Systematic review: association of polycystic ovary syndrome with metabolic syndrome and non-alcoholic fatty liver disease

Authors

  • A. Baranova,

    1. Betty and Guy Beatty Center for Integrated Research, Inova Health System, Falls Church, VA, USA.
    2. Center for the Study of Genomics in Liver Diseases, Molecular and Microbiology Department, George Mason University, Fairfax, VA, USA.
    Search for more papers by this author
  • T. P. Tran,

    1. Betty and Guy Beatty Center for Integrated Research, Inova Health System, Falls Church, VA, USA.
    2. Center for the Study of Genomics in Liver Diseases, Molecular and Microbiology Department, George Mason University, Fairfax, VA, USA.
    Search for more papers by this author
  • A. Birerdinc,

    1. Betty and Guy Beatty Center for Integrated Research, Inova Health System, Falls Church, VA, USA.
    2. Center for the Study of Genomics in Liver Diseases, Molecular and Microbiology Department, George Mason University, Fairfax, VA, USA.
    Search for more papers by this author
  • Z. M. Younossi

    1. Center for the Study of Genomics in Liver Diseases, Molecular and Microbiology Department, George Mason University, Fairfax, VA, USA.
    2. Center for Liver Diseases, Inova Fairfax Hospital, Falls Church, VA, USA.
    Search for more papers by this author

  • This uncommissioned systematic review was subject to full peer-review.

Dr Z. M. Younossi, Betty and Guy Beatty Center for Integrated Research, Inova Health System, 3300 Gallows Road, Falls Church, VA 22042, USA.
E-mail: zobair.younossi@inova.org

Abstract

Aliment Pharmacol Ther 2011; 33: 801–814

Summary

Background  Polycystic ovary syndrome (PCOS) is a common disorder for women of child-bearing age and is associated with metabolic syndrome (MS).

Aim  To assess the literature for associations between polycystic ovary syndrome and non-alcholic fatty liver disease (NAFLD).

Methods  We performed a systematic review using PubMed-search for peer-reviewed articles related to polycystic ovary syndrome and NAFLD. Articles were summarised and grouped according to different sections defining interactions of polycystic ovary syndrome with metabolic syndrome and non-alcholic fatty liver disease as well as risk factors, pathogenic pathways and treatment options.

Results  Obesity is a common factor involved in both polycystic ovary syndrome and non-alcholic fatty liver disease. Obesity causes non-alcholic fatty liver disease and aggravates hirsutism and menstrual disorders in polycystic ovary syndrome. Insulin resistance, a hallmark of metabolic syndrome is observed in 50–80% of women with polycystic ovary syndrome and patients with non-alcholic fatty liver disease. Recent findings suggest that women with polycystic ovary syndrome may be at risk for developing non-alcholic fatty liver disease and conversely, non-alcholic fatty liver disease may be a risk for polycystic ovary syndrome. Based on the association of polycystic ovary syndrome and other metabolic abnormalities, such as insulin resistance, hyperandrogenism, obesity and non-alcholic fatty liver disease, the candidate genes have been speculated for polycystic ovary syndrome. Closer scrutiny of these genes placed most of their proteins at the crossroads of three highly inter-related conditions: metabolic syndrome, obesity and non-alcholic fatty liver disease. In most studies, the prevalence of both polycystic ovary syndrome and non-alcholic fatty liver disease rises proportionally to the degree of insulin resistance and increases in the mass of adipose tissue.

Conclusions  Non-alcholic fatty liver disease is considered as the hepatic manifestation of metabolic syndrome. Similarly, it seems appropriate to consider polycystic ovary syndrome as the ovarian manifestation of metabolic syndrome. Both these conditions can co-exist and may respond to similar therapeutic strategies.

Introduction

One of the most common disorders for women at child-bearing age is polycystic ovary syndrome (PCOS), which affects not only hormones that regulate the normal development of eggs in the ovaries but also other metabolic pathways.1 The prevalence of this disorder is estimated at 5% to 10% of women at reproductive age. On the other hand, the true prevalence rates for PCOS could be even higher with the use of the recently adopted Rotterdam criteria, increasing this rate, in community settings, to around 18%.2 Importantly, PCOS has been noted to affect 28% of obese women, but only 5% of lean women.2, 3

According to the Rotterdam criteria, a positive diagnosis of PCOS can be made if at least two of the following criteria are met: chronic oligo-ovulation, clinical or biochemical hyperandrogenism and ultrasonic evidence of polycystic ovaries.1 The cause of PCOS is currently unknown; therefore no preventive measures can be implemented. PCOS is thought to be initiated by low levels of follicle stimulating hormone (FSH) and high levels of androgen, which are detected in the majority of cases. Hyperandrogenism causes the follicles in the ovaries to develop poorly. Consequently, the eggs in these follicles cannot mature and give rise to the cysts seen in patients with PCOS.1

Methods

We performed a systematic review using Medline search (1985–2010) for peer-reviewed articles related to PCOS and NAFLD. The articles were summarised and grouped according to different sections defining interactions of PCOS with metabolic syndrome and NAFLD. In addition, articles dealing with common risk factors for both PCOS and NAFLD were summarised. Finally, articles related to pathogenic pathways and treatment options are reviewed.

