Transcriptomic signature of the follicular somatic compartment surrounding an oocyte with high developmental competence

During antral folliculogenesis, developmental competence of prospective oocytes is regulated in large part by the follicular somatic component to prepare the oocyte for the final stage of maturation and subsequent embryo development. The underlying molecular mechanisms are poorly understood. Oocytes reaching the advanced stage of follicular growth by administration of exogenous follicle-stimulating hormone (FSH) possess higher developmental competence than oocytes in FSH-untreated smaller follicles. In this study, the transcriptomic profile of the cumulus cells from cows receiving FSH administration (FSH-priming) was compared, as a model of high oocyte competence, with that from untreated donor cows (control). Ingenuity Pathway Analysis showed that cumulus cells receiving FSH-priming were rich in down-regulated transcripts associated with cell movement and migration, including the extracellular matrix-related transcripts, probably preventing the disruption of cell-to-cell contacts. Interestingly, the transcriptomic profile of up-regulated genes in the control group was similar to that of granulosa cells from atretic follicles. Interferon regulatory factor 7 was activated as the key upstream regulator of FSH-priming. Thus, acquisition of developmental competence by oocytes can be ensured by the integrity of cumulus cells involved in cell-to-cell communication and cell survival, which may help achieve enhanced oocyte-somatic cell coupling.


Results
FSH-priming prior to oocyte retrieval enhances developmental competence of oocytes matured with gonadotrophin. Although there were no changes in the total number of follicles, a significantly higher number of large antral follicles (>8 mm) and lower number of small antral follicles (2-4 mm) were observed in cows following FSH-priming (P < 0.05) ( Table 1). The COCs were morphologically similar between the control and FSH-priming groups ( Table 2). The acquisition of EGF responsiveness by cumulus cells is a milestone of oocyte developmental competence 20 . To confirm EGF-like peptide responsiveness, we measured the cumulus expansion of COCs cultured with AREG between the control and FSH-priming groups. In both groups, cumulus expansion was induced by AREG; however, the level was significantly higher in the FSH-priming group (P < 0.05) (Fig. 1a). This result indicates that COCs from cows subjected to FSH-priming are more responsive to EGF ligands, than COCs from control cows. Post maturation and fertilization, there was no significant difference in embryo cleavage across all groups (Fig. 1b), suggesting oocytes derived from control and FSH-priming groups have similar competences on oocyte maturation and fertilization. In the presence of FSH in oocyte maturation medium, FSH-priming group increased blastocyst development by 2.3-fold, compared to the control group (control: 22.5%, FSH-priming: 52.5%; P < 0.05). However, in the absence of FSH in oocyte maturation medium, there was no significant difference between FSH-priming and control groups on blastocyst development (control: 12.2%, FSH-priming: 31.2%; Fig. 1c). Furthermore, blastocyst development was observed in all 12 cows subjected to FSH-priming when the COCs were matured with FSH, suggesting that the combined use of FSH-priming and FSH during the oocyte maturation stage yields maximum oocyte competence. Based on the result, COCs exposed to FSH-priming are defined as a high oocyte developmental competence model.
Biofunctions in the cumulus cells from cows subjected to FSH-priming. In this study, 1216 transcripts were identified as FSH-priming-sensitive genes. Compared to control, 937 transcripts were down-regulated and 279 were up-regulated in the FSH-priming group, relative to controls, suggesting a tendency towards global suppression of cumulus cell transcription in the FSH-priming group ( Fig. 2a and Supplementary  Table S3). Cell movement, migration of cells, and development of vasculature identified as the top three biofunctions in the FSH-priming group (Fig. 2b). Cell movement and migration involved 338 and 305 molecules, respectively, which included well known cumulus cell extracellular matrix-related genes such as pentraxin 3 (PTX3), tumour necrosis factor alpha-induced protein 6 (TNFAIP6), secreted phosphoprotein 1 (SPP1), and hyaluronan synthase 2 (HAS2) as the top molecules decreased in the FSH-priming ( Fig. 2b and Supplementary Table S4). Morbidity mortality, organism death, and glucose metabolism disorder were identified as the top three biofunctions to undergo an increase (Supplementary Table S4). Morbidity or mortality biofunctions included interferon regulatory factor 7 (IRF7), interferon-stimulated gene 15 (ISG15), and signal transducers and activator of transcription 1 (STAT1) as increased target molecules in FSH-priming (Supplementary Table S4).

