del227.fm Page 1 Tuesday, May 30, 2006 12:23 PM ARTICLE IN PRESS Human Reproduction Page 1 of 12 doi:10.1093/humrep/del227

Cryopreservation of intact human with its vascular pedicle

Mohamed A.Bedaiwy1,2, Mahmoud R.Hussein3, Charles Biscotti4 and Tommaso Falcone1,5

1Department of Obstetrics and Gynecology, Minimally Invasive Surgery Center, The Cleveland Clinic Foundation, Cleveland, OH, USA, 5 2Department of Obstetrics and Gynecology, 3Department of Pathology, Assiut University Hospitals and School of Medicine, Assiut, Egypt and 4Anatomic Pathology Department, Minimally Invasive Surgery Center, The Cleveland Clinic Foundation, Cleveland, OH, USA 5To whom correspondence should be addressed at: Department of Obstetrics and Gynecology, A81, The Cleveland Clinic Foundation, 9500 Euclid Avenue, Cleveland, OH 44195, USA. E-mail: [email protected]

10 BACKGROUND: The aim of this study was to assess the immediate post-thawing injury to the human ovary that was cryopreserved either as a whole with its vascular pedicle or as ovarian cortical strips. MATERIALS AND METHODS: Bilateral oophorectomy was performed in two women (46 and 44 years old) undergoing vaginal hysterectomy and laparoscopic hysterectomy, respectively. Both women agreed to donate their for experimental research. In both patients, one of the harvested ovaries was sectioned and cryopreserved (by slow freezing) as ovarian cortical 15 strips of 1.0 ´ 1.0 ´ 5.0 mm3 each. The other ovary was cryopreserved intact with its vascular pedicle. After thawing 7 days later, follicular viability, histology, terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-digoxigenin nick-end labelling (TUNEL) assay (to detect apoptosis) and immunoperoxidase staining (to define Bcl-2 and p53 pro- tein expression profiles) of the ovarian tissue were performed. Tissues from non-cryopreserved ovaries served as con- trol specimens (two cases). RESULTS: The overall viability of the primordial follicles was 75 and 78% in intact 20 cryopreserved–thawed (C–T) ovaries and 81 and 83% in ovarian cortical strips in the 46- and 44-year-old patients, respectively. Comparable primordial follicle counts, absence of features of necrosis, mean values of apoptosis and weak Bcl-2 and p53 protein expressions were observed both in the intact C–T ovary and in the C–T ovarian cortical strips. CONCLUSIONS: Cryoperfusion and cryopreservation of entire human ovary can be achieved with the main- tenance of excellent viability of the superficial and the deeper tissues using a slow-freezing protocol. Cryopreserva- 25 tion injury is associated neither with significant alteration in the expression pattern of Bcl-2 and p53 proteins in the ovarian tissues nor with significant follicular damage.

Key words: apoptosis/Bcl-2/cryopreservation/follicular viability/intact human ovary

Introduction can be implanted after cancer treatment as an autograft to an 45 Although many reproductive-age patients can survive their orthotopic or a heterotopic site. Immature oocytes derived from 30 cancers and lead normal lives, they are at increased risk of thawed tissue can be matured in vitro if appropriate protocols impaired reproductive functions. Consequently, fertility pres- are developed in the future. Also, the maturation process of the ervation is an important quality-of-life issue for them. Fertility immature oocytes could be achieved by xenografting in immu- preservation strategies were introduced to protect and/or regain nodeficient mice (Gook et al., 2003). 50 reproductive function in patients exposed to cancer chemother- To date, ovulation and creation of a human embryo from 35 apy and/or radiotherapy. Advances in assisted reproductive heterotopic ovarian transplant have been reported (Oktay et al., technologies (ART), such as ovarian tissue cryopreservation 2004). Moreover, two live births from orthotopic ovarian trans- and transplantation, oocyte cryopreservation and novel ovula- plants following spontaneous (Donnez et al., 2004) and IVF tion induction regimens, have renewed interest in fertility pres- (Meirow et al., 2005) cycles were reported, following modified 55 ervation in women scheduled to receive gonadotoxic cryopreservation and surgical protocols pioneered, in sheep, by 40 chemotherapy and/or radiotherapy (Falcone et al., 2004). Of Gosden and associates (Gosden et al., 1994; Baird et al., these technologies, ovarian tissue cryopreservation may be a 1999). Despite these advances, ischaemic damage to the tissue viable option for women who cannot delay treatment to undergo and revascularization injury and the theoretical possibility of ovarian stimulation to create embryos or obtain oocytes for reintroducing malignant tumour cells remain as the main 60 freezing. In this entirely experimental process, thawed tissue limitations for this option (Bedaiwy and Falcone, 2004).

© The Author 2006. Published by Oxford University Press on behalf of the European Society of Human Reproduction and Embryology. All rights reserved. 1 For Permissions, please email: [email protected] del227.fm Page 2 Tuesday, May 30, 2006 12:23 PM ARTICLE IN PRESS M.A.Bedaiwy et al.

