Improving the maturation rate of human oocytes collected ex vivo during the cryopreservation of ovarian tissue

Abstract

Purpose

The aim of the present study was to improve the in vitro maturation (IVM) procedure using oocytes from surplus ovarian tissue after fertility preservation.

Methods

Twenty-five patients aged 17–37 years were included in the study. Maturation was compared between oocytes collected in HEPES-buffered medium or saline, and we determined whether transport on ice prior to oocyte collection affected maturation. Two different IVM media were used that were supplemented with and without recombinant human midkine. Mature oocytes were assessed for aneuploidy using next-generation sequencing (NGS).

Results

On average, 36 immature oocytes were collected from each patient (range 7–90, N = 895). Oocytes recovered from HEPES-buffered medium matured at a higher rate than oocytes recovered from saline (36% vs 26%, p < 0.01). Ovarian transportation on ice prior to the procedure negatively affected maturation compared with non-transported samples (42% vs 27%, p < 0.01). The addition of midkine improved maturation rate (34% vs 27%, p < 0.05). On average, 11 MII oocytes were obtained per patient (range 1–30). NGS of 53 MII oocytes and their first polar bodies indicated that 64% were euploid.

Conclusions

The study demonstrated unexpectedly high number of immature oocytes collected from surplus ovarian tissue without any stimulation. The overall MII rate was one in three, resulting in a total number of MII oocytes that was similar to the number obtained after ovarian stimulation. If these MII oocytes prove suitable for IVF, they will provide a substantial improvement in fertility preservation for patients and advance IVM as an interesting platform for further improvements in assisted reproduction.

Electronic supplementary material

The online version of this article (10.1007/s10815-020-01724-7) contains supplementary material, which is available to authorized users.

Introduction

Immature human oocytes are not widely used to treat infertility, particularly the immature oocytes collected from small antral follicles with diameters less than 6 mm [1]. In many fertility clinics, non-MII oocytes are discarded, while at some centres, immature oocytes might be employed in different in vitro maturation (IVM) protocols. Researchers have previously attempted to define IVM in terms of the follicular diameter from which oocytes are derived and other conditions, such as necessity of triggering and gonadotropin use prior to oocyte collection [2, 3]. In the present study, IVM was defined as advancement of the nuclear maturation from the germinal vesicle stage to the MII stage with the extrusion of a first polar body during culture.

Since 1991, when the first pregnancy was obtained after in vitro fertilization (IVF) using in vitro matured oocytes, the clinical use of IVM has increased, but at a slow pace. This situation might reflect the technical challenges related to the procedure and the necessity for a different timing compared with normal IVF. The introduction of fertility-preserving techniques in which ovarian tissue is cryopreserved, however, provided a possibility to develop a new IVM model based on oocytes obtained from very small antral follicles (SAF) [4]. In connection with ovarian tissue cryopreservation (OTC), one whole ovary is often surgically excised and the cortical tissue is frozen. The remaining medulla tissue is normally discarded, since it contains the majority of growing follicles, which are too large and their architecture is too complicated to sustain freezing. Importantly, immature oocytes collected from surplus medulla tissue can be matured in vitro without previous hormonal stimulation of the patient. This option is particularly interesting in prepubertal girls and for patients with cancer in whom ovarian stimulation is contraindicated or chemotherapy cannot be delayed [5]. Immature oocytes have been aspirated ex vivo from SAF, which normally exceed a diameter of 3–5 mm and are visible on the ovarian surface [6–8]. These oocytes may be able to increase the fertility potential in addition to the cryopreserved cortex. Three pregnancies have been reported from embryos obtained from IVM of these immature oocytes [9–11].

Small antral follicles in the medulla may span diameters ranging from less than 1 mm to approximately 10 mm, and several studies have attempted to collect all immature oocytes released from SAF during the preparation of ovarian cortical tissue [12–14]. The meiotic and developmental competence of an oocyte is normally considered associated with the diameter of the corresponding follicle [15, 16]. Therefore, oocytes from SAF, with a diameter of only a few millimetres, are likely to display reduced competence compared with follicles with diameters exceeding 10 mm [17, 18]. Nonetheless, OTC now provides new opportunities for optimizing IVM culture conditions and developing fertility treatments based on oocytes obtained from SAF [10]. Different factors might affect the maturation rates of ex vivo collected oocytes, including whether they were collected from ovaries transported on ice prior to the dissection of the tissue, the type of medium from which oocytes were collected after the dissection of ovarian tissue and the addition of chemical substances/growth factors capable of enhancing maturation rates. In a previous study, we collected an average of 11 immature oocytes from the surplus medulla of one ovary from each patient, with an average MII rate of approximately 30% [14]. Importantly, the aneuploidy rate of in vitro matured MII oocytes obtained from SAF is currently unknown.

The aim of the present study was to increase the number of immature oocytes collected from surplus tissue. Furthermore, we focused on increasing the nuclear maturation rate of oocytes from SAF, which was performed by optimizing the oocyte collection procedure and increasing their maturation rates by adding midkine—a growth factor belonging to the heparin-binding growth factor family that was reported to improve IVM in cattle [19]. The importance of midkine has been also confirmed in mice, where it appeared to be a crucial factor for follicular maturation [20]. Although the detailed mechanism of action and signalling pathways activated by midkine are currently unknown, midkine acts indirectly on cumulus cells and triggers the secretion of factors acting on the oocyte [21]. In addition, the morphometric characteristics of oocytes, including the oolemma diameter and zona pellucida thickness, were evaluated. The importance of cumulus cells surrounding the oocyte for sustaining maturation and the aneuploidy rate of the in vitro matured MII oocytes were assessed as well.

