Polycystic ovary syndrome (PCOS) is a highly heterogeneous reproductive system disorder of which the aetiology is not fully understood. Previous association studies have been conducted on >100 candidate genes, which principally related to reproductive hormones, cellular metabolism and chronic inflammation. Heritable tendencies have long been recognized for pathogeneses of PCOS, and recently a genome-wide association study (GWAS) in PCOS provides new clues to understand the genetic components and pathways in PCOS physiology. However, the current knowledge of the pathogenesis of PCOS is still in its infancy. Further studies using new technologies such as next-generation sequencing (NGS) shall be useful to understand more causal variants for PCOS.
Highlights
- •
This paper sorts out most of the polycystic ovary syndrome (PCOS)-related candidate genes and highlights some of the key genes.
- •
A review of the results of recent PCOS genome-wide association studies is presented.
- •
The study gives an outlook on genetic study of PCOS.
Introduction
Genetic association studies have been regarded as a popular method for studying relations between the susceptibility loci and complex genetic diseases , including polycystic ovary syndrome (PCOS). In the past, >100 PCOS-related candidate genes have been examined through candidate gene association studies (the majority of them were reviewed in Supplementary Table 1 ). Single-nucleotide polymorphisms (SNPs) are main markers in such genetic association studies and are used for case–control designs. To compare the frequency of SNP allele or genotypes, the associations are indicated between the SNPs and the disease or trait .
The genome-wide association study (GWAS) is based on the high-throughput genotyping and array technique, which provides a powerful tool to discover the relationship between common genetic variations and diseases. Combining the linkage disequilibrium (LD) analysis, which captures most of the genome’s variations, a representative group of SNPs (e.g., tag SNPs) can be applied to evaluate genome-wide variations. Based on the HapMap project, there are 10 million common variants represented by 300,000–600,000 tag SNPs ( www.hapmap.org/ ). To ensure that every independent variant is associated with the disease, the genome-wide threshold requires an association of P < 5 × 10 −8 . As the GWAS does not identify the association between a gene and a particular disease (instead, it identifies a variant that may have an effect on the close or related genes), the association between the specific loci and diseases must be independently tested after GWAS.
In this review, we will summarize some of the candidate gene association studies based on different categories. GWASs on PCOS will also be discussed.
Candidate-gene association studies in PCOS
Candidate-gene association study is the most commonly used approach among current genetic studies for human diseases. To study the genetic basis of PCOS, a number of candidate-gene studies in PCOS have been conducted. Major PCOS candidate genes can be categorized into the following four different groups: (1) genes involved in the biosynthesis and the action of androgens; (2) genes related to metabolism; (3) genes correlated with inflammatory cytokines; and (4) the other candidate genes ( Supplementary Table 1 ).
Genes involved in the biosynthesis and activity of androgen
Hyperandrogenism is one of the fundamental characteristics and clinical manifestations of PCOS. The genes related to steroidogenesis such as CYP17A1 ( cytochrome P450 family 17 subfamily A member 1), CYP19, CYP21, HSD17B5 and HSD17B6 are considered to be essential candidates of PCOS. Likewise, sex hormones and their receptors such as AR (androgen receptor) , SHBG (sex hormone-binding globulin) , FSH (follicle-stimulating hormone) , FSHR ( FSH receptor gene) and LHCGR (luteinizing hormone (LH)/human chorionic gonadotrophin (hCG) receptor) are also regarded as compelling candidate genes for PCOS, due to their fundamental roles in folliculogenesis.
Cytochrome P450 family 17 subfamily A member 1
CYP17A1 encodes the enzyme P450c17a (17-α-hydroxylase and 17-lyase), which is a rate-limiting enzyme that transforms progesterone into testosterone. Polymorphism in gene CYP17A1 promoter has been reported to be related with a rise in testosterone level . However, a recent meta-analysis failed to verify this association .
