior exposure to human chorionic gonadotropin (hCG) stimulus. In the first culture step (CAPA), the germinal vesicle (GV) stage oocytes progress in nucleolar conformation, to rearrange organelles and shut down transcription, while the interconnectivity between the oocyte and cumulus complement is maintained (Sánchez et al., 2017). Therefore, C-type natriuretic peptide (CNP), which is a major physiological factor in the human follicle, is added to the CAPA medium to generate cyclic guanosine monophosphate. This inhibits activation of phosphodiesterase type-3A in the oocyte so that meiotic arrest is conserved (Norris et al., 2009; Vaccari et al., 2009). CNP acts via the natriuretic peptide receptor 2 expressed on cumulus cells (CCs) under the influence of oestradiol (Zhang et al., 2011). Other factors that support the viability of CCs when present at a physiological concentration, such as follicle-stimulating hormone (FSH) and insulin, are also added during this capacitation step (Romero et al., 2016). In the second culture step, designed to provide the ingredients for a positive meiotic-inducing stimulus, a combination of FSH and amphiregulin was chosen based on preclinical research (Ashkenazi et al., 2005; Procházka et al., 2011; Peluffo et al., 2012; Romero et al., 2016).

One of the key milestones in the small antral follicle development in mammals is the acquirement of EGF responsiveness (Sugimura et al., 2015). The accumulation of amphiregulin (the main EGF-like ligand in the ovulatory cascade) in the follicular fluid (FF) was observed following a gonadotropin surge both in humans and rhesus monkeys (Inoue et al., 2009; Zamah et al., 2010; Peluffo et al., 2012). Amphiregulin has several roles, including increasing blastocyst rates and blastomere numbers (Procházka et al., 2011) and regulating the translation of maternal mRNA (Chen et al., 2013), which lays the foundation for early embryonic development (Luong et al., 2020). Amphiregulin supplementation in culture media had also shown promise for rescue IVM (Ben-Ami et al., 2011).

To evaluate the essential role of amphiregulin, which has so far been part of the meiotic trigger in CAPA-IVM (Sánchez et al., 2017; Vuong et al., 2020), this study aims to explore the effects of amphiregulin supplementation in IVM medium in comparison with the routine meiotic trigger (control), as was previously used in standard IVM. The primary endpoint was oocyte maturation rate; secondary endpoints were embryo quality and IVM outcomes. In addition, relationships between expression of key genes in CCs from <6-mm follicles as well as maturation and fertility outcomes of their corresponding oocytes were determined.

Materials and methods

Study design, allocation and randomization

This prospective, randomized cohort study was conducted at IVFMD, My Duc Hospital, Ho Chi Minh City, Vietnam from 17 April 2019 to 6 December 2019. The study was approved by the institutional ethics committee and conducted according to Good Clinical Practice and Declaration of Helsinki 2002 principles. All participants provided written informed consent. The full study protocol can be accessed at www.clinicaltrials.gov (NCT03915054).

Study flow chart is depicted in Fig. 1. In the first stage of the study, sibling oocytes from each patient were matured in medium with (CAPA-AREG) or without (CAPA-Control) amphiregulin 100 ng/ml. An equivalent number of COCs (from follicles <6 and ≥6 mm in diameter) were allocated to the CAPA-AREG and CAPA-Control groups. After the maturation check and fertilization using intracytoplasmic sperm injection (ICSI), all good quality Day 3 embryos were vitrified. On return for frozen embryo transfer (second stage), patients were randomized (1:1) to have either CAPA-AREG or CAPA-Control embryo(s) implanted (≤2 per patient). Where no embryo(s) from the randomized group were available, embryo(s) from the other group were transferred.

Figure 1.

Study flow chart. AREG, amphiregulin; CAPA, capacitation; COCs, cumulus–oocyte complexes; ICSI, intracytoplasmic sperm injection; IVM, in vitro maturation.

Study population

Women with polycystic ovary syndrome (PCOS) diagnosed according to Rotterdam criteria (≥24 antral follicles in both ovaries) and/or increased ovarian volume (>10 ml) in at least one ovary, who had an indication for ART, had undergone ≤2 previous IVM or IVF attempts, and agreed to have ≤2 embryos transferred, were eligible for inclusion. Oocyte donation and pre-implantation genetic diagnosis cycles, cases of high-grade endometriosis and cases with extremely poor sperm quality were excluded.

Interventions and assessments

At the first visit, patients were evaluated for eligibility and received information about the study. Patients were invited to return to the clinic on Day 2–3 of their menstrual cycle (amenorrheic women were treated with oral contraceptives for 2 weeks to induce menstrual bleeding). All patients had an ultrasound scan of the ovaries and endometrium, and serum hormone levels were determined to evaluate if they were basal.

