Apoptosis and ultrastructural assessment after cryopreservation of whole human ovaries with their vascular pedicle

Transcription

1 Apoptosis and ultrastructural assessment after cryopreservation of whole human ovaries with their vascular pedicle Belen Martinez-Madrid, V.M.D., Ph.D., a Alessandra Camboni, M.D., a,b Marie-Madeleine Dolmans, M.D., Ph.D., a Stefania Nottola, M.D., Ph.D., b Anne Van Langendonckt, Ph.D., a and Jacques Donnez, M.D., Ph.D. a a Department of Gynecology, Cliniques Universitaires Saint Luc, Université Catholique de Louvain, Brussels, Belgium; and b Department of Anatomy, University of Rome La Sapienza, Rome, Italy Objective: To investigate possible damage caused by freeze-thawing whole human ovaries. Design: Prospective experimental study. Setting: Academic gynecology research unit in a university hospital. Patient(s): Ovaries were obtained from three women (aged years). Intervention(s): Ovaries were perfused and bathed in cryoprotective solution, and slow freezing was performed. Rapid thawing was achieved by perfusion and bathing with a decreased sucrose gradient. Main Outcome Measure(s): Apoptosis was assessed by the terminal deoxynucleotidyl transferase-mediated biotinylated deoxyuridine triphosphates nick end-labeling (TUNEL) method and by immunohistochemistry for active caspase-3 in fresh ovaries, after cryoprotectant exposure, and after thawing. Morphometric analysis of TUNEL-positive surface area was performed. Ultrastructure was assessed by transmission electron microscopy (TEM) in the thawed tissue. Result(s): No primordial or primary follicles were found to be positive for either TUNEL or active caspase-3. No statistically significant difference in mean TUNEL-positive surface area values was found between the three groups: fresh, 0.05% 0.03%, with 134 high-power fields (HPFs); cryoperfused, 0.02% 0.01%, with 130 HPFs; and thawed, 0.09% 0.03%, with 622 HPFs. By means of TEM, follicles and vessels showed a well-preserved ultrastructure, with 96.7% (29/30) healthy-looking primordial and primary follicles, and 96.3% (180/187) healthy-looking endothelial cells. Conclusion(s): Cryopreservation of intact human ovary with its vascular pedicle, according to the freeze-thawing protocol described here, is not associated with any signs of apoptosis or ultrastructural alterations in any cell types. Whole-organ vascular transplantation may thus be a viable option in the future. (Fertil Steril 2007;87: by American Society for Reproductive Medicine.) Key Words: Cryopreservation, human, whole ovary, apoptosis, TUNEL, caspase-3, TEM Supported by the Fonds National de Recherche Scientifique (grant F); Televie (grant ); the Fondation St. Luc, the Belgian Federation Against Cancer; Baron Albert Frère; the Comte de Spoelberch; the Italian Ministry of Education, Universities, and Research (grant C26A049412); and the Italian Ministry of Foreign Affairs (VII Executive Program of Scientific Collaboration between Italy and the French-Speaking Community of Belgium ). Belen Martinez-Madrid, V.M.D., Ph.D., and Alessandra Camboni, M.D., contributed equally to this article. Received February 28, 2006; revised August 23, 2006; accepted November 1, Reprint requests: Jacques Donnez, M.D., Ph.D., Department of Gynecology, Cliniques Universitaires Saint Luc, Université Catholique de Louvain, Avenue Hippocrate 10, 1200 Brussels, Belgium (FAX: ; Donnez@gyne.ucl.ac.be). Survival rates of cancer patients are on the increase because of constant improvements in the diagnosis and treatment of the disease. However, patients requiring chemotherapy and/or radiotherapy for cancer or other benign pathologies are likely to experience premature ovarian failure and loss of fertility as a consequence of these potentially gonadotoxic treatments. Several options are currently available to preserve fertility in these patients, giving them the opportunity to become mothers when they have overcome their disease: cryopreservation of embryos, oocytes, or ovarian tissue (1, 2). The choice of the most suitable strategy for preserving fertility depends on different parameters: the type and timing of chemotherapy, the type of cancer, the patient s age, and her partner status. The only established method of fertility preservation is embryo cryopreservation (3), but this option requires the patient to be of pubertal age and to have a partner or use donor sperm, and to be in a position to undergo a cycle of ovarian stimulation, which is not possible when chemotherapy must be started immediately or when stimulation is contraindicated according to the type of cancer. Cryopreservation of oocytes can be performed in single women who are able to undergo a stimulation cycle, although the effectiveness of this technique is very low, with pregnancy and delivery rates ranging from 1% 5% per frozen oocyte (4). Cryopreservation of ovarian tissue is the only option for /07/$32.00 Fertility and Sterility Vol. 87, No. 5, May 2007 doi: /j.fertnstert Copyright 2007 American Society for Reproductive Medicine, Published by Elsevier Inc. 1153

2 prepubertal girls and women who cannot delay the start of chemotherapy. Ovarian tissue can be frozen and grafted in three different ways: as fragments of ovarian cortex, as an entire ovary with its vascular pedicle, or as isolated follicles (5). So far, ovarian cryopreservation and transplantation procedures have been almost exclusively limited to avascular cortical fragments in both experimental and clinical studies (6 8) and, for now, this is the only procedure that has yielded live births in humans after autologous transplantation (9, 10). The main drawback of this method is that the graft is completely dependent on the establishment of neovascularization and, as a result, a significant proportion of follicles are lost due to ischemic damage by the time neovascularization is achieved (6, 11 15). Reducing the ischemic interval between transplantation and revascularization is therefore essential to maintaining the follicular reserve and extending the life span and function of the graft. Transplantation of intact ovaries with