Biomarkers in Polycystic Ovary Syndrome

Abstract

Polycystic Ovary Syndrome (PCOS) is the most common endocrine disorder in reproductive-age women. PCOS is diagnosed by the presence of two of the following three characteristics: hyperandrogenemia and/or hyperandrogenism, oligo/amenorrhea, and polycystic ovarian morphology. PCOS is associated with reproductive and non-reproductive complications, including obesity, insulin resistance and diabetes, dyslipidemia, and increased blood pressure. There is an urgent need for biomarkers that address both the reproductive and non-reproductive aspects of this complex syndrome. This review focuses on biomarkers, or potential ones, associated with the reproductive and non-reproductive aspects of PCOS, including anthropometric and clinical biomarkers, insulin and the IGF-1 system, lipids, anti-Müllerian hormone and gonadotropins, steroids, inflammatory and renal injury biomarkers, oxidative stress, and non-coding RNAs. We expect that this review will bring some light on the recent updates in the field and encourage researchers to join the exciting and promising field of PCOS biomarkers.

Introduction

Polycystic Ovary Syndrome (PCOS) is the most common endocrine disorder in reproductive-age women with a prevalence of 5-15% [1, 2]. The etiology of the syndrome remains unknown, although several genome-wide association studies (GWAS) have recently pinpointed several candidate genes in this complex disease. There are three different sets of diagnostic criteria available to diagnose PCOS (Fig. 1). The Rotterdam criteria (2003), the most frequently used one, requires the presence of two of the following three characteristics to be present: hyperandrogenemia and/or hyperandrogenism, oligo/amenorrhea, and polycystic ovarian morphology [3]. Other less frequently used PCOS diagnosis criteria are the National Institute of Health (NIH, 1990) [4], which requires the presence of hyperandrogenemia and/or hyperandrogenism and oligo/amenorrhea, and the Androgen Excess PCOSSociety (2006) [5] which requires the presence of hyperandrogenemia and/or hyperandrogenism, and either oligo/amenorrhea or polycystic ovarian morphology. More recently, in 2012, the NIH sponsored an evidence-based methodology workshop on PCOS that endorsed the inclusionary Rotterdam criteria with the caveat that the PCOS phenotype, one of the four associated with the Rotterdam criteria, should be specified [6] (Fig. 2). Moreover, in all of the aforementioned diagnosis criteria, other infrequent etiologies of androgen excess must be excluded for the diagnosis of PCOS.

Figure 1.

Polycystic Ovary Syndrome (PCOS) diagnosis criteria.

Figure 2.

Polycystic Ovary Syndrome (PCOS) phenotypes from the Rotterdam diagnosis criteria.

PCOS is associated with reproductive and non-reproductive complications, including obesity, insulin resistance and diabetes, dyslipidemia, and increased blood pressure. There is an urgent need for biomarkers that address both the reproductive and non-reproductive aspects of this complex syndrome.

A biomarker is a biological characteristic that is objectively measured and evaluated as an indicator of normal biological or pathological processes, or a response to a therapeutic intervention. Biomarkers are critical for the diagnosis, stratification, prognosis prediction, progression, regression, or outcome of disease treatment. In this review, we focus on some recent developments in the field of biomarkers in PCOS.

Anthropometric and clinical biomarkers

Body mass index (BMI) is a commonly accepted first-level measure of body fat and used as a screening tool for assessing excess adiposity at the population level. However, modern imaging techniques have shown that BMI has limited predictive value for estimating body fat and lean mass on an individual level [7]. Visceral adipose index (VAI) and lipid accumulation product (LAP) are surrogate biomarkers of adipose tissue function and distribution which are highly correlated with cardiovascular risk factors. VAI is based on both anthropometric [BMI and waist circumference (WC)] and functional [triglycerides (TG) and high-density lipoprotein cholesterol (HDL-cholesterol)] parameters, while LAP is an index based on WC and TG. LAP and VAI have been shown to be positively correlated with bioavailable androgens, fasting glucose, TG, and the homeostatic model assessment for insulin resistance (HOMA-IR) and negatively correlated with HDL-cholesterol in lean women with PCOS [8]. Moreover, LAP and VAI are better predictors of insulin resistance in both obese and lean women with PCOS [9]. Reliable and affordable screening biomarkers of insulin resistance for lean women with PCOS are critical for addressing cardiometabolic dysregulations in this PCOS population. In a comparison of biomarkers predictive of metabolic syndrome across the four PCOS phenotypes derived from the Rotterdam criteria, VAI was the only independent predictor across all PCOS phenotypes [10]. Moreover, LAP and VAI are elevated in young women with PCOS with or without hyperandrogenemia when compared with controls, and both are superior predictors of metabolic syndrome [11]. Additionally, LAP is predictive of insulin resistance in women with classic PCOS (phenotypes A and B), while VAI is superior for women with ovulatory PCOS (phenotype C) [12]. LAP and VAI, alone or in combination with other biomarkers, could help broaden our repertoire of biomarkers to address endocrine and metabolic dysregulation across the multiple PCOS phenotypes.

