Hormonal and molecular characterization of follicular ﬂuid, cumulus cells and oocytes from pre-ovulatory follicles in stimulated and unstimulated cycles

background: The use of ovarian stimulation, to stimulate a multi-follicular response for assisted reproduction treatments, may force the production of oocytes from follicles that do not reach optimal maturation, possibly yielding oocytes that are not fully competent. The present study aimed to deﬁne the follicular environment and oocyte competence of unstimulated pre-ovulatory follicles, to compare it with that of similar-sized stimulated follicles. For this purpose, we analyzed the follicular hormonal milieu, the oocyte meiotic spindle, the embryo development and the cumulus cells gene expression (GE) proﬁles. methods and results: The study population was divided in two groups: (i) 42 oocyte donors undergoing unstimulated cycles and (ii) 18 oocyte donors undergoing controlled ovarian stimulation cycles (COS). Follicular ﬂuid was analyzed to quantify the concentrations of estradiol (E2), progesterone (P), FSH, LH, testosterone (T) and androstendione ( D 4). T was higher in the COS group, while D 4, E2 and LH were signiﬁcantly higher in unstimulated cycles. The cumulus oophorus cells (CC) surrounding the oocyte were removed and their GE proﬁles were analyzed with microarrays. There were 18 differentially expressed genes in CC: 7 were up-regulated and 11 were down-regulated in the COS cycles. The microarray was validated by qRT–PCR. The analysis of spindle structure revealed no signiﬁcant differences between the groups, except for the parameter of length which presented differences. The fertilization ability and embryo morphology on Days 2, 3 and 4 did not show any signiﬁcant differences between groups. conclusions: The use of ovarian stimulation induces changes in the follicular ﬂuid and in CC GE that may affect immune processes, meiosis and ovulation pathways. Although these differences do not seem to relate to early-stage embryo morphology, the implications of some of the molecules, especially ALDH1A2, CTSL and ZNF33B at the CC level, deserve to be addressed in future studies.


Introduction
Since the first assisted reproductive treatment (ART) was successfully performed in 1978 by Steptoe and Edwards, its use has exponentially increased and currently accounts for 1-3% of annual births (Santos et al., 2010). The first IVF was performed in a spontaneous natural cycle. However, the use of natural cycles was very soon superseded by ovarian stimulation protocols (Pelinck et al., 2002) to optimize the results of IVF treatments because more follicles are obtained per cycle so that embryo selection for transfer becomes possible (Templeton and Morris, 1998).
The effects of ovarian stimulation on embryo quality have been well characterized in mice, showing that very aggressive ovarian stimulation may reduce early embryo development potential and enhance the frequency of chromosomal abnormalities (Chang, 1977;Ertzeid and Storeng, 2001; Van der Auwera and D'Hooghe, 2001). In humans, no differences in terms of morphology have been found in embryos deriving from natural or stimulated cycles, although it has been suggested that gonadotrophins may increase the rate of embryo aneuploidies in a dose-dependent manner (Munne et al., 1997;Baart et al., 2007). Nevertheless, human embryo aneuploidies in IVF cycles are present even in the absence of ovarian stimulation (Verpoest et al., 2008).
In recent years, the concept of achieving as many oocytes as possible has been modified by an attempt to obtain a cohort of good quality embryos (Macklon et al., 2006). For this reason, there is a trend toward mild stimulation protocols in an attempt to minimize pharmacological intrusion of ovarian physiology during follicular recruitment and follicular growth (Macklon and Fauser, 2003;Macklon et al., 2006;Verberg et al., 2009a,b). In this sense, natural cycles for IVF may offer the best follicular environment, along with the additional advantages of having a close-to-zero multiple pregnancy rate and no risk of hyperstimulation syndrome. Natural cycles are also more user-friendly for patients and cheaper than stimulated cycles (Pelinck et al., 2002).
In comparison with the natural cycle, the use of ovarian stimulation in IVF forces the production of oocytes from follicles that do not reach optimal maturation, therefore, possibly yielding oocytes that are not fully competent (Pellicer et al., 1987Sutton et al., 2003;Nogueira et al., 2006). In light of this, it is known that the mere presence of a polar body is insufficient to evaluate oocyte competence, and that other sub-cellular and molecular features need to be finely tuned to permit fertilization and adequate embryo development (Mann et al., 2010;Alvarez Sedo et al., 2011;Evans and Robinson, 2011). Nevertheless, a spindle analysis is very useful for evaluating the physiological progression of meiosis; in fact its presence and structure reveal a correlation with fertilization following ICSI and with development to the blastocyst stage (Montag et al., 2006;Rama Raju et al., 2007).
Some attempts to measure follicular milieu in stimulated and natural cycles have revealed that although gonadotrophins might regulate ovarian secretion of certain cytokines, such as IL-1 beta, IL-6 and TNF-alpha as well as active rennin, concentrations of the steroid hormones, estradiol (E2) and progesterone (P), do not seem to present significant changes (Loret de Mola et al., 1998. No simultaneous association have been found with oocyte and embryo quality. Given this background, the present study aimed to define follicular development and oocyte competence in the natural cycle to compare it with that of stimulated cycles. For this purpose, we analyzed the follicular hormonal milieu, the meiotic spindle and the cumulus cells gene expression (GE) profiles of the associated oocytes developed in either unstimulated or stimulated cycles.