vascular anastomosis would allow immediate revascularization of the transplant, and would avoid problems related to ischemic injury. Ovarian vascular transplantation was successfully performed with the use of intact fresh ovaries in rats (16, 17), rabbits (18), sheep (19, 20), dogs (21), monkeys (22), and humans (23 25). In the last few years, attempts at freezing and grafting whole ovaries in rats (16, 17), rabbits (26), sheep (27 30), and pigs (31) also yielded encouraging results. It appears that, in large mammals and humans, anastomosis of the ovarian pedicle is technically feasible. However, in these species, cryopreserving such a large-sized intact ovary is problematic due to the difficulty of adequate diffusion of cryoprotective agents into large tissue masses and the risk of vascular injury caused by intravascular ice formation. The main challenge of ovarian vascular transplantation therefore lies in successfully cryopreserving whole human ovaries. This procedure needs to avoid cryoinjury due to the freezethawing procedure itself, as well as ischemic and toxic injury during cryoprotectant perfusion, while allowing adequate diffusion of cryoprotective agents into the deepest parts of the ovary. We previously described a cryopreservation protocol for the intact human ovary with its vascular pedicle, and demonstrated high survival rates of follicles (75.1%), small vessels, and stroma, and a normal histological structure in all ovarian components after thawing (32). Despite these positive results, it is essential to conduct a thorough examination of the effects of this protocol on ovarian tissue before attempting vascular autotransplantation of whole ovaries in humans. The aim of the present study is to further investigate possible injury caused by this freeze-thawing protocol, by analyzing apoptosis and ultrastructural damage to all ovarian cell types. MATERIALS AND METHODS Cryopreservation of the Intact Ovary The use of human tissue for this study was approved by the Institutional Review Board of the Université Catholique de Louvain, Brussels, Belgium. Ovaries were obtained from three women (29, 31, and 36 years of age) after oophorectomy during a Wertheim-Meigs procedure prior to pelvic radiotherapy. All three ovaries showed normal macroscopic features at the time of oophorectomy, i.e., size, vasculature, and surface appearance, with the presence of some antral follicles and corpora lutea on the cortex. The ovaries were cryopreserved as previously described (32). Briefly, the ovary was perfused via the ovarian artery with heparinized physiologic solution and transported to the laboratory, where it was perfused and immersed in a bath containing a cryoprotective solution of Leibovitz-15 medium (GIBCO, Paisley, Scotland), supplemented with 10% dimethyl sulfoxide (DMSO) (Sigma, St. Louis, MO) and 0.4% human serum albumin (Red Cross, Brussels, Belgium) for 5 minutes at 4 C. The ovary was placed in a cryovial where it was preequilibrated at 4 C in a bath with the cryoprotective solution for 10 minutes. The cryovial containing the ovary was placed in a 5100 Cryo 1 C Freezing Container (Nalgene, VWR, Leuven, Belgium) precooled at 4 C, and deposited in a 80 C freezer. This confers a theoretical cooling rate of 1 C/minute. After 24 hours at 80 C, the cryovial containing the ovary was transferred to liquid nitrogen (LN 2 ). Thawing of the Intact Ovary For thawing, the cryovial was directly transferred from the LN 2 to a water bath at 60 C, where it was immersed until the ice melted. To remove the cryoprotectant, the ovary was bathed and perfused in three steps, for 10 minutes each at room temperature (RT), with a reversed sucrose concentration gradient (0.25 M, 0.1 M, and 0 M sucrose) in L-15 medium, to avoid osmotic injuries. Perfusion was performed at the same flow rate as already described. Assessment of Apoptosis At three different times (with freshly removed ovary, after preequilibration with cryoprotective solution before freezing, and after thawing), samples were taken for assessment of apoptosis. Cortical and medullar tissues were fixed in 4% formalin in phosphate-buffered saline (PBS), and embedded in paraffin. Five-micron-thick sections were cut from the blocks and air-dried on slides. Apoptosis was analyzed by a terminal deoxynucleotidyl transferase-mediated biotinylated deoxyuridine triphosphates nick end-labeling (TUNEL) technology method to detect DNA fragmentation, and by immunohistochemistry (IHC) for active caspase-3 to detect cells programmed to undergo apoptosis Martinez-Madrid et al. Freezing whole human ovaries Vol. 87, No. 5, May 2007

3 Analysis of DNA Strand Breaks by TUNEL Sections were dewaxed with histosafe (NV Yvsolab, Antwerp, Belgium), rehydrated with isopropanol, and washed in running deionized water. The slides were then pretreated with 20 g/ml of proteinase K working solution (catalogue no ; Roche Applied Science, Penzberg, Germany) in 10 mm Tris-HCl, ph 7.5, for 30 minutes at 37 C in a humidified chamber. Strand breaks of DNA occurring during the apoptotic process were detected by means of the In Situ Cell Death Detection Kit, TMR Red (catalogue no ; Roche Applied Science), a TUNEL assay. After washing with PBS, slides were incubated with a TUNEL reaction mixture: 50 L enzyme solution (terminal deoxynucleotidyl transferase) and 450 L label solution (nucleotide mixture in reaction buffer) for 60 minutes at 37 C in a humidified chamber protected from light, followed by rinsing with PBS. Positive control sections were treated with 1,500 U/mL DNase I (catalogue no ; Roche Applied Science) in 50 mm Tris-HCl, ph 7.5, 1 mg/ml bovine serum albumin (BSA), for 10 minutes at RT in a humidified chamber, before incubation with the TUNEL reaction mixture. Negative control sections were incubated with label solution without enzyme solution. Finally, slides were covered with Vectashield Mounting Medium with 4=,6-diamino-2-phenylindole (DAPI) (Vector Laboratories, Burlingame, CA). This special formulation is intended to preserve fluorescence during prolonged storage and, at the same time, to counterstain