Neck circumference is another anthropometric biomarker that is associated with decreased insulin sensitivity and sex hormone binding globulin (SHBG), and increased free androgen index and PCOS prevalence in severely obese premenopausal women [13]. More recently, neck circumference was reported to be positively correlated with BMI, WC, waist-to-hip ratio, and insulin resistance among women with PCOS [14]. Anthropometric and clinical biomarkers, or a combination of both, are minimally invasive and affordable biomarkers that could advance the diagnosis and prognosis of women with PCOS.

Insulin and IGF-1 system biomarkers

Of primary importance in the metabolic consequences of PCOS, insulin resistance and hyperinsulinemia are key factors. Although insulin resistance is exacerbated by obesity in PCOS, the two should be considered independent factors. In a systematic review, women with PCOS have a 27% lower insulin sensitivity than healthy women [15]. Importantly, BMI independently exacerbates this reduction by an additional 15%, and this relationship was consistent using two different diagnostic criteria [15]. Although the etiology of insulin resistance in PCOS is unknown, several factors during puberty bring insights into its beginnings. Both growth hormone and insulin-like growth factor 1 (IGF-1) increase during puberty; ovarian receptors for both insulin and IGF-1 stimulate androgen production. IGF binding proteins (IGFBPs) modulate the availability of circulating players of the IGF-1 system; IGFBP-1 improves insulin sensitivity, lowers blood pressure, and protects against atherosclerosis [16]. Women with PCOS have decreased circulating levels of IGFBP-1, which has an inverse relationship with both insulin and free IGF-I [17]. Although IGFBP-1 may not serve as a direct biomarker of insulin resistance in PCOS, it may itself be a biomarker of abnormalities in the IGF-1 system, which are worsened with obesity, and can directly contribute to hyperandrogenemia and LH release in PCOS.

Lipid biomarkers

Dyslipidemia is a well-recognized cardiometabolic feature of PCOS, which includes fairly consistent lipid abnormalities including increased fasting triglycerides and decreased HDL-cholesterol [18]. Atherogenic dyslipidemia extends this definition to include elevated total ApoB lipoprotein remnants. ApoB lipoprotein remnants, specifically ApoB48-lipoprotein derived from the liver and small intestine, are 2-fold higher in adolescent women with PCOS and positively correlated with triglycerides [19]. Alternatively, ceramides, part of the sphingolipid family, are important plasma lipoproteins that make up the cell membrane structure. Recent studies in women with PCOS diagnosed by the Rotterdam criteria [20] and a hyperandrogenemic rodent model of PCOS [21] have shown similar increases in ceramides, which may serve as a previously unappreciated diagnostic biomarker of metabolic dysfunction particularly in hyperandrogenic women with PCOS.

Anti-Müllerian hormone and gonadotropins biomarkers

Gonadotropin-releasing hormone (GnRH) neurons in the hypothalamus ultimately govern output of gonadotropins follicle-stimulating hormone (FSH) and luteinizing hormone (LH) which control ovarian hormones, including androgens. Because patterns of the GnRH pulse secretion cannot be directly monitored in humans, LH can be used as a proxy measurement [22]. An increase in LH pulse frequency is associated with hyperandrogenemia in women with PCOS, and is an early signature of neuroendocrine dysfunction that can be observed in adolescent girls with PCOS [23]. Impairment of steroid negative feedback in PCOS further adds to the GnRH pulse frequency observed. Similarly, anti-Müllerian hormone (AMH) has actions throughout the hypothalamic-pituitary-gonadal axis. In females, AMH is secreted by the ovaries and is important in follicular development; AMH is approximately three times higher in women with PCOS. A recent study has shown that serum AMH, which reflects the number of developing ovarian follicles, closely correlates with polycystic ovarian morphology (PCOM) suggesting that serum AMH could replace PCOM determination by ultrasound in PCOS diagnosis [24]. In addition, AMH has signaling roles in GnRH neurons, acting centrally to increase GnRH activity and modulate LH [25]. This impairment, already observed in adolescents, can exacerbate excessive androgen production to further worsen the PCOS phenotype.

