Ovarian follicles are the basic functional units of female reproduction in the mammalian ovary. We show here that the protein a disintegrin and metalloproteinase domain 10 (ADAM10), a cell surface sheddase, plays an indispensable role in controlling primordial follicle formation by regulating the recruitment of follicle supporting cells in mice. We demonstrate that suppressing ADAM10 in vitro or deletion of Adam10 in vivo disrupts germline cyst breakdown and primordial follicle formation. Using a cell lineage tracing approach, we show that ADAM10 governs the recruitment of ovarian follicle cells by regulating the differentiation and proliferation of LGR5-positive follicle supporting progenitor cells. By detecting the development of FOXL2-positive pregranulosa cells, we found that inhibiting ADAM10 reduced the number of FOXL2-positive cells in perinatal ovaries. Furthermore, inhibiting ADAM10 suppressed the activation of Notch signaling, and blocking Notch signaling also disrupted the recruitment of follicle progenitor cells. Taken together, these results show that ADAM10–Notch signaling in ovarian somatic cells governs the primordial follicle formation by controlling the development of ovarian pregranulosa cells. The proper recruitment of ovarian follicle supporting cells is essential for establishment of the ovarian reserve in mice.
As the basic functional unit of female reproduction, primordial follicles serve as a source of developing follicles and oocytes over the entire reproductive lifespan of most mammalian species. Meanwhile, the non-renewable pool of primordial follicles, which is established in the perinatal ovary, directly represents the fecundity of each female (Mandl and Zuckerman, 1951; Zhang et al., 2014). Although the formation of primordial follicles has been well studied at a histological level (Pepling, 2012), the inner molecular mechanisms that regulate this process are still unclear.
In mice, most germ cells stop dividing, enter into meiosis and form special structures termed germline cysts in the ovary before 16.5 days post coitum (dpc) (Pepling and Spradling, 1998). At same time, progenitor cells of somatic supporting cells start being recruited from the ovarian surface (Maatouk et al., 2012; Mork et al., 2012). At 17.5 dpc, FOXL2-positive (FOXL2+) somatic supporting cells, also called pregranulosa cells, begin to invade into cysts and separate germ cells to form primordial follicles (Pepling et al., 2010). At ∼4 days post partum (dpp), almost all oocytes are surrounded by pregranulosa cells and have formed primordial follicles in the ovary.
The origins of pregranulosa cells have been discussed for decades. Previous studies have shown that FOXL2+ cells surround oocytes as pregranulosa cells and form primordial follicles (Mork et al., 2012; Schmidt et al., 2004; Zheng et al., 2014a). Recently, it has been shown that Lgr5, a marker for tissue stem cells in many organs, is expressed in the cortical region of mouse ovary during perinatal stages (Ng et al., 2014; Rastetter et al., 2014). The specific expression pattern of Lgr5 and Foxl2 suggested that LGR5-positive (LGR5+) cells in the ovarian surface epithelium are one of the major origins of FOXL2+ pregranulosa cells that contribute to the formation of primordial follicles (Rastetter et al., 2014). However, the molecular mechanisms involved in how such cells are recruited and become differentiated remain unclear. More importantly, the relationship between the process of follicle somatic cell development and primordial follicle formation is totally unknown.
The protein a disintegrin and metalloproteinase domain 10 (ADAM10) is a cell surface sheddase that is able to cleave the extracellular domains of membrane-bound proteins and release the soluble ectodomain in a process known as ‘ectodomain shedding’ (Hayashida et al., 2010; Reiss and Saftig, 2009; Weber and Saftig, 2012). ADAM10 is crucial for the regulation of various physiological processes such as cell proliferation, differentiation and death (Gibb et al., 2011; Glomski et al., 2011; Guo et al., 2015; Jorissen et al., 2010; Tsai et al., 2014; Weber et al., 2011; Zhang et al., 2010). It is well known that the regulatory roles of ADAM10 are closely related to Notch signaling (Weber and Saftig, 2012). ADAM10 functions by releasing the Notch intracellular domain (NICD) from the plasma membrane, which then enters the nucleus to activate Notch target genes such as Hes1 and Hey2 (Borggrefe and Oswald, 2009). Both ADAM10 and Notch signaling regulate the proliferation and differentiation of LGR5+ stem (or progenitor) cells in various mouse organs, such as the intestine, stomach and cochlea (Bramhall et al., 2014; Demitrack et al., 2015; Tsai et al., 2014; VanDussen et al., 2012). More specifically, an increasing body of evidence has shown that Notch signaling is indispensable for the formation of ovarian primordial follicles (Trombly et al., 2009; Vanorny et al., 2014; Xu and Gridley, 2013). However, whether and how ADAM10 plays a role in the establishment of the primordial follicle pool has not been addressed, and the relationship between ADAM10 and Notch signaling in primordial follicle formation remains unknown.
