Developmental origins of polycystic ovary syndrome: Everything starts in utero

Introduction

Polycystic ovary syndrome (PCOS) is a prevalent, complex, highly heritable, subfertility health disorder of premenopausal women with numerous comorbidities . PCOS is commonly diagnosed from either revised Rotterdam or NIH criteria. Rotterdam criteria require at least two out of three: oligo/anovulation, hyperandrogenism, and polycystic ovaries. NIH criteria, in contrast, require only the first two (NIH, ; Rotterdam: , Revised by ). A PCOS diagnosis, nevertheless, requires the exclusion of endocrine disorders that mimic PCOS pathophysiology, including congenital adrenal hyperplasia, hyperprolactinemia, and primary hypothyroidism .

Women with PCOS are dissatisfied with currently available pharmaceutical, surgical, and lifestyle interventions because of PCOS-related contraindications and unpleasant side effects, postsurgical complications, burdens of compliance, and absence of a cure . In addition, women with PCOS often have to endure transitioning between two and three different physicians before they receive a PCOS diagnosis and start PCOS-appropriate clinical management . Inadequate knowledge of PCOS pathophysiology, PCOS diagnostic criteria, and effective clinical management have slowed progress towards a more timely PCOS diagnosis and achieving a cure. Both are essential for ameliorating and eliminating distressing symptoms, improving psychological health, quality of life, and fertility in order to minimize long-term health problems . Recent development and dissemination of International Guidelines for the clinical care and management of PCOS , advocacy for harmonizing research outcomes for PCOS , increasingly effective national PCOS patient support groups, engagement of new generations of therapeutics such as incretin-based therapies , and growing awareness of familial and developmental origins for PCOS , however, encourage optimism for paradigm-changing advances.

PCOS comorbidities include subfertility, insulin resistance, type 2 diabetes (T2D), obesity, cardiovascular disease (CVD), excessive hair growth, acne, mood disorders, and sexual dysfunction , contributing to more than $267 billion annual healthcare costs in the USA, alone . Given the prevalence and economic burden of PCOS on patients and the healthcare system, it is imperative to understand its origins.

PCOS phenotypes: Reproductive and metabolic components

PCOS women diagnosed by Rotterdam criteria exhibit four phenotypes: Classic PCOS, including type A, hyperandrogenism or hirsutism (HA) with intermittent/absent menstrual cycles (ovulatory dysfunction, OD) and polycystic ovary morphology (PCOM), as well as type B, HA, and OD; type C, HA, and PCOM; and type D, OD, and PCOM. NIH criteria comprise Classic PCOS, types A and B, alone. PCOS women with nonclassic phenotypes diagnosed by Rotterdam criteria contribute the majority of women with PCOS in the general human population, whereas PCOS women with classic phenotypes comprise the majority of PCOS in clinical referrals . Not surprisingly, therefore, PCOS women with classic phenotypes more prominently exhibit comorbidities, including obesity, insulin resistance, and preferential accumulation of abdominal fat, key components of the metabolic syndrome contributing to T2D and CVD .

There are parallel increases in women's body mass index (BMI) and the incidence of PCOS . BMI, however, has been progressively increasing among all women across recent decades. Hence it is not surprising that 50%–60% of women with PCOS are obese or overweight . Interestingly, gene variants in women associated with increased BMI positively correlate with increased incidence of PCOS , suggesting that genetic predisposition to obesity increases the risk of a PCOS phenotype, probably because of accompanying compensatory hyperinsulinemia due to insulin resistance . The converse, however, that genetic predisposition to PCOS increases the risk of obesity, was not found . PCOS women with classic phenotypes, therefore, appear to represent a more metabolically compromised PCOS presentation, in contrast to nonclassic phenotypes that represent a more reproductively compromised presentation. Such differential phenotypes also change with increasing age, since younger women with PCOS (20–30 years) are more likely to present with reproduction-related issues, whereas older women with PCOS (30–40 years) are more likely to present with metabolic dysfunction .

PCOS genetics

Whole genome-wide association studies (GWAS), exploring which gene variants are associated with different PCOS diagnostic criteria (Rotterdam, NIH, or self-reports by women), suggest significant commonality among PCOS risk genes associated with all PCOS phenotypes . Such commonality of genetic architecture across diverse PCOS phenotypes implies shared molecular and developmental pathogenesis across PCOS phenotypes. GWAS and family-based whole-genome sequencing studies have identified > 20 PCOS gene variants, with at least 5 identical variants appearing in PCOS women within both European and Chinese populations . The most parsimonious explanation for such PCOS gene variant concurrence across two geographically disparate human populations suggests that PCOS-associated gene variants were likely present in ancient Homo sapien populations before their dispersal out of Africa ~ 150,000–300,000 years ago .

