Exosomal miR‐27 negatively regulates ROS production and promotes granulosa cells apoptosis by targeting SPRY2 in OHSS

Abstract Ovarian hyperstimulation syndrome (OHSS) is one of the most dangerous iatrogenic complications in controlled ovarian hyperstimulation (COH). The exact molecular mechanism that induces OHSS remains unclear. In recent years, accumulating evidence found that exosomal miRNAs participate in many diseases of reproductive system. However, the specific role of miRNAs, particularly the follicular fluid‐derived exosomal miRNAs in OHSS remains controversial. To identify differentially expressed follicular fluid exosomal miRNAs from OHSS and non‐OHSS patients, the analysis based on miRNA‐sequence was conducted. The levels of 291 miRNAs were significantly differed in exosomes from OHSS patients compared with normal control, and exosomal miR‐27 was one of the most significantly down‐regulated miRNAs in the OHSS group. By using MiR‐27 mimic, we found it could increase ROS stress and apoptosis by down‐regulating the expression of p‐ERK/Nrf2 pathway by negatively regulating SPRY2. These data demonstrate that exosomal miRNAs are differentially expressed in follicular fluid between patients with and without OHSS, and follicular fluid exosomal miR‐27 may involve in the pathological process of OHSS development.


| INTRODUC TI ON
Ovarian hyperstimulation syndrome (OHSS) is a grievous complication of assisted reproductive technologies (ART), which is ordinarily induced by abnormal reaction to human chorionic gonadotropin (hCG). 1 Statistical data reveals that 32% of the patients in the general IVF cycle suffered mild OHSS, 10%-15% exhibited moderate OHSS, while 5%-8% of the patients were diagnosed with severe OHSS. 2 Besides influencing clinical outcomes like embryo implantation rates and live birth rates, 3 OHSS could cause life-threatening conditions, such as acute renal insufficiency, acute respiratory distress, in severe cases, it causes venous thrombosis and mortality. 4,5 With the development of ovulating induction protocols, women undergoing GnRH antagonist stimulation protocols partially decreases the risk of OHSS, but some deficiencies still limit the application of this method. 6 So far, prevention strategies and the development of novel therapeutic targets for OHSS remains a challenge in reproductive medicine.
The exosome is a small vesicle with lipid bilayer structure secreted by cells with a diameter of about 50-150 nm, it carries DNA, mRNA, microRNA, and proteins. 7 Studies have shown that exosomes can be detected in various body fluids, such as blood, urine, 8 cerebrospinal fluid, 9 ascites, follicular fluid, 10 amniotic fluid and joint fluid. 11 Animal experiments have proved that the microRNAs in follicular fluid of mature follicles are distinct from those in immature follicles. When cultured in follicular fluid, these microRNAs could be absorbed by follicular cells. 12 Recent studies have reported that exosomal microRNAs derived from follicular fluid are directly linked to granulosa cells apoptosis, follicular development, and ovarian function. 13,14 Also, Wang and his colleagues sequenced exosomal cir-cRNAs in PCOS follicle fluids, they identified 167 up-regulated, 245 down-regulated circRNAs and constructed a circRNA-microRNA network. 15 However, there was no evidence on the expression profile of exosomal microRNAs in follicular fluid and their potential roles in OHSS development.
In this work, we identified differentially expressed exosomal miRNAs (follicular fluid derived) from women with or without OHSS.
We found that miRNA-27-3p was the most significantly reduced in OHSS group, and interacted with SPRY2 in granulosa cells. Further, we found that miR-27 inhibits Nrf2 expression via EGFR/ERK signal and aggravates cells apoptosis by increasing ROS production. Our study shows a possible mechanism of exosomal miR-27-3p on the microenvironment of the follicle, which probably involved in the OHSS development. Patients who participated in the study received long-term regimen therapy. The diagnostic criteria of OHSS were borrowed from Golan criteria. 16 For the OHSS patients, the inclusion criteria were as follows: serum AMH > 5 ng/mL, serum E2 > 3500 pg/mL on HCG day and oocytes > 20. The controls were diagnosed with fallopian tubal diseases via laparoscopy and hysteroscopy, or exhibited male factor infertility. Women suffering from malignancy, benign ovarian cyst including endometrioma, allergic diseases; pelvic inflammation, known chronic, systemic, metabolic, or endocrine disease excluding polycystic ovarian syndrome, were excluded from this study.

