Ablation of beta subunit of protein kinase CK2 in mouse oocytes causes follicle atresia and premature ovarian failure

Premature ovarian failure (POF), a major cause of female infertility, is a complex disorder, but the molecular mechanisms underlying the disorder are only poorly understood. Here we report that protein kinase CK2 contributes to maintaining follicular survival through PI3K/AKT pathway and DNA damage response pathway. Targeted deletion of CK2β in mouse oocytes from the primordial follicle stage resulted in female infertility, which was attributed to POF incurring by massive follicle atresia. Downregulated PI3K/AKT signaling was found after CK2β deletion, indicated by reduced level of phosphorylated AKT (S473, T308, and S129) and altered AKT targets related to cell survival. Further studies discovered that CK2β-deficient oocytes showed enhanced γH2AX signals, indicative of accumulative unrepaired DSBs, which activated CHK2-dependant p53 and p63 signaling. The suppressed PI3K/AKT signaling and failed DNA damage response signaling probably contribute to large-scale oocyte loss and eventually POF. Our findings provide important new clues for elucidating the mechanisms underlying follicle atresia and POF.


Introduction
Premature ovarian failure (POF) refers to amenorrhea in women of less than 40 years of age accompanied by elevated menopausal levels of serum gonadotropins (folliclestimulating hormone, FSH > 40 IU/l) and decreased estrogen 1 . POF is an ovarian dysfunction characterized by premature depletion of ovarian follicles in~1% of women under the age of 40 years and 0.1% under the age of 30 years 2 , which usually leads to female infertility. The causes of POF vary and are complex, and it includes genetic aberrations [3][4][5] , autoimmune ovarian damage 6,7 , therapeutic interventions such as radiotherapy 8 and chemotherapy 9 , with genetic factors being the main causes. Accumulating genes in the X chromosome such as FMR1 4,10 , FMR2 11 , BMP15 [12][13][14] and genes in the autosome such as FOXL2 15,16 , FSHR 17 , LH receptor 18 , and inhibin A 19,20 are known to be involved in POF; however, for many years, the underlying mechanisms of POF have largely remained unknown.
CK2α −/− embryos die in mid-gestation, with defects in heart and neural tube 50 . CK2α' −/− females show normal fertility, but males are infertile. Male mice lacking CK2α' show extensive germ cell apoptosis characterized by nuclear abnormalities ranging from spermatogonia to early spermatids 51,52 . CK2β −/− mice die shortly after implantation with no signs of apoptosis but reduced cell proliferation. CK2β −/− blastocysts cannot develop an inner cell mass in vitro 53 . Zygote-specific knockout of CK2β is destructive for embryonic stem cells and primary embryonic fibroblasts 53 . The above studies demonstrate that CK2β is indispensable for cell survival. However, the roles of CK2β in folliculogenesis/oogenesis are largely unknown. Here we targeted CK2β for deletion in oocytes from the primordial follicle stage by crossing Ck2β fl/fl mice with Gdf9-Cre mice. We found that CK2β was essential for female fertility and loss of CK2β caused ovarian follicle atresia and POF.