Results

Information retrieved from reviewed articles is summarised in the following sections.

PCOS and metabolic syndrome

Various epidemiological studies have established a close association between body fat mass and an increased risk of developing a variety of chronic disorders including cardiovascular diseases and non-alcoholic fatty liver disease (NAFLD). The excessive body fat, especially visceral fat, contributes to the development of a complex network of potentially serious clinical conditions such as insulin resistance, glucose intolerance, dyslipidaemia, elevated blood pressure, impaired fibrinolysis and endothelial dysfunction.4 This constellation of risk factors are recognised as components of what is now called Metabolic Syndrome (MS).4, 5 The most common description of MS includes visceral obesity, insulin resistance, dyslipidaemia and hypertension.6 The pathogenetic mechanism of MS resides within a triangle of the following factors: (i) insulin resistance, (ii) hyperinsulinemia and (iii) glucose intolerance, which are mainly due to defect(s) in the insulin signal transduction pathways.

Along with the epidemic of obesity, the prevalence of MS is increasing worldwide, both in the developing and developed countries. As noted previously, MS is associated with a risk of cardiovascular disease and is a common early abnormality in the development of type 2 diabetes. In addition, MS plays a well-recognised role in the development of obstructive sleep apnoea, erectile dysfunction, polycystic ovary syndrome and malignant tumours.

Insulin resistance (IR), a hallmark of metabolic syndrome, is observed in about 50% to 80% of women with PCOS.7 In PCOS, IR is more than a biomarker of the disease, but is rather an active contributor to its pathogenesis. Insulin receptors are abundant in ovaries; dysregulation of insulin signalling in theca cells augments the production of androgens. Obesity aggravates the clinical presentation of PCOS. In fact, the prevalence of hirsutism and menstrual disorders is greater in the obese as compared to non-obese PCOS subjects.8 However, even in lean women, PCOS is often accompanied by abnormalities of insulin secretion and higher basal blood glucose than weight-matched controls.9, 10 In cases when lean women with PCOS maintain normal sensitivity to insulin, adiponectin biosynthesis deficiency11 or chronic low level inflammation is often present.12

The molecular underpinnings of insulin resistance (IR) in PCOS are related to a variety of defects, including postbinding receptor failure13 and insulin signalling defect at the level of glucose transport in skeletal muscle.14 Specifically, the cause of IR in skeletal muscle might be due to low levels of Insulin receptor substrate 1 (IRS-1) expression, impaired IRS-1 phosphorylation, reduced activity of the serine/threonine kinase AKT2 and altered glucose transporter GLUT4 translocation to the plasma membrane.15 However, some studies have shown that muscular IR in PCOS results from both intrinsic factors (genetically determined defects in insulin signalling) and extrinsic conditions pertinent to environmental exposures (obesity, ovarian dysfunction).15

PCOS and non-alcoholic fatty liver disease

Non-alcoholic fatty liver disease (NAFLD) is now considered to be one of the most common forms of chronic liver disease in the Western world. It occurs in an estimated 25% to 30% of the US general population, whereas its potentially progressive form, non-alcoholic Steatohepatitis (NASH) is reported in 2–3% of the population. Although the most common explanation for the increased prevalence of NAFLD is the increased prevalence of obesity, the risk of developing NAFLD and NASH is not limited to the overweight and obese individuals.

Non-alcoholic fatty liver disease refers to a clinico-pathologic spectrum of conditions ranging from simple steatosis (simple fatty liver) to NASH, involving inflammation and some evidence of liver cell damage, and in some cases, cirrhosis, which is advanced scarring of the liver.16 From the liver disease standpoint, simple, or bland, steatosis is a relatively benign condition. Its clinical manifestations are usually absent or subtle and it usually comes to medical attention incidentally when aminotransferase levels are found to be elevated or a radiographic study reveals that the liver is fatty.17 It is often noted that simple steatosis rarely progresses to NASH and may even be reversible. A few recent longitudinal studies with paired liver biopsies showed that the progression to NASH may occur in a proportion of patients with bland steatosis.18, 19 NASH, in turn, can potentially progress to cirrhosis, decompensated liver disease and hepatocellular carcinoma. The long-term follow-up evaluation of NAFLD patients revealed that NASH patients have increased liver-related mortality compared with non-NASH patients. In addition, patients with both NAFLD and type II diabetes are especially at risk for liver-related mortality.20, 21

Insulin resistance (IR) is the key event linking NAFLD to MS.22 The epidemiology, pathogenesis and approach to treatment of NAFLD follow the same trends as all other metabolic disorders.23, 24 Clinical features of the metabolic syndrome (obesity, diabetes mellitus, or hypertriglyceridemia) are commonly observed in NAFLD. Moreover, primary NAFLD is now considered the hepatic manifestation of metabolic syndrome.24 NAFLD is an early predictor of metabolic disorders in general, particularly in the normal-weight population.25