Materials and methods

Study design and patients

In a prospective cohort study conducted from December 2017 to March 2018, the ovarian cortex of 26 consecutive patients (mean age 28 years, range 17–37) was cryopreserved at the Laboratory of Reproductive Biology, University Hospital of Copenhagen, Denmark. None of the patients had received any type of exogenous gonadotropins or chemo-/radiotherapy prior to the procedure, and patients’ characteristics are presented in supplementary Table 1. Petri dishes in which the cortical tissue was prepared and containing the remaining ovarian medulla were examined for the presence of immature oocytes. Twenty-five patients were included in the study based on following inclusion criteria: (1) two ovaries were present, and (2) OTC was indicated. One patient was excluded prior to the collection of oocytes due to a diagnosis of imminent premature ovarian insufficiency and thereby having a priori fewer oocytes than an average patient.

The ethical committee of the municipalities of Capital Region of Denmark approved the project (H-2-2011-044), and all the patients signed an informed consent form prior to the surgical procedure, in which they consented to donate their surplus ovarian tissue for research. The assessment of aneuploidy and General Data Protection Regulation (GDPR) were approved by the ethics committee and data protection agency.

Collection of immature oocytes

One entire ovary from each patient was excised surgically through a unilateral oophorectomy and delivered to the laboratory from the local or collaborating hospitals. Ovaries were delivered in either α-MEM (Invitrogen, UK) at 37 °C within 10 min (when the surgical procedure was performed at the local hospital) or in IVF flushing medium (Origio A/S, Denmark) on ice within 2–5 h (when the surgical procedure performed at collaborating hospitals) [22]. Upon arrival, each ovary was cut into two halves and the cortex was isolated using sterile scalpels and forceps, as previously described [23, 24].

The entire procedure for cortex preparation was performed in a flow hood at room temperature using dissection medium at ambient temperature or pre-cooled (4 °C), which corresponded to the temperature during transportation. Since the beginning of the OTC program at our facility, saline was used as the medium to prepare cortical tissue prior to freezing. We compared saline with HEPES-buffered HTF medium (Invitrogen, UK) to test whether results of IVM may be improved by using a medium with an optimal pH during collection.

Immediately after tissue preparation, all petri dishes containing surplus tissue were meticulously examined for immature oocytes. Pieces of medulla tissue were finely dissected once with a tissue chopper (McIlwain Tissue Chopper, Campden Instruments Ltd., UK) to optimize the harvest of immature eggs. Settings for the chopper ensured that tissue was cut into pieces of approximately 2 mm³. Thereafter, pieces of the tissue were rinsed with saline or the HEPES-buffered HTF medium and were also examined for the presence of oocytes. Collection was performed within a period of up to 1 h by two operators simultaneously using a stereomicroscope (Leica MZ12, Germany) on a 37 °C heating table in a flow hood, and petri dishes were evaluated twice by both operators. Immature oocytes were collected in the holding medium consisting of McCoy’s 5α-medium containing 25 mM HEPES (Invitrogen, UK) and supplemented with 5 mg/ml human serum albumin (HSA–CSL Behring 20%, Germany), 2 mM Glutamax (Invitrogen, UK), 0.05 mg/ml penicillin/streptomycin (Invitrogen, UK), 10 μg/ml insulin, 5.5 μg/ml transferrin and 6.7 ng/ml selenium (ITS, Invitrogen, UK) [13, 14]. The immature oocytes (Fig. 1a) were grouped as either cumulus-oocyte complexes (COCs) or naked oocytes (NOs) (Fig. 1b). COCs were further grouped according to the size of the surrounding cumulus mass: small COCs with 3–10 layers of cumulus cells (Fig. 1c) and large COSs with more than 10 layers of cells (Fig. 1d). NOs were either surrounded by small clumps of cumulus cells or completely naked. Oocytes or COCs were classified as having a healthy appearance if they had the following characteristics: an absence of dark coloured cumulus cells, a spherically shaped cytoplasm and integral oolemma. Only oocytes with a healthy appearance were selected for further culture, which corresponded to approximately 95% of all collected oocytes, while oocytes with dark cumulus cells, a dark/irregular ooplasm or mechanical damage were excluded.

Fig. 1.

a Immature oocytes collected from surplus medulla during the preparation of ovarian tissue for cryopreservation. Scale bar = 100 μm. b–d Different categories of immature oocytes: naked immature oocyte (b), small COC (c) and large COC (d). Scale bar = 100 μm. e–h Different stages of oocytes after 2 days of in vitro maturation; MII (e), MI (f), GV (g) and degenerated oocyte (h). Scale bar = 100 μm. DEG degenerated, COC cumulus-oocyte complex

