Cumulus cell antioxidant system is modulated by patients' clinical characteristics and correlates with embryo development.

Purpose To study whether the cumulus cell antioxidant system varies accordingly to patients clinical characteristics’ as age, infertility diagnosis, BMI, and stimulation protocol applied and if the antioxidant profile of cumulus cells could be used as a predictor of embryo development.

A prospective study including 383 human cumulus samples provided by 191 female patients undergoing intracytoplasmic sperm injection during in vitro fertilization treatments from a local in vitro fertilization center and processed in university laboratories. Catalase (CAT), superoxide dismutase (SOD), glutathione peroxidase (GPx), and glutathione S-transferase (GST) enzyme activity levels and reduced glutathione (GSH) levels were measured in cumulus oophorus cells individually collected from each aspirated cumulus-oocyte complex, and the results of each sample were compared considering the oocytes outcome after ICSI and patients clinical characteristics. A total of 223 other human cumulus samples from previous studies were submitted to a gene expression meta-analysis.

The antioxidant system changes dramatically depending on patients’ age, infertility diagnosis, stimulation protocol applied, and oocyte quality. SOD activity in cumulus cells revealed to be predictive of top-quality blastocysts for young patients with male factor infertility (P < 0.05), while GST levels were shown to be extremely influenced by infertility cause (P < 0.0001) and stimulation protocol applied (P < 0.05), but nonetheless, it can be used as a complementary tool for top-quality blastocyst prediction in patients submitted to intracytoplasmic sperm injection technique (ICSI) by male factor infertility (P < 0.05).

Through a simple and non-invasive analysis, the evaluation of redox enzymes in cumulus cells could be used to predict embryo development, in a personalized matter in specific patient groups, indicating top-quality oocytes and improving success rates in in vitro fertilization treatments. Trial registration The trial was registered at UFRGS Research Ethics Committee and Plataforma Brasil under approval number 68081017.2.0000.5347 in June 6, 2019. Supplementary Information The online version contains supplementary material available at 10.1007/s10815-022-02496-y.

Cumulus oophorus cells, Oocyte quality, Blastocyst formation, Antioxidant metabolism, Redox metabolism

The release of the cumulus-oocyte complex during ovulation is a remarkable process of the mature follicle, caused by several orchestrated signs which culminate with a physiological inflammatory reaction [1, 2]. Per se, this dynamic action triggers oxidative processes in the ovaries and the generation of antioxidant responses in each follicle maturation cycle [3]. While physiological amounts of oxidants are necessary for cell signaling and ovulation, elevated levels of reactive species with insufficient activity of antioxidant defenses can cause a detrimental imbalance and oxidative stress, affecting long-term oocyte quality and subsequent embryo development [4–8]. Physiological processes of cells, such as cellular respiration by mitochondrial oxidative phosphorylation (OXPHOS) activity and inflammation, generates reactive species that can damage essential compounds. However, cells have developed antioxidant mechanisms to neutralize those molecules, which can be enzymatic or non-enzymatic. The superoxide dismutase enzymes (SOD-1, 2, and 3 isoforms) catalyze the dismutation of superoxide anions (O2•−), a free radical, into molecular oxygen and hydrogen peroxide (H2O2), a slightly less reactive form [9]. SOD1 knockout mice are viable, but subfertile [10, 11], illustrating the importance of this defense enzyme for fertility. Resulting peroxide can either be catalyzed into H2O and O2 by catalase (CAT) [12] or H2O by glutathione peroxidase (GPx), through the oxidation of the tripeptide glutathione (GSH), the most abundant thiol in animal cells, to GSSG [13]. The oxidized glutathione can be regenerated through the action of glutathione reductase, using NADPH as electron donor (GSSG-Rx) [14]. In spite of these cellular antioxidant defenses, if biological molecules end up suffering oxidation, they can be recycled (reduced) by systems, such as thioredoxin/thioredoxin reductase [15], or eliminated from the cell. The latter is achieved by glutathione S-transferases (GST), which catalyzes the conjugation of several molecules and xenobiotics compounds with GSH [16], generating a less reactive compound that is then eliminated from the cell by efflux pumps (e.g., the P-glycoprotein 1) at the plasma membranes. Successive oxidative stress insults in the follicular microenvironment might provoke lower oocyte quality. Data derived from assisted reproduction clinics suggests that redox imbalances are a major cause of fertility failure [17, 18]. The cumulus oophorus cells (CCs) are connected to the female gamete, significantly influencing and being influenced by the oocyte through their direct paracrine communication [19]. CCs and the oocyte share a specific coordination of oxidant and antioxidant productions [8, 20] (Fig. 1). CCs are routinely discarded during intracytoplasmic sperm injection (ICSI) treatments, therefore representing a non-invasive source of valuable biological information [21]. Female health conditions [22, 23], age [24], lifestyle (smoking habits, physical activity routine, diet) [8, 25], and environmental factors (such as air quality) [26] are known to impact oocyte health and the success of fertility treatments [8]. While patient population in assisted reproduction clinics is highly heterogeneous, the majority of studies in search of oocyte quality biomarkers only analyzes a specific patient subgroup, considering a single stimulation protocol, or infertility diagnosis and age group (normally only including patients under 35 years old, which differ from world tendency of patients seeking infertility treatments). Although some studies analyzing the redox status of follicular fluid (FF) [27–31] showed positive correlations between antioxidant enzyme levels and patients’ reproductive health [32], oocyte maturation [33], fertilization [28, 31, 34], and successful pregnancy [29], the FF has a very similar composition to blood plasma [35, 36] and might be more representative of the patients’ overall health then of the oocytes’ condition. Likewise, several groups analyze gene expression in CCs in order to identify possible biomarkers of oocyte quality, but little consistency is seen between studies [37–43]. There are only a few studies addressing the activities of isolated antioxidant enzymes in human CCs and its relation with oocyte quality [44, 45]. Further elucidations of the impact of redox metabolism in human CCs and in assisted reproduction procedures are still needed. Here, we analyzed expression levels of human antioxidant genes (HAG) [46] in 223 individually collected human CC samples obtained from public datasets [43, 47–53]. Then, we experimentally assessed enzymatic activities of several components of antioxidant system in 383 CC samples with different patient profiles (such as distinct infertility diagnosis, age, and stimulation protocol applied), elucidating if changes in redox metabolism in CCs are associated with patient’s clinical characteristics. Furthermore, we searched for possible biomarkers to be used in the clinical environment, which could be applicable either to all patients or, in a personalized manner, to a specific patient subgroup. Next, we compared our findings on antioxidant enzyme activities with the metadata analysis of public microarray and RNAseq datasets to verify if gene expression could be used as a proxy of enzymatic activity pattern.

