Pathology of Hyperandrogenemia in the Oocyte of Polycystic Ovary Syndrome

Abstract

Polycystic ovary syndrome (PCOS) is the most common ovulatory disorder in the world and is associated with multiple adverse outcomes. The phenotype is widely varied, with several pathologies contributing to the spectrum of the disease including insulin resistance, obesity and hyperandrogenemia. Of these, the role of hyperandrogenemia and the mechanism by which it causes dysfunction remains poorly understood. Early studies have shown that androgens may affect the metabolic pathways of a cell, and this may pose hazards at the level of the mitochondria. As mitochondria are strictly maternally inherited, this would provide an exciting explanation not only to the pathophysiology of PCOS as a disease, but also to the inheritance pattern. This review seeks to summarize what is known about PCOS and associated adverse outcomes with focus on the role of hyperandrogenemia and specific emphasis on the oocyte.

Introduction

Polycystic ovary syndrome (PCOS) is the most common ovulatory disorder in the world, affecting at least 10% of women, translating to more than 100 million women worldwide [1]. It is a broad and complex disease, comprised of several aberrations in physiology leading to variable phenotypes. PCOS may be described as the common endpoint of several metabolic disturbances, including but not limited to dysregulated hypothalamic-pituitary-ovarian signaling, insulin resistance and glucose intolerance, obesity and metabolic syndrome, hyperandrogenemia, along with various environmental and genetic factors [1–4]. PCOS is associated with multiple adverse outcomes, including increased obstetric risks such as preeclampsia, gestational diabetes, preterm delivery and higher infant mortality [5]. Further, there are longstanding implications in chronic disease states such as hypertension, diabetes, depression, stroke and some cancers [5]. In the realm of fertility, PCOS patients have higher miscarriage rates [6, 7], higher risk of ovarian hyperstimulation (OHSS), and lower APGAR scores in babies born after assisted reproduction technology (ART) [8]. Various studies have shown PCOS patients tend to produce higher number of oocytes during fertility treatments [9–13]. However, this higher oocyte yield is contested by lower fertilization and blastulation rates [14] and higher risk of implantation failure [13]. Given these outcomes, patients with PCOS exhibit abnormal physiology that must be better understood. This review seeks to detail the role of hyperandrogenemia in the phenotypic manifestations of the disease, with emphasis at the oocyte and preimplantation embryo level, highlighting potential future research endeavors.

Diagnosis

The diagnosis of PCOS remains controversial, with a long history of proposals from various societies and expert panels suggesting a range of diagnostic criteria [15]. The current consensus is to use the Rotterdam criteria: polycystic ovary morphology, irregular menses and clinical or laboratory evidence of hyperandrogenemia [16]. At least two of these three criteria will qualify a patient for the diagnosis of PCOS provided routine workup excludes other etiologies such as prolactin or thyroid disorders [16]. However, it should be noted there are well recognized ethnic differences within these criteria, with certain populations showing variations in the prevalence of these clinical signs [17]. The relative weights of the three criteria are still debated. For instance, ultrasound evidence of polycystic ovarian morphology may be found in a large proportion of women without PCOS and may in some scenarios be normal [18, 19], though the recent changes in diagnostic criteria to the ultrasound findings has decreased this occurrence [20]. Additionally, the differential diagnosis for irregular menstrual cycles is extensive and must be weighed thoughtfully prior to assigning the diagnosis of PCOS to a patient [21]. On the other hand, hyperandrogenemia, the most persistent of the three criteria, presents interesting implications for further consideration [22]. The severity of the disease depends upon the combination of the Rotterdam criteria a patient exhibits and hence this disease has a spectrum of phenotypes [23]. Patients with all three criteria tend to fall on a more extreme end of the PCOS spectrum, and patients with only two of the three criteria will manifest a less pervasive phenotype [23]. Therefore, it is a challenge to tease apart the different etiologies that lead to this observed spectrum of the phenotype and may require delving into each facet of PCOS to quantify the weight of specific contributions and to understand the mechanisms of the disease.

