- Email: [email protected]
Parameters affecting successful transplantation of frozen-thawed human fetal ovaries into immunodeficient mice
FERTILITY AND STERILITY威 VOL. 80, NO. 2, AUGUST 2003 Copyright ©2003 American Society for Reproductive Medicine Published by Elsevier Inc. Printed on acid-free paper in U.S.A.
Parameters affecting successful transplantation of frozen-thawed human fetal ovaries into immunodeficient mice Received July 30, 2002; revised and accepted November 4, 2002. Supported in part by research grants from Rabin Medical Center: a grant from the Steering Committee for Research Promotion (Dr. Orvieto) and a basic sciences research grant (Dr. Raanani). Presented in part at the International Workshop on Early Folliculogenesis and Oocyte Development: Basic and Clinical Aspects, London, United Kingdom, June 10 and 11, 1999; the Annual Meeting of the European Society for Human Reproduction and Embryology, Bologna, Italy, June 24 –27, 2000; and the Twelfth International Workshop on the Development and Function of the Reproductive Organs—Advances in Reproduction Research: From Follicle Culture to Nuclear Transfer and Embryonic Stem Cells, Ma’ale Hachamisha, Israel, April 30 –May 3, 2001. Reprint requests: Ronit Abir, Ph.D., IVF Research Laboratory, Infertility and IVF Unit, Department of Obstetrics and Gynecology, Rabin Medical Center, Beilinson Campus, Petah Tikva 49100, Israel (FAX: 972-3-9240533; Email: [email protected]). a Infertility and IVF Unit, Department of Obstetrics and Gynecology, Rabin Medical Center, Beilinson Campus. b Department of Obstetrics and Gynecology, Rabin Medical Center, Beilinson Campus. 0015-0282/03/$30.00 doi:10.1016/S0015-0282(03) 00658-7
Ronit Abir, Ph.D.,a,b Raoul Orvieto, M.D., M.Sc.,a,b Hila Raanani, M.D.,a,b Dov Feldberg, M.D.,a Shmuel Nitke, M.D.,a and Benjamin Fisch, M.D., Ph.D.a,b Infertility and IVF Unit, Department of Obstetrics and Gynecology, Rabin Medical Center, Beilinson Campus, Petah Tikva, and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
Objective: To compare the development and survival of human fetal follicles frozen-thawed with dimethylsulfoxide (DMSO) and propandiol (PROH) in immunodeficient mice, to study the effects of host treatment with FSH, and to compare kidney and subcutaneous transplantation. Design: Controlled histologic study. Setting: Major tertiary care and referral academic center. Patient(s): Twenty-one women undergoing second-trimester pregnancy termination. Main Outcome Measure(s): Microscopic morphometric analysis and immunocytochemistry for proliferating-cell nuclear antigen in human fetal ovaries grafted into immunodeficient mice. Result(s): Renal grafts that were frozen-thawed with DMSO rather than PROH survived better in the hosts (79.6% compared with 58.8%), but significantly more follicles were identified in grafts frozen-thawed with PROH (P⬍.001). Follicular development was observed only in FSH-treated hosts, and follicular survival and development was better in the kidney than the subcutaneous site. Conclusion(s): This is the first report showing development of human fetal follicles in immunodeficient mice. Freezing-thawing with PROH seems to support development and survival better than with DMSO. The kidney is a better transplantation site than the subcutaneous site, probably because of its superior vascularization. Administration of FSH to the host is essential for follicular development. Follicular development and growth was better in ovarian grafts from older fetuses, as they contained more formed follicles. (Fertil Steril威 2003; 80:421– 8. ©2003 by American Society for Reproductive Medicine.) Key Words: Human fetal ovaries, primordial follicles, survival, proliferating-cell nuclear antigen, DMSO, PROH
The development of germ cells in the human fetal ovary has three phases: mitotic divisions, meiotic division, and follicular formation (1). Female germ cells, or oogonia, undergo many mitotic divisions until very close to term, with a peak during mid-pregnancy, followed by a drastic drop during the third trimester. Meiotic division commences gradually in the third month of pregnancy. The diplotene stage is achieved within weeks of initiation of meiosis, when the oogonia become oocytes. Oocytes are larger than oogonia and contain more intracellular organelles. They have completed genetic recombination and do not undergo additional nuclear maturation until after puberty. Follicular formation begins during the fourth month of pregnancy, after rapid prolif-
eration of the ovarian somatic cells to form a single flat layer of pregranulosa cells surrounding the oocytes (primordial follicles). At birth, the human ovary contains from 266,000 to 472,000 follicles, most of which remain nongrowing primordial follicles (1). The shortage of donated oocytes for IVF has prompted suggestions of the potential use of oocytes from frozen-stored ovaries of aborted fetuses. It is, however, unknown whether such oocytes have the same developmental capacity as those from adults. Furthermore, while FSH apparently induces follicular growth from the secondary stages, the signals that initiate the growth of earlier follicles are unknown (1), although recent studies suggest that some growth factors may be involved in this process 421
(2). Before frozen-thawed fetal oocytes are used clinically, their growth and maturation ability must be tested. Human ovarian slices transplanted into immunodeficient mice are not rejected, making these mice an excellent model for the study of follicular development (3–17). In most studies, the highly vascularized kidney capsule has served as the transplantation site (3–12, 16, 17). However, subcutanous transplantation is much easier to perform and sometimes permits external identification of developing follicles (12–15). We sought to examine survival and development of frozen-thawed human fetal germ cells in immunodeficient mice, to compare kidney with subcutaneous transplantation, and to determine whether FSH administration to the host benefits grafted fetal follicles.
MATERIALS AND METHODS Human Fetal Ovaries The study was approved by the local ethics committee of Rabin Medical Center. Ovaries were dissected from 23 human fetuses 19 to 22 gestational weeks of age after informed consent was obtained from the mothers. This fetal age was chosen because the number of germ cells peaks at 20 gestational weeks (1) and because our department’s pregnancy termination policy mandates feticide for all fetuses older than 22 gestational weeks, and we preferred to use ovaries from viable fetuses. Most of the abortions were performed because of fetal anatomic malformations or chromosomal abnormalities and were induced by prostaglandins. One slice 1 to 2 mm in length was fixed in Bouin’s solution immediately after ovarian dissection (fresh control). Because all abortions were performed late at night, the study did not include freshly dissected ovaries.
Cryopreservation and Thawing of Ovarian Tissue Ovarian samples were cut into slices 1 to 2 mm (thickness less ⬍ 1 mm) in size before freezing. Six to 12 slices were prepared from both ovaries of the same fetus, because the size of the ovaries varied and correlated with fetal age. Cryopreservation was carried out 1 to 2 hours after the abortion. Eight of the ovarian samples were frozen in a solution consisting of 1.5 M of 1,2 propandiol (PROH) (Sigma, St. Louis, MO) and 0.1 M of sucrose (Sigma) (18). The group included specimens from one fetus at 20 gestational weeks with central nervous system and pericardial malformations; five fetuses at 21 gestational weeks with renal and cutaneous malformations, trisomy 21, or 47,XXX (one each), or no abnormalities (pair of twins); and two fetuses at 22 gestational weeks both normal (one case of maternal syphilis). The other 15 samples were frozen in a solution consisting of 1.5 M of dimethyl sulfoxide (DMSO) (Sigma) and 0.1 M 422
Abir et al.