DNA by means of DAPI. Slides were coverslipped and sealed around the perimeter with nail polish, stored at 4 C, and protected from light until examination. Seven replicates were performed, each including one slide from the three different groups (freshly removed ovary, after cryoprotectant exposure, and after thawing) from all three patients, as well as positive and negative controls. TUNEL-stained and DAPI-counterstained slides were examined under an inverted fluorescence microscope (Leica; Van Hopplynus Instruments, Brussels, Belgium). Red fluorescence could be visualized in TUNEL-positive cells with the use of an excitation wavelength in the range of nm, and by observing the emitted light at a wavelength between nm. DAPI reached excitation at about 360 nm, and emitted at about 460 nm when bound to DNA, producing a blue fluorescence in all nuclei. Morphometric analysis of TUNEL-positive surface area was performed to quantify apoptosis. For this purpose, sections were examined at 200 magnification, and all highpower fields (HPFs) were digitalized, either for TUNEL staining or DAPI counterstaining, using a Leica DFC320 camera and IM50 program (Leica). ImageJ, a freely available image-processing and analysis program developed at the National Institutes of Health ( was used to delimit all TUNEL-positive cells and to measure their surface area, as well as to determine total surface area in each section (by measuring DAPI-counterstained surface area). Antral follicles and corpora lutea were excluded from the morphometric analysis of TUNEL-positive surface areas because the physiological occurrence of apoptosis in these structures, as well as their large size, could distort the results. Immunohistochemical Detection of Active Caspase-3 The active caspase-3 technique is an immunohistochemical assay for the detection of the enzyme caspase-3, which can be activated during the apoptotic process and which, in turn, eventually activates endonucleases that cause the characteristic morphology of apoptotic cells. After deparaffination and rehydratation of slides as already described, an immunoperoxidase method was performed. Briefly, slides were treated with 0.3% H 2 O 2 for 30 minutes at RT to inactivate endogenous peroxidase activity, heated in a solution of 10 mm sodium citrate at 95 C for 75 minutes to retrieve epitopes, and incubated with 10% normal goat serum and 1% BSA in Tris-buffered solution for 30 minutes at RT to block nonspecific staining. The slides were incubated in a 1:100 dilution of the primary antibody, an anti-human rabbit polyclonal antibody directed against a peptide from the p18 fragment of human caspase-3 (Anti-Active Caspase-3 pab, catalogue no. G7481; Promega Corp., Madison, WI) for 16 hours at RT. They were then incubated with a secondary antibody conjugated to peroxidase, EnVision System Labelled Polymer-HRP Anti-Rabbit (catalogue no. K4003; DakoCytomation, Carpenteria, CA), for 2 hours at RT. The presence of peroxidase was revealed by incubating with Liquid DAB Substrate Chromogen System (catalogue no. K3468; DakoCytomation) for 15 minutes at RT. Human menstrual endometrium was used as a positive control. Slides were counterstained with hematoxylin. Four replicates were performed, each including one slide from the three different groups (freshly removed ovary, after cryoprotectant exposure, and after thawing) from all three patients, as well as positive and negative controls. Ultrastructural Assessment After thawing whole ovaries, samples of the cortex and medulla were taken for ultrastructural evaluation. The tissue was fixed in 2.5% glutaraldehyde for 2 5 days at 4 C. After rinsing in PBS, samples were postfixed with 1% osmium tetroxide (Agar Scientific, Stansted, Essex, United Kingdom) in PBS, and rinsed again in PBS. The samples were then dehydrated through ascending series of ethanol, immersed in propylene oxide overnight for solvent substitution, embedded in Epon 812, and sectioned with a Reichert-Jung Ultracut E ultramicrotome. Semithin sections (1 7 m thick) were stained with toluidine blue (Sigma, St. Louis, MO), examined by light microscopy (LM) with a Zeiss Axioskop microscope (Zeiss, Munich, Germany), and photographed with a Leica camera (DFC230). Ultrathin sections (60 80 nm) were cut with a diamond knife, mounted on copper grids, and contrasted with saturated uranyl acetate followed by lead citrate. They were examined and photographed with the use of Zeiss EM109 and Zeiss EM 10 electron microscopes at 80 kv. Criteria established by Motta et al. (33) Fertility and Sterility 1155

4 were used to define the developmental stage of small preantral follicles: primordial (diplotene oocyte, surrounded by a single layer of flattened follicular cells) or primary (growing oocyte, surrounded by a single layer of cuboid follicular cells). The following elements were evaluated for qualitative assessment of ultrastructural preservation: nuclear content, membrane integrity, density of the cytoplasm and intramitochondrial matrix, cytoplasmic organelles (quality, type, and microtopography), and intercellular contacts (between oocytes and follicular cells, and between endothelial cells). Statistical Analysis The significance of differences observed in TUNEL-positive surface area between groups was tested by one-way analysis of variance. Probability values of at least P.05 were considered statistically significant. Analyses were carried out using SPSS 11.5 (SPSS Inc., Chicago, IL). RESULTS Assessment of Apoptosis Apoptosis was assessed in freshly removed ovaries, after DMSO exposure before freezing, and after thawing. In each group, apoptosis was quantified in primordial and primary follicles and evaluated in the rest of the ovarian components, i.e., antral follicles, corpora lutea, stromal cells, and cell types composing vessels (endothelial cells, pericytes, and smooth muscle cells). The percentage of apoptotic surface area was also analyzed by morphometry. Analysis of DNA Strand Breaks by TUNEL Fifty-six primordial and primary follicles were analyzed in fresh ovarian tissue, 131 after DMSO exposure, and 199 