Challenges to Investigate PCOS, Oocytes and Embryos

The diversity of the PCOS spectrum inherently presents a challenge in study design, and our ability to investigate and interpret. Scientific rigor requires that study of a specific cohort be both homogenous and representative of the population under study. As PCOS represents more of a compilation of sub-populations by varying combinations of diagnostic criteria, a homogenous study population is difficult to achieve. Thus, the general PCOS population is abound with confounders, particularly in the case of obesity. Up to 80% of PCOS patients are obese, and given the epidemic of obesity seen at present, it is known that this well studied comorbidity is also similarly associated with a wide range of adverse events [24–26]. Obesity has a large impact on patient outcomes with many far-reaching implications in the overall health of a patient from mental health disorders, to metabolic and cardiovascular disease, cancer, and all-cause mortality [27, 28]. Many of the outcomes associated with obesity are also closely associated with outcomes seen in PCOS literature, and in fact, there lies a complex interplay between the pathology of obesity and the pathology of PCOS [9, 29]. For example, the link between obesity and insulin resistance is well known and has been shown to exacerbate metabolic abnormalities in PCOS such as insulin resistance, further exacerbating adverse outcomes [29, 30]. To this end, obesity must be controlled for to understand the implications of PCOS in fertility. A subpopulation of PCOS patients that is not obese, coined “lean PCOS” has elicited interest among researchers and has increasingly been studied [30–35]. This group comprises 20–30% of the PCOS population and tends to have higher androgen levels than the more classic PCOS phenotype, thus providing researchers the opportunity to look at PCOS independent of obesity [36]. Studies to date have begun to show this lean PCOS cohort demonstrate a distinct phenotype of PCOS, different from the more classic obese phenotype as well as from patients without hyperandrogenemia [37, 38]. In the IVF setting, obesity is an independent negative predictor for various fertility outcome parameters. Various recent studies show that obesity decreases implantation rates and increases miscarriage rates [29, 39–41]. While there are also a plethora of studies on PCOS in IVF [5–7, 10, 42, 43], most do not differentiate between the hyperandrogenic subtype. Thus, most published studies do not control properly for obesity and hyperandrogenemia leading to difficulty in comparing different studies.

Although PCOS affects fertility, there is a large lacunae in the investigations related to oocytes and embryos in this population. Most of the studies have focused on the ovarian, hormonal axis or metabolic aspect of the disease. There is an urgent need to investigate the role of PCOS in the development, maturation, quality and competence of the oocytes and the early development, implantation and subsequent growth of the embryos. Further, there are also some unique technical challenges to investigate oocytes and embryos due to the amount of starting materials. Until recently due to technical difficulties, single cells could not be reliably studied. Recent advances in various techniques has allowed us to investigate a single cell and to understand its structure and function [44, 45]. These innovations have paved way to reliably investigate genomics, epigenomics, transcriptomics, proteomics, secretomics, metabolomics and cell communications at the single cell level. However, starting material and cell purity are still an issue in many of these techniques. Other technologies have been rapidly evolving to assist in single cell analysis include cell isolation, DNA/RNA sequencing, microscopy, mass-spectrometry, microfluidics, flow cytometry etc. Further, there are also severe ethical and legal challenges to investigate human embryos [46, 47].

Hyperandrogenemia and the Ovary

Androgens are a necessary component of normal female reproductive physiology [48] but, when increased to pathologic levels, they can result in significant insult. In normal physiology, luteinizing hormone (LH) stimulates the ovarian theca cell of the developing follicles to produce androstenedione from cholesterol [49]. Cholesterol is taken up and converted to the initial precursor pregnenolone by the StAR protein and desmolase, and then converted to androstenedione via 17-alpha hydroxylase, 17,20-lyase, and finally 3-beta hydroxysteroid dehydrogenase. This androgenic precursor is converted to estradiol in the granulosa cell as part of the elegant “two cell” pathway of folliculogenesis with the rate limiting enzyme aromatase [49]. This illustrates the importance of androgens in the female reproductive cycle as they are vital precursors for the production of estrogen [50]. Further, various studies shows that androgen is essential for follicular development and to promote apoptosis and inhibit proliferation of granulosa in mature antral follicles [51]. Detailed function and mechanism of action of androgens in ovary are elegantly reviewed in other publications and are beyond the scope of the current review [4, 52–54].