Although PCOS and NAFLD do share a common attribute in regards to their pathogenesis, insulin resistance, the link between the two diseases has not been obvious until the first case was described in 2005. A young female patient with PCOS, at an out-patient Gastroenterology Clinic, was found to have both insulin resistance and severe NASH in her liver biopsy.26 This interesting case opened a new path into the investigation of the association of PCOS and NAFLD through their link to increased risk of metabolic disorders, such as insulin resistance and abdominal adiposity.27 Elevated alanine aminotransferase (ALT) serum levels are a common finding in PCOS.27 Moreover, in PCOS women with abnormal ALT, insulin sensitivity is markedly decreased (P < 0.001). One study found that 55% (48/88) of PCOS women had both hepatic steatosis and high HOMA-IR scores (P = 0.033).28

Another study found that 41% (17/41) of women with PCOS had concomitant NAFLD as diagnosed by hepatic steatosis and abnormal ALT levels, whereas the incidence of NAFLD in the weight and age matched non-PCOS control group was only 19% (6/31, P < 0.05).29 Some studies have also shown that for obese PCOS patients, the rates of co-diagnoses with NAFLD or NASH are even higher.30 Even stronger associations between NAFLD and PCOS has been described when screening for PCOS was performed in the NAFLD cohort. A recently published study of 14 NAFLD female patients of reproductive age (22–45) revealed that 71% (10/14) of these patients matched the 2003 Rotterdam diagnosis criteria for PCOS.31 Nevertheless, it should be noted that in the cohort of young and lean PCOS patients, no evidence of NAFLD were found.32 Despite its important contribution to the literature, this study lacked follow-up and therefore, the question of whether PCOS patients are predisposed to NAFLD at older age remains open.

Although insulin resistance is primarily associated with obesity, clinical evidence has revealed that insulin resistance exists in both obese and lean PCOS patients.27 Importantly, when insulin resistance was studied in young, lean individuals, net muscle glycogen synthesis in the group with IR was found to be lower by 61% as compared with the age-BMI matched, insulin sensitive control group. On the other hand, net hepatic triglyceride synthesis was about 2.5-fold greater in the IR group as compared to the insulin sensitive group. Moreover, hepatic de novo triglycerides lipogenesis was increased by 2.2-fold in the IR group (15.7 ± 1.5%) compared with the insulin sensitive group (7.2 ± 0.7%, P = 0.00005).33 These data suggested that in young, lean insulin sensitive individuals, energy is stored mostly in the muscle and liver glycogen, whereas in young lean insulin-resistant individuals, the energy is mainly diverted from muscle glycogen synthesis into liver triglyceride synthesis, resulting in increased fat production in the liver and leading to hepatic steatosis. This finding also suggests that skeletal muscle’s insulin resistance, which is common in PCOS patients, predisposes them to having hepatic IR, which is strongly linked to non-alcoholic fatty liver disease (NAFLD).34

The above findings suggest that women with PCOS probably are at an increased risk for developing NAFLD and conversely, women with NAFLD may be at risk for having PCOS. In addition, it is reasonable to propose that women with central obesity are at a higher risk for both NAFLD and PCOS. However, for reasons yet unknown, some women develop obesity and never develop PCOS, whereas others develop obesity and then develop PCOS and/or NAFLD. Below, we will review the molecular genetics and environmental interactions in PCOS and try to dissect these facets for possible interplay between PCOS, NAFLD and metabolic syndrome (Figure 1).

Figure 1.

 An interplay between various components of the PCOS phenotype.

The role for environment factors in PCOS and NAFLD

The prevalence of PCOS is increased in obese as compared to lean women (30% vs. 10%). The strongest contributing factors to the phenotypes of obesity and insulin resistance are poorly balanced diet and physical inactivity. Although an increasing number of studies stress the very strong association between PCOS, insulin resistance and obesity as well as co-association of insulin resistance and obesity, these conditions, to some degree, remain independent from each other.35 Some obese patients are insulin sensitive and there are lean patients who have substantial resistance to insulin. However, being overweight with a body mass index (BMI) greater than 25 kg/m2, especially with visceral fat deposits, substantially increases the risk of insulin resistance, metabolic disorders and reproductive abnormalities consistent with a PCOS phenotype. Similar to having PCOS, being overweight or obese can predispose one to NAFLD. Depending on the particular diagnostic criteria, the prevalence of NAFLD in obese cohorts is estimated to be between 60% and 90%.36 Once excessive adiposity is established, the more profound the associated metabolic abnormalities the more likelihood of patients having the progressive form of NAFLD or non-alcoholic steatohepatitis (NASH).37

