Keywords

1 Introduction

Polycystic ovarian morphology (PCOM) and polycystic ovary syndrome (PCOS) are very common findings in an in vitro fertilization (IVF) centre population, explaining why the concept of oocyte quality is so challenging. Indeed, 18–25% of the infertile couples meet the diagnosis criteria for PCOS ([1]; http://www.ivf-worldwide.com/survey/pcos-results.htm), whereas the prevalence of PCOM has been estimated as high as 33% in asymptomatic patients [2,3,4]. With the advent of highly sensitive ultrasound machines, the PCOM, mainly based on the antral follicle count, has become one of the main diagnosis criteria of PCOS and is now well recognized to be the common base of the wide spectrum of clinical, hormonal and metabolic phenotypes of the PCOS. PCOM is characterized by a significantly enlarged cohort of early growing and recruitable follicles. This excessive follicle number is linked to disturbances in folliculogenesis which are thought to be the consequence of intra-ovarian hyperandrogenism [5,6,7] (see also Chaps. 2 and 8). During controlled ovarian hyperstimulation (COH), the cohort of growing follicles is frequently heterogeneous in size, with mature, intermediary and small follicles. In parallel, the number and quality of mature oocytes has been proposed as being poor [6, 8, 9] leading to lower pregnancy rates and higher abortion rates. Furthermore, recent data suggested that oocyte competence could be impaired in PCOS patients due to an inadequate dialogue between the cumulus cells and the oocyte and an impairment of the follicular microenvironment [10, 11].

Despite these assumptions, the paucity of clinical studies focusing on oocyte quality in PCOS women does not allow to make definitive conclusions. Most of these studies are retrospective, concerning PCOS patients diagnosed with heterogeneous criteria and with various clinical and metabolic phenotypes with a historical control group [12,13,14,15,16,17] (see also Chaps. 2 and 7).

Mostly, oocyte quality was only extrapolated through fecundation, implantation and pregnancy rates according to the number of mature oocytes, i.e. metaphase II oocytes [12,13,14,15,16,17]. Consequently, the results provided by these studies are conflicting, reporting either a better oocyte/embryo quality and pregnancy rates or vice versa.

The concept of oocyte quality represents the oocyte developmental competence, meaning the intrinsic ability of oocytes to undergo meiotic maturation, fertilization, proper embryonic development and successful pregnancy [18]. These competences are progressively acquired during the follicular development through the cross-talk between oocyte and somatic/surrounding granulosa cells (GCs) [19]. As follicular growth disturbances have been widely shown in PCOS, especially during COH, it has been postulated that the consecutive endocrine and/or paracrine follicular microenvironment modifications could have detrimental effects on the oocyte quality.

In this review, we will firstly focus on the main approaches proposed to investigate oocytes from PCOS, and then we will look over the clinical relevance of all the compiled results from these studies by examining the final oocyte competence during the IVF process in PCOS patients.

2 Evaluation of Oocyte Quality

Evaluating the oocyte quality is obviously complicated because of the few number of oocytes retrieved during the IVF process. Hence, there is various ways, invasive or non-invasive, direct or indirect, to investigate the oocyte competence. Three main approaches, morphological, genetic and OMICS, have been proposed to investigate oocytes from PCOS.

