Epigenetics in Polycystic Ovary Syndrome: A Pilot Study of Global DNA Methylation

Polycystic ovary syndrome (PCOS) is a heterogenic common genetic disorder, and features such as hyperandrogenemia, abnormal insulin resistance/secretion in relatives of PCOS women are heritable. However, the pathophysiology of PCOS is poorly understood. Although positive association results have been reported, no genes have been broadly accepted as causative for PCOS due to lack of replication (1). Environmental factors may play a role in PCOS (2).

Emerging evidence from prenatally androgenized animal models provides new insights into the pathogenesis of PCOS, and suggests that an environmental factor, androgen exposure during in utero development, may predispose individuals to this disorder (3). Female mammals (such as rhesus monkeys, sheep, mice or rats) exposed to androgen excess in utero develop masculinized phenotypes similar to PCOS as an adult (3, 4). This phenomenon has been formulated as the fetal origins hypothesis of PCOS.

The mechanism of the fetal origins hypothesis is unknown (5). During development, unfavorable prenatal environmental influences may lead to persistent changes in the epigenome, such as loss of imprinting of insulin-like growth factor 2 (IGF2), increased or decreased DNA methylation levels on CpG sites, which result in altered molecular pathways and increase the later risk of adult diseases (6). A rat model exposed to estrogen in the neonatal period was generated to elucidate the potential role of early estrogen exposure in prostate cancer; expression of the phosphodiesterase D4 (PDE4D4) gene was increased by hypomethylation (7). In mice, diethylstilbestrol exposure in utero resulted in hypermethylation of the homeobox A10 gene (HOXA10), and long-term altered HOXA10 expression (8). Similar to estrogens in rodent models, elevated androgens in utero may disturb regular biological signals, and can predispose to disease later in life (9). We and others therefore hypothesize that exposure to excess androgens during early development may possibly reprogram the epigenome, causing persistent DNA methylation modification and irregular gene expression predisposing to PCOS (10).

Epigenetics is the study of heritable changes in gene expression that are not caused by DNA sequence alterations, but are mitotically and transgenerationally heritable. Inappropriate epigenetic reprogramming has been identified as contributing to common diseases with fetal origins such as type 2 diabetes (11), and prostate cancer (12), suggesting it may also contribute to PCOS, given that PCOS is a common disease with both reproductive and metabolic abnormalities (5, 10). Additionally, epigenetic alterations have been observed as non-random X-chromosome inactivation in PCOS women, evidence that epigenetics may modulate the effect of the androgen receptor gene located on the X chromosome (13–15). Furthermore, DNA methylation, the principal mechanism of epigenetics, has been reported to play a role in cancer, aging, and complex chronic diseases (16). To investigate the role of epigenetics in PCOS, we conducted a pilot epigenetic study of DNA methylation in PCOS comparing the global methylation percentage between PCOS and matched controls.

Twenty unrelated white patients with PCOS and 20 healthy unrelated white control women were selected from an existing cohort of 335 cases and 198 controls, recruited at the University of Alabama at Birmingham (UAB; n=16) and Cedars-Sinai Medical Center (CSMC; n=24). We selected PCOS subjects who were young and had a BMI less than the median BMI (33.4 kg/m²) of the entire PCOS group to minimize the effects of obesity on epigenetic changes (17). For each PCOS subject a control subject was one-to-one matched for age (+/− 3 years) and BMI (+/− 3 kg/m²). The recruitment strategy (18) and methods of clinical and biochemical characterization (19) of this cohort have been previously described.

In brief, all cases and controls were not taking hormonal therapy for at least three months prior to enrollment. PCOS was defined by the 1990 National Institutes of Health criteria (20). All subjects had given informed consent, and the study was performed according to the guidelines of the Institutional Review Boards of UAB and Cedars-Sinai Medical Center.