Study population
The present study was approved by the Ethics Committee at IVI Valencia (E Spain). After signing an informed consent form, a total of 60 oocyte donors were included in the present study. The study population was divided into two groups according to the type of protocol received: (i) 42 oocyte donors undergoing unstimulated cycles (also participating in the clinical trial whose registration number is NCT00707525) and (ii) 18 oocytes donors undergoing controlled ovarian stimulated (COS) cycles.

Patient inclusion criteria
The selected oocyte donors had a normal BMI of 19.3 -28.9 kg/m 2 and the same type of natural or ovarian stimulation protocol which is explained below. Regarding age, all were under 35 year of age as stated by the Assisted Reproduction Spanish Law 14/2006.

Stimulation protocols
In the unstimulated cycle, there was no medical intervention except for 250 mg of recombinant choriogonadotrophin (rCG, Ovitrellew, Merck-Serono, Geneva, Switzerland) administered as soon as the spontaneous pre-ovulatory follicle reached 18 mm in diameter. Stimulated cycles were carried out by following the GnRH agonist long protocol with daily doses of 225 IU of recombinant rFSH w (Merck Serono, Geneva, Switzerland). When more than three follicles were .18 mm in diameter, rCG was administered. Doses were adjusted according to ovarian response as judged by ultrasound and by serum estradiol concentrations.

Oocytes and follicular fluid collection
Oocyte pick-up was performed 36 h after injecting rCG by transvaginal ultrasound. In the unstimulated cycles, the unique pre-ovulatory follicle was aspirated, while only one of the good-sized (.18 mm in diameter) dominant pre-ovulatory follicles was aspirated individually in the COS cycles for this study. Follicular fluid was centrifuged for 10 min at 2000g, and the supernatant was aliquoted and immediately frozen by immersion in liquid nitrogen to be stored at 2808C until the time of the biochemical analysis.
Between 2 and 4 h after follicular aspiration, the cumulus oophorus cells (CCs) surrounding the oocyte were removed by cutting needles to be later placed in 200 ml of TRizol and stored at 2808C until RNA extraction.

Determination of the hormonal concentration in follicular fluid
Follicular fluids of pre-ovulatory follicles of the oocyte donors undergoing COS and unstimulated cycles were analyzed to quantify the hormonal concentration of E2, P, FSH, LH and testosterone (T) with the Abbot AXSYM System. This method consists of commercial enzyme immunoassay kits of microparticles, 'MEIA'. The high concentration of E2 and P in the follicular fluid necessitated dilution of the follicular samples with specific diluents. Androstendione (D4) was analyzed by radioimmunoanalysis (RIA) (Bio-Source Androstendione-RIA-CT-KIT). All the measurements were taken following the manufacturer's instructions. Inter-and intra-assay coefficients of variation for measurements E2, P, T, D4, FSH and LH were 8.8 and 4.6%, 7.9 and 4.5%, 8.4 and 4.4%, 6.0 and 3.3%, 6.9 and 4.1% and 5.9 and 3.8%, respectively.