In the current study, we investigated the functional role of ADAM10 in regulating the formation of primordial follicles in the perinatal mouse ovary. We found that somatic-cell-expressed ADAM10 plays an indispensable role in the progression of germline cyst breakdown and primordial follicle formation in mice. This was achieved by regulating the recruitment of FOXL2+ pregranulosa cells from LGR5+ follicle supporting progenitor cells through Notch signaling. Our findings provide evidence showing that ADAM10–Notch signaling determines the establishment of the primordial follicle pool by regulating the development of follicle somatic cells in the mouse ovary.
Suppressing ADAM10 disrupts germline cyst breakdown and primordial follicle formation
To investigate the potential relationship between ADAM10 and primordial follicle formation, we first measured the expression of Adam10 mRNA by quantitative real-time reverse transcription PCR (qRT-PCR) in perinatal ovaries. As shown in Fig. 1A, mRNA levels of Adam10 changed markedly from 16.5 dpc to 4 dpp as primordial follicle formation progressed. From 16.5 dpc to 1 dpp, Adam10 mRNA expression increased considerably with the breakdown of germline cysts; the highest level of Adam10 mRNA was observed at 1 dpp, which was 2.8-fold higher than that at 16.5 dpc. With the establishment of the primordial follicle pool at 4 dpp, the expression of Adam10 mRNA decreased to a level comparable to that at 16.5 dpc. To detect the localization of Adam10 in the perinatal ovary, we performed in situ hybridization and found that Adam10 was mainly expressed in somatic cells, especially ovarian surface epithelium cells, in 1 dpp ovary (Fig. 1B, arrowheads). These results indicate that somatic-cell-expressed ADAM10 has the potential to be involved in the regulation of primordial follicle formation in the mouse ovary.
To determine the functional role of ADAM10 in primordial follicle formation, we collected ovaries from 16.5 dpc wild-type mice and cultured them with or without GM6001, an inhibitor of ADAM10 (Lemjabbar and Basbaum, 2002; Saftig and Reiss, 2011; Zhang et al., 2011), for 7 days in vitro. After 7 days of culture, most oocytes (8188.0±1152.3 oocytes per ovary, 86.6±12.2%, mean±s.d.; Fig. 1C) in the control group were surrounded by pregranulosa cells and formed primordial follicles (Fig. 1D, arrowheads). By contrast, only 14.2±1.8% of oocytes were surrounded by pregranulosa cells to form primordial follicles in GM6001-treated ovaries (1442.0±181.4 per ovary, P<0.001 versus control; Fig. 1C); the majority of oocytes (8744.0±530.7 oocytes, 85.8±5.2% per ovary, P<0.001 versus control; Fig. 1C) were still included in cysts without forming primordial follicles (Fig. 1D, broken circle). An identical number of oocytes were observed in control (9452.0±999.5 oocytes per ovary; Fig. 1C) and GM6001-treated ovaries (10186.0±355.1 oocytes per ovary; Fig. 1C), indicating that the survival of oocytes was not affected by inhibiting ADAM10. In addition, we cultured ovaries with another ADAM10 selective inhibitor GI254023X (Hoettecke et al., 2010; Hundhausen et al., 2003; Lo Sardo et al., 2012). The results showed that 7 days of GI254023X treatment also significantly decreased the number of follicles (1138.0±184.0 per ovary for GI254023X versus 8188.0±1152.3 per ovary for control; Fig. S1B) and most of the oocytes (5944.0±711.4 per ovary for GI254023X versus 1264.0±273.0 per ovary for control; Fig. S1B) remained included in cysts (Fig. S1A). Taken together, these results demonstrate that ADAM10 plays a crucial role in the formation of primordial follicles in the perinatal mouse ovary.
Suppressing ADAM10 inhibits the recruitment of Foxl2+ pregranulosa cells in perinatal ovaries
To find out how somatically expressed ADAM10 controls the formation of primordial follicles, we next assessed the functional role of ADAM10 in regulating the recruitment of pregranulosa cells. We first investigated the in vivo behavior of FOXL2+ pregranulosa cells in the perinatal ovary. Immunostaining confirmed a previous report (Rastetter et al., 2014) describing the localization of FOXL2+ cells that first emerged in the medulla and then gradually emerged in the cortical region of the ovary, from 16.5 dpc to 4 dpp, with the formation of primordial follicles (Fig. S2A).