All patients received highly purified human menopausal gonadotropin (HP-hMG) 150 IU/day (Menopur, Ferring, Switzerland) for 2 days from the third day of the menstrual cycle. Oocyte retrieval was performed 42 h after the last HP-hMG injection. Follicle size was measured before oocyte retrieval by ultrasound scan. Larger follicles (≥6 mm) were punctured first, then the needle was flushed, and smaller follicles (<6 mm) were punctured. COCs from the two different size follicles were collected in separate tubes, then processed and cultured separately to allow the effects of the two meiotic trigger types to be determined based on follicle size. All COCs were immediately washed and placed in the capacitation culture dish. Capacitation medium was prepared by supplementing Medicult IVM medium (Origio, Denmark) with 1 mIU/ml recombinant FSH (rFSH; Puregon, MSD, Australia), 5 ng/ml recombinant human insulin (Merck KGaA, Darmstadt, Germany), 10 nmol/l oestradiol (Sigma, Schnelldorf, Germany), 10 mg/ml human serum albumin (SAGE, Denmark) and 25 nM CNP (Tocris Bioscience, Abingdon, UK).

After 24 h of capacitation culture, COCs were washed and divided equally between two meiotic maturation media, and incubated under oil at 37°C with 6% carbon dioxide (CO₂) in air for an additional 30 h. Medicult IVM media (Origio) was used as the basal IVM culture medium for both types of meiotic maturation trigger. For the control trigger (CAPA-Control), basal IVM media was supplemented with 75 mIU/ml FSH (Gonal-f; Merck Serono, Switzerland), 100 mIU/ml hCG (Pregnyl; MSD, USA), 0.01 mg/ml growth hormone (Saizen; Merck Serono, Switzerland), and 10 mg/ml human serum albumin (SAGE), a formula with proven efficacy in standard IVM (Vuong et al., 2020). The earlier approach has been widely adopted by clinics who offer standard IVM. For the investigational trigger (CAPA-AREG), the basal medium was supplemented with 5 ng/ml insulin, 10 nmol/l oestradiol, 100 mIU/ml rFSH, 10 mg/ml human serum albumin (SAGE), and 100 ng/ml human recombinant amphiregulin (R&D Systems, Minneapolis, USA). This approach has been used in all CAPA-IVM clinical studies to date and the composition of this medium is based on the rationale that low physiological concentrations of oestradiol, insulin, and FSH (such as those in a normal small follicle) preserve cumulus viability during the prematuration culture, and that supra-physiological concentrations of FSH and amphiregulin subsequently drive final meiotic maturation (Sánchez et al., 2017).

Although the meiotic trigger media used in the CAPA-IVM and CAPA-Control groups of our study are not completely equivalent in composition (besides the presence versus absence of amphiregulin), it is recognized that meiosis in the intact COC is mainly driven by FSH (via the highly expressed FSH receptor in CC) while LH maturation effects are mediated by EGF receptor-dependent signalling (Downs and Chen, 2008), which might still be deficient in small/medium follicles (Sugimura et al., 2015). The doses of FSH used for meiotic trigger are 75–100 times higher than those in capacitation media, as the excess increases cyclic adenosine monophosphate to drive meiosis. The amphiregulin doses used for the trigger were based on data from preclinical studies, where doses of 10 to >100 ng/ml showed effects on maturation and embryo development (Procházka et al., 2011; Peluffo et al., 2012). These doses correspond to the physiological levels observed in FF after conventional IVF which correlated with a favourable outcome (Zamah et al., 2010; Cakmak et al.,2016).

Oocytes were fertilized using ICSI and cultured in 30 µl Global Total LP (Life Global, Canada) at 37°C with 6% CO₂ and 5% oxygen. A fertilization check was performed 16–18 h after ICSI. Embryo grading was done by two independent embryologists at 68 ± 1 h after fertilization based on the Istanbul consensus (Alpha Scientists in Reproductive Medicine and ESHRE Special Interest Group of Embryology, 2011). All transferable Day 3 embryos were vitrified using Cryotech (Japan).

Starting on the day of embryo freezing, patients received oral progesterone (Duphaston 10 mg; Abbott, Singapore) for 10 days to induce bleeding. In the subsequent cycle, endometrial preparation was performed using 2 mg oral oestradiol valerate (Valiera; Abbott) four times daily from Day 2 of the menstrual cycle. Progesterone (Cyclogest; Accord-UK Limited, UK) 200 mg intravaginally four times daily was started when patients had received oestradiol valerate for ≥10 days and when endometrial thickness was ≥8 mm.

Transfer of a maximum of two embryos was scheduled for 3 days after the start of progesterone administration. Based on the randomized group, patients had transfer of thawed embryo(s) that had been matured in either CAPA-AREG or CAPA-Control medium. Where no embryo(s) from the randomized group was available, embryo(s) from the other group were transferred. Embryo transfer was performed 2 h after thawing as per routine clinical practice at My Duc Hospital. The first pregnancy test was performed 14 days after embryo transfer; a positive pregnancy test was defined as serum beta hCG >5 mIU/ml. Ultrasound was performed 3 weeks after pregnancy testing.

Outcomes

The primary endpoint was meiotic maturation efficiency (proportion of metaphase II (MII) oocytes; maturation rate). A related exploratory endpoint was oocyte developmental outcomes in subgroups based on follicle diameter at COC retrieval (<6 mm and ≥6 mm).

A number of predefined secondary outcomes were also evaluated: positive pregnancy test rate; implantation rate; clinical pregnancy rate; ongoing pregnancy after the first frozen embryo transfer of the started treatment cycle; number of embryos on Day 3; number of transferable embryos on Day 3; number of good quality embryos on Day 3 and live birth rate.