In vitro maturation and evaluation parameters

The MediCult IVM system (Origio A/S, Denmark) was used as previously described [13, 14], with modifications. Briefly, IVM medium was supplemented with 75 mIU/ml rhFSH (Rekovelle, Ferring, Denmark), 100 mIU/ml rhLH (Luveris, Merck Serono, Germany) and 10 mg/ml HSA. After the collection and sub-division of oocytes into the three groups as described above, oocytes from each group were equally divided between the control medium (IVM medium) or experimental group (IVM medium supplemented with 1 μg/ml recombinant human midkine (SRP3114, Sigma, Germany)). Immature oocytes were washed with the respective IVM culture medium, placed individually into 25 μl drops of media and covered with liquid paraffin (Origio A/S, Denmark). The culture dishes were pre-equilibrated overnight prior to oocyte collection in an atmosphere containing 5% CO₂ and ambient O₂ at 37 °C, conditions that were maintained for the subsequent 44–48 h culture period. The expansion of the cumulus in large and small cumulus complexes was evaluated after 24 and 44–48 h under the stereomicroscope. After 44–48 h, all oocytes were denudated by repeated gentle aspiration using denudation glass pipettes with a diameter of 130–133 μm (Vitrolife, Sweden) and data were recorded for each oocyte separately. Denuded oocytes were examined under inverted microscope (Carl Zeiss Axiovert 135, Germany) equipped with Hoffman interference optics with × 20 objective lenses and classified into the germinal vesicle (GV) stage, metaphase-I (MI) or metaphase-II (MII). GV stage oocytes were identified based on their distinct germinal vesicle, refractive nucleolus, darkened centre and granular ooplasm (Fig. 1g). MI oocytes were identified as cells without either a GV or polar body (Fig. 1f), and MII oocytes were identified by the presence of the first polar body (Fig. 1e). Oocytes that remained immature (GV or MI) after 44–48 h were regarded as incapable of maturation.

Photos of all oocytes were captured with the clear oolemma in focus using the Carl Zeiss microscope and the AxioVision software (SE64 Rel.4.9.1). Data were stored for subsequent analyses, which included measurements of the oocyte diameter (ooplasm diameter) and zona pellucida thickness. The values were obtained by calculating the average values of the parameters measured twice perpendicularly for each specific oocyte (Fig. 2).

Fig. 2.

Predicted probability of maturation based on oocyte diameter for combinations of handling and treatments. Punctuated lines: oocytes cultured without midkine. Full lines: oocytes cultured with midkine. Black, HEPES/no transport. Red, saline/no transport. Green, HEPES/transport. Blue, saline/transport. Left bellow, illustration for the measurement of oocyte diameter. Left top, equivalent diameter of which 50% (ED50%) of oocytes collected from surplus ovarian tissue will mature to MII stage. Values with different letters are significantly different (p < 0.05)

Oocyte vitrification

All MII oocytes were vitrified after the IVM procedure in accordance with the protocol described by Lucena et al. [25], with modifications. Briefly, oocytes were washed with three drops of basic solution (BS), which contained 199 medium supplemented with 10% HSA. Then, the last third drop of BS was merged with the first drop of equilibration solution (ES) containing 7.5% (v/v) of dimethyl sulfoxide (DMSO, Sigma, Germany) and 7.5% (v/v) of ethylene glycol (EG, Merck, Germany) in the BS for 3 min. Then, the previous drop was merged with the second ES drop for another 3 min, and finally oocytes were moved into the third ES drop for 6 min. During this process, three drops of vitrification solution (VS) were prepared that contained 15% (v/v) EG, 15% (v/v) DMSO, 0.5 M sucrose (Sigma, Germany) and 10% HSA. After equilibration, oocytes were transferred to each of the three VS drops for 20 s, loaded on the Cryolock carrier device (Biotech, USA), and then plunged into liquid nitrogen. The total time in VS was up to 1 min. The volume of all drops was 30 μl and the procedure was performed at room temperature.

Warming of oocytes after vitrification

Some vitrified oocytes were selected for warming with a subsequent first polar body biopsy and genetic analysis, as described below. Because we were unable to perform this procedure for all vitrified oocytes, we selected oocytes from patients of different ages to cover the range of women included in the study.

One millilitre of thawing solution (TS, which is BS supplemented with 1 M sucrose) was pre-warmed in a central well dish at 37 °C [25]. When the carrier was removed from the liquid nitrogen, the tip of the carrier was quickly immersed in TS for 1 min. Then, oocytes were transferred to 1 ml of dilution solution (DS) containing 0.5 M sucrose in the BS for 3 min and finally transferred to 1 ml of BS for 5 min (the last step was repeated twice). DS and BS were pre-warmed at room temperature for 30 min, and the entire procedure was performed at room temperature.

First polar body biopsy and tubing

Vitrified/warmed oocytes were maintained in a cleavage medium (Origio A/S, Denmark) in individual drops under oil in an incubator with an atmosphere of 5% CO₂ and ambient O₂ at 37 °C for at least 1 h prior to biopsy. Individual oocytes were fixed in a micro-manipulation system (Narishige, Japan) with a holding pipet. A laser (Research Instruments, UK) was employed to dissect the zona pellucida. The polar body was aspirated such that it did not entirely enter the biopsy pipet (Origio A/S, Denmark) and was released in the same drop as the oocyte. Then, the zona pellucida was removed with the further assistance of the laser and biopsy pipet. The biopsied first polar body and the oocyte were placed in separate DNA/RNA-free test tubes with the stripper pipet (Origio A/S, Denmark) in in a volume of approximately 5 μl. Immediately thereafter, test tubes were flash-frozen in liquid nitrogen and stored at − 80 °C prior to further procedures.