We then cultured 16.5 dpc ovaries with GM6001 to examine the role of ADAM10 in recruiting pregranulosa cells. After 7 days of culture, we found that 475.4±31.4 FOXL2+ cells per section (mean±s.d.) (Fig. 2B) were detected in the cortical region of control ovaries (Fig. 2A). By contrast, 7 days of GM6001 treatment significantly decreased the number of FOXL2+ cells (221.2±16.8 cells per section; Fig. 2B; P<0.001) in the cortex of the ovary (Fig. 2A), and only a few FOXL2+ cells were localized around germline cysts. To confirm the role of ADAM10 in recruiting pregranulosa cells, we also measured the levels of Foxl2 mRNA after culturing in the presence or absence of GM6001. Foxl2 mRNA expression was found to be significantly decreased after 3 or 7 days in culture in the presence of GM6001 (Fig. 2C; P<0.001 for both). These results indicate that ADAM10 determines the recruitment of FOXL2+ pregranulosa cells to the cortex to regulate primordial follicle formation.
Suppressing ADAM10 inhibits ovarian LGR5+ progenitor cell differentiation
Given that ADAM10 determines the number of FOXL2+ pregranulosa cells, we next examined how ADAM10 disrupts the recruitment of FOXL2+ cells to the ovarian cortex. Recent studies have reported that FOXL2+ cells are derived from LGR5+ cells in the ovarian surface epithelium (Ng et al., 2014; Rastetter et al., 2014). Therefore, we used Lgr5-EGFP-ires-CreERT2 (Lgr5-KI) reporter mice to trace the recruitment of LGR5+ cells, and investigated the developmental dynamics of LGR5+ cells and FOXL2+ cells in perinatal Lgr5-KI ovaries by co-staining with anti-FOXL2 antibody. LGR5 was located by staining with anti-EGFP antibody.
As shown in Fig. 3A, LGR5+ cells were not only observed in the ovarian surface epithelium but also in the cortical region of ovaries at 16.5 dpc and 1 dpp, indicating that the recruitment of LGR5+ cells was activated from 16.5 dpc to 1 dpp, before primordial follicle formation. At 4 dpp, no LGR5+ cells were found in the ovarian cortex, indicating that the recruitment of follicle supporting cells had been terminated in vivo. By staining for FOXL2 in Lgr5-KI mice, we found that the transformation of LGR5+ cells to FOXL2+ cells mainly occurred around 1 dpp (Fig. 3A; Fig. S2E) because many LGR5 and FOXL2 double-positive cells were observed in 1 dpp ovaries (Fig. 3A, arrowheads; Fig. S2D, arrowheads). By counting the number of LGR5+ cells and FOXL2+ cells from 16.5 dpc to 4 dpp, we assessed the dynamics of cell recruitment, and found that the number of LGR5+ cells in the cortical region gradually decreased (from 186.4±17.6 to 4.8±2.3 cells per section), and the transformation of FOXL2+ cells markedly increased (from 89.2±19.8 to 571.2±55.3 cells per section) in vivo with primordial follicle formation (mean±s.d.; Fig. 3B).
We next investigated whether ADAM10 controls the recruitment and differentiation of LGR5+ cells in ovaries. Ovaries (16.5 dpc) from Lgr5-KI mice were collected and cultured in the presence or absence of GM6001 for 3 or 7 days, respectively. To validate the culture system, the development of LGR5+ cells and FOXL2+ cells was first observed in cultured ovaries (Fig. 3C), and similar dynamics to that for in vivo developed ovaries follicle were found (compare Fig. 3C with Fig. 3A). After 7 days of in vitro culture with GM6001, however, LGR5+ cells were not only observed at the ovarian surface but also in the cortical region of ovaries (Fig. 3E); the number of LGR5+ cells (105.4±3.6 cells per section; Fig. 3F) was considerably increased in the GM6001-treated group compared to control (9.8±2.5 cells per section, Fig. 3D). The direct comparison shows that the change is significant (Fig. S2G). A fluorescence-activated cell sorting (FACS) analysis of Lgr5-KI ovarian cells after 7 days of culture also showed that the proportion of LGR5+ cells was significantly increased in the GM6001-treated group (27.1±1.2%; Fig. S3B) compared to control (14.9±1.5%; Fig. S3A,B). The higher number of LGR5+ cells indicates that either the differentiation of LGR5+ cells is suppressed or the proliferation of LGR5+ cells is increased by GM6001 treatment.