Differential expression analysis and gene set enrichment analysis Microarray and RNAseq data were obtained in public repository Gene Expression Omnibus (GEO) under the accession numbers GSE34230, GSE81579, GSE37277, GSE113239, GSE10946, GSE40400, GSE155489, GSE31681, and GSE9526. The entire bioinformatics pipeline was conducted in R statistical environment [54]. For microarray data, uninformative probes were removed and duplicated probes were filtered according to their variance using genefilter package [55]. Differential expression analysis was implemented using the limma package (PMID 25,605,792). Statistical difference was considered for genes with P-value < 0.05, and logFC was taken as a proxy of up- or downregulated (Table 1). Table 1. | GEO ID | Comparison | n | Original study aim | Ref | DEGs | Gene description | LogFC | P value | |---|---|---|---|---|---|---|---|---| | GSE37277 | Blastocyst vs. arrested development | 80 | Identify at the level of cumulus cells (CCs) any genes related to oocyte developmental competence | [47] | CAT | Catalase | − 0.15 | 0.04 | | GPX3 | Glutathione peroxidase 3 | 0.28 | 0.04 | ||||| | GPX7 | Glutathione peroxidase 7 | 0.19 | 0.007 | ||||| | GSTM3 | Glutathione S-transferase mu 3 | 0.38 | 0.0008 | ||||| | GSTO1 | Glutathione S-transferase omega 1 | 0.18 | 0.03 | ||||| | GSTT2 | Glutathione S-transferase theta 2 | 0.63 | 0.014 | ||||| | MGST1 | Microsomal glutathione S-transferase 1 | − 0.27 | 0.014 | ||||| | GSE81579 | Younger vs. older | 20 | Identify differentially expressed genes (DEGs) between CCs from younger and older patients | [106] | GSTK1 | Glutathione S-transferase kappa 1 | 0.21 | 0.0009 | | GSTM1 | Glutathione S-transferase mu 1 | − 0.05 | 0.005 | ||||| | GSTZ1 | Glutathione S-transferase zeta 1 | 0.15 | 0.015 | ||||| | GSS | Glutathione synthetase | 0.17 | 0.004 | ||||| | GSE34230 | GnRH agonist vs. GnRH antagonist | 46 | Assess differences in gene expression profiling in cumulus cells retrieved from patients undergoing GnRH agonist and GnRH antagonist IVF treatments | [49] | GPX3 | Glutathione peroxidase 3 | − 0.44 | 0.03 | | MGST3 | Microsomal glutathione S-transferase 3 | − 0.47 | 0.05 | ||||| | GSE113239 | Positive vs. negative | 10 | Provide insights into the determination of clinical pregnancy and live birth outcomes in the context of differential gene expression analysis of cumulus cells as predictors of clinical pregnancy and/or live birth after single embryo transfers | Not published | GSTZ1 | Glutathione S-transferase zeta 1 | 0.70 | 0.04 | | SOD3 | Superoxide dismutase 3, extracellular | 0.83 | 0.04 | Transcriptomic datasets from several human CC samples were obtained from public repositories. CC samples from COCs containing oocytes that further successfully developed into blastocyst stage were compared to samples from follicles whose embryos presented an arrested development in GSE37277. Using GSE81579, it is possible to observe the changes in gene expression caused by patients’ age. GSE34230 highlighted differences caused by distinct stimulation protocols. The final comparison, GSE113239, compares CC samples from oocytes that successfully resulted into a pregnancy and samples with negative outcomes after embryo transfer For comparison between PCOS and non-PCOS patients, microarray datasets GSE10946, GSE40400, GSE9526, and GSE31681 were merged and collapsed onto unique gene symbol annotation identifiers. This was accomplished using the virtualArray package (version 18.0) [56]. In addition, dataset batch correction was performed using the vsn package (3.60.0) [57]. Differential expression was computed using the limma package [58] in R statistical environment. For RNAseq data, firstly, raw RNA-seq data from GSE155489 was downloaded using the SRA Toolkit (https://github.com/ncbi/sra-tools). Afterwards, transcript alignment was performed using Salmon (v1.3.0) [59], mapped to a reference genome with the index derived from Homo sapiens Ensembl build. Aligned reads were summarized using tximport (v1.12.3) [60], and genes with mean count < 2 were filtered out. Afterwards, processed expression data was analyzed with the DESeq2 (v1.28.1) [61] method for differential expression. Genes with FDR adjusted P-value < 0.05 were considered as differential expression genes (DEGs). Furthermore, the behavior of redox metabolism genes as a group was analyzed in all datasets through Gene Set Enrichment Analysis (GSEA) [62] using HAG network [46] and the logFC obtained from differential expression analysis of datasets. HAG network was constructed using STRINGdb (PMID23203871) and RedeR (PMID22531049) packages and STRING database version 10. Ethical considerations This study was approved by the Research Ethics Committee of the Federal University of Rio Grande do Sul (UFRGS) (#68,081,017.2.0000.5347). CCs were obtained as waste products of ICSI procedures and had no other destination beyond the experiments described here. Written informed consent was obtained from all patients participating in this study, and samples and patient clinical data were then supplied anonymously to the laboratory. Patients and samples Patient samples were retrieved from a local fertility clinic. CCs were isolated from women who were submitted to ICSI-ET. Three hundred and eighty-three samples from 191 patients were included in this study (Table 2). Since individual collection of CCs requires a longer manipulation time of oocytes, two types of sample collection were used to minimize the influence of the study: Pooled CC samples with all CCs being collected from all COCs of the patient during the same stimulation cycle, and individually collected cumulus complexes, with CCs being collected from each follicle individually. Therefore, pooled samples were used for analysis of patient-related characteristics, since those are not affected by the oocytes, and individually collected samples were used for analysis of oocyte-related and intrapatient characteristics. One hundred and forty-six patients provided pooled samples. Forty-five patients provided 237 individualized CC samples. Each patient sample included information about age, clinical infertility diagnosis, body mass index (BMI), ovarian stimulation protocol, number of retrieved, injected, and fertilized oocytes and blastocyst rates. Table 2. | Individual samples | Pooled samples | || |---|---|---|---| | Total samples | 237 | 146 | | | Total patients | 45 | 146 | | | Infertility diagnosis | Controls (male, tubarian) | 182 | 95 | | Polycystic ovary | 13 | 21 | | | Endometriosis | 13 | 15 | | | Age | ≤ 35 years | 102 | 79 | | > 35 years | 129 | 67 | | | Stimulation protocol | ANT + FSH | 50 | 27 | | ANT + HMG | 33 | 16 | | | ANT + FSH + HMG | 110 | 82 | | | AGON | 18 | 10 | Samples were retrieved from a local fertility clinic, along with patient clinical information. Numbers in bold format represent total number of patients and samples in each collection format (individual or pooled) ANT = GNRH (gonadotropin-releasing hormone) antagonist stimulation, FSH = follicle-stimulating hormone, HMG = human menopausal gonadotropin, AGON = GNRH agonist stimulation The corresponding embryos from distinct CC samples were tracked individually until day 5 of culture and analyzed for developmental capacity. Embryos were morphologically analyzed using Gardner’s blastocyst grading [63, 64], presenting a well-defined blastocoel at day 5 and with expansion rate equal or superior to 3 and inner cell mass/trophectoderm ratings between A and B were classified as blastocysts (B). Oocytes that were fertilized but did not reach blastocyst stage and did not present a blastocoel were classified as arrested development (AD). Ovarian stimulation and luteal phase support Controlled ovarian stimulation followed the methods adopted by the clinic, which consisted of different short stimulation protocols, with administration of gonadotropin releasing hormone (GnRH) antagonist (Orgalutran®, Schering-Plough, Brazil), with or without recombinant (Puregon®, Organon, Holland) or urinary follicle-stimulating hormone (FSH) (Fostimon®, IBSA Institut Biochimique S.A., Switzerland), and human highly purified menopausal gonadotropin (HP-hMG) (Menopur®, Ferring Pharmaceuticals, Copenhagen, Denmark), with or without clomiphene citrate (Clomid®, Medley, Brazil). Long protocol (approximately 30 days) was also used with some patients, with GnRH agonist. Ultrasonography follow-up of the cycle initiated on the seventh day of stimulation was performed daily or at every 2 days, and the gonadotropin dose was adjusted according to the follicular growth observed (between 225 and 300 IU). When the patient presented follicles with a diameter over 1.8 cm, the gonadotropin trigger was performed. About 34 to 36 h after administration of recombinant human chorionic gonadotropin (HCG) (Ovidrel®, Serono, Brazil), each patient underwent oocyte retrieval under intravenous sedation with propofol (Diprivan®, Astra- Zeneca, Brazil) and fentanyl citrate (Fentanyl, Janssen-Cilag, Brazil). The number of samples collected from patients using different stimulation protocols is informed in Table 2. Oocyte and cumulus cell retrieval All COCs retrieved from each patient were placed together either on cell culture plates (2004 FIV; Ingamed, Brazil) (for pooled samples) or in individual drops for individual samples, filled