Implantation was defined as the number of gestational sacs per number of embryos transferred at 3 weeks after frozen embryo transfer. Clinical pregnancy was defined as at least one gestational sac on ultrasound at 7 weeks of gestation with the detection of heartbeat activity. Ectopic pregnancy was defined as a pregnancy in which implantation occurred outside the uterine cavity. Miscarriage was defined as pregnancy loss at <12 weeks of gestation. Multiple pregnancies were confirmed when there was more than one sac at early pregnancy ultrasound (6–8 weeks of gestation). Ongoing pregnancy was defined as pregnancy with a detectable heart rate at ≥12 weeks of gestation after the completion of the first transfer. Live birth was defined as the birth of at least one newborn after 24 weeks of gestation and exhibiting any sign of life; twins were counted as a single birth.

Cumulus cell collection

CCs were removed from oocytes at 30 h after IVM using short exposure to hyaluronidase (Life Global, Canada). Cells were briefly washed in micro droplets with medium and snap frozen in liquid nitrogen as previously described (Guzman et al., 2013). For each group, the oocyte maturation rate and embryo development were followed, allowing analysis of the relationship between gene expression and outcome parameters (Guzman et al., 2013).

RNA extraction and reverse transcription

Total RNA was extracted using TRIzol reagent (Thermo Fisher Scientific, Belgium). Briefly, following sample homogenization with TRIzol and subsequent phase separation with chloroform (Merck, Belgium), RNA precipitation was achieved using 2-propanol and glycogen (both from Sigma-Aldrich, Belgium). Following the final wash with 70% ethanol (Sigma-Aldrich), pellets were air dried. RNA was dissolved in nuclease-free water and quantification was done with NanoDrop ND-1000 (ThermoFisher). Reverse transcription of the samples was completed using iScript cDNA Synthesis Kit (Bio-Rad Laboratories, Belgium) following the manufacturer’s instructions.

Quantitative reverse transcription polymerase chain reaction

All qRT-PCR analyses were performed on LightCycler 480 (384-well plate format; Roche, Belgium). The reaction mix included 5 μl SYBR Green I Master 2× (Roche), 1 μl primer mix and 2 μl cDNA. The qRT-PCR program consisted of a pre-incubation step (95°C for 10 min) followed by 55 amplification cycles (95°C for 10 s and 60°C for 30 s). Specificity of the amplification was checked using the melting curve analysis. In total, expression levels of 27 target genes belonging to several different pathways and functions were analysed in the CCs (sequences provided in Supplementary Table S1). UBC and RPL13A were used as reference genes for normalization of the expression of target genes. The 2^ΔΔCt method was used for the calculation of the fold changes. Only CCs obtained from follicles <6 mm in diameter were included in the gene expression analysis because not every patient had oocytes collected from ≥6-mm follicles.

Sample size calculation

Prior to the study, the maturation rate in standard IVM was 57% at IVFMD, My Duc Hospital (Vuong et al., 2020). To assess whether the amphiregulin trigger would increase the maturation rate by 12% over standard trigger in standard IVM, we needed to randomize 670 oocytes (335 per arm; 90% power, two-sided alpha of 0.05). Given that the mean number of oocytes retrieved per patient is around 22, we needed to recruit 30 patients.

Statistical analysis

Between-group differences in primary and secondary endpoints were analysed using nonparametric methods for paired samples or Fisher’s exact test (categorical variables), and were reported as relative risks and 95% confidence intervals. For gene expression analysis, fold change values were normalized through log2 transformation and differences between conditions were determined using the paired t-test (two-tailed, 95% confidence interval). Correlation analyses were performed using the Spearman correlation test (two-tailed, 95% confidence interval) to detect possible relationships between gene expression and IVM outcomes. Genes that did not exhibit any difference between CAPA-Control and CAPA-AREG were selected for the correlation analysis to allow identification of CC markers of competent IVM oocytes. This ensured that the correlation analysis reveals the true relations without being affected by the culture conditions. For all statistical tests, a P-value of <0.05 was considered statistically significant. Statistical analyses were performed using GraphPad Prism (GraphPad Software, La Jolla, CA, USA, www.graphpad.com) or R statistical program (v3.5.0, www.cran.r-project.org).

Results

Study population

Thirty women aged 18–37 (mean 29.1 ± 2.9) years were included in the study (Table I). Overall, patients had normal mean body mass index (BMI) and elevated anti-Müllerian hormone; the majority had primary infertility, polycystic ovarian morphology with oligomenorrhea/amenorrhea, and had undergone one previous ART attempt (Table I).

Table I

Patient characteristics and cycle parameters at baseline, and a summary of stimulation characteristics.

