Ovary isolation and back table perfusion

A total of 22 ovaries were obtained from 11 sexually mature ewes (8–12 months old) of mixed breeds at a local slaughterhouse between October 2021 and November 2022. Immediately after euthanasia, 5-mL blood samples were collected from each animal and kept on ice during transportation to the laboratory. Blood samples were then centrifuged at 1500 g for 20 min, and plasma was stored at − 20 °C for further analysis.

To minimize contamination, all ovaries with intact vascular pedicles were rinsed once at the slaughterhouse, by dipping in a chlorhexidine solution dissolved in normal saline (0.001%; Fresenius Kabi, Uppsala, Sweden), followed by three dips in a sterile saline solution (0.9%; Braun, Melsungen AB, Germany).

The collected ovaries were divided into an experimental group (n = 11) and a control group (n = 11) (Fig. 1). The ovarian artery was separated from the ovarian vein using a dissection microscope (Nikon, Tokyo, Japan), and all ovaries were initially perfused at the slaughterhouse with 2 mL of cold (4 °C) sterile phosphate-buffered saline (PBS, Gibco, London, UK) supplemented with heparin (50 IU/mL; Apoteket AB, Stockholm, Sweden), xylocaine (0.04 mg/mL; Astra Zeneca, Gothenburg, Sweden), piperacillin/tazobactam (10 µg/mL and 1.25 µg/mL; Stragen Nordic, Hillerød, Denmark), and Gibco’s antibiotic–antimycotic solution (1%; Thermo Fisher Scientific, Stockholm, Sweden) until the organ blanched and clear liquid was observed flowing through the ovarian vein. The venous outflow was collected and stored at − 20 °C for further analysis. For the ovaries included in the experimental group, the ovarian artery was then cannulated (24 G; Becton Dickinson Infusion Therapy AB, Helsingborg, Sweden) using a dissection microscope (Nikon, Tokyo, Japan). The cannula was then fixed using a sterile 4–0 silk suture while the venous outflow was left open. Ovaries allocated to the experimental group were placed in a cold (4 °C) sterile PBS solution and transported on ice to the laboratory. Ovaries in the control group after the flushing were immediately dissected, and a small piece (approximately 2 mm3) of cortex tissue was placed in RNAlater (Qiagen, Hilden, Germany) for subsequent gene expression analysis, while a larger piece (approximately 4–5 cm3) was placed in 4% formaldehyde serving as fresh control tissue for histomorphological and immunohistochemistry examination. The remaining half of each control ovary was transported to the laboratory with the same conditions used for the ovaries in the experimental group. At the laboratory, the remaining half of each control ovary was further divided into two pieces; a small cortex tissue piece was placed in RNAlater (Qiagen), while the other piece was placed in 4% formaldehyde [15].

Fig. 1

Study design. Twenty-two ovaries were obtained from sexually mature ewes and divided into two main groups: control (n = 11) and experimental (n = 11). All ovaries underwent cannulation and flushing at the collection site before transportation to the laboratory within a 3- to 4-h timeframe. In the control group, half of each ovary served as fresh control tissue, while the other half served as control tissue to assess potential cold ischemia damage that occurs during transportation to the laboratory. In the experimental group, all ovaries were connected to a bioreactor for ex vivo perfusion. These ovaries were further divided into two experimental subgroups, each undergoing distinct perfusion protocols. Both protocols (established for 24 h and then repeated until experiment termination) involved hormonal stimulation with follicle-stimulating hormone (FSH) and human menopausal gonadotropin (hMG), with concentration halved every 4 h (starting concentration 2 IU/mL FSH and 1 IU/mL hMG). However, after 12 h, only subgroup 2 maintained FSH/hMG stimulation overnight. hCG, human chorionic gonadotropin. Figure created with Biorender

The average time between euthanasia, ovarian artery isolation, cannulation, and vascular perfusion was between 30 and 60 min, while the average cold ischemia time during the transport to the laboratory ranged from 180 to 240 min.