Human fetal ovaries in immunodeficient mice
TABLE 1 Division of ovarian tissue according to the transplantation sites. Cryoprotectant Propandiol Dimethyl sulfoxide
Total
Renal only
Subcutaneous only
Renal and subcutaneous
8 15
7 9
1 1
5
Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
of sucrose (Sigma) (19). This group included specimens from one fetus at 19 gestational weeks with lung and larynx malformations; seven fetuses at 20 gestational weeks with anencephaly and renal and central nervous system malformations, trisomy 13, or trisomy 21 (one each), or no abnormalities (three fetuses); six fetuses at 21 gestational weeks with diaphragmatic hernia or trisomy 21 (one each) or no abnormalities (four fetuses, including one pair of twins); and one fetus at 22 gestational weeks with anencephaly. Tissue slices 1 to 2 mm were placed in cryogenic vials (Nalge Nunc International, Delta, Roskilde, Denmark) filled with 0.8 mL PROH or DMSO. Before freezing, PROH samples were kept at room temperature for 10 minutes (5, 18) and the DMSO samples were kept on ice for 30 minutes, because DMSO might be toxic at room temperature (19). All samples were frozen slowly in a programmable freezer (Kryo 10, series 10/20; Planer Biomed, Sunbury on Thames, United Kingdom) and immediately placed in liquid nitrogen. The slices were cryopreserved for 1 week to 5 months. They were thawed by washouts with decreasing concentration gradients of PROH (Sigma) and sucrose (Sigma) or DMSO (Sigma) and sucrose (Sigma). One slice was fixed in Bouin’s solution immediately after thawing (thawed control).
Grafting Into Immunodeficient Mice Immunodeficient nu/nu female BalbC mice 6 to 8 weeks of age (Harlan, Jerusalem, Israel) were used for tissue grafting. All frozen-thawed tissue, apart from the thawed control, was used for grafting. The mice were anesthesized with an intramuscular injection of 0.04 mL of a 1:1 solution of 0.5% xylazine base (prepared from 2% XYL-M; VMD, Arendonk, Belgium) diluted with sterile distilled water and 50 mg/mL of ketamine hydrochloride (Ketalar; Parke Davis, Hampshire, United Kingdom). For the kidney grafts, the mice underwent bilateral oophorectomy, and the kidney capsule was opened and the slices were pushed inside (5). For the subcutaneous grafts, a small dorso median transverse incision was made, and the ovarian tissue slices were placed into the subcutaneous space above the flank without suturing (13). Table 1 shows the distribution of the ovaries according to the grafting sites. Vol. 80, No. 2, August 2003
TABLE 2 Survival of ovarian grafts. PROH Site Kidney Subcutaneous a
DMSO
No. of ovaries
No. of mice
Survival rate (%)
No. of ovaries
No. of mice
Survival rate (%)
7 1
34 2
20 (58.8) 1 (50)
14 6
49 13
39 (79.6) 12 (92.3)a
Significantly higher compared with propandiol kidney grafts (P⬍.04).
Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
Two thirds of the mice were stimulated with 1.0 U of recombinant FSH (Gonal-F; Laboratoires Serono S.A., Aubonne, Switzerland) (5, 7, 10), every second day starting 1 month after grafting (5). The mice were sacrificed periodically within 7 months, and the grafts were removed and fixed in Bouin’s solution.
Histologic Preparation and Immunocytochemical Analysis Control samples and transplanted ovarian tissue specimens removed from the hosts were prepared for paraffin embedding, sectioning (5 m), and staining with hematoxylin– eosin. The number of follicles in control and transplant samples were counted. Unstained sections were placed on OptiPlus positivecharged microscope slides (BioGenex, San Ramon, CA) and prepared for immunocytochemical study of proliferating-cell nuclear antigen (5, 10, 11, 13). The sections were deparaffinized, rehydrated, and quenched in 3% H2O2 (prepared from Perhydrol 30% H2O2; Merck, Darmstadt, Germany) in methanol (Bio Lab) for 30 minutes to block endogenous peroxidase activity. They were then microwaved for 15 minutes at full power (800 W) and for a further 15 minutes at 75% power with citrate buffer (pH, 6.0) that was diluted in distilled water from a ⫻20 citrate buffer solution (Zymed Laboratories, San Francisco, CA) to enhance antigen retrieval. After rinsing, 10% normal rabbit serum (Dako, Glostrup, Denmark) was added for 30 minutes at room temperature to inhibit nonspecific binding. The sections were incubated with a mouse anti–proliferating-cell nuclear antigen (Dako) at a dilution of 1:50 overnight at room temperature. The negative controls were incubated with a normal mouse IgGa2 antibody (Santa Cruz Biotechnology, Santa Cruz, CA), diluted to the same concentration as the primary antibody. Commercially available labeled slides were used as positive controls for proliferating-cell nuclear antigen (PCNA, human tonsil, Dako). The following morning, the samples were rinsed and incubated with a rabbit antimouse biotinylated immunoglubulin (Dako) at a dilution of 1:200 for 30 minutes. Finally, an avidin– biotin complex (StreptABC complex/HRP; FERTILITY & STERILITY威
Dako) diluted in hydrochloride Tris buffer at pH 7.6 (Sigma), was added for 30 minutes at room temperature. The sections were then exposed to a diaminobenzidine urea H2O2 solution in distilled water (Sigma Fast tablets; Sigma) for 5 minutes (brown dye for proliferating-cell nuclear antigen expression). The sections were counterstained with hematoxylin (blue-purple dye). Unless otherwise stated, all dilutions were performed with phosphate-buffered saline at pH 7.6 (Biological Industries, Beit Ha’emek, Israel). Phosphate-buffered saline also served as the main rinsing solution. At least two slides were stained for every graft.
Statistical Analysis Data were statistically analyzed by using the unpaired Student t-test, 2 test, or Fisher test, as indicated. Values of at least P⬍.05 were considered statistically significant.
RESULTS Figures 1 through 5 show follicles during various stages of the experiment. Because of a high rate of host morbidity during the first postoperative month (usually shortly after the operation), only hosts that survived for at least 1 month were included (Tables 2, 3, and 4). Table 2 shows survival rates of ovarian grafts after at least 1 month in the hosts. Significantly more subcutaneous DMSO frozen-thawed grafts than PROH frozen-thawed renal transplants were recovered (P⬍.04). Of the 13 recovered subcutaneous grafts, 11 were identified within 2 months; no follicles survived in them. These included two necrotic pieces from hosts that were not treated with FSH. Two other subcutaneous grafts were identified in FSH treated mice after 4 and 6 months; they contained very early primordial follicles (four and three follicles, respectively) (Fig. 4). Table 3 shows the number of surviving follicles after kidney transplantation in FSH-treated hosts. The initial number of follicles in both freshly dissected and thawed samples was significantly higher than in the DMSO frozen grafts after a postoperative period of 3 to 4 months (P⬍.01). Among the grafts that survived 3 to 4 months, the PROH frozen transplants had significantly more surviving follicles 423
FIGURE 1
FIGURE 3
A propandiol frozen-thawed ovary from a normal fetus 21 gestational weeks of age (control). The arrow indicates primordial follicles. Hematoxylin– eosin, ⫻400.
A human fetal ovary 6.5 months after renal grafting. The ovaries were frozen-thawed in dimethyl sulfoxide and originated from a 21-gestational-week-old fetus with trisomy 21. The arrow indicates secondary follicles. Note the brown proliferating-cell nuclear antigen staining in granulosa cells and oocytes. Magnification, ⫻200.
Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
than did the DMSO frozen-thawed transplants (P⬍.001). In some cases, oogonia was identified in recovered grafts up to 6 months postoperatively.
FIGURE 2 A primary follicle (P) identified 4 months after grafting. The ovaries were frozen-thawed with propandiol and originated from a 47,XXX fetus 21 gestational weeks of age. Note the brown proliferating-cell nuclear antigen staining in most of the granulosa cells and in the nuclei of the oocyte. Magnification, ⫻400.
Abir. Human fetal ovaries in immunodeficient mice.Fertil Steril 2003.