Indeed, our counting results showed that the increase in FOXL2+ cells was significantly suppressed in the cortical region of ovaries by GM6001 treatment (Fig. S2E). This indicated that the differentiation of LGR5+ cells was suppressed by GM6001 treatment. Moreover, by counting the number of LGR5 and FOXL2 double-positive cells after 3 or 7 days in GM6001-treated ovaries, it was found that the progression of LGR5+ cell differentiation was significantly retarded (Fig. 3G,H). In addition, the levels of Lgr5 mRNA were significantly higher after 3 or 7 days for the GM6001-treated group compared to control (Fig. 3I), which is consistent with the results of cell counting.
ADAM10 controls the recruitment of ovarian surface LGR5+ progenitors
Our previous results demonstrated that the transformation of LGR5+ cells into FOXL2+ cells was suppressed when ADAM10 was inhibited; hence, we next examined whether GM6001 treatment increased the proliferation of LGR5+ cells. The ovarian surface cells were labeled by lentivirus-transfected EGFP, which integrated into genomic DNA, and the recruitment of LGR5+ cells was traced in vitro with or without GM6001. The experimental strategy is shown in Fig. 4A. After 12 h of infection and 12 h of lentivirus-free culture, surface cells in infected ovaries were successfully labeled by EGFP (white color) and no cells within the inner part of ovaries were labeled (Fig. 4B). The statistics of EGFP+ cells indicated that there was no significant difference in the infection efficiency of different ovaries (Fig. S2I). Infected ovaries were then cultured in the presence or absence of GM6001 for 7 days and the behavior of labeled cells was visualized. In control group, it was found that for untreated control cells, the population of EGFP+ cells had obviously increased, with many EGFP+ cells invading the cortex and surrounding oocytes to form primordial follicles after culture (Fig. 4B, control). In sharp contrast, with GM6001 treatment, EGFP+ cells were mainly observed at the ovarian surface and only a few labeled cells had invaded into the ovary (Fig. 4B, GM6001). This result indicates that the recruitment of ovarian surface cells is suppressed by inhibiting ADAM10. Moreover, the number of EGFP+ cells in the GM6001-treated group (39.7±5.7 cells per section; mean±s.d.; Fig. 4C) was significantly less than the number of cells in the control group (110.0±9.0 cells per section; Fig. 4C; P<0.001), indicating that the proliferation of ovarian surface epithelium was also inhibited by GM6001. Given that ovarian surface epithelium cells are LGR5+ (Fig. 3A), our results show that both the recruitment and proliferation of LGR5+ cells is suppressed by GM6001 treatment.
To confirm our findings, proliferating cells were labeled with BrdU in Lgr5-KI ovaries and the number of dividing LGR5+ cells on the ovarian surface was measured. We found that the proliferation of ovarian surface LGR5+ cells was significantly suppressed after 3 days of GM6001 treatment (Fig. 4D, arrowheads; Fig. 4E, control at 22.6±3.4 cells compared to GM6001 at 10.4±1.7 per section). Thus, our results indicate that ADAM10 plays a key role in regulating the recruitment and proliferation of LGR5+ progenitor cells that contributes to primordial follicle formation.
ADAM10 regulates the recruitment and differentiation of pregranulosa cells through Notch signaling
A functional role for Notch signaling in controlling primordial follicle formation has recently been reported (Trombly et al., 2009; Vanorny et al., 2014; Xu and Gridley, 2013). As an upstream regulator of Notch signaling, ADAM10 might play a similar role in regulating the formation of primordial follicles according to our findings. Thus, we next investigated the relationship between ADAM10 and Notch signaling in the advancement of primordial follicle formation, especially the recruitment of follicle supporting cells.
To examine whether Notch signaling plays a similar role in regulating the differentiation and proliferation of LGR5+ cells as ADAM10, 16.5 dpc ovaries from Lgr5-KI mice were collected and cultured in the presence of DAPT, a Notch signaling inhibitor. As reported previously (Trombly et al., 2009), 7 days of DAPT treatment disrupted germline cyst breakdown and primordial follicle formation (data not shown). For the development of follicle supporting cells, we found that a comparable phenotype was observed with both DAPT and GM6001 treatments. Specifically, the differentiation of LGR5+ cells to FOXL2+ cells was markedly blocked by 7 days of DAPT treatment (Fig. 5E). In addition, a decrease in FOXL2+ cell numbers (157.6±14.4 for DAPT versus 475.4±14.4 cells per section for control; mean±s.d.; Fig. 5B,D) and an increase in LGR5+ cell numbers (133.0±10.2 per section in DAPT versus 9.8±2.5 cells per section in control; Fig. 5B,D) were noted in the cortical regions of ovaries treated with DAPT for 7 days (Fig. 5A,C). The direct comparison shows that the change was significant (Fig. S3C,D). FACS analysis showed that the proportion of LGR5+ cells was significantly increased in the DAPT-treated group (28.8±2.8%; Fig. S3B) compared to control after 7 days of culture (14.9±1.5%; Fig. S3A,B). Gene expression patterns confirmed our counting data that significantly decreased levels of Foxl2 and increased expression of Lgr5 were found in ovaries after 3 or 7 days of DAPT treatment compared to control (Fig. S3E,F; P<0.001 for all). Moreover, the proliferation of LGR5+ cells labeled with BrdU was also significantly suppressed in the DAPT-treatment group (Fig. 5F, arrowheads; 18.2±2.2 for DAPT versus 9.0±1.2 cells per section for control; Fig. 5G; P<0.001). Lentivirus EGFP infection experiments also showed that the recruitment of ovarian surface cells was significantly inhibited by DAPT treatment after 7 days of culture (Fig. S3G and H; P<0.001). These results point to a similar functional role for ADAM10 and Notch signaling in controlling the recruitment of supporting cells of the ovarian follicle in mice.
