Amniotic Fluid Stem Cells Prevent Follicle Atresia and Rescue Fertility of Mice with Premature Ovarian Failure Induced by Chemotherapy

Chemotherapy used to treat cancer may cause irreversible premature ovarian failure (POF). Of late, amniotic fluid stem cells (AFSCs) provide a novel source for regenerative medicine because of their primitive stage, low immunogenicity, and easy accessibility. In this study, we isolated AFSCs from transgenic mice that ubiquitously express enhanced green fluorescence protein (EGFP). These AFSCs exhibited morphologies, immunophenotypes, and mesoderm trilineage differentiation potentials similar to mesenchymal stem cells (MSCs). Further, AFSCs proliferated faster than MSCs and expressed OCT4, a marker for pluripotency. To investigate their potential in recovering fertility in POF model, AFSCs were transplanted into the ovaries of mice with POF six weeks post induction using chemotherapeutic drugs, busulfan and cyclophosphamide. AFSCs could rescue the reproductive ability of mice with POF by preventing follicle atresia and sustaining the healthy follicles. Notably, the transplanted AFSCs did not differentiate into granulosa and germline cells in vivo. After one month, the decreased numbers of transplanted AFSCs accompanied with the reduced beneficial effects indicated that the therapeutic efficacy were directly from AFSCs. These findings demonstrated the therapeutic effects of AFSCs and suggested the promise of AFSCs for treating infertility and POF caused by chemotherapy.


Introduction
Chemotherapy (CTx) is commonly used for treating various malignancies and can improve the survival rate for cancer patients [1]. A potential adverse effect of chemotherapy is to decrease the number of ovarian follicles, disrupt the menstrual cycle, induce infertility, and cause irreversible premature ovarian failure (POF) [2][3][4]. In addition, early onset of menopause can result in the reduction in the quality of life. POF is associated with risks of cardiovascular disease [5], osteoporosis [6], and psychiatric diseases such as depression [7]. A recent study showed that bone marrow transplantation can rescue the fertility of mice with CTx [8]. Subsequently, more studies proved that transplantation of mesenchymal stem cells (MSCs) could restore partial ovarian function in a mouse model of CTx-induced POF [9,10]. These findings suggested the potential of MSCs to treat ovarian failure in humans, but its cellular and molecular mechanisms are still unknown. Furthermore, the proliferative capacity of MSCs in vitro is limited, and its acquisition is highly invasive [11]; therefore, an alternative cell source in clinical POF applications may be preferred.
To address this issue, we reasoned that amniotic fluid stem cells (AFSCs) may serve as an alternative option because they are readily harvested from amniotic fluid, undergo rapid self-renewal, maintain a normal karyotype in long-term culture [12], and tally with all the properties defined in the minimum criteria for MSCs [13][14][15][16]. Compared with other MSCs, AFSCs are easy to harvest, have a shorter doubling time and express the embryonic stem cells (ESCs) marker OCT4 [17], whereas unlike ESCs and induced pluripotent stem cells, they do not develop into teratomas in vivo [18]. Further, evidence indicated that transplanted allogeneic AFSCs have low immunogenicity and can escape immune rejection [19]. The potential advantages of AFSCs for treating POF are further suggested by their use in the studies of regenerative medicine to repair damaged tissues or organs, such as muscle [14], kidney [20], lung [21], or heart [22].
In this study, we evaluated the efficacy of AFSCs in a mouse model of POF and investigate the potential mechanism of this therapeutic effect. We established and characterized AFSCs from transgenic mice ubiquitously expressing enhanced green fluorescence protein (EGFP), which enables us to trace the fate of AFSCs after transplantation. To evaluate whether AFSCs can rescue the female fertility, we transplanted AFSCs expressing EGFP (EGFP-AFSCs) into the ovaries of mice with POF, assessed ovarian function and traced the fate of transplanted AFSCs within the ovaries of recipients.

Animals and Ethics Statement
ICR mice aged 4-6 weeks were purchased from the Laboratory Animal Center of Medical College of National Taiwan University (Taipei, Taiwan). EGFP transgenic mice that ubiquitously express EGFP were as previously described [23]. In brief, EGFPexpressing transgenic mouse lines were created by pronuclear microinjection of ICR strain zygotes with a construct (pCX-EGFP) that expresses EGFP under control of the b-actin promoter [24][25][26]. Mice were housed under a 14-10 h light-dark cycle at 25uC62uC with food and water provided ad libitum. The Institutional Animal Care and Use Committee of National Taiwan University approved the experimental protocol (NTU-102-EL-10).