Table 4 shows the effects of FSH host treatment on survival of follicles from fetal ovaries transplanted under the kidney capsule. Because of the small number of surviving grafts and follicles, the results from PROH frozen-thawed samples and DMSO frozen-thawed samples were combined. Significantly more follicles were found in the control samples (both freshly dissected and thawed) than in the grafts recovered from FSH-treated mice after 1 to 2 months (P⬍.008), after 3 to 4 months (P⬍.01), and after 5 to 7 months (P⬍.05). Graft survival rates did not significantly differ between FSH-treated hosts and nontreated hosts. A few primary, secondary, and antral follicles were identified after at least 4 months, but only in FSH-treated hosts: Five primary follicles were identified in renal grafts after 4 months, and secondary (five and eight follicles per graft) and antral follicles (two follicles per graft, up to 0.5 mm in diameter) were identified in two kidney grafts after 6 to 7 months. The ovaries in which multilayered follicles developed were obtained from chromosomally abnormal fetuses at 21 gestational weeks: one from a 47,XXX fetus whose ovaries were frozen-thawed with PROH (4), and the other from a trisomy 21 fetus whose ovaries were frozen-thawed with DMSO (Fig. 3). Of note, follicular development was better in ovarian grafts from the older fetuses, as they contained initially a larger number of formed follicles (Fig. 1).
Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
424
Abir et al.
Human fetal ovaries in immunodeficient mice
Positive proliferating-cell nuclear antigen staining was observed in all ovarian components of the freshly dissected Vol. 80, No. 2, August 2003
FIGURE 4
FIGURE 5
An ovary from a fetus after 6 months of subcutaneous grafting. The ovaries were frozen-thawed in dimethyl sulfoxide and originated from a normal fetus 21 gestational weeks of age. The arrow indicates the initial formation of primordial follicles. Hematoxylin– eosin, ⫻400.
A negative control of the same graft shown in Figure 3. Note the secondary follicle, the overall purple-blue staining, and the lack of brown proliferating-cell nuclear antigen staining. Magnification, ⫻400.
Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003. Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
and thawed control samples. It was also identified in the grafted growing follicles from primary stages onward (Fig. 2 and 3). No negative control stained positive for proliferatingcell nuclear antigen (Fig. 5), whereas positive controls stained positive for proliferating-cell nuclear antigen (data not shown).
DISCUSSION Apart from our previous case report (5), this is the first study showing survival and development of frozen-thawed human fetal follicles in immunodeficient mice. Although transplants frozen-thawed with DMSO had a higher survival rate, more follicles were identified in PROH frozen-thawed grafts 3 to 4 months postoperatively. Despite the high number of identified subcutaneous grafts, very few follicles survived in these transplants, and development beyond the primordial stage was achieved only in the kidney-grafted FSH-treated hosts. However, our results should be interpreted with caution, as only 11 of the 23 ovaries were from normal fetuses, and development to multilayered follicles occurred in ovarian grafts from two chromosomally abnormal fetuses. Fresh (5, 16) and frozen-thawed (5) fetal murine ovaries were transplanted under the kidney capsule (6, 17) or in the bursal cavity of adult mice (6). Follicular development was identified in the grafts (6, 17), and in some cases in the bursal cavity group, oophorectomized hosts conceived from donor FERTILITY & STERILITY威
fetal oocytes (6). When ovaries from fetal hamsters were transplanted under the kidney capsule of adult hamsters, the number of growing follicles increased after host treatment with estradiol or testosterone propionate (16). In a recent study (20), bovine fetal cortical ovarian slices were grafted to the chorioalantoic membrane of cultured chick embryos for up to 10 days. Although the grafts became rapidly infiltrated with blood vessels, no follicular growth was initiated. In a preliminary study, when freshly dissected fragments from human fetuses 16 gestational weeks of age were transplanted under the kidney capsule of immunodeficient mice, germ cells in mitosis and at all stages of meiotic prophase were abundant, and initial follicular formation was observed 6 months after grafting (3). In addition, a pilot report from our laboratory showed development of antral and secondary follicles 6 months after transplantation in a FSH-treated host grafted with PROH frozen-thawed ovaries from a human fetus 21 gestational weeks of age (5). Transplantation of fresh (7, 8, 10 –12) or frozen-thawed (7, 8, 11, 12) human ovarian fragments from women under the kidney capsule of immunodeficient mice has also been described. Antral development was achieved in grafts from fresh (7, 10) and PROH frozen-thawed tissue (7). In these studies (7, 8, 10 –12), survival and development of human adult follicles from both fresh and frozen-thawed tissue seemed to be better than that in our current study. This finding might be due to differences between human fetal and adult tissue. The human fetal ovaries we used included oogonia as well as follicles at initial stages of follicular formation, which accounts for the low number of follicles, 425
TABLE 3 Follicular count in kidney transplants of FSH-stimulated mice.
Propandiol Dimethyl sulfoxide
Fresh control
Thawed control
1 to 2 months
3 to 4 months
5 to 7 months
20.5 ⫾ 14.8 (7) 24.9 ⫾ 36.0 (14)b
20.0 ⫾ 15.3 (7) 22.4 ⫾ 23.4 (14)b
4 (1) 4 ⫾ 5.7 (18)
18.6 ⫾ 27.2 (11)a 1.8 ⫾ 1.3 (16)
22 (1) 5.2 ⫾ 5.4 (10)
Note: Values are means (⫾SD). Numbers in parentheses are of initial specimens and recovered grafts. a Significantly higher compared with dimethyl sulfoxide–frozen grafts that survived for 3 to 4 months (P⬍.001). b Significantly higher compared with dimethyl sulfoxide–frozen grafts that survived for 3 to 4 months (P⬍.01). Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
whereas ovaries from women contain formed follicles (1). Oogonia do not seem to undergo proper folliculogenesis after transplantation. Moreover, fetal germ cells might be more sensitive to ischemic changes and to inadequate vascularization after grafting. Of note, human fetal ovaries are structurally soft and delicate, whereas adult ovaries are tough and fibrous. Four studies have been published on subcutaneous transplantation of fresh (12–15) or frozen-thawed human ovaries from women (12, 14, 15). All showed active angiogenesis in the grafts and follicular survival. Administration of FSH to mice grafted with fresh (13) or PROH or DMSO frozenthawed tissue (12) resulted in growth to antral stages. However, the grafts were reduced in size (12, 14, 15) and fewer subcutaneous than kidney grafts were recovered (12). However, no differences were observed in follicular recovery, density, or ultrastructural quality at both grafting sites (14). We found poor graft vascualization in the subcutaneous site. Follicles failed to survive for long or to develop beyond the very early primordial stages, even in hosts treated with FSH. These differences might be explained by the dissimilarities between human adult and fetal ovaries. Our finding that the development of human follicles in immunodeficient mice can only be achieved beyond the primordial stage with FSH host treatment concurs with other
studies (7, 10, 12, 13, 15). In contrast, in animal studies, follicular multilayered development was achieved without FSH host treatment (3, 4). Although one group reported human follicular growth initiation in hosts that were not treated with FSH, multilayered follicles were not observed (11). However, the FSH doses used in the various studies differed (1.0 to 7.5 IU), as did the degree of purification of the commercial gonadotropin agents. Of note, in our study, the number of nontreated mice was small owing to the initial small number in this group and high morbidity during the experiment; thus, few grafts were recovered from nontreated mice. This might account for the nonsignificant differences in follicular survival in the FSH-treated groups vs. nontreated groups. Human ovarian tissue from adults has been frozen-thawed with DMSO or PROH (7, 8, 11, 12, 14, 15, 18, 19). In both cases, follicles appeared histologically normal (18). There were, however, more viable follicles in ovarian biopsies from DMSO frozen-thawed grafts than PROH frozenthawed grafts (8). Another study (12) used grafts frozenthawed with both DMSO and PROH, but because of their low overall survival rate, no conclusions could be drawn as to which cryoprotectant is optimal. Gook et al. (21) used transmission electron microscopy to evaluate the quality of thawed human follicles after the use
TABLE 4 Effects of FSH stimulation on the number of follicles in kidney transplanted fetal ovaries corresponding to graft period. 2 months
Fresh control 24.2 ⫾ 33.2 (21)a,b,c
Thawed control
With FSH
21.0 ⫾ 21.9 (21)a,b,c
4.0 ⫾ 5.7 (19)
3 to 4 months
Without FSH 4 (1)
5 to 7 months
With FSH
Without FSH
With FSH
Without FSH
8.6 ⫾ 11.9 (27)
0.5 ⫾ 1.2 (6)
4.3 ⫾ 4.6 (11)
1.4 ⫾ 1.3 (5)
Note: Values are means (⫾SD). Numbers in parentheses are initial specimens and recovered grafts. a Significantly higher compared with grafts recovered from stimulated mice after 2 months (P⬍.008). b Significantly higher compared with grafts from stimulated mice recovered after 3 to 4 months (P⬍.01). c Significantly higher compared with grafts from stimulated mice recovered after 5 to 7 months (P⬍.05). Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
426
Abir et al.