2.1 Morphological Approach

After having removed the cumulus-corona cells in preparation for intracytoplasmic injection (ICSI), oocyte evaluation is based on the nuclear maturation status, the morphology of the cytoplasm and on the appearance of the extracytoplasmic structures. A higher number but a same rate of metaphase II oocytes between PCOS patients and controls were highlighted in various retrospective studies [12,13,14], as well as a higher number but a same rate of top-quality embryos [13, 14]. These results were confirmed in the meta-analysis of Heijnen et al. and in a recent prospective study comparing PCOM patients versus non-PCOM [20, 21]. Despite being the main morphological indicator, nuclear maturity examination alone is not enough to determine the quality of an oocyte: the nucleus and the cytoplasm have to mature in synergy in order to reach the conditions for an optimal fertilization rate. In addition, it is now well recognized that some specific morphological oocyte abnormalities, such as the presence of a wide perivitelline space (PVS) or a granular cytoplasm, must be given attention since it has been reported that they are associated with a significant decrease in the chance of fertilization [22, 23]. Only three studies are interested in oocyte morphology in PCOS patients. Sahu et al. [13] reported similar oocyte morphology in isolated PCOM, PCOS and age-matched controls. Sigala et al. [21] performed a prospective comparative study with a systematic examination of oocyte morphology in PCOM versus non-PCOM patients. Nuclear maturation, extra- and intracytoplasmic oocyte abnormalities were assessed, i.e. fragmented or abnormal first polar body (IPB), abnormal zona pellucida (ZP), presence of an enlarge perivitelline space (PVS) or material in the PVS and an abnormal shape of the oocyte [21]. No specific morphological abnormalities as well as no difference regarding the incidence of these abnormalities were observed in the PCOM versus non-PCOM patients. In this study, PCOM was defined according to the threshold of 19 follicles per ovary. Among the PCOM group, there were 52.5% of PCOS and 47.5% of PCOM only. No difference was observed between oocytes from PCOS and PCOM-only patients, but it has to be mentioned that obese PCOM patients were excluded [21]. Piomboni et al. [24] compared the oocyte quality based on the precited morphologic criteria in three groups of nonobese PCOS patients: PCOS treated by d-chiro-inositol, PCOS treated by metformin and non-treated PCOS. They showed a significantly higher number of good-quality oocytes in the groups treated by insulin-sensitizing agents, as well as they highlighted in parallel a significant reduction of reactive oxygen species production in the follicular fluid [24]. Data on the effect of inositol on oocyte quality are extensively reviewed in Chap. 16.

A few studies using polarized light microscopy (PLM) have shown, by highlighting meiotic spindle abnormalities, that some of metaphase II oocytes may still be immature [18, 25]. Indeed, meiotic spindle, when detectable with polscope microscope, is not always aligned with the first polar body (PB1) in fresh metaphase II oocytes, which may adversely affect the fertilizing ability and in vitro/in vivo developmental competence [25]. In addition, it has recently been shown that oocytes with normal spindle morphology are significantly more likely to produce euploid embryos [26]. Only two studies are interested in the spindle and chromosome configurations in oocytes from PCOS patients. Li et al. [27] compared the incidence of abnormality in spindle and chromosomal configurations in both in vitro and in vivo-matured oocytes. A higher rate of abnormalities was found in the group of in vitro-matured oocytes [27]. Vieira et al. compared in vitro-matured oocytes from PCOS versus non-PCOS patients and didn’t find any difference [28]. To the point of view of the authors, it is more likely that the in vitro conditions of oocyte maturation might explain these meiotic abnormalities rather than the PCOS itself.

2.2 Genetic Approach

The decrease of fertilization rate and the increase in pregnancy loss in certain subgroups of PCOS patients have led to the hypothesis that oocyte and embryos could be of poorer quality due to a higher aneuploidy rate. Morphological examination is insufficient to detect genetic abnormalities such as aneuploidy. Currently, the only method yielding a definitive evaluation of oocytes/embryos is aneuploidy diagnosis, which provides information of normal or abnormal chromosomal constitution. Moreover, aneuploidy detection requires breach of the zona pellucida and biopsy of the polar bodies from the oocyte that is obviously invasive. Because of the difficulty to directly evaluate the ploidy of the oocytes, some studies investigated indirectly the genetic potential of the PCOM gamete. Sengoku et al. [29] performed cytogenetic analysis on the unfertilized oocytes with normal morphology from PCOS and control patients but didn’t find any difference in the incidence of aneuploidy or diploidy [29]. Weghofer et al. [30] examined the association between PCOS and embryonic aneuploidy. They compared the results of preimplantation genetic diagnosis (PGD) between properly documented PCOS patients and controls. Despite a statistically higher absolute number of euploid embryos in PCOS group due to the higher number of metaphase II oocytes retrieved, there was no difference regarding the rate of aneuploidy [30]. Wang et al. [31] conducted a prospective cohort study by performing genetic analysis on abortuses from PCOS and non-PCOS patients who conceived after IVF. The aneuploidy rate was not significantly higher in the PCOS group but significantly lower in comparison with the non-PCOS patients [31].

