Obesity-induced inflammatory miR-133a mediates apoptosis of granulosa cells and causes abnormal folliculogenesis

Abstract

Obesity has been reported to promote disordered folliculogenesis, but the exact molecular mechanisms are still not fully understood. In this study, we find that miR-133a is involved in obesity-induced follicular development disorder. After feeding with a high-fat diet (HFD) and fructose water for nine weeks, the mouse body weight is significantly increased, accompanied by an inflammatory state and increased expression of miR-133a in the adipose tissues and ovaries as well as accelerated follicle depletion. Although miR-133a is increased in the fat and ovaries of HFD mice, the increased miR-133a in the HFD ovaries is not derived from exosome transferred from obese adipose tissues but is synthesized by ovarian follicular cells in response to HFD-induced inflammation. In vivo experiments show that intrabursal injection of miR-133a agomir induces a decrease in primordial follicles and an increase in antral follicles and atretic follicles, which is similar to HFD-induced abnormal folliculogenesis. Overexpression of miR-133a modestly promotes granulosa cell apoptosis by balancing the expression of anti-apoptotic proteins such as C1QL1 and XIAP and pro-apoptotic proteins such as PTEN. Overall, this study reveals the function of miR-133a in obesity-induced ovarian folliculogenesis dysfunction and sheds light on the etiology of female reproductive disorders.

Introduction

Obesity is a growing epidemic that induces various disorders, such as cardiovascular diseases, type 2 diabetes, and fertility-related disorders [1]. The adverse effects of obesity on female reproduction include ovulation disorders, oligomenorrhea, and an increased rate of miscarriage [ 2– 4] . Polycystic ovary syndrome (PCOS), a reproductive disorder characterized by oligomenorrhea, hyperandrogenism and polycystic ovaries, may be exacerbated by obesity [5]. Studies in humans have found that weight management improves ovarian dysfunction in women with PCOS [6]. Additionally, different animal models have provided evidence of the relationship between obesity and ovarian dysfunction. In a rat model, obesity induced by a high-fat diet accelerates follicle development and loss [7]. In a murine model, obesity alters ovarian folliculogenesis, impacting proinflammatory and steroidogenic signaling [8]. Moreover, cell cycle arrest and excessive apoptosis of granulosa cells during follicular development are induced by an obesogenic diet [9]. These results highlight the deleterious effects of obesity on ovarian function, but the exact molecular mechanisms are still not fully understood.

MicroRNAs (miRNAs) are a family of small noncoding RNAs that cause mRNA degeneration or gene silencing by binding to target genes. They are critical regulators of various physiological processes, such as apoptosis, development, and inflammation. In previous studies, miRNAs have been reported to play crucial roles in ovarian development and function [ 10– 13] . However, little attention has been paid to whether miRNA plays a role in obesity-impaired ovarian function.

miR-133 is one of the most studied miRNAs and was initially found to be specifically expressed in cardiac and skeletal muscle [14]. Recent studies have found that miR-133 is also involved in the development of adipose tissue and its associated metabolic regulation [ 15– 18] . The miR-133 family contains two members, miR-133a and miR-133b, which share an almost identical sequence but are processed from different genes on different chromosomes [19]. Both miR-133a and miR-133b are significantly highly expressed in the adipose tissue of PCOS patients [20], and miR-133a was found to be upregulated in ovarian granulosa cells of obese women with PCOS compared with the PCOS group and the control group [21]. Based on the conclusion that obesity aggravates PCOS [5], miR-133a may be involved in regulating the ovarian pathology of obese PCOS. A study on miRNAs involved in goat follicular development showed that the expression of miR-133a-3p is increased, which may affect goat follicular development [22]. Furthermore, miR-133a was found to inhibit ovarian cancer cell proliferation by targeting IGFR1 [23], suggesting that miR-133a may be involved in regulating ovarian pathology. Given these previous results, we hypothesized that miR-133a may play a role in obesity-impaired ovarian function.

In this study, the relationships between obesity, miR-133a, and folliculogenesis were assessed in a mouse model of obesity. We analysed the expression of miR-133a in the adipose tissue and ovaries of obese mice induced by a high-fructose and high-fat diet (HFD) and investigated the role of miR-133a in follicular development disorders and the mechanism involved in HFD-induced obesity.

Materials and Methods

Animals

Female C57BL/6 mice were purchased from Guangdong Medical Laboratory Animal Centre (Guangzhou, China), maintained at 22 to 24°C with 14-h light and 10-h dark cycles, and given food and water ad libitum. All procedures were carried out following the Guidelines for the Care and Use of Laboratory Animals, and all animal experimental protocols were approved by the Animal Experimental Ethics Committee of Jinan University (protocol code IACUC-20221108-08).

Diet-induced obesity mouse model

Three-week-old female mice were fed with a negative control diet (NCD) or a high-fat diet (HFD) for nine weeks. The HFD consisted of a high-fat feed (60 kcal% fat; Research Diets, New Brunswick, USA) and 20% fructose (Macklin, Shanghai, China). The mice were weighed weekly. After feeding for nine weeks, mice were anaesthetized with 0.67% pentobarbital natrium and the mesenteric adipose tissue (mWAT), peri-ovarian adipose tissue (POAT) and ovaries were collected. The unilateral ovary was fixed in 4% paraformaldehyde (PFA) and embedded in paraffin. The contralateral ovary, mWAT and POAT were stored at –80°C for further RNA or protein assays. In addition, mWAT and POAT were also used for exosome isolation. To obtain the ruptured ovary sample, the bursae of the unilateral ovary were scratched with a needle and washed in RNase-free PBS. The contralateral ovary was dissected as an intact control. The ovaries were stored at –80°C for further RNA assays.