To confirm the relationship between ADAM10 and Notch signaling in primordial follicle formation, we measured gene or protein expression patterns of Notch signaling in 16.5 dpc ovaries after 3 days of GM6001 treatment. Gene expression analysis showed no significant changes of Notch1, Notch2 and Jagged1 mRNA levels (data not shown). However, the protein level of NICD, the Notch intracellular domain, was decreased in ovaries of the GM6001-treated group (Fig. 5H) compared with control. In addition, the expression of the Notch target genes Hey2 and Hes1 was significantly reduced (Fig. 5I,J; P<0.001 for all) by GM6001 treatment. We therefore conclude that ADAM10 controls the differentiation of LGR5+ cells and recruitment of FOXL2+ cells by regulating Notch signaling in ovarian somatic cells.
Deletion of ADAM10 in vivo disrupts the formation of primordial follicles in mice
Our data from in vitro experiments outlined above indicate that ADAM10 expressed by ovarian somatic cells plays a key role in controlling primordial follicle formation. To assess whether ADAM10 is required in vivo for folliculogenesis, we conditionally deleted ADAM10 in a somatic lineage of perinatal ovaries using an anti-Müllerian hormone type 2 receptor-Cre (Amhr2-Cre) mouse line (Jamin et al., 2002), which mediates the expression of CRE recombinase in somatic cells of ovaries from 12.5 dpc (Jamin et al., 2002; Jorgez et al., 2004; Vanorny et al., 2014). qRT-PCR analysis of ovaries showed that the levels of Adam10 mRNA in Amhr2-Cre;Adam10f/d ovaries was 52.8% that of control littermates (Fig. S4C). This result indicates that Adam10 is partly disrupted in ovaries. Histological analysis of ovaries taken from Amhr2-Cre;Adam10f/d mice at 7 dpp identified abnormalities in primordial follicle formation (Fig. 6A, yellow, dashed circles). Many germline cysts were observed in the ovaries of Amhr2-Cre;Adam10f/d mice, which was in sharp contrast to that observed during normal folliculogenesis in control ovaries. By counting the number of germline cysts and oocyte numbers in cysts, we found that the ovaries of Amhr2-Cre;Adam10f/d mice exhibited significantly more unbroken cysts (114.0±37.6 cysts per ovary; mean±s.d.; Fig. 6B) compared with littermate controls (13.0±9.2 cysts per ovary; Fig. 6B; P<0.001). Many oocytes (892.0±176.9 for mutant versus 52.0±36.7 for control oocytes per ovary; Fig. 6C) in Amhr2-Cre;Adam10f/d ovaries remained within the cysts. These data indicate that ADAM10 in somatic cells of the mouse ovary regulates primordial follicle formation under physiological conditions.
In this study, we demonstrate that ADAM10 plays a key role in primordial follicle formation in the perinatal mouse ovary. Furthermore, we find that ADAM10 controls the recruitment of FOXL2+ pregranulosa cells by regulating the proliferation and differentiation of LGR5+ progenitor cells. In addition, we show that Notch signaling is the pathway through which ADAM10 regulates the recruitment of pregranulosa cells and primordial follicle formation in mice.
