DUOX2 is a Generator of ROS in the Ovary and a Potential Mediator of Ovulation

Background: Ovulation is triggered by the preovulatory surge of the pituitary luteinizing hormone (LH). LH/hCG induction of reactive oxygen species (ROS) is required for successful ovulation. H 2 O 2 , one of ROS species, was shown to fully mimic the effect of LH/hCG in mice ovulation. However, the molecular process that generates H 2 O 2 in the ovary during ovulation remains largely unknown. DUOX2, a member of the NOX/DUOX family of NADPH oxidase, is capable of generating H 2 O 2 . Results: Using global transcriptome RNAseq, we identied that DUOX2 is one of the transcripts that was markedly upregulated in granulosa cells during ovulation. Treatment with human chorionic gonadotropin (hCG), an ovulatory trigger, signicantly increases the expression of DUOX2 mRNA and protein in human GCs both in vivo and in vitro. hCG-induced up-regulation of DUOX2 is mediated by the cAMP-PKA and the PKC pathway. A functional test reveals that DUOX2 chemical inhibitor, Diphenyleneiodonium (DPI), an NADPH oxidase inhibitor, decreased H 2 O 2 levels in MGCs (Mural Granulosa Cells) treated with hCG. The inhibition of H 2 O 2 by DPI suggests that DUOX2 activity is required for hCG-induced elevation of extracellular H 2 O 2 in MGCs. In vivo treatment of mice with DPI signicantly decreases the number of ovulated oocytes and markedly attenuates the expression of key ovulatory genes. These results support the putative role of DUOX2 in ovulation. Conclusions: DUOX2 is a ROS generator during the ovulatory process and is involved in the LH/hCG-induced signaling cascades leading to ovulation. Treatment with DUOX2 inhibitors may affect late folliculogenesis and ovulation and thus may serve for fertility control.


Background
Reactive oxygen species (ROS) are oxygen-derived small molecules, including superoxide (O 2 •), hydroxyl, peroxyl (RO 2 •), and alkoxyl (RO) and certain nonradicals that are either oxidizing agents and/or are easily converted into radicals, such as hydrogen peroxide (H 2 O 2 ) (1). The ROS are involved in a spectrum of functions including hormone biosynthesis, cell signaling, aging and microbial killing (2). Studies in mouse model revealed that ROS have a role in ovulation (3). In the corpus luteum, they are involved in steroid production (4). ROS dysregulation may lead to oxidative stress which has a detrimental effect on fertility including impaired oocyte quality, follicular atresia and reduced pregnancy rate in IVF (5,6).
Our laboratory has undertaken to systematically identify ovulation-associated genes (7,8). Differentially expressed candidate genes (n = 1746) were identi ed by comparing the transcriptome of compact cumulus cells (CCs) from germinal vesicle cumulus oocyte complexes (COCs) obtained from patients undergoing in vitro maturation (IVM) procedures (CCGV), and expanded CCs from metaphase 2 (CCM2) COCs were obtained from patients undergoing IVF/ICSI. The differentially expressed genes identi ed likely serve as regulators of ovulation, cumulus expansion, and/or oocyte maturation.
The analysis of the resultant transcriptome database revealed that the NADPH oxidase DUOX2 exhibit a 602-fold increase in expanded CCs (CCM2) compared to immature CCs (CCGV). DUOX2 is a major source of ROS (9). Other NADPH oxidases identi ed in cumulus granulosa cells include DUOX1 which exhibited a 14-fold increase in CCM2s. NOX4 and NOX5 were identi ed at negligible levels. Other NOX genes were not expressed.
Prompted by the observation that DUOX2 constitutes a highly expressed periovulatory transcript, we set out to investigate the physiological role of this key NADPH oxidase in the ovulatory process.

