Evaluation of a new freezing protocol containing 20% dimethyl sulphoxide concentration to cryopreserve human ovarian tissue

Accepted Manuscript

Evaluation of a new freezing protocol containing 20% dimethyl sulphoxide concentration to cryopreserve human ovarian tissue Miguel Gallardo , Fernanda Paulini , Ariadna Corral , Marcin Balcerzyk , Carolina M. Lucci , Jer Ambroise , ´ ome ˆ Marta Merola , Laura Fernandez-maza , Ramon Risco , ´ Marie-Madeleine Dolmans , Christiani A. Amorim PII: DOI: Reference:

S1472-6483(18)30488-7 https://doi.org/10.1016/j.rbmo.2018.09.012 RBMO 2020

To appear in:

Reproductive BioMedicine Online

Received date: Revised date: Accepted date:

26 September 2017 31 August 2018 4 September 2018

Please cite this article as: Miguel Gallardo , Fernanda Paulini , Ariadna Corral , Marcin Balcerzyk , Carolina M. Lucci , Jer Ambroise , Marta Merola , Laura Fernandez-maza , Ramon Risco , ´ ome ˆ ´ Marie-Madeleine Dolmans , Christiani A. Amorim , Evaluation of a new freezing protocol containing 20% dimethyl sulphoxide concentration to cryopreserve human ovarian tissue, Reproductive BioMedicine Online (2018), doi: https://doi.org/10.1016/j.rbmo.2018.09.012

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Evaluation of a new freezing protocol containing 20% dimethyl sulphoxide concentration to cryopreserve human ovarian tissue Miguel Gallardo,a,b,1 Fernanda Paulinia,c,1 Ariadna Corrald Marcin Balcerzykd Carolina M Lucci,c Jérôme Ambroise,e Marta Merola,a Laura Fernandez-mazad Rámon Riscod,f MarieMadeleine Dolmans,a,g Christiani A. Amorima,* a

AN US

CR IP T

Pôle de Recherche en Gynécologie, Institut de Recherche Expérimentale et Clinique, Université Catholique de Louvain, Avenue Mounier 52, bte. B1.52.02, 1200 Brussels, Belgium b Ginemed Clínicas Sevilla, 41010, Sevilla, Spain c Physiological Sciences Department, Institute of Biological Sciences, University of Brasília, Brasília, DF, Brazil d National Center for Accelerators, Seville, Spain e Institut de Recherche Expérimentale et Clinique and Centre de Technologies Moléculaires Appliquées, Université Catholique de Louvain, Brussels, Belgium f Engineering School of Seville, Seville, Spain g Gynecology Department, Cliniques Universitaires Saint-Luc, 1200, Brussels, Belgium

*Corresponding author. E-mail address: [email protected] (C. Amorim).

The first two authors should be considered joint first authors.

M

1

ED

Key message

A new freezing protocol using 1.8 ml of cryopreservation solution, 20% dimethyl sulphoxide

PT

and a seeding temperature of –11°C does not enhance the survival of preantral follicles after

CE

xenotransplantation, compared with the original protocol using a 0.8-ml cryopreservation

AC

volume, 10% dimethyl sulphoxide concentration and seeding temperature of –8°C.

Abstract

Research question: Could a modification in the ovarian tissue freezing protocol improve follicle survival after cryopreservation and xenotransplantation? Design: Ovarian tissue was used from 13 adult patients, frozen either with our original protocol, or a modified version involving a higher concentration of dimethyl sulphoxide (DMSO), larger volume of cryopreservation solution and lower seeding temperature. After 1

ACCEPTED MANUSCRIPT thawing, the ovarian fragments were xenotransplanted to six mice with severe combined immunodeficiency (SCID) for 3 weeks. Results: The proportion of primordial follicles decreased, and the proportion of growing follicles increased significantly (all P < 0.01) after cryopreservation and xenografting compared with fresh controls for both protocols. Follicle density, development, ultrastructure

CR IP T

and function were similar between treatments. Conclusions: This study showed that, although the higher DMSO concentration did not

improve survival of preantral follicles, it did not seem to induce any major toxicity in the

AN US

follicle population either.

KEYWORDS: fertility preservation, ovarian tissue cryopreservation, cancer patients, preantral follicles, xenotransplantation, freezing protocol.

M

Introduction

Cryopreservation and transplantation of human ovarian cortical tissue is a crucial strategy in

ED

fertility preservation, being the only option for patients with certain types of cancer, as well as prepubertal girls (Pfeifer et al., 2014; Wallace et al., 2016). Several studies have shown that

PT

this procedure leads to restoration of ovarian function and fertility (Dolmans et al., 2013;

CE

Donnez and Dolmans, 2013), and more than 100 live births have been reported worldwide

AC

using this technique (Jadoul et al., 2017; Jensen et al., 2017).

In most cases, conventional freezing protocols have been applied to cryopreserve human ovarian tissue. Although all live births but two (Suzuki et al., 2015) resulted from frozen– thawed ovarian tissue, it is well known that this procedure can be a source of iatrogenic damage, impairing tissue viability by causing follicle and stromal cell death and increased

2

ACCEPTED MANUSCRIPT fibrosis (Nisolle et al., 2000; Nottola et al., 2008; Keros et al., 2009; Amorim et al., 2011; David et al., 2012). Such cryoinjuries may also exacerbate ischaemic damage after grafting, acting in a synergic fashion (Lee et al., 2016) that will, in turn, negatively affect the population of grafted follicles (Israely et al., 2006; David et al., 2012). This is probably a result of the empirical nature of slow-freezing formulations; current protocols used for

CR IP T

cryopreservation of human ovarian tissue have been mostly adapted from protocols developed for ovarian tissue from other mammalian species, and human oocytes and embryos (Gosden et al., 1994; Hovatta et al., 2005). Indeed, most protocols use a similar concentration (around 10%) of permeable cryoprotectant (CPA) and comparable cooling curve (Meirow et al., 2007;

AN US

Van Eyck et al., 2010; Rosendahl et al., 2011; Van der Ven et al., 2016; Wallace et al., 2016).

In a series of experiments conducted with bovine ovarian tissue, we demonstrated that by

M

increasing the dimethyl sulphoxide (DMSO) concentration to 20%, increasing the CPA volume to 1.8 ml and decreasing the seeding temperature (–11°C), CPA permeation was

ED

improved and ice formation was reduced in ovarian tissue (Corral et al., 2018). It is known that DMSO concentrations above 10% may have a negative effect on different cell types

PT

(Best, 2015), but 20% DMSO did not decrease the percentage of morphologically normal

AC

2004).

CE

preantral follicles in bovine ovaries, nor did it affect their oocyte ultrastructure (Lucci et al.,

As bovine ovaries closely resemble those of humans, with a similar composition and comparable follicle size and growth patterns (Amorim et al., 2011), we hypothesized that CPA permeation in human ovarian tissue could be improved, which could then have a positive influence on the survival and development of preantral follicles after transplantation. In an attempt to confirm our hypothesis, we assessed CPA permeation (computed

3

ACCEPTED MANUSCRIPT tomography), follicle survival and growth (histology and immunohistochemistry), ultrastructure (transmission electron microscopy) and tissue quality (histology and immunohistochemistry), after cryopreservation and short-term xenotransplantation of human

CR IP T

ovarian cortex.