Molecular genetics of PCOS

Until recently, the molecular genetics of PCOS have been poorly understood. Pedigree studies indicate that the inheritance mode of PCOS is autosomal dominant: the disorder is transmitted to both sons and daughters, but the clinical phenotype only occurs in women.38 Incidentally, women from PCOS families have been shown to have both insulin resistance and high levels of androstenedione in their blood.39 Other factors such as diet, lifestyle and medications can contribute to the progression of the disorder as well.40, 41 By using several molecular methods, such as cDNA microarray analysis, linkage studies and mutation analysis as well as case–control association studies, some genes associated with PCOS have been identified. However, dysregulation of one of these candidate genes alone is not sufficient to cause PCOS. As the pathogenesis of PCOS usually involves multiple pathways, ranging from insulin synthesis, androgen hormone production, follicle development and obesity; so far, studies of PCOS have focused on candidate genes from these pathways. Among the genes, most often investigated in the study of PCOS are the steroid biosynthesis-related cytochrome genes CYP11A and CYP17, FEM1A gene encoding for ischaemia-related mitochondrial protein,42 the obesity associated gene (FTO);43 several genes involved in the insulin pathway (AKT2 and INSR), in leukotriene metabolism (ALOX15),44 and the androgen signalling pathway (SGTA).41 In addition, some other genes were found to be differentially expressed in relevant tissues of the PCOS patients, including ones encoding for insulin signalling components (IRS-1, pIRS-1Y612, GLUT4), sex hormone-binding globulin (SHBG), adiponectin45 and inflammatory cytokines (IL-6, IL-18, TNFα).42, 46 Functions of PCOS related genes are summarised in Table 1.

Table 1.   Genes evidently participating in PCOS pathogenesis are also involved in NAFLD
Pathways involvedGene namesPCOSNAFLD
Steroid biosynthesis pathwayCYP11AElevates testosterone levelsSuppressed in response to fenofibrate treatment in parallel with alleviation of steatosis
Both CYP11a transcription and the development of the endocrine pancreas is regulated by GATA-6
CYP17Is elevated in ovaries and associated with IRIs attenuated by adiponectin, known anti-NAFLD/NASH factor
SRD5AProduces hyperandrogenism
Lowers cortisol/cortisone ratio
Expression us s sensitive to pioglitazone.
Liver fat accumulation is associated with lower cortisol/cortisone ratio
Androgen signalling pathwaySHBGLevels are decreased in PCOS, thus, increasing bioavailablity of androgensLevels are decreased in steatosis
SGTAPromotes apoptosisPromotes apoptosis, a hallmark of NASH
Obesity associated geneFTOInfluences both hyperandrogenemia and impairment of glucose tolerance, particularly, in muscle IRPredisposes to NAFLD (in PCOS population)
Leukotriene metabolism related genesALOX15Contributes to inflammation seen in PCOS;
augments the IR
When disrupted, prevents NAFLD in model animals
FEM1AMutated in some PCOS patientsParticipates in anti-inflammatory signalling of prostaglandin E2 that alleviates liver damage by ROS.
FEM1BVariants are associated with both reduced likelihood of PCOS and lower IRBinds SREBPs that are central effectors in NAFLD and NASH inflammatory mediators. Mediates apoptosis.
Adipokines and cytokinesIL-6, IL-18, hs-CRP, TNFα,TNFR2Elevated (especially in obese PCOS)Known to increased the risk of NAFLD
AdiponectinBoth total and HMW adiponectin levels are lowered in PCOSBoth total and HMW adiponectin levels are lowered in NAFLD/NASH
Insulin signalling pathwayInsulin (INS), INSR, IRS-1, IRS-2, AKT2, GLUT4An involvement of insulin signalling is well known pathogenetic cornerstone for both PCOS and NAFLD

Enzymes participating in the steroid biosynthesis pathway

In PCOS, an increased androgen production is dependent on the increased activity of three enzymes, two cytochromes – CYP11A and CYP17 – and Steroid 5-α Reductase SRD5A. As can be seen from observations outlines below, all three of these enzymes possess steatogenic or diabetogenic properties.

Gene CYP11A encodes a cholesterol side chain cleavage enzyme participating in the androgen biosynthesis by converting cholesterol to pregnenolone, the common steroid precursor. CYP11A contains a common polymorphism, specifically, a variable number of tandem repeats (VNTR) site that is associated with elevated serum testosterone levels in PCOS.47 In addition, the upregulation of GATA-6, which is involved in the CYP11a transcription, is also associated with PCOS.48 GATA-6 also regulates the development of the endocrine pancreas and interacts with Nkx2.2, a critical islet transcription factor.49 Interestingly, CYP11A is suppressed in response to fenofibrate, a hypolipidemic ligand for PPARα used in the treatment of hypertriglyceridemia and combined hyperlipidemia, being particularly effective in lowering the plasma triglyceride and cholesterol levels.50 Incidentally, fenofibrate also alleviates steatosis in Zucker diabetic fatty (ZDF) rats, while not changing either energy intake or expenditure or the progression of diabetes.51

The levels of CYP17 mRNA encoding for 17-α hydroxylase, yet another steroidogenic enzyme, are increased in ovaries of PCOS women.52 This increase in expression may be caused by the activity of many factors that affect the transcriptional and post-translational regulation of CYP17 in the thecal cells;53 one of these factors is insulin. Interestingly, obese PCOS carriers of hyperactive A2 allele of CYP17 show an odds ratio of 9.1 (confidence interval, 3.0–27.4; P < 0.0001) for developing insulin resistance.54 Adiponectin, generally a beneficial adipokine, typically deficient in obese and insulin-resistant individuals, decreases insulin-induced androstenedione production by attenuation of IGF-I-induced LHR, CYP11A1 and CYP17A1 gene expression in theca cells.55