The ischaemic damage to ovarian tissues can induce a high Materials and methods 120 rate of follicular loss (Baird et al., 1999; Demirci et al., 2002). The study group included both control (non-cryopreserved) and cryo- Therefore, eliminating this ischaemic damage could maintain preserved human ovaries. Also tissue sections from lymph nodes, 65 both the viability and the functional integrity of the transplant. squamous cell carcinoma, leukaemia cell lines, normal skin and liver Ideally, this could be achieved by transplantation of an intact were used as tissue-specific positive controls for Bcl-2, p53, apoptotic ovary with vascular anastomosis. Transplantation of an intact cells, PAS and Masson’s trichrome staining. 125 ovary with its vascular pedicle using microvascular anastomo- sis was achieved both in murine species (Yin et al., 2003) and Patients in the cryopreservation group 70 in mammals (Jeremias et al., 2002; Bedaiwy et al., 2003). The Institutional Review Board of The Cleveland Clinic Foundation Partial restoration of hormonal functions (Bedaiwy et al., approved this study. The first patient, aged 46 years, presented with 2003) and pregnancy (Wang et al., 2002) after transplantation menorrhagia and symptomatic fibroid , who underwent total vaginal hysterectomy with bilateral salpingo-oophorectomy, McCall 130 of intact cryopreserved–thawed (C–T) ovaries were also culdoplasty and cystoscopy. Intraoperative findings included an 8–10 reported. Recently, cryopreservation of intact human ovary week-sized uterus and normal-appearing ovaries. The second patient, 75 with its vascular pedicle resulting in high post-thaw survival rates aged 44 years, presented with severe premenstrual dysphoric disor- of follicles, small vessels and stromal cells as well as a normal ders, who underwent laparoscopic bilateral salpingo-oophorectomy. histological structure in all the ovarian components was Normal-appearing uterus, tubes and ovaries were detected. Both ova- 135 reported using a slow-freezing protocol. Slow freezing was ries and their pedicles, from both patients, were processed immedi- achieved using a cryofreezing one-degree container (Martinez- ately as detailed below. 80 Madrid et al., 2004). The structural homeostasis of tissues is regulated by a delicate Patients in the control group balance between cell survival and apoptotic cell death. In tis- In the control group, formalin-fixed paraffin-embedded ovarian sues, several molecules are involved in the survival (survival tissue was obtained from the Archives of the Department of 140 molecules such as Bcl-2) or cell death (apoptotic molecules Pathology, Assiut University Hospitals. Four ovarian tissue speci- mens were obtained from two women, 44 and 47 years old. The first 85 such as p53) processes. Bcl-2 is a membrane-associated protein was diagnosed with irregular uterine bleeding due to uterine that resides in the nuclear envelope and mitochondria. It exerts fibroids and was treated by total abdominal hysterectomy with its survival functions by modulating the mitochondrial release bilateral salpingo-oophorectomy. The second underwent total vag- 145 of cytochrome c and antagonizing the effects of Bax gene (Bcl-2- inal hysterectomy with bilateral salpingo-oophorectomy due to associated X). Bcl-2 is expressed in granulosa cells of both pelvic organ prolapse. The ovaries obtained from both patients 90 fetal and adult ovaries. p53 gene encodes a 53-kDa oncosup- appeared normal with a smooth white convoluted surface and firm pressive nuclear protein that functions to antagonize Bcl-2 consistency. effects. In the ovary, p53 protein is expressed in the apoptotic granulosa cells of atretic follicles. In several organs such as the Cryopreservation 150 heart, ischaemia is associated with the induction of Bcl-2 and Cryopreservation of the intact ovary 95 p53 protein expressions. Similarly, the induction of these mol- We adopted our previous protocol described in 2003 (Bedaiwy ecules following ovarian ischaemia may have far-reaching et al., 2003). Briefly, immediately after oophorectomy, one ovary effects on the outcome of the subsequent ovarian transplantation (from each patient) was perfused with heparinized Ringer’s solu- (Hussein, 2005). tion followed by perfusion and immersion in a bath containing the 155 Although previous studies examined the long-term effects cryoprotective mixture which was composed of Leibovitz L-15 medium (Irvine Scientific, Santa Ana, CA, USA), 10% fetal calf 100 of cryopreservation–thawing procedure injury on the ovar- serum (FCS) (Irvine Scientific) and 1.5 M dimethylsulphoxide ian cortical strips (Gosden et al., 1994), our understanding (DMSO) (Sigma, St. Louis, MO, USA). Both the ovarian vessels of the injury caused to ovarian tissue in the period immedi- and excess hilar tissue were dissected, and ovarian were 160 ately after thawing is still incomplete. The aims of this trimmed. The ovaries were perfused via the ovarian artery with the study were to: (i) describe a cryopreservation protocol using cryoprotective mixture using Horizon Modular Infusion System 105 the programmable freezer allowing freezing and thawing of (MCGaw, Irvine, CA, USA) to maintain a flow rate at 1.3 ml/min an intact human ovary with its vascular pedicle; (ii) assess with continuous replenishment of the reservoir. After perfusion, the survival (viability) of the ovarian elements in the C–T the ovary was bisected and transferred to two cryovials 12.7 × 92 165 tissues using Trypan Blue method; (iii) examine the frozen– mm (Corning Coaster Corporation, Cambridge, MA, USA) con- thawed ovarian tissues for features of follicular health and taining the cryoprotective mixture for controlled freezing using 110 atresia using the histological methods; (iv) evaluate the Planer cryochamber (Planer Freezer, Middlesex, UK). Cooling ° ° apoptotic changes in the ovarian tissues using combined began at +4 C and continued at 2 C per minute until ice nucleation was induced at –7°C. The temperature was then reduced at 2°C 170 histological methods and terminal deoxynucleotidyl trans- per minute until –35°C and subsequently at 25°C per minute until ferase (TdT)-mediated dUTP-digoxigenin nick-end labelling –140°C before cryovials were plunged into liquid nitrogen. (TUNEL) assay, immunoperoxidase method (p53 and Bcl-2 115 protein alterations, i.e. apoptotic and survival proteins) and Cryopreservation of the ovarian cortical strips (v) examine the status of the blood vessels using both histo- The other ovary, for each patient, was divided into ovarian cortical chemical [periodic acid Schiff (PAS) and Masson’s tri- strips of 1.0 × 1.0 × 5.0 mm3 each. The ovarian cortical strips were 175 chrome] and immunohistochemical (CD34, marker of vascular prepared as described by Gosden et al. (1994) and cryopreserved endothelium) staining methods. using the previously mentioned protocol but without perfusion. 2 del227.fm Page 3 Tuesday, May 30, 2006 12:23 PM ARTICLE IN PRESS Cryopreservation of an entire human ovary