DUOX2 expression in human granulosa cells in vivo.
Illumina-based RNA-sequencing (RNAseq) was carried out on RNA extracted from compact (CCGV) and expanded (CCM2) cumulus cells and sequence counts were used to assess gene expression (8). As shown in Fig. 1A, the in vivo expression of DUOX2 transcripts in CCs of expanded post-ovulatory MII COCs after in vitro fertilization (IVF) was 602-fold higher (p < 0.0001) as compared with CCs of compact GV COCs obtained during in vitro maturation (IVM). DUOX1 transcript was induced 14-fold in CCs of expanded post-ovulatory MII COCs but NOX1-5 were below threshold detection level.
In humans, during follicular antrum formation, GCs produced two distinct lineages; Mural GCs that line the follicular wall and cumulus GC (CGC) that surround the oocyte (10). We compared the expression levels of DUOX2 mRNA in MGCs and CGC by qPCR. MGCs and CGCs were obtained from large preovulatory follicles (> 17 mm) during the IVF procedure. As shown in Fig. 1B, CGCs expressed signi cantly higher (2.6-fold) levels of DUOX2 mRNA relative to MGCs (p = 0.018).
The impact of the oocyte-cumulus cell relationships on both maturation and fertilization capacity has been well-established (11,12). CGCs re ect the characteristics of the oocyte, providing a noninvasive means to assess oocyte quality. Several researchers have reported that gene expressions of CGCs could be used as a tool to predict oocyte competence or embryo development (13,14). To reveal whether DUOX2 is involved in oocyte maturation processes, we evaluated the expression of DUOX2 in CGCs isolated from GV and M2 oocytes retrieved during IVF procedures.
As shown (Fig. 1C) DUOX2 was expressed in CGCs isolated from mature M2 COCs at higher levels than in immature GV COCs (p = 0.015).
DUOX2 expression pattern in vitro and in vivo during the ovulation process.
The increase of DUOX2 in our library suggests that it is an hCG target gene. To con rm our in vivo observations, we employ an in vitro system of cultured MGCs. Luteinized MGCs obtained from IVF procedure cultured for 4 days with a daily medium exchange to transform them into "early non-luteinized like MGCs" (15). MGCs were further cultured without hCG (control) or stimulated with 1 U/ml hCG for 24 h. The induction of DUOX2 protein by hCG was detected by western blot (Fig. 1D) and immuno uorescence assay (Fig. 1E). Western blot quanti cation using ImageJ software found that hCG induces DUOX2 protein signi cantly (Fig. 1D, p = 0.013), con rming DUOX2 as an hCG-induced gene. The in vitro system of cultured MGCs time response expression pattern following hCG stimulation may hint for the role of DUOX2 in the ovary. After hCG stimulation, the cultured MGCs cells were harvested at 0, 3, 6, 12, 24 and 36 h later. DUOX2 mRNA was observed to be upregulated in hCG-treated cells, peaking at 12 h after hCG treatment (Fig. 1F) and returning to basal levels at 36 h post-hCG treatment. We also examined the ovarian DUOX2 expression pattern in vivo using the superovulation mouse model. PMSG treated mice were injected with hCG and harvested after 0, 3, 6, 9, and 12 h. The increase in DUOX2 mRNA expression was observed at 9 h and reached maximal levels at 12 hours after hCG injection just before ovulation (Fig. 1G).
Regulation of DUOX2 expression in cultured human mural granulosa cells: intracellular signaling pathways The LHCGR activates multiple signaling pathways, including protein kinase A (PKA) (16) and protein kinase C (PKC) (17). To determine the role of these signaling systems in the hCG-dependent up-regulation of DUOX2 expression, cultured MGCs were treated with FSK (an activator of adenylyl cyclase that in turn leads to the activation of PKA) or PMA (a PKC activator) for 24 h. As shown in Fig. 2, FSK (10 µM) signi cantly increased DUOX2 transcripts in MGCs 10.2-fold compared to control (p < 0.001, Fig. 2A). The effect was abolished following pre-treatment with H89 for 1 hour (a selective PKA inhibitor; 10 µM, p < 0.001). PMA (20 nM) treatment signi cantly increased DUOX2 transcripts in MGCs 4.1-fold, compared to control (p = 0.042, Fig. 2B).
In addition to the PKA and PKC pathways other intracellular signaling cascades, such as the phosphatidylinositol 3-kinase (PI3-K)/Akt pathway may also be involved in the signal transduction of DUOX2. To examine the role of these signaling pathways in the regulation of DUOX2 expression, MGCs were pretreated for 30 min with pharmacological inhibitors for PI3K (10 µM LY294002) pathways, followed by a 24 h treatment with 1 U/ml hCG. The results show that LY294002 (Fig. 4C) did not affect the hCG induction of DUOX2 expression.
In summary, our results indicate that the PKA and PKC pathways are the main mediators of the LH/hCG induction of DUOX2 up-regulation.
Assessment of the role of DUOX2 in hCG-induced H 2 O 2 generation in cultured human mural GCs.
To con rm the role of DUOX2 in the generation of H 2 O 2 in the follicle, MGCs were cultured without hCG (control), hCG or with hCG + DPI (a known non-speci c DUOX2 inhibitor) for 10 h and the H 2 O 2 levels were quanti ed (Fig. 2D).
H 2 O 2 levels were 2.1-fold higher in MGC's that were treated with hCG than in control. The treatment with DPI inhibited H 2 O 2 generation by about 96%. The inhibition of H 2 O 2 by DUOX2 inhibitor DPI suggests that DUOX2 activity is a source for the hCG-induced elevation of extracellular H 2 O 2 in granulosa cells.
The effect of DUOX2 blockade on ovulation in vivo.
To evaluate the physiological role of DUOX2 in the ovulatory process, we tested the effect of DPI on ovulation in mice treated by the PMSG/hCG superovulation protocol. As shown in Table 1, 100% of the mice treated with hCG ovulated, releasing an average of 54.3 oocytes. Concurrent injection of hCG and 5 or 7 mg/kg of DPI decreased the number of ovulated oocytes of the mice with a mean of 43.7 and 28.3 oocytes per ovulating mouse respectively, which represent 19.5% and 47.8% respectively, reduction in ovulation e ciency compared to hCG (p = 0.0477 and p = 0.0001 respectively). Table 1 The effect of hCG and dose-response of DPI on ovulation in a superovulation mouse model.

