Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Evaluation of circulating microRNA profiles in Brazilian women with polycystic ovary syndrome: A preliminary study

  • Giovana De Nardo Maffazioli ,

    Roles Conceptualization, Data curation, Formal analysis, Writing – original draft

    giovanamaffazioli@hotmail.com

    Affiliations Faculdade de Medicina de Sao Paulo, Departamento de Obstetrícia e Ginecologia, Disciplina de Ginecologia, Hospital das Clínicas HCFMUSP, Sao Paulo, SP, Brazil, Laboratório de Ginecologia Estrutural e Molecular (LIM 58), Hospital das Clinicas da Faculdade de Medicina da Universidade de Sao Paulo, Departamento de Obstetricia e Ginecologia, Disciplina de Ginecologia, São Paulo, Brazil

  • Edmund Chada Baracat,

    Roles Conceptualization, Supervision, Writing – review & editing

    Affiliations Faculdade de Medicina de Sao Paulo, Departamento de Obstetrícia e Ginecologia, Disciplina de Ginecologia, Hospital das Clínicas HCFMUSP, Sao Paulo, SP, Brazil, Laboratório de Ginecologia Estrutural e Molecular (LIM 58), Hospital das Clinicas da Faculdade de Medicina da Universidade de Sao Paulo, Departamento de Obstetricia e Ginecologia, Disciplina de Ginecologia, São Paulo, Brazil

  • José Maria Soares,

    Roles Supervision, Writing – review & editing

    Affiliation Faculdade de Medicina de Sao Paulo, Departamento de Obstetrícia e Ginecologia, Disciplina de Ginecologia, Hospital das Clínicas HCFMUSP, Sao Paulo, SP, Brazil

  • Kátia Cândido Carvalho,

    Roles Data curation, Investigation, Methodology, Supervision, Writing – review & editing

    Affiliations Faculdade de Medicina de Sao Paulo, Departamento de Obstetrícia e Ginecologia, Disciplina de Ginecologia, Hospital das Clínicas HCFMUSP, Sao Paulo, SP, Brazil, Laboratório de Ginecologia Estrutural e Molecular (LIM 58), Hospital das Clinicas da Faculdade de Medicina da Universidade de Sao Paulo, Departamento de Obstetricia e Ginecologia, Disciplina de Ginecologia, São Paulo, Brazil

  • Gustavo Arantes Rosa Maciel

    Roles Conceptualization, Resources, Supervision, Writing – review & editing

    Affiliations Faculdade de Medicina de Sao Paulo, Departamento de Obstetrícia e Ginecologia, Disciplina de Ginecologia, Hospital das Clínicas HCFMUSP, Sao Paulo, SP, Brazil, Laboratório de Ginecologia Estrutural e Molecular (LIM 58), Hospital das Clinicas da Faculdade de Medicina da Universidade de Sao Paulo, Departamento de Obstetricia e Ginecologia, Disciplina de Ginecologia, São Paulo, Brazil

Abstract

Objective

Polycystic ovary syndrome (PCOS) is a heterogeneous endocrinopathy, which etiology encompasses complex genetic traits associated with epigenetic factors, including differences in microRNA (miRNA) expression in a variety of tissues. The circulating form of these molecules is raising attention in the syndrome not only as potential biomarkers of PCOS but also as possible therapeutic targets. The aim of this study was to explore the circulating miRNA profiles present in a cohort of Brazilian women with and without PCOS and to evaluate the potential role of miRNAs in the pathophysiology of the syndrome.

Methods

Cross-sectional study of 36 well-characterized PCOS women and 16 healthy controls. Clinical, hormone and metabolic data were recorded and evaluated. The expression profile of the 201 circulating miRNA selected were analyzed by taqman quantitative real time polymerase chain reactions (RT-PCR) using a customized Open Array platform. Statistical and bioinformatic analyzed were performed.

Results

Circulating miR-21-5p, miR-23a-3p and miR-26a-5p were upregulated, and miR-103a-3p, miR-376a-3p, miR-19b-3p and miR-222-3p were downregulated in women with PCOS compared to healthy normo-ovulatory controls. miR-21-5p, miR-103a-3p and miR-376a-3p levels correlated positively with androgen levels. These miRNAs, in combination, were related to pathways involved in insulin signaling, steroids biosynthesis and endothelial regulation as well as in folliculogenesis.

Conclusion

In this study, we identified a specific circulating miRNA signature in Brazilian women with PCOS. According to our data, circulating miR-21-5p, miR-23a-3p, miR-26a-5p, miR-103a-3p, miR-376a-3p, miR-19b-3p and miR-222-3p may represent potential candidates for differential diagnosis of PCOS in the future.

Introduction

Polycystic ovary syndrome (PCOS) is a complex endocrinopathy that affects between 5% and 17% of women of reproductive age, depending on the population studied and the diagnostic criteria used [15]. It is characterized by chronic anovulation, hyperandrogenism, and the presence of polycystic ovaries on ultrasound [6,7], and it is frequently associated with infertility, insulin resistance, obesity and an increased risk of metabolic diseases such as type 2 diabetes mellitus, metabolic syndrome, cardiovascular disease and dyslipidemia [810].

The etiologies of the syndrome and its comorbidities are not fully understood, but they are known to have a complex genetic component with a high degree of heritability [11,12] and to be associated with epigenetic changes [1315]. Thus, a greater understanding of the molecular mechanisms underlying the pathophysiology of the syndrome is needed and might lead to the identification of novel diagnostic and therapeutic targets.

MicroRNAs (miRNAs) are small noncoding RNAs that are produced in all the cells of the body; they are associated with the posttranscriptional epigenetic regulation of gene expression [16] and are involved in the maintenance of metabolic homeostasis [17], obesity and diabetes [17]. The study of miRNAs in women with PCOS has aroused great interest in recent years. Most of these studies have focused on analysis of differential expression of miRNAs in women with and without PCOS using various tissues (serum, adipocytes, follicular fluid, and ovarian cells) [1822].