Thawing QIA33TDT-FragEL™ kit (Oncogen Research Products, Boston, MA, Thawing of the intact ovary USA) following other groups (Hussein et al., 2006). Positive controls 240 180 One week later, the vials were removed from the dewar and held for 1 consisted of HL60 promyelocytic leukaemia cells and HL60 cells μ min at room temperature before plunging in a bath of water at 37°C incubated with 0.5 g/ml actinomycin D for 19 h to induce apoptosis. with gentle shaking. The contents of the vials were quickly emptied Some ovarian specimens were used as negative controls by substitut- into a Petri dish containing Leibovitz L-15 medium supplemented ing distilled water for deoxynucleotidyl transferase (Gavrieli et al., with 10% FCS. The ovaries were washed and immediately perfused 1992; Liu et al., 1995). 245 185 with Leibovitz L-15 medium supplemented with 10% FCS using a Results of the TUNEL assay were evaluated following other groups × flow rate of 1.3 ml/min for 20 min. The cryoprotectant was gradually (Hussein et al., 2006). All slides were examined at 400 magnifica- eliminated by pumping Leibovitz L-15 supplemented with 10% FCS tion. Ten different areas of each follicular wall were examined, and into the reservoir (Bedaiwy et al., 2003). apoptotic index was determined as mean number of positively stained cells (Hussein et al., 2006). Thawing of the ovarian cortical strips Immunohistochemical evaluation of p53, Bcl-2 and CD34 protein 250 190 Ovarian cortical strips were thawed using the same procedure as for expressions in the C–T and control ovarian tissues intact ovary. Then, the strips were washed and held in Leibovitz L-15 containing 10% FCS for 20 min. To examine the prosurvival, proapoptotic proteins and the status of the blood vessels, we used immunoperoxidase staining methods and In vitro assessments of the C–T and control ovarian tissues monoclonal antibodies targeting p53, Bcl-2 and CD34 proteins. Briefly, ovary sections mounted on glass slides were deparaffinized 255 Evaluation of the ovarian follicular viability in the C–T ovarian and rehydrated. Endogenous peroxidase activity was blocked with 195 tissue 0.6% H202. Sections were then immersed in the retrieval solution (10 The cryopreserved ovarian tissues were evaluated for follicular viabil- mM sodium citrate buffer, pH 6.0) and subjected to heat-induced anti- ity in terms of plasma membrane function and structural integrity by gen retrieval for 20 min (microwave at ∼750 W). Non-specific protein the Trypan Blue exclusion test. Ovarian fragments were thinly sec- binding was blocked with 10 min exposure to 10% normal goat 260 tioned in Leibovitz L-15 medium supplemented with 1 mg/ml (200 serum. Sections were then incubated with primary antibodies for 1 h ° 200 IU/ml) type 1 collagenase (Sigma), incubated at 37 C for 2 h and at room temperature. The following primary mouse monoclonal anti- pipetted every 30 min. Collagenase activity was inhibited by 50% bodies were used: clone DO-7, Catalogue No. M7001 and clone 124, μ FCS. The suspension was filtered through a 70- m nylon filter (Becton Catalogue No. M0889 for p53 and Bcl-2, respectively (DAKO, Dickinson Labware, Frankline Lakes, NJ, USA) and centrifuged at Carpinteria, CA, USA). Anti-CD34 monoclonal QBEnd/10 (Novocastro, 265 400 g for 5 min. The precipitate was diluted with 50 ml of Leibovitz Newcastle, UK) was used to examine the blood vessels. After brief ° 205 L-15 medium and kept in a water bath at 37 C. Trypan Blue (0.4%, rinsing in phosphate-buffered saline (PBS), a catalyzed signal amplifi- μ Sigma) was added to the suspension containing the follicles (20 l), cation system (K1500, DAKO) was used according to the manufac- deposited on a glass slide and examined under inverted microscope turer’s instructions. Two observers independently evaluated the × ( 400). For each fragment, 100 small intact follicles were examined stained slides. 270 while the partially or completely denuded oocytes were excluded. 210 Both the number of stained cells and the total number of cells were Positive controls counted. The percentage of viable cells was determined by calculating Sections from reactive lymphoid hyperplasia (lymph node), p53-pos- the percentage of unstained cells. itive squamous cell carcinoma and haemangioma were used as posit- Histological evaluation of the ovarian tissues for the features ive controls for p53, Bcl-2 and CD34 staining, respectively. of follicular health and atresia in the C–T and control ovarian 215 tissues. The maturing follicle is composed of an oocyte with granu- Negative controls 275 losa layer and lacks reticulum. The hallmarks of the healthy follicles Additional sections of the tissues were stained in parallel but with include intact oocyte, intact membrana granulosa, lack of necrosis and omission of the primary antibody. the presence of few pyknotic nuclei (<5% pyknotic nuclei) in this 220 layer. The atretic follicles are characterized by the presence of attenu- PAS and Masson’s trichrome stains ated membrana granulosa, loosely attached granulosa cells and PAS stains carbohydrate moieties and basement membrane reticulin increased number of pyknotic nuclei (>5% pyknotic nuclei) (Wood and therefore delineates the wall of the blood vessels. Masson’s tri- 280 et al., 1997). Initially, the ovarian follicles were sorted into two chrome stains myofibres, indicating skeletal or smooth muscle differ- groups, healthy and atretic, following Wood et al. (1997). entiation (versus non-muscle fibres). Thus, it stains the smooth muscle 225 Histological evaluation of apoptotic changes in the C–T and of the blood vessels. control ovarian tissues. Evaluation of apoptosis was performed using combined histological and TUNEL assays. The results were expressed Evaluation of Bcl-2 and p53 immunostaining as mean and SEM. The histological criteria of apoptosis included For the evaluation of Bcl-2 and p53 immunoreactivities, several 285 230 condensed nuclear fragments, nuclei with marginated chromatin, rules were followed: (i) corresponding sections stained by haema- multiple nuclear fragments, a single condensed nucleus, membrane- toxylin and eosin were examined; (ii) sections were examined to bound structures containing variable amounts of chromatin and eosi- detect the sites of the antibody positivity; (iii) a higher power mag- nophilic cytoplasm (Kerr et al., 1972). nification was used to evaluate the immunostaining; (iv) sections were examined independently by two observers and (v) Bcl-2 and 290 DNA fragmentation assay for the detection of apoptotic cells p53 positivity was identified as diffuse brown cytoplasmic and 235 in the C–T and control ovarian tissues nuclear staining, respectively. The numbers of follicles with posit- To evaluate the apoptotic changes, we examined cellular morphol- ive and negative reactivity were counted in the ovarian tissues ogy with the TUNEL assay, using the commercially available (Hussein et al., 2006). 3 del227.fm Page 4 Tuesday, May 30, 2006 12:23 PM ARTICLE IN PRESS M.A.Bedaiwy et al.