Results
Oocyte-specific deletion of Ck2β gene causes mice infertility Toward determining the potential roles of CK2β in the female reproductive system, expression patterns of CK2β in ovaries and oocytes at different developmental stages were assessed through immunoblotting and immunohistochemistry. Immunoblotting revealed that CK2β was stably and highly expressed in GV, GVBD, MI, and MII oocytes (Fig. 1a). Within the ovary, immunohistochemistry revealed that CK2β located in the nuclei of oocytes and granulosa cells from primordial follicles to antral follicles (Fig. 1b). These data suggest that CK2β potentially functions in folliculogenesis/oogenesis.
To confirm our hypothesis, we generated oocytespecific CK2β mutant mice by crossing Ck2β fl mice in which exon I-II were targeted with transgenic mice expressing Gdf9 promotor-driven Cre recombinase (Supplementary Figure1). In Gdf9-Cre mice, Cre was expressed from primordial to later follicular stages. Histological analysis of Ck2β fl/fl ;GCre + mouse ovaries showed loss of CK2β localization in nuclei of oocytes, indicating functional deletion of CK2β (Fig. 1c).
To observe the effect of oocyte-specific deletion of CK2β on fertility, a breeding assay was conducted by mating Ck2β fl/fl or Ck2β fl/fl ;GCre + female mice with wildtype males of tested fertility for 6 months. Continuous breeding assessment indicated that Ck2β fl/fl ;GCre + females were completely infertile (Fig. 2a). To determine whether the infertility was due to ovarian dysfunction and consequential functional oocyte loss, we first measured the size of ovaries from Ck2β fl/fl and Ck2β fl/ fl ;GCre + mice. As shown in Fig. 2b, the size of the ovaries of Ck2β fl/fl mice continued to increase from week 3 to week 8, with a mean ovarian weight ratio of 0.0316%, 0.0326%, 0.0424%, and 0.0598% corresponding to week 3, 4, 6, and 8, respectively. Compared to the control, the ovary size of Ck2β fl/fl ;GCre + mice decreased slightly in week 3 and decreased sharply from week 4 to week 8, with mean ovarian weight ratio of only 0.0221%, 0.0196%, 0.0124%, and 0.0071%. These data revealed that CK2β deletion resulted in atrophy of ovaries in mice.
To clarify the cause of ovarian atrophy, we next examined the morphology of ovaries from both Ck2β fl/fl and Ck2β fl/fl ;GCre + mice. At 3 weeks of age, histological assessment revealed that Ck2β fl/fl mice showed normal ovarian morphology characterized by the presence of primordial and activated follicles including primary, secondary, and antral follicles (Fig. 2c, c'). All of these structures were also found in the Ck2β fl/fl ;GCre + mouse ovaries. Although a few follicles underwent atresia, the ovaries looked healthy on the whole (Fig. 2d, d'). At 4 weeks of age, all types of follicles could be found in both GCre + female for 6 months. At least five mice of each genotype were used in this assay. b Ovary weight to body weight ratio of Ck2β fl/fl and Ck2β fl/fl ; GCre + mice at 3, 4, 6, and 8 weeks of age after birth. For each time point, at least three mice of each genotype were used for analysis. Data are presented as the mean ± SEM. P < 0.05(*), 0.01(**) or 0.001(***). c-j Representative ovarian histology of Ck2β fl/fl and Ck2β fl/fl ;GCre + mice of 3, 4, 6, and 8 weeks of age, respectively. Images c'-j' correspond to the partial magnification of images c-j. Yellow arrowheads in f', h', and j' indicate atretic follicles. For each time point, at least three mice of each genotype were used for analysis. Scale bar: 100 μm Ck2β fl/fl mice (Fig. 2e, e') and Ck2β fl/fl ;GCre + mice (Fig. 2f, f'). However, most follicles in the ovaries of Ck2β fl/fl ; GCre + mice showed signs of atresia ( Fig. 2f', yellow arrows) in contrast to control ovaries that contained substantial healthy-looking follicles (Fig. 2e'). At 6 weeks, the time of sexual maturity, massive atretic follicles ( Fig. 2h', yellow arrows) appeared in the ovaries of Ck2β fl/ fl ;GCre + mice (Fig. 2h, h') compared to Ck2β fl/fl mice ( Fig. 2g, g'). Although most atretic follicles maintained follicular structures, the oocytes in atretic follicles were eliminated and the space was filled with granulosa cells (Fig. 2h, h'). By 8 weeks of age, in contrast to Ck2β fl/fl ovaries (Fig. 2I, i'), almost all types of follicles had been depleted in Ck2β fl/fl ;GCre + ovaries, and it was difficult to find follicular structures in the ovaries of Ck2β fl/fl ;GCre + mice ( Fig. 2j, j').
Quantitative analysis revealed that the numbers of primary, secondary, and antral follicles in the ovaries of Ck2β fl/fl ; GCre + mice were similar to those of Ck2β fl/fl mice at 3 and 4 weeks, but the number of primordial follicles within the ovaries of Ck2β fl/fl ;GCre + mice was markedly reduced as compared to Ck2β fl/fl mice (Fig. 3a, b). At 6 weeks, in addition to primordial follicles, significant differences were also observed in the secondary and antral follicles (Fig. 3c). By 8 weeks of age, all types of follicles in ovaries of Ck2β fl/fl ; GCre + mice were significantly decreased (Fig. 3d). In general, the primordial follicles reduction occurred at 3 weeks of age and continued decreasing until depletion of the primordial follicle pool at young adulthood (8 weeks) in Ck2β fl/fl ;GCre + mice compared to Ck2β fl/fl mice (Fig. 3e). The activated follicle reduction appeared at the time of onset of sexual maturity (4 weeks) and displayed a similar decreased trend from week 4 to week 6 for primordial follicles in the ovaries of Ck2β fl/fl ;GCre + mice (Fig. 3f). This phenotype resembles premature ovarian failure (POF) in humans.
The histological analysis indicated that absence of CK2β in oocytes caused follicular atresia and POF. To confirm these observations, we performed immunohistochemistry of the germ cell marker MVH on 3-and 8-week-old ovarian sections. As shown in Fig. 4a, in ovaries of Ck2β fl/fl ;GCre + mice at 3 weeks of age, most follicles including primordial, primary, secondary, and antral follicles showed MVH-positive staining, suggesting that these follicles were healthy. However, at 8 weeks of age, there were few MVH-positive primordial or primary follicles scattering in the cortical region in ovaries of Ck2β fl/fl ;GCre + mice, which was symptomatic of POF. TUNEL assay on ovarian sections showed that increased granulosa cell apoptosis occurred in ovaries of Ck2β fl/fl ;GCre + mice at 4 weeks of age compared to ovaries in Ck2β fl/fl mice, which was caused by growing follicle atresia (Fig. 4b). The above data demonstrate that the accelerated demise of primordial follicles and defective survival of growing follicles may be responsible for infertility of Ck2β fl/fl ; GCre + female mice.