with human tubal fluid-HEPES culture medium (HTF) (Irvine Scientific, USA) supplemented with 10% synthetic serum substitute (SSS; Irvine Scientific), covered with mineral oil (Sigma-Aldrich, Brazil), and incubated at 37 °C in 5.8% CO2 and 95% humidity for 2 h. After this period, the oocytes were denuded with hyaluronidase (H4272 type IV-S, Sigma; 40 IU/mL) for 30 s, and CCs were mechanically removed in HTF-SSS with the aid of a stripper pipette (130 mm; Denuding Pipette, Cook). Samples were then placed in a centrifuge tube either individually from each COC or pooled from each patient. The media containing the CCs were centrifuged (2000 g/10 min) in 1.5-mL tubes. After centrifugation, the supernatants were discarded and the CC samples were conditioned in Eppendorf tubes containing 100 µL lysis buffer (12.1 M HCl, 5 mM EDTA, 0.5% Nonidet P40, 150 mM NaCl, protease inhibitor) and stored at − 80 °C until experimentation. Intracytoplasmic sperm injection and embryo development assessment Mature oocytes characterized by the extrusion of the first polar body were submitted to ICSI 2 to 4 h after oocyte retrieval. About 16 to 18 h after ICSI, fertilization was assessed on the basis of the presence of two pronuclei and two polar bodies. On day 5 after ICSI, the presence of blastocoel was determined. Gardner grading [65] was applied in all blastocysts analyzed in this study. Only embryos with expansion rate equal or superior to 3 and inner cell mass/trophectoderm ratings between A and B were considered in the “blastocyst” group. Embryos were cultivated in Global® Total® culture media (LifeGlobal®, Brazil). Reagents and equipment All reagents were obtained from Sigma-Aldrich (São Paulo, Brazil), except when indicated. Assays were conducted in appropriate 96-well plates, and readings were assessed in a SpectraMax i3 spectrophotometer (Molecular Devices). Redox metabolism analysis Each assay was standardized for human CCs using 96-well plates, previously to this study [20], and the optimal amount of protein for each enzymatic assay considering duplicates was determined: 4 µg for Bradford assay, 30 µg for GSH, 40 µg for GPx, 35 µg for SOD, and 10 µg for CAT. Determination of total protein content The protein content was measured by the modified method of Bradford [66] with the use of an albumin standard curve for protein determination. Protein levels were used for assay normalization. SOD (EC 1.15.1.1.) activity SOD activity was measured in a spectrophotometer at 480 nm by inhibition of 1 mM epinephrine auto-oxidation in 50 mM glycine–NaOH buffer (pH 10) at 35 °C. One unit is defined as the enzyme amount that inhibits the rate of reaction by 50% [67]. This method detects all SOD isoenzymes since it is based on substrate consumption on samples with lysed membranes. A standard curve was used for SOD activity determination. CAT (EC 1.11.1.6) activity CAT activity was determined in a spectrophotometer at 240 nm in 50 mM phosphate buffer (pH 7.0) by measuring the decomposition rate of 10–50 mM hydrogen peroxide for 20 min at 25 °C [68]. One unit is defined as the reaction velocity constant of the first order (k) in s − 1. A standard curve was used for CAT activity determination. Glutathione peroxidase (GPx) (EC 1.11.1.9) activity Glutathione peroxidase activity was recorded spectrophotometrically at 340 nm by measuring the oxidation of reduced glutathione (GSH) by tert-butyl hydroperoxide. The substrate is maintained at a constant concentration by the addition of reduced nicotinamide adenine dinucleotide phosphate (NADPH) in a reaction mixture containing 1 mM azide, 0.5 mM tert-butyl hydroperoxide, and 50 mM phosphate buffer (pH 7.0) at 30 °C in the presence of 1 mM reduced glutathione and 0.25 U/mL glutathione reductase. The decrease in absorbance of the reaction mixture at 340 nm is a measure of the NADPH oxidation. One unit is defined as 1 µmol of glutathione oxidized/min or 1 µmol NADPH oxidized/min [69], according to the method of Lawrence and Burk [70]. This method detects all GPx isoenzymes since it is based on substrate consumption on samples with lysed membranes. A standard curve was used for GPx activity determination. GST (EC 2.1.5.18) activity Total GST activity is measured upon conjugation of the thiol group of reduced glutathione to the CDNB substrate, which is read at 340 nm [71] and by measuring conjugation with MCB, producing a fluorescent molecule detected at Ex/Em = 380/460 nm. This method detects all GST isoenzymes since it is based on substrate formation on samples with lysed membranes. A standard curve was used for GST activity determination. GSH concentration assay GSH concentrations were measured according to Browne and Armstrong [72] with minor modifications. Lysed samples (1 µg protein/µL) were first deproteinized with metaphosphoric acid, centrifuged at 7000 g for 10 min, and supernatant was immediately used for GSH quantification. One hundred and eighty-five microliters of 100 mM sodium phosphate buffer (pH 8.0) containing 5 mM ethylenediaminetetraacetic acid and 15 μL of o-phthaldialdehyde (1 mg/mL) were added to 30 μL of supernatant previously deproteinized. This mixture was incubated at RT in a dark room for 15 min. Fluorescence was measured using excitation and emission wavelengths of 350 and 420 nm, respectively. The calibration curve was prepared with standard GSH (0.001–1 mM) and the concentrations, determined in triplicate for each experimental condition, and referred to as 10–12 mol GSH/mg protein. Statistical analysis Experimental data were expressed as means ± SD, and P values were considered significant for P < 0.05. For intrapatient analysis, individual CC samples from each follicle of the same patient were analyzed through a paired t test or Wilcoxon test, selected based on the normality of groups, for each parameter to test the correlation with embryo development. Patients that presented CC samples corresponding to COCs yielding an oocyte of good quality (B) and samples from follicles from poor quality oocytes (AD) were added in this phase of the study. Group normality were assessed through D’Agostino-Pearson and Shapiro–Wilk normality tests (GraphPad® Software 5.0). One hundred and forty-seven pooled samples from all punctured follicles from the same patient were also included in this study, to analyze the influence of patients’ characteristics in redox metabolism of CCs. To make the joined analysis of pooled and individual samples possible, the values were normalized to Z-score. When comparing more than two groups, Kruskal–Wallis test or one-way ANOVA analysis was implemented, based on group normality.

and discussion Although a healthy full-term pregnancy is the ultimate goal of in vitro fertilization (IVF), it involves several complex clinical variables (such as implantation window, uterus health, and endometrial receptivity and patients’ habits). Here, we chose blastocyst formation as the main outcome because it is the last stage possible to track the embryo in vitro in a controlled clinical environment. Even though less than 35% of the fertilized retrieved oocytes of a cycle undergo cavitation to blastocyst stage, the transfer of high-quality blastocysts to the maternal uterus can lead to pregnancy rates of over 80% [64]. Moreover, multiple embryo transfer often provokes twinning, which increases the risk of possible complications and morbidity [73]. A marker of top-quality blastocyst would also be useful in oocyte cryopreservation techniques, providing extra information regarding the quality of morphologically normal preserved oocytes. Therefore, anticipating blastocyst formation can significantly impact the success of IVF techniques and abbreviate the waiting for a take home baby. In this study, the redox metabolism of CCs and its relation with oocyte quality, blastocyst development, patients’ clinical characteristics (age, diagnosis, body mass index — BMI), and cycle characteristic (stimulation protocol applied) were assessed. DEG level analysis We first analyzed the differential gene expression levels from the human antioxidant gene (HAG) network [46] in public microarray datasets (Table 1). Nine datasets where obtained by a survey in public repositories: We compared (1) CC samples from oocytes that successfully generated top-quality blastocysts on day 5 after ICSI versus CCs from oocytes that presented embryos with arrested development (n = 80) (GSE37277); (2) samples from younger (≤ 35 years old) versus older (> 35 years old) patients (n = 20) (GSE81579); (3) samples from patients who received gonadotropin-releasing hormone (GnRH) agonist stimulation versus GnRH antagonist (n = 46) (GSE34230); (4) CC samples from oocytes that further generated pregnancy versus CCs from oocytes that generated embryos that failed to implant after transfer (n = 10) (GSE113239) and CC samples from PCOS versus non-PCOS patients (n = 77) (GSE10946, GSE40400, GSE155489, GSE31681, and GSE9526 were included as controls). Positive log fold change (logFC) values represent genes upregulated in the first group appearing in the comparison (e.g., blastocyst, young, GnRH agonist, positive, and PCOS) and negative logFC values in genes upregulated in the second group (arrested development, old, GnRH antagonist, negative and non-PCOS). These analyses revealed that, while CAT appears to be less