Androgen Receptor
An important association between short trinucleotide (CAG) repeat of AR and hyperandrogenism was reported by Ibanez L et al. It has been demonstrated that polymorphisms of short CAG repeats can modify the association between insulin resistance (IR) and testosterone . The relationship between AR and PCOS has been studied by different groups, yet they have produced controversial results . Recently, a meta-analysis including 1536 PCOS patients and 1807 controls revealed that the lengths of CAG repeats contributed to hyperandrogenism in PCOS .
Sex hormone-binding globulin
SHBG specifically transports testosterone, dihydrotestosterone (DHT) and estradiol. Previous studies have demonstrated that decreased serum SHBG is associated with increased risk of PCOS . However, the results of association between SHBG and PCOS are controversial. For example, some studies suggested that PCOS and SHBG have a correlation, whereas others (mostly in European and Asian populations) showed opposite results .
Genes related to metabolism
PCOS is a metabolic disorder that is closely related to IR, obesity, type 2 diabetes (T2D) and metabolic syndrome. Therefore, genes affecting IR such as INS (insulin gene) , INSR (insulin receptor) , IRS1 (insulin receptor substrate 1) , IRS2, IGF, PPAR-γ and CAPN10 are all possible candidates for PCOS. Besides, genes related to obesity and T2D such as FTO (fat and obesity-associated gene) and TCF7L2 are also regarded as important candidate genes.
Insulin gene
INS is located at chromosome 11p5.5, with variable number tandem repeat (VNTR) polymorphisms located at the 5’ regulatory element that affects INS transcription. Some studies have reported that the allele carrying the longest repeat has an association with PCOS ; but other studies have not verified . A meta-analysis, which included 13 case–control studies concluded that no significant association between INS VNTR polymorphisms and the risk of PCOS in Caucasian and Asian populations was observed . Hence, contributions of INS to the aetiology of PCOS needed to be examined.
Insulin receptor
INSR is considered as an important candidate gene for PCOS. Mutations in INSR locus may regulate hyperandrogenaemia, hyperinsulinaemia and IR . SNP rs2252673 has been identified to be related to PCOS in Korean women and these findings have been replicated in Caucasian and Han Chinese women with PCOS.
Insulin receptor substrate 1
IRSs are functionally plausible candidates for PCOS susceptibility. A number of studies on polymorphisms of IRSs have been conducted but the results are controversial . Among these studies, Gly972Arg (rs1801278) polymorphisms of IRS1 have been thoroughly examined. It is significantly associated with PCOS in both Japanese and Greek populations . A meta-analysis of 11 studies suggests that rs1801278 confers a significant risk to PCOS .
Fat and obesity-associated gene
Obesity is found in approximately 50% of PCOS patients . ∗ As an obesity-associated gene, FTO has been identified to be strongly associated with childhood and adult obesity . Based on these findings, FTO is considered to be a candidate gene for PCOS. SNP rs9939609 of FTO has been reported to have a significant association with PCOS , whereas in other studies FTO was found to be related to body mass index (BMI) in PCOS . In fact, the nature of the relationship between FTO and PCOS is undetermined because the phenotypic manifestations of PCOS are often confounded by BMI. After adjustment of BMI and age, FTO is still associated with PCOS, indicating that FTO is a risk factor to PCOS independent of BMI or obesity . Thus, the gene FTO has a profound influence on PCOS.
Genes correlated with inflammatory cytokines
PCOS is a pro-inflammatory state disease, and genes related to chronic inflammation might contribute to the incidence of PCOS via an intermediate role among hyperandrogenism, obesity, IR and anovulation . ∗ Hence, the genes correlated with inflammatory cytokines have also become possible candidate genes of PCOS, such as TNF-α (tumour necrosis factor-alpha) , interleukin (IL)-6, IL-1A, IL-1B and PAI .
Tumour necrosis factor-alpha
TNF-α is located at chromosome 6p21.3. Polymorphisms, rs1800629 (G308A) and rs1799724 (C850T), in the promoter region of TNF-α are known to be associated with chronic inflammatory diseases . Although the results were not consistent among different ethnic populations , the meta-analysis revealed that TNF-α is a potential susceptibility factor for PCOS .
