Oocyte-specific deletion of furin leads to female infertility by causing early secondary follicle arrest in mice

The process of follicular development involves communications between oocyte and surrounding granulosa cells. FURIN is a member of the family of proprotein convertases that is involved in the activation of a large number of zymogens and proproteins by cleavage at its recognition motif. To investigate the functions of FURIN in female fertility, furinflox/flox (furfl/fl) mice were crossed with Zp3-Cre mice and Gdf9-Cre, respectively, to achieve oocyte-specific disruption of FURIN. Here we report for the first time that FURIN is dispensable for primordial follicle maintenance and activation but important for early secondary follicular development, as ablation of FURIN in oocytes caused failure of follicle development beyond the type 4 and/or 5a follicles in mutant mice, resulting in increased number of early secondary follicles and the severely decreased number of mature follicles, thus anovulation and infertility. We also found that the developmental arrest of early secondary follicles might be rooted in the loss of the mature form of ADAMTS1 (85-kDa prodomain truncated) and compromised proliferation of granulosa cells in mutant mice. Taken together, our data highlight the importance of FURIN in follicle development beyond the early secondary follicle stage and indicate that compromised FURIN function leads to follicular dysplasia and female infertility in mice.

In most cases, female infertility can be caused in many processes, including folliculogenesis, ovulation, early embryogenesis, implantation, and so on. However, premature ovarian insufficiency or premature ovarian failure with defects in follicular development in the ovary has become the main cause of female fertility, which troubles nearly 1% women under 40 years old. [1][2][3][4] The follicle, which consists of an oocyte surrounded by one or more layers of granulosa cells, is the functional content of the ovary. Follicles are distinguished as primordial follicles, primary follicles, secondary follicles and antral follicles based on the developmental stage. According to the well-accepted standards established by Pederson and Peters, 5 the secondary follicles can be divided into early secondary and late secondary follicles according to the number of granulosa layers. 6 Many proteins are synthesized as proproteins; FURIN has a variety of proproteins as its substrates. [7][8][9][10][11] FURIN has been shown to have multiple functions on the proteolytic cleavage of HIV envelope polyprotein precursor gp160 to gp120 and gp41 before viral assembly, and it participates in cleaving other proteins into their mature/active forms. [12][13][14][15][16] FURIN is ubiquitously expressed and its null mutant mice die by embryonic day 11.5. 17 Conditional knockout model shows that FURIN has limited redundancy in mouse liver 18 and FURIN is involved in granular acidification in endocrine pancreas by processing pro-AC45. 19 Further, a previous report shows that in T cells without FURIN, ADAMTS6 is downregulated, which is in keeping with the theory that the PCSK-targeted molecules are coordinately regulated by their converting enzyme. 20 In spite of the increasing number of reports about FURIN, the function of furin in folliculogenesis and female fertility is unknown.
ADAMTS-1 (A disintegrin-like and metalloproteinase with thrombospondin type I motifs-1) is a metalloprotease of the family of genes that contain metalloprotease, disintegrin and thrombospondin-like motif. 21,22 The human ortholog of ADAMTS-1 is named METH-1. The ADAMTS1 is synthesized as zymogen that is 110 kDa. 23 The mouse ADAMTS-1 gene was first described as a transcript, which is highly expressed in a cachexigenic colon tumor cell line. 24 ADAMTS1 is implicated in many physiological processes. For example, ADAMTS1 has a negative role in angiogenesis, as proved by a previous study. 22,25 Furthermore, Shindo et al. 26 reported that ADAMTS-1-null mice showed renal anomalies and female infertility. Subsequently, it was shown that ADAMTS1 participated in normal follicular development, in the ovulatory process, and in the organization of the medullary vascular network in the ovary. 27 The number of follicles at type 5b (late preantral) and later stages was markedly reduced in ADAMTS-1-null mice. On the other hand, a previous study demonstrated that the cleavage of the prodomain from the ADAMTS1 precursor was impaired in the FURIN-deficient cell line. 28  The aim of our present study was to determine the roles of FURIN in oogenesis and embryo development, and we adopted the Gdf9-Cre/loxP system and Zp3-Cre/loxP system, in which furin was selectively deleted in mouse oocytes of primordial follicles and primary follicles. We found that furin deletion caused female infertility and follicle arrest before the type 4 and/ 5a stage by causing failure of pro-ADAMTS1 proteolytic processing to ADAMTS1 by FURIN.