Treatment
No. of ovulation mice / no. of treated (%) These results show that DPI is partially inhibiting ovulation in mice suggesting a role of DUOX2 in ovulation.
To further validate the ovulation-blocking activity of DPI, we performed a histological examination of ovaries removed from the vehicle-or DPI (7 mg/kg)-treated mice at 16 hours after hCG administration ( intact MGCs and thick multilayered follicular walls, with no signs of impending ovulation, resembling ovaries obtained from mice treated with PMSG alone. These histological results further demonstrate that the treatment of mice with a DUOX2 inhibitor inhibits ovulation, con rming the putative role of DUOX2 in ovulation.
Effect of inhibition of DUOX2 on the levels genes that regulate ovulation and corpus luteum function To characterize the ovulation-inhibiting effect of DUOX2 at the molecular level, we evaluated the effect of treatment with DPI on LHCGR-dependent activities. Speci cally, we set out to study key representative genes of the ovulatory cascade. These genes included those involved in oocyte maturation (Areg, Ereg) (18,19), progesterone receptor signaling (Pgr) (20), cumulus expansion (Tnfaip6) (21), and steroidogenesis (Star) (22).
As shown in Fig. 3A, treatment with 7 mg/kg DPI inhibitor and hCG; signi cantly inhibited the expression of all genes examined compared to hCG only.
To assess the corpus luteum function, we measured serum progesterone concentration in PMSGprimed/hCG-triggered mice in the absence or presence of 7 mg/kg of DPI. Concurrent treatment with DPI did not affect serum progesterone levels compared to hCG only (Fig. 3B).