Ck2β deletion causes PI3K/AKT signaling hypoactivation
According to the above-mentioned research, Ck2β fl/fl ; GCre + mice displayed defects in primordial follicle survival. Accumulating literature indicates that PI3K/AKT signaling plays a vital role in regulating the survival of primordial follicles during their long dormancy [54][55][56] . In view of the involvement of CK2 in the PI3K/AKT pathway at a cellular level, we first focus our attention on the PI3K/ AKT pathway. Accordingly, we performed immunoblotting analysis using ovaries from Ck2β fl/fl and Ck2β fl/fl ; GCre + mice at 2 weeks of age. It showed that the levels of phosphorylated AKT (S473 and S129) decreased slightly, whereas phosphorylated AKT (T308) was markedly reduced (Fig. 5a); when considering the elevated total level of AKT1/2/3 protein (Fig. 5a), the relative level of phosphorylated AKT (S473, T308, and S129) was decreased more obviously in CK2β mutant ovaries. PI3K/ AKT regulator PTEN was also detected and it showed decreased expression (Fig. 5a). Subsequently, the downstream of PI3K/AKT pathway was detected. We first examined the TSC2/mTOR signaling which was critical to oocyte survival. The results showed that the activity of mTOR/rpS6 signaling pathway was enhanced in mutant ovaries, as indicated by elevated levels of phosphorylated mTOR (S2448), phosphorylated S6K (T389), and phosphorylated rpS6 (Ser240/244) (Fig. 5b). However, the levels of phosphorylated TSC2 (S1387) displayed no difference between mutant and control ovaries (Fig. 5b). These results were inconsistent when considering that mTOR was negatively regulated by TSC2. The upregulated mTOR/S6K/rpSK signaling may result from feedback effects on defective follicle survival. Afterward, other AKT substrates contributing to cell survival were further detected. As the first reported AKT substrate which is important to cell survival, two isoforms of glycogen synthase kinase 3 (GSK3), GSK3α and GSK3β, were detected. The results showed that the level of GSK3α expressed in ovaries was higher than GSK3β (Fig. 5c), and the expression levels of both phosphorylated GSK3α/β (S21/ 9) and total GSK3α/β protein displayed no variation in control and mutant mice (Fig. 5c). We next detected FOXO proteins which are also important cell survivalrelated AKT substrates. The examination of the levels of phosphorylated FOXO1 (T24)/FOXO3a (T32) found that they were significantly decreased (Fig. 5c), while the levels of total FOXO1 protein did not show obvious changes (Fig. 5c).