Isolation and Culture of Bone Marrow MSCs
Bone marrow MSCs (BMMSCs) were isolated using a published method [29]. Femurs and tibias were isolated from euthanized ICR mice, and scissors were used to remove epiphyses. The marrow cells were harvested using 23-and 26-gauge hypodermic needles from the diaphysis of femurs and tibias, respectively, followed by flushing with MEM-a supplemented with 3.7 mg/mL NaHCO 3 , 10% FBS, 100 IU/mL penicillin and 100 mg/mL streptomycin. Cells (2610 5 cells/cm 2 ) were added to 60-cm 2 culture dishes and incubated as described above. Non-adherent cells were removed by changing the medium every three days. Cells (70% confluent) were removed from the dishes by incubation at 37uC with 0.25% trypsin/EDTA for 5 min and were purified using transient lower-density plastic adherence [23].

Cell Proliferation
To assess the proliferative ability of EGFP-AFSCs, we estimated the population doubling time (PDT) according to the equation: TD = t p log2/(logNt2logNo), where Nt and No are the numbers of cells harvested and inoculated, respectively, and t is the time of culture (in hours). Cell viability was determined using the trypan blue (#15250, Gibco) exclusion test, and the cells were counted using a hemocytometer [30].

Mouse Model of CTx-Induced POF
Six-week-old ICR female mice were administered busulfan (20 mg/kg) (#B2635) and cyclophosphamide (200 mg/kg) (#C0768) dissolved in dimethyl sulfoxide (DMSO) (#D2650) or DMSO only using a single intraperitoneal injection. All chemicals were purchased from Sigma-Aldrich. Mice were observed for 6 weeks. Mice treated with busulfan/cyclophosphamide or DMSO only are designated CTx-and non-CTx-mice, respectively.

Cell Transplantation
After six weeks of the POF mouse model established successfully, EGFP-AFSCs were transplanted into the ovaries of CTxmice using a modification of a published study [31]. In brief, a suspension of 5610 5 EGFP-AFSCs in 5 mL of sterilized PBS was injected through a glass pipette (100-mm tip) into each ovary of mice anesthetized with inhaled isoflurane (#B506, Abbott, Chicago, IL). CTx-mice with EGFP-AFSCs transplantation or PBS injection only are designated CTx-AFSCs-and CTx-salinemice, respectively.

Ovarian Morphology and Follicle Counts
The number of follicles in each ovary was estimated according to a modified published protocol [32]. Ovaries were fixed with 4% PFA at 4uC overnight, embedded in paraffin, serially sectioned (8 mm), and stained with hematoxylin-eosin. The number of primordial, primary, secondary, antral, or atretic follicles in every randomly selected tenth section was recorded. The different categories of follicles were determined as described previously [33]. Only follicles containing an oocyte with a clearly visible nucleus were scored.

Vaginal Cytology
To validate the reproduction function, the number of estrous cycles will be analyzed by vaginal smear assay every day after one week post AFSCs transplantation and for 14 consecutive days of test. Vaginal cells were collected using a sterile cotton swab, smeared on glass microscope slides, and stained with 0.5% (w/v) methylene blue (#M9140, Sigma-Aldrich), and the cells were classified using a standard criteria [34]. An estrus cycle was defined as the period between two successive proestrus phases (proestrus = day 1).

Mating Trial
Mating trial was initiated after one week post AFSCs transplantation and continued for 24 days. One fertile male were housed with one female mouse from each experimental group. The presence of a copulatory plug was defined as successful mating. Males were randomly rotated among cages after each pregnancy, and the number of offspring per litter was recorded.

Statistical Analysis
Three biological replicates were performed, and all values represent the mean 6 standard error of the mean (S.E.M). An unpaired Student t-test and Tukey's multiple comparison test were used to compare the data for two groups and three groups, respectively. P,0.05 was defined as a statistically significant difference. Log-rank (Mantel-Cox) test was used to compare the percentage of successful mating for three groups and the P,0.01 was defined as a statistically significant difference.
To further evaluate the proliferative ability of AFSCs, we performed growth curve and PDT assay compared to BMMSCs. We found that AFSCs showed higher cell population growth with shorter PDTs than BMMSCs (31.2261.10 h and 34.4160.73 in AFSCs and BMMSCs, respectively) ( Figure 1G). In accordance with previous studies, we confirmed that most of cultured AFSCs expressed OCT4 in the nucleus by immunocytochemistry staining (Figure 1H-J). These findings illustrated that the AFSCs we established exhibited higher proliferation ability than BMMSCs and expressed OCT4.