Currently, approximately 79 miRNAs have been identified as showing differential expression [23] in PCOS, and the expression profiles of these miRNAs have been related to their varied characteristics and comorbidities. However, the results obtained in different studies vary, and the role that these molecules play as markers or mediators of PCOS requires further elucidation. Some possible explanations for the divergences in the results obtained in the existing studies include insufficient sample size, variation in ethnicity, age, body mass index (BMI), inclusion criteria or comparison parameters, as well as differences in the technical aspects of the studies such as the type of analysis performed or the cutoff points used to define differential expression [23].

Due to the contrasting results and the potential role that epigenetic changes may play in the pathophysiology of PCOS, it is important that more detailed studies be conducted to replicate the findings reported in other populations and to make it possible to understand the mechanisms that these or other miRNAs may play in the syndrome and in its associated comorbidities. Thus, the aim of this study was to explore circulating miRNA profiles in a cohort of Brazilian women with and without PCOS. The Brazilian population is characterized by racial diversity and wide miscegenation, characteristics that may result in the presence of unique epigenetic traits.

Material and methods

Study population and setting

This cross-sectional study included 38 participants with PCOS who were selected from among the patients who attended their first clinical appointments between 2016 and 2018 at the Division of Ginecologia Endocrina da Disciplina de Ginecologia do Hospital das Clinicas da Faculdade de Medicina da Universidade de Sao Paulo, Brazil (HC-FMUSP). These patients complained of hyperandrogenism and/or menstrual disturbances and met the criteria for the diagnosis of PCOS [7]. The 21 participants in the control group were recruited from the general population by convenience sampling. All participants received guidance on the research in accordance with the Declaration of Helsinki and signed a Free and Informed Consent Form (ICF). This project was approved by the Ethics Committee for the Analysis of Research Projects of the Clinical Directorate of HC-FMUSP (number 2.622.258).

All control participants were healthy and had regular menstrual cycles (between 24 and 38 days in length) and hormonal profiles within the normal range for the age group. All participants with PCOS met the diagnostic criteria listed in the Rotterdam consensus [7]. This diagnostic criterion was chosen for the classification of participants with PCOS since it has been widely used in clinical studies [24].

The recruited participants were interviewed and subjected to detailed physical examination according to the clinical protocol. Individuals with diabetes mellitus, any type of cancer, coronary artery disease, stroke, liver cirrhosis, thyroid diseases or a history of autoimmune disease were excluded. Women who had been pregnant, had an abortion or undergone a surgical procedure during the year prior to the study were excluded. Women who had used medications that alter the hormonal and metabolic profile, including hormonal contraceptives, hypoglycemic agents, statins/fibrates and glucocorticoids, during the 3 months prior to the study were also excluded. Gynecological and obstetric history as well as family history were collected.

The data collected during the physical examination included weight, height, body mass index (BMI), waist circumference (WC), systemic blood pressure, and assessment of the presence of acne, alopecia, voice deepening, clitoral enlargement and acanthosis nigricans. WC measurement was performed according to the standards recommended by the World Health Organization (WHO) [25]. Evaluation of the presence of clinical hyperandrogenism (hirsutism) was performed using the Ferriman-Gallwey scale modified by Hatch [26].

Biochemical analyses

Peripheral blood samples were collected from all participants after the medical interview and used to assess their hormonal and metabolic profiles. The blood samples were obtained after the subjects had fasted for 12 hours and during the early follicular phase of the menstrual cycle if the participants had regular menstrual cycles or at any stage of the cycle in the case of participants who had oligo/amenorrhea. An oral glucose tolerance test 2 hours after ingestion of 75 g of glucose was also performed in the PCOS group. The reference values ​​for the hormonal and metabolic parameters were obtained through the core laboratory of HC-FMUSP.

Transvaginal ultrasonography

To assess the polycystic ovarian pattern, transvaginal ultrasonography was performed on the participants in the PCOS group. According to the Rotterdam consensus, ovaries were classified as a polycystic appearance if they met at least one of the following criteria: twelve or more follicles smaller than 10 mm in at least one ovary and/or ovarian volume greater than 10 cm3 in at least one ovary [24].

Blood collection, microRNA selection and sample quality control

Peripheral blood was collected by venipuncture of the forearm and stored in 5-mL BD Vacutainer® tubes containing clot activator, which accelerates the clotting process, and separator gel to obtain serum of the highest quality. After collection, the samples were homogenized manually, and the tubes were kept in a vertical position for 30 minutes at room temperature to retain the clot. The tubes were then subjected to centrifugation at 2,000 x g at 4°C for 15 minutes using a Sorvall ST 16R centrifuge (ThermoFisher Scientific, Waltham, MA, USA). The serum was transferred to 2-mL DNase/RNase-free tubes and stored in 1-mL aliquots at -80°C until RNA extraction.

A total of 201 miRNAs that have been reported in the literature to be related to PCOS or to the expression of genes related to insulin metabolism, pancreatic β cell function or type 2 diabetes mellitus were selected for study [23,2730]. The FirePlex Discovery Engine platform (available at www.fireflybio.com) was used as an aid in their selection. To avoid contamination with miRNA from blood cells, only samples without hemolysis that showed an absorbance of ≤0.2 on the Nanodrop 2000 spectrophotometer (Thermo Scientific, Waltham, MA, USA) were considered adequate for the extraction of genetic material [31].

MicroRNA extraction and preparation

The samples were thawed on ice and then centrifuged for 5 minutes at 1,6000 x g at a temperature of 4°C (Sorvall Legend Micro 17R, Thermo Scientific) to remove cryoprecipitates. Soon after, 200 μL of the supernatant was transferred to a new 2-mL microtube. The total RNA extraction process was conducted using the miRNeasy Serum/Plasma kit (Qiagen, Germany) according to the manufacturer’s instructions.

To monitor the efficiency of total RNA extraction, 0.02 pmol/μl synthetic miRNA-39-3p (Caenorhabditis elegans—cel-miR-39-3p; Thermo Scientific) was added to the sample during the extraction process, as recommended by the manufacturer.

The extracted RNA was eluted with 14 μL of nuclease-free water (Qiagen, Germany). The concentration and purity of the RNA were assessed using a spectrophotometer (Nanodrop 2000, Thermo Scientific). Due to the low concentration of circulating miRNA in serum, cDNA synthesis was performed using the Taqman® Advanced miRNA Assay kit for low-throughput samples (Thermo Scientific) according to the protocol provided by the manufacturer.