Results
Localization and oocyte-specific deletion of furin. To study the physiological roles of FURIN during oogenesis and embryogenesis, we first carried out immunofluorescent staining experiments by using an antibody against FURIN to show the subcellular localization of FURIN. As shown in Figure 1a, FURIN protein was localized in the cell membrane in GV oocyte. Meanwhile, immunohistochemical (IHC) analysis showed that FURIN was expressed in oocytes from the primordial follicles to mature follicles (Figure 1b).
FURIN is ubiquitously expressed in mice, and its null mutant mice die by embryonic day 11.5. To investigate the in vivo function of FURIN, we adopted the Cre-loxP system to achieve oocyte-specific disruption of FURIN from primary follicle stage. The mutant mice, in which exon 2 of the furin gene is targeted, 18 was generated by crossing FURIN flox/flox (fur fl/fl ) mice with Zp3-Cre transgenic mice, which expressed Zp3 promoter-meditated Cre recombinase in oocytes of primary follicles after postnatal day 5 and in later developmental stages. 29 Quantitative real time-PCR (qRT-PCR) test showed Deletion of furin in oocytes causes anovulation and infertility. To identify the effect of oocyte-specific ablation of FURIN in female fertility, we carried out a breeding assay by crossing fur fl/fl or fur fl/fl ;Zp3-Cre female mice with fertile males for 6 months. As shown in Figure 2a, female fur fl/fl ;Zp3-Cre mice were completely infertile, while the fur fl/fl mice possessed normal fertility. The mutant mice (postnatal day 42) had smaller ovaries and failed to ovulate during a natural ovulation assay (Figure 2b). We speculated that the follicular development was abnormal in fur fl/fl ;Zp3-Cre mice. When we carried out superovulation to induce ovulation with exogenous PMSG and hCG, the mutant mice at the age of 6 weeks ovulated fewer oocytes than the control (Figure 2c). The MII oocytes collected from mutant mice after superovulation displayed poor quality, showing degenerated polar bodies and severely abnormal spindle (Figure 2d). Further immunofluorescent staining showed abnormal spindles in fur fl/fl ;Zp3-Cre oocytes ( Figure 2e). Importantly, the mutant mice at the age of 8 weeks and older could not be induced to ovulate anymore (data not shown).
Oocyte-specific deletion of furin causes significant decrease of late growing follicles but increase of early secondary follicles. The mutant mice showed anovulation in natural ovulation and significantly decreased the number of ovulated oocytes in a superovulation assay after sexual maturation. Histological analysis of ovaries showed that there  The communication between oocytes and surrounding granulosa cells is achieved by gap junctions, which consists of connexin 37. 30,31 We suspected that the gap junctions between oocytes and granulosa cells were impaired in mutant mice. IHC analysis of connexin 37 confirmed this speculation. The connexin 37 showed normal presence in the early secondary follicles in fur fl/fl ovaries (Figures 4e and g). On the contrary, connexin 37 was almost absent in the early secondary follicles (Figures 4f and h). These results suggested that the gap junctions between oocytes and surrounding granulosa cells were impaired since the early secondary follicle stage.