2.3 OMICS Approach

The emergence of the—OMICS technologies, i.e. epigenomics, genomics, transcriptomics, proteomics and metabolomics, provides a huge amount of new information regarding the biological processes involved in the reproductive field. Although using an invasive technology, microarray-based transcription profiles of oocytes at various stage of growth and maturation have led to a better understanding of the genes expressed during oocyte development: disruption of transcription within an oocyte or any modification of their adequate transcriptomes could compromise its growth and development, as well as the resulting embryo, since oocyte mRNAs pool is strongly correlated with the ability to develop into the blastocyst stage [32].

Recent cluster of analysis revealed differences in global gene expression profile between normal and PCOS tissues and oocytes [33]. Wood et al. [34] identified 374 genes with different mRNA transcript levels when analyzing metaphase II oocytes from normal responders and PCOS patients. A subset of these genes found to be differentially expressed in PCOS is involved in spindle dynamics, homologous recombination/chromosome alignment, cell cycle checkpoint and centrosome function during mitosis and/or meiosis [34]. Furthermore, some of the other differentially expressed genes contain putative androgen receptors and/or peroxisome proliferating Y binding sites [34]. The authors make the assumption that these observations could be related to a lower oocyte quality.

Cai et al. [35] interested in the in vitro effects of overexpression of Hsp27 on oocyte maturation and development derived from PCOS patients as emerging evidence indicates this heat-shock protein has strong antiapoptotic properties and has been shown to be mainly expressed in human oocytes. Interestingly, the same team has previously shown that Hsp27 was downregulated in ovarian tissue and in oocytes isolated from women with PCOS [36]. The results of upregulation of Hsp27 expression were a lowered oocyte maturation rate, similar fertilization but high-quality embryo blastocyst formation rates leading the authors to postulate that Hsp27 could be involved in the apoptotic imbalance of the oocytes via growth and differentiation factor 9 (GDF9) and bone morphogenic protein (BMP15) [36].

It is now recognized that there is a continuous bidirectional cross-talk between oocytes and somatic cells during folliculogenesis through gap junctions and paracrine signalling [19]. If it is sure that oocyte is nurtured and supported by the closely associated somatic cells, especially those of the cumulus, the fact remains that oocyte itself plays an active role via secretion of paracrine signalling factors, such as GDF9 and BMP15, which maintain an appropriate microenvironment for a proper follicular growth [19, 33]. It is now possible to identify the transcriptome of GCs with the microarray technology, and evidence supporting GCs genes markers as valuable and non-invasive predictors of oocyte competence is rapidly emerging [37].

As folliculogenesis disturbances frequently occur in PCO patients, it seemed logical to investigate this oocyte/granulosa cells dialogue. Ouandaogo et al. [38] compared the transcriptome profiles of cumulus cells (CCs) isolated from in vivo and IVM cumulus-oocyte complexes (COCs) at different nuclear maturation stages from normal responders and PCOS patients undergoing ICSI following in vivo or IVM. In the PCOS subgroup, the authors found a strong alteration of the expression profile of the CCs derived from IVM metaphase II oocytes in comparison to in vivo metaphase II oocytes [37]. The expression profile also differed significantly between normal and PCOS patients, but the authors conclude that these significant differences were related to the culture condition, not to the PCOS per se [37]. The same team focused few months later on the gene expression profile in CCs of in vivo metaphase II oocytes from PCOS and non-PCOS patients using the DNA microarray technology [39]. There were significant differences between groups in the gene expression profile. In addition, CCs from PCOS patients were characterized by abnormal expression of many growth factors, including members of the epidermal growth factor-like and IGF-like families that are known to play a role in oocyte competence [39]. Likewise, mRNA transcripts of factors involved in steroid metabolism seem to be deregulated in PCOS CCs [39].

miRNAs are small, non-coding RNAs detected in biological fluids that are able to regulate gene expression at the post-transcriptional level and which may be involved in reproductive function [40]. A limited number of studies have aimed to extensively profile circulating microRNAs (miRNAs) expression and function in the follicular fluid within a PCOS study population, and the results are at times not yet conclusive [40, 41]. Both highlight different miRNA expression between PCOS and non-PCOS patients but with no clear correlation with oocyte maturation or fertilization competence [40, 41].

