Human ovarian cryopreservation: vitrification versus slow freezing from histology to gene expression

Abstract Cryopreservation of ovarian tissue is one of the strategies offered to girls and women needing gonadotoxic treatment to preserve their fertility. The reference method to cryopreserve is slow freezing; vitrification is an alternative method. The aim was to evaluate which of the two is the best method for human ovarian tissue cryopreservation. Each ovary was divided into three groups: (i) fresh; (ii) slow freezing; and (iii) vitrification. An evaluation of the follicular density, quality and the expression six genes (CYP11A, STAR, GDF9, ZP3, CDK2, CDKN1A) were performed. We observed no significant difference in follicular density within these three groups. Slow freezing altered the primordial follicles compared to the fresh tissue (31.8% vs 55.9%, p = 0.046). The expression of genes involved in steroidogenesis varied after cryopreservation compared to the fresh group; CYP11A was under-expressed in slow freezing group (p = 0.01), STAR was under-expressed in the vitrification group (p = 0.01). Regarding the expression of genes involved in cell cycle regulation, CDKN1A was significantly under-expressed in both freezing groups (slow freezing: p = 0.0008; vitrification: p = 0.03). Vitrification had no effect on the histological quality of the follicles at any stage of development compared to fresh tissue. There was no significant difference in gene expression between the two techniques.


Introduction
Ovarian tissue cryopreservation is a strategy to preserve fertility of young women suffering from cancer.Various treatments (chemotherapy, radiotherapy and surgery) can induce great damage to ovarian reserves (Jeruss & Woodruff, 2009) despite progress in oncology diagnosis and treatments.There are two methods for cryopreservation of ovarian tissue (OT): slow freezing and vitrification.Slow freezing is the reference method for cryopreservation of human OT, although it can reduce the ovarian follicular reserve and induce damage to stroma cells (Fabbri et al., 2010).Vitrification is a relatively new and alternative method which has attracted attention due to apparent advantages over slow freezing, for the preservation of both the cells and stroma matrix as well as the survival of a vast pool of viable quiescent follicles that represents the ovarian reserve of the graft (Chang et al., 2011;Herraiz et al., 2014;Keros et al., 2009;Sanfilippo et al., 2015;Ting et al., 2011;Xiao et al., 2010Xiao et al., , 2013)).
Over the last decade, many studies have attempted to compare slow freezing and vitrification but results have been inconsistent.Some studies have reported the superiority of slow freezing (Abir et al., 2017;Gandolfi et al., 2006;Lee et al., 2019;Oktem et al., 2011;Vatanparast et al., 2018), whilst others report no significant difference between the two methods (Amorim et al., 2012;Campos et al., 2016;Fabbri et al., 2016;Huang et al., 2008;Wang et al., 2008;Zhou et al., 2016).To our knowledge, these conflicting results appear to be due to conflicting methods employed to vitrify OT, as well as to differing evaluation criteria (follicular morphology, follicular vitality, apoptosis and gene expression).After literature review, we found a lack of information with regards to the impact of gene expression after cryopreservation.Our previous study highlighted that vitrification preserved follicular morphology better than slow freezing and led to genes being overexpressed, whilst slow freezing led to genes being under-expressed (Labrune et al., 2020).
The aim of the present study was to evaluate which of the two methods, slow freezing versus vitrification, is best for human ovarian tissue cryopreservation at histological and molecular levels.

Collection of ovaries
The use of human tissue for this study was approved by the Ethics Committee of Hospices Civils de Lyon .After obtaining informed consent, both oral and written, ovarian tissues were collected by hysterectomy and oophorectomy from five transsexual women (female to male) with an average age of 28 ± 6.6 years old (21-38), suffering from gender identity disorder at Centre Hospitalier Lyon Sud.Ovaries were immersed in Leibovitz L-15 V R (Eurobio, Couraboeuf, France) and transported to the laboratory at 10 C within 30 min.First, ovaries were cut in two hemi-ovaries using a scalpel in a Petri dish under sterile conditions.Then, the medulla was removed with a sterile chisel in order to obtain two hemi-cortexes.Finally, cortexes were cut into small pieces of 200 ± 20 mg, representing dimensions of approximately 5 mm (length)Â3 mm (width) 1 mm (thickness).The same ovary was cut into several pieces to be studied fresh (control tissue), after slow freezing and vitrification.

