Melatonin improves age-induced fertility decline and attenuates ovarian mitochondrial oxidative stress in mice

Increasing evidence shows that melatonin protected against age-related mitochondrial oxidative damage. However, the protective effects of melatonin against ovarian aging has not been explored. Young Kunming females (aged 2–3 months) were fed with melatonin added to drinking water for 6 or 12 months (mo). We found that long-term (12 mo) melatonin treatment significantly reduced ovarian aging, as indicated by substantial increases in litter size, pool of follicles, and telomere length as well as oocyte quantity and quality. Melatonin treatment suppressed ovarian mitochondrial oxidative damage by decreasing mitochondrial reactive oxygen species (mROS) generation, inhibiting apoptosis, repressing collapse of mitochondrial membrane potential and preserving respiratory chain complex activities. Female mice fed with melatonin had enhanced mitochondrial antioxidant activities, thus reducing the risk of mitochondrial oxidative damage cause by free radicals. Notably, melatonin treatment enhanced SIRT3 activity but not the protein expression level, and increased the binding affinity of FoxO3a to the promoters of both superoxide dismutase 2 (SOD2) and catalase (CAT). In conclusion, melatonin exerted protection against aging-induced fertility decline and maintenance of mitochondrial redox balance.

Estimation of mitochondrial DNA (mtDNA) copy numbers. Oocyte manipulation techniques are described in more detail in Supplemental materials. A single GV or MII oocyte was added to a PCR tube with 20 μ L of lysis buffer (50 mM Tris, 0.1 mM EDTA, 100 μ g/mL Proteinase K and 0.5% Tween-20) and incubated at 55 °C for 30 min, then 95 °C for 10 min. To obtain purified plasmid standard DNA, PCR products were amplified with a mouse mtDNA-specific primer as previously described 18 , and were ligated into PMD-19T vector (Takara).
Immunofluorescence. For chromosome and spindle analysis, oocytes from each group were washed twice in PBS containing 0.1% polyvinylpyrrolidone (PBS-PVP) and then fixed with freshly prepared 4% paraformaldehyde on ice for 30 min, permeabilized with 0.1% Triton X-100 for 30 min and washed three times. After blocking in 1% Bull Serum Albumin (BSA)-supplemented PBS for 1 h, oocytes were incubated overnight at 4 °C with fluorescein isothiocyanate-conjugated α -tubulin antibody to visualize microtubules in the spindle. Chromosomes were stained with DAPI for 5 min. Embryos were harvested for the detection of Oct4-positive cells with an anti-Oct4 mouse monoclonal antibody (Santa Cruz, CA, USA). DNA fragmentation was assessed in blastocysts using a DeadEnd Fluorometric TUNEL System (Promega, Madison, WI), according to the manuscript's instructions.
Telomerase activity assay. The telomerase activity in ovaries of was measured using the telomerase (TE) ELISA kit (CUSABIO, Wuhan, China).
Measurement of average telomere length. DNA was extracted from ovaries using TIAN amp Genomic DNA Kit (Tian Gen Biotech CO., Ltd., Beijing, China) as described by the manufacturer and absorbance at 260 nm measured with a spectrophotometric plate reader (BioTek, Winooski, Vermont, USA). Average telomere length was measured from genomic DNA using a quantitative real-time PCR (qRT-PCR) assay. The telomere signal was normalized to that from the reference control gene to generate a T/S ratio indicative of relative telomere length.

Measurement of mitochondrial antioxidant function.
The mitochondrial fractions were assayed for the superoxide dismutase 2 (SOD2), catalase (CAT) and glutathione (GSH/GSSG) activities. All activities were measured using the kit, following the manufacturer's instructions (Beyotime Institute of Biotechnology, Jiangsu, China).
Isolation of cytosolic, mitochondria and nuclear fractions. Mitochondria or nuclear fractions were immediately extracted using the Mitochondrial or Nuclear Isolation Kit (Beyotime), respectively. All manipulations were performed on ice. Protein concentrations were determined with the BCA Protein Assay Kit (Pierce Biotech, Rockford, IL, USA).
Western blotting. Protein samples were separated by gradient electrophoresis on 8-15% SDS/PAGE gels.