after thawing (Fig. 1A). All were negative for TUNEL. The TUNEL-positive cells were mainly found in antral follicles (Fig. 1B) and corpora lutea. Occasionally, a few isolated stromal cells stained positive for TUNEL. In all three groups, endothelial cells, pericytes, and smooth muscle cells of medullar and cortical blood vessels of different types (from capillaries to arteries and veins; Fig. 1C,D) were found to be negative for TUNEL. Only a few isolated endothelial cells were TUNEL-positive. None of the negative controls showed any TUNEL-positive signals, while TUNEL staining was observed in all positive controls. Morphometric analysis of TUNEL-positive surface area was carried out, excluding antral follicles and corpora lutea. We analyzed 134 HPFs in fresh ovary, 130 after DMSO exposure, and 622 after thawing. The mean values of TUNEL-positive surface area were 0.05% 0.03% in fresh ovary, 0.02% 0.01% after DMSO exposure, and 0.09% 0.03% after thawing (Table 1). The results revealed that the total surface area of cells with cleaved DNA was 0.1% in all three groups, which was not statistically significant. Immunohistochemical Detection of Active Caspase-3 Early apoptosis was detected by means of immunohistochemical analysis of active caspase-3. We analyzed 121 primordial and primary follicles in fresh ovarian tissue, 182 after cryoprotectant exposure, and 244 after thawing, and all were negative for active caspase-3. Active caspase-3- positive cells were found in antral follicles (Fig. 2D) and corpora lutea (Fig. 2C). Smooth muscle cells, pericytes, and the majority of endothelial cells were found to be active caspase-3-negative, irrespective of the size of the vessels (from capillaries to arteries and veins) (Fig. 2A,B). Some endothelial cells were occasionally slightly stained. Nevertheless, after counting 2,000 endothelial cells per group, the percentage of active caspase-3-positive endothelial cells was 1% in all groups: fresh tissue, after DMSO exposure, and after thawing. None of the negative controls showed any active caspase-3 expression, while positive signals were consistently observed in positive controls. The difference in the number of follicles analyzed per group (fresh, DMSO-exposed, and frozen-thawed), both for TUNEL analysis (56, 131, and 199 follicles, respectively) and for caspase detection (121, 182, and 244 follicles, respectively), was partly due to the variable density and uneven distribution of follicle clusters in adult women, both from patient to patient and between different fragments of the same ovary (34). Furthermore, ovarian samples from the fresh and DMSO-exposed groups were smaller than those from the frozen-thawed group, because the former consisted of biopsies of ovarian tissue taken from whole ovaries during the cryopreservation procedure, while the latter were the remaining whole ovaries. This also explains the difference in the number of HPFs assessed per group (134, 130, and 622, respectively) in the morphometric analysis. Ultrastructural Assessment Ovarian follicles and vessels were analyzed by LM and transmission electron microscopy (TEM) in frozen-thawed tissue. Follicular Compartment Thirty follicles (10 per patient) with the characteristics of primordial and primary follicles were investigated in frozenthawed tissue by LM and TEM. By LM, these follicles looked like regularly rounded structures, varying in diameter from m. The oocytes were surrounded by a single layer of follicular cells, flattened or cuboidal in shape, in primordial and primary follicles, respectively. At the periphery, all the follicles were observed along a continuous basal membrane. By TEM, 29 (96.7%) of these follicles appeared totally healthy-looking, presenting all the ultrastructural fea Martinez-Madrid et al. Freezing whole human ovaries Vol. 87, No. 5, May 2007

5 FIGURE 1 Assessment of apoptosis. Analysis of DNA strand breaks by TUNEL in frozen-thawed whole ovary. (A) Primary follicle and stromal cells: (A1) DAPI counterstaining, and (A2) TUNEL staining. (B) Antral follicle: (B1) DAPI counterstaining, and (B2) TUNEL staining. (C and D) Blood vessels of different sizes: (C1 and D1) DAPI counterstaining, and (C2 and D2) TUNEL staining. Bars 100 m (A, C, and D), and 50 m (B). Fertility and Sterility 1157

6 TABLE 1 Morphometric analysis of apoptosis by the TUNEL method before freezing, after DMSO exposure, and after thawing. HPFs analyzed Mean TUNEL-positive surface area Before freezing a After DMSO a exposure After thawing a a P.05 (nonsignificant). tures of normal follicles (Fig. 3): intact nuclear and cellular membranes, normally arranged chromatin, well-preserved cytoplasmic organelles, and a continuous normal thick basal membrane, according to criteria established by Motta et al. (33). In just one follicle out of 30 analyzed, some altered features were detected (Fig. 4). Healthy-looking follicles showed oocytes with a large vesicular nucleus, in which the chromatin appeared finely granular and dispersed. One or more nucleoli were seen in the nucleoplasm (Fig. 3A). In the oocyte cytoplasm, the majority of organelles were close to the nucleus, forming Balbiani s vitelline body. Rounded mitochondria with a pale matrix and a few peripheral cristae, membranes of endoplasmic reticulum, free ribosomes, and, occasionally, multivesicular bodies were found in the cytoplasm of oocytes (Fig. 3A,B). FIGURE 2 Assessment of apoptosis. Immunohistochemical detection of active caspase-3 in frozen-thawed whole ovary. (A) Arteriole negative for active caspase-3 staining. (B) Vein negative for active caspase-3 staining. (C) Corpus luteum with luteal cells positive for active caspase-3 staining. (D) Antral follicle with GCs positive for active caspase-3 staining. Bars 100 m (A and B), and 50 m (C and D) Martinez-Madrid et al. Freezing whole human ovaries Vol. 87, No. 5, May 2007

