Granulosa secreted factors improve the developmental competence of cumulus oocyte complexes from small antral follicles in sheep

Oocyte in vitro maturation can be improved by mimicking the intra-follicular environment. Oocyte, cumulus cells, granulosa cells, and circulating factors act as meiotic regulators in follicles and maintain oocyte in the meiotic phase until oocyte becomes competent and ready to be ovulated. In a randomized experimental design, an ovine model was used to optimize the standard in vitro maturation media by Granulosa secreted factors. At first, the development capacity of oocyte derived from medium (>4 to 6 mm) and small (2 to ≤4 mm) size follicles was determined. Differential gene expression of granulosa secreted factors and their receptors were compared between the cumulus cells of the two groups. Then, the best time and concentration for arresting oocytes at the germinal vesicle stage by natriuretic peptide type C (CNP) were determined by nuclear staining in both groups. Oocyte quality was further confirmed by calcein uptake and gene expression. The developmental competence of cumulus oocyte complexes derived from small size follicles that were cultured in the presence of CNP in combination with amphiregulin (AREG) and prostaglandin E2 (PGE2) for 24 h was determined. Finally, embryo quality was specified by assessing expressions of NANOG, SOX2, CDX2, OCT4, and TET1. The cumulus oocyte complexes derived from small size follicles had a lower capacity to form blastocyst in comparison with cumulus oocyte complexes derived from medium size follicles. Prostaglandin E receptor 2 and prostaglandin-endoperoxide synthase 2 had significantly lower expression in cumulus cells derived from small size follicles in comparison with cumulus cells derived from medium size follicles. Natriuretic peptide type C increased the percentage of cumulus oocyte complexes arresting at the germinal vesicle stage in both oocytes derived from medium and small follicles. Gap junction communication was also improved in the presence of natriuretic peptide type C. In oocytes derived from small size follicles; best blastocyst rates were achieved by sequential exposure of cumulus oocyte complexes in [TCM+CNP (6 h), then cultured in TCM+AREG+PGE2 (18h)] and [TCM+CNP (6 h), then cultured in conventional IVM supplements+AREG+PGE2 (18h)]. Increased SOX2 expression was observed in [TCM+CNP (6 h), then cultured in TCM+AREG+PGE2 (18h)], while decreased OCT4 expression was observed in [TCM+CNP (6 h), then cultured in conventional IVM supplements+AREG+PGE2 (18h)]. It seems that the natriuretic peptide type C modulates meiotic progression, and oocyte development is probably mediated by amphiregulin and prostaglandin E2. These results may provide an alternative IVM method to optimize in vitro embryo production in sheep and subsequently for humans.


Introduction
Perhaps the greatest challenge of assisted reproductive techniques is in vitro maturation. Edwards reported the first IVM in humans, and Cha showed the first live birth after IVM in a woman with premature ovarian failure (POF). Since then, a great result has been achieved by capacitation prematuration-IVM (CAPA-IVM). Despite this recent progress in CAPA-IVM, this method has not replaced the conventional IVM procedure in humans and other species [1][2][3].
Improved development of in vitro matured oocytes can only be achieved by expanding our knowledge regarding the complex dialog between the oocyte and its surrounding somatic cells within follicle [4,5]. Nevertheless, this has proven to be challenging because: 1) dealing with a heterogeneous population of oocytes harvested from follicles of different size (2-6 mm in sheep and goat, 2-8 mm in bovine and < 12mm in human) is a formidable issue; 2) cumulus oocyte complexes (COCs) derived from small antral follicles possess less competence in response to regulatory and specific ligands because of an immature signaling capacity [6][7][8][9][10]; and 3) Isolation of COCs from their natural follicular environment results in spontaneous meiotic progression and, thus, asynchronization of nuclear and cytoplasmic maturation [11].
vitro oocyte maturation, especially in COCs derived from small follicles (2 to � 4mm) with lower developmental competence compared to follicles of � 4 to 6mm or of greater size [29,30]. Our goal is to select ovine as a model for human IVM in future studies.
To achieve this aim, we initially exposed COCs from small follicles to CNP for 6 h and then to TCM 199 as a base of routine culture medium in the presence of AREG and/or PGE2, and the results of each group were compared with their corresponding group.

In vitro maturation (IVM)
Ovine ovaries from a local abattoir were transported to the laboratory at 6:00 pm in saline (15˚C-20˚C) and stored for an additional 12 h at 15˚C since the lab working hours start in the morning [31,32]. COCs were isolated from medium (� 4 to 6 mm) and small (2 to � 4 mm) size follicles with the aid of 21-G needles. COCs with more than three layers of cumulus cells and homogenous cytoplasm were washed with the HEPES tissue culture medium 199 (HTCM199) + 1 mg/mL PVA+ 4 mg/mL BSA. Finally, 50 COCs were cultured in 400 μl containing 1 mg/mL PVA+ 8 mg/mL BSA in each group, according to experimental design (Fig 1) in 4 well dishes without mineral oil for 24 h at 38.5˚C and 5% CO 2 in the air [33,34].