Analysis of MicroRNA gene expression

Real-time PCR was performed to test the amplification of the 201 miRNAs of interest using the Open Array platform (Custom-Configured TaqMan® Open ArrayTM Plates, Thermo Scientific). The reactions were performed in a final volume of 25.0 μL using Taqman Open Array Real-Time PCR Master Mix (Taqman® Advanced miRNA Assay, Thermo Scientific) and were transferred to open array chips in an automated manner using the AccuFill® device (Thermo Scientific). The Open Array chips were then sealed and incubated in the QuantStudio 12K Flex Real-Time PCR System (Thermo Scientific) for RT–PCR.

The amplification curves for each miRNA were analyzed using ThermoFisher Cloud software (https://www.thermofisher.com/account-center/cloud-signin-identifier.html). miRNAs that did not present adequate amplification profiles based on these curves were excluded according to the algorithm present in the software itself, as indicated by the manufacturer. At the conclusion of this analysis, 97 miRNAs remained. Of these, 3 miRNAs (let-7b-5p, miR486-5p and miR-92a-3p) were selected as endogenous controls. These miRNAs were selected, as recommended by the reagent manufacturer (Thermo Scientific), because they showed the most homogeneous amplification curves in all samples among all evaluated miRNAs, as determined from their amplification scores (≤1.5). The software considered the mean of the comparative threshold cycle (CT) values of these three genes when calculating the relative expression (ΔCT) of the other miRNAs. The ΔCT was calculated by subtracting the CT from the PCOS group or the control group from the average of the CT values of the endogenous controls (let-7b-5p, miR486-5p and miR-92a-3p).

Expression values were calculated using the ΔΔCT method and are presented as relative quantification (RQ) and fold regulation (FR). Values were obtained as described by Calvano et al. [32]. The cutoff expression values of + 1.0 and -1.0, based on reference values of expression, were used to identify the miRNAs that were hyperexpressed (1) and hypoexpressed (-1) in the PCOS group compared to the control group.

Interaction network analysis of the miRNAs and genes associated with PCOS and metabolic dysfunctions

To better understand the roles of the selected miRNAs with the most significant differential expression profiles between the PCOS group and the control group in the pathophysiology of PCOS, the interaction networks of these miRNAs with their potential targets were evaluated separately and together using the miRTargetLink Human platform (https://ccb-web.cs.uni-saarland.de/mirtargetlink/).

Subsequently, a gene ontology (GO) biological process and molecular function enrichment analysis was performed in relation to the interaction network of the genes targeted by at least two differentially expressed miRNAs using the GeneNetwork platform, available at https://genenetwork.nl/. Enrichment analysis of the main signaling pathways related to these genes was performed using the same platform. The GeneNetwork platform integrates data from the Kyoto Encyclopedia of Genes and Genomes (KEGG) and from the REACTOME pathways database. The raw data were then manually screened, and the biological and molecular information as well as information on the enriched signaling pathways, which may be involved in the pathophysiology of PCOS, were selected.

Statistical analysis

Analysis of the collected data was conducted using the STATA program (version 14.0). The graphs were prepared using GraphPad Prism (version 5.0). The data are presented as mean ± standard deviation, median (interquartile range) or percentage depending on the nature of the data and were checked for normality using distribution graphs and the Shapiro–Wilk test. For the analysis of differences between two groups, Student’s t test or Wilcoxon’s rank sum test was used according to normality for continuous variables. For the analysis of categorical variables, the chi-square test or Fisher’s exact test (when more appropriate) was used. As expected, the control participants had significantly lower BMI than did the participants with PCOS. Because of this, the P values of the laboratory and miRNA expression data were adjusted using covariance analysis (ANCOVA). Spearman’s rank correlation was used to relate the miRNA expression profiles to clinical and laboratory characteristics. Sample calculation was based on a 50% difference in the measures of relative expression of miRNAs between the PCOS and control groups, with a common standard deviation of 0.5, according to a previous study [33]. A sample of 17 participants per group showed 80% power with an α of 0.05. To make it possible to conduct an exploratory analysis of subgroups within the SOP, it was decided to recruit twice as many participants with PCOS as participants without PCOS.

Results

Clinical characteristics

A total of 58 women were recruited; 38 of these were in the PCOS group, and 21 were in the control group. After samples with insufficient amounts of genetic material or quality of amplification in RT–PCR were excluded, 36 women with PCOS and 16 controls remained. Analysis of the clinical characteristics of these women showed that the two groups were adequately matched in terms of age (p = 0.16) and ethnicity (p = 0.87). As expected, participants with PCOS had higher BMI (p = 0.03), WC (p = 0.001), higher incidence of acanthosis nigricans (p = 0.02) and significantly different Ferriman-Gallwey scores (p = 0.0001) in relation to the control group [Table 1].

thumbnail
Table 1. Clinical, hormonal and metabolic characteristics of women with PCOS versus controls.

https://doi.org/10.1371/journal.pone.0275031.t001

Among women with PCOS, 97.2% had oligo/amenorrhea, 68.7% had clinical and/or laboratory hyperandrogenism, and 77.2% had increased ovarian volume or a polycystic pattern on ultrasound. Of the participants who reported menstrual changes, 25.7% had amenorrhea, and 74.3% reported oligo/amenorrheic cycles.

Hormonal and metabolic characteristics

The hormonal characteristics of the two groups are described in Table 1. Women with PCOS had higher mean concentrations of LH (12.9 ± 6.4 IU/L vs. 7.0 ± 5.5 IU/L, p = 0.001), higher LH/FSH ratios (2.47 ± 0.98 vs. 1.10 ± 0.89, p <0.0001), higher total testosterone (53 ± 25 ng/dL vs. 34 ± 11 ng/dL, p = 0.003) and free testosterone (33 ± 18 pmol/L vs. 14 ± 5, p <0.0001) levels, and higher AMH (7.48 ± 4.16 ng/mL vs. 2.89 ± 2.05 ng/mL, p <0.0001) than the controls, even after adjustment for BMI. In contrast, mean SHBG concentrations were higher in the control participants than in those with PCOS (62.4 ± 29.5 nmol/L vs. 41.8 ± 24.7 nmol/L, p = 0.01), maintaining a trend after adjustment for BMI. FSH showed no statistically significant difference between the two groups.