Oocyte-specific deletion of furin causes the loss of the mature form of ADAMTS1 in oocytes. To investigate the mechanism underlying furin-deletion-caused early secondary follicle arrest, we started to investigate the substrates of FURIN based on its function implicated in processing latent precursor proteins into their biologically active products. It has been reported that ADAMTS-1-null mutant mice have defects in female fertility and decreased the number of growing follicles. 26,27 These results demonstrated that ADAMTS1 is important for follicle development beyond early secondary follicles. Therefore, we determined the mRNA and protein level of ADAMTS-1 by qRT-PCR and western blot, respectively. ADAMTS-1 mRNA level increased more than ninefold ( Figure 5a); however, the mature form of ADAMTS1 was nearly  chloromethylketone that binds to the catalytic site of FURIN to block its activity for 24 h. Consistent with our results in vivo, pro-ADAMTS1 cleavage was efficient in untreated oocytes. However, oocytes treated with decanoyl-RVKR-CMK showed complete inhibition of FURIN activity and only pro-ADAMTS1 was detected (Figure 5c and Supplementary Figure S1B). These data suggest that the phenotype of fur fl/fl ;Zp3-Cre might have resulted from the lack of mature ADAMTS1.
Deletion of furin in oocytes does not affect the activation of primordial follicles. In female fur fl/fl ;Zp3-Cre mice, when the furin gene is targeted in oocytes of primary follicles, the follicles are arrested at the early second follicle stage. Thus, we wondered whether FURIN had roles in primordial follicle activation. We generated fur fl/fl ;Gdf9-Cre female mice by crossing fur fl/fl with Gdf9-Cre transgenic mice to identify whether disruption of FURIN in oocytes had an effect on primordial follicles. IHC analysis of FURIN showed the successful deletion of furin in oocytes (Figure 6a). In accordance with infertility in fur fl/fl ;Zp3-Cre mice, the fur fl/fl ; Gdf9-Cre female mice were also infertile. Histological analysis of ovaries in fur fl/fl ;Gdf9-Cre mice showed apparent increase in early secondary follicles at 6 weeks of age, which was the same as the phenotype in fur fl/fl ;Zp3-Cre mice (Figure 6b). We then performed immunostaining for the germ cell marker mouse VASA homolo) on 2-month-old ovarian sections. As shown in Figure 7a, primordial follicles were mostly scattered around the cortical region in fur fl/fl ;Gdf9-Cre ovaries as well as in fur fl/fl and fur fl/fl ;Zp3-Cre mice. Consistent with immunohistological staining, the quantitative analysis demonstrated that there was no difference in the number of primordial follicles in the ovaries, where all the genotypes contained almost the same amount of primordial follicles (Figure 7b). These results suggested that deletion of furin in oocytes did not have an evident influence on the survival and activation of primordial follicles in ovaries of fur fl/fl ;Zp3-Cre mice and in fur fl/fl ;Gdf9-Cre mice. The early secondary follicle arrest was not caused by excessive activation of primordial follicles.
The proliferation of granulosa cells is compromised in mutant mice. Given the impaired gap junctions between oocytes and surrounding granulosa cells in mutant mice, we suspected that furin deletion from oocytes might impact the proliferation of granulosa cells. To test this, we performed   (Figure 8c). These data demonstrated that there was a profound defect in granulosa cell proliferation in fur fl/fl ; Zp3-Cre mice and in fur fl/fl ;Gdf9-Cre mice, which might lead to the early secondary follicles arrest.

Discussion
Follicular recruitment, frequently used by different investigators to describe two important points during follicular development, includes initial recruitment and cyclic recruitment. 32 The primordial follicles are continuously recruited into the growing follicle pool during initial recruitment, whereas a cohort of follicles is recruited to develop into antral follicles per menstrual cycle. The initial recruitment of primordial follicles and the development of secondary follicles to the antral stage are critical during folliculogenesis. 7 Although the former was recently intensively investigated by others and us, 4,6,33-35 the mechanism involved in the subsequent development of early secondary follicles remained poorly understood. Ogiwara et al. 36 reported the greatest expression of furin mRNA in oocytes of small growing follicles compared with large growing follicles and antral follicles, indicating that FURIN might have an important role in further development of small growing follicles. In the present study, by using a mutant mouse model with oocyte-specific deletion of furin, we demonstrated that FURIN is essential for follicular development beyond the early secondary follicles.
