Introduction
Polycystic ovary syndrome (PCOS) is a common endocrine and metabolic dysfunction condition in the women of reproductive age, resulting in irregular menstruation, hyperandrogenism, infertility, and insulin resistance(IR) [
1]. Patients with PCOS are at higher risk of diabetes [
2], cardiovascular disease [
3], and endometrial cance [
4]. There is no specific therapy for PCOS, only symptomatic treatment such as lifestyle management, taking combination oral contraceptives to regulate menstrual cycles and ameliorate hyperandrogenism, using insulin sensitizers to alleviate IR, taking clomiphene or letrozole to promote ovulation in the case of anovulatory infertility [
5‐
7].
The etiology of PCOS is not yet entirely understood. It is widely assumed that PCOS is caused by a combination of environmental and genetic factors. PCOS is associated with hypothalamic-pituitary-ovarian axis (HPOA) neuroendocrine abnormalities, hyperandrogenemia, IR, chronic inflammation, and circadian rhythm disorder [
5,
8]. At the same time, PCOS appears to be a highly genetic and polygenic disease. Despite the fact that candidate gene association studies and genome-wide association studies have identified several susceptibility genes and single-nucleotide polymorphism loci significantly associated with PCOS, such as follicle stimulating hormone receptor gene (
FSHR, rs2268361, rs2349415), insulin receptor (
INSR, rs2059807), it is still difficult to explain the complexity of PCOS etiology and clinical manifestations [
9‐
11]. Recently, epigenetic factors have received a lot of attention in the pathogenesis of PCOS. As far as we know,women with PCOS have different epigenetic regulation, including DNA methylation, histone acetylation, and changes in non coding RNA content [
12,
13]. The changes of DNA methylation in peripheral and umbilical cord blood, ovary and adipose tissue of PCOS patients indicate that this epigenetic modification is related to the pathogenesis of the disease [
14]. Perhaps, these defects in DNA methylation promote the disorder of genes involved in inflammation, hormone synthesis and signal transduction, as well as glucose and lipid metabolism [
15‐
19]. And the research on the role of DNA methylation in the pathogenesis of PCOS has just begun.
N
6-methyladenosine (m
6A) is the most prevalent internal modification of mRNAs, which is reversible and dynamically regulated by methyltransferase complexes (writers) and demethylases (erasers), and recognized by m
6A binding proteins (readers) [
20]. m
6A modification regulates post-transcriptional expression of m
6A-tagged genes by participating in RNA metabolism such as pre-mRNA splicing, mRNA translation, nuclear export, mRNA decay, and non-coding RNA biogenesis [
21]. Recently, studies have revealed the role of m
6A modifications and their protein machines in oogenesis and in female reproductive tumors, as well as in other female reproductive diseases [
22]. It has been demonstrated that the m
6A modification level was increased, and several m
6A modulators were dysfunction in PCOS patients [
23]. In oogenesis, lack of YTHDF2 leads to failure of m6A modified mRNA degradation, which affecting the oocyte quality [
24,
25]. The m
6A proteins are also involved in ovulation, and the m
6A writers play an important role in oogenesis, but whether they can be used as therapeutic targets for abnormal ovulation remains to be further investigated. Nevertheless, little attention has been paid to the molecular mechanisms of m
6A modification in PCOS.
In this study, we aim to investigate the altered m6A modification landscape of mRNAs in the ovaries of PCOS mice using the epitranscriptomic microarray, and to preliminarily explore the potential signal pathways involved in the PCOS process though KEGG analysis.
Materials and methods
The study was approved by the Institutional Review Board (No.2021-S087), and animal experiments were in accordance with the Guide for the Care and Use of Laboratory Animal by International Committees of The Third Xiangya Hospital of Central South University.