7 FIGURE 3 Primary follicle with well-preserved features in frozen-thawed whole ovary. (A) The oocyte (O) is surrounded by a single layer of cuboidal follicular cells (Fc) on a continuous basal membrane (bm). Note the mitochondria (m) clustered in clouds in the oocyte cytoplasm, and the nucleolus (nu) visible in the oocyte nucleus (N). *Two stromal cells are visible close to the basal membrane. (B) Detail of a rounded mitochondrion with a pale matrix and peripheral cristae in the oocyte cytoplasm. (C) Detail of close interdigitations observed between follicular cell (Fc) projections and oocyte (O) microvilli. Note a desmosome connecting two follicular cells (arrow). (D) Detail of follicular cells, showing indented nuclei with peripheral patches of heterochromatin and numerous rod-shaped mitochondria (m) in the cytoplasm. bm continuous basal membrane. Bars 1 m (A), 0.25 m (B), 0.43 m (C), and 0.62 m (D). At the oocyte-follicular cell interface, close interdigitations between oocyte microvilli and follicular cell projections were seen (Fig. 3C). Follicular cells showed a voluminous, indented nucleus, in which the chromatin was distributed in isolated peripheral patches (Fig. 3A,D). This is a typical feature of follicular cells, and does not represent a sign of apoptosis. Rod-shaped mitochondria were observed dispersed in the follicular cell cytoplasm (Fig. 3D). Numer- Fertility and Sterility 1159

8 FIGURE 4 Primordial follicle with altered features in frozenthawed whole ovary (in total, 3.3% of primordial and primary follicles showed ultrastructural alterations). In the oocyte (O), note rupture of the nuclear membrane and condensation of chromatin (arrow), and among the follicular cells (Fc), one altered follicular cell with condensation of chromatin (arrowhead). N oocyte nucleus. Bar 2.27 m. size, were analyzed in frozen-thawed tissue by LM and TEM. The majority of vessels were totally healthy-looking, with all vascular compartments (endothelial cells, basal membrane, smooth muscle cells, and pericytes) well preserved (Fig. 5A). They had a continuous endothelium and intact junctions between the endothelial cells, which showed a wellpreserved nucleus containing normally arranged chromatin (Fig. 5B). No vessels showing any loss of endothelial cells or exposed connective tissue were observed. Arterioles showed thick endothelial cells protruding into the lumen, with abundant cytoplasmic processes toward either the lumen or the basal membrane (Fig. 5B), while venules contained flat endothelial cells with an elongated nucleus and a thin muscular layer (Fig. 5A). In total, 187 endothelial cells of all vessel types were investigated, 180 (96.3%) of which were healthy-looking. In these endothelial cells, cytoplasm organelles, such as numerous abluminal and basal pinocytotic vesicles and mitochondria, appeared well preserved (Fig. 5B,C). A small number of endothelial cells (7/187) with altered features were encountered in a few arterioles and venules. These included endothelial cells with nuclear and cytoplasmic condensation (Fig. 6A), with a swollen and electrontranslucent cytoplasm typical of edematous cells (Fig. 6B), or with separations of intercellular contacts (Fig. 6A). No capillaries showed an altered ultrastructure. ous junctions (gap junctions, desmosomes, and tight junctions) connected the follicular cells to each other (Fig. 3C). A continuous basal membrane of normal thickness ( m) divided all the follicles from the stroma (Fig. 3A,D). Only one primordial follicle showed some morphological alterations. These included rupture of the oocyte nuclear membrane, condensation of chromatin, and oocyte mitochondria with swollen and disrupted cristae (Fig. 4), which suggested necrosis. The altered oocyte was surrounded by a continuous layer of flattened follicular cells, among which both normal and apoptotic cells were found (Fig. 4). Despite the alterations found in this follicle, normal-featured interdigitations were detected between the oocyte and the follicular cells, as well as a continuous basal membrane surrounding the follicle. Vascular Compartment Sixty blood vessels (20 per patient) of different types (arterioles, capillaries, and venules), ranging from m in DISCUSSION Cryopreservation of human ovarian cortex with the use of standard slow freezing and rapid thawing protocols appears to maintain the viability, morphology, and function of such tissue after thawing (35). However, follicular loss due to ischemia after transplantation significantly shortens the functional life span of a graft. Ovarian cortical strip transplantation may reverse the menopause and restore fertility in patients, but only for relatively short periods of time until the follicular reserve is depleted, with a subsequent return to menopausal levels (5, 36). Vascular transplantation of intact ovaries would reduce the period of posttransplantation ischemia to just a few minutes, allowing more follicles to survive and extending the period of time that ovarian grafts remain functional (32). Encouraging results were recently documented in animals. Wang et al. (16) and Yin et al. (17) described successful vascular transplantation of frozen-thawed rat ovaries and reproductive tract in 4 out of 7 (57%) transplants that survived for 60 days, were ovulatory, and resulted in one pregnancy. Chen et al. (26) showed that frozen-thawed rabbit ovaries remained functional for at least 7 months after microvascular transplantation in 13 of 15 (86.7%) animals. Three recent studies achieved good results after freezethawing whole ovaries in large mammal species, i.e., pigs (31) and sheep (29, 30), with the birth of a live lamb (30) Martinez-Madrid et al. Freezing whole human ovaries Vol. 87, No. 5, May 2007

9 FIGURE 5 Blood vessels showing normal ultrastructure in frozen-thawed whole ovary. (A) Venule with flattened endothelial cells. (B) Arteriole containing thickened endothelial cells protruding into the lumen, with abundant cytoplasmic processes toward either the lumen or the basal membrane. The cytoplasm is filled with organelles and intact connections between endothelial cells. (C) Detail of mitochondria (m) with intact membranes from an endothelial cell. N endothelial cell nucleus. L vascular lumen. Bars 1.56 m (A), 1 m (B), and 0.3 m (C). Fertility and Sterility 1161