The following antibodies were used to detect the proteins of interest: Glyceraldehyde-3-phosphate dehydrogenase (GAPDH), LAMIN A, β -TUBULIN, SIRT3, Cytochrome c, Bax, Bcl-2, and Cytochrome c oxidase (COX) IV from Cell Signaling Technology (Danvers, MA, USA) and FoxO3a, SOD2, CAT from Abcam (Cambridge, UK). Western blot analysis was performed as described previously 20 using the protocols provided by the primary antibody suppliers. Ovarian mitochondrial membrane potential (MMP) and respiratory chain complex activity (I-IV) measurements. Before analysis, mitochondria samples were subjected to three freeze-thaw cycles to disrupt membranes and expose the enzymes. All enzymatic activities were measured at 37 °C. MMP was determined using JC-1 as previously described 21 . Complex I (NADH-ubiquinone oxidoreductase) activity was assayed by measuring decreased absorbance at 340 nm, indicating reduction of NADH oxidation. Complex II (succinate-ubiquinone oxidoreductase) activity was assayed by monitoring reduction of 2,6-dichlorophenolindophenol (DCPIP) at 600 nm. Complex III (CoQ-cytochrome c oxidoreductase) was assayed by monitoring reduction of cytochrome c at 550 nm. Complex IV (cytochrome c oxidase) was assayed by monitoring the oxidation of reduced cytochrome at 550 nm 16 .
Mouse primary granulosa cell collection and culture. Mouse granulosa cells were collected from the ovaries of immature (3 weeks) Kunming mice using the follicle puncture method, as described previously 22 . The granulosa cells were cultured in DMEM/F12 supplemented with 10% FBS, 100 IU/mL penicillin, and 100 μ g/mL streptomycin at 37 °C and 5% CO 2 . Cells were incubated with 2 mM H 2 O 2 in the presence or absence of melatonin (100 μ g/mL) for 2 h. After incubation, cells were used for further analysis.  RNA interference of SIRT3. The siRNA for SIRT3 was purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). Granulosa cells were transfected with 100 nM SIRT3 small interfering and non-targeted control siRNA according to the manufacturer's protocol. Twenty-four hours after transfection, cells were harvested for further experiments.
Differences were analyzed using one-way analysis of variance (ANOVA) or t-test. For statistical analysis of the percentage values, the χ 2 test were performed. Differences at P < 0.05 were considered to be statistically significant.

Results
Melatonin counteracted age-related fertility decline. Initially, we first determined plasma melatonin levels in female mice treated with melatonin and in the mice of the same age without treatment. Plasma melatonin levels in treated-mice significantly increased at both times tested in compared to the respective control group ( Table 1). The values showed a peak of melatonin in the middle of the dark period, and the endogenous melatonin levels significantly decreased at the end of the dark period. However, in melatonin-treated mice, the plasma levels remained constantly higher compared with untreated mice (Supplementary Table S1).
Next, to determine whether supplementation with melatonin was able to rescue the reduced fertility associated with aging, mice treated with melatonin, which showed only a slightly increased litter size. In mice at the age of 14-16 mo, the average number of pups in the melatonin treated groups was significantly higher than from age-matched mice not receiving melatonin. It is notable that there was an increased pregnancy failure in the aged female mice following successful mating regardless of melatonin treatment, as indicated by the presence of mating plugs and mice bearing pups. These data indicate that the melatonin delays age-associated infertility, but could not completely prevent eventual reproductive aging (Table 2 and Supplementary Table S2).
Melatonin attenuates the loss of follicles and follicle atresia. Follicles at different developmental stages in the ovaries were shown in Fig. 1a. Mice treated with melatonin for 6 mo had more primordial follicles compared with untreated age-matched mice. The number of atretic follicles was significantly decreased in aged mice with melatonin treatment. Melatonin administration for 6 mo did not, however, improve the number of growing and mature follicles (Fig. 1b). The numbers of primordial and primary follicles, and few growing and mature follicles continued to decline with age, but were higher in mice treated with melatonin for 12 mo than in age-matched untreated mice, the number of atretic follicles was significantly lower in mice treated with melatonin at this age (Fig. 1c).