Ck2β depletion results in accumulated DNA damage in oocytes
Double-strand breaks derived from unrepaired meiotic or environmental stress could result in oocyte elimination and female infertility through CHK2-dependent activation of p53 or p63 57 . Considering that CK2 is involved in the DNA damage response pathway 32,34-36 , we wonder whether Ck2β −/− oocyte depletion relates to this pathway. Thus, we first performed immunoblotting using ovary lysates from Ck2β fl/fl and Ck2β fl/fl ;GCre + mice at 2 weeks of age. As shown in Fig. 6a, the levels of phosphorylated CHK2 (T68) were slightly increased and γH2AX was significantly upregulated in ovaries of Ck2β fl/fl ;GCre + mice, whereas the levels of p63 and phosphorylated p53 (S15) showed no difference in ovaries of Ck2β fl/fl and Ck2β fl/fl ;GCre + mice. However, by 4 weeks of age, the level of p63 was downregulated while phosphorylated p53 (S15) was upregulated in ovaries of Ck2β fl/fl ;GCre + mice (Fig. 6b), consistent with the antagonizing relationship between p53 and p63 57 . Moreover, we confirmed our findings using immunofluorescence analysis. As indicated in Fig. 6c, the γH2AX signals in small oocytes from Ck2β fl/fl ;GCre + mice at 2 weeks of age were remarkably enhanced. Taken together, these data suggest that CK2β depletion causes DSBs accumulation and failed activation of the DNA damage response pathway.
Oocyte-specific deletion of CK2β causes a striking reduction in CK2α but not CK2α' expression To explore the forms of CK2 functions in the mouse ovary, immunoblotting was carried out to detect protein levels of CK2α, CK2α', and CK2β in ovaries of Ck2β fl/fl and Ck2β fl/fl ;GCre + mice (Fig. 7). As expected, the level of CK2β protein was dramatically reduced in ovary extracts prepared from Ck2β fl/fl ;GCre + mice compared with Ck2β fl/fl mice. The low level of CK2β protein that was detected in the ovary extracts collected from Ck2β fl/fl ; GCre + mice likely came from granulosa cells in which CK2β was not deleted. Meanwhile, the level of CK2α protein was significantly downregulated in ovaries of Ck2β fl/fl ;GCre + mice, while the levels of CK2α' protein in ovaries of Ck2β fl/fl ;GCre + mice showed no variation compared with Ck2β fl/fl mice. Immunoblotting analysis of phospho-CK2 substrate using ovaries from control and mutant mice found that CK2 activity was largely reduced (Fig. 7b). These data suggest that CK2 is presumably functioning in oocytes in the forms of α 2 β 2 and the reduced expression of CK2α protein in ovaries of Ck2β fl/fl ; GCre + mice is probably due to degradation resulting from decreased stability of CK2α protein without CK2β.