LPS-induced systemic inflammation mouse model

Four-week-old female mice were divided randomly into cage and received a single intraperitoneal injection of 100 μL lipopolysaccharide (100 μg/mL, LPS; Sigma-Aldrich, St Louis, USA) or PBS once a day for five consecutive days. Then mice were anaesthetized with 0.67% pentobarbital natrium, and the ovaries were collected 6 h after the last injection and stored at –80°C for further RNA assays.

miR-133a agomir intrabursal injection

miR-133a agomir (sense: 5′-UUUGGUCCCCUUCAACCAGCUG-3′; antisense: 5′-CAGCYGGUUCAAGGGGACCAAA-3′) and micrON™ miRNA agomir NC (sense: 5′-UUUGUACUACACAAAAGUACUG-3′; antisense: 5′-CAGUACUUUUGUGUAGUACAAA-3′) were obtained from RiboBio (Guangzhou, China). Three-week-old female mice were anaesthetized with 0.67% pentobarbital natrium using intraperitoneal injection, and 0.8 nmol of miR-133a agomir or miRNA agomir NC was injected into the bursa of the ovary as described in a previous study [24]. Two weeks after injection, the five-week-old mice were sacrificed, and the ovaries were collected and weighed. The ovaries were stored at –80°C for further RNA or protein assays or fixed in 4% PFA and embedded in paraffin.

Ovarian morphology and follicle counting

Ovaries were fixed with 4% PFA for 24 h, rinsed with PBS, dehydrated in ethanol, and embedded in paraffin. The ovaries were cut into 5-μm continuous sections. The sections were then stained with hematoxylin and eosin for ovarian morphological analysis and follicle counting as previously described [24]. Every section was counted for preantral follicles and small atretic follicles, and every 5th section was counted for antral and large atretic follicles. A follicle that was roughly round in appearance with an intact, nonfragmenting oocyte was scored as an intact antral follicle. Only follicles with clearly stained oocyte nuclei were counted. Two independent individuals counted sections for comparison.

TUNEL staining

To further confirm the apoptotic level of the ovaries, ovaries from each group were collected and sectioned for TUNEL staining using a fluorescence TUNEL kit (Roche Diagnostics, Indianapolis, USA) according to the manufacturer’s instructions. Briefly, after deparaffinization and hydration, two slices near the maximum section from each ovary were incubated with a TUNEL reaction mixture for 1 h at 37°C in a humidified chamber. The sections were then washed with PBS, counterstained with DAPI, and visualized under a Leica DFC300 FX fluorescence microscope (Leica, Weztlar, Germany). The percentage of TUNEL-positive signals in the entire section was analysed using ImagePro Plus 6.0 (Media Cybernetics, Rockville, USA).

Reverse transcription-polymerase chain reaction (RT-PCR) and quantitative real-time PCR

Total RNA from each sample was extracted with TRIzol (Invitrogen, Carlsbad, USA) and treated with DNase I (Takara, Dalian, China) to eliminate genomic DNA contamination. To detect the expression of mature miR-133a and precursor miR-133a (premiR-133a), a miDETECT A Track miRNA qRT-PCR Starter Kit (RiboBio) was used to synthesize cDNA, and qPCR assays were performed according to the manufacturer’s instructions. Tissue-, cellular- and exosome-containing miRNAs were detected. The level of miRNA expression was normalized to the level of U6. The primers for miR-133a, premiR-133a, and U6 used in the experiment were synthesized by RiboBio. Synthesis of cDNA and qPCR were performed as previously outlined [25]. The expression of primary miR-133a (primiR-133a) was checked by RT-PCR and electrophoresis of PCR products on 2% agarose gels. The specific band intensities were digitally quantified using ImageJ software (NIH, Bethesda, USA), and the amount was normalized to Rps16. The primer sequences used for qPCR and RT-PCR are listed in Table 1. The primer sequences of inflammation-related genes were described by Xiao et al. [26].

Table 1 The sequences of the primers used in this study

Gene	Forward primer (5′→3′)	Reverse primer (5′→3′)
Ccl2	GGGATCATCTTGCTGGTGAA	AGGTCCCTGTCATGCTTCTG
Ccl3	GTGGAATCTTCCGGCTGTAG	ACCATGACACTCTGCAACCA
Ccl4	GCTCTGTGCAAACCTAACCC	GAAACAGCAGGAAGTGGGAG
Cxcl1	CTTGACCCTGAAGCTCCCTT	AGGTGCCATCAGAGCAGTCT
Cxcl10	CCTATGGCCCTCATTCTCAC	CTCATCCTGCTGGGTCTGAG
Tnfα	CATCTTCTCAAAATTCGAGTGACAA	CCAGCTGCTCCTCCACTTG
Il6	CACAGAGGATACCACTCCCAACA	TCCACGATTTCCCAGAGAACA
Nos2	GTGGTGACAAGCACATTTGG	AAGGCCAAACACAGCATACC
C1ql1	GCAGACCAGAACTACGACTATGCCA	AAGGATGAAGAGCCACGGATGA
primiR-133a	AAAGAGCATTTAACCTGTTT	GCTGTCCATGTGTAATCAAT
Rps16	AGGTCTTCGGACGCAAGAAA	TTGCCCAGAAGCAAAACAG