Establishment of a mouse model of PCOS
Female C57BL/6J mice aged 7 weeks were acquired from SJA Laboratory Animal Co. Ltd (Hunan China), and adaptively fed for 1 week on 12 h light/ 12 h dark cycle at room temperature (24 ± 3 °C) with a humidity of 45 ± 2%. During the feeding period, vaginal smears were performed daily at 8:00 a.m. to observe the estrous cycle of mice. Ten mice with normal estrous cycle were selected and randomly divided in two groups: the control and PCOS group. In the PCOS group, 6 mg/(100 g·d) dehydroepiandrosterone (DHEA) and 0.2 ml injectable soybean oil were injected subcutaneously into mice daily for 20 consecutive days. Similarly, mice in the control group were injected with 0.2 ml injectable soybean oil daily for 20 days.
Subsequently, five mice with continuous keratosis of vaginal epithelial cells in PCOS group and five mice in control group were randomly sacrificed on the first day. The two ovaries randomly selected mice from each group were embedded in paraffin and sliced for hematoxylin and eosin (HE) staining to observe the morphological changes and confirm the induction of PCOS. And the three remaining mice in each group, the left ovaries were preserved at -80 °C for following N6-methyladenosine detection, the right ovaries were preserved at -80 °C for following western blot.
RNA extraction and quality control
To isolate total RNA, ovary tissues were cut into small pieces and homogenized in TRIzol reagent before being quantified using a NanoDrop ND-1000 (Thermo Fisher Scientific, Waltham, MA, USA). Supplementary Table S
1 presents the quantification and quality of RNA.
m6A immunoprecipitation
Total RNA (1-3ug) mixed with m6A spike-in control was immunoprecipitated with anti-m6A rabbit polyclonal antibody (Synaptic Systems, Göttingen, Germany) at 4 °C for 2 h. Dynabeads™ M-280 Sheep Anti-Rabbit IgG suspension (20 μL per sample) (Ivitrogen) was blocked with 0.5% bovine serum albumin (BSA) at 4 °C for 2 h, washed three times with IP buffer (300 μL) for 5 min, and resuspended in the prepared RNA-antibody mixture. The RNA was bound to the m6A-antibody beads for 2 h at 4 °C, then the beads were washed with IP buffer (500 μL, three times), followed by Wash buffer (500 μL, twice). In this case, the adsorbed RNA was eluted with Elution buffer (200 μL) at 50 °C for 1 h. The immunoprecipitated (IP) RNA and supernatant (Sup) RNA were extracted by acid phenol–chloroform and ethanol precipitation.
Labeling and hybridization
The IP RNAs and Sup RNAs were mixed with an equal amount of calibration spike-in control RNA, amplified separately and labeled with Cy3 (for Sup) and Cy5 (for IP) using Arraystar Super RNA Labeling Kit (Arraystar). The synthesized cRNAs was purified by the RNeasy Mini Kit (QIAGEN), and the concentration and specific activity of cRNAs were detected by the NanoDrop ND-1000 (Thermo Fisher Scientific) (Supplementary Table S
2). 2.5 μg of Cy3 and Cy5-labeled cRNAs were mixed, added with 5 μL 10 × Blocking Agent and 1 μL of 25 × Fragmentation Buffer. It was heated to 60℃ for 30 min, and then mixed with 25 μL 2 × Hybridization buffer. 50 μl of hybridization solution was injected into the gasket slide and assembled on the m
6A-mRNA epitranscriptome microarray slide. The slides were incubated at 65 °C for 17 h in an Agilent Hybridization Oven (Agilent, CA, USA). The hybridized arrays were washed, fixed, and scanned with an Agilent scanner G2505C (Agilent).
Epitranscriptomic microarray data analysis
To analyze acquired array images, Agilent Feature Extraction software (version 11.0.1.1) was used. The raw IP and Sup intensities were normalized to the log2-scaled Spike-in RNA intensity average. Following Spike-in normalization, probe signals with Present (P) or Marginal (M) QC flags were retained in at least three of six samples for further m6A methylation, quantity, and expression level analyses. The m6A methylation level was calculated for the percentage of modification based on the IP and Sup normalized intensities, and the m6A quantity was calculated for the m6A methylation amount. The expression level was calculated by adding the IP and Sup normalized intensities, and an additional quantile normalization method from the limma package was used to normalize the RNA expression level between arrays before flagging probes.