Interleukin-6
IL-6 is a pro-inflammatory cytokine, and a higher level of serum IL-6 was found in PCOS patients . Its promoter region SNP rs1800795 (−174 C/G) has been suggested to be associated with PCOS , but the results from other studies were opposite. For example, two meta-analysis studies found that IL-6 gene may not have correlation with the susceptibility of PCOS .
Other candidate genes
There are other candidate genes for PCOS, such as FBN3 ( FIBRILLIN -3), BMP15 ( Bone morphogenetic protein 15), GDF9 (growth differentiation factor 9) and other members of transforming growth factor-β (TGF-β) signalling pathway, which may have an effect on follicle development ; HLA-A ( major histocompatibility complex, class I, A ) and VDR ( vitamin D receptor) may involve in the autoimmunity of PCOS .
The candidate gene association study has advantages to identify associations between a certain gene and PCOS. However, it has limitations. First, the candidate gene association study relies on the understanding of the candidate genes and related diseases. Second, when the ethnic backgrounds of cases and controls are not well matched, false-positive results may occur . In addition, a number of those studies only have limited statistical power due to inadequate sample sizes, and only a few of those studies have yielded sufficiently robust results that could be replicated in different populations or by different investigators .
Candidate-gene association studies in PCOS
Candidate-gene association study is the most commonly used approach among current genetic studies for human diseases. To study the genetic basis of PCOS, a number of candidate-gene studies in PCOS have been conducted. Major PCOS candidate genes can be categorized into the following four different groups: (1) genes involved in the biosynthesis and the action of androgens; (2) genes related to metabolism; (3) genes correlated with inflammatory cytokines; and (4) the other candidate genes ( Supplementary Table 1 ).
Genes involved in the biosynthesis and activity of androgen
Hyperandrogenism is one of the fundamental characteristics and clinical manifestations of PCOS. The genes related to steroidogenesis such as CYP17A1 ( cytochrome P450 family 17 subfamily A member 1), CYP19, CYP21, HSD17B5 and HSD17B6 are considered to be essential candidates of PCOS. Likewise, sex hormones and their receptors such as AR (androgen receptor) , SHBG (sex hormone-binding globulin) , FSH (follicle-stimulating hormone) , FSHR ( FSH receptor gene) and LHCGR (luteinizing hormone (LH)/human chorionic gonadotrophin (hCG) receptor) are also regarded as compelling candidate genes for PCOS, due to their fundamental roles in folliculogenesis.
Cytochrome P450 family 17 subfamily A member 1
CYP17A1 encodes the enzyme P450c17a (17-α-hydroxylase and 17-lyase), which is a rate-limiting enzyme that transforms progesterone into testosterone. Polymorphism in gene CYP17A1 promoter has been reported to be related with a rise in testosterone level . However, a recent meta-analysis failed to verify this association .
Androgen Receptor
An important association between short trinucleotide (CAG) repeat of AR and hyperandrogenism was reported by Ibanez L et al. It has been demonstrated that polymorphisms of short CAG repeats can modify the association between insulin resistance (IR) and testosterone . The relationship between AR and PCOS has been studied by different groups, yet they have produced controversial results . Recently, a meta-analysis including 1536 PCOS patients and 1807 controls revealed that the lengths of CAG repeats contributed to hyperandrogenism in PCOS .
Sex hormone-binding globulin
SHBG specifically transports testosterone, dihydrotestosterone (DHT) and estradiol. Previous studies have demonstrated that decreased serum SHBG is associated with increased risk of PCOS . However, the results of association between SHBG and PCOS are controversial. For example, some studies suggested that PCOS and SHBG have a correlation, whereas others (mostly in European and Asian populations) showed opposite results .
Genes related to metabolism
PCOS is a metabolic disorder that is closely related to IR, obesity, type 2 diabetes (T2D) and metabolic syndrome. Therefore, genes affecting IR such as INS (insulin gene) , INSR (insulin receptor) , IRS1 (insulin receptor substrate 1) , IRS2, IGF, PPAR-γ and CAPN10 are all possible candidates for PCOS. Besides, genes related to obesity and T2D such as FTO (fat and obesity-associated gene) and TCF7L2 are also regarded as important candidate genes.