Whole genome amplification and next-generation sequencing

To amplify the DNA from the single MII oocyte, we used the SureMDA (Illumina Inc.) according to the manufacturer’s instructions (User guide: 15052710, Revision B; Illumina Inc.) [26]. Briefly working in a pre-PCR area, we centrifuged the tube at 300×g for 3 min., immediately followed by lysis. After collecting the lysis buffer and single oocyte sample, we centrifuged the sample for 5 s at 300×g, followed by the addition of the stop solution and amplification reagents. The DNA was amplified at 30 °C using a reduced amplification time of 2 h. We assessed the quality of the SureMDA product using gel electrophoresis and generated sequencing libraries using a custom Nextera®-based library preparation workflow (Illumina Inc.). MDA products were quantified (Quant-IT™ dsDNA High-Sensitivity Assay Kit, Thermo Fisher Scientific Inc.), and 100 ng were tagmented according to the Nextera® DNA Library Prep Reference Guide (Part: 15027987 v01; Illumina Inc.). Tagmented DNA was cleaned with 1.8× Solid Phase Reversible Immobilization (SPRI) (Part: 15041032; Illumina Inc.) prior to a limited number of cycles of PCR amplification for the addition of indexed barcodes using the following modified program: 72 °C for 3 min; 98 °C for 3 min; 10 cycles of 98 °C for 30 s, 60 °C for 30 s and 72 °C for 30 s; 72 °C for 5 min; and a 10 °C hold. Amplified libraries were cleaned with 1× SPRI and normalized according to the VeriSeq® PGS Library Prep Reference Guide (Part: 15052877 v03; Illumina Inc.). Normalized libraries were pooled for 24-plex sequencing using a Miseq System using Reagent Kit V3 consumables (2 × 36 bp; Part: MS-102-3001; Illumina Inc.). The data were analysed for copy number variations by extracting the .bam files and importing them into BlueFuse Multi® software (Illumina Inc.) with a custom, MDA-specific reference database. The analysis pipeline was validated using single cells and genomic DNA from four cell lines obtained from the NIGMS Human Genetic Cell Repository at the Coriell Institute for Medical Research, NJ, USA, two of which contained structural abnormalities: GM00526 (47, XY, + 13), GM04927 (47, XY, + 21), GM50121 (46, XY, − 18p 15.5 Mb, + 18q 59.3 Mb) and GM10985 (46, XX, − 3p 10.3 Mb). Cell lines were cultured according to the supplier’s recommendations, and single cells were isolated manually using a previously described method [26]. The concordance per chromosome was greater than 99% for single cells (n = 2208).

Study design

We planned to include 25 consecutive ovaries surgically excised for OTC to obtain an average estimate of the number of oocytes that is possible to collect from surplus medulla tissue from one entire ovary. This approach prevented an equal distribution of ovaries that were delivered from the local hospital and ovaries that were transported on ice, because we were unable to control for the admission of patients. The oocytes collected from each ovary were allocated in equal numbers representing the three different groups of COCs to either one of the two media tested for IVM (i.e. with and without the addition of midkine). In order to not prolong the period of preparation for the cortical tissue, we decided to test the effect of the preparation medium on one ovary at a time and evaluated 12 and 13 ovaries, respectively in the two types of media. An overview of the patient’s material for IVM of oocytes obtained from surplus medulla tissue after the dissection of ovarian tissue is shown in Fig. 3. Aneuploidy studies were performed on a randomly selected group of oocytes that sustained the MII transition and remained intact after warming.

Fig. 3.

Overview of patients material for in vitro maturation of oocytes obtained from surplus medulla after the dissection of ovarian tissue. A total of 25 ovaries from 25 patients were included in the study. Seven patients were operated at the local hospital (no transport of ovaries on ice was needed), while other 18 patients were operated at collaborating hospitals and their ovaries were delivered on ice. For the first 13 consecutive patients, saline was used as a medium for dissection and preparation of cortex, while for remaining 12 patients, the HEPES buffered HTF medium was used

Statistical analysis

The frequency of maturation of immature oocytes to the MII stage was analysed using a single multivariate mixed logistic regression model with maturation to MII (yes/no) as the outcome and transport (yes/no), collection medium (saline/HEPES-buffer), culture medium (with midkine/without midkine), COC size (naked oocytes, small COCs and large COCs) and oocyte diameter (linear effect) as explanatory variables. The patient effect was included as a random intercept. A cut-off value for the oocyte diameter where a 50% maturation rate was achieved (MII = yes) was determined using the predicted values obtained from the mixed logistic regression analysis. The distribution of COCs between treatments was analysed using the chi-square test. Diameters were analysed with one-way analysis of variance (ANOVA) in which the patient effect was included as a random intercept. All analyses were performed using R version 3.4.3. All p values less than 0.05 were regarded as significant.

Results

The main outcome measure was the number and frequency of MII oocytes obtained after IVM. Secondary outcome measures included whether the maturation rate was affected by the following parameters: (1) the originating hospital (transported vs non-transported ovaries), (2) the type of medium used for the dissection of ovarian tissue (saline vs HEPES-buffered medium) and (3) the addition of midkine to the culture medium (IVM medium supplemented with or without midkine). Furthermore, euploidy/aneuploidy rates were calculated for IVM MII oocytes after the biopsy of the first polar body.

Immature oocytes

Eight hundred ninety-five immature oocytes from 25 ovaries (25 patients) were collected from the surplus tissue that remained after the isolation of the ovarian cortex for cryopreservation (Fig. 1). Seven ovaries were obtained from the local hospital with no transportation on ice and 18 ovaries were delivered from collaborating hospitals on ice (Fig. 3). The average number of oocytes collected per patient was 36 ± 5 (mean ± SEM, range 7–90). No significant correlation was observed between the number of oocytes collected and age of the patient (R² = 0.03) or the number of oocytes collected and ovarian volume (R² = 0.26). The distribution of oocytes among groups with different experimental conditions is presented in Table 1.

Table 1.