Recent studies have found that the ovarian surface epithelium, which is LGR5 positive, appears to be the major origin of pregranulosa cells (Mork et al., 2012; Ng et al., 2014; Rastetter et al., 2014). After recruitment from ovarian epithelium, LGR5+ supporting progenitor cells gradually differentiate into pregranulosa cells and become Foxl2 positive (Suzuki et al., 2015). Although the dynamics of FOXL2+ cell recruitment have been well established by cell lineage tracing experiments (Mork et al., 2012; Zheng et al., 2014a), the upstream molecular regulation of this process is still unclear. In the current study, we found that ADAM10, by regulating Notch signaling, was essential for this process. Protein expression analysis of the Notch intracellular domain showed that ADAM10 might be responsible for cleavage of the Notch receptor in both LGR5+ cells and FOXL2+ cells in perinatal ovaries. Moreover, we also show that the Notch target genes Hey2 and Hes1 (Trombly et al., 2009) are also indirectly controlled by ADAM10, given that the expression of Hey2 and Hes1 is mainly restricted to somatic cells in ovaries.
It is known that Notch signaling influences cell fates, such as cell differentiation and the cell cycle, particularly in conjunction with stem cell development, in mammals (Chiba, 2006). Inhibiting Notch in the stomach reduces the proliferation of LGR5+ gastric stem (progenitor) cells and the differentiation of mucous and endocrine cells (Demitrack et al., 2015). In the intestine, ADAM10–Notch signaling directly targeted the LGR5+ crypt base columnar (CBC) cells and is pivotal to the balance of stemness and differentiation of LGR5+ CBC stem cells (Demitrack et al., 2015; Tsai et al., 2014; VanDussen et al., 2012). Our current results show that ADAM10–Notch signaling is crucial in determining the fate of LGR5+ cells in the ovary. We hypothesize that the activation of ADAM10–Notch signaling stimulates the differentiation of LGR5+ cells to FOXL2+ cells in the fetal ovary, whereas the quiescence of this signal is essential in maintaining the dormancy of the LGR5+ ovarian surface epithelium in the adult ovary.
Previous work from our laboratory has demonstrated that the synchronized development of oocytes and somatic cells is essential for primordial follicle formation (Lei et al., 2006). After the deletion of a cell cycle inhibitor gene, p27Kip1 (Cdkn1b), another study has reported that the number of somatic pregranulosa cells is potentially related to the size of the primordial follicle pool (Rajareddy et al., 2007). The current study suggests that blocking the recruitment of either LGR5+ or FOXL2+ cells leads to a failure of primordial follicle formation. Therefore, we conclude that the proper differentiation and proliferation of ovarian supporting cells is essential for the establishment of the primordial follicle pool, and the recruited number and speed of somatic cells might decide the final size of female reproductive reserve. Previous studies have reported there are two waves of primordial follicle formation in the ovary. These two primordial follicle populations differ from each other in terms of their developmental dynamics and their roles in ovarian function (Mork et al., 2012; Zheng et al., 2014a,,b). In the current study, we found that inhibiting ADAM10–Notch signaling suppressed the formation of the majority of primordial follicle formation in the ovary. However, we still found that ∼20% of oocytes formed primordial follicles in the medulla when ADAM10–Notch signaling was blocked. Using a cell lineage tracing approach, Zheng et al. have shown that the follicle somatic cells from the first wave of primordial follicles are recruited before 16.5 dpc in fetal ovaries (Zheng et al., 2014a). Given that the inhibitors of ADAM10–Notch signaling were added into culture from 16.5 dpc in this study, here, we mainly focused on the adult wave of primordial follicles. It is, however, unknown whether the first wave of primordial follicles are controlled by similar molecular networks in their development, and more exploration is necessary.
In conclusion, we have supplied evidence showing that ADAM10 plays a key role in determining the recruitment of pregranulosa cells and early folliculogenesis in mice. This is achieved through ADAM10-mediated Notch signaling that regulates the proliferation and differentiation of LGR5+ follicle supporting progenitor cells in fetal ovaries. Our results indicate that the proper recruitment of ovarian follicle somatic cells is essential for the establishment of the primordial follicle pool in mice. These findings provide important new insights into the mechanisms of early ovarian development.
MATERIALS AND METHODS
All wild-type mice were obtained from the Laboratory Animal Center of the Institute of Genetics and Development Biology (Beijing, China). The Lgr5-EGFP-ires-CreERT2 (Lgr5-KI) reporter mice have been described in detail previously (Barker et al., 2007). Adam10loxp/loxp (Adam10f/f) mice were generously provided by Dr Xiaohui Wu from the Institute of Developmental Biology and Molecular Medicine at Fudan University, Shanghai, China (Tian et al., 2008). Amhr2-Cre mice (Jamin et al., 2002; Jorgez et al., 2004) and Ddx4 (Vasa)-Cre (Gallardo et al., 2007) mice were obtained from the Model Animal Research Center of Nanjing University, Nanjing, China. Ddx4-Cre mice were crossed with Adam10f/f mice to produce Adam10f/d mice. To generate conditional Adam10 knockouts in somatic cells, Amhr2-Cre mice were then crossed with Adam10f/d mice to produce Amhr2-cre;Adam10f/d mice (Fig. S4A). Genotyping was performed by PCR using the primer sets detailed in Table S1 (Fig. S4B). Female and male mice were caged at a ratio of 1:1 overnight and the presence of a vaginal plug was designated 0.5 dpc. The day of birth was considered 0.5 dpp. Birth usually occurred between 19.0 and 19.5 dpc. Protocols and the use of animals conformed to the guidelines and regulatory standards of the Institutional Animal Care and Use Committee of China Agricultural University.