Discussion
The ovary is a metabolically active organ and generates reactive oxygen species (ROS) and reactive nitrogen species (RNS) on an extraordinary scale during various physiological processes including ovulation and control of oocyte maturation. In mammals, free radicals are generated by H 2 O 2 generators enzymes and as by-products of normal cellular metabolism. Free radicals serve as key signal molecules in various physiological and pathological processes.
In vivo studies in combination with ex vivo and molecular analyses were used to provide the evidence that reactive oxygen species present in the preovulatory ovarian follicles are essential for the ovulatory process (3). However, knowledge of the molecular process that provides reactive oxygen species in the ovary and during the ovulatory process is only partial. Earlier reports show that Nox/Duox NADPH oxidases are considered the primary, regulated sources of reactive oxygen species (ROS). These enzymes are expressed in diverse cells and tissues, and their products are essential in several physiological settings (23).
Ovulation is a key event in mammalian female reproduction, therefore the discovery of key players involved in this process is extremely important. Our laboratory applied the RNA sequencing method to systematically isolate genes with an ovulation-selective pattern of expression (8). DUOX2 was among the highly upregulated genes in our library. Taking into account the importance of reactive oxygen species for the ovulatory process (3), DUOX2 was selected for further study of its ovulation-associated expression and function.
Duox enzymes were initially identi ed as H 2 O 2 sources involved in thyroperoxidase mediated organi cation of iodide during thyroid hormone biosynthesis (24,25). Duox enzymes are also detected in exocrine (salivary) glands and on mucosal surfaces (airways, gastrointestinal tract), where they were proposed to support the antimicrobial activity of lactoperoxidase (23). Whereas their producer in the ovary and its contribution to the regulation of ovulatory cascade remains largely unknown.
In vitro studies have shown that LH surge induces the generation of ROS that is required for successful ovulation; Moreover, H 2 O 2 was shown to fully mimic the effect of LH (3). However, except DUOX2, of the NADPH oxidase family only DUOX1 was present and upregulated in our library of ovulation associated genes (Fig. 1A) but at much lower expression than DUOX2 (40-fold), suggesting that the inhibitory effect of DPI is primarily mediated through DUOX2. Further studies using DUOX2 conditional-knockdown technology may con rm these results.
Recently, it was suggested that NOX4 is a major producer of H 2 O 2 in human GCs in vitro (27). They were able to show that NOX1/4 inhibitor (GKT137831) shows a partial decrease in H 2 O 2 release in culture MGCs, suggesting the function of NOX1/4 as a mediator of H 2 O 2 in the ovary. However, in accordance with our ndings, they show that hCG did not affect levels of NOX4 and NOX5, but did signi cantly increased DUOX1 and DUOX2. Therefore, we cannot rule out that NOX4 is a source for H 2 O 2 in the ovary but the fact that NOX4 levels are not under the control of FSH or LH/hCG in GCs, does not support their potential role in the ovulatory process, which is in agreement with our results.

Conclusions
In summary, our study reveals DUOX2 as an important gene in the ovulatory process. We showed that DUOX2 is upregulated in CCs of hCG treated GCs and that the cAMP-PKA pathway but not by the PKC pathway mediates DUOX2 regulation. Moreover, we have demonstrated DUOX2 functions as an H 2 O 2 generator in the ovary. These ndings were further substantiated by in vivo assays. The effect of DPI, a DUOX2 chemical inhibitor, on mice ovulation show a dose-dependent manner partial blocking effect on ovulation and resulted from an inhibition of the hCG-induced ovulatory cascade in mice, including oocyte maturation and follicular rupture but not the steroidogenic pathway. All these ndings mark DUOX2 as a novel ovulatory gene, lling the gap between hCG induction of ROS/ H 2 O 2 and ovulation.
A better understanding of the precise mechanism of H 2 O 2 generation inside the follicular microenvironment during nal follicular maturation by DUOX2 may enable us to de ne new targets for improvement of in vitro maturation (IVM) media, fertility control, and contraception. was administered to trigger ovulation. Oocyte retrieval was performed 36 h following hCG triggering by transvaginal ultrasound-guided needle aspiration.