In PCOS, physiology is intrinsically altered resulting in the overproduction of androgens from the ovary, and partially from the adrenal gland, through a self-perpetuating cycle of activity [3]. It has been shown that elevated androgen levels can augment the activity of several enzymes along the steroid pathway, including side chain cleavage enzyme (p450scc), 17α-hydroxylase/17–20 lyase (p450c17), 5-α-reductase, as well as 3β-hydroxysteroid dehydrogenase (3βHSD) [13, 55]. Of note, androgens are important in the early phases of follicular growth, likely in the recruitment of early follicles, though the androgen receptor decreases as the oocyte matures through folliculogenesis [56]. However, patients with PCOS display a tendency towards a larger recruitment of early follicles and thus a higher production of AMH, likely in part due to the presence of increased androgen production [56]. This increased follicular activity results in further overproduction of androgens in the ovary by the theca cells, which alters the hypothalamic-pituitary axis to favor LH production over FSH and thus diminishes the capacity of the granulosa cell to aromatize the androgenic precursors into estradiol, and spilling the excess androgens into the bloodstream [55]. The cycle of androgens begetting more androgens results in systemic hyperandrogenemia perpetuating abnormal glucose/insulin metabolism, decreased hepatic sex hormone binding globulin production, altered hypothalamic-pituitary-ovarian (HPO) signaling, and dysregulated growth factor activity (IGF1, GDF9, activin, inhibin, etc.), all exacerbating the sensitive feedback system of the reproductive cycle [13, 22, 53]. Indeed, it is a known hallmark of PCOS to overproduce LH compared to FSH as a result of altered steroid feedback to the HPO axis. This FSH “deficiency” results in decreased signal and support of folliculogenesis with insufficient aromatization of androgen precursors to estradiol and subsequent impairment of developing follicles. To this point, a prior IVF study demonstrated that higher testosterone concentrations were found in the follicular fluid of meiotically incompetent oocytes compared to meiotically competent ones [57]. Furthermore, hyperandrogenic PCOS women treated with androgen blockers such as flutamide have shown improvements in fertility [53].

In addition, there is an increase in progesterone receptor expression of individual follicles that also exacerbates premature luteinization and growth arrest instead of maturation and ovulation [53, 58, 59]. The ultimate downstream effects of these alterations seen in both animal models and in clinical practice include increased follicular recruitment as mentioned above along with concomitant maturational defects leading to high atresia rates and low ovulation rates manifesting as oligomenorrhea. Over time, cysts accrue, made up of follicles that started to mature but found an unsustainable hormonal environment and were thus unable to continue [13, 59].

To date, limited data are currently available on the effects of increased androgens at the level of the cell, especially in the oocyte, that can help to better understand the true implications of PCOS. While a balance of androgen concentration is essential for the normal function [53], it is becoming increasingly apparent that supra-physiologic levels can be more harmful than initially appreciated, particularly at the late stages of folliculogenesis just prior to ovulation. Therefore, it is essential to define the specific insults created by hyperandrogenemia to define more specific targets for treatment or prevention.

Hyperandrogenemia and Oocytes:

The oocyte is the largest cell in the body, though the study of this cell has proven challenging through the years. Advances in several methodologies have allowed for more information to be gleaned at the level of the individual oocyte, and more information is rapidly becoming available. To that end, there are several animal models of PCOS, created via administration of androgens, including testosterone, androstenedione, dehydroepiandrosterone (DHEA), or dihydrotestosterone (DHT) which may be given prenatally, pre-pubertally or post-pubertally, that have been instrumental in these early, foundational experiments [60–62]. Most PCOS animal models, like most PCOS patients, tend to be obese with similarly associated metabolic and reproductive derangements [60, 61]. In particular, a study using one such obese mouse model has demonstrated poor oocyte quality showing enlarged perivitelline spaces, increased cytoplasmic fragmentation, enlarged polar bodies and increased lipid content in oocytes compared with controls [63]. Moreover, the authors further described abnormal mitochondrial function, as measured by decreased mitochondrial DNA copy number and decreased ATP content in the oocyte [63]. Another study using a different obese PCOS mouse model with androstenedione obtained oocytes and studied them under in vitro maturation conditions [64]. The authors noted abnormal follicular growth and increased atresia as well as inhibition of the spindle assembly with concurrent misalignment of chromosomes.

Another mouse model using oral administration of DHEA supplemented in chow was utilized to demonstrate dose-dependent alterations in oocyte quality via effects in the pentose phosphate pathway [65]. By demonstrating decreased concentration of NADPH, the authors showed that oral administration of DHEA can impair this pathway which is critical in oocyte maturation and normal metabolism. The downstream effects include altered cyclicity, sub-fertility, decreased numbers of oocytes collected in superovulation, and decreased lipid content in oocytes [65].

