Clinically applied procedures for human ovarian tissue cryopreservation result in different levels of efficacy and efficiency

The effects of the two selected cryopreservation and thawing protocols: protocol A_CA_T using ethylene glycol as cryoprotectant [18, 19] and B_CB_T using DMSO as cryoprotectant [13, 20, 21] on overall ovarian tissue’s viability, follicle viability, and tissue morphology were investigated in a four-arm design:

For ethical reasons, we could not use tissue from young cancer patients applying for ovarian tissue cryopreservation. We therefore used ovarian tissue from premenopausal women aged ≤45 years who were considered eligible for this prospective cohort study if they underwent a prophylactic laparoscopic salpingo-oophorectomy at the Radboud university medical center (Radboudumc), Nijmegen, the Netherlands. Informed consent was obtained prior to surgery. When possible, cortex tissue from each patient was included in all four arms of the study. In addition, a fresh (positive) control was analyzed for each patient before cryopreservation/thawing. In case of insufficient material (due to size of the ovary, presence of large follicles, and/or extensive damage as a result of cauterization during surgery), the cortex tissue of a given patient was only used in three or less of the study arm protocols.

All study procedures were approved by the Radboudumc local ethics committee.

At the operating theater, the ovary dissected during surgery was collected in cold Custodiol (4 °C; Dr. Franz Köhler Chemie GmbH, Bensheim, Germany). The ovary was immediately transferred to the laboratory (within 5–10 min), where it was placed on a pre-cooled surface (0 °C). The medulla was removed from the cortex using precision forceps and scalpels, after which ovarian cortex fragments of approximately 5 mm in length, 5 mm in width, and 1.0–1.5 mm in depth were prepared. These 5 × 5 × 1 mm fragments were subsequently used to prepare 3-mm-diameter punch biopsies for the glucose uptake assay (overall tissue viability) and for preparing small tissue fragments for the neutral red stain (follicle viability).

The ovarian tissue cryopreservation and thawing protocols selected for this study differ considerably. Briefly, protocol A consists of a cryopreservation protocol using ethylene glycol combined with a thawing protocol based on three consecutive short washes to remove the cryoprotectant. Protocol B uses dimethyl sulfoxide (DMSO) as a cryoprotectant and has a significantly longer and more elaborate thawing procedure based on continuous dilution. Tissue was stored in liquid nitrogen for at least 1 week before thawing.

As we combined two clinically used protocols in our study, we measured the osmolality of the four cryopreservation and thawing solutions in order to determine any additional osmotic stress to which the tissue was exposed when applying cryopreservation and thawing solutions from different protocols.

Tissue fragments were equilibrated in 30 mL of cryomedium, consisting of 0.1 mol/L sucrose (Sigma-Aldrich, Zwijndrecht, the Netherlands) and 1.5 mol/L ethylene glycol (Merck Millipore, Schiphol-Rijk, the Netherlands) in phosphate-buffered saline (PBS; Braun, Melsungen, Germany) for 30 min on a tilting table at 4 °C. Instead of using a Planer Freezer (Planer K10; Planer Ltd, UK), ovarian cortical fragments were frozen in Nunc CryoTubes (Sigma-Aldrich, St. Louis, MO, USA) in 1 mL cryomedium in a CryoLogic programmable temperature controller (CL-3300, Cryosolutions, ‘s Hertogenbosch, the Netherlands) using Cryogenesis^TM V5 software (Cryosolutions). CryoTubes were transferred to the CryoLogic set at 0 °C. Subsequently, the temperature was lowered at a rate of 2 °C/min until a temperature of −9 °C was reached. At this stage, CryoTubes were seeded manually with a cotton swab dipped in liquid nitrogen. The temperature remained at −9 °C for 10 min after which seeding was checked. Next, the temperature was decreased at a rate of 0.3 °C/min to −40 °C and subsequently with 8.5 °C/min to −120 °C before the CryoTubes was stored in liquid nitrogen. This freezing protocol was comparable to the original protocol, expect for the duration at which the temperature remained at −9 °C (5 versus 10 min) and the speed at which the temperature was lowered after reaching −40 °C (8.5 °C/min to −120 °C versus 10 °C/min to −140 °C in the original protocol; [18]). The cryomedium had an osmolality of 2.3 Osmol/kg.