Ligation-mediated polymerase chain reaction (LMPCR)

The experimental steps of ligation-mediated polymerase chain reaction (LMPCR) were performed as in previous reports [ 27, 28] . Genomic DNA for examination by LMPCR was purified from granulosa cells treated with the Omega DNA extraction kit (Omega Biotek Inc, Norcross, USA) following the manufacturer’s instructions and then eluted and diluted with TE (10 mM Tris-HCl pH 8.5, 0.5 mM EDTA) at 25°C, quantified spectrophotometrically and stored at –80°C. Annealing/ligation reactions comprised 19.68 μL sample DNA (156 pg to 160 ng maximum), 0.96 μL each of oligonucleotides 24-mer DHApo1 and 12-mer DHApo2 (both at stock 0.005 nM in TE) and 2.4 μL of 10× ligation buffer (New England BioLabs, Ipswitch, USA) to make a final volume of 24 μL. The sequences of DHApo1 (5′-AGCACTCTCGAGCCTCTCACCGCA-3′) and DHApo2 (5′-TGCGGTGAGAGG-3′) are listed in Table 1. Oligonucleotides were annealed to form blunt-end, partially double-stranded linkers by stepwise cooling from 55°C to 15°C at 5°C/ 8 min increments and then at 10°C/ 20 min increments using a thermal cycler. At the 10 min point of 10°C, the program was paused, and 2.4 U T4 DNA ligase (diluted from 5 U/μL to 1 U/μL in water; New England BioLabs) was added and mixed, and the temperature program continued at 10°C/ 10 min and then increased to 16°C/ 16 h for ligation. Postligation products were diluted to 80 μL in TE. Twenty-five microliters of LMPCRs contained 7.5 μL diluted ligation products, 3.15 μL additional 10 μmol DHApo1, 2.5 μL 10× Taq DNA polymerase buffer (Takara), 2 μL dNTP (2.5 mM; Takara), 2 μL MgCl ₂(25 mM; Takara), and 0.5 μL rTaq DNA polymerase (5 U/μL; Takara). Samples were first heated to 94°C, increased to 72°C for 4 min, and then amplified for 25 cycles of 1 min at 94°C and 3 min at 72°C. PCR products (12 μL) were analysed by electrophoresis using 2% agarose gels.

Western blot analysis

Western blot analysis was performed as previously described [24]. Briefly, total protein lysates obtained from the ovaries, adipose tissues, cells, and exosomes in the presence of Halt Protease & Phosphatase Inhibitor Cocktail (Pierce, Rockford, USA) were subjected to 12% SDS-PAGE on a Tris-acrylamide Minigel. The separated protein was electrotransferred to polyvinylidene fluoride membranes (Merck Millipore, Billerica, USA) in a semidry blotting chamber (Bio-Rad Laboratories, Hercules, USA). Membranes were blocked with 5% BSA in a Tris-buffered solution (pH 7.6) containing 0.05% Tween 20 and probed with the antibodies listed in Table 2. Signals were detected with the Chemiluminescent HRP Substrate (Merck Millipore) and visualized with the GeneGenius Bioimaging System (Syngene, Cambridge, UK). The specific band intensities were digitally quantified using ImageJ software, and the amount of protein was normalized to β-tubulin. The C1QL1 antiserum used in this experiment was prepared using the C1QL1 globular domain through prokaryotic expression as the antigen to immunize New Zealand rabbits. It has been experimentally verified that it can specifically recognize C1QL1 [25].

Table 2 Antibodies used in western blot analysis

Antibodies target	Host	Catalog No.	Supplier	Dilution
Primary antibodies
Anti-PARP1	Rabbit	DF7198	Affinity	1:1000
Anti-Caspase-3	Rabbit	14220	Cell Signaling Technology	1:1000
Anti-PTEN	Rabbit	9552	Cell Signaling Technology	1:1000
Anti-XIAP	Mouse	610716	BD Science	1:1000
Anti-TSG101	Mouse	sc-7964	Santa Cruz	1:1000
Anti-CD63	Rabbit	bs-1523R	Bioss	1:1000
Anti-β-Tubulin	Rabbit	2128	Bioss	1:1000
Secondary antibodies
HRP-linked anti-rabbit IgG	Goat	7074	Cell Signaling Technology	1:2000
HRP-linked anti-mouse IgG	Horse	7076	Cell Signaling Technology	1:2000

Experimental treatment of granulosa cells

Granulosa cells were isolated from pregnant mare serum gonadotropin (PMSG)-primed animals as previously described [24]. Briefly, 3-week-old mice were intraperitoneally (IP) injected with 10 IU PMSG (Ningbo Secondary Hormone Factory, Ningbo, China) for 36 h to stimulate follicle growth. The ovaries were then collected, cleaned and incubated in DMEM/F12 medium to facilitate separation into individual cells. Granulosa cells were released from the follicles by poking the follicle with a 28-gauge needle. The cell suspensions were filtered through a 40-μm nylon mesh to remove oocytes and primordial follicles. The cells were then seeded in DMEM/F12 medium supplemented with 10% FBS, 100 U/mL penicillin, and 100 μg/mL streptomycin for subsequent experiments.