Upstream regulators and canonical pathways of cumulus cells from cows subjected to FSH-priming.
Transforming growth factor beta 1 (TGFB1), STAT3, tumour protein p53 (TP53), platelet-derived growth factor BB (PDGFBB), mitogen-activated protein kinase 1 (MAPK1), and EGF, predicted as the top 30 upstream regulators undergoing inhibition, are shown in Fig. 3a and Supplementary Table S4. PTX3, TNFAIP6, SPP1, HAS2, matrix metalloproteinase 2 (MMP2), and thrombospondin 1 (THBS1) were included as the target molecules of TGFB1. Interleukin 6 (IL6) and nuclear factor kappa B (NF-K B) are also listed in the top 30 upstream regulators. Based on these results, TGFB signalling, STAT3 pathway, IL6 signalling, and NF-K B signalling were included as canonical pathways down-regulated by FSH-priming ( Fig. 4a and Supplementary Table S6). In addition to these canonical pathways, dendritic cell maturation, inhibition of angiogenesis by thrombospondin 1 (TSP1), and p38 mitogen-activated protein kinases (p38 MAPK) were identified. On the other hand, interferon alpha 2 (IFNA2), IRF7, interferon beta 1 (IFNB1), poly [ADP-ribose] polymerase 9 (PARP9), and dual specificity protein phosphatase 1 (DUSP1) were predicted as activated upstream regulators in FSH-priming (Fig. 3b and  Supplementary Table S5). ISG15, IRF7, 2′-5′-oligoadenylate synthetize 1 (OAS1), and the interferon-induced GTP-binding protein Mx1 (MX1), which are interferon-stimulated genes, were included as the up-regulated target molecules of IFNA2. Conversely, HAS2, TGFB1, THBS1, HBEGF, PDGFB are the down-regulated target molecules of PD98059. Interferon signalling was identified as top of up-regulated canonical pathways in the FSH-priming group (Fig. 4b and Supplementary Table S6). These results suggest that activation of interferon signalling and inhibition of MEK signalling in cumulus cells were induced by FSH-priming. Follicular atresia is suppressed by FSH-priming. The population of atretic follicles before the LH surge is 50-80% of all follicles on the basis of stereomicroscopic evaluation, apoptosis, and oestradiol-and insulin growth factor-binding proteins (IGFBPs), suggesting that a relatively large population of follicles before the LH surge could be unhealthy 28 . To evaluate the health status of cumulus cells derived from the control and FSH-priming groups at the transcriptomic level, RNA-seq data were compared with the transcriptomic profile of granulosa cells of bovine ovarian atretic follicles, as reported by Hatziodos et al. 29 Table 3 presents a profile of up-regulated genes in the control group compared with the FSH-priming group which significantly overlapped with the genes from the granulosa cells of atretic follicles, including THBS1 and MMP2. Furthermore, the profile of up-regulated genes in the FSH-priming group significantly overlapped with the genes from healthy granulosa cells, including the cytochrome p450 family 19 subfamily A (CYP19A1).
The expression of all genes significantly differed between the control and FSH-priming groups (P < 0.05).