3 Clinical Data

Research on oocyte is extremely complex, especially in PCOS patients undergoing IVF. Firstly, the fact that PCOS patients need IVF, independently of tubal or sperm alterations, introduces some recruitment bias. Secondly, the high ovarian response under COH in these patients may have detrimental effects on the oocyte quality through vascular and inflammatory factors. Likewise, the in vitro/in vivo culture conditions may play a major role in the oocyte quality. PCOS underlies endocrine, ovulatory and/or metabolic dysfunction. These three components are solely or synergistically strong confounders regarding the interpretation of the studies on oocyte quality. Indeed, the source of oocytes is highly variable according to the different PCOS phenotypes, from the lean PCOM-only patient to the obese PCOS patient. In addition, the question as to whether asymptomatic women with PCOM constitute a heterogeneous population in terms of ovarian dysfunction ranging from entirely healthy ovulatory women to women with mild occult PCOS [2] or, alternatively, constitute a homogenous population representing the milder end of the PCOS spectrum remains debatable. More data about PCOS phenotypes are discussed in Chaps. 2 and 7.

Undoubtedly, oocytes from PCOM in IVF condition exhibit molecular specificities in comparison with oocytes from non-PCOM patients. But surely, the key point is to have sufficient oocytes of quality to give substantial chances of pregnancy. One has to recognize that the larger prospective studies regarding the IVF results in PCOS failed to highlight poor results in terms of pregnancy chances, at least in nonobese patients.

Heijnen et al. [20] reported in a meta-analysis the comparative IVF outcome of PCOS patients defined by the Rotterdam criteria to the one of matched non-PCOS controls. Except the higher number of oocytes in PCOS patients, they failed to find any difference between fertilization, pregnancy and take baby home rates between groups [20]. Likewise, Sigala et al. [21] in a large prospective comparative study have shown the same rate of metaphase II and morphologically normal oocytes in nonobese PCOM versus nonobese non-PCOM patients. The rate of top-quality embryo was equivalent in the two groups while the implantation and clinical pregnancy rates were even better in the PCOM group [21]. These results were also confirmed after having divided the PCOM group in PCOS and “sonographic-only” PCOM. Engmann et al. [15] previously reported same results in PCOM patients. Hence, the bad prognosis frequently argued regarding pregnancy rates and/or abortion risk could be more related to the metabolic profile than to PCOM per se. It is well known that high body mass index (BMI) and hyperinsulinaemia are main contributors to the follicular microenvironment disturbances [9]. Indeed, studies in follicular fluids from PCOS patients highlight high levels of interrelated endocrino-metabolic factors such as androgens, VEGF, AMH, insulin and IGF, all of them playing an active role in the oocyte-CCs dialogue [9]. Moreover, numerous studies highlighted benefits from taking insulin-sensitizing agents like metformin as a co-treatment before and during the IVF process [42, 43]. The use of insulin-sensitizing drugs in PCOS is reviewed in Chap. 11.

From another point of view, as mentioned above, it seems that the meiotic/mitotic cell cycle pathway is altered in PCOS oocytes [34], but no more aneuploid oocytes were detected in PCOS patients [29] as well as no more aneuploidy embryos were observed [30]. It is therefore difficult to make the connection between such molecular specificities and the reality of the clinical experience.

Conclusion

It remains difficult to conclude about the oocyte competence in PCOS. Molecular specificities have been properly highlighted in PCOS oocytes, but it doesn’t mean that there are abnormalities in their reproductive competence. Too few studies have taken into account the very wide spectrum of PCOS phenotypes and the potential influence of comorbidity factors such as obesity and insulin resistance in their analysis. However, the largest prospective studies performed in nonobese PCOM patients so far are in favour to good pregnancy chances due to a high number of good-quality oocytes. The coexistence of a metabolic syndrome to PCOM seems to impair the oocyte competence and the pregnancy rates. Prospective studies with consensual international diagnosis criteria are needed to allow in better understanding of the exact molecular mechanisms involved in the oocyte competence according to each phenotype of PCOS and would elucidate if the PCOS oocyte deserve its bad reputation.