Nuclear status
COCs were treated with 300 IU/mL hyaluronidase and vortexes to remove CCs. Denuded oocytes (DOs) were subsequently fixed for 20 minutes in 4% paraformaldehyde. To visualize chromatin, DOs were stained with Hoechst 33342 (10 μg/mL) for 5 minutes. After mounting, images of stained oocytes were captured and assessed by high-resolution digital camera (DP-72 Olympus, Japan) using DP2-BSW software [35]. The percentage of oocytes at the GV stage in each group was determined.

Cumulus expansion index (CEI)
After 24 h IVM, CEI was scored on a 0 to 4 scale, as described by Vanderhyden et al., 1999: score 0, no expansion; score 1, no CC expansion but cells appear as spherical; score 2, only the outermost layers of CCs have expanded; score 3, all cell layers have expanded except the corona radiata; and score 4, expansion has occurred in all cell layers including the corona radiate [36].

PLOS ONE
Sequential maturation of sheep oocyte 10 μg/ml, 6 mg/ml BSA) for 20 h at 38.5˚C under 5% CO 2 in humidified air overlaid with light mineral oil. On the next day, to remove the cumulus cells, the presumptive zygotes were vortexed in HTCM199 + 1mg/ml PVA+ 4mg/ml BSA for 3 minutes [37]. Then, they were cultured for 8 days in BO-IVC (Brackett-Oliphant in vitro culture) medium (Bioscience, UK) at 39˚C, 6% CO 2 , 5% O 2 in humidified air under mineral oil. Day 0 was defined as the day of insemination. Therefore, cleavage and blastocyst rates (over cleavage) were determined on days 3 and 7 post embryo culture.

Relative gene expression
In each group, at the desired time, CCs were collected from COCs after vortexing for 4 minutes. Then, CCs were separated from oocytes, and CCs were stored in RLT buffer at -70˚C until RNA extraction. RNA was extracted from CCs with the aid of an RNeasyMini Kit (Qiagen) followed by DNase I (Fermenas; EN0521) treatment. Total RNA (1000 ng) was reversetranscribed using a Takara cDNA Synthesis kit (Takara; #RR07A) according to the manufacturer's instructions. The primers were designed using Beacon and Oligo7 softwares. Efficiency correction for each primer was performed by serial dilution of positive control cDNA as a template. Please note that we are comparing the level of expression within different periods; we have no control group. Therefore, the transcripts abundance of 6 genes (AREG, EGFR, NPPC, NPR2, PTGER2, PTGS2, GJA4, and GJA1) ( Table 1) was normalized to B-actin as reference gene using 2^-(delta CT) rather than using 2^-(delta delta CT) .

Statistical analysis
Data percentages were modeled to a normal distribution by ArcSin transformation. Cleavage, blastocysts rates, and relative gene expression results among more than three groups were examined using a one-way ANOVA followed by Tukey's post hoc tests (Figs 3-5, Fig 7 and Fig  9). The t-test was used for cleavage, blastocysts rates, and relative gene expression data between two groups (Fig 2, Fig 6 and Fig 8). The differences were considered significant at P<0.05. All results were presented as means ± standard error of the mean (SEM). No significant difference was found for AREG gene expression between CCs derived from two groups at 3 time points. EGFR gene expression in CCs derived from medium size follicles at 12 h post-maturation was significantly higher than small size follicles (p<0.05). No significant difference was found for NPPC gene expression between CCs derived from two groups at 3 time points. NPR2 gene expression significantly declined during the first 12 h of maturation in CCs derived from small size follicles (p<0.05). Expressions of PTGER2 and PTGS2 genes between two groups at 3 time points were lower in CCs derived from small size follicles, except for PTGER2 that had higher expression in small compared with medium size follicles at 0h post-maturation (p<0.05, Fig 3). Developmental competency of in vitro conventional maturation of COCs derived from medium (>4 to 6 mm) compared to small (2 to �4 mm) size follicles. 5 replicates and minimum number of oocytes in each replicate were 50. Culture medium was conventional maturation medium containing TCM + LH (10μg/ml) + FSH (10μg/ml) + IGF (100ng/ml) + EGF (10ng/ml) + Cys (0.1mM) + FBS (15%) + E2 (1μg/ml). The asterisk represents a significant difference (P<0.05) within columns with the same color.

