Melatonin as an endogenous hormone to slow down the exhaustion of ovarian follicle reserve in mice–a novel insight into its roles in early folliculogenesis and ovarian aging

5 Previous studies have shown that long-term intake of exogenous melatonin 6 can effectively delay ovarian aging, but the mechanism has not been fully 7 elucidated. We observed that SNAT, the rate-limiting enzyme in the melatonin 8 synthetic pathway, is localized in primordial and early follicle, and that 9 granulosa cells isolated from follicle can synthesize melatonin. In vitro cultured 10 neonatal mice ovaries with melatonin inhibited primordial follicle activation and 11 early follicle growth. In vivo experiments further indicated that daily injections 12 of melatonin to neonatal mice during the primordial follicle activation phase can 13 reduce the number of activated follicles by inhibiting the PI3K-AKT-FOXO3 14 pathway; during the early follicle growth phase, injections of melatonin 15 significantly suppressed early follicle growth and atresia, and transcriptome 16 data showed that multiple pathways involved in folliculogenesis, including 17 PI3K-AKT, were suppressed. Further, SNAT knockout in mice resulted in a 18 significant increase in follicle activation and atresia, and eventually accelerated 19 ovarian aging. We also demonstrated that prolonged high-dose melatonin 20 intake had no obvious adverse effect on the health condition of mice. This 21 study confirms that endogenous melatonin is involved in the regulation of 22 ovarian aging, and reveals that melatonin delays ovarian aging by inhibiting 23 primordial follicle activation, early follicle growth and atresia. 24 pattern of SNAT . The above data suggested that MLT may be involved in the regulation of primordial follicle activation and early follicle growth.


Introduction
Ovarian aging, which refers to the physiological process of active loss of 28 ovarian function in females before they age, is considered as a gain-of-function 29 mutation that can effectively avoid the physiological burden and life 30 threatening of the maternal body brought by the elderly pregnancy 1 . However, 31 ovarian aging is not just the loss of reproductive capability, the lack of gonadal 32 hormones caused by it can trigger many diseases including cardiovascular 33 disease, cancer, osteoporosis, obesity and menopausal syndrome 2-5 . 34 Therefore, ovarian aging, which is called the pacemaker of female body aging, 35 indirectly accelerates the aging of multiple organs 6 . With the development of 36 medical technology, the various risks brought by elderly pregnancy can be 37 technically avoided. Therefore, the radical cure or delay of ovarian aging is 38 becoming an urgent need for women. 39 The mainstream view tends to suggest that there are no germline stem 40 cells in the ovary to replenish the continual decrease of the ovarian follicle 41 reserve. Therefore, the exhaustion of the ovarian follicle reserve in middle age of primordial follicles in PreOR greatly decrease, which indirectly leads to a 66 reduction in SAFs in DOR. Therefore, the degree of ovarian aging can be 67 reflected by the SAFs count 15,16 . Of note, the SAFs from the DOR will be 68 eliminated by means of atresia if without sufficient gonadotropins. Therefore, 69 the larger the DOR, the more SAFs will be eliminated to ensure the stability of 70 fecundity, and accordingly, more primordial follicles will be activated from 71 PreOR to supplement the loss of DOR. This bad circulation will eventually 72 result in a meaningless exhaustion of ovarian follicle reserve 13,17,18 . For these 73 reasons, the vast majority of follicles in ovary will be eliminated before puberty, 74 because the hypothalamic pituitary is not mature enough to produce enough 75 gonadotropins 19 . In a word, inhibition of early folliculogenesis (i.e. primordial glucose metabolism in the long evolutionary process 23,24 . As the most 87 representational antioxidant, MLT is also used to delay ovarian aging [25][26][27][28][29][30] . 88 Studies have demonstrated that shortening the sunshine time can delay 89 reproductive aging, which is considered to be related to MLT 31,32 ; long-term   112 folliculogenesis 113 We first detected the expression and localization of SNAT in the prepubertal 114 mice (PD5,9,17,19). Immunohistochemical data indicated that SNAT was 115 mainly distributed in the granulosa cells (GCs) of primordial follicle and early 116 growing follicle (Fig. 1A). We next examined whether MLT can be synthesized 117 in GCs. To this end, isolated GCs were incubated with or without 118 5-hydroxytryptamine (5-HT), the precursor of MLT and SNAT substrate. HLPC 119 was then used to detect the conversion of 5-HT into MLT. In three cases, 120 namely, no addition of GCs or 5-HT, the addition of GCs but no 5-HT, and the 121 addition of 5-HT, no MLT was detected in the culture medium. When GCs and 122 5-HT were added at the same time, MLT absorption peak appeared in the 123 chromatography. The above results indicated that the early growing follicle of 124 mice had the capacity to synthesize MLT (Fig.1B).