10 FIGURE 6 Arterioles with a few altered endothelial cells in frozen-thawed whole ovary (in total, 3.7% of endothelial cells showed ultrastructural alterations). (A) Separation of intercellular connections. *An endothelial cell with nuclear and cytoplasmic condensation (arrow) in an arteriole. (B) Swollen endothelial cell with electron-translucent cytoplasm in an arteriole. L vascular lumen. Bars 2.2 m (A) and 0.75 m (B). However, since sheep ovaries are about one tenth of the size of human ovaries, problems related to heat and mass transfer during freezing and thawing are presumed to be more marked in humans. The aim of this study was to investigate whether the present cryopreservation protocol would induce apoptosis in any human ovarian cells, either after cryoprotectant perfusion or after freeze-thawing, and to further analyze any ultrastructural alterations in the thawed ovary. The hallmark of apoptosis is the presence of typical intranucleosomal DNA fragmentation, which is identifiable by the DNA end-labeling method (TUNEL). Fragmentation of DNA is triggered by the activation of specific endonucleases, such as caspases. In particular, caspase-3 is the principal downstream effector enzyme of cell death, and it triggers the activation of endogenous endonucleases. Therefore, the presence of caspase-3 identifies the early stage of apoptosis in those cells that are due to undergo programmed cell death. It is well known that apoptosis, the mechanism by which cells are eliminated through programmed cell death, also occurs in the ovary (37 39), where the process of remodeling is continuous. Physiologically, apoptosis is seen in luteal and thecal cells of healthy corpora lutea, as well as in granulosa cells (GCs) of antral follicles. Despite this physiologic removal of cells through apoptosis, assessment of apoptosis recently proved to be a good method to evaluate and compare the outcome of different freeze-thawing protocols for human ovarian tissue (40). Indeed, some cryopreservation protocols induce apoptosis in frozen-thawed cells, which was detected by TUNEL (40) and by immunohistochemical expression of active caspase-3 (41). Although the mechanism of cell death during cryopreservation is not fully understood, Tirelli et al. (42) suggested that increased apoptosis in GCs after cryopreservation was probably caused by physical alterations due to low temperature, high salt concentrations, and an impaired antioxidant metabolism. Moreover, it was shown that cold preservation (43) and the ischemic process (44) may induce apoptosis. Cryopreservation of a whole ovary requires a relatively long period of ischemia and cold preservation during cryoprotectant perfusion, which might induce apoptosis. It was proved that apoptosis plays a crucial role after ischemia-reperfusion injury in organ transplantation, and specific inhibition of caspase-3 decreases the number of apoptotic endothelial cells in rat liver transplantation (45), improving the survival of grafts. Thus, it is likely that the absence of apoptosis activation in endothelial cells would be a good marker to predict the outcome of vascular transplantation. In the present study, TUNEL and immunohistochemical detection of active caspase-3 were applied to investigate induction of apoptosis in whole human ovaries as a result of cold storage and ischemia during perfusion before freezing (DMSO-exposed group), and as a result of the cryopreservation procedure itself (frozen-thawed group). We did not find any signs of induction of apoptosis, either by the TUNEL method or active caspase-3 expression, in primordial or primary follicles, either after DMSO exposure 1162 Martinez-Madrid et al. Freezing whole human ovaries Vol. 87, No. 5, May 2007

11 and perfusion or after thawing. Our results in humans are completely in line with those obtained by Dermici et al. (46) and Bedaiwy et al. (27) in animals. Indeed, using TUNEL, they did not observe any induction of apoptosis in primordial or primary follicles after thawing of sheep ovarian cortex (46) or whole sheep ovaries (27). Rimon et al. (40), however, reported TUNEL-detected induction of apoptosis in primordial and primary follicles after cryopreservation of human ovarian cortical pieces. We did not detect any induction of apoptosis in vessels (endothelial cells, smooth muscle cells, or pericytes) or stromal cells in any group, either by TUNEL or IHC for active caspase-3 detection. The absence of induction of apoptosis in vascular cells, especially endothelial cells, is of great importance because they are in the front line of exposure during cryoprotectant perfusion. Once again, our results in humans confirm those observed in sheep. Arav et al. (29) found no difference in vascular or follicular survival or tissue morphology before and after freeze-thawing whole sheep ovaries. Recovery of vascular function after thawing was proved by the production of factor VIII by endothelial cells. In our study, TUNEL-positive cells or active caspase-3- positive cells were mainly found in GCs of antral follicles and in luteal cells of corpora lutea, which is a physiological feature. Excluding these antral follicles and corpora lutea from the assessment, our morphometric analysis of TUNELpositive surface area revealed no significant differences in the total surface area of cells with cleaved DNA, which was 0.1% in all three groups. Transmission electron microscopy analysis was subsequently applied to complete our apoptosis evaluation by investigating possible ultrastructural damage to ovarian tissue after thawing. TEM is a well-known method for detecting any subcellular alterations (such as fine damage to membranes and organelles) that may occur in tissue after cryopreservation procedures. Indeed, TEM analysis of follicular integrity in human ovarian cortical fragments subjected to cryopreservation has already been used (11,47 50). We analyzed, by TEM, three cryopreserved whole human ovaries, paying particular attention to the follicular and vascular compartments. No discernible ultrastructural alterations were encountered in frozen-thawed ovaries, confirming our previous