Oocyte quality assessment by spindle evaluation
The oocyte spindle analysis was performed by means of polarized light CRi Oosight TM optics (Quermed, Madrid, Spain). The following parameters were analyzed: meiosis stage, and the spindle presence, position, retardance, area, perimeter, length, width and overall structure in both groups. A spindle was considered as good if it was barrel shaped, well positioned and showed strong birefringence.

Characterization of unstimulated cycles
Fertilization and embryo quality assessment All the oocytes obtained from both the unstimulated and COS cycles were microinjected by ICSI, and fertilization ability was assessed 24 h after ICSI completion. Normal fertilization was considered as when oocytes had two pronuclei and two polar bodies. Embryo morphology was evaluated on Days 2, 3 and 4 by taking into account the number, symmetry and granularity of blastomeres, the type and percentage of fragmentation, the presence of multinucleated blastomeres and the degree of compaction (Alikani et al., 1999).

RNA extraction and microarray hybridization
Microarrays analyses were done in a total of 8 CC, 4 from women undergoing unstimulated and 4 from women undergoing COS cycles. Since the CC GE can be affected by patient characteristics (as well as by gonadotrophin preparations, and even the response to them; Haouzi et al., 2009), we matched the characteristics in our study population.
The total RNA from individual CC was extracted using the TRIzolw protocol (Invitrogen, Barcelona, Spain) with a minimal volume. The total amount and quality of RNA was first tested in a Nanodrop ND-1000 UV-Vis spectrophotometer and then with a more sensitive method by RNA 6000 PicoLab Chip in the Bioanalyzer 2100. Only the RNA that had a series of characteristics was used for further analysis: a 260/280 ratio .1.80, no presence of genomic DNA, RNA ratio 28s/18s .1 and RIN .7.0 (RNA Integrity Number).
RNA was linearly amplified and reverse-transcripted using the WT-Ovation Pico RNA Amplification System (NuGEN Technologies). Subsequently, a concentration of 100 ng/ml of cDNA labelled with Cy3 was hybridized in the Whole Human Genome Oligo Microarray G411A/G4112F of Agilent Gene Expression Arrays (Agilent Technologies, Madrid, Spain), encompassing .44 000 human DNA probes. The sample preparation and hybridization protocols of the CC samples were adapted from those in the Agilent Technical Manual (http://www.chem.agilent. com/Library/usermanuals/Public/G4140-90040_One-color_GE_5.7.pdf).
Hybridized microarrays were scanned in an Axon 4100A scanner (Molecular Devices, Sunnyvale, CA, USA) and data were extracted with GenePix Pro 6.0 software (Molecular Devices).
Spot intensities (medians) without background subtraction were transformed to the log2 scale. Data were normalized by quantile normalization. The replicates by gene symbol were merged and the data were filtered in order to delete unknown sequences or probes without gene description. The R-statistical software system was used for these purposes and for downstream analysis (R Development Core Team, 2004).
The GE profile was determined by means of non-parametric tests which perform comparisons 2 by 2, and two criteria were used to define genes with altered mRNA abundance between the different sample sets: an absolute fold change (FC) ≥2.0 and a corresponding FC P value , 0.05.
The biological processes that might relate to female fertility were analyzed by systematically searching for key gene ontology (GO) terms. In addition, a DAVID bioinformatics data analysis [database for Annotation, Visualization and Integrated Discovery (DAVID, http://david.abcc.ncifcrf. gov/home.jsp] permits the identification of those genes expressed in a wide variety of tissues. It is completed by performing bibliographical searches by GO (http://www.geneontology.org/) and ExPASy (http:// www.expasy.org/; Dennis et al., 2003).