The distribution of oocytes, number of mature oocytes and maturation rate for all different experimental groups: oocytes originated from transported (TRANSPORT) and non-transported (LOCAL) ovaries, collected from saline or HEPES buffered HTF medium, oocytes matured in the medium with (MK) and without (No MK) addition of midkine

	Saline						HEPES
	No MK			MK			No MK			MK
	Total (n)	MII (n)	MII rate (%)	Total (n)	MII (n)	MII rate (%)	Total (n)	MII (n)	MII rate (%)	Total (n)	MII (n)	MII rate (%)
Local	42	9	21	40	16	40	70	34	49	67	33	49
Transport	217	51	23	209	58	28	128	21	24	122	42	34

Maturation rate

The overall IVM rate was 31% (274 MII oocytes of 895 total oocytes), but a considerable variability was observed (range 6–55%, supplementary Table 1). The IVM rate for patients aged less than 20 years (n = 2, 40%) did not differ from patients aged 20–30 years (n = 14, 29%) and greater than 30 years (n = 9, 31%; p > 0.1).

The size of the cumulus cell mass was significantly associated with the maturation rate: oocytes surrounded by a large cumulus cell mass were more likely to mature (52%) than oocytes enclosed in small cumulus complexes (26%) or NOs (12%) (p < 0.001, Table 2). The number of large COCs was similar, regardless of whether the ovaries were transported, and the maturation rate did not differ (33% vs 40%, p > 0.05). Furthermore, for large COCs, the handling medium did not affect the MII rate (34% vs 35%, p > 0.05).

Table 2.

Maturation and cumulus expansion for immature oocytes in relation to the initial morphology

Category of oocyte	No.	No. MII oocytes (day 2)	Maturation rate (%)	Day 1		Day 2
				No. COCs expandedcumulus	Cumulus expansion rate (%)	No. COCs expandedcumulus	Cumulus expansionrate (%)
Naked oocytes	212	25	12	–	–	–	–
Small COCs	399	102	26	54	14	72	18
Large COCs	284	147	52	56	20	155	55
Total	895	274	31	110	16	227	33

Effects of transportation, media and midkine supplementation on the oocyte maturation rate

Oocytes collected from ovaries originating from patients who received an operation at the local hospital (oocytes = 219, ovaries = 7) matured at a significantly higher rate (42%) than oocytes collected from ovaries transported on ice (oocytes = 676, ovaries = 18) prior to dissection (27%; p = 0.0011, Table 3).

Table 3.

Maturation rate in relation to transport, collection and culture media (mean ± SEM (range)). All p values present the difference in maturation rate estimated in a multivariate model using the parameters: transport, collection media, culture media, COC size and diameter as explanatory variables. Odds ratio from the model used to predict maturation rate shown in Fig. 2 for the different treatments

Group		Number of oocytes	Oocytes/patient	MII oocytes/patient	Maturation rate (%)	p value	Odds ratio (95% CI)
Transport	Transported	676	38 ± 6 (9–90)	10 ± ±2 (1–27)	27 ± 2 (6–38)	0.0011	0.5 (0.4–0.8)
	Local	219	31 ± 7 (7–56)	13 ± 4 (2–30)	42 ± 5 (23–55)		1 (ref)
Collection media	HEPES	387	33 ± 8 (7–90)	12 ± 3 (2–30)	36 ± 4 (13–55)	0.0001	1 (ref)
	Saline	508	39 ± 5 (13–88)	10 ± 2 (1–27)	26 ± 3 (6–39)		0.5 (0.4–0.7)
Culture media	Midkine	438	20 ± 3 (3–65)	7 ± 1 (1–23)	34 ± 3 (11–58)	0.04	1.4 (1.0–2.0)
	Control	457	18 ± 4 (4–90)	6 ± 1 (1–24)	27 ± 3 (0–55)		1 (ref)
Total		895	36 ± 5 (7–90)	11 ± 2 (1–30)	31 ± 2 (6–55)

Ref reference value for the odds-ratio

The maturation rate for oocytes collected from ovaries dissected in HEPES-buffered medium (oocytes = 387, ovaries = 12) was significantly higher (36%) than oocytes collected from saline (oocytes = 508, ovaries = 13), which matured at a rate of 26% (p = 0.0001) (Table 3).

The addition of midkine to the culture media significantly increased the maturation rate. Oocytes cultured in the presence of midkine (oocytes = 438) matured at a rate of 34%, whereas the control group (oocytes = 457) matured at a rate of 27% (Table 3, p = 0.04).

Consequently, oocytes collected in HEPES-buffered medium from the local hospital that were cultured in the presence of midkine matured at a significantly higher rate (49%) than oocytes collected from transported ovaries in saline and cultured without midkine (23%) (p < 0.001) (Table 1).

Oocyte diameter

The average diameter was 110 ± 7 μm for of MII oocytes (n = 274), 105 ± 9 μm for MI stage cells (n = 223), 96 ± 12 μm for GV stage cells (n = 117) and 88 ± 19 μm for degenerated oocytes (n = 281). The differences in diameters between different maturation stages were all significant (p < 0.001). In contrast, the thickness of the zona pellucida was not significantly different among the groups of oocytes after 44–48 h of IVM culture (20 ± 4, 21 ± 5, 20 ± 5 and 21 ± 4 μm, respectively) (p > 0.1) (Table 4).

Table 4.