Steroid 5-α reductase SRD5A encodes an enzyme that converts testosterone into the more potent androgen, dihydrotestosterone and, as well, reduces cortisol. The decrease of cortisol levels in the blood stimulates ACTH-dependent steroidogenesis and produces hyperandrogenism.56 In PCOS, an increase in activity of 5-alpha reductase in the liver, skin and follicles was observed.57 The levels of SRD5A mRNA are also elevated in patients with PCOS.58 The metabolic abnormalities frequently seen in PCOS patients are tightly linked to increased cortisol elimination. In particular, SRD5A activity correlated with BMI, insulin levels and HOMA scores.59 The activity of 5-αalpha reductase is sensitive to the antidiabetic drug pioglitazone60 and to weight loss.61 Even more interesting is that higher urinary excretion of 5α-reduced cortisol metabolites is associated with indices of obesity, and liver fat accumulation, a hallmark of NAFLD with a lowered ratio of cortisol/cortisone metabolites.62

Androgen signalling pathway: small glutamine-rich tetracopeptide repeat containing protein α gene and sex hormone-binding globulin

Sex hormone-binding globulin (SHBG) is a glycoprotein that binds to both testosterone and estradiol and influences the function of these hormones. The levels of SHBG have been shown to negatively correlate with insulin levels46 and even were suggested as a surrogate marker of insulin resistance. In PCOS patients, due to their increased insulin levels, the SHBG levels are reduced significantly, leading to the release of more androgens, which causes the characteristic hyperandrogenemia of the disorder.41, 63 Importantly, SHGBs levels are lower in IR PCOS vs. non-IR PCOS patients.63 Moreover, in patients with steatosis, SHBG levels are also decreased,27, 64 possibly reflecting known abnormalities associated with NAFLD that influence SHBG secretion by the liver, such as obesity, central adiposity and insulin resistance.

Small glutamine-rich tetracopeptide repeat containing protein α gene is a member of the androgen receptor-chaperone-co-chaperone complex in the androgen signalling pathway, which binds to the androgen receptor (AR) and restrains its activity by holding it in cytoplasm.65 Over expression of SGTA results in the inhibition of AR bioactivity whereas reduced SGTA expression increases AR activity and promiscuous activation of the AR by nonclassical ligands (e.g. progesterone).65 Goodarzi and colleagues investigated variants of the SGTA gene within the PCOS patient population and found that haplotype-1 was associated with PCOS risk and haplotype-2 was associated with increased insulin resistance, a feature of PCOS.42 Experimental evidence suggests that the SGTA gene may increase PCOS risk by promoting apoptosis through enhanced DNA fragmentation, chromosome misalignment and mitotic arrest.42, 66, 67 Importantly, enhanced apoptosis is a characteristic of both PCOS43, 68 and NAFLD, in particular, of NASH variety.43, 69

Fat mass and obesity associated gene

The fat mass and obesity associated gene (FTO) has been shown to be associated with obesity in humans in several genome-wide association scans. The relevance of FTO to PCOS is suggested by the high prevalence of obesity in PCOS, about 60% to 70%.70 The function of FTO was not known until recently, when a transcriptional coactivator activity of FTO was demonstrated for both unmethylated and methylation-inhibited CCAAT/enhancer binding proteins (C/EBPs)-dependent gene promoters.71 Consistent with its C/EBP enhancer function, mutation of FTO leads to temporal progressive loss of adipose tissue in experimental animals.72

Variants within the FTO gene influence both hyperandrogenemia and impairment of the glucose tolerance parameters in women with PCOS.73 Tan et al. determined that the FTO genetic variants appear to have a greater impact on obesity and related traits in PCOS than in other phenotypes.43 These and other74 studies demonstrated that the predisposition to common obesity also result in altered susceptibility to PCOS, confirming the mechanistic link between these conditions.

Importantly, FTO is equally involved in muscle IR. The expression of FTO in skeletal muscle from type 2 diabetic patients is elevated, but could be normalised by thiazolidinedione treatment.75 Given the link between muscle IR and liver IR/steatosis discussed above, FTO overexpression common in PCOS may contribute to the predisposition to NAFLD in this population.

Leukotriene metabolism related genes ALOX15, FEM1A, FEM1B

One of the studies of the skeletal muscles from PCOS women employed gene set enrichment analysis (GSEA) to get an insight into the early stages of IR and found the systemic IR-associated changes in the expression of genes involved in mitochondrial oxidative metabolism and evidence of abnormal lipid metabolism.76 12/15- Lipoxygenase ALOX15 is one of the lipid metabolism genes upregulated in the omental fat of PCOS patients.44 As a large number of lypoxygenase-oxidised fatty acids become leukotrienes, which are natural chemical substances in the body that promote inflammatory response to damage, an impaired lipoxygenase function can theoretically contribute to the inflammatory condition seen in PCOS.46

Incidentally, it was shown that the lipoxygenase encoded by ALOX15 augments the resistance to insulin, whereas its inhibition enhances the action of insulin in rat models of insulin resistance and type 2 diabetes.77 12/15 lipoxigenase deficient mice are resistant to streptozotocin-induced diabetes.78 Targeted deletion of ALOX15 also protects non-obese diabetic (NOD) mice from autoimmune diabetes.79 The antidiabetic action of ALOX15 deletion is explained by the failure to augment the production of IL-12 in macrophages and induce apoptosis in β-cells of the pancreas.79