Validation studies
For validation purposes, the quantitative qRT -PCR was carried out in an additional 19 CC samples. We made two pools of CC: one pool comprising 9 CC obtained from unstimulated cycles and the second pool containing a total of 10 CC from COS cycles.
First, the RNA was reverse-transcripted to cDNA by RT -PCR. PCR amplifications were performed using the LightCycler TM Fast Star (Roche Molecular Biochemicals) and SYBR Green I TM . A relatively stable transcript (GAPDH) was used to normalize the input cDNA.
For microarrays genes validation, we chose the most differentially expressed genes and those implicated in the ovarian biological process. These genes were MYH11 (Homo sapiens myosin heavy chain 11), SOX4 (Homo sapiens sex-determining region Y-box 4) and PRB2 (Homo sapiens proline-rich protein BstNI subfamily 2). Primers were designed with Genefisher (http://bibiserv.techfak.uni-bielefeld.de/genefisher2/). Table I presents the primer sequences, amplification product size and the accession number of the sequence used to design the primers for each gene.
The amplification products from qRT -PCR underwent electrophoresis using the Agilent DNA 1000 Kit (Agilent Technologies) following the manufacturer's instructions for the Bioanalyzer 2100.

Statistical analysis
The statistical analysis of the results was performed with the Statistical Package for Social Sciences, version 15.0 for windows (SPSS, Inc., Chicago, IL, USA). For the statistical analysis of the hormonal concentration in the follicular fluid, we used a x 2 comparison. The statistical analyses used for the assessment of spindles were the x 2 and Kruskal-Wallis comparison tests. A P-value of ,0.05 was considered statistically different. Difference in the GE profile was determined using non-parametric tests with two criteria to define the transcripts that had altered the mRNA abundance of the different sample sets: an absolute FC of 2.0 or more and P , 0. versus the unstimulated cycles and a negative value denotes underexpression in COS versus the unstimulated cycles.

Study population
The number of samples used for each analysis is presented in Table II. The study population characteristics [age, body mass index (BMI), gonadotrophin dose and number of oocytes] are presented in Table III. Age and BMI were comparable between both groups, but significant differences were observed, as expected, in terms of total doses of gonadotrophins and the number of oocytes retrieved.

Hormonal concentration in follicular fluid
The similar size of the pre-ovulatory follicles analyzed was demonstrated by similar follicular fluid volumes. Table IV presents the serum and follicular hormonal concentrations. As expected, the serum E2 concentration was higher in COS than in the unstimulated cycles. For follicular fluid, however, we observed that although the E2/T ratio was conserved, T was at a higher concentration in the COS group, while D4, E2 and LH were significantly higher in the unstimulated cycles.
The assessment of the meiotic state and the spindle structure of the oocytes generated in unstimulated and COS cycles (Table V) revealed no significant differences. This suggests that stimulation with gonadotrophins had no effect of any kind on these characteristics. All the values are mean + SD, ns, not significant. Serum concentration of E2 and, follicular E2, T and AD showed significant differences, and were higher in COS compared with unstimulated cycles, unlike follicular E2 and D4 which were greater in unstimulated cycles. no significant differences were found between the groups for spindle characteristics such as retardance, width, perimeter and area measures in Oosight. Length was the only parameter which presented differences between COS and the unstimulated cycles (Table VI).

GE profile by associated pathways
The microarray was done with the RNA that met the previously described characteristics. To analyze the relationship between the unstimulated and COS cycles in CC, the signal intensity of all the probes was compared. There were 18 genes in CC which were significantly differentially expressed, of which 7 were up-regulated and 11 were down-regulated in the COS cycles when compared with the unstimulated cycles (P , 0.05 and FC .2; Supplementary data, Tables S1, Fig. 1).
The DAVID software analysis revealed that these differentially expressed transcripts were significantly involved in biological processes, especially those involved in cellular developmental processes, cell activation, cell differentiation, immune system development and regulation of biological quality (P , 0.05). The supervised clustering of CC samples revealed that the pathways related to a positive regulation of cell proliferation and cell activation, and that leukocyte differentiation and activation were significantly affected. Table VII presents the significantly represented biological processes and the genes involved in each process. However, neither cellular components nor molecular functions were significantly represented (P . 0.05).