expressed in CCs from oocytes that further generated top-quality blastocysts, GPX and GST isoforms (GPX3, GPX7, GSTM3, GSTO1, GSTT2) are upregulated in this group. Likewise, younger patients showed significantly higher GST expression levels (GSTK1, GSTZ1, GSS), but not for all isoforms detected (GSTM1 was found to be downregulated). GnRH antagonist protocols seems to stimulate GPX3 and MGST3 expression when compared to agonist protocols. CCs from oocytes that further generated pregnancies appear to present higher levels of GSTZ1 and SOD3 expression than the ones that fail to implant. Microarray data analysis from PCOS and non-PCOS patients did not reveal any significant differences between groups. However, when analyzing RNA seq data, the differential expression analysis resulted in over 1500 genes (FDR adjusted P-value < 0.05). Among these, we observed seven HAG genes upregulated (PRDX6, GSR, TXNDC17, PDIA6, TXNL1, TXNRD1, PRDX2). GSEA Fig. 2A shows protein–protein interaction (PPI) network between products of HAG genes. Since the redox potential of cells reflects the balanced activities of several components of the antioxidant system, we also evaluated the global activity of the HAG network using GSEA in selected datasets. The goal of GSEA is context-specific as it aims to determine whether a group of genes is mostly correlated with a phenotypic class of interest using a gene set–based statistical framework [62] Fig. 2B shows HAG PPI network topology of P-value and logFC from GSE37277, indicating that in this study, HAG is mostly correlated with high-quality blastocysts. Indeed, this is further demonstrated in GSEA, which showed a significant enrichment score for this group of genes. These results suggest that the HAG gene set was overexpressed in oocytes that successfully generated top-quality blastocysts (B) on day 5 after ICSI versus CCs from oocytes that presented embryos with arrested development (AD) (Fig. 2B and 2D). Similarly, in Fig. 2C, HAG PPI network topology of P-value and logFC from GSE81579 also indicated that this process was upregulated in younger women (Fig. 2C). This was also showed statistically through GSEA (Fig. 2E), suggesting a global decrease in the expression levels of HAG components during CC aging. Core enrichment is the subset of genes in a gene set that drives the association with the phenotypic class of interest These analyses show and reinforce the hypothesis that even though individual genes may not be altered between groups, there is an imbalanced antioxidant system in CCs derived from follicles that generated embryos with arrested development and from older patients. This also suggests that evaluating oocyte quality using gene signatures such as HAG or a subset of it, such as the core enrichment genes (Fig. 2D and E heatmaps; Supplementary Tables 1 and 2), may be a powerful tool to improve IVF outcomes. In the original work by Feuerstein et al. (GSE27377), using a different approach, the authors found that differential gene expression between B and AD samples also indicate an imbalance in the cell redox homeostasis process [47]. GSE34230, GSE113239, GSE10946, and GSE40400 did not show any statistical differences between analyzed groups. Biochemical analysis In order to identify redox metabolism pattern in human CCs, a large clinical cohort was used to analyze in house the superoxide dismutase, catalase, glutathione peroxidase, glutathione S-transferase activities and reduced glutathione (GSH) levels. CC samples presented on average 1.21 µg (SD ± 0.92) of protein per µL. All enzymatic assays were performed at the same day to avoid multiple freeze and thaw cycles. SOD activity CCs from individual follicles containing oocytes that further generated top-quality blastocysts (so called good quality oocytes) presented significantly less SOD activity than their counterparts from the same patient (P = 0.004) (n = 33 pairs) (Fig. 3A). This significance was confirmed in a non-paired analysis considering all CCs related to oocytes that generated top-quality blastocysts on day 5 after ICSI compared to CCs related to oocytes that presented arrested developed embryos (P = 0.032) (n = 43) (Fig. 3B). To evaluate if this difference could be associated with patient’s clinical characteristics, we analyzed SOD activities in different age groups (Fig. 3C), different infertility diagnoses (Fig. 3D), and different stimulation protocols (Fig. 3E). None of these comparisons evidenced influence of those characteristics in CCs SOD activities, therefore indicating that the correlation between embryo development and SOD activities in CCs is not directly related to patients’ clinical characteristics and could thus be used as a biomarker of embryo development. Although this result seems promising, it is not suitable for defining a clinical threshold for SOD activity as a blastocyst development biomarker, since there are overlaps between the levels of SOD activity in both groups. Therefore, we further stratified the samples, using paired analysis, establishing comparisons between CCs from different follicles of the same patient, within each diagnosis, age, and stimulation group. This approach allows data adjustments for patient- or cycle-related characteristics, eliminating possible personal and procedural-related confounding factors. When comparing B versus AD samples from specific groups, we observed that this increase is not significant in older patients (> 35 years) and in other diagnosis besides male factor patients. Thus, SOD activities might be used as an indicator of blastocyst development potential for younger patients (≤ 35 years) (P = 0.0479) (n = 15 pairs) (Fig. 3F) and for those submitted to IVF by male factor infertility (P = 0.0464) (n = 31 pairs) (Fig. 3G). Table 3 explores the minimum, maximum, mean, and standard deviation values of B and AD samples in each group compared, revealing that for SOD, B samples fluctuate between 21.59 and 276.3 U SOD/µg of protein, while AD samples present values between 36 and 344.5 U SOD/µg of protein in any age range, confirming that higher levels of SOD are related to a worst predictor of embryo quality. It could represent an adaptive response of the COCs with poorer quality oocytes to defend the gamete from higher levels of reactive oxygen species due to the O2∙−-scavenging capacity of SOD. This result is supported by findings that observed higher levels of FF SOD in follicles with lower fertilization rates [74]. The authors concluded that, even though it represents an increase in the antioxidant defense mechanism, if SOD levels are not accompanied by a proper increase in peroxidase activities (CAT and/or GPx), high amounts of H2O2 accumulate and cause direct gamete impairments or can react with ferrous ion producing the high reactive hydroxyl (HO•) radicals through Fenton chemistry [75]. Table 3. | Comparison | Number of pairs | Blastocyst | Arrested development | |||||| |---|---|---|---|---|---|---|---|---|---| | Minimum | Maximum | Mean | Standard. deviation | Minimum | Maximum | Mean | Standard deviation | || | Superoxide dismutase | 33 | 21.59 | 276.3 | 68.76 | 41.48 | 36.00 | 344.5 | 103.9 | 82.63 | | Superoxide dismutase — less than 35 years old | 15 | 21.59 | 276.3 | 67.51 | 59.62 | 36.71 | 344.5 | 126.1 | 114.9 | | Superoxide dismutase — male factor | 5 | 40.50 | 73.96 | 58.37 | 16.47 | 69.90 | 100.8 | 84.80 | 15.52 | | Glutathione S-transferase | 75 | 1.006 | 44.57 | 23.10 | 15.17 | 2.028 | 49.13 | 25.49 | 14.88 | | Glutathione S-transferase — male factor | 33 | 1.006 | 39.12 | 17.73 | 14.68 | 2.619 | 48.73 | 20.60 | 15.94 | In healthy patients with male factor infertility, Matos et al. [76] correlated higher levels of SOD with successful ART outcomes. They also observed increased SOD activity in patients with ovulatory dysfunction and endometriosis and an age-related significant decrease. In our study, when analyzing younger patients (≤ 35 years old) and patients with male factor infertility causes, we observed the contrary: lower levels of SOD activities correlate with top-quality blastocyst formation (Fig. 3F and G). However, both studies used different methodologies (fresh CCs versus cultured CCs and single-follicle CCs versus pooled CCs) and different endpoints (blastocyst formation versus live birth). CAT activity CAT activity changes according to the patient’s age (P = 0.0078) (n = 221) (Fig. 4C). Several studies support the idea that there is a global decrease in follicles’ antioxidant defenses related to maternal age [24, 77–79]. CAT levels are also altered depending on the stimulation protocol applied (P = 0.0223) (n = 206) (Fig. 4E). The stimulation protocol consists of a combination of drugs used to promote the simultaneous maturation of several follicles, allowing further retrieval of mature oocytes suitable for ICSI [80]. While its choice depends on patients’ clinical characteristics and the physician’s experience, most fertility clinics apply several protocols on a daily basis. Still, the vast majority of studies do not take into consideration this heterogeneity, although it is established that different protocols provoke changes in follicle biology [81–83]. Our results reinforce the awareness of considering the influence of stimulation protocols when studying CCs and oocyte biology. When pairing blastocyst formation (B) versus