Cumulus granulosa cell collection
After COC retrieval, CCs of each oocyte were removed with the use of hyaluronidase (SAGE, Trumbull, CT, USA) and a glass denudation pipette (Swemed, Billdal, Sweden). The CGCs were washed in Phosphate-Buffered Saline (PBS) and centrifuged at 5000 × g for 5 minutes at room temperature. The resulting pellets were stored at −80°C until RNA isolation. CCs of individual oocytes were classi ed as per the corresponding oocyte maturation stage: CCs from GV oocytes (CCGV) and CCs from MII oocytes (CCM2). CCs obtained from individual oocytes were collected from individual subjects were pooled to generate a single replicate (n=3-4 different subjects). Each experiment was performed at least three times.

Mural Granulosa Cell Collection and Grouping
MGCs were collected from the aspirated follicular uid (Follicles size >= 17 mm) and re-suspended in a phosphate-buffered solution (PBS; Sigma-Aldrich-Aldrich, St Louis, MO, USA). After allowing the cells to settle by gravity for a few minutes, the top portion of the medium was repeatedly re-suspended and aspirated until such time that the medium proved clear. The cells were then centrifuged at 1000 rpm for 5 minutes at room temperature. The resulting pellets were stored at −80°C until RNA isolation. MGCs from 3 different subjects were pooled to generate a single replicate.
Mural Granulosa Cell Culture MGCs were collected as described above and placed on a percoll gradient and centrifuge at 3000 RPM for 15 min. The MGCs were collected and washed with PBS, counted and plated in 24-well plates at a density of 100,000 cells/well, and incubated at 37°C in a humidi ed atmosphere with 5% CO2 in the air.

RNA Extraction and qPCR
RNA extraction and qPCR. Total RNA was extracted from MGCs or CCs using a Mini/Micro RNA Isolation I kit (Zymo Research, CA, USA) according to the manufacturer's instructions. RNA purity and concentration were assessed using a NanoDrop spectrophotometer (NanoDrop 2000C, Thermo Scienti c Waltham, MA, USA). Total RNA (25ng) from each sample was used for cDNA synthesis by a high capacity reverse transcription kit (Applied Biosystems, Carlsbad, CA, USA) according to the manufacturer's instructions in a 10μl total volume reaction. mRNA levels were analyzed by real-time PCR using the StepOnePlus realtime PCR system (Applied Biosystems). The real-time PCR mix contained 1μl of cDNA, fast SYBR Green Master Mix (Applied Biosystems), and speci c primers for DUOX2 or another gene of interest and β-actin (housekeeping gene) in a total volume of 10μl. Cycling parameters were: 1 cycle at 95°C for 20 seconds and 40 cycles each at 95°C for 3 seconds and 60°C for 30 seconds. A melting curve analysis was performed at the end of each run to ensure a single amplicon. All samples were run in duplicates. Analysis of the qPCR results was carried out with StepOne software. Relative gene expression was calculated using the delta-delta Ct method. Details of the primers used are shown in Table S1.