125
To further examine the changes of ovarian MLT synthesis capacity during 126 prepubertal folliculogenesis, we collected ovaries at PD5,7,9,11,13,15,17,127 19, and 21. The qRT-PCR data showed that the overall expression level of 128 SNAT in the ovary continued to decrease; Fshr and Lhcgr were the marker 129 genes that reflected the process of folliculogenesis, and their expression levels 130 gradually increased during the above process, which is opposite to the 131 expression pattern of SNAT (Fig. 1C). Next, we examined the MLT level in 132 ovarian homogenates at PD7, 9, 15, 17, and 19. The changes in MLT level, 133 which gradually decreased with the increase of the age (Fig.1D), is similar to 134 the expression pattern of SNAT. The above data suggested that MLT may be  different superscript letters (a-b) represent a significant difference, P <0.05. "*" represents 151 significant differences, P <0.05; "**" represents extremely significant differences, P <0.01. 152 153 culture system 154 After the establishment of PreOR of mouse, primordial follicles started to 155 activate around PD3, which makes neonatal ovary an ideal model to study 156 primordial follicle activation. To investigate whether MLT is involved in the 157 regulation of primordial follicle activation in in vitro culture system, ovaries from 158 mice at PD3 were cultured in vitro for 96h, and then the number of activated 159 follicles with and without MLT treatment were counted ( Fig. 2A). The result 160 showed a remarkable decrease in the number of activated follicles after both 161 10 -8 and 10 -7 M MLT treatment. Additionally, 10 -7 M MLT can also reduce the 162 proportion of atresia follicles, but the difference was not significant (Fig. 2B).

163
The above data indicated that MLT can inhibit the primordial follicle activation.

164
To further study the effect of MLT on early follicle growth, ovaries from   administration did not affect the mTORC1 signaling. Therefore, the mechanism 213 of how MLT inhibits the PI3K-Akt-Foxo3 cascade remains to be further 214 elucidated. 215 We also studied the antioxidant capacity of MLT in mouse ovary, and 216 observed that short-term MLT intake significantly upregulated the expression 217 of SOD1 in the ovary (P =0.0421) (Fig. 3D), and elevated the serum levels of   243 To further study the effect of MLT on early follicle growth and atresia in vivo, we smaller than the control group (P =0.0354), which is also consistent with the 284 section data, and the embryo size was bigger than the control group (P <0.01),

285
Of note, the embryos in the MLT intake group showed unequal distribution on 286 both uterine horns, and the cause is still unclear.