A more recent “lean PCOS” mouse model is also well described, created using administration of androgens prenatally, during the critical window of oogenesis [66–71]. This model has proven to be an ideal tool to investigate PCOS without the confounder of obesity, and has since been well characterized with otherwise similar metabolic pathology [69]. Looking specifically at the level of the oocyte, studies have shown that hyperandrogenemia can have negative effects, particularly at the site of the mitochondria [63, 64, 72, 73]. We recently showed that oocytes obtained from the lean PCOS mouse model displayed compromised inner mitochondrial membrane (IMM) potential and increased reactive oxygen species (ROS) formation, whereas the obese model also displayed compromised IMM potential with no change in ROS, with an increase in ATP concentration and mtDNA copy number was observed. Transmission electron microscopy showed significantly abnormal mitochondrial ultrastructure. These findings may be explained in part by the timing of the intervention in that the lean model is created using a short prenatal window of DHT over 3 days, where the obese model has chronic exposure to DHT throughout the adult lifespan [60]. Therefore, it may be that the chronic exposure to DHT serves as a constitutive signal for proliferation of more mitochondria (as evidenced by increased ATP and mtDNA copy number) that may be able to buffer the ROS formation despite poor IMM potential and altered structure. In fact, prior research has shown androgen receptor binding elements in the mitochondrial genome with effects in transcription regulation [74]. Both obese and lean mouse models have shown alterations in mitochondrial structure and function as well, yet with subtle phenotypic differences [71, 75, 76]. Further, more recent human studies have also shown differentially expressed mitochondrial genes in PCOS oocytes compared to controls [77].

Beyond the alterations in the functional capacity of the oocyte, there are genetic changes as well. Importantly, mRNA from oocytes is unique in that its half-life is substantially longer than in other cells, a necessary adaptation to allow for the early growth of an embryo during the transition to the fetal genome [78]. In a landmark study in 2007, Wood, et al showed altered gene expression in MII oocytes from PCOS patients compared to controls, particularly involving genes related to the maternal zygotic transition and the mitotic cell cycle [79]. Further animal studies have shown altered mRNA transcripts in PCOS versus controls [80, 81], with differences noted in the lean PCOS phenotype versus the obese [75, 76]. This would be consistent with the understanding that obesity is an important confounder in PCOS, and indeed, other research has shown that obesity alone can alter mRNA transcripts in the oocyte [82]. Clearly, more research is needed in this area to better elucidate these findings.

Hyperandrogenemia and Embryo:

While research on embryos poses many challenges, studies are emerging that speak to the effects of hyperandrogenic environments and their impact on embryonic development and miscarriage rates in women with PCOS, though these mechanisms are still actively being explored. Early studies have demonstrated alterations in growth patterns via morphokinetics as well as morphology with higher trophectoderm (TE) to inner cell mass (ICM) ratios [63, 64, 83]. Further, blastomeres that lyse prior to compaction have lower ATP levels than blastomeres that do not, which taken with the findings presented in the oocyte above, may indicate a certain threshold of ATP necessary for continued survival of an embryo [84].

Several studies have been published on the growth pattern of embryos in patients with PCOS compared to controls. Advances in incubation and microscopy have allowed for time lapse imaging of developing embryos and have spawned the study of morphokinetics, whereby the rate of growth of embryos can be observed and measured [85–87]. While some findings are conflicting, it appears on the whole that embryos from PCOS patients develop faster to the morula stage compared to age and BMI matched controls [37, 88–93]. One of the larger studies of PCOS embryos to date was done by our group where we had performed a sub-analysis of PCOS patients with hyperandrogenemia compared with the PCOS patients without hyperandrogenemia and controls [37]. In this paper we showed that the hyperandrogenic population have a more pronounced alteration in embryo development than the PCOS cohort as a whole and also showed a higher miscarriage rate [37].

Interestingly, this finding is similarly observed in the primate model of PCOS with altered progression of morula to blastocyst [94]. The morula stage is defined by the compaction of the cells immediately following the early divisions that increase the cell number to 8–10 cells. Compaction is driven by cell to cell signaling via gap junctions, desmosomes, tight junctions and other adhesion molecules [95]. Connexin 43 (Cx43) has been shown to be important in both mouse and human compaction, and animal studies have demonstrated increased expression of Cx43 in PCOS mouse models [96]. This protein, also known as GJA1, is also important in folliculogenesis and has further been shown to be under expressed in the oocytes of PCOS patients compared with controls [80]. Additionally, E-cadherin becomes activated at the time of compaction and is thought to play a role, and there are studies showing increased expression of E-cadherin in PCOS patients [97]. How these alterations translate clinically remains to be fully elucidated but it has been observed that PCOS patients have increased adverse outcomes with IVF compared to controls including miscarriage and other downstream obstetric complications [37, 38].

Implications

Taken together, it is clear that exposure to supraphysiologic androgens has adverse effects on the oocyte and subsequent embryo, particularly at the level of the mitochondria and in several important pathways of gene expression. These findings are important to consider within the context of two important concepts. Firstly, mitochondrial growth and replication do not occur in an embryo until blastulation, and therefore, without adequate mitochondrial reserve prior to ovulation and fertilization, the energy required for all divisions up to blastulation and implantation will be insufficient and the only possible outcome must be cessation of further growth. Further, given that all mitochondria are inherited from the oocyte, the competence of these mitochondria is critical in the health of all cells in the ensuing progeny. It is easy to see the importance of a healthy environment for mitochondria to translate into developmental competence of both the oocyte and embryo.