Regarding the analysis of metabolic profiles [Table 1], women with PCOS had higher mean fasting glucose/insulin ratios (5.3±3.7 vs. 10.9±5.1, p = 0.0001), higher HOMA-IR (6.9±7.6 vs. 2.0±1.1, p = 0.001), and higher triglyceride levels (140±99 mg/dL vs. 77±35 mg/dL, p = 0.01) than controls, even after adjustment for BMI, and HOMA-IR maintained a trend after this adjustment. Although participants with PCOS also had lower mean HDL-C levels than controls, these differences did not persist after adjustment for BMI. The two groups did not differ in fasting glucose or total or LDL cholesterol levels.

Differences in circulating miRNA expression profiles between PCOS and controls

Of the 201 miRNAs analyzed, 103 could not be assessed due to the absence of expression in the target or control group samples. Let-7b-5p, miR-486-5p and miR92a-5p were selected as endogenous controls because they presented adequate stability patterns among all samples. A total of 95 miRNAs were included in the differential expression analysis in which the two groups were compared. Only 33 of these miRNAs showed differential expression profiles/values (4 were upregulated, and 29 were downregulated) in the PCOS group compared to controls [Table 2]. For these miRNAs, the mean differences in relative expression between the two groups were analyzed. To verify the statistical relevance of the findings and the adjustment for BMI, the remaining 15 miRNAs that showed significant differences in expression between the two groups (miR-21-5p, miR-23a-3p and miR-26a-5p (overexpressed in the PCOS group) and miR-103a-3p, miR-361-5p, miR376a-3p, miR-382-5p, miR-128-3p, miR181b-5p, miR-19b-3p, miR-21-3p, miR-222-3p, miR-33a-5p, miR-494-3p, and miR-885-5p (hipoexpressed in PCOS group) were analyzed [Table 3]. To decrease the chance of bias in our findings, only miRNAs that presented amplification in more than 80% of the study samples (miR-103a-3p, miR-21-5p; miR-23a-3p, miR-26a-5p, miR-376a-3p, miR-19b-3p and miR-222-3p) were chosen for correlation and network interaction analysis.

thumbnail
Table 2. Fold regulation (FR) of the 95 evaluated miRNAs in the PCOS group in relation to controls.

https://doi.org/10.1371/journal.pone.0275031.t002

thumbnail
Table 3. Comparison of the expression values of the 33 miRNAs selected between PCOS and controls.

https://doi.org/10.1371/journal.pone.0275031.t003

Correlations between clinical and laboratorial characteristics and circulating miRNA expression in women with PCOS

Correlation coefficients were obtained to determine whether the seven miRNAs identified as having differential expression in the PCOS group were related to any particular clinical and/or laboratory features of the syndrome. As depicted in Table 4, miR-103a-3p, miR-21-5p and miR-376a-3p were positively correlated with total testosterone level (rho = 0.39, p = 0.02; rho = 0.48, p = 0.0.004; and rho = 0.34, p = 0.047, respectively). Other relevant characteristics were not correlated with these miRNAs in our PCOS group.

thumbnail
Table 4. Correlations between clinical and laboratorial characteristics with miRNAs in PCOS women.

https://doi.org/10.1371/journal.pone.0275031.t004

Interaction networks and enrichment analysis for miR-103a-3p, miR-21-5p, miR-23a-3p, miR-26a-5p, miR-376a-3p, miR-19b-3p and miR-222-3p and their main molecular targets involved in the development of PCOS and its associated metabolic dysfunctions

An interaction network of potential target genes related to each miRNA was constructed. Including only interactions for which there is support in the literature, miR-103a-3p was predicted to interact with 16 target genes, and miR-21-5p, miR-23a-3p, miR-26a-5p, miR-376-3p, miR-19b-3p, and miR-222-3p were predicted to interact with 99, 32, 52, 10, 24, and 32 genes, respectively. Of those genes, 17 were related to at least two of these miRNAs [Fig 1].

thumbnail
Fig 1. Mapping the target genes (n = 17) that were common to at least two miRNAs with differential expression in the PCOS groups.

https://doi.org/10.1371/journal.pone.0275031.g001

Subsequently, enrichment analysis of biological processes and molecular functions in relation to the genetic ontology (GO) of these target genes was performed. Table 5 shows a selection of the main biological processes and molecular functions of the 17 genes that are related to at least two of the miRNAs that were found to be differentially expressed in the PCOS patients.

thumbnail
Table 5. Genetic ontology (GO) of biologic process and function and KEGG and REACTOME data banks pathway analysis associated to the seventeen target genes related to at least two miRNAs.

https://doi.org/10.1371/journal.pone.0275031.t005

To obtain a list of signaling pathways that are likely controlled by these genes, the main interaction pathways that are overrepresented in the analysis and that may be related to the pathophysiology of PCOS were selected; these are listed in Table 5.

Discussion

For this study, we selected 201 circulating miRNAs with potential contribution to the pathophysiology of the syndrome and evaluated their expression profiles in a well-characterized population of women with PCOS. Increased expression of miR-21-5p, miR-23a-3p, and miR-26a-5p and lower expression of miR-103a-3p, miR-376a-3p, miR-19b-3p and miR-222-3p were observed in PCOS, compared to control women. The differential expression of miRNAs has been suggested to target genes and to modulate pathways previously described as being part of PCOS physiopathology and its related metabolic disturbances [18,19,34,35].

Previous studies have also found abnormal expression patterns of circulating miR-21, miR-23a-3p, miR-19b and miR-222 in women with PCOS [23,36,37]. Regarding miR-21, the results were conflicting. Sorensen et al. [38] found hypoexpression of this miRNA in the serum of women with PCOS compared to healthy women, whereas Jiang et al. [39] and Butler et al. [37] found it to be overexpressed. However, the first study evaluated miR-21-3p, whereas the latter two studies evaluated miR-21-5p. In our samples, although the differential expression of miR-21-3p was not statistically significant, this miRNA showed hypoexpression in the serum of women with PCOS compared to controls, while miR-21-5p showed hyperexpression, corroborating those studies. Another study [35] found no difference in serum miR-21 expression between women with PCOS and healthy control women. However, that study did not specify which miR-21 was assessed.