Discussion
In humans, the primordial germ cells (PGCs) migrate to gonadal ridges and are enclosed by pregranulosa cells to form primordial follicles. Most ovarian primordial follicles ;GCre + at 2 weeks after birth. The ovary lysates were collected at least from three mice of each genotype and immunoblotted for p-AKT (S473), p-AKT (T308), p-AKT (S129), AKT1/2/3, PTEN, p-TSC2 (S1387), TSC2, p-mTOR (S2448), p-S6K (T389), p-rpS6 (S240/244), p-GSKα/β (S21/9), GSKα/β, p-FOXO1 (T24)/FOXO3a (T32), FOXO1, and β-actin. Levels of β-action were used as internal control. Each experiment was repeated at least 2-3 times. Molecular mass is given in kilo Daltons are maintained in a quiescent state, providing as a reserve for a woman's reproductive life 58 . Premature depletion of the ovarian reserve incurs cessation of ovarian function, resulting in POF 59 . Elucidating the mechanisms that control the dormancy and survival of primordial follicles is critical for understanding of ovarian biology. In this study, using Gdf9 promotor-driven Cre recombinase, we successfully deleted CK2β in oocytes from the primordial follicle stage, which facilitated investigation on the roles of CK2β in folliculogenesis. We found that CK2β mutant females showed defective follicular survival and sterility.
Morphological observation revealed that the ovary size of CK2β mutant mice started to decrease from 3 weeks after birth and eventually became reduced to about 1/4 of that in control mice at 8 weeks, which was consistent with histological observations and follicle counts. At 3 weeks, most follicles looked healthy but the primordial follicles in ovaries of CK2β mutant mice were markedly reduced compared to control mice. Starting at 4 weeks, the ovaries of CK2β mutant mice showed increasing numbers of atretic follicles, which were eliminated quickly, resulting in the decrease of the number of activated follicles.
Previous researches reveal that PI3K signaling is critical to control the survival of primordial follicles, since suppressed or elevated PI3K/AKT signaling leads to premature depletion of follicles, causing POF [54][55][56] . In this study, we systematically analyze the PI3K/AKT signaling. Immunoblotting results showed that Ck2β depletion caused downregulation of phosphorylated AKT (S473, Fig. 6 Impaired follicle survival in Ck2β fl/fl ;GCre + mice involved in DNA damage response. a Immunoblotting analysis of DNA damage response signaling in ovaries of Ck2β fl/fl and Ck2β fl/fl ;GCre + mice at 2 weeks after birth. The ovary lysates were collected at least from three mice of each genotype and immunoblotted for γH2AX, H2AX, p-CHK2 (T68), CHK2, p63, p-p53 (S15), p53 and β-actin. Level of β-actin was detected as internal control. Each experiment was repeated at least 2-3 times. Molecular mass is given in kilo Daltons. b Immunoblotting analysis of the expression of p53 and p63 in ovaries of Ck2β fl/fl and Ck2β fl/fl ;GCre + at 4 weeks after birth. The ovary lysates were obtained from at least three mice of each genotype and immunoblotted for p63, p-p53 (S15), p53 and β-actin. Level of β-actin was used as internal control. Each experiment was repeated at least three times. Molecular mass is given in kilo Daltons. c MVH and γH2AX immunofluorescent staining of 2-week-old ovarian sections from Ck2β fl/fl and Ck2β fl/ fl ;GCre + mice. Green: γH2AX; Red: MVH; Blue: DAPI. At least three mice of each genotype were used in this assay. Scale bar: 50 μm T308, and S129). Previous studies find that CK2 can phosphorylate AKT/PKB at Ser129 and such phosphorylation of AKT prevents AKT Thr308 dephosphorylation and enhances the catalytic activity of AKT/PKB 24,30,37 . Previous studies also show that CK2 can phosphorylate PTEN at several sites and such phosphorylation prevents PTEN degradation and inhibits its activity [38][39][40] . Consequently, deletion of CK2β causes low stability and high activity of PTEN, leading to reduced level of PTEN protein and downregulated AKT phosphorylation.
Studies of targets of AKT found that the activity of TSC2 was not changed but its downstream pathway mTOR/S6K/ rpSK signaling was significantly enhanced in CK2β mutant ovaries. These findings are in conflict with previous reports which suggest that elevated mTOR/S6K/rpSK signaling is responsible for oocyte growth and follicular activation and suppressed mTOR/S6K/rpSK signaling leads to loss of primordial follicles 54,55 . The upregulated mTOR/S6K/ rpSK signaling may result from feedback effects to defective follicle survival. Further studies of other targets of AKT found that the levels of phosphorylated GSK3α/β (S21/9) were not changed. However, phosphorylated FOXO1 (T24)/FOXO3a (T32) were downregulated in CK2β mutant mice. CK2β deletion downregulates activated AKT, which in turn inhibits FOXO phosphorylation and keeps FOXO1 staying in nucleus 60,61 . Unphosphorylated FOXO proteins trigger expression of genes that are crucial for the induction of apoptosis, such as FASL and BIM 62 . From above results, CK2β regulates follicular survival at least partly via PI3K/AKT/FOXO pathway, although details remain to be elucidated.
The oocytes derived from primordial follicle are arrested at the diplotene stage of prophase of first meiotic division for months, even years, after birth depending on species. The oocytes at the diplotene stage have finished synapsis that relies on homologous recombination, a highfidelity DNA double-strand breaks (DSBs) repair process. Any homolog synapsis or DSB repair errors would prompt DNA damage checkpoints to eliminate defective meiotic oocytes, which is mediated by the CHK2-p53/p63 pathway 57,63 . Considering the involvement of CK2 in the DNA damage response pathway in other species 32,34-36 , we then studied this pathway in CK2β mutant ovaries. In our study, oocyte-specific deletion of CK2β from the primordial follicle stage caused elevated γH2AX and phosphorylated CHK2 (T68) signals in ovaries and elevated γH2AX signals in small oocytes from mice at 2 weeks of age, indicating accumulated DSBs. By 4 weeks of age, the DNA damage response pathway was significantly elevated, indicated by the upregulated level of phosphorylated p53 (S15) in ovaries of Ck2β fl/fl ;GCre + mice. Accordingly, the follicular atresia at least early follicle atresia may be partly dependent on DNA damage response pathway.
In this study, as expected, the level of CK2β protein was significantly reduced in CK2β mutant ovaries compared to controls. However, the level of CK2α also showed an obvious decrease despite of no difference in the levels of CK2α' in CK2β mutant ovaries and control ovaries. This is consistent with a previous report in which CK2α' −/− females show normal fertility 51 . The above data reveal that CK2 presumably functions in oocytes in the forms of α 2 β 2 , but since CK2α −/− embryos die at the embryonic stage 50 , oocyte-specific knockdown of CK2α is necessary to validate the hypothesis.
In summary, we identified CK2β as a key protein safeguarding mouse follicle survival and fertility. Our data provide new insights into occurrence, diagnosis, and treatment of POF.