Effects of CNP on meiotic arrest, relative gene expression, and gap junction's communications
As the concentration of CNP increased (10, 100, 1000 nM), the percentage of oocytes derived from medium follicles remaining at the GV stage increased (P<0.05). In another group, there was no significant difference between 100 and 1000 nM of CNP (P>0.05, Fig 4). According to these results, concentrations of 1000 nM and 100 nM for CNP were selected for both medium and small size follicles, respectively [ Meiotic progression of oocytes derived from medium follicles occurred around 6 h after the onset of IVM in TCM. The percentage of oocytes derived from medium size follicles at the GV   Meiotic progression of oocytes derived from small size follicles occurred around 8 h after the onset of IVM in TCM. The percentage of oocytes derived from small size follicles at the GV stage after 24 h IVM was 20.9 ± 2 in TCM ( Fig 5B). In this group, in the presence of CNP (100 nM), the percentage of oocytes arrested at the GV stage after 24 h IVM was 37.3 ± 10.3 [ Fig 5B, see the experimental design (Fig 1, part II)]. Our results also showed E2 independently or in the presence of CNP did not affect meiotic arrest (Fig 5A and 5B).
Both gene expression of gap junction protein alpha 1 and 4 (GJA1, GJA4) in cumulus cells and calcein uptake through gap junctions in COCs derived from small and medium size follicles increased in the presence of CNP. But in the oocyte, GJA4 gene expression in the presence of TCM or CNP in each group did not show any significant difference [Fig 6, see the experimental design (Fig 1, part II)].