Implementation of a modified closed circulation perfusion system for ex vivo ovarian perfusion

The ovaries assigned to the experimental group underwent ex vivo perfusion in a modified closed circulation bioreactor utilized in our previous study [15]. The perfusion system involved keeping the cannulated ovary in the media reservoir, fully immersed in the perfusion medium as depicted in Fig. 2 (Fig. 2). Thus, the media reservoir also acted as the organ chamber. The bioreactor system enabled recirculation of the perfusate (closed system), ensuring a sterile environment with adjustable monitoring of temperature, oxygenation, perfusion pressure, and flow rate throughout the experiment. The bioreactor comprised of three distinct circuit loops: (1) the organ chamber, which was integrated with the media reservoir; (2) Carbogen 5 gas (gas mixture of 95% O2, 5% CO2; Linde, Solna, Sweden), which provided buffering but also contributed to oxygenation of the perfusion media; and (3) warm circulating water (38 °C) in a jacketed system that maintained a constant physiological temperature for both the perfusate and the organ throughout the experiment. A peristaltic pump facilitated the circulation of the perfusate through the oxygenator, including a bubble trap, and directed it to the ovary in the organ chamber. To prevent potential damage to the ovarian vasculature, the perfusion pressure was kept below 80 mmHg by adjusting the flow rate to 1.53 mL per min during the entire perfusion period [15].

Fig. 2

Schematic representation of a bioreactor system designed for organ perfusion. The setup includes a peristaltic pump for controlled fluid circulation, an oxygenator for maintaining oxygen levels in the perfusion medium, and a heat exchanger for regulating temperature. The organ perfusion chamber houses the perfused ovary, which is completely submerged within the perfusion medium. This system enables precise control of environmental conditions to support the viability and function of the perfused organ [15]

Experimental design and perfusion protocols for ovarian stimulation

The ovaries in the experimental group were divided into two subgroups, each receiving a distinct gonadotropin stimulation protocol. Based on the daily peaks of gonadotropin secretion in cycling ewes [16] and the findings from our previous study [15], both experimental subgroups received different daily doses of follicle-stimulating hormone (FSH) and human menopausal gonadotropin (hMG) media supplementation as detailed in the subsequent description. The difference between subgroups was that subgroup 1 received continuous gonadotropins including overnight administration while subgroup 2 had an overnight pause (Fig. 1).

Ovaries in subgroup 1 (n = 5) were perfused for 4 to 8 days with a medium composed of M199 (Thermo Fisher Scientific) supplemented with 2% bovine serum albumin (Roche Diagnostic GmbH, Mannheim, Germany), piperacillin/tazobactam (10 µg/mL and 1.25 µg/mL; Stragen Nordic), Gibco antibiotic–antimycotic solution (Thermo Fisher Scientific), 12 mL/L sodium bicarbonate (7.5%; Thermo Fisher Scientific), 10% fetal bovine serum (FBS) (Thermo Fisher Scientific), IGF-1 (50 ng/mL, Peprotech, Stockholm, Sweden), and HEPES (Sigma Aldrich, Stockholm, Sweden). After an initial perfusion of 1 h without gonadotropins, FSH (Gonal-F, Merck AB, Solna, Sweden) and hMG (Menopur, Ferring Pharmaceuticals, Malmö, Sweden) were added in the medium in a single dose of 2 + 1 IU/mL, respectively, for 4 h, followed by 1 + 0.5 IU/mL for the next 4 h and 0.5 + 0.25 IU/mL for an additional 15 h. This cycle was repeated every 24 h until the end of the perfusion experiments.

Ovaries in subgroup 2 (n = 6) were perfused for 4 to 8 days with the same medium and protocol used for subgroup 1, except for not receiving the 15-h low-dose gonadotropin stimulation that was replaced by a 4-h perfusion with low-dose of gonadotropins (0.5 + 0.25 IU/mL of FSH and hMG respectively) followed by 11 h overnight pause of hormonal exposure. These conditions were repeated every 24 h during the perfusion experiments.

To trigger final oocyte maturation in both experimental subgroups, human chorionic gonadotropin (hCG) 50 IU/mL (Pregnyl, MSD Sverige AB, Stockholm, Sweden or Ovitrelle, Merck AB, Solna, Sweden) was added to the medium during the last 24 h of perfusion before ovum pick-up (OPU). OPU was performed by retrieving newly visible follicles with a diameter between 3 and 5 mm through gentle manual aspiration using an 18-G needle connected to a tube with an inner diameter of 2 mm [17].