Insulin gene
INS is located at chromosome 11p5.5, with variable number tandem repeat (VNTR) polymorphisms located at the 5’ regulatory element that affects INS transcription. Some studies have reported that the allele carrying the longest repeat has an association with PCOS ; but other studies have not verified . A meta-analysis, which included 13 case–control studies concluded that no significant association between INS VNTR polymorphisms and the risk of PCOS in Caucasian and Asian populations was observed . Hence, contributions of INS to the aetiology of PCOS needed to be examined.
Insulin receptor
INSR is considered as an important candidate gene for PCOS. Mutations in INSR locus may regulate hyperandrogenaemia, hyperinsulinaemia and IR . SNP rs2252673 has been identified to be related to PCOS in Korean women and these findings have been replicated in Caucasian and Han Chinese women with PCOS.
Insulin receptor substrate 1
IRSs are functionally plausible candidates for PCOS susceptibility. A number of studies on polymorphisms of IRSs have been conducted but the results are controversial . Among these studies, Gly972Arg (rs1801278) polymorphisms of IRS1 have been thoroughly examined. It is significantly associated with PCOS in both Japanese and Greek populations . A meta-analysis of 11 studies suggests that rs1801278 confers a significant risk to PCOS .
Fat and obesity-associated gene
Obesity is found in approximately 50% of PCOS patients . ∗ As an obesity-associated gene, FTO has been identified to be strongly associated with childhood and adult obesity . Based on these findings, FTO is considered to be a candidate gene for PCOS. SNP rs9939609 of FTO has been reported to have a significant association with PCOS , whereas in other studies FTO was found to be related to body mass index (BMI) in PCOS . In fact, the nature of the relationship between FTO and PCOS is undetermined because the phenotypic manifestations of PCOS are often confounded by BMI. After adjustment of BMI and age, FTO is still associated with PCOS, indicating that FTO is a risk factor to PCOS independent of BMI or obesity . Thus, the gene FTO has a profound influence on PCOS.
Genes correlated with inflammatory cytokines
PCOS is a pro-inflammatory state disease, and genes related to chronic inflammation might contribute to the incidence of PCOS via an intermediate role among hyperandrogenism, obesity, IR and anovulation . ∗ Hence, the genes correlated with inflammatory cytokines have also become possible candidate genes of PCOS, such as TNF-α (tumour necrosis factor-alpha) , interleukin (IL)-6, IL-1A, IL-1B and PAI .
Tumour necrosis factor-alpha
TNF-α is located at chromosome 6p21.3. Polymorphisms, rs1800629 (G308A) and rs1799724 (C850T), in the promoter region of TNF-α are known to be associated with chronic inflammatory diseases . Although the results were not consistent among different ethnic populations , the meta-analysis revealed that TNF-α is a potential susceptibility factor for PCOS .
Interleukin-6
IL-6 is a pro-inflammatory cytokine, and a higher level of serum IL-6 was found in PCOS patients . Its promoter region SNP rs1800795 (−174 C/G) has been suggested to be associated with PCOS , but the results from other studies were opposite. For example, two meta-analysis studies found that IL-6 gene may not have correlation with the susceptibility of PCOS .
Other candidate genes
There are other candidate genes for PCOS, such as FBN3 ( FIBRILLIN -3), BMP15 ( Bone morphogenetic protein 15), GDF9 (growth differentiation factor 9) and other members of transforming growth factor-β (TGF-β) signalling pathway, which may have an effect on follicle development ; HLA-A ( major histocompatibility complex, class I, A ) and VDR ( vitamin D receptor) may involve in the autoimmunity of PCOS .
The candidate gene association study has advantages to identify associations between a certain gene and PCOS. However, it has limitations. First, the candidate gene association study relies on the understanding of the candidate genes and related diseases. Second, when the ethnic backgrounds of cases and controls are not well matched, false-positive results may occur . In addition, a number of those studies only have limited statistical power due to inadequate sample sizes, and only a few of those studies have yielded sufficiently robust results that could be replicated in different populations or by different investigators .