Materials and methods Experimental design

Ovarian biopsies from 13 adult women were cryopreserved using either our routine procedure

AN US

or a new freezing protocol. Because of the small size of the biopsies, they were alternately assigned to one or the other of the two protocols. Only in one patient did we obtain a biopsy large enough to divide equally into two pieces and use in both protocols. Thus, ovarian

M

fragments from six patients were subjected to our routine procedure and seven to the new freezing protocol (Supplementary Table 1). Frozen samples from both groups were analysed

ED

by computed tomography to calculate CPA concentrations in the tissue. After thawing and xenografting to mice with severe combined immunodeficient (SCID), the tissue fragments

PT

were fixed and evaluated. A longer period of transplantation (3 weeks) was chosen to

CE

investigate possible long-term effects of the freezing protocol on tissue viability and survival. Histological and immunohistochemical analyses for the purposes of follicle counting and

AC

classification, as well as calculation of fibrotic and vascularized areas, were conducted blindly by four observers (MG, FP, MM and CAA). The experimental design is presented in Supplementary Figure 1. Percentage values for solutions referred to throughout the ‘Materials and methods’ section are based upon volume percentages.

Ethics

4

ACCEPTED MANUSCRIPT Use of human ovarian tissue was approved by the Institutional Review Board of the Université Catholique de Louvain on 28 November 2016 (IRB reference 2012/23MAR/125, registration number B403201213872). Guidelines for animal welfare were approved by the Committee on Animal Research of the Université Catholique de Louvain on 19 June 2014

CR IP T

(reference 2014/UCL/MD/007).

Collection of ovarian tissue

Ovarian biopsies were taken from 13 women (between 26 and 35 years of age) after obtaining informed consent. All patients underwent laparoscopic surgery for benign gynaecological

AN US

disease. Biopsies were immediately transported on ice to the laboratory in minimal essential medium plus GlutaMAXTM (MEM; Gibco, Invitrogen, Merelbeeke, Belgium). Once in the laboratory, the medullary part of each biopsy was removed and the cortex was cut into two

M

small fragments (5 x 5 x 1 mm) for freezing, and a smaller piece for immediate fixation in

ED

formalin (fresh control).

Ovarian tissue freezing

PT

Freezing of ovarian tissue was conducted using our routine protocol (protocol 1) (Donnez et

CE

al., 2004) or a new protocol (protocol 2) developed by Corral et al. (2018). Protocol 1

AC

Freezing of ovarian strips was carried out as previously described (Amorim et al., 2009). The strips were suspended in CPA solution consisting of minimum essential medium (MEM) supplemented with 4 mg/ml human serum albumin (HSA), (Sanquin, Amsterdam, the Netherlands), and 10% DMSO (Sigma, Bornem, Belgium) at 4°C on a refrigerated plate (Omega Services, Belgium), before being transferred to 2-ml cryovials (reference number: T311-2; Simport, Quebec City, QC, Canada) containing 0.8 ml CPA solution (one strip per

5

ACCEPTED MANUSCRIPT cryovial). The cryovials were cooled in a programmable freezer (Kryo 10, Series III; Planer, Sunbury-on-Thames, UK) using the following programme: soaked at 0°C for 30 min (tissue equilibration in CPA solution); cooled from 0oC to –8oC at –2oC/min; seeded manually; cooled to –40o at –0.3oC/min; and cooled to –140°C hat –30°C/min and transferred to liquid

CR IP T

nitrogen (–196°C) for storage.

Protocol 2

The strips were suspended in CPA solution consisting of MEM supplemented with 4 mg/ml HSA and 20% DMSO at 4°C on a refrigerated plate, before being transferred to 2 ml

AN US

cryovials containing 1.8 ml CPA solution (one strip per cryovial). The cryovials were cooled in a programmable freezer (Freezer Control CL-8800i, CryoLogic, Victoria, Australia) using the following programme: soaked at 0°C for 30 min (tissue equilibration in CPA solution);

M

cooled from 0oC to –11oC at –2oC/min; seeded manually; cooled to –40oC at –0.3oC/min; cooled to –140°C at –30°C/min and transferred to liquid nitrogen (–196°C) for storage. The

ED

seeding temperature was selected based on tests conducted with different temperatures; –11°C

PT

generated ice formation in a similar way to –8°C when seeding with 10% DMSO solution.

CE

It is important to stress that both freezing machines were fully validated during our previous studies (Donnez et al., 2008; Martinez-Madrid et al., 2009; Van Eyck et al., 2010; Amorim et

AC

al., 2012).

Computed tomography scanning of frozen ovarian tissue To assess DMSO concentrations in cryopreserved fragments of ovarian tissue, cryovials containing the samples were analysed by X-ray computed tomography in a specially designed chamber cooled to cryogenic temperatures (–140°C; below the glass transition point), as

6

ACCEPTED MANUSCRIPT described by Corral et al. (2015). Attenuation of the signal recorded by the computed tomography scan (NanoCT scanner, Bioscan, USA; currently Mediso, Hungary) is directly proportional to the CPA concentration, providing information on the degree of DMSO diffusion into tissue and its homogeneity.

CR IP T

The acquisition process is described in depth elsewhere (Corral et al., 2015). In brief, each image required 3 min of acquisition time with a 106 mA current, a voltage of 75 kV and 500 ms of exposure time per projection, performing 360 projections per rotation. Images were reconstructed to obtain a spatial resolution of 0.1 mm. Nucline software (Mediso, Budapest,

reconstruction (Invicro, Boston, USA).

AN US

Hungary) was used for image acquisition, and IVS image processing software for

M

For image visualization, computed tomography attenuation coefficients were transformed to a PMOD cold scale (PMOD 3.7, PMOD Technologies LLC, Zurich, Switzerland); dark blue

ED

corresponds to the lowest attenuation, passing through green, orange and red, before reaching dark red for the highest attenuation. To locate the tissue in the cryovials, pictures were taken

PT

of each frozen cryovial and compared with computed tomography images. Moreover, a 3 x 3

CE

x 1 mm volume of interest (VOI) was chosen to ensure that all the VOI actually contained tissue, as the dimensions of tissue are larger (around 5 x 5 x 1 mm). After positioning the VOI

AC

in the selected area, we checked each section in every direction to confirm that all points contained within the VOI corresponded to the darker areas representing tissue. Four frozen ovarian tissue fragments from each protocol were analysed to determine DMSO permeation. After computed tomography scanning, ovarian tissue samples were again stored in liquid nitrogen. After thawing, one sample from each protocol was subjected to computed tomography scanning to assess the efficacy of the washing steps in removing the CPA.

7

ACCEPTED MANUSCRIPT

Ovarian tissue thawing The thawing procedure was the same for both groups. Cryovials were exposed to room temperature for 2 min and immersed in a water bath at 37°C until the ice completely melted. To remove the CPA solution, the ovarian tissue was immediately transferred from the

min per bath at room temperature) before xenografting.

CR IP T

cryovials to plastic Petri dishes containing 5 ml of MEM, where it was washed three times (5

Xenotransplantation to mice with severe combined immunodeficiency

AN US

Six 6-week-old female mice with SCID (Charles River Laboratories, L'Arbresle, France) were used for the study. Three to four mice were housed per cage under high-efficiency particulate air filtered hoods in rooms maintained at 28°C with a 12-h light–12-hour dark

M

cycle. They were fed laboratory chow and acidified water ad libitum. All housing materials,

ED

food and water were autoclaved before use.

The xenografting procedure was carried out as previously described (Amorim et al., 2011).

PT

Two ovarian grafts were stitched using non-absorbable sutures (6/0 Prolene; Ethicon, Johnson

CE

and Johnson International, Diegem, Belgium) to the anterior wall of the peritoneum at the level of the bladder, one from protocol 1 on the right, and one from protocol 2 on the left

AC

(Supplementary Figure 1). After 3 weeks, the animals were killed by CO2 asphyxiation and the grafts were recovered and fixed in formalin overnight at 4°C.

Histological analysis Fresh and frozen–thawed xenografted tissue were used for histological analysis. After fixation. The ovarian fragments were dehydrated, embedded in paraffin and serially sectioned

8

ACCEPTED MANUSCRIPT (5-µm-thick sections) to evaluate follicular morphology. Every fourth slide was stained with haematoxylin and eosin (Merck, Darmstadt, Germany) for histological evaluation; the other slides (Superfrost Plus slides, Menzel-Glaser, Braunschweig, Germany) were kept for immunostaining.