295 Evaluations of the status of the blood vessels in the C–T and control Histological changes in the control and C–T ovarian tissues 325 ovarian tissues In the control ovarian tissues, evaluation of the cortex and We chose CD34, as it is more sensitive than other markers for the medulla revealed the absence of the histological features of detection of endothelial cells in the ovary. Microvessel density necrosis such as: (i) swelling and clumping of the chromatin, (MVD) was assessed by light microscopy following Weidner et al. (ii) pyknosis (condensation of the chromatin and shrinkage of 300 (1991). Ovarian tissue was scanned at low magnification (×40 and ×100) to select the areas that showed the most intense vascularization the nucleus), (iii) karyorrhexis (fragmentation of the nucleus), 330 (hot spots). Individual microvessels were counted in three fields at (iv) karyolysis (dissolution of the nucleus) and (v) cytoplasmic ×200 magnification (×20 objective lens and ×10 ocular lens; 0.7386 opacification or eosinophilia (Figure 1). Similarly, in the cryo- mm2/field). The final MVD was the mean value obtained from preserved ovaries, evaluation of the cortex and medulla of C–T 305 the counts of the three fields. MVD was expressed as mean (SD) (vessel/ intact ovaries and cortical strips revealed the absence of these 2 mm ). Any immunostained endothelial cells or endothelial cell clus- histological features (Figure 2). Moreover, the blood vessels 335 ters that were clearly separated from the adjacent microvessels were were not only viable and patent but also had intact muscle coat considered as a single and countable microvessel. Vessel lumens were and endothelium (Figure 2). not a prerequisite for a structure to be defined as a microvessel, and red In every sample, the primordial follicles were counted in 10 310 blood cells were not used to define a vessel lumen (Weidner et al., 1991). different fields and mean values were reported for the cryop- reservation and control groups. Histological evaluation of the 340 Statistical analysis count of the primordial follicles revealed a trend towards higher Analysis of variance (ANOVA) and Student’s t-tests with a statistical values in the intact C–T ovary as compared with the cortical significance of P < 0.05 were performed. Calculations were done with the Statistical Package for the Social Sciences for Windows, strips (4.5/high-power field versus 3/high-power field, respec- 315 version 10.0. tively) (Figure 2). However, differences among these values did not reach the level of statistical significance, probably due to 345 small sample size (two patients). Histological features of apop- Results tosis were seen in the granulosa cells of atretic follicles either Follicular viability in the C–T ovarian tissue within the central region of the membrana granulosa layer or In the 46-year-old patient, the primordial follicular viability loosely attached to the membrana granulosa near its antral sur- was 75 and 81% in C–T intact ovary and C–T ovarian corti- face. A summary of these data is presented in Table I. 350 320 cal strips, respectively. In the 44-year-old patient, these val- ues were 78 and 83%, respectively. The average value for Bcl-2 and p53 protein expressions in the control and C–T both intact ovaries (76.5%) was comparable with that for ovarian tissues cortical strips (82%). A summary of these data is presented The positive and negative controls were positive and negative, in Table I. respectively, indicating the validity of our results. In both

Table I. Morphological analysis of the non-cryopreserved (control), cryopreserved intact ovary and cortical ovarian tissue strips

Aspects Cryopreserved ovarian tissue Non-cryopreserved ovarian tissue P-value

Ovarian cortical Intact ovary Cortical strips and Control group strips intact ovaries

Follicular viability 82/100 (82) 76/100 (76.5) 79/100 (79) NA NA [unstained cells/total cells (%)] Healthy follicles Present Present Present Present NS Atretic follicles Present Present Present Present NS Apoptotic index 1.7 ± 0.1 1.8 ± 0.1 1.7 ± 0.1 1.8 ± 0.1 NS Apoptotic changes Healthy follicles Absent Absent Absent Absent NA Atretic follicles Present Present Present Present NS Bcl-2 protein expression Healthy follicles (mean ± SEM) 1.75 ± 0.3 1.70 ± 0.3 1.70 ± 0.3 1.76 ± 0.3 NS Atretic follicles (mean ± SEM) 0.06 ± 0.7 0.06 ± 0.7 0.06 ± 0.7 0.06 ± 0.7 NS p53 protein expression Healthy follicles Absent Absent Absent Absent NA Atretic follicles Absent Absent Absent Occasional NA Status of the blood vessels Persistent vascularity Persistent vascularity Persistent vascularity Persistent vascularity Microvessel density (mean ± SEM) 5.3 ± 0.8 5.3 ± 0.8 5.3 ± 0.8 5.6 ± 0.9 NS Periodic acid Schiff stain Persistent vascularity Persistent vascularity Persistent vascularity Persistent vascularity NA Masson’s trichrome stain Persistent vascularity Persistent vascularity Persistent vascularity Persistent vascularity NA

The healthy follicles include primordial, primary and secondary follicles. Plasma membrane function and structural integrity were determined by the Trypan Blue exclusion test (viability assay), in which both the number of stained follicles and the total numbers of follicles (100 follicles) were counted. The percentage of viable cells was determined by calculating the percentage of unstained cells. Follicular health and atresia were determined using the histological methods and were reported in descriptive terms. Apoptotic index was evaluated using combined terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-digoxigenin nick-end labelling and histological methods where 10 different areas of each follicular wall were examined. The apoptotic index was calculated as mean number of positively stained cells. For the evaluation of Bcl-2 and p53 immunoreactivities, the numbers of follicles with positive and negative reactivity were counted in the ovarian tissues. NA, not applicable; NS, not significant (P > 0.05). 4 del227.fm Page 5 Tuesday, May 30, 2006 12:23 PM ARTICLE IN PRESS Cryopreservation of an entire human ovary

Figure 1. Histological features of the control (non-cryopreserved) ovarian tissue. The figures show primordial follicle (A, ×200), primary (B and C, ×400), secondary (D, ×400), Graafian (E, ×200) and atretic follicles (F, ×400).