Slow freezing protocol and thawing procedure
OT was frozen according to the method described by our team that allowed live birth in ewes (Demirci et al., 2001)  Vitrification protocol and warming procedure OT were vitrified in a solution composed of DMSO, ethylene glycol (Sigma-Aldrich), SSS and sucrose (Dutscher, Brumath, France).The protocol was based on several vitrification protocols described in the literature (Huang et al., 2008;Suzuki et al., 2015).Fragments were incubated in equilibration solution (BM1 containing 5.58% (1 M) ethylene glycol, 3.55% (1 M) DMSO 2.50% SSS and 0.125 M (sucrose) for 5 min at room temperature between 20 and 23 C. Then they were placed into a second bath (BM1 containing 11.16% (2 M) ethylene glycol, 7.10% (1 M) DMSO, 5.00% SSS and 0.25 M sucrose) for 7 min at room temperature between 20 and 23 C, after which they were placed in the vitrification solution (BM1 containing 22.32% (4 M) ethylene glycol, 14.20% (2 M) DMSO, 10% SSS and 0.5 M sucrose) at 4 C for 10 min.OT were placed on a piece of semi-rigid 1 mm thick absorbent paper developed by our team, then cooled by direct contact with liquid nitrogen.For warming, OT were removed from straws and placed in a warming solution of 0.4 M sucrose for 5 min at room temperature, and then in sucrose-free BM1 for 5 min at room temperature.Warmed OT were placed into 4% (v/v) formaldehyde for investigation of follicle morphology.

Histological evaluation
Five fresh, five frozen/thawed and five vitrified/ warmed OT were fixed in 4% (v/v) formaldehyde for 24 h at room temperature, paraffin embedded after dehydration, and cut into 4 mm serial sections.Ten sections were stained with haematoxylin (Millipore, Burlington, VT), eosin (Sigma-Aldrich) and safran (RAL Diagnostic, Martillac, France).The entire section was photographed to make a count and classification of the follicles using the Image J software.Then sections were checked by light microscopy to evaluate follicle morphology.The follicles were counted by two blind operators and were classified according to the description reported by Gougeon (1986) and counted as altered or intact (Figure 1).Follicles were classified as altered if there were at least one sign of oocyte or granulosa cell degeneration: the presence of pycnotic oocyte or follicular cell nuclei, detachment of the oocyte from surrounding granulosa cells, vacuolization of ooplasm, partially degenerated granulosa cells, or detachment of the basal membrane.

RNA extraction and complementary DNA synthesis for molecular assessment
Total RNA was extracted from fresh OT, frozen/thawed OT and vitrified/warmed OT using a RNeasy V R Plus Mini kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions.We used the whole warmed tissue.Before the RNA samples were treated with DNase to remove any genomic DNA contamination prior to proceeding DNA synthesis.RNA was eluted twice in 30 mL of RNase free-water after extraction and consecutively collected in a final volume of 30 mL and was conserved at À80 C. The RNA concentration was determined by spectrophotometry; the ratio of the readings at 260 and 280 nm (A260/A280) provided an estimate of the purity of RNA.A total of 500 ng of the extracted RNA was used for cDNA synthesis using the High-Capacity RNA-to-cDNA V R kit (Applied Biosystem, Waltham, Massachusetts, US).The cDNA synthesis was performed at 37 C for 60 min and stopped by heating to 95 C for 5 min.The obtained cDNA was stored at À20 C and prepared for real-time PCR analyses.