Characteristic	Patients (n = 30)
Age, years	29.1 ± 2.9
Body mass index, kg/m2	23.3 ± 3.68
Anti-Müllerian hormone, ng/ml	10.5 ± 4.3
Duration of infertility, years	2 [2, 5]
Type of infertility, n (%)
Primary	22 (73.3)
Secondary	8 (26.7)
Number ART attempts, n, (%)
1	29 (96.7)
2	1 (3.3)
Diagnosis, n, (%)
PCO morphology + normal menstrual cycle length (Group 1)	7 (23.3)
PCO morphology + oligomenorrhea/amenorrhea (Group 2)	23 (76.7)
Hormonal profile at first visit
Testosterone (nmol/l)	1.29 [0.72, 1.81]
SHBG (nmol/l)	41.01 [30.33, 50.38]
TSH (µIU/ml)	2.04 [1.50, 3.03]
DHEA-SO4 (µg/ml)	1.98 [1.60, 2.64]
17-hydroxyprogesterone (ng/ml)	1.29 [1.06, 1.61]
Days on gonadotropins	2
Total dose of gonadotropin used, IU	300
Number of follicles at last ultrasound	29 [24, 35]
Total number of COCs, n	18.5 [16.3, 25.8]
Endometrial thickness on day of oocyte retrieval (mm)	5.7 ± 1.2

Values are mean ± standard deviation, median [interquartile range], or number of patients (%).

ART, assisted reproductive technology; COCs, cumulus–oocyte complexes; DHEA-SO4, dehydroepiandrosterone sulphate; PCO, polycystic ovary; SHBG, sex hormone-binding globulin; TSH, thyroid-stimulating hormone.

In vitro maturation and embryology outcomes

In the by-patient analysis, the only statistically significant difference between the CAPA-AREG and CAPA-Control groups was in the number of MII oocytes, which was significantly higher in the CAPA-AREG versus CAPA-Control group (Table II). Similarly, when the outcomes were analysed for each COC, the proportion that reached the MII stage was significantly higher in the CAPA-AREG versus CAPA-Control group, with no other statistically significant between-group differences (Table III). The findings were again similar in the analysis by follicle size, with the proportion of COCs reaching the MII stage being significantly higher for follicles of <6 or ≥6 mm in diameter (Table IV). Overall, COCs from larger follicles (≥6 mm) had high maturation, fertilization and normal development rates (Table IV).

Table II

In vitro maturation and embryology outcomes (per patient).

	Sibling oocytes from 30 patients		P-value
	CAPA-AREG	CAPA-Control
Total number of COCs, n	10.0 [8.0, 13.8]	10.0 [8.0, 13.8]	0.53
Total number of MII oocytes, n	7.0 [5.3, 8.0]	6.0 [4.0, 7.0]	0.01
Total number of pronuclear stage, n	5.0 [4.0, 5.8]	4.0 [2.0, 5.0]	0.15
Total number of Day 3 embryos, n	3.0 [2.0, 4.0]	2.0 [1.3, 3.8]	0.27
No embryo on Day 3, n (%)	3 (10.0)	1 (3.3)	0.61
Total number of grade 1 and 2 embryos, n	2.0 [1.0, 3.0]	1.5 [0.0, 2.8]	0.14
Total number of transferrable embryos, n	3.0 [2.0, 4.0]	2.0 [1.3, 3.8]	0.16

Values are median [interquartile range] or number of patients (%). P-value Wilcoxon test (paired samples).

CAPA, capacitation; COCs, cumulus–oocyte complexes.

Table III

In vitro maturation and embryology outcomes (per cumulus–oocyte complex).

	CAPA-AREG	CAPA-Control	RR (95% CI)	P-value
	(n = 336)	(n = 335)
Maturation rate (MII), n (%)	227/336 (67.6)	185/335 (55.2)	1.22 (1.08, 1.38)	0.001
Pronuclear stage rate, n (%)	148/207 (71.5)	130/180 (72.2)	0.99 (0.87, 1.12)	0.91
Day 3 embryos rate, n (%)	96/148 (64.9)	81/130 (62.3)	1.04 (0.87, 1.24)	0.71
EQ1 + EQ2 rate, n (%)	66/148 (44.6)	49/130 (37.7)	1.18 (0.89, 1.57)	0.27
Transferrable embryos rate, n (%)	90/148 (60.8)	73/130 (56.1)	1.08 (0.89, 1.32)	0.47

CAPA, capacitation; CI, confidence interval; RR, relative risk.

Table IV

In vitro maturation and embryology outcomes in follicles with diameter <6 or ≥6 mm.

Oocyte developmental outcomes, n (%)	COCs <6 mm			COCs ≥6 mm
	CAPA-AREG	CAPA-Control	P-value	CAPA-AREG	CAPA-Control	P-value
	(n = 314)	(n = 300)		(n = 22)	(n = 35)
Maturation rate (MII)	210/314 (66.9)	172/300 (57.3)	0.02	17/22 (77.3)	13/35 (37.1)	0.006
Pronuclear stage rate	135/191 (70.7)	121/168 (72.0)	0.82	13/16 (81.3)	9/12 (75.0)	0.99
Day 3 embryos rate	87/135 (64.4)	76/121 (62.8)	0.80	9/13 (69.2)	5/9 (55.6)	0.66
EQ1 + EQ2 rate	60/135 (44.4)	44/121 (36.4)	0.20	6/13 (46.2)	5/9 (55.6)	0.99
Transferrable embryos rate	81/135 (60.0)	68/121 (56.2)	0.61	9/13 (69.2)	5/9 (55.6)	0.66

CAPA, capacitation; COCs, cumulus–oocyte complexes; EQ1, grade 1 embryos; EQ2, grade 2 embryos.