Methylated RNA immunoprecipitation-quantitative PCR (MeRIP-qPCR)
The m
6A epitranscriptomic microarray data was validated using IP (
n = 3, each group). MeRIP-qPCR was then used to quantify the RNA enrichment via 2
−ΔΔct analysis. Supplementary Table S
3 describes the primer used. Furthermore, the mRNA m
6A sites were predicted using the sequence-based RNA adenosine methylation site predictor (SRAMP) program (
https://www.cuilab.cn/sramp) [
26].
Western blot
The protein concentration of right ovary tissue was determined using the Bradford method after protein extraction (M&C Gene Technology Ltd.). SDS–polyacrylamide gel electrophoresis (SDS-PAGE) was used to separate the protein samples, which were then transferred to a polyvinylidene difluoride membrane. Antibodies of interest were used to probe the membranes. The antibodies used were as follows: phosphatidylinositol 3-kinase (PI3K) (1:1000; ABclonal), phosphorylated PI3K (p-PI3K) (1:1000; ABclonal), protein kinase B (AKT) (1:1000; Proteintech), phosphorylated AKT (ser473) (p-AKT) (1:1000; Proteintech), skeletal muscle and kidney-enriched inositol polyphosphate 5-phosphatase (SKIP) (1:1000; Proteintech), and β-actin (1:1000; Abcam).
Statistical analysis
Filtering with the fold change ≥ 1.5 and statistical significance thresholds (P < 0.05) revealed differentially m6A-methylated or differentially expressed RNAs between two comparison groups. Hierarchical clustering was carried out using the R software (version 4.02). To perform GO analysis, the topGO package in the R environment for statistical computing and graphics was used, and the Fisher's exact test was used to perform pathway analysis. The western blot data were presented as mean ± SD and compared between groups using the Student's t test or the Mann–Whitney U test. P < 0.05 was regarded as significant. SPSS version 21.0 was used for all statistical analysis (SPSS 21.0, Inc.,Chicago, IL, USA).
Discussion
Polycystic ovary syndrome (PCOS) has numerous adverse effects on women. However, the pathogenesis of PCOS is currently unclear, and there is no specific therapy for PCOS. Given the continuously increasing incidence of PCOS in recent years, it is important to explore the mechanisms of PCOS development and to develop effective treatments for patients with PCOS. PCOS is commonly associated with aberrant DNA methylation, and several genes are epigenetically dysregulated, and are associated with the pathological consequences of PCOS and metabolic comorbidities [
27]. However, the methylation status of specific genes and the extent to which genes are dysregulated in terms of methylation patterns are unknown.Due to the reversibility of epigenetic modifications, "druable" regions can be screened to target or correct abnormalities in gene expression, so PCOS methylation promises the development of novel chromatin methylation therapies targeting PCOS [
28].Herein, we used epitranscriptomic microarray for the first time to investigate the altered m
6A modification of mRNAs and preliminarily explore the potential molecular mechanisms of m
6A modification in the mouse model of PCOS induced by hyperandrogenism, and to screen new molecular targets and develop effective targets for treating PCOS or inhibiting its progression.
According to our microarray results, the m
6A modification levels of mRNAs were increased, and the m
6A ‘readers’ (
Ythdf3,
Hnrnpa2b1) are overexpressed in PCOS mice ovaries. A similar methylation trend was confirmed using MeRIP sequencing (MeRIP-seq) in luteinized granulosa cells (GCs) of PCOS patients [
23]. In contrast to MeRIP-seq, epitranscriptomic microarray can determined the percentage of modified and unmodified RNA of each transcript [
29]. As a result, 437 RNAs with differentially expressed and differentially methylated levels were identified. Further bioinformatics analysis revealed that the m
6A modification may mainly participate in the insulin signaling pathway and type II diabetes mellitus pathway, which were closely related to IR and secondary hyperinsulinemia [
30]. The PI3K/AKT pathway and the mitogen activated protein kinase (MAPK) pathway were two major insulin-related signal transduction pathway, regulating the glucose metabolism, cell proliferation and differentiation, respectively [
30,
31]. In mouse cumulus-oocyte complexes, activated PI3K/AKT signaling can increase glucose uptake by mediating the translocation of GLUT4 to the GCs membrane, which provides energy substrate for follicular development [
32]. The dysfunction of PI3K/AKT signaling is not only linked to IR, inflammation, and oxidative stress, but it also inhibits proliferation and promotes apoptosis, all of which may contribute to PCOS [
33,
34]. Similarly, the inhibition of PI3K/AKT pathway in PCOS mice was confirmed in our study by assessing p-AKT expression levels.