Tissue was thawed rapidly at 37 °C and then rinsed in 25 mL of three solutions (PBS containing 0.25 mol/L sucrose and 0.75 mol/L ethylene glycol; PBS containing 0.25 mol/L sucrose; and PBS, respectively) for 10 min each under continuous agitation to wash out the cryoprotectant. The first washing solution had an osmolality of 1.5 Osmol/kg. In total, this resulted in a thawing procedure of approximately 35 min.

According to cryopreservation protocol B_C [13, 20, 21], a freezing solution with an osmolality of 2.1 Osmol/kg was prepared consisting of Leibovitz’s L-15 GlutaMAX cryomedium (Gibco, Carisbad, CA, USA), containing 10 % dimethyl sulfoxide (1.4 mol/L; CryoSure-DMSO; Wak-Chemie Medical GmbH, Steinbach, Germany), and 10 % serum substitute supplement (SSS; Irvine Scientific, Santa Ana, CA, USA). Nunc CryoTubes were filled with 1.7 mL cold medium (0 °C) and a cortex fragment. Instead of using the program as described for the IceCube (14S-A, SY-LAB, Neupurkersdorf, Austria) in the original protocol [13], the CryoTubes were frozen in the CryoLogic programmable temperature controller using the program as stated under “Cryopreservation protocol A_C.” This freezing protocol was comparable to the original protocol, expect for an incubation step in the programmable freezer at 2 °C for 40 min, the seeding temperature (−9 instead of −6 °C, allowing us to freeze tissue according to protocol A_C and B_C simultaneously and in the same Cryologic freezing device) and the speed at which the temperature was lowered after reaching −40 °C (8.5 °C/min to −120 °C versus 10 °C/min to −140 °C in the original protocol).

Thawing protocol B_T was based on continuous dilution of the cryoprotectant [20]. The CryoTubes were removed from the liquid nitrogen, warmed for 30 s at room temperature, and placed in a 37 °C water bath for 2 min. Directly after thawing, the ovarian cortex fragments were transferred to a solution of 10 mL Dulbecco’s Phosphate-Buffered Saline (DPBS; Gibco), 0.75 M sucrose (MP Biomedicals, Eschwege, Germany), 10 % serum substitute supplement (SSS; Irvine Scientific, Santa Ana, USA), and 0.1 mg/mL Pen/Strep (Lonza, Basel, Switzerland) and incubated under continuous agitation for 15 min at room temperature. This first washing solution had an osmolality of 1.2 Osmol/kg. After this first incubation step, 50 mL of a solution consisting of DPBS with 10 % SSS and 0.1 mg/mL Pen/Strep was added using a pump set at 100 mL/h. At complete infusion (30 min), the cortex fragments were transferred into a 5 mL pre-warmed (37 °C) Hepes-buffered medium (Gamete, Cook Medical Europe LTD, Limerick, Ireland) and incubated for another 15 min. After repeating this step once, the tissue was washed three times for 5 min in tissue culture medium used for the glucose uptake assay (described below). In total, this resulted in a thawing procedure of approximately 1.5 h.

For the control as well as the four study arms, the overall ovarian tissue viability, follicle viability, and tissue morphology were examined using the following methods.

The ovarian tissue’s overall viability was assessed using a glucose uptake assay we described previously [22]. This assay quantitatively measures the glucose uptake by ovarian tissue during in vitro culture. Glucose uptake has been shown to be inversely correlated with the extent of (cryo)damage the tissue has sustained [22]. This assay has also been successfully used for determining the viability of human ovarian cortex tissue after cryopreservation [21].

For each individual measurement, four cortex biopsies were prepared using a 3-mm-diameter biopsy punch (Pfm Medical ag, Cologne, Germany). For each of these four ovarian biopsies, glucose uptake was determined by culturing them separately in 2 mL Dulbecco’s Modified Eagle Medium High Glucose (4.5 g/L) with L-Glutamine (DMEM, PAA Laboratories) supplemented with 10% fetal bovine serum (FBS; PAA Laboratories GmbH, Cölbe, Germany) and 0.1 % Pen/Strep (GIBCO; 10.000 units penicillin/mL and 10.000 μg streptomycin/mL) in a 24-well plate (TPP, Trasadingen, Switzerland) at 37 °C in humidified air with 5 % CO₂. The 2 mL of conditioned culture medium was collected from each individual biopsy, and replaced by fresh medium at day 4 and again collected at day 7, when the culture was ended. After culture, the biopsies were weighed separately and the glucose content of unconditioned medium (control) and of the conditioned medium after 4 or 7 days of culture were measured using an Architect i2000 system (Abbott Diagnostics, IL, USA). The glucose uptake of each biopsy during the culture period from day 0–4 and day 4–7 was determined by subtracting the glucose content of the conditioned medium from the unconditioned control medium. Glucose uptake was corrected for the weight (mg) of the biopsy and the duration (h) of the culture period [21, 22]. Then, for each series of four biopsies cultured under the same conditions, the mean glucose uptake/milligram/hour (nmol) was calculated representing an individual measurement on day 0–4, day 4–7, and (combined) day 0 to 7.