Effect of CTx on Ovarian Function and Induced POF
The ovaries of CTx-mice were atrophic compared with those of non-CTx-mice and contained fewer developing follicles at the sixth week post CTx ( Figure S1). Further, primordial follicles were significantly fewer in the ovaries of CTx-mice and almost disappeared after six weeks post CTx (Figure 2A). The numbers of developing and total healthy follicles were also significantly reduced in the ovaries of CTx-mice, particularly after six weeks ( Figure 2B-E). In contrast, the numbers of atretic follicles were not significantly different between the ovaries of CTx-and non-CTxmice ( Figure 2F).

AFSCs Prevent Follicular Atresia and Sustained the Development of Follicles in the Ovaries of CTx-mice
To assess the therapeutic effect of AFSCs, we transplanted EGFP-AFSCs directly into bilateral ovaries of mice with POF six weeks after CTx. Although there were few primordial follicles in the ovaries of CTx-saline-and CTx-AFSCs-mice, during the course of the experiment, the numbers of primordial follicles were significantly higher in CTx-AFSCs-mice during the first four weeks post-transplantation (wpt) ( Figure 3A). The results were similar for the numbers of primary and secondary follicles among the three groups ( Figure 3B and C). The number of antral follicles of CTx-AFSCs-mice was significantly higher compared with that of CTx-saline-mice, whereas lower compared with that of non-CTx-mice at all times ( Figure 3D). The number of total healthy follicles in CTx-AFSCs-mice was higher compared with CTxsaline-mice within 4 wpt and was significantly lower compared with that of non-CTx-mice at all times ( Figure 3E). Injection of EGFP-AFSCs significantly decreased the number of atretic follicles up to 4 wpt. In contrast, there was no significant difference between non-CTx-and CTx-saline-mice at all times ( Figure 3F). There was no significant difference between the sum of total healthy and atretic follicles of CTx-AFSCs-and CTx-saline-mice ( Figure 3G). These results indicated that AFSCs could transiently prevent follicular atresia and sustain the follicular development in the ovaries of mice with POF.

AFSCs Restore the Fertility of Mice with POF
The number of estrous cycles of CTx-AFSCs-mice (3.460.22) was significantly higher compared with that of CTx-mice (0.460.22) but not significantly different compared with that of non-CTx-mice (460.28) ( Figure 4A). Although the CTx-AFSCsmice reached 100% successful mating rate significantly slower than non-CTx control group (P,0.01), the CTx-saline-mice reached only 40% successful mating in the 24-day test period and was significantly worse than CTx-AFSCs-mice (P,0.01) (Figure 4B). Moreover, consistent with the numbers of antral follicles, the average number of litters produced by CTx-AFSCs-mice (6.960.85) was significantly higher compared with those of CTxmice (0) and lower compared with those of non-CTx-mice (13.261.09) ( Figure 4C).

EGFP-AFSCs do not Contribute to the Germline in Ovaries of CTx-Mice
The progeny of CTx-AFSCs-mice appeared normal ( Figure 4D and E), and none expressed EGFP ( Figure 4F), indicating that offspring were derived from the recipient's germline (n = 69). Therefore, we tracked the fate of the transplanted EGFP-AFSCs within the ovaries of recipients. EGFP-AFSCs were detected in ovarian tissue at 1 wpt and persisted for at least four weeks ( Figure 5A-D). Neither the germ cell marker VASA [35] ( Figure 5D) nor the granulosa cell marker follicle stimulating hormone receptor (FSHR) [36] (data not shown) was detectably expressed by EGFP-AFSCs. Moreover, the level of the mRNA encoding EGFP in the ovaries of CTx-AFSCs-mice decreased gradually after transplantation, and little was detected after 5 wpt ( Figure 5E).