Human fetal ovaries in immunodeficient mice
Vol. 80, No. 2, August 2003
of various PROH freezing protocols. In the second stage of the study, human ovarian tissue cryopreserved by their optimized PROH method was transplanted, and follicles were developed to antral stages in FSH-treated hosts (7). However, only 50% of the grafts were recovered (7), which agrees with the proportion of PROH kidney transplants that we identified (58.8%). Three other studies described transplantation of DMSO frozen-thawed human ovarian tissue into immunodeficient mice (11, 14, 15). Reduced follicular density in the recovered grafts was found in two studies (11, 14), whereas the third reported survival of most of the primordial follicles (15). Transmission electron microscopy studies revealed similarities in the ultrastructural quality of the grafted follicles from fresh and DMSO frozen tissue (14). Proliferatingcell nuclear antigen staining in granulosa cells from DMSO frozen-thawed grafts was also shown (11). Proliferating-cell nuclear antigen is a protein involved in the cell cycle. Its expression in granulosa cells of primordial follicles correlates with initiation of folliculogenesis (10, 11, 13). Thus, proliferating-cell nuclear antigen expression has been found in granulosa cells of growing human follicles in grafts from fresh tissue (10, 13) or DMSO cryopreserved tissue (11). In our study, proliferating-cell nuclear antigen was expressed not only in growing grafted follicles from primary stages onwards but also in all ovarian components of control samples from fresh and thawed tissue. This apparent discrepancy can probably be explained by the proliferation of the stromal cells at mid pregnancy, concomitant with peak germ-cell mitotic activity (1). Proliferating-cell nuclear antigen staining in oocytes and granulosa cells has also been observed in two studies in which ovarian tissue from women was transplanted into imunodeficient mice (10, 13). The expression of proliferating-cell nuclear antigen in oocytes cannot be attributed to cell division, since the oocyte is not engaged in mitotic activity. However, the mammalian oocyte is not quiescent. Although no new DNA is created, DNA polymerase delta might be activated to repair possible damage to the nuclear and mitochondrial DNA in the oocyte, and during its activity, it might incorporate proliferating-cell nuclear antigen. The fetal follicles from PROH frozen samples showed better survival, indicating a benefit of PROH for freezing human fetal tissue. The low survival of grafted fetal germ cells suggests that it might be worthwhile to use grafted ovaries from older fetuses, which contain oocytes in formed primordial follicles. Moreover, to achieve better vascularization to grafted human fetal ovaries, highly vascularized transplantation sites (such as the kidney) should be chosen. Methods to speed up vascularization and reduce ischemic damage are needed. Nugent et al. (9) showed that vitamin E administration reduces ischemic changes to murine and human ovarian grafts. FERTILITY & STERILITY威
Our data are still preliminary, and to date few studies have been conducted on human fetal ovaries. More information is needed before human fetal oocytes can possibly be used for donations. Moreover, studies on human fetal ovaries are prohibited in many countries, as they are considered unethical. The overall worldwide availability of human fetal ovaries for research is thus very limited. To make progress in this field, further studies on fetal ovaries from other mammalian species should also be conducted.
Acknowledgments: The authors thank Ms. G. Ganzach from the Editorial Board of Rabin Medical Center for English-language editing, the staff at the Gynecology Ward for helping to locate suitable patients, and the Ultrasound Unit for identifying fetal gender.
References 1. Gosden RG. Oocyte development throughout life. In: Grudzinshes JG, Yovici JL, (eds). Gametes: the oocyte. Cambridge (UK): Cambridge University Press, 1995:119 –49. 2. Van den Hurk R, Abir R, Telfer EE, Bevers MM. Preantral and antral follicles as possible source for fertilizable oocytes in human and bovine. Hum Reprod Update 2000;6:457–74. 3. Gosden RG. Transplantation of ovaries and testes. In: Edwards RG, (ed). Fetal tissue transplants in medicine. Cambridge (UK): Cambridge University Press, 1992:253–79. 4. Nugent D, Meirow D, Brook PF, Aubard Y, Gosden RG. Transplantation in reproductive medicine: previous experience, present knowledge and future prospects. Hum Reprod Update 1997;3:267–80. 5. Abir R, Orvieto R, Raanani H, Fisch B, Schoenfeld A, Ginton D, et al. Development of human fetal follicles in an immunodeficient mouse. J Assist Reprod Genet 2000;17:393–6. 6. Cox SL, Shaw J, Jenkin G. Transplantation of cryopreserved fetal ovarian tissue to adult recipients in mice. J Reprod Fertil 1996;107: 315–22. 7. Gook DA, McCully BA, Edgar DH, McBain JC. Development of antral follicles in human cryopreserved ovarian tissue following xenografting. Hum Reprod 2001;16:417–22. 8. Newton H, Aubard Y, Rutherford A, Sharma V, Gosden R. Low temperature storage and grafting of human ovarian tissue. Hum Reprod 1996;11:1487–91. 9. Nugent D, Newton H, Gallivan L, Gosden RG. Protective effect of vitamin E on ischaemia-reperfusion in ovarian grafts. J Reprod Fertil 1998;114:341–6. 10. Oktay K, Newton H, Mullan J, Gosden RG. Development of human primordial follicles to antral stages in Scid/hpg mice stimulated with follicle stimulating hormone. Hum Reprod 1998;13:1133–8. 11. Oktay K, Newton H, Gosden RG. Transplantation of cryopreserved human ovarian tissue results in follicle growth initiation in SCID mice. Fertil Steril 2000;73:599 –603. 12. Van den Broecke R, Liu J, Handyside A, Van den Elst JC, Krausz T, Dhont M, et al. Follicular growth in fresh and cryopreserved human ovarian cortical grafts transplanted in immunodeficient mice. Europ J Obstet Gynecol Reprod Biol 2001;97:193–201. 13. Weissman A, Gotlieb L, Colgan T, Jurisicova A, Greenblatt EM, Casper RF. Preliminary experience with subcutaneous human ovarian cortex transplantation in the NOD-SCID mouse. Biol Reprod 1999;60: 1462–7. 14. Nisolle M, CasanasRoux 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;74:122–9. 15. Van den Broecke R, Van den Elst J, Liu J, Hovatta O, Dhont M. The female-to-male transsexual patient: a source of human ovarian cortical tissue for experimental use. Hum Reprod 2001;16:145–7. 16. Arrau J, Roblero L, Cury M, Gonzalez R. Effect of xenogenous sex steroids upon the number of germ cells and the growth of foetal grafted under the kidney capsule of adult ovariectomized hamsters. J Embryol Exp Morphol 1983;78:33–42. 17. Cox SL, Shaw J, Jenkin G. Follicular development in transplanted fetal and neonatal mouse ovaries is influenced by the gonadal status of the adult recipient. Fertil Steril 2000;74:366 –71.
427
18. Hovatta O, Silye R, Krausz T, Abir R, Margara RA, Trew G, et al. Cryopreservation of human ovarian tissue by using dimethylsulphoxide and propandiol-sucrose as cryoprotectants. Hum Reprod 1996;11:1268 –72. 19. Newton H, Fisher J, Arnold JRP, Pegg DE, Faddy MJ, Gosden RG. Permeation of human ovarian tissue with cryoprotective agents in preparation for cryopreservation. Hum Reprod 1998;13:376 –80.
428
Abir et al.
Human fetal ovaries in immunodeficient mice
20. Cushman RA, Wahl CM, Fortune JE. Bovine ovarian cortical pieces grafted to chick embryonic membranes: a model for studies on the activation of primordial follicles. Hum Reprod 2002;17:48 –54. 21. 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 propandiol. Hum Reprod 1999;14: 2061–8.