Importantly, disruption of ALOX15 protects apolipoprotein E-deficient (ApoE−/−) mice against the development of NAFLD. These mice show reduced serum alanine aminotransferase levels; decreased hepatic steatosis, inflammation and macrophage infiltration; and decreased fatty acid synthase, TNF-α, monocyte chemoattractant protein-1 (MCP-1), interleukin (IL)-18 and IL-6 expression.80

The FEM1A gene is a homologue of fem-1 sex determination gene of C. elegans; it is highly expressed in human skeletal and cardiac muscle, brain, liver and in the ovaries; it localises within mitochondria. Mouse homologue of FEM1A is expressed in androgen-producing secondary interstitial cells, with a marked increase in expression after puberty, consistent with a key feature of PCOS – ovarian hyperandrogenism.81 In C. elegans, fem-1 determines the development of the male phenotype. Therefore, it is likely that human homologue of fem-1 contributes to the hyperandrogenic features of PCOS.81 Moreover, the FEM1A gene locates to chromosome 19p13.3 and it has been confirmed by several studies that the microsatellite D19S884 on chromosome 19p13.2 is a potential PCOS susceptibility locus.82 In a study performed on five Caucasian PCOS patients from Mississippi, a heterozygous germline missense mutation in a conserved amino acid within FEM1A, H500Y, was found.81

Interestingly, FEM1A was independently discovered in a yeast two-hybrid based screening for proteins that directly interact with the cytoplasmic tail of the EP4 receptor for prostaglandin E2 (PGE2) in human macrophages, where it participates in anti-inflammatory signalling of prostaglandin E2.83 PGE2 has a prominent antioxidant effects in the liver where it decreases the production of hydroxyl radicals, lipoperoxides, conjugated dienes, malonic dialdehyde and carbonyl-containing products of lipid peroxidation. In addition, PGE2 normalises activity of the liver microsomal monooxygenase system components responsible for the free oxygen radical production84 and alleviates liver damage in cirrhotic rats.85 In addition to its PGE2 augmenting action, FEM1A directly interacts with NF-kappaB1 p105/p50 in macrophages, and, in a concentration-dependent manner, inhibits NF-kappaB activation induced by various proinflammatory stimuli.86

FEM1B gene, an orthologue of the C. elegans feminization factor 1 (FEM-1) is a binding partner for PHTF1, a transcription factor encoding a gene identified as susceptibility factors for type 1 diabetes in humans.87 FEM1B is a proapoptotic protein that interacts with apoptosis-inducing proteins Fas, tumour necrosis factor receptor-1 (TNFR1) and apoptotic protease activating factor-1 (Apaf-1) and mediates proteasome inhibitor-induced apoptosis.88 It also can serve as an adaptor protein that links CHK1 and Rad9 thus facilitating checkpoint signalling induced by replication stress.89 FEM1B polymorphisms are associated with both a reduced likelihood of PCOS and lower insulin resistance.82 This fact is not surprising as Fem1b-knockout mice display abnormal glucose tolerance that is due predominantly to defective glucose-stimulated insulin secretion.90

Inflammatory mediators (hs-CRP, IL-6, IL-18)

As mentioned above, obese PCOS patients have high levels of proinflammatory mediators, such as high sensitive C-reactive protein (hs-CRP), interleukin 6 (IL-6) and IL-18, which make them vulnerable to an increased risk of non-alcoholic fatty liver disease (NAFLD).91 C-reactive protein (hs-CRP), one of the acute phase proteins known as a vascular inflammatory marker, rises dramatically in response to inflammation due to an increase in the circulating level of interleukin 6 (IL-6), which is produced mainly by macrophages and adipocytes. The level of hs-CRP is significantly elevated in PCOS patients.92 However, other studies reported that serum hs-CRP levels increase with obesity rather than with the presence of PCOS itself. Therefore, an elevation of hs-CRP in PCOS women is mostly related to their increased BMI.93

Other inflammatory mediators, such as interleukin 18 (IL-18), interleukin 6 (IL-6), tumour necrosis factor α (TNFα) and type 2 TNF receptor (TNFR2) are also reportedly elevated in PCOS,41, 94 and are strongly associated with central fat excess in PCOS patients.95 As both hs-CRP and IL-6 are considered strong risk markers for NASH and severity of liver fibrosis, these findings may point towards a mechanism for increased risk of NAFLD and NASH in PCOS patients.96, 97 Furthermore, the increased levels of TNFα induce insulin resistance is a common feature in NASH and PCOS.97, 98 All together, these observations indicate that multifaceted inflammatory processes may be the underlying causes of the high prevalence of type 2 diabetes and steatohepatitis seen in PCOS patients.

