Introduction
Ovarian stimulation (OS) has been recognized for many years as one important part of clinically assisted reproduction techniques (ART), which induces the development of supernumerary follicles and multiple ovulations in animal and human subjects [
1]. Despite the increasing success rate of ART, most infertile couples need more than one cycle of controlled ovarian hyperstimulation to achieve pregnancy.
Animal studies have revealed detrimental effects of gonadotropin stimulation [
2-
6]. Concern has been raised on whether repeated OS cycles might alter the oocyte quality [
7], decrease the number of retrieved oocytes in the subsequent cycles, increase embryonic mortality, and result in fetal retardation [
8,
9]. Given the potential abnormalities in embryo and offspring elicited by repeated OS, it is unclear whether consecutive OS is a complete risk-free procedure in terms of ovarian reserves for adult females treated by these repeated pharmacological therapies. Also, the current study demonstrated that the ovarian surface epithelium (OSE) cells which continuously proliferated in culture eventually became transformed [
10-
12]. The gonadotropins responsible for ovulation may stimulate both cancer initiation and progression. It is unknown whether repeated ovulation might induce ovarian cancer
in vivo.
In this study, we employed mouse models for further study. Experiments designed here addressed the changes of repeated cycles of stimulation on ovarian tissue of the mice, which were assessed by using the percentages of primordial, primary and secondary follicles in mouse ovarian tissue, oocyte retrieval rate and embryo developmental competence for ovarian reserves, and the profile of the OSE cells for ovarian cancer initiation, thus covering a range of intraovarian and postfertilization parameters. Our aim was to assess the effect of multiple OS on ovarian tissues of adult females.
Materials and methods
Subjects
Kunming mice (the Centre of Animals of Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China), 4 wk of age, were kept in a room in a 14 h light/10 h dark cycle with ambient temperature being (20±2)°C. Food and water were provided ad libitum. Typically, five females were kept per cage, and a total of 150 females were enrolled to complete the present studies.
Throughout the present study, cycles of OS were initiated in 4-wk-old females that received intraperitoneal (i.p.) injections of 5 IU pregnant mare serum gonadotrophin (PMSG, Tianjin, China) with predominantly follicle stimulating hormone (FSH)-like activity, 48 h later being followed by 5 IU i.p. of human chorionic gonadotropin (hCG, Tianjin, China), with luteinizing hormone (LH) activity [
13]. Hormones were prepared in 0.9 g/mL sterile NaCl (vehicle) and originated from the same lot number for the entire study. Eighty female mice were induced to superovulate. Forty mice for the repeated ovarian stimulated group were stimulated with PMSG and hCG at 6-day intervals for 5 cycles. Another 40 of single ovarian stimulated group were injected with PMSG and hCG at the same time when the repeated ovarian stimulated group was injected with PMSG and hCG for the last cycle. When the repeated ovarian stimulated group and single ovarian stimulated group were injected with hCG, 70 female mice at oestrus stage without any treatment served as the control group. Firstly, 35 females in the repeated and single ovarian stimulated groups and 65 females in control group were placed with same strain male mice, and successful mating was identified by the presence of a vaginal plug the following morning. In this stage, we obtained 20 mating females in each group for all embryo collections and morphology studies of ovarian tissue. Secondly, the repeated ovarian stimulated group and the single ovarian stimulated group were injected with bromodeoxyuridine (BrdU) at 100 mg/kg at the time of the last hCG injection. Injections containing BrdU were given in the control group at the oestrus stage. Each experimental group contained five animals for the proliferation study.
Analysis of ovarian follicular population and morphology
To assess the morphological change of follicles after repeated gonadotropin stimulation, we examined the follicle morphology by histology. The ovaries were collected from female Kunming mice of the repeated ovarian stimulated group, single ovarian stimulated group and the control group. Six samples from six different ovaries in each group were detected. The specimens were embedded in paraffin blocks and serially sectioned into 5-μm slices. They were stained with haematoxylin and eosin (H&E) and observed by light microscopy (×400). Follicles were classified as follows: (1) primordial follicles with one layer of flattened granulosa cells surrounding the oocyte; (2) primary follicles with one layer of cuboid granulosa cells, and (3) secondary follicles with two or three layers of granulosa cells [
14]. Antral follicles were not included in this study. To avoid counting follicles more than once, one section was selected from every three sections, and only follicles with a visible nucleus were counted [
15].