Discussion
Oocyte competence to support embryo development is regulated by somatic cells surrounding the oocyte. These somatic cells are substantially modified by gonadotrophins during folliculogenesis. From a practical perspective, exogenous FSH treatment before OPU enhances bovine oocyte developmental competence 27,32 . Furthermore, in mouse and human studies, equine chorionic gonadotropin/FSH-priming before oocyte retrieval enhance oocyte developmental competence during preimplantation embryo development and also to term 33,34 . Global changes in follicle development and function are responsible for the improved subsequent oocyte quality and, due to their intimate association with the oocyte, the cumulus cells transmit and/or are responsible for these key changes in the oocyte during the final stages of oocyte development. As such developmental changes are critical to healthy early development, understanding the underlying molecular mechanisms in follicular somatic cells such as cumulus cells is important 23 . Here, we showed that FSH-priming drastically modulates gene expression in the cumulus cells, in particular in relation to increasing cell-to-cell communication and anti-inflammatory response, which may be implicated in the acquisition of oocyte developmental competence. This is the first report that describes the effect of FSH-priming on the RNA-seq profile of cumulus cells. This finding is likely to be important in efforts to enhance oocyte developmental competence not only in vivo, but also in in vitro maturation systems such as assisted reproduction technologies (ART). The final stage of oocyte maturation and ovulation is mediated by EGF-like peptides stimulated by the surge in gonadotrophins. Functional EGFR signalling is developmentally regulated in the ovarian follicles, and the acquisition of somatic cell signalling capability is likely to be an important developmental milestone for oocytes 20 . In pigs, oocytes in growing small antral follicles with inherently low developmental competence are poor responders to EGF or EGF-like peptides [20][21][22] . In the current study, in a bovine model, although the COCs from small antral follicles (control group) were responsive to EGF-like peptides in terms of cumulus expansion, the extent of responsiveness was higher in the COCs from large antral follicles following FSH-priming, despite the fact that EGFR expression did not differ between the control and FSH-priming groups (Table S3). Promoting the responsiveness of cumulus cells to EGF-like peptides increases oocyte developmental competence 22 . Previously, we have suggested that improved oocyte competence may be aided by EGF-like peptide-based stimulation of cumulus cells by facilitating oocyte mitochondrial activity and energy production required for subsequent development 8,35 . Furthermore, EGF-like peptides stimulate RNA translation in matured oocytes, which is critical for embryo development 36 . The developmental competence of oocytes matured in vitro with FSH was increased by FSH-priming. The pre-maturation in vivo of cumulus cells subjected to FSH-priming may allow efficient production of nutritional molecules, thanks to the surge in gonadotrophin levels via a functional EGFR signalling network.
Analysing the transcriptomic landscape in the cumulus cells from small (low oocyte competence) and large antral follicles (high oocyte competence) may provide more details about the underlying functional and molecular changes which participate in the development of competence in oocytes during antral folliculogenesis. RNA-seq analysis showed drastic transcriptomic changes (change in 1274 genes) in the cumulus cells subjected to . Upstream regulator analysis of FSH-priming-sensitive transcripts in cumulus cells. Ingenuity Pathway Analysis was used for the prediction of upstream molecules including cytokines, growth factors, transcription factors and chemicals. Selected upstream molecules, from the top 30, whose predicted activation status was "inhibited" (a) or "activated" (b) with corresponding molecules are shown. Inhibited and activated upstream molecules were defined as an activation z-score ≤ −1.5 and −log(P-value) > 1, and as an activation z-score ≥ 1.5 and −log(p-value) > 1, respectively. The upstream molecules were ranked by their −log(p-value). The significance was defined as −log(p-value) > 1. Brown, pink, and blue lines indicate upstream regulators, and up-regulated and down-regulated genes stimulated by each upstream molecule, respectively.
SCIENTIfIC REPORTS | 7: 6815 | DOI:10.1038/s41598-017-07039-5 FSH-priming; however, down-regulated genes constituted a major population of the altered transcriptomic profile, contrary to our expectation. These down-regulated genes were associated with cell movement and migration, and included extracellular matrix-related genes such as HAS2, TNFIP6, and PTX3. These genes are characteristically up-regulated in cumulus cells during oocyte maturation, and this is also associated with enhanced oocyte competence 37,38 , however it should be noted that in the current study, cumulus cells were collected immediately prior to this stage (i.e. from immature unexpanded COCs). As a well-characterised action, expansion of the extracellular matrix during oocyte maturation, which involves the movement/migration of cumulus cells within the matrix, disrupts oocyte-cumulus and cumulus-cumulus GJC. Bidirectional communication between the oocyte and cumulus cells via GJC is essential for enabling the transfer of nutrients and other small molecules, along with ensuring that the oocyte acquires the molecular machinery required to support embryo development. Hence, an important feature of follicular growth treatment associated with improved oocyte quality may be to prevent the disruption of cell-to-cell communication, such as GJC, in the immature COC (Fig. 6a).