During the experiments, 10 mL of the perfusion medium was sampled from the organ chamber prior to medium changes at ten different time points: 20 min, 1 h, 5 h, 9 h, 24 h, 36 h, 48 h, 60 h, 72 h, and 80 h, until the end of perfusion (prior to any medium change) and immediately frozen at − 20 °C for further biochemical analysis, as described below.

At the end of the ex vivo perfusion, a small ovarian piece of cortical tissue, about 2 mm3, was directly placed in RNAlater for future gene expression analysis, while the remaining ovarian tissue was fixed in 4% formaldehyde for later histopathological analysis.

Assessment of follicular steroidogenic activity and reproductive cycle phase in ewes

To assess the follicular steroidogenic activity and reproductive cycle phase of each ewe, concentrations of hormones were measured in the undiluted sampled media, plasma, and liquid collected after gently flushing the organ at the slaughterhouse. The following ELISA kits, specific for sheep, from Reagent Genie (Dublin, Ireland) were used: estrogen (#SHFI00063; sensitivity, 9.375 pg/mL; detection range, 15.625–1000 pg/mL), progesterone (#SHFI00062; 18.75 pg/mL, 31.25–2000 pg/mL), FSH (#SHFI00026; 1.875 mIU/mL, 3.125–200 mIU/mL), and luteinizing hormone (LH; #SHFI00032; 0.563 mIU/mL, 0.938–60 mIU/mL).

Measurement of ovarian metabolic activity

To assess ovarian metabolic activity, concentrations of glucose, lactate, potassium, sodium, calcium, and chloride were measured in the perfusion medium using a standard blood-gas analysis using the RapidPoint 500 Blood Gas System (Siemens Healthcare, Erlangen, Germany), following the manufacturer’s protocols.

Evaluation of oocyte maturation stages

Retrieved oocytes were immediately examined by two experienced embryologists, who were blinded to the experimental subgroups. The oocytes were classified based on their maturation stages into germinal vesicle (GV), metaphase I (MI), and metaphase II (MII) oocytes. Oocyte maturity assessment was performed using a stereomicroscope. The number of retrieved oocytes, recovery rates, and their distribution across these maturation stages were meticulously recorded.

Assessment of ovarian tissue integrity through histomorphology and immunohistochemistry

Formaldehyde-fixed ovarian samples included fresh control tissue (collected and fixed at the slaughterhouse), cold ischemia control tissue (collected and fixed after transport to the laboratory), and the experimental ex vivo perfused ovaries. These biopsies underwent dehydration and paraffin embedding followed by 4 µm sectioning and stained with hematoxylin and eosin (H&E) as per standard protocols. Each specimen was examined by an experienced pathologist blinded to the sample groups. To prevent duplicate follicle counting, a minimum of ten sections from each ovary were reviewed, with at least a 50-µm separation between each section. Digital scanning of all slides was performed using a NanoZoomer S210 scanner (Hamamatsu Photonics, Hamamatsu, Japan), and the surface area of the ovarian cortex in each section was measured using NanoZoomer Digital Pathology view 2 analysis software (Hamamatsu Photonics). The degree of injury was assessed digitally, based on established pathology parameters [18, 19]. Non-damaged primordial follicles were identified as those with a spherical oocyte and a non-pyknotic nucleus, surrounded by one layer of flattened granulosa cells. Non-damaged primary follicles were small, spherical follicles containing a GV stage oocyte encircled by one to two layers of cuboidal granulosa cells. Secondary non-damaged follicles were larger with several layers of granulosa cells, sometimes having small spaces between them, and a GV stage oocyte. Non-damaged antral follicles were recognized by their fluid-filled antral cavity, with the oocyte situated at the edge in a mound of granulosa cells (cumulus oophorus). Damaged follicles were categorized into grade 1 (G1) when theca and granulosa cells were separated from the follicular edge, showing disruptions and apparent loss of granulosa cells while the oocyte maintained its spherical shape. Grade 2 follicles (G2) exhibited more severe disruptions, loss of granulosa cells, theca cell detachment from the follicle edge, pyknotic nuclei in granulosa cells, and misshapen oocytes, with or without vacuolation or pyknotic nuclei. The overall ovarian histomorphological architecture was evaluated for abnormalities in arteries and veins, such as endothelial detachment, internal elastic membrane rupture, or smooth muscle cell bloating [20, 21].