CR IP T

Quality of follicles were evaluated on the basis of the integrity of the basement membrane, cellular density, presence or absence of pyknotic bodies and integrity of the oocyte. These criteria were used to classify preantral follicles as morphologically normal or atretic. The percentage of each type of follicle was calculated in both groups. Morphologically normal

AN US

follicles (MNF) were further classified according to stage into primordial or growing

(primary, secondary and antral) follicles. Primordial follicles are composed of an oocyte surrounded by a single layer of flattened granulosa cells, whereas growing follicles have one

Follicle density

ED

M

or more layers of cuboidal granulosa cells around the oocyte.

PT

Follicle density was estimated by counting ovarian follicles in three randomly selected 1-mm² areas using at least three sections (extremities and middle) of each fresh and grafted ovarian

CE

tissue fragment from every animal (Amorim et al., 2013). Sections were scanned using the Mirax scanner (Zeiss, Jena, Germany), and pictures were taken at X100 magnification. Using

AC

the grid tool from ImageJ, a freely available image-processing and analysis programme developed at the National Institutes of Health (http://rsb.info.nih.gov/ij/), 1-mm² squares were randomly distributed across the picture, and captured follicles were counted and classified as primordial, growing or atretic.

Fibrosis

9

ACCEPTED MANUSCRIPT Relative areas of fibrosis were evaluated on 3-5 Masson’s trichrome-stained sections (Amorim et al., 2012), depending on the size of the graft: one section from each extremity and one to three sections from the middle of the graft. Fibrotic areas were characterized by poor cellularity, as shown by a reduced number of cell nuclei and collagen deposits, as previously described (Dath et al., 2010). Masson’s trichrome staining turns tissue green, indicating that it

CR IP T

has been replaced with collagenous connective tissue, rendering fibrotic areas easily recognizable.

Sections were scanned using the Mirax scanner, and measurement of fibrotic and total section

AN US

areas was carried out using the Mirax Viewer programme. Areas were delimited with the freehand tool and then measured (Amorim et al., 2012).

M

Immunohistochemistry Follicle survival and development

ED

Several markers were selected to assess follicle survival (caspase-3), growth (Ki67) and function (anti-Müllerian hormone [AMH], kit ligand and its receptor, c-kit). Paraffin sections

PT

were deparaffinized with Histosafe (Yvsolab SA, Beerse, Belgium) and rehydrated in alcohol

CE

series. After blocking endogenous peroxidase activity with 0.3% H2O2 diluted in demineralized water (for Ki67 and caspase-3), 3% H2O2 diluted in demineralized water (for

AC

kit ligand and c-kit), or methanol (for AMH), a demasking step was carried out for 75 min at 98°C with citrate buffer and Triton X100 before the sections were subjected to an antigen retrieval step. Antibodies, incubation conditions and positive and negative control tissues are presented in Table 1. Tissue sample slides used as controls were donated by our university biolibrary. Diaminobenzidine was used as a chromogen (SK 4100, Vector Laboratories, Peterborough, UK). The slides were then counterstained with haematoxylin and mounted with

10

ACCEPTED MANUSCRIPT DPX neutral mounting medium (Prosan, Merelbeke, Belgium). FLEX Negative Control mouse (IS75061-2; Dako, Glostrup, Denmark) and rabbit Flex Universal negative control (IS60061-2; Dako) were used for negative controls (human ovarian tissue) with, and in place of,h the primary antibody, as appropriate.

CR IP T

Follicles containing at least one granulosa cell staining positive for Ki67 were classified as proliferative follicles. For quantitative analysis of AMH, kit ligand and c-kit expression, a minimum of 10 follicles were evaluated for each patient and treatment group. Follicles were considered AMH- or kit-ligand-positive when at least one granulosa cell was immunostained,

AN US

and c-kit-positive when the oocyte was immunostained. For caspase-3 immunostaining,

follicles were considered positive when more than 50% of granulosa cells, the oocyte, or both,

Graft vascularization

M

were caspase-3-positive.

ED

To concomitantly visualize human and murine vessels in grafts and analyse graft vascularization, double anti-human and anti-mouse CD34 immunohistochemistry was carried

PT

out as previously described by Soares et al. (2015). After deparaffinizing and rehydrating

CE

paraffin sections, endogenous peroxidase activity was blocked with 0.3% H2O2 diluted in demineralized water. The slides were incubated with 10% normal goat serum (NGS) plus 1%

AC

bovine serum albumin (BSA) to block non-specific binding sites (30 min at room temperature), before incubation with the primary antibody, rat anti-mouse CD34 (Table 1). They were then incubated for 60 min at room temperature with the first secondary antibody, rabbit anti-rat, coupled to biotin (Vector). Diaminobenzidine (Dako) was used as a chromogen after incubation with a solution of streptavidin–horseradish peroxidase for 30 min. The slides were incubated again with 10% normal goat serum plus 1% bovine serum albumin to block

11

ACCEPTED MANUSCRIPT non-specific binding sites, before incubation with the second (anti-human) primary ntibody, mouse anti-human CD34 (Table 1), overnight at 4°C. Sections were then incubated with the second secondary antibody, goat anti-mouse (1:300 dilution; Jackson Immunoresearch, Suffolk, UK). Fast Red TR (Sigma) was used as a chromogen and nuclei were counterstained with haematoxylin. Human liver and mouse ovary were used as positive control and rabbit

CR IP T

Flex Universal negative control (IS60061-2; Dako) and IgG from rat serum (I4131; Sigma) were used for negative controls, in place of the primary antibody, as appropriate.

Depending on the size of grafts, three to five slides were scanned using the Mirax scanner and

AN US

visualized using Mirax Viewer software. The area around each ovarian tissue graft was

defined and vessel area was measured. Total vessel area and total graft area were calculated in each graft, yielding a single vessel area value and single graft area value for each graft, based

M

on the sum of the analysed fragments.

ED

Ultrastructure analysis

The ultrastructure of preantral follicles was evaluated in fresh controls and in ovarian tissue

PT

after freezing, thawing and xenografting. After fixation in Karnovsky solution, samples were

CE

postfixed with 1% osmium tetroxide in Dulbecco’s phosphate-buffered saline (PBS), en bloc contrasted with uranyl acetate, dehydrated through an ascending series of ethanol, immersed

AC

in propylene oxide (solvent substitution), embedded in Epon 812 (Agar Scientific, Brussels, Belgium), and then sectioned using a Leica EM UC7 ultramicrotome (Wetzlar, Germany). Semithin (3-μm-thick) sections were stained with toluidine blue and examined by light microscopy (Zeiss Axioskop, NY, USA) to detect follicles. Ultrathin (70-nm) sections were examined under a JEM-2100 transmission electron microscope (Jeol, Tokyo, Japan). The general aspect of oocytes and granulosa cells, the distribution and appearance of organelles,

12

ACCEPTED MANUSCRIPT the integrity of membranes, and connections between granulosa cells, as well as between granulosa cells and the oocyte, were evaluated.

Statistical analysis Statistical analyses were carried out to compare the effect of each protocol on each outcome

CR IP T

(follicle density, survival and development and graft fibrosis and vascularization) using a specially constructed linear mixed-effects regression model. The protocol used (protocol 1 versus protocol 2) and the follicle source (fresh versus cryopreserved and xenografted tissue)

take account of interpatient variability.