Figure 2. Upper panel: Histological features of both cryopreserved–thawed intact ovary and ovarian cortical strips. The figures show primordial follicle (A, ×400) and intact intraovarian blood vessels (B, ×400). The primordial follicle is formed of primary oocyte, surrounded by a single layer of flattened mitotically inactive, granulosa cells resting on a thin basal lamina. Lower panel: Histological features of the human ovarian fol- licles in both cryopreserved intact ovary and ovarian cortical strips. The follicles include (A) primary, (B) Graafian and (C) atretic follicles. A (×200), the primary follicle is formed of enlarged oocyte, surrounded by a single layer of cuboidal to columnar mitotically inactive granulosa cells. B (×400), the Graafian follicle is formed of an antrum, several layers of granulosa cells and theca layers. C (×200), the atretic follicle is formed of thin inner layer of small, exfoliating granulosa cells and an outer theca layer.

355 control and cryopreserved ovarian tissues, rare weak Bcl-2 theca cells were never immunoreactive regardless of the devel- expression was seen both in the healthy follicles and in the opmental stage of the follicles (Figures 3 and 4). 365 stromal cells (Figures 3 and 4). p53 expression (nuclear) was virtually absent in the healthy primordial, primary and secondary follicles, germinal epithelium and stromal cells subjected to DNA fragmentation assay (TUNEL) in the control and C–T 360 warm ischaemia (Figures 3 and 4). Bcl-2 immunoreactivity ovarian tissues was observed more frequently in the secondary follicles, fol- None of the negative controls stained with TUNEL technique lowed by primary and primordial follicles. In these follicles, showed any immunoreactivity, whereas signal-positive TUNEL Bcl-2 immunoreactivity was restricted to granulosa cells. The staining was observed in the positive controls (Figure 5). In the 370 5 del227.fm Page 6 Tuesday, May 30, 2006 12:23 PM ARTICLE IN PRESS M.A.Bedaiwy et al.

Figure 3. Bcl-2 and p53 protein expressions in the control (non-cryopreserved) ovarian tissue. Bcl-2 protein expression in the primary (A, ×400), secondary (B, ×400) and atretic (C, ×400) follicles and the ovarian stroma (C). The expression appears as golden yellow cytoplasmic staining. Note the lack of p53 protein expression in the primary (D, ×400), Graafian (E, ×200) and atretic (F, ×400) follicles and the ovarian stroma (F) (arrowheads).

control ovarian tissues, granulosa cells with the histologic fea- cortical ovarian tissues, respectively. A summary of these tures of apoptosis were observed in the atretic follicles, either results is presented in Table I and Figures 6–8. within the central region of membrana granulosa layer or loosely attached with membrana granulosa near its antral sur- Discussion 405 375 face or in the antral . The healthy (primordial, primary and secondary) follicles were TUNEL negative. In this investigation, the immediate post-thawing injury to the TUNEL-positive signals were detected in the atretic follicles. human ovary (cryopreserved both as a whole with its vascular This was evident as punctuated brown staining of the frag- pedicle and as ovarian cortical strips) was assessed using both mented nuclei of granulosa cells. The positively stained nuclei histological and immunohistological methods. To achieve our 380 were observed either in the central layers of the membrana goals, we established a study group formed of four patients. It 410 granulosa, at the antral surface or floating in the follicular included both control (non-cryopreserved) and cryopreserved atrium. DNA fragmentation was absent both in the interstitial human ovaries. Tissue sections from lymph nodes, squamous cells and in the theca cells of the atretic follicles. Evaluation of cell carcinoma, leukaemia cell lines, normal skin and liver were apoptosis in the cryopreserved ovaries revealed similar find- used as tissue-specific positive controls for the histochemical 385 ings: (i) the healthy (primordial, primary and secondary) folli- (PAS and Masson’s trichrome staining) and immunohistochem- 415 cles were TUNEL negative; (ii) TUNEL-positive signals were ical (Bcl-2, p53 and apoptotic cells) parameters. Our investiga- detected in the atretic follicles as indicated by the punctuated tion demonstrated the following findings: (i) the mean values of brown staining of the fragmented nuclei of granulosa cells MVD, follicle counts and Bcl-2 protein expression were com- (Figure 5) and (iii) the mean values of apoptotic cells in the parable between the cryopreserved and the control groups and 390 intact cryopreserved ovarian tissues were comparable with (ii) no differences were seen between the two groups for the fol- 420 those in the ovarian cortical strips and control ovarian tissues lowing parameters: healthy and atretic follicle counts, apoptotic (1.8 ± 0.1 versus 1.7 ± 0.1 versus 1.8 ± 0.1, respectively). The index and Bcl-2 protein expression in the atretic follicles results are summarized in Table I. Ovarian tissue banking is more popular as a fertility preser- vation option for patients without a partner or those who cannot delay gonadotoxic chemotherapy for the sake of ovarian stimula- 425 Evaluation of CD34 protein expression in the control tion to produce oocytes for freezing or embryos for cryopreserva- 395 and C–T ovarian tissues tion. With the limited long-term survival upon reimplantation with Endothelial immunostaining with anti-CD34 as well as evaluation ovarian tissue strips, ovarian transplantation with microvascular of the basement membrane (PAS) and muscle coat (Masson’s anastomosis could help avoid accelerated follicular loss and trichrome) of ovarian blood vessels demonstrates persistent improve the longevity of the ovarian grafts. This helps to guar- 430 vascularity in both cryopreserved and control ovarian tissues. antee an immediate revascularization and establishment of the 400 Microvessels are heterogeneously distributed within the ovar- graft function. Although successful in rat and sheep animal ian tissue. The mean values of ovarian MVD were 5.6 ± 0.9, models (Wang et al., 2002; Bedaiwy et al., 2003), microvascular 5.3 ± 0.8 and 5.3 ± 0.8 for control, cryopreserved intact and anastomosis of intact frozen–thawed ovaries remains technically 6 del227.fm Page 7 Tuesday, May 30, 2006 12:23 PM ARTICLE IN PRESS Cryopreservation of an entire human ovary

Figure 4. Upper panel: Bcl-2 and p53 protein expressions in both cryopreserved intact ovary and ovarian cortical strips. Bcl-2 protein expression in the primordial (A, ×200), atretic (B, ×400) follicles and the ovarian stroma (C, ×200). The expression appears as golden yellow cytoplasmic staining. Note the lack of p53 protein expression in the primordial (D, ×400), atretic follicles (E, ×400) and the ovarian stroma/Graafian follicle (F, ×400) (arrowheads). Lower panel: Bcl-2 and p53 protein expressions in positive controls. Bcl-2 protein expression in the lymph node with reactive hyperplasia (G, ×200) and p53 protein expression in the cells of squamous cell carcinoma (H, ×400) (arrowheads).