Vol. 80, No. 2, August 2003
Parameters affecting successful transplantation of frozen-thawed human fetal ovaries into immunodeficient mice Received July 30, 2002; revised and accepted November 4, 2002. Supported in part by research grants from Rabin Medical Center: a grant from the Steering Committee for Research Promotion (Dr. Orvieto) and a basic sciences research grant (Dr. Raanani). Presented in part at the International Workshop on Early Folliculogenesis and Oocyte Development: Basic and Clinical Aspects, London, United Kingdom, June 10 and 11, 1999; the Annual Meeting of the European Society for Human Reproduction and Embryology, Bologna, Italy, June 24 –27, 2000; and the Twelfth International Workshop on the Development and Function of the Reproductive Organs—Advances in Reproduction Research: From Follicle Culture to Nuclear Transfer and Embryonic Stem Cells, Ma’ale Hachamisha, Israel, April 30 –May 3, 2001. Reprint requests: Ronit Abir, Ph.D., IVF Research Laboratory, Infertility and IVF Unit, Department of Obstetrics and Gynecology, Rabin Medical Center, Beilinson Campus, Petah Tikva 49100, Israel (FAX: 972-3-9240533; Email: [email protected]). a Infertility and IVF Unit, Department of Obstetrics and Gynecology, Rabin Medical Center, Beilinson Campus. b Department of Obstetrics and Gynecology, Rabin Medical Center, Beilinson Campus. 0015-0282/03/$30.00 doi:10.1016/S0015-0282(03) 00658-7
Ronit Abir, Ph.D.,a,b Raoul Orvieto, M.D., M.Sc.,a,b Hila Raanani, M.D.,a,b Dov Feldberg, M.D.,a Shmuel Nitke, M.D.,a and Benjamin Fisch, M.D., Ph.D.a,b Infertility and IVF Unit, Department of Obstetrics and Gynecology, Rabin Medical Center, Beilinson Campus, Petah Tikva, and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel
Objective: To compare the development and survival of human fetal follicles frozen-thawed with dimethylsulfoxide (DMSO) and propandiol (PROH) in immunodeficient mice, to study the effects of host treatment with FSH, and to compare kidney and subcutaneous transplantation. Design: Controlled histologic study. Setting: Major tertiary care and referral academic center. Patient(s): Twenty-one women undergoing second-trimester pregnancy termination. Main Outcome Measure(s): Microscopic morphometric analysis and immunocytochemistry for proliferating-cell nuclear antigen in human fetal ovaries grafted into immunodeficient mice. Result(s): Renal grafts that were frozen-thawed with DMSO rather than PROH survived better in the hosts (79.6% compared with 58.8%), but significantly more follicles were identified in grafts frozen-thawed with PROH (P⬍.001). Follicular development was observed only in FSH-treated hosts, and follicular survival and development was better in the kidney than the subcutaneous site. Conclusion(s): This is the first report showing development of human fetal follicles in immunodeficient mice. Freezing-thawing with PROH seems to support development and survival better than with DMSO. The kidney is a better transplantation site than the subcutaneous site, probably because of its superior vascularization. Administration of FSH to the host is essential for follicular development. Follicular development and growth was better in ovarian grafts from older fetuses, as they contained more formed follicles. (Fertil Steril威 2003; 80:421– 8. ©2003 by American Society for Reproductive Medicine.) Key Words: Human fetal ovaries, primordial follicles, survival, proliferating-cell nuclear antigen, DMSO, PROH
The development of germ cells in the human fetal ovary has three phases: mitotic divisions, meiotic division, and follicular formation (1). Female germ cells, or oogonia, undergo many mitotic divisions until very close to term, with a peak during mid-pregnancy, followed by a drastic drop during the third trimester. Meiotic division commences gradually in the third month of pregnancy. The diplotene stage is achieved within weeks of initiation of meiosis, when the oogonia become oocytes. Oocytes are larger than oogonia and contain more intracellular organelles. They have completed genetic recombination and do not undergo additional nuclear maturation until after puberty. Follicular formation begins during the fourth month of pregnancy, after rapid prolif-
eration of the ovarian somatic cells to form a single flat layer of pregranulosa cells surrounding the oocytes (primordial follicles). At birth, the human ovary contains from 266,000 to 472,000 follicles, most of which remain nongrowing primordial follicles (1). The shortage of donated oocytes for IVF has prompted suggestions of the potential use of oocytes from frozen-stored ovaries of aborted fetuses. It is, however, unknown whether such oocytes have the same developmental capacity as those from adults. Furthermore, while FSH apparently induces follicular growth from the secondary stages, the signals that initiate the growth of earlier follicles are unknown (1), although recent studies suggest that some growth factors may be involved in this process 421
(2). Before frozen-thawed fetal oocytes are used clinically, their growth and maturation ability must be tested. Human ovarian slices transplanted into immunodeficient mice are not rejected, making these mice an excellent model for the study of follicular development (3–17). In most studies, the highly vascularized kidney capsule has served as the transplantation site (3–12, 16, 17). However, subcutanous transplantation is much easier to perform and sometimes permits external identification of developing follicles (12–15). We sought to examine survival and development of frozen-thawed human fetal germ cells in immunodeficient mice, to compare kidney with subcutaneous transplantation, and to determine whether FSH administration to the host benefits grafted fetal follicles.
MATERIALS AND METHODS Human Fetal Ovaries The study was approved by the local ethics committee of Rabin Medical Center. Ovaries were dissected from 23 human fetuses 19 to 22 gestational weeks of age after informed consent was obtained from the mothers. This fetal age was chosen because the number of germ cells peaks at 20 gestational weeks (1) and because our department’s pregnancy termination policy mandates feticide for all fetuses older than 22 gestational weeks, and we preferred to use ovaries from viable fetuses. Most of the abortions were performed because of fetal anatomic malformations or chromosomal abnormalities and were induced by prostaglandins. One slice 1 to 2 mm in length was fixed in Bouin’s solution immediately after ovarian dissection (fresh control). Because all abortions were performed late at night, the study did not include freshly dissected ovaries.
Cryopreservation and Thawing of Ovarian Tissue Ovarian samples were cut into slices 1 to 2 mm (thickness less ⬍ 1 mm) in size before freezing. Six to 12 slices were prepared from both ovaries of the same fetus, because the size of the ovaries varied and correlated with fetal age. Cryopreservation was carried out 1 to 2 hours after the abortion. Eight of the ovarian samples were frozen in a solution consisting of 1.5 M of 1,2 propandiol (PROH) (Sigma, St. Louis, MO) and 0.1 M of sucrose (Sigma) (18). The group included specimens from one fetus at 20 gestational weeks with central nervous system and pericardial malformations; five fetuses at 21 gestational weeks with renal and cutaneous malformations, trisomy 21, or 47,XXX (one each), or no abnormalities (pair of twins); and two fetuses at 22 gestational weeks both normal (one case of maternal syphilis). The other 15 samples were frozen in a solution consisting of 1.5 M of dimethyl sulfoxide (DMSO) (Sigma) and 0.1 M 422
Abir et al.
Human fetal ovaries in immunodeficient mice
TABLE 1 Division of ovarian tissue according to the transplantation sites. Cryoprotectant Propandiol Dimethyl sulfoxide
Total
Renal only
Subcutaneous only
Renal and subcutaneous
8 15
7 9
1 1
5
Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
of sucrose (Sigma) (19). This group included specimens from one fetus at 19 gestational weeks with lung and larynx malformations; seven fetuses at 20 gestational weeks with anencephaly and renal and central nervous system malformations, trisomy 13, or trisomy 21 (one each), or no abnormalities (three fetuses); six fetuses at 21 gestational weeks with diaphragmatic hernia or trisomy 21 (one each) or no abnormalities (four fetuses, including one pair of twins); and one fetus at 22 gestational weeks with anencephaly. Tissue slices 1 to 2 mm were placed in cryogenic vials (Nalge Nunc International, Delta, Roskilde, Denmark) filled with 0.8 mL PROH or DMSO. Before freezing, PROH samples were kept at room temperature for 10 minutes (5, 18) and the DMSO samples were kept on ice for 30 minutes, because DMSO might be toxic at room temperature (19). All samples were frozen slowly in a programmable freezer (Kryo 10, series 10/20; Planer Biomed, Sunbury on Thames, United Kingdom) and immediately placed in liquid nitrogen. The slices were cryopreserved for 1 week to 5 months. They were thawed by washouts with decreasing concentration gradients of PROH (Sigma) and sucrose (Sigma) or DMSO (Sigma) and sucrose (Sigma). One slice was fixed in Bouin’s solution immediately after thawing (thawed control).