AN US

were included as fixed effects in each model, whereas a random patient effect was included to

First-order interaction effects between the follicle source (fresh versus cryopreserved and

M

xenografted tissue) and the protocol (protocol 1 versus protocol 2) were also incorporated in each model and discarded when non-significant. For each outcome, we evaluated if log

ED

transformation was needed to meet the assumptions of the statistical model (namely residuals with normal distribution and homogeneity of variance). When log transformation was not

PT

required, the model’s coefficients represented the additive effect of the corresponding

CE

predictor. When log transformation was required, the model’s coefficients were backtransformed to compute the multiplicative effects (namely fold change) of the corresponding

AC

predictor.

To compare DMSO concentrations in tissues cryopreserved using the two different protocols, Student’s t-test was applied. Values were considered statistically significant when P < 0.05.

13

ACCEPTED MANUSCRIPT Results Computed tomography scanning Computed tomography scans of each protocol in PMOD cold scale are presented in Figure 1; the colour ranges from dark blue for low attenuation to intense red for high attenuation, with a 3 x 3 x 1 mm3 VOI in pink located where the sample was placed. The colour of these areas

CR IP T

corresponds to intratissular DMSO concentrations. In both images of frozen tissues, extracellular ice, formed during the slow-freezing procedure, is clearly visible in dark blue with a striped pattern (Figure 1). In the middle of these ice crystals, some islands (in green, yellow and red) show higher attenuation, and hence higher DMSO concentrations, that have

AN US

formed because of the extracellular ice. In tissue areas, the striped pattern is much less

pronounced, indicating a low probability of ice formation in these tissues. Even though the concentration (colour) is not completely homogeneous (blue, green and yellow) in tissues

M

cryopreserved with the two protocols, tissue frozen using protocol 2 showed a higher DMSO concentration than tissue frozen using protocol 1. Therefore, formation of ice is less likely in

ED

tissues cryopreserved with an initial concentration of 20% DMSO. Average DMSO concentrations in tissues cryopreserved with the two protocols (protocol 1: 11.0%; protocol 2:

PT

19.2%) were calculated from VOI. The distribution of computed tomography attenuation

CE

values of pixels within the VOI for one tissue fragment frozen per protocol were compared. The average value was lower (P < 0.05) with protocol 1 (1.57 ± 0.10 computed tomography

AC

units) than protocol 2 (1.76 ± 0.08 computed tomography units), indicating a lower DMSO concentration, as expected.

Computed tomography scans were also applied to evaluate the presence of DMSO in tissue fragments after thawing. With both protocols, whole vials were a homogeneous dark blue colour, showing the lowest possible attenuation (same as that of water). This demonstrates

14

ACCEPTED MANUSCRIPT that the thawing procedure was able to effectively remove DMSO from samples in both protocols (Figure 1).

Follicle survival and development On average, follicle density dropped by one-half after ovarian tissue cryopreservation and

CR IP T

xenotransplantation (protocol 1: 1.5 ± 1.3 follicles/mm²; protocol 2: 2.3 ± 1.8 follicles/mm²) compared with fresh controls (protocol 1: 2.6 ± 3.6 follicles/mm²; protocol 2: 4.9 ± 4.9

follicles/mm²). This was not significant in the mixed-effects model, owing to wide inter-

patient variations in treatment effect and limited sample size. Similarly, differences between

AN US

the two protocols were not found to be significant. Proportions of atretic follicles did not differ, whereas the proportion of primordial follicles decreased, and the proportion of growing follicles increased significantly after cryopreservation with both protocols (P < 0.01 in all

M

cases) (Figure 2). The proportion of caspase-3-positive growing follicles was significantly higher after ovarian tissue cryopreservation and xenotransplantation compared with fresh

ED

controls for both protocols (both P < 0.05) (Figure 3), resulting in a significant increase of 21.6-fold (P = 0.02) for protocol 1 and protocol 2 combined, No difference was found

PT

between the two protocols. Interestingly, proportions of caspase-3-positive primordial

CE

follicles did not differ between groups (Figure 3).

AC

When both protocols were pooled together, a large proportion of primordial follicles were activated after ovarian tissue cryopreservation and xenotransplantation, which significantly decreased (–17.4%; P = 0.03) in number, with a borderline significant increase (+15.6%; P = 0.05) in the proportion of growing follicles (Figure 2). Such a change in the follicle population was confirmed by Ki67 immunostaining, which showed a higher proportion of positive primordial (x45.4%; P < 0.01) and growing (+39.1%; P < 0.01) follicles after

15

ACCEPTED MANUSCRIPT cryopreservation and xenotransplantation (Figure 3). No difference was observed between protocol 1 and protocol 2 in follicle classes or Ki67 staining (Figure 3).

Functional status of follicles To evaluate the functional status of follicles after cryopreservation and xenotransplantation,

CR IP T

expression of different factors involved in follicle development was investigated. As shown in Figure 4, immunostaining for AMH was observed in granulosa cells of

primordial and growing follicles, ranging from faint to strong in intensity. The AMH

expression in primordial (protocol 1: 2.1% ± 2.4%; protocol 2: 3.3% ± 3.4%) and growing

AN US

(protocol 1: 94.7% ± 6.5%; protocol 2: 71.6% ± 14.6%) follicles did not significantly change after ovarian tissue cryopreservation and xenotransplantation compared with fresh controls (primordial follicles: protocol 1: 2.6% ± 3.8%; protocol 2: 2.5% ± 3.3%; growing follicles:

M

protocol 1: 51.1% ± 50.0%; protocol 2: 58.1% ± 26.1%). It was also found to be comparable

ED

between protocols.

As kit ligand and c-kit play an essential role in early folliculogenesis (Driancourt et al., 2000),

PT

immunostaining was carried out to evaluate if their expression was altered by a higher DMSO

CE

concentration. No significant difference was noted between protocol 1 and protocol 2 in kit ligand expression in oocytes and granulosa cells from preantral follicles (Figure 5); however,

AC

a significant increase in kit ligand expression in frozen-thawed and xenografted tissue compared with fresh controls was observed for both protocols (both P < 0.05), resulting in an overall increase of 19.0% (P = 0.04) for the combined data of both protocols. C-kit expression was observed in the oolemma of primordial and growing follicles (Figure 5). Unlike kit ligand, c-kit expression did not increase after cryopreservation and xenografting with either protocol; however, c-kit expression was significantly different (P = 0.03) in cryopreserved

16

ACCEPTED MANUSCRIPT and xenografted between the protocols protocol 1 and protocol 2. Although cryopreservation and xenografting were not associated with a significant change in c-kit expression in protocol 1, these treatments resulted in a significant (P = 0.02) decrease in protocol 2 (Figure 5).

Follicle ultrastructure

CR IP T

In general, after cryopreservation, follicle ultrastructure remained similar to fresh follicles. After freeze-thawing, follicles from both protocol 1 and protocol 2 showed preserved

mitochondria and other organelles, intact cell membranes and nuclear envelope, and close contact between granulosa cells and the oocyte (Figure 6). With both protocols, the ooplasm

AN US

had a granulated appearance. After xenografting, follicles from both protocols had wellpreserved mitochondria and intact membranes. The ooplasm still displayed a granulated

aspect and some granulosa cells were less electron-dense, suggesting loss of content (Figure

M

6).

ED

Stromal tissue fibrosis

ovarian fragments from fresh and cryopreserved–xenotransplanted tissue stained with

PT

Masson’s trichrome to show fibrotic areas, marked by the presence of collagen and reduced

CE

density of stromal cells, are shown in Figure 7. Combined data from both protocols showed a significant increase (P = 0.04) in the proportion of fibrotic areas in frozen–thawed and

AC

xenografted tissue (protocol 1: 66.2% ± 14.3%; protocol 2: 52.8% ± 23.2%) compared with fresh controls (protocol 1: 11.6% ± 16.9%; protocol 2: 16.5% ± 8.4%), irrespective of protocol used.