435 challenging. Moreover, cryopreservation of an entire organ protocol of an entire ovary with its vascular pedicle. Adequate with its vascular pedicle is much more complicated than ovar- delivery of the cryoprotective agent (CPA) to almost every cell ian cortical strips or suspended cell (oocyte) or group of cells of the ovary is mandatory to guarantee an adequate post-thaw (embryo). survival. Aiming at adequate CPA permeation to virtually 450 In our previous study, we reported a 30% long-term patency every cellular component of the ovary, we used the Horizon 440 rate of the ovarian vessels upon reimplantation of an intact Modular Infusion pump to ensure adequate CPA distribution frozen–thawed ovary in merino sheep (Bedaiwy et al., 2003). throughout the ovarian substance. Infusing the CPA slowly However, human ovaries are larger and may not be amenable through the ovarian artery and allowing it to come out of the to the same protocol. The technical challenges of the microvas- ovarian vein ensures adequate use of the ovarian vascular 455 cular anastomosis could be circumvented adequately by appro- channel as a delivery vehicle. To avoid intracellular ice forma- 445 priate training. On the contrary, the main determinant of tion, we used a slow-freezing protocol using the Planer freezer. success would be to develop an appropriate cryopreservation The use of constant infusion pressure during the freezing and 7 del227.fm Page 8 Tuesday, May 30, 2006 12:23 PM ARTICLE IN PRESS M.A.Bedaiwy et al.

Figure 5. Apoptotic changes in the HL60 promyelocytic leukaemia cells, cryopreserved intact ovary and ovarian cortical strips. A (×200) Apop- tosis in the HL60 promyelocytic leukaemia cells (control), with positive staining (inset, arrowhead). B (×400) Apoptosis in atretic ovarian folli- cles, terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-digoxigenin nick-end labelling-positive cells (arrowheads).

thawing phases ensured adequate distribution and clearance of are open for future investigations. Apoptosis in the atretic folli- 460 the CPA. cles can effectively be present before the cryopreservation, but The selection of DMSO as a CPA of choice was based on it does not seem to be increased by the cryopreservation pro- 500 the initial observation by Newton et al. (1998) that exposure cess. Moreover, the presence of apoptotic changes in the atretic for 30 min to 1.5 M solutions of DMSO at 4°C achieved a follicles supports the link between apoptosis and ovarian follic- mean tissue concentration that was almost 80%. In our study, ular atresia. 465 1.5 M DMSO was tolerated comparably by both an intact It is also possible that substances which stimulate apoptosis ovary and ovarian cortical strips. Interestingly, in a recent may have atretogenic effects on the ovarian follicles. The pres- 505 report where ovarian tissue was exposed to 10% DMSO for 15 ence of Bcl-2 protein expression in the healthy ovarian follicles min at 4°C, no significant differences in follicular viability concurs with other reports (Tilly, 1996a,b; Felici et al., 1999) before (99.4%) or after cryoprotectant exposure (98.1%) were and suggests its possible physiologic role in the maintenance of 470 documented. This could probably mean that ovarian tissue follicular integrity. In support, mice lacking functional Bcl-2 could tolerate adequately varying concentrations of DMSO protein possess reduced numbers of primordial follicles rela- 510 with comparable post-thaw survival (Martinez-Madrid et al., tive to their wild-type sisters (Ratts et al., 1995). The lack of 2004). p53 protein expression in the healthy follicles versus its pres- After thawing, viability assessment of the primordial folli- ence of atretic ones concurs with previous studies and suggests 475 cles was comparable between intact ovaries and ovarian corti- lack of underlying DNA damage and therefore implies integ- cal strips. This demonstrates that delivering the CPA using an rity of the genome in the healthy follicles (Hussein, 2005). 515 infusion pump was as efficient as the permeation of the ovarian During organ transplantation, tissues are subjected to varia- cortical strips with the CPA avoiding cryoinjury to the follicles ble duration of ischaemic injury. In the heart, retina, liver and as well as the intraovarian vessels. The viable primordial fol- brain, ischaemic injury is associated with apoptosis as well as 480 licular count in the intact ovaries in both of our patients (75 and Bcl-2 and p53 alterations. Namely, the induction of Bcl-2 and 78%) was comparable with the 75.1% follicular viability p53 can occur as early as 0.5 and 1 h, respectively, following 520 reported by Martinez-Madrid et al. (2004), who tried a differ- ischaemia (Peralta et al., 2002; Wang et al., 2003). Apoptosis ent freezing protocol for intact human ovary. requires active cell metabolism that could not be complete In this investigation, the presence of apoptotic activity in the within 30 min and could therefore not alter the status of the fol- 485 ovarian follicles may be attributed to the presence of apoptosis- licles within an ovary. A hypothesis to be tested is that the dif- triggering signals and/or degradation of apoptosis-related pro- ferences observed in this investigation are the result of outright 525 teins (caspase-3, caspase-8, Bcl-2 and actin) (Zhu et al., 1999; lysis (ischaemia could make cells more prone to lysis). It Schmidt-Mende et al., 2000). At the molecular level, cryop- would be interesting for future investigations to examine this reservation and thawing had no effects on the values of apop- hypothesis. 490 totic cell death and Bcl-2 and p53 protein expressions in both Our investigation revealed the presence of persistent vascu- ovarian cortical strips and intact ovaries. A hypothesis to be larity and comparable MVD in both control and cryopreserved 530 tested is that cryopreservation and thawing can induce some groups. These findings indicate that our freeze–thawing proto- effects on these values (apoptotic cell death and Bcl-2 and p53 col has no detrimental effects on the integrity of the vascular protein expressions) over longer time intervals. This hypothesis network of the ovarian tissue. Of note, the status of the ovarian 495 stems from the fact that apoptosis is a dynamic process that vascular supply is central to dynamic changes occurring during requires active metabolism and usually takes several hours or the normal ovarian cycle. Follicular growth and the develop- 535 even days to be executed (Hussein et al., 2006). These questions ment of the are dependent on intact vasculature.