Real-time reverse transcription polymerase chain reaction
The primers for real-time RT-PCR were found by the literature review (Table 1) or by our own design.Each primer was verified by using the University California, Santa Cruz (UCSC) Genome browser (https//genome.ucsc.edu) to check their specificity, target region and size.Only verified primers were used for the PCR analyses.A total of six genes were analyzed: (i) CYP11A (Cytochrome P450 Family 11 Subfamily A Member 1); (ii) STAR (Steroidogenic Acute Regulatory Protein); (iii) GDF9 (Growth Differentiation Factor 9); (iv) ZP3 (Zona Pellucida Glycoprotein 3); (v) CDK2 (Cyclin-Dependant Kinase 2); and (vi) CDKN1A (Cyclin-Dependant Kinase Inhibitor 1A).We chose to evaluate GDF9 and ZP3 due to their dominant role in follicular development, in which they code for proteins produced by the oocyte.CYP11A and STAR encode proteins secreted by granulosa cells and are essential for the proper functioning of steroidogenesis.The expression of CYP11A allows the production of pregnenolone by enzymatic cleavage of the cholesterol side chain.Similarly, STAR expression plays a role in the production of pregnenolone by controlling the entry of cholesterol into the mitochondria.Finally, CDK2 and CDKN1A are involved in cell cycle regulation.CDK2 is involved in the control of the cell cycle; essential for meiosis, but non-essential for mitosis.
CDKN1A encodes an inhibitory protein to regulate cell growth and cell response to DNA damage.One-step RT-PCR was performed using the StepOneplus real-time thermal cycler (Applied Biosystems) and using the Quantitect SYBR Green RT-PCR kit (Thermofisher, Waltham, USA).The reference gene was GAPDH (Glyceraldehyde-3-Phosphate Dehydrogenase).Prior to quantitative analysis, optimization procedures were performed by running real-time RT-PCR with or without a template to verify the reaction condition, including the annealing temperatures of the primers and specific products.The real-time thermal conditions included a holding step at 50 C for 2 min and 95 C for 2 min, a cycling step at 95 C for 15 s, 60 C for 15 s, followed by a melt curve step 95 C for 15 s, 60 C for 1min and 95 C for 15 s.Each sample was analyzed in triplicate; water was used as negative control.

Statistical analysis
The analysis was performed with the R software.In order to compare non-independent proportions a Generalised Linear Mixed Model (GLMM) was used.In this way, the ratios estimated by the GLMM were not a simple division of two values but the inverse of a logit function.However, the estimated ratios by the GLMM were near the ratios calculated by a simple division.Moreover, in order to take account of the overdispersion of the values, a quasibinomial distribution was used for the calculation and the test of the difference amongst the ratios.Thus, a precise analysis was performed, with ratios correctly calculated.To perform this study the so-called glmmPQL function of the MASS package was used.The Kruskal-Wallis non-parametric test was used to analyze the results of real-time RT-PCR data.A test was considered statistically significant when the p value was under 0.05.

Follicular density
There was no significant difference observed in the initial follicular density within the three groups (fresh, slow freezing and vitrification).The percentage of primordial follicle was higher in the slow freezing and vitrification group compared to the fresh group without significant difference, it was secondary to the heterogeneity of the ovarian tissue (Table 2).

Follicle morphology
We classified follicle quality using two conditions: intact or damaged at different stages of development: primordial, primary, intermediate and secondary follicle.Slow freezing significantly altered the pool of primordial follicles compared to the fresh tissue (31.8% vs 55.9% p ¼ 0.046, respectively) whilst this effect was not statistically significant after vitrification (48.8%, p ¼ 0.508).The percentage of intact primordial follicle was higher but not significantly after vitrification compared to slow freezing (p ¼ 0.129).The analysis of intermediate follicles was not possible because of the low number of follicles in the slow-freezing group (0.3 follicles/mm 2 ± 0.7).Moreover, we did not observe intermediate follicles in the vitrification group.There was no negative impact on the percentage of primary follicle after slow freezing and vitrification compared to the fresh group (slow freezing 27.6% p ¼ 0.085; vitrification 89.05% p ¼ 0.690).The percentage of intact primary follicle was higher in the vitrification group than slow freezing but not significant (p ¼ 0.064).The secondary follicle results were difficult to study because of the low number of follicles.We observed secondary follicles were more damaged after slow freezing compared to vitrification.(Table 2).

Gene expression
Concerning the genes expressed by the granulosa cells, the mean ± SEM fold change of CYP11A was 0.53 ± 0.18 in frozen OT and 0.46 ± 0.50 in vitrified OT; the change after slow freezing was significantly lower (p ¼ 0.01) and there was no significant difference between vitrified OT and fresh tissue (p ¼ 0.12).The mean ± SEM fold change of STAR mRNA was 2.25 ± 3.41 in frozen OT and this was 0.38 ± 0.22 in vitrified OT; there was no significant difference between frozen OT and fresh tissue (p ¼ 0.47) and the change after vitrification was significantly lower (p ¼ 0.01).Concerning the genes expressed by the oocytes, the mean ± SEM fold change of GDF9 was 1.81 ± 1.90 in frozen OT and this was 1.47 ± 1.43 in vitrified OT; there was no significant difference between cryopreserved OT and fresh tissue (p ¼ 0.39 and p ¼ 0.50).The mean ± SEM fold change of ZP3 mRNA was 1.47 ± 0.73 in frozen OT and this was 1.82 ± 2.55 in vitrified OT; there was no significant difference between cryopreserved OT and fresh tissue (p ¼ 0.28 and p ¼ 0.51).Concerning cell cycle gene, the mean fold change ± SEM of CDK2 was 1.38 ± 1.53 in frozen OT and this was 1.41 ± 2.01; there was no significant difference between cryopreserved OT and fresh tissue (p ¼ 0.60 and p ¼ 0.67).The mean ± SEM fold change of CDKN1A mRNA was 0.45 ± 0.13 frozen OT and this was 0.58 ± 0.29 in vitrified OT; the change after slow freezing and vitrification were significantly lower (p ¼ 0.0008 and p ¼ 0.03), as per Figure 2.