The definition of normality of follicles was based on the criteria proposed by Lucci
et al [
16]. The follicles classified as morphologically normal were those containing an intact oocyte surrounded by well-organized granulosa cells (Fig. 1a, b). In contrast, follicles were defined as degenerated if they had one or more of the following aspects: a pyknotic oocyte nucleus, shrunken ooplasm or disorganized granulosa cells (Fig. 1b, c).
Examination of ovarian response and embryo developmental ability
To evaluate the ovarian response and embryo developmental ability in mice, we collected the embryos from each group. Every group had 15 mating females. One-cell embryos were flushed from oviducts at 16-18 h post-hCG injection. All culture reagents were purchased from Life Technologies, Gibco BRL (USA) unless stated otherwise. All culture media were freshly prepared. Embryos were obtained in the collection medium and cultured in the drops of media overlaid with embryo-tested washed mineral oil (Sigma Biosciences, USA) for 5 days in a humidified atmosphere of 5% CO
2 at 37°C [
17]. Evaluation of embryonic development (cleavage, good quality embryos, blastocysts) was continued every 24 h until 120 h, at which time cultures were terminated.
Experimental design of proliferation study
Each experimental group contained five animals. Mice were sacrificed by cervical dislocation. Ovaries, including fat pads and bursae, were collected 12 h after the last injection of hCG. Ovaries were fixed in 4% paraformaldehyde for 8-12 h, and then dehydrated with ethanol, embedded in paraffin, and serially sectioned at 4 μm.
The proliferation of ovarian surface epithelium were evaluated by immunohistochemistry using the antibodies against BrdU (sheep; 1∶50 dilution; Santa Cruz, USA). The sections were deparaffinized in xylene and rehydrated with subsequent ethanol dilutions. Antigen retrieval was performed using 1 mmol/L sodium citrate by microwaving. Endogenous peroxidase was blocked with 3% H2O2 at room temperature for 10 min. After blocking for 45 min with 1.5% normal horse serum in PBS, the sections were incubated overnight at 4°C with primary antibody in PBS. The sections were then incubated at 37°C for 30 min in a secondary antibody (ZDR-5308, Zhongshan Biotechnology, China) at 1∶50 and incubated subsequently with avidin and biotinylated peroxidase at room temperature for 45 min, and finally with diaminobenzidine (DAB, 400 mg/mL) at room temperature for 3 min. Hematoxylin was used for counterstaining. The staining of 200 consecutive cells in five non-adjacent microscopic fields was evaluated. The total numbers of cells and positive cells in the three groups were counted separately. Counts for each group were averaged. Control slides received serum block instead of the primary antibody.
Statistical analysis
Data are presented as ±s, and variations are indicated with the standard error mean (SEM, errors bars on graphs). The data were analyzed by using SPSS 11.5 Statistics Package for Social Sciences (SPSS). Comparison of proportions of oocyte category for samples was performed with a Chi-square test. The retrieval embryos following superovalution were evaluated using a nonparametric Kruskal-Wallis test followed by the Mann-Whitney test for two independent samples. Number of cleavages, good quality embryos and blastocyst cell counts were analyzed using one-way ANOVA followed by least significant difference (LSD) and Tukey post hoc tests. When in disagreement, results using the more stringent Tukey post hoc were reported unless noted otherwise. Assumptions that the populations were normal (Shapiro-Wilk and Kolmogorov-Smirnov tests for normality) and population variances were all equal (Levene test) were checked prior to performing ANOVA. Furthermore, if homogeneity of variances could not be assumed, a Tamhane post hoc was used. Difference was considered significant at a P<0.05.