Ovary isolation and culture
Ovaries were dissected from mice and ovarian capsules were removed in pre-cooled PBS (10 mM, pH 7.4) under a stereomicroscope. Ovaries were cultured in six-well culture dishes in 1.1 ml basic Dulbecco's modified Eagle's medium (DMEM) with F12 medium (GIBCO, Life Technologies) at 37°C in a 5% CO2, 95% air atmosphere with saturated humidity. The ovaries were turned every 12 h to ensure the all of the ovary cells spent half the time in the medium and half in the air.
To assess the role of ADAM10 in primordial follicle formation, we cultured 16.5 dpc ovaries (five ovaries/group) for 3 to 7 days in either medium alone, or in medium supplemented with the ADAM10 inhibitor GM6001 (sc-203979; Santa Cruz Biotechnology) or GI254023X (A4436; Apexbio). To analyze Notch signaling during primordial follicle formation, 16.5 dpc ovaries were cultured in the presence or absence of the Notch inhibitor DAPT (D5942; Sigma-Aldrich) for 3 to 7 days. For measurements in the bromodeoxyuridine (BrdU) incorporation assay, isolated ovaries from different days, or cultured ovaries with inhibitors, were cultured with BrdU (B5002; Sigma) for 1 h before ovaries were collected for analysis.
Histology and follicle characterization
Ovaries were fixed in 4% paraformaldehyde for over 24 h at 4°C and then processed to obtain 5-μm paraffin sections. Sections were stained with hematoxylin, and the numbers of oocytes and follicles were counted in every fifth section. To estimate the total number of oocytes and follicles in each ovary, the sum of counted oocytes and follicles was multiplied by five.
Immunofluorescence and labeled cell characterization
Ovaries were fixed in 4% paraformaldehyde for over 24 h at 4°C and processed to obtain 5-μm paraffin sections. To ensure the comparability of different ovaries, all ovaries were sliced through the long plane and were cut into kidney-shaped sections by adjusting the orientation of the ovary when embedded. Whole-ovary serial sections were made and ten sections in the middle of the band were selected for subsequent experiments. Sections were de-paraffinized, rehydrated and subjected to microwave antigen retrieval in 0.01% sodium citrate buffer (pH 6.0) for 16 min. Ovarian sections were then blocked with 10% normal donkey serum in PBS for 1 h at room temperature and incubated overnight at 4°C with primary antibody. Primary antibodies were: chicken anti-EGFP (ab13970; Abcam) used at 1:300; goat anti-FOXL2 (IMG-3228; Imgenex/Novus Biologicals) used at 1:400; rabbit anti-DDX4 (ab13840; Abcam) used at 1:300; and mouse anti-BrdU [G3G4; Developmental Studies Hybridoma Bank (DSHB)] used at 1:300 antibodies. After washing thoroughly with PBS, slides were incubated with Alexa-Fluor-350-, -488- or -555-conjugated donkey secondary antibodies against mouse, rabbit, goat or chicken IgG (1:100, Invitrogen/Life Technologies) for 1 h at 37°C. Slides were subsequently washed with PBS again and stained with Hoechst 33342 (B2261; Sigma) or propidium iodide (421301; BioLegend) as a nuclear counterstain. Sections were examined and photographed using a Nikon Eclipse 80i digital fluorescence microscope.
To count labeled cells, whole ovarian sections were photographed. The cortical region was defined as 25% of the whole ovary from the ovarian surface to the ovarian center in this study, which was based on the localization of LGR5+ cells, FOXL2+ cells and germline cysts (arrowheads) in fetal ovaries at 16.5 dpc (Fig. S2H). Other days of ovarian sections in vivo or in vitro also used this method to define regions for counting. All positive cells in the cortical region of each section were counted and analyzed. The images were merged to analyze double-labeled cells. At least three different sections were analyzed for each ovary.