To analyse the effects of LPS on miR-133a expression in granulosa cells, after the cells reached 90% confluence, they were treated with 1 μg/mL LPS (Sigma-Aldrich) for 6 h. The treated cells were harvested and stored at –80°C for further RNA assays.

To analyse the effect of overexpression of miR-133a on apoptosis of granulosa cells, miR-133a mimic (sense: 5′-UUUGGUCCCCUUCAACCAGCUG-3′; antisense: 5′-CAGCYGGUUCAAGGGGACCAAA-3′) or miRNA mimic negative control (sense: 5′-UUUGUACUACACAAAAGUACUG-3′; antisense: 5′-CAGUACUUUUGUGUAGUACAAA-3′) (50 nM; RiboBio) were transfected into granulosa cells with a riboFECT CP Transfection Kit (RiboBio). For western blot analysis, cells were harvested 48 h after transfection. In some assays, the cells were cultured in DMEM/F12 with 2% FBS to induce starvation. For RNA assays, the cells were harvested 24 h after transfection. For LMPCR, the cells were divided into four groups: miR-133a mimic group, miRNA mimic negative control group, normal control group (without treatment), and starved group (as a positive control of apoptosis, cultured in DMEM/F12 with 2% FBS). The cells in each group were harvested 24 h after treatment.

Isolation of exosomes from POAT and mWAT

POAT and mWAT were dissected and washed with cold sterilized PBS to remove blood and connective tissue. Minced pieces were weighed and incubated with DMEM/F-12 medium (250 mg wet tissue/mL) containing 10% exosome-depleted FBS (centrifugation at 100,000 g for 12 h) and 1% penicillin-streptomycin. After incubation for 24 h, the supernatants were stored at ‒80°C and then used to isolate exosomes using a Hieff® Quick exosome isolation kit (Yeasen, Shanghai, China) according to the manufacturer’s instructions. Briefly, the supernatants were centrifuged at 3000 g for 15 min and filtered through a 0.22-μm filter. Subsequently, the exosome isolation solution was added to the centrifuged supernatants, vortexed for 1 min, and incubated at 4°C for 2 h. Pellets of exosomes were collected by centrifuging the mixture at 10,000 g for 60 min. The final exosome pellet was resuspended in PBS.

Transmission electron microscopy (TEM)

Transmission electron microscopy was used to observe the morphology of the isolated exosomes. Briefly, 5 μL of exosome suspension was loaded onto a formvar-coated copper grid and fixed for at least 5 min. The exosomes were then negatively stained with 2% uranyl acetate for 5 min. After being air-dried, the grids were visualized with a transmission electron microscope (TECNAI 10; PHILIPS, Amsterdam, Holland) at 80 kV.

Nanoparticle tracking analysis

A NanoSight 3000 (Malvern’s, Malvern-Worcestershire, UK) was used to analyse the size of the exosomes. Briefly, the sample chamber was first cleaned using particle-free distilled water. The exosome samples were then diluted 5000 times with sterilized PBS. Subsequently, the diluted exosomes were slowly injected into the chamber and quantified using nanoparticle tracking analysis software.

Luciferase reporter constructs and luciferase assay

A wild-type murine C1ql1 mRNA 3′UTR segment was amplified and cloned into the pGEM-T-easy vector (Promega, Madison, USA). Mutation in the miR-133a binding site of C1ql1 was generated using the Fast Mutagenesis System (TransGen Biotech, Beijing, China) from the seed region of miR-133a. HEK293T cells were cotransfected with either miR-133a mimics or negative control and the 3′UTR of C1ql1 (with either wild-type or mutant miR-133a binding sites) using Lipofectamine 2000 provided with the Dual-Luciferase Reporter Assay System (Promega), and were then cultured in DMEM supplemented with 10% FBS for 48 h.

Bioinformatics analysis

Potential miRNA targets were identified using TargetScan and starBase v2.0 ( http://starbase.sysu.edu.cn/) [29]. Gene Ontology (GO) analysis was used to study biological processes for the target genes of differentially expressed miRNAs. The GO annotations from Gene Ontology ( http://www.geneontology.org/) were downloaded. To identify the significant GO categories, Fisher’s exact test was applied to calculate the P values, and the results of multiple hypotheses were used to correct the FDR. In addition, miR-133a and apoptosis-related mRNA network analysis was performed based on the significant GO analysis.

Statistical analysis

Data are presented as the mean±SEM. Statistical significance was determined using GraphPad Prism 6 (GraphPad Software, La Jolla, USA). The data were analysed using unpaired or paired Student’s t test. P values of <0.05 compared with the appropriate control were considered statistically significant.