8 del227.fm Page 9 Tuesday, May 30, 2006 12:23 PM ARTICLE IN PRESS Cryopreservation of an entire human ovary

Figure 6. Upper panel: Evaluation of the status of the blood vessels in the control (non-cryopreserved) ovarian tissue. CD34 protein expression in the ovarian stroma. The ovarian blood vessels show persistent vascularity and moderate density of microvessels. The endothelial cells of the microvessels are stained brown by anti-CD34 antibodies. The inset represents CD34 staining in the control (haemangioma) (arrowheads) (×400). Lower panel: Evaluation of the status of the blood vessels in the control (non-cryopreserved) ovarian tissue. Periodic acid Schiff (PAS) (A–C) and Masson’s trichrome (D–F) stain within the ovarian stroma. The ovarian blood vessels show persistent vascularity and moderate density of microvessels. The muscle coat (Masson’s trichrome) and basement membrane (PAS stain) outline the blood vessels (arrowheads) (A, C and D: ×200; B, E and F: ×400).

The selection of a dominant follicle in monovular species is cholesterol to luteal cells for progesterone biosynthesis (Carr associated with angiogenesis. Moreover, selected follicles et al., 1981; Goede et al., 1998). possess a more elaborate microvascular network than other fol- Many studies indicate that the ovaries of older women are 540 licles. The vasculature also plays a key role in the delivery of smaller than those of young women. They are composed 9 del227.fm Page 10 Tuesday, May 30, 2006 12:23 PM ARTICLE IN PRESS M.A.Bedaiwy et al.

Figure 7. Upper panel: Evaluation of the status of the blood vessels both in cryopreserved intact ovary and in ovarian cortical strips. CD34 protein expression within the ovarian stroma. The ovarian blood vessels show persistent vascularity and moderate density of microvessels. The endothe- lial cells of the microvessels are stained brown by anti-CD34 antibodies (arrowheads) (×400). Lower panel: Evaluation of the status of the blood vessels both in cryopreserved intact ovary and in ovarian cortical strips. Periodic acid Schiff (PAS) and Masson’s trichrome stain within the ovar- ian stroma. The ovarian blood vessels show persistent vascularity and moderate density of microvessels. The muscle coat (Masson’s trichrome) and basement membrane (PAS stain) outline the blood vessels (arrowheads). (All pictures are on scale of ×200 except the lower right one, ×400.) 10 del227.fm Page 11 Tuesday, May 30, 2006 12:23 PM ARTICLE IN PRESS Cryopreservation of an entire human ovary