histology and viability findings (32), as well as the results of the present study of apoptosis. All the follicles but one (96.7%) were healthy-looking, and exhibited normal ultrastructural features. Our results with thawed whole human ovaries are in total agreement with those of Nisolle et al. (11), Eyden et al. (50), and Hreinsson et al. (49) with human ovarian cortical pieces. They reported similarities in the ultrastructural quality of primordial and primary follicles before and after freeze-thawing avascular human ovarian cortical fragments. Nevertheless, in these studies, a few altered follicles containing either vacuolated oocytes (11, 49) or follicular cells with pyknotic nuclei or a swollen cytoplasm (49, 50) were occasionally detected. In animals, only one study has so far used TEM to evaluate ovarian tissue after cryopreservation of whole ovaries, in this case porcine (31). It demonstrated the feasibility of intact ovary cryopreservation in this model, showing good viability (84.4%) and a well-preserved ultrastructure of the follicular compartment after cryostorage. In our opinion, proving the ultrastructural integrity of the vascular compartment is of crucial importance in any cryopreservation protocol for whole organ vascular transplantation. Indeed, endothelial cell loss and apoptosis were also demonstrated by TEM in the iliac artery (51) and liver vessels (43) after cryopreservation procedures. However, to our knowledge, no ultrastructural study has been performed on cryopreserved whole ovaries in any species to evaluate their vascular compartment, even though cryoinjury to the vascular compartment could result in organ failure after transplantation. Therefore, in the present study, blood vessels of different types and sizes were scrupulously analyzed. All were found to be well preserved, with no discernible ultrastructural alterations in the endothelial cells, basal membrane, pericytes, or smooth muscle cells. Indeed, we did not observe any subcellular alterations, cell-cell contact detachment, or loss of endothelial cells, features which were described in ischemia-reperfusion-injured or cryoinjured vessels in other organs (43, 51, 52). In conclusion, the present study shows that cryopreservation of an intact human ovary with its vascular pedicle is not associated with induction of apoptosis or ultrastructural alterations in any of the ovarian components. These data, together with our previous findings of high survival rates and normal histological structure in all ovarian elements after thawing (32), lead us to conclude that the present freezethawing protocol does not cause significant injury to human ovaries, and may be a suitable technique for the cryopreservation of whole ovaries before transplantation. Acknowledgments: The authors thank Mira Hryniuk, B.A., for reviewing the manuscript, and G.. Macchiarelli, M.D., for fruitful discussions and critical review. The authors also thank the Department of Anatomopathology, Université Catholique de Louvain, Brussels, Belgium, for advice in the setup of caspase-3 IHC, for help with TEM procedures, and for specimenembedding and hematoxylin-eosin staining. REFERENCES 1. Donnez J, Bassil S. Indications for cryopreservation of ovarian tissue. Hum Reprod Update 1998;4: Donnez J, Godin PA, Qu J, Nisolle M. Gonadal cryopreservation in the young patient with gynaecological malignancy. Curr Opin Obstet Gynecol 2000;12: The Ethics Committee of the American Society for Reproductive Medicine. Fertility preservation and reproduction in cancer patients. Fertil Steril 2005;83: Stachecki JJ, Cohen J. An overview of oocyte cryopreservation. Reprod Biomed Online 2004;9: Donnez J, Dolmans MM, Martinez-Madrid B, Demylle D, Van Langendonckt A. The role of cryopreservation for women prior to treatment of malignancy. Curr Opin Obstet Gynecol 2005;17: Fertility and Sterility 1163

12 6. Baird DT, Webb R, Campbell BK, Harkness LM, Gosden RG. Longterm ovarian function in sheep after ovariectomy and transplantation of autografts stored at 196 C. Endocrinology 1999;140: Oktay K, Buyuk E, Veeck L, Zaninovic N, Xu K, Takeuchi T, et al. Embryo development after heterotopic transplantation of cryopreserved ovarian tissue. Lancet 2004;363: Silber SJ, Lenahan KM, Levine DJ, Pineda JA, Gorman KS, Friez MJ, et al. Ovarian transplantation between monozygotic twins discordant for premature ovarian failure. N Engl J Med 2005;353: Donnez J, Dolmans MM, Demylle D, Jadoul P, Pirard C, Squifflet J, et al. Livebirth after orthotopic transplantation of cryopreserved ovarian tissue. Lancet 2004;364: Meirow D, Levron J, Eldar-Geva T, Hardan I, Fridman E, Zalel Y, et al. Pregnancy after transplantation of cryopreserved ovarian tissue in a patient with ovarian failure after chemotherapy. N Engl J Med 2005; 353: Nisolle M, Casanas-Roux F, Qu J, Motta P, Donnez J. Histologic and ultrastructural evaluation of fresh and frozen-thawed human ovarian xenografts in nude mice. Fertil Steril 2000;74: Liu J, Van der Elst J, Van den Broecke R, Dhont M. Early massive follicle loss and apoptosis in heterotopically grafted newborn mouse ovaries. Hum Reprod 2002;17: Newton H, Aubard Y, Rutherford A, Sharma V, Gosden R. Low temperature storage and grafting of human ovarian tissue. Hum Reprod 1996;11: Candy CJ, Wood MJ, Whittingham DG. Effect of cryoprotectants on the survival of follicles in frozen mouse ovaries. J Reprod Fertil 1997;110: Gunasena KT, Villines PM, Critser ES, Critser JK. Live births after autologous transplant of cryopreserved mouse ovaries. Hum Reprod 1997;12: Wang X, Chen H, Yin H, Kim SS, Tan SL, Gosden RG. Fertility after intact ovary transplantation. Nature 2002;415: Yin H, Wang X, Kim SS, Chen H, Tan SL, Gosden RG. Transplantation of intact rat gonads using vascular anastomosis: effects of cryopreservation, ischemia and genotype. Hum Reprod 2003;18: Winston RM, Browne JC. Pregnancy following autograft transplantation of fallopian tube and ovary in the rabbit. Lancet 1974;2: Jeremias E, Bedaiwy MA, Gurunluoglu R, Biscotti CV, Siemionow M, Falcone T. Heterotopic autotransplantation of the ovary with