Activation of cAMP-PKA signalling is one of the reasons for the decrease in cell movement/migration within cumulus cells in cows subjected to FSH-priming. A recent study showed that cAMP, as the key secondary messenger of FSH, suppresses ERK1/2, which is part of the MAPK system 39 . Cyclic AMP stimulates a decrease in ERK1/2 phosphorylation via DUSP1 phosphatase, consequently causing the down-regulation of extracellular matrix-related genes such as HAS2. In the present data obtained by IPA, MAPK1, also known as ERK2, was predicted to be an inhibited upstream regulator. On the other hand, DPSP1 and the inhibitor of MEK1/2, which is the kinase of ERK1/2, were predicted as activated upstream regulators. Furthermore, HAS2 was included as the . Canonical pathway analysis in bovine cumulus cells with FSH-priming. Ingenuity Pathway Analysis was used for predicting canonical pathways. The canonical pathways are identified and ranked by their z-score and -log(P-value). Down-regulated (a) and up-regulated canonical pathways (b) by FSH-priming were defined as a z-score ≤ −1.5 and −log(P-value) > 1, and as a z-score ≥ 1.5 and −log(P-value) > 1, respectively. The significance of canonical pathway was defined as −log(P-value) > 1.    target molecule of the inhibitor of MEK1/2. Thus, decreased cell movement and migration may be implicated in the inhibition of extracellular matrix expansion in cumulus cells, which could be induced by the negative regulation of MEK signalling, probably via cAMP-PKA signalling (Fig. 6a). FSH also stimulates enhancement in GJC, at least in part by a cAMP-PKA-based mechanism 5 . Using cAMP-elevating agents such as forskolin and IBMX for in vitro pre-maturation, GJC is enhanced and helps improve oocyte competence 40 . The enhanced GJC affects oocyte chromatin remodelling and transcription 41 , oocyte metabolism 8 , and accumulation of intra-oocyte GSH 42 , which are necessary for the acquisition of developmental competence. Thus, enhancement of GJC by elevating and/or sustaining cAMP concentration in cumulus cells subjected to FSH-priming may also be a reason for improved oocyte competence. Cumulus cells subjected to FSH-priming were enriched in transcripts associated with interferon signalling and IRF activation, including interferon-stimulated genes. Moreover, IRF7 ranked first in the list of activated upstream regulators in the FSH-priming group. IRF7 is a master regulator of the IFNα/β immune response against viruses 43 , whose expression is crucial for myeloid cell conversion from a pro-to an anti-inflammatory state. IRF7 expression is regulated by TGFB1, and negative regulation of IRF7 is stimulated by prolonged exposure to TGFB1 44 . In granulosa cells, increased TGFB1 is a hallmark of the activated inflammatory process, which may stimulate follicular atresia 29 . In the results obtained by IPA, TGFB1 has been predicted as an inhibited upstream regulator, which was dependent on a suppressed TGFB pathway, and the expression of TGFB1 and the receptor TGFR2 were lower in the FSH-priming group. Furthermore, we found that follicular atresia-like transcriptomic profile in cumulus cells, which was characterized as activation of TGFB and TP53 signalling and inhibition of angiogenesis by TSP1 45 , is suppressed in the FSH-priming group. The high expression level of CYP19A1 observed in the FSH-priming group corroborates a healthy large follicle 30,31 . In the natural oestrus cycle in mono-ovulatory species, only a large dominant follicle is ultimately selected for ovulation and other follicles with inhibited angiogenesis are induced into atresia by apoptosis signalling in granulosa cells; this has been attributed to the depletion of survival factors, in particular FSH 46 . Therefore, most COCs from a natural oestrus cycle ovary are derived from follicles in varying degrees of atresia. Artificial manipulation of follicular development by oestrus synchronization and FSH administration rescues small antral follicles from atresia via suppression of STAT3 signalling 47 . STAT3 was predicted as an inhibited upstream regulator in the FSH-priming group. Moreover, it has been shown that STAT3 knockout or knockdown cells display enhanced expression of IFNα/β response genes, including OAS and IRF7 48 . Hence, the cumulus cells in cows subjected to FSH-priming may sustain cell integrity because of an induced anti-inflammation system, probably via increased IRF7 expression resulting from inhibited TGFB1 and/ or STAT3 signalling (Fig. 6a).