| Collection of follicular fluid
All patients received 10 000 IU human chorionic gonadotropin (hCG) administration on trigger day. The follicular fluid from the large follicle (>18 mm) of the patient first sampled and collected in a 15 mL centrifugal tube for centrifugation (Falcon). After 2500 g centrifugation for 15 minutes, we collected the supernatant and stored at -80°C until use. Granulosa cells were purified using density centrifugation from follicular fluid at 2000 g for 10 minutes. The cell pellets were re-suspended in percoll (GE Healthcare) and centrifuged at 2000 g for 20 minutes to separate red blood cells. The interface cells were collected for future use.

| Exosome isolation and identification
Follicular fluid exosomes were isolated and characterized according to a previously published protocol. 17 The follicular fluid was gradually melted on ice and diluted by adding 20 mL PBS (pH 7.4) (Thermo Fisher Scientific). The follicular fluid was isolated by centrifugation at 2500 g for 30 minutes, and the supernatant was then transferred to a new centrifugal tube and centrifuged at 12 000 g for 5 minutes to eliminate large particles. After centrifugation, we filtered the supernatant using a 0.22 μm filter. Finally, the filtered samples were transferred to ultracentrifuge tubes for centrifugation at 120 000 g and 4°C for 4 hours in an ultracentrifuge (Beckman Coulter). Thereafter, exosome pellets were re-suspended in RIPAlysate (Thermo Fisher Scientific) for western blot analysis or in PBS for nanoparticle tracking analysis (NTA). Then, the exosomes were collected for further treatment. Through the dynamic light scattering method, the particle size distribution of exosomes was evaluated. We entrusted the NTA analysis to Shanghai XP Biomed Ltd.

| Exosomal RNA extraction and sequencing
The exosomes were added into 1 mL Trizol (Invitrogen, USA), blown uniformly with RNA-free gun-head repeatedly, and then placed at room temperature for 5 minutes. A total of 200 μL chloroform was added and incubated at room temperature for 10 minutes with severe shock for 15 seconds. Then, the exosomes were centrifuged at 12 000 g and 4°C for 15 minutes and 400 μL supernatant collected.
Additional 400 μL isopropanol (Sangon Biotech) was added, mixed well, and left to dry at room temperature for 10 minutes. Afterwards, the mixture was centrifuged at 12 000 g and 4°C for 10 minutes and the supernatant discarded. Again, 1 mL 75% ethanol was added, gently washed and precipitated, centrifuged it at 4°C for 5 minutes and the supernatant discarded. After drying at room temperature, an appropriate amount of DEPC water was added to dissolve and kept on ice for the next experiments. Follicular fluid samples from ten OHSS patients and ten normal controls were used for exosomal RNA sequencing. Library preparation and sequencing were conducted at Anoroad. A total of 1 μg total RNA per sample was used for the small RNA library.
After the total RNA samples were quantified, the RNA fragments were fractionated on a 15% polyacrylamide gel (Invitrogen) and small RNAs ranging between 15 and 30 nucleotides (nt) were used for library preparation. The two ends of the separated RNA fragments were connected respectively. Small RNAs were reverse transcribed by RT primers and amplified by PCR. The PCR products were sequenced using the Illumina HiSeq 2500 platform (Illumina Inc).

| Detection of microRNAs
The samples were isolated and reversed by reverse transcription AMV enzyme reverse transcription system (Taraka). The 10 μL of a master mix contains 2 μL of AMV 5X buffer, 0.5 μL AMV Enzyme, 1 μL dNTPs, 1 μL reverse transcription primer, 1 μL RNA template, and 3.5 μL DEPC Water. The samples were incubated at 16°C for 30 minutes followed by 42°C for 30 minutes and 85°C for 15 minutes. The resulting cDNA was collected for real-time PCR.
Real-time PCR was performed using 10 μL SYBR Premix ExTaq (2X), 0.8 μL forward primers (5 pmol/mL), 0.8 μL reverse primers (5 pmol/mL), 0.4 μL ROX, 1 μL reverse transcription reaction products, and 7 μL ddH 2 O. The following two-step PCR amplification was used: 95°C for 5 seconds and 60°C for 30 seconds for a total of 40 cycles. After the reaction, the amplification and melting curves of Real-time PCR were confirmed, and the data were analysed by the 2 −ΔΔCT method.