Microarray GE data validation by the qRT -PCR analysis
Microarray validation was performed with an additional 19 CC by means of qRT-PCR. Using this technique, a relative quantification of the mRNA from these genes: MYH11 (the two spliced variants), SOX4 and PRB2 in CC from unstimulated and stimulated oocyte donors, was performed to corroborate the data obtained by microarray technology. GAPDH was used as the housekeeping gene. Fig. 1 depicts the qRT-PCR results; the results are very similar to the microarray data. In both techniques, MYH11 was up-regulated in the unstimulated cycles and down-regulated in the COS cycles, SOX4 was up-regulated in the unstimulated cycles and downregulated in the COS cycles and PRB2 was up-regulated in COS and down-regulated in the unstimulated cycles. The size of the qRT -PCR products was checked by chip electrophoresis to verify their identity.

Discussion
Acquisition of oocyte competence relies on the well-controlled events accompanying follicular development (Eppig et al., 1997;Eppig, 2001;Barret and Albertini, 2010). The use of exogenous gonadotrophins may disturb this process by two possible mechanisms: by overriding the normal follicular physiology or by production of the small-sized follicles carrying incompetent oocytes. The present study was designed to describe the individual changes occurring in the hypothetically mature pre-ovulatory follicles in COS and to compare them with a more natural approach via an unstimulated cycle using oocyte donors as a gold standard patient. As far as we know, this is the first report to simultaneously explore the hormonal changes at the follicular level, as well as the CC and the oocyte perspectives.
Follicular fluid steroids are produced by granulosa and theca cells under the control of gonadotrophins, and the hormonal microenvironment may be of certain relevance in the subsequent pre-ovulatory development of the follicle, oocyte quality and IVF outcome (Pellicer et al., 1987). In the present study, even though higher serum E2 levels were observed in the COS cycles if compared with the unstimulated ones, the intrafollicular E2 concentration was lower. Conversely, despite unstimulated cycles having higher intrafollicular E2 and D4 and a lower T concentration, the E2/T ratio was conserved. Other  authors have shown that the E2/T ratio correlates with fertilization and cleavage rates (Andersen, 1993;Xia and Younglai, 2000). Moreover, significant differences were found in terms of the intrafollicular concentration of LH, which was lower in the COS cycles, probably due to the use of the GnRH agonist long protocol for down-regulation to avoid a premature LH surge. As expected, no relationship was found between the LH and P intrafollicular levels in both groups (Schoolcraft et al., 1991;Bosch et al., 2003). Although higher, this is in concordance with the fact that serum P level increases despite of low LH concentrations observed after the use of GnRH analog. LH and E2 serum levels have been previously associated with good embryo morphology and cycle success in stimulated cycles (Mendoza et al., 2002), we were unable to find this association in our study as the embryo morphology was similar in both cycles regardless of the steroid intrafollicular levels.
Spindle analysis has been proposed as a marker of oocyte competence (De Santis et al., 2005;Rama Raju et al., 2007;Kilani et al., 2009). We evaluated the effect of ovarian stimulation on the organization of the meiotic spindle and meiosis progression, but our results, with the exception of spindle length which was longer in COS cycles, indicated no statistical differences in the spindle features for either regime. Although it has been reported that more blastocysts are obtained from oocytes with spindle lengths of .12 nm than from oocytes with spindle lengths 10 -12 nm or ,10 nm, we cannot test this hypothesis since not all embryos were cultured till blastocyst stage (Rama Raju et al., 2007). Nevertheless, other studies looking at implantation potential did not reveal the spindle length as a marker of oocyte quality (Kilani et al., 2009).
Regarding our last objective in CC GE profile, it is well known that somatic cell contact optimizes meiotic maturation by regulating spatial organization and the function of the meiotic spindle, and that it is related with development and acquisition of competence (Eppig et al., 1997;Eppig, 2001;Albertini and Barrett, 2003;Hutt and Albertini, 2007;Barrett and Albertini, 2010).