Apart from survival factors such as FSH, paracrine signalling from oocytes participate in regulating follicular atresia. This signalling could be triggered by OSFs such as BMP15 or cumulin via the activation of AKL6 (BMPR1B)-SMAD1/5/8, but not OSFs such as GDF9 12,49 . In fact, BMPR1B transcripts are up-regulated in atretic follicles in cattle 50 . In sheep as a mono-ovulatory animal model, loss-of-function mutation of BMPR1B stimulates multi-ovulation by suppressing atresia of small antral follicles 51 . On the other hand during late folliculogenesis, BMPR1B signalling works as a key player for cumulus expansion and ovulation following the surge in gonadotrophin levels. The present study showed that although the expression levels of BMP15 and GDF9 in oocytes did not differ between the control and FSH-priming ( Supplementary Fig. S1), BMPR1B expression was down-regulated in cumulus cells in the FSH-priming group. A diminished role for BMP and SMAD1/5/8 signalling in the FSH-priming group may be consistent with natural mammalian polyovulation and high fecundity being caused, at least to a certain extent, by a diminished role for BMP15 relative to GDF9 52,53 .
Another interesting finding is that dendritic cell maturation was predicted as an inhibited pathway in cumulus cells in the FSH-priming group. Dendritic cell maturation is triggered by pathogens, tissue damage, and local inflammation 54,55 and the cells present in the ovarian environment 56 . The small ovarian dendritic cell population is essential for LH-stimulated up-regulation of specific ovulatory genes that are crucial for cumulus mucification and expansion and subsequent ovulation 57 . Ovulation has long been likened to an inflammation-like proses 58 . Hence, inhibition of dendritic cell maturation by FSH-priming, which may be associated with the suppression of initiation of spontaneous ovulation-like cascades before the gonadotrophin surge, probably regulates the responsiveness to intrinsic cumulus mucification and expansion needed for ovulation and acquisition of developmental competence.
In conclusion, the present study showed that cumulus cells undergoing advanced follicular growth prepare the oocyte for final maturation; in part by promotion of cell-to-cell communication and anti-inflammation as well as EGF-like peptide responsiveness, probably allowing efficient transfer of nutritional molecules from the somatic component of the follicle to the oocyte. These processes are likely to constitute important components of oocyte developmental competence (Fig. 6b). This concept and further IVM improvement may help to obtain high competence oocytes for animal production and human ART.

Materials and Methods
Chemical. Unless specified, all chemicals were purchased from Sigma-Aldrich (St Louis, MO, USA).
Animal care and use. This study was approved by the Ethics Committee for the Care and Use of Experimental Animal at the National Livestock Breeding Center located in Nishigo, Japan. All animals received human care according to guideline numbers 6, 22 and 105 of the Japanese Guidelines for Animal Care and Use.

Collection of cumulus-oocyte complexes (COCs) via ovum pick-up (OPU). We performed 4 series
of OPU throughout the present study. Four cows were used for each OPU series. As described previously 27 , COCs were collected from total of 16 pubertal Japanese Black cows of 45 to 74 months old by OPU using an ultrasound scanner (SSD-900; ALOKA, Tokyo, Japan) and 7.5-MHz convex array transducer (UST-9109P-7.5; ALOKA) with a 17-gauge stainless steel needle guide. Follicles were categorized as small (2-4 mm in diameter), medium (5-8 mm), and large (>8 mm). Follicles >2 mm in diameter were vacuum-aspirated (120 mmHg and 22 mL/min aspiration rate) through a disposable aspiration needle (COVA Needle; Misawa Medical, Tokyo, Japan). (FSH-priming). The first OPU session (OPU without stimulation) was performed at arbitrary days of the estrous cycle (Day 0), yielding the control COCs from small and medium follicles (Fig. 7). Then on day 5, dominant follicles were ablated (DFA) by aspirating follicles larger than 8 mm in diameter (these COCs were discarded) and a progesterone-release controlled internal drug releasing (CIDR) device (CIDR 1900; Pfizer Animal Health, Hamilton, New Zealand) was inserted intravaginally.