Western blot analyses were conducted as described previously (Wang et al., 2015). Cultured ovaries were collected and lysed in WIP (Beijing CellChip Biotechnology Co.), according to the manufacturer's protocol. Electrophoresis was performed using 50 μg of total protein per sample separated by 15% SDS-PAGE and transferred to polyvinylidene fluoride membranes (IPVH00010, Millipore). Membranes were incubated overnight at 4°C with anti-Notch2 antibody (Abcam) diluted 1:200, which detects the Notch2 full-length band at 120 kDa and the Notch2 intracellular domain at 85 kDa. The secondary antibodies (ZB-2301, ZB-2305 from ZSGB-BIO) were diluted 1:5000 in TBST (TBS plus 0.5% Tween 20). The membranes were visualized using a Super Signal Chemiluminescent Detection System (34080; Thermo Scientific). The level of GAPDH (anti-GAPDH AM4300; Life Technologies; 1:4000) was used as an internal control.
Quantitative real-time RT-PCR
RNA was isolated from eight ovaries for each sample using TRIZOL (Invitrogen/Life Technologies) according to the manufacturer's protocol. Reverse transcription reactions were carried out using 1 μg RNA and a Promega Reverse Transcription System according to the manufacturer's instructions. qRT-PCR reactions were undertaken and analyzed using an ABI 7500 Sequence Detection System (Applied Biosystems). Data were normalized to results with Gapdh. The primers used for testing genes are listed in Table S2.
Infection of ovaries
Lentivirus EGFP vector was kindly provided by Professor Haibin Wang from the Institute of Zoology, Chinese Academy of Sciences, Beijing, China. Packaged virus (≥1×108 transducing units/ml) was diluted in DMEM/F12 (1:1) and 15.5 dpc ovaries were incubated in drops of 20 μl DMEM/F12 medium with lentiviral vectors for 12 h. Infected ovaries were then transferred into six-well culture dishes with 1.1 ml basic DMEM/F12 medium for 12 h to ensure full virus infection. Finally, the infected ovaries from one drop were, respectively, cultured in the presence and absence of inhibitors to examine the behavior of labeled cells.
LGR5–EGFP-positive ovaries from Lgr5-EGFP-ires-CreERT2 (Lgr5-KI) reporter mice were used in FACS analysis. After 7 days of culture in the control or inhibitor treatment group, ovaries were incubated with pre-warmed 0.25% trypsin (Invitrogen) at 37°C and 5% CO2 for 15 min. Then ovaries were dissociated into single-cell suspensions. Following trypsin inactivation with PBS supplemented with 10% fetal bovine serum (FBS), ovarian cells were pelleted at 1500 g for 5 min at 4°C, re-suspended in PBS with 5% FBS and were immediately analyzed in a FACS Calibur (BD Biosciences) machine.
In situ hybridization
In situ hybridization was performed as described previously (Zhang et al., 2015). Briefly, ovaries with ovarian capsules were dissected from mice and were embedded in optimal cutting temperature (OCT) compound and frozen in liquid nitrogen until analysis. Frozen sections (10 μm) of ovaries were hybridized with a mouse-specific DIG-labeled Adam10 probe according to the manufacturer's protocol. Primers used to generate the Adam10 amplicon were: forward 5′-TATGGTGGAGTTGGTTCTTA-3′ and reverse 5′-GCAGGCTTGAAGTATATGG-3′. Sections hybridized with sense probes served as negative controls and did not show positive signals.
All culture, immunochemical and immunofluorescence analyses were repeated at least three times using ovaries from different fetuses. Data are expressed as means±s.d. with each experiment performed in triplicate. Data were analyzed by t-test. Values of P<0.05 were considered statistically significant.
We are grateful to Professor Haibin Wang's laboratory (State Key Laboratory of Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, China) for help in experimental techniques. We are also grateful to Dr. Xiaohui Wu (Institute of Developmental Biology and Molecular Medicine, Fudan University, Shanghai, China), Professor Enkui Duan (State Key Laboratory of Reproductive Biology, Institute of Zoology, Chinese Academy of Sciences, China) and Dr. Ting Chen (National Institute of Biological Sciences, Beijing, China) for providing transgenic mice.
L.F., H.Z. and G.X. designed the work, with input from other authors. L.F., Y.W., H.C., G.S., X.T. and J.Z. performed the experiments. L.F, Y.W., W.N., Q.X. and H.Z. analyzed the data and contributed reagents, materials and/or analytical tools. The manuscript was written by L.F. and revised by C.W., H.Z. and G.X.
This study was supported by grants from the National Natural Science Foundation of China [grant numbers 31571542 to H.Z., 31571540 to G.X.]; and the National Basic Research Program of China (973, part of the Ministry of Science and Technology of the People's Republic of China) [grant numbers 2013CB945501, 2012CB944701 to G.X.].
The authors declare no competing or financial interests.