The viability of ovarian follicles was assessed in fresh tissue, or immediately after thawing, using a neutral red (NR) assay modified from Kristensen et al. [23]. With this metabolic assay, based on the ability of living cells to incorporate and bind neutral red in their lysosomes [24], living follicles were stained red, whereas dead follicles remained transparent (Fig. 1). A total of 10–15 fragments (<1 mm³) were cut from 3–4 ovarian cortex fragments of 5 × 5 × 1 mm, and incubated in 5 mL Ultraculture (Lonza) supplemented with 1 mg/mL Collagenase type IA (Sigma-Aldrich, Steinheim, Germany) at 37 °C for 1 h to soften and partially dissolve the tissue. Subsequently, the suspension was centrifuged at 400×g for 5 min after which the tissue pellet was resuspended in NR solution, consisting of 4.8 mL McCoy medium (Gibco, Life Technologies, New York, USA), 75 μL neutral red (to a final concentration of 50 mg/mL; Sigma-Aldrich), 2 μL Albuman (200 g/L; Sanquin, Amsterdam, the Netherlands), 50 μL Insulin-Transferrin-Selenium (ITS-G, Gibco), 200 units penicillin (Gibco), and 200 μg streptomycin (Gibco). The NR suspension containing the ovarian tissue was incubated for 90 min at 37 °C in humidified air with 5 % CO₂. Next, the partly dissolved tissue fragments were removed from the bottom of the tube and placed on a glass slide. By gently pressing a cover slip on the softened tissue fragments, a squash preparation was obtained. Living (stained red) as well as dead (non-stained) follicles were counted in this preparation using light microscopy. The percentage of living follicles was documented for each condition for which a NR viability assay was performed, provided that at least 20 follicles could be analyzed. Results obtained from samples with less than 20 follicles were considered unreliable and therefore excluded from further analysis. Using these criteria, the percentage of living follicles could be determined in the tissue of 16 patients while the tissue of the five remaining patients was excluded.

For histology, cortex fragments of 5 × 5 × 1 mm from 18 patients were fixed in Bouin’s solution (Sigma-Aldrich) and subsequently embedded in paraffin wax. Eight μm hematoxylin- and eosin-stained sections were examined using light microscopy and photographed. For each fragment, three sections were evaluated. Sections were separated by 100 μm of tissue to prevent counting the same pre-antral follicle twice. Follicles were scored as morphologically normal or degenerated (in case of cytoplasmatic shrinkage, disorganized granulosa cells, pyknotic nuclei) and categorized as primordial/primary, secondary, or antral according to a predefined criteria [25]. We focused at the percentage of degenerated/dead follicles rather than the total number of follicles observed per study arm as the density of follicles in various parts of the cortex of the same ovary may vary considerably [26].

A linear-mixed model for repeated measurements was used to study differences between protocols used for cryopreservation and for thawing on each of variables of the viability of human ovarian tissue, separately. In this way, it is possible to find small within-subject differences relative to a large between-subject (biological) variation. The dependent variable was the glucose uptake and the percentage living follicles, respectively. The independent categorical variables were cryopreservation (two protocols A_C, B_C), thawing (two protocols A_T, B_T), and the time point of measurement (two levels day 0–4, day 4–7). The intercept of each patient was treated as a random variable, in order to allow different levels for different patients. Initially, all interaction terms between the categorical variables were included in the linear part of the model. However, these terms were omitted from the final model presented as these did not statistically significant improve the fit to the data (Likelihood-Ratio test). The estimated mean differences between the levels of each variable with the appropriate 95 % confidence interval are presented. Statistical analyses were performed by using SAS® version 9.2 for Windows (SAS institute Inc. Cary, NC, USA).