| MicroRNAs target gene ontology and pathway analysis
Differential expression of known and unknown microRNAs was analysed and target genes of differentially expressed microRNAs were predicted on the miRNA-database. Further, GO annotation analysis and KEGG signalling pathway enrichment analysis was performed to establish the enriched GO term and Pathway. Eventually, the interaction of microRNAs with genes in the GO term and Pathway was analysed, and the key pathways and regulatory networks in OHSS were mapped by combining the known signalling pathways. The regulatory relationship between the two was obtained on the mi-croRNA database, and the microRNA-RNA interaction network was established.

| Cell proliferation assay
Cell proliferation Assay was tested using the CCK8 Assay. All processes were according to the manufacture's protocol (Beyotime).

| RNA extraction and qRT-PCR
Total RNA was extracted from the granulosa cells and 293T cells using Trizol following the manufacturer's protocol and quantified using the NanoDrop ND-2000 spectrophotometer. Total RNA was reverse transcribed into cDNA using RNA reverse transcription kit (Thermo). For the detection of genes, qPCR was performed as previously described. The primers for genes are listed in Table 1.

| Dual-luciferase reporter assay
The relationship between the SPRY2 and miR-27-3p was verified on the database Tarcan. Next, the luciferase reporter plasmid containing wild-type 3'-UTR of SPRY2 was purchased from Genomeditech.
Wild-type and the mutated vectors were co-transfected with miR-27-3p mimics or negative control into HEK-293T cells using Lipofectamine2000. After 48 hours, Luciferase activity of cultured supernatant was measured using a kit as per the manufacturer's instructions.

| Immunofluorescence
Granulosa cells were collected and washed with PBS three times at room temperature. They were then placed in PBS containing 0.1% Triton X-100 at room temperature for 30 minutes. After blocking in PBS-5% bovine serum albumin (BSA) at 37°C for 1 hour, the granulosa cells were incubated with the special antibodies described above at 4°C overnight. Cells were further incubated with goat antimouse and goat anti-rabbit at 37°C in the dark followed by staining of nuclei with DAPI for 5 minutes. Finally, all granulosa cells were placed on slides with mounting medium (Beyotime) and gently covered with ethanol-primed coverslips and kept in the dark until confocal scanning.

| Co-Immunoprecipitation
The cells were washed with PBS at room temperature and collected by centrifugation. The PBS was drained carefully and 100 μL nondenaturing lysis buffer added. Further, we added a 10 μL antibody and topped with Lysis Buffer containing protease inhibitors to make 500 µL volume. The mixture was gently mixed overnight at 4°C on a rotary mixer. The Protein-A/G beads were washed twice with 1 mL Wash Buffer and centrifuged at 2000 g for 2 minutes aspirating the supernatant in between washes. The sample was suspended as a 50% slurry in Wash Buffer. After Antibody Binding, 40 µL of protein-A/G beads slurry was added to each tube and incubated at 4°C for 1 hour.
The beads were collected by low-speed centrifugation at 2000 g and 4°C for 2 minutes. Then, the beads were washed 3 times with 1 mL Wash Buffer and collected by low-speed centrifugation at 4°C aspirating the supernatant in between washes. Afterwards, the Wash Buffer was removed and the beads kept wet. A total of 40 µL 2x SDS-PAGE loading buffer was added to the beads and boiled for 5 minutes to elute the complex. Eluent collected after centrifugation was stored on ice for same-day analysis or frozen at −80˚C for future analysis by SDS-PAGE.

| Caspase 3/7 activity analysis
To detect the activity of caspase-3/7 in KGN cells, we used the Caspase3/7 activity apoptosis assay kit (Ribobio) according to the
Granulosa cells were incubated in the dark with 5uM staining solution in PBS at 37°C for 30 minutes and harvested with a 0.05% trypsin-EDTA solution, then immediately analysed with a flow cytometer. 18

| Statistical analysis
For data analysis, we used SPSS 19.0 statistical software. Chi-square test was used for counting data, while one-way ANOVA for comparison between groups. The measurement data were expressed in the form of mean ± SD. The comparison between the two groups was analysed by the Student's t-test. The difference at P <0.05 was statistically significant.

| Participant characteristics
The characteristics of all the enrolled participants are summarized in

| Characteristics of follicular fluid exosomes
Exosomes have a typical round or elliptical cup-disc structure with a diameter of 30-100 nm ( Figure 1A). We further characterized the follicular exosomes size and the purity by using NanoSight analysis (NTA). NTA of exosome size showed that the diameter of exosomes in the follicular fluid of the two groups ranged between 75 to 150 nm. Two colours were used two to distinguish them. The red line represents the OHSS group while the black line represents the normal controls ( Figure 1B). A total of 3 positive marker proteins (Alix, CD9, Tsg101) and two negative marker proteins (Calnexin, GM130) were selected to validate the follicular fluid exosomes of the two groups by western blot. Alix, CD9, and Tsg101 were all expressed in the exosome samples ( Figure 1C), but exosome lysates showed no expression of Calnexin and GM130 ( Figure 1D). These results attested that the purified particles shared the typical features of exosomes.