In contrast with our results, Xiong et al. [36] found downregulation of circulating miR-23a-3p in PCOS, and that it was positively associated with BMI. However, in agreement with our findings, there was no association of this molecule with androgen or gonadotrophin levels. Serum miR-19b-3p and miR-222-3p were also previously studied in women with PCOS, and both displayed increased expression compared to age- and BMI-matched healthy controls [20,37]. Although our study found that these miRNAs were downregulated in PCOS, we agree with the conclusions presented in previous reports that these molecules are not related to BMI, androgen or AMH levels in PCOS.

miR-21-5p, miR-103a-3p and miR-376a-3p were positively correlated with androgen levels in our PCOS population. Murri et al. [40] also demonstrated a potential role of testosterone concentration in the positive regulation of miR-21-5p and miR-103. In the work of Naji et al. [35], decreased levels of miR-21-5p were found in the serum of women with hyperandrogenic PCOS, although the difference was not significant, whereas Butler et al. [37] found no relationship between androgen levels and this miRNA. In silico analysis of miR-21-5p demonstrates that it may be involved both in steroid biosynthesis and in androgen receptor signaling pathways. The conflicting findings reported in these studies may be related to the different roles miR-21-5p may play in the manifestation of hyperandrogenism. In agreement with our results, miR-103 has been previously positively correlated with testosterone levels [40] in PCOS, whereas to the best of our knowledge, our study is the first to demonstrate this correlation for miR-376a-3p. Interestingly, both of these miRNAs target the cyclin-dependent kinase 2 (CDK2) gene, which has been associated with phosphorylation of the androgen receptor [41].

Additionally, in women with type 2 diabetes, a condition related to PCOS [9,10], circulating miR-23a was suggested as a potential marker of this condition [42], whereas an increase in miR-21-5p was associated with the severity of the disease as well as with related vascular dysfunctions such as retinopathy and diabetic nephropathy [43,44]. This may be justified based on this miRNA role in angiogenesis, that culminate with the PI3K-AKT [45] signaling pathway, dysregulation, which is common in women with PCOS and insulin resistance [46]. Furthermore, another study reported a decrease in miR-21-5p in vascular smooth muscle cells after metformin treatment under diabetic conditions. Similarly, metformin treatment was also associated with decreased expression of miR-26a-5p and miR-222-3p in pancreatic stem cells and arterial mammary arteries of women with type 2 diabetes, respectively [47,48]. Furthermore, miR-222-3p levels were positively correlated with fasting glucose and with glucose 2 hours after OGTT in prediabetic patients, suggesting its potential use as a new biomarker for this condition [49]. Our in silico analysis showed that, together with miR-21-5p, miR-26a-5p and miR-222-3p targeted the PTEN gene, suggesting that the PI3K-AKT insulin signaling pathway should be further explored as a potential therapeutic target for insulin resistance in PCOS.

Other studies of populations of women with PCOS sought to assess the relationship between miRNAs and insulin resistance in the syndrome. Jiang et al. [50] found that miR-122, miR-194, and miR-193b were differentially expressed in glucose-tolerant and glucose-intolerant women with PCOS. In this study, these miRNAs were related to HOMA-IR. Increased circulating concentrations of miR-122 are also strongly associated with the risk of developing type 2 diabetes mellitus in the general population [51]. Song et al. [52] found that miR-6767 is downregulated in the serum of South Korean women with PCOS compared to healthy controls. This miRNA correlated positively with SHBG concentrations and negatively with glucose concentrations. In this current work, we did not find any correlation between the miRNAs studied and HOMA-IR or serum glucose levels.

The use of HOMA-IR instead of hyperinsulinemic euglycemic clamps, which are considered the gold standard for the diagnosis of insulin resistance, may have been a limiting factor in our study in the evaluation of the relationship between the miRNAs identified in this work and PCOS. However, HOMA-IR is the model most commonly used in clinical practice to assess insulin resistance, and the results obtained using this model show a good correlation with the standard [53]. In addition, most studies conducted in women with PCOS have used HOMA-IR to assess insulin resistance, making our results comparable [54].

Although in the current study we did not find significant correlations between most of the miRNAs studied and the main hormonal and metabolic features of PCOS, miR-21-5p, miR-23a-3p, miR-222-3p, and miR-376a-3p are highly expressed in ovarian tissue [5559] and are associated with biological functions and signaling pathways that participate in the pathophysiological processes associated with the syndrome. miR-21-5p was originally identified as an LH/hCG regulatory gene in a murine model of ovarian granulosa cells [60], in which it was shown to promote follicular survival during ovulation. Mase et al. [61] found that silencing miR-21-5p reduced ovulation rates. Another study found miR-21-5p to be an important regulator of granulosa cell apoptosis and proliferation in PCOS [59]. Moreover, miR-21-5p was not identified in ovarian theca cells or in oocytes [62,63], suggesting its specific action in the granulosa cells or other cumulus cells [64]. miR-376a-3p was also identified in granulosa cells in murine models, and associated with primordial follicular development and granulosa cell proliferation [56], whereas miR-23a-3p promoted follicular granulosa cell apoptosis [58]. On the other hand, miR-222-3p was found to promote the progression of PCOS [57].

PCOS ovaries display a disordered folliculogenesis, but the underlying mechanisms are not clear yet [6568]. Signaling pathways, as MAPK, FOXO, PI3K-AKT, mTOR, SMAD, and TGFß [46,6972], may be involved in this process. In the prediction of biological processes using bioinformatics tools in our study, the molecular functions and signaling pathways related to miR-21-5p, miR-23a-3p, miR-222-3p and miR-376a-3p included functions and pathways involved in the physiology of folliculogenesis as well as in the processes of cell proliferation and apoptosis. In one sense, our results may bring insights into the mechanisms of this phenomenon. However, the manner through those miRNAs regulates those pathways during abnormal folliculogenesis in PCOS requires further investigation that is beyond the scope of the present study.