The extracted peripheral leukocyte DNA was quantified using the NanoDrop™ 1000 Spectrophotometer (Thermo Scientific, Wilmington, DE). For each sample, methylation analysis was performed in triplicate and averaged (100 ng DNA each) using a global methylation quantification kit (Methylamp™, Epigenetek, Brooklyn, NY), following manufacturer’s instructions. Briefly, the methylated fraction was recognized by 5-methylcytosine antibody and quantified through an ELISA-like reaction. The total amount of methylated DNA is proportional to the optical density (OD). A standard curve was generated by plotting the OD values of a dilution series made from a 100% methylated DNA standard (supplied with the kit), the slope of which was used to calculate the total amount of methylated DNA for each sample. Methylation percentage was obtained by dividing the methylated DNA amount into the initial loaded DNA CpG amount. Purified DNA extracted from MCF-7 human breast adenocarcinoma cells was provided with the kit and included in each experiment as an internal control. Global methylation percentage values were normalized to this internal control.

Clinical characteristics and methylation percentages of women with PCOS and controls were compared using the Mann-Whitney U test. Clinical and methylation data are expressed as the median (interquartile range). A P-value <0.05 was considered significant.

The clinical characteristics of the subjects are presented in Supplemental Table 1. Overall, the subjects were young (median age 23.0 (4.5) years) and lean (median BMI 23.3 (4.7) kg/m²). Cases and controls differed only in terms of androgen levels and hirsutism score. These subjects had normal insulin sensitivity and insulin secretion. The median global methylation percentages were 6.7% (IQR 5.6%) for PCOS women and 7.1% (IQR 6.2%) for controls, a non-significant difference (P=0.79) (Figure 1).

This is the first epigenetic study to investigate whether global DNA methylation is altered in PCOS patients compared to matched controls. The most accessible and achievable bio-resource is peripheral blood, so we first asked the question of whether the overall methylation percentage is altered in genomic DNA extracted from the peripheral blood in PCOS patients.

A strength of our study was the selection of matched cases and controls. Epigenetic status can be modified by a combination of environment and nutrition; age and BMI may affect methylation patterns. Remarkable differences between older monozygotic twins have been observed in their overall content and genomic methylated DNA and histone acetylation (21). We aimed to match PCOS and controls by age and BMI, and excluded older subjects and extremely obese subjects, avoiding confounding methylation differences arising from these factors.

There are biologic reasons that may explain the negative results of this study. First, epigenetic modification such as methylation on CpG islands is likely to occur at tissue-specific differentially methylated regions (22), although global methylation changes over time have been detected in peripheral leukocyte DNA in a few studies (23, 24). Second, our experiment measuring the total methylated DNA fraction across the entire genome is not designed to detect individual methylation alterations at specific sites. Hypomethylation or hypermethylation in specific genomic regions may exist that affect PCOS, but are undetectable by the methodology used in this study (hyper- and hypomethylation would neutralize each other in this analysis). In the absence of knowledge of candidate genes for methylation in PCOS, global DNA methylation analysis was an appropriate first step to study epigenetics; however, our negative results suggest future studies should examine specific candidate genes/regions implicated by other evidence (e.g. expression studies, genetic association studies). Alternatively, site-specific methylation analysis on a genome-wide scale would enable us to systematically discover individual sites differentially methylated in PCOS versus controls.

In part, the small sample size may have been responsible for the negative results. However, a post-hoc power calculation determined that over a thousand subjects would be necessary to establish significance for the observed methylation percentage difference between cases and controls, suggesting that if a difference truly exists, it is likely to be subtle.

In summary, this pilot study found no significant difference in the global methylation of peripheral leukocyte DNA between PCOS and matched controls. Our preliminary study indicates the need to further investigate methylation in key tissues other than peripheral leukocytes, such as human ovaries, adipose tissue or adrenals, and specific target genes or regions. Aberrant DNA methylation patterns could serve as epigenetic biomarkers for early detection of PCOS, and understanding the epigenetic mechanisms involved in PCOS may provide novel avenues for diagnosis and treatment of this common disorder. Finally, it is also possible that epigenetics does not play a critical role in the development of PCOS.