Results

HFD-induced obesity upregulates the mRNA expressions of chemokines and cytokines and accelerates follicle depletion

Female pubertal mice were fed with a HFD for nine weeks to induce obesity. HFD significantly increased body weight ( Figure 1A) and resulted in the accumulation of mesenteric adipose tissue (mWAT) and peri-ovarian adipose tissue (POAT) ( Figure 1B). Compared with those in the NCD group, the transcript levels of chemokines ( Ccl4 and Cxcl10) and the proinflammatory cytokine Tnfα were increased in POAT ( Figure 1C). In addition, the transcript levels of chemokines ( Ccl2, Ccl4 and Cxcl10) and Nos2 were increased in mWAT ( Figure 1C). Moreover, HFD increased the expressions of chemokines ( Ccl2 and Ccl4) and the proinflammatory cytokines Il6 and Nos2 in ovaries ( Figure 1C). HFD-mice showed a tendency of increased Ccl3, Cxcl10 and Tnfα mRNA expressions in ovaries compared with their littermates ( Figure 1C). Histological analysis of the ovaries and follicle counting showed that HFD markedly decreased the proportion of primordial follicles and increased the proportion of atretic follicles ( Figure 1D,E). These data show that HFD-induced obesity results in inflammatory status in the adipose tissues and ovaries, and that HFD accelerates follicle depletion.

Figure 1 .

HFD induced obesity and affected follicular development

(A) Changes in the body weight of mice over 9 weeks of a high-fructose and high-fat diet (HFD) or negative control diet (NCD) ( n=10). HFD mice were significantly heavier than NCD littermates from 1 to 9 weeks of feeding. (B) Anatomical and histological changes associated with NCD and HFD. The left panel shows hypertrophy of the abdominal fat pads, and the right panel shows hypertrophy of POAT. (C) Relative mRNA expressions of inflammation-related factors in HFD murine adipose tissues and ovaries ( n=3). (D) Representative images of ovarian HE staining. (E) The proportion of follicles of different stages ( n=3). Data are presented as the mean±SEM. * P<0.05, *** P<0.001, **** P<0.001.

HFD increases the expression of miR-133a in adipose tissues and ovaries

qPCR analysis showed that the level of miR-133a was significantly increased in the mWAT, POAT, and ovaries of HFD-mice compared with that in the NCD group ( Figure 2A). The ovaries of many mammals lie within membranous sacs called bursae ovaricae [30]. The bursae ovaricae is a fluid-filled space. In HFD-mice, miR-133a level was significantly decreased after the bursae were scratched with a needle ( Figure 2B). These data demonstrate that the highly expressed miR-133a is partly located in the bursa.

Figure 2 .

HFD increased the expression of miR-133a, and the highly expressed miR-133a in the ovary was not transported from adipose tissue by exosomes

(A) Relative expression of miR-133a in murine ovaries and adipose tissues in NCD and HFD mice ( n=3). (B] Relative expression of miR-133a in intact ovaries and ruptured ovaries of HFD mice ( n=3). (C) Morphology of exosomes observed by transmission electron microscopy and size distribution of exosomes shown by NanoSight analysis. (D) Western blot analysis showing that isolated vehicles expressed exosomal markers, including CD63 and TSG101, and the adipocyte marker Perilipin-1. CM: control medium. (E) Relative expression of miR-133a in adipose tissue-derived exosomes. Data are presented as the mean±SEM. * P<0.05, ** P<0.01.

The increased miR-133a in the HFD ovaries is not derived from exosome transferred from obese adipose tissues

As adipose tissue constitutes an essential source of circulating exosomal miRNAs that can regulate gene expression in distant tissues [31], we speculated that miR-133a may be transported into the ovary through exosomes secreted by adipose tissues. To confirm this, exosomes were isolated from the control medium (CM) of adipose tissue explants. Transmission electron microscopy and NanoSight analysis revealed adipose tissue-derived exosomes with a diameter of 50-200 nm ( Figure 2C). Western blot analysis revealed the presence of the exosome-specific protein markers CD63 and TSG101 and the adipocyte marker Perilipin-1 in adipose tissue-derived exosomes ( Figure 2D). However, the qPCR analysis showed that miR-133a expression was not increased in either POAT-EVs (71.6% reduction) or mWAT-EVs (52.5% reduction) from HFD-mice ( Figure 2E). These data demonstrate that increased miR-133a in the HFD ovaries is not derived from exosome transferred from obese adipose tissues.

Increased expression of miR-133a in HFD ovaries is a response to the inflammatory status caused by HFD

To determine the sources of the increased miR-133a in the HFD ovaries, we examined the levels of premiR-133a and primiR-133a. Significantly increased expressions of premiR-133a and primiR-133a were found in the ovaries ( Figure 3A,B), indicating that increased mature miR-133a is synthesized in HFD ovaries.

Figure 3 .

Highly expressed miR-133a was synthesized in granulosa cells in response to the inflammatory status caused by HFD

(A) Relative expression of precursor premiR-133a in murine ovaries ( n=3). (B) Relative expression of primiR-133a ( n=3). (C) Relative mRNA expressions of inflammation-related factors in ovaries after intraperitoneal injection of LPS ( n=3). (D) Relative expression of miR-133a in ovaries after intraperitoneal injection of LPS ( n=3). (E) Relative mRNA expressions of inflammation-related factors in granulosa cells treated with LPS ( n=3). (F) Relative expression of miR-133a in granulosa cells treated with LPS ( n=3). Data are presented as the mean±SEM. * P<0.05, ** P<0.01, **** P<0.0001.