Grafting Into Immunodeficient Mice Immunodeficient nu/nu female BalbC mice 6 to 8 weeks of age (Harlan, Jerusalem, Israel) were used for tissue grafting. All frozen-thawed tissue, apart from the thawed control, was used for grafting. The mice were anesthesized with an intramuscular injection of 0.04 mL of a 1:1 solution of 0.5% xylazine base (prepared from 2% XYL-M; VMD, Arendonk, Belgium) diluted with sterile distilled water and 50 mg/mL of ketamine hydrochloride (Ketalar; Parke Davis, Hampshire, United Kingdom). For the kidney grafts, the mice underwent bilateral oophorectomy, and the kidney capsule was opened and the slices were pushed inside (5). For the subcutaneous grafts, a small dorso median transverse incision was made, and the ovarian tissue slices were placed into the subcutaneous space above the flank without suturing (13). Table 1 shows the distribution of the ovaries according to the grafting sites. Vol. 80, No. 2, August 2003
TABLE 2 Survival of ovarian grafts. PROH Site Kidney Subcutaneous a
DMSO
No. of ovaries
No. of mice
Survival rate (%)
No. of ovaries
No. of mice
Survival rate (%)
7 1
34 2
20 (58.8) 1 (50)
14 6
49 13
39 (79.6) 12 (92.3)a
Significantly higher compared with propandiol kidney grafts (P⬍.04).
Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
Two thirds of the mice were stimulated with 1.0 U of recombinant FSH (Gonal-F; Laboratoires Serono S.A., Aubonne, Switzerland) (5, 7, 10), every second day starting 1 month after grafting (5). The mice were sacrificed periodically within 7 months, and the grafts were removed and fixed in Bouin’s solution.
Histologic Preparation and Immunocytochemical Analysis Control samples and transplanted ovarian tissue specimens removed from the hosts were prepared for paraffin embedding, sectioning (5 m), and staining with hematoxylin– eosin. The number of follicles in control and transplant samples were counted. Unstained sections were placed on OptiPlus positivecharged microscope slides (BioGenex, San Ramon, CA) and prepared for immunocytochemical study of proliferating-cell nuclear antigen (5, 10, 11, 13). The sections were deparaffinized, rehydrated, and quenched in 3% H2O2 (prepared from Perhydrol 30% H2O2; Merck, Darmstadt, Germany) in methanol (Bio Lab) for 30 minutes to block endogenous peroxidase activity. They were then microwaved for 15 minutes at full power (800 W) and for a further 15 minutes at 75% power with citrate buffer (pH, 6.0) that was diluted in distilled water from a ⫻20 citrate buffer solution (Zymed Laboratories, San Francisco, CA) to enhance antigen retrieval. After rinsing, 10% normal rabbit serum (Dako, Glostrup, Denmark) was added for 30 minutes at room temperature to inhibit nonspecific binding. The sections were incubated with a mouse anti–proliferating-cell nuclear antigen (Dako) at a dilution of 1:50 overnight at room temperature. The negative controls were incubated with a normal mouse IgGa2 antibody (Santa Cruz Biotechnology, Santa Cruz, CA), diluted to the same concentration as the primary antibody. Commercially available labeled slides were used as positive controls for proliferating-cell nuclear antigen (PCNA, human tonsil, Dako). The following morning, the samples were rinsed and incubated with a rabbit antimouse biotinylated immunoglubulin (Dako) at a dilution of 1:200 for 30 minutes. Finally, an avidin– biotin complex (StreptABC complex/HRP; FERTILITY & STERILITY威
Dako) diluted in hydrochloride Tris buffer at pH 7.6 (Sigma), was added for 30 minutes at room temperature. The sections were then exposed to a diaminobenzidine urea H2O2 solution in distilled water (Sigma Fast tablets; Sigma) for 5 minutes (brown dye for proliferating-cell nuclear antigen expression). The sections were counterstained with hematoxylin (blue-purple dye). Unless otherwise stated, all dilutions were performed with phosphate-buffered saline at pH 7.6 (Biological Industries, Beit Ha’emek, Israel). Phosphate-buffered saline also served as the main rinsing solution. At least two slides were stained for every graft.
Statistical Analysis Data were statistically analyzed by using the unpaired Student t-test, 2 test, or Fisher test, as indicated. Values of at least P⬍.05 were considered statistically significant.
RESULTS Figures 1 through 5 show follicles during various stages of the experiment. Because of a high rate of host morbidity during the first postoperative month (usually shortly after the operation), only hosts that survived for at least 1 month were included (Tables 2, 3, and 4). Table 2 shows survival rates of ovarian grafts after at least 1 month in the hosts. Significantly more subcutaneous DMSO frozen-thawed grafts than PROH frozen-thawed renal transplants were recovered (P⬍.04). Of the 13 recovered subcutaneous grafts, 11 were identified within 2 months; no follicles survived in them. These included two necrotic pieces from hosts that were not treated with FSH. Two other subcutaneous grafts were identified in FSH treated mice after 4 and 6 months; they contained very early primordial follicles (four and three follicles, respectively) (Fig. 4). Table 3 shows the number of surviving follicles after kidney transplantation in FSH-treated hosts. The initial number of follicles in both freshly dissected and thawed samples was significantly higher than in the DMSO frozen grafts after a postoperative period of 3 to 4 months (P⬍.01). Among the grafts that survived 3 to 4 months, the PROH frozen transplants had significantly more surviving follicles 423
FIGURE 1
FIGURE 3
A propandiol frozen-thawed ovary from a normal fetus 21 gestational weeks of age (control). The arrow indicates primordial follicles. Hematoxylin– eosin, ⫻400.
A human fetal ovary 6.5 months after renal grafting. The ovaries were frozen-thawed in dimethyl sulfoxide and originated from a 21-gestational-week-old fetus with trisomy 21. The arrow indicates secondary follicles. Note the brown proliferating-cell nuclear antigen staining in granulosa cells and oocytes. Magnification, ⫻200.
Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
than did the DMSO frozen-thawed transplants (P⬍.001). In some cases, oogonia was identified in recovered grafts up to 6 months postoperatively.
FIGURE 2 A primary follicle (P) identified 4 months after grafting. The ovaries were frozen-thawed with propandiol and originated from a 47,XXX fetus 21 gestational weeks of age. Note the brown proliferating-cell nuclear antigen staining in most of the granulosa cells and in the nuclei of the oocyte. Magnification, ⫻400.
Abir. Human fetal ovaries in immunodeficient mice.Fertil Steril 2003.
Table 4 shows the effects of FSH host treatment on survival of follicles from fetal ovaries transplanted under the kidney capsule. Because of the small number of surviving grafts and follicles, the results from PROH frozen-thawed samples and DMSO frozen-thawed samples were combined. Significantly more follicles were found in the control samples (both freshly dissected and thawed) than in the grafts recovered from FSH-treated mice after 1 to 2 months (P⬍.008), after 3 to 4 months (P⬍.01), and after 5 to 7 months (P⬍.05). Graft survival rates did not significantly differ between FSH-treated hosts and nontreated hosts. A few primary, secondary, and antral follicles were identified after at least 4 months, but only in FSH-treated hosts: Five primary follicles were identified in renal grafts after 4 months, and secondary (five and eight follicles per graft) and antral follicles (two follicles per graft, up to 0.5 mm in diameter) were identified in two kidney grafts after 6 to 7 months. The ovaries in which multilayered follicles developed were obtained from chromosomally abnormal fetuses at 21 gestational weeks: one from a 47,XXX fetus whose ovaries were frozen-thawed with PROH (4), and the other from a trisomy 21 fetus whose ovaries were frozen-thawed with DMSO (Fig. 3). Of note, follicular development was better in ovarian grafts from the older fetuses, as they contained initially a larger number of formed follicles (Fig. 1).
Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
424
Abir et al.
Human fetal ovaries in immunodeficient mice
Positive proliferating-cell nuclear antigen staining was observed in all ovarian components of the freshly dissected Vol. 80, No. 2, August 2003
FIGURE 4
FIGURE 5
An ovary from a fetus after 6 months of subcutaneous grafting. The ovaries were frozen-thawed in dimethyl sulfoxide and originated from a normal fetus 21 gestational weeks of age. The arrow indicates the initial formation of primordial follicles. Hematoxylin– eosin, ⫻400.
A negative control of the same graft shown in Figure 3. Note the secondary follicle, the overall purple-blue staining, and the lack of brown proliferating-cell nuclear antigen staining. Magnification, ⫻400.
Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003. Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
and thawed control samples. It was also identified in the grafted growing follicles from primary stages onward (Fig. 2 and 3). No negative control stained positive for proliferatingcell nuclear antigen (Fig. 5), whereas positive controls stained positive for proliferating-cell nuclear antigen (data not shown).
DISCUSSION Apart from our previous case report (5), this is the first study showing survival and development of frozen-thawed human fetal follicles in immunodeficient mice. Although transplants frozen-thawed with DMSO had a higher survival rate, more follicles were identified in PROH frozen-thawed grafts 3 to 4 months postoperatively. Despite the high number of identified subcutaneous grafts, very few follicles survived in these transplants, and development beyond the primordial stage was achieved only in the kidney-grafted FSH-treated hosts. However, our results should be interpreted with caution, as only 11 of the 23 ovaries were from normal fetuses, and development to multilayered follicles occurred in ovarian grafts from two chromosomally abnormal fetuses. Fresh (5, 16) and frozen-thawed (5) fetal murine ovaries were transplanted under the kidney capsule (6, 17) or in the bursal cavity of adult mice (6). Follicular development was identified in the grafts (6, 17), and in some cases in the bursal cavity group, oophorectomized hosts conceived from donor FERTILITY & STERILITY威
fetal oocytes (6). When ovaries from fetal hamsters were transplanted under the kidney capsule of adult hamsters, the number of growing follicles increased after host treatment with estradiol or testosterone propionate (16). In a recent study (20), bovine fetal cortical ovarian slices were grafted to the chorioalantoic membrane of cultured chick embryos for up to 10 days. Although the grafts became rapidly infiltrated with blood vessels, no follicular growth was initiated. In a preliminary study, when freshly dissected fragments from human fetuses 16 gestational weeks of age were transplanted under the kidney capsule of immunodeficient mice, germ cells in mitosis and at all stages of meiotic prophase were abundant, and initial follicular formation was observed 6 months after grafting (3). In addition, a pilot report from our laboratory showed development of antral and secondary follicles 6 months after transplantation in a FSH-treated host grafted with PROH frozen-thawed ovaries from a human fetus 21 gestational weeks of age (5). Transplantation of fresh (7, 8, 10 –12) or frozen-thawed (7, 8, 11, 12) human ovarian fragments from women under the kidney capsule of immunodeficient mice has also been described. Antral development was achieved in grafts from fresh (7, 10) and PROH frozen-thawed tissue (7). In these studies (7, 8, 10 –12), survival and development of human adult follicles from both fresh and frozen-thawed tissue seemed to be better than that in our current study. This finding might be due to differences between human fetal and adult tissue. The human fetal ovaries we used included oogonia as well as follicles at initial stages of follicular formation, which accounts for the low number of follicles, 425
TABLE 3 Follicular count in kidney transplants of FSH-stimulated mice.
Propandiol Dimethyl sulfoxide
Fresh control
Thawed control
1 to 2 months
3 to 4 months
5 to 7 months
20.5 ⫾ 14.8 (7) 24.9 ⫾ 36.0 (14)b
20.0 ⫾ 15.3 (7) 22.4 ⫾ 23.4 (14)b
4 (1) 4 ⫾ 5.7 (18)
18.6 ⫾ 27.2 (11)a 1.8 ⫾ 1.3 (16)
22 (1) 5.2 ⫾ 5.4 (10)
Note: Values are means (⫾SD). Numbers in parentheses are of initial specimens and recovered grafts. a Significantly higher compared with dimethyl sulfoxide–frozen grafts that survived for 3 to 4 months (P⬍.001). b Significantly higher compared with dimethyl sulfoxide–frozen grafts that survived for 3 to 4 months (P⬍.01). Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
whereas ovaries from women contain formed follicles (1). Oogonia do not seem to undergo proper folliculogenesis after transplantation. Moreover, fetal germ cells might be more sensitive to ischemic changes and to inadequate vascularization after grafting. Of note, human fetal ovaries are structurally soft and delicate, whereas adult ovaries are tough and fibrous. Four studies have been published on subcutaneous transplantation of fresh (12–15) or frozen-thawed human ovaries from women (12, 14, 15). All showed active angiogenesis in the grafts and follicular survival. Administration of FSH to mice grafted with fresh (13) or PROH or DMSO frozenthawed tissue (12) resulted in growth to antral stages. However, the grafts were reduced in size (12, 14, 15) and fewer subcutaneous than kidney grafts were recovered (12). However, no differences were observed in follicular recovery, density, or ultrastructural quality at both grafting sites (14). We found poor graft vascualization in the subcutaneous site. Follicles failed to survive for long or to develop beyond the very early primordial stages, even in hosts treated with FSH. These differences might be explained by the dissimilarities between human adult and fetal ovaries. Our finding that the development of human follicles in immunodeficient mice can only be achieved beyond the primordial stage with FSH host treatment concurs with other
studies (7, 10, 12, 13, 15). In contrast, in animal studies, follicular multilayered development was achieved without FSH host treatment (3, 4). Although one group reported human follicular growth initiation in hosts that were not treated with FSH, multilayered follicles were not observed (11). However, the FSH doses used in the various studies differed (1.0 to 7.5 IU), as did the degree of purification of the commercial gonadotropin agents. Of note, in our study, the number of nontreated mice was small owing to the initial small number in this group and high morbidity during the experiment; thus, few grafts were recovered from nontreated mice. This might account for the nonsignificant differences in follicular survival in the FSH-treated groups vs. nontreated groups. Human ovarian tissue from adults has been frozen-thawed with DMSO or PROH (7, 8, 11, 12, 14, 15, 18, 19). In both cases, follicles appeared histologically normal (18). There were, however, more viable follicles in ovarian biopsies from DMSO frozen-thawed grafts than PROH frozenthawed grafts (8). Another study (12) used grafts frozenthawed with both DMSO and PROH, but because of their low overall survival rate, no conclusions could be drawn as to which cryoprotectant is optimal. Gook et al. (21) used transmission electron microscopy to evaluate the quality of thawed human follicles after the use
TABLE 4 Effects of FSH stimulation on the number of follicles in kidney transplanted fetal ovaries corresponding to graft period. 2 months
Fresh control 24.2 ⫾ 33.2 (21)a,b,c
Thawed control
With FSH
21.0 ⫾ 21.9 (21)a,b,c
4.0 ⫾ 5.7 (19)
3 to 4 months
Without FSH 4 (1)
5 to 7 months
With FSH
Without FSH
With FSH
Without FSH
8.6 ⫾ 11.9 (27)
0.5 ⫾ 1.2 (6)
4.3 ⫾ 4.6 (11)
1.4 ⫾ 1.3 (5)
Note: Values are means (⫾SD). Numbers in parentheses are initial specimens and recovered grafts. a Significantly higher compared with grafts recovered from stimulated mice after 2 months (P⬍.008). b Significantly higher compared with grafts from stimulated mice recovered after 3 to 4 months (P⬍.01). c Significantly higher compared with grafts from stimulated mice recovered after 5 to 7 months (P⬍.05). Abir. Human fetal ovaries in immunodeficient mice. Fertil Steril 2003.