Steroid biomarkers

There are multiple factors that contribute to the different PCOS phenotypes, and each subtype may include different hormonal profiles; however an overwhelming clinical feature of the syndrome is high circulating androgens in about 80% of women [26]. Free androgen index (FAI), a ratio of total testosterone to SHBG, is one of the most commonly clinically used surrogate biomarkers of free testosterone. FAI is significantly elevated in women with PCOS in multiple ethnicities, as shown in a recent meta-analysis [27]. Androgens in women with PCOS should be measured preferably by liquid chromatography-mass spectrometry (LC-MS) or by high-quality assays to overcome the lack of sensitivity and/or specificity of other commonly used, but unreliable, assays [28]. The testosterone to dihydrotestosterone ratio is higher in women with PCOS with metabolic syndrome, and may indicate higher peripheral synthesis of androgens in the more severe metabolic phenotypes [29]. In a population of adolescent PCOS girls, the odds of metabolic syndrome increase approximately five times for every quartile increase in testosterone, even after adjustment for insulin resistance and obesity [30]. This highlights the impact of hyperandrogenemia on early cardiovascular risk, independent of other metabolic derangements. There has also been recent attention given to the clinical importance of 11-oxygenated androgens, as adolescent females with PCOS have higher 11β-hydroxyandrostenedione (11OHA4) and 11β-hydroxytestosterone (11OHT), even after controlling for obesity [31]. Several 11-oxygenated androgens correlate with the severity of hirsutism, which is one of the most common symptoms and concerns for women with PCOS.

SHBG, a glycoprotein that binds to and decreases the bioavailability of androgens in humans, is both consistently lower in women with PCOS and predictive of the disease [32]. Low SHBG in PCOS is associated with metabolic co-morbidities including insulin resistance and obesity, which can lead to increased risk of cardiovascular disease [32].

The anogenital distance (AGD) is the distance measured from the anus to the genital tubercle. Clinical and preclinical studies have shown that AGD is a reliable biomarker of prenatal androgen exposure. Women with PCOS have increased AGD compared with controls; furthermore, women with AGD in the upper AGD tertile have a several-fold increased prevalence of PCOS compared with women in the lower AGD tertile [33]. Moreover, newborn daughters of women with PCOS have increased AGD [34]. Those studies highlight that the anthropometric biomarker AGD, which reflects prenatal hormonal imbalance exposure, could be critical for early diagnosis and treatment of PCOS.

Inflammation biomarkers

PCOS is usually associated with a state of chronic low-grade inflammation and an altered immune cell profile [35, 36]. This inflammatory state is proposed to be involved in the pathophysiology of both PCOS-associated ovarian dysfunction and long-term metabolic and cardiovascular complications such as obesity, cardiovascular disease, insulin resistance, and diabetes [37]. These co-morbidities are associated with increased circulating inflammatory cytokines [38]. Multiple inflammatory and immune cell infiltration biomarkers are differentially modulated in PCOS. Compared to healthy women, women with PCOS have elevated pro-inflammatory cytokines such as TNF-α, IL-6, IL-8, IL-18, and IL-33, while they have lower levels of the anti-inflammatory cytokine IL-10 [39]. C-reactive protein (CRP), an acute-phase protein produced by the liver in response to inflammatory cytokines such as TNF-α and IL-6 following tissue injury, predicts systemic inflammation. Circulating CRP levels are elevated in PCOS patients and are associated with an increased risk of type 2 diabetes mellitus and cardiovascular disease in these patients [40]. According to a recent study, the two biomarkers associated with immune cell inflammation, HD domain containing 3 (HDDC3) and syndecan 2 (SDC2), are significantly downregulated in granulosa cells from women with PCOS and may be novel biomarkers for the diagnosis of PCOS [36].

Renal injury biomarkers

Women with PCOS present with albuminuria, an early indicator of renal injury [41]. In women with PCOS, serum testosterone positively correlates with urinary levels of the renal injury biomarkers KapU, LamU, α1-MU, and β2-MU) [41]. As a result, these biomarkers may be an effective diagnostic tool for renal injury in PCOS. The kidney injury molecule-1 (KIM-1) is a novel immune factor in kidney pathophysiology, with urinary levels upregulated in many acute and chronic kidney diseases and involvement in the progression of kidney fibrosis [42]. Urinary KIM-1 is a biomarker of early renal injury that predicts long-term renal outcome [42]. A preclinical study reported that KIM-1 urinary excretion is higher in aging hyperandrogenemic female rats than in placebo controls [43]. The use of these biomarkers to detect kidney injury in PCOS is uncommon, but they could undoubtedly be used for prognosis and patient monitoring to prevent the development of PCOS renal complications later in life. Urinary angiotensin converting enzyme 2 (ACE2) levels are higher in patients with chronic kidney disease than in healthy controls, particularly those with diabetic nephropathy [44]. We recently reported an increase in urinary ACE2 in a hyperandrogenemic mouse model of PCOS, implying that urinary ACE2 could be an early novel biomarker to predict renal injury in PCOS independent of the diet [45, 46].