287
The above data collectively indicate that exogenous MLT intake restricts 288 the size of DOR and inhibits its atresia. To further explore the possible 289 molecular mechanism, we performed RNA-Seq. The results showed that 598 290 genes in ovaries at PD17 exhibited differential expression (Fig. 4E,F). GO 291 enrichment analysis showed that the differentially expressed genes were 292 mainly clustered into "Negative regulation of cell proliferation and 293 differentiation", "Regulation of cellular response to growth factor stimulus", and 294 "Blood vessel development", (Fig. 4G). We further validated the accuracy of 295 GO analysis. The proliferation activity of GCs was suppressed by MLT, as 296 PCNA, the proliferation marker, was down-regulated in MLT-treated group (P 297 =0.0291) (Fig. 4G). The KEGG heatmap showed that the differentially 298 expressed genes were mainly enriched in the typical pathways associated with 299 folliculogenesis, including "PI3K-AKT signaling", "MAPK signaling", "VEGF 300 signaling", "Pathways in cancer" and "TGF-beta signaling pathway" (Fig. 4H). It  were deleted (Fig. 5A). We then counted the effect of SNAT knockout on 343 follicle activation and atresia (Fig. 5B). It was found that the number of 344 activated follicles could be slightly increased after SNAT knockout, and the  Since MLT can affect the activation, growth and atresia of follicles, will it 361 affect ovarian aging? To this end, we first analyzed the changes of litter size in 362 mice with age after SNAT knockout (Fig. 5D). At the age of 2-3 months, SNAT 363 knockout had no significant effect on the litter size; as the age increased to 5-6 364 months, the litter size in SNAT knockout mice was significantly smaller than 365 that in the wild type (P =0.0398); and at 8-9 months the litter size in SNAT 366 knockout mice at was extremely significantly smaller than that in the wild type Zp3 in SNAT knockout ovary were group was significantly lower than that in 375 the wild type (P <0.05) (Fig. 5F), which suggested that the number of 376 primordial follicles in SNAT knockout ovary was lower than wild type. The  represent a significant difference. "*" represents significant differences, P <0.05; "**" 401 represents extremely significant differences, P <0.01. 402 403 404 In the present study, both 1.0 mg/kg and 15 mg/kg MLT were effective in 405 inhibiting follicle activation. However, these doses are far above the 406 physiological dose of MLT in the plasma 38 , and this is a problem that cannot 407 be ignored. Therefore, the safety of high-dose MLT was assessed in this 408 section. We firstly studied the effect of MLT on mice physical growth. It was 409 observed that long-term intake of 15mg/kg MLT neither affected body 410 development nor visceral index (Fig. 6A, B). Subsequently, we evaluated the  438 folliculogenesis and ovarian aging 439 Cisplatin is a commonly used chemotherapeutic drug that excessively  467 ovarian aging 468 Cell culture and gene knockout data indicated that follicular cells had the ability 469 to secrete MLT (Fig. 1B), the loss of follicular reserve in MLT-deficient mice 470 was faster than that of the wild type, and the phenotype of ovarian aging was 471 more obvious in middle age (Fig. 5). These results collectively indicated that  488 Studies have found that preovulatory follicular fluid contains high levels of MLT demonstrated that follicular GCs can synthesize MLT (Fig. 1B). Therefore, 496 even though the MLT that regulates follicle activation, growth, and atresia in 497 this study is unlikely to be entirely derived from GCs, we can still reasonably 498 speculate that the ovary is another MLT secreting organ based on the above 499 research progress. market. This study also demonstrated that high-dose MLT intake that lasted for 519 one month had no significant effect on the estrous cycle, reproductive rhythm, 520 and physical development of experimental animals (Fig. 6). Nevertheless, the 521 question of whether MLT can be used as a human ovarian health product 522 needs further investigation. In this study, the minimum dose of MLT to inhibit 523 follicle activation is 1 mg/kg, if converted to a human dose, it is about 50 mg 524 per day, but the current recommended maximum intake of MLT products on 525 the market is 10 mg/day. Therefore, the dosage intake of mice cannot be 526 applied directly to humans. In addition, the average effective reproductive life 527 of humans is 40 years, while the average of mice is 1.5 years. Therefore, it is 528 technically difficult to determine whether MLT can delay ovarian aging in 529 humans. Even if MLT can delay ovarian aging in humans, the using method, 530 optimal dose and potential side effects at that dose should also be reassessed.

532
In summary, this study demonstrates for the first time that endogenous 533 melatonin is involved in the regulation of ovarian aging, and presents a novel 534 insight into the roles of melatonin in early folliculogenesis and ovarian aging, 535 that is, in addition to its well-known antioxidant properties, melatonin as an 536 endogenous hormone slows down the exhaustion of ovarian follicle reserve 537 and ovarian aging by directly inhibiting primordial follicle activation, early 538 follicle growth and atresia (Fig. 7). This discovery helps us to understand the 539 functions of melatonin in reproductive regulation comprehensively. It will be of 540 great importance in future studies to ascertain whether MLT has the similar 541 effects in humans and whether it can be used as a health product for delaying 542 ovarian aging.  were quantified based on peak height. 645 Blood samples of each group were collected according to the above design.

646
After clotting for 30 min, the serum was obtained by centrifugation at 3000 rpm   665 The whole blood samples in each group were collected in anti-coagulation  676 The sequencing and data annotation were completed by Novogene In vivo treatment with MLT suppressed the activation of primordial follicle though PI3K AKT FOXO3 pathway In vivo treatment with MLT inhibited early follicle growth and atresia SNAT knockout in mice accelerated the exhaustion of ovarian follicle reserve and age related fertility decline Figure 6 Long term ingestion of excessive MLT did not disturb the reproductive rhythm, growth and health condition of mice