We next challenged the mice with LPS to mimic the inflammatory status caused by obesity. The mRNA levels of chemokines (Ccl2, Ccl4 and Cxcl10) and inflammatory factors ( Tnfα and Il6) were markedly increased after LPS stimulation ( Figure 3C) and accompanied by significantly increased miR-133a expression level in the ovaries ( Figure 3D). Moreover, elevated miR-133a expression was also found in LPS-treated granulosa cells ( Figure 3F), in which the transcription of chemokines (CCL2, CCL4 and CXCL10) and inflammatory factors (TNFα and IL6) was increased ( Figure 3E). These data suggest that the overexpression of miR-133a in the HFD ovaries is a response to the inflammatory status caused by HFD.

miR-133a arrests ovarian growth and follicular development

To confirm the effect of miR-133a on the ovary, miR-133a agomir was injected into the bursa of the ovary. Two weeks later, the size and weight of miR-133a agomir-injected ovaries were decreased significantly ( Figure 4A,B). Hematoxylin and eosin staining showed that miR-133a agomir-injected ovaries exhibited a disordered morphology, characterized by an evident accumulation of atretic follicles ( Figure 4C). Detailed counting of follicles at different stages showed that the miR-133a agomir significantly decreased the proportion of primordial follicles and increased the proportion of antral and atretic follicles ( Figure 4D). These results demonstrated that a large proportion of primordial follicles were recruited and developed into antral follicles but underwent atresia in miR-133a-agomir-injected ovaries, which suggests that overexpression of miR-133a accelerates the depletion of the ovarian reserve.

Figure 4 .

miR-133a arrested ovarian growth and follicular development

(A,B) Comparison of ovarian size (A) and weight (B) after intrabursal injection of miR-133a agomir (paired t test, n=5). (C) Representative images of ovarian hematoxylin and eosin staining. (D) The proportion of follicles at different stages ( n=3). ProF: primordial follicles; SF: secondary follicles; AF: antral follicles; ATF: atretic follicles. Data are presented as the mean±SEM. * P<0.05, ** P<0.01.

miR-133a follows a low-intensity pattern to promote granulosa cell apoptosis in vitro

Apoptotic cell death is the molecular mechanism underlying follicle atresia [32]. TUNEL staining showed that the apoptosis signals in miR-133a agomir-injected ovaries were detected in both atretic follicles and the granulosa cells of large antral follicles, and the signals were more prominent than those in the negative control-injected ovaries ( Figure 5A). Given that the increased apoptosis signals were mostly present in granulosa cells, the granulosa cells were isolated and transfected with a miR-133a mimic. Western blot analysis showed that the protein level of cleaved caspase-3 increased, but not significantly, after transfection of granulosa cells with the miR-133a mimic ( Figure 5B). However, one of the caspase-3 substrates, PARP1, was cleaved significantly in miR-133a-overexpressing granulosa cells ( Figure 5B). We next used LMPCR, which is sensitive, to detect the production of DNA fragments in the nucleus [ 27, 28] . LMPCR analysis revealed extensive nucleosomal DNA fragmentation in granulosa cells starved and transfected with miR-133a mimic ( Figure 5C). These results demonstrate that miR-133a promotes granulosa cell apoptosis in a low-intensity pattern in vitro.

Figure 5 .

miR-133a promoted granulosa cell apoptosis in a low-intensity pattern in vitro

(A) TUNEL staining of apoptosis in ovaries after intrabursal injection with miR-133a agomir. Representative ovarian tissue sections are shown. The right panel is a magnification of the left panel. Relative areas stained with TUNEL were quantified using Image-Pro Plus 6.0 and are shown as the mean±SEM in statistical charts ( n=3). ** P<0.01. AF: antral follicles; ATF: atretic follicles. (B) The protein levels of caspase-3 and PARP1 were analysed by western blot analysis. Representative blots of three mice in each group with the amount of protein normalized to β-tubulin, which is presented as the mean±SEM ( n=3). * P<0.05. (C) Ligation-mediated PCR analysis of apoptotic DNA after granulosa cells were transfected with miR-133a mimic. Con: control group (without treatment); NC: negative control group (transfected with NC); miR-133a: miR-133a mimic group (transfected with miR-133a); Starved: starvation treatment for 12 h.

miR-133a participates in the regulation of granulosa cell survival through multiple pathways

To further confirm the mechanism by which miR-133a regulates granulosa cell apoptosis, we examined the effect of miR-133a on apoptosis-related proteins in granulosa cells. Our previous study showed that C1QL1 deficiency promoted granulosa cell apoptosis [25]. TargetScan showed that C1ql1 is one of the target genes of miR-133a, and the predicted binding sequences are shown in Figure 6A. We performed luciferase assays to investigate the direct targeting of the C1ql1 3′UTR by miR-133a. HEK293T cells transfected with reporter plasmids containing the C1ql1 3′UTR showed markedly decreased luciferase activity in the presence of miR-133a mimic. Mutation of the C1ql1 3′UTR abrogated the miR-133a mimic-induced repression of luciferase activity ( Figure 6B). Furthermore, the mRNA and protein levels of C1ql1 were significantly decreased in granulosa cells transfected with the miR-133a mimic ( Figure 6C,D). These data indicate that miR-133a could target C1ql1 expression to promote granulosa cell apoptosis.