Immuno uorescence
Immuno uorescence staining and confocal analysis. MGCs were seeded on coverslips. The cells were xed with cold 4% paraformaldehyde for 15 minutes, permeabilized with 0.1% Triton X-100 for 10 minutes, blocked in 1% BSA and 10% normal donkey serum in PBS for 10 minutes, incubated with anti-DUOX2 antibody (Thermo Fisher) diluted 1:100 for 1 hour at room temperature, washed 3 times, stained with FITC-conjugated donkey anti-mouse antibody for 1 hour at room temperature, and washed again 6 times. The slides were mounted with coverslips using GelMount (Biomeda), and the cells were analyzed with a Leica SP5 (Leica) confocal laser scanning microscope.
Superovulation protocol 25-day old C57BL female mice were injected with 10U of Pregnant mare's serum gonadotropin (PMSG, Sigma, St. Louis, MO, USA) to stimulate follicle growth, and 48 h later with an ovulatory dose of 10U hCG that mimicking an endogenous LH surge, stimulating follicle development and ultimately ovulation 12-16 h later. The animals were sacri ced 48 h after PMSG treatment as well as 3, 6, 9, 12, 16 and 24 hours after hCG administration.
All mice were sacri ced by CO 2 asphyxiation, and the ovaries were removed and either frozen or paraformaldehyde-xed until used, or punctured to collect entrapped oocytes. Blood samples were collected at the time of euthanasia for progesterone measurement, and the number of oocytes within the ampullas of all oviducts was recorded. DPI (Almog diagnostic 5 or 7 mg/kg) was injected at the same time as hCG to PMSG-primed mice. Mice treated with hCG and vehicle (DMSO) served as controls. Mice were sacri ced 16 hours after the administration Mouse ovarian morphology Fixed ovaries (4% formalin) were embedded in para n blocks, mounted on slides, and stained with hematoxylin/eosin. Mouse ovarian morphology was assessed by examining 4-μm serial histological sections cut by a microtome.

Measurement of progesterone concentrations
Blood samples for hormone assays in female mice were obtained at the time of euthanasia by cardiac puncture. Sera were separated from whole blood and frozen until the time of analysis. Progesterone concentrations were measured in duplicate by the American Medical Laboratories in Herzliya, Israel.
MGCs aspirated during IVF procedures were cultured for 4 days with daily medium exchange (Medium 199 Sigma Aldrich St Louis Mo). MGCs were stimulated with hCG 1 U/ml. 15 hours after hCG administration, the medium was exchanged to DMEM medium (DMEM/F-12, no Phenol Red) 1% FBS (Invitrogen Grand Island, NY) and 1% penicillin/streptomycin (Sigma Aldrich St Louis Mo). MGCs were stimulated either with vehicle (control), 1 U/ml hCG, with or without DPI inhibitor 10 μM. CaCl 2 1 mM was added to all cells.
Detection of H 2 O 2 using the AmplexR Red Hydrogen Peroxide/Peroxidase Assay Kit (Invitrogen Catalog no. A22188). The reaction was started by adding a working solution of 100 μM Amplex® Red reagent and 0.2 U/mL HRP to each microplate well containing the standards, controls, and samples. The plate was incubated for 1 hour at room temperature protected from light. Fluorescence was then measured with a uorescence microplate reader. Fluorescence emission detection at 570 nm. Background uorescence, determined for a no-HRP control reaction, has been subtracted from each value.

Statistics
Each experiment was carried out at least three times. Data, expressed as mean ± SEM, were evaluated with Student's t-test (two-tailed) or with ANOVA for more than 2 variables using the post hoc Tukey test assuming equal variances or the Games-Howell test for unequal variances. When appropriate, Kruskal-Wallis non-parametric comparison test was used. SPSS 20 software (IBM) was used for all analyses. P values < 0.05 were considered statistically signi cant.

Declarations
Ethics approval and consent to participate The study was approved by the local Institutional Review Board (IRB) committee of Chaim Sheba Medical Center, Tel Hashomer (ethical approval number SMC-11-8707 and SMC-12-9342). Written informed consent was obtained from each patient who provided samples. All experiments involving mice were conducted in compliance with the principles of the National Research Council (NRC) and were approved by the institutional animal care and use committee (IACUC) #919/14/ANIM.

Consent for publication
Not applicable.

Availability of data and materials
The datasets generated and analyzed during the current study are available in the GEO repository, GSE50174. All other data generated or analyzed during this study are included in this published article.

Competing interests
The authors declare that they have no competing interests. H. contributed to the study design and analysis, the interpretation of the data, and the writing of the manuscript. O. S. and Y.Y. were involved in sample preparation and in vivo and in vitro experimental work. All the authors have contributed to data analysis and to nalizing the manuscript.