Results
Follicular population and morphology
The percentages of primordial, primary and secondary follicles in mouse ovarian tissue after treatment with different superovulation cycles and that of the control group are presented in Fig. 2. In the control, single ovarian stimulated and repeated ovarian stimulated groups, the percentages of primordial follicles were 13.08%, 12.30%, and 16.38%, respectively; the percentages of primary follicles were 18.45%, 23.77%, and 23.28%, respectively; the percentages of secondary follicles were 68.46%, 63.93%, and 60.34%, respectively. No significant difference was observed in the proportions of different stages of follicles in the three groups. Values were the percentages of different stage follicles in the total count of observed follicles.
The percentages of morphologically normal primary and secondary follicles are presented in Figure 3. The percentages in the control group (91.67%; 94.38%) and single ovarian stimulated group (89.66%; 91.03%) were significantly higher than those in the repeated ovarian stimulated group (66.67%; 72.86%).
Ovarian response and embryo developmental ability
In the present study, the ovulatory responsivity of females to single or repeated super-ovulation was evidenced by the presence of one-cell embryos in the oviducts 16-18 h post-hCG. Among animals that responded to superovulation treatment with PMSG and hCG, the mean number of ovarian oocytes or embryos remained increased in the single ovarian stimulated group (39.25±10.77 one-cell embryos/female), compared with the control group (15.90±2.75 one-cell embryos/female; P=0.000); however, following the repeated stimulation of PMSG and hCG, fewer oocytes or embryos were retrieved per female compared with the single ovarian stimulated and control groups (5.15±2.81 eggs/female; P<0.001). Numbers reported here included all eggs besides morphologically abnormal ones showing signs of degeneration, lysis, or fragmentation. We tested the ability of embryo development from cleavage to blastocyst in vitro. Repeated gonadotropin stimulation significantly decreased the cleavage rate (67.47%), good quality embryo rate (2.41%) and blastocyst reproduction rate (0) compared with the other groups. Meanwhile, there was no significant difference on the above-mentioned analysis between the control group (85.57%; 70.10%; 65.64%) and the single ovarian stimulated group (81.76%; 68.54; 64.74%). Thus, following repeated OS, a progressive decrease may reduce embryo developmental ability in mice (Fig. 4, P<0.05).
Cumulative proliferation of OSE in response to superovulation
To investigate proliferative changes of OSE cells in response to ovulation, BrdU incorporation of the OSE was quantified in mice. In this study, animals were injected with BrdU at the time of injection of hCG to label the proliferation that occurred during the time of follicular rupture and repair, until the animals were sacrificed, for a total of 12 h. Measured proliferation was defined as postovulatory proliferation. Although the timing of ovulation after hCG varied slightly from animal to animal, the optimized time point when the animals were sacrificed was at the 12th h after hCG (datum not shown).
The proliferation of the OSE in response to gonadotropins occurred primarily within certain anatomical regions of the ovary, primarily near antral follicles and corpus luteum (CL), compared with regions distal from follicular development [
18].
Proliferation of the OSE adjacent to antral follicles and CL was detected in all sections analyzed. The incorporation levels of BrdU in granulosa cells of developing follicles served as an internal positive control to monitor proliferation. OSE cells of both squamous and cuboidal morphologies were found to have proliferated in all of the ovaries analyzed (Fig. 5).
The OSE cells adjacent to antral follicles and CL in the repeated ovarian stimulated group (81.8%) had a significantly higher proliferation rate than the other groups, and this difference was also significant between the single ovarian stimulated group and the control group. The proliferation rate of the OSE cell in the single ovarian stimulated group (56.4%) was significantly higher than that in the control group (37.5%).
Discussion
Superovulation by injection of exogenous gonadotropin is an elementary method to produce increasing follicles in ART. Despite widespread utilization, OS has been associated with a number of detrimental reproductive effects on the oocyte and embryo [
19-
22]. However, the unpredictability of the effects of repeated on the ovarian tissue still remains a major problem following repeated OS. This study was designed to test the effect of repeated OS on the ovarian reserve, ovarian response and ovarian cancer genesis.