Discussion
Vitrification is a simple alternative method for cryopreservation of ovarian tissue.In our study we worked on human ovaries from transsexual women.This model has also been used in the literature (Ramezani  et al., 2017;Shams Mofarahe et al., 2015, 2017).The neutrality of androgenic treatments on follicular and tissue quality has been reported (Walters, 2015;Walters et al., 2019).Few studies have studied women without prior androgenic treatment (Abdollahi et al., 2013;Isachenko et al., 2009;Wang et al., 2016).We have shown a greater alteration of the primordial follicles after slow freezing compared to fresh tissue.Our study showed slow freezing would seem to affect the normal morphology of primary and secondary follicle.Vitrification had no effect on the histological quality of the follicles on the human ovary at any stage of development compared to the fresh tissue which agrees with previous studies (Abdollahi et al., 2013;Herraiz et al., 2014;Keros et al., 2009;Shams Mofarahe et al., 2015, 2017;Shi et al., 2017;Wang et al., 2016).Abdollahi et al. (2013) reported the absence of apoptosis at the histological and molecular level after vitrification compared to the fresh tissue (Abdollahi et al., 2013).
We also found a decrease in granulosa cell gene expression after slow freezing and vitrification.In Wang's study the results are similar: they observed a decrease in CYP11A and STAR expression after vitrification and slow freezing (Wang et al., 2016).
We also observed a decrease in the expression of the CDKN1A gene, a gene that inhibits cell cycle initiation.This gene was under-expressed by both cryopreservation techniques.To our knowledge, no studies have evaluated the expression of cell cycle genes after vitrification and slow freezing on human ovarian tissue.In the bovine model, CDKN1A expression was decreased after cryopreservation techniques (Wang et al., 2017).Under-expression can trigger the cell cycle.An initiation of primordial follicles into growth follicles after tissue transplantation is reported in the literature, which is secondary to an interruption of the PI3K/Akt/mTOR when dissecting ovarian tissue (Ayuandari et al., 2016;Maidarti et al., 2020;Masciangelo et al., 2020).This pathway plays a role in the cell cycle (Grosbois & Demeestere, 2018;Masciangelo et al., 2020).The CDKN1A gene is not reported as a target for this signalling pathway.Its change in expression may also be an explanation for this entry into reserve follicle growth.Functional analyses are required to complete these results.Oocyte gene expression was not altered after cryopreservation techniques.This result was similar in the ovine model.The absence of effect of slow freezing and vitrification on oocyte genes is rather reassuring.Wang et al. (2016) observed a decrease in ZP3 expression after cryopreservation techniques.Our study only focussed on two genes, which is not enough to know the real impact on the oocyte.Two authors studied the effect of vitrification after in vitro tissue culture with histological and molecular analysis.
They observed an over-expression of FSH and GDF9 genes, which reflect the quantity of primary and secondary follicles (Shams Mofarahe et al., 2015, 2017) and under-expression of FIGLA and KIT-L which reflect the quantity of primordial follicles (Ramezani et al., 2017).These results show a resumption of folliculogenesis after vitrification of the ovarian tissue.Vitrification of human ovarian tissue is equivalent to slow freezing.This initial work highlights the absence of deleterious effects of vitrification compared to slow freezing on human ovarian tissue just after warming.The longterm effect with functional studies remains to be defined and will be the subject of future work.

pp
Value for slow freezing versus fresh.bValue for vitrification versus fresh.

Table 1 .
Oligonucleotide primer sequences for PCR.