Bedaiwy MA and Falcone T (2004) Ovarian tissue banking for cancer patients: reduction of post-transplantation ischaemic injury: intact ovary freezing and transplantation. Hum Reprod 19,1242–1244. Bedaiwy MA, Jeremias E, Gurunluoglu R, Hussein MR, Siemianow M, Biscotti C and Falcone T (2003) Restoration of ovarian function after autotransplantation of intact frozen–thawed sheep ovaries with microvascu- lar anastomosis. Fertil Steril 79,594–602. Carr BR, Sadler RK, Rochelle DB, Stalmach MA, MacDonald PC and Simpson ER (1981) Plasma lipoprotein regulation of progesterone biosynthesis by human corpus luteum tissue in organ culture. J Clin Endocrinol Metab 52,875–881. Couzinet B, Meduri G, Lecce MG, Young J, Brailly S, Loosfelt H, Milgrom E and Schaison G (2001) The postmenopausal ovary is not a major androgen- producing gland. J Clin Endocrinol Metab 86,5060–5066. Demirci B, Salle B, Frappart L, Franck M, Guerin JF and Lornage J (2002) Morphological alterations and DNA fragmentation in oocytes from primor- dial and primary follicles after freezing–thawing of in sheep. Fertil Steril 77,595–600. Donnez J, Dolmans MM, Demylle D, Jadoul P, Pirard C, Squifflet J, Martinez- Madrid B and van Langendonckt A (2004) Livebirth after orthotopic trans- Figure 8. Morphological analysis of the control and cryopreserved plantation of cryopreserved ovarian tissue. Lancet 364,1405–1410. intact ovary and cortical ovarian tissue strips. As compared with the Falcone T, Attaran M, Bedaiwy MA and Goldberg JM (2004) Ovarian func- control group, the mean values of microvessel density, follicle counts tion preservation in the cancer patient. Fertil Steril 81,243–257. and Bcl-2 protein expression tended to be higher in the cryopreserved Felici MD, Carlo AD, Pesce M, Iona S, Farrace MG and Piacentini M (1999) group; however, the differences were not significant (P > 0.05). No Bcl-2 and Bax regulation of apoptosis in germ cells during prenatal oogene- differences were seen between the two groups for the following sis in the mouse embryo. Cell Death Differ 6,908–915. parameters: healthy and atretic follicle counts, apoptotic index and Gavrieli Y, Sherman Y and Ben-Sasson SA (1992) Identification of Bcl-2 protein expression in the atretic follicles. programmed cell death in situ via specific labeling of nuclear DNA frag- mentation. J Cell Biol 119,493–501. Goede V, Schmidt T, Kimmina S, Kozian D and Augustin HG (1998) Analysis 545 mainly of stromal cells that have receptors for, and respond to, of blood vessel maturation processes during cyclic ovarian angiogenesis. gonadotrophins and secrete testosterone and lesser amounts of Lab Invest 78,1385–1394. Gook DA, Edgar DH, Borg J, Archer J, Lutjen PJ and McBain JC (2003) other androgens and estrogens. These ovaries remain gonado- Oocyte maturation, follicle rupture and luteinization in human cryopre- trophin-driven, androgen-producing glands. Also, they are served ovarian tissue following xenografting. Hum Reprod 18,1772–1781. characterized by the presence of prominent cystic follicles, Gosden RG, Baird DT, Wade JC and Webb R (1994) Restoration of fertility to oophorectomized sheep by ovarian autografts stored at –196 degrees C. 550 thick-walled vessels, granulomas, hyaline scars and low Hum Reprod 9,597–603. number of follicles. However, these follicles are not necessar- Hussein MR (2005) Apoptosis in the ovary: molecular mechanisms. Hum ily more sensitive to the freezing–thawing process (Couzinet Reprod Update 11,162–177. et al., 2001; Longcope, 2001). In this investigation, the study Hussein M, Bedaiwy M and Falcone T (2006) Analysis of apoptotic cell death, Bcl-2, and p53 protein expression in freshly fixed and cryopreserved ovarian group was formed of older patients knowing in advance the tissue after exposure to warm ischemia. Fertil Steril, in press. 555 histopathological and follicular implications of ageing on the Jeremias E, Bedaiwy MA, Gurunluoglu R, Biscotti CV, Siemionow M and ovarian structure. Whether the aforementioned age-related Falcone T (2002) Heterotopic autotransplantation of the ovary with micro- vascular anastomosis: a novel surgical technique. Fertil Steril 77,1278–1282. changes have an impact on our results awaits further confirma- Kerr JF, Wyllie AH and Currie AR (1972) Apoptosis: a basic biological phe- tion in younger age group upon ethical approval. nomenon with wide-ranging implications in tissue kinetics. Br J Cancer In this study, we assessed injury of the ovarian tissue in the 26,239–257. Liu TJ, el-Naggar AK, McDonnell TJ, Steck KD, Wang M, Taylor DL and 560 period immediately after thawing. Our investigation provides Clayman GL (1995) Apoptosis induction mediated by wild-type p53 adeno- further evidence that intact human ovary could be cryopre- viral gene transfer in squamous cell carcinoma of the head and neck. Cancer served using a slow-freezing protocol. We have demonstrated Res 55,3117–3122. comparable survival rates of follicles and limited molecular Longcope C (2001) Endocrine function of the postmenopausal ovary. J Soc Gynecol Invest 8,S67–S68. alterations between C–T intact ovaries and C–T ovarian corti- Martinez-Madrid B, Dolmans MM, Van Langendonckt A, Defrere S and 565 cal strips. The information herein reported is provided for the Donnez J (2004) Freeze–thawing intact human ovary with its vascular pedi- period immediately after thawing. The changes associated with cle with a passive cooling device. Fertil Steril 82,1390–1394. long-term injury mandates further investigations. Meirow D, Levron J, Eldar-Geva T, Hardan I, Fridman E, Zalel Y, Schiff E and Dor J (2005) Pregnancy after transplantation of cryopreserved ovarian tissue in a patient with ovarian failure after chemotherapy. N Engl J Med 353,318–321. Acknowledgements Newton H, Fisher J, Arnold JR, Pegg DE, Faddy MJ and Gosden RG (1998) The financial support for this study was a research grant from the Min- Permeation of human ovarian tissue with cryoprotective agents in prepara- tion for cryopreservation. Hum Reprod 13,376–380. 570 imally Invasive Surgery Center, The Cleveland Clinic Foundation, Oktay K, Buyuk E, Veeck L, Zaninovic N, Xu K, Takeuchi T, Opsahl M and Cleveland, OH, USA. This work was presented at ESHRE 20th Rosenwaks Z (2004) Embryo development after heterotopic transplantation Annual meeting, June 27–July 1, Berlin, Germany 2004. of cryopreserved ovarian tissue. Lancet 363,837–840. Peralta C, Bulbena O, Xaus C, Prats N, Cutrin JC, Poli G, Gelpi E and Rosello- References Catafau J (2002) Ischemic preconditioning: a defense mechanism against the reactive oxygen species generated after hepatic ischemia reperfusion. Baird DT, Webb R, Campbell BK, Harkness LM and Gosden RG (1999) Transplantation 73,1203–1211. Long-term ovarian function in sheep after ovariectomy and transplantation Ratts VS, Flaws JA, Kolp R, Sorenson CM and Tilly JL (1995) Ablation of of autografts stored at –196°C. Endocrinology 140,462–471. Bcl-2 gene expression decreases the numbers of oocytes and primordial

11 del227.fm Page 12 Tuesday, May 30, 2006 12:23 PM ARTICLE IN PRESS M.A.Bedaiwy et al.

follicles established in the post-natal female mouse gonad. Endocrinology Weidner N, Semple JP, Welch WR and Folkman J (1991) Tumor angiogenesis and 136,3665–3668. metastasis – correlation in invasive breast carcinoma. N Engl J Med 324,1–8. Schmidt-Mende J, Hellstrom-Lindberg E, Joseph B and Zhivotovsky B (2000) Wood TC, Montali RJ and Wildt DE (1997) Follicle-oocyte atresia and temporal Freezing induces artificial cleavage of apoptosis-related proteins in human taphonomy in cold-stored domestic cat ovaries. Mol Reprod Dev 46,190–200. bone marrow cells. J Immunol Methods 245,91–94. Yin H, Wang X, Kim SS, Chen H, Tan SL and Gosden RG (2003) Transplan- Tilly JL (1996a) Apoptosis and ovarian function. Rev Reprod 1,162–172. tation of intact rat gonads using vascular anastomosis: effects of cryopreser- Tilly JL (1996b) The molecular basis of ovarian cell death during germ cell vation, ischaemia and genotype. Hum Reprod 18,1165–1172. attrition, follicular atresia, and luteolysis. Front Biosci 1,d1–d11. Zhu L, Yu X, Akatsuka Y, Cooper JA and Anasetti C (1999) Role of mitogen- Wang X, Chen H, Yin H, Kim SS, Lin Tan S and Gosden RG (2002) Fertility activated protein kinases in activation-induced apoptosis of T cells. Immu- after intact ovary transplantation. Nature 415,385. nology 97,26–35. Wang F, Hu YL and Li YH (2003) The expression of Bcl-2 protein on acute myocardial ischemia/reperfusion and the significance in forensic pathology. Submitted on December 4, 2005; resubmitted on March 15, 2006, May 11, Fa Yi Xue Za Zhi 19,207–208, 211. 2006; accepted on May 16, 2006

12