arrested development (AD) samples (n = 81 pairs) from patients of the same age group and same protocol group (Fig. 4A), no significance was observed, showing that CAT levels indeed depend on patient’s age and protocol applied and not on oocyte’s developmental potential (Fig. 4B). GPx activity GPx activity is significantly higher in younger patients (P = 0.0052) (n = 153) (Fig. 5C). GPx levels are also altered depending on the stimulation protocol applied (AGON. vs. ANT + FSH, P = 0.0015; ANT + HMG vs. ANT + FSH, P = 0.0066) (n = 147) (Fig. 5E). Previous studies observed lower glutathione peroxidase activity in FF from women with unexplained causes of infertility when compared to FF from fertile women [28, 84]. No relationships between GPx activity and oocyte quality or embryo developmental potential were observed in our study (Fig. 5A and B). GST activity CCs from patients with polycystic ovarian syndrome (PCOS) presented an extremely low GST activity, compared to patients with endometriosis (P < 0.0001) or male factor (P = 0.0002) as an infertility cause (n = 117) (Fig. 6D). It is known that PCOS influences directly on oocyte quality and CCs redox metabolism [85]. CCs from PCOS patients have previously shown mitochondria dysfunction, imbalanced redox potential, and increased oxidative stress [86]. Our result also matches previous studies showing that these patients present diminished antioxidant potential in serum [85, 87]. On the contrary, in samples of patients diagnosed with endometriosis, we detected a higher level of GST (Fig. 6D). This could be a compensatory mechanism against the chronic inflammatory scenario of ovarian tissue in the presence of endometriosis. CCs from endometriotic patients were previously shown to present higher levels of oxidative stress, negatively correlated with embryo quality [88] and pregnancy rates [22]. Redox metabolism pattern in women with endometriosis is correlated differently with oocyte quality and pregnancy rates [89]. GST activity also varied greatly depending on the stimulation protocol applied. Patients that received agonist protocols have higher GST levels (P = 0.0126) (n = 123) (Fig. 6E). GnRH agonist is administered to promote suppression of pituitary function and prevention of premature ovulation. GSTs comprehend a large family of detoxification enzymes known to catalyze reactions with a wide variety of compounds. These enzymes can interact with several drugs, including oncologic treatment substances, eliminating these compounds from the organism and interfering with the patients’ treatment [90]. When analyzing specifically CCs from healthy patients submitted to IVF for male factor causes, the paired analysis comparing samples from the same patient revealed that CCs related to oocytes that successfully developed into top-quality blastocysts showed slightly lower GST activity levels than oocytes that failed to develop from the same patient (P = 0.0193) (n = 33 pairs) (Fig. 6G). This tendency is not observed in any other group. This parameter could be used carefully as a prediction factor only in this patient subgroup. Table 3 shows that while there is an important overlap between GST activity values in B and AD samples, B samples fluctuate between 1 and 44.57 U GST/µg of protein, with the maximum value of 39.12 for younger women, while AD samples present values between 2.03 and 49.13 U GST/µg of protein in any age range and 2.62 and 48.73 U GST/µg of protein in younger women. Reduced GSH levels Previous studies report that FF from endometriotic patients present significantly lower levels of GSH and that follicles with higher GSH levels correlate with high-quality embryos [33]. The same tendency was not confirmed in CCs in our study. No differences were observed in GSH levels, in any of the comparisons (Fig. 7). The proportion of fertilized oocytes from pooled samples was assessed. Samples representative of 80% or more of fertilized oocytes were considered the fertilized group, while samples with 50% or less fertilized oocytes were considered the non-fertilized group. No significant differences were observed between groups for any of the SOD, CAT, GPx, GST, and GSH assays (data not shown). Samples were also analyzed according to patients’ body mass index (BMI), defined as a person’s weight in kilograms divided by the square of the person’s height in meters (kg/m2), and classified as normal weight, overweight, and obese according to the Worlds Health Organization [91]. None of the analyzed parameters significantly varied between groups (data not shown). Here, we highlight the complex relationship of cause and effect between the CCs redox metabolism, the environment, and the oocytes’ developmental potential, revealing different metabolism patterns in distinct patients’ groups and pointing SOD and GST activity levels as possible oocyte quality biomarkers. While it is already known that the redox state of CCs influences directly on oocyte quality[8], until the present study, little was elucidated on human CC redox metabolism and the contribution of patients’ characteristics and oocyte quality. Here, we state that patients’ clinical characteristics influence directly not only on oocyte quality and embryo developmental potential, but also in CC biological functioning, which may or may not be related to oocyte quality. Thus, it is urgent that potential biomarkers should be tested, validated, and implemented for specific patient subgroups, seeking for a more personalized treatment and efficient outcome. Prospective, large-scale, randomized clinical trials should be designed to evaluate these ideas. Moreover, uncovering the redox dynamics of follicles in different patient groups and how it correlates with oocyte quality may contribute to the development of new approaches to improve oocyte quality in vivo, like the administration of specific supplements for each patient profile [92–98], before oocyte retrieval procedures, and possibly increase success rates after fertilization. These results can also contribute to in vitro studies that might observe different results when administering antioxidants in vitro for distinct patient profiles [5, 99–105]. Our meta-analysis of publicly available microarray data revealed a series of differential expression genes (DEGs) between groups (Table 2). In fact, a handful of previous studies suggest several different genes as oocyte quality biomarkers, but there is great inconsistency between studies [37–42]. In summary, none of the significance we observed individually in the meta-analysis corresponded to the ones observed experimentally, suggesting that single gene expression cannot reliably represent patterns of enzyme activity in CCs. However, when applying GSEA to assess if there is consistent expression behavior of antioxidant genes as a group, we observed that HAG network was upregulated in CCs related to top-quality blastocysts, especially regarding GST isoforms. The same pattern was observed experimentally in CCs from younger patients that are known to present best quality oocytes. Analysis of SOD activity in different subgroups showed that while it can be used as an oocyte quality biomarker, its predictive strength will depend on the patients’ characteristics. Nonetheless, it presented itself as a good complementary tool for oocyte selection for younger patients and the ones submitted to ICSI for male infertility causes. Likewise, GST activity levels in CCs can be used as a complementary tool for oocyte selection only for male factor patients, since its levels are influenced by patients’ infertility causes and stimulation protocol applied. Thus, this study presents, for the first time, a robust analysis of the antioxidant system components in human CCs, reveals important differences between patient groups, highlights differences between redox gene expression levels and enzyme activities, and suggests promissory biomarkers to be used in specific patient subgroups, reinforcing the need for personalized treatments in the clinical environment. These biomarkers can be very useful in oocyte cryopreservation procedures, giving the embryologist and the patient an idea of the quality of the gametes in the moment of preservation. Although this study analyzed 383 human cumulus samples for enzymatic measurement, it was not performed in a randomized controlled matter. Further studies are needed to assess the viability of the proposed markers in the clinical environment. Supplementary Information Below is the link to the electronic supplementary material. Author contribution All authors contributed to the study conception and design. Lucia von Mengden developed the project, collected and processed the samples, performed the experiments, analyzed the data, and wrote the manuscript. Marco Antônio De Bastiani performed the experiments, performed data analyses and experimental design construction, and participated in the discussion and writing of the manuscript. Letícia Arruda participated in the development of the project, carried out the collection of the samples, and participated in the discussion of the results. Carlos Alberto Link participated in the development of the project, provided the clinical data of the patients, supervised the collections, and participated in the discussion of the results. Fábio Klamt funded, developed, and supervised the project, data analyses, discussion, and writing of the manuscript. Funding This study was supported by the Brazilian funds MCTI/CNPq INCT-TM/CAPES/FAPESP (465458/2014–9) and PRONEX/FAPERGS (16/2551–0000499-4). LM received fellowships from PDSE-CAPES 47/2017 (88881.188914/2018–01) and CNPq 143934/2019–8 during the course of this research. FK received a fellowship from MCT/CNPq (306439/2014‐0). Data availability The authors declare that data supporting the findings of this study are available within the paper and its supplementary information files. Codes and intermediary files of the analyses are available on request. Declarations Conflict of Interest The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Lucia von Mengden, Email: [email protected]. Marco Antônio De Bastiani, Email: [email protected]. Leticia Schmidt Arruda, Email: [email protected]. Carlos Alberto Link, Email: [email protected]. Fábio Klamt, Email: [email protected].