| Follicular fluid exosomes from non-OHSS group inhibit cell viability and increase apoptosis in human KGN cells
To explore the effects of exosomes on KGN cells viability, we transfected cells with exosomes from the two groups for 48 hours. We Conclusively, these findings suggest that exosomes from the normal control group inhibit viability and enhance apoptosis of KGN cells.

| Differential expression of miRNAs in follicular fluid exosomes derived from OHSS and non-OHSS patients
To identify differentially expressed miRNAs in exosomes between two groups, we isolated total RNA from two groups of follicular fluid  Figure S1A,B). Of the differently expressed miRNAs, miR-27-3p was significantly downregulated ( Figure 3D), and was significantly higher in exosomes compared with follicular fluid supernatant in normal control ( Figure 3E).

| SPRY2 is a potential direct target of miR-27-3p
After searching at the online target gene databases (TargetScan, miRanda, miRWalk, and miRTarBase) ( Figure 4A), we downloaded 4 potential targets which have the highest comprehensive score after analysing the predicted results (SPRY2, RECK, STAB1 and SOX5). We compared the mRNA expression levels of these 4 genes in patients' granulosa cells between the two groups, and SPRY2 is speculated to be a target gene of miR-27-3p (8.49 ± 0.75 vs 1.05 ± 0.050, P <0.05) ( Figure 4B). Furthermore, to demonstrate a direct interaction between the SPRY2 3'UTR and miR-27-3p, the WT SPRY2 3'UTR region (predicted to interact with miR-27-3p) was cloned into a luciferase reporter vector. We also cloned a SPRY2 3'UTR fragment with a mutation of the complementary base ( Figure 4C). As expected, luciferase activity suppression was significantly rescinded when interaction between miR-27-3p and its target 3'UTR was disrupted in cells transfected with SPRY2 3'UTR Mut. Luciferase activity was reduced by 40% in cells co-transfected with miR-27-3p ( Figure 4D).

| SPRY2 interacts with EGFR in KGN cells
SPRY2 belongs to Sprouty protein family and regulates the signalling downstream of multiple growth factor receptors. A search at the STRING database (string-db.org) for possible interaction proteins of SPRY2 found that EGFR, FGFR2, and GRB2 might interact with SPRY2 ( Figure S1C). EGFR was a key receptor in OHSS. 19 Despite a few literature reporting the interaction between SPRY2 and EGFR in some cells, 20 it is not clear whether SPRY2 and EGFR interact in granulosa cells. We detected the expression of SPRY2 and EGFR in granulosa cells of two groups by RT-QPCR and western blot and found that the expression levels of SPRY2 and EGFR in mRNA and protein among the OHSS group was higher than that in the normal control group (Figure 5A,B). Immunofluorescence, co-localization, and immunoprecipitation results reported an interaction in granulosa cells ( Figure 5C,D). According to the previous literature,

| Exosomal miR-27-3p promotes KGN cells ROS and apoptosis
The Through H2DCFDA staining, we detected the ROS levels in granulosa cells. When granulosa cells were transfected with miR-27 mimic, they showed high oxidative stress ( Figure 7C,D). High levels of ROS apoptosis are known to cause apoptosis. 21 By Tunel apoptosis assay and cleaved caspase 3 expression, we found that miR-27 increases granulosa cells apoptosis ( Figure 7E,F).

| D ISCUSS I ON
Ovarian hyperstimulation syndrome (OHSS) is an iatrogenic condition culminating from the abnormal ovarian response, and it is also closely linked with increased miscarriages and decreased live birth. 22 Interestingly, OHSS is a good indicator of ovarian function in a sense, since most patients with OHSS are younger than 35 years. 23 Our statistics also show that OHSS group patients have a higher AMH level, which could reflect the state of ovarian function. 24 Previously, many studies indicated that the principal reason for OHSS occurrence is

ACK N OWLED G EM ENTS
This work was Supported by the National Natural Science Foundation of China (No 81774357, 82074479).

CO N FLI C T O F I NTE R E S T
The authors confirm that there are no conflicts of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available from the corresponding author upon reasonable request.