To study follicular cell proliferation, 3,3ʹ-diaminobenzidine (DAB) immunohistochemistry for Ki-67 (#15,580, 1:100, Abcam, Cambridge, UK) was performed. Apoptosis was confirmed by immunostaining for cleaved caspase 3 (#9661, 1:300, Cell Signaling Technology, BioNordika, Stockholm, Sweden). Antigen retrieval was conducted in a pressure cooker using Diva Decloaker (Biocare Medical, HistoLab, Gothenburg, Sweden) for Ki-67 and citric acid buffer (pH = 6.0) for cleaved caspase 3. Immunohistochemistry analysis was performed using the Mach 3 and Vulcan fast red kits (Biocare Medical) after overnight incubation with the respective primary antibodies. Finally, all slides were cover-slipped with Pertex (Histolab) and scanned using a microscope slide scanner. Proliferation and apoptotic indexes were determined by calculating the percentage of Ki-67- and cleaved caspase 3-positive cells in hotspot areas of the ovarian cortex, considering each follicular subclass separately.

Quantitative gene expression analysis using digital droplet PCR

For gene expression analysis, we adhered to the Minimum Information for Publication of Quantitative Experiments (MIQE) guidelines for digital droplet PCR [22]. Biopsies kept in RNAlater were weighed, segmented into 30-mg pieces, and homogenized. Total RNA extraction was performed using the RNeasy micro-Kit (Cat No. #74,034; Qiagen). Reverse transcription was performed using the iScript cDNA Synthesis Kit (Cat No. #170–8891; BioRad, Stockholm, Sweden). The PCR mixtures were prepared with EvaGreen primer mix (Cat No. #164–4033; BioRad) and converted into droplets following the manufacturer’s instructions using QX200™ droplet generation oil for EvaGreen (Cat No. #1,864,006; BioRad) and the QX200 droplet generator (BioRad). During partitioning, each reaction was divided into 16,000–21,000 droplets, followed by PCR amplification within a C1000 thermal cycler (BioRad). Data analysis involved measuring the fluorescence intensity of each droplet using a droplet reader and employing Quantasoft software (BioRad) for comprehensive analysis. Negative and positive droplet populations were identified, and the number of copies per microliter was standardized to the reference gene peptidyl‐prolyl cis‐trans isomerase H (PPIH). Subsequently, the relative expression levels were determined. Additionally, the primers used in this study were designed and synthesized by Integrated DNA Technologies, with specific sequences detailed in Supplementary Table 1. In this study, we analyzed a set of genes, including FSHR, LHCGR, AMH, IGF-1, BCL-2, BAX, HIF1A, SOD1, and TNFA, focusing on their roles in follicular development, ovarian function, cell growth, apoptosis, hypoxia response, and endothelial cell-related pro-inflammatory and antioxidant processes.

Statistical analysis

Gene expression analyses entailed group comparisons, utilizing log-transformed outcome values to enhance statistical robustness. The non-parametric Kruskal–Wallis test was selected for these comparisons, with adjustments for multiple comparisons based on Dunn’s criterion. This consistent approach extended to the assessment of histomorphological and immunohistochemistry differences across follicle categories, specifically targeting non-damaged, G1, and G2 follicles. The categorization of blood vessel characteristics into normal or injured states prompted an investigation into their associations with experimental groups, control ovary, and cold ischemia by applying Fisher’s exact test. P-values between groups were determined through logistic regression analysis. In exploring changes during ovarian perfusion, mixed models were deployed to analyze observed values of metabolic and electrolytic biomarkers, along with steroid hormones. This approach facilitated the examination of alterations within each experimental subgroup as well as between them across perfusion periods. The reported P-values underwent adjustment using Tukey’s criterion to account for multiple comparisons, ensuring the robustness of the statistical findings. All statistical analyses were conducted using GraphPad Prism v. 9.0 (CA, USA), with a predetermined significance level of 0.05 for all tests.

Ethical approval

Ethical approval for the study was not required, as the ovaries were obtained from ewes bred for food production with the possibility to use discarded organs for research purposes, aligning with the Swedish and European Union ethical regulations.