COCs morphological assessment was slightly modified from previous report 59 . COCs collected by OPU were categorized into 4 groups according to morphology of oocyte cytoplasm and cumulus cells: category A; COCs with homogenous oocyte cytoplasm and a complete, compacted, and more than 6 layers of cumulus; category B, the cumulus was smaller than in category A COCs but had 3 to 5 layers of compact cumulus cells; category C, the COCs had 1 to 2 layers of cumulus cells; category D, the cumulus was completely part absent or expanded, or oocyte cytoplasm was degenerated. In both treatment groups (control, FSH-priming), only COCs of category A and B were used in all experiments. Category A and B COCs were then subjected to either (Figure 1), (1) in vitro culture (end-points: cumulus expansion and blastocyst production), or (2) immediate (0 hours) cumulus cell stripping for RNA analysis (see below). COCs expansion assay. Cumulus expansion assay that is widely used was performed as described 24 .
Cumulus expansion was scored at 22 h of IVM with 100 ng/ml recombinant human amphiregulin (AREG; R&D systems, Mineapolis, MN, USA). Degree of cumulus expansion was scored into 0 to +4. A score of 0 indicate no detectable response, +1 indicating the minimum response with cells in the peripheral two layers beginning to expand, +2 indicating expansion extending inwards to several layers, +3 indicating expansion of all cumulus layers except the corona radiata cells, and +4 indicating expansion of the entire cumulus including the corona radiate cells.
In vitro fertilization (IVF). IVF was performed as reported previously 60 . At the end of IVM, ejaculated sperm samples from Japanese black bulls were thawed and then centrifuged in 3 ml of 90% Precoll solution (GE Heathcare, Uppsala, Sweden) at 750 × g for 10 min. After centrifugation, the pellet was re-suspended and centrifuged in 6 ml of sperm washing solution (Brackett and Oliphant solution, BO) 61 , supplemented with 10 mM hypotaurine (Sigma) and 4 U/mL heparin (Novo-Heparin Injection 1000; Aventis Pharma Ltd., Tokyo, Japan) at 550 × g. Then the pellet was re-suspended in sperm-washing solution and BO solution supplemented with 20 mg/ml BSA (Sigma) to achieve a final concentration of 3 × 10 6 sperm/ml. 100 µl droplets of this suspension were aliquoted in 35-mm dishes under paraffin oil as fertilization droplets. COCs were washed twice in BO supplemented with 10 mg/ml BSA and cultured in the fertilization droplets for 6 h at 38.5 °C in 5% CO 2 in air with saturated humidity.
IVC of embryos was performed at 38.5 °C in 5% O 2 , 5% CO 2 , and 90% N 2 with saturated humidity for 7 days in 125 μl CR1aa placed in a microwell culture dish 63 . The microwell culture dish allows embryos to be cultured in an individually identifiable manner without affecting embryo density 64 . After 6 h of insemination, cumulus cells and sperm were completely removed from zygotes by pipetting with a glass pipette in IVC medium. Zygotes were placed in microwells of the culture dish.
RNA extraction. Cumulus cells were removed from immature COCs (0 hours) by pipetting with a grass pipette in PBS. Cumulus cells were lysed in 300 µl of RTL buffer containing 10 µl/ml of 2-mercaptoethanol and stored at −80 °C. Total RNA was extracted from each sample using the RNeasy Micro Kit (Qiagen, CA, USA). Genomic DNA was removed by digestion with recombinant RNase-free DNase I (Qiagen). RNA was quantified using a NanoDrop spectrophotometer (Thermo Fisher Scientific, MA, USA). To identify oocyte specific genes with RNA sequence, RNA from the oocytes of four cows was extracted in the same manner as for cumulus cells.