DNA microarray technology permits the analysis of global GE profiles that are quite useful to understand the basis of cellular biology. The CC analysis can provide more information about the factors controlling follicular development, and it has been proposed as a noninvasive method for assessing embryo quality (Assou et al., 2006(Assou et al., , 2010van Montfoort et al., 2008). For example, the genes involved in the extracellular matrix, cell communication and cell adhesion, inflammation mediated by chemokine and cytokine signaling, angiogenesis or EGF signaling pathways provide an idea of the different potentially related pathways in oocyte and embryo quality (van Montfoort et al., 2008).
In our experiments, only a total of 18 genes were seen to be differentially expressed between the unstimulated and the COS cycles; this may be due to the hybridization strategy utilized (one color protocol) or perhaps because no huge differences exist between COS cycles and unstimulated ones.
The microarray was validated with three genes (MYH11, PRB2 and SOX4) which represent around 15% of the genes found to be differentially expressed between the two conditions studied validating therefore the rest of genes.
One interesting finding in the present study is that some of the genes which were differentially expressed after ovarian stimulation treatments were associated with leukocyte differentiation, T cell activation and regulation. These genes were Bloom syndrome (BLM) RecQ helicase-like, GTPase IMAP family (GIMAP) and SRY box 4 (SOX4), which were down-regulated in stimulation regimes.
Similar results have been reported with a comparable approach performed in mice, demonstrating that ovulation induction treatments induce a large repertoire of immune cell-like functions which could be involved in innate immune responses during the ovulation Characterization of unstimulated cycles process (Hernandez-Gonzalez et al., 2006;Shimada et al., 2006a,b). Genes that were thought to belong exclusively to immune cells, such as the toll-like receptors, TLR2 and TLR4, are known to be present on the surface of cumulus cells and are induced by LH. A similar situation may occurred in our case; in the present study unstimulated cycles had greater values of intrafollicular LH and increased expression of Bloom syndrome (BLM) RecQ helicase-like, GTPase IMAP family (GIMAP) and SRY box 4 (SOX4); those genes are not only related to T cell differentiation and activation, but also to other process such as organ development, cell differentiation, cell proliferation and regulation of biological quality.
A certain risk of white blood cells contamination of CC during follicular aspiration may exist, especially if very bloody follicles are obtained, however, this risk was minimized; on very few occasions the presence of highly bloody follicles were observed, moreover CC were picked-up very rapid from follicular fluid and thoroughly washed before analysis. However, the presence of CD4+ T cells in CC has been observed in women undergoing IVF (Piccinni et al., 2001). Exogenous gonadotrophins have been shown to regulate the ovarian secretion of cytokines by CC and T cells, such as IL-1 beta, IL-6, TNF-alpha levels, as well as other cytokines and immune cells (Loret de Mola et al., 1998. Among the repertoire of the genes differentially expressed between the unstimulated and the stimulated cycles, there were five which proved to be of interest to us. These genes were: MYH11 (myosin heavy chain), SOX4, a member of the SRY-related genes, ALDH1A2 (aldehyde dehydrogenase), CTSL2 (cathepsin L) and ZNF33B (zinc finger protein 33B).
MYH11 encodes for those molecules related to the regulation of angiotensin II (AngII) and angiogenesis by inducing the synthesis of AngII. An endothelin -angiotensin -atrial natriuretic peptide functional system at the ovarian level is involved in oocyte maturation and ovulation, and seems to be controlled by LH . The higher expression of MYH11 in an unstimulated CC suggested the possibility of its down-regulation through the use of exogenous gonadotrophins during stimulation and of its relation with ovulation and oocyte maturation processes.
SOX4 has been shown to be expressed in mice ovaries and testes, and is involved in the regulation of embryonic development and in the determination of the cell fate playing a key role in human sex determination during gonadal development. Here we demonstrate its expression in CC in humans, as well as its regulation by gonadotrophins, which also occurs in mice (Hunt and Clarke, 1999;Sim et al., 2011).
ALDH1A2 is the enzyme that catalyzes the formation of retinoic acid (RA) from vitamin A. RA is a meiosis-inducing factor which increases the mRNA levels for ALDH1A1 in the human ovary when meiosis begins and causes the requirement to synthesize more RA, and thus sustain meiosis (Le Bouffant et al., 2010). Moreover, vitamin A and its metabolites are important regulators of cell growth, embryonic morphogenesis and differentiation in many cell types. Moreover, both retinol-binding protein and cellular-binding protein mRNA have been detected in bovine oocytes in CC and follicular fluid (Mohan et al., 2001;Hidalgo et al., 2005) and, although it can stimulate oocyte cytoplasmic maturation and bovine blastocyst development in vitro (Mohan et al., 2001;Duque et al., 2002;Hidalgo et al., 2005), it can become teratogenic at a high concentrations (Hidalgo et al., 2005). Since ovarian stimulation is able to downregulate the expression of this enzyme in CC, we hypothesized that the final oocyte meiotic stages occur through interaction with the CC.