- 1.Espey LL. Ovulation as an inflammatory reaction a hypothesis. Biol Reprod. 1980;22:73–106. doi: 10.1095/biolreprod22.1.73. [DOI] [PubMed] [Google Scholar] - 2.Duffy DM, Ko C, Jo M, Brannstrom M, Curry TE. Ovulation: parallels with inflammatory processes. Endocr Rev. 2019;40:369–416. doi: 10.1210/er.2018-00075. [DOI] [PMC free article] [PubMed] [Google Scholar] - 3.Cicinelli E, Ignarro LJ, Lograno M, Galantino P, Balzano G, Schonauer LM. Circulating levels of nitric oxide in fertile women in relation to the menstrual cycle. Fertil Steril. 1996;66:1036–1038. doi: 10.1016/S0015-0282(16)58706-8. [DOI] [PubMed] [Google Scholar] - 4.Miyamoto K, Sato EF, Kasahara E, Jikumaru M, Hiramoto K, Tabata H, et al. Effect of oxidative stress during repeated ovulation on the structure and functions of the ovary, oocytes, and their mitochondria. Free Radic Biol Med. 2010;49:674–681. doi: 10.1016/j.freeradbiomed.2010.05.025. [DOI] [PubMed] [Google Scholar] - 5.Lin X, Dai Y, Tong X, Xu W, Huang Q, Jin X, et al. 2020 Excessive oxidative stress in cumulus granulosa cells induced cell senescence contributes to endometriosis-associated infertility. Redox Biol. 30 101431 [DOI] [PMC free article] [PubMed] - 6.Eichenlaub-Ritter U, Wieczorek M, Lüke S, Seidel T. Age related changes in mitochondrial function and new approaches to study redox regulation in mammalian oocytes in response to age or maturation conditions. Mitochondrion. 2011;11:783–796. doi: 10.1016/j.mito.2010.08.011. [DOI] [PubMed] [Google Scholar] - 7.Kishi I, Ohishi M, Akiba Y, Asada H, Konishi Y, Nakano M, et al. Thioredoxin, an antioxidant redox protein, in ovarian follicles of women undergoing in vitro fertilization. Endocr J. 2016;63:9–20. doi: 10.1507/endocrj.EJ15-0210. [DOI] [PubMed] [Google Scholar] - 8.Von Mengden L, Klamt F, Smitz J. Redox biology of human cumulus cells: basic concepts, impact on oocyte quality, and potential clinical use. Antioxidants Redox Signal. 2020;32:522–535. doi: 10.1089/ars.2019.7984. [DOI] [PMC free article] [PubMed] [Google Scholar] - 9.Ramasarma T. Many faces of superoxide dismutase, originally known as erythrocuprein. Curr Sci. 2007;92:184–191. [Google Scholar] - 10.Matzuk MM, Dionne L, Guo Q, Kumar TR, Lebovitz RM. Ovarian function in superoxide dismutase 1 and 2 knockout mice. Endocrinology. 1998;139:4008–4011. doi: 10.1210/endo.139.9.6289. [DOI] [PubMed] [Google Scholar] - 11.Ho YS, Gargano M, Cao J, Bronson RT, Heimler I, Hutz RJ. Reduced fertility in female mice lacking copper-zinc superoxide dismutase. J Biol Chem. 1998;273:7765–7769. doi: 10.1074/jbc.273.13.7765. [DOI] [PubMed] [Google Scholar] - 12.Chelikani P, Fita I, Loewen PC. Diversity of structures and properties among catalases. Cell Mol Life Sci. 2004;61:192–208. doi: 10.1007/s00018-003-3206-5. [DOI] [PMC free article] [PubMed] [Google Scholar] - 13.Flohe L, Günzler WA, Schock HH. Glutathione peroxidase: a selenoenzyme. FEBS Lett. 1973;32:132–134. doi: 10.1016/0014-5793(73)80755-0. [DOI] [PubMed] [Google Scholar] - 14.Mannervik B. The enzymes of glutathione metabolism: an overview. Biochem Soc Trans. 1987;15:717–718. doi: 10.1042/bst0150717. [DOI] [PubMed] [Google Scholar] - 15.Lu J, Holmgren A. The thioredoxin antioxidant system. Free Radic Biol Med. 2014;66:75–87. doi: 10.1016/j.freeradbiomed.2013.07.036. [DOI] [PubMed] [Google Scholar] - 16.Hayes JD, McLellan LI. Glutathione and glutathione-dependent enzymes represent a co-ordinately regulated defence against oxidative stress. Free Radic Res. 1999;31:273–300. doi: 10.1080/10715769900300851. [DOI] [PubMed] [Google Scholar] - 17.Agarwal A, Gupta S, Sikka S. The role of free radicals and antioxidants in reproduction. Curr Opin Obstet Gynecol. 2006;18:325–332. doi: 10.1097/01.gco.0000193003.58158.4e. [DOI] [PubMed] [Google Scholar] - 18.Gupta S, Malhotra N, Sharma D, Chandra A, Ashok A. Oxidative stress and its role in female infertility and assisted reproduction: clinical implications. Int J Fertil Steril. 2009;2:147–164. [Google Scholar] - 19.Russell DL, Gilchrist RB, Brown HM, Thompson JG. Bidirectional communication between cumulus cells and the oocyte: old hands and new players? Theriogenology. 2016;86:62–68. doi: 10.1016/j.theriogenology.2016.04.019. [DOI] [PubMed] [Google Scholar] - 20.Akin N, von Mengden L, Herta A-C, Billooye K, Leersum J, Cava-Cami B, et al. 2021 Glucose metabolism characterization during mouse in vitro maturation identifies alterations in cumulus cells. Biol Reprod. 1–12. [DOI] [PubMed] - 21.Mizuno S, Ishikawa Y, Matsumoto H, Sato M, Ida M, Fukuda A, et al. The timing of cumulus cell removal for intracytoplasmic sperm injection influences the capability of embryonic development. Reprod Med Biol. 2019;18:111–117. doi: 10.1002/rmb2.12257. [DOI] [PMC free article] [PubMed] [Google Scholar] - 22.Donabela FC, Meola J, Padovan CC, De Paz CCP, Navarro PA. Higher SOD1 gene expression in cumulus cells from infertile women with moderate and severe endometriosis. Reprod Sci. 2015;22:1452–1460. doi: 10.1177/1933719115585146. [DOI] [PubMed] [Google Scholar] - 23.Paszkowski T, Traub AI, Robinson SY, McMaster D. Selenium dependent glutathione peroxidase activity in human follicular fluid. Clin Chim Acta. 1995;236:173–180. doi: 10.1016/0009-8981(95)98130-9. [DOI] [PubMed] [Google Scholar] - 24.Ito M, Imai M, Muraki M, Miyado K, Qin J, Kyuwa S, et al. GSTT1 is upregulated by oxidative stress through p38-mk2 signaling pathway in human granulosa cells: possible association with mitochondrial activity. Aging (Albany NY) 2011;3:1213–1223. doi: 10.18632/aging.100418. [DOI] [PMC free article] [PubMed] [Google Scholar] - 25.Sinkó I, Mórocz M, Zádori J, Kokavszky K, Raskó I. Effect of cigarette smoking on DNA damage of human cumulus cells analyzed by comet assay. Reprod Toxicol. 2005;20:65–71. doi: 10.1016/j.reprotox.2004.12.007. [DOI] [PubMed] [Google Scholar] - 26.Younglai EV, Holloway AC, Foster WG. Environmental and occupational factors affecting fertility and IVF success. Hum Reprod Update. 2005;11:43–57. doi: 10.1093/humupd/dmh055. [DOI] [PubMed] [Google Scholar] - 27.Oyawoye O, Gadir AA, Garner A, Constantinovici N, Perrett C, Hardiman P. Antioxidants and reactive oxygen species in follicular fluid of women undergoing IVF: Relationship to outcome. Hum Reprod. 2003;18:2270–2274. doi: 10.1093/humrep/deg450. [DOI] [PubMed] [Google Scholar] - 28.Paszkowski T, Traub AI, Robinson SY, McMaster D. Selenium dependent glutathione peroxidase activity in human follicular fluid. Clin Chim Acta. 1995;236:173–80. doi: 10.1016/0009-8981(95)98130-9. [DOI] [PubMed] [Google Scholar] - 29.Pasqualotto EB, Agarwal A, Ph D. Effect of oxidative stress in follicular fluid on the outcome of assisted reproductive procedures. Fertil Steril. 2004;81:973–976. doi: 10.1016/j.fertnstert.2003.11.021. [DOI] [PubMed] [Google Scholar] - 30.Das S, Chattopadhyay R, Ghosh S, Ghosh S, Goswami SK, Chakravarty BN, et al. Reactive oxygen species level in follicular fluid - embryo quality marker in IVF? Hum Reprod. 2006;21:2403–2407. doi: 10.1093/humrep/del156. [DOI] [PubMed] [Google Scholar] - 31.Revelli A, Canosa S, Bergandi L, Skorokhod OA, Biasoni V, Carosso A, et al. Oocyte polarized light microscopy, assay of specific follicular fluid metabolites, and gene expression in cumulus cells as different approaches to predict fertilization efficiency after ICSI. Reprod Biol Endocrinol. 2017;15:1–12. doi: 10.1186/s12958-017-0265-2. [DOI] [PMC free article] [PubMed] [Google Scholar] - 32.Kishi I, Ohishi M, Akiba Y, Asada H, Konishi Y, Nakano M, et al. Thioredoxin, an antioxidant redox protein, in ovarian follicles of women undergoing in vitro fertilization. Endocr J. 2015;63:9–20. doi: 10.1507/endocrj.EJ15-0210. [DOI] [PubMed] [Google Scholar] - 33.Choi YS, Cho SH, Seo SK, Park JH, Kim SH, Lee BS. Alteration in the intrafollicular thiol-redox system in infertile women with endometriosis. Reproduction. 2015;149:155–162. doi: 10.1530/REP-14-0438. [DOI] [PubMed] [Google Scholar] - 34.Sabatini L, Wilson C, Sc B, Lower A, Al-shawaf T, Grudzinskas JG. Superoxide dismutase activity in human follicular fluid after controlled ovarian hyperstimulation in women undergoing in vitro fertilization. Fertil Steril. 1999;72:1027–1034. doi: 10.1016/S0015-0282(99)00411-2. [DOI] [PubMed] [Google Scholar] - 35.Rodgers RJ, Irving-Rodgers HF. Formation of the ovarian follicular antrum and follicular fluid. Biol Reprod. 