RNA sequence. Total RNA derived from four cows per treatment (one biological replicate/treatment) was used for library construction using the TruSeq Stranded mRNA LT Sample Prep Kit (Illumina, CA, USA) Figure 7. Schematic representation of experimental design. The first OPU from small to medium follicles was performed on unstimulated animals on arbitrary days of the oestrus cycle (control). Then, follicles larger than 8 mm in diameter were ablated (DF-A), and CIDRs were inserted on day 5 (the day of first OPU = day 0). The cows then received 30 mg FSH twice a day from days 7 to 10, in decreasing doses (6, 6, 4, 4, 3, 3, 3, and 2 mg) by intramuscular injection. Cloprostenol (PGF2α; 0.75 mg) was administered in the morning of day 9. The second OPU was performed 48 h after PGF2α administration (day 11), and only follicles larger than 5 mm in diameter (medium and large) were aspirated. CIDRs were removed from cows just before OPU (FSH-priming). COCs of only morphological category A and B were used for investigating cumulus expansion and developmental competence of oocytes and analysing the RNA-seq and real-time qPCR of cumulus cells.
according to the manufacturer's protocol. RNA purity and integrity were assessed using TapeStation 2200 (Agilent Technologies, CA, USA), and all samples had a RNA Integrity Number Equivalent (RINe) value of >8.0. The libraries were sequenced on the Illumina HiSeq. 2500 platform (Illumina) with 101-bp paired-end reads. The reads were aligned to the reference genome (NCBI UMD_3.1) using TopHat (v.2.0.13) (Trapnell et al. 2012) with the Refseq gene annotation. Expression levels (FPKM) of Refseq genes were calculated using Cufflinks (v. 2.2.1) 65 .

Analysis of biofunction, upstream regulators and canonical pathways.
Ingenuity pathway analysis (IPA) was used in differentially expressed genes. Lists of mRNA differentially expressed between control and FSH-priming grouped, which was defined as (FSH-priming_FPKM + 0.1)/(Control_FPKM + 0.1) ≥ 2 and FSH-priming_FRKM ≥ 1 and (FSH-priming_FPKM + 0.1)/(Control_FPKM + 0.1) ≤ 0.5 and Control_ FPKM ≥ 1, were uploaded in the ingenuity package. In order to eliminate the possibility of oocyte contamination, oocyte-specific genes were deleted from the list of mRNA. Oocyte-specific genes was defined as Control_Cumulus cell_FPKM < 1 and Control_Oocyte_FPKM ≧ 10 (Supplementary Table S2). Significantly affected diseases, biofunctions, upstream regulators and canonical pathways were defined based on the P-value of Fischer's exact test and the activation Z-score. The predicted upstream regulators and their downstream targets were visualized using Circos 66 .
Comparison between transcriptomic profile of the present RNA-seq data and atretic follicular somatic cells. Present RNA-seq data were compared with microarray probe data of granulosa cells derived from well characterised bovine ovarian atretic follicles, which was reported by Hatzirodos et al. 29 . Statistical significance of number of overlapped genes between data sets of present RNA-seq and Hatziodos et al. were analyzed by Chi-squared test using R statistical software version 2.15.0 (R Foundation for Statistical Computing). mRNA expression by real-time RT-PCR analysis. For validation of RNA-seq data, cumulus cell mRNA expression from eight cows was analysed individually with quantitative real-time PCR as described 8 . The cumulus cell samples were from different animals that were used to generate the RNA-seq data. Beta-actin (ACTB) was used as the endogenous control. Total RNA (200 ng) extracted from cumulus cells from individual cows was reverse transcribed with random primers (Invitrogen, CA, USA) using Super-Script III (Invitrogen). The real-time PCR analysis was performed on a StepOne TM instrument (Applied Biosystems, Foster City, CA, USA) in a 20-µl reaction volume containing 3 µl cDNA, 2.5 µl each of forward and reverse primers (Supplementary  Table S1), 2 µl nuclease-free water and 10 µl of SYBR Green PCR Master Mix (Applied Biosystems). PCR reactions were performed in duplicate. Universal thermal cycling parameters (initial step of 2 min at 50 °C and 10 min at 95 °C, followed by 40 cycles of 15 s at 95 °C and 60 s at 60 °C) were used. Melting curve analysis was carried out on the real-time cycler to check the specificity of the reaction. A standard curve was generated for the genes in every PCR run by using a serial 5-fold dilution of amplified cDNA derived from cumulus cells. Results were normalized to ACTB and expressed relative to a mean value of control which was set at 1.