Role of GWASs in PCOS
Since the first report on GWASs in 2005, >2000 studies have been added to the catalogue of published GWASs by 2013 ( http://www.genome.gov/gwastudies/ ). GWASs have been successfully used to identify susceptibility loci for a number of common diseases including metabolic diseases such as T2D , mental disorders such as schizophrenia and autoimmune diseases such as Graves’ disease . In recent years, GWASs have been used to examine female reproductive disorders such as PCOS, premature ovarian failure and endometriosis . ∗ To date, five GWASs have been performed on women with PCOS, which were conducted in Han Chinese ∗ Korean ∗ and European ancestry populations . ∗ From these studies, 15 risk loci ( P < 5 × 10 −8 ) have been identified ( Table 1 ).
| Population | Chromosome | SNP | Nearby Genes | P- value | OR | References |
|---|---|---|---|---|---|---|
| Han Chinese | 2p21 | rs13429458 | THADA | 1.73 × 10 −23 | 1.49 | (Chen et al., 2011) |
| 2p21 | rs12478601 | THADA | 3.48 × 10 −23 | 1.39 | ||
| 2p16.3 | rs13405728 | LHCGR | 7.55 × 10 −21 | 1.41 | ||
| 9q33.3 | rs2479106 | DENND1A | 8.12 × 10 −19 | 1.34 | ||
| 9q33.3 | rs10818854 | DENND1A | 9.40 × 10 −18 | 1.51 | ||
| Han Chinese | 12q13.2 | rs705702 | RAB5B/SUOX | 8.64 × 10 −26 | 1.27 | (Shi et al., 2012) |
| 11q22.1 | rs1894116 | YAP1 | 1.08 × 10 −22 | 1.27 | ||
| 12q14.3 | rs2272046 | HMGA2 | 1.95 × 10 −21 | 1.43 | ||
| 9q22.32 | rs3802457 | C9orf3 | 5.28 × 10 −14 | 1.3 | ||
| 2p21 | rs13429458 | THADA | 4.17 × 10 −13 | 1.49 | ||
| 2p16.3 | rs2268361 | FSHR | 9.89 × 10 −13 | 1.15 | ||
| 16q12.1 | rs4784165 | TOX3 | 3.64 × 10 −11 | 1.15 | ||
| 9q33.3 | rs2479106 | DENND1A | 5.14 × 10 −10 | 1.35 | ||
| 20q13.2 | rs6022786 | SUMO1P1 | 1.83 × 10 −9 | 1.13 | ||
| 2p16.3 | rs13405728 | LHCGR | 3.77 × 10 −9 | 1.35 | ||
| 19p13.2 | rs2059807 | INSR | 1.09 × 10 −8 | 1.14 | ||
| Korean | 8q24.2 | rs10505648 | KHDRBS3 | 5.46 × 10 −8 | 0.52 | (Lee et al., 2015) |
| Caucasians | 9q22.32 | rs10993397 | c9orf3/FANCC | 4.6 × 10 −13 | 0.72 | (Hayes et al., 2015) |
| 8p32.1 | rs804279 | GATA4/NEIL2 | 8.0 × 10 −10 | 0.74 | ||
| 11p14.1 | rs11031006 | KCNA4/FSHB | 1.9 × 10 −8 | 1.12 | ||
| Caucasians | 2q34 | rs1351592 | ERBB4/HER4 | 1.2 × 10 −12 | 1.18 | (Day et al., 2015) |
| 11q22.1 | rs11225154 | YAP1 | 7.6 × 10 −11 | 1.22 | ||
| 2p21 | rs7563201 | THADA | 3.3 × 10 −10 | 1.13 | ||
| 11p14.1 | rs11031006 | FSHB | 1.3 × 10 −9 | 1.16 | ||
| 5q31.1 | rs13164856 | RAD50 | 3.5 × 10 −9 | 1.13 | ||
| 12q21.2 | rs1275468 | KRR1 | 1.9 × 10 −8 | 1.13 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