Fertility and obstetric outcomes

Of the 29 randomized patients in the second stage of the study, the majority had two embryos transferred. All patients had embryos from the combination of both triggers. Although the rates of clinical pregnancy, ongoing pregnancy and live birth were numerically higher in the patients who had CAPA-AREG versus CAPA-Control embryos implanted, all of the fertility and obstetric outcomes did not differ significantly between the groups (Table V).

Table V

Fertility and obstetric outcomes.

	CAPA-AREG	CAPA-Control	RR (95% CI)	P-value
	(n = 15)	(n = 14)
Number of embryos transferred, n (%)				0.99
1	2 (13.3)	1 (7.1)
2	13 (86.7)	13 (92.9)
Number of EQ1 + EQ2 transferred, n (%)				0.76
0	1 (6.7)	2 (14.3)
1	3 (20.0)	2 (14.3)
2	11 (73.3)	10 (73.3)
Positive beta hCG, n (%)	10 (66.7)	6 (42.9)	1.56 (0.77, 3.14)	0.272
Implantation, n	46.7 ± 39.9	28.6 ± 37.8	–	0.222
Clinical pregnancy, n (%)	10 (66.7)	6 (42.9)	1.56 (0.77, 3.14)	0.272
Miscarriage (before 12 weeks), n (%)	2 (13.3)	2 (14.3)	0.93 (0.15, 5.76)	0.99
Ectopic pregnancy, n (%)	0 (0)	0 (0)	–	–
Ongoing pregnancy, n (%)	8 (53.3)	4 (28.6)	1.87 (0.72, 4.85)	0.264
Miscarriage (after 12 weeks), n (%)	1 (6.7)	0 (0)	–	–
Live birth, n (%)	7 (46.7)	4 (28.6)	1.63 (0.61, 4.39)	0.45
Singleton	6 (85.7)	3 (75.0)
Twin	1 (14.3)	1 (25.0)
Birth weight, grams
Singleton	3100.0 ± 442.7	3200.0 ± 529.2	–	0.79
Twin*	2700/2500	3000/2600	–

CAPA, capacitation; CI, confidence interval; EQ1, grade 1 embryos; EQ2, grade 2 embryos; hCG, human chorionic gonadotropin; RR, relative risk.

*

One case each group had twin delivery.

Gene expression profiles

Although RNA was extracted from all 30 patients, due to cDNA volume limits reflected by the low RNA concentrations, the final paired analysis was performed with 25 patients for 13 genes, 20–24 patients for 2 genes, 16–19 patients for 11 genes and 14 patients for 1 gene. The expressions of several genes belonging to the ovulatory cascade including DUSP16 (P = 0.0149), EGFR (P < 0.0001), HAS2 (P = 0.0168), PTGER2 (P = 0.0002), PTGS2 (P = 0.001) and RPS6KA2 (P = 0.0241) were significantly higher in the CCs collected from the CAPA-Control group (Fig. 2A). Conversely, LHR expression was significantly higher in CCs from the CAPA-AREG versus CAPA-Control group (P < 0.00001), and there was no between-group difference in the expression of AREG and VCAN mRNA (Fig. 2A).

Figure 2.

Cumulus cell gene expression analysis. Cumulus cells were collected at the end of in vitro maturation, and 27 genes involved in the ovulatory signalling cascade (A), steroidogenesis (B), glucose and redox metabolism (C), transcription factors (D) and miscellaneous functions (E) were analysed. Paired t-test was performed on the log2 transformed fold change values. Data are presented as mean ± 95% confidence interval. *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001 for between-group comparisons.

For genes involved in steroidogenesis, expression of 3BHSD and STAR were significantly higher in the CAPA-AREG compared with the CAPA-Control group (P < 0.0001 and P < 0.0055, respectively), while no differences were observed in PR expression (Fig. 2B).

Genes involved in glucose and redox metabolism showed significantly lower expression of SLC2A1 (P = 0.0003) and SOD2 (P < 0.0001), and significantly higher expression of CAT (P < 0.0001), GPX1 (P = 0.0029) and LDHA (P = 0.0491) in the CAPA-AREG group versus the CAPA-Control group (Fig. 2C).

Of the transcription factors studied, only the expression of NRF2 was significantly different between the two groups, being higher in the CAPA-Control versus the CAPA-AREG group (P = 0.0017; Fig. 2D). For the remaining genes analysed, only expression of CDC42 mRNA was significantly higher in the CAPA-AREG versus the CAPA-Control group (P = 0.0298; Fig. 2E).

Correlation between gene expression patterns and IVM outcome

There was a significant negative association between expression of JUN and the total number of embryos per mature oocyte (P = 0.0204) and the number of good quality embryos per COC (P = 0.0325; Fig. 3A and D). Similarly, VCAN expression was negatively correlated with the number of good quality embryos per mature oocyte (P = 0.0258; Fig. 3F). TRPM7 expression was significantly and positively correlated with the total number of embryos both per MII and fertilized (2PN) oocyte (P = 0.0053 and P = 0.0232, respectively; Fig. 3B and C), and with the number of good quality embryos per MII oocyte (P = 0.0413; Fig. 3E).