To explore the reason for the rapid decrease of late growing follicles from PD30 to PD42, we conducted an apoptosis test. The TUNEL staining showed that the large follicles beyond the early secondary follicles were induced to undergo apoptosis, suggesting a defective effect in those follicles. Furthermore, the gap junctions between oocytes and surrounding granulosa cells were impaired from the early secondary follicles, which may affect the normal follicular development beyond the early secondary follicles.
Although the apoptosis staining explained the decreased number of large follicles, it could not account for the increased number of early secondary follicles. To investigate the mechanisms underlying the developmental arrest of early secondary follicles, we designed experiments to find potential molecules responsible for this phenotype. It has been reported that FURIN is involved in processing latent precursor proteins into their biologically active products. It has further been reported that female ADAMTS-1-null mutant mice have severe defects in fertility and significantly decreased the number of late growing follicles, suggesting that ADAMTS1 is important for follicle development beyond the type 4 and/or type 5a (early secondary follicles). 27 Interestingly, ADAMTS1 is synthesized as pro-ADAMTS1. In accordance with ADAMTS-1-null phenotype in the ovary, oocyte-specific ablation of FURIN caused the developmental arrest of early secondary follicles. Based on this, we determined whether the mRNA and protein level of ADAMTS-1 was affected in oocytes in mutant mice. Surprisingly, although the ADAMTS-1 mRNA level was increased up to ninefold, the mature ADAMTS1 protein was nearly lost upon the disruption of FURIN in oocytes of mutant mice. Therefore, we suggest that a defect in the mature ADAMTS1 could be the most probable cause. In addition, we should note that there is no accumulation of pro-ADAMTS1. In fact, we did not know the reason why there is no accumulation of pro-ADAMTS1 and even decrease. To our knowledge, Louagie et al. 19 investigated the role of FURIN in the endocrine pancreas by conditional knockout model and found that there was no accumulation of proinsulin II and pro-PC2 in spite of impaired processing of proproteins. We suggest that there might be one possible mechanism for control of the translation efficiency in vivo, which remains to be studied.
To investigate whether ablation of FURIN also affects the primordial follicles recruited into the growing follicle pool during the initial recruitment, we generated fur fl/fl ;Gdf9-Cre female mice to achieve deletion of furin from primordial follicles. The fur fl/fl ;Gdf9-Cre female mice were also infertile and showed developmental arrest of early secondary follicles, similar to the phenotype in fur fl/fl ;Zp3-Cre female mice. The IHC staining of the germ cell marker VASA confirmed that there was no influence on the primordial follicle pool.
Normal communication between oocytes and surrounding granulosa cells as well as the granulosa cell proliferation are indispensable for normal folliculogenesis. ADAMTS1 is implicated in many physiological events, such as angiogenesis inhibition, various inflammatory processes and development of cancer cachexia. It has been reported that ADAMTS-1-null mice showed female infertility and impaired follicular development. 26,27 Subsequently, Brown et al. 37 demonstrated that ADAMTS1 was required for extracellular matrix (ECM) remodeling during ovarian follicular development and formation of lymphatic vessels. 38 In accordance with this, Russell et al. 39 showed that pro-ADMATS1 was localized to cytoplasmic secretory vesicles and the mature form of ADAMTS1 was accumulated in ECM. Particularly, it has be demonstrated that ADAMTS1 is an ECM-bound metalloproteinase, 26 and ECM components and ECM-associated molecules, such as proteoglycans, aggrecan and versican, are targets of ADAMTS1. [40][41][42] Meanwhile, Horiguchi et al. 43 reported that ECM component laminin promoted the formation of gap junction in the rat anterior pituitary gland, which inspired us to investigate the gap junction between oocytes and granulosa cells. IHC analysis of connexin 37 showed impaired gap junctions between oocytes and granulosa cells. IHC staining of the cell proliferation marker PCNA indicated that the proliferation of granulosa cells was compromised, which might partially explain the developmental arrest of early secondary follicles.