Melatonin increases the quantity and quality of oocytes. Next, we observed the oocytes retrieved from ovaries and oviducts of mice. As shown in Table 3 and Supplementary Table S3, the maternal age had a striking effect on the mean number of oocytes retrieved from ovaries and oviducts per mouse. The mean numbers of both ovarian GV oocytes and ovulated MII oocytes were higher in mice treated with melatonin for 12 mo than in age-matched untreated mice.
In vitro maturation of GV oocytes were shown in Table 4 and Supplementary Table S4. In mice at the age of 14-16 mo, maturation rate of oocytes from mice treated with melatonin was significantly lower than those of oocytes from the age-matched untreated mice.
Pronuclear formation of oocytes from mice of different ages in 4-7 h of IVF were shown in Table 5 and Supplementary Table S5. Pronuclear formation was accomplished in 7 h of IVF and the rate of oocytes forming pronuclei has no significant difference in each group. In mice at the age of 14-16 mo, the pronuclei formed in oocytes from aged mice without treatment later than that from mice treated with melatonin. Early embryonic development of oocytes after IVF is also shown in Table 5. Mice treated with melatonin for 12 mo increased the rates of oocytes developing to 2-cell and blastocyst than that of age-matched mice without treated.
Reports indicate that insufficient mtDNA levels and ATP contents could be directly associated with diminished oocyte competence 24 . The data showed that mtDNA copy numbers in GV oocytes in the vehicle group showed no difference from that of mice treated with melatonin for 12 mo (Fig. 2a). However, the average mtDNA copy numbers in MII oocytes from melatonin-treated mice were significantly higher compared with the age-matched mice without treatment (Fig. 2b). Mice treated with melatonin for 12 mo significantly increased the ATP levels both in GV (Fig. 2c) and MII oocytes (Fig. 2d).
Oocyte quality was further determined by spindle morphology and chromosome alignment (Fig. 3a). The frequency of normal oocytes in the untreated group showed no difference from that of mice treated with melatonin for 6 mo (Fig. 3b). Mice treated with melatonin for 12 mo exhibited a higher frequency of normal oocytes than age-matched untreated control (Fig. 3c). Oct4 immunostaining not only marks pluripotent cells but also indicates the inner cell mass of blastocysts 25 . Mice treated with melatonin for 12 mo did not influence the total Numbers of different types of follicles in the ovaries of female mice exposed orally to melatonin for 6 (b) or 12 mo (c) Bars represent means ± SD (n = 5/group). *P < 0.05; **P < 0.01 versus the Vehicle group.

Length of treatment (months) Groups
No. of mice examined number of cells in blastocysts (Fig. 3d,e), but significantly increased the number of Oct4-positive cells (Fig. 3d,f) and reduced the frequency of apoptosis (Fig. 3d,g) in blastocysts.
Melatonin suppressed ovarian mitochondrial oxidative stress and apoptosis. Mice treated with melatonin significantly decreased mitochondrial ROS production (Fig. 4a). Telomeres erode rapidly in response to oxidative stress. Ovaries from mice treated with melatonin for 12 mo had longer telomeres than those from age-matched mice without treatment (Fig. 4b). Telomerase activity in mice treated with melatonin for 12 mo was higher than in age-matched untreated mice (Fig. 4c). Aged mice treated with melatonin for 12 mo had markedly reduced 8-OHdG level in mitochondria compared with those of age-matched untreated mice (Fig. 4d).
Long-term treatment with melatonin prevented aging-associated oxidative stress in the ovarian mitochondria, as assessed by the recovery of GSH levels and GSH/GSSG ratios (Fig. 4e). Melatonin treatment significantly suppressed Bax protein expression and decreased the Bax/Bcl-2 ratio (Fig. 4f) and decreased cytoplasmic cytochrome c levels (Fig. 4g)     Melatonin attenuated ovarian mitochondrial dysfunction. In the ovarian mitochondria from female mice treated with melatonin for 12 mo, ATP content (Fig. 5a) and MMP (Fig. 5b) were significantly higher than in age-matched untreated females. Mitochondria, especially mitochondrial respiratory chain, has been shown to be major sources of ROS 26 . We measured the activities of complexes I-IV of mitochondrial electron transport chain (ETC). Aging did not affect activity of complex IV (Fig. 5f), but there was a significant decrease in activities of complex I (Fig. 5c), II (Fig. 5d) and III (Fig. 5e) in untreated mice at 14-16 mo of age compared with in young mice. Administration of melatonin resulted in a remarkable augmentation of the ETC activities, as compared with age-matched mice without treatment. In agreement with the reduced mitochondrial ROS production, long-term treatment with melatonin significantly elevated the ETC activities.  (Fig. 6a). In vitro analysis showed that melatonin blocked mitochondrial ROS elevation induced by H 2 O 2 in granulosa cells (Fig. 6b). Moreover, co-incubation with melatonin prevented H 2 O 2 -induced oxidative stress (Fig. 6c), DNA damage (Fig. 6d) and apoptosis (Fig. 6e,f).