2010;82:1021–1029. doi: 10.1095/biolreprod.109.082941. [DOI] [PubMed] [Google Scholar] - 36.Shalgi R, Kraicer P, Rimon A, Pinto M, Soferman N. Proteins of human follicular fluid: the blood follicle barrier. Fertil Steril. 1973;24:429–434. doi: 10.1016/S0015-0282(16)39730-8. [DOI] [PubMed] [Google Scholar] - 37.Adriaenssens T, Segers I, Wathlet S, Smitz J. The cumulus cell gene expression profile of oocytes with different nuclear maturity and potential for blastocyst formation. J Assist Reprod Genet. 2011;28:31–40. doi: 10.1007/s10815-010-9481-9. [DOI] [PMC free article] [PubMed] [Google Scholar] - 38.Adriaenssens T, Wathlet S, Segers I, Verheyen G, De VA, Van Der EJ, et al. Cumulus cell gene expression is associated with oocyte developmental quality and influenced by patient and treatment characteristics. Hum Reprod. 2010;25:1259–1270. doi: 10.1093/humrep/deq049. [DOI] [PubMed] [Google Scholar] - 39.Artini PG, Tatone C, Sperduti S, D’Aurora M, Franchi S, Di Emidio G, et al. Cumulus cells surrounding oocytes with high developmental competence exhibit down-regulation of phosphoinositol 1,3 kinase/protein kinase B (PI3K/AKT) signalling genes involved in proliferation and survival. Hum Reprod. 2017;32:2474–2484. doi: 10.1093/humrep/dex320. [DOI] [PMC free article] [PubMed] [Google Scholar] - 40.Anderson RA, Sciorio R, Kinnell H, Bayne RAL, Thong KJ, De Sousa PA, et al. Cumulus gene expression as a predictor of human oocyte fertilisation, embryo development and competence to establish a pregnancy. Reproduction. 2009;138:629–637. doi: 10.1530/REP-09-0144. [DOI] [PubMed] [Google Scholar] - 41.Papler TB, Bokal EV, Lovrecic L, Kopitar AN, Maver A. No specific gene expression signature in human granulosa and cumulus cells for prediction of oocyte fertilisation and embryo implantation. PLoS ONE. 2015;10:1–13. doi: 10.1371/journal.pone.0115865. [DOI] [PMC free article] [PubMed] [Google Scholar] - 42.Green KA, Franasiak JM, Werner MD, Tao X, Landis JN, Scott RT, et al. Cumulus cell transcriptome profiling is not predictive of live birth after in vitro fertilization: a paired analysis of euploid sibling blastocysts. Fertil Steril. 2018;109:460–466. doi: 10.1016/j.fertnstert.2017.11.002. [DOI] [PubMed] [Google Scholar] - 43.van Montfoort APA, Geraedts JPM, Dumoulin JCM, Stassen APM, Evers JLH, Ayoubi TAY. Differential gene expression in cumulus cells as a prognostic indicator of embryo viability: a microarray analysis. Mol Hum Reprod. 2008;14:157–168. doi: 10.1093/molehr/gam088. [DOI] [PubMed] [Google Scholar] - 44.Bergandi L, Basso G, Evangelista F, Canosa S, Dalmasso P, Aldieri E, et al. Inducible nitric oxide synthase and heme oxygenase 1 are expressed in human cumulus cells and may be used as biomarkers of oocyte competence. Reprod Sci. 2014;21:1370–1377. doi: 10.1177/1933719114525268. [DOI] [PubMed] [Google Scholar] - 45.Matos L, Stevenson D, Gomes F, Almeida H. Superoxide dismutase expression in human cumulus oophorus cells. Mol Hum Reprod. 2009;15:411–419. doi: 10.1093/molehr/gap034. [DOI] [PubMed] [Google Scholar] - 46.Gelain DP, Dalmolin RJS, Belau VL, Moreira JCF, Klamt F, Castro MAA. A systematic review of human antioxidant genes. Front Biosci. 2009;14:4457–4464. doi: 10.2741/3541. [DOI] [PubMed] [Google Scholar] - 47.Feuerstein P, Puard V, Chevalier C, Teusan R, Cadoret V, Guerif F, et al. 2012 Genomic assessment of human cumulus cell marker genes as predictors of oocyte developmental competence: impact of various experimental factors. PLoS One. 7. [DOI] [PMC free article] [PubMed] - 48.Molinari E, Bar H, Pyle AM, Patrizio P. Transcriptome analysis of human cumulus cells reveals hypoxia as the main determinant of follicular senescence. Mol Hum Reprod. 2016;22:566–576. doi: 10.1093/molehr/gaw038. [DOI] [PMC free article] [PubMed] [Google Scholar] - 49.Devjak R, FonTacer K, Juvan P, VirantKlun I, Rozman D, BokalVrtačnik E. Cumulus cells gene expression profiling in terms of oocyte maturity in controlled ovarian hyperstimulation using GnRH agonist or GnRH antagonist. PLoS One. 2012;7:e47106. doi: 10.1371/journal.pone.0047106. [DOI] [PMC free article] [PubMed] [Google Scholar] - 50.Li J, Chen H, Gou M, Tian C, Wang H, Song X, et al. Molecular features of polycystic ovary syndrome revealed by transcriptome analysis of oocytes and cumulus cells. Front Cell Dev Biol. 2021;9:1–16. doi: 10.3389/fcell.2021.735684. [DOI] [PMC free article] [PubMed] [Google Scholar] - 51.Kenigsberg S, Bentov Y, Chalifa-caspi V, Potashnik G, Ofir R, Birk OS. Gene expression microarray profiles of cumulus cells in lean and overweight-obese polycystic ovary syndrome patients. Mol Hum Reprod. 2009;15:89–103. doi: 10.1093/molehr/gan082. [DOI] [PubMed] [Google Scholar] - 52.Huang X, Hao C, Shen X, Liu X, Shan Y, Zhang Y, et al. Differences in the transcriptional profiles of human cumulus cells isolated from MI and MII oocytes of patients with polycystic ovary syndrome. Reproduction. 2013;145:597–608. doi: 10.1530/REP-13-0005. [DOI] [PubMed] [Google Scholar] - 53.Ouandaogo ZG, Haouzi D, Assou S, Dechaud H, Kadoch IJ, de Vos J, et al. Human cumulus cells molecular signature in relation to oocyte nuclear maturity stage. PLoS One. 2011;6:e27179. doi: 10.1371/journal.pone.0027179. [DOI] [PMC free article] [PubMed] [Google Scholar] - 54.R Core Team. R: a language and environment for statistical computing. Vienna: R Found Stat Comput; 2014. Available at: http://www.R-project.org/. - 55.Gentleman AR, Carey V, Huber W, Hahne F, Bioconductor M, Maintainer P. Genefilter: methods for filtering genes from high-throughput experiments. Version 1.63.0. 2018. - 56.Heider A, Alt R. VirtualArray: a R/bioconductor package to merge raw data from different microarray platforms. BMC Bioinformatics. 2013;14:75. doi: 10.1186/1471-2105-14-75. [DOI] [PMC free article] [PubMed] [Google Scholar] - 57.Huber W, Von Heydebreck A, Sültmann H, Poustka A, Vingron M. Variance stabilization applied to microarray data calibration and to the quantification of differential expression. Bioinformatics. 2002;18(SUPPL.1): S96–S104. [DOI] [PubMed] - 58.Ritchie ME, Phipson B, Wu D, Hu Y, Law CW, Shi W, et al. Limma powers differential expression analyses for RNA-sequencing and microarray studies. Nucleic Acids Res. 2015;43(7):e47. doi: 10.1093/nar/gkv007. [DOI] [PMC free article] [PubMed] [Google Scholar] - 59.Patro R, Duggal G, Love MI, Irizarry RA, Kingsford C. Salmon provides fast and bias-aware quantification of transcript expression. Nat Methods. 2017;14:417–419. doi: 10.1038/nmeth.4197. [DOI] [PMC free article] [PubMed] [Google Scholar] - 60.Soneson C, Love MI, Robinson MD 2015 Differential analyses for RNA-seq: transcript-level estimates improve gene-level inferences. F1000Research 4:1521 [DOI] [PMC free article] [PubMed] - 61.Walter W, Sánchez-Cabo F, Ricote M. GOplot: An R package for visually combining expression data with functional analysis. Bioinformatics. 2015;31:2912–2914. doi: 10.1093/bioinformatics/btv300. [DOI] [PubMed] [Google Scholar] - 62.Subramanian A, Tamayo P, Mootha V. 2014 GSEA : Gene set enrichment analysis Gene set enrichment analysis: a knowledge-based approach for interpreting genome-wide expression profiles. PNAS. 102. [DOI] [PMC free article] [PubMed] - 63.Matsuura K, Hayashi N, Takiue C, Hirata R, Habara T, Naruse K. Blastocyst quality scoring based on morphologic grading correlates with cell number. Fertil Steril. 2010;94:1135–1137. doi: 10.1016/j.fertnstert.2009.11.003. [DOI] [PubMed] [Google Scholar] - 64.Gardner DK, Lane M, Stevens J, Schlenker T, Schoolcraft WB. Blastocyst score affects implantation and pregnancy outcome: towards a single blastocyst transfer. Fertil Steril. 2000;73:1155–1158. doi: 10.1016/S0015-0282(00)00518-5. [DOI] [PubMed] [Google Scholar] - 65.Gardner DK, Schoolcraft WB. Culture and transfer of human blastocysts. Curr Opin Obstet Gynecol. 1999;11:307–311. doi: 10.1097/00001703-199906000-00013. [DOI] [PubMed] [Google Scholar] - 66.Zor T, Selinger Z. Linearization of the Bradford protein assay increases its sensitivity: theoretical and experimental studies. AnalBiochem. 1996;236:302–308. doi: 10.1006/abio.1996.0171. [DOI] [PubMed] [Google Scholar] - 67.Misra HP, Fridovich I. The role of superoxide anion in the autoxidation of epinephrine and a simple assay for superoxide dismutase. J Biol Chem. 1972;247:3170–3175. doi: 10.1016/S0021-9258(19)45228-9. [DOI] [PubMed] [Google Scholar] - 68.Aebi H. Catalase in vitro. Methods Enzymol. 1984;105:121–126. doi: 10.1016/S0076-6879(84)05016-3. [DOI] [PubMed] [Google Scholar] - 69.Günzler WA, Kremers H, Flohé L. An improved coupled test procedure for glutathione peroxidase (EC 1.11.1.9.) in blood. Clin Chem Lab Med. 1974;12:444–8. doi: 10.1515/cclm.1974.12.10.444. [DOI] [PubMed] [Google Scholar] - 70.Lawrence RA, Burk RF. Glutathione peroxidase activity in selenium-deficient rat liver. Biochem Biophys Res Commun. 1976;71:952–958. doi: 10.1016/0006-291X(76)90747-6. [DOI] [PubMed] [Google Scholar] - 71.Habig WH, Pabst MJ, Jakoby WB. Glutathione S transferases. The first enzymatic step in mercapturic acid formation. J Biol Chem. 1974;249:7130–9. doi: 10.1016/S0021-9258(19)42083-8. [DOI] [PubMed] [Google Scholar] - 72.Browne RW, Armstrong D. Reduced glutathione and glutathione disulfide. Methods Mol Biol. 1998;108:347–352. doi: 10.1385/0-89603-472-0:347. [DOI] [PubMed] [Google Scholar] - 73.Gerris J. Single-embryo transfer versus multiple-embryo transfer. Reprod Biomed Online. 