Despite the limitations of this study, we hypothesize that the differential expression of miR-21-5p, miR-23a-3p, miR-222-3p and miR-376a-3p found in the serum of women with PCOS may be related to the dysfunctional proliferation of granulosa cells, especially at more advanced stages of follicular development, or to the proliferation of other cumulus cells and ovarian vascular smooth muscle cells. The identification of miRNAs in blood and other cellular fluids is relatively recent and has stimulated the development of research with two objectives: first, to determine whether these molecules are useful as biomarkers of disease and, second, to determine whether these molecules serve as a previously unknown form of intercellular and intertissue communication [73].

In summary, our preliminary study demonstrated that miR-21-5p, miR-23a-3p and miR-26a-5p were upregulated and that miR-103a-3p, miR-376a-3p, miR-19b-3p and miR-222-3p were downregulated in serum samples obtained from a well-characterized cohort of women with PCOS compared to healthy normo-ovulatory controls. We also demonstrated that serum levels of miR-21-5p, miR-103a-3p and miR-376a-3p are positively correlated with androgen levels in the PCOS population. These miRNAs, in combination, are related to pathways involved in insulin signaling, steroids and endothelial regulation as well as in folliculogenesis. Our study may open a new line of investigation into the role of these regulatory molecules in PCOS and in its metabolic and reproductive consequences. Furthermore, it may also bring insights into possible therapeutic tolls based on molecular strategies.

Acknowledgments

The authors thank the patients and individuals who participated in this study and the Grupo Fleury for assistance with the anti-Mullerian hormone assay analysis. The current article is part of the PhD thesis of Giovana De Nardo Maffazioli from Faculdade de Medicina da Universidade de Sao Paulo.