Vascularization

17

ACCEPTED MANUSCRIPT Both human and murine vessels were detected in ovarian tissue after 3 weeks of xenografting, with human vascularization significantly more extensive than murine vascularization (P < 0.01 for both protocols) (Figure 8). The area of vascularization identified after cryopreservation and xenografting did not differ from fresh tissue (Figure 8), nor was any

CR IP T

difference observed between protocols.

Discussion

A number of studies have reported the negative effects of slow-freezing on stromal cells and

AN US

follicles (Nottola et al., 2008; Keros et al., 2009; Amorim et al., 2011; David et al., 2012), emphasizing the need to improve this procedure. On the basis of these concerns, and our computed tomography scan analyses of human ovarian tissue frozen using protocol 1, which

M

revealed significant formation of macroscopic ice crystals, we decided to initiate this study. Here, two controlled-rate freezing protocols designed to cryopreserve human ovarian tissue

ED

were directly compared to determine whether modification of our conventional protocol

PT

would offer better cryoprotection.

CE

As macroscopically visible ice causes mechanical damage as well as an undesirable increase in electrolyte concentrations (Bakhach, 2009), our aim was to reduce the amount of

AC

intratissular ice. A number of strategies can be used to achieve this, with favourable consequences in depression of the equilibrium freezing point of biological samples. For instance, we can decrease cooling rates, add non-permeable solutes or augment concentrations of permeable CPA to enhance dehydration and directly increase CPA concentrations in tissue samples during the incubation period (Routledge and Armitage, 2003; Jin et al., 2014). We decided on the latter option and, because DMSO toxicity diminishes with the decrease in its

18

ACCEPTED MANUSCRIPT temperature of exposure (Best, 2015), we chose to maintain the CPA incubation period. On the other hand, we had to reduce the seeding temperature. The phase diagram for DMSO represents the right half of an inverted parabola, where the freezing temperature (in °C) = 0−1.38 m to 0.177 m² (Kleinhans and Mazur, 2007), ‘m’ being the molality of the solution. This means that at higher (negative) temperatures (close to 0°C), the depression of the melting

CR IP T

point is lower than at lower temperatures. For 10% volume, for example, the freezing temperature is around –4°C. Given the necessary degree of undercooling required when

seeding (to allow ice to grow and to compensate for warming during manipulation), we added another –4°C, so the standard seeding temperature, as used in protocol 1, was –8°C. For 20%,

AN US

the freezing point is around –7°C, so we compensated in a similar way adding –4°C, yielding a seeding temperature of –11°C, which produced a similar effect on the growth of ice. With such modifications, we reached our goal of achieving a reduction in the quantity of ice inside

M

tissue, while doubling the final average amount of DMSO (from 11% to 19.2%). Although a higher DMSO concentration could potentially have a positive or negative effect on follicle

ED

survival, development, or both, either through better penetration of ovarian tissue or higher toxicity levels, our study did not reveal any major difference compared with the lower DMSO

CE

PT

concentration used in our routine cryopreservation solution.

As previously mentioned, computed tomography scan analysis showed that these changes

AC

increased CPA permeation of tissue but did not have affect the follicle population or viability, despite decreasing the amount of ice inside the tissue. From our point of view, this reduction is insufficient. Indeed, we suspect that only complete elimination of ice would have any significant effect, particularly when accounting for the fact that ice will grow upon rewarming. Hence, the reduction in the amount of ice formed during freezing may well be eclipsed by its growth during thawing.

19

ACCEPTED MANUSCRIPT

The only difference between protocol 1 and protocol 2 was the proportion of c-kit-positive follicles after xenografting, which was significantly higher with our routine protocol than with the new procedure. Although one could hypothesize that protocol 2 might have a negative effect on c-kit expression, it is important to bear in mind that the proportion of c-kit-positive

CR IP T

follicles found in protocol 2 after xenografting was statistically similar to results in fresh tissue fragments. Hence, despite this difference between protocols, the higher DMSO concentration does not seem to affect expression of the kit ligand receptor.

Interestingly, in the present study, follicle density did not differ between fresh and

AN US

cryopreserved-xenografted tissue fragments, contradicting the results of our previous study (David et al., 2011) and those of other teams (Oktem et al., 2011; Herraiz et al., 2014), which reported a noticeable decrease in follicle density after xenografting of frozen-thawed ovarian

M

tissue. Such a discrepancy is caused by variations in the follicle population present in ovarian cortex (Qu et al., 2000; Dath et al., 2010) and not to any modification of the cryopreservation

ED

procedure, as the same findings were observed with our routine freezing method (protocol 1). While follicle density has served as a tool to evaluate follicle survival after transplantation,

PT

the best evaluation criterion to calculate total follicle death is counting the number of follicles

CE

present throughout the entire tissue fragment. Indeed, to estimate follicle density, three to five tissue sections are randomly selected, but to calculate the total follicle population, all

AC

haematoxylin-eosin slides are taken into account, which decreases the odds of picking a section with too many or too few follicles.

A small number of atretic follicles were found with both cryopreservation protocols after xenografting. This could have been due to the low rate of injury caused by these cryopreservation procedures or simply because of the removal of dead follicles and their

20

ACCEPTED MANUSCRIPT cellular debris in the first days after transplantation. On the other hand, a higher proportion of caspase-3-positive growing follicles were found with both treatments compared with fresh controls. This confirms that primordial follicles are more cryoresistant than their growing counterparts (Eyden et al., 2004), and our increased DMSO concentration was not able to

CR IP T

protect primary and secondary follicles from damage.

Follicle activation was also similar between our routine protocol and the new

cryopreservation method. The higher proportion of growing follicles detected after

xenografting was probably caused by hypoxic stress and a lack of AMH in ovarian tissue at

AN US

the time of transplantation (David et al., 2012). Secondary follicles produce AMH and are known to exert an inhibitory effect on primordial follicles, keeping them at a quiescent stage (Kim et al., 2012). The high activation observed in our histological analysis was confirmed by

M

the greater proportion of Ki67-positive primordial and growing follicles encountered after 3

ED

weeks of xenografting.

An interesting finding in our study concerns the proportion of kit-ligand -positive follicles

PT

detected after freezing, thawing and xenografting, which was similar to our previous findings

CE

(David et al., 2012) in fresh ovarian tissue, but significantly higher after transplantation. David et al. (2012) reported that, after cryopreservation with our routine procedure, followed

AC

by 28 weeks of xenografting, kit-ligand expression was statistically lower than in fresh controls. As the grafting period was much longer in this previous study, we can hypothesize that after an initial increase in kit-ligand expression during the first weeks after transplantation, it may subsequently decline with time.

21

ACCEPTED MANUSCRIPT To assess possible damage to the ultrastructure of follicles caused by the new cryopreservation protocol, ovarian tissue samples were fixed for transmission electron microscopy after cryopreservation and transplantation and also soon after thawing. No difference in follicle ultrastructure, however, was found between protocol 1 and protocol 2. The only ultrastructural change observed in follicles issuing from tissue previously

CR IP T

cryopreserved with both protocols was the granulated aspect of the ooplasm, which is a characteristic often observed in cryopreserved cells (Lucci et al., 2004; Borges et al., 2009). Indeed, after xenografting, follicles from both protocols had a comparable ultrastructure,

AN US

suggesting a similar ability to develop.

In assisted reproductive techniques, DMSO is the most commonly used CPA and is probably the least toxic and most protective (Vajta et al., 2013). Despite a widespread belief that high

M

CPA concentrations can be toxic to reproductive cells, it is important to point out that some of

CPA itself (Best, 2015).