microvascular anastomosis: a novel surgical technique. Fertil Steril 2002;77: Goding JR, McCracken JA, Baird DT. The study of ovarian function in the ewe by means of a vascular autotransplantation technique. J Endocrinol 1967;39: Paldi E, Gal D, Barzilai A, Hampel N, Malberger E. Genital organs. Auto and homotransplantation in forty dogs. Int J Fertil 1975;20: Scott JR, Keye WR, Poulson AM, Reynolds WA. Microsurgical ovarian transplantation in the primate. Fertil Steril 1981;36: Leporrier M, Von Theobald P, Roffe JL, Muller G. A new technique to protect ovarian function before pelvic irradiation. Cancer 1987;60: Hilders CG, Baranski AG, Peters L, Ramkhelawan A, Trimbos JB. Succesful human ovarian autotransplantation to the upper arm. Cancer 2004;101: Mhatre P, Mhatre J, Magotra R. Ovarian transplant: a new frontier. Transplant Proc 2005;37: Chen C, Chen S, Chang F, Wu G, Liu J, Yu C. Autologous heterotopic transplantation of intact rabbit ovary after cryopreservation. Hum Reprod 2005;20(Suppl 1):i Bedaiwy A, Jeremias E, Gurunluoglu R, Hussein MR, Siemianow M, Biscotti C, et al. Restoration of ovarian function after autotransplantation of intact frozen-thawed sheep ovaries with microvascular anastomosis. Fertil Steril 2003;79: Revel A, Elami A, Bor A, Yavin S, Natan Y, Arav A. Whole sheep ovary cryopreservation and transplantation. Fertil Steril 2004;82: Arav A, Revel A, Nathan Y, Bor A, Gacitua H, Yavin S, et al. Oocyte recovery, embryo development and ovarian function after cryopreservation and transplantation of whole sheep ovary. Hum Reprod 2005; 20: Imhof M, Bergmeister H, Lipovac M, Rudas M, Hofstetter G, Huber J. Orthotopic microvascular re-anastomosis of whole cryopreserved ovine ovaries resulting in pregnancy and live birth. Fertil Steril 2006; 85: Imhof M, Hofstetter G, Bergmeister H, Rudas M, Kain R, Lipovac M, et al. Cryopreservation of a whole ovary as a strategy for restoring ovarian function. J Assist Reprod Genet 2004;21: Martinez-Madrid B, Dolmans MM, Van Langendonckt A, Defrere S, Donnez J. Freeze-thawing intact human ovary with its vascular pedicle with a passive cooling device. Fertil Steril 2004;82: Motta PM, Nottola SA, Familiari G, Makabe S, Stallone T, Macchiarelli G. Morphodynamics of the follicular-luteal complex during early ovarian development and reproductive life. Int Rev Cytol 2003;223: Schmidt KLT, Byskov AG, Nyboe Andersen A, Muller J, Yding Andersen C. Density and distribution of primordial follicles in single pieces of cortex from 21 patients and in individual pieces of cortex from three entire human ovaries. Hum Reprod 2003;18: Hovatta O. Methods for cryopresevation of human ovarian tissue. RBM Online 2005;6: Donnez J, Dolmans MM, Demylle D, Jadoul P, Pirard C, Squifflet J, et al. Restoration of ovarian function after orthotopic (intraovarian and periovarian) transplantation of cryopreserved ovarian tissue in a woman treated by bone marrow transplantation for sickle cell anaemia: case report. Hum Reprod 2006;21: Tilly JL. Apoptosis and the ovary: a fashionable trend or food for thought? Fertil Steril 1997;67: Johnson AL, Bridgham JT. Caspase-mediated apotosis in the vertebrate ovary. Reproduction 2002;124: Depalo R, Nappi L, Loverro G, Bettocchi S, Caruso ML, Valentini AM, et al. Evidence of apoptosis in human primordial and primary follicles. Hum Reprod 2003;18: Rimon E, Cohen T, Dantes A, Hirsh L, Amit A, Lessing JB, et al. Apoptosis in cryopreserved human ovarian tissue obtained from cancer patients: a tool for evaluating cryopreservation utility. Int J Oncol 2005;27: Yagi T, Hardin JA, Valenzuela YM, Miyoshi H, Gores GJ, Nyberg SL. Caspase inhibition reduces apoptosis death of cryopreserved porcine hepatocytes. Hepatology 2001;33: Tirelli M, Basini G, Grasselli F, Bianco F, Tamanini C. Cryopreservation of pig granulosa cells: effect of FSH addition to freezing medium. Domest Anim Endocrinol 2005;28: Moriga T, Arii S, Takeda Y, Furuyama H, Mizumoto M, Mori A, et al. Protection by vascular endothelial growth factor against sinusoidal endothelial damage and apoptosis induced by cold preservation. Transplantation 2000;69: Kim SS, Yang HW, Kang HG, Lee HH, Lee HC, Ko DS, et al. Quantitative assessment of ischemic tissue damage in ovarian cortical tissue with or without antioxidant (ascorbic acid) treatment. Fertil Steril 2004;82: Mueller TH, Kienle K, Beham A, Geissler EK, Jauch KW, Rentsch M. Caspase 3 inhibition improves survival and reduces early graft injury after ischemia and reperfusion in rat liver transplantation. Transplantation 2004;15: Dermici B, Salle B, Frappart L, Franck M, Guerin JF, Lornage J. Morphological alterations and DNA fragmentation in oocytes from primordial and primary follicles after freezing-thawing of ovarian cortex in sheep. Fertil Steril 2002;77: Gook DA, Edgar DH, Stern C. Effect of cooling rate and dehydration regimen on the histological appearance of human ovarian cortex following cryopreservation in 1,2-propanediol. Hum Reprod 1999;14: Gook DA, Edgar DH, Stern C. The effects of cryopreservation regimens on the morphology of human ovarian tissue. Mol Cell Endocrinol 2000;169: Martinez-Madrid et al. Freezing whole human ovaries Vol. 87, No. 5, May 2007

13 49. Hreinsson J, Zhang P, Swahn ML, Hultenby K, Hovatta O. Cryopreservation of follicles in human ovarian cortical tissue. Comparison of serum and human serum albumin in the cryoprotectant solutions. Hum Reprod 2003;18: Eyden B, Radford J, Shalet SM, Thomas N, Brison DR, Lieberman BA. Ultrastructural preservation of ovarian cortical tissue cryopreserved in dimethylsulfoxide for subsequent transplantation into young female cancer patients. Ultrastruct Pathol 2004;28: Pascual G, Rodriguez M, Corrales C, Turegano F, Garcia-Honduvilla N, Bellon JM, et al. New approach to improving endothelial preservation in cryopreserved arterial substitutes. Cryobiology 2004;48: Chiavarelli R, Macchiarelli G, Familiari G, Chiavarelli M, Macchiarelli P, Del Basso P, et al. Ultrastructural changes of coronary artery endothelium induced by cardioplegic solutions. Thorac Cardiovasc Surg 1989;37: Fertility and Sterility 1165