Figure 6 .

miR-133a participated in the regulation of granulosa cell survival through multiple pathways

(A) The predicted binding sequences between C1ql1 and miR-133a. (B) Validation of miR-133a binding to the C1ql1 3′UTR by dual-luciferase reporter assay. (C) Relative mRNA expression of C1ql1 after miR-133a mimic transfection. (D) The protein levels of C1QL1, XIAP and PTEN were analysed by western blot analysis. Representative blots of three mice in each group and the amount of protein normalized to β-tubulin ( n=3). Data are presented as the mean±SEM. * P<0.05.

XIAP, an X-linked inhibitor of apoptosis protein, is a member of the inhibitor of apoptosis protein (IAP) family known for directly suppressing apoptotic cell death pathways [33]. The XIAP protein was significantly decreased in granulosa cells transfected with the miR-133a mimic ( Figure 6D). Unexpectedly, PTEN, a negative regulator of the Pl3K/Akt pathway that can negatively regulate the expression of XIAP, was also decreased in granulosa cells transfected with the miR-133a mimic ( Figure 6D). We searched starBase v2.0 ( http://starbase.sysu.edu.cn/) [29] and found that XIAP and PTEN are the predicted target genes of miR-133a. Therefore, we speculated that miR-133a may decrease XIAP and PTEN expression through direct targeting. miR-133a targets antiapoptotic C1QL1 and XIAP expression to promote granulosa cell apoptosis. In addition, miR-133a downregulates the expression of PTEN, thereby regulating the Pl3K/Akt pathway and affecting cell survival.

To further explore the mechanism of miR-133a-involved granulosa cell apoptosis, we constructed a miR-133a-apoptosis-related mRNA network. The results of the network analysis are shown in Figure 7 and elucidate the role of miR-133a in regulating granulosa cell apoptosis. We first analysed the potential targets of miR-133a using TargetScan and starBase v2.0. Subsequent GO analysis revealed that 163 genes among the potential targets of miR-133a were associated with the regulation of apoptosis. Among them, 87 genes were involved in the negative regulation of apoptosis, 48 genes were involved in the positive regulation of apoptosis, and 30 genes were involved in both the negative and positive regulation of apoptosis. These results indicate that miR-133a regulates granulosa cell apoptosis through multiple targets to balance cell survival and apoptosis.

Figure 7 .

miR-133a-apoptosis-related mRNA analysis

The yellow cycle nodes represent genes related to the negative regulation of apoptosis. The blue cycle nodes represent genes associated with the positive regulation of apoptosis. The green cycle nodes represent genes related to negative or positive regulation of apoptosis. All represented genes were identified by GO analysis.

Discussion

With the gradual improvement in living standards, the number of overweight and obese people has gradually increased [34]. Numerous clinical studies have shown that obesity causes female reproductive dysfunction [35]. The clinical manifestations of reproductive dysfunction include PCOS, oligomenorrhea, and ovulation disorders, which are closely related to follicular development disorders [ 5, 36] . Follicular growth and development in obese women are stunted, but the molecular mechanism of obesity-induced abnormal folliculogenesis is still unclear. In this study, we demonstrated the role of miR-133a in obesity-induced abnormal folliculogenesis by tracing its origin and investigating the effects of miR-133a overexpression on follicle development in vivo and apoptosis of granulosa cells in vitro.

It has been suggested that obesity leads to follicular development disorder and even the PCOS phenotype in different animal models. To find the link between obesity and abnormal folliculogenesis, we established an obese mouse model by feeding mice with an obesogenic diet consisting of high-fat feed and fructose water (HFD) for nine weeks. HFD significantly reduced the number of primordial follicles and increased the number of atretic follicles in mice, resulting in abnormal follicle development. However, unlike other studies, PCOS features such as thickening of the follicular parietal layer and the appearance of cystic follicles were not found [ 37, 38] , which may be due to the shorter duration of obesogenic diet feeding in our study. In rat models, HFD feeding of prepubertal rats for 105 d results in PCOS disturbances [37].

Although the ovarian phenotype after nine weeks of HFD was slightly different from that of typical PCOS, we found that the expression of miR-133a in adipose tissue and ovaries of obese mice increased, which is consistent with the increase in miR-133a in adipose tissue of PCOS patients and granulosa cells of obese PCOS patients [ 20, 21] . The above results suggest that miR-133a may be a link between obesity and PCOS follicular dysplasia, and the increase in miR-133a in a short-term HFD may be related to the final formation of the PCOS follicular dysplasia phenotype. We found that mice showed typical PCOS symptoms after eight months of HFD feeding (results not shown). Based on the speculation that miR-133a may link obesity to derangements in follicular development, we detected a significant increase in miR-133a expression in the ovaries of HFD mice, but its origin is still unclear.