Immunoblotting
Ovary extracts were prepared using a homogenizer in RIPA buffer supplemented with protease and phosphatase inhibitor cocktail (Roche Diagnostics). After transient ultrasound, the ovary lysates were incubated on ice for 30 min and then centrifuged at 4°C, 12,000 rpm for 20 min. The supernatant was transferred to a new tube and equal volume loading buffer was added. After being boiled at 95°C for 10 min, the protein lysates were used for immunoblotting analysis. Immunoblotting was performed as described previously 65 . Briefly, the separated proteins in SDS-PAGE were electrically transferred to a polyvinylidene fluoride membrane. After incubation with primary and secondary antibodies, the membranes were scanned with Bio-Rad ChemiDoc XRS+.

Hematoxylin and eosin staining and quantification of ovarian follicles
Ovaries were dissected from Ck2β fl/fl and Ck2β fl/fl ; GCre + mice immediately after killing. The ovaries were fixed in 4% formaldehyde overnight at 4°C, dehydrated in an ethanol series and embedded in paraffin. Paraffinembedded ovaries were cut into sections of 8-μm thickness and mounted on glass slides. After 48°C overnight drying, the sections were deparaffinized in xylene, hydrated by a graded alcohol series and stained with hematoxylin and eosin for histological analyses. Ovarian primordial, primary, secondary, and antral follicles were counted in every fifth section of an ovary. Quantification of ovarian follicles was performed as previously reported 66,67 . In each section, follicles that contained oocytes with clearly visible nuclei were scored and the cumulative number of follicles were multiplied by a correction factor of 5 to represent the estimated number of total follicles in an ovary.

Immunohistochemistry and immunofluorescence
Ovaries used for immunostaining were fixed in 4% paraformaldehyde (pH 7.4) overnight at 4 ℃, dehydrated, and embedded in paraffin. Paraffin-embedded ovaries were cut into sections of 5-μm thickness. Then, the sections were deparaffinized, immersed in sodium citrate buffer (pH 6.0), and heated for 15 min in a microwave for antigen retrieval. After blocking with 5% donkey serum albumin, sections were incubated with primary antibodies at 4°C overnight. For immunohistochemistry, the sections were treated with 3% H 2 O 2 to eliminate internal peroxidase activity and incubated with an appropriate horseradish peroxidase (HRP)-conjugated secondary antibody. Finally, the signal of primary antibody was detected by the Vectastain ABC kit (Vector Laboratories, CA, USA) and the sections were counterstained with hematoxylin. Images were captured using a Nikon DS-Ri1 CCD camera. For immunofluorescence, the sections were incubated with an appropriate FITC-conjugated secondary antibody. The nuclei were stained with DAPI. Images were captured using a laser scanning confocal microscope (Zeiss 780 META).

TUNEL assay
TUNEL assay was carried out in accordance to the DeadEnd TM Fluorometric TUNEL System (Promega BioSciences, Madison, WI, USA). Images were captured using a laser scanning confocal microscope (Zeiss 780 META).

Breeding assay
In the breeding assay, 6-8 week-old Ck2β fl/fl and Ck2β fl/ fl ;GCre + female mice were mated to 8-week-old C57BL/6J wild-type male mice with known fertility. At least five mice of each genotype were used in this assay. For 6 months, the cages were monitored daily for recording the number of pups and litter size.

Statistical analysis
All experiments were performed at least three times. Paired two-tailed Student's t test was used for statistical analysis. Data were presented as mean ± SEM and P < 0.05 (*), 0.01(**), or 0.001(***) was considered statistically significant.