ED

the cell damage is caused by osmotic shock, oxidative stress or chilling injury rather than the

PT

In conclusion, our study revealed that a new freezing protocol using a higher volume of

CE

cryopreservation solution, an increased DMSO concentration and a lower seeding temperature does not enhance survival of preantral follicles after xenotransplantation. As our

AC

modifications did not affect the results, we need to test a different approach if we hope to improve our freezing protocol. It is also important to stress that, although the higher DMSO concentration did not increase survival of preantral follicles, it did not induce any major damage to the population of human ovarian follicles at early stages of development. These findings are important for vitrification of human ovarian tissue, as they indicate that 20% DMSO, a concentration widely used in vitrification solutions (Kagawa et al., 2009; Amorim

22

ACCEPTED MANUSCRIPT et al., 2012; Herraiz et al., 2014), does not seem to be toxic to the follicle population. On the other hand, it would be of great interest to assess whether increased DMSO concentrations could affect the further development of these follicles.

Acknowledgements

CR IP T

The authors thank Mira Hryniuk, BA, for reviewing the English language of the manuscript, and Dolores Gonzales, Olivier Van Kerk, Angel Parrado-Gallego and Isabel FernandezGomez for their technical assistance. Declaration

AN US

This study was supported by grants from the Fonds National de la Recherche Scientifique de Belgique (FNRS) (CA Amorim is a Research Associate, FRS – FNRS; grant 5/4/150/5 awarded to M.M. Dolmans), Fonds Spéciaux de Recherche, Fondation St Luc, Foundation

M

Against Cancer, Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES Brazil) (grant number 013/14 CAPES/WBI awarded to CM Lucci; F Paulini received a post-

ED

doc fellowship), Wallonie-Bruxelles International, Association for the Study of Reproductive Biology (mobility grant awarded to M.Gallardo), and Ministry of Economy and

PT

Competitiveness of Spain (Retos grant RTC-2016-4733-1 awarded to M Gallardo and R

CE

Risco). The authors report no financial or commercial conflicts of interest.

AC

References

Amorim, C.A., Curaba, M., Van Langendonckt, A., Dolmans, M.M., Donnez, J., 2011. Vitrification as an alternative means of cryopreserving ovarian tissue. Reprod. Biomed. Online 23, 160-186.

23

ACCEPTED MANUSCRIPT Amorim, C.A., David, A., Dolmans, M.M., Camboni, A., Donnez, J., Van Langendonckt, A., 2011. Impact of freezing and thawing of human ovarian tissue on follicular growth after long-term xenotransplantation. J. Assist. Reprod. Genet. 28, 1157-1165. Amorim, C.A., Dolmans, M.M., David, A., Jaeger, J., Vanacker, J., Camboni, A., Donnez, J., Van Langendonckt, A., 2012. Vitrification and xenografting of human ovarian tissue.

CR IP T

Fertil. Steril. 98, 1291-1298 e1-2. Amorim, C.A., Jacobs, S., Devireddy, R.V., Van Langendonckt, A., Vanacker, J., Jaeger, J., Luyckx, V., Donnez, J., Dolmans, M.M., 2013. Successful vitrification and autografting of baboon (Papio anubis) ovarian tissue. Hum. Reprod. 28, 2146-2156.

AN US

Amorim, C.A., Van Langendonckt, A., David, A., Dolmans, M.M., Donnez, J., 2009.

Survival of human pre-antral follicles after cryopreservation of ovarian tissue, follicular isolation and in vitro culture in a calcium alginate matrix. Hum. Reprod. 24, 92-99.

M

Bakhach, J., 2009. The cryopreservation of composite tissues: Principles and recent advancement on cryopreservation of different type of tissues. Organogenesis. 5, 119-126.

ED

Best, B.P., 2015. Cryoprotectant Toxicity: Facts, Issues, and Questions. Rejuvenation Res. 18, 422-436.

PT

Borges, E.N., Silva, R.C., Futino, D.O., Rocha-Junior, C.M., Amorim, C.A., Báo, S.N., Lucci,

CE

C.M., 2009. Cryopreservation of swine ovarian tissue: effect of different cryoprotectants on the structural preservation of preantral follicle oocytes. Cryobiology 59, 195-200.

AC

Corral, A., Balcerzyk, M., Amorim, C.A., Paulini, F., Parrado-Gallego, A., Gallardo, M., Dolmans, M.M., Risco, R., 2018. Ovarian tissue cryopreservation by stepped vitrification and monitored by X-ray computed tomography. Cryobiology 81, 17-26.

Corral, A., Balcerzyk, M., Parrado-Gallego, A., Fernandez-Gomez, I., Lamprea, D.R., Olmo, A., Risco, R., 2015. Assessment of the cryoprotectant concentration inside a bulky organ for cryopreservation using X-ray computed tomography. Cryobiology 71, 419-431.

24

ACCEPTED MANUSCRIPT Dath, C., Van Eyck, A.S., Dolmans, M.M., Romeu, L., Delle Vigne, L., Donnez, J., Van Langendonckt, A., 2010. Xenotransplantation of human ovarian tissue to nude mice: comparison between four grafting sites. Hum. Reprod. 25, 1734-1743. David, A., Dolmans, M.M., Van Langendonckt, A., Donnez, J., Amorim, C.A., 2011. Immunohistochemical localization of growth factors after cryopreservation and 3 weeks'

CR IP T

xenotransplantation of human ovarian tissue. Fertil. Steril. 95, 1241-1246. David, A., Van Langendonckt, A., Gilliaux, S., Dolmans, M.M., Donnez, J., Amorim, C.A., 2012. Effect of cryopreservation and transplantation on the expression of kit ligand and anti-Mullerian hormone in human ovarian tissue. Hum. Reprod. 27, 1088-1095.

AN US

Dolmans, M.M., Jadoul, P., Gilliaux, S., Amorim, C.A., Luyckx, V., Squifflet, J., Donnez, J., Van Langendonckt, A., 2013. A review of 15 years of ovarian tissue bank activities. J. Assist. Reprod. Genet. 30, 305-314.

M

Donnez, J., Dolmans, M.M., 2013. Fertility preservation in women. Nat. Rev. Endocrinol. 9:735-749.

ED

Donnez, J., Dolmans, M.M., Demylle, D., Jadoul, P., Pirard, C., Squifflet, J., MartinezMadrid, B., van Langendonckt, A., 2004. Livebirth after orthotopic transplantation of

PT

cryopreserved ovarian tissue. Lancet. 364:1405-1410.

CE

Donnez, J., Squifflet, J., Van Eyck, A.S., Demylle, D., Jadoul, P., Van Langendonckt, A., Dolmans, M.M., 2008. Restoration of ovarian function in orthotopically transplanted

AC

cryopreserved ovarian tissue: a pilot experience. Reprod. Biomed. Online. 16:694-704.

Driancourt, M.A., Reynaud, K., Cortvrindt, R., Smitz, J., 2000. Roles of KIT and KIT LIGAND in ovarian function. Rev. Reprod. 5:143-152.

Eyden, B., Radford, J., Shalet, S.M., Thomas, N., Brison, D.R., Lieberman, B.A., 2004. Ultrastructural preservation of ovarian cortical tissue cryopreserved in dimethylsulfoxide

25

ACCEPTED MANUSCRIPT for subsequent transplantation into young female cancer patients. Ultrastruct. Pathol. 28, 239-245. Gosden, R.G., Baird, D.T., Wade, J.C., Webb, R., 1994. Restoration of fertility to oophorectomized sheep by ovarian autografts stored at -196 degrees C. Hum. Reprod. 9, 597-603.

CR IP T

Hovatta O., 2005. Methods for cryopreservation of human ovarian tissue. Reprod. Biomed. Online 10, 729-734.

Israely, T., Nevo, N., Harmelin, A., Neeman, M., Tsafriri, A., 2006. Reducing ischaemic damage in rodent ovarian xenografts transplanted into granulation tissue. Hum. Reprod.

AN US

21, 1368-1379.