Second, as mitochondria uniquely contain their own genetic code for key enzymes related to their function, they also present an exciting hypothesis to explain the inheritance pattern seen in PCOS. To date, researchers have been unable to identify gene candidates that can explain any significant portion of this genetic relationship, despite the knowledge that at least 60% of daughters to PCOS patients will also manifest PCOS [4, 98–100]. As mitochondria are strictly maternally inherited, all mitochondrial function in all ensuring cells of the progeny rests on the health and vitality of the mother at the time of development and maturation of the oocyte. If there were to be significant pathology in an oocyte’s mitochondria, this would be a plausible explanation for the propagation of this complex disorder. Indeed, animal studies have shown that abnormal mitochondria can be passed from generation to generation, continuing to show pathology, even if no further interventions are encountered in the offspring [101–103]. More specifically, studies involving the Indian rhesus macaque, known to have a high DNA sequence homology with humans and to naturally overproduce androgens, have shown that prenatal exposure to these supraphysiologic androgen levels leads to offspring with clinical signs and symptoms consistent with PCOS [104, 105]. Moreover, it is certain that androgens in the human maternal bloodstream cross the placenta, and a PCOS woman exposes her fetus to more androgens than a non-PCOS patient [106–108].

Precisely what molecular mechanisms are at play in affecting these changes are not clearly understood yet. It may be a direct effect from excess androgens, or from resulting hyperinsulinemia secondary to insulin resistance. Furthermore, there could be changes further upstream. While genome wide association studies have not demonstrated strong candidate genes to explain the strong heritability of PCOS, there are promising data regarding epigenetics that may impart a higher risk for developing PCOS, resulting in excess androgens, insulin resistance, and ultimate mitochondrial dysfunction [103, 109].

Regardless of the etiology, it appears that if mitochondrial changes were to occur in the maternal oocyte, directly or indirectly from exposure to androgens, then one should observe abnormal mitochondria in multiple tissue types of the offspring, given the maternal inheritance pattern of mitochondria, and that all mitochondria of an offspring are singly derived from the fertilized oocyte and nowhere else. In fact, studies in a rat model have noted mitochondrial dysfunction in the pancreas [110] and kidney [111], and human studies have now reported altered mitochondrial function in skeletal muscle [112] as well as follicular fluid and cumulus cells [113]. Therefore, it is likely that these changes in the oocyte mitochondria are transferred to the offspring, and thus result not only in the adverse reproductive outcomes seen in the PCOS population such as miscarriage [6], but also in the increased prevalence of PCOS in offspring [54]. To date, studies investigating the role of supplements to ameliorate adverse effects seen in PCOS have demonstrated limited benefit. Supplements include compounds that improve insulin sensitivity such as inositols, antioxidants such as melatonin or N-acetylcysteine, vitamins and minerals such as Vitamin D, B12, folate, selenium, Zinc, and chromium, and even probiotics [114, 115]. In summary, many of these compounds have been shown to improve certain serum markers of PCOS such as insulin sensitivity or cholesterol, but no one supplement has been shown to conclusively improve oocyte or embryo quality or fertility outcomes, though further research is certainly warranted.

Concluding Remarks

Hyperandrogenemia significantly impacts multiple organ systems, affecting hypothalamic-pituitary-ovarian signaling, glucose and insulin metabolism, and steroid regulation and production. Integral to these effects may be mitochondrial dysfunction at the level of the oocyte, which could then be propagated to fertilized embryos and ultimately offspring. These implications merge several fields of study in reproduction. It is known that PCOS patients have abnormal reproductive outcomes similar to patients with other metabolic diseases. It is also now known that abnormal metabolism affects offspring not only through genetics/epigenetics, but also through mitochondrial dysfunction due to the maternal inheritance of the organelle. To the best of our knowledge, there are no data to explain a possible mechanism linking hyperandrogenemia and mitochondrial dysfunction, though early work linking the two sheds light on this possibility. Perhaps the pathophysiology involves abnormal insulin/glucose signaling in the oocyte or via alterations in gene expression by manipulation of transcription, mRNA transcript stability, or even miRNA expression. Regardless, the scope of mitochondrial health in this complex disease warrants close inspection, particularly given its significant implications of inheritance pattern, and the widespread compromise in cellular function throughout the offspring.