Oocyte diameter and zona pellucida thickness in relation to the maturation stage after 48 h of IVM. Data are mean ± SEM (range)

Oocyte stage	No. oocytes	Diameter of oocyte (μm) (mean ±SEM)	ZP thickness (μm) (mean ±SEM))
MII	274	110 ± 7 (83–148)	20 ± 4 (10–33)
MI	223	105 ± 9 (75–153)	21 ± 5 (10–51)
GV	117	96 ± 12 (49–118)	20 ± 5 (8–32)
DEG	281	88 ± 19 (37–142)	21 ± 4 (8–38)

We developed a linear model to understand the effects of factors such as the collection site, collection medium and culture medium on the diameter. The equivalent diameter of the 50% MII transition (ED50% MII transition value, i.e. the diameter at which 50% of the oocytes will mature to the MII stage) under different conditions was calculated based on a linear model of diameters for all oocytes included in the study (Fig. 2). Clearly, oocytes that originated from ovaries delivered from the local hospital, dissected in HEPES-buffered medium and cultured with midkine showed the lowest ED50% value (102.5 μm) compared with any other group (Fig. 2). In contrast, the oocytes that originated from ovaries that were transported, dissected in saline and cultured without midkine had the highest ED50% value (117.9 μm). The diameter of the oocytes did not change during the culture period (data not shown).

Chromosomal error rates among MII oocytes after IVM of GV oocytes obtained from SAF

Next-generation sequencing was used to assess the aneuploidy rate of the mature MII oocytes after IVM (Fig. 4). For this experiment, we removed the first polar body and isolated the MII oocyte, which was subjected to whole-genome amplification. Seventy oocytes from 13 patients were warmed for analysis. After warming, 53 oocytes were identified as alive and of good quality based on their morphology (survival rate of 76%) and were subsequently biopsied. An analysis of the shallow next-generation sequencing profiles (0.01×) revealed that 36% ± 3.3% of the 53 oocytes sampled from 13 patients aged 19 to 33 years were aneuploid (Fig. 4b and c). Thus, we estimated that approximately two-thirds of the MII oocytes obtained after IVM of GV oocytes from SAF of women aged less than 35 years were euploid with a normal chromosomal constellation. The distribution of the oocytes among the different parameters is presented in supplementary Table 2.

Fig. 4.

Aneuploidy rates in oocytes from small antral follicles after in vitro maturation (IVM). a The first polar body was removed and the MII oocyte was used as template for single-cell wholegenome amplification (WGA) and shallow next-generation sequencing (NGS, 0.01×). b Aneuploidy rates in oocytes matured in the IVM protocol with or without midkine 1 mg/ml. The fraction of MII oocytes that were euploid shown in green, whereas the fraction that were aneuploidy is represented by grey. Error bars represent standard error. In total, 53 MII oocytes were assessed after IVM: 33 oocytes in the IVM medium + midkine group and 20 oocytes in the IVM medium group. It was found that 22 and 12 oocytes in these groups respectively were euploid. c The frequencies of euploid (green) and aneuploid (grey) MII oocytes per participant, ordered according to chronological age. The number on the green or grey part of the bar is the number of oocytes analysed

Discussion

To our knowledge, the present study is the largest to examine immature oocytes isolated from surplus medulla tissue and thus from SAF. This is a new source of human oocytes deriving from antral follicles with a very small diameter that has not been studied in great detail. However, this material provides for the first time an opportunity to explore how and if such oocytes can be matured in vitro employing new growth factors and culture conditions to potentially form a basis for a new platform for IVM and clinical applications in the future. The main finding was that a large number of immature oocytes was obtained from the surplus ovarian medulla tissue of each patient, which is the tissue that previously was discarded. Thus, this source of human oocytes has not previously been studied in great detail. On average, 36 immature oocytes were obtained from one ovary of an unselected group of patients, with a few women providing approximately 100 immature oocytes each, thereby increasing the average number. The IVM procedure yielded an average of 11 MII oocytes, with one patient producing 30. The ability to obtain such a high number of oocytes from a single ovary is a result of a meticulous search and recovery process. This procedure mainly included oocytes from SAF with a small diameter, which are rarely visible with the naked eye and most probably having diameters well below 2–3 mm. In previous IVM studies performed in our laboratory [13, 14], the average number of COCs collected per patient was 10 to 11. In the present study, we have tripled the average number of immature oocytes collected to 36 per patient. This improvement was possible partially due to the preparation of the surplus tissue, which was cut into small (approximately 2 mm³) pieces with a tissue chopper prior to oocyte collection. During this procedure, more oocytes are released from SAF compared with previous studies where the tissue was minced into many small fragments with a scalpel by hand. In the present study, two experienced operators collected the oocytes and double checked one another, while only one operator performed the collection procedure in the previous studies.

The genomic analysis of MII oocytes revealed that 64% were euploid, which is 9% lower than estimates for MII oocytes retrieved from large and mature follicles after ovarian stimulation [27]. However, the wide range of aneuploidy rates reported for women aged less than 35 years prevents a ready comparison with our data, although the aneuploidy rate appears to be increased [28, 29]. Despite the slight increase in the aneuploidy rate, the large number of oocytes retrieved from the ovarian tissue provides MII oocytes in numbers that are comparable with 1 cycle of ovarian stimulation. On average, eight euploid MII oocytes were recovered and potentially used for fertility treatment. The potential for the embryonic development of oocytes matured in vitro is unknown but likely to be lower than oocytes matured in vivo during stimulated cycles, and the efficiency of embryo formation and further development remains to be determined. Currently, the Danish health authorities consider IVM an experimental procedure and additional ethical approval is needed before MII oocytes obtained after IVM are fertilized and the chromosomal constitution of embryos is assessed locally. This requirement for additional approval is an obvious limitation of the present study since the MII oocytes originating from SAF are likely to possess an attenuated developmental competence. Furthermore, the aneuploidy rate in the group of oocytes exposed to midkine was similar to oocytes that were not exposed to midkine, while the MII rate was significantly increased. Therefore, this study is the first to confirm that oocytes obtained from small follicles mature in vitro to the MII stage, most of them are euploid and the addition of human recombinant midkine to the culture medium significantly increases the maturation rate of these oocytes.