All but one of the GWAS identified PCOS gene variants, however, account for < 10% of PCOS. The exception, DENND1A , is involved in regulating androgen biosynthesis . In family-based, whole-genome sequencing studies, rare gene variants of DENND1A were found in 50% of PCOS families . A posttranscriptionally truncated form of DENND1A , DENND1A.v2, is over-expressed in women with PCOS, and over-expression of DENND1A.v2 in human ovarian follicle theca cells enhances ovarian androgen biosynthesis . In a separate, family-based whole-genome sequencing study, rare gene variants of antimullerian hormone ( AMH ) and of its cognate receptor were found in ~ 3% of PCOS families . AMH is produced by ovarian follicle granulosa cells, is a crucial intra-ovarian regulator of follicle development , and stimulates hypothalamic GnRH release . Elevated circulating AMH levels, and thus increased antral follicle populations (ovarian defect), in women with PCOS may thus contribute to LH hypersecretion (neuroendocrine defect) and enhance intrinsic ovarian hyperandrogenism .

In a recent Mendelian randomization genetic study, specific combinations of gene variants in women associated with high bioavailable (unbound) testosterone levels were also associated with T2D and PCOS , suggesting a causal effect between hyperandrogenism in human females and its likelihood to induce T2D or PCOS.

Recent genetic studies re-categorize PCOS women into etiologically disparate reproductive and metabolic phenotypes

More recently, a statistical unsupervised clustering approach was employed using reproductive and metabolic traits in PCOS women in order to identify PCOS phenotypes from quantitative parameters, rather than from discrete diagnostic criteria, as well as their association with PCOS risk genes . Two distinct phenotypes were identified in women with PCOS using this approach and were termed “reproductive” (higher luteinizing hormone (LH) and sex hormone-binding globulin (SHBG) levels, but not BMI or basal insulin levels) and “metabolic” (higher BMI, basal glucose and insulin levels, but not LH or SHBG). They are reminiscent of previous dichotomies between nonclassic and classic PCOS, as well as between younger and older women with PCOS ( . The reproductive and metabolic phenotypes comprised 23% and 37% of women with PCOS, respectively, with the remaining 40% termed as “intermediate” and demonstrating no distinct pattern . A comparable outcome was achieved when the analysis was limited to a family-based cohort of women, suggesting such distinctions of PCOS phenotype are robust across generations. This PCOS study, in contrast to prior GWAS, identified two genetically distinct reproductive and metabolic phenotypes within PCOS, each with its relatively homogenous developmental origin .

Characterizing women with PCOS into two genetically and phenotypically distinct pathophysiological phenotypes with differing etiologies suggests that PCOS is a heterogeneous disorder that may have varied developmental origins. This re-engages a long-standing debate about whether PCOS is a discrete entity, with phenotypically diverse pathophysiology, or is at least two, and maybe three, pathophysiologically and genetically distinct entities, each with different molecular and developmental origins, and likely requiring different clinical interventions . Future research in this area is vital to provide more tailored and effective clinical management of PCOS, and perhaps effective therapeutic or procedural cures.

Hyperandrogenism in PCOS redefined

Hyperandrogenism is found in > 80% of women with PCOS and is the most commonly inherited PCOS trait . In women with PCOS, testosterone (T), a conventional biopotent androgen produced in the ovaries, adrenals, and multiple somatic organs and tissues, including adipose, skin, and hair follicles, provides the highest circulating levels of conventional biopotent androgens , notwithstanding the even higher circulating levels of one of its weakly bipotent precursors, androstenedione (A ₄ ) . While the ovary has long been identified as the primary source of conventional androgen excess in women with PCOS , a subgroup of PCOS women also exhibits adrenal androgen excess . The adrenal cortex is also the major source of circulating 11-oxygenated metabolites of T and A _4, contributing T-comparable circulating levels of the equally bipotent 11-ketotestosterone (11-ketoT) from subsequent peripheral tissue metabolism . Circulating and peripheral tissue concentrations of 11-ketoT are elevated in women with PCOS , as well as in hyperandrogenic or obese adolescent girls , likely due to increased expression of the steroidogenic enzyme aldo-keto reductase family 1 member C3 (AKR1C3, or 17beta-hydroxysteroid dehydrogenase type 5). 11-ketoT, moreover, is a nonaromatizable androgen, has low affinity for SHBG, and crosses the placenta barrier to the fetus . Thus, both conventional and 11-oxygenated androgens contribute to bioeffective hyperandrogenism in women with PCOS.