Oxidative stress biomarkers

Oxidative stress is defined as an imbalance between oxidants and antioxidants caused by the excessive production of reactive oxygen species (ROS). Circulating biomarkers of oxidative stress that are elevated in PCOS are associated with an increased risk for several cardiovascular risk factors and may be involved in the pathogenesis of PCOS [47]. Independent of obesity, circulatory oxidative stress biomarkers such as xanthine oxidase, homocysteine, malondialdehyde, and asymmetric dimethylarginine are increased, while levels of anti-oxidative biomarkers such as glutathione, vitamin E, vitamin C, and paraoxonase-1 activity are decreased in women with PCOS [48, 49]. Lipid peroxidation is a well-known mechanism of cellular injury caused by oxidative stress-induced damage. In human studies, the prostaglandin-like molecule isoprostane, which is generated by free radical-induced peroxidation of arachidonic acid, has been proposed as the gold standard for assessing lipid peroxidation in vivo [50]. Circulating 8-isoprostanes levels are two-fold higher in women with PCOS than in controls and are associated with obesity, hyperlipidemia, and insulin resistance [47]. Moreover, women with PCOS have higher follicular fluid 8-isoprostane levels that are associated with spontaneous abortions in pregnant women with PCOS [51]. Therefore, due to its high sensitivity and chemical stability, 8-isoprostanes may be a reliable, quantifiable biomarker of oxidative stress in PCOS which may also be useful in predicting a higher risk of miscarriage in this population [52].

Non-coding RNAs biomarkers

microRNAs are small non-coding RNAs that regulate protein levels of specific genes by translational repression or mRNA degradation. There are multiple microRNAs that have been reported to be regulated in PCOS or have a significant role in PCOS pathology [53, 54]. Moreover, the family of non-coding RNAs has recently expanded beyond microRNAs to include other regulatory RNAs, such as long non-coding RNAs and circular RNAs, which have also been reported to have a role in PCOS pathology [55–57]. Despite intensive research in the field of microRNAs, or more broadly non-coding RNAs, in PCOS, this is still and emerging field in which most proposed biomarkers need further research before their translation into the clinic. Nevertheless, non-coding RNAs in PCOS is a rapidly evolving field worth to be closely watched and we refer the readers to some of the excellent recent reviews in the field mentioned above.

Conclusions

PCOS is a highly prevalent disorder in women, associated with reproductive and non-reproductive complications. In this review, we present a series of potential biomarkers of PCOS for the diagnosis, stratification, prognosis prediction, progression, regression, or outcome of treatment (Fig. 3). The field of biomarkers in PCOS is rapidly evolving, continuously bringing new developments for this complex and challenging disease. Unfortunately, there are multiple useful, minimally invasive, and relatively inexpensive biomarkers that have been widely adopted in clinical research, but have not yet been incorporated into clinical use. We expect that this review will encourage researchers to join the promising field of biomarkers in PCOS.

Figure 3.

Potential novel biomarkers in Polycystic Ovary Syndrome (PCOS).

Abbreviations

11OHA4

11β-hydroxyandrostenedione

11OHT

11β-hydroxytestosterone

ACE2

angiotensin converting enzyme 2

AGD

anogenital distance

AMH

anti-Müllerian hormone

BMI

body mass index

CRP

C-reactive protein

FSH

stimulating hormone

GnRH

gonadotropin-releasing hormone

GWAS

genomewide association studies

HDDC3

HD domain containing 3

HDL-cholesterol

high-density lipoprotein cholesterol

HOMA-IR

Homeostatic model assessment for insulin resistance

IGFBPs

IGF binding proteins

IGF-1

insulin-like growth factor 1

KIM-1

kidney injury molecule-1

LAP

lipid accumulation product

LH

luteinizing hormone

PCOM

polycystic ovarian morphology

PCOS

Polycystic ovary syndrome

ROS

reactive oxygen species

SDC2

syndecan 2

SHBG

sex hormone binding globulin

TG

triglycerides

VAI

visceral adipose index

WC

waist circumference