Adiponectin

Adiponectin, a peptide hormone, is secreted by adipocytes and is the only adipokine downregulated in obesity.45 Clinical studies have shown that PCOS patients, especially those with obesity, type 2 diabetes and insulin resistance have significantly decreased levels of adiponectin as compared to weight-matched controls.45 Importantly, the high-molecular weight (HMW) form of adiponectin that most closely correlates with insulin sensitivity was even more substantially downregulated in PCOS women than total adiponectin.98 The lower levels of total and, specifically, HMW adiponectin in PCOS patients may explain the very high risk for PCOS patients of being prone to metabolic syndrome and diseases of the NAFLD spectrum and point at a possibility of adipocyte dysfunction underlying the pathology of PCOS.

Insulin signalling pathway: INS, INRS, AKT2, GLUT4 and IRS genes

It has been known that hyperinsulinemia and insulin resistance are common features of PCOS patients with or without obesity. Hence, several genes involved in insulin secretion and action have been proposed as candidate genes in PCOS pathogenesis.41

Insulin, encoded by gene INS, is a hormone secreted by the β cells of the pancreas to maintain normal blood glucose levels by regulating cellular glucose uptake, carbohydrate, lipid and protein metabolism.99 In PCOS, these functions of insulin may be disturbed by defects in insulin signalling, resulting in insulin resistance and hyperinsulinemia. Through a family-based association study using DNA obtained from 1723 individuals in 412 families with 412 index cases and 43 PCOS sisters, Urbanek et al. found that there was an association between D19S884 allele 8 with higher fasting insulin levels and homoeostasis model assessment for IR (HOMA-IR)) in lean PCOS women BMI <25 kg/m².100 In addition to regulating blood glucose levels, insulin also stimulates the growth and replication of cells in the ovaries. With excess of insulin in the bloodstream, the cells of the ovaries are stimulated to overproduce androgens, which lead to the hyperandrogenism seen in PCOS.

The defects of INSR gene that encodes for the insulin receptor are an uncommon cause of diabetes; however, conditional knockouts of this gene in various murine tissues served as an indispensable tool in understanding insulin resistance, often producing surprising effects. For example, knockout of INSR in white adipose tissue protects mice against obesity, whereas its elimination in brown adipose lead to the development of β-cell failure (see101 for detailed analysis). By using microsatellite markers, analysis of linkage and family-based association, it was concluded that the INSR gene marker D19S884, which is located 1 cM telometric to the INSR gene, is significantly associated with PCOS (χ2 = 11.85; P < 0.0006).100, 102 Another study carried out by Siegel et al. found an association between a C/T single nucleotide polymorphism (SNP) at the tyrosine kinase domain of INSR with PCOS in lean patients.103

As compared to the insulin receptor itself, two of its substrate proteins, IRS-1 and IRS-2, critical to signal transduction in insulin target tissues, garnered larger attention as both PCOS and NAFLD candidates. For example, IRS-1 polymorphisms were associated with increased susceptibility to PCOS in many independent studies.104 One suggested mechanism linking IRS-1 and PCOS involves an co-stimulatory interaction between testosterone and insulin that promotes the serine phosphorylation of IRS-1 Ser(636/639), that, in turn, influence the phosphorylation of its downstream targets, particularly, Akt, mTOR and S6K.105

Both IRS-1 and IRS-2 also play important roles in the control of hepatic metabolism, with IRS-1 more closely linked to glucose homoeostasis and IRS-2 more closely linked to lipid metabolism.106 Importantly, decreased IRS-1 was also associated with a trend towards increased blood glucose, whereas knockdown of IRS-2 resulted in the upregulation of lipogenic enzymes and increased hepatic lipid accumulation.106 In our own study of phosphoproteomic endpoints in morbidly obese patients who underwent liver biopsy, two components of the insulin signalling pathway, AKT kinase and IRS1 were identified as independent predictors of NASH.69

AKT2 is one of the three conserved genes that encode isoforms of the protein kinase B, which participates in the insulin pathway, mitogenic signalling and apoptosis. Out of three isoforms, AKT2 is the one most relevant to metabolic syndrome and the related chronic diseases. Experimental evidence shows that AKT2 is involved in insulin signalling in adipose tissue and is required for the relocation of GLUT4 to the cell membrane in response to insulin.107 In diabetes type 2, activation of Akt by insulin is impaired. Overexpression of activation-impaired Akt2 blocks insulin signalling and reduces glucose disposal,108 both of which are features of PCOS.

Upon analysis of the variants of the AKT2 gene in PCOS and control populations, it was revealed that the minor alleles of rs3730051 and rs8100018 as well as their haplotypes are significantly associated with PCOS.109 The leading hypothesis that connects AKT2 defects with PCOS, although, is not insulin related, but rather focused on its apoptosis suppressing qualities. Overexpression of AKT2 reduces the rate of apoptosis in many tissues, including ovaries. This feature may contribute to the abnormal growth of follicle cells, eventually, leading to the forming of polycystic ovaries. The effect of Akt2 on the rate of apoptosis may also increase the number of steroidogenic cells,110 thus, increasing total androgen production by the ovary and resulting in hyperandrogenism, a predominant characteristic of PCOS.