Adipose tissue is an important site of energy storage and regulates metabolism by releasing adipokines [ 39, 40] . A recent study showed that adipose tissue is an essential source of circulating exosomal miRNAs in both mice and humans. Circulating exosomal miRNAs derived from adipose tissue may regulate whole-body metabolism and mRNA translation in other tissues [ 31, 41] . Among adipose tissues, POAT surrounds the ovaries in rodents, and our research showed that POAT is essential for follicular development, and its secreted substances may play an important role [42]. Given the increased accumulation of miR-133a in the POAT and visceral mWAT, we speculated that the high expression of miR-133a in the HFD ovaries may be derived from exosomes secreted from adipose tissues. Unexpectedly, in adipose-derived exosomes of HFD mice, the expression of miR-133a was reduced, indicating that the increased expression of miR-133a in the HFD ovaries is not derived from exosome transferred from obese adipose tissues.

In addition to exogenous transfer, intraovarian miRNA may be produced by ovarian follicular cells as a response to changes in the ovarian microenvironment. We found that the expressions of primiR-133a and premiR-133a were significantly increased in HFD ovaries, indicating that the biosynthesis of intraovarian mature miR-133a takes place in the ovaries. Obesity leads to chronic systemic inflammation, and this chronic inflammatory state contributes to the progression of PCOS [ 43, 44] . In our study, the increased production of miR-133a in the inflammatory adipose tissues, ovaries of obese and LPS-treated mice, and LPS-treated granulosa cells indicated that increased expression of miR-133a is associated with inflammatory status. Accordingly, in two recent clinical studies, high miR-133a levels were found in the serum of sepsis patients and LPS-treated human bronchial epithelial cells [ 45, 46] . Therefore, HFD induced the hypertrophy of adipose cells and the proinflammatory polarization state of macrophages, which led to the release of several inflammatory cytokines and other factors. These inflammatory cytokines and other factors may circulate to the ovary and create the inflammatory status of the ovaries. miR-133a is then upregulated in granulosa cells in response to inflammatory stress.

Intrabursal injection of miR-133a agomir caused an ovarian phenotype similar to nine weeks of feeding with an HFD, including accelerated recruitment of primordial follicles and an increased number of atretic follicles. Granulosa cell apoptosis was confirmed to be the main reason for follicular atresia [47]. Cleavage of PARP1 by caspase-3 is a universal hallmark of apoptotic cell death [48]. Overexpression of miR-133a did not significantly increase the expression of activated caspase-3 but slightly increased the level of cleaved PARP1. Additionally, TUNEL staining showed strong apoptosis signals in the granulosa cells of the large antral follicles after miR-133a agomir treatment. Moreover, extensive nucleosomal DNA fragmentation was found in granulosa cells overexpressing miR-133a. These data indicated that miR-133a is involved in the apoptosis of granulosa cells, not strongly but at a modest level. Genetic inactivation of miRNAs often does not have obvious phenotypic consequences [49]. miRNAs lead to very modest repression of their direct targets, and one miRNA can target multiple targets simultaneously [50]. Here, we believe that miR-133a does not act as a switch for cell apoptosis that strongly represses one or several targets but as a rheostat that buffers cell apoptosis by targeting hundreds of genes through a different mechanism. Phenotypic consequences can only occur after long-term regulation, which may also explain increased apoptosis and atresia follicles in mouse ovaries only after long-term HFD consumption.

miR-133a may target many genes related to cell survival. Our previous studies found that C1QL1 deficiency leads to apoptosis of granulosa cells [25], and here, we revealed an inverse relationship between the expression level of miR-133a and its target gene C1ql1. Additionally, XIAP is a member of the inhibitor of the apoptosis protein family. Several studies have shown that XIAP is expressed and regulated at different stages of follicular development [ 51– 53] . XIAP expression was decreased in granulosa cells after overexpression of miR-133a. In addition, other studies have confirmed that miR-133a targets TAGLN2 in intestinal epithelial cells to promote apoptosis of intestinal epithelial cells [54] and the anti-apoptotic protein Bcl-xL [55]. The PTEN/Akt pathway is a crucial regulator of cell proliferation and survival. PTEN is a negative regulatory protein of AKT, whereas overexpression of miR-133a increased PTEN in granulosa cells. This may be another mechanism by which miR-133a regulates the apoptosis of granulosa cells. With hundreds of potential targets related to the apoptosis of miR-133a, we speculate that miR-133a may regulate the apoptosis of granulosa cells by targeting multiple targets to balance cell survival and apoptosis. Eventually, caspase-3 is slightly activated and granulosa cells are in a state of mild apoptosis, consistent with the chronicity of obesity-induced abnormal folliculogenesis.

In summary, we established a diet-induced obesity mouse model and a miR-133a agomir intrabursal injection model to demonstrate a possible link between obesity and follicular development disorders mediated by the increased production of miR-133a in the ovary. The findings of this study facilitate an improved understanding of obesity-induced reproductive dysfunction.

COMPETING INTERESTS

The authors declare that they have no conflict of interest.

Funding Statement

This work was supported by the grants from the Science and Technology Program of Guangzhou, China (No. 201904010041), the Guangdong Basic and Applied Basic Research Foundation (No. 2020A1515010033), the National Natural Science Foundation of China (No. 30900232), the Guangdong Science and Technology Department Foundation (No. 2020B1212060031), and the Research Grant of Key Laboratory of Regenerative Medicine, Ministry of Education, Jinan University (No. ZSYXM202207).