References

  1. 1. March WA, Moore VM, et al. The prevalence of polycystic ovary syndrome in a community sample assessed under contrasting diagnostic criteria. Hum Reprod. 2010;25(2):544–51. pmid:19910321
  2. 2. Lauritsen MP, Bentzen JG, et al. The prevalence of polycystic ovary syndrome in a normal population according to the Rotterdam criteria versus revised criteria including anti-Mullerian hormone. Hum Reprod. 2014;29(4):791–801. pmid:24435776
  3. 3. Lizneva D, Suturina L, et al. Criteria, prevalence, and phenotypes of polycystic ovary syndrome. Fertil Steril. 2016;106(1):6–15. pmid:27233760
  4. 4. Azziz R, Sanchez LA, et al. Androgen excess in women: experience with over 1000 consecutive patients. J Clin Endocrinol Metab. 2004;89(2):453–62. pmid:14764747
  5. 5. Yildiz BO, Bozdag G, et al. Prevalence, phenotype and cardiometabolic risk of polycystic ovary syndrome under different diagnostic criteria. Hum Reprod. 2012;27(10):3067–73. pmid:22777527
  6. 6. Teede HJ, Misso ML, et al. Recommendations from the international evidence-based guideline for the assessment and management of polycystic ovary syndrome. Hum Reprod. 2018.
  7. 7. Rotterdam EA-SPCWG. Revised 2003 consensus on diagnostic criteria and long-term health risks related to polycystic ovary syndrome. Fertil Steril. 2004;81(1):19–25.
  8. 8. Ehrmann DA. Polycystic ovary syndrome. N Engl J Med. 2005;352(12):1223–36. pmid:15788499
  9. 9. Maffazioli GDN, Lopes CP, et al. Prevalence of metabolic disturbances among women with polycystic ovary syndrome in different regions of Brazil. Int J Gynaecol Obstet. 2020;151(3):383–91. pmid:32931592
  10. 10. Ovalle F, Azziz R. Insulin resistance, polycystic ovary syndrome, and type 2 diabetes mellitus. Fertil Steril. 2002;77(6):1095–105. pmid:12057712
  11. 11. Hayes MG, Urbanek M, et al. Genome-wide association of polycystic ovary syndrome implicates alterations in gonadotropin secretion in European ancestry populations. Nat Commun. 2015;6:7502. pmid:26284813
  12. 12. Azziz R. Introduction: Determinants of polycystic ovary syndrome. Fertil Steril. 2016;106(1):4–5. pmid:27238627
  13. 13. Ilie IR, Georgescu CE. Polycystic Ovary Syndrome-Epigenetic Mechanisms and Aberrant MicroRNA. Adv Clin Chem. 2015;71:25–45. pmid:26411410
  14. 14. Xu N, Azziz R, et al. Epigenetics in polycystic ovary syndrome: a pilot study of global DNA methylation. Fertil Steril. 2010;94(2):781–3 e1. pmid:19939367
  15. 15. Li S, Zhu D, et al. Differential DNA methylation patterns of polycystic ovarian syndrome in whole blood of Chinese women. Oncotarget. 2016.
  16. 16. Bartel DP. Metazoan MicroRNAs. Cell. 2018;173(1):20–51.
  17. 17. Karolina DS, Tavintharan S, et al. Circulating miRNA profiles in patients with metabolic syndrome. J Clin Endocrinol Metab. 2012;97(12):E2271–6. pmid:23032062
  18. 18. Sang Q, Yao Z, et al. Identification of microRNAs in human follicular fluid: characterization of microRNAs that govern steroidogenesis in vitro and are associated with polycystic ovary syndrome in vivo. J Clin Endocrinol Metab. 2013;98(7):3068–79. pmid:23666971
  19. 19. Roth LW, McCallie B, et al. Altered microRNA and gene expression in the follicular fluid of women with polycystic ovary syndrome. J Assist Reprod Genet. 2014;31(3):355–62. pmid:24390626
  20. 20. Long W, Zhao C, et al. Characterization of serum microRNAs profile of PCOS and identification of novel non-invasive biomarkers. Cell Physiol Biochem. 2014;33(5):1304–15. pmid:24802714
  21. 21. Jiang L, Huang J, et al. MicroRNA-93 promotes ovarian granulosa cells proliferation through targeting CDKN1A in polycystic ovarian syndrome. J Clin Endocrinol Metab. 2015;100(5):E729–38. pmid:25695884
  22. 22. Wu HL, Heneidi S, et al. The expression of the miR-25/93/106b family of micro-RNAs in the adipose tissue of women with polycystic ovary syndrome. J Clin Endocrinol Metab. 2014;99(12):E2754–61. pmid:25243570
  23. 23. Deswal R, Dang AS. Dissecting the role of micro-RNAs as a diagnostic marker for polycystic ovary syndrome: a systematic review and meta-analysis. Fertil Steril. 2020;113(3):661–9 e2. pmid:32192599
  24. 24. Azziz R, Kintziger K, et al. Recommendations for epidemiologic and phenotypic research in polycystic ovary syndrome: an androgen excess and PCOS society resource. Hum Reprod. 2019;34(11):2254–65. pmid:31751476
  25. 25. WHO. Waist circunference and Waist/Hip Ratio http://wwwwhoint/nutrition/publications/obesity/WHO_report_waistcircumference_and_waisthip_ratio/en/. 2011.
  26. 26. Ferriman D, Gallwey JD. Clinical assessment of body hair growth in women. J Clin Endocrinol Metab. 1961;21:1440–7. pmid:13892577
  27. 27. Zhang A, Li D, et al. Islet beta cell: An endocrine cell secreting miRNAs. Biochem Biophys Res Commun. 2018;495(2):1648–54.
  28. 28. Calderari S, Diawara MR, et al. Biological roles of microRNAs in the control of insulin secretion and action. Physiol Genomics. 2017;49(1):1–10. pmid:27815534
  29. 29. Nigi L, Grieco GE, et al. MicroRNAs as Regulators of Insulin Signaling: Research Updates and Potential Therapeutic Perspectives in Type 2 Diabetes. Int J Mol Sci. 2018;19(12). pmid:30469501
  30. 30. Mononen N, Lyytikainen LP, et al. Whole blood microRNA levels associate with glycemic status and correlate with target mRNAs in pathways important to type 2 diabetes. Sci Rep. 2019;9(1):8887. pmid:31222113
  31. 31. Shah JS, Soon PS, et al. Comparison of Methodologies to Detect Low Levels of Hemolysis in Serum for Accurate Assessment of Serum microRNAs. PLoS One. 2016;11(4):e0153200. pmid:27054342
  32. 32. Calvano Filho CM, Calvano-Mendes DC, et al. Triple-negative and luminal A breast tumors: differential expression of miR-18a-5p, miR-17-5p, and miR-20a-5p. Tumour Biol. 2014;35(8):7733–41. pmid:24810926
  33. 33. Sathyapalan T, David R, et al. Increased expression of circulating miRNA-93 in women with polycystic ovary syndrome may represent a novel, non-invasive biomarker for diagnosis. Sci Rep. 2015;5:16890. pmid:26582398
  34. 34. Sorensen AE, Wissing ML, et al. MicroRNA Species in Follicular Fluid Associating With Polycystic Ovary Syndrome and Related Intermediary Phenotypes. J Clin Endocrinol Metab. 2016;101(4):1579–89. pmid:26771704
  35. 35. Naji M, Aleyasin A, et al. Differential Expression of miR-93 and miR-21 in Granulosa Cells and Follicular Fluid of Polycystic Ovary Syndrome Associating with Different Phenotypes. Sci Rep. 2017;7(1):14671. pmid:29116087
  36. 36. Xiong W, Lin Y, et al. Circulatory microRNA 23a and microRNA 23b and polycystic ovary syndrome (PCOS): the effects of body mass index and sex hormones in an Eastern Han Chinese population. J Ovarian Res. 2017;10(1):10. pmid:28193283
  37. 37. Butler AE, Ramachandran V, et al. microRNA Expression in Women With and Without Polycystic Ovarian Syndrome Matched for Body Mass Index. Front Endocrinol (Lausanne). 2020;11:206.