Melatonin upregulates SOD2 and CAT through the interaction of SIRT3 with FoxO3a in vitro.
We firstly examined whether the expression of aging-related SIRT3 is modulated by oxidative stress in granulosa cells. Interestingly, the results revealed that granulosa cells stressed with H 2 O 2 induced a significant increase in SIRT3 expression (Supplementary Fig. S1a) and activity ( Supplementary Fig. S1b), however, it dropped significantly after two hours. Cell viability was further determined using the CCK-8 assay (Supplementary Fig. S1c).
Overexpression of SIRT3 restore the cell viability ( Supplementary Fig. S1e), while SIRT3-siRNA exhibited a  Fig. S1d) in response to H 2 O 2 , indicating that SIRT3 is necessary to prevent granulosa cells against senescence-related oxidative stress. Our data showed that melatonin enhanced the SIRT3 activity (Fig. 7a) without significantly affecting SIRT3 protein levels (Fig. 7b) in response to oxidative stress. FoxO3a protein expression significantly increased in the nuclear fraction of granulosa cells treated with melatonin compared to control (Fig. 7c) under oxidative stress. In ChIP assay, we found that SIRT3 promoted the binding of FoxO3a to the promoter of SOD2 and CAT (Fig. 7d). The increased protein expression of SOD2 and CAT were further confirmed with western blot analysis.

Melatonin increased the SIRT3 and FoxO3a expression in vivo.
Mice receiving melatonin had significantly higher activities of SIRT3 (Fig. 9a), CAT (Fig. 9b) and SOD2 (Fig. 9c) in the ovaries than did the untreated aged mice. We further investigated the role of FoxO3a in melatonin-induced CAT and SOD2 expression (Fig. 9d) using chromatin immunoprecipitation. In ChIP assay, we found that melatonin promoted the binding of FoxO3a to the promoters of CAT and SOD2 (Fig. 9e).

Discussion
In the present study we show that exogenous melatonin on the protects against the reduction during the aging of mice as indicated by litter size. Furthermore, melatonin improves the healthy follicle number, telomerase activity and telomere length, as well as the oocyte quality and quantity, indicating that it staves off the ovarian aging.
The well-established paradigm of reproduction in mammals holds that females are born with a fixed number of oocytes which continuously decline until few or none remain 27 . Our results are in agreement with studies, that aging ovaries exhibit reduced numbers, as well as reduced quantity and quality of oocytes. We show that melatonin may provide long-term beneficial effect by maintaining female reproductive capacity in mice. The beneficial effects of melatonin on ovarian aging might be contribute to the slowed loss of non-renewable female germ cells. But, we should note that melatonin rescue experiments showed that melatonin cannot prevent but can significantly delay declined fertility associated with reproductive age.