Figure 3.

Cumulus cell gene expression correlation with embryology outcome. Correlation analysis was performed using Spearman correlation test (two-tailed, 95% confidence interval) to detect possible relationships between gene expression and in vitro maturation outcomes. P-value <0.05 indicates a significant correlation. 2PN, fertilized oocyte; COC, cumulus–oocyte complex; GQE, good quality embryo; MII, mature oocyte.

Discussion

In this study of women with PCOS undergoing IVM, a meiosis trigger including amphiregulin significantly improved oocyte maturation rates from follicles both <6 and ≥6 mm in diameter; hence, the study met its primary endpoint. Subsequent frozen transfer of embryos from the oocytes matured in vitro with amphiregulin tended to result in higher clinical pregnancy, ongoing pregnancy and live birth rates, but differences compared with the controls did not achieve statistical significance because the study was not powered for secondary endpoints.

This exploratory randomized trial used a sibling oocyte setup that provides a sensitive model to show the effects of amphiregulin while controlling for inter-patient variability in oocyte maturation and developmental potential. Similar to prior preclinical and clinical studies, amphiregulin had a significant positive effect on maturation (Zamah et al., 2010; Ben-Ami et al., 2011; Procházka et al., 2011; Cakmak et al., 2016). COCs retrieved from follicles ≥6 mm in diameter in this study showed high maturation and fertilization potential, with a similar potential to form good quality embryos (per fertilized oocyte) as those of <6 mm in diameter incubated with amphiregulin.

Embryos from both the CAPA-AREG and CAPA-Control groups showed respectable implantation rates after vitrification and transfer in artificial cycles (46.7% and 28.6%, respectively). Eight of 15 patients who had transfer of embryos from the CAPA-AREG group had an ongoing pregnancy, while 4 of 14 who had transfer of embryos from the CAPA-Control group achieved an ongoing pregnancy.

Our previous experience using hCG-free IVM in patients with PCOS showed that the majority of COCs obtained originated from the smallest follicles (diameter of 2–6 mm; Guzman et al., 2013; Sánchez et al., 2019). This was also the case in the current study, where nearly 95% of COCs came from follicles <6 mm in diameter.

The COCs retrieved from these small-sized follicles contain GV oocytes enclosed in a dense cumulus mass, which need to reside for some time (24–48 h) in a capacitation medium to increase their developmental competence (Sánchez et al., 2017, 2019). The prematuration culture strategy has been pioneered in a variety of animal models using different invitro set-ups, as summarized by Gilchrist et al. (2016). Maturation results from minimally stimulated hCG-free IVM for follicles of <6 mm in diameter obtained from patients with PCOS and observed after different time-points between 28 and 48 h were reported to generally yield no >50% polar body (PB) oocytes after meiosis trigger with gonadotrophins (FSH, LH, hCG) (Guzman et al., 2012; Lin et al., 2020). However, several controlled studies reported that culturing these COCs in a prior capacitation step uniformly increased the maturation rate to PB by 10–20% (relative increase of 30–40%; Sánchez et al., 2017, 2019; Vuong et al., 2020). As explained in the Introduction, capacitation culture is dependent on the addition of physiological concentrations of CNP, oestradiol, FSH and insulin (Romero et al., 2016; Sánchez et al., 2017).

Data from a porcine model showed that using amphiregulin and epiregulin (without FSH) to stimulate meiosis did not result in more MII oocytes (unlike here), but produced significantly more blastocysts (Procházka et al., 2011). Given that EGF-like peptides were shown not to mimic all effects of FSH on cultured COCs in Procházka’s model, we decided to complement the amphiregulin meiotic stimulus with FSH. Nevertheless, another study showed that normal piglets were born from COC treated with amphiregulin only (Akaki et al., 2009), showing its potential to induce full oocyte developmental competence. Our study provides the first data on the effects of amphiregulin on COCs from small human follicles.

In addition to investigating the effects of amphiregulin during meiotic maturation on embryo maturation and quality, this study also included a molecular exploration of a series of candidate genes on the stored cumulus complements from oocytes from <6-mm follicles. Knowing that most of the transcriptional activity of the oocyte occurs during the meiotic arrest before GV breakdown (Hyttel et al., 2001), evaluating gene expression at 30 h after the start of the meiotic trigger could reveal correlates of oocyte quality. Encouraged by this, several important functions were studied in the CCs to understand the potential roles of amphiregulin in CAPA-IVM. Genes involved in the ovulatory cascade, steroidogenesis and redox metabolism were found to be differentially expressed between the groups (Fig. 4).

Figure 4.

Summary of the studied pathways and the results. FSH triggers the LHR to produce AREG. Similarly, amphiregulin in the culture media stimulates both LHR and EGFR for the activation of downstream pathways. At the same time, amphiregulin becomes a part of the loop and stimulates its own production. Amphiregulin in the culture media regulates the expressions of genes involved in cumulus expansion, progesterone production and redox metabolism. It also regulates temporal activation of maternal mRNA translation. Straight lines indicate direct stimulations on the receptors by the given ligands. Dashed lines indicate actions resulting by the stimulations of the receptors. AREG, amphiregulin; EGFR, epithelial growth factor receptor; FSH, Follice stimulating hormone; LHR, luteinising hormone receptor.