In summary, FURIN is indispensable for female fertility and FURIN-ADAMTS1 is potentially partnering as an enzyme substrate in oocytes, which may be responsible for follicular development beyond the early secondary follicles. Our study may provide an important in vivo model for further understanding the mechanism of early-stage follicular development. Fertility assessment, natural ovulation and superovulation analysis. To assess the reproductive activity, seven individually housed control or mutant mice were mated with a male mouse validated to be fertile at the age of 6 weeks. Cages were examined two times a week and the number of pups was recorded up to 6 months. For the natural ovulation assay, 6-9-week-old female mice were examined for the estrous cycle every day. Females in the estrus were mated with a fertile male. The next day in the morning, the oviducts were dissected from the female with the plug, and zygotes were collected and counted. To induce synchronized Figure 8 The proliferation of granulosa cells is compromised in mutant mice. Proliferating cell nuclear antigen (PCNA) immunohistochemistry showing the status of granulosa cell proliferation in follicles. Black arrowheads point to the PCNA staining in granulosa cells in early secondary follicles. Scale bar = 100 μm follicular growth and ovulation, each 6-9-week-old female mouse was injected with 10 IU PMSG, followed by 10 IU hCG after 48 h to promote ovulation. The oviducts were dissected and cumulus-oocyte complexes were collected at 16 h of hCG treatment. After a 3-min treatment with 0.5 mg/ml hyaluronidase (Sigma, St. Louis, MO, USA) in the M2 medium, oocytes were collected, counted and then analyzed.
Oocyte collection and culture. The GV oocytes were isolated from ovaries of 3-to 4-week-old female mice. MII stage oocytes were isolated from 3-to 6-weekold female mice. Oocytes were cultured in the M2 medium (Sigma) under paraffin oil at 37°C with 5% CO 2 in air.
Histological analysis and IHC staining of ovary. Ovaries were fixed in 4% paraformaldehyde (pH 7.5) overnight at 4°C, dehydrated in 50%, 70%, in a graded ethanol series, cleared in xylene three times and finally embedded in paraffin wax. The treated ovaries were sectioned consecutively at 8 μm for hematoxylin and eosin (H&E) staining and IHC staining. Serial ovarian sections were photographed and follicles were counted in every fifth sections using the Image J software and cell counter plugin.
Quantification of ovarian follicles. Quantification of ovarian follicles was performed as previously described by Liu et al. 4 Briefly, to count the number of follicles, paraffin-embedded ovaries were serially sectioned at 8-μm thickness and every fifth section was mounted on slides. Then, these sections were stained with H&E for morphological analysis. Ovarian follicles at different developmental stages, including primordial follicles, primary follicles (type 3), early secondary and late secondary follicles (type 4 and type 5) and antral follicles (type 6 and type 7) were counted in collected sections of an ovary, based on the well-accepted standards established by Pederson and Peters. 5 In each section, only those follicles in which the nucleus of the oocyte was clearly visible were scored and the cumulative follicle counts were multiplied by a correction factor of 5 to represent the estimated number of total follicles in an ovary.
Quantitative real time-PCR. RNA was isolated from~100 oocytes for each group and reverse transcription reaction was carried out using the RNeasy Kit TUNEL assay. Analysis of apoptosis in ovarian follicles was carried out by TUNEL assay using the ApopTag Plus In Situ Apoptosis Detection Kit (Chemicon International, Temecula, CA, USA). At least four different specimens from each genotype were analyzed in parallel.
FURIN-specific inhibition. Decanoyl-RVKR-CMK was obtained from Merck Millipore (no. 344930, USA). The FURIN inhibitor was dissolved in dimethyl sulfoxide at a concentration of 25 mM, and this stock solution was added directly to the M2 medium. The GV oocytes were treated with 200 μM decanoyl-RVKR-CMK for 24 h.
Western blot analysis. A total of 200 oocytes per sample were mixed with 2 × SDS sample buffer and boiled for 5 min at 100°C for SDS-PAGE. Western blotting was performed as described previously, 44  Statistical analysis. All data presented were collected in at least three independent experiments and analyzed using SPSS (SPSS, IBM, USA). Data were expressed as mean ± S.E.M. and significance of differences was evaluated with Student's t-test.