Discussion
Cancer CTx using busulfan/cyclophosphamide are designed to destroy highly proliferating cells by inducing cross-linking of DNA [37], which affect mitochondria and result in the activation of apoptotic pathway [38]. Accordingly, it is widely believed that chemotherapeutic drugs may massively eliminate granulosa cells [39], which are required for oocyte survival and follicle development [40]. The ensuing damage to growing follicles by CTx increases recruitment of primordial follicles and, in turn, the number of all types of follicles reduced [41]. Consistent with this hypothesis and previous studies [42,43], our results showed that CTx treatment significantly reduced various types of follicles and number of total healthy follicles (Figure 2A-E). Nonetheless, although CTx largely diminished number of total healthy follicles, the number of atretic follicles was not significantly different at any point during the 6 weeks observation ( Figure 2F). According to the previous studies, chemotherapeutic drugs significantly increased follicle atresia predominantly within 72 h after the treatment and the number of atretic follicle quickly returned to the basal level within one week after the treatment. Subsequently, the degenerative follicles were removed from the ovary within a 3-day period [32,44]. This should be the reason of the discordance in the observed numbers of follicle loss versus atresia. Consequently, the exhaustion of primordial follicles and continued follicular atresia severely impaired the ovarian function and resulted in POF after six weeks of CTx.
In this study, we demonstrated that AFSCs increased healthy ovarian follicles and restored fertility in a mouse model system of POF induced by CTx. In principle, AFSCs may sustain the number of healthy ovarian follicles in POF mice either by promoting de novo folliculogenesis or by inhibiting follicular CTx-AFSCs-mice. Data represent the mean 6 standard error of mean (S.E.M.) of three experiments. Lower case letters (a, b, and c) represent significant differences (P,0.05) among groups. doi:10.1371/journal.pone.0106538.g003 atresia or both of these two mechanisms. Folliculogenesis is the process that the primitive oocytes, which originated from primordial germ cells [45], are then surrounded by pregranulosa cells in fetal ovary to form primordial follicles during embryogenesis [46]. Previous study reported that AFSCs expressed oocyte marker in vitro [47], and therefore it might be possible that the transplanted AFSCs have the potential to differentiate into primitive oocyte and participate in the following folliculogenesis within ovaries of POF mice. However, we found no progeny of CTx-AFSCs-mice were derived from the EGFP-AFSCs and none expressed EGFP. Moreover, transplanted EGFP-AFSCs were distributed within the ovarian interstitium rather than follicles and express neither a marker of differentiated germ cells ( Figure 5D) nor of granulosa cells (data not shown). These results indicated that the transplanted AFSCs per se did not participate in the folliculogenesis.
We further showed a significant decrease in the number of atretic follicles in the ovaries of CTx-AFSCs-mice that persisted for four weeks (Figure 3F), which is in correlation with the existence period of EGFP-AFSCs ( Figure 5E). Notably, the numbers of total healthy and atretic follicles were not significantly different between CTx-AFSCs-mice and mice that were not engrafted ( Figure 3G). These data strongly suggested that AFSCs exerted therapeutic effect on increasing the number of healthy ovarian follicles by preventing follicular atresia.
According to previous research, follicular atresia is predominantly mediated by the apoptosis of follicular cells, especially granulosa cells which are necessary for follicular development [48][49][50]. Nonetheless, we found no evidence that transplanted AFSCs directly differentiated into granulosa cells to replace damaged cells within ovary of POF mice. It has been hypothesized that the restorative effects of MSCs transplantation on impaired ovarian function resulted from secretion of trophic factors, which are beneficial to follicle growth [9,10]. Previous studies reported that AFSCs can secrete TGF-b [51], vascular endothelial growth factor [52], and glia cell-derived neurotrophic factor [53], which are required for follicular development [54,55], and can inhibit follicular cell apoptosis [56] and follicle atresia [57]. Taken together, we infer that the favorable effects of AFSCs on CTxdamaged ovary is derived from the paracrine secretion of certain beneficial factors, which support ovarian follicle survival. Since ovarian granulosa cells are essential to sustain follicle survival, it is likely that the interaction between AFSCs and granulosa cells plays a key role in this favorable effect. To address this issue, establishment of the granulosa cells and AFSCs in vitro co-culture system for screening the candidate factors shall be performed in the future.
In this study, we performed AFSCs transplantation at six weeks post CTx that the ovaries have been impaired seriously, when compared to the healthy control, to evaluate the therapeutic effect of AFSCs apply in POF. Our findings indicated that the transplanted AFSCs rescued the fertility by preventing atresia of the remaining follicle. Since the CTx induced follicular atresia peaked within the first week after CTx and the continued decrease in the primordial follicle pool, it is reasonable to expect an attenuated therapeutic effect with the delay in timing of transplantation. In the clinical applications, it might be favorable to administer the AFSCs closer to or even prior to CTx to better preserve follicle reserve from destruction and prevent POF from occurrence so that the maximal efficacy shall be achieved.
Collectively, our findings demonstrated that AFSCs may prevent follicle atresia in the ovaries of mice with POF and restore fertility. These findings suggest the potential of AFSCs to treat infertility. Figure S1 Appearance and histological analyses of ovaries of mice with CTx. (A) The appearance of ovary between non-chemotherapy (CTx)-mouse (Con, left) and CTxmouse (CTx, right) showed the severely atrophic ovary in CTxmice followed by CTx for six weeks. Scale bar = 1 mm. (B) The weight of ovaries of CTx-mice compared with those of non-CTxmice (Con) (n = 3). Data represent the mean 6 standard error of mean (S.E.M.) *P,0.05. (C, D) Hematoxylin-eosin staining of ovaries of non-CTx-mice (C) and CTx-mice (D). Arrows indicate developing follicles. Scale bar = 100 mm. (TIF)