It has been suspected that the reproductive failure can be attributed to the functional and structural qualities of the oocytes 28 . Reports indicate that sub-normal mtDNA levels and ATP contents could be directly associated with poor oocyte quality and detrimental embryonic development 29 . It seems that the reduced number of oocytes retrieved with age may be of less importance than the decline in oocyte quality. Age-related fertility decline of the mice is not only due to a reduced oocyte pool, but also the result of a decrease in oocyte quality and embryonic development potential. Melatonin treatment both improves the quantity and quality of oocytes with age.   Various theories propose an essential role of mitochondria in cellular events associated with the aging process, via accumulation of mitochondrial ROS and oxidative damage to mitochondrial and cytoplasmic components 30,31 . There is evidence that mitochondrial respiratory complex activity and mitochondrial membrane potential decline during aging and that endogenous antioxidant system activities exhibit an age-dependent decrease, making redox imbalance play a more predominant role in the aged cells and tissues. Here we showed that melatonin was effective in ameliorating mitochondrial oxidative damage, apoptosis and preserving mitochondrial function in the aged ovarian. There is increasing evidence for beneficial effects of the antioxidant melatonin in preventing ROS-induced damage and pathology 32,33 . Reducing oxidative stress by supplementation of melatonin could potentially reduce mitochondrial ROS-induced damage, thus maintaining the number and quality of oocytes and follicles. Telomere shortening is considered as a biomarker of cellular senescence and is highly sensitive to oxidative stress 34,35 . Telomere dysfunction may contribute to reproductive aging-associated meiotic defects, miscarriage and infertility 36 . Thus, protection against oxidative stress by melatonin ( Supplementary Fig. S2) could directly contribute to maintenance of telomeres in aging mouse ovaries.
As one of the main components of ovary, granulosa cell is a appropriate object for in vitro experiments. Ovarian ageing involves the accumulation of damage by oxidative stress associated with an increasingly compromised microcirculation around the leading follicle 37 . Previous study have established that advanced maternal age have altered expression of genes involved in mitochondrial metabolic pathways in granulosa and cumulus cells compared with young 38,39 . Communication between the oocyte and its surrounding cells is essential for the development of a viable oocyte, and impaired mitochondria ( Supplementary Fig. S3) within the follicular cells ultimately affect oocyte development. Moreover, follicular atresia is the main process responsible for the loss of follicles and oocytes from the ovary, and it is the root cause of ovarian aging. More than 99% of follicles undergo atresia that has been associated with apoptosis of granulosa cells 40 . Therefore primary ovarian granular cells were further selected for in vitro experimentation.
SIRT3 is the primary mitochondrial deacetylase, and it regulates biological functions directly involved in mitochondrial function and limits the generation of mitochondrial ROS 41 . Residing predominantly in the mitochondria, SIRT3 is known to deacetylate and activate complex I of the respiratory chain in mitochondria, leading to an increase in levels of ATP and thereby protecting the cells from ROS-mediated oxidative damage 42 . Ectopic SIRT3 expression has antioxidant regulatory effects in primary cardiomyocytes, suggesting that SIRT3 may regulate antioxidant systems 43 . The activity of sirtuin 3 (SIRT3) was decreased in granulosa and cumulus cells from The binding of FoxO3a to the promoters of SOD2 and CAT were performed by ChIP-qPCR in ovary. Data are presented as the mean ± SD (n = 7/group). *p < 0.05, **p < 0.01 versus the Vehicle group.
women of advanced maternal age 44 . SIRT3 attenuation or ablation is associated with accelerated development of several diseases of aging 45 . SIRT3 has also been shown to interact with FoxO3a, the transcription factor implicated in ROS regulation and to activate FoxO3a-dependent gene expression 46,47 . Increased nuclear translocation of FoxO3a strengthens antioxidant defense systems, which leads to extended longevity, in several experimental organisms 48,49 . An age-mediated decrease in FoxO3a activity was observed in aged rats, which led to a decrease in the mitochondrial antioxidant enzymes 50 . In this study as FoxO3a was found to bind to the promoter regions of SOD2 and CAT genes. These results indicate that melatonin enhances the SIRT3 activity and activates FoxO3a nuclear translocation which leads to transactivation of the anti-oxidant genes, thereby resulting in inhibition of mitochondrial oxidative damage. Notably, though melatonin could scavenge the H 2 O 2 directly, H 2 O 2 radical scavenging ability of melatonin works mainly through the SIRT3 pathway ( Supplementary Fig. S4). Overall, SIRT3 and its target FoxO3a appear to function against mitochondrial oxidative damage.
In mammals, overall reproductive lifespan is predetermined at birth by the number of oocytes in the ovary and the ovarian aging manifestations of menopause occur when the finite supply of oocytes is depleted. However, our results using mice as a model suggest that this depletion can be significantly delayed, with potential clinical benefits for women who wish to postpone childbearing. More extensive research will be needed, however, prior to conducting clinical trials to test this proposal.