Amphiregulin is a critical mediator of the LH-induced EGF network and, upon binding to EGFR on CCs, it stimulates the dynamic expression of several downstream genes for COC expansion, oocyte meiotic maturation, ovulation and luteinization (Richani and Gilchrist, 2018). PTGS2 and HAS2 have been identified as fast-response genes in CCs, with rapidly peaking and gradually decreasing expression patterns following the ovulatory stimulus (Shimada et al., 2006; Adriaenssens et al., 2011; Procházka et al., 2011). Given the consistently lower expression of several genes downstream of EGFR in CCs, including PTGS2 and HAS2 from the CAPA-AREG group, we propose that direct addition of amphiregulin to the culture media activates the ovulatory signalling cascade more quickly, shunting peak expression to an earlier timepoint. The constant abundant availability of amphiregulin would also compensate for heterogenous constitutions of COCs. Similar to previous studies (Ritter et al., 2015), a negative correlation between EGFR mRNA levels and cumulus expansion was observed in this study. The presence of LHR in CCs from smaller follicles has already been confirmed in PCOS patients, with an up-regulated pattern during IVM (Guzman et al., 2013), and levels were correlated with maturation and blastocyst rates (Yang et al., 2005). Likewise, high CC LHR expression in the CAPA-AREG group was associated with high maturation levels.

Progesterone production through the actions of 3BHSD and STAR is classically attributed to mural granulosa cells (Miller and Auchus, 2011) and increased CC progesterone synthesis has been related to distorted functionality due to the invitro culture. However, transcriptomic studies showed that post-ovulatory CCs perform progesterone synthesis by default (Chaffin et al., 2012). Consistent observations were derived from porcine IVM when CC expansion was associated with progesterone production by CCs, which was also directly proportional to oocyte competence (Grupen and Armstrong, 2010). Moreover, an inverse correlation between the ratio of apoptotic CCs and progesterone production implicates the hormone as an indirect measure of CC health (Grupen and Armstrong, 2010; Cajas et al., 2020). In the current study, expressions of both 3BHSD and STAR were significantly higher in CCs from the CAPA-AREG group. In addition, higher expression of CDC42, a small GTPase of the Rho family that plays a regulatory role in cell apoptosis (Tu and Cerione, 2001), in the same group might support the concept that progesterone has anti-apoptotic functions on CCs. Based on available data, improved maturation rates in the current study could be attributed to amphiregulin-stimulated production of progesterone by CCs, leading to higher stamina in these cells, enabling them to be better companions for the corresponding oocytes. Further precise correlations could be obtained by measuring the progesterone output and apoptotic signals of CCs.

PCOS is a multifactorial disorder that includes a metabolic component which contributes to CC malfunction, leading to imbalanced redox potential and increased oxidative stress in these cells (Zhao et al., 2015). In fact, infertile women with ovulatory dysfunction show higher superoxide dismutase activity in CCs compared with those with male factor or tubal factor infertility (Matos et al., 2009). Furthermore, PCOS patients with high oxidative stress markers in CCs also have low ongoing pregnancy rates (Zhao et al., 2015). Reactive oxygen species (ROS) homeostasis is required to prevent oxidative stress (Riley and Behrman, 1991; Agarwal et al., 2006; Combelles et al., 2009), which can only be achieved through the support and protection provided by CCs (von Mengden et al., 2020). An animal model of IVM involving antioxidant showed higher cleavage and blastocyst rates, along with reduced ROS levels and downregulated SOD2, both in oocytes and CCs (Cajas et al., 2020). Thus, lower SOD2 expression in CCs from the CAPA-AREG group, together with higher expression of GPX1 and CAT, might suggest a role for amphiregulin in reducing oxidative stress by inducing ROS homeostasis and antioxidant capacity during IVM.

CCs divert media glucose to glycolysis and/or lactic acid fermentation, generating byproducts such as pyruvate and lactate, for oocytes during IVM (Biggers et al., 1967; Clark et al., 2006; Akin et al., 2021). Dumollard et al. showed that lactate supplementation stimulated antioxidant defense in mouse oocytes (Dumollard et al., 2007). Our data showed higher CC LDHA levels in the CAPA-AREG group compared with the CAPA-Control group. Together with the higher SLC2A1 expression in CAPA-Control CCs, reflecting high glucose intake by CCs (Uldry and Thorens, 2004), this could be attributed to an oocyte’s attempt to boost lactate levels through increased glycolysis, which would possibly improve antioxidant defenses. Similarly, higher levels of NRF2, an important transcription factor known to activate transcription of several antioxidant genes to maintain the redox homeostasis (Ma, 2013), might possibly be an attempt to compensate for the poor antioxidant capacity of CCs.