Importantly, AKT2 is the predominant isoform of AKT in the liver. AKT2 was recently shown to be the key mediator for PIP3-induced accumulation of lipids in the liver, manifesting as NAFLD in model animals.111 Humans with post insulin receptor defects in AKT2 signalling also manifest increased lipogenesis and elevated liver fat content accompanied by an increase in triglyceride-enriched VLDLs, hypertriglyceridemia and low HDL cholesterol levels.112 Interestingly, other studies showed that Akt2 is required for hepatic lipid accumulation in obese, insulin-resistant states induced by high-fat diet feeding; thus indicating that Akt2 is a requisite component of the insulin-dependent regulation of lipid metabolism.113 In either instance, deregulation of AKT2 seems detrimental in both NAFLD and PCOS.

In addition to the components of the insulin signalling cascade described above, PCOS patients demonstrate changes in the levels or the functionality of the downstream components of the insulin signalling cascade, including the glucose transporter 4 (GLUT4), which is necessary for glucose uptake into the cell. In particular, GLUT4 levels are reduced both in endometrial tissue13 and in adipose.114 It is well known that insulin resistance in adipocytes leads to downregulation of GLUT4; the SLC2A4 gene that encodes GLUT4 is a part of the intra-hepatic gene network associated with NAFLD in obese subjects.115

Bariatric surgery in PCOS and NAFLD

Both PCOS and NAFLD/NASH are confounded by insulin resistance and by obesity. The impact of the removal of one or another of these factors on each of these two disorders is worthwhile to note. In PCOS patients, use of insulin sensitizers, primarily metformin, results in the reduction of glucose levels and the attenuation of insulin resistance that are often paralleled by improvement of the distinctive features of PCOS – hyperandrogenism, irregularity of menses and ovulatory dysfunctions.116 Likewise, weight loss-centred interventions often results in similar improvements; however, responsiveness to weight loss in overweight/obese PCOS women varies considerably.117 In a number of studies, it has been reported that obese women with PCOS, on average, display more severe phenotype and experience more profound impairment of fertility as compared to the normal-weight PCOS patients, even if the latter group has the same degree of insulin resistance. It seems that the primary underlying cause of low menstrual frequency observed in a subset of PCOS patients is obesity and not insulin resistance. In line with this hypothesis, a randomised, placebo-controlled, double-blind study unequivocally showed that metformin did not improve the degree of weight loss or menstrual frequency in obese patients with PCOS, while weight loss alone through lifestyle changes substantially improves menstrual frequency.118

In obese patients, weight loss can be achieved by lifestyle modification/behavioural therapy or by bariatric surgery. Out of these two options, bariatric surgery is the most informative one as it produces relatively fast resolution of insulin resistance that precedes actual loss of adiposity.

Increasingly, bariatric surgery is being applied to women of reproductive age who fail to achieve a loss of weight through dietary modification and exercise. The majority of the weight loss occurs in the first year, when pregnancy avoidance is recommended.119 Therefore, in the PCOS population, the effects of insulin resistance abatement and loss of adiposity are difficult to discern. Nevertheless, bariatric surgery was shown to improve many aspects of reproduction, including menstrual regularity,120 clinical and biochemical hyperandrogenism121 and successful conception.122

Similar to PCOS, both liver steatosis and NASH improve after successful weight loss. This improvement is usually accompanied by alleviation of both metabolic parameters and levels of inflammatory mediators. Weight loss improves histological disease activity in NASH in a dose dependent manner; however, more than 50% of patients fail to achieve their BMI targets and more than 50% of patients fail to achieve target weight loss (reviewed by123, 124). Although lack of randomised controlled trials evaluating any bariatric procedure may have led to some bias,125 there are increasing evidence that weight loss induced by bariatric procedures could be beneficial for patients with NASH.20, 126 As NAFLD was recently proposed as the hepatic manifestation of metabolic syndrome, it seems appropriate to establish PCOS as the ovarian manifestation of the same spectrum of metabolic diseases and expect the improvement of its manifestations after bariatric surgery.

Conclusions

Polycystic ovary syndrome is a complex disorder, in which multiple genetic, metabolic and hormonal controls fail to interact properly and produce the hallmark symptoms of the disease. Based on the high association of PCOS and other metabolic abnormalities, such as insulin resistance, hyperandrogenism, obesity and non-alcoholic fatty liver disease, the candidate genes, which are related to those metabolic pathways, have been proposed as PCOS causative genes. Close scrutiny of these PCOS candidates identified that most of their protein products are located at the crossroad of three highly relevant conditions, metabolic syndrome, obesity and NAFLD. In all studied populations, the prevalence of both PCOS and NAFLD rises proportionally to the degree of insulin resistance and gain of adiposity. It seems that both NAFLD and PCOS are often accompanied or augmented by two inter-related physiological states i.e. metabolic syndrome and obesity. As NAFLD was recently proposed as the hepatic manifestation of metabolic syndrome, it seems highly plausible that PCOS is the ovarian manifestation of the same spectrum of metabolic diseases (Figure 2) and consequently, expect improvement of this condition following bariatric surgery.

Figure 2.

 Both NAFLD and PCOS are at the crossroad of two highly relevant conditions, metabolic syndrome and obesity.

Acknowledgement

Declaration of personal and funding interests: None.

Ancillary