We first examined the follicle morphology by histology. There was no difference in the percentages of primordial, primary and secondary follicles in mouse ovarian tissue between the mice after treatments with different superovulation cycle and controls. The oocytes retrieved in one cycle seemed to come from the antral pool that many follicles destined to become atretic due to dominant follicle selection escape from this end [
23]. In the OS cycle, obtaining more follicles may be due to gonadotropins altering the physiologic selection of one single dominant follicle but not accelerating the recruitment of follicles from further cycles. The findings of this study confirmed that there was no detrimental effect on ovarian reserve after repetitive controlled ovarian hyperstimulation.
However, during repeated OS cycles, there was an increase in the percentage of morphologically abnormal primary and secondary follicles, suggesting that repetitive ovarian hyperstimulation may induce damage in follicles of mouse ovaries. Recent available data indicated that more aggregated mitochondria were found in the cytoplasm of the repetitively stimulated embryos. Higher amounts of oxidative substances including 8-OH-dG, lipoperoxides, and carbonyl proteins were revealed in the ovaries with more cycle numbers of ovarian stimulation [
24,
25]. Repetitive gonadotropin injections may result in high gonadotrophin milieu in mice receiving repeated OS. Maybe it is the hormone accumulation induces oxidative damage and mitochondrial DNA mutations in mouse ovaries.
An understanding of the relationship between embryonic developmental competence and ovarian responses to stimulation in the mouse may provide insights into oocyte defects. Our experimental design allowed comparison of the embryonic developmental competences derived from the repeated ovarian stimulated group, single ovarian stimulated group, and control group. The follicular/embryonic counts were significantly greater in single OS cycles than those in repeated OS cycles and the control group. However, the numbers of oocytes/embryos were decreased in the group with more ovarian stimulation. During embryo culture, we also found that embryos in the repeated ovarian stimulated group failed to develop into the blastocyst stage. The reason why repeated OS compromises gametic and embryonic development is not clear. Perhaps accelerated follicular growth induced by repeated OS would yield oocytes with decreased developmental competence because of incomplete oogenesis [
26].
One hypothesis regarding ovarian cancer is that an increased number of ovulations contributes to the formation of transformed cells. The ovarian surface epithelium (OSE) is a single cell layer of squamous and cuboidal cells and is important for the integrity of the ovary and serves as the regulatory barrier at the time of ovulation [
27]. The OSE receives attention because these cells are considered to be the progenitors of 90% of ovarian cancers [
28]. In this study, we also observed the proliferation of OSE in superovulation mice and in controls. We found that OSE cells in superovulated mice proliferated more rapidly than those in controls. The proliferation rate of OSE adjacent to antral follicles and CL in repeated superovulated mice was higher than that in a single cycle, which may demonstrate that the repeated gonadotropins responsible for ovulation may stimulate OSE cellular proliferation. Because the ovary must repair the surface after each ovulation, proliferation was primarily thought to occur after rupture in response to the wound, allowing the ovary to heal the exposed area. Investigators have speculated that OSE cells continuously proliferated in culture eventually became transformed [
29-
32]. Recently, Burdette
et al. confirmed that ovulation dramatically increases the rate of OSE proliferation without simultaneously increasing programmed cell death, and potentially creating a system in which damaged cells would be retained and could contribute to the formation of transformed cells and the progress of ovarian cancer [
18].
In conclusion, these studies show significant effects of repeated OS on oocyte quality parameters based on follicle morphology and embryonic developmental competence and on the profiling of OSE, covering a range of intraovarian and postfertilization development outcomes. These studies suggest that repeated OS may affect the coordination of oogenesis with folliculogenesis and have potential interactions with ovarian cancer.
Higher Education Press and Springer-Verlag Berlin Heidelberg
PDF
(279KB)