Discussion
One of the limiting factors in the application of IVM is a heterogeneous stage of follicles developing within or between ovaries. The heterogeneous populations of COCs derived from one ovary or different ovaries or from different abattoir have various developmental capacities or outcomes [10,38]. Differences in morphology, adenosine triphosphate (ATP) content, metabolism, mitochondria distribution, protein, mRNA pattern, and methylation lead to different outcomes in in vitro production of the embryo (IVP) [10,39]. Consistent with previous reports, in the present study, a reduction was observed in the developmental potential of ovine oocytes derived from small (2 to � 4 mm) compared to medium (>4 to 6 mm) follicles, concluding that the follicular size affects developmental competency of COCs. It is well defined that the isolation of COCs from the natural follicular environment results in spontaneous meiotic progression [40]. Granulosa cells under physiological conditions regulate the expression of transcripts like NPPC/NPR2, PTGER2 /PTGS2, AREG/EGFR involved in oocyte maturation, cumulus expansion, and ovulation through autocrine or paracrine mediators rather than circulating hormones [10, 28]. As we have shown, expression of these transcripts, especially PTGER2 /PTGS2, is lower in CCs isolated from small size follicles compared to medium ones. Also, our data showed that the relative NPR2 transcript level decreased by passing of time during maturation in CCs derived from small follicles. Thus, the addition of these factors during IVM may improve IVP outcomes.
When CNP binds to guanylyl cyclase (GC)-coupled NPR2, cGMP production increases, and oocytes are maintained at the GV stage [41]. Therefore, it is not surprising to see that the CNP or Npr2 knockout rodent model results in an early resumption of meiosis, concluding the role of CNP in meiotic arrest. Similar functions for NPs were reported in mice, human, goat, cattle, pig, sheep, rat, and cat [42,43]. In summary, the higher mRNA abundance of NP receptors in the dominant or medium follicles compared to the subordinate or smaller size follicles may be used as an indicator of follicle health and suggests that NPs signaling may regulate steroidogenesis and/or cell proliferation and differentiation [44]. The present study showed that the addition of 1000 nM and 100 nM CNP could arrest the majority of COCs derived from medium and small follicles for 24 h and 12 h, respectively. Recently, some studies have shown 6 h of culture with CNP (100 nM) can enhance oocyte development in cattle and goat [45,46]. Furthermore, markers of gap junction communication (GJA4 and GJA1) increased in the presence of CNP. Moreover, based on literature, chromatin configuration changes from non-surrounded nucleolus (NSN) to surrounded nucleolus (SN) during oocyte maturation, and this is regulated in vitro by NPs [47,48]. Recently, a novel mechanism for the CNP-induced oocyte meiotic arrest has been introduced in bovine. Based on these results, bovine oocytes have NPR2 receptors and can mediate meiotic arrest [23]. Furthermore, inhibition of meiotic resumption with NPs instead of synthetic reagents like forskolin, PDE inhibitors and specific inhibitors of cyclin-dependent kinases (CDKs) and meiosis promoting factor (MPF) have been shown to have better outcomes as they maintain gap junction activity, and also support key gene expression, which are critical for oocyte development [49][50][51]. In addition, it has been stated that E2 can mediate NPR2 gene expression and may affect meiotic arrest [22]. Our results showed E2 independently or in the presence of CNP did not affect the meiotic arrest.
On the other hand, the final stage of oocyte maturation and ovulation is mediated by EGFlike peptides after surging with gonadotrophins. The mRNA levels of CNP in granulosa and Npr2 in cumulus cells reduce after LH/hCG treatment despite stimulation of EGF-like factors and activation of EGFR [44,52]. But, according to our observation and other studies, the level of EGFR transcript decreases in the final stage of in vitro maturation in CCs derived from small size follicles. Therefore, it has been stated that the addition of AREG during maturation enhances bovine and porcine oocyte developmental competence [17,42].
In ovulating follicles, the EGFR signaling cascade involves many other signaling networks in cumulus and granulosa cells, which participate in the development of oocyte competence. Follicle-stimulating hormone (FSH) mediates the induction of AREG mRNA via P38 mitogen-activated protein kinases (p38MAPK). AREG also induces PTGS2 expression via ERK1/2. PGs, also acting via PTGER2 in cumulus cells, provide a secondary, autocrine pathway to regulate expression of AREG in COCs. PGE2 acts on a group of G-protein-coupled receptors and can maintain high cAMP levels. PGE2 in our study resulted in mild cumulus expansion, similar to cow [53]. Previous findings indicate that the addition of PGE2 during IVM improves embryonic cell survival of blastocysts and post-hatching development. In this regard, PTGS2 -/mice present severe failure in the expansion of cumulus cells and extrusion of the first polar body. Our results also indicated that the addition of PGE2 improves cumulus expansion, percentage of MII oocytes, and developmental competency. Besides, we showed that all three factors (AREG, PGE2, and CNP) independently improved the development of COCs derived from small size follicles with TCM as a base medium [8,14,17,43].
We also revealed that sequential exposure of ovine COCs to CNP then to AREG and/or PGE2 improved rate of blastocyst formation in [TCM+CNP, then cultured in TCM+AREG], [TCM+CNP, then cultured in TCM+AREG+PGE2], and [TCM+CNP, then cultured in conventional IVM supplements + AREG + PGE2] compared to TCM, with the best result observed in the latter group. Amongst the treatment groups, the latter also showed higher CEI compared to conventional IVM and [TCM + CNP, then cultured in TCM+AREG+PGE2].
Expansion of the extra-cellular matrix during oocyte maturation is essential for acquiring molecular machinery competence in vivo but in vitro overexpansion by promoting the hexosamine biosynthesis pathway (HBP) negatively influences in vitro oocyte competence [47,[54][55][56]. Therefore, despite higher CEI in this group, super developmental capacity was not observed.
Expression of embryo quality markers at RNA level revealed that their patterns were not different between conventional IVM, [TCM+CNP, then cultured in TCM+AREG+PGE2], and [TCM+CNP, then cultured in conventional IVM supplements + AREG+PGE2] for NANOG, CDX2, and TET1. But SOX2 significantly increased in [TCM+CNP (6 h), then cultured in AREG+PGE2], and OCT4 decreased in [TCM+CNP (6 h), then cultured in conventional IVM supplements + AREG + PGE2]. In mammals, despite specific species differences, the transcription of triad factors (NANOG, SOX2, and OCT4) is governed through feedback loops in a steady state. Therefore, knock-down of one factor results in the up-regulation of one or two of the other triad factors [57]. In the studies assessing temporal expression of the triad factors in ruminants, it was shown that the temporal expression of triads is stage-specific dependent, and the expression of OCT4 is induced during oocyte maturation and declines following development to blastocyst stage, while SOX2 and NANOG are transcribed during maternal/zygote transition and among triads, SOX2 presents the highest expression relative to references gene [58]. Therefore, the high expression of SOX2 in the treated group is consistent with the literature, and higher expression of SOX2 suggests improved zygote genomics activation in the treated group. Further emphasized, improved in vitro maturation in this group reflects itself in an improved effect on SOX2 expression. But, these propositions need further experiments and verifications. One of the shortcomings of this study is that we did not perform concomitant differential staining to observe whether this differential expression may have any effect on total cell numbers and cell allocations.

Conclusion
Despite interspecies differences among sheep with human and other farm animals, the sequential maturation of sheep oocyte by stimulating NP/NPR2, AREG/EGFR, and PGE2/PTGSR2 system can improve the quality of COCs that have reduced expression of these functional pathways, meaning that COCs derived from small size follicles are less competent due to lower expression of above genes. Therefore, based on the results of this study and others, the supplementation of IVM medium with products of COCs [59] and GCs secreted factors improves developmental capacity and is an alternative to synthetic chemical treatment.