Herraiz, S., Novella-Maestre, E., Rodrígues, B., Díaz, C., Sánchez-Serrano, M., Mirabet, M., Pellicer, A., 2014. Improving ovarian tissue cryopreservation for oncologic patients: slow

M

freezing versuns vitrification, effect of different procedures and devices. Fertil. Steril. 101, 775-784.e1.

ED

Jadoul, P., Guilmain, A., Squifflet, J., Luyckx, M., Votino, R., Wyns, C., Dolmans, MM., 2017. Efficacy of ovarian tissue cryopreservation for fertility preservation: lessons learned

PT

from 545 cases. Hum. Reprod. 32, 1046-1054.

CE

Jensen, A.K., Macklon, K.T., Fedder, J., Ernst, E., Humaidan, P., Andersen, C.Y., 2017. 86 successful births and 9 ongoing pregnancies worldwide in women transplanted with

AC

frozen-thawed ovarian tissue: focus on birth and perinatal outcome in 40 of these children. J. Assist. Reprod. Genet. 34, 325-336.

Jin, B., Kleinhans, F.W., Mazur, P., 2014. Survivals of mouse oocytes approach 100% after vitrification in 3-fold diluted media and ultra-rapid warming by an IR laser pulse. Cryobiology. 68, 419-430.

26

ACCEPTED MANUSCRIPT Kagawa, N., Silber, S., Kuwayama, M., 2009. Successful vitrification of bovine and human ovarian tissue. Reprod. Biomed. Online 18, 568-577. Keros, V., Xella, S., Hultenby, K., Pettersson, K., Sheikhi, M., Volpe, A., Hreinsson, J., Hovatta, O., 2009. Vitrification versus controlled-rate freezing in cryopreservation of human ovarian tissue. Hum. Reprod. 24, 1670-1683.

CR IP T

Kim, J.Y., 2012. Control of ovarian primordial follicle activation. Clin. Exp. Reprod. Med. 39, 10-14.

Kleinhans, F.W., Mazur, P., 2007. Comparison of actual vs. synthesized ternary phase diagrams for solutes of cryobiological interest. Cryobiology. 54, 212-222.

AN US

Lee, J., Kong, H.S., Kim, E.J., Youm, H.W., Lee, J.R., Suh, C.S., Kim, S.H., 2016. Ovarian injury during cryopreservation and transplantation in mice: a comparative study between cryoinjury and ischemic injury. Hum. Reprod. 31, 1827-1837.

M

Lucci, C.M., Kacinskis, M.A., Lopes, L.H., Rumpf, R, Báo, S.N., 2004. Effect of different cryoprotectants on the structural preservation of follicles in frozen zebu bovine (Bos

ED

indicus) ovarian tissue. Theriogenology 61, 1101-1114. Martinez-Madrid, B., Donnez, J., Van Eyck, A.S., Veiga-Lopez, A., Dolmans, M.M., Van

PT

Langendonckt, A., 2009. Chick embryo chorioallantoic membrane (CAM) model: a useful

CE

tool to study short-term transplantation of cryopreserved human ovarian tissue. Fertil. Steril. 91:285-92.

AC

Meirow, D., Levron, J., Eldar-Geva, T., Hardan, I., Fridman, E., Yemini, Z., Dor, J., 2007. Monitoring the ovaries after autotransplantation of cryopreserved ovarian tissue: endocrine studies, in vitro fertilization cycles, and live birth. Fertil. Steril. 87, 418.e7-418.e15.

Nisolle, M., Casanas-Roux, F., Qu, J., Motta, P., Donnez, J., 2000. Histologic and ultrastructural evaluation of fresh and frozen-thawed human ovarian xenografts in nude mice. Fertil. Steril. 74, 122-129.

27

ACCEPTED MANUSCRIPT Nottola, S.A., Camboni, A., Van Langendonckt, A., Demylle, D., Macchiarelli, G., Dolmans, M.M., Martinez-Madrid, B., Correr, S., Donnez, J., 2008. Cryopreservation and xenotransplantation of human ovarian tissue: an ultrastructural study. Fertil. Steril. 90, 2332. Oktem, O., Alper, E., Balaban, B., Palaoglu, E., Peker, K., Karakaya, C., Urman, B., 2011.

CR IP T

Vitrified human ovaries have fewer primordial follicles and produce less antimüllerian hormone than slow-frozen ovaries. Fertil. Steril. 95, 2661-2664.e1.

Pfeifer, S., Goldberg, J., Lobo, R., Pisarka, M., Thomas, M., Widra, E., Sandlow, J., Licht, M., Rosen, M., Vernon, M., Catherino W, Davis O, Dumesic D, Gracia C, Odem R,

AN US

Thornton K, Reindollar R, Rebar R, La Barbera A., 2014. Ovarian tissue cryopreservation: a committee opinion. Fertil. Steril. 101, 1237-1243.

Qu, J., Godin, P.A., Nisolle, M., Donnez, J., 2000. Distribution and epidermal growth factor

M

receptor expression of primordial follicles in human ovarian tissue before and after cryopreservation. Hum. Reprod. 15, 302-310.

ED

Rosendahl, M., Schmidt, K.T., Ernst, E., Rasmussen, P.E., Loft, A., Byskov, A.G., Andersen, A.N., Andersen, CY., 2011. Cryopreservation of ovarian tissue for a decade in Denmark: a

PT

view of the technique. Reprod. Biomed. Online 22, 162-171.

CE

Routledge, C., Armitage, W.J, 2003. Cryopreservation of cornea: a low cooling rate improves functional survival of endothelium after freezing and thawing. Invest. Ophthalmol. Vis.

AC

Sci. 44, 3326-3331.

Soares, M., Sahrari, K., Chiti, M.C., Amorim, C.A., Ambroise, J., Donnez, J., Dolmans, M.M., 2015. The best source of isolated stromal cells for the artificial ovary: medulla or cortex, cryopreserved or fresh? Hum. Reprod. 30, 1589-1598.

28

ACCEPTED MANUSCRIPT Suzuki, N., Yoshioka, N., Takae, S., Sugishita, Y., Tamura, M., Hashimoto, S., Morimoto, Y., Kawamura, K., 2015. Successful fertility preservation following ovarian tissue vitrification in patients with primary ovarian insufficiency. Hum. Reprod. 30, 608-615. Vajta, G., Reichart, A., Ubaldi, F.L.R., 2013. From a backup technology to a strategyoutlining approach: the success story of cryopreservation. Expert Rev. Obstet. Gynecol. 8,

CR IP T

181-190. Van der Ven, H., Liebenthron, J., Beckmann, M., Toth, B., Korell, M., Krussel, J., Frambach, T., Kupka, M., Hohl, M.K., Winkler-Crepaz, K., Seitz S, Dogan A, Griesinger G, Häberlin F, Henes M, Schwab R, Sütterlin M, von Wolff M, Dittrich R, FertiPROTEKT network,

AN US

2016. Ninety-five orthotopic transplantations in 74 women of ovarian tissue after cytotoxic treatment in a fertility preservation network: tissue activity, pregnancy and delivery rates. Hum. Reprod. 31, 2031-2041.

M

Van Eyck, A.S., Bouzin, C., Feron, O., Romeu, L., Van Langendonckt, A., Donnez, J., Dolmans, M.M., 2010. Both host and graft vessels contribute to revascularization of

ED

xenografted human ovarian tissue in a murine model. Fertil. Steril. 93, 1676-1685. Wallace, W.H., Kelsey, T.W., Anderson, R.A., 2016. Fertility preservation in pre-pubertal

AC

CE

PT

girls with cancer: the role of ovarian tissue cryopreservation. Fertil. Steril. 105, 6-12.

29

ACCEPTED MANUSCRIPT Author biography Professor Christiani Amorim is a veterinarian working in the Gynaecology Research Unit of Université Catholique de Louvain. She has been conducting research on ovarian tissue engineering, biomaterials and preantral follicle cryopreservation, in-vitro culture and

AC

CE

PT

ED

M

AN US

CR IP T

xenotransplantation.