The isolation of immature oocytes from surplus tissue in connection with fertility preservation has previously been performed with highly variable recovery of oocytes ranging from 1 to 15 cells [7, 9, 10, 11, 12, 30, 31]. Some of these studies excluded naked oocytes, which may partially explain the lower numbers of immature oocytes collected than in the present study. Some studies did not mention specific details on whether naked oocytes were included for IVM. Interestingly, preliminary NGS results suggest that MII oocytes derived from naked oocytes at collection are likely to be euploid, despite the limited number of cells included. In our study, 3 of 3 oocytes derived from naked cells were normal (supplementary Table 2). Therefore, despite the low maturation rate of naked oocytes (12%), we considered every opportunity to improve the reproductive outcome for the patient.

IVM studies of prepubertal girls [4, 7, 32] revealed an attenuation of the average maturation rate compared with adults [8, 30, 33]. The maturation rate in our study was similar to or lower than other studies including large groups of patients [4, 10, 12, 30, 32, 34, 35], but we included oocytes from all diameters of follicles and the number of MII oocytes obtained per patient after IVM is the important measure. Previous studies [4, 10, 12, 30, 32, 34, 35] reported as maximum of 5 MII oocytes per patient, and the value more than doubled in the present study to an average of 11 MII oocytes.

Based on our findings, optimizing the collection procedure and using a more appropriate HTF HEPES-buffered media during the preparation of cortical tissue improves the resumption of meiosis. Immature oocytes that were cooled to approximately zero degrees during transportation for 2–5 h showed a significant reduction in their capacity to resume meiosis compared with oocytes obtained from the local hospital. We were unable to determine whether the cause is a reduced temperature or increased apoptosis in the follicles. Nevertheless, the number of oocytes recovered justifies their collection and putative use for fertility preservation. Some collected oocytes were present in very large COCs with many layers of somatic cells, which is a characteristic that is seldom observed in normal IVF. Surprisingly, the category of large COCs displayed the highest maturation rate (52%). We hypothesise that many cumulus cells are needed to sustain oocyte development, but we are currently unable to pinpoint the specific follicle diameter from which these large COCs derive. Collectively, immature oocytes obtained from the surplus medulla tissue may increase fertility opportunities for women receiving OTC to a larger extent than previously assumed.

The presence of the growth factor midkine significantly increased the maturation rate. Midkine was previously reported to affect maturation of immature bovine oocytes [19] by acting indirectly via granulosa cells and/or cumulus cells, as no effect was observed on naked oocytes [21]. An effect on somatic cells surrounding oocytes is supported by the findings from the present study, where a beneficial effect was only observed on large and small COCs. Furthermore, midkine has been shown to reduce the apoptosis of bovine cumulus cells during IVM [21]. Midkine is present in human granulosa and theca cells of the preovulatory follicles [36]. Therefore, based on our results and published data, we postulate that midkine is suitable for use in human IVM, but further studies are needed to verify this hypothesis. Taken together, the present study supports previous investigations of the ability of midkine to promote IVM and showed for the first time that recombinant human midkine significantly increased meiotic progression to the MII stage of human immature oocytes during a 44–48 h culture period. We recommend the potential inclusion of midkine in IVM medium in routine practice to increase the maturation rate of human oocytes.

The classification of human oocytes is not yet well defined, as oocytes will often merely be classified either as mature (metaphase II stage) or immature (germinal vesicle or metaphase I stage) cells [37]. However, immature oocytes now require further classification, as they are collected from follicles with a range of diameters and their meiotic competence varies. The average diameter of IVM MII oocytes from SAF (110 ± 7 μm) is similar to the mean diameter of oocytes retrieved at IVF clinics after ovarian stimulation (112 ± 3 μm) [38]. At the same time, the diameter of oocytes originating from SAF appears to reflect their meiotic capacity: the larger the diameter, the better chance of supporting the MII transition. This observation confirms and extends an earlier study that obtained a similar result: 75% of 49 oocytes collected from ovaries excised from six women underwent the MII transition when the diameter exceeded 105 μm, while only 25% matured when the diameter was smaller [12]. Different factors have been employed to improve the conditions for oocyte collection and maturation, as reflected in the equivalent diameter of the 50% MII transition. The ED50% value is significantly reduced when ovaries from a local hospital were compared with ovaries transported at zero degrees. The same result was obtained when we compared HEPES-buffered media with saline and when midkine was added to the culture medium. Collectively, the diameter of immature oocytes from SAF reflects their meiotic capacity, and the ED50% MII transition is a useful parameter that can be employed to optimize conditions for human immature oocytes, including collection and culture. According to a previous study by Sánchez et al. [39], oocytes with a diameter less than 109 μm have a higher frequency of dispersed chromatin. Consistent with these results, the oocyte diameter is important as a first parameter to subdivide or categorize the large heterogeneous group of human immature oocytes from follicles with different diameters. In this context, the MII rate of oocytes with a diameter of less than 100 μm was as modest as 3% (9 MII oocytes of 274 oocytes). We, therefore, recommend that immature oocytes should be distinguished based on the diameter and propose two categories based on their developmental competence: immature human oocytes with diameters less than or greater than 100 μm. However, further investigations of the developmental competence based on these findings are warranted.