426
Abir et al.
Human fetal ovaries in immunodeficient mice
Vol. 80, No. 2, August 2003
of various PROH freezing protocols. In the second stage of the study, human ovarian tissue cryopreserved by their optimized PROH method was transplanted, and follicles were developed to antral stages in FSH-treated hosts (7). However, only 50% of the grafts were recovered (7), which agrees with the proportion of PROH kidney transplants that we identified (58.8%). Three other studies described transplantation of DMSO frozen-thawed human ovarian tissue into immunodeficient mice (11, 14, 15). Reduced follicular density in the recovered grafts was found in two studies (11, 14), whereas the third reported survival of most of the primordial follicles (15). Transmission electron microscopy studies revealed similarities in the ultrastructural quality of the grafted follicles from fresh and DMSO frozen tissue (14). Proliferatingcell nuclear antigen staining in granulosa cells from DMSO frozen-thawed grafts was also shown (11). Proliferating-cell nuclear antigen is a protein involved in the cell cycle. Its expression in granulosa cells of primordial follicles correlates with initiation of folliculogenesis (10, 11, 13). Thus, proliferating-cell nuclear antigen expression has been found in granulosa cells of growing human follicles in grafts from fresh tissue (10, 13) or DMSO cryopreserved tissue (11). In our study, proliferating-cell nuclear antigen was expressed not only in growing grafted follicles from primary stages onwards but also in all ovarian components of control samples from fresh and thawed tissue. This apparent discrepancy can probably be explained by the proliferation of the stromal cells at mid pregnancy, concomitant with peak germ-cell mitotic activity (1). Proliferating-cell nuclear antigen staining in oocytes and granulosa cells has also been observed in two studies in which ovarian tissue from women was transplanted into imunodeficient mice (10, 13). The expression of proliferating-cell nuclear antigen in oocytes cannot be attributed to cell division, since the oocyte is not engaged in mitotic activity. However, the mammalian oocyte is not quiescent. Although no new DNA is created, DNA polymerase delta might be activated to repair possible damage to the nuclear and mitochondrial DNA in the oocyte, and during its activity, it might incorporate proliferating-cell nuclear antigen. The fetal follicles from PROH frozen samples showed better survival, indicating a benefit of PROH for freezing human fetal tissue. The low survival of grafted fetal germ cells suggests that it might be worthwhile to use grafted ovaries from older fetuses, which contain oocytes in formed primordial follicles. Moreover, to achieve better vascularization to grafted human fetal ovaries, highly vascularized transplantation sites (such as the kidney) should be chosen. Methods to speed up vascularization and reduce ischemic damage are needed. Nugent et al. (9) showed that vitamin E administration reduces ischemic changes to murine and human ovarian grafts. FERTILITY & STERILITY威
Our data are still preliminary, and to date few studies have been conducted on human fetal ovaries. More information is needed before human fetal oocytes can possibly be used for donations. Moreover, studies on human fetal ovaries are prohibited in many countries, as they are considered unethical. The overall worldwide availability of human fetal ovaries for research is thus very limited. To make progress in this field, further studies on fetal ovaries from other mammalian species should also be conducted.
Acknowledgments: The authors thank Ms. G. Ganzach from the Editorial Board of Rabin Medical Center for English-language editing, the staff at the Gynecology Ward for helping to locate suitable patients, and the Ultrasound Unit for identifying fetal gender.
References 1. Gosden RG. Oocyte development throughout life. In: Grudzinshes JG, Yovici JL, (eds). Gametes: the oocyte. Cambridge (UK): Cambridge University Press, 1995:119 –49. 2. Van den Hurk R, Abir R, Telfer EE, Bevers MM. Preantral and antral follicles as possible source for fertilizable oocytes in human and bovine. Hum Reprod Update 2000;6:457–74. 3. Gosden RG. Transplantation of ovaries and testes. In: Edwards RG, (ed). Fetal tissue transplants in medicine. Cambridge (UK): Cambridge University Press, 1992:253–79. 4. Nugent D, Meirow D, Brook PF, Aubard Y, Gosden RG. Transplantation in reproductive medicine: previous experience, present knowledge and future prospects. Hum Reprod Update 1997;3:267–80. 5. Abir R, Orvieto R, Raanani H, Fisch B, Schoenfeld A, Ginton D, et al. Development of human fetal follicles in an immunodeficient mouse. J Assist Reprod Genet 2000;17:393–6. 6. Cox SL, Shaw J, Jenkin G. Transplantation of cryopreserved fetal ovarian tissue to adult recipients in mice. J Reprod Fertil 1996;107: 315–22. 7. Gook DA, McCully BA, Edgar DH, McBain JC. Development of antral follicles in human cryopreserved ovarian tissue following xenografting. Hum Reprod 2001;16:417–22. 8. Newton H, Aubard Y, Rutherford A, Sharma V, Gosden R. Low temperature storage and grafting of human ovarian tissue. Hum Reprod 1996;11:1487–91. 9. Nugent D, Newton H, Gallivan L, Gosden RG. Protective effect of vitamin E on ischaemia-reperfusion in ovarian grafts. J Reprod Fertil 1998;114:341–6. 10. Oktay K, Newton H, Mullan J, Gosden RG. Development of human primordial follicles to antral stages in Scid/hpg mice stimulated with follicle stimulating hormone. Hum Reprod 1998;13:1133–8. 11. Oktay K, Newton H, Gosden RG. Transplantation of cryopreserved human ovarian tissue results in follicle growth initiation in SCID mice. Fertil Steril 2000;73:599 –603. 12. Van den Broecke R, Liu J, Handyside A, Van den Elst JC, Krausz T, Dhont M, et al. Follicular growth in fresh and cryopreserved human ovarian cortical grafts transplanted in immunodeficient mice. Europ J Obstet Gynecol Reprod Biol 2001;97:193–201. 13. Weissman A, Gotlieb L, Colgan T, Jurisicova A, Greenblatt EM, Casper RF. Preliminary experience with subcutaneous human ovarian cortex transplantation in the NOD-SCID mouse. Biol Reprod 1999;60: 1462–7. 14. Nisolle M, CasanasRoux 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;74:122–9. 15. Van den Broecke R, Van den Elst J, Liu J, Hovatta O, Dhont M. The female-to-male transsexual patient: a source of human ovarian cortical tissue for experimental use. Hum Reprod 2001;16:145–7. 16. Arrau J, Roblero L, Cury M, Gonzalez R. Effect of xenogenous sex steroids upon the number of germ cells and the growth of foetal grafted under the kidney capsule of adult ovariectomized hamsters. J Embryol Exp Morphol 1983;78:33–42. 17. Cox SL, Shaw J, Jenkin G. Follicular development in transplanted fetal and neonatal mouse ovaries is influenced by the gonadal status of the adult recipient. Fertil Steril 2000;74:366 –71.
427
18. Hovatta O, Silye R, Krausz T, Abir R, Margara RA, Trew G, et al. Cryopreservation of human ovarian tissue by using dimethylsulphoxide and propandiol-sucrose as cryoprotectants. Hum Reprod 1996;11:1268 –72. 19. Newton H, Fisher J, Arnold JRP, Pegg DE, Faddy MJ, Gosden RG. Permeation of human ovarian tissue with cryoprotective agents in preparation for cryopreservation. Hum Reprod 1998;13:376 –80.
428
Abir et al.
Human fetal ovaries in immunodeficient mice
20. Cushman RA, Wahl CM, Fortune JE. Bovine ovarian cortical pieces grafted to chick embryonic membranes: a model for studies on the activation of primordial follicles. Hum Reprod 2002;17:48 –54. 21. 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 propandiol. Hum Reprod 1999;14: 2061–8.
Vol. 80, No. 2, August 2003