  38. 38. Sorensen AE, Udesen PB, et al. Hyperandrogenism and Metabolic Syndrome Are Associated With Changes in Serum-Derived microRNAs in Women With Polycystic Ovary Syndrome. Front Med (Lausanne). 2019;6:242. pmid:31737638
  39. 39. Jiang L, Li W, et al. Ciculating miRNA-21 as a Biomarker Predicts Polycystic Ovary Syndrome (PCOS) in Patients. Clin Lab. 2015;61(8):1009–15. pmid:26427146
  40. 40. Murri M, Insenser M, et al. Non-targeted profiling of circulating microRNAs in women with polycystic ovary syndrome (PCOS): effects of obesity and sex hormones. Metabolism. 2018. pmid:29410349
  41. 41. Jorda R, Buckova Z, et al. Selective inhibition reveals cyclin-dependent kinase 2 as another kinase that phosphorylates the androgen receptor at serine 81. Biochim Biophys Acta Mol Cell Res. 2018;1865(2):354–63. pmid:29157894
  42. 42. Yang Z, Chen H, et al. Serum miR-23a, a potential biomarker for diagnosis of pre-diabetes and type 2 diabetes. Acta Diabetol. 2014;51(5):823–31. pmid:24981880
  43. 43. Taheri M, Eghtedarian R, et al. Non-coding RNAs and type 2 diabetes mellitus. Arch Physiol Biochem. 2020:1–10. pmid:33155514
  44. 44. Olivieri F, Spazzafumo L, et al. MiR-21-5p and miR-126a-3p levels in plasma and circulating angiogenic cells: relationship with type 2 diabetes complications. Oncotarget. 2015;6(34):35372–82. pmid:26498351
  45. 45. Jiang Q, Han Y, et al. Ursolic acid induced anti-proliferation effects in rat primary vascular smooth muscle cells is associated with inhibition of microRNA-21 and subsequent PTEN/PI3K. Eur J Pharmacol. 2016;781:69–75. pmid:27085898
  46. 46. Diamanti-Kandarakis E, Dunaif A. Insulin resistance and the polycystic ovary syndrome revisited: an update on mechanisms and implications. Endocr Rev. 2012;33(6):981–1030. pmid:23065822
  47. 47. Bao B, Wang Z, et al. Metformin inhibits cell proliferation, migration and invasion by attenuating CSC function mediated by deregulating miRNAs in pancreatic cancer cells. Cancer Prev Res (Phila). 2012;5(3):355–64. pmid:22086681
  48. 48. Coleman CB, Lightell DJ Jr., et al. Elevation of miR-221 and -222 in the internal mammary arteries of diabetic subjects and normalization with metformin. Mol Cell Endocrinol. 2013;374(1–2):125–9. pmid:23648338
  49. 49. Sadeghzadeh S, Dehghani Ashkezari M, et al. Circulating miR-15a and miR-222 as Potential Biomarkers of Type 2 Diabetes. Diabetes Metab Syndr Obes. 2020;13:3461–9. pmid:33061506
  50. 50. Jiang L, Huang J, et al. Identification of several circulating microRNAs from a genome-wide circulating microRNA expression profile as potential biomarkers for impaired glucose metabolism in polycystic ovarian syndrome. Endocrine. 2016;53(1):280–90. pmid:26860517
  51. 51. Willeit P, Skroblin P, et al. Circulating MicroRNA-122 Is Associated With the Risk of New-Onset Metabolic Syndrome and Type 2 Diabetes. Diabetes. 2017;66(2):347–57. pmid:27899485
  52. 52. Song DK, Sung YA, et al. The Role of Serum MicroRNA-6767-5p as a Biomarker for the Diagnosis of Polycystic Ovary Syndrome. PLoS One. 2016;11(9):e0163756. pmid:27677182
  53. 53. Malita FM, Karelis AD, et al. Surrogate indexes vs. euglycaemic-hyperinsulinemic clamp as an indicator of insulin resistance and cardiovascular risk factors in overweight and obese postmenopausal women. Diabetes Metab. 2006;32(3):251–5. pmid:16799402
  54. 54. Luo Y, Cui C, et al. The role of miRNAs in polycystic ovary syndrome with insulin resistance. J Assist Reprod Genet. 2021;38(2):289–304. pmid:33405004
  55. 55. Hossain MM, Sohel MM, et al. Characterization and importance of microRNAs in mammalian gonadal functions. Cell Tissue Res. 2012;349(3):679–90. pmid:22842772
  56. 56. Zhang H, Jiang X, et al. microRNA 376a regulates follicle assembly by targeting Pcna in fetal and neonatal mouse ovaries. Reproduction. 2014;148(1):43–54. pmid:24686458
  57. 57. Huang X, She L, et al. MiR-222 promotes the progression of polycystic ovary syndrome by targeting p27 Kip1. Pathol Res Pract. 2019;215(5):918–23. pmid:30718101
  58. 58. Nie M, Yu S, et al. miR-23a and miR-27a promote human granulosa cell apoptosis by targeting SMAD5. Biol Reprod. 2015;93(4):98. pmid:26400397
  59. 59. Yu Y, Li G, et al. MicroRNA-21 regulate the cell apoptosis and cell proliferation of polycystic ovary syndrome (PCOS) granulosa cells through target toll like receptor TLR8. Bioengineered. 2021;12(1):5789–96. pmid:34516355
  60. 60. Carletti MZ, Christenson LK. MicroRNA in the ovary and female reproductive tract. J Anim Sci. 2009;87(14 Suppl):E29–38. pmid:18791135
  61. 61. Mase Y, Ishibashi O, et al. MiR-21 is enriched in the RNA-induced silencing complex and targets COL4A1 in human granulosa cell lines. Reprod Sci. 2012;19(10):1030–40. pmid:22573493
  62. 62. Xu YW, Wang B, et al. Differentially expressed micoRNAs in human oocytes. J Assist Reprod Genet. 2011;28(6):559–66. pmid:21647640
  63. 63. Virant-Klun I, Stahlberg A, et al. MicroRNAs: From Female Fertility, Germ Cells, and Stem Cells to Cancer in Humans. Stem Cells Int. 2016;2016:3984937. pmid:26664407
  64. 64. Assou S, Al-edani T, et al. MicroRNAs: new candidates for the regulation of the human cumulus-oocyte complex. Hum Reprod. 2013;28(11):3038–49. pmid:23904466
  65. 65. Stubbs SA, Stark J, et al. Abnormal preantral folliculogenesis in polycystic ovaries is associated with increased granulosa cell division. J Clin Endocrinol Metab. 2007;92(11):4418–26. pmid:17698906
  66. 66. Das M, Djahanbakhch O, et al. Granulosa cell survival and proliferation are altered in polycystic ovary syndrome. J Clin Endocrinol Metab. 2008;93(3):881–7. pmid:18073308
  67. 67. Maciel GA, Baracat EC, et al. Stockpiling of transitional and classic primary follicles in ovaries of women with polycystic ovary syndrome. J Clin Endocrinol Metab. 2004;89(11):5321–7. pmid:15531477
  68. 68. Chang RJ, Cook-Andersen H. Disordered follicle development. Mol Cell Endocrinol. 2013;373(1–2):51–60. pmid:22874072
  69. 69. Haouzi D, Assou S, et al. Altered gene expression profile in cumulus cells of mature MII oocytes from patients with polycystic ovary syndrome. Hum Reprod. 2012;27(12):3523–30. pmid:22951915
  70. 70. Yang JL, Zhang CP, et al. Testosterone induces redistribution of forkhead box-3a and down-regulation of growth and differentiation factor 9 messenger ribonucleic acid expression at early stage of mouse folliculogenesis. Endocrinology. 2010;151(2):774–82. pmid:20032061
  71. 71. Knight PG, Glister C. TGF-beta superfamily members and ovarian follicle development. Reproduction. 2006;132(2):191–206. pmid:16885529
  72. 72. Li Q, Pangas SA, et al. Redundant roles of SMAD2 and SMAD3 in ovarian granulosa cells in vivo. Mol Cell Biol. 2008;28(23):7001–11. pmid:18809571
  73. 73. Mori MA, Ludwig RG, et al. Extracellular miRNAs: From Biomarkers to Mediators of Physiology and Disease. Cell Metab. 2019;30(4):656–73. pmid:31447320