30

ACCEPTED MANUSCRIPT Figure 1. Computed tomography images of ovarian tissue samples inside cryovials at – 140°C. The spatial resolution is 0.1 mm and the colour scale ranges from dark blue for low attenuation (1.2 CT) to intense red for high attenuation (3.0 CT). The pink cube is a volume of interest measuring 3 x 3 x 1 mm3 located within the tissue. (A) Tissue sample cryopreserved with protocol 1 (P1). The ice structure is shown in dark blue, surrounded by small islands of

CR IP T

high dimethyl sulphoxide (DMSO) concentration. The tissue cannot be distinguished from the solution around it, although the presence of ice is less pronounced in the sample area; (B) tissue sample cryopreserved with protocol 2 (P2). The ice structure is shown in dark blue and the tissue area is green in colour corresponding to a higher DMSO concentration than seen in

AN US

(A). Thawed tissue sample at room temperature previously cryopreserved with (C) P1 and (D) P2. No ice structure is visible and a homogeneous dark blue colour can be seen along the whole vial, corresponding to the colour of water attenuation, and hence to a low DMSO

M

concentration. Pictures shown on the bottom right of each figure are maximum intensity projections, two-dimensional parallel projections of three-dimensional spaces that show

AC

CE

PT

ED

maximum image attenuation. RT, room temperature.

31

ACCEPTED MANUSCRIPT Figure 2. Preantral follicle analysis after haematoxylin eosin staining. (A) Proportion of primordial, growing (primary and secondary) and atretic follicles before and after cryopreservation and xenografting of human ovarian tissue. Xenografted primordial and primary follicles from (B) protocol 1 and (C) protocol 2, and a primordial follicle surrounded by two atretic follicles (red arrows) in an ovarian tissue fragment after freezing and thawing

AC

CE

PT

ED

M

AN US

CR IP T

with protocol 2 followed by (D) xenografting.

32

ACCEPTED MANUSCRIPT Figure 3. Immunohistochemical staining of preantral follicles after cryopreservation and xenografting. (A) Proportion (mean ± SD) of caspase-3- and Ki67-positive follicles before and after cryopreservation and xenografting. Ki67 immunostaining: Ki67-positive granulosa cells in a secondary follicle from the (B) protocol 1 group and (C) positive and (D) negative controls (human proliferative endometrium). Caspase immunostaining: caspase-3-positive

CR IP T

oocyte (nucleus in dark brown) from the (E) protocol 2 group and (F) positive and (G)

AC

CE

PT

ED

M

AN US

negative controls (human tonsil).

33

ACCEPTED MANUSCRIPT Figure 4. Anti-Müllerian hormone (AMH) immunohistochemical staining of preantral follicles. (A) AMH-negative granulosa cells in a primordial follicle (from fresh ovarian tissue); (B) AMH-positive granulosa cells in a primary follicle (from fresh ovarian tissue); and (C) secondary follicle (from ovarian tissue after freezing, thawing and xenografting (protocol 1 group); (D) positive (human ovary with antral follicles); and (E) negative controls

AC

CE

PT

ED

M

AN US

CR IP T

(human ovarian tissue).

34

ACCEPTED MANUSCRIPT Figure 5. Kit ligand (KL) and c-kit immunohistochemical staining of preantral follicles after cryopreservation and xenografting. (A) Proportion (mean ± SD) of KL- and c-kit-positive follicles before and after cryopreservation and xenografting. Kit ligand immunostaining: Dispersed KL staining in granulosa cells of preantral follicles from the (B) protocol 2 group and (C) positive and (D) negative controls (human colon). C-kit immunostaining: a c-kit-

CR IP T

positive oolemma in a (E) primary follicle, and (F) positive and (G) negative controls (human

AC

CE

PT

ED

M

AN US

brain).

35

ACCEPTED MANUSCRIPT Figure 6. Ultrastructure of preantral follicles. (A) Representative electron micrographs of human follicles from fresh ovarian tissue; (B) tissue frozen-thawed with protocol 1; (C–D) tissue frozen–thawed with protocol 2; and (E) frozen-thawed tissue with protocol 1 and xenografted; and (F) tissue frozen-thawed with protocol 2 and xenografted. GC, granulosa cell; Nu, oocyte nucleus; m, mitochondria; O, oocyte; *Granulosa cell with lower electron

AC

CE

PT

ED

M

AN US

CR IP T

density.

36

ACCEPTED MANUSCRIPT Figure 7. Evaluation of ovarian tissue after cryopreservation and xenografting. Fibrotic areas shown by Masson’s trichrome staining. Fragments from (A) fresh and (B) frozen–thawed and xenografted tissue (protocol 2 group) stained green (centre of the grafted tissue) owing to (B)

AC

CE

PT

ED

M

AN US

CR IP T

collagen fibres present in higher numbers among stromal cells.

37

ACCEPTED MANUSCRIPT Figure 8. Analysis of vascularization before and after cryopreservation and xenografting of human ovarian tissue. To calculate the vessel area, the area around each ovarian tissue graft was defined and vessel area (stained in pink or brown) was measured. (A) Human vascularization was significantly more extensive than murine counterpart in protocol 1 (P1) and protocol 2 (P2) (both P < 0.01); (B) CD34 immunostaining: CD34 human vessels in fresh

CR IP T

ovarian tissue; and (C) double CD34 immunostaining (human vessels in pink and murine vessels in brown) in human ovarian tissue from the P1 group; (D) positive and (E) negative controls for human CD34 immunostaining (human liver) and (F) positive and (G) negative

AC

CE

PT

ED

M

AN US

controls for mouse CD34 immunostaining (mouse ovary).

38

ACCEPTED MANUSCRIPT Table 1. Immunohistochemical protocols applied to assess follicle growth and function, and graft vascularization. Antibody

Antigen

Dilution

Positive

Negative

control

control

4°C

Human

Human

overnight

proliferative

proliferative

Incubation

Ki67 (M7240;

Monoclonal 1:100 in TBS

Dako, Glostrup,

+ 1% NGS +

Denmark)

0.1% BSA

Promega,

Polyclonal

1:200 in TBS

1 h room

(rabbit)

+ 4% NGS +

temperature

Madison, USA)

Human tonsil

Human tonsil

0.4% BSA

AMH (MCA2246;

Monoclonal 1:20 in TBS + 1% NGS +

Brussels,

0.1% BSA

4°C

Human ovary

Human ovary

overnight

with antral

with antral

follicles

follicles

Human colon

Human colon

Human brain

Human brain

Human liver

Human liver

M

Serotec, Gentaur,

ED

Belgium) KL (SC-13126;

Monoclonal 1:50 in TBS + 2% NGS +

PT

Santa Cruz Biotechnology,

endometrium endometrium

AN US

Caspase-3 (G748A;

CR IP T

type

4°C overnight

0.2% BSA

CE

Santa Cruz, USA)

AC

c-kit (A4502; Dako)

CD34 (CM084B;

Polyclonal

1:200 in TBS

4°C

(rabbit)

+ 2% NGS +

overnight

0.2% BSA Monoclonal 1:500 in TBS

Biocare Medical,

+ 1% NGS +

Pacheco, USA)

0.1% BSA

4°C overnight

39

ACCEPTED MANUSCRIPT CD34 (HM 1015;

Monoclonal 1:100 in TBS

1h room

Mouse ovary

Hycult Biotech,

+ 1% NGS +

temperatu

Uden, the

0.1% BSA

re

Mouse ovary

Netherlands)

AC

CE

PT

ED

M

AN US

CR IP T

AMH, anti-Müllerian hormone; BSA, bovine serum albumin; KL, kit ligand; NGS, normal goat serum; TBS, tris-buffered saline.

40