Notably, the activity of upstream and downstream factors in the PI3K/AKT pathway were opposite. In contrast to p-AKT, p-PI3K expression levels were elevated in the PCOS group, which may be related to inositol phosphate metabolism pathway, another enriched pathway. SKIP, a PIP3 5-phosphatase, was localized at endoplasmic reticulum under resting conditions. Insulin stimulation induced its translocation to the plasma membrane, and binding with activated p21-activated protein kinase 1 (PAK1), thereby activating SKIP’s PIP3 phosphatase activity. The rapid and efficient hydrolysis of PIP3, resulting in decreased AKT2 phosphorylation, inhibits membrane ruffle and GLUT4 translocation, thus negatively regulates insulin signaling in skeletal muscle [
35,
36]. As expected, SKIP protein was confirmed to be significantly overexpressed in PCOS mice in this study. Furthermore, due to the non-statistically significant difference in mRNA expression levels between two groups, SKIP overexpression appears to be linked to higher m
6A modification in PCOS mice. The m
6A sites in humans and mice are highly conserved, and mainly enriched in the 3’untranslated region (UTR) and around stop codons [
37]. However, Shen Zhang et al. discovered increased m
6A peaks in the coding sequence and transcription start regions, but less prominent enrichment near stop codons in PCOS patients' GCs compared to controls [
23]. Similarly, the m
6A sites of
Skip mRNA was also mainly predicted in the coding sequence by SRAMP program in this study [
26]. In addition, previous research revealed that the METTL3-induced m
6A modification in the coding sequence may alleviate ribosome stalling and thus improve mRNA translation in acute myeloid leukemia [
38]. And in Arabidopsis thaliana, increased mRNA expression levels were associated with a modification change around the start codon [
39]. Therefore, the altered m
6A modification of
Skip mRNA may explain its translation upregulation. Therefore, it is possible that the higher m
6A modification of
Skip mRNA result in the overexpression of SKIP protein, then reversed the activation of p-PI3K on the downstream AKT.
In addition, Fc epsilon RI signaling pathway and GnRH secretion pathway, another two enriched pathways, may indicated that the m
6A modification also participant in the etiological mechanism of chronic inflammation and neuroendocrine in PCOS. Increasing number of studies demonstrated that low-grade chronic inflammation can induce IR, obesity, and hyperandrogenemia through related pathways, leading to ovulation disorders in PCOS [
40‐
42]. The HPOA plays an important role in the regulation of female reproductive endocrine as an integrated and coordinated neuroendocrine system. Hypothalamic gonadotropin-releasing hormone (GnRH) neurons secrete GnRH, which regulates the secretion of gonadotropins FSH and LH, then regulate the secretion of sex hormones and reproductive function [
43]. The GnRH neuronal firing activity was identified to be increased in PCOS mice induced by prenatal androgenization compared to normal mice, which may be closely linked to PCOS development [
44,
45].
Inevitably, there were several limitations of this study. First, several acknowledged inhibiting factors for PI3K/AKT pathway were not investigated, such as Phosphatase and tensin homolog (PTEN) and c-Jun N-terminal kinase (JNK). Then, we did not further validate our hypothesis that the hypermethylation of SKIP mRNAs enhances its expression, then inhibits the PI3K/AKT signaling pathway.
In conclusion, our study demonstrates that the altered m6A modification of mRNAs might play a critical role in PCOS process. And we emphasized the changes in the activity of upstream and downstream factors in the PI3K/AKT signaling pathway. Moreover, the role of m6A modification of Skip mRNA in the pathogenesis of PCOS warrants further studies.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.