2009;18:S63–70. doi: 10.1016/S1472-6483(10)60451-8. [DOI] [PubMed] [Google Scholar] - 74.Sabatini L, Wilson C, Lower A, Al-Shawaf T, Grudzinskas JG. Superoxide dismutase activity in human follicular fluid after controlled ovarian hyperstimulation in women undergoing in vitro fertilization. Fertil Steril. 1999;72:1027–1034. doi: 10.1016/S0015-0282(99)00411-2. [DOI] [PubMed] [Google Scholar] - 75.Prousek J. Fenton chemistry in biology and medicine. Pure Appl Chem. 2007;79:2325–2338. doi: 10.1351/pac200779122325. [DOI] [Google Scholar] - 76.Matos L, Stevenson D, Gomes F, Silva-carvalho JL, Almeida H. Superoxide dismutase expression in human cumulus oophorus cells. Mol Hum Reprod. 2009;15:411–419. doi: 10.1093/molehr/gap034. [DOI] [PubMed] [Google Scholar] - 77.Cree LM, Hammond ER, Shelling AN, Berg MC, Peek JC, Green MP. Maternal age and ovarian stimulation independently affect oocyte mtDNA copy number and cumulus cell gene expression in bovine clones. Hum Reprod. 2015;30:1410–1420. doi: 10.1093/humrep/dev066. [DOI] [PubMed] [Google Scholar] - 78.Carbone MC, Tatone C, Monache SD, Marci R, Caserta D, Colonna R, et al. Antioxidant enzymatic defences in human follicular fluid: characterization and age-dependent changes. Mol Hum Reprod. 2003;9:639–643. doi: 10.1093/molehr/gag090. [DOI] [PubMed] [Google Scholar] - 79.Lim J, Luderer U. Oxidative damage increases and antioxidant gene expression decreases with aging in the mouse ovary. Biol Reprod. 2011;84:775–782. doi: 10.1095/biolreprod.110.088583. [DOI] [PMC free article] [PubMed] [Google Scholar] - 80.Gallos ID, Eapen A, Price MJ, Sunkara SK, Macklon NS, Bhattacharya S, et al. 2017 Controlled ovarian stimulation protocols for assisted reproduction: a network meta-analysis. Cochrane Database Syst Rev. 3. - 81.Tulić L, Vidaković S, Tulić I, Ćurčić M, Stojnić J, Jeremić K. Oxidative stress markers in GnRH agonist and antagonist protocols in IVF. J Med Biochem. 2017;36:163–170. doi: 10.1515/jomb-2017-0001. [DOI] [PMC free article] [PubMed] [Google Scholar] - 82.Shrestha D, La X, Feng HL. Comparison of different stimulation protocols used in in vitro fertilization: a review. Ann Transl Med. 2015;3:1–7. doi: 10.3978/j.issn.2305-5839.2015.04.09. [DOI] [PMC free article] [PubMed] [Google Scholar] - 83.Assou S, Haouzi D, Dechaud H, Gala A, Ferrières A, Hamamah S. 2013 Comparative gene expression profiling in human cumulus cells according to ovarian gonadotropin treatments. Biomed Res Int. [DOI] [PMC free article] [PubMed] - 84.Carbone MC, Tatone C, Delle Monache S, Marci R, Caserta D, Colonna R, et al. Antioxidant enzymatic defences in human follicular fluid: characterization and age-dependent changes. Mol Hum Reprod. 2003;9:639–643. doi: 10.1093/molehr/gag090. [DOI] [PubMed] [Google Scholar] - 85.Fenkci V, Fenkci S, Yilmazer M, Serteser M. Decreased total antioxidant status and increased oxidative stress in women with polycystic ovary syndrome may contribute to the risk of cardiovascular disease. Fertil Steril. 2003;80:123–127. doi: 10.1016/S0015-0282(03)00571-5. [DOI] [PubMed] [Google Scholar] - 86.Zhao H, Zhao Y, Li T, Li M, Li J, Li R, et al. Metabolism alteration in follicular niche: the nexus among intermediary metabolism, mitochondrial function, and classic polycystic ovary syndrome. Free Radic Biol Med. 2015;86:295–307. doi: 10.1016/j.freeradbiomed.2015.05.013. [DOI] [PubMed] [Google Scholar] - 87.Dincer Y, Akcay T, Erdem T, Ilker Saygili E, Gundogdu S. DNA damage, DNA susceptibility to oxidation and glutathione level in women with polycystic ovary syndrome. Scand J Clin Lab Invest. 2005;65:721–728. doi: 10.1080/00365510500375263. [DOI] [PubMed] [Google Scholar] - 88.Seino T, Saito H, Kaneko T, Takahashi T, Kawachiya S, Kurachi H. Eight-hydroxy-2′-deoxyguanosine in granulosa cells is correlated with the quality of oocytes and embryos in an in vitro fertilization-embryo transfer program. Fertil Steril. 2002;77:1184–1190. doi: 10.1016/S0015-0282(02)03103-5. [DOI] [PubMed] [Google Scholar] - 89.Bedaiwy MA, Elnashar SA, Goldberg JM, Sharma R, Mascha EJ, Arrigain S, et al. Effect of follicular fluid oxidative stress parameters on intracytoplasmic sperm injection outcome. Gynecol Endocrinol. 2012;28:51–55. doi: 10.3109/09513590.2011.579652. [DOI] [PubMed] [Google Scholar] - 90.Townsend DM, Tew KD. The role of glutathione-S-transferase in anti-cancer drug resistance. Oncogene. 2003;22:7369–7375. doi: 10.1038/sj.onc.1206940. [DOI] [PMC free article] [PubMed] [Google Scholar] - 91.Physical status: the use and interpretation of anthropometry. Report of a WHO Expert Committee. World Health Organ Tech Rep Ser. 1995; [PubMed] - 92.Griesinger G, Franke K, Kinast C, Kutzelnigg A, Riedinger S, Kulin S, et al. Ascorbic acid supplement during luteal phase in IVF. J Assist Reprod Genet. 2002;19:164–168. doi: 10.1023/A:1014837811353. [DOI] [PMC free article] [PubMed] [Google Scholar] - 93.Eryilmaz OG, Devran A, Sarikaya E, Aksakal FN, Mollamahmutoğlu L, Cicek N. Melatonin improves the oocyte and the embryo in IVF patients with sleep disturbances, but does not improve the sleeping problems. J Assist Reprod Genet. 2011;28:815–820. doi: 10.1007/s10815-011-9604-y. [DOI] [PMC free article] [PubMed] [Google Scholar] - 94.Scutiero G, Iannone P, Bernardi G, Bonaccorsi G, Spadaro S, Volta CA, et al. Oxidative stress and endometriosis: a systematic review of the literature. Oxid Med Cell Longev. 2017;2017:1–7. doi: 10.1155/2017/7265238. [DOI] [PMC free article] [PubMed] [Google Scholar] - 95.Verit FF, Erel O, Kocyigit A. Association of increased total antioxidant capacity and anovulation in nonobese infertile patients with clomiphene citrate-resistant polycystic ovary syndrome. Fertil Steril. 2007;88:418–424. doi: 10.1016/j.fertnstert.2006.11.172. [DOI] [PubMed] [Google Scholar] - 96.Richards JS. 2005 Ovulation: new factors that prepare the oocyte for fertilization. Mol Cell Endocrinol. 75–9. [DOI] [PubMed] - 97.Elgindy EA, El-Huseiny AM, Mostafa I, Gaballah AM, Ahmed TA. N-acetyl cysteine: could it be an effective adjuvant therapy in ICSI cycles? A preliminary study. Reprod Biomed Online Reprod Healthcare Ltd. 2010;20:789–796. doi: 10.1016/j.rbmo.2010.03.001. [DOI] [PubMed] [Google Scholar] - 98.Hosseini FS, Shamsipour M, Yazdekhasti H, Akbari-Asbagh F, Shahraki Z, Aghaee-Bakhtiari SH. 2021 The effect of oral melatonin supplementation on MT-ATP6 gene expression and IVF outcomes in Iranian infertile couples: a nonrandomized controlled trial. Naunyn. Schmiedebergs. Arch. Pharmacol. 1487–95. [DOI] [PubMed] - 99.Tan DX, Manchester LC, Hardeland R, Lopez-Burillo S, Mayo JC, Sainz RM, et al. Melatonin: a hormone, a tissue factor, an autocoid, a paracoid, and an antioxidant vitamin. J Pineal Res. 2003;34:75–78. doi: 10.1034/j.1600-079X.2003.02111.x. [DOI] [PubMed] [Google Scholar] - 100.Kim MK, Park EA, Kim HJ, Choi WY, Cho JH, Lee WS, et al. Does supplementation of in-vitro culture medium with melatonin improve IVF outcome in PCOS. Reprod Biomed Online Reprod Healthcare Ltd. 2013;26:22–29. doi: 10.1016/j.rbmo.2012.10.007. [DOI] [PubMed] [Google Scholar] - 101.Brenjian S, Moini A, Yamini N, Kashani L, Faridmojtahedi M, Bahramrezaie M, et al. Resveratrol treatment in patients with polycystic ovary syndrome decreased pro-inflammatory and endoplasmic reticulum stress markers. Am J Reprod Immunol. 2020;83:1–10. doi: 10.1111/aji.13186. [DOI] [PubMed] [Google Scholar] - 102.Barrozo LG, Paulino LRFM, Silva BR, Barbalho EC, Nascimento DR, Neto MFL, et al. 2021 N-acetyl-cysteine and the control of oxidative stress during in vitro ovarian follicle growth, oocyte maturation, embryo development and cryopreservation. Anim Reprod Sci. 231. [DOI] [PubMed] - 103.Khajeh M, Nouri M, Ghasemzadeh A, Mehdizadeh A, Shanehbandi D, Yousefi S, et al. Arachidonic acid alleviates the detrimental effects of acetylsalicylic acid on human granulosa cells performance in vitro. Mol Reprod Dev. 2020;87:607–619. doi: 10.1002/mrd.23343. [DOI] [PubMed] [Google Scholar] - 104.Batnasan E, Xie S, Zhang Q, Li Y. Observation of parthanatos involvement in diminished ovarian reserve patients and melatonin’s protective function through inhibiting ADP-ribose (PAR) expression and preventing AIF translocation into the nucleus. Reprod Sci. 2020;27:75–86. doi: 10.1007/s43032-019-00005-8. [DOI] [PubMed] [Google Scholar] - 105.Yu K, Wang RX, Li MH, Sun TC, Zhou YW, Li YY, et al. 2019 Melatonin reduces androgen production and upregulates heme oxygenase-1 expression in granulosa cells from PCOS patients with hypoestrogenia and hyperandrogenia. Oxid Med Cell Longev. 2019. [DOI] [PMC free article] [PubMed] - 106.Molinari E, Patrizio P, Pyle AM, Bar H. Transcriptome analysis of human cumulus cells reveals hypoxia as the main determinant of follicular senescence. Mol Hum Reprod. 2016;22:566–576. doi: 10.1093/molehr/gaw038. [DOI] [PMC free article] [PubMed] [Google Scholar] Associated Data This section collects any data citations, data availability statements, or supplementary materials included in this article. Supplementary Materials Data Availability Statement The authors declare that data supporting the findings of this study are available within the paper and its supplementary information files. Codes and intermediary files of the analyses are available on request.