CTSL2, or cathepsin 2, expression was also up-regulated in CC from the unstimulated cycles. CTSL2 is a lysosomal cysteine protease required for ovulation and is a transcriptional target of P4 action after LH stimulation (Robker et al., 2000b). As no differences in serum or intrafollicular P were observed between the treatment groups, most of the CTSL2 expression observed in the unstimulated cycles may be due to the direct action of LH on the CTSL2 expression, as the intrafollicular LH concentration was higher in the unstimulated cycles, or the effect of cell recovery timing; this has been formerly shown in rats where the maximal levels of these proteases are observed at 12-16 h after an LH surge since oocyte retrievals in unstimulated cycles were performed 2 h earlier to evade risk of ovulation, and this timing may coincide with the maximal expression in cumulus cells (Robker et al., 2000a).
Very recently, some transition metals like zinc have been revealed to be involved in the regulation of meiosis in mammalian oocytes (Kim et al., 2010. In mice, total zinc content of maturing oocytes increases by over 50% during meiotic resumption, and insufficiency leads to failure to establish maturation promoting factor activity  causing meiotic arrest. Moreover, zinc finger protein expression has been detected in both mice testes and ovaries (Zhou et al., 2010) and its expression, at least in testes, seems to be associated with spermatogenesis events. The present report is the first to demonstrate the differential expression of the transcriptional factor ZNF33B in human CC with up-regulation in unstimulated cycles. Therefore, its role in human oocyte maturation and its down-regulation in stimulated cycles need to be further addressed.
Collectively, our results suggest that the use of ovarian stimulation induces changes in both follicular fluid and the CC GE. However, from the gamete competence perspective, these changes may be of no relevance to oocyte quality since no differences were observed in the fertilization ability, embryo morphology or meiotic status, although longer spindles were seen in the COS group. In our study, the spindle analysis reveals that the oocytes from the unstimulated cycles were found in an earlier meiosis II phase. This might eventually lead to embryos with chromosomal abnormalities because of premature penetration of sperm into the oocyte cytoplasm and the subsequent premature release of signals for oocyte activation . Indeed, unstimulated cycles do not rule out the possibility of yielding slow-developing embryos or embryos carrying chromosomal abnormalities (Verpoest et al., 2008). However, in terms of embryo quality, we found no differences in embryo morphology on Days 2 and 3 or the follicular growth regimes of those oocytes obtained from pre-ovulatory follicles and the oocytes picked at random from a pre-ovulatory follicle achieved after ovarian stimulation. The impact of controlled ovarian stimulation on oocyte and subsequent embryo quality has been addressed (Ziebe et al., 2004) and obtained similar results to those described in this study; this enables us to conclude that administration of exogenous gonadotrophins does not reflect impairments in cleavage capacity or embryo quality.
It is important to point out though, that in the COS group, we were analyzing only one follicle out of the usual multiple follicle population generated after ovarian stimulation, the results related to oocyte fertilization ability and embryo quality need to be interpreted with caution. We were only comparing one to one oocytes/embryos obtained from similar size follicles originated from both unstimulatd and COS cycles, not with the total population of embryos produced from COS cycles.
In conclusion, the similar-sized pre-ovulatory follicles of the unstimulated cycles have different hormonal milieu when compared with the follicles growing in COS cycles, and the CC experiment GE changes not only affect immune processes, but also meiosis and ovulation pathways. Nevertheless, although these differences do not seem to relate to early-stage embryo morphology. The implications of some of the molecules, especially ALDH1A2, CTSL2 and ZNF33B at the CC level, deserve to be addressed in future studies in order to understand their function in follicular growth, and the meiotic progression of oocytes and generation of euploid embryos.