The relation between CC gene expression and ART outcome has been studied extensively (McKenzie et al., 2004; Feuerstein et al., 2007; van Montfoort et al., 2008; Anderson et al., 2009; Assidi et al., 2011; Gebhardt et al., 2011). While these biomarkers are valid for CCs collected after conventional IVF stimulation with a meiotic trigger by a high hCG dose, their validity for IVM CCs still needs confirmation. Follicles of IVM oocytes are generally <10 mm in diameter without a prior hCG trigger, and thus CC gene expression after IVM culture follows a unique pattern (Brown et al., 2017). CC expressions of TRPM7 and VCAN have been previously correlated to the best Day-3 embryos after ICSI (positive correlation; Wathlet et al., 2011, 2012) and good quality embryos (negative correlation; Wathlet et al., 2011). Data from the current study confirm the significance of TRPM7 as a marker of oocyte quality in IVM CCs because it was positively correlated with both good quality embryo rates (per fertilized oocyte) and total embryo rates (per mature and fertilized oocyte). The negative correlation between CC VCAN expression and good quality embryo morphology was preserved in IVM CCs. Furthermore, a negative relation between CC JUN expression and good quality embryo rate was found. JUN is a member of AP-1 transcription factor family and inhibits aromatase expression in granulosa cells (Ghosh et al., 2005). While the exact role in the CCs is unknown, Ghosh et al. (2005) did not exclude the possibility that JUN might be active in other parts of steroidogenesis. Nevertheless, additional studies are needed to reveal the exact mechanism of JUN in oocyte maturation and competency.

Key strengths of this study are the insights into the two follicular diameter compartments, the sibling oocyte design (which facilitates better control for inter-patient differences), and the deferred transfer (which allows a focus on the evaluation of embryo potential without the influence of endometrial factors). However, the study also had its limitations. The sample size was calculated based on the maturation rate per oocyte, meaning that the study was underpowered to detect between-group differences in secondary fertility outcomes. Furthermore, the patients included were all of Vietnamese origin and had a normal BMI despite the presence of PCOS, which limits the generalizability of the findings to other ethnicities. Another limitation is that there were differences in the meiotic maturation media (second culture step) other than amphiregulin (for reasons of clinical acceptance) that might have influenced some gene profiles. Also, the use of hyaluronidase (which is standard of care before ICSI) might have had an influence on the studied genes. Finally, the exploratory study was performed using CCs from COCs that had been cultured in groups, which reduced the power of our findings. Follow-up investigation of the potential of the embryos that were frozen might strengthen our findings.

Conclusions

These findings highlight the important role of amphiregulin in oocyte maturation. Although numerical differences between groups did favour CAPA-AREG over CAPA-Control, these benefits were not translated into significantly better fertility outcomes. Gene expression analysis showed a number of potential mechanisms that would explain a beneficial role for amphiregulin in oocyte maturation and development, including activation of the ovulatory cascade sooner, advanced redox homeostasis and an appropriate steroidogenesis profile. Moreover, expression of three genes showed a significant correlation with embryo quality and therefore may have potential as biomarkers in IVM. However, larger studies are needed to further investigate and validate the current findings.

Supplementary data

Supplementary data are available at Human Reproduction online.

Data availability

The data underlying this article are not publicly available but are available from the corresponding author on reasonable request.

Acknowledgements

Mrs Heidi Van Ranst, senior research embryologist at Follicle Biology Research Laboratory (FOBI-VUB), is greatly acknowledged for preparing the ingredients of CAPA-IVM and for arranging the transport of materials under appropriate conditions.

Authors’ roles

L.N.V., J.S. and T.M.H. designed the clinical study and N.A., J.S., S.R. and E.A. designed the molecular analysis. The monitoring of the clinical data collection was done by M.H.N.N., A.H.L., T.D.P., L.N.V., U.D.T.H. and F.S. The molecular analysis from sample selection to final result was done by N.A. The statistical analysis of clinical data was done by T.D.P. and L.N.V.; N.A. did the data reduction and statistical evaluation of the molecular analyses. The draft of the study was written by N.A., J.S. and L.N.V. All authors were involved in the decision to publish the paper and in the revisions of the manuscript. L.N.V. and J.S. act as guarantor and accept the full responsibility for the work and controlled the decision to publish.

Funding

FWO Flanders (grant number FWO G.OD97.18N) and the Vietnam National Foundation for Science and Technology Development (NAFOSTED) under (grant number FWO.106-YS.2017.02). The study sponsors were not involved in the study design, the collection, analysis and interpretation of the data, the writing of the publication, or in the decision to submit the paper for publication. L.N.V. and J.S. had full access to the data and took the final decision to submit for publication.

Conflict of interest

L.N.V. has received speaker and conference fees from Merck, grant, speaker and conference fees from Merck Sharpe and Dohme and speaker, conference and scientific board fees from Ferring; T.M.H. has received speaker fees from Merck, Merck Sharp and Dohme and Ferring. J.S. reports lecture fees from Ferring Pharmaceuticals and Biomérieux Diagnostics and grants from FWO Flanders, is co-inventor on granted patents on CAPA-IVM methodologies in USA (US10392601B2), Europe (EP3234112B1) and Japan (JP 6806683 registered 08-12-2020) and is a co-shareholder of Lavima Fertility Inc., a spin-off Company of the Vrije Universiteit Brussel (VUB, Brussels, Belgium). N.A., T.D.P., A.H.L., F.S., S.R., M.N.H.N., E.A. and U.D.T.H. have no financial relationships with any organizations that might have an interest in the submitted work in the previous three years, and no other relationships or activities that could appear to have influenced the submitted work.