Although the collection of immature oocytes was intended to increase the potential fertility of women who had one ovary excised for OTC, the unexpectedly high number of immature oocytes collected may provide an opportunity to justify the procedure itself. Women in whom the transplantation of ovarian tissue is not recommended because of a risk of reintroducing malignant cells may benefit from this procedure, including women with ovarian cancer, borderline ovarian cancer, granulosa cell tumours or cancers with a high risk of infiltration of malignant cells into the ovarian tissue, such as leukaemia or colon cancer. OTC is currently not offered to many women with ovarian cancer because of the risk of reintroducing the malignancy during transplantation. However, the collection of immature oocytes would eliminate this risk since the oocytes will not include malignant cells. Therefore, this procedure is new and may provide a group of women who previously had no opportunity to preserve fertility with a chance of having their genetically own children for the first time.

The present study has weaknesses, and most pronounced limitation is the inability to follow the subsequent developmental competence of fertilized MII oocytes. Furthermore, an evaluation of the effect of the collection of oocytes in the two types of media was not able to be performed in a strictly randomized manner for practical reasons. Regardless of these limitations, the present study advances the use of immature oocytes collected from surplus medulla tissue, mainly due to the large number of oocytes included. The MII maturation rate was evaluated after a culture period of 44–48 h. The oocytes were not destined for use in the clinic, and for convenience, the evaluation was not performed at 36 h, which would be used in a clinical setting.

The longer-term perspective of the present findings is potentially the dawn of fertility treatment without stimulation with exogenous hormones in certain groups of women. Up to 100 immature oocytes may be harvested from one ovary of women with, for instance, PCOS. The collection of oocytes from SAF is envisioned to be optimized by developing new procedures for the aspiration of very small follicles. In women with PCOS and two ovaries, enough immature oocytes may be collected without ovarian stimulation to create equal or even higher numbers of MII oocytes for treatment than conventional ovarian stimulation. If necessary, repeated aspirations of oocytes may be performed with stockpiling of frozen oocytes between retrieval sessions. Prior to the use of immature oocytes from SAF on a larger scale, safety studies are required, and one aspect to evaluate will be the epigenetic status. To the best of our knowledge, the methylation status of human IVM oocytes collected ex vivo from SAF has not been evaluated. As shown for human IVM oocytes collected in vivo after hormonal stimulation [40–42], the IVM procedure may modify the methylation pattern under some conditions. However, an analysis of IVM-induced epigenetic changes in 11 new-borns after IVM in stimulated cycles [43] revealed relatively few changes, if any.

Conclusions

As shown in the present study, an unexpectedly high number of immature oocytes was obtained from surplus medulla tissue during cortex preparation and fertility preservation. Although many of these oocytes were derived from antral follicles with a diameter of only a few millimetres, the meiotic competence in terms of the MII transition was maintained at an overall rate of approximately one in three MII oocytes. The euploidy rate was acceptable, although the majority of ovaries were transported on ice with a significantly reduced MII rate than ovaries obtained from the local hospital.

Using this new source of human oocytes from very small antral follicles for IVM is currently an experimental exercise to form a platform for improving the IVM procedure, especially because the true clinical potential cannot be evaluated in our studies. An important component of this improvement is the identification of midkine as a growth factor that significantly increased the maturation rate when it was present in concentrations comparable with the preovulatory follicular fluid [32]. We hypothesised that oocytes from these small follicles undergo an accelerated meiotic competence after quick physical handling and placing them in an environment with high levels of FSH, whereby they avoid degenerative changes that they might have encountered in vivo. The study suggests a sub-division of the large ill-defined category of immature human oocytes based on an oocyte diameter, and a diameter of 100 μm separates oocytes that possess the capacity to undergo the MII transition from oocytes that are unable to sustain this transition. The maturation rate of 31% described in the present study results in the generation of an average of 11 MII oocytes, of which approximately two-thirds are euploid, similar to the results obtained with most ovarian stimulation regimes.

Authors’ contributions

CYA, DN, CJ, SEP and SGK planned the study. DN, CJ, VS, RB, SEP, SGK, MLG performed the experiments and collected the data. DN, CJ, SEP, SGK, JC, VS, RB, ZG, YW and EH analysed the data and produced figures. KTM, LC, EE and AMBB recruited patients and collected the tissues. DN, SEP, SGK and CYA wrote the manuscript. All authors approved the manuscript.

Funding information

The grants from the organizations listed below are gratefully acknowledged as sources of financial support. ERH, VS and RB are grateful for the support from The Independent Research Fund Denmark, the NNF Young Investigator Award (grant number 16662), the ERC (grant number 724718) and the Danish National Research Foundation (DNRF115, 6110-00344B). The project ITN REP-BIOTECH 675526 funded by the European Union, European Joint Doctorate in Biology and Technology of the Reproductive Health is acknowledged for supporting DN. The University Hospital of Copenhagen, the EU Interregional Project ReproUnion and Vera and Carl Johan Michaelsens Legat are acknowledged for providing support for CYA, SGK and SEP. All personnel, including the staff involved in clinical activities in fertility preservation, are acknowledged for their passionate work.

Compliance with ethical standards

Conflict of interest

The authors declare that they have no conflicts of interest.